Supramolecular detection of geometrical differences of azobenzene carboxylates

Article abstract of DOI:10.1016/j.tetlet.2016.03.043

In dynamic combinatorial chemistry, the geometry of a template can be translated into the composition of a library of interchanging components. In this study, such a dynamic combinatorial library was used for the first time to detect and evaluate differences in the geometry of isomers of photoswitchable azobenzene based templates.

Full text of DOI:10.1016/j.tetlet.2016.03.043

Tetrahedron Letters 57 (2016) 1820–1824
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Supramolecular detection of geometrical differences of azobenzene
carboxylates
y
y
⇑
Filip Ulatowski , Kajetan Da˛browa , Janusz Jurczak
Institute of Organic Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland
a r t i c l e i n f o
a b s t r a c t
Article history:
In dynamic combinatorial chemistry, the geometry of a template can be translated into the composition
of a library of interchanging components. In this study, such a dynamic combinatorial library was used for
the first time to detect and evaluate differences in the geometry of isomers of photoswitchable
azobenzene based templates.
Received 21 February 2016
Revised 9 March 2016
Accepted 14 March 2016
Available online 15 March 2016
Ó 2016 Published by Elsevier Ltd.
Keywords:
Dynamic combinatorial chemistry
Host–guest systems
Molecular recognition
Azo compounds
Macrocycles
The modus operandi of molecular machines consists in utilising
switchable moieties which change their geometry when an appro-
priate stimulus is applied.1 Among the various triggers that can be
used for this purpose, light has unique features that render it partic-
ularly useful, i.e. its unmatched spatial resolution and electrically
neutral character.2 Moreover, light-triggered transformations are
generally reversible and can easily be fine-tuned to selectively affect
only the chosen molecules. The photoactive moieties that are often
applied in such transformations include diaryl- and dithienylethene,
spiropyrane, and azobenzene (AB) derivatives.2b,3 Among these, the
latter appear to be the most useful owing to their synthetic
availability and robustness.4 One challenging aspect of such light-
induced changes in geometry lies in accurately detecting and evalu-
ating them. Simple UV–vis and 1D NMR measurements are typically
employed to track such isomerisation, but they do not measure
changes in the molecule geometry directly, usually providing infor-
mation only about chromophore moiety and its close proximity.5
Such geometrical changes are indeed difficult to spot and quantify,
and this is even more troublesome when one isomer is thermody-
namically unstable, e.g. (Z)-isomer of AB. There are few examples
in the literature that deal with the detection of such differences in
the geometry of photoisomers, the methods used being limited to
NOE6 and diffusion-based7 NMR, conducting atomic force
microscopy8 (c-AFM), FRES, and in rare cases single crystal X-ray
diffraction.9
In this Letter we present a novel concept for sensing photo-
chemical isomerisation phenomena using dynamic combinatorial
chemistry (DCC).10 A dynamic combinatorial library (DCL) consists
of species that interchange via reversible reactions. The composi-
tion of such a DCL is a consequence of the thermodynamic stability
of its components, which can be modified through the addition of a
template. The template may bind to selected components, result-
ing in the amplification of their abundances. In general, the content
of a library component exhibiting the highest affinity towards the
template is increased at the cost of all other DCL members.11 Bind-
ing usually occurs between two compatible groups, enabling weak
non-covalent interactions, the most important of these being
hydrogen bonds (H-bonds). The strongest binding and therefore
amplification are obtained when there is a good geometrical match
between the anchoring points in the template and in the DCL
member. If the template is a switchable molecule the spatial
arrangement of the anchoring groups can be changed upon iso-
merisation, and different amplification profiles will be obtained
for each isomer. Thus, the composition of the DCL can be translated
into the geometrical parameters of the template.
Recently, we described
a DCL relying on disulfide bond
exchange, composed of cyclooligomers (1n) incorporating dipicol-
inic acid diamide subunits equipped with H-bond donors suitable
for anion complexation (Fig. 1).12 Under basic conditions these
macrocycles are in equilibrium, which can be altered through the
introduction of anionic guests. This library proved to be strongly
⇑
Corresponding author. Fax: +48 22 8230944.
E-mail address: jurczak@icho.edu.pl (J. Jurczak).
These authors contributed equally to this work.
y
http://dx.doi.org/10.1016/j.tetlet.2016.03.043
0040-4039/Ó 2016 Published by Elsevier Ltd.

F. Ulatowski et al. / Tetrahedron Letters 57 (2016) 1820–1824
1821
Table 1
Kinetic data for thermal (Z)?(E)a
S
S
O
O
O
O
Compound
%Z@PSS
k
Τꢀ10ꢁ6 (s)
s1/2 (h)
D
NH
NH
HN
HN
2
3
4
85 (68)
89 (70)
90 (82)
8.4
2.7
5.1
23.0
71.5
37.8
N
N
Measured at T = 298 0.1 K; PSS are given for host concentration 5 ꢀ 10ꢁ5 M and
a
5 ꢀ 10ꢁ2 M (in parentheses), respectively; k Τ was predicted using
D
Gà derived from
D
thermodynamic data (see ESIà).
S
S
n-1
1n
meta-substituted AB derivatives should resemble that of the par-
ent AB, due to weak -conjugation of substituents with an azo
p
Figure 1. Dynamic combinatorial library 1n.
group. Surprisingly, however, template 3 most resembles the par-
ent AB. Notably, the rates of isomerisation k T for the correspond-
ing carboxylic acids are 2.2–8 times higher (see ESI), which
supports the recently published hypothesis that increased electron
D
influenced by the shape and size of the various carboxylates acting
as templates. For example, benzene-1,3,5-tricarboxylate predomi-
nantly amplifies the tetrameric 14 whereas structurally related
cyclohexane-1,3,5-tricarboxylate increases mainly the abundance
of trimeric 13.
For the present study, we decided to use as template model
photoswitchable carboxylates 2, 3, and 4 in the form of tetra-n-
butylammonium (TBA) salts equipped with two, three and four
carboxylic groups, respectively (Scheme 1).
The corresponding compounds were synthesised according to
known procedures (see ESI for details). Upon irradiation of com-
pounds 2–4 with light, a photostationary state (PSS) is established
between the near planar (E)-isomer and the more compact V-
shaped (Z)-isomer.
The corresponding (E)?(Z) and (Z)?(E) isomerisation for a
diluted solution (5 ꢀ 10ꢁ5 M) of templates was completed within
1 min by irradiation with UVA light (368 nm, 60W) or blue light
(410 nm, 5W LED), respectively. In UVA light driven PSS, the major-
ity of each template exists in the respective (Z)-form. Relevant data
describing (Z)?(E) isomerisation of the templates studied are
shown in Table 1.13 The rates of the thermal (Z)?(E) isomerisation
indicate that the number of carboxylate groups is just as important
as their substitution pattern in the arene ring. For example, the
symmetrical di-para-substituted 2 back-isomerises faster than
the unsymmetrical di-meta- and para-substituted 3, whereas the
symmetrical tetra-meta-substituted 4 is situated in between.
Reports in the literature indicate that thermal-isomerisation of
density in the
p-system of the AB scaffold accelerates this pro-
cess.14 For our purposes, the rates of thermal isomerisation of all
templates are sufficiently low. The degree of thermal (Z)?(E) iso-
merisation was negligible (s1/2 > 20 h @ 25 °C) during the templa-
tion experiment, since the equilibration of the library takes only
3 h. Templation experiments with a library of 1n were conducted
using tetrabutylammonium (TBA) salts of 2–4 in (E)-form and their
PSS mixtures of (E)- and (Z)-isomers (further named in the text as
(Z)-form). TBA salts of benzoic (5) and isophthalic acids (6), as well
as their equimolar mixture, were used as references, as they
resemble in terms of their structure the appropriate halves of tem-
plates 2–4. Templation experiments were run in DMSO (+0.5%
H2O), with a total concentration of 1 being 0.01 M. Compositions
of the library were determined by HPLC analyses. For each tem-
plate and each library component we calculated a normalised
amplification factor (AFn), a useful parameter introduced recently
by Otto and co-workers,15 which we modified in the range of neg-
ative values to limit and normalise its range to ꢁ1.
(
)
AnꢁA0
for An > A0
for An < A0
AmaxꢁA0
AFn ¼
A0ꢁAn
A0
where A0 is the concentration of library component, A in the
absence of template, An – in the presence of template, and Amax
N
N
-
N
CO2
-O2C
N
-
-O2C
•2TBA+
CO2
•2TBA+
(E)-2
(Z)-2
- •TBA+
CO2
N
N
-
CO2
5
368nm
N
N
-O2C
-
CO2
-
CO2
410nm (LED)
-
-O2C
CO2
•3TBA+
•3TBA+
-
ΔT
CO2
(E)-3
(Z)-3
•2TBA+
-
N
N
CO2
-
CO2
-O2C
6
N
N
-
-O2C
CO2
-
CO2
-
-
CO2
-O2C
CO2
•4TBA+
•4TBA+
(E)-4
(Z)-4
Scheme 1. Structures of templates 2–6 and isomerisation of 2–4.

1822
F. Ulatowski et al. / Tetrahedron Letters 57 (2016) 1820–1824
(a)
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
15
14
13
12
none
E-3
E-4
E-2
5+6
Z-3
Z-4
5
6
Z-2
Template
(b)
0.4
0.3
0.2
15
14
13
12
0.1
0.0
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
5
E-2
Z-2
5+6
E-3
Z-3
6
E-4
Z-4
Template
Figure 2. Distribution of components in templated DCLs. (a) %mass of oligo-macrocycles, (b) normalised amplification factors induced by templates.
is the maximum possible concentration of component A. AFn ranges
from ꢁ1 to 1. Positive values indicate amplification of the given
component, while negative values mean the reduction of its abun-
dance. The extreme values of ꢁ1 and 1 indicate a total absence of
the component and the maximum attainable concentration, respec-
tively. The library compositions in the templation experiments and
amplification profiles are presented in Figure 2.
In the case of template 3, as a reference we employed an equimolar
mixture of benzoate and isophthalate (5 + 6). This mixture induced
changes very similar to those seen with 6 alone, with the trimer as
the only amplified guest. Since photoswitchable host 3 in its
(E)-form amplifies mainly the trimer, we attribute this to the inter-
action of the guest with the isophthalic part of the host. However,
compared with the reference mixture 5 + 6, (E)-3 also amplifies to
some extent 14 and 15, which definitely requires cooperation of all
three anchoring points. DCL templated with photoisomerised guest
(Z)-3 possesses quite a different distribution of oligomers. Tetramer,
havinga significantly higher contribution(25 vs 10 mass%), seems to
be the optimal host for the contracted (Z)-3. Concurrently, the
amount of dimer is reduced from 47 to 28 mass%. In addition, the
amount of the trimer remains virtually the same regardless of the
geometry of template 3 applied. The amplification factors for 13
and 14 with (Z)-3 are nearly the same. This means that the tetramer
possesses higher affinity towards the template than the trimer, since
the higher oligomers are much more difficult to amplify due to
entropic reasons.11c,12a
The last template examined, namely tetra-anionic 4 in (E)-form,
in contrast to the reference isophthalate (6) is capable of effective
amplification of the trimer, tetramer, and pentamer. Amplification
of the trimer, which engages only one half of the template, is less
pronounced than for model 6. The significant amplification of 14
and 15 is a result of the cooperation of the four carboxylate groups.
Quite surprisingly, the (Z)-4 mixture exhibits all amplification
factors more weakly than the (E)-isomer. We expected the anion
A monofunctional template (benzoate) amplifies the smallest
library component (the dimer) at the expense of all other macrocy-
cles. The changes induced by (E)-2 are quite similar to that of
benzoate, the dimer being the most amplified macrocycle, though
the AFn is far lower. Yet unlike benzoate, (E)-2 also exhibits a slight
amplification of the tetramer, which seems to be the smallest host to
interact with both distant anchoring points of 2. The binding is quite
weak and, according to simple modelling, the supermolecule’s
formation requires the host to adopt an unfavourable expanded con-
formation. Amplification of larger hosts, which may accommodate
(E)-2 more easily, is not observed since binding of just two anionic
groups does not provide enough enthalpy to overcome the negative
entropy of the formation of higher oligomers. The distribution of
macrocycles induced by photoisomerised (Z)-2 is characterised by
a negligible amplification of the dimer, but stronger amplification
of the tetramer. This means that (Z)-2 with a shorter distance
between carboxylate groups is a better suited template for 14 and
binding does not require the aforementioned receptor stretching.
The weak amplification of 12 indicates that in the (Z)-form of tem-
plate 2 the two binding points are less likely to act independently.

F. Ulatowski et al. / Tetrahedron Letters 57 (2016) 1820–1824
1823
at the level of >104 Mꢁ1, which indicates a strong binding of the
templates by the macrocyclic hosts.
In the next step, we attempted to employ light as a factor influ-
encing the composition of the DCL with the templates as media-
tors. So far, there is only one paper by Waters and co-workers18
describing light induced changes in a DCL in which the library sub-
strate is based on an azobenzene core. The equilibrated mixtures of
1n and one of the templates 2–4 in the (E)-form were irradiated
with 368 nm light and PSSs were reached, as indicated by UV–vis
analyses. We expected the library to modify the distribution of
macrocycles according to aforementioned templation experiments.
However, for all anions we observed no change in composition,
even 12 h after irradiation – the library was ‘‘frozen”. Analogous
freezing was observed for (Z)?(E) isomerisation of compounds 2
and 3 with blue light. This effect was never observed when 1n
and templates were irradiated individually before mixing them.
We eventually found that it is possible to influence the library
composition by light stimulus when 5 mol% of unoxidised mono-
mer 1-H2 (dithiol) was added to the mixture of DCL and photo-
switchable template 3 (see ESI). The complex behaviour of this
system requires further investigation.
In summary, for our three model anionic templates having
switchable functionality, the differences in the geometry of the iso-
mers can be detected and semi-quantitatively evaluated by analys-
ing their templation effects on the dynamic combinatorial library.
Information about the template’s geometry is, in such an experi-
ment, translated into the composition of the DCL. In addition, we
proved that it is possible to modify the composition of a DCL by
a light stimulus via the photoswitchable template acting as a
mediator.
Acknowledgment
This work was supported by the Poland’s National Science Cen-
tre (project 2011/02/A/ST5/00439).
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.tetlet.2016.03.
043.
Figure 3. Changes in the chemical shift (
(yellow) upon addition of template (E)-3 (top) and (Z)-3 (bottom) in DMSO-d6.
Corresponding correlations between d for both macrocycles (inset).
Dd) of amide protons of 13 (red) and 14
D
References and notes
in the (Z)-form to be particularly effective in the amplification of
the tetramer. A possible explanation for this observation may
involve the formation of a strong ion pair between tetraanionic
(Z)-4 and the TBA cation, which hampers any interactions with
the macrocyclic hosts. This assumption is, however, not supported
by NMR experiments.
1. Balzani, V.; Credi, A.; Raymo, F. M.; Stoddart, J. F. Angew. Chem. Int. Ed. 2000, 39,
3348–3391.
2. (a) Balzani, V.; Credi, A.; Venturi, M. Chem. Soc. Rev. 2009, 38, 1542–1550; (b)
Dürr, H.; Bouas-Laurent, H. Photochromism: Molecules and Systems: Molecules
and Systems; Elsevier Science, 2003.
´
3. (a) Szymanski, W.; Beierle, J. M.; Kistemaker, H. A. V.; Velema, W. A.; Feringa, B.
L. Chem. Rev. 2013, 113, 6114–6178; (b) Brieke, C.; Rohrbach, F.; Gottschalk, A.;
Mayer, G.; Heckel, A. Angew. Chem. Int. Ed. 2012, 51, 8446–8476.
4. (a) Bandara, H. M. D.; Burdette, S. C. Chem. Soc. Rev. 2012, 41, 1809–1825; (b)
Beharry, A. A.; Woolley, G. A. Chem. Soc. Rev. 2011, 40, 4422–4437; (c) Merino,
E. Chem. Soc. Rev. 2011, 40, 3835–3853.
In order to provide further confirmation of the differences in
affinities of the photoisomers towards macrocyclic hosts we per-
formed an NMR competitive titration.16 Such an experiment is par-
ticularly useful in determining the ratio of two association
constants – the selectivity. We carried the measurement on a mix-
ture of 13 and 14 with (E)- and (Z)-3 as guests in two separate
experiments (see ESI for details). Under neutral conditions there
was no disulfide bond exchange, the composition remained con-
stant during the course of titration. The changes of chemical shifts
5. For example, blue shift of
p–
p⁄ band of the azo chromophore as well as large
shielding of the protons adjacent to the N@N linkage are observed upon (E)?
(Z) isomerisation of the ABs in UV–vis and 1H NMR spectra, respectively.
6. (a) Ulysse, L.; Cubillos, J.; Chmielewski, J. J. Am. Chem. Soc. 1995, 117, 8466–
8467; (b) Kusukawa, T.; Fujita, M. J. Am. Chem. Soc. 1999, 121, 1397–1398; (c)
Chen, X.; Hong, L.; You, X.; Wang, Y.; Zou, G.; Su, W.; Zhang, Q. Chem. Commun.
2009, 1356–1358.
7. Nguyen, T.-T.-T.; Türp, D.; Wagner, M.; Müllen, K. Angew. Chem. Int. Ed. 2013,
52, 669–673.
8. Mativetsky, J. M.; Pace, G.; Elbing, M.; Rampi, M. A.; Mayor, M.; Samorì, P. J. Am.
Chem. Soc. 2008, 130, 9192–9193.
9. Note that majority of them is stabilised in the macrocyclic scaffold or by direct
interaction of the N@N bond with the metal cation.
10. (a) Reek, J. N.; Otto, S. Dynamic Combinatorial Chemistry; John Wiley & Sons,
2010; (b) Miller, B. L. Dynamic Combinatorial Chemistry: In Drug Discovery,
Bioorganic Chemistry, and Materials Science; John Wiley & Sons, 2009; (c) Lehn,
J.-M. Chem. Soc. Rev. 2007, 36, 151–160; (d) Otto, S.; Furlan, R. L. E.; Sanders, J.
(
D
d) of amide protons are presented in Figure 3. For (E)-3 the cor-
relation between the d of the two macrocycles is nearly linear,
D
which indicates that both hosts possess similar affinity towards
the anions guest. In the case of photoisomerised (Z)-3 the correla-
tion line is bent which indicates that host 14 binds the introduced
guest more efficiently than host 13.17 These results are in perfect
agreement with the observed amplification in DCL. Additionally,
titration of pure 13 with (E)-3 provided an association constant

1824
F. Ulatowski et al. / Tetrahedron Letters 57 (2016) 1820–1824
K. M. Drug Discovery Today 2002, 7, 117–125; (e) Rowan, S. J.; Cantrill, S. J.;
13. For additional data for corresponding carboxylic acids as well as
thermodynamic data see ESI.
14. Dabrowa, K.; Niedbala, P.; Jurczak, J. Chem. Commun. 2014, 15748–15751.
15. Hamieh, S.; Saggiomo, V.; Nowak, P.; Mattia, E.; Ludlow, R. F.; Otto, S. Angew.
Chem. Int. Ed. 2013, 52, 12368–12372.
16. Whitlock, B.; Whitlock, H. J. Am. Chem. Soc. 1990, 112, 3910–3915.
17. (Z)-3 stands for a PSS mixture of photoisomers, therefore during titration a
mixture of four components is obtained (two hosts and two guests) and so the
exact value of ratio of association constants cannot be calculated.
18. Ingerman, L. A.; Waters, M. L. J. Org. Chem. 2008, 74, 111–117.
Cousins, G. R.; Sanders, J. K.; Stoddart, J. F. Angew. Chem. Int. Ed. 2002, 41, 898–
952; (f) Lehn, J.-M.; Eliseev, A. V. Science 2001, 291, 2331–2332; (g) Lehn, J. M.
Chem. Eur. J. 1999, 5, 2455–2463.
11. (a) Corbett, P. T.; Sanders, J. K.; Otto, S. Chem. Eur. J. 2008, 14, 2153–2166; (b)
Ludlow, R. F.; Otto, S. J. Am. Chem. Soc. 2008, 130, 12218–12219; (c) Corbett, P.
T.; Sanders, J. K.; Otto, S. J. Am. Chem. Soc. 2005, 127, 9390–9392; (d) Corbett, P.
T.; Otto, S.; Sanders, J. K. M. Org. Lett. 2004, 6, 1825–1827; (e) Severin, K. Chem.
Eur. J. 2004, 10, 2565–2580; (f) Grote, Z.; Scopelliti, R.; Severin, K. Angew. Chem.
Int. Ed. 2003, 42, 3821–3825.
12. a) Ulatowski, F.; Sadowska-Kuziola, A.; Jurczak, J. J. Org. Chem. 2014, 79, 9762–
9770; b) Ulatowski, F.; Lichosyt, D.; Jurczak, J. Org. Biomol. Chem. 2015, 13,
10451–10455.