Combinatorial libraries of diamidopyridine-derived 'tweezer' receptors and sequence selective binding of peptides

Article abstract of DOI:10.1016/S0040-4020(01)01111-5

A library of diamidopyridine derived tweezer receptors with peptidic side-arms has been prepared. Screening experiments with the library and a dye-labelled tripeptide L-Glu-L-Ser-L-Val-OH showed excellent selectivity (strong binding of the dye-labelled guest to ～0.2% of resin-bound library members was observed) and led to the identification of a tweezer receptor structure, which was shown to bind the tripeptide with good selectivity over related tripeptide sequences. Analogous screening experiments with the side-chain protected tripeptide L-Glu(O'Bu)-L-Ser('Bu)-L-Val-OH, showed little selectivity for any particular library member. The results are discussed in relation to previous screening experiments, using the same tripeptide guests, with related tweezer receptor libraries, which differ by one methylene in the separation of the diamidopyridine unit from the peptidic side-arms of the tweezer receptors.

Full text of DOI:10.1016/S0040-4020(01)01111-5

TETRAHEDRON
Pergamon
Tetrahedron 58 (2002) 711±719
Combinatorial libraries of diamidopyridine-derived `tweezer'
receptors and sequence selective binding of peptides
Rosa Arienzo and Jeremy D. Kilburnp
Department of Chemistry, University of Southampton, Southampton, Hants SO17 1BJ, UK
Received 5 June 2001; revised 2 August 2001; accepted 3 August 2001
AbstractÐA library of diamidopyridine derived tweezer receptors with peptidic side-arms has been prepared. Screening experiments with
the library and a dye-labelled tripeptide l-Glu±l-Ser±l-Val±OH showed excellent selectivity (strong binding of the dye-labelled guest to
,0.2% of resin-bound library members was observed) and led to the identi®cation of a tweezer receptor structure, which was shown to bind
the tripeptide with good selectivity over related tripeptide sequences. Analogous screening experiments with the side-chain protected
tripeptide l-Glu(OtBu)±l-Ser(tBu)±l-Val±OH, showed little selectivity for any particular library member. The results are discussed in
relation to previous screening experiments, using the same tripeptide guests, with related tweezer receptor libraries, which differ by one
methylene in the separation of the diamidopyridine unit from the peptidic side-arms of the tweezer receptors. q 2002 Elsevier Science Ltd.
All rights reserved.
1. Introduction
selective recognition of many biological molecules remains
a considerable challenge.
Since the beginnings of supramolecular chemistry, chemists
have sought to design and synthesise receptors capable of
binding selectively to speci®c substrates.1 In many cases
receptors for biological substrates have been targeted, in
part driven by the desire simply to mimic nature's own
ability to produce highly selective receptors, but also to
provide model systems and insights into natural molecular
recognition systems, and because of the practical appli-
cations that such receptors could haveÐe.g. as sensors,
therapeutics and catalysts. The ability to design a receptor
rationally, based on our ever-increasing understanding of
the non-covalent interactions responsible for molecular
recognition, and using the huge range of sophisticated
methodology now available to the synthetic chemist, has
always been an intellectually satisfying challenge. Much
of our own research over the last 10 years has focused on
the development of receptors for peptides, particularly for
peptides with a free carboxylic acid terminus. We have had
considerable success in this area, developing individual
macrocyclic receptors, which incorporate a speci®c recog-
nition site for the carboxylic acids2 or carboxylates,3 which
show good sequence selectivity (e.g. for the important
Ala±Ala±OH dipeptide sequence2b). However, as most
practitioners in the ®eld have found, the rational design
and synthesis of novel receptorsÐand a stepwise, iterative
approach to producing increasingly selective receptorsÐ
can be both time-consuming and frustrating, and the
In the 1990s, an alternative to the iterative approach to the
identi®cation of peptide receptors was pioneered by Still
and co-workers.4 Mirroring developments in medicinal
chemistry, the combinatorial approach was used to prepare
large libraries of possible receptor structures, which could
then be screened to identify a receptor for a given substrate.
In practice, Still et al. used a steroidal core to attach two
variable peptide strands to produce libraries, which were
screened to identify receptors for enkephalin-like peptides.4
Since the original publications by Still there have been a few
reports of the approach being used by other research
groups,5 and the reverse processÐi.e. screening a single
receptor with a library of peptide guests to identify possible
substrates for the receptorÐhas been used quite exten-
sively.6 The latter strategy has been applied particularly to
the identi®cation of peptide substrates for acyclic `tweezer'
receptors or `two-armed' receptors.7 The structure of such
tweezer receptors generally consists of a `head group' or
`hinge', which is typically a conformationally restricted
moiety that directs two functionalised substrate-binding
arms. The advantage of such structures is that they are
relatively easy to synthesise (in comparison with more
conformationally restricted macrocyclic structures) and
the work of several groups has demonstrated that, despite
their conformational ¯exibility, such structures are capable
of peptide recognition, both in organic4,5a±c,f,6c±h and
aqueous5c,6b solvents. Whereas in many of these systems
the head group plays only a limited role in the binding of
the guest, we became interested in developing tweezer
receptors5a,b,6b which incorporated a head group that can
Keywords: tweezer; receptor; peptide; combinatorial; library.
Corresponding author. Tel.: 144-2380-593596; fax: 144-2380-596805;
e-mail: jdk1@soton.ac.uk
p
0040±4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(01)01111-5

712
R. Arienzo, J. D. Kilburn / Tetrahedron 58 (2002) 711±719
tweezers identi®ed from the screening process could be
resynthesised, and binding studies revealed that the
tweezers bind peptides with binding constants 104±
105 M21 in CHCl3. In order to investigate these systems in
more detail, and to probe the effect of small structural
changes to the library, we decided to create similar libraries,
but using a diamidopyridine 5 with two b-alanine units,
instead of phenylalanine units, as the starting point for the
generation of the variable peptidic arms. By so doing the
peptidic arms of the tweezer structure would be one
methylene unit further removed from the CBS provided
by the diamidopyridine. In this paper, we describe the
results obtained in this new investigation.
Figure 1. Schematic of tweezer receptor with a CBS.
bind speci®cally to carboxylic acid functionality (in analogy
to our earlier macrocyclic peptide receptors), and with two
peptidic arms. Such tweezers should be receptors for
peptides with a carboxylic acid terminus, and it was antici-
pated that the peptidic arms might form (parallel) b-sheet-
like interactions with the backbone of the peptidic guest
(Fig. 1).
2. Synthesis
For the synthesis of the new tweezer libraries the previously
described5a Boc protected diaminopyridine 3 was prepared
in six steps from chelidamic acid (Scheme 2). After removal
of the Boc protecting groups, the resulting diaminopyridine
was preactivated with bis(trimethylsilyl)acetamide (BTSA)
and then coupled with N-Boc b-alanine acid ¯uoride, using
a methyl trimethylsilyl dimethylketene acetal (MTDA)8 as a
scavenger, to give the diamidopyridine 4 in excellent yield.
Hydrogenolysis of the benzyl ester gave the corresponding
acid 5, suitable for attachment to the solid phase.
In pursuit of this idea, we have described the synthesis
and screening of libraries of tweezer receptors, such as 2,
featuring a diamidopyridine as a carboxylic acid binding
site (CBS).5a,b A diamidopyridine 1, suitable for attachment
to the solid phase, was prepared and incorporated two
phenylalanine units as the starting point for the generation
of the variable peptidic arms (Scheme 1).
The libraries were successfully screened with various dye-
labelled tripeptides to identify tweezer receptors. Individual
Acid 5 was attached to TentaGel resin (Rapp Polymere, 130
mm) prepared with a 10% loading of N-Boc phenylalanine
Scheme 1.
Scheme 2.

R. Arienzo, J. D. Kilburn / Tetrahedron 58 (2002) 711±719
713
Scheme 3.
to serve as a coding strand.9 Removal of the Boc protecting
groups yielded a resin 6 ready for library generation.
In a typical screening experiment a sample of ,10 mg
(,9000 beads) of the library was equilibrated in a chosen
solvent system for 24 h. Dye-labelled tripeptide, as a solu-
tion in the same solvent system, was added and incubated
for a further 24 h. Beads were analysed in ¯at-bottomed
glass pots under a Leica inverted DML microscope (mag-
ni®cation, 40£). The concentration of dye-labelled peptide
guest could be increased to provide optimal selectivity as
adjudged by the number of highly stained beads against a
background of non- or lightly stained beads. Control experi-
ments were carried out as in previous screening studies.5a,b
Thus, we incubated the peptide guests with a simple peptide
library directly attached to TentaGel resin (analogous to the
coding strand). No evidence for any selective binding was
observed in these experiments. Thus we conclude that any
observed binding selectivity (of the dye-labelled guest
peptide for the resin-bound library members) is not a conse-
quence of interaction of the tripeptide guest simply with the
peptide side arm of the tweezer receptor, or with the coding
strand on the library beads. Furthermore, the simple acetyl-
ated red-dye 10 was incubated with the tweezer library, but
again no selective binding to any of the resin-bound library
A 1728 membered library of Fmoc protected tweezers 7 was
prepared by a three-fold coupling of 12 Fmoc protected
amino acids (Gly, l-Ala, l-Val, l-Phe, l-Leu, l-Lys(Boc),
l-Pro, l-Glu(OtBu), l-Ser(tBu), l-Met, l-Trp, l-Gln) to the
free amine groups of 6, using the `split and mix' strategy10
(Scheme 3).
3. Screening experiments and binding studies
Screening experiments were carried out with the tweezer
library using the dye-labelled peptide guests Red dye-
spacer±l-Glu(OtBu)±l-Ser(tBu)±l-Val±OH 8, and Red
dye-spacer±l-Glu±l-Ser±l-Val±OH 9 (Fig. 2). These
peptides were chosen as they had been used in our previous
studies5a,b with the earlier libraries, and would allow a direct
comparison of the screening results from both libraries 2
and 7.
Figure 2. Peptides used in the screening experiments with library 2.

714
R. Arienzo, J. D. Kilburn / Tetrahedron 58 (2002) 711±719
Table 2. Sequencing data for ®ve dyed beads selected from screening
experiment of Red dye-spacer±l-Glu±l-Ser±l-Val±OH 9 with library 2
in CHCl3
Bead
AA1
AA2
AA3
1
2
3
4
5
Ala
Ala
Gln
Gln
Met
Leu
Leu
Leu
Val
Val
Ala
Ala
Ala
Ala
Ala
third position for four beads, with the sequence Gln±xxx±
Gln clearly favoured, and Val as the most common residue
at the second position (Table 1).
In comparison, previous screening experiments with peptide
9 and library 2 had given reasonable binding selectivity
(,1% of beads were highly stained in the screening experi-
ment),5a but not as high as that observed for peptide 9
with library 7. The sequences identi®ed from the former
screening experiment (Table 2) have some similarities
with the sequences identi®ed from the latter (Table 1).
Thus for peptide 9 and library 2, Gln was found at the
®rst position for two out of the ®ve beads sequenced and
Val, or structurally very similar Leu, was found consistently
at the second position. At the third position, however, Ala
was found for each bead. The similarity in sequences, at the
®rst and second position, suggest that binding of the peptide
9, by tweezers from libraries 2 and 7, involve similar
interactions with the tweezer side-arms.
Figure 3. Portion of library 7, after incubation with dye-labelled peptide 9.
members was observed, con®rming that the selectivity
observed with the dye-labelled peptide guests are not a
consequence of binding to the dye moiety alone.
In contrast no binding selectivity was observed when library
7 was incubated with the protected peptide Red dye-spacer±
l-Glu(OtBu)±l-Ser(tBu)±l-Val±OH 8 at a peptide concen-
tration of 18 mM, and even at a peptide concentration
of 100 mM only slight selectivity could be observed by
inspection of the beads under the microscope.
When library 7 was incubated with the deprotected peptide
red dye-spacer±l-Glu±l-Ser±l-Val±OH
9
in 1%
DMSO:99% chloroform excellent binding selectivity was
observed. Thus, at a peptide concentration of 18 mM,
,0.2% of beads were highly coloured after a 24 h incu-
bation (Fig. 3). It was essential that the receptor library1
peptide guest was left for .12 h to reach equilibrium,
indicating that binding with the resin-bound receptors has
relatively slow kinetics. The observed binding selectivity
was also very solvent dependent as no beads remained
highly stained after addition of 10% DMSO.
These latter results are in marked contrast to the screening
results obtained previously using the same peptide guest, 8,
with receptor library 2.5a In that case excellent selectivity
was observed and four beads were sequenced. Three beads
gave identical results for the tweezer side-arm structure:
Val±Ala±Pro; and the fourth bead yielded the closely
related sequence Val±Met±Pro (Table 3).
Seven highly coloured beads were taken from the screening
experiment and sequenced by Edman degradation to
identify the structures of the tweezer receptors on each
bead. The sequencing results gave a high level of consensus
with glutamine at the ®rst position for ®ve beads and at the
As in previous studies, we sought to establish that the
observed binding with resin bound tweezers was also
operating in free solution. Thus tweezer 13 was prepared
by solution-phase synthesis (Scheme 4), as a single
compound, with tetrapeptide side-arms b-Ala±Gln±Val±
Gln, corresponding to the most commonly found sequence
in the screening experiments.
Table 1. Sequencing data for seven dyed beads selected from screening
experiment of Red dye-spacer±l-Glu±l-Ser±l-Val±OH 9 with library 7 in
CHCl3
Table 3. Sequencing data for four dyed beads selected from screening
experiment of Red dye-spacer±l-Glu(OtBu)±l-Ser(tBu)±l-Val±OH 8
with library 2 in CHCl3
Bead
AA1
AA2
AA3
1
2
3
4
5
6
7
Gln
Gln
Gln
Gln
Gln
Val
Val
Val
Val
Val
Lys
Lys
Ala
Gln
Gln
Gln
Gln
Gln
Gln
Glu(OtBu)
Leu
Bead
AA1
AA2
AA3
1
2
3
4
Val
Val
Val
Val
Ala
Ala
Ala
Met
Pro
Pro
Pro
Pro

R. Arienzo, J. D. Kilburn / Tetrahedron 58 (2002) 711±719
715
Scheme 4.
The synthesis of 13, using sequential coupling of Boc-amino
acids and TFA deprotection, was straightforward since
each of the intermediates could be isolated cleanly by
precipitation, without use of chromatography. However,
and not too surprisingly, the ®nal compound 13 proved to
be completely insoluble in CDCl3, which precluded any
binding studies in this solvent. Tweezer 13 could be
dissolved in DMSO, but there was no evidence for any
binding in this solvent, which is consistent with the
screening experiments where addition of as little as
10% DMSO to the chloroform solution resulted in a
complete loss of binding. Derivatives 14 and 15 of the
tweezer receptor were also prepared, incorporating a
Glu±d-Ser±d-Val±OH, and the extent to which the peptides
were absorbed onto the resin beads was determined.11 Thus
the receptor was resynthesised on solid phase, to give 16
with an estimated loading of 0.05 mmol of receptor/g resin
(Scheme 5).
Addition of 3.0 mg of resin-bound tweezer receptor 16
(,0.15 mmol of receptor) to a solution of Red dye-
spacer±l-Glu±l-Ser±l-Val±OH
9 (0.5 mL, 0.3 mM,
,0.15 mmol of peptide) in CHCl3 led to absorption of
90% (^5%) of the peptide as adjudged by HPLC (using
the acetylated red dye 10 as an internal standard) after
24 h incubation. (After only 2 h the amount of peptide
absorbed by the resin was ,50%, again indicating the rela-
tively slow rate at which binding equilibrium is reached
with the resin-bound receptors). Since the addition of
acetyl-capped tentagel resin beads led to no signi®cant
absorption of the peptides or acetylated red dye 10, this
provides an estimate12 of the binding constant of
Kass,3£105 M21. Addition of the same amount of resin-
bound tweezer receptor (,0.15 mmol of receptor) to an
equivalent solution of the side-chain protected peptide
8 (Red dye-spacer±l-Glu(OtBu)±l-Ser(tBu)±l-Val±OH,
long (dodecyl)alkyl chain or
a
diethyleneglycol
hexyl ether, but these derivatives were also insoluble in
CHCl3.
As an alternative to binding studies with the tweezer recep-
tor in free solution, we established that the resin bound
receptor 16, with the same tetrapeptide side-arms, b-Ala±
Gln±Val±Gln, as 13, is genuinely selective. Thus portions
of the resin-bound tweezer were added to solutions of the
peptides 8 and 9, and of the enantiomer Red dye-spacer±d-
Scheme 5.

716
R. Arienzo, J. D. Kilburn / Tetrahedron 58 (2002) 711±719
Figure 4.
0.5 mL, 0.3 mM, ,0.15 mmol of peptide) led to a
absorption of ,20% (^3%) of the peptide, giving an
estimated Kass,1£103 M21. Thus it can be concluded that
the resin-bound tweezer receptor 16 binds the peptide 9
.100 times more strongly than side-chain protected ana-
logue 8, which is consistent with the screening experiments.
Identical experiments with the enantiomeric peptide Red
dye-spacer±d-Glu±d-Ser±d-Val±OH, led to absorption
of ,55% (^10%) of the peptide, giving an estimated
Kass,9£103 M21, suggesting that resin-bound tweezer
receptor 13 binds the l-con®gured peptide 9 ,30 times
more strongly than the enantiomer.
alignment of the backbone of the peptide guest in relation
to the side-arms of the receptor, and no selective binding is
observed.
Binding of the deprotected peptide 9, on the other hand, is
complicated by the fact that either the terminal carboxylic
acid moiety of the peptide or the side-chain carboxylic acid
moiety from the glutamic acid may bind to the amido-
pyridine. If the glutamic acid side-chain carboxylic acid is
used, the link from the carboxylic acid unit to the rest of the
peptide backbone is both longer and more ¯exible, such that
the increased separation of the amidopyridine unit and the
receptor side-arms in library 7 may now be more appro-
priate for binding the peptide guest, but the increased
¯exibility of the side-chain carboxylic acid also allows
binding with tweezers from library 2 (Fig. 4b). Unfortu-
nately, although we have been able to prepare individual
tweezer receptors identi®ed from the screening experi-
ments, the poor solubility of the receptors off-resin, and in
the non-polar solvents suitable for binding (i.e. chloroform)
has precluded any detailed studies (e.g. by NMR) on the
structure of the receptor-substrate complex, so that the
proposed differences in binding can not be con®rmed
experimentally. Indeed, in the present case the complete
insolubility of tweezer receptor 13 in chloroform means
that no binding in free solution can be demonstrated. This
is clearly a limitation of this type of receptor and begs
the question as to whether the binding properties of
these systems, while resin-bound, is in¯uenced by the
environment created by the solid support.
4. Discussion
The screening of the combinatorial library of diamido-
pyridine derived tweezer receptors 7 led to the identi®cation
of a relatively limited number of tweezer structures from a
sizeable library, which appear to bind the chosen peptide
guest 9, while attached to the solid-phase. The selective
binding properties of one of these receptors 16, while
attached to the solid-phase, is demonstrated by the ability
to discriminate between peptide 9 and the side-chain
protected analogue 8 by a factor of ,100:1, and by the
ability to discriminate between peptide 9 and its enantiomer
by a factor of ,30:1. The screening results using the same
peptide guest 9, with the two receptor libraries 2 and 7, give
sequences with clear similarities at the ®rst and second
positions, which suggest that there are similar interactions
between the backbone of the peptide guest and the tweezer
receptor side-arms. The side-chain protected peptide 8,
however, behaves very differently with the two libraries 2
and 7, with excellent selectivity observed in the screening
experiments with library 2 and no apparent selectivity in the
screening experiments with library 7. A possible expla-
nation for these observations is that, for the side-chain
protected peptide 8 with receptors derived from amido-
pyridine derivative 1 (library 2), binding involves a strong
amidopyridine±carboxylic acid interaction, which places
the backbone of the peptide guest in a suitable position to
bind with the tweezer side-arms of the receptorÐpossibly
using parallel b-sheet-like interactions (Fig. 4a). In
comparison, introduction of an additional methylene
group to the side-arm of the receptors (i.e. using library 7
derived from amidopyridine derivative 5), disrupts the
5. Conclusions and outlook
Our studies described here further demonstrate the power of
the combinatorial approach for identifying receptors for a
chosen substrate, but also expose some of the limitations of
the approach when using diamidopyridine derived tweezers.
It is clear, from this and earlier work,5a,b that such diamido-
pyridine derived tweezers, despite their inherent ¯exibility,
can act as receptors for speci®c peptide sequences, particu-
larly when attached to the solid-phase. These receptors may
therefore have applications, particularly in the development
of chromatographic supports for separation or puri®cation
of peptides. However, the diamidopyridine±carboxylic acid
interaction is probably too weak to promote peptide binding

R. Arienzo, J. D. Kilburn / Tetrahedron 58 (2002) 711±719
717
by the tweezer receptors other than in relatively non-polar
solvents, in which the receptors may have poor solubilityÐ
thus limiting their usefulness. There is also very little
information regarding the actual mode of binding either in
free solution or while the receptor is still attached to
the solid-phase. Studies in our laboratory are aimed at
answering the unresolved questions regarding the mode of
binding of this class of tweezer receptors, and we are also
evaluating the binding properties of monolayers of such
receptors while attached to a surface. However, it is clearly
desirable to extend the concept of screening receptor
libraries to libraries suitable for peptide recognition in
water. Successful realisation of this objective, and a clear
understanding of the structure of resulting tweezer-substrate
complexes formed in water, has considerable potential
for providing model systems of peptide±peptide inter-
oxycarbonyl)amino]-4-pyridyloxy}acetate 3 (1.5 g, 2.2
mmol) in CH2Cl2 (32 mL) and the resulting mixture was
stirred for 2 h. The solution was concentrated and the
trituration of the resulting oil with diethyl ether gave the
bis-TFA salt as a white solid. The salt was suspended in an
aqueous solution of potassium carbonate (10%) and
extracted with dichloromethane to give the free 2,6 diamino
pyridine. The diamine (0.74 g, 2.7 mmol) was dissolved in
dry acetonitrile (63 mL). After addition of N,O-bis(tri-
methylsilyl)acetamide (0.84 mL, 2.7 mmol) the reaction
mixture was strirred at room temperature for 2 h. N-tert-
Butoxycarbonyl-b-alanyl ¯uoride (1.58 g, 13.5 mmol) and
MTDA (1.68 mL, 13.5 mmol) were added and the reaction
mixture was stirred for 20 h at room temperature. The
solvent was removed under reduced pressure and the residue
was redissolved in dichloromethane (300 mL). The organic
layer was washed with an aqueous solution of sodium
hydrogencarbonate (5%, 300 mL) and water (300 mL) and
dried over magnesium sulfate. The solvent was evaporated
and the resulting residue was puri®ed by column chroma-
tography (dichloromethane12% methanol) to give diamido-
pyridine 4 as colourless foam (1.53 g, 92%). Analytical
data: 1H NMR (400 MHz, [D6] DMSO):10.00 (s, 2H,
NHCO), 7.39±7.33 (m, 7H, ArH, C4H2N), 6.79 (br s, 2H,
NHBoc), 5.21 (s, 2H, OCH2Ar), 4.89 (s, 2H, OCH2CO),
3.21 (q, Jˆ6.5 Hz, 4H, CH2NHBoc), 2.55, (t, Jˆ6.5 Hz,
4H, NHCOCH2), 1.36 (s, 18H, (CH3)3C); 13C NMR
(100 MHz, [D6] DMSO) dˆ170.5(0), 167.9(0), 166.3(0),
155.5(0), 151.3(0), 135.5(0), 128.4(1), 128.1(1), 128.0(1),
95.6(1), 79.2(0),77.6(0), 66.2(2), 64.5(2), 36.6(2), 36.2(2),
28.2(3); LRMS (ES1): m/z (%)ˆ616.2 (100) [M1H]1,
638.2 (10) [M1Na]1; HRMS calcd for C30H41N5O9
615.29043, Found 615.29035; elemental analysis calcd
(%) for C30H41N5O9´MeOH: C 57.48, H 7.00, N 10.81, O
23, Found: C 57.26, H 6.54, N 11.23.
actions and application to
a range of therapeutic
problems. Work in our laboratories is addressing this
question, using libraries of tweezer receptors incorporating
a guanidinium as the CBS,6b and results will be reported in
the near future.
6. Experimental
6.1. Materials and general methods
Whenever possible all solvents and reagents were puri®ed
according to literature procedures. Solvents for peptide
syntheses were purchased from Rathburn Chemicals,
È
HPLC grade solvents from Riedel-de-Haen. TentaGel-S-
NH2 resin was used as solid support in screening experi-
ments and in peptide syntheses and purchased from Rapp
Polymere, Germany. Peptide and library syntheses on solid
phase were performed in glass vessels with scinter frits or
polypropylene ®ltration tubes with polyethylene frits on a
Visiprep SPE Vacuum Manifold from Supelco. The reaction
containers were agitated either on a shaker (Stuart Scienti®c
Flash Shaker SF1) or on a blood tube rotator (Stuart
Scienti®c Blood Tube Rotator SB1). Thin layer chroma-
tography (TLC) was performed on aluminium-backed plates
Merck silica gel 60 F254. Sorbsil C60, 40±60 mesh silica was
used for column chromatography. Infrared spectra were
recorded on a Perkin±Elmer 1600 FT-IR spectrophotometer
or on a Bio-Rad FT-IR spectrometer. Proton NMR spectra
were obtained at 300 MHz on a Bruker AC 300 and at
400 MHz on a Bruker DPX 400. Carbon NMR spectra
were recorded at 75 MHz on a Bruker AC 300 and at
100 MHz on a Bruker DPX 400. Chemical shifts are
reported in ppm on the d-scale relatively to TMS as internal
standard or to the solvent signal used. Coupling constants
are given in Hz. The multiplicities of the signals were
determined using the distortionless enhancement by phase
transfer (DEPT) spectral editing technique. Mass spectra
were obtained on a VG analytical 70-250-SE normal
geometry double focussing mass spectrometer. All electro-
spray (ES) spectra were recorded on a Micromass Platform
quadrupole mass analyser with an ES ion source using
acetonitrile as solvent.
6.1.2.
2-[2,6-Bis(N-tert-butoxycarbonyl-b-alaninyl-4-
pyridyloxy)acetic acid (5). 10% Palladium on charcoal
(1.1 g) was added to a solution of benzyl ester 4 (1.22 g,
2 mmol) in ethanol (25 mL). The reaction mixture was
stirred vigorously for 18 h at room temperature under a
hydrogen atmosphere. The catalyst was separated by ®l-
tration through a plug of celite. Evaporation of the
solvent and drying of the residue at high vacuum
yielded the corresponding carboxylic acid 5 as a colourless
solid (1.05 g, 99%). Analytical data: mp 608C; IR
1
nmaxˆ3400, 1686.1, 1654.3, 1583, 1158.7, 848.5 cm21; H
NMR (400 MHz, [D6] DMSO): dˆ10.14 (s, 2H, NHCO),
7.32 (s, 2H, C4H2N), 6.81 (br s, 2H, NHBoc), 4.71 (s, 2H,
OCH2CO), 3.21 (4H, t, Jˆ6.5 Hz, CH2NHBoc), 2.55 (4H,
t, Jˆ6.5 Hz, NHCOCH2), 1.36 (18H, s, (CH3)3C); 13C NMR
(100 MHz, [D6] DMSO) dˆ171.0(0), 169.4(0), 167.0(0),
155.7(0), 151.1(0), 95.6(1), 77.9(0), 64.6(2), 38.2(2),
36.4(2), 28.5(3); LRMS (ES1): m/z (%)ˆ526.3 (100)
[M1H]1, 548.2 (20) [M1Na]1, 1051.4 (10) [2M1H]1;
HRMS calcd for C23H35N5O9: 525.24348, Found
525.24341
6.1.3. Synthesis of the tweezer receptors library. A solu-
tion of Boc±Phe (9.8 mg, 0.037 mmol), PyBOP (19.2 mg,
0.037 mmol), HOBt (6 mg, 0.037 mmol) in DMF (1 mL)
was added to Tentagel±NH2 resin (1.28 g, 0.37 mmol,
NH2), preswollen in DMF (2 mL) followed by DIPEA
6.1.1. Benzyl 2-[(2,6-bis(N-tert-butoxycarbonyl-b-alanyl-
amino)-4-pyridyloxy]acetate (4). TFA (6.4 mL) was added
dropwise to a solution of benzyl 2-{(2,6-bis[bis(tert-butyl-

718
R. Arienzo, J. D. Kilburn / Tetrahedron 58 (2002) 711±719
(29 mL, 0.166 mmol), and the resulting suspension was
shaken for 20 h. A solution of acid 5 (220 mg, 0.417
mmol), PyBOP (217 mg, 0.417 mmol), HOBt (56.3 mg,
0.417 mmol) in DMF (2 mL) was added to the resin,
followed by DIPEA (303 mL, 1.74 mmol) and the suspen-
sion was shaken for 20 h. Any remaining amine residues
were capped by treating the reisn with an excess of acetic
anhydride. The resin was ®ltered and washed with CH2Cl2
(3£5 mL), DMF (3£5 mL), CH2Cl2 (3£5 mL). A qualitative
ninhydrin test was negative. Subsequent Boc deprotection
was achieved with a solution of TFA (30%) in DCM for 2 h.
The resin was washed with CH2Cl2 (3£5 mL), DMF
(3£5 mL), CH2Cl2 (3£5 mL) and washed with a solution
of DIPEA (10%) in DCM, MeOH (3£5 mL) and Et2O
(5£5 mL) and dried in vacuo. The tweezer library was
then prepared, using three cycles of split and mix synthesis.
Thus the resulting resin was divided in 12 equal portions. To
each resin portion one of the following Fmoc amino acids
was added: l-Ala, l-Gln, l-Glu(OtBu), Gly, l-Leu,
l-Lys(Boc), l-Met, l-Phe, l-Pro, l-Ser(tBu), l-Trp, l-Val
(0.08 mmol amino acid per resin portion), along with HBTU
(30 mg, 0.08 mmol) and DIPEA (47 mL, 0.27 mmol) in
DMF (2 mL). The reaction mixtures were shaken for 18 h.
Qualitative ninhydrin test were carried out to check all
transformations were complete. The resin was mixed and
the terminal Fmoc-groups removed with 20% piperidine in
DMF. The resin was split again into 12 equal portions and
the procedure repeated twice in order to build up the tweezer
receptor library 7.
and the resulting mixture was stirred for 2 h. The solution
was concentrated and the trituration of the resulting oil with
diethyl ether gave the bis-TFA salt as a white solid, which
was dissolved in DMF (2 mL). Boc±Val (0.12 g, 0.55
mmol), HBTU (0.21 g, 0.55 mmol) and DIPEA (0.16 mL,
0.93 mmol) were added and the resulting mixture stirred for
18 h. The precipitate formed was isolated by centrifugation
to give the diamidopyridine 12 as a white solid (0.14 g,
1
73%). Analytical data: mp decompose .2108C; H NMR
(400 MHz, [D6] DMSO) dˆ10.04 (s, 2H, NHCO), 7.96 (br
s, 2H, br CH2CH2NHCO), 7.82 (d, Jˆ7 Hz, 2H, CHNHCO),
7.42±7.32 (m, 7H, ArH1pyrH), 7.19 (s, 2H, CONHaHb),
6.74 (m, 4H, NHBoc1CONHaHb), 5.21 (s, 2H, CH2Ar),
4.89 (s, 2H, OCH2CO), 4.22 (br s, 2H, CHCH(CH3)2),
3.79 (br s, 2H, CHCH2CH2CONH2), 3.3 (m, 4H,
NHCOCH2CH2), 2.57 (m, 4H, NHCOCH2.CH2), 2.05 (m,
4H, CH2CONH2), 1.94 (m, 2H, CH(CH3)2), 1.82 (m, 2H,
CHaHbCH2CONH2), 1.72 (m, 2H, CHaHbCH2CONH2), 1.38
(s, 18H, C(CH3)3), 0.82 (d, Jˆ6.5 Hz, 6H, CH(CH3)2), 0.8
(d, Jˆ6.5 Hz, 6H, CH(CH3)2); 13C NMR (100 MHz, [D6]
DMSO) dˆ173.7(0), 171.2(0), 170.6(0), 168.0(0),
166.4(0), 162.4(0), 155.6(0), 151.4(0), 135.6(0), 128.5(1),
128.2(1), 128.1(1), 95.7(1), 78.2(0), 64.7(2), 62.9(2),
57.1(1), 51.8(1), 35.6(2), 34.3(2), 31.0(2), 30.3(1),
30.1(2), 28.2(3), 19.3(3), 18.1(3); MS (ES1): m/z (%)ˆ
1070.1 (100) [M1H]1, 1091.9 (40) [M1Na]1.
6.1.6. Benzyl 2-{3,5-di[(3-{[5-amino-2-({2-[(5-amino-2-
{[(9H-9-¯uorenylmethoxy)carbonyl]amino}-5-oxopenta-
noyl)amino]-3-methylbutanoyl}amino)-5-oxopentanoyl]-
amino}propanoyl)amino]phenoxy}acetate (13). TFA
(1 mL) was added dropwise to a solution of the diamido-
pyridine 12 (0.08 g, 0.075 mmol) in CH2Cl2 (3 mL) and the
resulting mixture was stirred for 2 h. The solution was
concentrated and the trituration of the resulting oil with
diethyl ether gave the bis-TFA salt as a white solid, which
was dissolved in DMF (2 mL). Fmoc±Gln (0.06 g, 0.16
mmol), HBTU (0.06 g, 0.16 mmol) and DIPEA (0.05 mL,
0.3 mmol) were added and the resulting mixture stirred for
18 h. The precipitate formed was isolated by centrifugation
to give the diamidopyridine 13 as a white solid (0.08 g,
6.1.4. Benzyl 2-(3,5-di{[3-({5-amino-2-[tert-butoxycarbo-
nyl)amino]-5-oxopentanoyl}amino)propanoyl]amino}-
phenoxy)acetate (11). TFA (2 mL) was added dropwise to
a solution of the diamidopyridine 4 (0.16 g, 0.26 mmol) in
CH2Cl2 (4 mL) and the resulting mixture was stirred for 2 h.
The solution was concentrated and the trituration of the
resulting oil with diethyl ether gave the bis TFA salt as a
white solid, which was dissolved in DMF (2 mL). Boc±Gln
(0.196 g, 0.76 mmol), HBTU (0.3 g, 0.76mmol) and DIPEA
(0.23 mL, 1.32 mmol) were added and the resulting mixture
stirred for 18 h. The precipitate formed was isolated by
centrifugation to give the diamidopyridine 11 as a white
solid (0.16 g, 70%). Analytical data: mp .2408C; 1H
NMR (400 MHz, [D6] DMSO) dˆ10.05 (s, 2H, NHpyr),
7.82 (br s, 2H, NHCH2CH2), 7.41±7.31 (m, 7H, ArH1
pyrH), 7.22 (s, 2H, CONHaHb), 6.82 (d, Jˆ7.5 Hz, 2H,
NHBoc), 6.72 (s, 2H, CONHaHb), 5.21 (s, 2H, CH2Ar),
4.88 (s, 2H, OCH2CO), 3.84 (m, 2H, CHNHBoc), 3.38
(m, 4H, CH2CH2NH), 2.57 (m, 4H, CH2CH2NH), 2.07 (m,
4H, CH2CONH2), 1.81 (m, 2H, CHaHbCHNHBoc), 1.66 (m,
2H, CHaHbCHNHBoc), 1.35 (s, 18H, C(CH3)3). 13C NMR
(100 MHz, [D6] DMSO) dˆ173.2(0), 171.2(0), 170.1(0),
167.36(0), 165.7(0), 154.7(0), 150.8(0), 134.9(0), 127.8(1),
127.6(1), 127.4(1), 95.1(1), 77.5(0), 65.7(2), 63.9(2),
54.3(2), 53.5(1), 35.6(2), 34.2(2), 31.03(2), 27.6(3).
LRMS (ES1): m/z (%)ˆ872.3 (80) [M1H]1, 894.3 (10)
[M1Na]1.
1
70%). Analytical data: mp decompose .2408C; H NMR
(400 MHz, 313 K, [D6] DMSO) dˆ9.94 (s, 2H, pyrNH),
7.96 (d, Jˆ7.5 Hz, 2H, NH), 7.90 (d, 4H, Jˆ7.5 Hz,
¯uorenylH), 7.85 (t, Jˆ6 Hz, 2H, CH2CH2NHCO), 7.75±
7.70 (m, 6H, NH1¯uorenylH), 7.53 (d, 2H, NH), 7.45±7.32
(m, 15H, ArH1¯uorenylH1pyrH), 7.24 (s, 2H,
CONHaHb),7.16 (s, 2H, CONHcHd), 6.72 (s, 2H,
CONHaHb), 6.67 (s, 2H, CONHcHd), 5.10 (s, 2H,
OCH2Ph), 4.79 (s, 2H, OCH2CO), 4.16±4.07 (m, 10H,
CHCH(CH3)2, CHCH2CH2CONH2, CHCH2¯uorenyl), 3.94
(m, 2H, CHCH2CH2CONH2), 3.40 (m, 4H, CH2CH2NH),
2.20±1.70
(m,
14H,
CH2CH2NH1CH(CH3)21
CH2CH2CONH2), 0.85 (d, 6H, CH3), 0.83 (d, 6H,
CH3); 13C NMR (100 MHz, [D6] DMSO) dˆ173.9(0),
173.7(0), 171.1(0), 170.7(0), 170.5(0), 167.9(0), 166.3(0),
157.3(0), 155.9(0), 151.3(0),143.8(0), 140.7(0), 135.5(0),
128.4(1), 128.1(1), 128.0(0), 127.6(0), 127.1(1), 125.3(1),
120.1(0), 95.7(1), 66.2(2), 65.7(2), 64.5(2), 57.5(1),
54.3(1), 52.3(1), 46.6(1), 36.1(2), 34.7(2), 31.6(2),
31.4(2), 30.5(1), 27.9(2), 27.7(2), 19.2(3), 17.8(3), MS
(TOF LD1): m/z (%)ˆ 1571.07 (100) [M1H]1, 3301.65
(30) [2M1H]1.
6.1.5. Benzyl 2-(3,5-di[(3-{[5-amino-2-({2-[(tert-butoxy-
carbonyl)amino]-3-methylbutanoyl}amino)-5-oxopenta-
noyl]amino}propanoyl)amino]phenoxy}acetate
(12).
TFA (2 mL) was added dropwise to a solution of the
diamidopyridine 2 (0.16 g, 0.18 mmol) in CH2Cl2 (4 mL)

R. Arienzo, J. D. Kilburn / Tetrahedron 58 (2002) 711±719
719
8253±8256. (h) Gennari, C.; Nestler, H. P.; Salom, B.; Still,
W. C. Angew. Chem., Int. Ed. Engl. 1995, 34, 1763±1765.
7. For original work on tweezer receptors, see: (a) Chen, C.-W.;
Whitlock, Jr., H. W. J. Am. Chem. Soc. 1978, 100, 4921±4922.
(b) Zimmerman, S. C.; Wu, W. M.; Zeng, Z. J. J. Am. Chem.
Soc. 1991, 113, 196±201.
References
1. For recent reviews see: (a) Hartley, J. H.; James, T. D.; Ward,
C. J. J. Chem. Soc., Perkin Trans. 1 2000, 3155±3184.
(b) Peczuh, M. W.; Hamilton, A. D. Chem. Rev. 2000, 100,
2479±2493. (c) Davis, A. P.; Wareham, R. S. Angew. Chem.,
Int. Ed. Engl. 1999, 38, 2979±2996.
8. The use of MTDA as a scavenger provides better yields for the
acid ¯uoride couplings with diaminopyridine than reported in
our earlier procedures (Ref. 5a,b). For this modi®cation we are
indebted to Dr Enrique Botana, Dipartimento di Chimica
2. See for example: (a) Henley, P. D.; Waymark, C. P.; Gillies, I.;
Kilburn, J. D. J. Chem. Soc., Perkin Trans. 1 2000, 1021±
1031. (b) Henley, P. D.; Kilburn, J. D. Chem. Commun.
1999, 1335±1336. (c) Dowden, J.; Edwards, P. D.; Flack,
S. S.; Kilburn, J. D. Chem. Eur. J. 1999, 79±89.
Á
Organica e Industriale, Universita di Milano. See: (a) Botana,
E.; Ongeri, S.; Arienzo, R.; Demarcus, M.; Frey, J. G.;
Piarulli, U.; Potenza, D.; Gennari, C.; Kilburn, J. D. Chem.
Commun. 2001, 1358±1359. (b) Gennari, C.; Gude, M.;
Potenza, D.; Piarulli, U. Chem. Eur. J. 1998, 4, 1924±1931.
(c) Gude, M.; Piarulli, U.; Potenza, D.; Salom, B.; Gennari, C.
Tetrahedron Lett. 1996, 37, 8589±8592.
3. (a) Pernia, G. J.; Kilburn, J. D.; Essex, J. W.; Mortishire-
Smith, R. J.; Rowley, M. J. Am. Chem. Soc. 1996, 118,
10220±10227. (b) Pernia, G. J.; Kilburn, J. D.; Rowley, M.
J. Chem. Soc., Chem. Commun. 1995, 305±306.
4. (a) Cheng, Y.; Suenaga, T.; Still, W. C. J. Am. Chem. Soc.
1996, 118, 1813±1814. (b) Boyce, R.; Li, G.; Nestler, H. P.;
Suenaga, T.; Still, W. C. J. Am. Chem. Soc. 1994, 116, 7955±
7956.
9. Previous studies (Ref. 5a,b) have shown that a coding strand
is necessary with these diamidopyridine-derived tweezer
libraries because the conditions used for Edman sequencing
leads to cleavage of the amide bonds to the pyridine core,
leading to loss of the peptide side-arms.
5. (a) Braxmeier, T.; Demarcus, M.; Fessmann, T.; McAteer, S.;
Kilburn, J. D. Chem. Eur. J. 2001, 7, 1889±1898.
(b) Fessmann, T.; Kilburn, J. D. Angew. Chem., Int. Ed.
Engl. 1999, 38, 1993±1996. (c) Dong, D. L.; Liu, R. P.;
Sherlock, R.; Wigler, M. H.; Nestler, H. P. Chem. Biol.
1999, 6, 133±141. (d) Hioki, H.; Yamada, T.; Fujioka, C.;
Kodama, M. Tetrahedron Lett. 1999, 40, 6821±6825.
(e) Xu, R.; Greiveldinger, G.; Marenus, L. E.; Cooper, A.;
Ellman, J. A. J. Am. Chem. Soc. 1999, 121, 4898±4899.
(f) Nestler, H. P. Mol. Diversity 1996, 2, 35±40. (g) Sasaki,
S.; Takagi, M.; Tanaka, Y.; Maeda, M. Tetrahedron Lett.
1996, 37, 85±88.
10. (a) Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G. Int. J.
Pept. Prot. Res. 1991, 37, 487±493. (b) Lam, K. S.; Salmon,
S. E.; Hersh, E. M.; Hruby, V. J.; Kazmierski, W. M.; Knapp,
R. J. Nature 1991, 354, 82±84.
11. For a related approach to the quanti®cation of binding
constants for resin-bound receptors, see: Smith, P. W.;
Chang, G.; Still, W. C. J. Org. Chem. 1988, 53, 1587±1590.
12. With a starting concentration of peptide of 0.3 mM, and an
equivalent amount of resin-bound receptor present (0.15 mmol
in 0.5 mL solution), if y is the fraction of peptide absorbed,
then concentration of free guest [G]ˆ0.3(12y) mM; concen-
tration of host guest complex [HG]ˆ0.3y mM; concentration
of free host [H]ˆ0.3(12y) mM; and Kassˆ[HG]/[H][G]ˆy/
0.3(12y)2 mM21. Thus for peptide Red dye-spacer±l-Glu±
l-Ser±l-Val±OH 9, yˆ0.9 and Kass,3£105 M21. For peptide,
Red dye-spacer±l-Glu(OtBu)±l-Ser(tBu)±l-Val±OH 8,
yˆ0.2 and Kass,1£103 M21. For peptide Red dye-spacer±d-
6. (a) van Wageningen, A. M. A.; Liskamp, R. M. J. Tetrahedron
Lett. 1999, 40, 9347±9351. (b) Davies, M.; Bonnat, M.;
Guillier, F.; Kilburn, J. D.; Bradley, M. J. Org. Chem. 1998,
63, 8696±8703. (c) Lowik, D. W. P. M.; Weingarten, M. D.;
Broekema, M.; Brouwer, A. J.; Still, W. C.; Liskamp, R. M. J.
Angew. Chem., Int. Ed. Engl. 1998, 37, 1846±1850.
(d) Torneiro, M.; Still, W. C. Tetrahedron 1997, 53, 8739±
8750. (e) Still, W. C. Acc. Chem. Res. 1996, 29, 155±163.
(f) Shao, Y. F.; Still, W. C. J. Org. Chem. 1996, 61, 6086±
6087. (g) Lowik, D. W. P. M.; Mulders, S. J. E.; Cheng, Y.;
Shao, Y. F.; Liskamp, R. M. J. Tetrahedron Lett. 1996, 37,
Glu±d-Ser±d-Val±OH 9, yˆ0.55 and Kass,9£103 M21
.
These derived binding constants should only be considered
as estimates given the experimental errors in determining
the loading of the receptor on the resin, and in determining
the amount of peptide absorbed.