Chemical triple-mutant boxes for quantifying cooperativity in intermolecular interactions

Article abstract of DOI:10.1002/1521-3765(20021202)8:23<5435::AID-CHEM5435>3.0.CO;2-V

Chemical double-mutant cycles have been used to quantify intermolecular functional-group interactions in H-bonded zipper complexes in chloroform. If the same interaction is measured in zippers of different overall stability, the double-mutant cycles can b

Full text of DOI:10.1002/1521-3765(20021202)8:23<5435::AID-CHEM5435>3.0.CO;2-V

FULL PAPER
Chemical Triple-Mutant Boxes for Quantifying Cooperativity in
Intermolecular Interactions
Christopher A. Hunter,*[a] Philip S. Jones,[b] Pascale Tiger,[a] and Salvador Tomas[a]
Abstract: Chemical double-mutant cy-
cles have been used to quantify inter-
molecular functional-group interactions
in H-bonded zipper complexes in
chloroform. If the same interaction is
measured in zippers of different overall
stability, the double-mutant cycles can
be combined to produce a triple-mutant
box. This construct quantifies coopera-
tivity between the functional group
interaction of interest and the other
interactions that are used to change the
overall stability of the complexes. The
sum of two edge-to-face aromatic inter-
actions (À2.9 Æ 0.5 kJmolÀ1) is shown to
be insensitive to changes of up to 13.7 Æ
0.2 kJmolÀ1 in the overall stability of the
complex. In principle, enthalpic cooper-
ative effects caused by entropy enthal-
py compensation could perturb the
measurement of intermolecular interac-
tions when using the double-mutant
cycle approach, but these experiments
show that, for this system, the magnitude
of the effect lies within the error of the
measurements.
Keywords: cooperative phenomena
¥ enthalpy entropy compensation ¥
supramolecular chemistry
interactions
¥ weak
Introduction
The characterisation of weak, noncovalent interactions is an
issue of fundamental interest in supramolecular and biological
chemistry,[1] since these interactions control a range of
processes, such as protein folding, molecular recognition and
the formation of crystalline solids. If we are to formulate
molecular-design strategies based on such processes, a de-
tailed quantitative understanding of intermolecular interac-
tions is essential. One of the most difficult questions to
address is the way in which any given interaction is affected by
the presence of other nearby interactions, that is, cooperative
effects.
The enthalpy of an interaction can be altered by nearby
interactions due to a change in conformation or electronic
structure. Inductive effects cause a change in internal
electronic structure that leads to the cooperative reinforce-
ment of polar interactions such as hydrogen bonds (Fig-
8]
ure 1a).[2
If a receptor binds two different molecules,
cooperativity can be mediated by direct interactions between
15]
the two substrates (Figure 1b).[9
For a conformationally
Figure 1. Different forms of cooperativity (DG2=DG1). a) Cooperativity
mediated by a change in electronic structure. b) Cooperativity mediated by
a conformational change. c) The classic chelate effect.
[a] Prof. C. A. Hunter, Dr. P. Tiger, Dr. S. Tomas
Centre for Chemical Biology
Krebs Institute for Biomolecular Science, Department of Chemistry
University of Sheffield, Sheffield, S3 7HF (UK)
Fax : (‡44)114-273-8673
mobile molecule with several interaction sites, the formation
of one binding interaction may affect a subsequent interac-
tion, by altering the accessible conformational states (Fig-
ure 1b). This can be manifested in the enthalpy if a high-
energy conformation is required for binding, or in the entropy
E-mail: c.hunter@sheffield.ac.uk
[b] Dr. P. S. Jones
Roche Discovery Welwyn, 40 Broadwater Road
Welwyn Garden City, Hertfordshire, AL7 3AY (UK)
Chem. Eur. J. 2002, 8, No. 23
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5435 $ 20.00+.50/0
5435

C. A. Hunter et al.
FULL PAPER
if a flexible molecule must adopt a more organised confor-
single point interaction between two functional groups i and j.
The strongly bound complex (A') has a deep narrow enthalpic
well, and so the thermally populated intermolecular vibra-
tional states sample a very small range of i j separations (Dx).
In contrast, in the weakly bound complex (A), the broad
shallow potential well leads to thermal population of many
intermolecular vibrational states in which the i j separation
varies over a large amplitude. Thus the mean functional-group
separation is larger in a weakly bound complex, and the
interaction enthalpy will be correspondingly reduced (by
DDHi-j) relative to the maximum value that is observed in
strongly bound complexes. Evidence for this enthalpic chelate
effect has been obtained from studies of complexes of the
vancomycin family of antibiotics with peptides.[1, 3, 4, 20, 27]
We have been working on the quantification of weak
intermolecular functional-group interactions using a chemical
21]
mation for binding.[1, 3, 16
The entropy cost of restricting the overall molecular
translational and rotational degrees of freedom in an inter-
molecular association does not apply to an intramolecular
interaction (Figure 1c), and this gives rise to cooperativity,
generally termed the chelate effect.[22 24] Changes in solvation
may also play an important role in mediating cooperative
effects, especially with respect to the interplay between the
enthalpic and entropic contributions.
When a molecule interacts with a receptor, a significant
amount of residual intermolecular motion still remains, and so
the entropy lost is generally less than the theoretical
maximum. If another intermolecular interaction is now
added, the residual intermolecular motion will be restricted,
so the favourable enthalpy due to the new interaction will be
compensated for to some extent by an entropy loss. Williams
has suggested that this gives rise to the commonly observed
phenomenon of enthalpy entropy compensation.[1, 25, 26]
There is a finite entropy associated with the intermolecular
degrees of freedom that can be lost when a complex is formed.
Thus, as intermolecular interactions become more enthalpi-
cally favourable, the entropy loss should approach this limit,
that is, the compensation between enthalpy and entropy
should follow an asymptotic curve rather than a straight line.
In practice, this means that maximal chelate effects will only
be realised in systems where all of the intermolecular degrees
of freedom have already been removed by other interactions.
This has consequences for the enthalpies associated with
the individual interactions that hold a complex together.
When there is a lot of residual motion in a weakly bound
complex, all of the interactions will be weakened compared
with a strongly bound complex. Williams has proposed a
model based on differences in the shapes of the enthalpic
wells in these two situations (Figure 2).[1, 25, 26] Consider a
version of the double-mutant cycles originally devised to
36]
investigate interactions in proteins.[28
Entropy enthalpy
compensation has important implications for such studies. If
enthalpic chelate effects are large, then any attempts to
experimentally quantify intermolecular functional-group in-
teraction energies will be prone to errors, unless the experi-
ments are carried out at the strong binding limit. Conversely,
the double-mutant approach for measuring functional group
interaction energies provides us with a convenient tool with
which to directly probe the magnitude of the enthalpic chelate
effect experimentally. Thus chemical double-mutant cycles
can be used to measure the magnitude of a particular
functional-group interaction in both weakly and strongly
bound complexes. Any difference between the functional-
group interaction energies in the two systems provides a
measure of the magnitude of the enthalpic chelate effect in
these complexes.
This kind of approach has been used previously to measure
cooperative interactions at protein protein interfaces.[2, 5, 7, 8]
Two double-mutant cycles can be formally combined to
produce a triple-mutant box, as illustrated in Figure 3, and this
provides a general method for quantifying cooperative effects.
Complexes A D are used to measure the i j interaction in a
weakly bound complex by using the usual double-mutant
cycle method, which enables us to dissect out one isolated
functional-group interaction from the array of other inter-
actions and secondary effects that contribute to the overall
Figure 2. The enthalpic chelate effect. The intermolecular vibrational
modes are damped in a strongly bound complex (A') compared with a
weakly bound complex (A), because the potential well is deeper and
narrower. Thus on average, the separation between two interacting groups i
and j (Dx) is reduced in a strongly bound complex, and this leads to an
increase in the enthalpy of interaction (DDHi-j).
Figure 3. A triple-mutant box for determining the magnitude of the
cooperativity between the interaction of two functional groups, i and j, and
additional interactions that stabilise the core of the complex.
5436
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5436 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 23

Quantifying Cooperativity
5435 5446
binding energy of Complex A (back face of the cube in
Figure 3) [Eq. (1)].
does not impinge on the triple-mutant analysis presented in
this paper. Complex A in Figure 4 is the first member of a
family of related zipper complexes that are amide oligomers
composed of the alternating repeats of the bisaniline (B) and
isophthalic acid (I) subunits, TBT¥ AIA, (AIBT)2, TBIBT¥
AIBIA, and so on (see Figure 5).[18, 37] These complexes are
ideally adapted for the triple-mutant box experiment shown in
DDGi-j ˆ DGA À DGB À DGC ‡ DGD
(1)
Complex A' contains exactly the same i j interaction in
exactly the same environment, but with some additional
remote interactions that increase
the overall stability of this system
relative to Complex A. The dou-
ble-mutant cycle constructed with
Complexes A' D' therefore
measures the
i j functional-
group interaction in the more
strongly bound complex, A'
(front face of the cube in Fig-
ure 3) [Eq. (2)].
DDG'i-j ˆ DGA' À DGB' À DGC' ‡ DGD'
(2)
The difference between the two
measurements of the i j interac-
tion is the magnitude of the
cooperativity between the addi-
tional interactions present in
Complex A' and the i j interac-
tion [Eq. (3)].
DDGcoop ˆ DDG'i-j À DDGi-j
(3)
Figure 4. A chemical double-mutant cycle for determining the magnitude of the two terminal edge-to-face
interactions in complex A. Inset: the cartoon relationship to Figure 3.
This procedure is clearly based
on measurements of free energy
rather than enthalpy, but an en-
thalpic chelate effect must mani-
fest itself as a cooperativity in free
energy in this scheme. Thus the
triple-mutant box will allow us to
directly probe enthalpic coopera-
tivity.
Approach: The system that we
have used previously to quantify
edge-to-face aromatic interac-
tions is shown in Figure 4.[35] The
interaction between the terminal
tert-butylbenzoyl (T) and diiso-
propyl aniline (A) groups was
found to be À1.4 Æ 0.8 kJmolÀ1
by mutating T to trimethyacetyl
(X) and A to n-hexylamine (H)
and constructing the double-mu-
tant cycle. The cycle in Figure 4
actually measures the sum of the
two terminal T¥ A interactions,
and, to obtain the value above,
we have previously assumed that
the two interactions have the
same energy, but this assumption
Figure 5. The family of zipper complexes used in this study. The cartoon relationship to Figure 3 is also shown.
Chem. Eur. J. 2002, 8, No. 23
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5437 $ 20.00+.50/0
5437

C. A. Hunter et al.
FULL PAPER
Figure 3. The stability of the zipper complex increases by an
order of magnitude for every B ¥ I subunit added, and
intermolecular NOEs together with the complexation-in-
duced changes in 1H NMR chemical shift show that the
structures of the terminal A ¥ T edge-to-face interactions are
essentially identical in these systems. Thus T! X and A ! H
mutations on the zipper complexes in Figure 5 will allow us to
construct three triple-mutant boxes: by combining the TBT¥
AIA and (AIBT)2, the TBT¥ AIA and TBIBT¥ AIBIA, and
the (AIBT)2 and TBIBT¥ AIBIA double-mutant cycles.
The components of the shortest zippers based on AIA ¥
TBT were obtained in a straightforward fashion by single-
step coupling reactions. The syntheses of AIA and HIH have
been described previously.[37, 38] TB1T and XB1X were ob-
tained by treating H2N-B1-NH2 with an excess of 4-tert-
butylbenzoyl chloride or trimethylacetyl chloride respectively
(Scheme 2).
O
O
O
N
Cl
H2N-B1-NH2
O
O
Results and Discussion
N
H
N
H
O
Synthesis: The first problem we faced in implementing this
approach was solubility. The longer zipper components are
almost insoluble in chloroform. They can be solubilised with
small amounts of methanol, but this destabilises the short
zipper complexes to the extent that it is not possible to
measure accurate binding constants. We therefore developed
an analogue of the bisaniline building block equipped with a
site for the attachment of solubilising groups (Scheme 1). The
compound 2,6-dimethylaniline was heated under reflux with
piperidin-4-one and HCl to give the piperidine analogue of
H2N-B-NH2, 1. Condensation of the secondary amine with
benzylchloroformate gave H2N-B1-NH2. Amide oligomers
incorporating this subunit showed significantly improved
solubility, but the carbobenzoxy (Cbz) group was not suffi-
cient to solubilise AIBIA or the associated mutant XIBIX.
For these compounds, the gallic acid derivative 2 shown in
Scheme 1 was required in order to achieve good solubility in
pure chloroform. The fact that solubilising groups are differ-
ent in different classes of compound does not affect the
thermodynamic analysis, since all four complexes in every
double-mutant cycle have an identical set of solubilising
groups.
Cl
XB1X
O
O
N
O
O
N
H
N
H
TB1T
Scheme 2.
The components of the zipper complexes required for the
(AIB1T)2 double-mutant cycle were prepared by using the
routes shown in Schemes 3 5, below. Excess H2N-B1-NH2
was treated with 4-tert-butylbenzoyl chloride and trimethyl-
acetyl chloride to obtain TB NH2 and XB NH2 respectively
after purification by column chromatography (Scheme 3).
1
1
À
À
À
Monofunctionalised acid chloride AI Cl was readily prepared
from isophthaloyl chloride and 2,6-diisopropylaniline as
described previously (Scheme 3),[37] but it proved impossible
to obtain the n-hexyl analogue
1
À
HI Cl by this method. AIB T
and AIB1X were therefore ob-
tained by direct coupling of
O
O
H
N
À
N
AI Cl with the appropriate
O
N
NH2
1
À
monoamines TB NH2 and
CbzCl
HCl, ∆
XB NH2 (Scheme 4). HIB1T
1
À
+
and HIB1X were obtained by a
H2N
NH2
H2N
NH2
O
different procedure. The appro-
priate monoamine, TB NH2
or XB NH2, was added to an
1
H2N-B1-NH2
À
1
1
À
excess of isophthaloyl chloride,
the reaction was then quenched
with excess n-hexylamine, and
the products were separated
by column chromatography
(Scheme 5).
Similar methodology was
used to obtain the components
of the longer zippers based on
TBIBT¥ AIBIA. Isophthaloyl
chloride was added to excess
OC14H29
OC14H29
O
N
OC14H29
O
O
i. BrC14H29
K2CO3
,
OC14H29
OC14H29
i. SOCl2
OH
OH
HO
O
ii. 1
ii. KOH
OC14H29
H2N
NH2
OH
2
H2N-B2-NH2
Scheme 1.
5438
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5438 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 23

Quantifying Cooperativity
5435 5446
NH2
O
O
O
O
O
O
Cl
Cl
N
H
Cl
N
AI-Cl
+
O
O
O
AI-Cl
N
H
N
H
N
H
TB1-NH2
AIB1T
O
O
O
Cl
H2N-B1-NH2
N
O
N
H
NH2
O
O
O
N
Cl
XB1-NH2
AI-Cl
+
O
O
O
N
H
N
H
N
H
XB1-NH2
O
O
AIB1X
N
Scheme 4.
O
N
H
NH2
n, experience a downfield shift due to H bonding. The signals
due to t and d on the I subunit experience large upfield shifts,
while the signals due to a and m on the B subunit show small
downfield shifts; this is characteristic of the edge-to-face
interaction that docks I into the aromatic pocket of B. As
reported previously, the t CIS value is lower in the HIH
complexes, due to an increase in conformational flexibility
relative to the AIA complexes.[38] In the AIA complexes, the
geometry is locked by steric interactions with the isopropyl
groups, but in the HIH complexes, the I subunit can flex to
some extent inside the B binding pocket, and this produces
large chemical-shift differences. However, this conforma-
tional difference has no impact on the analysis, because the
thermodynamic effects on the core of the complex appear in
both of the HIH complexes used in the double-mutant cycle,
and they therefore cancel out.
TB1-NH2
Scheme 3.
H2N-B1-NH2, and the resulting amide oligomers were sepa-
rated by column chromatography to give H2N-B1IB1-NH2 as
the major product. This was then capped with an excess of
4-tert-butylbenzoyl chloride or trimethylacetyl chloride to
give TB1IB1T and XB1IB1X respectively (Scheme 6). AIB2IA
was obtained directly by coupling the highly solubilised
bisaniline H2N-B -NH2 with AI Cl (Scheme 7). HIB IH was
obtained adding H2N-B2-NH2 to an excess of isophthaloyl
chloride, quenching the reaction with excess n-hexylamine,
and then separating the products by column chromatography
(Scheme 7).
2
2
À
The self-complementary (AIB1T)2 zippers were character-
ised by using dilution experiments; the dimerisation constants,
Kd, and limiting CIS values are listed in Table 1. The CIS
Binding studies: Formation of
the zipper complexes was in-
vestigated by using 1H NMR
titrations and dilutions in
O
O
O
O
CDCl3 at room temperature.
N
Dilution experiments with the
shortest zippers showed that
dimerisation is negligible for
these systems (Table 1), and
the complexes could therefore
be characterised in straightfor-
ward manner by titration ex-
periments. The association con-
stants, Ka, and complexation-
induced changes in chemical
shift (CIS) are listed in Table 2.
The CIS pattern is similar to
that reported previously for the
zipper complexes.[37] The sig-
nals due to the amide protons,
i.
Cl
Cl
O
O
O
TB1-NH2
N
H
N
H
N
H
ii.
NH2
HIB1T
O
O
O
O
N
i.
Cl
Cl
O
O
O
XB1-NH2
N
H
N
H
N
H
ii.
NH2
HIB1X
Scheme 5.
Chem. Eur. J. 2002, 8, No. 23
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5439 $ 20.00+.50/0
5439

C. A. Hunter et al.
FULL PAPER
patterns are very similar to
those observed in the shorter
zippers. Intermolecular NOEs
observed in ROESY experi-
ments are listed in Table 3,
and
the
proton-labelling
scheme is given in Figure 6.
These confirm that the com-
plexes adopt the expected
structures with interactions be-
tween the B and I subunits and
between the terminal groups, T,
X, A and H.
The longer zippers were
more complicated systems to
characterise. Dilution experi-
ments showed that all of the
compounds self-associate to a
significant extent, and this equi-
librium competes with the for-
mation of the zipper complex.
We found previously that these
systems tend to form dimers
rather than polymeric aggre-
gates in solution.[37] The dilu-
tion experiments were there-
fore analysed by using a dimer-
isation model, but the self-
association constants are not
significantly different if an ag-
gregation model is used. The
dimerisation constants, Kd, and
the CIS values are listed in
Table 1. The dimerisation con-
stants are an order of magni-
Scheme 6.
tude lower than the association constants for the formation of
the TBIBT¥ AIBIA complexes, but the competing dimerisa-
tion equilibria broaden the 1H NMR spectra and significantly
complicate the analysis of titration experiments. These com-
plexes were therefore studied by diluting 1:1 mixtures of the
two components. This minimises the amount of dimer present
throughout the experiment. The free chemical shifts of both
species are known from the dilution experiments, and the Ka
OC14H29
OC14H29
O
OC14H29
N
AI-Cl
O
O
O
O
H2N-B2-NH2
N
N
N
N
H
H
H
H
AIB2IA
OC14H29
OC14H29
O
OC14H29
O
O
N
i.
Cl
Cl
O
O
O
O
H2N-B2-NH2
N
N
H
N
H
N
H
H
ii.
NH2
HIB2IH
Scheme 7.
5440
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5440 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 23

Quantifying Cooperativity
5435 5446
1
Table 1. Dimerisation constants, Kd [mÀ1], and limiting dimerisation-induced changes in H NMR chemical shift [ppm] in CDCl3 at 298 K.[a]
amides
A/H subunit
I subunit
B subunit
T/X subunit
dimer
Kd
n
h/i
j
t
d
a
m
e/f
g
(AIA)2
< 1
< 1
< 1
< 1
(TB1T)2
(HIH)2
(XB1X)2
(AIB1T)2
210 Æ 30
‡ 1.7
‡ 1.3
‡ 1.2
‡ 1.6
‡ 1.1
‡ 0.7
‡ 1.6
‡ 1.0
‡ 0.9
‡ 1.5
‡ 1.1
‡ 0.5
‡ 1.5
‡ 0.7
‡ 1.5
‡ 0.5
‡ 1.6
‡ 1.0
‡ 1.4
‡ 0.4
À 0.1
À 0.1
À 0.3
À 0.3
À 0.1
À 0.2
À 0.1
À 1.7
À 1.7
À 1.0
À 0.9
À 0.5
‡ 0.1[b]
À 0.1[b]
À 0.3
À 0.3
0.0
À 0.5
À 0.1
À 0.5
(AIB1X)2
(HIB1T)2
(HIB1X)2
39 Æ 5
49 Æ 3
33 Æ 5
À 0.6
À 0.4
‡ 0.1[b]
À 0.3
À 0.1
À 0.5
À 0.3
0.0
À 0.1
À 0.3
À 0.1
À 0.5
À 0.3
0.0
À 0.1
À 0.2
À 0.1
0.0
(AIB2IA)2
(TB1IB1T)2
(HIB2IH)2
(XB1IB1X)2
420 Æ 40
81 Æ 23
À 0.1
À 1.3
À 1.3
À 0.7
À 1.5
À 0.5
‡ 0.2[b]
À 0.1[b]
À 0.4
À 0.5[b]
À 0.1
0.0
0.0[b]
À 0.3
À 0.1
À 0.2
À 0.2
À 0.2
110 Æ 30
150 Æ 30
À 0.3
À 0.3
À 0.2
À 0.6
À 0.6
À 0.2[b]
0.0
0.0
À 0.2[b]
[a] Errors in CIS are of the order of 20%. Where more than one proton was observed in each category, they are listed from the highest to the lowest CIS
observed regardless of the position in the molecule. [b] Composite value for multiple unresolved signals. Protons not listed were unaffected by dimerisation
(CIS < 0.1).
1
Table 2. Association constants, Ka [mÀ1], and limiting complexation-induced changes in H NMR chemical shift [ppm] in CDCl3 at 298 K.[a]
amides
n
A/H subunit
I subunit
B subunit
T/X subunit
complex
Ka
h/i
j
t
d
a
m
e/f
g[b]
titrations
AIA ¥ TB1T
38 Æ 2
6 Æ 1
‡ 1.3
‡ 1.1
‡ 1.3
‡ 0.5
‡ 1.0
‡ 0.7
‡ 1.2
‡ 0.6
0.0
0.0
À 0.1
À 0.1
À 1.6
À 0.4
À 0.4
À 0.3
À 0.4
‡ 0.2
0.0
0.0
0.0
À 0.3
À 0.2
À 0.5
AIA ¥ XB1X
HIH ¥ TB1T
HIH ¥ XB1X
À 1.5
À 0.8
À 0.9
‡ 0.1
0.0
13 Æ 1
6 Æ 1
À 0.3
À 0.3
‡ 0.1
0.0
[c]
0.0
À 0.1
dilutions
AIB2IA ¥ TB1IB1T
[c]
10500 Æ 980
1750 Æ 90
0.0
À 0.1
À 0.1
À 1.7
À 1.6
À 0.4[d]
À 0.5[d]
‡ 0.2
0.0
0.0
0.0
À 0.1
À 0.1
0.0
À 0.3
À 0.3
À 0.5
‡ 0.1[d]
AIB2IA ¥ XB1IB1X
HIB2IH ¥ TB1IB1T
HIB2IH ¥ XB1IB1X
‡ 1.4
‡ 1.4
‡ 1.1
‡ 0.8
‡ 1.4
‡ 1.4
‡ 1.1
‡ 1.1
‡ 1.7
‡ 1.0
‡ 1.0
‡ 0.7
À 0.1
À 1.7
À 1.6
‡ 0.2
‡ 0.2
‡ 0.1
1850 Æ 75
1000 Æ 75
À 0.1
À 0.2
À 1.6
À 1.5
À 0.5
À 0.5
À 0.4
‡ 0.1
‡ 0.1
0.0
À 0.2
À 0.1
À 0.1
‡ 0.1
0.0
À 1.5
À 1.4
À 0.6
À 0.5
À 0.4
‡ 0.1
‡ 0.1
0.0
À 0.2
À 0.1
0.0
0.0
[a] Errors in CIS are of the order of 20%. Where more than one proton was observed in each category, they are listed from the highest to the lowest CIS
regardless of the position in the molecule. [b] Composite value for multiple unresolved signals. [c] These signals were not sufficiently resolved during the
titration/dilution experiment to obtain reliable chemical-shift changes. Protons not listed were unaffected by complexation (CIS < 0.1).
values are so large that the experiments start from almost
100% bound, so that the bound chemical shifts are also well-
defined. Thus, the dilution data are analysed to determine one
major unknown, Ka. The experiments were analysed with a
model that also allowed for dimerisation of the two compo-
nents by using the previously determined Kd and CIS values
from Table 1. The results are listed in Table 2. The CIS values
for formation of the TBIBT¥ AIBIA complexes show the
Chem. Eur. J. 2002, 8, No. 23
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5441 $ 20.00+.50/0
5441

C. A. Hunter et al.
FULL PAPER
Table 3. Intermolecular NOEs observed in ROESY experiments in CDCl3
at 298 K.
values are used to construct double-mutant cycles, we find
that, although the overall stability of the complexes varies by
nearly three orders of magnitude, the sum of the two terminal
T¥ A interactions is similar in the all three systems (Figure 7).
If we assume that the two T¥ A interactions are identical, then
complex
core NOEs
terminal NOEs
(AIB1T)2
t
t
t
t
t
t
t
t
m, d a, d
a, t m, d a, d
m
j
f, j g, k o, l
o
o
(AIB1X)2
m
(HIB1T)2
m, d
m, d
m
m
(HIB1X)2
e
j
h
AIB2IA ¥ TB1IB1T
AIB2IA ¥ XB1IB1X
HIB2IH ¥ TB1IB1T
HIB2IH ¥ XB1IB1X
a, t m, d a, d
a, t m, d a, d
a, t m, d a, d
a, t m, d a, d
m
m
m
m
f, j g, k o, l
R
N
m
a
n
H
O
N
Figure 7. The sum of the two terminal A ¥ T interactions measured by using
double-mutant cycles (DDG) plotted as a function of the overall stability of
the zipper complex (DG). The dotted line corresponds to DDG ˆ
À2.9 kJmolÀ1, the mean interaction energy.
B
I
N
t
O
H
d
H
O
N
s
N
H
O
on average each T¥ A interaction contributes À1.4 Æ
0.2 kJmolÀ1 to the overall stability of all of these complexes.[39]
This value is identical to that determined previously in slightly
different zipper complexes (À1.4 Æ 0.5 kJmolÀ1).[35, 38] When
we construct the triple-mutant boxes, the following results are
obtained [Eqs. (4) (6)].
f
g
O
O
e
o
T
X
H
N
N
H
H
O
O
h
A
l
N
N
H
n
k
H
n
i
j
TBT¥ AIA ! (AIBT)2
DDGcoop ˆ À0.5 Æ 0.8 kJmolÀ1
DDGcoop ˆ À0.3 Æ 0.7 kJmolÀ1
DDGcoop ˆ ‡0.2 Æ 0.6 kJmolÀ1
(4)
(5)
(6)
Figure 6. Representative NOEs observed in ROESY experiments. Each
complex shows a subset of these NOEs as detailed in Table 3. The 1H NMR
labelling scheme and subunit nomenclature are also illustrated. The signals
due to protons that are not labelled in this diagram are unaffected by
complexation.
TBT¥ AIA ! TBIBT¥ AIBIA
(AIBT)2 ! TBIBT¥ AIBIA
There is no detectable cooperativity in this system.
For these weak interactions (2.9 kJmolÀ1) in the zipper
system, the magnitude of the enthalpic chelate effect lies
within the experimental error (0.6 0.8 kJmolÀ1). Therefore,
in double-mutant experiments over this range of complex
stability (DG ˆ 4 23 kJmolÀ1), the effects of enthalpic coop-
erativity can be ignored, and free energy differences can be
attributed to differences in functional-group interaction
energies.[38]
same pattern as the other two systems; this indicates that the
expected zipper structure is obtained. This is supported by
NOEs from ROESY experiments (Table 3).
Thus all of the complexes have similar structures and are
suitable for the construction of double-mutant cycles and
triple-mutant boxes. Table 4 summarises the stability-constant
data from the titration and dilution experiments. When these
Experimental evidence for enthalpic cooperativity in the
vancomycin peptide system is the large increase in limiting
CIS values observed as the stability of the complex increas-
es.[1, 3, 4, 20, 27] These chemical-shift changes are evidence for
structural tightening related to a corresponding increase in the
intrinsic functional-group interaction energy. In our experi-
ments, we directly measure the functional-group interaction
energies and see no differences in different zipper complexes.
If we plot the limiting CIS values as a function of the overall
stability of the complex, we also see no changes (Figure 8);
that is, there is no structural tightening in this system,
consistent with the thermodynamic measurements. Thus the
zipper complexes described here appear to have very different
properties from the vancomycin peptide system reported by
Williams, and further work is required to clarify the origins of
this difference.
Table 4. Summary of stability constants, K [mÀ1], free energies of binding,
DG [kJmolÀ1], and double-mutant cycle interaction energies for the two
terminal T¥ A interactions, DDG [kJmolÀ1].
complex
K
DG
À 8.9 Æ 0.1
DDG
AIA ¥ TB1T
38 Æ 2
6 Æ 1
À 2.6 Æ 0.6
AIA ¥ XB1X
HIH ¥ TB1T
À 4.4 Æ 0.4
À 6.3 Æ 0.2
À 4.4 Æ 0.4
À 13.1 Æ 0.3
À 9.0 Æ 0.3
À 9.5 Æ 0.1
À 8.5 Æ 0.3
À 22.6 Æ 0.2
À 18.2 Æ 0.1
À 18.4 Æ 0.1
À 16.9 Æ 0.2
13 Æ 1
6 Æ 1
HIH ¥ XB1X
(AIB1T)2
210 Æ 30
39 Æ 5
49 Æ 3
33 Æ 5
À 3.1 Æ 0.5
À 2.9 Æ 0.3
(AIB1X)2
(HIB1T)2
(HIB1X)2
AIB2IA ¥ TB1IB1T
AIB2IA ¥ XB1IB1X
HIB2IH ¥ TB1IB1T
HIB2IH ¥ XB1IB1X
10500 Æ 980
1750 Æ 90
1850 Æ 75
1000 Æ 75
5442
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5442 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 23

Quantifying Cooperativity
5435 5446
(m, 4H; CH2N), 2.14 (m, 4H; CH2), 2.00 (s, 12H; m); 13C NMR (62.5 MHz,
[D6]DMSO, 218C): d ˆ 154.7, 141.7, 137.3, 134.8, 128.5, 127.9, 127.5, 126.1,
120.5, 66.10, 4.13, 4.12, 35.6, 18.2; MS (‡ve, FAB): m/z (%): 457, (100)
‡
[M ‡H]; elemental analysis calcd (%) for C29H34N3 ¥ 1³2 H2O: C 74.60, H
7.78, N 9.01; found: C 74.48, H 7.61, N 8.82.
Synthesis of 2: n-Propyl gallate (5.55 g, 26.2 mmol), 1-bromotetradecane
(22 mL, 78.6 mmol) and anhydrous K2CO3 (32.22 g, 236 mmol) were
suspended in acetone/DMSO (90:10, v/v; 100 mL). The mixture was stirred
for 30 min at room temperature and then under reflux for 10 h. The
reaction mixture was then poured into water (1.5 L), and the pH was
brought to 6 with a mixture of HCO2H/water (90:10, v/v). The product was
extracted with CH2Cl2 (200 mL), and the organic layer was dried over
Na2SO4 (anhydrous). The volume of solvent was reduced to 50 mL under
reduced pressure, and the resulting solution was purified by flash column
chromatography on basic alumina with CH2Cl2 as eluant. The resulting
solid was suspended in a solution of KOH in EtOH/water (90:10, v/v; 0.5m,
130 mL) and stirred for 1 h under reflux. The reaction was then cooled to
room temperature and brought to pH 6 by adding a mixture of HCO2H/
water (90:10, v/v). The desired product precipitated upon addition of water
(200 mL). It was filtered off, dried and recrystallised from hot EtOH and
Figure 8. Complexation-induced changes in chemical shift (CIS) for the
*
^
signals due to n ( ), d (*) and t ( ) plotted as a function of the overall
stability of the zipper complex (DG) (t data for the HI ¥¥¥ complexes are not
included as the values are perturbed by conformational flexibility as
discussed in the text).
then dried under vacuum for 48 h (11 g, 56% yield). M.p. 43 458C;
Conclusion
1
À
H NMR (250 MHz, CDCl3, 218C): d ˆ 7.31 (s, 2H; Ar CH), 4.02 (m, 6H;
OCH2), 1.81 (m, 6H; OCH2CH2), 1.47 (m, 6H; OCH2CH2CH2), 1.25 (m,
60H; CH2), 0.87 (t, 9H; Me); elemental analysis calcd (%) for C49H90O5: C
77.52, H 11.95; found: C 77.87, H 12.12.
We have shown that triple-mutant boxes provide a powerful
tool for experimentally investigating cooperativity in inter-
molecular interactions. Three double-mutant cycles have been
constructed to quantify the same intermolecular functional
group interaction in three different situations, in which the
overall stability of the complexes varies by 13.7 Æ 0.2 kJmolÀ1.
The aromatic interaction we have quantified is worth À2.9 Æ
0.5 kJmolÀ1 in all three systems, and there is no detectable
cooperative effect. Thus the simple additive approach used in
our double-mutant cycle analysis of functional group inter-
actions in the zipper complexes appears to be justified.
Synthesis of H2N-B2-NH2: Oxalyl chloride (80 mL, 783 mmol) was added
slowly to a suspension of acid 2 (6.6 g, 8.7 mmol) in CH2Cl2 (100 mL). The
reaction mixture was stirred for 12 h, and then the solvent was removed
under reduced pressure. To eliminate unreacted oxalyl chloride, the solid
was dissolved in CH2Cl2 (100 mL), and the solvent was removed under
reduced pressure. The last traces of oxalyl chloride were eliminated under
high vacuum. The remaining solid was then dissolved in CH2Cl2 (40 mL)
and added dropwise to a solution of 1 (2.5 g, 7.74 mmol) and Et3N (1.63 mL,
11.6 mmol) in CH2Cl2 (40 mL), at 58C. The reaction mixture was stirred for
16 h at room temperature. The solution was then washed with aqueous HCl
(1m, 200 mL) and aqueous NaOH (1m, 200 mL), and then dried with
Na2SO4, filtered, and the solvent was removed under reduced pressure.
Flash column chromatography on silica eluting with a mixture of CH3Cl2/
EtOH (98:2, v/v) yielded H2N-B2-NH2 as a white solid (5.34 g, 65% yield).
M.p. 69 718C; 1H NMR (250 MHz, CDCl3, 218C): d ˆ 6.79 (s, 4H; a), 6.52
(s, 2H; ArCH), 3.93 (t, 6H; OCH2), 3.75 (m, 2H; CH2N), 3.45 (m, 2H;
CH2N), 2.57 (m, 4H; CH2), 2.20 (s, 12H; m), 1.75 (m, 6H; OCH2CH2), 1.43
(m, 6H; OCH2CH2CH2), 1.25 (m, 60H; CH2), 0.83 (t, 9H; Me); MS (‡ve,
Experimental Section
À
The preparation of AIA, HIH and AI Cl has been described previous-
ly.[37, 38] All other chemicals were purchased from Aldrich and used without
further purification. The 1H NMR signals are assigned by using the proton-
labelling scheme in Figure 6.
‡
FAB) m/z (%): 1066, (65) [M ‡2H]; elemental analysis calcd (%) for
C70H117N3O4: C 78.97, H 11.08, N 3.95; found: C 78.85, H 11.09, N 3.87.
Synthesis of TB1T: 4-tert-butylbenzoyl chloride (1.52 mL, 7.8 mmol) was
dissolved in CH2Cl2 (20 mL), and the mixture was cooled to 58C. H2N-B1-
NH2 (1.18 g, 2.6 mmol) and Et3N (1.1 mL, 7.8 mmol) dissolved in CH2Cl2
(15 mL) were then added dropwise, and the resulting solution was stirred
for 12 h at room temperature. Then CH2Cl2 (100 mL) was added, and the
solution was washed with aqueous HCl (1m, 5 Â 100 mL), aqueous NaOH
(1m, 5 Â 100 mL), water (200 mL) and brine (30 mL). The organic phase
was dried over Na2SO4 (anhydrous), and the solvent was removed under
reduced pressure. Recrystallisation from CH2Cl2/pet.ether (40 60) (50:50)
yielded TB1T as a white solid. (2.90 g, 97%). M.p. 250 2528C; 1H NMR
(250 MHz, [D6]DMSO, 218C): d ˆ 9.58 (s, 2H; n), 7.91 (d, 4H; f), 7.52 (d,
4H; g), 7.34 (m, 5H; ArCH), 7.18 (s, 4H; a), 5.08 (s, 2H; OCH2), 3.45 (m,
4H; CH2N), 2.38 (m, 4H; CH2), 2.13 (s, 12H; m), 1.28 (s, 18H; o); 13C NMR
(62.5 MHz, CDCl3, 218C): d ˆ 165.7, 155.4, 145.2, 136.8, 135.5, 132.0, 131.6,
128.5, 127.8, 127.2, 127.1, 126.9, 125.7, 67.0, 44.0, 41.0, 35.0, 31.2, 18.9. MS
Synthesis of 1: A mixture of 2,6-dimethylaniline (119 mL, 1.24 mol), N-
acetyl-4-piperidone (55.4 mL, 0.45 mol) and concentrated hydrochloric
acid (150 mL) was stirred under reflux. After 24 hours, more 2,6-
dimethylaniline (13.1 mL, 0.12 mol) was added, and the same amount
was added again after 48 hours. After a further 72 hours, the reaction
mixture was cooled to room temperature and diluted with water (1.2 L).
The resulting mixture was neutralised by addition of solid Na2CO3. The
precipitate was then filtered off, washed with water (500 mL), Et2O
(500 mL) and pentane (500 mL) and dried under vacuum, yielding the
desired product as
a white powder (87.5 g, 60%). M.p. 203 2058C;
1H NMR (250 MHz, [D6]DMSO, 218C): d ˆ 8.85 (s, 1H; NH), 6.71 (s,
4H; ArCH), 4.41 (s, 4H; NH2), 2.94 (m, 4H; CH2NH), 2.40 (m, 4H; CH2),
2.09 (s, 12H; Me); 13C NMR (62.5 MHz, [D6]DMSO, 218C): d ˆ 141.8,
134.9, 127.8, 120.6, 115.9, 41.4, 41.1, 32.8, 18.22; MS (‡ve, FAB): m/z (%):
‡
324 (90) [M ].
‡
Synthesis of H2N-B1-NH2: Compound 1 (14.14 g, 44 mmol) and Et3N
(9.13 mL, 66 mmol) were dissolved in MeOH (250 mL). A solution of
benzyl chloroformate (6.87 mL, 48 mmol) in CH2Cl2 (30 mL) was added
dropwise to the resulting suspension and the mixture was stirred for two
hours. The volume of solvent was reduced by half under reduced pressure,
and the mixture was poured into water (200 mL) and extracted with CH2Cl2
(3 Â 100 mL). The organic layer was dried over Na2SO4 (anhydrous),
evaporated under reduced pressure and then washed with pentane
(500 mL), yielding the desired product as a pale yellow powder (13.6 g,
68% yield). M.p. 132 1348C; 1H NMR (250 MHz, CDCl3, 218C): d ˆ 7.34
(m, 5H; ArCH), 6.68 (s, 4H; a), 5.03 (s, 2H; OCH2), 3.32 (s, 4H; NH2), 2.48
(‡ve, FAB) m/z (%): 778 (100) [M ‡H]; elemental analysis calcd (%) for
C51H59N3O4: C 78.21, H 7.61, N 5.17; found: C 78.73, H 7.64, N 5.40.
Synthesis of XB1X: Trimethylacetyl chloride (1.85 mL, 15 mmol) was
dissolved in CH2Cl2 (20 mL), and the mixture was cooled to 58C. H2N-B1-
NH2 (2.3 g, 5 mmol) and Et3N (2.1 mL, 15 mmol) in CH2Cl2 (50 mL) were
then added dropwise. The solution was stirred for 12 h at room temper-
ature. The product was isolated as a white powder after workup of the
reaction by using the procedure described for TB1T. XB1X was obtained as
a white solid (2.00 g, 66% yield). M.p. 237 2398C; 1H NMR (250 MHz,
CDCl3 218C): d ˆ 7.33 (m, 4H; ArCH), 6.88 (s, 4H; a), 6.80 (s, 2H; n), 5.10
(s, 2H; OCH2), 3.52 (m, 4H; CH2N), 2.28 (m, 4H; CH2), 2.12 (s, 12H; m),
Chem. Eur. J. 2002, 8, No. 23
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5443 $ 20.00+.50/0
5443

C. A. Hunter et al.
FULL PAPER
1.33 (s, 18H; e); 13C NMR (62.5 MHz, CDCl3, 218C): d ˆ 176.5, 154.5,
period, the mixture was added dropwise to a solution of n-hexylamine
(5.50 g, 52 mmol) and Et3N (7.29 mL, 52 mmol) in CH2Cl2 (200 mL) at 58C.
The resulting solution was stirred for 16 h at room temperature, and then
washed with aqueous HCl (1m, 4 Â 150 mL), aqueous NaOH (1m, 2 Â
150 mL) and brine (2 Â 50 mL). The organic phase was dried over Na2SO4
(anhydrous), and the solvent was removed under reduced pressure. Flash
chromatography on silica eluting with a mixture of CHCl3/EtOH (97.5:2.5,
v/v) yielded HIB1X as a white solid (0.48 g, 48% yield). M.p. 194 1968C;
1H NMR (250 MHz, CDCl3, 218C): d ˆ 8.35 (s, 1H; s), 8.07 (s, 1H; n), 7.90
(d, 1H; d), 7.68 (d, 1H; d), 7.33 (m, 5H; ArCH), 7.11 (t, 1H; t), 7.06 (s, 1H;
n), 6.97 (s, 2H; a), 6.92 (s, 2H; a), 6.76 (s, 1H; n), 5.10 (s, 2H; OCH2), 3.54
(m, 4H; CH2N), 3.51 (m, 2H; h), 2.33 (m, 4H; CH2), 2.18 (s, 6H; m), 2.12 (s,
6H; m), 1.77 (m, 2H; NHCH2CH2), 1.30 (m, 15H; e/CH2), 0.84 (m, 3H;
CH3); 13C NMR (62.5 MHz, [D6]DMSO, 218C): d ˆ 175.6, 167.0, 163.5,
154.7, 148.4, 147.4, 135.1, 134.9, 134.6, 132.6, 132.0, 130.0, 128.9, 128.4, 127.6,
127.5, 126.8, 126.3, 125.6, 66.3, 45.4, 41.0, 40.5, 31.9, 29.4, 27.5, 26.3, 22.8,
145.0, 135.3, 131.9, 128.5, 127.9, 127.8, 126.8, 66.9, 44.0, 41.0, 39.3, 27.8, 18.7.
‡
MS (‡ve, FAB) m/z (%): 626, (100) [M ]; elemental analysis calcd (%) for
C39H51N3O4: C 74.85, H 8.21, N 6.71; found: C 75.44, H 8.15, N 6.61.
1
1
À
Synthesis of XB NH2: A mixture of H2N-B -NH2 (11 g, 24 mmol) and
Et3N (5.0 mL, 36 mmol) in CH2Cl2 (60 mL) was cooled to 58C. Trimethyl-
acetyl chloride (0.74 mL, 6 mmol) dissolved in CH2Cl2 (300 mL) was added
dropwise over a period of 2 h. The solution was then allowed to warm up to
room temperature and stirred for a further 12 h. The solution was then
washed with aqueous HCl (1m, 10 Â 100 mL), water (200 mL) and brine
(30 mL). The organic phase was dried over Na2SO4 (anhydrous), and the
solvent was removed under reduced pressure. Flash column chromatog-
1
À
raphy with CH2Cl2 as eluant yielded XB NH2 as a white powder (3.19 g,
98% yield). M.p. 222 2258C; 1H NMR (250 MHz, [D6]DMSO, 218C): d ˆ
8.60 (s, 1H; n), 7.38 (m, 5H; ArCH), 6.93 (s, 2H; a), 6.77 (s, 2H; a), 5.09 (s,
2H; OCH2), 4.40 (s, 2H; NH2), 3.43 (m, 4H; CH2N), 2.18 (m, 4H; CH2),
2.05 (s, 6H; m), 2.02 (s, 6H; m), 1.23 (s, 9H; e); 13C NMR (62.5 MHz,
[D6]DMSO, 218C): d ˆ 176.4, 155.0, 146.3, 142.4, 137.5, 135.6, 133.6, 133.4,
128.9, 128.3, 127.9, 126.6, 126.1, 120.9, 66.5, 43.1, 41.4, 35.7, 27.9, 18.7; MS
‡
18.6, 18.4, 13.97; HRMS (FAB) m/z ˆ 773.4612, C48H61N4O5 requires
773.4642.
Synthesis of HIB1T: HIB1T was prepared and purified in the same way as
1
1
1
‡
À
HIB X, starting from XB NH2 (0.80 g, 1.30 mmol). HIB Twas obtained as
a white solid (0.51 g, 46% yield). M.p. 180 1828C; 1H NMR (250 MHz,
CDCl3 218C): d ˆ 8.31 (s, 1H; s), 7.95 (s, 1H; n), 7.91 (d, 1H; d), 7.86 (d, 2H;
f), 7.76 (d, 1H; d), 7.56 (s, 1H; n), 7.49 (d, 2H; g), 7.34 (m, 5H; ArCH), 7.15
(t, 1H; t), 7.01 (s, 2H; a), 6.97 (s, 2H; a), 6.61 (t, 1H; n), 5.11 (s, 2H; OCH2),
3.45 (m, 4H; CH2N), 3.38 (m, 2H; h), 2.33 (m, 4H; CH2), 2.12 (s, 6H; m),
2.16 (s, 6H; m), 1.76 (m, 2H; NHCH2CH2), 1.34 (s, 9H; o), 1.29 (m, 6H;
CH2), 0.90 (m, 3H; CH3); 13C NMR (62.5 MHz, CDCl3, 218C): d ˆ 166.5,
166.0, 165.3, 155.2, 145.1, 144.8, 136.6, 135.6, 134.9, 134.4, 133.7, 132.3, 132.0,
130.1, 128.3, 127.8, 127.6, 127.1, 126.4, 126.3, 125.4, 66.9, 43.5, 40.8, 40.0, 35.5,
34.8, 31.3, 31.0, 29.3, 26.5, 22.4, 18.7, 18.5, 13.9; HRMS (FAB) m/z ˆ
(‡ve, FAB) m/z (%): 541, (100) [M ]; elemental analysis calcd (%) for
C34H43N3O3: C 75.38, H 8.00, N 7.76; found: C 74.49, H 8.04, N 7.35.
1
1
À
À
Synthesis of TB NH2: TB NH2 was prepared in the same way as
1
1
À
XB NH2, with H2N-B -NH2 (16.0 g, 35 mmol) and 4-tert-butylbenzoyl
chloride (2.3 mL, 11 mmol). The desired product was obtained as a white
solid (5.49 g, 81% yield) after flash column chromatography with CH2Cl2/
MeOH (98:2, v/v) as eluant. M.p. 262 2658C; 1H NMR (250 MHz,
[D6]DMSO, 218C): d ˆ 9.53 (s, 1H; n), 7.93 (d, 2H; f), 7.52 (d, 2H; d, g),
7.34 (m, 5H; ArCH), 7.03 (s, 2H; a), 6.78 (s, 2H; a), 5.07 (s, 2H; OCH2),
4.42 (s, 2H; NH2), 3.48 (m, 4H; CH2N), 2.29 (m, 4H; CH2), 2.16 (s, 6H; m),
2.02 (s, 6H; m), 1.32 (s, 9H; o); 13C NMR (62.5 MHz, [D6]DMSO, 218C):
d ˆ 165.3, 155.0, 154.7, 146.71, 142.4, 137.5, 135.6, 133.5, 133.2, 132.1, 128.9,
128.2, 127.9, 127.8, 126.6, 126.3, 125.6, 121.0, 66.5, 43.1, 41.4, 35.1, 31.4, 18.9,
‡
849.4957, C48H61N4O5 requires 849.4955.
Synthesis of H2N-B1IB1-NH2: Isophthaloyl dichloride (0.93 g, 4.58 mmol)
was dissolved in CH2Cl2 (85 mL), and the solution was added dropwise to a
mixture of Et3N (1.30 mL, 9.16 mmol) and H2N-B1-NH2 (2.5 g, 4.6 mmol) in
CH2Cl2 (150 mL). The resulting solution was stirred for 16 h, then the
solvent was removed under reduced pressure. Flash chromatography on
silica, with a gradient elution with a mixture CHCl3/THF (98:2 to 90:10,
v/v) allowed separation of unreacted H2N-B1-NH2 from the desired H2N-
‡
18.7; MS (‡ve, FAB) m/z (%): 618 (60) [M ‡H].
1
1
À
Synthesis of AIB X: XB NH2 (0.638 g, 1.18 mmol) and Et3N (0.210 mL,
1.47 mmol) were dissolved in CH2Cl2 (50 mL), and the mixture was cooled
À
to 58C. Freshly prepared AI Cl (0.506 g, 1.47 mmol) dissolved in CH2Cl2
(25 mL) was added dropwise. The reaction mixture was then stirred for 16 h
at room temperature. The solution was washed with HCl (1m, 3 Â 100 mL),
water (200 mL) and brine (30 mL), the organic phase was dried over
Na2SO4 (anhydrous), and the solvent was removed under reduced pressure.
Flash column chromatography on silica (eluant CH2Cl2/EtOH 99.5:0.5, v/v)
yielded AIB1X as a white powder (0.88 g, 88% yield). M.p. 153 1558C;
1H NMR (250 MHz, CDCl3 218C): d ˆ 8.40 (s, 1H; s), 8.22 (s, 1H; n), 7.86
(s, 1 H; n), 7.77 (d, 1H; d), 7.73 (d, 1H; d), 7.34 (t, 1H; t), 7.38 (m, 5H;
ArCH), 7.20 (t, 2H; l), 7.10 (s, 1H; n), 7.02 (d, 1H; k), 6.95 (s, 2H; a), 6.83 (s,
2H; a), 5.10 (s, 2H; OCH2), 3.53 (m, 4H; CH2N), 3.12 (m, 2H; i), 2.18 (m,
4H; CH2), 2.10 (s, 6H; m), 2.05 (s, 6H; m), 1.17 (s, 9H; e), 1.02 (d, 12H; j);
13C NMR (62.5 MHz, [D6]DMSO, 218C): d ˆ 176.5, 166.3, 165.3, 155.3,
147.7, 147.0, 146.5, 135.7, 135.6, 133.6, 133.0, 132.9, 131.1, 129.3, 128.9, 128.4,
128.2, 127.7, 126.9, 126.6, 126.1, 122.0, 66.7, 45.1, 36.0, 29.1, 28.2, 26.9, 24.6,
B1IB1-NH2, which was obtained as
a white powder (3.20 g, 67%).
M.p. 231 233 C; 1H NMR (250 MHz, [D6]DMSO, 218C): d ˆ 9.83 (s,
2H; n), 8.53 (s, 1H; s), 8.32 (d, 2H; d), 7.68 (t, 1H; t), 7.32 (m, 10H; ArCH),
7.05 (s, 4H; a), 6.80 (s, 4H; a), 5.08 (s, 4H; OCH2), 4.38 (s, 4H; NH2), 3.50
(m, 8H; CH2N), 2.30 (m, 8H; CH2), 2,16 (s, 12H; m), 2.08 (s, 12H; m);
13C NMR (62.5 MHz, CDCl3, 218C): d ˆ 165.3, 154.7, 146.6, 142.0, 137.2,
135.2, 135.1, 133.1, 132.7, 130.2, 128.5, 127.8, 127.5, 127.0, 126.3, 125.9, 125.0,
120.7, 67.1, 66.2, 42.8, 35.4, 18.5, 18.2; MS (‡ve, FAB) m/z (%): 1047 (100)
‡
[M ‡H].
Synthesis of XB1IB1X: H2N-B1IB1-NH2 (2.97 g, 2.85 mmol) and Et3N
(0.85 mL, 6.27 mmol) were dissolved in CH2Cl2 (25 mL) and added
dropwise to a solution of trimethylacetyl chloride (0.7 mL, 5.70 mmol) in
CH2Cl2 (25 mL). The mixture was stirred for 16 h and then washed with
aqueous HCl (1m, 5 Â 100 mL), aqueous NaOH (1m, 5 Â 100 mL), water
(200 mL) and brine (50 mL). The organic layer was dried over anhydrous
Na2SO4, and the solvent was removed under reduced pressure. Flash
chromatography on silica eluting with CH2Cl2/EtOH (97:3, v/v) yielded the
desired product (3.10 g, 90% yield). M.p. 206 2078C; 1H NMR (250 MHz,
CDCl3, 218C): d ˆ 8.70 (s, 2H; n), 8.46 (s, 1H; s), 7.83 (s, 2H; n), 7.79 (d,
2H; d), 7.37 (m, 10H; ArCH), 7.05 (s, 4H; a), 6.95 (s, 4H; a), 6.50 (t, 1H; t),
5.08 (s, 4H; OCH2), 3.50 (m, 8H; CH2N) 2.25 (m, 8H; CH2), 2.13 (s, 24H;
m), 1.26 (s, 18H; e); 13C NMR (62.5 MHz, CDCl3, 218C): d ˆ 176.8, 165.1,
147.2, 146.8, 135.0, 134.9, 133.6, 131.6, 131.5, 130.8, 128.8, 128.2, 127.9, 126.6,
125.6, 66.5, 45.1, 41.4,. 38.9, 35.9, 27.6, 18.6, 18.4; MS (‡ve, FAB) m/z (%):
‡
24.0, 23.3, 18.7, 18.6; HRMS (FAB) m/z ˆ 849.4907, C54H65N4O5 requires
849.4955.
Synthesis of AIB1T: AIB1T was prepared in the same way as AIB1X,
1
À
À
starting from TB NH2 (0.67 g, 1.08 mmol) and AI Cl (0.47 g, 1.35 mmol).
AIB1T was isolated as
a white powder (0.97 g, 89% yield) after
chromatography (eluant CH2Cl2/EtOH 99.5:0.5, v/v). M.p. 197 1998C;
1H NMR (250 MHz, [D6]DMSO, 218C): d ˆ 9.85 (s, 1H; n), 9.72 (s, 1H n),
9.55 (s, 1H; n), 8.48 (s, 1H; s), 8.16 (d, 1H; d), 8.08 (d, 1H; d), 7.87 (d, 2H;
f), 7.66 (t, 2H; t), 7.49 (d, 2H; g), 7.38 (m, 5H; ArCH), 7.31 (t, 1H; l), 7.20 (d,
2H; k), 7.09 (s, 4H; a), 5.07 (s, 2H; OCH2), 3.55 (m, 4H; CH2N), 3.10 (m,
2H; i), 2.22 (m, 4H; CH2), 2.13 (s, 6H; m), 2.08 (s, 6H; m), 1.33 (s, 9H; o),
1.20 (d, 12H; j); 13C NMR (62.5 MHz, [D6]DMSO, 218C): d ˆ 167.5, 166.9,
164.3, 155.8, 154.7, 147.2, 146.9, 146.6, 136.9, 135.7, 135.6, 135.2, 134.8, 134.6,
133.7, 132.3, 132.1, 130.9, 128.9, 128.0, 127.5, 126.1, 124.2, 122.8, 45.7, 36.2,
31.5, 28.7, 26.0, 24.4, 24.0, 22.7, 18.7; HRMS (FAB) m/z ˆ 925.5229,
‡
1213 (100) [M ‡H]; elemental analysis calcd (%) for C76H88N6O8: C 75.22,
H 7.31, N 6.92; found: C 75.92, H 7.41, N 6.80.
Synthesis of TB1IB1T: TB1IB1T was synthesised and purified in the same
way as XB1IB1X, by using 4-tert-butylbenzoyl chloride (0.69 mL,
3.46 mmol), and H2N-B1IB1-NH2 (1.2 g, 1.16 mmol). TB1IB1Twas obtained
‡
C60H69N4O5 requires 925.5268.
1
1
Synthesis of HIB X: XB NH2 (0.70 g, 1.30 mmol) and Et3N (0.22 mL,
1.56 mmol) were dissolved in CH2Cl2 (100 mL) and added dropwise to a
solution of isophthaloyl dichloride (5.28 g, 26 mmol) in CH2Cl2 (200 mL) at
58C. The solution was then stirred for 2 h at room temperature. After this
as a
white powder (1.20 g, 76% yield). M.p. 250 2528C; 1H NMR
À
(250 MHz, [D6]DMSO, 218C): d ˆ 9.82 (s, 2H; n), 9.58 (s, 2H; n), 8.50 (s,
1H; s), 8.25 (d, 2H; d), 7.91 (d, 4H; f), 7.72 (t, 1H; t), 7.52 (d, 4H; g), 7.33
(m, 10H; ArCH), 7.12 (s, 8H; a), 5.09 (s, 4H; OCH2), 3.44 (m, 8H; CH2N),
5444
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5444 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 23

Quantifying Cooperativity
5435 5446
p
2.36 (m, 8H; CH2), 2.15 (s,
12H; m), 2.11 (s, 12H; m),
1.30 (s, 18H; o); 13C NMR
(62.5 MHz, CDCl3, 218C):
d ˆ 165.7, 155.1, 147.4, 147.1,
135.1, 133.65, 131.6, 131.4,
128.9, 128.1, 127.8, 127.3,
1 ‡ 4KdGꢀ‰GŠ0 À ‰HGŠ† À 1 ‡ 8KdGꢀ‰GŠ0 À ‰HGŠ†
8KdG
[GG] ˆ
(8)
(9)
q
2
1 ‡ Kꢀ‰GŠ0 À ‰GGŠ†ꢀ‰HŠ0 À ‰HHŠ† À ꢀ1 ‡ Kꢀ‰GŠ0 À ‰GGŠ†ꢀ‰HŠ0 À ‰HHŠ†† À 4K2ꢀ‰GŠ0 À ‰GGŠ†ꢀ‰HŠ0 À ‰HHŠ†
[HG] ˆ
2K
[H] ˆ [H]0 À 2[HH] À [HG]
(10)
(11)
127.0, 125.9, 66.4, 41.2, 35.7, 34.9, 31.1, 18.7, 18.5. MS (‡ve, FAB) m/z
‡
(%): 1363 (100) [M ‡H]; elemental analysis calcd (%) for C88H96N6O8 ¥
[G] ˆ [G]0 À 2[GG] À [HG]
2H2O: C 75.40, H 7.19, N 6.00; found: C 75.69, H 7.17, N 5.84.
Synthesis of AIB2IA: H2N-B2-NH2 (2.54 g, 2.38 mmol) and Et3N (1.00 mL,
7.14 mmol) dissolved in CH2Cl2 (25 mL) were added dropwise to a solution
‰HGŠ
‰HHŠ
‰HŠ
‰HŠ
dobs
ˆ
db ‡
dd ‡
df
(12)
‰HŠ
‰HŠ
0
0
0
À
of freshly prepared AI Cl (2.46 g, 7.14 mmol) in CH2Cl2 (25 mL) at 58C.
in which [HH] is the concentration of host dimer, [GG] is the concentration
of guest dimer, KdG is the guest dimerisation constant, KdH is the host
dimerisation constant and dd is the limiting bound chemical shift of the host
dimer.[41]
The solution was then stirred for 16 h at room temperature. The solvent was
removed under reduced pressure. Flash chromatography by using
gradient elution with a mixture CH2Cl2/THF (98:2 to 90:10, v/v) yielded
AIB2IA as white solid (0.80 g, 20% yield). 1H NMR (250 MHz,
a
a
[D6]DMSO, 218C): d ˆ 8.82 (s, 2H; n), 8.67 (s, 2H; n), 8.54 (s, 2H; s),
8.11 (d, 4H; d), 7.45 (t, 2H; t), 7.22 (t, 2H; l), 7.08 (d, 4H; k), 6.91 (s, 4H; a),
6.45 (s, 2H; ArCH), 3.85 (m, 6H; OCH2), 3.71 (m, 4H; CH2N), 3.06 (q, 4H;
i), 2.57 (m, 4H; CH2), 2.17 (s, 12H; m), 1.65 (m, 6H; OCH2CH2), 1.36 (m,
6H; OCH2CH2CH2), 1.12 (m, 84H; CH2/j), 0.76 (t, 9H; CH3); 13C NMR
(62.5 MHz, CDCl3, 218C): d ˆ 170.4, 166.1, 165.4, 153.2, 146.5, 145.0, 139.2,
136.3, 134.0, 133.8, 132.5, 131.2, 131.0, 130.8, 129.0, 128.6, 126.5, 125.9, 123.6,
105.4, 73.5, 69.2, 43.8, 31.9, 30.3, 29.7 29.4 (CH2 solubilising group), 28.9,
All experiments were performed at least twice. The association constant for
a single run was calculated as the mean of the values obtained for each of
the signals followed during the titration weighted by the observed changes
in chemical shift. The association constants from different runs were then
averaged. Errors are quoted at the 95% confidence limits (twice the
standard error). For a single run, the standard error was determined by
using the standard deviation of the different association constants
determined by following different signals.
‡
26.1, 23.6, 22.7, 18.9, 14.1; HRMS (FAB) m/z ˆ 1679.2267, C110H160N5O8
requires 1679.2296.
Synthesis of HIB2ICH: H2N-B2-NH2 (1.5 g, 1.41 mmol) and Et3N (0.40 mL,
2.82 mmol) were dissolved in CH2Cl2 (250 mL) and added dropwise to a
solution of isophthaloyl dichloride (11.45 g, 56.4 mmol) in CH2Cl2 (100 mL)
at 58C. The solution was then stirred for 2 h at room temperature. After this
period, the resulting solution was added dropwise to a solution of n-
hexylamine (11.40 g, 112.8 mmol) and Et3N (15.82 mL, 112.80 mmol) in
CH2Cl2 (300 mL) at 58C and then stirred for 16 h at room temperature. The
solution was then washed with aqueous HCl (1m, 4 Â 300 mL), aqueous
NaOH (1m, 2 Â 200 mL), and brine (2 Â 75 mL). The organic phase was
dried over anhydrous Na2SO4, and concentrated under reduced pressure.
Flash column chromatography on silica by using a gradient elution with
CHCl3/EtOH (100:0 to 97.5:2.5, v/v) yielded HIH (17.5 g, 53 mmol) and the
desired HIB2IH, which was obtained as a white solid. (0.42 g, 22% yield).
M.p. 53 568C; 1H NMR (250 MHz, CDCl3 218C): d ˆ 8.35 (s, 2H; s), 8.00
(s, 2 H; n), 7.80 (d, 2H; d), 7.73 (d, 2H; d), 7.23 (t, 2H; t), 6.98 (s, 4H; a), 6.67
(t, 2H; n), 6.52 (s, 2H; ArCH),), 3.96 (t, 6H; OCH2), 3.42 (m, 4H; CH2N),
3.35 (q, 4H; h), 2.32 (m, 4H; CH2), 2.18 (s, 12H; m), 1.75 1.50 (m, 10H;
CH2), 1.50 1.30 (m, 10H; CH2), 1.25 (m, 68H; CH2), 0.86 (t, 15H; CH3);
13C NMR (62.5 MHz, [D6]DMSO, 218C): d ˆ 166.8, 165.7, 155.5, 145.2,
136.7, 135.9, 134.6, 133.9, 132.2, 130.4, 128.5, 128.0, 127.8, 126.3, 125.9, 67.1,
43.2, 40.2, 31.4, 32.9, 32.6, 22.5, 18.7, 14.0; MS (‡ve, FAB) m/z (%): 1527
Acknowledgements
We thank the BBSRC (S.T.), Roche Discovery Welwyn (P.T.), the
University of Sheffield (P.T.) and the Lister Institute (C.A.H.) for funding.
[1] D. H. Williams, M. S. Westwell, Chem. Soc. Rev. 1998, 27, 57 63.
[2] S. Albeck, R. Unger, G. Schreiber, J. Mol. Biol. 2000, 298, 503 520.
[3] S. R. Griffiths-Jones, M. S. Searle, J. Am. Chem. Soc. 2000, 122, 8350
8356.
[4] D. H. Williams, A. J. Maguire, W. Tsuzuki, M. S. Westwell, Science
1998, 280, 711 714.
[5] G. Schreiber, A. R. Fersht, J. Mol. Biol. 1997, 270, 111 122.
[6] J. P. Mackay, U. Gerhard, D. A. Beauregard, R. A. Maplestone, D. H.
Williams, J. Am. Chem. Soc. 1994, 116, 4573 4580.
[7] A. Horovitz, A. R. Fersht, J. Mol. Biol. 1992, 224, 733 740.
[8] A. Horovitz, A. R. Fersht, J. Mol. Biol. 1990, 214, 613 617.
[9] D. A. Bell, S. G. Diaz, V. M. Lynch, E. V. Anslyn, Tetrahedron Lett.
1995, 26, 4155 4158.
[10] N. Colocci, P. B. Dervan, J. Am. Chem. Soc. 1995, 117, 4781 4787.
[11] N. Colocci, M. D. Distefano, P. B. Dervan, J. Am. Chem. Soc. 1993,
115, 4468 4473.
‡
(100) [M ‡H]; elemental analysis calcd (%) for C88H96N6O8: C 77.07, H
9.97, N 4.59; found: C 76.70, H 10.03, N 4.39.
1H NMR binding studies: The procedures used for the dilution and titration
experiments have been described previously.[38] The association constants
for the longer zipper complexes were determined by dilution of 1:1
mixtures, and the data was analysed by using purpose-written software on
an Apple Macintosh microcomputer, NMRDil HG HH GG. This pro-
gram requires a previous determination of the dimerisation parameters
(Kd, dd and df) for the two components and fits the data to a 1:1 binding
isotherm, taking into account the dimerisation equilibria for both the host
and guest. The method starts by assuming that [HG] ˆ 0, so that
Equations (7) and (8) can be solved exactly for [HH] and [GG]. These
values of are then used to solve Equation (9) for [HG]. Equations (10) and
(11) give the concentrations of free host [H] and free guest [G]. At this
point, [H] ‡ [HH] ‡ [HG]=[H]0, and [G] ‡ [GG] ‡ [HG]=[G]0, so the
value of [HG] from Equation (9) is used in Equations (7) and (8) to re-
evaluate [HH] and [GG], and the procedure is carried out repetitively until
[H] ‡ [HH] ‡ [HG] ˆ [H]0, and [G] ‡ [GG] ‡ [HG] ˆ [G]0. This allows the
set of simultaneous equations to be solved for the concentrations of all
species present.
[12] M. D. Distefano, P. B. Dervan, Proc. Natl. Acad. Sci. USA 1993, 90,
1179 1183.
[13] C. A. Hunter, D. H. Purvis, Angew. Chem. 1992, 104, 779 782;
Angew. Chem. Int. Ed. Engl. 1992, 31, 792 795.
[14] J. S. Lyndsey, New J. Chem. 1991, 15, 153 180.
[15] K. Onan, J. Rebek, T. Costello, L. Marshall, J. Am. Chem. Soc. 1983,
105, 6759 6760.
[16] R. B. Prince, J. G. Saven, P. G. Wolynes, G. S. Moore, J. Am. Chem.
Soc. 1999, 121, 3114 3121.
[17] P. Groves, M. S. Searle, M. S. Westwell, D. H. Williams, J. Chem. Soc.
Chem. Commun. 1994, 1519 1520.
[18] A. P. Bisson, C. A. Hunter, Chem. Commun. 1996, 1723 1724.
[19] A. Pfeil, J.-M. Lehn, J. Chem. Soc. Chem. Commun. 1992, 838 840.
[20] M. S. Searle, G. J. Sharman, P. Groves, B. Benhamu, D. A. Beaure-
gard, M. S. Westwell, R. J. Dancer, A. J. Maguire, A. C. Try, D. H.
Williams, J. Chem. Soc. Perkin Trans. 1 1996, 2781 2786.
[21] M. S. Searle, D. H. Williams, J. Am. Chem. Soc. 1992, 114, 10690
10697.
p
[22] W. P. Jenks, Proc. Nat. Acad. Sci. USA 1981, 78, 4046.
[23] W. P. Jenks, Adv. Enzymol. 1975, 43, 219 410.
[24] M. I. Page, Chem. Soc. Rev. 1973, 2, 295.
1 ‡ 4KdHꢀ‰HŠ0 À ‰HGŠ† À 1 ‡ 8KdHꢀ‰HŠ0 À ‰HGŠ†
[HH] ˆ
(7)
8KdH
Chem. Eur. J. 2002, 8, No. 23
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5445 $ 20.00+.50/0
5445

C. A. Hunter et al.
FULL PAPER
[25] M. S. Searle, M. S. Wetswell, D. H. Williams, J. Chem. Soc. Perkin
Trans. 2 1995, 2, 141.
[36] Y. Kato, M. M. Conn, J. Rebek, J. Am. Chem. Soc. 1994, 116, 3279
3284.
[26] M. S. Westwell, M. S. Searle, J. Klein, D. H. Williams, J. Phys. Chem.
1996, 100, 16000 16001.
[27] D. H. Williams, T. F. Gale, B. Bardsley, J. Chem. Soc. Perkin Trans. 2
1999, 1331 1334.
[28] M. S. Searle, S. R. Griffiths-Jones, H. Skinner-Smith, J. Am. Chem.
Soc. 1999, 121, 11615.
[29] P. J. Carter, G. Winter, A. J. Wilkinson, A. R. Fersht, Cell 1984, 38,
835 840.
[30] G. K. Ackers, F. R. Smith, Annu. Rev. Biochem. 1985, 54, 597 629.
[31] L. Serrano, M. Bycroft, A. R. Fersht, J. Mol. Biol. 1991, 218, 465 475.
[32] G. Schreiber, A. R. Fersht, J. Mol. Biol. 1995, 248, 478 486.
[33] Y. Aoyama, M. Asakawa, H. Yamagishi, H. Toi, H. Ogoshi, J. Am.
Chem. Soc. 1990, 112, 3145 3151.
[37] A. P. Bisson, F. J. Carver, D. S. Eggleston, R. C. Haltiwanger, C. A.
Hunter, D. L. Livingstone, J. F. McCabe, C. Rotger, A. E. Rowan, J.
Am. Chem. Soc. 2000, 122, 8856 8868.
[38] F. J. Carver, C. A. Hunter, P. S. Jones, D. J. Livingstone, J. F. McCabe,
E. M. Seward, P. Tiger, Chem. Eur. J. 2001, 7, 4854 4862.
[39] The magnitude of the aromatic interactions in this system are small.
The experiments are carried out in chloroform, and desolvation
presumably reduces the value from that expected from gas phase
calculations[40] and experiments in water, where the hydrophobic
effect makes an important contribution.[31] In addition, the interaction
is sensitive to the nature of the substituents and the precise geo-
metrical arrangement.[38]
[40] S. Tsuzuki, K. Honda, T. Uchimaru, M. Mikami, K. Tanabe, J. Am.
Chem. Soc. 2002, 124, 104 112.
[41] The curve fitting programme is available from the author on request.
[34] Y. Aoyama, M. Asakawa, Y. Matsui, H. Ogoshi, J. Am. Chem. Soc.
1991, 113, 6233 6240.
[35] H. Adams, F. J. Carver, C. A. Hunter, J. C. Morales, E. M. Seward,
Angew. Chem. 1996, 108, 1628 1631; Angew. Chem. Int. Ed. Engl.
1996, 35, 1542 1544.
Received: January 11, 2002
Revised: July 16, 2002 [F3790]
5446
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/02/0823-5446 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 23