On the role of guests in enforcing the mechanism of action of gated baskets

Article abstract of DOI:10.1039/c3ob41511b

We designed and prepared a spacious and gated basket of type 2 (V = 318 A³) in ten synthetic steps. With the assistance of ¹H NMR spectroscopy, we found that the pyridine gates at the rim of 2 form a seam of N-H...N hydrogen bonds, t

Full text of DOI:10.1039/c3ob41511b

Organic&
BiomolecularChemistry
www.rsc.org/obc
Volume 11 | Number 44 | 28 November 2013 | Pages 7633–7804
ISSN 1477-0520
PAPER
Jovica D. Badjić et al.
On the role of guests in enforcing the mechanism of action of gated baskets

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
On the role of guests in enforcing the mechanism of
action of gated baskets†
Cite this: Org. Biomol. Chem., 2013, 11,
7667
Yian Ruan, Bao-Yu Wang, Jeremy M. Erb, Shigui Chen, Christopher M. Hadad and
Jovica D. Badjić*
We designed and prepared a spacious and gated basket of type 2 (V = 318 Å3) in ten synthetic steps.
With the assistance of 1H NMR spectroscopy, we found that the pyridine gates at the rim of 2 form a
seam of N–H⋯N hydrogen bonds, thereby adopting right- (P) and left-handed (M) helical arrangements.
The recognition characteristics of the smaller basket 1 (V = 226 Å3) and the larger 2 for various solvents
as guests were quantified by 1H NMR spectroscopy in CD2Cl2 (61 Å3), CDCl3 (75 Å3), CFCl3 (81 Å3) and
CCl4 (89 Å3); the apparent guest binding equilibria Ka were found to be inversely proportional to the
affinity of bulk solvents KS for populating each host. The rate of the P/M racemization (krac, s−1) was, for
both 1 and 2, studied in all four solvents using dynamic NMR spectroscopy. From these experiments, two
isokinetic relationships (ΔS‡P/M vs. ΔHP‡/M) were identified with each one corresponding to a different
mechanism of P/M racemization. A computational study (B3LYP/6-31+G**//PM6) of 1 and 2 in the gas
phase indicates two competing racemization pathways: (a) RM1–2 describes a pivoting of a single gate
followed by the rotation of the remaining two gates, while (b) RM3 depicts simultaneous (geared)
rotation of all three gates. The racemization of the larger basket 2, in all four solvents (packing coeffi-
cient, PC = 0.19–0.28), conformed to one isokinetic relationship, which also coincided with the operation
of the smaller basket 1 in CD2Cl2 (PC = 0.27). However, in CDCl3, CFCl3 and CCl4 (PC = 0.33–0.39), the
mode of action of 1 appears to correlate with a different isokinetic relationship. Thus, we propose that
the population of the basket’s inner space (PC) determines the mechanism of P/M racemization. When
PC < 0.3, the mechanism of operation is RM1–2, whereas, a greater packing, represented when PC > 0.3,
enforces the geared RM3 mechanistic alternative.
Received 7th July 2013,
Accepted 3rd September 2013
DOI: 10.1039/c3ob41511b
www.rsc.org/obc
concave hosts while competing with other guests.13 Markedly,
sizeable molecules incapable of penetrating the capsular host
Introduction
A complete encirclement of guests by covalent/self-assembled and effectively solvating its concave surface14 permit the for-
hosts gives rise to encapsulation complexes.1,2 A conformation- mation of more stable encapsulation assemblies.15 In fact,
al change within the structure of the host,3 of such complexes, one could show (vide infra) that organic solvents capable
could govern the in/out trafficking of guests,4 and thus the of occupying the capsular host with an affinity of KS > 10−2
–
encapsulation persistency (ΔG‡in/out).5 For the entrapment to 10−3 M−1 weaken the binding to the point that there is
take place in the liquid phase, the size and shape of guests almost no measurable complexation.16 In a similar manner,
ought to be complementary6 to the interior of the host.7,8 Fur- polar water molecules are incompatible with the interior of
thermore, ∼55% population (packing coefficient, PC = Vguest
/
hydrophobic organic hosts, thereby assisting guest complexa-
Vhost)9 of the cavitand’s inner space10 has been found to con- tion via the hydrophobic effect.19,20 Evidently, the effective
tribute to the formation of stable van der Waals encapsulation formation of encapsulation complexes21 is a function of the
complexes.11 The role of bulk solvent is important12 as solvent solvophobic effect,22 yet finding a suitable medium for explor-
molecules are in a constant quest to occupy the inner space of ing encapsulation is still
investigation.23
a
matter of experimental
Our mechanistic studies of gated baskets of type 1 (226 A3,
Fig. 1A) have, so far, been completed in dichloromethane
(CD2Cl2, 61 Å3), and the affinity for guests was satisfactory.24
Accordingly, in this environment, the basket can be classified
as solvophobic: one CD2Cl2 (61 A3, PC = 28%) is not enough to
populate its cavity, two CD2Cl2 are not complementary to the
Department of Chemistry and Biochemistry, 100 West 18th Avenue, Columbus, Ohio,
USA. E-mail: badjic@chemistry.ohio-state.edu; Fax: +614-292-1685;
Tel: +614-247-8342
†Electronic supplementary information (ESI) available: Additional experimental/
computational results and spectra for new compounds. See DOI: 10.1039/
c3ob41511b
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7667–7675 | 7667

View Article Online
Paper
Organic & Biomolecular Chemistry
Results and discussions
Design and preparation of basket 2
To make a bigger gated host, one could extend the phthali-
mide groups in basket 1 (Fig. 1A) by incorporating quinoxaline
moieties into the tris-norbornane framework (Fig. 1B).18
Indeed, we already optimized a template-directed protocol for
the preparation of bowl-shaped 4syn (Fig. 2A).18 In accord with
this methodology, compound 3 was cyclotrimerized to give syn
diastereomer of 4 with Cu(I)/Cs(I) as templating cations and
Pd(0)/Cu(I) as catalysts (Fig. 2A). The synthetic strategy could
therefore be useful for the preparation of more spacious
basket 2 (Fig. 1B), although we had to experimentally test the
cyclotrimerization of 9 under the same or a similar set of con-
ditions (Fig. 2B).18 Dimethyl-4,5-diaminophthalate 5 (Fig. 2B)
was obtained in gram quantities following already reported
procedures (Scheme S1†).35,36 The condensation of this com-
pound with 5-norbornene-2,3-dione 6 (CH2Cl2) was facile and
provided 7 in 92% yield. The bromination/dehydrobromina-
tion of 7 gave bromoalkene 8 (60%), which was further stanny-
Fig. 1 Chemical structures of basket 1 (A) and 2 (B) forming intramolecular N–
H⋯N hydrogen bonds and their corresponding energy-minimized forms (PM6).
The volume of each basket’s cavity was computed by the 3 V computational
method17 while their van der Waals surfaces are visualised using UCSF Chimera
software.
lated at low temperature (195 K) to form compound
(Fig. 2B). The cyclotrimerization of bromo(trimethylstannyl) 9
was attempted with Pd(0)/Cu(I)/CsF catalytic system
9
a
(Table S1†). As discussed earlier, Cu(I)/Cs(I) metal cations were
expected18 to template the formation of 10syn by coordinating
to quinoxaline nitrogen atoms within the basket’s framework.
inner concave surface (122 Å3, PC = 55%), while three CD2Cl2 Interestingly, diastereomeric 10syn and 10anti were obtained in
are simply too much for occupying the interior (183 Å3, PC = a nearly statistical ratio (∼1 : 3) and overall 60% yield. Given
83%); note that dissolved gases (N2, O2, Ar, etc.)25 could poten- the absence of templation, the reaction’s outcome is attributed
tially fill the cavity of cavitands as well.26 Importantly, numer- to both electronic and steric factors: (1) less basic quinoxaline
ous haloalkanes (∼80–125 Å3)27 were found to substitute nitrogen atoms in 9, carrying two electron-withdrawing ester
weakly bound solvent within 1 in a process that is first-order functionalities, ought to have a lower coordination affinity
in both guest and basket (vin = kin [basket] [guest]).28 The recog- toward metal cations, and (2) six ester groups at the rim of
nition is driven by enthalpy (ΔH° < 0)27 while the entropic 10syn could hinder the formation of a more compact transition
contributions are adverse in most cases (ΔS° < 0).5 In this state that is required for its formation. The base-catalyzed
basket, the three hydrogen-bonded pyridines (gates) revolve at (LiOH) hydrolysis of hexamethylester 10syn followed by treat-
the rim adopting right- (P) and left-handed (M) orientations.29 ment of the hexaacid with trifluoroacetic anhydride (TFAA,
The rotation creates a sizeable aperture at the basket’s north- Fig. 2B) gave tris-anhydride 11. The conjugation of diamine 12
ern side to permit the solvent/guest exchange.30 Regulation of to 11 gave tris-amine 13, which upon condensation with TFAA
the dynamics of the fluctuating entrance is important for con- afforded basket 2 in 41% yield.
trolling the chemoselectivity29,31 by which molecules access
NMR characterization of basket 2
the basket’s cavity as well as the lifetime of encapsulation com-
plexes.32 Importantly, the hydrogen-bonded gates could rotate 1H NMR spectrum of basket 2 (400 MHz, CDCl3) showed a set
without guest species entering/departing the basket’s inner of signals corresponding to
a C3v symmetric molecule
space.30 Will a more sizeable basket 2 (318 A3, Fig. 1B) also (Fig. 3A); 1H–1H NOESY spectroscopic correlations (Fig. S1†)
form a gated host? How will a bigger cavity affect the mechan- allowed us to assign all of the basket’s resonances. Notably,
ism of in/out guest exchange? Does the P/M racemization the singlet corresponding to the three amide N–H protons is
mechanism in any way relate to the size of the basket’s positioned downfield (δ = 11.9 ppm, Fig. 2C), confirming the
interior? Does the guest size have any effect on the mechanism formation of N–H⋯N hydrogen bonds. In fact, a dilution of
by which baskets operate? To address these questions, we used CDCl3 solution of 2 (from 0.8 to 0.1 mM, Fig. S2†) caused a
methods of both experimental and computational chemistry negligible shift of the N–H resonance to indicate the absence
and examined the inner solvation, conformational dynamics, of intermolecular aggregation. In line with this result, NMR
and encapsulation characteristics of 1 and 2. Interestingly, it diffusion measurements (2D DOSY, 298.0 K) of 0.8–0.3 mM
appears that the filling of the basket’s cavity determines the solution of 2 in CDCl3 showed a set of signals with similar
mechanism by which each gated host opens and closes its diffusion coefficients (D = 5.4–7.4 × 10−10 m2 s−1, Fig. S3†); the
gates to control the encapsulation kinetics.
computed Stokes–Einstein radius RH = 5.4–7.5 Å of 2 bodes
7668 | Org. Biomol. Chem., 2013, 11, 7667–7675
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 2 (A) Template-directed synthesis of cup-shaped 4syn as reported previously.18 (B) The preparation of basket 2 was completed in ten synthetic steps.
well with the approximate diameter (12–15 Å, Fig. 1) of this
basket. In conclusion, host 2 is monomeric in CDCl3 with
pyridine gates forming intramolecular N–H⋯N contacts.
Despite a slightly greater distance between the “hinge” CH2
groups at the rim of 2 with respect to 1 (Δd = 0.5 Å, Fig. 1), the
pyridine gates are apparently close enough to make non-
covalent contacts: for both baskets, the N–H⋯N distances are
almost identical, 3.13–3.16 Å, while the corresponding bond
angles are somewhat different (148–154°, Fig. 1). Importantly,
a comparison of the computed structural parameters for 1 and
2 also shows a small contraction of the cup-shaped platform of
2 to perhaps assist the formation of the N–H⋯N contacts.
The inner space of baskets 1 and 2
Assessing the volume enclosed by the van der Waals surface of
a concave host is a challenging task,37,38 as one has to account
for the host’s corners/dimples as well as arbitrate on the mole-
Fig. 3 (A) Chemical structure of basket 2 and 1H NMR spectrum (400 MHz) of
cule’s boundary surface.10 A rapid computational method for
this compound (298.0 K) in CDCl3. (B) Pyridine gates in 2 adopt P and M stereo-
completing this job would certainly be useful to experimental-
chemical arrangements with rate coefficient krac characterizing the 2P/M inter-
ists as the encapsulation selectivity is a function of the popu-
lation of the host’s cavity.10 That is to say, stable van der Waals
encapsulation complexes are typically obtained when PC =
conversion. (C)
A
segment of simulated (bottom, WinDNMR)33,34 and
experimental (top) VT 1H NMR spectra of basket 2 in CD2Cl2 showing the
coalescence of an AB quartet, corresponding to Hd protons, into a singlet.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7667–7675 | 7669

View Article Online
Paper
Organic & Biomolecular Chemistry
0.55 0.09, which is in the range of the packing density of
common organic liquids (PC = 0.51–0.63).10
Table 1 The inner volume of various molecular capsules computed with the
3 V and cavity-filling methods
The optimal distance between neutral atoms forming a
stable noncovalent complex is, from a Lennard-Jones potential
curve, 1.12 times greater (re = 21/6σ; −dU/dr = Force = 0)39 than
the sum of their van der Waals radii.8 Given such an analysis,
envision a system of two atoms as spheres, each having 1 Å
radius and at the necessary equilibrium distance of re = 2.24 Å.
If the radius of the first sphere is kept constant (r1 = 1 Å) while
that of the second is increased to r2 = 1.24 Å, then we can
enclose the smaller into the bigger sphere to efficiently popu-
late the space. Next, we can calculate that the volume occupied
by the smaller sphere (V1/V2 = (r1/r2)3) amounts to 52%, which
is within the range of reported packing coefficient of liquids!10
To compute the inner space of baskets 1 and 2 (Fig. 1), we
used the recently developed 3 V (Voss Volume Voxelator) com-
putational protocol;17 this freeware has been recommended
for investigating drug-binding sites in biological molecules
and is readily available at http://3vee.molmovdb.org. In
essence, the method employs two variously sized rolling
probes elucidating the shape of the molecule of interest: the
bigger probe (typically r1 > 10 Å) must not populate the cavity
to establish the outer molecular surface, while the smaller
probe (depending on its size) delineates the accessible (r2 > 0)
or van der Waals (r2 = 0) surface of the molecule. The subtrac-
tion of two computed surfaces, comprising a 3D grid of 0.5 Å
voxels, gives the volume.
3 V method17 Cavity-filling
Variation
Capsule
(Å3)
method10 (Å3) (%)
Tennis ball I40
Tennis ball II41
Softball42
50
83
362
114
205
52
68
322
97
197
159
117
420
−4
18
11
15
4
1
8
Cryptophane A16
Calix[4]arene dimer38
Calix[4]arene carceplex43 161
Carceplex44
127
423
Cylindrical capsule45
1
Scheme 1
KS (M−1) for residing in the host are also more competitive.
That is, solvent molecules can reduce the experimentally
measured and apparent affinity Ka (eqn (1)) of a guest [G] occu-
pying the host in its free and solvated forms ([B]′ = ([B]free
[S⊂B]). The apparent affinity Ka is, therefore, given as:
+
½Gꢀ þ ½Bꢀ′ ¼ ½G,Bꢀ
The distance between two carbonyl oxygen atoms across the
aperture of energy-minimized 1 and 2 (MMFFs, Spartan) is
∼6.2–6.8 Å and the 10 Å (r1) virtual probe shall not penetrate
the cavity. The size of the smaller probe (r2 = 0.2 Å), however,
was chosen to obtain the volume of the smaller basket 1
(226 Å3) close to the value calculated with the cavity filling
method (221 9 Å3).27 Following this same protocol (with r1 =
10 Å and r2 = 0.2 Å), the cavity of the larger basket 2 was esti-
mated to be 318 Å3 (Fig. 1B). Importantly, varying the radius of
the bigger probe (r1 = 9–11 Å) had a small effect on the com-
puted volumes ( 2%). The variation of the smaller probe (r2 =
0.1–0.3 Å), however, had a greater effect on the volume vari-
ation ( 9%). To further assess the utility of this rapid methodo-
logy for determining the volume of artificial cavitands, we
examined the inner space of eight variously sized/shaped cap-
sular hosts (Table 1). Importantly, the computed 3 V volume of
these molecules is comparable to the values obtained from the
cavity-filling method (Table 1).10
Ka ¼ ½G,Bꢀ=ð½Gꢀ½Bꢀ′Þ
½Bꢀ′ ¼ ð½Bꢀfree þ ½S,BꢀÞ
From Scheme 1, one can derive:
K ¼ ½G,Bꢀ=ð½Gꢀ½Bꢀfree
ð1Þ
Þ
ð2Þ
ð3Þ
KS ¼ ½S,Bꢀ=ð½Sꢀ½BꢀfreeÞ
By dividing eqn (1) and (2), we obtain:
Ka=K ¼ ½Bꢀfree=½Bꢀ′
Ka=K ¼ ½Bꢀfree=ð½Bꢀfree þ ½S,BꢀÞ
ð4Þ
From eqn (3) into (4), however, it follows:
½S,Bꢀ ¼ KS½Sꢀ½Bꢀfree
Ka=K ¼ ½Bꢀfree=ð½Bꢀfree þ KS½Sꢀ½Bꢀfree
Ka=K ¼ 1=ð1 þ Ks½SꢀÞ
ð5Þ
On the encapsulation thermodynamics
Finally, the relationship between three stability constants16 is
obtained:
Numerous solvent molecules surround a concave host with a
tendency to occupy its interior.45 Thus, bulk solvent [S] could
act as guest14 and reside in the cavity of unoccupied molecular
capsules [B]free to give the [S⊂B] complex (KS, Scheme 1). The
greater the stability of the [S⊂B] complex (KS), the lower the
Ka ¼ K=ð1 þ Ks½SꢀÞ
ð6Þ
The concentration of organic solvents is typically in the
concentration of the desired host–guest complex [G⊂B] (K, range of [S] = 10–15 M−1, and if KS < 0.01 M−1, then eqn (6)
Scheme 1). In essence, two competing equilibria determine reduces to Ka ∼ K: the apparent affinity Ka of the basket for
the outcome of the recognition process and thereby the for- trapping the targeted guest is close to the intrinsic affinity K
mation of [G⊂B]. It follows that solvents with a greater affinity and is hardly affected by the solvent. However, when the
7670 | Org. Biomol. Chem., 2013, 11, 7667–7675
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Axial P/M chirality in gated baskets
Previously, we found that the formation of N–H⋯N hydrogen
bonds at the northern portion of basket 1 would limit the tor-
sional motion about each of its three C–NH(CvO) single
bonds (Fig. 2B).24 In accord with this conformational bias, we
measured NOE magnetization transfer (298.0 K, Fig. S1†)
between resonances corresponding to N–H and Ha protons in
the larger basket 2 (Fig. 2B). It follows that the pyridine gates
in 2 are also adopting right- (P) and/or left-handed (M) orien-
tations29 with a pair of diastereotopic Hd hydrogen atoms at
the “hinge” (Fig. 3B). Indeed, variable temperature 1H NMR
spectra (400 MHz, CDCl3) of 2 supported the existence of P/M
stereoisomers: a singlet corresponding to Hd protons at higher
temperatures would change into an AB quartet at lower
temperatures (Fig. 3C). The classical band-shape analysis
(WinDNMR)33,34 of the NMR spectroscopic data provided rate
coefficient krac (Fig. 3B) characterizing the rotation of the gates
and thereby the racemization of 2P/M in CDCl3. Importantly, we
completed a series of VT-NMR measurements for baskets 1
and 2 in four differently sized solvents: CD2Cl2, 61 Å3; CDCl3,
75 Å3; CFCl3, 81 Å3 and CCl4, 89 Å3. The activation parameters
ΔH‡P/M and ΔSP‡/M (Table 2), characterizing the racemization,
were extracted from the corresponding Eyring plots
(Fig. S12–19†).
Fig. 4 (A) The apparent binding affinity Ka (M−1) of (CH3)2CBr2 for occupying
basket 1 was measured (VT 1H NMR, Fig. S4–S7†) in four different solvents. (B)
The apparent binding affinity Ka (M−1) of C2Cl6 for occupying basket 2 was
quantified (VT 1H NMR, Fig. S8–S11†) in CDCl3, while it was below the detection
limit in CD2Cl2, CFCl3 and CCl4.
affinity of solvent molecules for occupying the basket is greater
(KS > 0.1 M−1), then the apparent binding Ka decreases and
eqn (6) gives Ka ≤ K/2.
1
We used VT H NMR spectroscopy to measure the apparent
affinity Ka of 2,2-dibromopropane (107 Å3, PC = 0.47) towards
basket 1 (226 Å3) in four differently sized solvents: CD2Cl2,
61 Å3; CDCl3, 75 Å3; CFCl3, 81 Å3 and CCl4, 89 Å3 (Fig. 4A).27
On the basis of size, these solvents were expected to fill the
basket’s interior (PC = 0.27–0.39, Table 2) with an increasingly
greater affinity Ks for occupying its cavity along the series.
Indeed, the stability of [(CH3)2CBr2⊂1] was found to be greater
in CD2Cl2 (ln Ka > 5.3, Fig. 4A) than in CDCl3/CFCl3 (ln Ka <
2.3, Fig. 4A) while practically non-measurable in CCl4.
The revolving of pyridine gates and the inner solvation of
baskets
From eqn (6), we can deduce the following trend: KsCD2Cl2
KCs DCl3/CFCl3 < KsCCl4. Since KaCD2Cl2 is ∼102–103 M−1 (K >
102 −1), it is reasonable to assume KsCCl4 to be >10 M−1 for
<
The revolving of pyridine gates is, in both baskets 1 and 2,
characterized with negative entropy of activation for all four
solvents (ΔS‡P/M = −6 to −23 e.u., Table 2). Thus, in the mech-
anism for the P/M conversion (vide infra), the seam of N–H⋯N
hydrogen bonds must loosen up and break in the rate-control-
ling step of the transformation (Fig. 3B). This, in turn, should
augment the framework’s flexibility contributing to its greater
disorder (ΔS‡1 > 0, Scheme 2). Concurrently, the same unfold-
ing of the basket’s framework contributes to the extension
of its surface area to prompt a more excessive solvation.
Additional solvent molecules ought to surround the partly
open capsular host and interact with it, to presumably increase
the order of the system (ΔS‡2 < 0, Scheme 2). To explain the
experimental result (Table 2), the adverse entropy of solvation
M
shutting off the encapsulation (KCa Cl4 < 1) in this environment.
The apparent affinity Ka of basket 2 (318 Å3) toward hexa-
chloroethane C2Cl6 (137 Å3, PC = 0.43) was, as in the case
above, measured with 1H NMR spectroscopy in four differently
sized solvents (Fig. 4B). Markedly, the binding affinity could be
quantified in CDCl3 while it is difficult to measure (too weak)
in CD2Cl2, CFCl3, and CCl4. For the more sizeable basket 2,
there is a smaller affinity for chloroform than the remaining
three solvents for occupying the cavity. The basket can there-
fore be characterized as solvophobic in CDCl3 (KsCDCl3
<
KCs D2Cl2/CFCl3/CCl4); we realize that for future recognition
studies, we need to survey other solvents.
Table 2 The interconversion of 1P/M and 2P/M stereoisomers is, in four differently sized solvents, characterized with ΔH‡P/M/ΔSP‡/M activation parameters
(Fig. S12–19). Packing coefficients were computed for one solvent molecule occupying each basket’s interior. Note that it is difficult to computationally (MMFFs) fit
more than one solvent molecule in the interior of each basket without breaking the seam of hydrogen bonds at the northern side
Basket 1 (226 Å3)
Basket 2 (318 Å3)
CD2Cl2
Parameters
CD2Cl2
CDCl3
9.90 0.14
CFCl3
CCl4
CDCl3
CFCl3
CCl4
ΔH‡P/M
8.98 0.19
7.30 0.10
7.97 0.15
6.40 0.12
7.75 0.13
4.60 0.10
6.82 0.07
(kcal mol−1
ΔS‡P/M (e.u.)
)
−6.61 0.85 −7.80 0.57 −18.45 0.42 −16.32 0.61 −16.82 0.56 −11.14 0.62 −23.49 0.50 −14.50 0.36
Volume (Å3)
Packing
61
0.27
75
0.33
81
0.36
89
0.39
61
0.19
75
0.24
81
0.25
89
0.28
coefficient (PC)
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7667–7675 | 7671

View Article Online
Paper
Organic & Biomolecular Chemistry
The mechanism of P/M racemization – computational study
For the racemization of the C3 symmetric basket 1, we con-
sidered two mechanistic alternatives RM1–2 and RM3
(Fig. 5).46–48 In the one/two-gate RM1–2 transformation
(Fig. 5A), the racemization begins with one pyridine gate
“breaking away” from the N–H⋯N hydrogen bonding in 1 to
form an intermediate state. The concomitant flip of the
remaining two gates completes the P/M conformational inter-
conversion; note that this particular mechanism is reminis-
cent of Mislow’s two-ring flip within molecular propellers.46
However, for the three-gate RM3 racemization (Fig. 5B), all
three gates undergo a synchronized revolving motion that
reverses the basket’s helicity in one elementary step.49 In our
Scheme 2 The entropy of activation ΔS‡P/M of basket 1 (A: CD2Cl2; B: CCl4) is
partitioned into more favourable ΔS‡1, pertaining to the flexibility of the mole-
cular framework, and adverse ΔS2‡, describing the solvation changes.
must be dominant in the course of the P/M change: ΔSP‡/M
ΔS‡1 + ΔS‡2 with |ΔS2‡| > |ΔS1‡|.
=
The entropy of activation, characterizing 1P/M interconver- prior work, we examined the departure of CBr4 from gated
sion, is “smaller” for CD2Cl2 (ΔS‡P/M = −6.61 0.85 e.u.) than baskets of type 1 using steered molecular dynamics (SMD).28
any other solvent (Table 2). We surmise that host 1 undergoes The cleavage of all three intramolecular N–H⋯N hydrogen
a less dramatic solvation change in CD2Cl2 (smaller |ΔS‡2| bonds would, in this particular case, take place with the depar-
term, Scheme 1A) as turning into the corresponding transition ture of CBr4 (108 Å3, PC = 0.48) from the basket. Given the
state, in the course of the P/M interconversion. The greater interdependence between the rotation of the gates and the in/
entropic change in CCl4 (ΔS‡P/M = ↕16.32 0.61 e.u., Table 2), out guest exchange,24,30,32 we deduced that the three-gate RM3
however, suggests a more significant solvent reorganization mechanism dominates in the case of the 1/CBr4 system.
(with a bigger |ΔS‡2| term, Scheme 1B) along the P/M tran-
In line with the above considerations, we completed a
sition. As discussed earlier, basket 1 is solvophobic in CD2Cl2 series of dihedral-driver computations of baskets 1 and 2 (PM6
and solvophilic in CCl4 (KCs Cl4 ≫ KsCD2Cl2, Fig. 4A). Accordingly, for geometries, followed by B3LYP/6-31+G** single-point ener-
CCl4 solvent molecules are more attracted to the interior of gies) containing both methyl (R = CH3, Table 3) and trifluoro-
1P/M to contribute to a more ordered ground state (Scheme 2B). methyl (R = CF3, Table 3) amide groups and undergoing two
This, in turn, renders a greater solvation change during the racemisation pathways, RM1–2 and RM3 (Fig. 5). Importantly,
racemization (Scheme 2B). In the case of 1P/M conversion for all of these calculations, the baskets were empty (i.e., no
in CD2Cl2, however, the ground state of 1 is poorly solvated guest), and calculations were performed in the gas phase. For
(i.e. more disordered) and a smaller change in its order is the RM1–2 process, we drove the rotation of one pyridine gate
necessary to reach the transition state.
by constraining its OvC–N–C dihedral angle, φ1. Specifically,
The activation entropy for 2P/M interconversion is most favour- the rotation of the pyridine gate from the inner to the outer
able in CDCl3 (ΔS‡P/M = −11.14 0.62 e.u., Table 2). It follows side of the basket was evaluated in 5° increments; upon each
that the basket should be more solvophobic in this medium change in φ1, the structure was subjected to energy minimi-
with a poorly solvated ground state. This is in agreement with zation and the overall energy was computed. To complete the
our encapsulation studies in which CDCl3 is indeed the least inversion of chirality, we drove the rotation of the remaining
competitive solvent with KCs DCl3 < KCs D2Cl2/CCl4/CFCl3 (Fig. 4B).
two gates by restricting their dihedral angle φ2 (Fig. 5A). The
Fig. 5 (A) One/two-gate racemization mechanism (RM1–2) of 1 was computed (B3LYP/6-31+G**//PM6) to proceed via rotation of one gate (φ1) followed by simul-
taneous rotation of the remaining two gates (φ2). (B) Three-gate racemization mechanism (RM3) of 1 was computed (B3LYP/6-31+G**//PM6) to proceed via simul-
taneous rotation of all three gates (φ2).
7672 | Org. Biomol. Chem., 2013, 11, 7667–7675
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 3 The activation energy difference (ΔE, kcal mol−1) between two racemi-
zation pathways, RM1–2 and RM3 (see text), was computed for the interconver-
sion of 1P/M and 2P/M stereoisomers at the PM6//PM6 and B3LYP/6-31+G**//
PM6 levels of theory
the opening/closing of the smaller basket 1 is fastest in CD2Cl2
(Fig. 6A) relative to the other three solvents (ΔG‡P/M < 11 kcal
mol−1, Fig. 6A). The nature of bulk solvent had, however, a less
dramatic effect on the dynamics of basket 2 (Fig. 6B). The
interconversion rate of 2P/M is similar in all four environments
but also akin to basket 1 in CD2Cl2 (Fig. 6A)! Another way to
present the data in Fig. 6A/B is by plotting the interdepen-
dence between enthalpy ΔH‡P/M and entropy ΔS‡P/M characteri-
zing the racemization (Fig. 6C).50 Evidently, there are two
isokinetic relationships51 with similar isokinetic temperatures
β = (ΔH‡P/M − ΔH‡o)/ΔSP‡/M = 240–260 K. In essence, the P/M race-
mization of (a) basket 1 in CDCl3, CFCl3 and CCl4 appears to
follow one mechanism while (b) basket 2 in all four solvents
another mechanism, which as well operates for 1 in CD2Cl2
(Fig. 6C).51,52
In line with this interpretation, we observe that the size of
the solvent molecules should matter in opening/closing of
gated baskets by imposing on the operation of gates rotating at
the rim. That is to say, the enthalpic component ΔH‡P/M of the
racemization energy ΔG‡P/M is greater for the P/M interconver-
sion of basket 1 in CHCl3, CFCl3 and CCl4 (∼2 kcal mol−1 per
solvent, Fig. 6C). Presumably, a considerable van der Waals
steric strain is created in the operation of 1 as the guest
(solvent) molecules press against the revolving pyridine gates.
In contrast, more spacious and gated host 2 creates a
sufficiently large aperture to permit in/out guest exchange and
racemization with a smaller enthalpy of activation ΔHP‡/M
(Fig. 6C).
PM6//PM6
B3LYP/6-31+G**//PM6
Basket Preferred
1R/2R
Preferred
mechanism ΔE (kcal mol−1
)
mechanism ΔE (kcal mol−1
)
1CH3
2CH3
1CF3
2CF3
RM1–2
RM1–2
RM1–2
RM3
4.3
1.3
0.1
0.7
RM1–2
RM1–2
RM3
0.2
5.6
4.7
9.8
RM3
potential energy landscape of the three-gate mechanism RM3
was examined by simultaneous variation of all three dihedral
angles φ2 (Fig. 5B). The computed difference of activation
energies (ΔE) for two mechanistic alternatives is summarized
in Table 3.
Energies from the PM6 optimizations predict that both
(empty) acetamide baskets (1CH3, 2CH3) would prefer the RM1–2
mechanism. The same preference for RM1–2 exists for single-
point energies at the B3LYP/6-31+G** level of theory, although
the stronger preference for RM1–2 is with 2CH3 instead of 1CH3
.
The PM6 optimizations of 1CF3 showed a slight preference
(∼0.1 kcal mol−1) for RM1–2, while those done with 2CF3
support the RM3 pathway. It is important to remember that
the experimental results with different solvents were obtained
for the CF3 baskets of 1 and 2. Single-point energy calculations
on 1CF3 and 2CF3 point to RM3 as the preferred mechanism
(Table 3). Importantly, these calculations were completed in
the gas phase, and without consideration of any solvent in the
cavity; indeed, the inclusion of a guest and solvation could
alter the behavior of the baskets. Given such considerations,
we conclude that our computational data suggest two compet-
ing pathways with comparable energies of activation.
As noted in an earlier study,30 the gates revolve at the
basket’s rim with and without in/out guest exchange; our
experimental measurements, however, do not reveal the pro-
portion of these two alternatives. In further discussion, we will
therefore contemplate on the P/M racemization following
RM1–2 or RM3 mechanisms with a guest (a) sitting inside
baskets or (b) departing from their inner space (four mechan-
istic pathways).
The mechanism of the P/M racemization – experimental study
In the case of the larger basket 2, CHCl3/CH2Cl2/CFCl3/CCl4
guests are too small to effectively populate its inner space
The free energy of activation ΔG‡P/M (ΔG‡P/M = ΔH‡P/M − TΔS‡P/M
,
Table 2) for the P/M racemization of baskets 1 and 2 is plotted (PC = 0.19–0.28) and affect the rotation of the pyridines at the
over the 180–340 K temperature range (Fig. 6A/B). Interestingly, rim. In this case, we presume that the preferred mechanism of
Fig. 6 (A) The activation energy ΔG‡P/M for the racemization of 1P/M in CD2Cl2 (blue), CDCl3 (red), CFCl3 (black, 30% CD2Cl2) and CCl4 (green, 20% CD2Cl2) as a func-
tion of temperature. (B) The activation energy ΔG‡P/M for the racemization of 2P/M in CD2Cl2 (blue), CDCl3 (red), CFCl3 (black, 30% CD2Cl2) and CCl4 (green, 20%
CD2Cl2) as a function of temperature; note that ΔG‡P/M was calculated as ΔHP‡/M − TΔSP‡/M using data from Table 2. (C) The enthalpy/entropy compensation relation-
ships (R2 = 0.999, SigmaPlot) corresponding to the racemization of 1P/M (red) and 2P/M (blue) in four different solvents.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7667–7675 | 7673

View Article Online
Paper
Organic & Biomolecular Chemistry
operation is RM1–2 since it is easier to break two hydrogen
bonds than three for opening the host; Table 3 shows that
RM1–2 dominates racemisations in the gas phase, yet our com-
putations did not consider the swapping of entrapped guest
with solvent molecule(s) and/or entropic changes in the
process. When dichloromethane enters/exits basket 1, a rather
insignificant population of the basket’s cavity (PC = 0.27) also
enforces the RM1–2 mechanistic pathway with only one
pyridine gate opening to permit the racemization with/without
the guest exchange.
Notes and references
1 D. J. Cram, Nature, 1992, 356, 29–36.
2 F. Hof and J. Rebek Jr., Proc. Natl. Acad. Sci. U. S. A., 2002,
99, 4775–4777.
3 K. N. Houk, K. Nakamura, C. Sheu and A. E. Keating,
Science, 1996, 273, 627–629.
4 L. C. Palmer and J. Rebek Jr., Org. Biomol. Chem., 2004, 2,
3051–3059.
5 S. Rieth, K. Hermann, B.-Y. Wang and J. D. Badjic, Chem.
Soc. Rev., 2011, 40, 1609–1622.
6 J. B. Wittenberg and L. Isaacs, Supramol. Chem. Mol. Nano-
mater., 2012, 1, 25–43.
7 R. G. Chapman, G. Olovsson, J. Trotter and J. C. Sherman,
J. Am. Chem. Soc., 1998, 120, 6252–6260.
8 A. Varnek, S. Helissen, G. Wipff and A. Collet, J. Comput.
Chem., 1998, 19, 820–832.
9 L. Garel, J. P. Dutasta and A. Collet, Angew. Chem., 1993,
105, 1249–1251 (Angew. Chem., Int. Ed. Engl., 1993,
1232(1248), 1169–1271).
10 S. Mecozzi and J. Rebek Jr., Chem.–Eur. J., 1998, 4, 1016–
1022.
For smaller basket 1 (226 Å3), guests CHCl3/CFCl3/CCl4
populate a greater portion of its cavity (PC = 0.33–0.39). The
guest exchange would necessitate enough room, thereby en-
forcing the rotation of all three gates28 and therefore the RM3
mechanism. The sole P/M rotation of the gates, without the
guest exchange, might perhaps follow the RM1–2 alternative
since it is energetically more favourable to break apart a single
gate (Table 3). On the basis of our measurements (Fig. 6C),
however, the guest exchange via the RM3 mechanism must be
dominating the racemisations with 1 carrying more sizeable
guests.28,30
11 J. Canceill, M. Cesario, A. Collet, J. Guilhem, L. Lacombe,
B. Lozach and C. Pascard, Angew. Chem., 1989, 101, 1249–
1251.
Conclusions
Two differently sized baskets employ functionalized pyridine 12 M. Rekharsky and Y. Inoue, Supramol. Chem. Mol. Nano-
gates at the rim to form a seam of hydrogen bonds and mater., 2012, 1, 117–133.
thereby adopt P and M helical orientations. The P/M racemiza- 13 C. A. Hunter, Angew. Chem., Int. Ed., 2004, 43, 5310–5324.
tion mechanism appears to be a function of the guest’s 14 K. T. Chapman and W. C. Still, J. Am. Chem. Soc., 1989, 111,
volume and could be elucidated by considering the population
of the basket’s cavity. Thus, larger guests that more efficiently 15 J. Canceill, L. Lacombe and A. Collet, J. Am. Chem. Soc.,
fill the host (PC > 0.30) enforce the geared RM3 pathway: the 1986, 108, 4230–4232.
3075–3077.
entrance/exit of guests necessitate the movement of all three 16 A. Collet, Compr. Supramol. Chem., 1996, 2, 325–365.
gates. When the guest occupies a small portion of the inner 17 N. R. Voss and M. Gerstein, Nucleic Acids Res., 2010, 38,
space of the gated host (PC < 0.27), however, the rotation of
W555–W562.
one gate creates a large enough aperture for facile in/out guest 18 Z. Yan, T. McCracken, S. Xia, V. Maslak, J. Gallucci,
exchange via the RM1–2 mechanism.
C. M. Hadad and J. D. Badjic, J. Org. Chem., 2008, 73, 355–
Understanding the action of dynamic hosts, akin to 1 and
363.
2, is important for creating functional systems resembling 19 A. Godec and F. Merzel, J. Am. Chem. Soc., 2012, 134,
those found in nature.53 Our study about subtle mechanistic
17574–17581.
variations of molecular gating could be useful for improving 20 B. C. Gibb, Chemosensors, 2011, 3–18.
catalysis54 and/or controlling the trafficking of molecules in 21 F. Hof, S. L. Craig, C. Nuckolls and J. Rebek Jr., Angew.
artificial environments.55
Chem., Int. Ed., 2002, 41, 1488–1508.
22 I. A. Sedov, M. A. Stolov and B. N. Solomonov, J. Phys. Org.
Chem., 2011, 24, 1088–1094.
23 N. Graulich, H. Hopf and P. R. Schreiner, Chem. Soc. Rev.,
2010, 39, 1503–1512.
Acknowledgements
This work was financially supported with funds obtained from 24 B.-Y. Wang, X. Bao, Z. Yan, V. Maslak, C. M. Hadad
the National Science Foundation (CHE-1012146, to J.D.B.) and
the Department of Defense, the Defense Threat Reduction
and J. D. Badjic, J. Am. Chem. Soc., 2008, 130, 15127–
15133.
Agency (HDTRA1-11-1-0042, to J.D.B. and C.M.H.). The content 25 D. A. Evans and J. S. Evans, J. Org. Chem., 1998, 63, 8027–
of the information does not necessarily reflect the position or 8030.
the policy of the federal government, and no official endorse- 26 S. Tartaggia, A. Scarso, P. Padovan, O. De Lucchi and
ment should be inferred. The Ohio Supercomputer Center is F. Fabris, Org. Lett., 2009, 11, 3926–3929.
gratefully acknowledged for providing generous computational 27 B.-Y. Wang, X. Bao, S. Stojanovic, C. M. Hadad and
resources.
J. D. Badjic, Org. Lett., 2008, 10, 5361–5364.
7674 | Org. Biomol. Chem., 2013, 11, 7667–7675
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
28 S. Rieth, X. Bao, B.-Y. Wang, C. M. Hadad and J. D. Badjic, 42 R. Meissner, X. Garcias, S. Mecozzi and J. Rebek Jr., J. Am.
J. Am. Chem. Soc., 2010, 132, 773–776. Chem. Soc., 1997, 119, 77–85.
29 B.-Y. Wang, S. Stojanovic, D. A. Turner, T. L. Young, 43 P. Timmerman, W. Verboom, F. C. J. M. van Veggel,
C. M. Hadad and J. D. Badjic, Chem.–Eur. J., 2013, 19,
4767–4775.
30 K. Hermann, S. Rieth, H. A. Taha, B.-Y. Wang, C. M. Hadad
and J. D. Badjic, Beilstein J. Org. Chem., 2012, 8, 90–99.
31 S. Rieth and J. D. Badjic, Chem.–Eur. J., 2011, 17, 2562–
2565.
J. P. M. van Duynhoven and D. N. Reinhoudt, Angew.
Chem., 1994, 106, 2437–2440 (Angew. Chem., Int. Ed. Engl.,
1994, 2433(2422), 2345–2438).
44 R. G. Chapman, N. Chopra, E. D. Cochien and
J. C. Sherman, J. Am. Chem. Soc., 1994, 116, 369–370.
45 A. Shivanyuk and J. Rebek, Chem. Commun., 2002, 2326–
2327.
32 B.-Y. Wang, S. Rieth and J. D. Badjic, J. Am. Chem. Soc.,
2009, 131, 7250–7252.
46 K. Mislow, Acc. Chem. Res., 1976, 9, 26–33.
33 N. Bampos and A. Vidal-Ferran, J. Chem. Educ., 2000, 77, 47 B. Laleu, G. Bernardinelli, R. Chauvin and J. Lacour, J. Org.
130–133. Chem., 2006, 71, 7412–7416.
34 H. J. Reich, J. Chem. Educ. Software, Ser. D, 1996, 3D, 48 H. Ito, T. Abe and K. Saigo, Angew. Chem., Int. Ed., 2011, 50,
17–19. 7144–7147.
35 J. Svoboda, H. Schmaderer and B. Koenig, Chem.–Eur. J., 49 S. Rieth, Z. Yan, S. Xia, M. Gardlik, A. Chow, G. Fraenkel,
2008, 14, 1854–1865.
36 R. L. Williams and S. W. Shalaby, J. Heterocycl. Chem., 1973,
10, 891–899.
C. M. Hadad and J. D. Badjic, J. Org. Chem., 2008, 73, 5100–
5109.
50 J. D. Dunitz, Chem. Biol., 1995, 2, 709–712.
37 C. N. Eid Jr. and D. J. Cram, J. Chem. Educ., 1993, 70, 349–351. 51 W. Linert, Chem. Soc. Rev., 1994, 23, 429–438.
38 B. C. Hamann, K. D. Shimizu and J. Rebek Jr., Angew. 52 Chemical Kinetics and Reaction Mechanisms, ed.
Chem., Int. Ed. Engl., 1996, 35, 1326–1329.
J. H. Espenson, 1995.
39 J. N. Israelachvili, Intermolecular and Surface Forces, 2011.
40 C. Valdes, L. M. Toledo, U. Spitz and J. Rebek Jr., Chem.–
Eur. J., 1996, 2, 989–991.
53 A. Gora, J. Brezovsky and J. Damborsky, Chem. Rev., 2013,
113, 5871–5923.
54 B.-Y. Wang, T. Zujovic, D. A. Turner, C. M. Hadad and
J. D. Badjic, J. Org. Chem., 2012, 77, 2675–2688.
41 R. Wyler, J. de Mendoza and J. Rebek Jr., Angew. Chem.,
1993, 105, 1820–1821 (Angew. Chem., Int. Ed. Engl., 1993, 55 H.-X. Zhou and J. A. McCammon, Trends Biochem. Sci.,
1832(1812), 1699–1701).
2010, 35, 179–185.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7667–7675 | 7675