Dynamic combinatorial libraries of 2,5-diformylfuran-derived macrocycles

Article abstract of DOI:10.1021/jo501966q

A series of polyazamacrocycles, containing a furan moiety, have been prepared using the all-in-solution approach of dynamic combinatorial chemistry. The methodology involves the use of a range of simple, fully soluble inorganic salts as templates and fast imine-to-amine reduction followed by high-performance liquid chromatography screening for the best reaction conditions. It offers an elegant and labor-efficient alternative to the classical methodology of imine trapping via crystallization of complexes. For all the presented 2,5-diformylfuran-derived libraries, the templates provided control over the libraries behavior, which was reflected in increased isolated yields of the corresponding macrocyclic amines, compared to those of nontemplated libraries. The key parameters for achieving true thermodynamic control over the system, which are macrocyclization kinetics and imine reduction kinetics using NaBH₄ accompanied by various protic additives, have been discussed.

Full text of DOI:10.1021/jo501966q

Article
pubs.acs.org/joc
Dynamic Combinatorial Libraries of 2,5-Diformylfuran-Derived
Macrocycles
Krzysztof Ziach,† Aleksandra Obrocka-Hrycyna,† and Janusz Jurczak*,‡
†Faculty of Chemistry, University of Warsaw, Pasteura 1, 02-093 Warsaw, Poland
‡Institute of Organic Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland
S
* Supporting Information
ABSTRACT: A series of polyazamacrocycles, containing a
furan moiety, have been prepared using the all-in-solution
approach of dynamic combinatorial chemistry. The method-
ology involves the use of a range of simple, fully soluble
inorganic salts as templates and fast imine-to-amine reduction
followed by high-performance liquid chromatography screen-
ing for the best reaction conditions. It offers an elegant and
labor-efficient alternative to the classical methodology of imine
trapping via crystallization of complexes. For all the presented
2,5-diformylfuran-derived libraries, the templates provided
control over the libraries’ behavior, which was reflected in
increased isolated yields of the corresponding macrocyclic
amines, compared to those of nontemplated libraries. The key
parameters for achieving true thermodynamic control over the system, which are macrocyclization kinetics and imine reduction
kinetics using NaBH4 accompanied by various protic additives, have been discussed.
However, only a few papers in which DCC principles are
applied to azamacrocycle synthesis via the imination reaction
have been published, focusing on the use of aromatic
dialdehydes such as phthalic aldehydes,11,12 2,6-diformylpyr-
idines,13−16 diformylbipyrrole,17 and 2,5-diformylfuran.18 The
ability to control the behavior of the dynamic combinatorial
library (DCL) is obviously limited by the ability to reliably
analyze the composition of the library at equilibrium.19 The
historically first and now routinely used approach involves
“freezing” of imine libraries by reduction to amines using
variety of reducing agents such us NaBH3CN,20 BH3·2,6-
lutidine,21 BH3·THF,22 or NaBH4.11−16 Recently, we have
found that careful NaBH4 reduction can transform the whole
primary library of imines into a resulting secondary library of
amines in <1 min,23 which is much faster than of any of the
reequilibration processes that may take place.24 The secondary
library of amines can easily be analyzed using reversed phase
HPLC.
INTRODUCTION
■
The imination reaction was introduced early1 to supramolecular
chemistry as a promising way of synthesizing azamacrocycles,
because it requires readily available substrates. The reaction
proceeds under mild conditions and offers compatibility with a
range of ionic templates. However, the rapid reversibility of this
reaction was considered a serious drawback, because of the
instability of the products. Control over imine-based systems
was thus gained by suppressing the reversibility. Typically, this
was achieved by trapping macrocyclic imine complexes in a
solid form,2,3 which always led to the least soluble species, but
not necessarily to those intended. A new paradigm was
introduced by dynamic combinatorial chemistry (DCC).4−9
DCC treats the imine mixtures as a library of interconverting
species (Scheme 1), the composition of which is governed by
the thermodynamic stability of particular library members.10
Scheme 1. General Scheme of Macrocyclic Imine Library
Formation from Dialdehydes and Diamines
In this contribution, we would like to present an application
of this methodology to the study of the dynamics of the
nontemplated and templated libraries derived from 2,5-
diformylfuran (1) and diaminoethers (2−5) that vary in length
(shown in Figure 1). We have chosen fully soluble alkali metal
salts to act as templates that can amplify the formation of small-
and medium-ring polyazamacrocycles.
Received: August 25, 2014
Published: October 30, 2014
© 2014 American Chemical Society
10334
dx.doi.org/10.1021/jo501966q | J. Org. Chem. 2014, 79, 10334−10341

The Journal of Organic Chemistry
Article
The addition of methanol as a reduction cosolvent shortened
the reduction time to ∼3 min for the nontemplated (1+2)
library (Figure 2c), which should be still considered too slow
for some rapidly equilibrating libraries.24−27 Water was also
found to accelerate imine reduction (Figure 2c), but even less
efficiently than MeOH. Finally, the best additive we found was
1% (v/v) TFA either in water (Figure 2d) or in MeOH (see
Figure S1 of the Supporting Information). Both reduction
systems were able to transform (1+2) libraries in MeCN into
the corresponding amines in <1 min. In this project, we decided
to use a H2O/TFA solvent for reduction in MeCN-based
libraries. UV-monitored reduction kinetic experiments, using
MeOH or MeCN/H2O/TFA protocols, were performed for all
dialdehyde/diamine reagent pairs in both solvents. In all the
cases, reduction took <1 min, which is consistent with previous
observations made for phthalic12 and bisfurylmethyl alde-
hydes23,24 and practically ensures the freezing of the
equilibrium imine composition into a resulting secondary
library of amines.
Although direct measurements of the composition of imine
DCLs by NMR13−15,28 or mass spectrometry14 were proposed
for some substrate systems, we have found both techniques
unsuitable for multimembered libraries such as those presented
here. 1H NMR spectra were complicated, and it was practically
impossible to distinguish, assign, and integrate signals from
particular library members. On the other hand, ESI mass
spectrometry showed a tendency to oversimplify the libraries by
presenting only macrocyclic species and also suggested false
templation,18 a problem we have previously encountered with
phthalaldehyde-based libraries.12 However, having in hand a
fast reduction protocol, we applied it to the monitoring of DCL
evolution with a nearly 1 min resolution on the time scale.
Aliquots of the library mixtures were collected over the reaction
time and reduced and subjected to HPLC analysis. Figure 3
presents the chromatograms obtained for nontemplated (1+3)
libraries in methanol and acetonitrile. In the first case, the DCL
reaches equilibrium in approximately 30 min, while in the
second case, it takes ∼6 h; however, the changes are already
minor after 3 h.
Figure 1. Imine library building blocks.
RESULTS AND DISCUSSION
■
In the case of imine libraries that are subsequently reduced to
amines, the key parameters that need to be determined prior to
the actual library experiments are the kinetics of macro-
cyclization and the kinetics of imine-to-amine reduction. This is
to ensure that the protocols applied do not bias the libraries’
responses to templates and that the concentrations measured
reflect the situation at equilibrium. Kinetics of reduction can be
easily monitored by UV spectroscopy because of the distinct
absorption maxima of 2,5-diconjugated CO and CN
derivatives (280−300 nm) and reduced species (∼220 nm). In
this project, the DCL created from dialdehyde 1 and diamine 2
[abbreviated (1+2)] was used as a model to examine the
behavior of the system under the influence of various additives
to satisfy the need for an external source of protons.18,25 In
particular, when methanol was used as a solvent, NaBH4
reduction of either the (1+2) library (nontemplated or
templated) or pure aldehyde 1 took <1 min (Figure 1a).
In acetonitrile, however, only dialdehyde 1 was reduced (in
10−15 min), while the (1+2) library stayed intact for at least 1
h (Figure 2b) if no external source of protons was provided.
HPLC analysis also revealed that library formation proceeded
through many intermediate compounds that, after being
initially formed, were then consumed as the DCL evolution
progressed. In particular, initial ESI-MS studies, performed 15
min after mixing substrates 1 and 3 in MeCN, suggested a
significant contribution from the [1+1] macrocycle, the
analogue of diaza-15-crown-5, which could possibly be a kinetic
product that was later unsupported at equilibrium. Indeed,
HPLC monitoring of the reaction during the first minutes of
library formation confirmed that there was a new substance that
vanished over time. However, after the reaction mixture had
been frozen at 15 min, instead of the [1+1] macrocycle we
isolated a linear compound 6 (Figure 3) in 52% yield but could
not identify and isolate the [1+1] macrocycle. The other
isolated compound was diol 7, the product of reduction of
unreacted dialdehyde 1.
Similar macrocyclization kinetic studies were performed for
all reactant pairs, in both solvents, in the absence or presence of
the inorganic salt (2 equiv of NaClO4·H2O), and these data are
summarized in Table 1.
In methanol, as well as in acetonitrile, equilibrium kinetics
varied by ∼1 order of magnitude, depending on the amine
used, despite the fact that their reactivity is supposed to be very
similar. In all cases, the reaction in MeOH was significantly
Figure 2. Kinetics of (1+2) imine library reduction with NaBH4,
measured by UV spectroscopy: (a) reactions in MeOH without
additives, (b) reactions in MeCN without additives, (c) reactions in
MeCN with water or MeOH added subsequent to NaBH4, and (d)
reactions in MeCN with 1% aqueous TFA added subsequent to
NaBH4. In the legends, protic additives are given in italics and [H]
stands for NaBH4 reduction.
10335
dx.doi.org/10.1021/jo501966q | J. Org. Chem. 2014, 79, 10334−10341

The Journal of Organic Chemistry
Article
Scheme 2. Macrocycles Isolated from the Nontemplated
a
Libraries (upon NaBH4 reduction)
Figure 3. (1+3) imine library macrocyclization kinetics measured by
HPLC of a secondary library of amines: (a) nontemplated library in
MeOH and (b) nontemplated library in MeCN.
Table 1. Equilibration Times of the Libraries, Estimated by
HPLC Analysis of Secondary Libraries of Amines
a
substrates
solvent
nontemplated NaClO4-templated acceleration
1 and 2
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
MeOH
MeCN
60 min
30 min
120 min
15 min
60 min
30 min
120 min
5 min
2-fold
1.5-fold
2-fold
6-fold
4-fold
4-fold
6-fold
3-fold
180 min
30 min
1 and 3
1 and 4
1 and 5
360 min
120 min
480 min
30 min
360 min
120 min
a
The ratio of the equilibration time of the templated reaction to that
of the nontemplated one.
a
For the yields, the first value corresponds to the reaction in MeOH
while the value in parentheses to the reaction in MeCN.
faster than in MeCN. Further, for all the libraries, the presence
of inorganic salt increased the reaction rate several-fold, even if
the additive was later found not to be an active template
promoting any of the DCL members. We can speculate that
inorganic salts may act as weak Lewis acids activating aldehyde
to nucleophilic attack. However, no particular acceleration by
the inactive template was observed in the previously described
imine system,24 which altogether suggests that there is no
simple mode of inorganic salt action. Taking into account the
differences in macrocyclization kinetics, and also the slow
dissolution of some of the templates, we conducted all the
reactions for 16−24 h.
We started our studies of the library dynamics by analyzing
nontemplated DCLs (Scheme 2) upon reduction to the
corresponding secondary libraries of amines as described
above. The most abundant products were isolated and
characterized and served as a reference material for subsequent
studies. The isolated yields were calculated taking into account
only the pure fractions collected upon single-column
chromatography on silica gel and so reflect the combined
chemical yield and purification efficiency.
In all the DCLs but one, (1+5), [2+2] macrocycles
dominated and were the only ones isolated in pure form.
Some other aspects are worth mentioning. First, the solvent
used (MeOH or MeCN) seemed not to play any important
role in the compositions of nontemplated libraries because the
yields were virtually the same. Second, furan analogues of diaza-
15-crown-5 [[1+1], in the (1+3) library] and diaza-18-crown-6
[[1+1], in the (1+4) library] were not detected. The (1+4)
library also yielded a [3+3] macrocycle, although it was impure
and present in only a small quantity. Third, and most strikingly,
despite the small difference in amine partner length (13 vs 15
atoms), there was a qualitative difference in the reaction of
10336
dx.doi.org/10.1021/jo501966q | J. Org. Chem. 2014, 79, 10334−10341

The Journal of Organic Chemistry
Article
a
amines 4 and 5. The latter was the only case in which
nontemplated DCL produced a [1+1] macrocycle (60−63% for
amine 17), whereas the only compound we managed to isolate
from the (1+4) library was a [2+2] macrocycle (22−25% for
amine 15) not previously described in the literature.
Table 2. Template Effects in (1+4) Libraries
MeOH
MeCN
template (2 equiv)
[1+1]
[2+2]
[1+1]
[2+2]
−
5
40 (25)
traces
39
50 (22)
30
b
The structure of the library building blocks and thus the
structure of the products made alkali metal cations the
templates of choice. The following salts from this group were
used: LiClO4·3H2O, NaClO4·H2O, KSCN, RbCl, and CsSCN,
accompanied by PbCl2·2H2O, SrCl2·6H2O, and CaCl2. The
latter three cations proved to be effective in the crystallization
approach described for the (1+2) and (1+3) reagent pairs by
Nelson29,30 and for the (1+4) pair by Fenton.31,32 We use no
common counterion. Instead, the particular metal salts were
chosen for their best solubility in MeOH and MeCN, especially
considering that all the experiments were conducted at
relatively high concentrations (50 mM with respect to the
substrates) in the presence of 2 molar equiv of inorganic salt.
Our previous studies did not indicate any important role of the
kind of anion in library equilibria, other than providing
solubility.12,23,24
LiClO4
traces
80 (80)
76
40
0
c
NaClO4
65 (70)
−
21 (10)
−
ce
,
NaClO4
KSCN
0
49
24
22
25
61
22
d
RbCl
42
16
41
CsSCN
34
47
24
a
Data obtained by HPLC analysis of the corresponding secondary
libraries of amines, presented as relative concentrations of key library
members (percent of the corresponding HPLC peak with respect to
the whole chromatogram at 222 nm) [isolated yields (percent) in
b
c
d
parentheses]. As LiClO4·3H2O. As NaClO4·H2O. Salt not fully
e
soluble in MeCN. One equivalent was used.
Scheme 4. [2+2] and [1+1] Macrocycle Selection in the
Nontemplated and Na+-Templated (1+4) Library
In the (1+2) DCLs, no qualitative change in the composition
of the libraries in comparison to those of the nontemplated
ones was observed using HPLC. However, several templates
increased the isolated yield of [2+2] macrocycle 11. The most
effective were NaClO4·H2O (75% of 11 in MeCN; in MeOH,
the template was inactive) and PbCl2·2H2O (67% of 11 in
MeOH; in MeCN, the template was not soluble). The (1+3)
library exhibited behavior toward the templates similar to that
of (1+2) DCL. No new species were formed, but the yield and
purity of the macrocycles observed previously in the non-
templated libraries increased. Templation by KSCN in
acetonitrile was most effective and allowed isolation of [2+2]
tetraamine 8 and [3+3] hexaamine 9 in 25 and 14% yields,
respectively (Scheme 3). The latter 45-member macrocycle was
not isolated in a pure form from the nontemplated DCLs.
somewhat unexpected, because in the previously studied
systems,23,24 acetonitrile was always found to be a less
demanding solvent in terms of templation. A stoichiometric
amount of NaClO4 in MeOH was already found to be efficient.
The possibility of “switching” the DCL response between
[1+1] and [2+2] macrocycles by templation (Scheme 4) was
used to demonstrate thermodynamic control over the library
composition (Figure 4 and Table S2 of the Supporting
Information). To a pre-equilibrated (1+4) DCL in methanol,
which consisted of 40% [2+2] and 5% [1+1] as determined by
HPLC (Figure 4a), was added 1 molar equiv of NaClO4·H2O.
A stoichiometric amount of template already led to conversion
of most of the library material into the [1+1] macrocycle [72%
as determined by HPLC (Figure 4b)]. Then, sodium cations
were “removed” by addition of a known effective Na+ ligand,
namely 15-crown-5, in 5 molar equiv, and the new equilibrium
was established in which the [2+2] macrocycle dominated
again [38% as determined by HPLC, accompanied by 11%
[1+1] (Figure 4c)]. Then, an additional 5 equiv of NaClO4·
H2O was added, and because at least 1 equiv of Na+ became
accessible for the library, it produced the [1+1] macrocycle
[73% as determined by HPLC (Figure 4d)]. As described
above, an additional amount of template was “removed” by an
excess of 15-crown-5 [35% [2+2] and 11% [1+1] as
determined by HPLC (Figure 4e)]. In the sixth equilibration
step, sodium salt was added again to the DCL and the library
reequilibrated toward the [1+1] macrocycle [70% as
determined by HPLC (Figure 4f)]. To the best of our
knowledge, this is so far the longest sequence of multiple
Scheme 3. Templation by KSCN in the (1+3) Library in
MeCN
In the (1+4) library, alkali metal cations were able to strongly
influence the equilibrium (Table 2), amplifying the formation
of [1+1] macrocycle 18 (Scheme 4), which was not isolated
from the nontemplated DCLs. The smallest cation, Li+, was
able to amplify formation of only the [1+1] macrocycle in
acetonitrile (39% by HPLC, while that of the [2+2] macrocycle
decreased to 30%) and was completely inactive in methanol.
We have previously described a similar solvent dependence on
Li+ action.24 In contrast, Na+ was found to be the best template
for the [1+1] macrocycle in MeOH, where the [2+2]
macrocycle disappeared completely and the isolated yield of
macrocyclic [1+1] diamine 19 (Scheme 4) was 80%, while in
MeCN, some [2+2] macrocycle was still present. This is
10337
dx.doi.org/10.1021/jo501966q | J. Org. Chem. 2014, 79, 10334−10341

The Journal of Organic Chemistry
Article
Table 3. Content of [1+1] Macrocycle 17 in the (1+5)
a
DCLs
template (2 equiv)
MeOH
MeCN
−
51 (63)
57
57 (60)
80
b
LiClO4
c
NaClO4
43
85 (60)
92 (94)
58
KSCN
98 (98)
90
d
RbCl
CsSCN
92 (85)
63
a
b,c,d
See footnotes
of Table 2.
The similar sizes of amines 4 and 5 used as DCL building
blocks remain in contrast with their preference for the
formation of variably sized macrocycles under template-free
conditions. In addition, different cations, Na+ and K+, were
found to be the most effective templates for (1+4) and (1+5)
DCLs, respectively. This prompted us to conduct competition
experiments in which 1 equiv of each diamine 4 and 5 was
reacted with only 1 equiv of dialdehyde 1. MeOH was chosen
as a solvent. The HPLC traces of the corresponding secondary
library of amines are shown in Figure 5.
Figure 4. Multiple changes of the equilibrium in (1+4) DCL by
addition and “removal” of Na+ cations: (a) formation of nontemplated
DCL in MeOH, to which was added in succession (b) 1 equiv of
NaClO4, (c) 5 equiv of 15-crown-5, (d) an additional 5 equiv of
NaClO4·H2O, (e) an additional 25 equiv of 15-crown-5, and (f) an
additional 27 equiv of NaClO4·H2O. Data are presented as
chromatograms of the corresponding secondary libraries of amines.
templations in a single dynamic library. In this case, it was not
possible to continue, because the amount of NaClO4 and 15-
crown-5 already exceeded 30% of the total mass of the mixture,
yet the library responded as expected. Although the binding
constant of [1+1] imine 18 with Na+ was not established,
experiments suggest it is at least comparable with that of 15-
crown-5, because a large excess of crown ether was required. In
the presence of 2 equiv of 15-crown-5, 27% of the [1+1]
macrocycle remained [accompanied by 33% of the [2+2]
macrocycle (Table S1 of the Supporting Information)]. As
much as a 10-fold excess of 15-crown-5 reduced the content of
the [1+1] macrocycle to 6% (accompanied by 38% of the
[2+2] macrocycle), which is virtually the same as in the
nontemplated DCL.
In light of these results, we reinvestigated the results of (1+4)
macrocyclization kinetic experiments. Unlike in the (1+2) and
(1+3) libraries, the [1+1] macrocycle was actually formed in
the nontemplated (1+4) library as the kinetic product, which
was later nearly completely consumed to build up the
thermodynamically favorable [2+2] macrocycle.
Figure 5. Competition experiments in the (1+4+5) DCL in MeOH
measured by HPLC of the corresponding secondary library of amines.
The case with an insufficient amount of the carbonyl
component altered the DCL responses. First, [1+1] macrocycle
18, unsupported in a nontemplated (1+4) library (Table 2),
now became one of the most abundant, even under template-
free conditions, while the previously dominant [2+2] tetraimine
14 almost disappeared. This can be rationalized in light of the
work by Roelens33 and later Severin.34 For a system governed
by thermodynamics, in a situation in which there is an
insufficient amount of a certain building block, those library
members that “use” it in smaller amounts (one molecule of 1 is
needed for the [1+1] macrocycle) gain additional stability over
those that use it extensively (two molecules of 1 for the [2+2]
macrocycle is needed). Stoichiometry is thus another important
factor to take into account while working with imines. Second,
the templation effects observed in the individual (1+4) and
(1+5) DCLs were preserved. The amplification of [1+1]
macrocycle 18 (amine 19 after reduction) was strongest when
NaClO4 was used as a template, whereas [1+1] 16 (amine 17
after reduction) derived from amine 5 dominated in the DCL
templated with KSCN or CsSCN. Although in the (1+4)
libraries K+ and Cs+ effectively templated formation of [1+1]
The reaction of 1 with the longest amine 5 was the only case
in which a 20-member [1+1] macrocycle dominated in the
nontemplated library. Alkali metal templates did not give rise to
any new species but increased the yield of the [1+1]
macrocycle to nearly quantitative with KSCN as a template,
which makes potassium-templated synthesis the method of
choice for this compound (Table 3).
10338
dx.doi.org/10.1021/jo501966q | J. Org. Chem. 2014, 79, 10334−10341

The Journal of Organic Chemistry
Article
hydrogen ceased (typically 30−60 min). Then, the resulting secondary
libraries of amines were subjected to HPLC analysis.
imine 18, in (1+4+5) DCL under competition conditions these
cations preferred the larger [1+1] macrocycle 17. Lithium
stayed inactive, as in the simple (1+4) and (1+5) DCLs in
MeOH. Third, despite the fact that aldehyde-to-amine
stoichiometry is formally biased toward the formation of linear
compounds, the macrocyclization reaction prevails, as no new
significant peaks that may correspond to linear compounds
were detected.
(1+4) DCL Multiple-Templation Experiments. The procedure
follows the “analytical scale” general procedure. The initial library was
made on a 3.5 mL scale; after equilibration, a 0.5 mL aliquot was taken
and reduced, and to the remaining imine DCL was added NaClO4·
H2O. Then, after equilibration, a 0.5 mL aliquot was taken and
reduced, and to the remaining imine was added 15-crown-5 and the
mixture left to equilibrate. This sequence was repeated, and the
collected amine samples were then analyzed by HPLC.
Competition Experiments in the (1+4+5) DCL. The procedure
follows the “analytical scale” general procedure. Libraries were
prepared in MeOH; the concentration of aldehyde 1 was 50 mM,
and 1 equiv of amine 4 and 5 and 2 equiv (with respect to aldehyde 1)
of inorganic salts were used.
HPLC Conditions. The following settings were used: spectral
range of 190−400 nm, detection at λ = 222 nm, flow rate of 2 mL/
min, and injection volume of 20 μL. Crude reaction mixtures of
amines were diluted 50-fold (to 1 mM with respect to the substrates)
with the “A” mobile phase and injected directly without any sample
pretreatment on the column and run in a gradient of solvent mixtures
“A” and “B”, where “A” was a 100:0.1 (v/v) H2O/TFA mixture and
“B” was a 90:10:0.1 (v/v/v) MeCN/H2O/TFA mixture.
(1+3) Library Macrocyclization Kinetic Studies (Figure 3). A
precolumn-guarded Discovery HS C18 column (250 mm × 4.6 mm, 5
μm) was used, with the following gradient program: 95% “A” and 5%
“B” at 0 min to 90% “A” and 10% “B” at 5 min to 80% “A” and 20%
“B” at 15 min to 80% “A” and 20% “B” at 18 min to 95% “A” and 5%
“B” at 21 min.
(1+4) Library (Figure 4). A precolumn-guarded Polaris C-18A
column (250 mm × 4.6 mm, 5 μm) was used, with the following
gradient program: 100% “A” at 0 min to 80% “A” and 20% “B” at 15
min to 80% “A” and 20% “B” at 25 min to 100% “A” at 30 min.
(1+4+5) Library (Figure 5). A precolumn-guarded Polaris C-18A
column (250 mm × 4.6 mm, 5 μm) was used, with the following
gradient program: 95% “A” and 5% “B” at 0 min to 90% “A” and 10%
“B” at 5 min to 80% “A” and 20% “B” at 15 min to 80% “A” and 20%
“B” at 20 min to 95% “A” and 5% “B” at 25 min.
CONCLUSIONS
■
We have presented an application of the all-in-solution
approach of dynamic combinatorial chemistry to the synthesis
of azamacrocycles via an imination reaction. The methodology,
validated in this project for furan-derived macrocycles, involves
the use of a range of simple, fully soluble inorganic salts as
templates and fast imine-to-amine reduction followed by HPLC
screening for the best reaction conditions. It offers an elegant
and labor-efficient alternative to the classical methodology of
imine trapping via crystallization of complexes. For all the
presented 2,5-diformylfuran-derived libraries, the templates
provided control over the libraries’ behavior, which was
reflected in increased isolated yields of the corresponding
macrocyclic amines, as compared to those of nontemplated
libraries. Similarly, DCL dynamics can be controlled by the
choice of reaction solvent or carbonyl-to-amine(s) stoichiom-
etry. Small- and medium-ring azamacrocycles, containing a
furan moiety, can further serve as versatile supramolecular
scaffolds that could be easily modified using, independently,
alkylation/acylation of secondary amines and/or furan
chemistry, such as a Diels−Alder reaction, oxidative ring
opening, or diastereocontrolled ring reduction that may quickly
deliver advanced supramolecular structures.
EXPERIMENTAL SECTION
■
General Procedure for the Formation of Imine Libraries and
Their Transformation into Corresponding Secondary Amine
Libraries on the Preparative Scale. The synthesis follows the
“analytical scale” general procedure. Reactions were conducted on a
1−2 mmol scale. Because of intense foaming in the reduction step, in
the case of MeOH-based libraries, NaBH4 was added in two or three
portions, and in the case of MeCN-based libraries, a water/TFA
solvent was added slowly. When the reduction was finished, the
General Information. All of the reagents and solvents were
purchased from commercial sources and used without further
purification. All the libraries were prepared in HPLC-grade MeOH
or MeCN. Chemical shifts in 1H and 13C NMR spectra were
referenced to tetramethylsilane, used as an internal reference, or to the
residual solvent peak. Column chromatography was performed with
silica gel (100−200 mesh) as the stationary phase. For all HPLC
measurements, an HPLC apparatus equipped with a DAD UV−vis
detector and a low-pressure gradient pump was used. HPLC-grade
MeCN, distilled and further deionized water (18.2 MΩ), and
analytical-grade TFA were used as a mobile phase constituents.
HPLC analyses were performed at room temperature (nonthermo-
stated column).
General Procedure for the Formation of Imine Libraries and
Their Transformation into Corresponding Secondary Amine
Libraries on the “Analytical Scale”. All the reaction mixtures used
for initial screening of libraries’ behavior were prepared on a 1 mL
scale, typically in a series of several experiments. A mixture of
dialdehyde 1 and the proper diamine (2−5) was prepared for the
whole series by mixing solutions of dialdehyde 1 and diamine in a
proper solvent (MeOH or MeCN) to obtain the final concentration of
50 mM with respect to each substrate. This solution was added (1 mL)
to vials containing proper inorganic templates (typically 2 equiv)
immediately upon mixing. The vials, protected from air by a stopper
and parafilm, were then left for 1 day for equilibration. Then, stirring
rods were placed in reaction vials; mixtures were cooled with an ice−
water bath, and powdered NaBH4 was added (5 equiv with respect to
substrates, 0.25 mmol/mL of solution) while stirring was continued. In
the case of MeCN-based libraries, this was followed by immediate
addition of 1 mL of 1% (v/v) TFA in water. The vials (MeOH- or
MeCN-based libraries) were corked with a needle-punctured stopper
and left, while being stirred, in an ice/water bath until the evolution of
mixture was carefully concentrated on the rotavap (possible foaming!)
1
to approximately
/ of the volume (in the case of MeOH-based
4
libraries water, 50% with respect to methanol volume, was added prior
to concentration). Then the residue was acidified with 5% HCl to pH
2, then basified with 25% ammonia followed by 20% NaOH to pH 12,
and extracted five times with CHCl3. Extracts were washed with 2%
NaOH and dried with Na2SO4. The drying agent and solvents were
removed, and the crude was purified by silica gel column
chromatography in a CHCl3/MeOH/NH4OH(25% aq) gradient. Col-
lected samples were evaporated several times with CHCl3 to remove
water and then dissolved in a small volume of MeOH, and CH2Cl2 was
added and filtered via a small plug of Celite to remove dissolved silica
gel. Finally, for tetra- and hexaamines, products were dissolved in
CHCl3 and dried with Na2SO4. All solvents were rotoevaporated, and
traces of solvents were removed under high vacuum (<1 mmHg).
Macrocyclic amines were stored in a freezer.
Yields and Analytical Data. (1+2) Library. Conditions and
yields: nontemplated, in MeCN, 50% (98 mg) of 11; nontemplated, in
MeOH, 49% (96 mg) of 11; NaClO4·H2O templated (2 equiv), in
MeOH, 75% (147 mg) of 11; PbCl2·2H2O templated (2 equiv), in
MeOH, 67% (131 mg) of 11.
1
[2+2] Tetraamine 11. Yellowish, slowly solidifying oil: H NMR
(CDCl3, 200 MHz) δ 2.16 (4H, brs), 2.75 (8H, t, J = 5.0 Hz), 3.52
(8H, t, J = 5.0 Hz), 3.73 (8H, s), 6.07 (4H, s); 13C NMR (50 MHz,
10339
dx.doi.org/10.1021/jo501966q | J. Org. Chem. 2014, 79, 10334−10341

The Journal of Organic Chemistry
Article
CDCl3) δ 46.4, 48.5, 70.4, 107.8, 153.3; LR ESIMS m/z 393.3 [M +
H]+, 415.2 [M + Na]+, 431.2 [M + K]+; HRMS (ESI-TOF) m/z [M +
Na]+ calcd for C20H32N4O4Na 415.2321, found 415.2318.
crystallized form DCM to afford 60 mg (94%) of pure 2,5-
1
dihydroxymethylfuran (7): off-white solid; mp 72−75 °C; H NMR
(300 MHz, MeCN-d4) δ 3.25 (2H, brs), 4.47 (2H, s), 6.23 (2H, s);
(1+3) Library. Conditions and yields: nontemplated, in MeCN,
14% (34 mg) of 8 and traces of 9 (impure); nontemplated, in MeOH,
14% (34 mg) of 8 and traces of 9 (impure); NaClO4·H2O templated
(2 equiv), in MeCN, 18% (42 mg) of 8 and traces of 9 (impure);
NaClO4·H2O templated (2 equiv), in MeOH, 18% (43 mg) of 8 and
traces of 9 (impure); KSCN templated (2 equiv), in MeCN, 25% (118
mg) of 8 and 14% (68 mg) of 9; nontemplated in MeCN, reaction
stopped after 15 min, 52% (70 mg) of 6.
13C NMR (75 MHz, MeCN-d4) δ 56.3, 107.9, 117.4; LR ESIMS m/z
151.2 [M + Na]+, 167.0 [M + K]+, 279.2 [2M + Na]+; HRMS (EI-
TOF) m/z [M+•] calcd for C6H8O3 128.0473, found 128.0475.
Macrocyclization Kinetic Experiments. The protocols follow
the general procedure. Reactions were conducted on a several milliliter
scale to afford 250 μL of solution for every measuring point. Aliquots
were added to precooled (4 °C) vials containing 5 equiv of NaBH4
and a stirring bar, and reduction was further conducted in an ice/water
bath. Then, the resulting secondary libraries of amines were subjected
to HPLC analysis.
[2+2] Tetraamine 8. Yellowish oil: 1H NMR (300 MHz, CDCl3) δ
2.21 (4H, brs), 2.81 (8H, t, J = 5.1 Hz), 3.59−3.63 (16H, m), 3.77
(8H, s), 6.11 (4H, s); 13C NMR (75 MHz, CDCl3) δ 46.2, 48.3, 70.2,
70.4, 107.6, 153.0; LR ESIMS m/z 481.4 [M + H]+, 503.3 [M + Na]+;
HRMS (ESI-TOF) m/z [M + Na]+ calcd for C24H40N4O6Na
503.2846, found 503.2836.
UV−Vis Determination of Imine-to-Amine Reduction Ki-
netics. All measurements were taken using the following settings of
the UV spectrophotometer: optical path of 10 mm (quartz cuvette),
spectral range of 200−335 nm, scan speed of 240 nm/min, bandwidth
of 1 nm, data interval of 1 nm, and temperature of 20−22 °C. The
libraries were reduced according to the common protocols. In
particular, for MeCN-based libraries, protic additives (H2O, MeOH,
H2O/1% TFA, and MeOH/1% TFA) were added immediately after
NaBH4. After a given period of reduction, 30 μL of the reaction
mixture was added to 270 μL of pure solvent (MeOH or MeCN, the
same that was used to generate the library). Then, 30 μL (MeOH
libraries) or 60 μL (MeCN libraries) of the solution described above
was injected into the UV cuvette filled with 3000 μL of the
corresponding solvent (to a final concentration of ∼5 × 10−5 M with
respect to the substrates); the solution was mixed, and spectra were
recorded. Two dilution steps were used to avoid working with very
small volumes of the bubbling reaction mixture. The whole operation
took 20−30 s, which was not included in the “reaction time”.
1
[3+3] Hexaamine 9. Yellow slowly solidifying oil: H NMR (200
MHz, CDCl3) δ 2.15 (4H, brs), 2.79 (12H, t, J = 5.2 Hz), 3.56−3.61
(24H, m), 3.75 (12H, s), 6.08 (6H, s); 13C NMR (50 MHz, CDCl3) δ
46.4, 48.5, 70.4, 70.6, 107.6, 153.2; LR ESIMS m/z 721.4 [M + H]+,
743.4 [M + Na]+; HRMS (ESI-TOF) m/z [M + H]+ calcd for
C36H61N6O9 721.4500, found 721.4471.
1
Compound 6. Yellow slowly solidifying oil: H NMR (200 MHz,
CD3OD) δ 3.08−3.16 (4H, m), 3.69−3.77 (8H, m), 4.16 (2H, s), 4.52
(2H, s), 6.34 (1H, d, J = 3.2 Hz), 6.51 (1H, d, J = 3.2 Hz); 13C NMR
(50 MHz, CD3OD) δ 40.1, 44.9, 47.9, 57.3, 67.9, 68.1, 71.3, 109.6,
113.1, 148.5, 157.2; LR ESIMS m/z 259.2 [M + H]+, 281.2 [M +
Na]+; HRMS (ESI-TOF) m/z [M + H]+ calcd for C12H23N2O4 [M +
H]+ 259.1658, found 259.1679.
(1+4) Library. Conditions and yields: nontemplated, in MeCN,
22% (62 mg) of 15; nontemplated, in MeOH, 25% (72 mg) of 15;
NaClO4·H2O templated (2 equiv), in MeCN, 80% (114 mg) of 19;
NaClO4·H2O templated (2 equiv), in MeOH, 70% (200 mg) of 19
and 10% (29 mg) of 15.
ASSOCIATED CONTENT
* Supporting Information
■
S
Additional UV−vis data, additional data for (1+4) multiple-
templation experiments, and copies of NMR spectra. This
material is available free of charge via the Internet at http://
pubs.acs.org.
1
[1+1] Diamine 19. Yellow oil: H NMR (200 MHz, CDCl3) δ
2.66−2.70 (6H, m), 3.46−3.57 (12H, m), 3.67 (4H, s), 5.99 (2H, s);
13C NMR (50 MHz, CDCl3) δ 46.1, 48.1, 69.8, 69.9, 70.1, 107.4,
152.5; LR ESIMS m/z 285.2 [M + H]+, 307.2 [M + Na]+; HRMS
(ESI-TOF) m/z [M + H]+ calcd for C14H25N2O4 285.1814, found
285.1805.
AUTHOR INFORMATION
Corresponding Author
*E-mail: jurczak@icho.edu.pl.
■
1
[2+2] Tetraamine 15. Brownish, slowly solidifying oil: H NMR
(200 MHz, CDCl3) δ 2.42 (4H, brs), 2.79 (8H, t, J = 5.1 Hz), 3.57−
3.66 (24H, m), 3.76 (8H, s), 6.10 (4H, s); 13C NMR (50 MHz,
CDCl3) δ 46.4, 48.5, 70.47, 70.54, 70.7, 107.8, 153.29; LR ESIMS m/z
569.3 [M + H]+, 591.3 [M + Na]+; HRMS (ESI-TOF) m/z [M + H]+
calcd for C28H49N4O8 569.3550, found 569.3575.
Notes
The authors declare no competing financial interest.
(1+5) Library. Conditions and yields: nontemplated, in MeCN,
60% (188 mg) of 17; nontemplated, in MeOH, 63% (197 mg) of 17;
NaClO4·H2O templated (2 equiv), in MeCN, 60% (187 mg) of 17;
KSCN templated (2 equiv), in MeCN, 94% (293 mg) of 17; KSCN
templated (2 equiv), in MeOH, 98% (307 mg) of 17; CsSCN
templated (2 equiv), in MeOH, 85% (264 mg) of 17.
ACKNOWLEDGMENTS
We acknowledge the National Science Centre (Project 2011/
02/A/ST5/00439) for financial support.
■
DEDICATION
■
1
[1+1] Diamine 17. Dark yellow, slowly solidifying oil: H NMR
This paper is dedicated to Prof. M. Makosza on the occasion of
his 80th birthday.
̨
(300 MHz, CDCl3) δ 1.74−1.82 (4H, m), 2.55 (2H, brs), 2.73 (4H, t,
J = 6.4 Hz), 3.54−3.65 (12H, m), 3.77 (4H, s), 6.10 (2H, s); 13C
NMR (75 MHz, CDCl3) δ 29.4, 46.1, 46.8, 70.3, 70.4, 70.7, 107.7,
153.0; LR ESIMS m/z 313.2 [M + H]+, 351.2 [M + K]+; HRMS (ESI-
TOF) m/z [M + H]+ calcd for C16H29N2O4 313.2127, found
313.2117.
REFERENCES
■
(1) Curtis, N. F. Coord. Chem. Rev. 1968, 3, 3−47 and references
cited therein.
(2) Vigato, P. A.; Tamburini, S. Coord. Chem. Rev. 2004, 248, 1717−
2128 and references cited therein.
(3) Vigato, P. A.; Peruzzo, V.; Tamburini, S. Coord. Chem. Rev. 2012,
256, 953−1114 and references cited therein.
(4) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.;
Stoddart, J. F. Angew. Chem., Int. Ed. 2002, 41, 898−952 and references
cited therein.
(5) Corbett, P. T.; Leclaire, J.; Vial, L.; West, K. R.; Wietor, J.-L.;
Sanders, J. K. M.; Otto, S. Chem. Rev. 2006, 106, 3652−3711 and
references cited therein.
Synthesis of 2,5-Dihydroxymethylfuran (7). To the solution of
62 mg (0.5 mmol) of dialdehyde 1 in 5 mL of methanol, cooled in an
ice/water bath, was added 38 mg of NaBH4 (2 equiv, 1 mmol) in three
portions. When the evolution of hydrogen ceased (approximately 45
min), the ice bath was removed and 5% HCl was added dropwise until
the pH reached 2. Then, the mixture was alkalized with 25% NH3aq
until the pH reached 12 and extracted seven times with DCM. Organic
extracts were washed with brine and dried with MgSO4. After removal
of the drying agent and solvents, the crude product was filtered via
silica with a DCM/2-propanol mixture (97:3) as the eluent and then
10340
dx.doi.org/10.1021/jo501966q | J. Org. Chem. 2014, 79, 10334−10341

The Journal of Organic Chemistry
Article
(6) Cougnon, F. B. L.; Sanders, J. K. M. Acc. Chem. Res. 2012, 45,
2211−2221 and references cited therein.
(7) Li, J.; Nowak, P.; Otto, S. J. Am. Chem. Soc. 2013, 135, 9222−
9239 and references cited therein.
(8) Herrmann, A. Chem. Soc. Rev. 2014, 43, 1899−1933 and
references cited therein.
(9) Matache, M.; Bogdan, E.; Had
2106−2131.
̆
ade, N. D. Chem.Eur. J. 2014, 20,
(10) Recent review devoted to DCC of imines: Belovich, M. E.;
Stooddart, J. F. Chem. Soc. Rev. 2012, 41, 2003−2024.
(11) Bru, M.; Alfonso, I.; Burguete, M. I.; Luis, S. V. Angew. Chem.,
Int. Ed. 2006, 45, 6155−6159.
(12) Ceborska, M.; Tarnowska, A.; Ziach, K.; Jurczak, J. Tetrahedron
2010, 66, 9532−9537.
(13) Storm, O.; Luning, U. Chem.Eur. J. 2002, 8, 793−798.
̈
́
(14) Gonzal
́
ez-Alvarez, A.; Alfonso, I.; Lop
́
ez-Ortiz, F.; Aguirre, A.;
García-Granda, S.; Gotor, V. Eur. J. Org. Chem. 2004, 1117−1127.
(15) Gonzalez-Alvarez, A.; Alfonso, I.; Gotor, V. Chem. Commun.
2006, 2224−2226.
(16) Fischmann, S.; Luning, U. Isr. J. Chem. 2013, 53, 87−96 and
̈
references cited therein.
(17) Katayev, E. A.; Pantos, D. G.; Reshetova, M. D.; Khrustalev, V.
N.; Lynch, V. M.; Ustynyuk, Y. A.; Sessler, J. L. Angew. Chem., Int. Ed.
2005, 44, 7386−7390.
(18) Obrocka, A.; Ziach, K.; Jurczak, J. Polish J. Chem. 2006, 80,
1915−1918.
(19) Misuraca, M. C.; Moulin, E.; Ruff, Y.; Giuseppone, N. New J.
Chem. 2014, 38, 3336−3349.
(20) Huc, I.; Lehn, J.-M. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 2106−
2110.
(21) Rowan, S. J.; Stoddart, J. F. Org. Lett. 1999, 1, 1913−1916.
(22) Wu, J.; Leung, K. C.-F.; Stoddart, J. F. Proc. Natl. Acad. Sci.
U.S.A. 2007, 104, 17266−17271.
(23) Ziach, K.; Jurczak, J. Org. Lett. 2008, 10, 5159−5162.
(24) Ziach, K.; Kulesza, A.; Jurczak, J. Org. Biomol. Chem. 2014, 12,
3827−3830.
(25) Wigfield, D. C.; Gowland, F. W. J. Org. Chem. 1977, 42, 1108−
1109.
(26) Ciaccia, M.; Cacciapaglia, R.; Mencarelli, P.; Mandolini, L.; Di
Stefano, S. Chem. Sci. 2013, 4, 2253−2261.
(27) Ciaccia, M.; Pilati, S.; Cacciapaglia, R.; Mandolini, L.; Di
Stefano, S. Org. Biomol. Chem. 2014, 12, 3282−3287.
(28) Ulrich, S.; Lehn, J.-M. Angew. Chem., Int. Ed. 2008, 47, 2240−
2243.
(29) Drew, M. G. B.; Yates, P. C.; Murphy, B. P.; Nelson, J.; Nelson,
S. M. Inorg. Chim. Acta 1986, 118, 37−47.
(30) Nelson, S. M.; Knox, C. V.; McCann, M.; Drew, M. G. B. J.
Chem. Soc., Dalton Trans. 1981, 1669−1667.
(31) Fenton, D. E.; Cook, D. H. J. Chem. Soc., Chem. Commun. 1977,
623−624.
(32) Cook, D. H.; Fenton, D. E. J. Chem. Soc., Dalton Trans. 1979,
810−813.
(33) Ercolani, G.; Mandolini, L.; Mencarelli, P.; Roelens, S. J. Am.
Chem. Soc. 1993, 115, 3901−3908.
(34) Severin, K. Chem.Eur. J. 2004, 10, 2565−2580.
10341
dx.doi.org/10.1021/jo501966q | J. Org. Chem. 2014, 79, 10334−10341