Binding trimethyllysine and other cationic guests in water with a series of indole-derived hosts: Large differences in affinity from subtle changes in structure

Article abstract of DOI:10.1039/c2ob25882j

The binding of a series of indole-derived hosts to various ammonium cations in pure, buffered water is investigated using both solution phase ¹H NMR studies and computational modeling. These hosts can engage their targets via the cation-pi interaction, electrostatic attraction, and the hydrophobic effect. The hydrophobic effect is shown to be a dominant influence in the strength of the binding interactions, both in terms of the hydrophobicity of the host and of the guest. Our findings show that small changes that reduce the host hydrophobic surface area without reducing either the number of negative charges or amount of aromatic surface area are found to significantly decrease binding. Additionally, the position of solubilizing charges is also shown to influence the preferred host geometry and resulting binding constants.

Full text of DOI:10.1039/c2ob25882j

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 6885
www.rsc.org/obc
PAPER
Binding trimethyllysine and other cationic guests in water with a series of
indole-derived hosts: large differences in affinity from subtle changes in
structure†
Amanda L. Whiting and Fraser Hof*
Received 8th May 2012, Accepted 23rd June 2012
DOI: 10.1039/c2ob25882j
The binding of a series of indole-derived hosts to various ammonium cations in pure, buffered water is
investigated using both solution phase 1H NMR studies and computational modeling. These hosts can
engage their targets via the cation–pi interaction, electrostatic attraction, and the hydrophobic effect. The
hydrophobic effect is shown to be a dominant influence in the strength of the binding interactions, both in
terms of the hydrophobicity of the host and of the guest. Our findings show that small changes that reduce
the host hydrophobic surface area without reducing either the number of negative charges or amount of
aromatic surface area are found to significantly decrease binding. Additionally, the position of solubilizing
charges is also shown to influence the preferred host geometry and resulting binding constants.
functionality from one host to another (Fig. 1). All four hosts
provide the same amount of aromatic surface area and the same
Introduction
Molecular recognition arises from the contributions of many
kinds of weak, intermolecular forces. In pure water, additional
considerations such as solvation and the hydrophobic effect
become important. Synthetic host–guest systems that function in
pure water remain relatively underrepresented in the supramole-
cular literature.1–4 While nature has perfected the art of encoding
strong and selective binding in water, chemists still find it
difficult to generate simple synthetic receptors that can reproduce
this level of mastery. This is largely because the structure-func-
tion relations that have become relatively easy to predict in
organic solvents still present a challenge to the designer who
wishes to create a system that functions in pure water.
Host 1, decorated with three indole rings appended with car-
boxylates, is a mimic of tryptophan-rich protein binding pockets
that have evolved to engage quaternary and tertiary ammonium
ions such as acetylcholine (ACh) and trimethyllysine
(Kme3).5–13 Preliminary studies of host 1 show that it binds
organic ammonium cations in water, and that the hydrophobic
effect plays a dominant role that far exceeds the energetic
influence of, for example, cation–pi interactions, in determining
its binding affinities and selectivities.14 We sought to explore the
extent of this effect in a series of related hosts – each containing
three indole rings and three carboxylic acids attached to a central
benzene ring but with variation in the exact positioning of
number of negative charges to engage a cation.
In host 2, we chose to retain the 3-propionic carboxylate tails
on the indoles and append methoxy groups to the core benzene
ring. It was thought by us and others15 that the electron-donating
ability of the oxygen atoms could contribute to an increase in the
electron density on the core. This could potentially lead to an
increase in binding due to an enhanced cation–pi effect at the
central ring. Hosts 3 and 4 were obtained by shortening the pro-
pionate side chain of 1 to a simple carboxylate at the 3- and
2-position of the indoles, respectively, and were envisioned as
less hydrophobic versions of 1. The effect of removing the
–CH2CH2– linker could be determined directly by comparing 1
and 3, while altering the position of the carboxylate (host 4)
would allow us to probe a different type of geometric variation
within this family.
University of Victoria, Department of Chemistry, PO Box 3065 STN
CSC, Victoria, Canada, V8W3V6. E-mail: fhof@uvic.ca
†Electronic supplementary information (ESI) available. See DOI:
10.1039/c2ob25882j
Fig. 1 Host studied in this work (indole numbering guide for
reference).
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 6885–6892 | 6885

View Article Online
2. Host geometries in organic solution and water
Results and discussion
Hosts 1–4 are inherently flexible due to the nature of the single
bonds connecting the indole arms to the central benzene. As a
result, they are likely able to adopt a number of conformations
separated by low energetic barriers. This is especially true in
organic solutions like dimethyl sulfoxide (DMSO) where the
greasy aromatic elements of the hosts are well solvated. In water
however, the geometries adopted will depend on the inherent
bond rotational preferences of the host structures (as in DMSO)
as well as aromatic clustering driven by the hydrophobic
effect.17,18 The preferred geometry prior to binding will not
necessarily resemble an ideal open and bowl-like conformation.
We examined the aqueous host conformations without guests
by comparing the chemical shifts of each host in DMSO to that
in buffered D2O (1 mM solutions) to see which protons on the
hosts were most affected by the solvent change. The resulting
spectra of the aromatic region protons of host 1 are shown in
Fig. 2. Protons which were found to have ≥0.2 ppm change in
chemical shift are highlighted in Fig. 3 (see Fig. S1–S5 in ESI†
for additional full spectra and chemical shift data for all protons).
Protons that significantly shift upfield upon the move to D2O
indicate a closer association with aromatic surfaces.
1. Synthesis
The synthesis of host 1 has been previously reported.14 Host 2 is
prepared starting with 1,3,5-trimethoxybenzene 5, and treating
with paraformaldehyde and HBr to provide the 1,3,5-tris(bromo-
methyl)-2,4,6-trimethoxybenzene core
6
in 30% yield
(Scheme 1).16 This methoxy core is then reacted with three
equivalents of methyl indole-3-propionate 7 that has been
pre-treated with NaH in DMF to give 8 in 34% yield. Deprotec-
tion of the methyl esters in basic aqueous solution gives the final
host 2. Hosts 3 and 4 are similarly synthesized starting with the
protection of commercial 3- and 2-indole carboxylic acids to
give methyl esters 9 and 10. Coupling to 1,3,5-tris(bromo-
methyl)benzene 11 and deprotection with KOH in MeOH gives
hosts 3 and 4. All four hosts were isolated as carboxylic acids,
and converted to their tri-sodium salts by treatment with stoi-
chiometric NaOMe prior to their dissolution in buffered water
for use in NMR studies. Our studies included both a detailed
analysis of host conformations based on experimental and com-
putational data, and NMR-based determinations of association
constants for various ammonium ions of interest.
Scheme 1 Synthesis of hosts.
6886 | Org. Biomol. Chem., 2012, 10, 6885–6892
This journal is © The Royal Society of Chemistry 2012

View Article Online
Fig. 4 Equilibrium geometry in implicit water for hosts 1, 2, 3 and 4
(HF/6-31G* as implemented in Spartan ’10).19
conformation for 2 that is consistent with the (more limited)
experimental data available for this host.
Fig. 2 Aromatic protons for host 1 (7.8–6.2 ppm) showing chemical
shift upon solvent change in (a) DMSO and (b) H2O. Proton identities
as follows: 4-H (●), 7-H (○), 2-H (♦), 6-H (■), 5-H (□), and Ar–H (of
central benzene ring, ◊).
Host 4, with a 2-carboxy substituent, shows overall smaller
changes in chemical shift than the other hosts, but follows a
similar pattern to the shifts observed in hosts 1 and 3 again indi-
cating a closer association with aromatic surfaces. Interestingly,
the energy minimized structure for 4 in implicit water suggests
an open conformation, unlike the other hosts in this series. If
accurate, the open conformation of 4 (Fig. 4) would be expected
to have downfield chemical shifts for the 6-H and 7-H indole
protons arising from CH⋯O contacts with the carboxylates.
Experimentally, this was observed for proton 6-H on host
4 which at 7.08 ppm in D2O and 7.14 ppm in DMSO-d6, is
located furthest downfield of all hosts in both water and also
DMSO (see Fig. S1–S4† for NMR spectra).
Further NMR evidence that helps us to draw conclusions
about the conformational changes experienced by all four hosts
in DMSO and in buffered water comes from 1D ROESY experi-
ments carried out in each solvent. In DMSO, key ROEs indicat-
ing through-space interactions are noted between the N–CH2
methylene protons and indole H-7 for all hosts (“indole-out,”
Fig. 5a), and between N–CH2 and indole H-2 for hosts 1, 2 and
3 (“indole-in,” Fig. 5b; host 4 has no H-2 proton). These two
close contacts are mutually exclusive and occur in two different
conformations that differ only by rotation about the indole
N–CH2 single bond. We conclude that the ROE data in DMSO
are best explained by the rapid exchange between both confor-
mations. In buffered water, hosts 1, 2, and 3 continue to show
N–CH2 to H-2 contacts that arise from the “indole-in” confor-
mation shown in Fig. 5b, but show no evidence of contacts
between H-7 and N–CH2 that would arise from the “indole-out”
conformation shown in Fig. 5a. Of note is that the “indole-in”
rotamer, with H-7 closely packed over the central ring, is con-
sistent with the calculated closed, propeller-like structures for hosts
1–3 shown in Fig. 4. Host 4 maintains H-7 to N–CH2 contacts in
D2O, as well as showing H-7 – ArH contacts that are consistent
with the calculated open conformation for 4 shown in Fig. 4.
Overall, the chemical shift, computational, and ROE data
suggest geometries for hosts 1–3 in which the aromatic rings are
further compressed onto each other and the charged carboxylate
arms are mostly directed outwards into solution. This is con-
sistent with the expected influence of the hydrophobic effect on
these radially amphiphilic molecules. The collapsed form in
Fig. 3 Significant chemical shift changes from DMSO to D2O. Magni-
tude of shift is indicated (ppm).
Hosts 1 and 3 each have several protons that move upfield to
the same degrees, indicating a similar conformation is adopted
for these two hosts in D2O. Energy minimizations in implicit
water (HF/6-31G* as implemented in Spartan ‘10)19 suggest that
the three indole rings are collapsed into a closed, propeller-like
aromatic cluster, with each of the protons that are observed to be
upfield-shifted in the NMR data located in the shielding cone of
a neighboring aromatic ring (Fig. 4). The methoxy-containing
host 2 has only one proton that experiences significant
solvent-induced shifts (7-H indole) with all other shifts being
negligible. Energy minimizations again suggest a propeller-like
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 6885–6892 | 6887

View Article Online
water also allows for stronger edge-to-face aromatic interactions
between the rings than would be present in DMSO, where the
hosts maintain significant mobility.
introduced. This suggested an opening of the aromatic rings of
the indoles upon binding. The greatest downfield change in
chemical shift (Δδ) was observed for the protons closest to the
benzene core: Ar–H (for hosts 1, 3 and 4) and Ar–OCH3 (for
host 2). These shifts are consistent with a cationic guest posi-
tioned directly above the central ring, interacting both with it and
the now face-on indole rings in an ideal bowl or “cage-like”
conformation.
3. Solution-phase binding studies
The binding properties of hosts 1–4 were probed by NMR
studies in phosphate-buffered D2O (Table 1). Dilution titrations
conducted on each host alone gave concentration-dependent
chemical shifts that were fit to determine a self-association con-
stant for each host (Kdimerization). Subsequent titrations with
various ammonium ions provided data that could be fit to a
1 : 1 host–guest binding isotherm to determine Kassoc while
taking into account the effects of host homodimerization using
HypNMR20 (see ESI† for experimental and calculation details).
Similar trends in chemical shifts upon binding were observed
across the hosts irrespective of guest. Host signals identified in
Fig. 3 as being most shielded (upfield) in water (Ar–H core,
N–CH2, 7-H indole and 2-H indole) were generally found to
have downfield chemical shifts when a cationic guest was
3.1 Comparison of 1 vs. 2 – effect of methoxy substituents.
Host 1 exhibits a significant range in its association constants.
+
For those guests containing a trimethyllysine-like R-NMe3
group (entries 2–6) there is variation ranging from 40 M−1 to
250 M−1. The simplest binding of a quaternary ammonium ion
would be tetramethylammonium (entry 2); this could be con-
sidered a baseline of interaction with these types of hosts. For
guests containing a single, simple alkyl substitution (nBuNMe3I,
entry 4), binding to host 1 in water becomes stronger as the
length of the non-polar alkyl chain increases (from one carbon
+
+
in NMe4 to four carbons in nBuNMe3 ). Given that the overall
charge between these guests has not changed, this reflects an
increased contribution from the hydrophobic effect. When the
single alkyl substitution contains heteroatoms, such as the ester
in AChCl (entry 5) and the amino acid head group in Kme3
(entry 6), the binding affinity to host 1 is again increased. While
it is likely that there is an increased hydrophobic force associated
with these larger guests, we cannot rule out additional polar
interactions between host and guest. Binding to Kme3, for
example, is significantly stronger than the other guests of this
type, leading us to suggest additional favourable electrostatic
interactions between the ammonium head group of Kme3 and
the host. Interestingly, the aromatic quaternary ammonium
Fig. 5 Key ROESY interactions arising from the (a) “indole-out”
rotamer (H-7 to CH2 contact) and (b) “indole-in” rotamer (H-2 to CH2
contact). The “closed” indole rotamer in 5b is the same as that adopted
by all three indoles of hosts 1–3, as shown in Fig. 4.
+
BnNMe3 (entry 3), does not result in a significant increase in
+
the binding constant compared to NMe4 .
To examine the binding of larger quaternary ammonium ions,
we increased the length of all four alkyl substituents across the
series from methyl to n-butyl (entries 7–9). We observed a sig-
nificant increase in binding, this time exclusively due to an
Table 1 Binding affinities for host 1–4 in phosphate-buffered D2Oa
assoc (M−1
)
+
K
increased hydrophobic contribution. The largest guest, NBu4 ,
had the strongest effect on the host propionate chains, shifting
the methylene protons 0.1 ppm downfield and indicating the
strongest involvement of the alkyl chain in binding. These small
shifts are most consistent with the small deshielding influence of
a cationic alkylammonium binding partner. Binding to primary,
secondary and tertiary ammonium ions, (entries 10–12) showed
weaker binding constants that diminished as methyl groups were
removed. This was attributed to increased competition with the
aqueous environment that could better solvate these guests as the
number of hydrogen bonding NH’s increases.
The addition of methoxy substituents to the benzene core in
host 2 was intended to increase the electron-density available in
the pi system and strengthen the resulting cation–pi interaction.
However, the methoxy substituents were found to have signifi-
cantly diminished binding relative to host 1, especially with
hydrophobic guests. A comparison of dimerization constants
revealed 2 to be an order of magnitude less than 1 (entry 1). Pre-
sumably, the methoxy groups contribute to steric gearing22 and
can bias the indole arms of 2 to favor the closed “propeller” con-
formation (Fig. 4) regardless of solvation effects or the presence
Entry
1
2
3
4
1
2
3
4
5
6
7
8
Kdimerization
NMe4Cl
BnNMe3Cl
nBuNMe3Ib
AChCl
Kme3Cl
NEt4Cl
NPr4Cl
NBu4Cl
HNMe3Cl
H2NMe2Cl
H3NMeCl
330 50
40 15 27 12
47 44
100 20 35 10
44
22 19
180 10 55 15
1100 210 60 20
22
6
22
32 14
41
40 10
30
55 27
80 85
70 30
2
<1
1
1
0
0
1
1
1
1
22
55
22
24
24
28
28
23
80
3
1
2
120
250
8
9
1
6
9
7060 2100 90 10 145 55
10
11
12
96 10
50 20
16 27
8
8
20 10
26 33 50 20
n.b. n.b.
9
n.b.c
n.b.
a Phosphate buffered D2O (50 mM Na2HPO4/NaH2PO4) at pH 7.0 (pD
7.4). Kdimerization is defined as the association constant (βHH) for host
dimerization while Kassoc is defined as βHG, the association constant for
the 1 : 1 host–guest complex. In all cases, Kdimerization was first
determined via dilution titrations of host alone, and then used as a
constant in the fitting of host–guest titration data to determine the true
value of Kassoc. See ESI for details.† b The iodide salt of this guest was
used. The identity of counter ion for NMR studies in water has been
shown to be negligible.21 c n.b. = no binding.
6888 | Org. Biomol. Chem., 2012, 10, 6885–6892
This journal is © The Royal Society of Chemistry 2012

View Article Online
+
of guests. This is consistent with the observation that host 2 has
the fewest protons changing chemical shift upon moving from
DMSO to water – its conformation in both solvents is fixed into
the propeller shape. This also decreases the ability of 2 to associ-
ate with larger, hydrophobic guests (entries 7–9) that require an
open and flexible host conformation. Similar diminished
affinities (relative to 1) are seen with almost all of the guests
studied. It is clear that the added methoxy groups do not favor a
conformation suitable for guests binding.
constants are approximately the same as for NMe4 , the baseline
guest molecule. This differs from host 3, which has a small
dimerization constant and a weak but measurable preference for
hydrophobic guests. For host 4, binding with BnNMe3 stands
out as having the highest binding constant for guests with the
+
R-NMe3+ motif.
In looking at Fig. 4, carboxylates at the 2-positions of host 4
have the effect of drastically changing the geometric preferences
of the indole arms compared to host 3. Host 4 also has the smal-
lest change in proton chemical shifts of all the similar, unrest-
ricted hosts (1, 3 and 4) indicating that it likely adopts a
geometry that is highly similar in aqueous and organic solution.
By moving the polar carboxylate groups to the indole-2 position,
we have effectively excluded them from the binding pocket
making them less able to participate in any electrostatic inter-
action with an incoming cation. For smaller R-NMe3+ guests that
require little adjustment of the binding pocket, this translates into
slightly smaller binding constants than for host 3 that could use
its carboxylates in a favourable manner on the upper rim of the
binding pocket. As we branch into larger quaternary ammonium
cations, there is little change in binding as now host 4 must
disrupt its preferred conformation and adjust to accommodate a
larger guest. So while there is an increase in the hydrophobic
association between host and the larger guests, the entropic
penalty associated with opening the host binding pockets could
negate any gain in binding affinity.
3.2 Comparison of 1 vs. 3 – effect of methylenes in propio-
nate chains. Host 3, with the water-solubilizing carboxylates
attached directly to the indole 3-position, was studied in order to
observe the effect that the –CH2CH2– units in the indole arms of
1 would have on the strength of its binding interactions. Guest
binding to host 3 in all cases is diminished relative to 1
(Table 1), and the effects are largest for the most hydrophobic
guests. This is consistent with the assertion that the hydrophobic
effect plays a dominant role in driving cation binding to 1.
Interestingly, host 3 binds all quaternary R-NMe3+ guests with
approximately the same affinity (entries 2–6), regardless of their
degree of hydrophobic character. Given the similarity in struc-
ture, the aromatic skeletons of hosts 1 and 3 should have very
similar conformational preferences and abilities to bind guests
through their identical aromatic rings. Thus, for guests that do
not cause a large hydrophobic effect, the magnitude of binding
should be similar. This is clearly evident for guests with little
hydrophobic contribution such as NMe4 (entry 2; 40 M−1 vs.
+
32 M−1, for 1 and 3, respectively). For greasier guests where the
hydrophobic effect is likely to be larger, the disparity in binding
between 1 and 3 increases to a 50-fold difference for the largest
4. Calculated binding geometries
Energy minimizations in water were also used to evaluate the
geometry of each host upon guest binding. For the typical guest
NMe4 , the collapsed conformations of the empty hosts in each
+
cation, NBu4 (entry 9; 7060 M−1 vs. 145 M−1). The difference
+
in surface area per –CH2CH2– group between 1 and 3 is calcu-
lated to be 41 Å2. Assuming that all surface area of the CH2CH2
in 1 is buried upon complexation (which must be an over-esti-
mate of the surface area actually involved), this would lead to a
prediction of 10 kcal mol−1 difference in binding energy
(3.3 kcal mol−1 × 3 –CH2CH2–) between the two,23 which is
case convert to a more open conformation capable of making
favourable contacts with guests by rotating the indole arms about
the N–C single bond (Fig. 6). The process of creating a pocket
for the guest changes the chemical environment experienced by
the host protons as shown in the solution phase data. Protons
which were most shielded in the collapsed form are now
deshielded as they are moved out of shielding regions. In
addition, the aromatic rings are now involved in interacting with
a cation. While this would shield any protons attached to the
cation, the aromatic rings themselves (and thus their protons)
experience some deshielding (or a return to the more normal
deshielded state).
+
larger than that observed for NBu4 (2.3 kcal mol−1). The small
+
downfield shifts of –CH2CH2– protons upon binding NBu4 are
consistent with the formation of close contacts with the large cat-
ionic guest. Thus, although the deletion of the CH2CH2 groups
seems intuitively to be a modest change of a peripheral group,
the observed large changes in affinities are best interpreted as
arising from the participation of those arms as important hydro-
phobic binding elements.
The models of hosts 1 and 2 indicate that the methylene
region of the propionate arms come into contact with cationic
R-NMe3 region of the guests. While the resulting change in
+
3.3 Comparison of 3 vs. 4 – effect of carboxylate position.
We were also interested in what effect geometry would have on
tripodal hosts of this nature. In all cases, the three indole rings
need to adopt an open conformation to allow a cation to interact
in a “face-on” manner with all four aromatic surfaces. Hosts 3
and 4 are structural isomers; they have the same number and
identity of atoms and therefore the same degree of hydrophobi-
city. Their difference lies only in the position of the carboxylate
and thus, the preferred geometry of the molecule. Experimen-
tally, host 4 has no tendency to self-associate in water and binds
all quaternary cations with approximately the same value regard-
less of size, shape or hydrophobic character. All binding
NMR chemical shift might not be great for these protons, this
Fig. 6 Equilibrium geometry in implicit water for 1, 2, 3 and 4 with
NMe4+ (HF/6-31G* as implemented in Spartan ’10).19
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 6885–6892 | 6889

View Article Online
would be an additional weak interaction that would give
these hosts an advantage over the others. Indeed, the largest
observed shift for these protons was only 0.1 ppm downfield
Experimental section
General considerations
+
and only for host 1 with NBu4 , the most hydrophobic guest.
Proton (1H) NMR and carbon (13C) NMR spectra were recorded
on a Brüker AC300 (300 MHz) spectrometer, a Brüker
AVANCE360 (360 MHz) and a Brüker AVANCE500 (500 MHz).
Proton (1H) NMR spectra for NMR titration studies were
recorded on a Brüker AVANCE360 (360 MHz) spectrometer.
Chemical shifts (δ) are given in parts per million (ppm) and
referenced to residual protonated solvent (CHCl3: δH 7.26 ppm,
δC 77.16 ppm; (CH3)2SO: δH 2.50 ppm, δC 39.52 ppm).
J values are given in Hz. Abbreviations used are s (singlet), d
(doublet), t (triplet), q (quartet), and m (multiplet). Infrared
spectra were measured on a Thermo-Nicolet Nexus 670 FT-IR
spectrometer with a resolution of 2 cm−1 using a Pike MIRacle
attenuated total reflection (ATR) sampling accessory. Melting
points were obtained using a Gallenkamp Melting Point Appar-
atus and are uncorrected. Accurate mass determination
(HR-ESI-MS) was performed at the UVIC Genome BC Proteo-
mics Center on a Thermo Scientific LTQ Velos Orbitrap. The
synthesis of compounds 1 and 7 was performed as previously
reported.14 All final tri-acid compounds (1, 2, 3 and 4) were con-
verted to their tri-sodium salts by reaction with stoichiometric
NaOMe in MeOH before their dissolution in buffered water for
use in NMR titration studies.
Still, the increased hydrophobic surface area was shown to play a
significant role in the overall binding strength of these hosts
given the differences in binding affinity, especially between
hosts 1 and 3. As well, these models show that the carboxylate
groups of these hosts are associated with the upper rim of the
host pocket, in position to interact favourably with cationic
guests.
The models of hosts 3 and 4 demonstrate the physical differ-
ence of shifting the carboxylate from position 3 to 2. The car-
boxylates in 3 are more exposed, similar to those in hosts 1 and
2 and likely able to weakly participate in electrostatic inter-
actions. In host 4, those polar regions are buried in the base of
the host cavity and less able to make contact with a cationic
guest.
Conclusions
Within this series of highly similar indole-based hosts, two
factors stand out as having the biggest influences on the strength
of the host–guest interaction: (1) the amount of hydrophobic
surface area available for binding (particularly for hydrophobic
guests) and (2) the effect of small changes in functional group
position on both geometric preference and binding affinity. The
complementary and beneficial result of having additional hydro-
phobic regions on both host and guest is understandable as a
consequence of a strong hydrophobic contribution within the
aqueous environment. The more subtle effect of a small change
in polar group position leading to a weaker electrostatic inter-
action was less expected and demonstrates our lack of ability to
predict the outcome of such changes.
Though these hosts exhibit weak-to-moderately strong attrac-
tion to cations in water, there are important lessons here for
future host designs targeting biologically relevant cations in
water. Though we normally think of trying to build hosts that
take advantage of as many attractive interactions as possible
(hydrogen bonding and cation–pi interactions, for example), the
complications that can be introduced by carrying out studies in
pure water should not be underestimated. The current
examples show that, in pure water, even subtle changes in
pendant solubilizing groups can have dramatic effects on host–
guest affinities. While past studies have been concerned with the
possible role of, for example, an anionic water-solubilizing
group vs. a neutral water-solubilizing group,24 the current results
show that even apparently innocent changes in a small number
of linker CH2 groups can have a strong and determinant role in
guest binding. Related structural elements in proteins, such as
the methylene stretches of lysine, arginine, and glutamate side
chains, are increasingly appreciated to be binding elements that
play a strong role in determining the structures and interactions
of folded proteins.25–29 Designs for aqueous host–guest systems
will need to continue to consider the hydrophobic aspects of
small functional group changes as a key driver of their molecular
recognition performance.
1,3,5-Tris(bromomethyl)-2,4,6-trimethoxybenzene (6). Syn-
thesized by reported procedure for exact compound.16 1H NMR
(CDCl3, 300 MHz): δ 4.14 (s, 9H), 4.60 (s, 6H). 13C NMR
(CDCl3, 75 MHz): δ 22.7, 62.9, 123.5, 160.3.
Compound 8. 60% NaH (110 mg, 2.7 mmol) and indole ester
7 (550 mg, 2.7 mmol) were suspended in anhydrous DMF
(5 mL) and stirred under N2. After 40 minutes, 1,3,5-tris(bromo-
methyl)-2,4,6-trimethoxybenzene 6 (200 mg, 0.46 mmol) in
DMF (2 mL) was added drop-wise and the reaction left to stir at
room temperature for 17 hours. After complete reaction of com-
pound 6, most solvent was removed via vacuum and the reac-
tion quenched with hexanes and THF. After salt removal via
filtration, column chromatography (15 to 50% EtOAc/hexanes)
yielded product 8 (127 mg, 34%) as a tan oil. IR (neat),
ν (cm−1): 3728, 3698, 3625, 3013, 2951, 2369, 2349, 2327,
1
1737, 1728, 1673, 1585 1467, 1193, 1107, 1093, 747. H NMR
(CDCl3, 300 MHz): δ 2.73 (t, J = 7.7, 6H), 3.13 (t, J = 7.7, 6H),
3.45 (s, 9H), 3.68 (s, 9H), 5.28 (s, 6H), 6.97 (s, 3H), 7.16 (dt,
J = 3.5, J = 0.6, 3H), 7.29 (dt, J = 4.0, J = 0.3, 3H), 7.61 (t, J =
8.9, 6H). 13C NMR (CDCl3, 75 MHz): δ 20.7, 35.1, 39.4, 51.5,
62.9, 109.8, 113.8, 118.9, 119.1, 121.4, 121.8, 125.1, 127.7,
136.6, 160.2, 173.7. HR-ESI-MS: 836.3519 ([M + Na]+;
C48H51N3O9Na; calc: 836.3523).
Compound 2. Tri-substituted methyl ester indole 8 (127 mg,
0.16 mmol) was dissolved in THF (4 mL) and distilled H2O
(2 mL). Excess NaOH(s) (91.5 mg) was added and the reaction
was stirred at room temperature for 18 hours under N2. Reaction
was diluted with an equal volume of 1 M HCl(aq), extracted with
EtOAc (3 × 30 mL), and dried over MgSO4 before concentrating
under vacuum. Precipitation from a minimum of DCM with
hexanes and sonication gave triacid 2 (90 mg, 75%) as an
6890 | Org. Biomol. Chem., 2012, 10, 6885–6892
This journal is © The Royal Society of Chemistry 2012

View Article Online
off-white solid. Mp: 117–120 °C (dec.) IR (neat), ν (cm−1):
3024, 2957, 1708, 1581, 1463, 1249, 740. H NMR (CDCl3,
drop-wise and the reaction left to stir for 18 hours. Most solvent
was removed via vacuum and the reaction quenched with THF.
After salt removal via filtration, column chromatography (50 to
80% DCM/hexanes) yielded 13 (172 mg, 27% yield) as a beige
solid. Mp: 174–176 °C (dec.) IR (neat), ν (cm−1): 3050, 2948,
1712, 1709, 1249, 1197, 742. 1H NMR (CDCl3, 300 MHz):
δ 3.65 (s, 9H), 5.52 (s, 6H), 6.44 (s, 3H), 6.96–7.14 (m, 9H),
7.17 (s, 3H), 7.57 (dd, J = 7.0, J = 1.1, 3H). 13C NMR (CDCl3,
75 MHz): δ 47.5, 51.5, 110.7, 111.2, 120.7, 122.6, 123.4, 125.2,
126.0, 127.1, 138.9, 139.3, 162.1. HR-ESI-MS: 662.2263
([M + Na]+; C39H33N3O6Na; calc: 662.2267).
1
300 MHz): δ 2.71 (t, J = 6.5, 6H), 3.08 (t, J = 6.5, 6H), 3.20 (s,
9H), 5.24 (s, 6H), 6.90 (s, 3H), 7.11 (t, J = 7.2, 3H), 7.24 (t, J =
7.2, 3H), 7.48 (d, J = 8.2, 3H), 7.55 (d, J = 7.8, 3H). 13C NMR
(CDCl3, 125 MHz): δ 20.4, 34.7, 39.7, 62.7, 109.8, 113.4,
118.8, 119.3, 120.9, 122.0, 124.8, 127.8, 136.8, 160.8, 179.3.
HR-ESI-MS: 774.3051 ([M + Na]+; C45H45N3O9Na; calc:
774.3054).
Methyl indole-3-carboxylate (9). Indole-3-carboxylic acid
(2.00 g, 12.5 mmol) was dissolved in MeOH (23 mL) and con-
centrated H2SO4 (0.3 mL, 5.6 mmol) and heated to reflux for
3.5 hours. The reaction was cooled, poured into 75 mL ice, and
extracted with DCM (3 × 60 mL). The combined organics were
washed with saturated brine and saturated NaHCO3(aq), dried
over MgSO4, filtered and condensed. Column chromatography
(20 to 25% ethyl acetate/hexanes) gave indole 9 (1.25 g, 57%)
as a tan powder. Spectra matched known reference.30 1H NMR
(DMSO, 500 MHz): δ 3.80 (s, 3H), 7.19 (m, 2H), 7.47 (m, 1H),
7.99 (m, 1H), 8.07 (d, J = 3.0, 1H), 11.92 (br s, 1H). 13C NMR
(DMSO, 125 MHz): δ 50.6, 106.3, 112.3, 120.4, 121.2, 122.4,
125.6, 132.4, 136.4, 164.8.
Compound 3. Tri-substituted methyl ester indole 12 (122 mg,
0.19 mmol) was suspended in MeOH (12 mL). KOH(s) (221 mg)
was dissolved in distilled H2O (8 mL) and added drop-wise to
the suspension. The reaction was heated to reflux for 16 hours
until complete conversion observed by TLC (1 : 1 EtOAc/
hexanes). Reaction was diluted with 1 M HCl(aq), extracted with
EtOAc (3 × 30 mL), and dried over MgSO4 before concentrating
under vacuum. Solid was suspended in CHCl3, filtered and air-
dried to give triacid 3 (106 mg, 93% yield) as a tan solid. Mp:
261–263 °C (dec.) IR (neat), ν (cm−1): 3055, 2938, 1665, 1659,
1536, 1531, 1278, 1253, 1190, 751. 1H NMR (DMSO,
360 MHz): δ 5.38 (s, 6H), 7.08 (t, 3H, J = 7.6), 7.16 (t, 3H, J =
7.4), 7.23 (s, 3H), 7.32 (d, 3H, J = 8.1), 8.00 (d, 3H, J = 7.9),
8.16 (s, 3H), 12.01 (br s, 3H). 13C NMR (DMSO, 90 MHz):
δ 49.4, 106.8, 111.0, 120.8, 121.3, 122.2, 126.4, 126.6, 135.4,
136.1, 138.1, 165.6. HR-ESI-MS: 620.1797 ([M + Na]+;
C36H27N3O6Na; calc: 620.1797).
Methyl indole-2-carboxylate (10). Indole-2-carboxylic acid
(3.00 g, 18.7 mmol) was dissolved in MeOH (30 mL) and con-
centrated H2SO4 (0.3 mL, 5.6 mmol) and heated to reflux for
15 hours. The reaction was cooled, diluted with saturated
NaHCO3(aq) (50 mL), and extracted with EtOAc (2 × 40 mL).
The combined organics were dried over Na2SO4, filtered and
condensed. The crude solid was suspended in hexanes, soni-
cated, and then filtered to give indole 10 (2.80 g, 86% yield) as a
tan powder. Spectra matched known reference.30 1H NMR
(CDCl3, 300 MHz): δ 3.96 (s, 3H), 7.16 (t, J = 7.5, 1H), 7.23 (s,
1H), 7.33 (t, J = 7.6, 1H), 7.43 (d, J = 8.3, 1H), 7.70 (d, J = 8.0,
1H), 8.98 (br s, 1H). 13C NMR (CDCl3, 75 MHz): δ 52.2, 109.0,
112.1, 121.1, 122.9, 125.7, 127.3, 127.7, 137.1, 162.7.
Compound 4. Tri-substituted methyl ester indole 13 (26 mg,
0.04 mmol) was suspended in MeOH (5 mL). KOH(s) (108 mg,
1.9 mmol) was dissolved in distilled H2O (4 mL) and added
drop-wise to the suspension. The reaction was heated to reflux
for 17 hours until complete conversion observed by TLC
(DCM). Reaction was diluted with 1 M HCl(aq), extracted with
EtOAc (3 × 30 mL), and dried over MgSO4 before concentrating
under vacuum. Solid was suspended in CHCl3, filtered and air-
dried to give triacid 4 (23 mg, 93% yield) as a white solid.
Mp: 262–264 °C (dec.) IR (neat), ν (cm−1): 3035, 1721, 1687,
1678, 1673, 1519, 1267, 1200, 1171, 1136, 744. 1H NMR
(DMSO, 500 MHz): δ 5.65 (s, 6H), 6.64 (s, 3H), 7.10 (td, J =
3.9, J = 0.87, 3H), 7.15 (td, J = 3.5, J = 1.1, 3H), 7.19–7.23 (m,
6H), 7.65 (d, J = 7.9, 3H), 12.81 (br s, 3H). 13C NMR (DMSO,
125 MHz): δ 46.8, 110.4, 111.0, 120.5, 122.3, 123.6, 124.7,
125.5, 127.7, 138.7, 139.0, 162.8. HR-ESI-MS: 620.1795
([M + Na]+; C36H27N3O6Na; calc: 620.1797).
Compound 12. 60% NaH (245 mg, 6.1 mmol) and indole
ester 9 (1.03 g, 5.9 mmol) were suspended in anhydrous DMF
(8 mL) and stirred under N2. After 25 minutes, 1,3,5-tris(bromo-
methyl)benzene 11 (359 mg, 1.0 mmol) in DMF (2 mL) was
added drop-wise and the reaction left to stir at room temperature
for 18 hours. Most solvent was removed via vacuum and the
reaction quenched with hexanes and THF. After salt removal via
filtration, column chromatography (15 to 50% EtOAc/hexanes)
yielded 12 (320 mg, 50% yield) as a light yellow solid. Mp:
165–167 °C (dec.) IR (neat), ν (cm−1): 2925, 1703, 1692, 1530,
1536, 1243, 1180, 1092, 757, 748. 1H NMR (CDCl3,
300 MHz): δ 3.93 (s, 9H), 5.18 (s, 6H), 6.74 (s, 3H), 7.05 (d,
J = 8.2, 3H), 7.18 (td, J = 4.1, J = 1.1, 3H), 7.27 (td, J = 7.5, J =
0.9, 3H), 7.73 (s, 3H), 8.18 (d, J = 7.9, 3H). 13C NMR (CDCl3,
75 MHz): δ 50.2, 51.0, 108.0, 110.0, 121.9, 122.2, 123.2, 124.9,
126.7, 134.2, 136.4, 138.1, 165.2. HR-ESI-MS: 662.2264
([M + Na]+; C39H33N3O6Na; calc: 662.2267).
Acknowledgements
This work was funded by the Michael Smith Foundation for
Health Research (MSFHR) and NSERC. FH is a MSFHR Career
Scholar and Canada Research Chair. ALW is a MSFHR Junior
Graduate Research Trainee.
Compound 13. 60% NaH (245 mg, 6.2 mmol) and indole 10
(1.05 g, 6.0 mmol) were suspended in anhydrous DMF (8 mL)
and stirred under N2. After 30 minutes, 1,3,5-tris(bromomethyl)-
benzene 11 (356 mg, 1.0 mmol) in DMF (2 mL) was added
Notes and references
1 F. Diederich and K. Dick, J. Am. Chem. Soc., 1984, 106, 8024–8036.
2 C. S. Beshara, C. E. Jones, K. D. Daze, B. J. Lilgert and F. Hof, Chem-
BioChem, 2010, 11, 63–66.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 6885–6892 | 6891

View Article Online
3 Y. Ma, X. Ji, F. Xiang, X. Chi, C. Han, J. He, Z. Abliz, W. Chen and
F. Huang, Chem. Commun., 2011, 47, 12340–12342.
4 L. A. Ingerman, M. E. Cuellar and M. L. Waters, Chem Commun, 2010,
46, 1839–1841.
17 P. Martín-Gago, M. Gomez-Caminals, R. Ramón, X. Verdaguer,
P. Martin-Malpartida, E. Aragón, J. Fernández-Carneado, B. Ponsati,
P. López-Ruiz, M. A. Cortes, B. Colás, M. J. Macias and A. Riera,
Angew. Chem., Int. Ed., 2011, 1820–1825.
5 D. L. Beene, G. S. Brandt, W. Zhong, N. M. Zacharias, H. A. Lester and
D. A. Dougherty, Biochemistry, 2002, 41, 10262–10269.
6 J. D. Schmitt, C. G. V. Sharples and W. S. Caldwell, J. Med. Chem.,
1999, 42, 3066–3074.
7 X. Shi, I. Kachirskaia, K. L. Walter, J.-H. A. Kuo, A. Lake, F. Davrazou,
S. M. Chan, D. G. E. Martin, I. M. Fingerman, S. D. Briggs, L. Howe,
P. J. Utz, T. G. Kutateladze, A. A. Lugovskoy, M. T. Bedford and
O. Gozani, J. Biol. Chem., 2007, 282, 2450–2455.
8 S. A. Jacobs and S. Khorasanizadeh, Science, 2002, 295, 2080–2083.
9 J. F. Flanagan, L.-Z. Mi, M. Chruszcz, M. Cymborowski, K. L. Clines,
Y. Kim, W. Minor, F. Rastinejad and S. Khorasanizadeh, Nature, 2005,
438, 1181–1185.
10 Y. Huang, J. Fang, M. T. Bedford, Y. Zhang and R.-M. Xu, Science,
2006, 312, 748–751.
18 A. G. Cochran, N. J. Skelton and M. A. Starovasnik, Proc. Natl. Acad.
Sci. U. S. A., 2001, 98, 5578–5583.
19 Spartan ‘10, Wavefunction, Inc, Irvine, CA, 2010.
20 P. Gans, A. Sabatini and A. Vacca, HypNMR – http://www.hyperquad.co.
uk/hypnmr.htm 2006.
21 T. B. Gasa, C. Valente and J. F. Stoddart, Chem. Soc. Rev., 2011, 40,
57–78.
22 G. Hennrich, V. M. Lynch and E. V. Anslyn, Chem.–Eur. J., 2002, 8,
2274–2278.
23 D. Mendel, J. Ellman, Z. Chang, D. Veenstra, P. Kollman and P. Schultz,
Science, 1992, 256, 1798–1802.
24 S. M. Ngola, P. C. Kearney, S. Mecozzi, K. Russell and D. A. Dougherty,
J. Am. Chem. Soc., 1999, 121, 1192–1201.
25 R. M. Hughes and M. L. Waters, J. Am. Chem. Soc., 2005, 127,
6518–6519.
26 H. J. Dyson, P. E. Wright and H. A. Scheraga, Proc. Natl. Acad.
Sci. U. S. A., 2006, 103, 13057–13061.
11 K. D. Daze, T. Pinter, C. S. Beshara, A. Ibraheem, S. A. Minaker,
M. C. F. Ma, R. Courtemanche, R. E. Campbell and F. Hof, Chem. Sci.,
2012, DOI: 10.1039/c2sc20583a.
12 S. A. Minaker, K. D. Daze, M. C. F. Ma and F. Hof, J. Am. Chem. Soc.,
2012, DOI: 10.1021/ja303465x.
27 G. Némethy, I. Z. Steinberg and H. A. Scheraga, Biopolymers, 1963, 1,
43–69.
13 K. D. Daze and F. Hof, Acc. Chem. Res., 2012, DOI: 10.1021/ar300072g.
14 A. L. Whiting, N. M. Neufeld and F. Hof, Tetrahedron Lett., 2009, 50,
7035–7037.
15 P. C. Kearney, L. S. Mizoue, R. A. Kumpf, J. E. Forman, A. McCurdy
and D. A. Dougherty, J. Am. Chem. Soc., 1993, 115, 9907–9919.
16 H. Li, E. A. Homan, A. J. Lampkins, I. Ghiviriga and R. K. Castellano,
Org. Lett., 2005, 7, 443–446.
28 A. Fernandez and H. A. Scheraga, Proc. Natl. Acad. Sci. U. S. A., 2003,
100, 113–118.
29 C. D. Tatko and M. L. Waters, Protein Sci., 2004, 13, 2515–
2522.
30 AIST, Spectral Database for Organic Compounds (SDBS) maintained by
the National Institute of Advanced Industrial Science and Technology
(Japan).
6892 | Org. Biomol. Chem., 2012, 10, 6885–6892
This journal is © The Royal Society of Chemistry 2012