Comprehensive experimental study of N-heterocyclic π-stacking interactions of neutral and cationic pyridines

Article abstract of DOI:10.1021/jo400370e

A comprehensive experimental study was carried out by measuring the relative strengths of parallel π-stacking interactions of N-heterocycles with nonheterocycles. A versatile and rigid model system was developed, which was in equilibrium between a "closed" conformation that forms an intramolecular π-stacking interaction and an "open" conformation that cannot form the interaction. First, the formation and geometries of the intramolecular N-heterocyclic π-stacking interactions were verified by X-ray crystallography. Next, the closed/open ratios were measured in solution via integration of the ¹H NMR spectra, providing an accurate comparison of the N-heterocyclic π-stacking interactions. The synthetic versatility of this model system enabled the systematic and comprehensive comparison of the influences of position, charge, and substituent effects of the nitrogen atom of the N-heterocycles within a single model system. The π-stacking interactions of the neutral N-heterocyclic rings were slightly stronger than that of nonheterocyclic rings. Cationic N-heterocycles formed significantly stronger π-stacking interactions than neutral N-heterocycles. The position of the nitrogen atom also had a strong influence on the stability of N-heterocyclic π-stacking complexes. Interestingly, opposite stability trends were observed for neutral and cationic N-heterocycles. For neural N-heterocycles, geometries with the nitrogen away from the π-face of the opposing ring were the more stable. For cationic N-heterocycles, geometries with the nitrogen close to the π-face of the opposing ring were the more stable. Finally, N-methylated heterocycles consistently formed stronger π-stacking interactions than N-protonated heterocycles.

Full text of DOI:10.1021/jo400370e

Subscriber access provided by Cape Breton University | Library
Article
A comprehensive experimental study of N-heterocyclic #-
stacking interactions of neutral and cationic pyridines
Ping Li, Chen Zhao, Mark D Smith, and Ken Daniel Shimizu
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/jo400370e • Publication Date (Web): 15 May 2013
Downloaded from http://pubs.acs.org on May 23, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
The Journal of Organic Chemistry is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
A comprehensive experimental study of N-heterocyclic π-stacking interactions of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
neutral and cationic pyridines
Ping Li, Chen Zhao, Mark D. Smith, Ken D. Shimizu*
Department of Chemistry and Biochemistry, University of South Carolina,
Columbia SC 29208
Fax: (803)7779521; Tel: (803)7776523
E-mail: shimizu@mail.chem.sc.edu
ABSTRACT: A comprehensive experimental study was carried out measuring the
relative strengths of parallel π-stacking interactions of N-heterocycles with non-
heterocycles. A versatile and rigid model system was developed, which was in
equilibrium between a “closed” conformation that forms an intramolecular π-stacking
interaction and an “open” conformation that cannot form the interaction. First, the
formation and geometries of the intramolecular N-heterocyclic π-stacking interactions
were verified by X-ray crystallography. Next, the closed/open ratios were measured
1
in solution via the integration of the H NMR spectra, providing an accurate
comparison of the N-heterocyclic π-stacking interactions. The synthetic versatility of
this model system enabled the systematic and comprehensive comparison of the
influences of position, charge, and substituent effects of the nitrogen atom of the N-
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 34
1
2
3
4
5
6
7
8
9
heterocycles within a single model system. The π-stacking interactions of the neutral
N-heterocyclic rings were slightly stronger than that of non-heterocyclic rings.
Cationic N-heterocycles formed significantly stronger π-stacking interactions than
neutral N-heterocycles. The position of the nitrogen atom also had a strong influence
on the stability of N-heterocyclic π-stacking complexes. Interestingly, opposite
stability trends were observed for neutral and cationic N-heterocycles. For neural N-
heterocycles, geometries with the nitrogen away from the π-face of the opposing ring
were the more stable. For cationic N-heterocycles, geometries with the nitrogen close
to the π-face of the opposing ring were more stable. Finally, N-methylated
heterocycles were consistently observed to form stronger π-stacking interactions than
N-protonated heterocycles.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
■ INTRODUCTION
Attractive non-covalent interactions between aromatic surfaces play an
important role in many key molecular recognition processes.1,2 The majority of π-
stacking interaction studies, to date, have focused on the interactions of non-
heterocycles such as benzene3-7 and substituted benzene dimers.8-19 Yet, N-
heterocycles are ubiquitous in synthetic20,21 and biological systems22 and their π-
stacking interactions are often key to their function and utility.23,24 The fewer
experimental studies with N-heterocycles can be attributed to the greater synthetic
challenge and the greater number of possible π-stacking geometries. Consequently,
the majority of N-heterocyclic π-stacking studies have been theoretical computational
studies.25-30
In this paper, we describe the development of a versatile and rigid molecular
model system that can fix an N-heterocyclic and a non-heterocyclic ring into specific
π-stacking geometries. The model system also provides a quantitative measure of the
2
ACS Paragon Plus Environment

Page 3 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
strengths of these interactions via their conformational equilibrium ratios (Figure 1).
These “molecular balances” are in conformational equilibrium between open and
closed conformations due to restricted rotation around the Caryl-Nimide single bond. In
the closed conformation, the rigid bicyclic framework forces the aromatic surfaces of
the arm and shelf into an off-set parallel π-stacking geometry. In the open
conformation, these two surfaces are held apart. Therefore, the closed/open ratio
provides a sensitive measure of the relative strengths of the intramolecular π-stacking
interactions in solution.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. A general depiction of the molecular balance model system utilized in this
study to measure the intramolecular N-heterocyclic π-stacking interactions formed in
the closed conformer between an N-heterocyclic and a non-heterocyclic aromatic ring.
The structural rigidity and synthetic versatility of our molecular balance model
system enabled the systematic study of the influences of the position of the nitrogen
atom, charge, and substituent on the nitrogen atom within a single molecular
framework. This is important, as much greater structural and geometric diversity of
π-stacking complexes can form with N-heterocycles (Figure 2a) than with simple
benzenes (Figure 2b). First, the nitrogen atom disrupts the symmetry of the π-
stacking complexes giving rise to multiple geometries. These can be grouped into
two distinct classes: the distal geometry where the pyridine nitrogen is far from the π-
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 34
1
2
3
4
5
6
7
8
9
face of the opposing benzene ring, and the proximal geometry where the pyridine
nitrogen is close to the π-face of the opposing benzene ring. Second, N-heterocyclic
complexes can be neutral or charged via the absence or presence of a substituent on
the nitrogen. Finally, the cationic N-heterocyclic complexes can have a proton or
alkyl N-substituent.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Comparison highlighting the greater structural and geometric diversity of
the π-stacking complexes of (a) N-heterocycles versus (b) benzene rings.
Similar small molecule model systems that use conformational equilibriums to
study non-covalent interactions have been widely employed with great success.31
This includes the study of the π-stacking interactions of N-heterocycles. However,
these studies have not been comprehensive and have only focused on specific
pairwise comparisons. For example, Gellman and co-workers used a flexible
4
ACS Paragon Plus Environment

Page 5 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
secondary amide-based model system to demonstrate the enhanced affinity of a
neutral six-membered heterocycle versus a phenyl ring.32 Gung and co-workers also
observed the stronger π-stacking interactions of an N-heterocyclic versus non-
heterocyclic aromatic rings using a three-state triptycene-based model system.33
Dougherty and co-workers used an electron-rich cyclophane aromatic host to measure
the enhanced affinities of cationic versus neutral quinoline guests.34 Waters and co-
workers used a flexible rotamer model system to demonstrate that the position of the
nitrogen atom in cationic π-stacking interactions had a profound influence on the
interaction energies.35 However, the results from these pioneering studies cannot be
combined to produce an accurate comprehensive quantitative comparison of all of the
N-heterocyclic π-stacking complexes shown in Figure 2.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To simplify the discussions, the following abbreviations will be used to
describe the different complexes shown in Figure 2. The aromatic groups are
abbreviated as follows: phenyl (Ph), pyridine (Py), pyridinium (PyH+), N-
methylpyridinium (PyMe+). Thus, the interactions and complexes shown in Figure 2
(from top to bottom) are Py–Ph, PyH+–Ph, PyMe+–Ph and Ph–Ph. In addition, we
will use the term, “π-stacking”, to describe the interactions shown in Figure 2 in order
to maintain consistency with the literature. However, we realize this term is
somewhat misleading as the π-π interaction is repulsive.36-39 The attractive
components of π-stacking interactions are generally attributed to dispersion
interactions of molecular surfaces and electrostatic attraction between dipoles and/or
quadrupoles of aromatic rings.
■ RESULTS AND DISCUSSION
Molecular Balance Design. The molecular balance model system developed in
this study has the same rigid bicyclic N-arylimide framework that we have
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 34
1
2
3
4
5
6
7
8
9
successfully used to study face-to-face aromatic π-stacking (balance 1)40 and aliphatic
CH–π41 interactions (not shown). This similarity provided a number of advantages
for the present study. First, the N-heterocyclic and non-heterocyclic aromatic surfaces
could be predictably and precisely positioned relative to each other. We have
previously demonstrated that the bicyclic framework fixes the arm and shelf aromatic
surfaces into an off-set parallel π-stacking geometry in the closed conformer.40 The
proximity of the two interacting surfaces in the closed conformers also prevents the
formation of the other commonly observed edge-to-face, perpendicular geometry.
Second, a series of molecular balances with N-heterocyclic units incorporated into
either the arm or shelf surfaces can be rapidly assembled owing to the highly modular
syntheses of this framework. Third, the closed/open ratios can be easily and
accurately measured via the integration of the 1H NMR spectra. The two conformers
are in slow exchange at room temperature (23 ˚C), and differences as small as ±0.03
kcal/mol can be accurately measured. Fourth, the crystallinity of these molecular
balances provided a unique opportunity to simultaneously characterize and study the
geometry of N-heterocyclic π-stacking interactions in the solid-state.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For this study, six new molecular balances (2-7) were developed (Figure 3).
The intramolecular interactions in 2-7 mirror the different types of the N-heterocyclic
π-stacking complexes shown in Figure 2a. The relative position of the nitrogen atom
in the complexes was controlled by placing the N-heterocyclic ring into either the arm
or shelf. Incorporation of a pyridine ring into the aromatic arm of 2 yielded a
molecular balance that adopts a distal geometry. The pyridine ring of the arm extends
beyond the edge of the aromatic shelf so the pyridine nitrogen is away from the π-face
of the aromatic shelf. Conversely, incorporating pyridine rings into the arene shelf in
3 yielded a molecular balance that adopts a proximal geometry. In this case, the
6
ACS Paragon Plus Environment

Page 7 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
benzene arm extends out and over one of the nitrogens in the aromatic shelf.
Similarly, cationic N-protonated and N-methylated balances with the N-heterocycles
in the arm (4 and 6) adopt distal geometries and with the N-heterocycles in the shelf
(5 and 7) adopt proximal geometries.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R
O
N
O
O
Z
Ph!Ph
stacking
O
Z
1 (R = H, Z = Ph)
1' (R = OPh, Z= Ph)
1'' (R = H, Z = Me)
distal
proximal
R
Py!Ph
stacking
O
O
N
N
O
O
O
O
N
N
N
Me
Ph
N
O
Me
3 (R = H)
3' (R = OPh)
O
Ph
2
PyH+!Ph
O
N
O
stacking
N
O
O
O
H
O
N
N
Me
Ph
O
H
O
Me
Ph
4
5
PyMe+!Ph
O
O
N
N
stacking
O
O
Me
O
O
N
N
Me
Ph
Me
N
O
O
Me
Ph
6
7
Figure 3. The closed conformers of molecular balances 1-7 that form intramolecular
Ph–Ph, Py–Ph, PyH+–Ph and PyMe+–Ph stacking interactions. The two interacting
rings are highlighted in red (N-heterocycle) and blue (non-heterocycle). Balances 1’
and 3’ are two-armed analogues used in the crystallographic analysis and balance 1’’
is a control used to examine influence of the difference in bridgehead Z group over
the closed/open ratios in solution.
Synthesis. The syntheses of balances with the N-heterocyclic ring in the arm
(2, 4 and 6) followed the previously reported the two-step convergent route.40,41
A
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 34
1
2
3
4
5
6
7
8
9
representative synthesis of 2 is shown in Scheme 1. First, the SNAr reaction between
3-hydroxypyridine (8) and 2-fluoronitrobenzene (9) followed by Pd/C catalyzed
reduction of the nitro group yielded aniline 10.42 The thermal condensation of aniline
10 with the endo-bicyclic anhydride 11 yielded the endo-bicyclic balance 2. Pyridyl
arms with the nitrogen in the meta-position were prepared because the SNAr reaction
with the para-pyridyl precursor, 4-hydroxypyridine, was not successful due to
preferential N- versus O-arylation.43 Despite its lack of symmetry, the meta-pyridyl
arm was able to adopt the desired distal geometry, as was later confirmed in the X-ray
crystallographic study and the solution 1H NMR studies.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Syntheses of balances 2 and 6 a
aKey: (a) i. Cs2CO3, 23 °C; ii. H2, Pd/C, MeOH, 23 °C, 76% for two steps (b)
HOAc, reflux, 81% (c) MeI, acetone, reflux, quantitative.
Incorporation of the N-heterocyclic rings into the shelf of balances (3, 5 and 7)
required the synthesis of a new diene building block (13) that contained an N-
heterocyclic phenanthroline shelf. Diene 13 was formed in two-steps from the
Knovenagel cyclization reaction of dione 12 and 2-pentanone.44 The Diels-Alder
reaction between diene 13 and maleimide 14 gave the endo-bicyclic balance 3
(Scheme 2). The choice of 1,10-phenanthroline with two heterocyclic nitrogens as the
8
ACS Paragon Plus Environment

Page 9 of 34
The Journal of Organic Chemistry
1
2
3
4
shelf motif was intended to ensure the symmetry of the arene shelf, as the arene of
5
6
arm can form π-stacking interactions with either of the two outer rings of the shelf.
7
8
9
Scheme 2. Syntheses of balances 3 and 7a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aKey: (a) i. K2CO3, MeOH, 23 °C; ii. py-SOCl2, 23 °C 37% for two steps (b)
CH2Cl2, pressure tube, 90 °C, 45% (c) MeI, acetone, reflux, quantitative.
In addition, two new control balances 3’ and 1’’ were prepared. Balance 3’ is
a “two-armed” analogue of 3, which was synthesized from diene 11 and two-armed
N-phenylmaleimide 15 following the previously reported method.40 Balance 1’’ is a
structural analogue of 1 with methyl groups instead of phenyl groups attached to the
bridgeheads (Figure 3). Balance 1’’ was prepared by a similar route as 3 via the
Diels-Alder reaction between maleimide 14 and the known diene 16.45
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 34
1
2
3
4
5
Figure 4. Maleimide 15 and diene 16 precursors used in the syntheses of the control
balances 3’ and 1’’.
6
7
8
9
The cationic balances (4-7) were formed by methylation or protonation of the
N-heterocycles in the arms or shelves. The N-methylated balances 6 and 7 were
obtained through methylation of the neutral balances 2 and 3 using methyl iodide in
acetone (Scheme 1 and 2). The protonated balances 4 and 5, on the other hand, were
obtained by in situ treatment of balances 2 and 3 with methanesulfonic acid.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
X-ray Crystal Structure Analysis. To confirm the formation and correct
geometry of the intramolecular N-heterocyclic π-stacking interactions in our model
systems, the molecular balances were characterized using X-ray crystallography. X-
ray quality single crystals of balances 2, 3, 4, 6, and 7 were successfully obtained
from chloroform-acetonitrile mixtures; only balance 5 did not form X-ray quality
crystals. All but one of the molecular balances crystalized in the closed conformation.
This observation provided the first indication of the greater strength of the N-
heterocyclic π-stacking interactions. In contrast, the non-heterocyclic π-stacking
balance 1 consistently crystalized in the open conformation.40 The only N-
heterocyclic balance that crystalized in the open conformation was balance 3 that
contains the weakest N-heterocyclic π-stacking interaction, as was later confirmed by
the solution study. To characterize the intramolecular π-stacking interaction in
balance 3, a two-armed version 3’ was synthesized and crystalized. The presence of
two identical phenyl ether arms attached to both ortho-position of the N-aryl rotor
ensured that one arm would always adopt the closed conformation.
The crystal structures of 2, 3’, 4, 6, and 7 all showed intramolecular parallel π-
stacking interactions between the aromatic arm and shelf surfaces. Representative
crystal structures of balances 2 and 3’ are shown in Figure 5. The rigid C-shaped
endo-bicyclic framework brings the aromatic arm surface in close proximity to the
10
ACS Paragon Plus Environment

Page 11 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
aromatic shelf in the closed conformation. Examination of the geometries of the
interacting six-membered rings confirmed the formation of the expected off-set face-
to-face π-stacking interactions in balances 2, 3’, 4, 6 and 7. The center-to-plane
distance distances (D) were 3.31–3.63 Å, which were well within the typical range of
3.3–3.8 Å for face-to-face aromatic stacking interactions.46 The two rings were
roughly parallel as shown by the relatively small dihedral angles (α) (4.81°–17.04°)
between the planes of the interacting rings, which were below the cutoff (< 20°) for
the face-to-face π-stacking geometry.47 Finally, the centroid-to-centroid horizontal
off-set (H) of the two interacting rings (0.82–1.98 Å) compared favorably with the
reported average value (1.30 Å) for the off-set geometry in the literature.46
(Definition for these geometric parameters is provided in Figure S2 in the SI)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. X-ray crystal structures of balances 2 and 3’ (the bridgehead phenyl and
methyl functional groups are omitted for viewing clarity). The intramolecular N-
heterocyclic π-stacking interactions are highlighted: N-heterocycle in red and non-
heterocycle in blue.
11
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 34
1
2
3
4
5
6
Table 1. Measured plane-to-plane angle (α), centroid-to-plane distance (D) and
horizontal centroid-to-centroid off-set (H) between the two π-stacking six-membered
rings in the crystal structures of balances 1’, 2, 3’, 4, 6, and 7.
7
π-stacking
interaction
Ph–Ph
Py–Ph
8
9
Balance.
geometry
α
7.8°a
9.3°
15.3°
17.0°
D [Å]
H [Å]
1’
2
3’
4
6
7
–
3.75a
3.63
3.60
3.37
0.82a
1.67
1.82
1.27
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
distal
proximal
distal
Py–Ph
PyH+–Ph
PyMe+–Ph
distal
4.8°, 6.4°b 3.29, 3.36b 1.78, 1.98b
PyMe+–Ph proximal
10.3°
3.31
1.65
aRef. 40 bThe crystal structure of 6 contains two crystallographically
independent molecules.
Analysis of the crystal structures also provided verification of the formation of
the predicted proximal and distal π-stacking geometries. To better visualize the
geometries, top views of the interacting rings in the crystal structures were examined
(Figure 6). In particular, we were concerned that the mobility of the arene arm in the
plane of the arene shelf might lead to alternative geometries. However, examination
of each structure in Figure 7 shows the expected distal geometries 2, 4 and 6, and the
proximal geometries in the closed conformers 3’ and 7. We were also concerned with
the ability of balances 3’, 4 and 6 to adopt the less favorable geometries (proximal,
distal and distal) for their respective N-heterocyclic π-stacking interactions.
However, examination of these structures shows the N-heterocyclic nitrogen has been
positioned into the designed geometry. This study confirmed the ability of the rigid
framework to control the intramolecular interaction geometries.
12
ACS Paragon Plus Environment

Page 13 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Top views of the crystal structures of balances 1’, 2, 3’, 4, 6, and 7
highlighting the intramolecular π-stacking interactions in the closed conformers. The
corresponding ChemDraw representations are also provided for viewing clarity.
Third, analysis of the crystal structures for balances 1’, 2, 3’, 4, 6 and 7
confirmed the structural continuity of the intramolecular π-stacking interactions
providing support for the ability to accurately compare their respective interaction
energies. These new heterocyclic balances contained a much wider diversity of
functional groups, charges, and geometries which, we were concerned, might lead to
different secondary interactions and steric effects and thus mask the π-stacking
interactions of interest. The similarity in the structures of the balances was
demonstrated by the conformity of the positions of the ether oxygen linkers to the
arene shelves. Figure 7 shows the overlaid structures of the ether oxygen linkers and
the arene shelf of the closed conformers 1’, 2, 3’, 4, 6, and 7. The position of the
oxygen linkers was remarkably consistent among the seven crystal structures,
confirming that the geometric and steric constraints in the close conformers were very
similar.
13
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 34
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Top and front views of the overlaid crystal structures for balances 1’, 2, 3’,
4, 6, and 7 highlighting the relative positions of the oxygen linkers on the arm over
the arene shelf. Only the arene shelves and arm ether oxygens are shown, and the
overlapping center six-membered aromatic ring of the shelves is highlighted in
yellow.
Characterization in solution. The relative strengths of the intramolecular π-
stacking interactions in the N-heterocyclic molecular balances were assessed by their
closed-open equilibrium ratios. These were measured for balances 1-7 in solution via
1
integration of their H NMR spectra at 23 °C. The polar solvent DMSO-d6 was
chosen for these studies due to its ability to dissolve both the neutral and cationic
balances. The equilibrium ratios were measured from the areas of the separate peaks
for the open and closed succinimide methylene protons, which were singlets at ~4.9
ppm. The assignment of the open succinimide peak was made by matching its area to
the distinct doublet of doublet for the open conformer’s ortho-proton of the N-aryl
rotor, which was shifted drastically upfield (~ 4.5 ppm) due to its proximity over the
aromatic shelf. The closed/open equilibrium ratios were converted to energies using
the equation: ΔG = –RTln([closed]/[open]). The closed/open ratio for balance 7 was
estimated to be greater than 19:1 because only the closed conformer was observed.
For some of the cationic balances, the closed/open ratios were measured
indirectly. The cationic N-protonated balances 4 and 5 were formed by adding MsOH
to solutions of balances 2 and 3, respectively. Unfortunately, in the presence of 1
equiv. of MsOH, the broad water peak obscured the key succinimide singlets that
were used to measure the closed/open ratios (See SI Figure S17). The closed/open
ratios, however, could be accurately measured at lower equiv. (< 1.0) of MsOH and
14
ACS Paragon Plus Environment

Page 15 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
the equilibrium energies (ΔG’s) were observed to change linearly with the amount of
added MsOH (See SI Figure S17 and S18). Therefore, ΔG’s for balances 4 and 5
were obtained from the linear extrapolation to 1 equiv. of added MsOH.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Another concern was that the differences in the substituents at the bridgehead
carbons of the balances might not allow accurate comparison. Balances 1, 2, 4, and 6
have bridgehead phenyl rings and balances 3, 4, and 7 had bridgehead methyl groups.
Thus balance 1” with a bridgehead methyl group was prepared and compared with
balance 1 with a bridgehead phenyl group. Balances 1 and 1” had idential
open/closed ratios in DMSO-d6 (0.58 verus 0.53), suggesting that the bridgehead
group did not influence the open/closed equilibrium ratio.
Quantitative analysis of the stability trends. Comparison of the measured
ΔG’s for balances 1-7 is shown in Figure 8. Stronger π-stacking interactions in the
closed conformer led to higher closed/open ratios and more negative ΔG values. In
general, the ΔG values were consistent with the expected N-heterocyclic π-stacking
trends, confirming the ability of the balances to accurately compare these π-stacking
interactions. For example, neutral N-heterocyclic balances 2-3 showed stronger π-
stacking interactions than the non-heterocyclic balance 1.29 Cationic N-heterocyclic
balances 4-7 also showed stronger π-stacking interactions than the neutral N-
heterocyclic balances 2-3.27
15
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 34
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. 1H NMR measured ΔG’s (kcal/mol) for balances 1-7 at 23 °C in DMSO-d6
at 23 °C. The ChemDraw representation of the π-stacking interaction in each balance
is shown. The uncertainty in 1H NMR integration measurements gives an error of less
than ± 0.03 kcal/mol.
To facilitate the comparison of the respective π-stacking interactions, ΔΔG’s
were calculated for the N-heterocyclic balances 2-7 relative to the non-heterocyclic
balance 1, which contained the weakest π-stacking interaction (Table 2).
Table 2. Measured conformational equilibrium energies (ΔG, kcal/mol) for balances
1-7 in DMSO-d6 at 23 °C and relative energies (ΔΔG, kcal/mol) of the N-heterocyclic
balances 2-7 relative to the non-heterocyclic balance 1.
π-stacking
interaction
Balance
geometry
ΔGa
ΔΔGb
1
2
3
4
5
6
7
Ph–Ph
–
0.32
-0.22
0.00
-0.54
-0.37
-0.77
-1.75
-1.08
distal
Py–Ph
proximal -0.05
distal -0.45
proximal -1.43
distal -0.76
PyH+–Ph
PyMe+–Ph
proximal ≤-1.74c ≤-2.06
a DMSO-d6, 23 °C; uncertainty less than ± 0.03 kcal/mol b ΔΔG = (ΔG[x] –
c
ΔG[1]); x = 2-7; uncertainty less than ± 0.06 kcal/mol Only the closed conformer of
balance 7 was observed on the H NMR spectrum, and thus ΔG[7] was estimated to
1
be ≤ -1.74 kcal/mol.
The neutral N-heterocyclic (Py–Ph) π-stacking interactions in balances 2 and 3
were found to be modestly stronger (ΔΔG = -0.4 to -0.5 kcal/mol) than the non-
heterocyclic (Ph–Ph) π-stacking interaction in balance 1. This was consistent with the
influence of the electronegative heterocyclic nitrogen that enhances the electrostatic
attraction between rings due to the formation of a permanent dipole.29 The magnitude
16
ACS Paragon Plus Environment

Page 17 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
of this trend compared favorably with the values measured by Gung and co-workers
using their triptycene-based molecular model system.33
7
8
9
The cationic N-heterocyclic π-stacking interactions (PyH+–Ph and PyMe+–Ph)
in balances 4-6 were found to be much stronger than the non-heterocyclic (Ph–Ph) π-
stacking interaction in balance 1 (ΔΔG = -0.8 to ≤-2.1 kcal/mol), and also stronger
than the neutral Py–Ph π-stacking interactions in balances 2-3 (ΔΔG = -0.2 to ≤-1.7
kcal/mol). This significant increase in strength was consistent with the greater
electrostatic attraction arising from cationic nitrogen ring and the π-cloud of the
opposing aromatic surface.27 This observation helped explain the prevalence of
cationic N-heterocyclic π-stacking interactions in applications that rely on π-stacking
interactions such as biomimetic catalysis of SN2 reactions,48 conformational molecular
switches,49 diastereotopic control of stereoselective reactions,50 and the threading of
rotaxanes and catanenes.51
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Influence of the position of N-heterocyclic nitrogen. The rigidity of the
bicyclic framework of this model system provided the unique opportunity to
investigate how the relative position of the N-heterocyclic nitrogen atom influenced
the stabilities of the π-stacking interaction. However, the ability of the molecular
balances to maintain the distal and proximal geometries in solution was first verified.
Balances 2-7 adopted the expected distal and proximal geometries in the solid-state.
However, these geometries are more dynamic in solution. Computational studies of
our model system have shown that the arene arm can sweep back and forth over the
arene shelf.52 Thus, the conformational equilibrium energies in solution of the
proximal (2, 4 and 6) and distal (3, 5 and 7) balances were compared. Specifically,
for each type of N-heterocyclic π-stacking interaction (Py–Ph, PyH+–Ph and PyMe+–
Ph), the energies of balances designed to adopt distal and proximal geometries were
17
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 34
1
2
3
4
5
6
7
8
9
compared. In each case, the distal and proximal balances had distinctly different ΔG
values. More importantly, the distal versus proximal stability trends for each type of
interaction matched well the predictions of Hunter and Sander’s electrostatic model.53
This suggests that the balances were adopting the expected geometries in solution.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For the Py–Ph stacking interaction, the distal geometry (balance 2) was found
to be slightly stronger than the proximal geometry (balance 3) by ΔΔG = -0.17
kcal/mol. Figure 9 shows the electrostatic origins of this geometric preference. The
electronegative nitrogen of the pyridine ring bears a partial negative charge (δ–) and
creates a partial positive charge (δ+) on the opposite end of the ring. Thus, the distal
geometry provides better electrostatic complementarity to the quadrupole of the
opposing phenyl ring. The proximal geometry, on the other hand, lacks an
electropositive hydrogen on the pyridine nitrogen which eliminates one of the
attractive electrostatic interactions. In addition, the proximal geometry contains an
additional repulsive interaction between the electronegative pyridine nitrogen and the
π-surface of the opposing phenyl ring. The weaker proximal Py–Ph stacking
interaction in balance 3 was also evident in the solid-state studies, as balance 3 was
the only balance that did not crystalize in the closed conformation.
"
!
+
N
!
N
"
!
"
"
!
+
+
!
!
Py-Ph stacking
!
distal
"
!
+
N
!
N
"
!
"
"
!
!
+
+
!
!
Py-Ph stacking
proximal
Figure 9. Schematic representations comparing the neutral Py–Ph π-stacking
interactions in the proximal and distal geometries using Hunter and Sander’s
electrostatic quadruple model.
18
ACS Paragon Plus Environment

Page 19 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
For the cationic N-heterocyclic (PyH+–Ph and PyMe+–Ph) π-stacking
interactions, the opposite geometric preference was observed. The proximal geometry
(balances 5 and 7) were considerably more stable than the distal geometry (balances 4
and 6) by ΔΔG = ≤-0.98 kcal/mol. The preference for the proximal geometry in the
cationic N-heterocyclic π-stacking interactions can also be explained using Hunter’s
electrostatic model. The positive charge caused by protonation or alkylation of the
pyridine is electrostatically attracted to the π-cloud of the opposing phenyl ring. This
attractive force is stronger in the proximal geometry as the positively charged nitrogen
is closer to the face of opposing phenyl ring. The observed preference for the
proximal geometry in PyH+–Ph stacking interactions is also consistent with the
“orientation effect” of cationic π-stacking interactions observed by Waters and co-
workers in their flexible folding model system.35 Furthermore, the larger difference in
energy between the proximal and distal geometries for the cationic (|ΔΔG| ≥ 0.98
kcal/mol) versus neutral (|ΔΔG| = 0.17 kcal/mol) π-stacking interactions reinforced
the electrostatic origins of these geometrical preferences.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Influence of the N-substituent. Next the influence of different N-substituents
(methyl versus proton) in the cationic balances was analyzed. To isolate this variable
from the geometric preference noted above, the proximal (5 and 7) and distal (4 and
6) balances were compared separately. For the proximal balances, the N-methylated
balance 7 had a stronger π-stacking interaction than the N-protonated balance 5 by
ΔΔG ≤ -0.31 kcal/mol. The greater stability of the proximal N-methylated
heterocycle complex was attributed to the ability of the N-methyl group to form an
additional stabilizing CH–π interaction. Evidence for the additional CH–π interaction
was provided by the crystal structure of balance 7. As shown in Figure 10, a proton
of the N-methyl group points towards the opposing phenyl ring with two short H to C
19
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 34
1
2
3
4
5
6
7
8
9
distances (atom-to-atom distance) of 3.02 Å and 3.03 Å, respectively, which falls
within the common cut-off distance (3.05 Å) for a CH–π interaction.54 We have also
observed that similar intramolecular CH–π interactions can stabilize the closed
conformer of our N-arylimide bicyclic balances by up to -1.0 kcal/mol in organic
solvents.41 The similar supporting role of CH–π interactions in the cation–π
interactions of alkylated ammonium has been mentioned in several studies, such as
the folding of a β-hairpin peptide and the in vitro binding of trimethyllysine to the
HP1 chromodomain.55
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A similar analysis was performed on the distal cationic balances (4 and 6).
However, the equilibrium ratios of these distal balances were much more susceptible
to solvent effects which made the N-methyl versus N-proton preference solvent
dependent. These solvent effects are explained in the next section.
Figure 10. Truncated side-view (left) of the crystal structure of closed-7 highlighting
the presence of a CH-π interaction between the proximal N-methylated heterocycle on
the shelf and the phenyl ring on the arm. A ChemDraw representation is also provided
(right). The short distance contacts between N-methyl proton and phenyl carbon are
highlighted with red dotted lines.
Solvent effects. The ability to measure the closed/open ratios in different
solvents provided an additional method of probing the origins of the stability trends.
Thus, the above analyses were also carried out in two additional organic solvents:
CDCl3 and CD3CN. The measured relative energies are shown in Table 3. Overall,
the similar stability trends were observed in all three solvents, providing further
confirmation of the observed N-heterocyclic stability trends.
20
ACS Paragon Plus Environment

Page 21 of 34
The Journal of Organic Chemistry
1
2
3
4
5
Table 3. Measured relative energies (ΔΔG, kcal/mol) a of the N-heterocyclic balances
2-7 relative to the non-heterocyclic balance 1.
6
7
8
9
ΔΔGa
CD3CN DMSO-d6
π-stacking
interaction
Ph–Ph
Balance
geometry
CDCl3
0.00
1
2
3
4
5
6
–
0.00
-0.54
-0.21
-1.06
-1.66
-0.98
≤-1.89
0.00
-0.54
-0.37
-0.77
-1.75
-1.08
≤-2.06
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
distal
proximal
distal
proximal
distal
proximal
-0.60
-0.19
-0.89
-1.43
-0.40
≤-2.23
Py–Ph
PyH+–Ph
PyMe+–Ph
7
a
The relative energy were quantified as ΔΔG = (ΔG[x] – ΔG[1]); x = 2-7;
propagated uncertainty less than ± 0.06 kcal/mol. The measured ΔG’s for balances 1-
7 in CDCl3 and CD3CN were included in Table S10
The only inconsistency was in the trends for the distal N-protonated and N-
methylated cationic balances 4 and 6, as noted in the last section. In CDCl3, the N-
protonated balance 4 had a stronger interaction, whereas in DMSO-d6, the N-
methylated balance 6 had a stronger interaction. In CD3CN, the two balances had
very similar interaction energies. The greater solvent effects for these distal balances
may be due to the greater interaction of the solvents with the charged cationic surfaces
in 4 and 6. This may be due to the more solvent exposed positive charge of the distal
geometry. By comparison, the more shielded positive charge of the proximal
balances (5 and 7) show much smaller solvent effects and consistent N-methyl versus
N-proton trends in each solvent. These differences could also be due to differences in
solvation of the different counter-ions (I- and MsO-) for the N-methylated and N-
protonated distal balances (4 and 6), respectively.
Correlation with benchmark-quality computational calculations. Finally,
the experimentally measured values in this study provided an opportunity to test the
accuracy of current computational methods as well as the performance of our model
system in accurately assessing N-heterocyclic π-stacking interactions. Benchmark
quality calculations (CCSD(T) or SCS–MP2 methods) from the literature for π-
21
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 34
1
2
3
4
5
6
7
8
9
stacking interactions were used in the correlation analysis (Figure 11).27,29,56 The
geometries of the calculations were matched to the geometries of the π-stacking
interactions observed in the crystal structures of the balances, and the calculated
interaction energies were matched to the experimentally measured relative energies
(ΔΔG) in DMSO-d6 from Table 2.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 11. Correlation plot between the calculated energies from the benchmark
quality computations in literature and the experimentally observed relative energies
(ΔΔG’s) in DMSO-d6 for the π-stacking interactions in balances 1-7. The
computational data were from ref. 27 (open diamonds), ref. 29 (open squares) and ref.
56 (open triangles).
The correlation plot showed a good general agreement between the
computational and experimentally measured energies. For example, both methods
provide similar stability trends: Ph–Ph < Py–Ph < (PyH+–Ph and PyMe+–Ph). There
was also good agreement for the geometry-related stability trends (distal versus
proximal) for the both neutral and cationic N-heterocyclic π-stacking interactions. For
example, both methods found that the distal geometry was more stable for the neutral
Py–Ph stacking interaction; whereas the proximal geometry was more stable for the
cationic PyH+–Ph and PyMe+–Ph stacking interactions.
22
ACS Paragon Plus Environment

Page 23 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Two minor deviations in the experimental and calculated trends were
observed. The first deviation was that the neutral and cationic N-heterocyclic π-
stacking interactions appeared to form separate trend lines on the correlation plot.
This may be explained by the different environments (vacuum versus solvent) of the
two methods. The solvent in the experimental studies screen the electrostatic
interactions of charged groups, which reduces their interaction strengths as compared
to the in vacuo computational studies.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The second inconsistency was that the experimental study found that the
proximal Py–Ph interaction in 3 was slightly stronger than the Ph–Ph stacking
interaction in 1. This trend is in opposition to the computational studies that predicted
that positioning an electronegative nitrogen over an opposing aromatic π-surface in
the proximal geometry should be a destabilizing interaction. We attribute this
deviation to the mobility of the phenyl arm in solution, which allows the balance to
also adopt geometries that avoid this destabilizing interaction. This minimizes the
contribution of the destabilizing interaction to the closed/open ratio and leads to an
overestimation of the strength of the proximal Py–Ph stacking interaction in balance
3. It should be pointed out that the mobility of the arene arm will have less of an
impact on balances that only contain attractive stabilizing interactions, as the arm will
preferentially seek out and adopt these more stable geometries.
Evaluation of the design of the molecular balances. Overall, our N-
arylimide bicyclic balance model system appeared to provide an effective platform to
systematically and quantitatively study many aspects of the N-heterocyclic π-stacking
interaction. However, as with all model systems, we had a number of concerns. In
this section, we re-analyzed the data to check whether each concern was warranted.
23
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 34
1
2
3
4
5
6
7
8
9
The first concern was that the repulsive O-π interaction from the oxygen linker
of the arm with the arene shelf might obscure or prevent accurate comparisons of the
intramolecular π-stacking interactions of interest. However, this repulsive interaction
was actually advantageous feature, which enabled more accurate measurement of the
stronger N-heterocyclic π-stacking interaction. The strength of the intramolecular
interactions that can be measured by our balances is limited by the highest
closed/open ratio that can be accurately measured. Based on a conservative detection
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
limit of 5% of H NMR spectra,57,58 the highest closed/open ratio that can be
measured is 19:1. If the closed/open ratio in the absence of the interaction was 1:1,
this would correspond to a maximum measureable interaction energy of only -1.74
kcal/mol. This low limit is particularly important for this study because the cationic
N-heterocyclic π-stacking interactions were all stronger than -2.0 kcal/mol. However,
the repulsive interactions of the oxygen linkers in our balances acts like a “spring”,
which shifts the resting state of the balance toward the open conformer to give a
closed/open ratios of approximately 1:7 in DMSO-d6.40 This spring effect allows for
a larger change in closed/open ratio from 1:7 to 19:1, which corresponds to a
maximum measurable interaction of up to -2.88 kcal/mol.
A second design concern was whether we could compare balances with N-
heterocycles in the arm against those with N-heterocycles in the shelf. These two
surfaces are not identical and possess different degrees of π-conjugation. However,
analysis of the ΔG’s for balances 1-7 suggested that this concern was unwarranted or
was very minor, as the expected stability trends; Ph–Ph < Py–Ph < (PyH+–Ph and
PyMe+–Ph), were clearly observed. More importantly, the expected and more subtle
geometry-related stability trends (proximal versus distal) were accurately reproduced
24
ACS Paragon Plus Environment

Page 25 of 34
The Journal of Organic Chemistry
1
2
3
4
by the molecular balances. These geometry-related trends relied on directly
5
6
comparing balances with N-heterocycles in the arm and shelf.
7
8
9
A third design concern was whether the mobility of the arm arene in the closed
conformer balances would compromise their ability to distinguish the distal and
proximal geometries of the N-heterocyclic π-stacking interaction. Generally, the
expected distal versus proximal stability trends were observed. However, the ability
of the arene arm to slide back and forth across the arene shelf could lead to
underestimations of the difference in energy of proximal and distal geometries when
destabilizing repulsive interactions are involved. To address this problem, we are
currently utilizing advanced NMR methods such as through-space coupling and
nuclear Overhauser effects to investigate the mobility of the arene arm and measuring
the population of various sub-conformations to refine the energy calculations.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The last design concern was the asymmetry of the cationic aromatic shelves in
balances 5 and 7. Ideally, both heterocyclic nitrogens would be charged so that the
arm arene could form the same π-stacking interactions with either outer ring of the
arene shelf, but this was not synthetically feasible. Clearly, these singly charged
balances are still able to form the expected cation-π interactions, as these balances had
significantly stronger π-stacking interactions than their neutral counterparts. This
highlights one advantage of arm mobility in the study of attractive interactions, as the
arms will preferentially form the more stabilizing π-stacking interaction with the
cationic versus neutral outer ring. However, we cannot eliminate the possibility that
the ability of these balances to also form the neutral π-stacking interaction may lead to
an underestimation of the energies of the cationic π-stacking interactions in balances 5
and 7. Again, advanced NMR techniques will be applied in future studies to quantify
the contribution of these less stable geometries.
25
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 34
1
2
3
4
■ CONCLUSIONS
5
6
7
8
9
A new series of molecular balance model systems (2-7) were designed and
used to perform a comprehensive study of the influences of position, charge, and
substitution of the nitrogen atom on N-heterocyclic π-stacking interactions in the
solid-state and in solution. The formation and geometry of the intramolecular N-
heterocyclic π-stacking interactions in the closed conformation models were verified
via X-ray crystallographic analysis in the solid-state. The relative stabilities of these
N-heterocyclic π-stacking interactions in solution were systematically studied by
comparing the closed/open ratios of the corresponding molecular balances that were
measured via integration of the 1H NMR spectra. The stability trend: Ph–Ph < Py–Ph
< (PyH+–Ph and PyMe+–Ph) was quantitatively measured for the first time within a
single model system. The position of the heterocyclic nitrogen was found to influence
the strengths of both the neutral and cationic N-heterocyclic π-stacking interactions
leading to opposite stability trends. For the neutral Py–Ph stacking interaction, the
distal geometry was found to be slightly more favorable. For the cationic PyH+–Ph
and PyMe+–Ph stacking interactions, the proximal geometry was found to be
significantly more favorable. The influence of N-methylation and N-protonation was
also compared. For the proximal complex, the PyMe+–Ph interaction was found to be
significantly stronger than the PyH+–Ph stacking interaction. This was attributed to
the formation of additional CH-π interaction between the N-methyl proton and the
opposing phenyl ring. The measured trends were consistently observed in three
common organic solvents (chloroform, acetonitrile and DMSO) confirming the
generality of these trends. The experimentally measured stabilities for the N-
heterocyclic π-stacking interactions in balances 1-7 were found to generally agree
with the corresponding benchmark quality computational values in literature.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
26
ACS Paragon Plus Environment

Page 27 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Discrepancies were mainly attributed to the presence of solvent in the experimental
studies and the difficulties in experimentally measuring interactions strength of
unfavorable geometries.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The quantitatively measured stability trends for the N-heterocyclic π-stacking
interactions provide guidance in the design and optimization of applications that rely
on these molecular recognition processes. For example, insertion of N-heterocycles
will strengthen attractive aromatic π-stacking interactions but only if the nitrogen
atoms are appropriately positioned away from the π-face of the opposing aromatic
surface. The insertion of positively charged N-heterocycles also provide a simple
means to significantly strengthen the π-stacking interaction, but these charged
heteroarenes also impose an even stronger directional preference. Finally, to study
the stability trends for the non-covalent interactions of aromatic surfaces in more
biologically relevant aqueous environments, we have developed a water-soluble
version of these molecular balances and will be reporting these results in future
studies.
■ EXPERIMENTAL SECTION
General Experimental. All chemicals were purchased from commercial
suppliers and used as received unless otherwise noted. All reactions were carried out
under a dry N2 atmosphere in oven–dried glassware. Dry organic solvents were
obtained by passing the degassed solvents through activated alumina columns. Flash
chromatography was carried out using either silica gel (60 Å, 200–400 mesh) or
aluminum oxide (60 Å, 20–200 mesh) as noted. Thin layer chromatography (TLC)
for monitoring the reaction progress was performed using either precoated 0.25 mm
silica gel 60 F254 plates or precoated 0.25 mm aluminum oxide 60 F254 plates.
NMR spectra were recorded on 300 MHz, 400 MHz spectrometers. Chemical shifts
27
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 34
1
2
3
4
5
6
7
8
9
are reported in ppm (δ) and were referenced to TMS. HRMS was recorded with a
magnetic sector spectrometer using EI sources and the Q-TOF 1 spectrometer using
ESI sources.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Balances 1 and 1’, anhydride 11, N-(2-phenyloxyl)phenyl maleimide (17), and
N-(2,6-diphenoxylphenyl) maleimide were prepared utilizing our previously described
routes for the analogous parent compounds.40 Diene 1344 and ortho-pyridin-3-
ylaniline (10)42 were obtained using literature procedures.
(9R,9aS,12aR,13S)-9,13-Dimethyl-11-(2-phenoxyphenyl)-12a,13-dihydro-9H-
9,13-methanophenanthro[9,10-f]isoindole-10,12,14(9aH,11H)-trione (1’’). Diene 16
(0.20 g, 0.77 mmol) and N-(2-phenyloxylphenyl) maleimide (17) (0.24 g, 0.92 mmol)
were dissolved in dry toluene (10 mL). The mixture was stirred and heated at reflux
under N2 for 24 h. The solvent was removed under vacuum to give a light yellow
solid. Recrystallization from ethanol gave diene 1’’ as a white crystalline solid (0.36
1
g, 87%). mp 310 ºC dec. H NMR (400 MHz, CDCl3) δ 8.79 (dd, J = 8.3 Hz, J = 1.2
Hz, 2H major), 8.38 (dd, J = 8.2 Hz, J = 1.3 Hz, 2H major), 8.26 (dd, J = 8.3 Hz, J =
1.0 Hz, 2H minor), 8.14 (dd, J = 8.4 Hz, J = 1.0 Hz, 2H minor), 7.61–7.72 (m, 4H
major), 6.81–7.49 (m, 5H major, 12H minor), 6.65 (dd, J = 8.3 Hz, J = 1.2 Hz, 1H
major), 6.44 (td, J = 7.7 Hz, J = 1.3 Hz, 1H major), 5.96 (dd, J = 8.4 Hz, J = 1.2 Hz,
1H minor), 5.73–5.78 (m, 1H major), 4.68 (dd, J = 7.9 Hz, J = 1.6 Hz, 1H major),
13
3.60 (s, 2H minor), 3.42 (s, 2H major), 2.31 (s, 3H major), 2.28 (s, 3H minor). C
NMR (100 MHz, CDCl3): 201.1, 199.6, 174.2, 173.7, 156.2, 153.0, 133.3, 133.1,
130.9, 130.8, 130.4, 129.9, 129.3, 128.9, 128.2, 127.5, 127.4, 127.3, 126.7, 126.5,
125.3, 125.0, 124.2, 123.5, 123.4, 123.3, 121.8, 121.7, 121.6, 119.5, 118.8, 116.1,
77.5, 77.2, 77.2, 76.8, 55.2, 55.1, 48.4, 47.9, 14.8. HRMS (EI) m/z calcd for
C35H25NO4 (M+): 523.1784; found: 523.1778.
28
ACS Paragon Plus Environment

Page 29 of 34
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
(9R,9aS,12aR,13S)-9,13-Diphenyl-11-(2-(pyridin-3-yloxy)phenyl)-12a,13-
dihydro-9H-9,13-methanophenanthro[9,10-f]isoindole-10,12,14(9aH,11H)-trione (2).
To a solution of anhydride 11 (0.29 g, 0.60 mmol) in acetic acid (25 mL), ortho-
pyridine-3-ylaniline (10) (0.15 g, 0.81 mmol) was added. The mixture was stirred and
heated at reflux under N2 for 24 h. The solvent was removed under vacuum to give a
brown solid. The solid was suspended in Na2CO3 aqueous solution (50 mL), and then
extracted with CH2Cl2 (3 x 30 mL). The organic layer was collected, washed with
brine (30 mL) and dried over MgSO4. The solvent was removed under vacuum to
give a brown solid. Flash chromatography (aluminum oxide, MeOH/CH2Cl2 = 1/100)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
provided 2 as a pale white solid (0.29 g, 81%). mp 280 ºC dec. H NMR (400 MHz,
CDCl3) δ 8.73 (d, J = 8.5 Hz, 2H minor), 8.45–8.35 (m, 2H minor, 2H major), 8.33–
8.26 (m, 1H major, 2H minor), 8.16 (d, J = 8.5 Hz, 2H major), 6.86–7.75 (m, 19H
major and 16H minor), 6.65 (dd, J = 8.3 Hz, J = 1.1 Hz, 1H minor), 6.47 (td, J = 7.9
Hz, J = 1.1 Hz, 1H minor), 6.18 (ddd, J = 8.3 Hz, J = 2.6 Hz, J = 1.4 Hz, 1H major),
5.93 (dd, J = 8.3 Hz, J = 1.1 Hz, 1H major), 4.68 (s, 2H major), 4.64 (dd, J = 7.9 Hz,
13
J = 1.6 Hz, 1H minor), 4.58 (s, 2H minor). C NMR (100 MHz, CDCl3): 197.0,
195.7, 173.5, 173.0, 154.4, 152.5, 150.5, 145.7, 145.5, 143.8, 142.4, 133.9, 133.7,
133.7, 133.5, 131.3, 131.2, 131.1, 130.9, 130.7, 129.5, 129.4, 129.3, 129.2, 129.1,
128.9, 128.8, 128.6, 128.5, 128.4, 127.4, 127.0, 126.8, 126.7, 126.4, 126.3, 126.2,
126.0, 125.9, 124.4, 124.1, 123.9, 123.1, 122.9, 122.6, 122.2, 120.8, 118.4, 116.0,
77.5, 77.2, 77.1, 76.8, 63.7, 63.6, 45.6, 45.2. HRMS (EI) m/z calcd for C44H28N2O4
(M+): 648.2049, found: 648.2059.
(9R,9aS,12aR,13S)-9,13-Dimethyl-11-(2-phenoxyphenyl)-12a,13-dihydro-9H-
9,13-methanoisoindolo[5,6-f][1,10]phenanthroline-10,12,14(9aH,11H)-trione
(3).
Diene 13 (0.20 g, 0.77 mmol) and N-(2-phenyloxylphenyl) maleimide (17) (0.24 g,
29
ACS Paragon Plus Environment