A pH-, light- and redox-responsive flavylium-bipyridinium molecular dyad (bioinspired in natural anthocyanins) was synthesized and employed to devise a pseudorotaxane with the macrocycle cucurbituril (CB7) in aqueous solution. The inclusion complex was characterized by UV-Vis absorption, fluorescence emission, NMR and electrochemical techniques which demonstrate formation of a stable binary complex between the dyad and CB7 both under acidic and neutral conditions. It is noteworthy that the flavylium-bipyridinium tricationic dyad is only stable in highly acidic media, undergoing a reversible hydration reaction at slightly acidic or neutral pH to give a trans-chalcone-bipyridinium dication. 1H NMR experiments showed that in this last species the CB7 binds to the bipyridinium unit while in the tricationic species the macrocycle is positioned between the flavylium and the bipyridinium moieties. The different location of the CB7 wheel in the two dyad states allows control of the shuttling movement using light and pH stimuli that trigger the interconversion between these two species.
Synthetic flavylium compounds comprise an important family of dyes bioinspired in their natural polyphenolic counterparts that include important pigments such as anthocyanins and anthocyanidins responsible for the most of red and blue colors found in flowers and fruits , . Both natural and synthetic dyes from this family are usually represented by the flavylium cation. Nevertheless, this species is only stable at very acidic conditions (pH<1). At higher pH values, deprotonation and hydration of the positively charged electrophilic species trigger a dynamic network of reversible reactions that in slightly acidic/neutral conditions comprise at least five species (Scheme 1) . The quinoidal base (A) is immediately formed from the deprotonation of the flavylium cation (AH+) upon a direct pH jump from acidic to higher pH values. However, in a competitive slower pathway, AH+ undergoes hydration to form a hemiketal (B) followed by a ring-opening tautomerization process to give the cis-chalcone (Cc). Finally, the last process consists in the isomerization of the cis-chalcone to the trans-chalcone (Ct). At slightly acidic/neutral conditions, all basic species (A, B, Cc and Ct) coexist in dynamic equilibrium with relative stabilities (mole fraction distribution) that depend on the nature and position of substituents decorating the flavylium skeleton.
Despite of its apparent complexity, flavylium multistate systems can be understood as a simple acid-base equilibrium between the flavylium cation and a conjugate base (CB) that comprises all the remaining species. Its concentration is given by the sum of the concentrations of all basic species: [CB]=[A]+[B]+[Cc]+[Ct] and an apparent acidity constant can be defined to account for the stability of the flavylium cation against the pH.
An important feature of flavylium compounds is associated with their photochromic properties. These properties arise from the photoreactivity of the uncolored/pale yellow Ct species that upon irradiation undergoes a photoinduced trans-cis photoisomerization, populating the Cc intermediate species that in turn, under appropriate pH conditions, drives the equilibrium towards the deep yellow/red AH+ , , , , . Systems for which Ct is by far the most stable amongst the basic species are particularly interesting in what photochromic properties is concerned. This arises from the fact that in these cases the photoactive Ct predominates over a wide range of acidic pH values enhancing photochromic properties of the system that usually are favored under acidic conditions where AH+ is formed. In such optimized situations, Ct can be almost quantitatively photoconverted into the AH+ which thermally reverts in the dark back to the most stable Ct species, defining a photochromic system (Scheme 2).
The photoinduced interconversion between the neutral Ct and the positively charged AH+ are particularly attractive to develop photoresponsive supramolecular systems based on host-guest complexes. The distinct structural and electronic properties of the two species are expected to result in significantly different or even contrasting affinities toward host receptors. In fact, previous works carried out in our lab showed that Ct displays much higher affinity towards cyclodextrins than AH+ while the reverse is observed for cucurbiturils, allowing the development of photoresponsive host-guest complexes that can be dissociated or formed upon light exposure, respectively , , , , , , , , , , . The selectivity of cucurbiturils, and in particular cucurbituril (CB7), towards AH+ was also exploited to devise multistate pH-responsive and pH-gated photoresponsive pseudorotaxanes , .
Cucurbituril receptors stand out as a highly appealing class of water soluble macrocyclic molecules displaying extraordinary affinity and selectivity for complementary guest molecules , , , . Their outstanding recognition properties have been widely investigated for stimuli-responsive supramolecular systems including rotaxanes and pseudorotaxanes , , . However, owing to the often observed selectivity of these receptors for positively charged recognition motifs , , , pH and redox stimuli are more commonly employed to control the complexation and co-conformational changes in cucurbituril-based rotaxanes and pseudorotaxanes , , , , , . Despite of the recognized advantages of using light as a stimulus, examples of photoresponsive co-conformational switching in CB-based rotaxanes and pseudorotaxanes remain scarce , , .
In this manuscript, we report a pseudorotaxane assembled from a new flavylium-bipyridinium molecular axle (1) and CB7 macrocyclic wheel (Scheme 3). This system will be used to highlight some advantages and limitations of the application of flavylium switches for the construction of supramolecular systems, in particular pseudorotaxanes, and discuss some key aspects that must be taken into account during the design and synthesis of new molecules for these purposes.
All solvents and chemicals employed for synthesis and for sample preparation were of reagent grade and used as received. Ultrapure Millipore grade water was used. Cucurbituril was available from previous studies .
1-Methyl-1′-[(acetophenone-4-il)methyl]-4,4′-bipyridinium hexafluoro-phosphate (0.25 mmol; 150 mg) and 2-hydroxy-4-methoxybenzaldehyde (0.25 mmol; 38.4 mg) were dissolved in 5 mL de glacial acetic acid. 1.5 mL of sulphuric acid was added and the resulting mixture was allowed to stir overnight. A red-brownish solid was obtained by treating the solution with H2O and perchloric acid. The solid was filtered off, washed with ethyl acetate and dried, yielding 1-methyl-1′-[(7-methoxyflavylium-4′-il)methyl]-4,4′-bipyridinium perchlorate (1) quantitatively. 1H-NMR (DCl/D2O, pD≈1.0, 400.13 MHz) δ 9.16 (d, J=8.4 Hz, 1H), 9.08 (d, J=6.4 Hz, 2H), 8.88 (d, J=6.3 Hz, 3H), 8.45 (d, J=6.4 Hz, 2H), 8.41 – 8.30 (m, 5H), 8.08 (d, J=9.2 Hz, 1H), 7.73 – 7.60 (m, 3H), 7.43 (dd, J=9.7 Hz, 1H), 5.96 (s, 2H), 4.33 (s, 3H), 3.99 (s, 3H).
The pH of the solutions was adjusted with citrate buffer, HCl and NaOH and measured with a Crison basic 20+ pH meter. UV-Vis absorption spectra were recorded with a Varian Cary 100 Bio or a Varian Cary 5000 spectrophotometer and fluorescence spectra were acquired on a SPEX Fluorolog-3 Model FL3-22 spectrofluorimeter.
The titration experiments were carried out in 1 cm path length quartz cells. Two stock solutions are prepared for these titrations: one with only compound 1 in water at desired pH and a second one with exactly the same concentration of 1, the same pH and known concentration of CB7 in excess. Then, a known volume of the first solution is placed in a quartz cuvette and small aliquots of the second solution are added to the cuvette and the spectra registered. Using this method, the concentration of dye is kept constant during the titration while the concentration of host is increased until the no more spectral variations are observed. For the fluorescence experiments, solutions with absorbance values below 0.1 at the excitation and higher wavelengths were always used. The excitation wavelength was set at λex=455 nm, and the slits were set at 1.75 nm for emission and 1.5 nm for excitation.
Cyclic voltammetry (CV) measurements were performed on an Autolab PGSTAT 12 potentiostat/galvanostat, controlled with GPES software version 4.9 (Eco-Chemie), using a 5 mL cylindrical three-electrode cell. A glassy carbon electrode (MF-2013, f=1.6 mm, BAS Inc.) was used as the working electrode and a Pt wire was used as an auxiliary electrode. All potentials refer to a SCE (3 M KCl) reference electrode (Metrohm). Prior to use, the working electrode was polished in aqueous suspensions of 1.0 and 0.3 mm alumina (Beuhler) over 2–7/8″ micro-cloth (Beuhler) polishing pads, then rinsed with water and methanol. This cleaning procedure was always applied before any electrochemical measurements. Electrochemical cleaning was also employed, submitting the electrode to −1/+1 V currents for 60 s. Cyclic voltammetry (CV) at a scan rate of 20, 50 and 100 mV s−1 was used to characterize the electrochemical responses between 0 and −1.250 V vs. SCE.
Indicator displacement assays
Binding constants from indicator displacement assays were determined according to the following set of coupled equations:
[I] and [IC] are the equilibrium concentrations of free and complexed indicator dye, respectively, while [I]0=[I]+[IC] is the total concentration of indicator. εI and εC are the molar extinction coefficients of free and complexed dye at the monitorization wavelength, respectively.
Equation 1 assumes that only the free and complexed dye absorb at the monitorization wavelength (i.e. the guest competitor does not absorb either free or complexed).
According to the mass balance and equilibrium expressions, the following equations can be written:
and combined to give eqs. 7–9:
The experimental data can be fitted to eq. 10 (using solver tool in an Excel spreadsheet for example) coupled with the cubic eq. 9 that can be solved using the Newton-Raphson algorithm. In indicator displacement assays, one of the binding constants is usually known and used as reference and kept constant (KI or KG) while the other can be optimized through data fitting.
NMR experiments were run on a Bruker AMX 400 instrument operating at 400 MHz (1H) and 101 MHz (13C). The solutions for NMR were prepared in D2O and the pD adjusted with DCl or NaOD solutions. Corrections due to isotope effects were applied using the equation pD=pH*+0.4, where pH* is the reading taken from the pH meter .
Results and discussion
When compared with 4′-hydroxyflavylium (see Scheme 1) it can be readily verified that flavylium-based molecular axle 1 lacks hydroxyl groups and therefore its characteristic network of chemical reactions in acidic medium is attributed to four species instead of five (Scheme 4).
The flavylium cation (1AH+) is the most stable species under very acidic conditions (pH<1). However, upon a pH jump to higher values the disappearance of this species (band centered at 425 nm) and the concomitant formation of the other species can be monitored by UV-Vis absorption spectroscopy (Fig. 1a). Representation of the spectra registered for equilibrated solutions of 1 at different pH values allows determination of a closer examination of the spectra shown in Fig. 1 suggests that 1AH+ is almost quantitatively converted into the trans-chalcone species (1Ct, band centered at 376 nm) and the other species, 1B and 1Cc, are minor components in the final equilibrium mixture. This observation is compatible to what has been observed with other synthetic flavylium compounds with similar substitution pattern and was confirmed by 1H NMR experiments vide infra.
The existence of only two main species in equilibrium greatly simplifies the investigation of CB7 binding towards 1. The interaction between the macrocyclic host and the 1AH+ species can be conveniently investigated under acidic conditions while the complexation of 1Ct can be studied at pH values slightly below 7. Figure 2a shows the UV-Vis spectral variations observed for 1AH+ upon gradual addition of CB7 at pH=1. Plots of the absorbance at 425 nm against the CB7 concentration shows a variation profile that can be accounted for by a 1:1 host:guest binding stoichiometry. However the existence of two binding sites for CB7, the bipyridinium and the flavylium units, is expected to allow the sequential formation of 1:1 and 2:1 host:guest complexes. In fact, 1H NMR experiments (see Supporting Information, Fig. S1) show evidence for the formation of the ternary complex although with a low K2:1 binding constant. Fitting the UV-Vis titration data allows estimation of the 1:1 binding constant: K1:1>5×106 M−1. Owing to the high stability of the 1:1 complex only the lower limit of the respective binding constant can be estimated under the experimental conditions while the 2:1 complex is not observed in this concentration range. The low stability of this species can be attributed to electrostatic repulsion generated by the carbonyl portals of two CB7 molecules in close proximity.
The host-guest titration was also monitored by fluorescence spectroscopy. As can be observed from Fig. 2b, the fluorescence intensity of 1AH+ shows a significant and gradual increase upon stepwise addition of CB7. The results can be satisfactorily fitted to the previous model showing that even in more diluted conditions the complex is quantitatively formed precluding the accurate determination of the binding constant.
The interaction between 1Ct and CB7 was investigated using an equilibrated solution of compound 1 at pH=6.3, Fig. 3a. As can be observed, similarly to 1AH+, the spectral variations upon addition of CB7 are compatible with the formation of 1:1 host:guest complexes within the investigated concentration range with a high binding constant of K1:1>3×106 M−1 (again NMR experiments shows the formation of 2:1 complexes in the mM concentration range, see Supporting Information, Fig. S2). It is also worth noting that the band centered at 265 nm (assigned to the bipyridinium unit) shows larger spectral variations while the chalcone absorption is hardly affected suggesting that the bipyridinium binding site is preferentially recognized by the CB7 host.
In order to obtain the value of K1:1 for the binding of 1Ct towards CB7, competitive binding experiments using the indicator displacement assay (IDA) were employed . First, we employed dye 2 as an indicator to obtain the binding constant of tetramethylammonium bromide (3) competitor (see Scheme 3). A binding constant of 2.3×105 M−1 for dye 2 towards CB7 was previously determined in our group . This dye is particularly useful for IDA monitored by UV-Vis absorption spectroscopy owing to its large bathochromic shift of ca. 50 nm upon complexation with CB7. Figure 4 shows the competitive binding (IDA) experiment of 3 towards 2:CB7 which afforded a binding constant of 1.5×106 M−1 for 3 (Fig. 4a). With this value in hand, we used 3 as a competitor for the 1Ct:CB7 complex and obtained a binding constant of 5.3×107 M−1 for 1Ct towards CB7 (Fig. 4b).
Following the determination of the binding constant for 1Ct and the lower limit for 1AH+ towards CB7, the pH-dependent 1AH+/1Ct interconversion of the 1:1 pseudorotaxane complex was investigated. As can be anticipated, the high stability of the binary complexes ensured their predominance under stoichiometric conditions even at μM concentrations. Figure 3b shows the pH-dependent spectra obtained for 1 in the presence of 1 equiv. of CB7. Fitting the pH-dependent absorbance data (at 425 nm) afforded an apparent value of 4.1 which is two units higher than what is observed for free 1. The higher stability of the acidic species in the presence of CB7 implies that the binding constant for 1AH+ must be two orders of magnitude higher than the one for 1Ct and allows the estimation of a nanomolar value of K1:1≈5×109 M−1 for the binding constant of 1AH+ towards CB7 using the well stablished relationship: (free species) (complex) , .
The structure of the pseudorotaxane formed between 1AH+ and CB7 can be investigated by NMR. Based on the well-established complexation induced chemical shifts of proton signals of guests inside the cucurbituril cavities, which led to upfield shifts when included within the hydrophobic cavity and downfield shifts when located near outside the carbonyl portals, it is possible to obtain detailed structural information from simple 1H NMR experiments . As can be observed from Fig. 5, addition of 0.5 equiv. of CB7 to 1AH+ in D2O (pD=1) leads to the appearance of new set of signals concomitantly with those observed in the absence of CB7. This behavior suggests that exchange between free and complexed 1AH+ is slow on the NMR chemical shift time scale and is supported by the fact that for 1 equiv. of CB7 the signals of the free guest disappear from the spectrum. Further support was obtained from ROESY experiment with 0.5 equiv. of CB7 (see Supporting Information, Fig. S3) which shows several signals with the same phase of the diagonal due to the exchange and NOE signals with opposite phase. This experiment in combination with COSY allowed straightforward assignment of all the protons of 1AH+ and of 1AH+:CB7 pseudorotaxane.
The high upfield complexation induced chemical shifts observed for protons i, h and g clearly shows that CB7 predominantly includes the central part of the molecule. However, smaller upfield shifts observed for protons j and f and downfield shifts observed for protons k, l, a and b suggests that the CB7 wheel is not restricted and undergoes a small amplitude shuttling between the viologen and the flavylium moieties along the axle (Scheme 5). It should also be noted that based on the sharp 1H NMR signals of the complex it can be expected that this shuttling movement is faster than the exchange between free and complexed species.
The 1H NMR spectrum of an equilibrated solution of 1 at pD=6 (Fig. 6) shows the presence of a single predominant species. The pair of doublets appearing at 7.50 and 7.94 with coupling constant of ca. J=16 Hz confirms that this species is the trans-chalcone 1Ct. In the presence of 1 equiv. of CB7 it can be clearly observed that the NMR signals of the bipyridinium protons l and k are subjected to large (ca. 1 ppm) upfield displacement while the remaining protons show lower complexation induced chemical shifts. This spectral signature can be assigned to the inclusion of the bipyridinium unit in the cavity of the macrocycle in the 1Ct:CB7 pseudorotaxane (Scheme 6).
The different positions of the CB7 wheel along the molecular axis for 1AH+ and 1Ct offers a necessary condition to demonstrate stimuli-responsive shuttling for the 1:CB7 pseudorotaxane. As already demonstrated above, 1AH+:CB7/1Ct:CB7 pseudorotaxanes can be interconverted by pH stimuli which, therefore, promotes shuttling of the CB7 host from the bipyridinium unit in the 1Ct:CB7 form to the central region of the axle in the 1AH+:CB7 species. The photochemical-induced transformation of 1Ct:CB7 into 1AH+:CB7 was also investigated to evaluate the possibility of controlling the shuttling movement with light stimulus. Figure 7 shows the spectral variations observed upon 365 nm light irradiation of 1Ct:CB7 at pH=4.48. The absorbance decrease at 370 nm and the concomitant increase at 428 nm is compatible with the photoinduced formation of 1AH+:CB7 form 1Ct:CB7. At this pH, and considering a a fraction of 1Ct:CB7 of 70% can be estimated before irradiation, while at the photostationary state this species decreases to 58% which corresponds to a modest 12% conversion at this pH. The efficiency of the Ct to AH+ conversion of the photochromic system depends on different thermodynamic and kinetic parameters that define the apparent of the photostationary state (PSS) . For an efficient and reversible Ct to AH+ photoinduced conversion the must be considerably higher than the (for instance 80% conversion requires a difference of 2 pKa units). Considering that the 1Ct:CB7 is 58% at the PSS for pH=4.48, the can be estimated to be 4.34 Thus, the small difference between the two pKa values (0.24 units) accounts for the modest photoinduced 1Ct:CB7 into 1AH+:CB7 conversion. The observed rate constant for the thermal conversion of 1AH+:CB7 into 1Ct:CB7, under similar conditions, was determined to be kobs=1.5×10−3 s−1 (t1/2=468 s) upon a pH jump from pH=1 to pH=4.6 in the presence of 1 equiv. of CB7 (see Supporting Information, Fig. S4).
The presence of a redox responsive viologen unit in the structure of 1AH+ prompted us to characterize its electrochemical properties by cyclic voltammetry . As can be observed from Fig. 8, 1AH+ shows two reduction half-wave potentials at ca. −0.5 V and 1.0 V (vs. SCE) which can be assigned to the two consecutive mono-electronic reductions of the viologen unit. A closer look at the first reduction wave in Fig. 8a also reveals a shoulder centered at −0.4 V that is tentatively assigned to the irreversible reduction of the flavylium unit , .
In contrast with the parent methylviologen , the reduction waves in 1AH+ show up as chemically irreversible electron transfer. However, as demonstrated in the inset of Fig. 8a, when the cyclic voltammetry was restricted to the first reduction wave a “quasi-reversible” behavior (together with some adsorption) was observed suggesting that 1AH+ may undergo degradation coupled with the second reduction process.
In the presence of increasing concentrations of CB7, the reduction waves are shifted to more negative potentials in agreement with higher affinity of this host for the more oxidized forms . The second reduction wave could not be detected in the electrochemical window used probably due to the larger shifts induced by complexation with CB7 . In contrast to 1AH+, 1Ct (Fig. 8b) shows the quasi-reversible two-electron reduction behavior, as expected for a viologen derivative, and the typical shifts of the first reduction wave to more negative potentials upon addition of increasing concentrations of CB7. However, unexpectedly, the position of the second reduction wave hardly changed with the concentration of CB7.
A new flavylium-viologen dyad was synthesized and its assembly with CB7 to form a pseudorotaxane was characterized in detail using different spectroscopic and electrochemical techniques. The pseudo-rotaxane was shown to display high stability constants both at neutral and acidic conditions. The interconversion from the chalcone to flavylium species, using pH or light stimuli, is accompanied from a displacement of the CB7 wheel from the peripheric bipyridinium station to the center of the axle allowing control over the shuttling process with these two stimuli. However, the photo-induced conversion of the chalcone into the flavylium was found to be relatively low for this compound. Systematic studies are ongoing in our laboratories to rationalize the reasons for the modest photochemically-induced trans-chalcone–flavylium interconversion observed for compound 1, which contrast with previously reported analogs, and new derivatives are being designed and prepared to improve photo-switching properties of this water-soluble pseudorotaxanes.
A collection of invited papers based on presentations at the 4th International Conference on Bioinspired and Biobased Chemistry & Materials (NICE-2018), Nice, France, 14–17 October 2018.
Funding source: H2020-MSCA-RISE-2016
Award Identifier / Grant number: 734834
Funding statement: This work was supported by LAQV-REQUIMTE, which is financed by the Portuguese FCT/MCTES (UID/QUI/50006/2019) and co-financed by the ERDF under the PT2020 (POCI-01-0145-FEDER-007265). FCT/MCTES is also acknowledged for supporting the National Portuguese NMR Network (RECI/BBB-BQB/0230/2012), projects PTDC/QUI-COL/32351/2017, PTDC/QUI-QFI/30951/2017 and CEECIND/00466/2017 (N.B.) (Funder Id: http://dx.doi.org/10.13039/501100001871). EC is acknowledged for the INFUSION project grant no. 734834 under H2020-MSCA-RISE-2016 (Funder Id: http://dx.doi.org/10.13039/100010665). H.C. acknowledges postdoctoral grant SFRH/BPD/102705/2014.
 F. Pina, M. J. Melo, A. J. Parola, M. Maestri, V. Balzani. Chem. Eur. J. 4, 2001 (1998).10.1002/(SICI)1521-3765(19981002)4:10<2001::AID-CHEM2001>3.0.CO;2-PSearch in Google Scholar
 F. Pina, A. Roque, M. J. Melo, M. Maestri, L. Belladelli, V. Balzani. Chem. Eur. J. 4, 1184 (1998).10.1002/(SICI)1521-3765(19980710)4:7<1184::AID-CHEM1184>3.0.CO;2-6Search in Google Scholar
The online version of this article offers supplementary material (https://doi.org/10.1515/pac-2019-0225).
©2020 IUPAC & De Gruyter. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. For more information, please visit: http://creativecommons.org/licenses/by-nc-nd/4.0/