Synthesis and binding properties of a macrobicyclic receptor for N-protected peptides with a carboxylic acid terminus

Article abstract of DOI:10.1039/a908090b

A novel macrobicyclic receptor, 3, has been synthesised by linking together a diaminopyridine with suitable amino acids, followed by a double intramolecular cyclisation of a suitably activated precursor. Macrobicycle 3 features a diamidopyridirie unit, designed to serve as a specific binding site for carboxylic acid functionality, at the base of an open, bowl-shaped cavity. Incorporation of additional amide functionality around the rim of the bowl-shaped structure provides further hydrogen bonding sites to interact with pcptidic guests. The binding properties of 3 with N-protected amino acid and peptide derivatives have been investigated by NMR titration experiments, which reveal that 3 is a strong and selective receptor for peptides with a carboxylic acid terminus in CDCl₃ solution, the strongest binding being observed with Cbz-β-alanyl-D-alanine (-ΔG_ass = 22.8 kJ mol^-1). The macrobicycle is reasonably enantioselective (Cbz-β-alanyl-L-alanine, -ΔG_ass = 19.1 kJ mol^-1) and notably the binding of Cbz-β-alanyl lactic acids is considerably weaker than the binding of the corresponding Cbz-β-alanyl alanines (ΔΔG_ass ～ 8-9 kJ mol^-1). Molecular modelling and 2D NMR studies have been carried out on the free macrobicycle and the 1:1 complex formed with the most strongly bound substrate (Cbz-β-alanyl-D-alanine). These studies provide a consistent picture of the macrobicycle as a flexible receptor, which is able to bind the Cbz-β-alanyl-D-alanine substrate in the macrobicyclic cavity with a series of well defined hydrogen bonds to the alanylalanine amide, and less well defined hydrogen bonds to the benzylcarbamate functionality. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/a908090b

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis and binding properties of a macrobicyclic receptor for
N-protected peptides with a carboxylic acid terminus
Peter D. Henley,a Christopher P. Waymark,a Iain Gilliesb and Jeremy D. Kilburn*a
1
a Department of Chemistry, University of Southampton, Southampton, UK SO17 1BJ
b Chemical Development Laboratory, Glaxo–Wellcome Ltd, Temple Hill, Dartford,
UK DA1 5AH
Received (in Cambridge, UK) 8th October 1999, Accepted 17th January 2000
A novel macrobicyclic receptor, 3, has been synthesised by linking together a diaminopyridine with suitable amino
acids, followed by a double intramolecular cyclisation of a suitably activated precursor. Macrobicycle 3 features a
diamidopyridine unit, designed to serve as a specific binding site for carboxylic acid functionality, at the base of
an open, bowl-shaped cavity. Incorporation of additional amide functionality around the rim of the bowl-shaped
structure provides further hydrogen bonding sites to interact with peptidic guests. The binding properties of 3 with
N-protected amino acid and peptide derivatives have been investigated by NMR titration experiments, which reveal
that 3 is a strong and selective receptor for peptides with a carboxylic acid terminus in CDCl3 solution, the strongest
binding being observed with Cbz-β-alanyl--alanine (Ϫ∆Gass = 22.8 kJ molϪ1). The macrobicycle is reasonably
enantioselective (Cbz-β-alanyl--alanine, Ϫ∆Gass = 19.1 kJ molϪ1) and notably the binding of Cbz-β-alanyl lactic
acids is considerably weaker than the binding of the corresponding Cbz-β-alanyl alanines (∆∆Gass ~ 8–9 kJ molϪ1).
Molecular modelling and 2D NMR studies have been carried out on the free macrobicycle and the 1:1 complex
formed with the most strongly bound substrate (Cbz-β-alanyl--alanine). These studies provide a consistent
picture of the macrobicycle as a flexible receptor, which is able to bind the Cbz-β-alanyl--alanine substrate in the
macrobicyclic cavity with a series of well defined hydrogen bonds to the alanylalanine amide, and less well defined
hydrogen bonds to the benzylcarbamate functionality.
to have a more 3-dimensional architecture, and which feature a
Introduction
specific binding site for carboxylic acid functionality at the base
The development of synthetic receptors for peptides and amino
of an open, bowl-shaped cavity. By incorporating additional
acid derivatives is of considerable interest because the inter-
amide functionality around the rim of such a bowl-shaped
molecular interactions involved in small molecule–peptide
structure we intended to provide further hydrogen bonding
complexes are of direct relevance to many biological peptide–
sites suitably preorganised to interact with guests, such as
protein interactions, and may also lead to new bio-sensors,
amino acid derivatives, bound within the macrobicyclic
therapeutics and catalysts for peptide hydrolysis.1 We have
focused on developing receptors specifically for peptides with
a free carboxylic acid terminus, by incorporating a carboxylic
acid binding site into macrocyclic structures with additional
cavity. Thus, for example, we have described macrobicycle 23 in
which a carboxylate binding site is provided by a thiourea
moiety.
Macrobicycle 2 binds a range of N-acetyl amino acids (as
tetrabutylammonium carboxylates), with N-Ac--amino acid
substrates binding through a carboxylate–thiourea interaction
on the outside of the cavity, while N-Ac--amino acid sub-
hydrogen bonding functionality to bind to the backbone of
peptidic guests. We have described one such macrocyclic struc-
ture 12 which utilises a diamidopyridine unit to serve as a carb-
strates are bound in the cavity of the macrobicyclic receptor,
but with the N-acetyl unit in a cis amide configuration. In order
to bind larger substrates than the derivatised amino acids
accommodated by 2, we wished to prepare macrobicycle 3 with
a deeper cavity and, in addition, to investigate the use of a
diamidopyridine moiety to serve as a carboxylic acid (as
opposed to carboxylate) binding site in this type of system. In
order to deepen the cavity we chose to link the rim of the
macrobicycle, formed from the same biaryl unit used previously
in 2, through glutamic acid and phenylalanine residues, to a
diamidopyridine unit at the base of the cavity. We chose to
incorporate phenylalanine into the structure to increase the
amount of chiral information in the receptor (which might
therefore lead to greater enantioselective binding properties) to
aid the solubility of the receptor in non-polar solvents such as
oxylic acid binding site in a cavity lined on one side with amide
functionality, and on the other by a rigid biaryl unit. This
macrocycle selectively binds to amino acid and peptide deriv-
atives with a carboxylic acid terminus, in chloroform solution.
As an alternative approach to recognition of such peptidic
substrates we have prepared a range of macrobicycles designed
chloroform, and, assuming that the benzyl residues would
prefer to be on the outside of the macrobicycle, this should aid
in the preorganisation of the cavity. In this paper we describe
the synthesis of macrobicycle 3, which indeed shows strong
and selective binding to amino acid and dipeptide derivatives in
chloroform solution.4
DOI: 10.1039/a908090b
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031
This journal is © The Royal Society of Chemistry 2000
1021

View Article Online
Synthesis
The synthesis of macrobicycle 3 follows the same strategy as for
the synthesis of 2 and related macrobicycles,3,5 with a double
intramolecular cyclisation of a suitably activated acyclic pre-
cursor 4 as the key step (Scheme 1). Assembly of the cyclisation
Scheme 2
used previously in receptor 2, and so in our first attempts at the
synthesis of 3 the previously described amino methyl ester 73a
was coupled with tert-butoxycarbonyl--glutamic acid γ-allyl
ester in 67% yield, followed by removal of the allyl protecting
Scheme 1
precursor then requires coupling together of the various amino
acid building blocks around the carboxylic acid binding site, in
this case diaminopyridine.
Coupling of diaminopyridine to activated carboxylic acid
derivatives to generate diamidopyridines can be complicated by
the low reactivity of the amino groups and, in our case, the
normal use of acid chlorides as the coupling partner was pre-
cluded if racemisation of the phenylalanine was to be avoided.
We found that tert-butoxycarbonylphenylalanine could be
coupled to diaminopyridine in ~50% yield, using the coupling
agent benzotriazol-1-yloxytripyrrolidinophosphonium hexa-
fluorophosphate (PyBop), and carrying out the reaction at
50 ЊC (Scheme 2). Alternatively stirring diaminopyridine with
an excess of Boc-phenylalanine and dicyclohexylcarbodiimide,
as the coupling agent, in DMF for 72 hours gave the desired bis
adduct in 85% yield.6 Treatment of 5 with TFA then gave the
bis-TFA salt 6. Macrobicycle 3 uses the same biarylmethane
fragment as the rigidifying unit in the rim of the structure as
Scheme 3
group, using Pd(0) catalysis, to give acid 9 (Scheme 3).7 Two
equivalents of 9 were then coupled to the bis-TFA salt 6 to give
the protected macrocyclisation precursor 10. However, all
attempts at the selective cleavage of the methyl esters of 10 were
unsuccessful. Thus alkaline hydrolysis led to competitive cleav-
age of the amide bonds of the amidopyridine unit, giving diacid
1022
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031

View Article Online
1
fragment 11 (stereochemistry unproven) in good yield, and
treatment of diester 10 with alkyl thiolates,8 iodide9 or cyanide
anion10 similarly failed to give the desired diacid. We therefore
prepared the biarylmethane unit as the benzyl ester 13 by a
Suzuki coupling11 of benzyl 4-bromomethylbenzoate 12 with
4-cyanophenylboronic acid3 and reduction of the resulting
nitrile to give the amine 14 (Scheme 4). Coupling of 14 with
which could be unambiguously assigned. The H NMR spec-
trum also indicated that macrobicycle 3 was formed as a single
diastereoisomer, which ruled out the possibility that racemis-
ation of any of the chiral centres might have occurred during
the synthesis. Earlier work on a related, but smaller, macro-
bicycle 215 had shown that in solution this macrobicycle had
a preference for a conformation with the amidopyridine unit
inverted inside the rim of the macrobicycle so that the amido-
pyridine (and hence the carboxylic acid binding site) was point-
ing away from the intended binding cavity, which was clearly
unfavourable for binding of amino acids or peptidic guests. For
macrobicycle 3, a 2D ROESY12 spectrum revealed no NOE’s
from the pyridine ring protons to the rim of the macrobicycle,
suggesting that the pyridine ring is not inverted into the macro-
bicyclic cavity in this case. This notion was supported by
detailed molecular modelling studies on the macrobicycle (vide
infra). The 1H NMR spectrum of 3 also revealed that the signal
for the carboxamide proton NH5 (for the atom labelling used,
see Fig. 2) occurs at 8.16 ppm in CDCl3 solution. In com-
parison the other amide protons in macrobicycle 3, NH3 and
NH7, resonate at 6.37 and 6.47 ppm respectively in CDCl3 solu-
tion. Furthermore, for the related macrobicyclic structure 22,
incorporating an identical biaryl rim portion, the carboxamide
proton resonates at 6.42 ppm in CDCl3 solution.13 This strongly
suggests that for macrobicycle 3 there is an intramolecular
hydrogen bond to NH5, and this is borne out by the molecular
modelling results for 3 (vide infra) which showed structures in
which a consistent feature was the presence of intramolecular
Scheme 4
tert-butoxycarbonyl--glutamic acid γ-allyl ester, followed by
removal of the allyl protecting group now gave acid 16, which in
turn was coupled to the bis-TFA salt 6 to give the protected
macrocyclisation precursor 17. Now debenzylation to give the
diacid 18 was successfully accomplished using Pd/C with
ammonium formate in essentially quantitative yield. Diacid 18
was converted to the bis(pentafluorophenyl ester) 19 and the
amine protecting groups were removed with trifluoroacetic
acid. Finally, slow addition of the resulting bis(trifluoroacetate
salt) 20 to a solution of diisopropylethylamine, in acetonitrile,
gave the desired macrobicycle 3 in ~30% overall yield from 18.
Macrobicycle 3 is soluble in a range of organic solvents,
including chloroform, and gave well resolved NMR spectra
hydrogen bonds, particularly from NH5 to C᎐O4. In the case of
᎐
macrobicycle 22, such an intramolecular hydrogen bond is not
possible with the rigid aromatic ring, which links the biaryl rim
to the diamidopyridine unit.
Binding studies
Binding studies with macrobicycle 3 were carried out with a
series of substrates in deuteriochloroform, using a standard
NMR titration experiment, by monitoring the shift of
the amidopyridine NH signal (H1) and analysing the result-
ing binding curves using the Hostest software.14 Examples of
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031
1023

View Article Online
Table 1 Binding constants (Kass) and free energies of complexation
(Ϫ∆Gass) for the 1:1 complexes formed between macrobicycle 3 and
various substrates in CDCl3 at 20 ЊC
a
b
Kass
MϪ1
/
Ϫ∆Gass /
Entry
Substrate
kJ molϪ1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
PhCH2CO2H
CH3(CH2)4CO2H
Cbz-Gly-OH
Cbz--Ala-OH
Cbz--Ala-OH
Boc--Phe-OH
Boc--Phe-OH
Cbz--Phe-OH
Ac--Phe-OH
Boc--Val-OH
Boc--Val-OH
Boc--Ser-OH
Boc--Ser-OH
350
470
800
620
1370
400
14.2
15.0
16.3
15.6
17.6
14.6
15.8
17.6
19.5
12.6
10.7
17.6
14.8
16.5
13.4
13.6
19.2
15.4
14.6
17.7
19.1
22.8
11.3
13.2
660
1370
2970
180
85
1360
450
870
280
300
2670
570
Cbz-β-Ala-OH
Cbz-Gly--Phe-OH
Cbz--Phe-Gly-OH
Cbz-Gly--Ser-OH
Cbz-Gly--Ser-OH
Cbz--Ala--Ala-OH
Cbz--Ala--Ala-OH
Cbz-β-Ala--Ala-OH
Cbz-β-Ala--Ala-OH
Cbz-β-Ala--Lac-OH
Cbz-β-Ala--Lac-OH
410
1420
2220
12200
110
260
a Errors for Kass were estimated as <10%. b Errors for Ϫ∆Gass were
estimated as <0.3 kJ molϪ1. Errors were estimated from the quality of
the fit of the experimental to the calculated data and by carrying out
several titration experiments to obtain several estimates of Kass and
averaging the values obtained.
each of the substrates the largest shift in the 1H NMR spectrum
was for the amidopyridine NH1 signal (for the atom labelling
used, see Fig. 2), which was consistent with our expectation that
the primary binding interaction would be between the carb-
oxylic acid terminus of the guests and the diamidopyridine of
the host. Other shifts in the 1H NMR spectrum of 3 were noted
on addition of most substrates and, in particular, we recorded
the shift of all the macrobicycle amide NH’s in the titration
experiments, since some tentative inferences may be drawn con-
cerning hydrogen bonding interactions in the complexes from
the magnitude of these shifts.15 Small shifts in the aromatic
signals for the pyridine and phenylalanine rings were also
consistently observed, presumably reflecting conformational
changes of the macrobicycle on binding, although the shifts of
these aromatic protons overlapped with other signals in the
course of the titration, or were not generally large enough (<0.1
ppm), to provide additional data for determination of associ-
ation constants.16
Fig. 1 Binding isotherms showing the shift of NH1 (ppm) of macro-
bicycle 3, in CDCl3 (3.11 × 10Ϫ3 M) against guest concentration (M)
on titration with (i) Cbz-β-alanyl--alanine, and (ii) Cbz-β-alanyl--
alanine. The ×’s represent the experimental data points and —— the
calculated fit.
Titration of macrobicycle 3 with phenylacetic acid (Table 1,
entry 1) gave Ϫ∆Gass = 14.2 kJ molϪ1 with a limiting downfield
1
shift of the amidopyridine NH1 signal in the H NMR of 0.45
ppm, but with no significant changes to the rest of the signals
for 3. Similarly, titration with hexanoic acid (entry 2) gave
Ϫ∆Gass = 15.0 kJ molϪ1 with a limiting downfield shift of the
amidopyridine NH1 signal in the 1H NMR of 0.28 ppm. Again,
these results are consistent with the anticipated binding of the
carboxylic acid functionality by the amidopyridine unit, and
give a good indication of the strength of this interaction (~14–
15 kJ molϪ1), in this system, against which to compare the bind-
ing of the other substrates.17 Binding of simple amino acid
derivatives (entries 3–14) generally showed improved binding
over phenylacetic acid and hexanoic acid, with correspondingly
larger shifts of the amidopyridine NH1 signal. In addition,
binding of these substrates resulted in downfield shifts of the
benzylic amide proton H7. For alanine and phenylalanine sub-
strates (entries 4–9) there is a small preference for the -amino
Fig. 2 Atom labelling for macrobicycle 3 and guest Cbz-β-alanyl-
-alanine, and shifts (ppm) for the various protons indicated, on
formation of the 1:1 complex between the two, relative to the signals
for uncomplexed materials.
titration isotherms for the most strongly bound guest Cbz-β-
alanyl--alanine and its enantiomer Cbz-β-alanyl--alanine
are shown in Fig. 1. A number of simple carboxylic acids,
α- and β-amino acid derivatives, and dipeptides were studied to
allow a thorough investigation of how stereochemistry, relative
disposition of functionality and the steric bulk of side chains
affected the binding properties of the receptor (Table 1). In each
binding experiment a 1:1 binding stoichiometry has been
assumed which was supported by the good fit of the measured
data to the theoretical model, on analysis. On titration with
1024
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031

View Article Online
acid, but there is a reversal of this selectivity for the valine and
serine substrates (entries 10–13). For the -phenylalanine sub-
strates we investigated the effect of varying the N-protecting
group (tBoc, Cbz, Ac). The binding of these substrates
increased significantly with decreasing size of the N-protecting
group (entries 7–9) which may be attributed to a binding inter-
action between the N-protecting group functionality and the
macrobicycle, which is weakened with the increasing steric
demands of the protecting group (although the change from
carbamate (tBoc, Cbz) to amide (Ac) may also influence the
relative binding of these three substrates). In comparison to the
tBoc phenylalanine substrates, the poor binding of the tBoc
valine substrates (entries 10, 11) may be attributed to steric
demands of the side-chain as well as of the bulky tBoc group,
while the tBoc serine substrates (entries 12, 13) showed substan-
tially better binding, indicating a positive interaction between
the hydroxymethyl moiety of the substrate and the receptor.
Binding of the tBoc serine substrates was accompanied by a
substantial downfield shift of the benzylic amide proton H7
(>0.3 ppm) and upfield shift of amide proton H5 (>0.3 ppm),
which was not evident on binding of the other tBoc amino
acids. One possible explanation for this latter observation is
that binding the serine substrates requires a conformational
change in the macrobicycle or involves formation of a hydrogen
macrobicycle 3. This result is of some interest since certain
bacteria have mutated their cell wall precursor structure from a
terminal Cbz--alanyl--alanine sequence to a Cbz--alanyl--
lactic acid sequence, thus providing immunity to conventional
antibacterials such as the natural -alanyl--alanine receptor,
vancomycin.18
NMR studies
The binding studies show that macrobicycle 3 is a potent and
selective receptor for certain dipeptide guests, as it was designed
to be. In order to try to visualise the mode of binding of such
guests we undertook further NMR and molecular modelling
studies. In particular we wished to ascertain whether the bind-
ing of dipeptide guests was indeed within the cavity of the
macrobicycle, with the dipeptide backbone threading through
the rim formed by the biarylmethane units, or whether the
observed binding might simply involve interactions between the
guests and the exterior of the macrobicyclic structure. The
mode of binding of the most strongly bound guest (Cbz-β-
alanyl--alanine) was therefore investigated by variable tem-
perature 1D and 2D 1H NMR and molecular modelling.
1D and 2D TOCSY19 1H NMR spectra of the 1:1 complex
(3.0 mM concentration) at 25 ЊC allowed assignment of most
protons for both the host and the guest, and showed that many
of the signals are shifted significantly on complexation relative
to the free host and free guest (Fig. 2). Thus, for example, three
of the NH’s of the host (NH1, NH3 and NH7) are shifted sig-
nificantly downfield (0.5–1 ppm) whereas NH5 is shifted down-
field by only 0.2 ppm, but this latter signal is considerably
broadened on complexation. For the guest the alanyl amide
bond to C᎐O4, which in turn breaks the intramolecular hydro-
᎐
gen bond (found in both molecular modelling and NMR
studies) to NH5.
Dipeptide substrates did not generally show substantially
better binding compared with the single amino acid derivatives
[although strong binding was observed for Cbz-glycinyl--
serine (entry 17) with large downfield shifts of H1 (~1 ppm), H3
(>0.4 ppm) and H7 (>0.35 ppm), and an upfield shift of H5
(~0.3 ppm)] until the β-alanylalanine substrates were used
(entries 21 and 22). Macrobicycle 3 again showed a preference
for the -substrate (Cbz-β-alanyl--alanine, Ϫ∆Ga = 22.8 kJ
molϪ1, Cbz-β-alanyl--alanine, Ϫ∆Ga = 19.1 kJ molϪ1), and
both the β-alanylalanine substrates were significantly better
bound than the corresponding alanine substrates (entries 4, 5)
or the alanylalanine substrates (entries 19 and 20). The extent
of this increase in binding suggests that strong additional
binding interactions had been formed.
p
NHr is shifted downfield by 0.6 ppm. The alanyl methyl (CH3 )
and alanyl α-proton (Hq) of the guest are shifted upfield upon
complexation by 0.2 and 0.5 ppm respectively, whilst the phenyl-
alanyl α-proton (H2) of the host was shifted downfield by 0.4
ppm indicating a significant change in environment for these
protons on complexation that is consistent with binding of the
substrate within the cavity of the macrobicyclic receptor 3 and
is supported by the molecular modelling results (vide infra). The
signals for the β-alanyl portion of the guest in the 1:1 complex
could not be assigned with certainty, either by inspection of the
1
We also measured the binding of the ester linked Cbz-β-
alanyllactic acids (entries 23 and 24) as substrates for macro-
bicycle 3, effectively replacing a hydrogen bond donor substitu-
ent (NH) with a hydrogen bond acceptor substituent (O-lone
pair) in the guest structure. The binding energies for both
enantiomers of Cbz-β-alanyllactic acid (11.3, 13.2 kJ molϪ1) are
smaller than the binding energy measured for phenylacetic acid
(14.2 kJ molϪ1) indicating that these substrates apparently gain
no stabilisation on binding other than from the carboxylic acid
interaction with the amidopyridine binding site. The substan-
tially lower binding energies for these esters compared to the
corresponding Cbz-β-alanylalanines strongly suggest that a
hydrogen bond from the guest amide NHr (for the atom label-
ling, see Fig. 2) to the receptor side wall is a key interaction in
the binding of the latter—a result which is strongly supported
by the molecular modelling studies (vide infra), which suggest
that binding of Cbz-β-alanyl--alanine involves a hydrogen
1D H NMR or from the 2D TOCSY where no cross peaks
attributable to the β-alanyl portion were observed. However a
reverse titration experiment (addition of a solution of macrobi-
cycle 3 to a solution of Cbz-β-alanyl--alanine) showed an
t
u
upfield shift of the signals for CH2 CH2 and a downfield shift
for NHv, which could be extrapolated to give the shifts in the
1:1 complex.
1
At an elevated temperature (40 ЊC), the H NMR spectrum
showed little change from that at 25 ЊC, indicating that the guest
is still bound at the higher temperature. 2D ROESY spectra of
the 1:1 mixture at 25 and 40 ЊC showed only one assignable
intermolecular cross peak between host and guest, that being
p
between the alanyl methyl (CH3 ) of the guest and the phenyl-
alanyl α-proton (H2) of the host. Comparison of the intensity
of this cross peak with other intramolecular cross peaks
allowed an estimate of the upper limit of the distance between
these signals as 3.6 Å.
bond from NHr to C᎐O4 (although the weaker binding might
It was evident from these NMR studies that various exchange
processes were occurring on the NMR timescale which were
leading to poor resolution of, in particular, the signals for the
β-alanyl portion of the guest. In order to overcome this, a series
of 1D NMR spectra at Ϫ60, Ϫ40, Ϫ20 and 0 ЊC was obtained
of free host, free guest and 1:1 complex in an attempt to freeze
out any exchange between bound and unbound material. As the
temperature was lowered for the mixtures, resolution was
improved, but many more peaks appeared indicating that a
number of conformations of the complex were being observed.
An example of this was clearly seen in the region above 8.5
ppm, where at 40 ЊC, only one signal was seen for the amido-
᎐
alternatively result from the replacement of the relatively good
hydrogen bond acceptor provided by the alanyl amide carbonyl
with the less effective lactate ester carbonyl, and as a con-
sequence of the ester being inherently more flexible than the
amide). Coincidentally, a very similar result was observed with
receptor 1,1d which also bound Cbz-β-alanylalanine substrates
strongly (Cbz-β-alanyl--alanine, Ϫ∆Ga = 19.9 kJ molϪ1, Cbz-
β-alanyl--alanine, Ϫ∆Ga = 16.6 kJ molϪ1) and Cbz-β-alanyl-
lactic acids much more weakly (Cbz-β-alanyl--lactic acid,
Ϫ∆Ga = 13.1 kJ molϪ1, Cbz-β-alanyl--lactic acid, Ϫ∆Ga = 11.2
kJ molϪ1), although the selectivity is more pronounced with
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031
1025

View Article Online
pyridyl NH1 but at Ϫ40 ЊC, at least four separate signals could
be observed.
Thus the data from the variable temperature and 2D NMR
studies indicated a flexible host–guest system with several
low energy conformations and in particular the signals for the
β-alanyl portion of the guest were poorly resolved. However,
the significant shifts in the various signals for the host and the
guest on complexation are consistent with complexation of the
guest within the cavity of the receptor utilising several binding
interactions. The location of the alanyl methyl of the guest
close to the phenylalanyl α-proton of the host also provides a
useful constraint for subsequent molecular modelling studies.
Molecular modelling
Molecular modelling was also carried out in an attempt to
visualise the structure of the complex. The geometry of the free
macrobicycle 3 and its complex with Cbz-β-alanyl--alanine
was examined using a combination of simulated annealing cal-
culations and molecular dynamics using the MacroModel pro-
gram.20 The AMBER*21 force field was used22 as implemented
in MacroModel V5.0 and the effect of solvent was included
through the use of the GB/SA chloroform model.23 Initially the
free macrobicycle was energy-minimised using a conjugate
gradient (PRCG) and ten simulated annealing calculations of
1 ns, involving slow cooling from 600 to 0.01 K. Three of the
structures from these simulated annealing calculations were
selected for further 1 ns molecular dynamics calculations at 300
K to examine the behaviour of 3 at room temperature. The
simulations generated a large number of low energy structures,
which indicated that the free macrobicycle is a very flexible
structure. Thus many conformations were observed, including
structures in which the pyridine unit was inverted inside the rim
of the macrobicycle (cf. structure 21) or one of the phenyl-
alanine-derived benzyl groups was folded inside the rim, but
these structures were significantly higher in energy than the
predominant form of the macrobicycle, with an open cavity and
the diamidopyridine pointing inwards, suitable for binding sub-
strates as desired. A consistent feature of these structures was
the presence of intramolecular hydrogen bonds, particularly
Fig. 3 Typical structure of a 1:1 complex between macrobicycle 3 and
Cbz-β-alanyl--alanine from molecular dynamics calculations, showing
hydrogen bond interactions between receptor and substrate.
preference for a hydrogen bond from NHr of the guest to C᎐O4
᎐
of the host—maintained for 81% of the structures sampled dur-
ing the course of the molecular dynamics—and for a hydrogen
bond from C᎐Os of the guest to NH5Ј of the host—maintained
᎐
for 86% of the structures sampled (Fig. 3). The NMR studies
revealed that NHr is shifted 0.6 ppm downfield on complex-
ation, which is entirely consistent with formation of a hydrogen
bond to C᎐O4 of the host (with concomitant breaking of the
᎐
intramolecular hydrogen bond to NH5). The NMR studies also
revealed that NH5 is only shifted by 0.2 ppm on complexation,
but this, again, is consistent with breaking the intramolecular
hydrogen bond from NH5 to C᎐O4 and forming an inter-
᎐
molecular hydrogen bond to C᎐Os. (Conversely, binding with
᎐
serine substrates was observed in the titration experiments to
involve an upfield shift of NH5, which can be explained by for-
mation of a hydrogen bond from the hydroxy side chain of the
serine substrate to C᎐O4, which breaks the intramolecular
᎐
from NH5 to C᎐O4, which is entirely consistent with the
hydrogen bond to H5.)
᎐
1
observed downfield position for the signal for NH5 in the H
The Cbz-β-alanyl portion of the guest, however, adopted
many positions relative to the receptor throughout the molecu-
lar dynamics calculations, allowing various hydrogen bond
NMR of macrobicycle 3 (vide supra).
A structure from the room temperature simulation was
chosen and the dipeptide Cbz-β-alanyl--alanine was docked
within the cavity with the peptide chain threaded through the
rim of the macrobicyclic receptor, such that the carboxylic acid
proton was constrained at 1.8 Å from the pyridyl nitrogen atom
and the distance from the alanyl methyl carbon of the guest to
the phenylalanyl α-proton (H2) of the host was constrained at
2.5 2.5 Å, based on distance constraint determined from the
2D NMR studies on the same complex. Similarly, in a separate
calculation, the dipeptide Cbz-β-alanyl--alanine was docked
with the same constraints on the carboxylic acid proton and on
the alanyl methyl carbon, but with the peptide chain on the
outside of the cavity of the macrobicyclic receptor. The result-
ing complexes were both subjected to 10 cycles of the same
energy minimisation and simulated annealing protocol (from
600 to 0.01 K) used for the free macrobicycle.
With the peptide docked initially within the rim of the
macrobicyclic cavity, the simulated annealing procedure led to a
series of related structures with several hydrogen bond inter-
actions between host and guest. The two lowest energy struc-
tures were selected for a 5 ns molecular dynamics calculation at
300 K (sampling 100 structures for each) to examine the
behaviour of the complex at room temperature. The molecular
dynamics calculations for both of these two starting structures
generated a series of closely related structures which consist-
ently involved the expected motif of two hydrogen bonds from
the carboxylic acid to the amidopyridine unit, as well as a clear
interactions from the benzyl carbamate NHv to C᎐O6 of the
᎐
receptor, or from the carbamate C᎐Ow or Ox to NH7 of the
᎐
receptor. This observation is entirely consistent with the NMR
1
studies, which showed that, in the 1:1 complex with 3, the H
signals for the β-alanyl portion of the guest were poorly
resolved.
In contrast, structures generated by the simulated annealing
protocol, but starting with the peptide guest initially docked on
the outside of the cavity, gave a series of structures devoid of
any hydrogen bond interactions between host and guest, other
than those between the carboxylic acid and the amidopyridine
unit. (For one of the calculations the peptide chain was
observed to work its way from the outside of the cavity and
through the rim to produce a structure closely related to those
found in the initial annealing calculations starting with the
guest docked within the cavity.) The structures generated with
the peptide guest on the outside of the cavity were therefore not
consistent with the experimentally determined binding selec-
1
tivities or observed changes to proton signals in the H NMR
spectra on complexation.
Thus the modelling and experimental data produce a consist-
ent picture of a flexible receptor, which binds the most strongly
bound substrate Cbz-β-alanyl--alanine within the cavity of
the receptor as hoped, utilising hydrogen bonds to both the NH
and C᎐O of the substrate amide (which are, of course, lost
᎐
on changing to the substrate Cbz-β-alanyl--lactic acid) and
1026
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031

View Article Online
a less well-defined interaction with the benzyl carbamate
functionality.
(CH), 127.5 (C), 127.4 (CH), 117.6 (CH2), 78.2 (C), 64.4 (CH2),
53.8 (CH), 52.1 (CH3), 41.8 (CH2), 40.6 (CH2), 30.2 (CH2), 28.2
(CH3), 27.2 (CH2); LRMS (FAB) m/z 525 (Mϩ ϩ H, 21%), 469
(12), 425 (32), 407 (20), 154 (100) (Found C, 66.74; H, 7.18; N,
5.67. C29H36N2O7 requires C, 66.40; H, 6.92; N, 5.34%).
Based on this work it is clear that the basic design of an
amidopyridine-derived macrobicycle is an appropriate struc-
ture for peptide recognition, but more selective binding proper-
ties can be anticipated for structures with greater rigidity—and
thus greater levels of preorganisation—as has been borne out
by more recent work with macrobicycle 22 which is indeed more
selective (particularly for Cbz--alanyl--alanine) than 3.13
(7S)-14-Methoxycarbonyl-7-tert-butoxycarbonylamino-6-oxo-5-
aza-1(1),3(1,4)-dibenzenanonaphane-9-carboxylic acid 9
Tetrakis(triphenylphosphine)palladium(0) (77 mg, 0.067 mmol)
was added to a stirred solution of 8 (710 mg, 1.35 mmol),
triphenylphosphine (43 mg, 0.17 mmol) and pyrrolidine (0.284
mL, 3.37 mmol) in degassed CH2Cl2 (7 mL). The solution was
stirred for 30 minutes and the reaction was shielded from light.
1 M HCl (10 mL) was added and the aqueous layer extracted
with EtOAc (2 × 20 mL). The combined organic layers were
washed with 1 M HCl (10 mL) and brine (10 mL), dried
(Na2SO4), filtered and concentrated under reduced pressure to
give an off-white solid. The product was purified by column
chromatography [CH2Cl2–MeOH (9:1 v/v)] to yield the acid 9
as a white solid (622 mg, 95%); Rf [CH2Cl2–MeOH (9:1 v/v)]
0.20; νmax(CHCl3)/cmϪ1 3700–3350, 3016, 2930, 1714, 1666;
δH(270 MHz, (CD3)2SO) 8.33 (1H, br s, NHCH2), 7.90 (2H, d,
J 8, ArH), 7.33 (2H, d, J 8, ArH), 7.19 (4H, s, ArH), 7.01 (1H,
d, J 6, NHCH), 4.24 (2H, d, J 6, ArCH2NH), 4.00 (2H, s,
ArCH2Ar), 3.93 (1H, m, NHCH), 3.82 (3H, s, OCH3), 2.12
(2H, m, CH2COO), 1.80 (1H, m, CHAHBCH2CO), 1.72 (1H,
m, CHAHBCH2CO), 1.39 (9H, s, C(CH3)3); δC(75.5 MHz,
(CD3)2SO) 174.8 (C), 172.0 (C), 166.2 (C), 155.4 (C), 147.0 (C),
138.8 (C), 137.3 (C), 129.4 (CH), 128.9 (CH), 128.7 (CH), 127.6
(C), 127.4 (CH), 78.1 (C), 54.7 (CH), 51.9 (CH3), 41.9 (CH2),
40.8 (CH2), 31.4 (CH2), 28.1 (CH3), 27.7 (CH2); LRMS (FAB)
m/z 485 (Mϩ ϩ H, 14%), 423 (24), 385 (38), 335 (73), 239 (100).
Experimental
General methods
Whenever possible all solvents and reagents were purified
according to literature procedures.24 Thin layer chromato-
graphy (TLC) was performed on aluminium-backed sheets
(CamLab) coated with either silica gel (SiO2; 0.25 mm) or
neutral alumina, containing fluorescent indicator UV254. Unless
otherwise indicated column chromatography was performed on
Sorbsil C60, 40–60 mesh silica. All melting points were deter-
mined in open capillary tubes using Gallenkamp Electro-
thermal Melting Point Apparatus and are uncorrected. Infrared
spectra were recorded on a Perkin-Elmer 1600 Fourier Trans-
1
form Spectrophotometer. H NMR spectra at 270 MHz were
obtained on a JEOL GX 270, at 300 MHz on a Bruker AC 300,
at 360 MHz on a Bruker AM 360 and at 500 MHz on a Varian
VXR 500 spectrometer. 13C NMR spectra were recorded at 68
MHz on the JEOL GX 270, at 75 MHz on the Bruker AC 300
and at 90 MHz on the Bruker AM 360. Chemical shifts are
reported in ppm on the δ scale, coupling constants in Hz. The
multiplicities of the signals were determined using the dis-
tortionless enhancement by phase transfer (DEPT) spectral
editing technique. Microanalytical data were obtained from
Wellcome Research Laboratories, Beckenham, Kent or The
University College of London. Mass spectra were recorded
either on a Micromass Platform quadrupole mass analyser with
an electrospray ion source, or on a VG Analytical 70-250-SE
normal geometry double focusing mass spectrometer at
Southampton University. Optical rotations were measured on an
(1S,7S)-1,7-Dibenzyl-1,7-bis(tert-butyloxycarbonylamino)-3,5-
diaza-4(2,6)-pyridinaheptaphane-2,6-dione 5
Dicyclohexylcarbodiimide (4.8 g, 23 mmol) was added to a
solution of 2,6-diaminopyridine (0.52 g, 4.8 mmol) and N-tert-
butyloxycarbonylphenylalanine (6.2 g, 24 mmol) in DMF (30
mL) and the mixture was stirred for 72 hours. The resulting
yellow suspension was filtered and the filtrate concentrated to
half of the original volume. EtOAc (75 mL) was added and the
solution was washed with 1 M HCl (3 × 50 mL), 1 M NaOH
(3 × 50 mL), brine (2 × 50 mL), dried (MgSO4), filtered and the
solvent removed under reduced pressure to give a foam, which
was purified by flash column chromatography [petroleum ether
(bp 40–60 ЊC)–EtOAc (7:3 v/v)] to yield the diamidopyridine 5
as a foam (2.42 g, 85%), mp 88–90 ЊC; [α]D ϩ32.2 (c = 1.0 in
MeOH); νmax(KBr)/cmϪ1 2980, 1685, 1585, 1510, 1448; δH(300
MHz, (CD3)2SO) 10.22 (2 H, s, pyrNH), 7.77 (3 H, br s, pyrH),
7.39–7.16 (12 H, m, PhH, NHBoc), 4.48 (2 H, m, CHCO), 3.03
(2 H, dd, J 13 and 5, CHAHBPh), 2.82 (2 H, dd, J 13 and 11,
CHAHBPh), 1.31 (18 H, s, CH3); δC(75.5 MHz, (CD3)2SO) 171.8
(C), 155.6 (C), 150.2 (C), 140.2 (CH), 138.0 (C), 129.4 (CH),
128.1 (CH), 126.4 (CH), 109.3 (CH), 78.3 (C), 56.5 (CH), 37.1
(CH2), 28.2 (CH3); LRMS (ES) m/z 626.2 (Mϩ ϩ Na, 13%),
604.2 (Mϩ ϩ H, 25) (Found C, 65.27; H, 6.92; N 11.28.
C33H41N5O6 requires C, 65.65; H, 6.85; N, 11.60%).
AA-100 Polarimeter, and are given in units of 10Ϫ1 deg cm2 gϪ1
.
1-Methyl 9-allyl (7S)-7-tert-butoxycarbonylamino-6-oxo-5-aza-
1(1),3(1,4)-dibenzenanonaphane-1,9-dicarboxylate 8
Dicyclohexylcarbodiimide (467 mg, 2.27 mmol) was added to a
stirred solution of amine salt 7 (668 mg, 2.27 mmol), N-tert-
butoxycarbonylglutamic acid γ-allyl ester (650 mg, 2.27 mmol),
HOBT (311 mg, 2.27 mmol), and diisopropylethylamine
(DIPEA) (397 µL, 2.28 mmol) in DMF (20 mL) at 0 ЊC. The
mixture was allowed to warm to room temperature and stirred
for a further 24 hours. The resulting suspension was filtered and
the filtrate was concentrated under reduced pressure to give a
crude orange solid which was dissolved in EtOAc (10 mL) and
washed with 1 M HCl solution (2 × 10 mL) and brine (10 mL).
The organic layer was dried (Na2SO4), filtered, and concen-
trated under reduced pressure to give an orange solid which was
purified by column chromatography [petroleum ether (bp 40–
60 ЊC)–EtOAc (3:2 v/v)] to give 8 as a white powder (797 mg,
67%); Rf [petroleum ether (bp 40–60 ЊC)–EtOAc (3:2 v/v)] 0.47;
νmax(CHCl3)/cmϪ1 3429, 3327, 3018, 1716, 1672, 1512; δH(300
MHz, (CD3)2SO) 8.30 (1H, t, J 6, NHCH2), 7.85 (2H, d, J 8,
ArH), 7.35 (2H, d, J 8, ArH), 7.18 (4H, s, ArH), 6.95 (1H, d,
Bis(trifluoroacetate) salt 6
Trifluoroacetic acid (2.0 mL) was added to a solution of 5 (0.57
g, 0.95 mmol) in CH2Cl2 (10 mL) and the reaction stirred for 2
hours. The mixture was concentrated under reduced pressure
and the resulting yellow–brown oil was triturated with ether (10
mL) and filtered to give the ammonium salt 6 as a pale yellow
powder (0.485 g, 85%), mp 140–143 ЊC; νmax(KBr)/cmϪ1 3032,
2935, 1669, 1587, 1540, 1499; δH(300 MHz, CDCl3/(CD3)2SO,
95:5 v/v) 10.56 (s, 2 H, pyrNH), 8.45 (s, 6 H, NH3), 7.74 (d,
J 7.4, 2 H, pyrH), 7.60 (t, J 7.4, 1 H, pyrH), 7.23–7.08 (m, 10 H,
J 8, NHCH), 5.90 (1H, m, CH᎐CH ), 5.28 (1H, d, J 16,
᎐
᎐
2
CH᎐CH H ), 5.18 (1H, d, J 9, CH᎐CH H ), 4.50 (2H, d, J 7,
᎐
A
B
A
B
CH CH᎐CH ), 4.21 (2H, d, J 6, ArCH NH), 3.95 (2H, s,
᎐
2
2
2
ArCH2Ar), 3.98 (1H, m, NHCH), 3.81 (3H, s, OCH3), 2.36
(2H, t, J 7, CH2COO), 1.90 (1H, m, CHAHBCH2CO), 1.76 (1H,
m, CHAHBCH2CO), 1.38 (9H, s, C(CH3)); δC(75.5 MHz,
(CD3)2SO) 172.0 (C), 171.6 (C), 166.2 (C), 155.4 (C), 147.2 (C),
138.9 (C), 137.4 (C), 132.8 (CH), 129.4 (CH), 129.0 (CH), 128.7
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031
1027

View Article Online
PhH), 4.41 (br s, 2 H, NCHCO), 3.19 (dd, J 14.0 and 7.0, 2 H,
CHAHBPh), 3.08 (dd, J 14.0 and 7.0, 2 H, CHAHBPh).
sure. The residue was dissolved in EtOAc (180 mL) and washed
with saturated NaHCO3 (2 × 80 mL), brine (80 mL) and dried
(Na2SO4). Solvent was removed under reduced pressure to give
a colourless oil which solidified upon standing. The product
was triturated with cold hexane and filtered to give ester 12 as
a white solid (23.9 g, 80%); mp (hexane–EtOAc): 63–65 ЊC;
νmax(CHCl3)/cmϪ1 3033, 2955, 1712, 1608, 1575, 1493, 1447;
δH(300 MHz, CDCl3) 8.07 (dt, J 8 and 1, 2 H, ArH), 7.49–7.36
(m, 7 H, ArH), 5.38 (s, 2 H, CH2Ph), 4.51 (s, 2 H, ArCH2Br);
δC(75.5 MHz, CDCl3) 166.0 (C), 142.9 (C), 136.1 (C), 130.4
(CH), 130.2 (C), 129.2 (CH), 128.1 (CH), 128.5 (CH), 128.3
(CH), 67.0 (CH2), 32.3 (CH2); LRMS (CI) m/z 322 (Mϩ ϩ NH4,
100%), 305 (Mϩ ϩ H, 6), 244 (52) (Found C, 58.92; H, 4.31.
C15H13BrO2 requires C, 59.04; H, 4.29%).
(7S,12S,18S,23S)-Dimethyl 12,18-dibenzyl-7,23-bis(tert-
butoxycarbonylamino)-6,10,13,17,20,24-hexaoxo-5,11,14,16,
19,25-hexaaza-1,29(1),3,27(1,4)-tetrabenzena-15(2,6)-
pyridinanonacosaphane-14,294-dicarboxylate 10
DIPEA (0.07 mL, 0.4 mmol) was added to a solution of acid 9
(100 mg, 0.207 mmol), bis-TFA salt 6 (44 mg, 0.068 mmol) and
PyBOP (108 mg, 0.207 mmol) in DMF (1 mL) and the mixture
was stirred at 50 ЊC for 18 hours. The reaction mixture was
diluted with EtOAc (10 mL) and washed with 2 M HCl solution
(8 × 5 mL), sat. NaHCO3 (5 mL), brine (5 mL), dried (Na2SO4)
and concentrated under reduced pressure. The crude product
was purified by column chromatography [CH2Cl2–MeOH (97:3
v/v)] to give diester 10 as a pale brown powder (66 mg, 71%); Rf
[CH2Cl2–MeOH (95:5 v/v)] 0.63; [α]D ϩ6.2 (c = 1.0 in MeOH);
νmax(CHCl3)/cmϪ1 3427, 1715, 1663, 1530; δH(270 MHz, (CD3)2-
SO) 10.34 (br s, 2 H, pyrNH), 8.29 (br s, 4 H, CH2NH and
CH2CONH), 7.91 (d, J 8, 4 H, ArH), 7.78 (s, 3 H, pyrH), 7.42–
7.14 (m, 22 H, ArH), 6.94 (d, J 6, 2 H, BocNH), 4.84 (m, 2 H,
PhCH2CH), 4.26 (d, J 6, 4 H, ArCH2NH), 4.02 (s, 4 H,
ArCH2Ar), 3.91 (m, 2 H, CHCH2CH2), 3.84 (s, 6 H, CO2CH3),
3.10 (m, 2 H, PhCHAHB), 2.85 (m, 2 H, PhCHAHB), 2.24–2.04
(m, 4 H, CH2CH2CO), 1.82–1.62 (m, 4 H, CH2CH2CO), 1.39 (s,
18 H, C(CH3)); δC(67.9 MHz, (CD3)2SO) 171.8 (C), 171.7 (C),
171.1 (C), 166.0 (C), 155.2 (C), 149.9 (C), 147.0 (C), 140.1
(CH), 138.7 (C), 137.5 (C), 137.1 (C), 129.2 (CH), 129.1 (CH),
128.8 (CH), 128.5 (CH), 127.9 (CH), 127.1 (CH), 126.2 (CH),
109.2 (CH), 78.0 (C), 54.5 (CH), 54.0 (CH), 51.9 (CH3), 41.6
(CH2), 40.4 (CH2), 37.0 (CH2), 31.6 (CH2), 28.0 (CH3), 27.7
(CH2); LRMS (FAB) m/z 1336 (Mϩ ϩ H, 4%), 335 (63), 239
(91), 120 (100) (Found C, 67.39; H, 6.46; N, 9.38. C75H85N9O14
requires C, 67.40; H, 6.41; N, 9.43%).
Benzyl 4-(4Ј-cyanophenyl)methylbenzoate 13
Bromo ester 12 (8.4 g, 28 mmol), 4-cyanophenylboronic acid
(5.3 g, 36 mmol) and tetrakis(triphenylphosphine)palladium(0)
(20 mg, 0.018 mmol) were stirred in degassed dimethoxyethane
(100 mL) for 5 minutes to give a clear yellow solution. Degassed
2 M Na2CO3 (25 mL, 50 mmol) was added and the reaction was
refluxed for 4 hours. After cooling, water (100 mL) was added
and the mixture stirred for 5 minutes. The organic layer was run
off and the aqueous layer extracted with CH2Cl2 (2 × 100 mL).
The organics were combined, dried (MgSO4), filtered and the
solvent removed under reduced pressure to give an oil, which
was purified by flash column chromatography [petroleum ether
(bp 40–60 ЊC)–EtOAc (5:95 v/v)], to give the cyano ester 13 as
white crystals (6.10 g, 68%), mp 69–70 ЊC; νmax(KBr)/cmϪ1
3031, 2957, 2224, 1717, 1604, 1494; δH(300 MHz, CDCl3) 8.04
(dt, J 8 and 2, 2 H, ArH), 7.60 (dt, J 8 and 2, 2 H, ArH), 7.48–
7.35 (m, 7 H, ArH), 7.28 (d, J 8, 2 H, ArH), 7.25 (d, J 8, 2 H,
ArH), 5.37 (s, 2 H, CH2Ph), 4.10 (s, 2 H, ArCH2Ar); δC(75.5
MHz, CDCl3) 166.3 (C), 145.8 (C), 144.9 (C), 136.2 (C), 132.6
(CH), 130.4 (CH), 129.8 (CH), 129.2 (CH), 128.9 (C), 128.8
(CH), 128.4 (CH), 128.3 (CH), 119.0 (C), 110.6 (C), 66.9 (CH2),
42.1 (CH2) (Found C, 80.58; H, 5.11; N, 4.16. C22H17NO2
requires C, 80.71; H, 5.23; N, 4.28%).
Attempted hydrolysis of bis-methyl ester 10
1 M LiOH (1 mL) was added to a solution of diester 10 (28 mg,
0.02 mmol) in 1,4-dioxane (1 mL) and the mixture stirred at
room temperature for 22 hours. The reaction mixture was
concentrated under reduced pressure to give a white solid,
which was partitioned between EtOAc (5 mL) and water (5 mL).
The aqueous layer was acidified to pH 1 with 1 M HCl
and extracted with EtOAc (3 × 5 mL). The combined organic
extracts were washed with brine (5 mL), dried (Na2SO4), filtered
and concentrated under reduced pressure to give diacid 11 as a
white foam (23 mg, 92%); Rf [CH2Cl2–MeOH (9:1 v/v)] 0.36;
νmax(CHCl3)/cmϪ1 3500–3200, 1711, 1642; δH(270 MHz,
(CD3)2SO) 8.21 (t, J 6, 1 H, NHCH2), 8.15 (d, J 8, 1 H,
NHCOCH2), 7.85 (d, J 8, 2 H, ArH), 7.33 (d, J 8, 2 H, ArH),
7.28–7.10 (m, 9 H, ArH), 6.90 (d, J 8, 1 H, CH2CHNH), 4.40
(m, 1 H, CHCH2Ph), 4.22 (d, J 6, 2 H, ArCH2NH), 3.99 (s, 2 H,
ArCH2Ar), 3.87 (m, 1 H, CH2CH2CH), 3.15 (dd, J 5 and 11,
1 H, PhCHAHB), 2.83 (dd, J 9 and 11, 1 H, PhCHACHB), 2.10
(m, 2 H, CH2CH2CO), 1.75 (m, 1 H, CHAHBCH2CO), 1.65 (m,
1 H, CHAHBCH2CO), 1.36 (s, 9 H, CH3); δC(67.9 MHz,
(CD3)2SO) 173.0 (C), 171.9 (C), 171.6 (C), 171.5 (C), 167.1 (C),
155.2 (C), 146.5 (C), 138.8 (C), 137.5 (C), 137.1 (C), 129.4
(CH), 128.9 (CH), 128.7 (CH), 128.5 (CH), 128.0 (CH), 127.1
(CH), 126.2 (CH), 77.9 (C), 54.0 (CH), 53.2 (CH), 41.6 (CH2),
40.4 (CH2), 36.6 (CH2), 31.7 (CH2), 28.0 (CH3), 27.7 (CH2);
LRMS (FAB): m/z 618 (Mϩ ϩ H, 47%), 518 (58), 335 (100), 225
(85).
Benzyl 4-(4Ј-aminomethylphenyl)methylbenzoate hydrochloride
salt 14
Borane–dimethyl sulfide complex (3.6 mL of a 2 M solution in
THF, 7.2 mmol) was added over 5 min to a stirred solution of
benzyl 4-(4Јcyanophenyl)methylbenzoate 13 (2.00 g, 6.12
mmol) in THF (60 mL) at reflux. The resulting mixture was
stirred at reflux for 5 hours to give a clear solution. The reaction
was cooled to room temperature and quenched by adding HCl
(9.2 mL of a 0.8 M solution in MeOH). The resulting white
suspension was heated at reflux for 15 minutes and concen-
trated under reduced pressure to give a white solid, which was
triturated with dry MeOH (15 mL). The MeOH was removed
and the resulting white solid suspended in boiling MeOH (10
mL) for 5 minutes before being isolated by filtration to give the
amine hydrochloride salt 14 as a white solid (1.61 g, 72%); mp
220–222 ЊC; νmax(KBr)/cmϪ1 3031, 2580, 2360, 1715, 1609, 1591,
1482; δH(300 MHz, (CD3)2SO) 8.48 (s, 3 H, NH3), 7.93 (d, J 8, 2
H, ArH), 7.47–7.34 (m, 9 H, ArH), 7.28 (d, J 8, 2 H, ArH), 5.34
(s, 2 H, CH2Ph), 4.04 (s, 2 H,CH2N), 3.95 (s, 2 H, ArCH2Ar);
δC(75.5 MHz, (CD3)2SO) 165.6 (C), 147.1 (C), 140.9 (C), 136.3
(C), 132.1 (C), 129.6 (CH), 129.3 (CH), 129.2 (CH), 129.0 (CH),
128.6 (CH), 128.2 (CH), 128.0 (CH), 127.6 (C), 66.1 (CH2), 41.9
(CH2), 40.7 (CH2); LRMS (ESϩ) m/z 332.4 (Mϩ ϩ H) (Found
C, 72.13; H, 5.99; N, 3.64. C22H22ClNO2 requires C, 71.83; H,
6.03; N, 3.81%).
Benzyl ꢀ-bromo-p-toluate 12
A solution of α-bromo-p-toluic acid (21.5 g, 0.10 mol), benzyl
alcohol (14.7 mL, 0.13 mol), p-TsOHؒH2O (1.9 g, 10 mmol) in
toluene (250 mL) was heated to reflux and water removed in a
Dean–Stark trap. After 18 hours, the mixture was allowed to
cool and most of the toluene was removed under reduced pres-
1-Benzyl 9-allyl (7S)-7-tert-butoxycarbonylamino-6-oxo-5-aza-
1(1),3(1,4)-dibenzenanonaphane-14,9-dicarboxylate 15
DIPEA (6.3 mL, 36 mmol) was added to a suspension of amine
hydrochloride salt 14 (3.4 g, 9.2 mmol), N-tert-butoxycarbonyl-
1028
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031

View Article Online
-glutamic acid γ-allyl ester (3.07 g, 10.6 mmol) and PyBOP
(5.9 g, 11 mmol) in DMF (200 mL) and the resulting solution
was stirred at room temperature for 18 hours. The mixture was
concentrated under reduced pressure to give an orange oil
which was dissolved in EtOAc (50 mL), washed with 1 M HCl
(10 mL), water (5 × 10 mL), brine (10 mL) and dried (Na2SO4).
The solvent was removed under reduced pressure to give a solid
which was purified by column chromatography [petroleum
ether (bp 40–60 ЊC)–EtOAc (3:2 v/v)] to afford the allyl ester 15
as a white solid (5.10 g, 93%); mp 56–58 ЊC (EtOAc–hexane);
[α]D ϩ3.6 (c = 1.0 in CH2Cl2); νmax(KBr)/cmϪ1 3030, 2979, 2940,
1720, 1682, 1655, 1609, 1521; δH(300 MHz, CDCl3) 8.01 (d, J 8,
2 H, ArH), 7.47–7.12 (m, 12 H, ArH and NHCH2), 6.53 (br s,
1 H, NHBoc), 5.91 (ddt, J 17, 10 and 6, 1 H, CH=CH2), 5.36 (s,
2 H, CH2Ph), 5.32 (dq, J 17 and 1, 1 H, CH=CHAHB), 5.24 (dq,
J 10 and 1, 1 H, CH=CHAHB), 4.58 (dt, J 6 and 1, 2 H,
CH2CH=CH2), 4.42 (d, J 5, 2 H, NCH2), 4.17 (m, 1 H,
NCHCO), 4.02 (s, 2 H, ArCH2Ar), 2.55 (dt, J 17 and 7, 1 H,
CHAHBCO2), 2.43 (dt, J 17 and 7, 1 H, CHAHBCO2), 2.17 (m,
1 H, CHAHBCH2CO2), 1.96 (m, 1 H, CHAHBCH2CO2), 1.41
(s, 9 H, CH3); δC(75.5 MHz, CDCl3) 176.2 (C), 172.1 (C),
166.5 (C), 156.4 (C), 146.6 (C), 139.3 (C), 136.1 (C), 136.0
(C), 131.9 (CH), 130.0 (CH), 129.1 (CH), 128.9 (CH), 128.6
(CH), 128.2 (CH), 128.1 (CH), 128.0 (C), 127.7 (CH), 118.5
(CH2), 80.7 (C), 66.6 (CH2), 65.4 (CH2), 53.3 (CH), 43.1
(CH2), 41.5 (CH2), 30.0 (CH2), 28.3 (CH3), 27.8 (CH2);
LRMS (ESϩ) m/z 623.6 (Mϩ ϩ Na), 601.6 (Mϩ ϩ H) (Found
C, 69.96; H, 6.72; N, 4.66. C35H40N2O7 requires C, 69.98; H,
6.71; N, 4.66%).
Solvent was removed under reduced pressure and the resulting
red–brown oil was taken up in EtOAc (20 mL) and washed with
2 M HCl (8 × 10 mL), saturated NaHCO3 (10 mL) and brine
(10 mL), and dried (Na2SO4). The crude product was filtered,
concentrated under reduced pressure and purified by column
chromatography [CH2Cl2–MeOH (97:3 v/v)] to give a brown
powder which was recrystallised from EtOAc–petroleum ether
(bp 40–60 ЊC) to give diester 17 as a pale brown solid (264 mg,
64%); mp 156–158 ЊC [EtOAc–petroleum ether (bp 40–60 ЊC)];
[α]D ϩ46.7 (c = 1.0 in DMF); νmax(KBr)/cmϪ1 3305, 2974, 2927,
1715, 1684, 1653, 1608, 1585; δH(300 MHz, (CD3)2SO) 10.35 (br
s, 2 H, pyrNH), 8.27–8.22 (m, 4 H, NHCOCH2 and NHCH2),
7.91 (d, J 8, 4 H, ArH), 7.76 (s, 3 H, pyrH), 7.46–7.13 (m, 32 H,
ArH), 6.90 (d, J 8, 2 H, NHBoc), 5.33 (s, 4 H, OCH2Ph), 4.82
(m, 2 H, PhCH2CH), 4.23 (d, J 5, 4 H, NHCH2Ar), 3.99 (s, 4 H,
ArCH2Ar), 3.87 (m, 2 H, CH(CH2)2), 3.06 (dd, J 13 and 3, 2 H,
PhCHAHB), 2.83 (dd, J 13 and 11, 2 H, PhCHAHB), 2.12 (m,
4 H, CH2CH2CO), 1.72 (m, 4 H, CH2CH2CO), 1.36 (s, 18 H,
CH3); δC(75.5 MHz, (CD3)2SO) 171.8 (C), 171.6 (C), 171.1 (C),
165.3 (C), 155.1 (C), 149.9 (C), 147.2 (C), 140.0 (CH), 138.7
(C), 137.5 (C), 137.1 (C), 136.1 (C), 129.3 (CH), 129.1 (CH),
128.9 (CH), 128.5 (CH), 128.4 (CH), 127.9 (CH), 127.7 (CH),
127.2 (C), 127.1 (CH), 126.2 (CH), 109.2 (CH), 77.9 (C), 65.8
(CH2), 54.5 (CH), 54.0 (CH), 41.6 (CH2), 40.4 (CH2), 37.0
(CH2), 31.6 (CH2), 28.0 (CH3), 27.7 (CH2); LRMS (FAB) m/z
1488 (Mϩ ϩ H, 5%), 335 (8), 315 (12), 91 (100) (Found C, 69.11;
H, 6.25; N, 8.23. C87H93N9O14ؒH2O requires C, 69.35; H, 6.36;
N, 8.37%).
(7S,12S,18S,23S)-12,18-Dibenzyl-7,23-bis(tert-butoxy-
carbonylamino)-6,10,13,17,20,24-hexaoxo-5,11,14,16,19,25-
hexaaza-1,29(1),3,27(1,4)-tetrabenzena-15(2,6)-pyridina-
nonacosaphane-14,294-dicarboxylic acid 18
(7S)-14-Phenylmethoxycarbonyl-7-tert-butoxycarbonylamino-6-
oxo-5-aza-1(1),3(1,4)-dibenzenanonaphane-9-carboxylic acid 16
Tetrakis(triphenylphosphine)palladium(0) (0.52 g, 0.45 mmol)
was added to a solution of allyl ester 15 (5.0 g, 8.3 mmol),
pyrrolidine (1.6 mL, 19 mmol) and triphenylphosphine (0.22 g,
0.85 mmol) in degassed CH2Cl2 (100 mL) and the resulting
yellow solution was stirred at room temperature for 22 hours.
1 M HCl (50 mL) was added and the CH2Cl2 layer was separated.
The aqueous phase was extracted with EtOAc (30 mL) and the
organics were combined, washed with 1 M HCl (70 mL), brine
(50 mL), H2O (50 mL), dried (Na2SO4) and filtered. The solu-
tion was concentrated under reduced pressure to give an off-
white solid. Recrystallisation from EtOAc–petroleum ether (bp
40–60 ЊC) gave the title compound as a white solid (4.42 g,
95%), mp 103–106 ЊC [EtOAc–petroleum ether (bp 40–60 ЊC)];
[α]D ϩ2.0 (c = 1.0 in CH2Cl2); νmax(KBr)/cmϪ1 3323, 2978, 2929,
1717, 1653, 1609, 1575; δH(300 MHz, (CD3)2SO) 12.11 (s, 1 H,
CO2H), 8.28 (t, J 6, 1 H, CH2NH), 7.91 (d, J 8, 2 H, ArH),
7.47–7.34 (m, 7 H, ArH), 7.16 (s, 4 H, ArH), 6.94 (d, J 8, 1 H,
NHCO2), 5.33 (s, 2 H, CH2Ph), 4.23 (d, J 6, 2 H, CH2N), 4.00
(s, 2 H, ArCH2Ar), 3.93 (m, 1 H, NHCHCO), 2.23 (t, J 8, 2 H,
CH2CO2), 1.87 (m, 1 H, CHAHBCH2CO2), 1.72 (m, 1 H,
CHAHBCH2CO2), 1.36 (s, 9 H, CH3); δC(75.5 MHz, (CD3)2SO)
173.9 (C), 171.7 (C), 165.4 (C), 155.3 (C), 147.3 (C), 138.8 (C),
137.3 (C), 136.2 (C), 129.4 (CH), 129.0 (CH), 128.6 (CH), 128.5
(CH), 128.0 (CH), 127.8 (CH), 127.3 (C), 127.2 (CH), 78.0
(C), 65.9 (CH2), 53.8 (CH), 41.7 (CH2), 40.5 (CH2), 30.2 (CH2),
28.1 (CH3), 27.1 (CH2); LRMS (ESϩ) m/z 583.2 (Mϩ ϩ Na),
561.7 (Mϩ ϩ H) (Found C, 68.51; H, 6.39; N, 4.92. C32H36N2O2
requires C, 68.56; H, 6.47; N, 5.00%).
A solution of diester 17 (260 mg, 0.175 mmol) with NH4CO2H
(313 mg, 5 mmol) and 10% Pd/C (50 mg) in DMF (4 mL) was
stirred for 3 hours. The reaction was filtered through Celite and
concentrated under reduced pressure to give the diacid 18 as a
white solid (225 mg, 98%); mp 176–178 ЊC (EtOAc–hexane);
[α]D ϩ11.7 (c = 1.0 in DMF); νmax(KBr)/cmϪ1 3295, 2975, 2928,
1711, 1653, 1608, 1584, 1513; δH(300 MHz, CDCl3) 10.35 (s,
2 H, pyrNH), 8.27–8.22 (m, 4 H, CH2NH and NHCOCH2),
7.86 (d, 4 H, J 8, ArH), 7.76 (s, 3 H, pyrH), 7.34–7.14 (m, 22 H,
ArH), 6.91 (d, J 8, 2 H, NHBoc), 4.82 (m, 2 H, CHCH2Ph),
4.23 (d, J 6, 4 H, CH2N), 3.97 (s, 4 H, ArCH2Ar), 3.87 (m, 2 H,
CHNHBoc), 3.06 (dd, J 13 and 4, 2 H, PhCHAHB), 2.83 (dd,
J 13 and 11, 2 H, PhCHAHB), 2.23–1.99 (m, 4 H, CH2CO),
1.82–1.58 (m, 4 H, CH2CH2CO), 1.36 (s, 18 H, CH3); δC(67.9
MHz, (CD3)2SO) 172.3 (C), 172.1 (C), 171.4 (C), 167.5 (C),
155.6 (C), 150.3 (C), 146.9 (C), 140.4 (C), 139.2 (C), 137.8 (C),
137.4 (C), 129.8 (CH), 129.5 (CH), 129.3 (CH), 129.0 (CH),
128.9 (CH), 128.3 (CH), 127.5 (CH), 126.6 (CH), 109.6 (CH),
78.4 (C), 54.9 (CH), 54.4 (CH), 42.0 (CH2), 40.8 (CH2), 37.4
(CH2), 32.0 (CH2), 28.4 (CH3), 28.1 (CH2); LRMS (FAB) m/z
1309 (Mϩ ϩ H, 19%), 335 (61), 238 (100), 225 (76) (Found C,
66.36; H, 6.26; N, 9.38. C73H81N9O14ؒH2O requires C, 66.10; H,
6.31; N, 9.50%).
Bis(pentafluorophenyl) (7S,12S,18S,23S)-12,18-dibenzyl-7,23-
bis(tert-butyloxycarbonylamino)-6,10,13,17,20,24-hexaoxo-
5,11,14,16,19,25-hexaaza-1,29(1),3,27(1,4)-tetrabenzena-
15(2,6)-pyridinanonacosaphane-14,294-dicarboxylate 19
Dibenzyl (7S,12S,18S,23S)-12,18-dibenzyl-7,23-bis(tert-
butoxycarbonylamino)-6,10,13,17,20,24-hexaoxo-5,11,14,16,
19,25-hexaaza-1,29(1),3,27(1,4)-tetrabenzena-15(2,6)-
pyridinanonacosaphane-14,294-dicarboxylate 17
Dicyclohexylcarbodiimide (113 mg, 0.548 mmol) was added to
a solution of diacid 18 (230 mg, 0.176 mmol), pentafluoroph-
enol (100 mg, 0.543 mmol) and DMAP (14 mg, 115 mmol) in
THF (5 mL) at 0 ЊC. The reaction was allowed to warm to room
temperature and stirred for 16 hours. The reaction was filtered
and the filtrate diluted with CH2Cl2 (10 mL). The organics were
washed with 0.5 M HCl (10 mL), dried (Na2SO4), filtered and
concentrated under reduced pressure to give the crude title
DIPEA (0.30 mL, 1.7 mmol) was added to a solution of acid 16
(464 mg, 0.859 mmol), bis-TFA salt 6 (180 mg, 0.286 mmol)
and PyBOP (449 mg, 0.859 mmol) in DMF (5 mL) and the
mixture was stirred under nitrogen at 50 ЊC for 18 hours.
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031
1029

View Article Online
compound as a white solid (285 mg) which was used directly in
placed in an NMR tube along with a molecular sieve (4 Å,
the next step without further purification.
Aldrich). Guest stock solutions in deuteriochloroform were
made to concentrations such that 10 µL of that solution con-
tained typically 0.25 equivalents of guest with respect to the
host. The 1H NMR of 3 was recorded for the starting solution
and after addition of each aliquot of guest, and the data was
analysed using the Hostest software.14 A dilution experiment
was also carried out in order to verify that there were no effects
on the shift of the signals due to the dilution of the host
throughout the titration experiment.
Bis(trifluoroacetate) salt 20
Trifluoroacetic acid (3 mL) was added to a suspension of crude
activated ester 19 (285 mg, ~0.18 mmol) in CH2Cl2 (6 mL) and
the resulting solution was stirred at room temperature for 2
hours. Solvent and most of the trifluoroacetic acid was removed
under reduced pressure to give a brown gum, which upon
trituration with diethyl ether (3 × 10 mL) yielded a white solid,
which was dried under reduced pressure to give the crude bis-
TFA salt 20 (315 mg) which was used directly in the next step
without further purification; δH(of crude 20) (270 MHz,
(CD3)2SO) 10.47 (s, 2 H, pyrNH), 8.90 (t, J 3, 2 H, NHCH2),
8.46 (d, J 6, 2 H, NHCOCH2), 8.23 (br s, 6 H, NH3ϩ), 8.13 (d,
J 8, 4 H, ArH), 7.78 (s, 3 H, pyrH), 7.55 (d, J 8, 4 H, ArH),
7.42–7.18 (m, 22 H, ArH), 4.87 (m, 2 H, PhCH2CH), 4.41 (dd,
J 3 and 13, 2 H, NHCHAHBAr), 4.32 (dd, J 5 and 13, 2 H,
NHCHAHBAr), 4.12 (s, 2 H, ArCH2Ar), 3.78 (m, 2 H,
CHCH2), 3.11 (dd, J 3 and 13, 2 H, PhCHAHB), 2.87 (dd, J 11
and 13, 2 H, PhCHAHB), 2.22 (m, 4 H, CH2CONH), 2.00 (m, 4
H, CH2CH2CONH).
Acknowledgements
We thank the EPSRC for studentships for P. H. and C. P. W.,
and we thank the Welcome Foundation for a CASE award for
C. P. W.
References
1 Recent examples of synthetic receptors for peptides, see: (a) T.
Fessmann and J. D. Kilburn, Angew. Chem., Int. Ed., 1999, 38,
1993; (b) R. Xu, G. Greiveldinger, L. E. Marenus, A. Cooper and
J. A. Ellman, J. Am. Chem. Soc., 1999, 121, 4898; (c) M. Sirish and
H.-J. Schneider, Chem. Commun., 1999, 907; (d) J. Dowden, P. D.
Edwards, S. S. Flack and J. D. Kilburn, Chem. Eur J., 1999, 5, 79;
(e) Md. A. Hossain and H.-J. Schneider, J. Am. Chem. Soc., 1998,
120, 11208; ( f ) R. Breslow, Z. Yang, R. Ching, G. Trojandt and
F. Odobel, J. Am. Chem. Soc., 1998, 120, 3536; (g) D. W. P. M.
Löwik, M. D. Weingarten, M. Broekema, A. J. Brouwer, W. C. Still
and R. M. J. Liskamp, Angew. Chem., Int. Ed., 1998, 37, 1846;
(h) M. Davies, M. Bonnat, F. Guillier, J. D. Kilburn and M. Bradley,
J. Org. Chem., 1998, 63, 8696; (i) R. J. Pieters, J. Cuntze, M. Bonnet
and F. Diederich, J. Chem. Soc., Perkin Trans. 2, 1997, 1891; (j)
W. C. Still, Acc. Chem. Res., 1996, 29, 155.
2 See reference 1(d) and (a) J. Dowden, P. D. Edwards and J. D.
Kilburn, Tetrahedron Lett., 1997, 38, 1095; (b) S. S. Flack and
J. D. Kilburn, Tetrahedron Lett., 1995, 36, 3409.
3 (a) G. J. Pernia, J. D. Kilburn, J. W. Essex, R. J. Mortishire-Smith
and M. Rowley, J. Am. Chem. Soc., 1996, 118, 10220; (b) G. J.
Pernia, J. D. Kilburn and M. Rowley, J. Chem. Soc., Chem.
Commun., 1995, 305.
4 For a preliminary communication of a part of this work, see: C. P.
Waymark, J. D. Kilburn and I. Gillies, Tetrahedron Lett., 1995, 36,
3051.
5 (a) R. S. Wareham, J. D. Kilburn, N. H. Rees, D. L. Turner, A. R.
Leach and D. S. Holmes, Tetrahedron Lett., 1995, 36, 3047; (b) R. S.
Wareham, J. D. Kilburn, N. H. Rees, D. L. Turner and D. S. Holmes,
Angew. Chem., Int. Ed. Engl., 1995, 34, 2660.
6 Although the dicyclohexylcarbodiimide coupling proved the
method of choice in this instance, we have had difficulties coupling
other amino acids with diaminopyridine and have found that the
most reliable method involves preactivation of the diaminopyridine
with bis(trimethylsilyl)acetamide (S. Rajeswari, R. J. Jones and M. P.
Cava, Tetrahedron Lett., 1987, 28, 5099) and coupling with the acid
fluoride derived from the amino acid by treatment with cyanuric
fluoride (L. A. Carpino, D. Sadat-Aalaee, H. G. Chao and R. H.
DeSelms, J. Am. Chem. Soc., 1990, 112, 9651).
7 H. Kunz, H. Waldmann and C. Unverzagt, Int. J. Pept. Protein Res.,
1985, 26, 493.
8 G. I. Feutrill and R. N. Mirrington, Aust. J. Chem., 1972, 25, 1731.
9 J. McMurry, Org. React., 1976, 24, 187.
10 P. Müller and B. Siegfried, Helv. Chim. Acta, 1974, 57, 987.
11 (a) A. Suzuki, Acc. Chem. Res., 1982, 15, 178; (b) A. R. Martin and
Y. Yang, Acta Chem. Scand., 1993, 47, 221.
12 A. Bax and D. G. Davis, J. Magn. Reson., 1985, 63, 207.
13 P. D. Henley and J. D. Kilburn, Chem. Commun., 1999, 1335.
14 Hostest 5 software: C. S. Wilcox and N. M. Glagovich, University of
Pittsburgh, USA, © C. S. Wilcox and the University of Pittsburgh,
1993. See also (a) C. S. Wilcox, Design, Synthesis and Evaluation of
an Efficacious Functional Group Dyad. Methods and Limitations in
the Use of NMR for Measuring Host–Guest Interactions, in Frontiers
in Supramolecular Organic Chemistry and Photochemistry, ed. H.-J.
Schneider and H. Durr, VCH, Weinheim, 1991; (b) C. S. Wilcox,
J. C. Adrian, Jr., T. H. Webb and F. J. Zawacki, J. Am. Chem. Soc.,
1992, 114, 10189.
(1S,6S,12S,17S)-6,12-Dibenzyl-5,8,10,13,18,24,26,32-octaza-
20,22,28,30(1,4)-tetrabenzena-9(2,6)-pyridinabicyclo[15.8.8]-
tritriacontaphan-4,7,11,14,19,25,27,33-octone 3
A solution of crude bis-TFA salt 20 (315 mg) in DMF (30 mL)
was added dropwise over 9 hours to a solution of DIPEA (0.60
mL) in MeCN (150 mL) at 70 ЊC. The mixture was concen-
trated under reduced pressure, the residue was dissolved in
EtOAc (100 mL) and the resulting solution washed with 0.5 M
HCl (2 × 20 mL), water (2 × 20 mL), brine (2 × 20 mL), dried
(MgSO4), filtered and concentrated under reduced pressure.
Purification by chromatography [CH2Cl2 to CH2Cl2–MeOH
(97:3 v/v)] gave a solid which was recrystallised from CHCl3 to
give macrobicycle 3 as a white solid (63 mg, 33% from diacid
18); mp 213–14 ЊC (CHCl3); [α]D Ϫ5.3 (c = 1.0 in DMF);
νmax(CHCl3)/cmϪ1 3323, 1684, 1652, 1635, 1604, 1544; δH(500
MHz, CDCl3) 8.16 (br s, 2 H, NHCOAr), 8.08 (br s, 2 H,
pyrNH), 7.82 (d, J 5, 2 H, pyrH), 7.73 (t, J 5, 1 H, pyrH), 7.60
(d, J 7, 4 H, ArH), 7.19 (m, 4 H, ArH), 7.16 (m, 2 H, ArH), 7.08
(d, J 7, 4 H, ArH), 6.99 (d, J 7, 4 H, ArH), 6.88 (d, J 7, 4 H,
ArH), 6.80 (d, J 7, 4 H, ArH), 6.48 (d, J 8, 2 H, NHCH2), 6.37
(br s, 2 H, NHCHCH2Ph), 4.85 (dd, J 9 and 15, 2 H, CHAHB-
NH), 4.73 (m, 2 H, NHCH(CH2)2), 4.35 (m, 2 H, CHCH2Ph),
3.83 (dd, J 3.0 and 15, 2 H, CHAHBNH), 3.77 (d, J 14, 2 H,
ArCHAHBAr), 3.63 (d, J 14, 2 H, ArCHAHBAr), 2.98 (dd,
J 5 and 13, 2 H, PhCHAHBCH), 2.74 (dd, J 7 and 13, 2 H,
PhCHAHBCH), 2.45 (m, 2 H, CHAHBCH2CO), 2.42 (m, 4 H,
CH2CH2CO), 2.20 (m, 2 H, CHAHBCH2CO); δC(90.6 MHz,
CDCl3) 174.9 (C), 171.1 (C), 169.0 (C), 167.6 (C), 149.3 (C),
145.6 (C), 141.2 (CH), 139.7 (C), 136.1 (C), 135.8 (C), 130.7
(C), 129.4 (CH), 129.3 (CH), 129.2 (CH), 129.1 (CH), 128.5
(CH), 128.0 (CH), 127.8 (CH), 110.1 (CH), 56.6 (CH), 53.7
(CH), 43.3 (CH2), 41.7 (CH2), 37.4 (CH2), 31.9 (CH2), 25.3
(CH2); MS (MALDI-TOF): m/z 1072.9 (MHϩ, 20%), 247.6
(67), 207.3 (92) [Found (FAB-MS) MHϩ = 1072.4880. C63H62-
N9O8 requires MH, 1072.4887] (Found C, 69.07; H, 5.62; N,
11.32. C63H61N9O8ؒH2O requires C, 69.41; H, 5.82; N, 11.56%).
Measurement of binding constants
All titration experiments were conducted on a JEOL GX 270 or
Bruker AM300 NMR machine at 293 K. In order to minimise
effects of acid impurities in the solvent on the position of the
NH signal, all deuteriochloroform used was passed through a
column of alumina prior to use and collected over molecular
sieves (4 Å, Aldrich). For the host, macrobicycle 3, 600 µL of a
solution of 10 mg in 3 mL deuteriochloroform (3.11 mM) was
15 As has been pointed out by a referee, the observed shifts in the
signals for the NH protons in many of the titration experiments are
quite small in comparison to the often large shifts reported for
1030
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031

View Article Online
amide–amide hydrogen bonding in chloroform solution, in many
other host–guest systems. In the present case this may be explained
by several factors, including i) the breaking of intramolecular
hydrogen bonds in 3 on complexation (as is particularly suggested
for NH5 by its downfield starting position in the NMR and by the
molecular modelling), ii) the general flexibility of the host–guest
system and consequent loose binding fit, as revealed by the NMR
and molecular modelling studies and iii) the macrobicyclic host 3
possesses C2 symmetry, so that while a hydrogen bond is only
formed with one of two symmetrically related protons, the NMR
reports the time-averaged signal for the hydrogen-bonded and non-
hydrogen bonded partner.
acid may be partly due to the difference in their acidities. However,
the differences in acidities, in chloroform solution, of closely related
amino acid substrates are likely to be small and are unlikely to lead
to differences in the interaction with the amidopyridine unit which
exceed the error in the experimentally determined binding constants
and clearly for enantiomers of the same substrate the acidities will
be the same.
18 (a) C. T. Walsh, S. L. Fisher, I.-S. Park, M. Prahalad and Z. Wu,
Chem. Biol., 1996, 3, 21; (b) R. J. Dancer, A. C. Try, G. J. Sharman
and D. H. Williams, Chem. Commun., 1996, 1445.
19 M. Rance, J. Magn. Reson., 1987, 74, 557.
20 F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp,
M. Lipton, C. Caufield, G. Chang, T. Hendrickson and W. C. Still,
J. Comput. Chem., 1990, 11, 440.
16 In the case of the titration experiment with what turned out to be
the most tightly bound substrate, Cbz-β-alanyl--alanine, we did
monitor the shift of the phenylalanyl α-proton (∆δ = 0.4) and the
aromatic rim proton ortho to the carboxamide (∆δ = Ϫ0.12).
Analysis of the resulting data gave association constants in agree-
ment with those determined from monitoring the amidopyridine
NH, but the quality of the fit for the data from the phenylalanyl
α-proton and aromatic rim proton was not as good, due to loss of
resolution and the presence of overlapping signals from both host
and guest protons as the titration progressed.
21 D. Q. McDonald and W. C. Still, Tetrahedron Lett., 1992, 33, 7742.
22 (a) W. D. Cornell, P. Ceiplak, C. I. Baoyly, I. R. Gould, K. M. Merz,
Jr., D. M. Ferguson, D. C. Spellmeyer, T. Fox, J. W. Caldwell and
P. A. Kollman, J. Am. Chem. Soc., 1995, 117, 5179; (b) S. J. Weiner,
P. A. Kollman, D. A. Case, U. C. Singh, C. Chio, G. Alagona,
S. Profeta, Jr. and P. Weiner, J. Am. Chem. Soc., 1984, 106, 765;
(c) S. J. Weiner, P. A. Kollman and D. A. Case, J. Comput. Chem.,
1986, 7, 230.
17 Comparison of binding constants of structurally different
carboxylic acids should be treated with some caution since the
strength of the interaction with the amidopyridine unit will be
influenced by the acidity of the carboxylic acid moiety (see e.g.,
M. H. Abraham, Chem. Soc. Rev., 1993, 22, 73; see also S. Shan,
S. Loh and D. Herschlag, Science, 1996, 272, 97). Thus, for example,
the difference in binding strength of hexanoic and phenylacetic
23 W. C. Still, A. Tempczyk and R. C. Hawley, J. Am. Chem. Soc.,
1990, 112, 6127.
24 D. D. Perrin and W. L. F. Armarego, Purification of Laboratory
Chemicals, Pergamon Press, Oxford, 3rd edn., 1989.
Paper a908090b
J. Chem. Soc., Perkin Trans. 1, 2000, 1021–1031
1031