Organic Letters
Letter
62 and 72, each at 5 mM), thus confirming the thermodynamic
nature of the investigated equilibria.
Following this characterization of the single-level dynamic
libraries, we moved toward a more complex system, where the
simultaneous formation and exchange processes within double-
level dynamic systems were investigated. These experiments
were intended to provide compounds containing both boronic
ester and imine dynamic linkages within a single molecular
component. For this purpose, two separate DCLs were
prepared that, depending on the type of the studied exchange
reaction (imine exchange and/or boronic transesterification),
differed in the composition of the components used. The first
doubly dynamic library was obtained using a stoichiometric
mixture of boronic acid 1, diol 3, and two different amines, 4
1
and 5. The generated DCL was analyzed after 24 h by H
NMR spectroscopy and revealed the presence of two doubly
dynamic compounds 13 (marked blue) and 14 (marked red)
in the ratio 24/76 along with unreacted substrate (Figure S11
in the Supporting Information). A similar library composition
was observed upon the addition of amine 5 to the solution of
preformed molecule 13, which induced the imine exchange
reaction, yet left the ester bond intact.
Figure 1. Schematic representation of the experiments performed to
establish the thermodynamic equilibrium of boronic esters 8 and 9
formation through (a) self-sorting of starting materials and (b)
components exchange between isolated ester 8 and diol 3 (the
unreacted materials constituted 22% of the postreaction mixture).
The same procedure was exploited in the second DCL,
involving a transesterification in the vicinity of the imine bond.
Analogously to the previous experiment, diols 2 and 3 were
1
mixed with boronic acid 1 and p-hydroxyaniline 5. The H
NMR spectrum showed a strong preference toward the
formation of product 14 possessing the more stable six-
membered boronate ester ring (88% of the products library)
compared to its structural analogue 15 (12% of the products
library, Figure S12 in the Supporting Information). Sub-
sequently, a transesterification between 3 and isolated 15 was
found to give a DCL with the same distribution. These
observations highlight again the influence of the structural
features of the molecular building blocks on the ultimate
constitution of the generated DCLs. The dynamic processes in
the second system proceeded orthogonally without the
decomposition of a reversible bond not involved in the
exchange reaction.
The final step of the present work involved the investigation
of a triply dynamic library within the same reaction flask.
Although the formation of disulfide, boronate ester, and
hydrazone exchange has been previously established for
multicomponent surface architectures,5,11 a reaction set
consisting of the first two linkages and an imine has not
been studied. Two approaches were used to evaluate the
possibility of simultaneous formation of the three dynamic
bonds under the conditions applied. First, the “one pot”
approach was performed, where building blocks 1−6 (each at 5
mM) were combined together in an equimolar ratio in DMSO-
d6 at a temperature adjusted to 50 °C. After 24 h, the resulting
mixture was analyzed by 1H NMR spectroscopy (Figure S13 in
the Supporting Information), which confirmed the presence of
a complex dynamic library of structurally distinct compounds,
among which the desired triply dynamic compound 16 was
found in trace amounts. In an attempt to reduce the library
complexity, which would in turn allow for an easier isolation
and a full characterization of 16, a stepwise approach was
performed using only those components that constitute the
desired molecule, that is, 1, 3, and 6 (each at 5 mM). Although
the DCL generated from these building blocks remained
Information), it allowed the isolation of 16, which precipitated
The last linkage studied within a single-level dynamic library
was that of the disulfide bond. Unlike the two previously
discussed reversible bonds, a symmetrical disulfide formation
may result from the autoxidation of an individual reaction
component. Thus, the rate of the autoxidation process was
monitored by 1H NMR spectroscopy, separately for each thiol
building block (6 and 7), which differ in the type of substituent
in the para position. While the signal from the p-
bromothiophenol 7 disappeared within 9 h as a result of 72
formation, component 6 (p-aminothiophenol) required 28 h
for the complete oxidation to homodimer 62, indicating a
major influence of the thiol structural features on the rate of
Information). To assess that the disulfide exchange operates
under thermodynamic control, two distinct pathways for
generating the DCL were explored. In the first, the DCL was
engendered by mixing equimolar amounts of thiol components
6 and 7, while in the second pathway the preoxidized disulfides
62 and 72 were mixed together (Figure S15). In the case of a
thermodynamically controlled system, the composition of the
generated DCL (a mixture consisting of heterodimer (12) and
two homodimers (62 + 72) was expected) should be identical,
regardless of the pathway used. The first DCL was prepared by
dissolving 6 and 7 (each at 5 mM concentration) in DMSO-d6
at a temperature adjusted to 50 °C. The library was stirred for
24 h in air in a capped NMR tube to allow oxidation of the
1
thiol building blocks. An analysis of the H NMR spectrum of
the reaction mixture (Figure S15b, top) revealed a fully
oxidized library with several sets of doublets corresponding to
the aromatic protons of homodimeric products 62 (marked
red) and 72 (marked blue), respectively (each accounted for
30% of the library material). The remaining set of peaks was
assigned to the heterodimeric disulfide compound 12 (marked
green), the predominant species in the mixture (40% of the
library material). An essentially identical product distribution
(Figure S15b, bottom) was observed in the DCL generated by
the alternative pathway (starting from preformed homodimers
3643
Org. Lett. 2021, 23, 3641−3645