Guangyun Yu1, Anna Chiara Vicini1, Roland J Pieters1. 1. Department of Chemical Biology & Drug Discovery , Utrecht Institute for Pharmaceutical Sciences, Utrecht University , P.O. Box 80082, 3508 TB Utrecht , The Netherlands.
Abstract
Divalent ligands were prepared as inhibitors for the adhesion protein of the problematic Pseudomonas aeruginosa pathogen. Bridging two binding sites enables simultaneous binding of two galactose moieties, which strongly enhances binding. An alternating motif of glucose and triazole and aryl groups was shown to have the right mix of rigidity, solubility, and ease of synthesis. Spacers were varied with respect to the core unit as well as the aglycon portions in an attempt to optimize dynamics and enhance interactions with the protein. Affinities of the divalent ligands were measured by ITC, and Kd's as low as 12 nM were determined, notably for a compounds with either a rigid (phenyl) or flexible (butyl) unit at the core. Introducing a phenyl aglycon moiety next to the galactoside ligands on both termini did indeed lead to a higher enthalpy of binding, which was more than compensated by entropic costs. The results are discussed in terms of thermodynamics and theoretical calculations of the expected and observed multivalency effects.
Divalent ligands were prepared as inhibitors for the adhesion protein of the problematic Pseudomonas aeruginosa pathogen. Bridging two binding sites enables simultaneous binding of two galactose moieties, which strongly enhances binding. An alternating motif of glucose and triazole and aryl groups was shown to have the right mix of rigidity, solubility, and ease of synthesis. Spacers were varied with respect to the core unit as well as the aglycon portions in an attempt to optimize dynamics and enhance interactions with the protein. Affinities of the divalent ligands were measured by ITC, and Kd's as low as 12 nM were determined, notably for a compounds with either a rigid (phenyl) or flexible (butyl) unit at the core. Introducing a phenyl aglycon moiety next to the galactoside ligands on both termini did indeed lead to a higher enthalpy of binding, which was more than compensated by entropic costs. The results are discussed in terms of thermodynamics and theoretical calculations of the expected and observed multivalency effects.
Protein–carbohydrate
interactions are involved in many biological
processes and diseases.[1−6] In this context, it is an important goal to find new specific molecular
ligands for carbohydrate-binding or carbohydrate-processing proteins,[7−9] to be used as chemical probes,[10] or leads
for therapeutic application.[11,12] One specific aspect
of protein–carbohydrate interactions is the widespread prevalence
of multivalency.[13−18] Numerous carbohydrate binding proteins of biological or medicinal
interest contain more than one binding site, either identical or not.
Bridging such binding sites by divalent or higher valency ligands
may lead to greatly enhanced binding or inhibitory potencies.[19,20] Increasingly higher potency enhancements are being reported, and
larger distances are also being covered by spacers.[21] In this context, it is likely that rigidified spacers are
beneficial with potential for high potency and specificity. More flexible,
often PEG-based spacers exhibit more shallow affinity optima.[22] Recent calculations involving effective molarity
calculations and experiments revealed that PEG-based spacers will
have no enhancing effect for bridging distant weak sites of millimolar
binding affinities.[21] Nucleic acid based
spacers seem to be preferred for the bridging of long distances,[19,23−27] while for shorter distances various structural types have been reported
including, e.g., polyproline[28] and phenylene-ethynylene.[29] Notably, such spacers should have a persistent
linear overall shape, but they should also allow for overall aqueous
solubility of the multivalent construct. Previously, we described
a modular spacer based on directly equatorially 1,4-linked glucose
and 1,4-linked triazole moieties.[30,31] This system
was used for optimization and yielded divalent ligand 1 (Figure ) of the Pseudomonas aeruginosa adhesion lectin LecA with two galactose
specific binding sites separated by ca. 26 Å when measured between
the anomeric oxygens of the bound galactosides in pdb entry 1OKO.[32,33] The spacer length of 1 was optimized on the basis of
inhibition and binding data (Kd = 28 nM),
and it is the most potent reported divalent ligand for LecA. Its divalent
binding mode was confirmed by X-ray crystallography, in which the
whole spacer was visible, a feat not seen before for a synthetic spacer
of this length (Figure ).[34] The structure was also largely predicted
by molecular modeling.[31] Interestingly,
besides the interactions between the terminal galactoside ligands
and the protein, additional interactions were observed between the
protein and the spacer, possibly adding to the binding affinity. While 1 was clearly a potent LecA inhibitor, it was not clear which
of its structural features were contributing significantly to its
potency. We here describe the synthesis and detailed thermodynamic
evaluation of a series of variants.
Figure 1
X-ray structure of 1 bound
to LecA (pdb 4YWA(34)).
X-ray structure of 1 bound
to LecA (pdb 4YWA(34)).The aim was to first explore the synthetic possibilities
of rigid
spacers composed of glucose, triazole, and phenyl units and second
to study the structure activity relationships between the spacer moieties
and the binding to LecA. The syntheses were modular in all cases,
but different strategies were explored involving building blocks based
on azido-glucose derivatives and 1,4-diethynyl benzene linked together
by CuAAC. ITC was used to shed more light on the effect of various
components of the spacers, producing thermodynamic parameters. Multivalent
LecA inhibitors have been reported in the literature, e.g., based
on fullerenes,[35] β-peptoids,[36] peptide dendrimers,[37−40] calixarenes,[41,42] cyclic carbohydrates,[43] perylene,[44] tetraphenylethylene,[45] a carbohydrate core,[46−48] and gold nanoparticles.[49] In addition, among other potent divalent ligands,[50,43] a potent divalent ligand with a Kd of
82 nM was found by library screening, and although less potent than 1, it achieved its potency while being considerably more flexible.[51]The lectin LecA is of medicinal interest
as a virulence factor
for P. aeruginosa, involved in the adhesion of the
pathogen, biofim formation, and causing lung injury.[33,52,53]P. aeruginosa is a Gram-negative pathogen involved in diseases such as dermatitis,
pancreatitis, urinary tract infections, keratitis, and respiratory
tract infections.[54] It is regarded as a
primary cause of death in immuno-compromised patients, notably those
with cystic fibrosis.[55] Treating P. aeruginosa infection is becoming more difficult because
of the increasing spread of drug-resistant strains,[56,57] which made it one of the highest priorities targets for intervention.[58] Another reason for its difficult eradication
is its tendency to form biofilms.[59] In
these biofilms, the bacteria are protected from the host defense system
and the action of antibiotics. It was estimated that within a biofilm,
bacteria are upward of 1000 times more resistant to conventional antibiotic
treatment.[60−63] These issues combined make the search for P. aeruginosa therapeutics an urgent one. Bacterial adhesion is often a prelude
to infection.[64,65] For P. aeruginosa, lectin LecA has been identified to play an important role in the
internalization of the pathogen by binding to glycosylated targets
displayed on the cell surface.[66] Therefore,
inhibition of LecA is aimed at affecting adhesion of the bacteria
at an early stage of the infection process and may provide an alternative
to conventional antibiotics.[67] This concept[68,65] was supported by the therapeutic effect of a galactose solution
against P. aeruginosa pneumonia in mouse models and
cystic fibrosispatients through inhibiting the binding of LecA to
its glycosylated targets.[53,69]
Results and Discussion
From previous research, we knew that the length of the divalent
ligand is a very important factor for the binding affinity.[31] For this reason, ligand 2 and 3 were designed with the same numbers of atoms in the spacer
as the previously optimized 1 (Figure ). For ligand 2, a phenyl group
replaces the central glucose moiety of 1 and maintains
the number of atoms in the spacer (in terms of distance between the
two galactosides). Furthermore, the two remaining glucose units in
the spacer of 2 are linked in the opposite direction;
i.e., the C(4) is linked to the core instead of C(1). The molecule
is now also symmetrical just like its target protein. The consequences
of the modification are that this synthesis does not require the use
of a glucose building block with a C(1) alkyne, which is a more difficult
to prepare building block. The strategy for the synthesis of 2 relied on the construction of the diazido-functionalized
spacer 13 (Scheme ). To this end, the two hemiacetals in 12 were
converted to two β-azides using 2-azido-1,3-dimethylimidazolinium
hexafluorophosphate (ADMP).[70] CuAAC conjugation
of 13 and 14, followed by Zemplén
deprotection, yielded 2. Next, a totally unconstrained
central unit was introduced in the design of 3 in order
to evaluate the importance of the constraint in 1 and 2. For ligand 3, octa-1,7-diyne was used to introduce
the central unit. For the synthesis, a different strategy was used
than for 2. Here, the galactoside ligand was first coupled
to the spacer unit, and the resulting compound was linked to the core
structure at the end. The partially benzoylated building block 16 was “clicked” with 14 to yield 17. After activation as a triflate, the axial hydroxyl at
C(4) was displaced by sodium azide leading to equatorial azide 18. CuAAC conjugation to the central dialkyne, followed by
the Zemplén deprotection afforded ligand 3. Overall,
the advantage of this strategy was to avoid the relatively low yielding
ADMP step. The synthesis is now highly efficient with only nine steps
from commercial peracetylated sugars and an overall yield of 13%.
Figure 2
Structures
of mono- and divalent LecA inhibitors used in this study.
Scheme 1
(a) CuSO4·5H2O, Na-ascorbate, DMF/H2O 9:1, microwave, 80 °C,
40 min, 65–85%; (b) D2O/CH3CN 4/1, Et3N, 0 °C, 3 days, 50%; (c) MeONa, MeOH, 40–50%
after prep HPLC; (d) (i) Tf2O, pyridine, CH2Cl2, 0 °C, 1 h; (ii) NaN3, DMF, 14 h,
80% over two steps.
Structures
of mono- and divalent LecA inhibitors used in this study.(a) CuSO4·5H2O, Na-ascorbate, DMF/H2O 9:1, microwave, 80 °C,
40 min, 65–85%; (b) D2O/CH3CN 4/1, Et3N, 0 °C, 3 days, 50%; (c) MeONa, MeOH, 40–50%
after prep HPLC; (d) (i) Tf2O, pyridine, CH2Cl2, 0 °C, 1 h; (ii) NaN3, DMF, 14 h,
80% over two steps.The next aim was to introduce
a phenyl group as the aglycon part
of the terminal galactoside ligands, as this moiety is known to enhance
the LecA binding by a factor of ca. 5–10 fold,[71,43,41,72,47,73] benefiting
from CH−π interactions.[74] In
the first approach, 13 was linked to 20a (Scheme S1) by CuAAC to give 21 and after acetyl removal 4a was obtained (Scheme ).
Scheme 2
(a)
CuSO4·5H2O, Na-ascorbate, DMF/H2O 9:1, microwave, 80 °C,
40 min, 65–85%; (b) MeONa, MeOH, 20–40% after prep HPLC;
(c) (i) Tf2O, pyridine, CH2Cl2, 0
°C, 1 h; (ii) NaN3, DMF, 14 h, 73% over two steps
(a)
CuSO4·5H2O, Na-ascorbate, DMF/H2O 9:1, microwave, 80 °C,
40 min, 65–85%; (b) MeONa, MeOH, 20–40% after prep HPLC;
(c) (i) Tf2O, pyridine, CH2Cl2, 0
°C, 1 h; (ii) NaN3, DMF, 14 h, 73% over two stepsUnfortunately, 4a proved to be insoluble
in water.
For this reason, the central benzene ring was outfitted with two short
PEG units in the synthesis of 4b. As before, a C4-bridged
version of the molecule (5) was also prepared using the
same synthetic strategy. The sequence started with a “click”
coupling between 16 and 20a, resulting in 22a. Its axial C(4) hydroxyl was converted to an equatorial
azide to give 23a. CuAAC coupling of 23a to bis-alkyne 24(75) yielded 25 and, after deprotection 4b, which exhibited
sufficient aqueous solubility. Similarly, 23a was coupled
to octa-1,7-diyne, yielding 26, and after acetyl removal 5 was obtained.While the structure of compound 4b was very close
to that of compound 2, the modification also introduces
an extra six atoms, and the spacer is therefore longer. In order to
keep a similar length to 2 while introducing the phenyl
units adjacent to the galactose moiety, compound 6 was
designed. In this compound, the shortest path between the two anomeric
oxygens of the galactose ligands involves 25 atoms (Table ), while this number is 26 for
compounds 1–3. To evaluate the sensitivity
of the binding to both length and flexibility, compounds 7 and 8 were added that contain additional flexible units
in the spacer.
Table 1
LecA Binding by ITCa
compd
atoms in spacerf
Kd
n
ΔH
–TΔS
ΔG
rel pot (per sugar)
1
d-dalactose[71]
87500
1.1
–7.9
2.3
–5.5
2
Gal-β-OMe[43]
70000
0.8
–9.3
3.6
–5.7
3
Gal-β-OPh[71]
8800
0.9
–11.2
4.8
–6.9
4
9a(31)
22000
0.92
–8.3
2.0
–6.3
0.25 (0.25)
5
9b
6200 ± 400
1.00 ± 0.01
–11.6 ± 0.9
4.5 ± 0.9
–7.1 ± 0.1
1 (1)
6
10
8400 ± 1,700
0.94 ± 0.06
–8.7 ± 1.0
–1.8 ± 0.8
–6.9 ± 0.1
1 (1)
7
10
7300 ± 800
0.95 ± 0.01
–9.5 ± 0.6
2.5 ± 0.6
–7.0 ± 0.1
1 (1)d
Methylene Seriesb
7
1
26
28[31]
0.55
–11.6
1.3
–10.3
221 (111)
8
2
26
12 ± 8
0.41 ± 0.09
–22.9 ± 1.9
12 ± 1.8
–10.9 ± 0.4
517 (258)
9
3
26
13 ± 3
0.58 ± 0.06
–14.3 ± 0.2
3.6 ± 0.3
–10.8 ± 0.14
477 (238)
Phenylene Seriesc
10
4b
32
61 ± 10
0.43 ± 0.03
–10.2 ± 0.5
0.4 ± 0.1
–9.9 ± 0.2
120 (60)d,e
11
5
32
92 ± 15
0.43 ± 0.03
–16.8 ± 0.5
7.2 ± 0.5
–9.6 ± 0.2
91 (46)
12
6
25
87 ± 17
0.46 ± 0.01
–16.8 ± 0.5
7.1 ± 0.6
–9.6 ± 0.1
97 (48)
13
7
29
94 ± 13
0.43 ± 0.01
–15.2 ± 0.2
5.6 ± 0.1
–9.6 ± 0.1
89 (45)
14
8
31
35 ± 15
0.50 ± 0.01
–16.1 ± 0.9
5.9 ± 1.0
–10.2 ± 0.3
240 (120)
Kd in
nM, ΔΗ, −ΤΔS, and ΔΗ in
kcal/mol, Standard deviations are given over two or more experiments.
Relative potency determined
vs 9b.
Relative
potency determined vs 10 in buffer.
Determined in buffer with 5% DMSO.
Relative potency determined vs 10 in buffer with 5% DMSO.
Number of atoms between the two
anomeric oxygens of the galactosides of divalent ligands using the
shortest path.
Kd in
nM, ΔΗ, −ΤΔS, and ΔΗ in
kcal/mol, Standard deviations are given over two or more experiments.Relative potency determined
vs 9b.Relative
potency determined vs 10 in buffer.Determined in buffer with 5% DMSO.Relative potency determined vs 10 in buffer with 5% DMSO.Number of atoms between the two
anomeric oxygens of the galactosides of divalent ligands using the
shortest path.The syntheses
of 6–8 started with
alkynes 20 (Schemes S1 and S3) that were linked to azide 27 giving 28 (Scheme ). For 27, a new route was developed (Schemes S2). Previously, we used a galactose moiety where the C(4)
OH was inverted to the glucosideazide.[31] As the alkyne introduction chemistry typically works better on glucoside
derivatives, this was used here, and a double-inversion strategy led
to 27. Removal of the TIPS groups from 28 by TBAF yielded the free alkynes 29. The other half
of the target compounds was prepared similarly, starting again with
alkynes 20, now coupled to azidosugar 16, affording 22. Installing the required equatorial azido
groups in 23, set the stage for the coupling with 29. After the CuAAC coupling and removal of the acetyl and
benzoyl groups, divalent ligands 6–8 were obtained.
Scheme 3
(a) CuSO4·5H2O, Na-ascorbate, DMF/H2O 9:1, microwave, 80 °C,
40 min, 65–85%; (b) TBAF, Et3N, THF, 14 h, 80–90%;
(c) (i) Tf2O, pyridine, CH2Cl2, 0
°C, 1 h; (ii) NaN3, DMF, 14 h, 80% over two steps;
(d) MeONa, MeOH, 30–40% after prep HPLC.
(a) CuSO4·5H2O, Na-ascorbate, DMF/H2O 9:1, microwave, 80 °C,
40 min, 65–85%; (b) TBAF, Et3N, THF, 14 h, 80–90%;
(c) (i) Tf2O, pyridine, CH2Cl2, 0
°C, 1 h; (ii) NaN3, DMF, 14 h, 80% over two steps;
(d) MeONa, MeOH, 30–40% after prep HPLC.The compounds were evaluated for their binding abilities of the
LecA lectin using isothermal titration calorimetry (ITC) as previously
reported by us and others.[31,41,76] The bivalent compounds were compared to a number of monovalent ligands
reported in the literature and a few relevant reference compounds,
and the numbers are shown in Table . Entries 1 and 2 show similar Kd’s for free galactose versus its β-OMe derivative
at around 70–90 μM. Entry 3 shows the effect of an aromatic
aglycon moeity, which enhances the binding to a Kd of ca. 9 μM. The divalent inhibitors (1–8) were divided into two groups. In the first,
the methylene series (1–3) refers
to the methylene group between the galactose ligand and the spacer
(triazole). In the phenylene series (4b–8) this is a phenyl group. Both series showed an n value of around 0.5, consistent with bivalent binding. In the methylene
series, both compounds 2 and 3 that contain
a phenyl and an n-butyl unit in the center, respectively,
were more potent than the glucose-bridged 1. The Kd’s of the divalent 2 and 3 were essentially identical and reached unprecedented levels
of 12–13 nM. Interestingly, the thermodynamic parameters in
the methylene series varied widely. In the phenylene series, affinities
varied but were lower with Kd’s
between 35 and 94 nM. With the exception of PEG-containing 4b which was measured in the presence of 5% DMSO, thermodynamic parameters
of the phenylene group members 5–8 were relatively close.One of the notable thermodynamic features
of 1 was
the low entropic loss associated with its binding event. The TΔS component was smaller than that
of monovalent ligands. While it is tempting to attribute this to the
rigid spacer, naturally other factors such as solvation also play
important roles in the entropic component. A surprising notion was
the observation that the flexible 3 was such a good ligand
surpassing 1 in terms of Kd. Furthermore, this compound’s entropic loss was relatively
low and similar to that of the monovalent ligands. Besides this, it
also has a larger favorable enthalpy than 1, but the
differences are small. Another notable aspect is the fact that turning
the two glucosides around in 3 versus 1 and
removing the central glucose did not have a deleterious effect. In
the X-ray structure of 1 bound to the protein LecA,[31] three water-bridged hydrogen bonds and a direct
hydrogen bond involving the C(3) and C(4) OH’s of the central
and adjacent spacer glucoside were observed.While the flipped
glucose moieties in 3 could similarly
make these hydrogen bonds too, the central sugar is obviously missing
in 3. These protein–spacer interactions may have
contributed some to the binding energy, but it seems their contributions
were minor as the removal of the central glucose had no negative effect
on binding.Compound 2 bound with essentially the
same affinity
as compound 3. Interestingly, the binding energy of compound 2 has a much higher enthalpic contribution (ca. 2-fold) and
a higher entropic loss (ca. 9-fold) than 1. The origin
of this effect is not obvious but may involve the differential solvation
between bound and free state. Even so, the large and opposite enthalpic
and entropic components of the binding of 2 are not extreme
and fall between the following two reported divalent ligands: (1)
a flexible, peptide-based divalent ligand GalAG1[38] showed a ΔH of −29 kcal/mol
with a Kd 83 nM, and (2) and the mentioned
flexible structure discovered by library screening[51] showed a ΔH of −18 kcal/mol
for a Kd 82 nM.In the aryl-linked
galactoside series (4–8), Kd differences were relatively
moderate, and none of the compounds were as potent as the starting
point 1. Compound 4b showed a remarkably
low entropic loss (−TΔS = 0.4 kcal/mol), even lower than that of 1, and in
stark contrast with 2, but they may not be directly compared
because it was measured in the presence of 5% DMSO. In the phenylene
series, the enthalpic contribution is typically larger than that of
the methylene-linked series (1–3),
but the entropic loss is also larger leading to a weaker overall binding.In terms of the benefit of the divalent presentation and the usefulness
of building a spacer between the two ligands, a reference monovalent
ligand needs to be chosen. This is a delicate issue, as no reference
molecule is perfect in this regard. After previously using 9a, we here use 9b for the methylene series as the extended
monovalent ligand. It is clear that it benefits somewhat from interactions
to the protein with a lower Kd down from
22 μM for 9a to 6.2 μM for 9b. The affinity is now comparable with that of the phenyl aglycon
containing 10 and related compounds. Possibly, the observed
interactions between the spacer glucose moiety is the cause of this
enhancement,[34] while the phenyl-linked
compounds were reported to benefit from the interactions of the phenyl
with the nearby CH group of a histidine.From our results, clearly
the methylene series was more potent
than the phenylene series. Large relative potencies (and relative
potencies per sugar) were obtained for 2 with a 542-fold
(271-fold per sugar) multivalency enhancement. Note that this number
is 1822 when compared to 9a. The 271-fold Kd-based number is among the highest enhancements in the
literature for LecA to the best of our knowledge. For Reymond’s
tetravalent peptidic system, the Kd was
2.5 nM, which calculates to a ca. 300-fold enhancement per sugar when
compared to the monovalent ligand (Kd ca.
3 μM). Similarly, for a calixarene-based tetramer (Kd 90 nM) a ca. 300-fold enhancement per sugar was calculated,
when compared to the reference arm.[41]Based on recent mathematical models we calculated whether the observed
enhancements were in line with expectations and possibly whether they
were close to optimum or not. The recently reported models for rigid
spacers indicate that rigidity is key for multivalency effects, especially
when the binding sites have a weak monovalent affinity.[21] In that case, a flexible PEG spacer will not
induce a measurable multivalency effect, while a rigid spacer of appropriate
length does. The effective molarity is key in these discussions, which
is the concentration of the second ligand around a second binding
site after the first one is bound and should be higher than the monovalent Kd to have a gain in potency. For a PEG-based
system, this concentration was calculated to be in the micromolar
range in contrast to the rigid spacer where it was millimolar and,
thus, much more likely to bind with enhanced affinity. In the present
system, the binding sites have a relatively high affinity with monovalent
ligands binding below 100 μM. Even a PEG-based divalent system
in our hands bound with a 30-fold enhancement.[30] We calculated the effective concentration according to
the paper using a 30 Å distance between the center of the two
nearby binding sites in LecA and, subsequently, the predicted Kd’s as a function of spacer length for
rodlike ligands, with a variability of its length of 4 Å, since
no spacer is perfectly rigid. The Kd’s
were calculated according to eq .[21]This equation contains the mentioned effective
concentration ceff. The monovalent Kd is 70 μM, i.e., Kmono, based on Table , entry 2, the affinity
of the ligand without any parts of the spacer (Gal-β-OMe). The
experimentally determined interaction with the spacer (ΔGspacer) is based on entry 5, where the ligand 9b contains a sizable part of the spacer which does contribute
to the affinity (Kd = 6.2 μM).The graph (Figure ) shows an optimum Kd for the ideal 30
Å spacer, the spacing between the binding site centers. The Kd was predicted to be ca. 1 nM. Considering
that the best compound in this study has a Kd of 12 nM, our compounds are ca. 1 order of magnitude below
their theoretical optimum. It should be pointed out that the experimental
approach has an impact on multivalency effects, e.g., due to protein
concentration. This was previously noted in the lower inhibition concentrations
for 1 (IC50 = 2.7 nM) by ELISA,[31] and similar effects were seen for the multivalent
inhibition of cholera toxin.[77]
Figure 3
Calculated
dissociation constants of a divalent ligands of various
lengths (rete is end-to-end distance)
for LecA, according to eq (see the SI for details).
Calculated
dissociation constants of a divalent ligands of various
lengths (rete is end-to-end distance)
for LecA, according to eq (see the SI for details).The type of modeling described above makes it clear
that major
advances due to the chelation type of bivalent binding are possible
and favorable for rigid spacers of the right length but would unlikely
be able to distinguish between the subtle structural variations in
the present series, which nevertheless show significant binding differences.Can we look to the thermodynamic parameters for answers? Overall,
the thermodynamics indicate an enthalpically driven binding that can
be associated with an induced fit model.[78] If we take the profile of 2 as exceptional, the rest
behaves in the following way. The addition of the phenyl aglycon does
indeed help the binding enthalpy as it does for the monovalent Gal-β-OPhe
(entry 3); however, unlike the case of the monovalent Gal-β-OPhe,
the gains are more than balanced by increased entropic losses, resulting
in overall weaker binding. Surface burial could be the source of favorable
entropy, often associated with hydrophobic surfaces like the phenyl
group. In the present compounds this factor does not dominate. More
likely is the option that conformational rearrangements were needed
in the ligand, and possibly the protein, to accommodate the galactoside
ligand with its properly oriented phenyl aglycon to take advantage
of the additional CH−π interactions.[71] The large difference between 2 and 3 with respect to thermodynamic parameters, while exhibiting essentially
the same Kd, is intriguing. Building a
CPK model and handling both compounds reveals the large difference
in rigidity. Compound 2 is quite rigid, while 3 has a central hinge region that allows a range of conformations.
The large entropic costs for the binding of 2 suggest
major reorganization of the protein to enable all intermolecular contacts.
Ironically, binding the flexible 3 costs far less entropy,
as the single degree of freedom caused by the hinge is not overall
very costly when compared to rearranging the protein for accomodating 2.
Conclusions
Rigid, well-defined spacers were synthesized
that were based on
equatorially 1,4-linked glucose moieties or 1,4-linked phenyl rings
alternated with 1,4 triazole moieties. Variations were made in the
central unit and in the part linked to the galactose ligand with additional
flexibility-enhancing units. Synthetic strategies varied accordingly,
with the most successful synthesis of one the best ligands 3 being only nine steps from commercial peracetylated sugars and an
overall yield of 13%. This synthesis coupled an azido-galactose moiety
to the terminal propargyl galactoside ligand by CuAAC. After introducing
an equatorial azido group at the galactoseC(4)–OH, two of
these units were linked to the central bis-alkyne.Overall,
a major affinity improvement was obtained by linking either
octa-1,7-diyne or 1,4-diethynylbenzene to 19a, which
yielded divalent ligands 2 and 3 with ca.
500-fold binding enhancements.Thermodynamic parameters were
evaluated in detail, and surprisingly
large differences were observed, while the differences in Kd’s were relatively minor. The compounds
in the phenylene series did generally show more favorable enthalpy
as expected for additional CH−π interactions, but this
advantage was more than erased by additional entropic costs possibly
caused by protein rearrangement. This phenomenon made the methylene
series the more effective ligands. Within this series, the large differences
in thermodynamic parameters between 2 and 3 were intriguing and tentatively attributed to differences on rigidity
and solvation as caused by replacing a phenyl with a butyl group.A recent spacer modeling approach was applied and led to the conclusion
that more improvements should theoretically be possible, but also
that the method was too coarse grain to predict the subtle effects
that were seen here. In that sense, possibly a full modeling approach
may eventually become successful. Bridging ligands is a common theme
in the carbohydrate recognition realm but also interfaces in general[18] or in noncarbohydrate ligands that were linked
together to achieve improved properties.[79]
Experimental Section
General Methods
Unless stated otherwise, chemicals
were obtain from commercial sources and were used without further
purification. Compounds 11,[80]14,[2]16,[3] and 24(4) were synthesized following literature procedure. Solvents were purchased
from Biosolve (Valkenswaard, The Netherlands). All moisture-sensitive
reactions were performed under nitrogen atmosphere. Anhydrous THF
was dried over Na/benzophenone and freshly distilled prior to use.
All of the other solvents were dried over molecular sieves 4 or 3
Å. TLC was performed on Merck precoated silica 60 plates. Spots
were visualized by UV light, 10% H2SO4 in MeOH,
and triphenylphosphine in THF followed by ninhydrin (for azides).
Microwave reactions were carried out in a Biotage microwave Initiator
(Uppsala, Sweden). The microwave power was limited by temperature
control once the desired temperature was reached. Sealed vessels of
2–5 and 10–20 mL were used. Analytical HPLC runs were
performed on a Shimadzu automated HPLC system with a reversed-phase
column (Phenomenex, C4, 250 × 4.60 mm 5 μm 140087-2
for ligand 5, C18, 250 × 2.00 mm 5 μm
132174-4 for ligands 2, 3, 4a,b, 6–8, and 9b) that was equipped with an evaporative light scattering
detector (PLELS 1000, Polymer Laboratories, Amherst, MA) and a UV/vis
detector operating at 220 and 254 nm. Preparative HPLC runs were performed
on an Applied Biosystems workstation. Elution was effected by using
a linear gradient of 5% MeCN/0.1% TFA in H2O to 5% H2O/0.1% TFA in MeCN. 1H NMR spectra were recorded
at 400, 500, and 600 MHz and 13C at 101, 126, and 151 MHz.
Electrospray mass experiments were performed in a Shimadzu LCMS QP-8000.
High-resolution mass spectrometry (HRMS) analysis was performed using
an ESI-QTOF II spectrometer (Bruker, Billerica, MA).
Isothermal
Titration Microcalorimetry (ITC)
The lectin
LecA was obtained from Sigma-Aldrich and was dissolved in buffer (0.1 M
Tris–HCl, 6 mM CaCl2, pH 7.5) and degassed. Protein
concentration (between 10 and 40 μM depending on the ligand
affinity) was checked by measurement of optical density by using a
theoretical molar extinction coefficient of 28000 units. Carbohydrate
ligands were dissolved directly into the same buffer, degassed, and
placed in the injection syringe. ITC was performed using a MicroCal
Auto ITC200 (Malvern, Worcestershire, UK). LecA (0.01–0.04
mM) was placed into the 200 μL sample cell at 25 °C.
Titration was performed with injections of carbohydrate ligands (5–20
times of LecA, 2.5 μL) every 120 s. Data were fitted using the
“one-site model” using MicroCal Origin 7 software according
to standard procedures. Fitted data yielded the stoichiometry (n), the association constant (Ka), the enthalpy (ΔH) and the entropy of binding.
The Kd value was calculated as 1/Ka, and T = 298 K.
Diglucoside
(12)
Compound 11 (63 mg, 500 μmol)
and 1,4-diethynylbenzene (256 mg, 1.25 mmol)
were dissolved in 0.9 mL of DMF. Then the aqueous solution of CuSO4·5H2O (18 mg in 25 μL of water, 75 μmol)
and Na-ascorbate (30 mg in 25 μL of water, 150 μmol) was
added to the resulting mixture. Finally, TBTA (40 mg, 75 μmol)
was added, and the reaction system was heated by microwave irradiation
at 80 °C for 40 min. TLC indicated complete conversion of the
reaction. Then Cuprisorb was added, stirred for 30 min, and filtered.
The filtrate was dried under vacuum, and the residue was purified
by column chromatography (MeOH/DCM 1:1) to afford 12 as
a colorless syrup (174 mg, 348 μmol, 65%). 1H NMR
(400 MHz, D2O): δ 8.54 (s, 1H), 8.52 (s, 1H), 7.99–7.91
(m, 4H), 5.43 (d, J = 3.7 Hz, 1H), 4.92 (d, J = 8.0 Hz, 1H), 4.70 (td, J = 10.4, 3.9
Hz, 2H), 4.58 (ddd, J = 10.6, 4.2, 2.2 Hz, 1H), 4.46
(t, J = 9.9 Hz, 1H), 4.31–4.20 (m, 2H), 3.80
(dd, J = 9.6, 3.7 Hz, 1H), 3.59 (ddd, J = 12.7, 7.5, 2.2 Hz, 2H), 3.54–3.47 (m, 1H), 3.41–3.33
(m, 2H). 13C{1H} NMR (151 MHz, D2O): δ 146.8, 129.1, 125.8, 122.3, 122.1, 96.1, 92.3, 74.5,
74.1, 73.4, 71.8, 70.2, 69.7, 62.3, 62.2, 59.9, 59.8. HRMS (ESI, Q-TOF): m/z calcd for C22H 29N6O10 [M + H]+ 537.1945, found 537.1956.
Bis-azide (13)
To a D2O/CH3CN (4:1) solution of 12 (1.3 g, 2.4 mmol) was
added triethylamine (3.4 mL, 24 mmol) dropwise and the solution cooled
to 0 °C. Then 2-azido-1, 3-dimethylimidazolinium hexafluorophosphate
(ADMP 4 g, 14.2 mmol) was added, and the mixture was stirred at 0
°C until most of the starting material was converted to the azide
compound. Initially, two new spots formed on the TLC plate (developing
eluent n-BuOH/H2O/Acetic acid 6:3:1, 13 highest new spot, mono azide lower new spot). Within 3
days, conversion to 13 was complete. The solvent was
removed under vacuum, and the residue was purified by column chromatography
to give compound 13 as a white solid (703 mg, 1.2 mmol,
50%). 1H NMR (400 MHz, D2O): δ 8.44 (s,
2H, H-trizole), 7.92 (s, 4H, ArH), 4.82 (d, J = 8.7
Hz, 2H, H-1), 4.62 (t, J = 10.3 Hz, 2H, H-4), 4.26–4.17
(m, 4H, H-5, H-3), 3.62 (dd, J = 12.7 Hz, 1.8 Hz,
2H, H-6a), 3.37 (t, J = 8.8 Hz, 2H, H-2), 3.34–3.25
(m, 2H, H-6b). 13C{1H} NMR (101 MHz, CD3OD): δ 147.9, 131.5, 127.2, 123.7, 92.2, 78.0, 75.4,
75.4, 63.3, 61.4. HRMS (ESI, Q-TOF): m/z calcd for C22H27N12O8 [M + H]+ 587.2075, found 587.2073.
General Procedure
for the “Click Reaction”, Step
a. Preparation of Compounds 15, 17, and 19
The alkyne compound, CuSO4·5H2O (0.15 equiv), and sodium ascorbate (0.3 equiv) were added
to a solution of the azide compound in DMF containing 10% water. The
mixture was heated under microwave irradiation at 80 °C for 40
min. After evaporation of the solvent, the residue was dissolved in
CH2Cl2. The organic solution was washed three
times with water and brine and dried over sodium sulfate. The solvent
was removed, and the residue was purified by column chromatography.
General Procedure for Removal of Acetyl and Benzoyl Protecting
Groups, Preparation of Compounds 2, 3, 4a,b, and 5
The protected
substrate was suspended or dissolved in methanol. Sodium methoxide
was added to obtain a basic pH (pH ≈ 8). The reaction was stirred
at rt, and it was monitored by HPLC. After disappearance of the substrate,
the reaction was neutralized with DowexH+ resin. The mixture
was filtered, and the solvent was evaporated under vacuum, which was
subjected to purification by Preparative-HPLC.
Compound 17 (240
mg, 266 μmol) was first
dissolved in DCM/pyridine (10:1, 11 mL), and then triflic anhydride
(0.75 g, 2.66 mmol) was added dropwise at 0 °C and reacted for
1 h at this temperature. The reaction was quenched with 1 M KHSO4, and then DCM (20 mL) was added to extract the triflate intermediate.
The solution was washed with water and brine and dried by sodium sulfate.
The solvent was removed under vacuum to afford the crude triflate
intermediate, which was used directly for the next step. To the solution
of the intermediate in DMF (10 mL) was added sodium azide (87 mg,
1.33 mmol), and the mixture reacted at room temperature overnight.
After removal of the solvent, DCM was added to dilute the product.
The organic phase was washed with water and brine and dried with sodium
sulfate. The residue was purified by column chromatography (toluene/ethyl
acetate 3:1) to afford the compound as a white solid (217 mg, 234
μmol, 88%). 1H NMR (400 MHz, CDCl3): δ
8.11–8.01 (m, 2H, ArH), 7.98–7.90 (m, 2H, ArH), 7.88
(s, 1H, H-triazole), 7.80–7.70 (m, 2H, ArH), 7.62–7.23
(m, 9H, ArH), 6.13 (d, J = 9.2 Hz, 1H, H-1), 5.90
(t, J = 9.6 Hz, 1H, H-3), 5.80 (t, J = 9.4 Hz, 1H, H-2), 5.37 (dd, J = 3.5, 1.1 Hz,
1H, H-4′), 5.16 (dd, J = 10.4, 7.9 Hz, 1H,
H-2′), 4.98 (dd, J = 10.4, 3.4 Hz, 1H, H-3′),
4.88 (d, J = 12.9 Hz, 1H, −OCH2−), 4.82–4.72 (m, 2H, −OCH2–,
H-6a), 4.65 (dd, J = 12.5, 4.4 Hz, 1H, H-6b), 4.44
(d, J = 7.9 Hz, 1H, H-1′), 4.26–4.02
(m, 4H, H-6′a, H-4, H-6′b, H-5), 3.95 (ddd, J = 7.2, 6.0, 1.2 Hz, 1H, H-5′), 2.12 (s, 3H, CH3COO−), 2.04 (s, 3H, CH3COO−), 1.94
(s, 3H, CH3COO−), 1.76 (s, 3H, CH3COO−). 13C{1H} NMR (101 MHz, CDCl3): δ
170.7, 170.4, 170.1, 169.6, 166.0, 165.5, 165.0, 144.0, 134.0, 134.0,
133.7, 130.0, 129.9, 129.9, 129.4, 128.7, 128.6, 128.4, 127.8, 122.1,
98.9, 86.2, 76.1, 73.7, 71.2, 71.0, 70.6, 68.7, 67.2, 63.0, 61.4,
61.4, 60.6, 20.9, 20.8, 20.7, 20.6. HRMS (ESI, Q-TOF): m/z calcd for C44H45N6O17 [M + H]+ 929.2841, found 929.2859.
20a.1. To the solution of pentaacetyl β-d-galactopyranoside (200 mg, 513 μmol) in CH2Cl2 (3 mL) with 4 Å molecular sieves were added 4-iodophenol
(147 mg, 667 μmol) and BF3·Et2O (124
mg, 872 μmol) slowly at 0 °C. Then the mixture was allowed
to warm to room temperature and reacted for further 32 h. The reaction
was quenched with water (0.5 mL) and diluted with ethyl acetate. The
organic layer was washed with 1 M HCl, saturated NaHCO3, water, and brine and dried with sodium sulfate. The solvent was
removed, and the residue was purified by column chromatography (PE/EtOAc
4:1) to obtain the pure β isomer as a white solid (169 mg, 308
μmol, 60%). 1H NMR (400 MHz, CDCl3): δ
7.62–7.53 (m, 2H, ArH), 6.81–6.72 (m, 2H, ArH), 5.50–5.41
(m, 2H, H-2, H-4), 5.09 (dd, J = 10.4, 3.4 Hz, 1H,
H-3), 4.99 (d, J = 7.9 Hz, 1H, H-1), 4.25–4.07
(m, 2H, H-6a, H-6b), 4.04 (m, 1H, H-5), 2.17 (s, 3H, CH3COO−), 2.05 (s, 6H, 2 × CH3COO−), 1.98
(s, 3H, CH3COO−). 13C{1H}
NMR (101 MHz, CDCl3): δ 170.4, 170.3, 170.2, 169.4,
156.8, 138.6, 119.3, 99.6, 86.2, 71.2, 70.8, 68.6, 66.9, 61.5, 20.8,
20.8, 20.7. The spectral data are in accordance with literature data.[5]
Galactoside (20a.2)
Compound 20a.1 (150 mg, 273 μmol), Pd(PPh3)2Cl2 (5.74 mg, 8.2 μmol), and
CuI (1.56 mg, 8.2 μmol) were added to a round-bottomed flask
and degassed for 30 min. Then the previously degassed triethylamine
(3 mL) was added to the flask, and finally ethynyltrimethylsilane
(40.3 mg, 410 μmol) was added via syringe. The resulting system
reacted at rt overnight. Triethylamine was removed under vacuum, and
CH2Cl2 was added to extract the product. The
organic layer was washed with water and brine and dried with sodium
sulfate. After removal of the solvent, the residue was purified by
column chromatography (PE/EtOAc 4:1) to obtain 20a.2 as
a white solid (114 mg, 218 μmol, 80%). 1H NMR (400
MHz, CDCl3): δ 7.41–7.33 (m, 2H, ArH), 6.93–6.85
(m, 2H, ArH), 5.50–5.39 (m, 2H, H-2, H-4), 5.09 (dd, J = 10.4 Hz, 3.4 Hz, 1H, H-3), 5.02 (d, J = 8.0 Hz, 1H, H-1), 4.21–4.11 (m, 2H, H-6a, H-6b), 4.06–4.03
(m, 1H, H-5), 2.15 (s, 3H, CH3COO−), 2.03 (s, 6H,
2 × CH3COO−), 1.98 (s, 3H, CH3COO−),
0.21 (s, 9H, Si(CH3)3). 13C{1H} NMR (101 MHz, CDCl3): δ 170.4, 170.3,
170.2, 169.4, 156.9, 133.6, 118.2, 116.7, 104.5, 99.3, 93.7, 71.3,
70.9, 68.7, 67.0, 61.5, 20.8, 20.8, 20.7. HRMS (ESI, Q-TOF): m/z calcd for C25H 32O10SiNa [M + Na]+ 543.1663, found 543.1659.
Galactoside (20a)
Compound 20a.2 (108 mg, 219 μmol) was
dissolved in THF (10 mL). TBAF·3H2O (83 mg, 263 μmol)
was added, and the mixture was stirred at room temperature for 1 h.
The solvent was removed under vacuum, and the residue was purified
by column chromatography (PE/EtOAc 2:1) to afford 20a as a brown solid (74 mg, 164 μmol, 75%). 1H NMR
(400 MHz, CDCl3): δ 7.46–7.38 (m, 2H, ArH),
6.98–6.89 (m, 2H, ArH), 5.52–5.42 (m, 2H, H-2, H-4),
5.10 (dd, J = 10.4, 3.4 Hz, 1H, H-3), 5.05 (d, J = 8.0 Hz, 1H, H-1), 4.24–4.12 (m, 2H, H-6a, H-6b),
4.08–4.05 (m, 1H, H-5), 3.03 (s, 1H, CH ≡ C−),
2.17 (s, 3H, CH3COO−), 2.05 (s, 6H, 2 × CH3COO−), 2.00 (s, 3H, CH3COO−). 13C{1H} NMR (101 MHz, CDCl3): δ
170.4, 170.3, 170.2, 169.4, 157.1, 133.7, 117.1, 116.8, 99.3, 83.1,
77.4, 71.3, 70.9, 68.6, 66.9, 61.5, 20.8, 20.8, 20.7. HRMS (ESI, Q-TOF): m/z calcd for C22H 24O10Na [M + Na]+ 471.1267, found 471.1267.
General procedure for the “Click Reaction”, Preparation
of Compounds 21, 22a, 25, and 26
The compounds were prepared following the procedure
previously described for the synthesis of compound 12.
In a two-neck round-bottom flask under nitrogen atmosphere,
a solution of 2,3,4,6-tetra-O-benzyl-d-glucopyranose
(7.007 g, 12.96 mmol) in dry CH2Cl2 (65 mL)
was treated with Dess–Martin periodinane (DMP) (6.596 g, 15.55
mmol) at rt and the reaction mixture was stirred at the same temperature
until complete conversion of the starting material (monitored by TLC,
PE/EtOAc 2:1). After 1 h and 45 min, the reaction mixture was filtered
through a silica pad and rinsed with 1.4 L of CH2Cl2 to afford the lactone 27.1 as a pale yellow
oil (6.388 g, 11.86 mmol, 92%) which was used in the next step without
any further purification. 1H NMR (400 MHz, CDCl3): δ 7.41–7.15 (m, 20H, ArH), 4.99 (d, J = 11.4 Hz, 1H, PhCH), 4.76–4.43 (m, 8H, PhCH, H-5), 4.12
(d, J = 6.5 Hz, 1H, H-2), 3.88–3.98 (m, 2H,
H-3, H-4), 3.73 (dd, J = 11.0, 2.4 Hz, 1H, H-6a),
3.67 (dd, J = 11.0, 3.3 Hz, 1H, H-6b). 13C{1H} NMR (101 MHz, CDCl3): δ 169.5,
137.7, 137.7, 137.6, 137.1, 128.6, 128.6, 128.5, 128.3, 128.2, 128.1,
128.1, 128.0, 81.1, 78.3, 77.6, 76.2, 74.1, 73.9, 73.7, 68.4. MS (ESI): m/z calcd for C34H34O6Na [M + Na]+ 561.23, found 561.55, m/z calcd for C34H38NO6 [M+NH4]+ 556.27 found 556.55.
Spectroscopic data were in accordance with literature data.[6]
In a round-bottom flask under
argon atmosphere, a solution
of triisopropylsilyl acetylene (6.5 mL, 29.0 mmol) in dry THF (35
mL) was treated dropwise at −78 °C with a 2.5 M solution
of n-BuLi in hexane (7.0 mL, 17.4 mmol) and stirred for 15 min. Subsequently,
a solution of compound 27.1 (6.256 g, 11.61 mmol) in
dry THF (12 mL) was added dropwise within 2 min at −78 °C
and stirred for an hour at −78 °C (monitored by TLC, PE/EtOAc
2:1). When the conversion of 27.1 was complete, the reaction
was neutralized by the addition of Amberlite IR120 H+ form
resin (checked using pH-paper), allowing at the same time the temperature
to increase slowly from −78 °C to rt. The resin was filtered,
washed with CH2Cl2, and the solvent was removed in vacuo. The residue, a yellow oil, was dissolved in dry
CH3CN/CH2Cl2 1:1 (130 mL in total)
and transferred to a three-necks round-bottom flask. The solution
was cooled to −15 °C by means of an ice and salt bath,
and Et3SiH (11.0 mL, 68.9 mmol) was added at once, followed
by dropwise addition of BF3·OEt2 (8.5 mL,
68.9 mmol), while the temperature inside the flask was monitored to
avoid it exceeding −10 °C. The reaction mixture was stirred
for 1h at −15 °C, and then, as the reaction was not complete,
overnight at −20 °C. The reaction was quenched by pouring
the mixture into Et2O/NaHCO3 (satd) 1:1 (200
mL in total). Et2O (150 mL) and H2O (100 mL)
were added and the layers were separated. The aqueous phase was extracted
with Et2O (3 × 150 mL) and the combined organic layers
were dried over Na2SO4 and concentrated in vacuo. Purification by column chromatography (PE 95%,
EtOAc 3%, CHCl3 2%) gave 6.45 g (9.15 mmol, 79% in 2 steps)
of 27.2 as white needles. 1H NMR (400 MHz,
CDCl3): δ 7.39–7.16 (m, 20H, ArH), 5.11 (d, J = 10.7 Hz, 1H, PhCH), 4.90 (d, J = 11.1
Hz, 1H, PhCH), 4.85–4.81 (m, 3H, PhCH), 4.66–4.52 (m,
3H, PhCH), 4.07–4.01 (m, 1H, H-2), 3.77 (dd, J = 11.2 Hz, 2.1 Hz, 1H, H-6a), 3.70 (dd, J = 11.2,
4.5 Hz, 1H, H-6b), 3.68–3.59 (m, 3H, H-1, H-3, H-4), 3.47–3.41
(m, 1H, H-5), 1.10 (s, 21H, SiCH(CH3)2 and SiCH(CH3)2). 13C{1H} NMR (101 MHz, CDCl3): δ
138.7, 138.4, 138.3, 138.2, 128.5, 128.5, 128.5, 128.4, 128.1, 128.0,
128.0, 127.9, 127.8, 127.8, 127.7, 104.8, 87.5, 86.3, 82.7, 79.4,
77.9, 75.8, 75.4, 75.2, 73.6, 70.4, 68.8, 18.8, 11.4. HRMS (ESI, Q-TOF): m/z calcd for C45H60NO5Si [M+NH4]+ 722.4235, found 722.4245.
To a solution of 27.2 (6.230 g, 8.84 mmol)
in acetic anhydride (45 mL) was added slowly BF3·OEt2 (9 mL, 74 mmol) at 0 °C. The solution was then warmed
to rt and stirred for 3 days at rt. After the solution was cooled
to 0 °C, the reaction was neutralized by addition of NaHCO3 (satd) (100 mL). The mixture was diluted with EtOAc (200
mL) and H2O (200 mL), and the layers were separated. The
aqueous phase was extracted with EtOAc (200 mL), and then the combined
organic layers were washed with H2O (3 × 200 mL),
NaHCO3 (satd) (3 × 200 mL), and brine (200 mL), dried
over Na2SO4, and concentrated in vacuo. The
crude product was purified by column chromatography (PE/EtOAc 5:1)
to obtain compound 27.3 as a thick yellow syrup (3.527
g, 6.88 mmol, 78%). 1H NMR (400 MHz, CDCl3):
δ 5.19–5.04 (m, 3H), 4.30–4.19 (m, 2H), 4.10 (dd,
12.4 Hz, 2.3 Hz, 1H, H-6a) 3.64 (m, 1H, H-5), 2.09 (s, 3H, CH3COO−), 2.01 (s, 3H, CH3COO−), 2.01
(s, 3H, CH3COO−), 1.99 (s, 3H, CH3COO−),
1.04 (br s, 21H, SiCH(CH3)2 and SiCH(CH3)2). 13C{1H} NMR (101 MHz, CDCl3): δ 170.9,
170.5, 169.5, 169.1, 100.7, 89.3, 76.0, 74.0, 71.5, 69.2, 68.3, 62.2,
20.9, 20.8, 20.8, 20.7, 18.6, 11.2. HRMS (ESI, Q-TOF): m/z calcd for C25H41O9Si [M + H]+ 513.2520, found 513.2544; C25H40O9SiNa [M + Na]+ 535.2340, found 535.2351.
Compound 27.3 (3.515 g, 6.86 mmol) was dissolved
in MeOH (40 mL). The minimum amount of dioxane necessary to obtain
a clear solution was added, and the reaction mixture was treated with
an aqueous solution of NaOH (1 M, 500 μL) to obtain a basic
pH (pH ≈ 8). The reaction mixture was stirred at rt until complete
conversion of the starting material (checked by TLC PE/EtOAc 3:1),
and then it was neutralized with Amberlite IR120 H+ form
resin (monitored using pH paper). After filtration, the resin was
washed with MeOH, and removal of the solvent under reduced pressure
gave 2.36 g of 27.4 as a white foam (quantitative yield).
The crude compound was used in the next step without any further purification. 1H NMR (400 MHz, CDCl3): δ 4.03 (d, J = 9.1 Hz, 1H, H-1), 3.92 (ddd, 12.2 Hz, 5.6 Hz, 3.5 Hz,
1H, H-6a), 3.80 (dd, J = 12.0, 5.2 Hz, 1H, H-6b),
3.61 (t, J = 9.0 Hz, 1H, H-4), 3.54 (t, J = 8.7 Hz, 1H, H-3), 3.47 (t, J = 9.1 Hz, H-2),
3.36 (m, 1H, H-5), 1.08 (s, 21H, SiCH(CH3)2 and SiCH(CH3)2). 13C{1H} NMR (101 MHz, CDCl3):
δ 103.3, 88.5, 79.5, 77.6, 74.2, 71.4, 69.8, 62.0, 18.8, 18.8,
11.2. HRMS (ESI, Q-TOF): m/z calcd
for [M + H]+ 345.2097, found 345.2093.
Compound 27.4 (2.356 g, 6.84 mmol) and CSA
(450 mg, 1.94 mmol) were dissolved in DMF (15 mL) in a 50 mL round-bottom
flask and reacted with anisalaldehydedimethyl acetal (2.4 mL, 13.68
mmol) at 60 °C under reduced pressure on a rotary evaporator.
After 1 h, the reaction was completed (monitored by TLC, PE/EtOAc
1:3). The reaction mixture was cooled to rt and neutralized with triethyl
amine (5 mL), which changed the color of the solution from red to
bright yellow. The mixture was concentrated under reduced pressure
and the crude product was purified by column chromatography (PE/EtOAc
3:1) to afford 27.5 as a white foam (2.56 g, 5.54 mmol,
81%). 1H NMR (400 MHz, CDCl3): δ 7.44–7.38
(m, 2H, ArH), 6.92–6.85 (m, 2H, ArH), 5.49 (s, 1H, p-OMe-C6H4-CH), 4.34
(dd, J = 10.5, 4.9 Hz, 1H, H-6a), 4.12 (d, J = 9.4 Hz, 1H, H-1), 3.82–3.70 (m, 5H, OCH, H-6b and H-3), 3.64 (t, J = 9.3
Hz, 1H, H-2), 3.55 (t, J = 9.2 Hz, 1H, H-4), 3.45
(m, 1H, H-5), 2.87(s, 1H, OH), 2.50 (s, 1H, OH), 1.09 (s, 21H, SiCH(CH3)2 and SiCH(CH)2). 13C{1H} NMR (101 MHz, CDCl3): δ 160.4, 129.5,
127.7, 113.8, 102.7, 102.0, 89.2, 80.6, 75.0, 74.3, 71.9, 70.8, 68.7,
55.5, 18.7, 11.2. HRMS (ESI, Q-TOF): m/z calcd for C25H39O6Si [M + H]+ 463.2510, found 463.2515.
To the solution of 26.7 (2.887 g, 5.28 mmol)
in MeOH/THF (2:1, 30 mL) was added pyridinium paratoluensulfonate
(PPTS, 132.7 mg, 0.528 mmol). After being stirred for 3 days at rt,
the mixture was diluted with Et2O (100 mL) and neutralized
with NaHCO3 (satd, 75 mL). The reaction was slightly exothermic
and led to the precipitation of a white solid, which remained in the
aqueous phase. The phases were separated and the aqueous layer was
extracted three times with Et2O (100 mL). The combined
organic layers were dried over Na2SO4 and concentrated
under reduced pressure to give a yellow oil, which was purified by
column chromatography (PE/EtOAc 3:2). Pure 27.7 was recovered
as a colorless oil (2.053 g, 4.79 mmol, 91%). 1H NMR (400
MHz, CDCl3): δ 5.04 (t, J = 9.5
Hz, 1H, H-2), 4.97 (t, J = 9.2 Hz, 1H, H-3), 4.23
(d, J = 9.6 Hz, 1H, H-1), 3.91 (dd, J = 12.1, 3.1 Hz, 1H, H-6a), 3.80 (dd, J = 12.2 Hz,
4.5 Hz, 1H, H-6b), 3.72 (t, J = 9.3 Hz, 1H, H-4),
3.37 (dt, J = 9.7, 3.4, 3.4 Hz, 1H, H-5), 3.02 (br
s, 2H, OH), 2.07 (s, 3H, CH3COO−), 2.01 (s, 3H,
CH3COO−), 1.03 (br s, 21H, SiCH(CH3)2 and SiCH(CH3)2). 13C{1H} NMR (101 MHz, CDCl3): δ 171.8, 169.3, 101.3, 88.8, 79.7, 77.0, 71.6, 69.3,
69.0, 62.2, 21.0, 20.8, 18.6, 11.2. HRMS (ESI, Q-TOF): m/z calcd for C21H37O7Si [M + H]+ 429.2303, found 429.2318.
Benzoyl chloride (834 μL, 7.18 mmol) was added slowly
to a solution of 27.7 (2.05 g, 4.79 mmol) in dry pyridine
(45 mL) at 0 °C, and the reaction mixture was stirred for another
30 min at 0 °C. The reaction was quenched by the addition of
methanol, and after dilution with EtOAc, it was washed successively
with water (2 × 150 mL), HCl (2M) (2 × 150 mL), NaHCO3(satd) (2 × 150 mL), and brine (150 mL). Drying over
Na2SO4 and evaporation of the solvent in vacuo
gave a pale yellow oil which was purified by column chromatography
(PE/EtOAc 3:1) to afford pure 27.8 as a white foam (1.966
g, 3.69 mmol, 77%). 1H NMR (400 MHz, CDCl3):
δ 8.11–8.05 (m, 2H, ArH), 7.63–7.56 (m, 1H, ArH),
7.50–7.43 (m, 2H, ArH), 5.09–5.03 (m, 2H, H-2, H-3),
4.75 (dd, J = 12.3, 3.8 Hz, 1H, H-6a), 4.55 (dd, J = 12.3, 2.2 Hz, 1H, H-6b), 4.24 (d, J = 9.6 Hz, 1H, H-1), 3.67 (t, J = 9.2 Hz, 1H, H-4),
3.59 (m, 1H, H-5), 2.07 (s, 3H, CH3COO−), 2.03 (s,
3H, CH3COO−), 1.04 (br s, 21H, SiCH(CH3)2 and SiCH(CH3)2). 13C{1H} NMR (101 MHz, CDCl3): δ 171.5, 169.3, 167.5, 133.6, 130.1, 129.6, 128.6,
101.3, 88.8, 78.5, 76.4, 71.5, 69.3, 68.9, 63.5, 21.0, 20.8, 18.6,
11.2. HRMS (ESI, Q-TOF): m/z calcd
for C28H 41O8Si [M + H]+ 533.2565, found 533.2578.
A solution of 27.8 (1. 960 g, 3.68 mmol) in
dry CH2Cl2 (25 mL) was cooled to −15
°C, and dry pyridine (2.5 mL, 10% v/v) was added at once, followed
by neat triflic anhydride (2 mL, 11.09 mmol), which was added dropwise
at −15 °C. The solution was stirred at −15 °C
for 30 min and then quenched by addition of KHSO4 (1 M).
The reaction mixture was allowed to reach rt and then diluted with
CH2Cl2 and water. The phases were separated,
and the organic layer was washed with water (2 × 40 mL) and brine
(40 mL) and dried over Na2SO4. Removal of solvent
under reduced pressure gave a dark yellow oil, which was dissolved
in a minimum amount of DMF (10 mL) and reacted with NaNO2 (890 mg, 12.9 mmol) at rt for 19 h. Brine (20 mL) was added, and
the mixture was stirred for another 30 min to hydrolyze the nitro
ester intermediate. After dilution with CH2Cl2 (50 mL) and separation of the phases, the organic layer was washed
two more times with brine (2 × 40 mL), dried over Na2SO4, and concentrate in vacuo. The crude
product was purified by column chromatography (PE/EtOAc 4:1) to afford 27.9 as a white foam. 1.38 g, 4.15 mmol, 70% in two steps. 1H NMR (400 MHz, CDCl3): δ 8.06–7.96
(m, 2H, ArH), 7.58–7.50 (m, 1H, ArH), 7.46–7.37 (m,
2H, ArH), 5.44 (t, J = 10.0 Hz, 1H, H-2), 4.96 (dd, J = 10.0, 3.2 Hz, 1H, H-3), 4.62 (dd, J = 11.0, 5.5 Hz, 1H, H-6a), 4.49 (dd, J = 11.4,
6.0 Hz, 1H, H-6b), 4.22 (d, J = 10.0 Hz, 1H, H-1),
4.15–4.11 (m, 1H, H-4), 3.86 (t, J = 6.5 Hz,
1H, H-5), 2.53 (br s, 1H, OH), 2.10 (s, 3H, CH3COO−),
2.04 (s, 3H, CH3COO−), 1.06 (s, 21H, SiCH(CH3)2 and SiCH(CH3)2). 13C NMR (101 MHz, CDCl3): δ 170.3, 169.2, 166.6, 133.5, 129.9, 129.7, 128.6,
101.2, 88.7, 76.0, 74.0, 69.6, 69.1, 67.6, 62.7, 21.0, 20.9, 18.6,
11.2. HRMS (ESI, Q-TOF): m/z calcd
for C28H44NO8Si [M+NH4]+ 550.2836, found 550.2849.
A solution of 27.9 (871 mg, 1.64 mmol) in dry
CH2Cl2 (10 mL) was cooled to −15 °C,
and dry pyridine (1 mL, 10% v/v) was added at once, followed by neat
triflic anhydride (800 μL, 4.67 mmol) which was added dropwise
at the same temperature. The solution was stirred at −15 °C
for 20 min and then quenched by the addition of KHSO4 (1M,
10 mL). The reaction mixture was allowed to reach rt and then was
diluted with CH2Cl2 and water. The phases were
separated, and the organic layer was washed with water (2 × 30
mL) and brine (30 mL) and dried over Na2SO4.
Removal of the solvent under reduced pressure gave a yellow oil, which
was dissolved in acetone (8 mL), treated with an aqueous solution
of NaN3 (533.0 mg, 8.2 mmol, in 2 mL of water), and stirred
at rt for 24 h. The acetone was subsequently removed under reduced
pressure, and the residue was dissolved in EtOAc (40 mL), washed twice
with NaHCO3 (satd) (30 mL), once with water (30 mL)
and once with brine (30 mL), dried over Na2SO4, and concentrated in vacuo. Purification by column chromatography
(PE/EtOAc 92:8) gave 616.5 mg of 27 as a white solid
(1.11 mmol, 67% in two steps). 1H NMR (400 MHz, CDCl3): δ 8.09–8.04 (m, 2H, ArH), 7.62–7.55
(m, 1H, ArH), 7.50–7.43 (m, 2H, ArH), 5.18–5.10 (m,
2H, H-2, H-3), 4.64 (dd, J = 12.3, 2.3 Hz, 1H, H-6a),
4.50 (dd, J = 12.3, 4.4 Hz, 1H, H-6b), 4.24 (d, J = 10.0 Hz, 1H, H-1), 3.74 (t, J = 9.9
Hz, 1H, H-4), 3.54 (m, 1H, H-5), 2.09 (s, 3H, CH3COO−),
2.03 (s, 3H, CH3COO−), 1.03 (s, 21H, SiCH(CH3)2 and SiCH(CH3)2). 13C{1H} NMR (101
MHz, CDCl3): δ 170.1, 169.5, 166.2, 133.4, 129.9,
129.8, 128.6, 100.8, 89.3, 76.3, 74.9, 71.7, 69.2, 63.6, 60.4, 20.8,
20.7, 18.6, 11.1. HRMS (ESI, Q-TOF): m/z calcd for C28H 40N3O7Si [M + H]+ 558.2635, found 558.2662; C28H39N3O7SiNa [M + Na]+ 580.2455,
found 580.2473.
4-(Prop-2-yn-1-yloxy)phenol (20b.1)
To the solution of hydroquinone (0.220
g, 2.0 mmol) in DMF
(10 mL) were added sequentially potassium carbonate (0.138 g, 1.0
mmol) and propargyl bromide (0.117 g, 1.0 mmol). The resulting system
reacted at 60 °C. After 4 h, dichloromethane (50 mL) was added.
The organic layer was washed with 10% HCl and water and dried with
sodium sulfate. After removal of the solvent, the compound was purified
by column to afford the product as a yellowish syrup (121 mg, 820
μmol, 41%). 1H NMR (400 MHz, CDCl3): δ
6.91–6.82 (m, 2H, ArH), 6.82–6.73 (m, 2H, ArH), 5.41
(br s, 1H, −OH), 4.66–4.60 (d, J =
2.3 Hz, 2H, −CH2−), 2.51 (t, J = 2.3 Hz, 1H, −C≡CH). 13C{1H}
NMR (101 MHz, CDCl3): δ 151.8, 150.4, 116.5, 116.2,
78.9, 75.6, 56.8. The spectrum was in accordance with a published
paper.[7]
Authors: Ruxana T Sadikot; Timothy S Blackwell; John W Christman; Alice S Prince Journal: Am J Respir Crit Care Med Date: 2005-02-01 Impact factor: 21.405
Authors: Ciaran O'Reilly; Salvador Blasco; Bina Parekh; Helen Collins; Gordon Cooke; Thorfinnur Gunnlaugsson; Joseph P Byrne Journal: RSC Adv Date: 2021-05-04 Impact factor: 4.036
Authors: Pouya Zaree; Javier Sastre Torano; Cornelis A M de Haan; Richard A Scheltema; Arjan Barendregt; Vito Thijssen; Guangyun Yu; Frits Flesch; Roland J Pieters Journal: Glycobiology Date: 2021-12-18 Impact factor: 4.313
Authors: Wenjing Lu; Wenjuan Du; Victor J Somovilla; Guangyun Yu; Diksha Haksar; Erik de Vries; Geert-Jan Boons; Robert P de Vries; Cornelis A M de Haan; Roland J Pieters Journal: J Med Chem Date: 2019-06-28 Impact factor: 7.446
Authors: Xuan Wei; Wenjuan Du; Margherita Duca; Guangyun Yu; Erik de Vries; Cornelis A M de Haan; Roland J Pieters Journal: J Med Chem Date: 2022-05-12 Impact factor: 8.039
Figure 1
Figure 2
Scheme 1
Scheme 2
Scheme 3
Figure 3