Substituent effects in synthetic lectins - Exploring the role of CH-π interactions in carbohydrate recognition

Article abstract of DOI:10.1021/jo200755z

Contacts between aromatic surfaces and saccharide CH groups are common motifs in natural carbohydrate recognition. These CH-π interactions are modeled in "synthetic lectins" which employ oligophenyl units as apolar surfaces. Here we report the synthesis a

Full text of DOI:10.1021/jo200755z

ARTICLE
pubs.acs.org/joc
Substituent Effects in Synthetic Lectins - Exploring the Role of CHꢀπ
Interactions in Carbohydrate Recognition
Nicholas P. Barwell and Anthony P. Davis*
School of Chemistry, University of Bristol, Cantock’s Close, Bristol, U.K. BS8 1TS
S Supporting Information
b
ABSTRACT: Contacts between aromatic surfaces and sacchar-
ide CH groups are common motifs in natural carbohydrate
recognition. These CHꢀπ interactions are modeled in “syn-
thetic lectins” which employ oligophenyl units as apolar sur-
faces. Here we report the synthesis and study of new synthetic
lectins with fluoro- and hydroxy-substituted biphenyl units,
designed to explore the role of π-electron density in carbohy-
drate CHꢀπ interactions. We find evidence that recognition
can be moderated through electronic effects but that other
factors such as cavity hydration are also important and some-
times predominant in determining binding strengths.
’ INTRODUCTION
The binding of carbohydrates in water by proteins such as
lectins is one of the less understood aspects of biomolecular
recognition.1 It is generally accepted that persuading saccharides
to displace water molecules from a binding site is intrinsically
difficult. The hydroxyl groups on carbohydrates have similar
binding properties to those of the solvent, and exchanging one
for the other is unlikely to yield much energetic benefit. Binding
free energy must be garnered from subtle differences in hydrogen
bonding patterns, the release of confined water molecules (raising
entropy), and noncovalent bonding to groups other than hydro-
xyl (e.g., the CH groups). The CHꢀπ interaction,2 in particular,
is thought to play an important role.3 Many saccharide units
possess patches of codirected CH groups which together can
make contact with a flat surface. Proteins often provide aromatic
surfaces to complement these patches, a well-known example
being the E. coli glucose/galactose chemoreceptor protein
(Figure 1).4 Such contacts should lead to the release of high-
energy water molecules, and thus hydrophobic binding, but there
is an additional more specific effect. Both theory and experiment
suggest that the electron-poor CH hydrogen atom should
interact favorably with the electron-rich surface of the π-system.
The attraction should increase with π-electron density,5 and it is
notable that the electron-rich tryptophan indole seems especially
widely used in nature.
Model studies can be useful in clarifying the role of specific
noncovalent interactions in biological processes. In the case of
carbohydrate CHꢀπ interactions, a number of investigations
have implied that increased π-electron density does indeed lead
to stronger binding.6 However, these have not included fully
synthetic lectin analogues in which the geometry of substrate
binding is well-established. We have recently developed a family
of receptors capable of binding carbohydrates in water with good
Figure 1. Binding site of the E. coli glucose/galactose chemoreceptor
protein, with glucose substrate.4 Aromatic tryptophan and phenylala-
nine residues are shown in blue, and polar residues in gold.
affinities and excellent selectivities.7,8 As illustrated in Figure 2,
the general design employs oligophenyl units, to define the roof
and floor of the binding cavity, and isophthalamide pillars, to
hold them apart. The cavities are complementary to “all-equa-
torial” carbohydrates such as β-linked glucose or N-acetylgluco-
samine (GlcNAc) 1, or oligomers such as cellobiose 4. The roof
and floor are positioned to contact the axial CH groups in the
substrates, contributing hydrophobic and CHꢀπ interactions to
binding. The isophthalamide spacers can bind to substrate OH or
Received: April 11, 2011
Published: July 06, 2011
r
2011 American Chemical Society
6548
dx.doi.org/10.1021/jo200755z J. Org. Chem. 2011, 76, 6548–6557
|

The Journal of Organic Chemistry
ARTICLE
Figure 2. Synthetic lectins for all-equatorial carbohydrates. Monosac-
charide receptor 2 is shown binding β-glucosyl 1 (Z = OH) or β-GlcNAc
1 (Z = NHAc). Disaccharide receptor 3 is similarly complementary to
cellobiose 4.
Figure 3. NMR/molecular modeling structures for the complexes between
(top) 2 and GlcNAcβ-OMe 1 (Z = NHAc, R = Me),8d and (bottom) 3 and
cellobiose 4(βanomer).8c The bi- and terphenyl units are highlighted in pale
cyan (space-filling mode), and the carbohydrates are shown as pink.
Intermolecular NOE contacts are shown as red broken lines. The water-
solubilizing tricarboxylate groups are omitted. The structure of 2.1 (Z =
NHAc, R = Me) is supported by 51 NOE contacts and is essentially
unambiguous. The structure of 3.4 is supported by fewer cross-peaks, but the
positioning of the substrate is nonetheless certain.
NHAc groups through hydrogen bonding. NOE studies confirm
the geometry of binding and can give detailed structures for the
complexes (Figure 3).7bꢀe The performance of these systems can
be remarkably similar to that of lectins. For example, prototype 2
binds GlcNAc derivative 1(Z=NHAc, R=Me) withKa =650Mꢀ1
,
comparable to Wheat Germ Agglutinin (the lectin traditionally
binding in carbohydrate receptors but that differences are small
and may be masked by other effects.
used for GlcNAc moieties).7d
As plausible lectin models, receptors 2 and 3 have previously
been employed to study the hydrophobic effect in carbohydrate
recognition.9 The role of CHꢀπ interactions is another issue
which should, in principle, be addressable. As discussed above,
the binding geometries are clearly established such that there is
no doubt that contacts exist between carbohydrate CH groups
and the π systems of the oligophenyl units. By introducing
substituents into the oligophenyls, it should be possible to vary
the electron density in their π-systems. The effects on binding
constants would then shed light on the nature of the interactions.
Meanwhile affinities would hopefully be raised in some cases, a
useful outcome from the viewpoint of applications.
’ RESULTS AND DISCUSSION
The opportunities for substitution in the oligophenyl units are
limited by steric effects; any group added ortho to an ArꢀAr
bond will promote a twisted conformation unsuitable for bind-
ing, while the positions between CH2 groups are quite hindered.
Nonetheless, small substituents will fit in these latter positions
without disruption to the structure. In previous work we prepared a
series of receptors 5 with alkoxy groups in the biphenyl 4,40
positions and were pleased to find that affinities for β-glucosides
were raised considerably.8e However interpretation of these
results in terms of π-electron density was complicated by the
hindered nature of the position, which forces the alkoxy group
out of plane and disrupts nꢀπ conjugation. We now report a
more informative study, featuring substituents which are electro-
nically more diverse and less affected by crowding. The results
suggest that π-electron density can indeed be used to modulate
Receptor Design and Synthesis. For this study we required
analogues of 2 in which the biphenyl 4,40 positions were occupied
by small, electronically active substituents. The groups chosen
were F, as in 6, and OH, as in 7. F was expected to reduce the
electron density in the biphenyl π-system, while OH is generally
seen as electron-donating (but see later).10 Moreover deproto-
nation is possible to give a third species 8, with greatly increased
π-electron density. Unlike the alkoxy groups in 5, the hydroxyls
in 7 can exert their normal effect, being able to lie coplanar to the
benzene ring such that the oxygen lone pairs can interact optimally
with the π-cloud. Molecular modeling was employed to assess the
effects of the F, OH, and Oꢀ substituents on the receptor
conformations. Monte Carlo Molecular Mechanics searches11 on
unsubstituted receptor 2 and the new systems 6ꢀ8 revealed similar
6549
dx.doi.org/10.1021/jo200755z |J. Org. Chem. 2011, 76, 6548–6557

The Journal of Organic Chemistry
ARTICLE
ground state conformations with inward-directed NH groups and
twisted biphenyl units.12 Differences were subtle and seemed unlikely
to have major effects on the binding properties.
with NBS under irradiation with visible light to give tribromide
10. Conversion to diamine 12 was accomplished via Gabriel
synthesis. Two different N-protection methods gave 14 and 13,
which were coupled via SuzukiꢀMiyaura methodology to give
biphenyl 16. The overall yield of 16 from 9 was ∼17%. N-Boc
deprotection and macrocyclization with spacer component
188b under high dilution gave mainly the [2 + 2] product 19.
N-Deprotection, a second macrocylization with 18, and then
O-deprotection gave 6. The yields for the macrocyclizations
were low, at 17% and 25% respectively, but were sufficient to
provide material for testing.
Tetrahydroxy receptor 7 was prepared via biphenyl 23,
obtained from the previously described diol 218e (Scheme 2).12
The protection scheme in 23 was chosen to allow exposure of
two amino groups via N-Boc cleavage, unmasking of the second
two by reduction of the azides with PPh3/H2O, and then deprotec-
tion of the phenolic hydroxyls by hydrogenolysis. Cyclizations with
18 after the first and second deprotections gave macrotricycle 26.
Tetrafluoro receptor 6 was prepared as shown in Scheme 1.12
Commercially available 5-bromo-2-fluoro-m-xylene 9 was treated
Scheme 1a
a Reagents and conditions: (a) NBS, methyl acetate, hυ, 62%; (b) potassium phthalimide, DMF, 100%; (c) hydrazine, chloroform/methanol (1:4),
reflux, 96%; (d) Boc2O, DIPEA, THF, 88%; (e) Benzyl chloroformate, DIPEA, THF, 52%; (f) bispinacolato diboron, Pd(dppf)Cl2 (3 mol %), KOAc
(4 equiv), DMF, 80 °C, 37%; (g) Pd(dppf)Cl2 (5 mol %), Na2CO3 (aq., 2 M), DMSO, 60 °C, 86%; (h) TFA, DCM, 0 °C then 18, DIPEA, THF, 17%;
(i) Pd/C, H2, THF/methanol/NH3; (j) 18, DIPEA, THF, 25%; (k) TFA, DCM, 0 °C, 95%.
6550
dx.doi.org/10.1021/jo200755z |J. Org. Chem. 2011, 76, 6548–6557

The Journal of Organic Chemistry
ARTICLE
Scheme 2a
a Reagents and conditions: (a) CBr4, PPh3, DCM, 73%; (b) NaN3, DMF, 60 °C, 100%; (c) TFA, DCM, 0 °C; (d) 18, DIPEA, THF, 29%; (e) PPh3,
THF, 60 °C then H2O, 60 °C, 94%; (f) 18, DIPEA, THF, 28%; (g) Pd/C, H2, THF/methanol/NH3, 91%; (h) TFA, DCM, 0 °C, 96%.
The hydrogenolysis of the benzyl groups proceeded cleanly in
91% yield. Finally, cleavage of the side-chain tert-butyl esters
gave 7. The deprotonation of the phenolic hydroxyls in 7 was
investigated by UV spectroscopy. Between pH = 6.5 and 8.5 the
UV spectrum remained constant, implying full protonation. As
the pH was raised further the spectrum changed, stabilizing at
pH = 11.5.12 We therefore presumed that the hydroxyl groups
would be completely deprotonated at this and higher pH values.
This is consistent with the behavior of 4,40-biphenol where the
pKa of both hydroxyl groups is around 10.13
Measurement and Analysis of Binding Properties. The
binding constants of 6ꢀ8 to glucose 28, galactose 29, mannose
30, methyl β-D-glucoside 31, and methyl R-D-glucoside 32 were
measured by 1H NMR titrations in D2O.12 For the titrations on 7
the pH was adjusted to 7.5 by addition of sodium hydroxide. For
protons also moved substantially and could be readily analyzed.
Where two signals were analyzed, an average value was calculated
unless one seemed obviously less accurate. A typical set of
spectra with the corresponding analysis plot are shown in
Figure 4. In the great majority of cases, errors were estimated
at (5% or better. For favorable substrates such as glucose and
methyl β-D-glucoside, the limiting spectra were remarkably
similar to each other and to those previously observed for 2
and 5. We therefore infer that the binding geometries for all
these tricyclic receptors are essentially similar, involving the
carbohydrate entering the cavity as depicted in Figure 3.14
To assist in the interpretation of the binding data, the surface
charge distributions were calculated for tetramethyl biphenyls
33ꢀ36, models for the biaryl units in 2 and 6ꢀ8 respectively.
Surface charge is expected to correlate quite well with the relative
strengths of CHꢀπ interactions.15 The results for 33ꢀ35 are
depicted in Figure 5; the surface of 36 is far more negative than
the others and cannot be usefully represented on the same color
scale. Some aspects of the charge distributions may seem
unexpected. Although hydroxyl groups are normally considered
electron-donating, the π-surface of dihydroxybiphenyl 35 is very
similar to that of the parent 33. While hydroxyl groups promote
electrophilic attack through π-donation, their effect on the
ground state charge distribution is almost neutral (due to
opposing σ and π effects). Nonetheless, the series covers a good
range of electron densities. Difluoride 34 is significantly
the titrations on 8, further sodium hydroxide was added to bring
the pH to 12. As previously observed for 2 and 5, addition of the
carbohydrates caused movements of signals due to receptor
aromatic protons. The motions were consistent with 1:1 binding
and were analyzed by nonlinear curve-fitting to obtain association
constants Ka. In the case of 6 and 7, the signals due to the internally
directed isophthalamide protons moved the greatest amounts and
were used to determine Ka values. In the case of 8 the biphenyl
6551
dx.doi.org/10.1021/jo200755z |J. Org. Chem. 2011, 76, 6548–6557

The Journal of Organic Chemistry
ARTICLE
Figure 4. (Top) Partial 1H NMR spectra from the addition of 28 to 7 in D2O (pH = 7.5). Peak B is the internally directed isophthalamide proton; for
other assignments see Figure S1. (Bottom) Binding analysis based on the movements of peak B, assuming 1:1 complex formation. Experimental and
calculated values are shown. Ka = 43 Mꢀ1. Limiting Δδ = 0.348 ppm.
more electron-poor than 33 or 35, while 36 is much more
electron-rich. The strengths of CHꢀπ interactions involving 2
and 6ꢀ8 should therefore increase in the order 6 < 2 ≈ 7 < 8.
Overall the order of binding strengths for preferred substrates
28 and 31 may be summarized as 2 < 6 ≈ 8 < 7 (i.e., H <
F ≈ Oꢀ < OH).
The data can be rationalized by assuming that the CHꢀπ
effect is operative and can be controlled by varying the electron
density, but that differences are small and easily overwhelmed by
other factors. Thus F and OH are sterically similar but electro-
nically different, providing arguably the best test for the role of
π-electron density. As expected, tetrafluoride 6 is significantly
weaker than tetraol 7. The difference is modest (a factor of
∼2), but this is not too surprising. Modeling of the complex
between β-glucose and 7 suggests that only two CH groups may
be well-positioned to contact the aromatic π-surfaces (the
remainder are closer to aromatic hydrogens; see Figure 6).
Calculations imply that the CHꢀπ interactions to ArꢀO units
should be ∼1.5 kJmolꢀ1 stronger than those to ArꢀF units.5a
The binding constants measured for the receptors to carbo-
hydrates 28ꢀ32 are listed in Table 1. All showed the expected
selectivity for all-equatorial substrates, binding glucose 2816 in
preference to galactose 29 or mannose 30, and methyl β-D-
glucoside 31 more strongly than R anomer 32. Binding
strengths varied significantly, by factors of 5 and 6 for sub-
strates 28 and 31 respectively. The sequence, however, was not
as expected on the basis of CHꢀπ effects alone. Thus, all the
new receptors, including tetrafluoride 6, proved stronger than
prototype 2. Although unanticipated this result is potentially
useful. Indeed, tetraol 7 is among the most powerful mono-
saccharide receptors we have studied. On the other hand
the exceptional affinities expected for 8 did not materialize.
6552
dx.doi.org/10.1021/jo200755z |J. Org. Chem. 2011, 76, 6548–6557

The Journal of Organic Chemistry
ARTICLE
Figure 6. Molecular modeling of β-D-glucose in the binding site of 7.
This structure is the global minimum from a Monte Carlo Molecular
Mechanics calculation in which random variations were made to the
carbohydrate’s position and conformation.12 The conformation of the
receptor was taken from the NOE study of the complex between 2 and
MeO-β-D-GlcNAc.8d Transparent surfaces highlight the two unambig-
uous CHꢀπ interactions within the complex.
linked by a chain of three water molecules running through the
binding site.12 Displacement of these solvent molecules could
be relatively difficult and might thus contribute little to (or even
inhibit) binding.17 In the case of 2 the biphenyl units are slightly
smaller than those of 6 and 7 and may be less hydrophobic.
Certainly it is expected that hydrophobicity should increase
between 2 and 6, as substituting H with F is known to have
this effect.18
Figure 5. Calculated surface potentials for 33 (top), 34 (middle), and
35 (bottom). Computations were performed using Spartan, Hartreeꢀ
Fock method, 6-31G* basis set. The colors represent the energy required
to bring a point positive charge to the surface of each molecule and are
plotted on the same scale: red e ꢀ110 kJ molꢀ1 (i.e., negatively charged
surface), blue g 68 kJ molꢀ1 (i.e., positively charged surface).
’ CONCLUSIONS
Contact between aromatic rings and saccharide CH groups
clearly plays an important role in carbohydrate recognition. With
this research we provide further evidence that this CHꢀπ effect
can be moderated by changing the electron density of the π-
system. However, the work also highlights the fact that the
changes in binding energy tend to be small. Enhancing the π-
electron density can be a useful strategy in optimizing synthetic
lectins, but all consequences of a design change, especially effects
on binding site hydration, must be taken into account.
Table 1. Binding Constants of Monosaccharides 28ꢀ32 to
Receptors 2a and 6ꢀ8, As Measured by 1H NMR Titration in
D2O
Ka/Mꢀ1
Carbohydrate
D-Glucose 28
ꢀH 2a
ꢀF 6
ꢀOHb 7 ꢀOꢀ c 8
’ EXPERIMENTAL SECTION
9
20
9
43
8
17
4
General Methods. Proton and carbon NMR spectra were recorded
at 400 MHz or at 500 MHz. Chemical shifts are reported in ppm
downfield from tetramethylsilane, for proton and carbon. Solvents for
synthesis were dried by passing through a modified Grubbs system.19
Routine monitoring of reactions was performed using precoated silica
gel TLC plates. Spots were visualized by UV light, ethanolic solution of
phosphomolybdic acid, potassium permanganate, or ninhydrin. Flash
column chromatography20 was performed using silica gel (particle size
35ꢀ70 μm) as the absorbent.
D-Galactose 29
2
v.s.d
D-Mannose 30
2
3
5
Methyl β-D-glucoside 31
Methyl R-D-glucoside 32
a Data from ref 8a. b pH = 7.5. pH = 12. d Very small. Minor signal
27
66
16
155
26
90
15
7
c
movements, almost linear with substrate concentration.
Two such interactions would therefore result in a difference of
3 kJ molꢀ1 in the gas phase, equivalent to an ∼3-fold ratio of
binding constants, quite similar to that observed.
On the other hand both 2 and 8 are weaker than expected on
the basis of CHꢀπ effects, and this may relate to the ease of
hydration of their cavities. In the case of 8, the basic phenoxide
centers should act as strong H-bond acceptors and could
nucleate networks of H-bonded water molecules in the vicinity
of the cavity. Indeed, modeling shows that (H2O)2 bridges can
form across either end of the cavity and that these can be
5-Bromo-1,3-bis-(bromomethyl)-2-fluorobenzene 10. 5-
Bromo-2-fluoro-m-xylene (5.0 g, 24.6 mmol) was dissolved in methyl
acetate (25 mL), and N-bromosuccinimide (9.2 g, 52 mmol) was added
under nitrogen. The reaction mixture was then refluxed for 3.5 h using a
100 W lamp, causing the mixture to change from a colorless suspension
to a deep red solution. The solvent was evaporated, and the residue
obtained was washed with boiling hexane. The solvent was evaporated,
and the product was recrystallized from hexane to yield 10 as white
needle-shaped crystals (5.43 g, 15.2 mmol, 62%). Rf = 0.27 (hexane); 1H
4
NMR (400 MHz, CDCl3) δ 4.45 (d, J(H,F) = 1.2 Hz, 4H), 7.50
6553
dx.doi.org/10.1021/jo200755z |J. Org. Chem. 2011, 76, 6548–6557

The Journal of Organic Chemistry
ARTICLE
(d, 4J(H,F) = 6.3 Hz, 2H); 13C NMR (100 MHz, CDCl3) δ 24.0
saturated aqueous NaHCO3 (50 mL), and brine (50 mL). The organic
phase was dried over Na2CO3 and evaporated. The white solid obtained
was then recrystallized from hexane/EtOAc (24:1, 50 mL) to yield 14 as
white needle-shaped crystals (876 mg, 2.02 mmol, 88%). Rf = 0.57 (3:2,
hexane/EtOAc); 1H NMR (400 MHz, CDCl3) δ 1.46 (s, 18H), 4.33 (d,
3J(H,H) = 5.8 Hz, 4H), 4.90 (bs, 2H), 7.37 (d, 4J(H,F) = 6.4 Hz, 2H);
13C NMR (100 MHz, CDCl3) δ 28.3, 38.2, 80.0, 116.7 (d, 4J(C,F) = 3.0
Hz), 128.2 (d, 2J(C,F) = 16.2 Hz), 131.2 (d, 3J(C,F) = 3.9 Hz), 155.7,
157.8 (d, 1J(C,F) = 245 Hz); 19F NMR (283 MHz, CDCl3) δ ꢀ127.20
(s); HRMS (ESI): m/z calculated for C18H26BrFN2O4Na+ [M + Na]+,
455.0952; found, 455.0949.
1,3-Bis-(tert-butoxycarbonylamino-methyl)-2-fluoro-5-
(40,40,50,50-tetramethyl-1,3,2-dioxaborolan-2-yl)-benzene 15.
Bis-carbamate 14 (734 mg, 1.69 mmol), bis-(pinacolato) diboron (450
mg, 1.78 mmol), and KOAc (580 mg, 5.93 mmol) were dried under
vacuum for 25 min before being dissolved in DMF (30 mL). Pd(dppf)Cl2
(41 mg, 0.05 mmol) was added under nitrogen. The reaction mixture was
heated at 60 °C overnight before the solvent was evaporated. The residue
obtained was dissolved in water (100 mL) and extracted with ethyl acetate
(3 ꢁ 70 mL), dried over sodium sulfate, filtered, and evaporated to yield a
black oil. The oil was purified by flash chromatography (eluent DCM to
DCM/EtOAc 17:3) producing the arylboronate 15 as a white solid (300
mg, 0.62 mmol, 37%). Rf = 0.59 (3:2, hexane/ethyl acetate); 1H NMR
(400 MHz, CDCl3) δ 1.33 (s, 12H), 1.46 (s, 18H), 4.37 (d, 3J(H,H) =
5.1 Hz, 4H), 4.85 (bs, 2H), 7.69 (d, 4J(H,F) = 7.3 Hz, 2H); 13CNMR (100
MHz, CDCl3) δ 24.8, 28.4, 38.7, 79.6, 84.0, 125.2 (d, 2J(C,F) = 14.6 Hz),
135.7, 155.7, 161.4 (d, 1J(C,F) = 253 Hz); 19F NMR (283 MHz, CDCl3)
δ ꢀ120.33 (s); HRMS (ESI): m/z calculated for C24H38BFN2O6Na+
[M + Na]+, 503.2699; found, 503.2702.
4
2
(d, 3J(C,F) = 4.6 Hz), 116.6 (d, J(C,F) = 3.8 Hz), 127.7 (d, J(C,F)
= 16.2 Hz), 134.2 (d, 3J(C,F) = 3.8 Hz), 157.5 (d, 1J(C,F) = 254 Hz); 19
F
NMR (283 MHz, CDCl3) δ ꢀ122.72 (t, 4J(H,F) = 6.4 Hz, 1F); HRMS
(EI): m/z calculated for C8H6Br3F+ [M]+, 357.8004; found, 357.8016.
1-Bromo-4-fluoro-3,5-bis-(phthalimido-methyl)benzene 11.
Dibromide 10 (1.67 g, 4.6 mmol) and potassium phthalimide (1.90 g,
10.7 mmol) were suspended in DMF (20 mL) and stirred for 3 h. H2O
(100 mL) was added, and the reaction mixture was then filtered. The
solid obtained was dissolved in chloroform (100 mL) and washed with
water (3 ꢁ 100 mL), dried over Na2CO3, and evaporated to yield the
1
bis-phthalimide 11 as a white solid (2.30 g, 4.67 mmol, 100%). H
NMR (400 MHz, CDCl3) δ 4.90 (s, 4H), 7.35 (d, 4J(H,F) = 6.3 Hz,
2H), 7.76 (m, 4H), 7.89 (m, 4H); 13C NMR (100 MHz, CDCl3) δ
34.8, 116.7 (d, 4J(C,F) = 3.8 Hz), 123.6, 125.6 (d, 2J(C,F) = 16.2 Hz),
131.9, 132.1 (d, 3J(C,F) = 3.9 Hz), 134.2, 157.4 (d, 1J(C,F) = 251 Hz),
167.6; 19F NMR (283 MHz, CDCl3) δ ꢀ123.30 (t, 4J(H,F) = 6.0 Hz);
HRMS (ESI): m/z calculated for C24H14BrFN2O4Na+ [M + Na]+,
515.0013; found, 515.0019.
1,3-Bis-(amino-methyl)-5-bromo-2-fluorobenzene 12. Bis-
phthalimide 11 (2.30 g, 4.67 mmol) was suspended in chloroform/
methanol (1:4, 50 mL) under nitrogen. Hydrazine (1.17 mL, 23.3 mmol,
5 equiv) was added, and the reaction mixture was heated at reflux
overnight. The reaction mixture was allowed to cool to room temperature,
and water (20 mL) was added. The organic solvents were evaporated, and
concentrated HCl (20 mL) was added. The aqueous solution was stirred
for 45 min forming a white precipitate which was removed by filtration.
The pH of the solution obtained was adjusted to pH 14 by the addition
of NaOH pellets and was extracted with DCM (3 ꢁ 50 mL). The organic
phases were combined, dried over Na2CO3, and evaporated to yield 12
as a white solid (0.85 g, 3.66 mmol, 78%). Rf = 0.69 (10% MeOH sat.
3,5-Bis-(benzyloxycarbonylamino-methyl)-30,50-bis-(tert-
butoxycarbonylamino-methyl)-4,40-difluorobiphenyl 16.
Bis-carbamate 13 (303 mg, 0.60 mmol) and arylboronate 15 (290 mg,
0.60 mmol) were dissolved in DMSO (20 mL) and degassed, after which
PdCl2(dppf) (15 mg, 0.02 mmol) and Na2CO3 aq. (2 M, 0.60 mL,
1.20 mmol) were added and the reaction mixture was heated at 60 °C
overnight. The solvent was then removed, and residue obtained was
dissolved in EtOAc (25 mL) and washed with H2O (2 ꢁ 25 mL) and
brine (25 mL), dried over sodium sulfate, filtered, and evaporated to
yield a black oil. The oil was purified by flash chromatography (eluent
DCM/EtOAc 9:1 to 7:3) producing 16 as a white solid (403 mg,
0.52 mmol, 86%). Rf = 0.31 (3:2, hexane/ethyl acetate); 1H NMR
(400 MHz, CDCl3) δ 1.45 (s, 18H), 4.35 (d, 3J(H,H) = 5.6 Hz,
1
NH3/DCM); H NMR (400 MHz, CDCl3) δ 3.88 (s, 4H), 7.38 (d,
4J(H,F) = 6.1 Hz, 2H); 13C NMR (100 MHz, CDCl3) δ 40.1 (d, 3J(C,F) =
4.8 Hz), 116.7 (d, 4J(C,F) = 3.8 Hz), 130.4 (d, 3J(C,F) = 4.6 Hz), 132.3 (d,
2J(C,F) = 16.9 Hz), 157.9 (d, 1J(C,F) = 244 Hz); 19F NMR (283 MHz,
CDCl3) δ ꢀ128.8 (t, 4J(H,F) = 6.1 Hz); HRMS (EI): m/z calculated for
C8H10BrFN2+ [M]+, 232.0011; found, 232.0002.
1,3-Bis-(benzyloxycarbonylamino-methyl)-5-bromo-2-
fluorobenzene 13. Diamine 12 (695 mg, 2.98 mmol) and DIPEA
(1.17 mL, 6.70 mmol) were dissolved in THF (50 mL) under nitrogen
and cooled to 0 °C, after which benzyl chloroformate (0.95 mL, 6.70 mmol)
was added. The reaction mixture was allowed to warm to room
temperature and stirred overnight before the solvent was evaporated,
and the oil obtained was dissolved in EtOAc (50 mL). This solution was
washed with saturated aqueous NH4Cl (50 mL), water (50 mL),
saturated aqueous NaHCO3 (50 mL), and brine (50 mL). The organic
phase was dried over Na2CO3 and evaporated. The white solid obtained
was then recrystallized from hexane/EtOAc (1:1, 30 mL) to yield 13 as
white needle-shaped crystals (0.77 g, 1.5 mmol, 52%). Rf = 0.43 (3:2,
3
4HNHBoc), 4.42 (d, J(H,H) = 5.9 Hz, 4HNHCbz), 5.12 (s, 4HPh),
5.34 (bs, 2H CH2NHBoc), 5.68 (bs, 2H CH2NHCbz), 7.15ꢀ7.38 (m,
14H); 13C NMR (100 MHz, CDCl3) δ 28.4, 38.6, 39.0, 67.0, 79.8, 125.9
(d, 2J(C,F) = 15.4 Hz), 126.3 (d, 2J(C,F) = 14.7 Hz), 127.3, 127.5, 128.1,
128.2, 128.4, 135.7, 136.0, 136.2, 155.8, 156.4, 158.6 (d, 1J(C,F) = 248.3 Hz);
19F NMR (283 MHz, CDCl3) δ 131.49 (m); HRMS (ESI): m/z calculated
for C42H48F2N4O8Na+ [M + Na]+, 797.3332; found, 797.3352.
1
3
3,5-Bis-(amino-methyl)-30,50-bis-(benzyloxycarbonylamino-
methyl) 4,40-difluorobiphenyl 17. Biaryl 16 (336 mg, 0.43 mmol)
was dissolved in dry DCM (40 mL) under nitrogen and cooled to 0 °C,
and TFA (12 mL) was added slowly. This solution was stirred for 45 min
after which the solvent was evaporated. The residue obtained was
dissolved in DCM (30 mL) and washed with aqueous sodium hydroxide
(20 mL, 2 M). The aqueous solution was extracted with DCM (3 ꢁ
30 mL), and the organic fractions were combined, dried over Na2SO4,
and reduced under reduced pressure producing the diamine 17 as a
white solid (248 mg, 0.43 mmol, 100%). 1H NMR (400 MHz, CDCl3) δ
1.52 (bs, 4H), 3.93 (s, 4HNH2), 4.46 (d, 3J(H,H) = 5.9 Hz, 4HNHCbz),
hexane/EtOAc); H NMR (400 MHz, CDCl3) δ 4.38 (d, J(H,H) =
6.1 Hz, 4HN), 5.13 (s, 4H), 5.22 (bs, 2H), 7.30ꢀ7.42 (m, 12H); 13C
NMR (100 MHz, CDCl3) δ 38.6, 67.1, 116.7, 127.8 (d, 2J(C,F) = 16.1
Hz), 128.1, 128.2, 128.5, 131.6 (d, 3J(C,F) = 3.8 Hz), 136.2, 156.3, 157.8
1
(d, J(C,F) = 246 Hz); 19F NMR (283 MHz, CDCl3) δ ꢀ126. 93 (t,
4J(H,F) = 6.0 Hz); HRMS (ESI): m/z calculated for C24H22BrFN2O4Na+
[M + Na]+, 523.0639; found, 523.0648.
1,3-Bis-(tert-butoxycarbonylamino-methyl)-5-bromo-2-
fluorobenzene 14. Diamine 12 (538 mg, 2.31 mmol) and DIPEA
(0.90 mL, 5.16 mmol) were dissolved in THF (50 mL) under nitrogen
and cooled to 0 °C, after which di-tert-butyl carbonate (1.13 g, 5.16 mmol)
was added. The reaction mixture was allowed to warm to room
temperature and stirred overnight before the solvent was evaporated,
and the oil obtained was dissolved in EtOAc (50 mL). This solution was
washed with saturated aqueous NH4Cl (50 mL), water (50 mL),
4
5.11 (s, 4H), 5.39 (bs, 2H), 7.20ꢀ7.37 (m, 12H), 7.44 (d, J(H,F) =
6.6 Hz); 13C NMR (100 MHz, CDCl3) δ 39.2, 40.6, 66.9, 126.0 (d, 2J(C,
F) = 15.3 Hz), 130.6 (d, 2J(C,F) = 16.1 Hz), 126.5 (d, 3J(C,F) = 4.6 Hz),
127.9 (d, 2J(C,F) = 2.3 Hz), 128.1, 128.1, 128.5, 135.9 (d, 4J(C,F) =3.1 Hz),
6554
dx.doi.org/10.1021/jo200755z |J. Org. Chem. 2011, 76, 6548–6557

The Journal of Organic Chemistry
ARTICLE
4
1
136.3, 136.7 (d, J(C,F) = 3.1 Hz), 156.3, 158.7 (d, J(C,F) = 245.2
Tetrafluoro Receptor 6. Tricycle 20 (18 mg, 5.52 μmol) was
dissolved in DCM (10 mL) under a nitrogen environment and cooled in
ice water. TFA (2 mL) was added dropwise, and the reaction mixture
was stirred for 5 h at which point the solvent was removed in vacuo. The
white solid obtained was dissolved in methanol/water (6:4, 10 mL), and
the pH of the solution was adjusted to pH 7 with aqueous sodium
hydroxide (0.5 M) which was monitored with a pH meter. The solvent
was removed to yield the salt 6 as a white solid (15 mg, 5.27 μmol, 95%).
1H NMR (500 MHz, D2O) δ = 2.45 (t, 3J(H,H) = 6.4 Hz, 24H), 3.75 (t,
3J(H,H) = 6.4 Hz, 24H), 3.85 (s, 24H), 4.34 (d, 2J(H,H) = 15.2 Hz, 8H),
4.85 (d, 2J(H,H) = 15.2 Hz, 8H), 7.77 (d, 4J(H,F) = 6.1 Hz, 8H), 7.84 (s,
Hz), 158.8 (d, J(C,F) = 246.7 Hz); 19F NMR (283 MHz, CDCl3) δ
1
ꢀ127.40 (s, 1F), ꢀ128.94 (s, 1F); HRMS (ESI): m/z calculated for
C32H33F2N4O4+ [M + H]+, 575.2464; found, 575.2484.
Tetra(benzyloxycarbonyl) Protected Macrocycle 19. Di-
amine 17 (246 mg, 0.43 mmol) and DIPEA (450 μL, 2.58 mmol, 6 equiv)
were dissolved in dry THF (500 mL) under nitrogen, and a solution of
bis-pentafluorophenyl ester 187b (441 mg, 0.43 mmol) in dry THF
(10 mL) was added dropwise over a 50 h period. The reaction mixture
was stirred for a further 24 h before the solvent was evaporated. The oil
obtained was dissolved in DCM (30 mL) and washed with saturated
aqueous NH4Cl (25 mL), H2O (25 mL), saturated aqueous NaHCO3
(25 mL), and brine (25 mL) before being dried over Na2SO4 and
evaporated. The residue was purified by flash chromatography (eluent
DCM to DCM/EtOAc 1:1) resulting in a white solid which was further
purified by preparative high performance liquid chromatography
(Hichrom Kromasil column, 150 mm ꢁ 21.2 mm, 5 μm, eluent:
methanol/water 80:20 to 95:5 in 14 min, then to 100:0 after 50 min,
2+
4H), 8.22 (s, 8H); HRMS (ESI): m/z calcd for C120H132F4N12O48Na2
[M + 2Na]2+, 1315.3989; found, 1315.4028.
3,5-Bis-[(tert-butyloxycarbonyl)aminomethyl]-30,50-bis-
(bromomethyl)-4,40-bis-(benzyloxy)biphenyl 22. Diol 21 (400
mg, 0.49 mmol) and PPh3 (388 mg, 1.48 mmol, 3 equiv) were dissolved
in dry THF (40 mL) under nitrogen and cooled to 0 °C, after which CBr4
(491 mL, 1.48 mmol, 3 equiv) was added. The reaction was warmed to
room temperature after 10 min, and the reaction mixture was then left to stir
overnight, after which the solvent was evaporated. The residue obtained was
purified by flash chromatography (eluent DCM to DCM/ethyl acetate 9:1)
producing 22 as a white solid (219 mg, 0.27 mmol, 79%). Rf = 0.68 (3:2,
hexane/ethyl acetate); 1H NMR (400 MHz, CDCl3) δ 1.48 (s, 18H), 4.42
(d, 3J(H,H) = 5.6 Hz, 4H), 4.59 (s, 4H), 4.91 (bs, 2H), 4.92 (s, 2H), 5.25 (s,
2H), 7.38ꢀ7.49 (m, 8H), 7.45 (s, 2H), 7.57ꢀ7.60 (m, 2H), 7.58 (s, 2H);
13C NMR (100 MHz, CDCl3) δ 27.7, 28.4, 39.9, 76.3, 76.3, 79.7, 127.1,
128.1, 128.3, 128.5, 128.6, 128.7, 128.8, 130.7, 132.6, 133.0, 136.0, 136.4,
136.6, 137.5, 154.5, 154.5, 155.9; HRMS (ESI): m/z calculated for
C40H46Br2N2O6Na+ [M + Na]+, 831.1615; found, 831.1599.
flow rate 15 mL minꢀ1). The macrocycle 19 (89 mg, 0.04 mmol, 17%)
3
was isolated as a white solid. Retention time = 19 min; 1H NMR
(400 MHz, 92:8 CDCl3/CD3OD) δ 1.29 (s, 54H), 2.37 (t, 3J(H,H) =
6.2 Hz, 12H), 3.59 (t, 3J(H,H) = 6.2 Hz, 12H), 3.73 (s, 12H), 4.33 (d,
3J(H,H) = 5.4 Hz, 8H), 4.54 (s, 3J(H,H) = 4.7 Hz, 8H), 4.98 (s, 8H),
6.16 (bs, 4H), 6.84 (bs, 2H), 7.10ꢀ7.25 (m, 20H, Cbz-ArCH), 7.33 (d,
4J(H,F) = 6.3 Hz, 4H), 7.45 (d, 4J(H,F) = 5.9 Hz, 4H), 8.10 (s, 2H), 8.24
(bs, 4H), 8.26 (s, 4H); 13C NMR (100 MHz, 92:8 CDCl3/CD3OD) δ
27.7, 36.1, 38.2, 38.7, 60.4, 66.6, 68.7 (C(CH2O)3), 80.6, 125.4 (d, 2J(C,
F) = 12.6 Hz), 125.9 (d, 2J(C,F) = 12.7 Hz), 127.5, 127.8, 127.9, 128.3,
129.0, 129.6, 134.5, 135.7, 136.1, 136.2, 156.6, 159.2 (d, 1J(C,F) = 249.1
Hz), 166.3, 166.6, 171.2; 19F NMR (283 MHz, 92:8 CDCl3/CD3OD) δ
ꢀ126.48 (s, 1F), ꢀ126.86 (s, 1F); HRMS (ESI): m/z calculated for
C132H158F4N10O32Na+ [M + Na]+, 2494.0872; found, 2494.0907.
tert-Butyl Protected Tetrafluoro Receptor 20. Pd on C (100 mg)
was activated by heating at 200 °C under vacuum for 4 h after which a
solution of macrocycle 19 (86 mg, 0.04 mmol) in THF and MeOH
saturated with NH3 (1:1, 100 mL) was added. The flask was evacuated
and filled with hydrogen (1 atm), and the reaction mixture was stirred
for 1.5 h. The reaction mixture was then filtered over Celite and washed
with ethyl acetate before the filtrate was evaporated to yield the
tetraamine as a white solid (64 mg) which was then dissolved in dry
THF (150 mL) containing DIPEA (23 μL, 0.13 mmol). A solution of
bis-pentafluorophenyl ester 187b (48 mg, 0.05 mmol) in dry THF
(10 mL) was added dropwise over 50 h to the tetraamine and DIPEA
solution under a nitrogen environment at room temperature. After the
addition the reaction mixture was allowed to stir for a further 24 h
before the solvent was evaporated. The oil obtained was dissolved in
DCM (30 mL) and washed with saturated aqueous NH4Cl (25 mL),
H2O (25 mL), saturated aqueous NaHCO3 (25 mL), and brine
(25 mL) before being dried over Na2SO4 and evaporated. This was
purified by preparative high performance liquid chromatography
(Hichrom Kromasil column, 150 mm ꢁ 21.2 mm, 5 μm, eluent:
methanol/water 80:20 to 95:5 in 14 min, then to 100:0 after 50 min,
3,5-Bis-[(tert-Butyloxycarbonyl)aminomethyl]-30,50-bis-
(azidomethyl)-4,40-bis-(benzyloxy)biphenyl 23. Dibromide 22
(219 mg, 0.26 mmol) was dissolved in DMF (15 mL). NaN3 (85 mg,
1.30 mmol, 5 equiv) was added. The reaction mixture was heated at
60 °C under nitrogen for 30 h before the solvent was removed by
evaporation. The solid obtained was dissolved in ethyl acetate (30 mL)
and washed with water (3 ꢁ 30 mL) and brine (30 mL). The organic
phase was dried over Na2SO4 and evaporated under reduced pressure to
yield 23 as a white solid (199 mg, 0.26 mmol, 100%). Rf = 0.36
1
(3:2, hexane/ethyl acetate); H NMR (400 MHz, CDCl3) δ 1.46 (s,
18H), 4.43 (m, 8H), 4.88 (bs, 2H), 4.92, 4.99, 7.36ꢀ7.49 (m, 12H), 7.53
(s, 2H); 13C NMR (100 MHz, CDCl3) δ 28.4, 39.9, 49.8, 76.3, 77.2,
79.6, 127.2, 128.0, 128.3, 128.5, 128.6, 128.7, 129.1, 130.1, 132.9, 136.3,
136.4, 136.5, 137.2, 154.5, 154.8, 155.9; HRMS (ESI): m/z calculated
for C40H46N8O6Na+ [M + Na]+, 757.3433; found, 757.3430.
3,5-Bis-(aminomethyl)-30,50-bis-(azidomethyl)-4,40-bis-
(benzyloxy)biphenyl 24. Diazide 23 (199 mg, 0.26 mmol) was
dissolved in dry DCM (10 mL) under nitrogen and cooled to 0 °C. TFA
(2 mL) was added slowly. This solution was stirred for 2 h after which
the solvent was evaporated, and the residue obtained was dissolved in
ethyl acetate (30 mL) and washed with a sodium hydroxide solution
(20 mL, 2 M). The aqueous solution was extracted with DCM (3 ꢁ
30 mL), and the organic fractions were combined, dried over Na2SO4,
and evaporated under reduced pressure. The oil obtained was purified by
flash chromatography (eluent DCM to DCM/NH3-saturated methanol,
49:1) producing 24 as a white solid (139 mg, 0.26 mmol, 100%). Rf =
0.81 (9:1, DCM/NH3-saturated methanol); 1H NMR (400 MHz, CDCl3)
δ 1.54 (bs, 4H), 3.99 (s, 4H2), 4.46 (s, 4H), 4.97, 5.01, 7.36ꢀ7.51 (m,
10H), 7.50 (s, 2H), 7.59 (s, 2H); 13C NMR (100 MHz, CDCl3) δ 42.0,
49.8, 76.3, 77.4, 126.4, 127.9, 128.0, 128.4, 128.6, 128.7, 128.8, 129.2,
130.0, 136.3, 136.4, 137.0, 137.3, 137.7, 154.6, 154.7; HRMS (ESI): m/z
calculated for C30H31N8O2+ [M + H]+, 535.2564; found, 535.2584.
Tetra(azido) Tetrabenzyloxy Macrocycle 25. Diamine 24
(139 mg, 0.26 mmol) and DIPEA (283 μL, 1.63 mmol, 6 equiv)
were dissolved in dry THF (300 mL) under nitrogen. A solution of
flow rate 15 mL minꢀ1). The pure tricycle 20 (18 mg, 5.5 μmol, 25%)
3
was isolated as a white solid. Retention time = 21 min; 1H NMR (400
MHz, CDCl3) δ = 1.40 (s, 108H), 2.46 (t, 3J(H,H) = 6.2 Hz, 24H), 3.69
(t, 3J(H,H) = 6.2 Hz, 24H), 3.83 (s, 24H), 4.46 (dd, 2J(H,H) = 13.8 Hz,
3J(H,H) = 4.8 Hz, 8H), 4.74 (dd, 2J(H,H) = 13.8 Hz, 3J(H,H) = 4.8 Hz,
8H), 6.75 (s, 4H), 7.22 (bs, 8H), 7.78 (d, 4J(H,F) = 6.3 Hz, 8H), 8.06
(s, 4H), 8.30 (s, 8H); 13C NMR (75 MHz, CDCl3) δ 28.0, 36.4, 39.5,
60.4, 67.1, 69.0, 80.6, 125.5 (d, 2J(C,F) = 15.5 Hz), 128.7, 128.9, 129.0
(d, 3J(C,F) = 3.1 Hz), 133.5 (d, 4J(C,F) = 2.5 Hz), 135.0, 136.5, 160.0 (d,
1J(C,F) = 251.1 Hz), 166.3, 166.5, 171.2; 19F NMR (283 MHz, CDCl3)
2+
δ ꢀ124.77 (s); HRMS (ESI): m/z calcd for C168H228F4N12O48Na2
[M + 2Na]+, 1651.7745; found, 1651.7794.
6555
dx.doi.org/10.1021/jo200755z |J. Org. Chem. 2011, 76, 6548–6557

The Journal of Organic Chemistry
ARTICLE
bis-pentafluorophenyl ester 187b (279 mg, 0.27 mmol) in dry THF
(50 mL) was added dropwise over a 30 h period. The reaction mixture
was stirred for a further 10 h before the solvent was evaporated under
reduced pressure. The oil obtained was purified to by flash chromatog-
raphy (eluent DCM to ethyl acetate) resulting in a white solid which was
further purified by preparative high performance liquid chromatography
(Hichrom Kromasil column, 150 mm ꢁ 21.2 mm, 5 μm, eluent:
methanol/water 80:20 to 95:5 in 14 min, then to 100:0 in a further
153.9, 166.2, 168.3, 171.2; MS (ESI): m/z calcd for C168H232N12O52-
Na22+ [M + 2Na]2+, 1647.7832; found, 1647.7872.
Tetrahydroxy Receptor 7. Macrotricycle 27 (8 mg, 2.5 μmol)
was dissolved in DCM (5 mL) under nitrogen and cooled to 0 °C in an
ice bath. TFA (1 mL) was added dropwise, and the reaction mixture was
stirred for 5 h at which point the solvent was evaporated. The white solid
obtained was dissolved in methanol/water (6:4, 10 mL). The pH of the
solution was adjusted to pH 7 by the addition of a sodium hydroxide
solution (0.5 M); this was monitored using a pH meter. The solvent
was removed by lyophilization to yield the salt 7 as a white powder
(7 mg, 2.4 μmol, 96%). 1H NMR (500 MHz, D2O, pH = 7.5) δ 2.46
(t, 3J(H,H) = 6.3 Hz, 24H), 3.75 (t, 3J(H,H) = 6.3 Hz, 24H), 3.85 (s,
24H), 4.36 (d, 2J(H,H) = 13.8 Hz, 8H), 7.71 (s, 8H), 7.82 (s, 4H), 8.17 (s,
8H); 1H NMR (500 MHz, D2O, pH = 7.5) δ 2.46 (t, 3J(H,H) = 6.8 Hz,
24H), 3.76 (t, 3J(H,H) = 6.8 Hz, 24H), 3.87 (s, 24H), 4.19 (d, 2J(H,H) =
13.6Hz, 8H), 7.42(s,8H), 8.07(s,4H), 8.28(s,8H);MS(ESI):m/zcalcd
for C120H136N12O52Na22+ [M + 2Na]2+, 1311.4075; found, 1311.4101.
36 min, flow rate 15 mL minꢀ1). The macrocycle 25 was isolated as a
3
1
white solid (90 mg, 0.04 mmol, 29%). Retention time = 27 min; H
NMR (400 MHz, 92:8 CDCl3/CD3OD) δ 1.39 (s, 54H), 2.48 (t, 3J(H,
H) = 6.3 Hz, 12H), 3.70 (t, 3J(H,H) = 6.3 Hz, 12H), 3.84 (s, 12H), 4.38
(s, 8H), 4.62 (s, 8H), 4.94 (s, 4H), 4.99 (s, 4H), 7.28ꢀ7.46 (m, 24H),
7.50 (s, 4H), 8.12 (t, 4J(H,H) = 1.7 Hz, 2H), 8.40 (d, 4J(H,H) = 1.7 Hz,
4H); 13C NMR (100 MHz, 92:8 CDCl3/CD3OD) δ 27.8, 36.1, 39.1,
49.5, 60.3, 66.9, 68.8, 77.1, 77.2, 80.6, 127.5, 128.3, 128.9, 129.2, 130.0,
132.2, 134.4, 127.8, 128.1, 128.5, 128.6, 136.1, 136.2, 136.2, 136.5, 136.6,
154.8.2, 154.9, 166.0, 166.6, 171.1; HRMS (ESI): m/z calculated for
C
128H154N18O28Na22+ [M + 2Na]2+, 1218.5482; found, 1218.5453.
’ ASSOCIATED CONTENT
tert-Butyl Protected Tetrabenzyloxy Macrotricycle 26.
S
PPh3 (491 mg, 1.87 mmol) was added to a solution of tetraazide 25
(224 mg, 0.09 mmol) dissolved in THF, and the reaction mixture was
heated at 60 °C under nitrogen overnight. H2O (2 mL) was added. The
reaction mixture was heated at 60 °C for a further 4 h, after which the
solvent was evaporated to give a crude tetraamine which was used
directly after flash chromatography (eluent DCM to 19:1 DCM/
methanol saturated with NH3) in the subsequent cyclization process.
A solution of bis-pentafluorophenyl ester 187b (208 mg, 0.20 mmol) in
dry THF (50 mL) was added dropwise over 30 h to a solution of the
foregoing tetraamine (210 mg, 0.09 mmol) and DIPEA (94 μL,
0.54 mmol) in dry THF (250 mL) under nitrogen at room temperature.
After the addition the reaction mixture was allowed to stir for a further
24 h, after which the solvent was evaporated. The oil obtained was dissolved
in DCM (30 mL) and washed with saturated aqueous NH4Cl (25 mL)
and brine (25 mL) before being dried over Na2SO4 and evaporated. The
residue was purified by flash chromatography (eluent DCM to ethyl
acetate) resulting in a white solid which was subjected to further
purification by preparative high performance liquid chromatography
(Hichrom Kromasil column, 150 mm ꢁ 21.2 mm, 5 μm, eluent: methanol/
water 80:20 to 95:5 in 14 min, then to 100:0 in a further 36 min, flow rate
Supporting Information. NMR spectra for characteriza-
b
tion, details of binding experiments including spectra and
analysis plots, UV spectra of 7 at different pH values, details of
molecular modeling. This material is available free of charge via
the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: Anthony.Davis@bristol.ac.uk
’ ACKNOWLEDGMENT
This work was supported by the EPSRC (EP/D060192/1,
and DTA award to N.P.B.).
’ REFERENCES
(1) (a) Davis, A. P.; Wareham, R. S. Angew. Chem., Int. Ed. 1999,
38, 2978. (b) Walker, D. B.; Joshi, G.; Davis, A. P. Cell. Mol. Life Sci.
2009, 66, 3177. (c) Mazik, M. Chem. Soc. Rev. 2009, 38, 935. (d) Gabius,
H. J. Naturwissenschaften 2000, 87, 108. (e) Lemieux, R. U. Acc. Chem.
Res. 1996, 29, 373.
(2) Meyer, E. A.; Castellano, R. K.; Diederich, F. Angew. Chem., Int.
Ed. 2003, 42, 1210. Takahashi, O.; Kohno, Y.; Nishio, M. Chem. Rev.
2010, 110, 6049.
(3) Leading references: (a) Ramirez-Gualito, K.; Alonso-Rios, R.;
Quiroz-Garcia, B.; Rojas-Aguilar, A.; Diaz, D.; Jimꢀenez-Barbero, J.;
Cuevas, G. J. Am. Chem. Soc. 2009, 131, 18129. (b) Bautista-Ibanez,
L.; Ramirez-Gualito, K.; Quiroz-Garcia, B.; Rojas-Aguilar, A.; Cuevas, G.
J. Org. Chem. 2008, 73, 849. (c) Kiehna, S. E.; Laughrey, Z. R.; Waters,
M. L. Chem. Commun. 2007, 4026. (d) Spiwok, V.; Lipovova, P.; Skalova, T.;
Buchtelova, E.; Hasek, J.; Kralova, B. Carbohydr. Res. 2004, 339, 2275.
(4) Vyas, N. K.; Vyas, M. N.; Quiocho, F. A. Science 1988, 242, 1290.
(5) Theoretical studies: (a) Amicangelo, J. C.; Gung, B. W.; Irwin,
D. G.; Romano, N. C. Phys. Chem. Chem. Phys. 2008, 10, 2695. (b)
Ringer, A. L.; Figgs, M. S.; Sinnokrot, M. O.; Sherrill, C. D. J. Phys. Chem.
A 2006, 110, 10822. Experimental confirmation:(c) Carver, F. J.;
Hunter, C. A.; Livingstone, D. J.; McCabe, J. F.; Seward, E. M. Chem.—
Eur. J. 2002, 8, 2848.
(6) For example, see: (a) Chavez, M. I.; Andreu, C.; Vidal, P.;
Aboitiz, N.; Freire, F.; Groves, P.; Asensio, J. L.; Asensio, G.; Muraki, M.;
Canada, F. J.; Jimꢀenez-Barbero, J. Chem.—Eur. J. 2005, 11, 7060. (b)
Laughrey, Z. R.; Kiehna, S. E.; Rierrien, A. J.; Waters, M. L. J. Am. Chem.
Soc. 2008, 130, 14625. (c) Yanagihara, R.; Aoyama, Y. Tetrahedron Lett.
1994, 35, 9725.
15 mL minꢀ1). Macrotricycle 26 (86 mg, 0.02 mmol, 26%) was isolated as
3
a white solid. Retention time = 27 min; 1H NMR (400 MHz, CDCl3) δ =
1.39 (s, 108H), 2.45 (t, 3J(H,H) = 6.3 Hz, 24H), 3.70 (t, 3J(H,H) = 6.3 Hz,
24H), 4.41 (dd, 2J(H,H) = 13.8 Hz, 3J(H,H) = 5.3 Hz, 8H), 4.69 (dd, 2J(H,
H) = 13.8 Hz, 3J(H,H) = 4.9 Hz, 8H), 5.00 (s, 8H), 6.63 (t, 3J(H,H) = 5.2
Hz, 8H), 6.68 (bs, 4H), 7.34ꢀ7.42 (m, 20H, 7.74 (s, 8H), 7.90 (s, 4H), 8.21
(s, 8H); 13C NMR (100 MHz, CDCl3) δ 28.0, 36.2, 41.1, 60.4, 67.1, 69.0,
76.5, 80.5, 128.1, 128.7, 128.9, 132.1, 134.3, 135.1, 136.4, 136.7, 156.3,
2+
166.1, 166.3, 170.9; HRMS (ESI): m/z calcd for C196H256N12O52Na2
[M + 2Na]2+, 1827.8771; found, 1827.8829.
tert-Butyl Protected Tetrahydroxy Receptor 27. Pd on C
(40 mg) was activated by heating at 140 °C under vacuum overnight,
after which a solution of tricycle 26 (45 mg, 12.5 μmol) in THF and
MeOH saturated with NH3 (1:1, 10 mL) was added. The flask was
evacuated and filled with hydrogen (1 atm), and the reaction mixture
was stirred for 3 h. The reaction mixture was then filtered over Celite and
washed with ethyl acetate, after which the filtrate was evaporated to yield
27 as a white solid (37 mg, 11.3 μmol, 91%). Retention time = 26 min;
3
1H NMR (400 MHz, CDCl3) δ = 1.41 (s, 108H), 2.48 (t, J(H,H) =
6.3 Hz, 24H), 3.69 (t, 3J(H,H) = 6.3 Hz, 24H), 3.84 (s, 24H), 4.29 (dd,
2J(H,H) = 13.8 Hz, 3J(H,H) = 6.0 Hz, 8H), 4.71 (dd, 2J(H,H) = 13.8 Hz,
3J(H,H) = 5.3 Hz, 8H), 6.80 (s, 4H), 7.16 (bs, 8H), 7.52 (s, 8H), 7.88 (s,
4H), 8.23 (s, 8H), 9.26 (s, 4H); 13C NMR (100 MHz, CDCl3) δ 28.1,
36.4, 41.2, 60.5, 67.2, 69.1, 80.7, 125.6, 127.5, 128.4, 129.6, 135.6, 136.5,
6556
dx.doi.org/10.1021/jo200755z |J. Org. Chem. 2011, 76, 6548–6557

The Journal of Organic Chemistry
ARTICLE
(7) (a) Davis, A. P. Org. Biomol. Chem. 2009, 7, 3629. (b) Klein, E.;
Crump, M. P.; Davis, A. P. Angew. Chem., Int. Ed. 2005, 44, 298. (c)
Ferrand, Y.; Crump, M. P.; Davis, A. P. Science 2007, 318, 619. (d)
Ferrand, Y.; Klein, E.; Barwell, N. P.; Crump, M. P.; Jimꢀenez-Barbero, J.;
Vicent, C.; Boons, G. J.; Ingale, S.; Davis, A. P. Angew. Chem., Int. Ed.
2009, 48, 1775. (e) Barwell, N. P.; Crump, M. P.; Davis, A. P. Angew.
Chem., Int. Ed. 2009, 48, 7673.
(8) For background on synthetic carbohydrate receptors, see refs
1aꢀ1c. For recent examples, see: (a) Pal, A.; Berube, M.; Hall, D. G.
Angew. Chem., Int. Ed. 2010, 49, 1492. (b) Arda, A.; Venturi, C.; Nativi,
C.; Francesconi, O.; Gabrielli, G.; Canada, F. J.; Jimenez-Barbero, J.;
Roelens, S. Chem.—Eur. J. 2010, 16, 414. (c) Mazik, M.; Hartmann, A.;
Jones, P. G. Chem.—Eur. J. 2009, 15, 9147. (d) Palde, P. B.; Gareiss,
P. C.; Miller, B. L. J. Am. Chem. Soc. 2008, 130, 9566. (e) He, C.; Lin,
Z. H.; He, Z.; Duan, C. Y.; Xu, C. H.; Wang, Z. M.; Yan, C. H. Angew.
Chem., Int. Ed. 2008, 47, 877. (f) Goto, H.; Furusho, Y.; Yashima, E.
J. Am. Chem. Soc. 2007, 129, 9168. (g) Waki, M.; Abe, H.; Inouye, M.
Angew. Chem., Int. Ed. 2007, 46, 3059.
(9) Klein, E.; Ferrand, Y.; Barwell, N. R.; Davis, A. P. Angew. Chem.,
Int. Ed. 2008, 47, 2693.
(10) Substituent electronic effects could also change the H-bonding
properties of the spacer amides. However, the differences are likely to be
small given the distance involved (five bonds) and the likelihood of
canceling effects (increased binding to substrate will usually be accom-
panied by increased binding to competing water).
(11) Conformational searches were performed using Macromodel
9.1, accessed via Maestro 7.5, employing the MMFs force field with
water GB/SA solvation.
(12) For details, see Supporting Information.
(13) Jonsson, M.;Lind, J.;Merenyi, G. J. Phys. Chem. A 2002, 106, 4758.
(14) For NOE evidence of glucose enclosure by 2 and 5 (R = Pr), see
refs 7b and 7e respectively.
(15) Cockroft, S. L.; Hunter, C. A. Chem. Soc. Rev. 2007, 36, 172.
(16) Note that the binding constants reported to reducing sugars are
averaged values, with contributions from both R and β anomers. For
further discussion of this issue see ref 7e.
(17) Affinities to 8 might also be lowered by ionization of the
carbohydrate substrates, giving alkoxide species which cannot bind to
the tetra-anionic cavity. The pKa of glucose in water has been measured
as 12.45 (see El Rassi, Z.; Nashabeh, W. In Carbohydrate analysis - high
performance liquid chromatography and capillary electrophoresis; El Rassi,
Z., Ed.; Elsevier: Amsterdam, 1995; Vol. 58, p 267), so at pH = 12 (as
used to study 8) a substantial minority of glucose molecules should
be ionized. By reducing the concentration of neutral glucose mol-
ecules, this effect could lower the apparent binding constant by a
factor approaching 2. We are grateful to a referee for raising this
possibility.
(18) Dalvi, V. H.; Rossky, P. J. Proc. Natl. Acad. Sci. U.S.A. 2010,
107, 13603.
(19) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518.
(20) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923.
6557
dx.doi.org/10.1021/jo200755z |J. Org. Chem. 2011, 76, 6548–6557