Merging of a Supramolecular Ligand with a Switchable Luminophore – Light-Responsiveness, Photophysics and Bioimaging

Graphical Abstract

A series of novel luminophores with aggregation-induced emission is reported. The design concept is based on merging of cyanostilbenes with guanidiniocarbonyl-pyrroles yielding hybrid luminophores, capable to undergo photoisomerization and are able to be used in biomedical imaging.

Abstract

In this contribution we report on a novel approach towards luminescent light-responsive ligands. To this end, cyanostilbene- guanidiniocarbonyl-pyrrole hybrids were designed and investigated. Merging of a luminophore with a supramolecular bioactive ligand bears numerous advantages by overcoming the typical drawbacks of drug-labelling, influencing the overall performance of the active species by attachment of a large luminophore. Here we were able to establish a simple and easily accessible synthesis route to different cyanostyryl-guanidininiocarbonyl-pyrrole (CGCP) derivatives. These compounds were investigated regarding their light-responsive double bond isomerisation, their molecular structures in single crystals by means of X-ray diffractometry, their emission properties by state of the art photophysical characterisation as well as bioimaging and assessment of cell toxicity.

Introduction

The design of small supramolecular ligands as artificial receptors for anions,¹ organic ammonium cations,² metal ions³ e. g. is still a very challenging field and numerous groups worldwide are working in this direction. Especially their use in biomedical chemistry⁴ or in the development of novel materials⁵ plays a key role and the search towards novel leitmotifs remains a challenging task.

Ever since the discovery of the well-known and widely applied oxo-anion-binding guanidiniocarbonyl-pyrrole (GCP)⁶ motif by Schmuck et al. back in the year 1999,⁷ this unique supramolecular ligand has become one of the working horses in numerous application fields ranging from materials science⁸ to biomedical chemistry.⁹

His initial design was based on the established binding of guanidinum-ions to oxo-anions in organic solvents.¹⁰

He aimed for improved binding affinities even in aqueous media by introducing additional hydrogen-bond donors (amides, pyrrole) leading to the development of the GCP binding motif.^{7a, 11} Nevertheless, tracking of GCP as well as other supramolecular receptors, in e. g. tissue, inside a cell or in environmental samples, remains a significant problem. In this regard especially luminophores, showing emission upon excitation are well-known compounds to label ligands enabling a real-time tracking using spectroscopy or microscopy techniques.¹²

Up to now, numerous classes of emissive compounds have been described and used for biomedical imaging, which either show fluorescence¹³ and/or phosphorescence¹⁴ emission, respond towards external stimuli¹⁵ or are sensitive towards their state of aggregation.¹⁶

The latter has become a very popular concept since the group of Tang et al. claimed the emission induction upon aggregation as aggregation-induced emission (AIE) in 2001.¹⁷ This phenomenon is based on a restriction of motion (RIM) of the typically attached phenyl rotors leading to a suppression of non-radiative decay based on motional freedom.¹⁸ Well-known examples of rotor-based AIE's are tetraphenylethenes,¹⁹ aromatic thioethers,²⁰ hexaphenylsiloles^{17, 21} or cyano-stilbenes.²² The ability of selected compounds to emit in the aggregated or solid state has been known for more than a century²³ and was described by scientists²⁴ as e. g. emission from solid solutions.²⁵ In any case, this phenomenon has experienced a renaissance and is widely applied in numerous fields ranging e. g. from organic light emitting diodes²⁶ to polymer sciences,²⁷ liquid crystals,²⁸ bio-imaging^16b as well as drug design.²⁹ Nevertheless, independently from the use of classic emitters such as rhodamine, fluoresceine or luminophores featuring aggregation-induced emission properties, labelling with luminophores often goes along with a tremendous influence on the properties of the ligand. This leads e. g. to a change in solubility, loss in binding affinity, increased toxicity, reduced cell permeability or altered physical properties, rendering labelled compounds hard to compare with the initial ligand.³⁰ In our approach, we planned to overcome this issue by merging a luminophore with the supramolecular ligand GCP. Here we used cyanostilbenes, known compounds featuring aggregation-induced emission properties. This class can be easily tuned regarding their photophysical properties by functional group variations and is known to undergo (Z) to (E) conversion under UV light irradiation (note: due to the higher priority of the -CN group (CIP rule), the isomer with the aromatic rings facing in opposite directions is attributed to the (Z)-isomer).³¹ Only very few cyanostilbenes are known, bearing pyrroles as heteroaromatic units.³² Therefore, we decided to use a cyanostilbene base to be merged with the GCP subunit leading to cyanostyryl-guanidininiocarbonyl-pyrroles (CGCP, Figure 1) featuring benefits from both compound classes: 1. aggregation-induced emission (AIE), 2. light responsive (Z) to (E) conversion, 3. capable for bio-imaging and 4. able to penetrate cellular membranes. This unique combination of multiple features leads to an all-in-one probe. The photoswitching leads to adaptive materials, which will be for sure interesting for materials science as well as for biomedical applications in terms of addressing different hot spots on e. g. proteins and cells with the same compound by means of photoexcitation. This design elegantly compromises an extension of the π-system, maintenance of the GCP features and easy synthetic access. With these compounds in hand, we aimed to fully characterize these novel compounds in terms of the above-mentioned parameters: photophysical performance, light-responsiveness, toxicity and potential applications in biomedical imaging.

Figure 1

A) Depiction of the merging between cyanostilbenes and guanidiniocarbonyl-pyrrols leading to CGCP, B) molecular structures of the five investigated compounds in this study, C) photographs of the five compounds under daylight and under UV-light excitation (λex=395 nm).

Results and Discussion

Synthesis

The synthesis of the CGCP library consisting of five variants, featuring electron donating and withdrawing groups, is based on the literature known tert-butoxycarbonyl (BOC) protected guanidiniocarbonyl-pyrrole-aldehyde (GCPA),³³ which was coupled to phenyl acetonitrile derivatives. 4-Methoxy- and 3,5- dimethoxyphenyl acetonitril were directly coupled in a Knoevenagel condensation under basic conditions yielding (Boc)CGCP-4-OMe and (Boc)CGCP-3,5-OMe in good yields. Although attempted, the free hydroxyl derivative of (Boc)CGCP-4-OMe could not be obtained directly from 4-hydroxyphenylacetonitrile. Hence a methoxymethyl (MOM) protected 4-hydroxyphenylacetonitrile was used, yielding (Boc)CGCP-4-OMOM.

To our surprise, the two nitro compounds could not be obtained via this well-known route and hence a non-conventional coupling, using boronic acid in methanolic solution, yielded the protected (Boc)CGCP-4-NO₂ and (Boc)CGCP-3-NO₂ derivatives. After deprotection of the BOC-groups, using 6 M HCl in THF under reflux, the hydrochloride salts of all five CGCP derivatives, CGCP-3,5-OMe, CGCP-4-OMe, CGCP-4-OH, CGCP-4-NO₂ and CGCP-3-NO_2, were obtained in good yields and as pure Z-isomers (Figure S1–27). It is worth mentioning that under these conditions, the MOM-group of (Boc)CGCP-4-OMOM was also cleaved.

Photophysical Properties and Light-Induced Photoisomerisation

All five compounds have been investigated regarding their photophysical properties in their solid state. It was found that all isolated compounds showed the expected excitation and emission properties (Table 1, Figure 2, S33 and S34) with photoluminescence quantum yields (Φ_PL) of up to 21 % as solids, but only weak emission was observed in the molecularly dissolved state (Figure S35 and S37), which can be attributed to the efficient non-radiative decay in the form of intramolecular vibration or rotation of the phenyl or pyrrole rotors. The lower Φ_PL of the nitro-compounds can be attributed to the well-known promotion of intersystem crossing, leading to triplet formation with a concomitant loss of emission efficiency.³⁴ Upon aggregation, the emission intensity is enhanced, underlining the AIE character of these compounds. Only for Z-CGCP-4-OH the photoluminescence efficiency does not change drastically in the aggregated state, suggesting dual state emission properties (Figure S36–37). Interestingly the compounds featuring electron donating groups like methoxy or hydroxy at the peripheral aromatic ring emit merely in the blue to green portion of the electromagnetic spectrum (488–559 nm). Whereas the nitro-substituted compounds show a significant bathochromic shift (λ_em up to 589 nm), compared to the previously described compounds, most likely due to the drastically changed electronic properties of Z-CGCP-4-NO₂ and Z-CGCP-3-NO₂. In all cases lifetimes in the nanoseconds range (Figure S38–42) were observed, which can be attributed to fluorescence from excited singlet states (Table 1). This approach allowed us to observe the variability in the values of radiative and radiationless deactivation rates (k_r and k_nr, respectively) with different substituents in various positions (Table S1). Electron-donating groups, such as methoxy in the para-position, resulted in high k_r and lower k_nr compared to the same group in the meta-position. The hydroxy group in the para-position showed a lower k_r and higher k_nr than the methoxy group in the same position. When electron-withdrawing groups, such as nitro (−NO₂), are in the para-position, both k_r and k_nr decrease compared to the analogous group in the meta-position.

Table 1. Overview regarding the photophysical properties of the investigated compounds in the solid state. a Absolute quantum yields were obtained by using an integrated sphere, b amplitude-weighted average lifetime.

Compound	λex [nm]	ΦPL ±0.02^[a]	λem [nm]	τav(amp) [ns]^[b]
Z-CGCP-4-OMe	458	0.21	488	2.21±0.06
Z-CGCP-4-OH	459	0.04	488	0.69±0.05
Z-CGCP-3,5-OMe	458	0.02	559	1.15±0.03
Z-CGCP-3-NO2	458	0.08	567	7.66±0.07
Z-CGCP-4-NO2	500	0.07	589	1.92±0.02

Figure 2

Emission spectra of all investigated compounds in the solid state. λex: Z-CGCP-3,5-OMe (420 nm), Z-CGCP-4-OMe (420 nm), Z-CGCP-4-OH (440 nm), Z-CGCP-4-NO2 (500 nm) and Z-CGCP-3-NO2 (458 nm).

However, repeated measurements of the compounds’ emission led to changes in the photoluminescence spectra, but without evidence of a change in the compound at the molecular level. An explanation is provided by the mechanochromism of the system, which undergoes a change in emission between (micro)crystalline arrangement and the amorphous state due to the measurement setup (a sample is placed between glass-slides and fixated with clips). By grinding Z-CGCP-4-OMe, exactly this change in emission was observed (Figure S29 and S31–32). and the change in morphology – loss in crystallinity – was detected by using PXRD (see discussion in the paragraph about powder X-ray diffractometry vide infra).

In addition to the photophysical properties, we were especially interested in the light-induced isomerisation of the double bond, which is known for cyanostilbenes. A first indication that both diastereomers, namely (Z) and (E), are present in solution, was found in the ¹H-NMR spectrum of (Boc)CGCP-4-OMe. Here a double set of signals (Figure S3) was observed, which was first attributed to the two isomers present in solution with a preference for the (Z)-isomer with oppositely facing rings, reducing the overall sterical hindrance (ratio (Z) : (E)=80 : 20).

However, after the cleavage of the Boc-groups under acidic and thermal conditions (reflux), just one set of signals and thus a single isomer was obtained. In order to determine whether the ¹H-NMR signals are related to the (E)- or the (Z)-isomer, a ¹H-coupled ¹³C-NMR (Figure S7) was measured, which confirmed the Z-Isomer of Z-CGCP-4-OMe by comparing the obtained coupling constant (J_CH=14.4 Hz) between the nitrile carbon and the vinylic proton, with literature known examples.³⁵

To our surprise the isomerically pure sample of Z-(Boc)CGCP-4-OMe converts under room light conditions (storage on the laboratory bench, quartz cuvettes) to the E-(Boc)CGCP-4-OMe.

To determine the static equilibrium of pure Z-CGCP-4-OMe in solution under room-light conditions, a sample of the pure (Z)-Isomer was exposed to ambient irradiation and ¹H-NMR spectra were recorded in intervals over a total period of 42 hours of illumination. To calculate the corresponding isomer ratio, the integrals of the vinyl H-signals of the (Z)-isomer at δ=7.76 ppm and its analogues signal of the (E)-isomer at δ=7.35 ppm were considered. By using an exponential fit function, the calculated equilibrium of the isomers in DMSO-d₆ under room light conditions, was determined to be an approximate ratio of 93 : 7 (E/Z).

Since Z-CGCP-4-OMe showed the highest Φ_PL, we focussed on this compound to evaluate the influence of double bond isomerisation on the emission properties.

To this end, subsequent irradiation (460 nm, 0.85 W LED, 5 h.) of a quantitative amount of Z-(Boc)CGCP-4-OMe in solution (DCM /MeOH), followed by solvent removal and chromatographic separation on SiO₂, led to the isolation of the pure Boc-protected (E)-isomer in 65 % yield (Figure 3).

Figure 3

Room-temperature isomerisation of Z-CGCP-4-OMe to E-CGCP-4-OMe, A) excerpt of the aromatic region in the time-dependent 1H-NMR spectra showing the decrease of the Z- and increase of the E-isomer, B) plotting the time dependent ratio changes of Z- to E-isomer by integration of the corresponding double bond protons, while the sample was exposed to room light.

In addition to the NMR spectroscopic confirmation of the (E)-isomer, the structure was also proven by X-ray diffractometry on single crystals (see Figure 5). The hydrochloride salt as (E)-isomer was obtained after acidic deprotection of E-(Boc)CGCP-4-OMe. To our surprise, E-CGCP-4-OMe was found to be extremely sensitive towards irradiation with light (405 nm). After 1.5 seconds of irradiation of a solid sample in the fluorescence spectrometer, a well detectable bathochromic shift (32 nm) was monitored, which then showed, after further 9 seconds of irradiation, no further changes (Figure 4B). Deviations from the emission wavelength in Table 1 can easily be explained by the sensitivity towards mechanical stress, leading to a mechanochromic shift, which has been proven (Figure S29 and S31–32). The sample on the quartz-plate (Figure 4A) was irradiated (0.67 W LED, 405 nm) for three more minutes to ensure full conversion followed by ¹H-NMR investigation.

Figure 4

A) Solid sample of E-CGCP-4-OMe after determination of the emission wavelength. The irradiated area shows a bathochromic shift which was attributed the E−Z conversion of the compound. B) Time dependent changes in the emission wavelength and intensity.

A regression of the (E)-isomer signals was monitored, while the corresponding signals of Z-CGCP-4-OMe increased (Figure S30). However, it should be mentioned that also additional signals can be monitored in the NMR spectrum, which are attributed to the product of a photochemical [2+2]-cycloaddition.^31b

X-ray Diffractometry on Single Crystals

E-(Boc)CGCP-4-OMe and Z-(Boc)CGCP-4-OMe (Figure 5) crystalize in the triclinic space group P$$ with one independent molecule in the asymmetric unit (Figure S43–47).³⁶ The Z-isomer is accompanied by an acetone molecule. Both isomers of the methoxy compound show protonation at the nitrogen atom in beta position of the pyrrole ring. The resulting intramolecular hydrogen bonds to the C=O moieties of the Boc-groups lead to a bent conformation of the molecules. The central packing motif in both isomers is a ring formed by two guanidinium groups which is comparable to the motif of a carboxylic acid dimer (blue in the packing diagrams, see ESI: Figure S45 and 47). In the (E) isomer the nitrile groups’ orientation allows the formation of another motif with the pyrrol's NH as a donor (red in the packing diagram, Figure S47). This leads to strings in the packing compared to isolated dimers in the (Z) isomer. Z-(Boc)CGCP-4-NO₂ crystalizes in the triclinic space group P$$ (Figure 6A and S50–53) with three independent molecules in the asymmetric unit while Z-(Boc)CGCP-3-NO₂ crystalizes in the monoclinic space group P2₁/n with Z’=1 (Figure S6B, 6 C and S52–54).³⁶ The NO₂ compounds are protonated at the nitrogen neighbouring the Boc-groups resulting in two intramolecular NH⋅⋅⋅O hydrogen bonds from the amino groups to the adjacent C=O groups. The consequence is a straight overall conformation of the pyrrole's side chain ending in the Boc-group. However, the differing protonation patterns increase the single bond character of the initial C−C bond – indicated by the increased bond length (Figure S52–53) – which allows the C=O group to rotate away from the pyrrole's NH group in Z-(Boc)CGCP-3-NO₂. The overall conformation of the molecule is crescent-shaped and leads to a set of intermolecular interactions, which allows a firm connection of two molecules. The properties of this self-recognition motif are currently investigated in more detail. In Z-(Boc)CGCP-4-NO₂, the C=O and the NH group point in the same direction, which follows a straight conformation of the whole molecule. Since growing of single crystals failed for the deprotected hydrochloride salts, attempts to grow the corresponding picrates succeeded in obtaining single crystals of (Z)-CGCP-4-OMe_Pic suitable for X-ray diffractometry. The conformation observed for the picrate of (Z)-CGCP-4-OMe (Figure 7)³⁶ was found to be similar to that of Z-(Boc)CGCP-4-NO₂. It crystallises in the monoclinic space group C2/c with one ion pair and two additional solvent molecules in the asymmetric unit (Figure S48–49). The crystallisation of (Z)-CGCP-4-OMe as a hydrochloride proved to be difficult thus the picrate (Z-CGCP-4-OMe_Pic) supplying numerous acceptors for hydrogen bonds was chosen. It is well-known that hydrogen bonding strongly influences the packing in the crystal and the pK_a value of the anion will have an impact on the protonation state.³⁷ Unfortunately, any attempts to crystallise other salts for comparison or the (E) isomer failed.

Figure 5

Molecular structure of Z-(Boc)CGCP-4-OMe (top) and E-(Boc)CGCP-4-OMe (bottom, solvent molecule omitted for clarity). Displacement ellipsoids are shown at 50 % probability levels. Irradiation was performed in dichloromethane/methanol. For details see ESI.

Figure 6

A) Molecular structure of Z-(Boc)CGCP-4-NO2 (only one of the independent molecules displayed) and B) Z-(Boc)CGCP-3-NO2. Displacement ellipsoids are shown at 50 % probability levels. C) Self- recognition motif of Z-(Boc)CGCP-3-NO2 (bottom).

Figure 7

Molecular structure of Z-CGCP-4-OMe as picrate salt (Z-CGCP-4-OMePic). Displacement ellipsoids are shown at 50 % probability levels and solvent molecule omitted for clarity.

X-ray Diffractometry on Powders

For this analysis, a powder sample of Z-CGCP-4-OMe was investigated by PXRD before (bg) and after grinding (ag). The measurements showed, that the mechanical treatment influenced the crystallinity of the samples switching from a micro- (bg) to nanocrystalline (ag) nature, whereas the crystal structure remained unchanged, as can be seen from the broadening and positions of the peaks, respectively (Figure S32). Furthermore, this also affected the intensities of the reflection peaks indicating a decreasing preferred orientation (texturing) by a factor appr. 10 after applying grinding, as can be verified for the peaks at smaller 2θ region (e. g. 5.2 or 10.5°), where low-indexed Miller indices hkl of an organic unit cell are typically located. Such decreasing peak intensity with a simultaneously increasing peak broadening (or decreasing crystallite size) during ag to bg transition can be explained by the larger bg crystals with the initially texture orientated layers, which were finally destroyed after grinding resulting in much smaller ag crystals without any preferred orientations (Figure 8). In order to quantify a change in the crystallite size (CS), a single reflection peak at 2θ=13.7° for both bg and ag samples was selected and refined (Figure 8), with the advantage that no other neighbour peaks would affect the peak profile due to undesirable convolution. The quantification showed that the crystallite size in the grinded samples decreased by a factor about 8.5 (from CS_bg=170 to CS_ag =20 nm) and this agreed well with the values for texturing which also goes along with the aforementioned changes in emission properties (Figure S30–31).

Figure 8

PXRD of Z-CGCP-4-OMe A) before and B) after grinding – changes in particle size/crystallite size. Representative X-ray powder diffractograms with Rietveld refinement (left) before (A) and after grinding (B) with a corresponding peak broadening as a result of changing crystallite size (right).

Cellular Internalization and Toxicity

Since molecules incorporating GCP-units have been reported to show high levels of internalization in cells^6b our compounds were analyzed regarding cellular uptake (Figure 9) and toxicity (Figure S55-57).

Figure 9

Internalization by Hela Kyoto cells, imaged using confocal fluorescence microscopy after 24 h treatment with the indicated compound (10 μM). Cells were imaged after fixation and staining with CellMask™ Deep Red. All investigated compounds are depicted in cyan color (independently from their actual emission wavelength) to ensure good contrast between the cell mask and the compound in the overlay images.

To investigate the cellular internalization, HeLa Kyoto cells were incubated with the respective compounds for 24 h (10 μM) and analyzed by confocal fluorescence microscopy. Briefly, cells were fixed and stained using CellMask^® Deep Red to allow robust and reliable cell segmentation despite the broad emission range of the employed compounds. The samples were then imaged by excitation of the luminophores at 405 nm and detection of the emission between 480 nm and 620 nm. This enabled equal treatment of all samples, although the compounds differed significantly in their emission maxima.

First, all compounds were detected in vesicles, which hints towards an uptake via endocytosis as a common mode of internalization. Z-CGCP-4-OH and Z-CGCP-4-NO₂ exhibit a strong vesicular localization and accumulation at the Golgi apparatus. Both compounds cannot be detected outside of vesicles suggesting a very low bioavailability. Z-CGCP-3,5-OMe also shows a profound accumulation in vesicles around the Golgi apparatus; however, it is also present throughout the cell in minor amounts and can even be detected within the nucleus. In comparison, the localization of Z-CGCP-4-OMe and Z-CGCP-3-NO₂ is more diffuse, although both can also be detected in vesicles. Both compounds are present in the nucleus and seem to accumulate around nuclear sub-structures resembling nucleoli. As such, Z-CGCP-4-OMe and Z-CGCP-3-NO₂ feature the highest level of bioavailability. The accumulation of Z-CGCP-4-OMe at nucleoli-like structures is however much more pronounced. It is worth to notice that high solubility of our compounds in DMSO/water mixtures leads to a decrease in emission properties of AIE-emitters. Interestingly we observed bright emission in cellular studies, which is easy to understand, since entrapment in vesicular structures or membranes leads to a motion hindrance and hence a drastic emission increase. To proof this hypothesis, we dissolved Z-CGCP-3-NO₂ compound in water/DMSO 99 : 1 and added TritonX at a concentration above the CMC, which leads to a drastic emission increase (see Figure S58) supporting our hypothesis.

Due to their ability to locate to substructures within the nucleus Z-CGCP-4-OMe and Z-CGCP-3-NO₂ are considered to have the highest level of bioavailability. The similarity between all compounds makes different internalization mechanisms unlikely; therefore, the differences in bioavailability most probable result from different rates of endosomal escape. Especially the striking differences between Z-CGCP-4-NO₂ and Z-CGCP-3-NO₂ are very interesting in this regard.

Next, HeLa Kyoto cells were treated with the respective compounds in concentrations ranging from 0.1–400 μM to investigate their effects on the viability of cells over 24 h (Figure S55–57). As all compounds were solved in DMSO, it was used as a control as well as the pure GCP aldehyde (GCPA). Both show only marginal toxicity at higher concentration (Figure S55). Cell viability was assessed using the Cell titer® AQ_ueous One kit (Promega). The compounds’ cytotoxicity considerably varied, which is interesting considering that most of them only differ in one closely related functional group (Table S13 and Figure S55–57). The most prominent toxic effect was observed in cells treated with Z-CGCP-4-OMe (Figure S56); here, a half maximal effective concentration (EC₅₀) of 10.27 μM was calculated (R²=0.98). Substitution of the methoxy group from Z-CGCP-4-OMe by a hydroxy group yields Z-CGCP-4-OH, which exhibits only minor cytotoxicity at concentrations above 100 μM (Figure S56). Z-CGCP-3,5-OMe, which carries two methoxy groups in meta position, manifests slightly less pronounced toxic effects than Z-CGCP-4-OMe (Figure S56). The two compounds carrying nitro groups only differ in their substitution pattern. While Z-CGCP-4-NO₂ (para) exhibits only a minor toxicity at concentrations above 100 μM, Z-CGCP-3-NO₂ clearly affects cell viability even at lower concentrations (EC₅₀=42.5 μM; R²=0.99). The observed cytotoxicity strongly seems to correlate with the compounds’ bioavailability as assessed by confocal microscopic analyses of the intracellular distribution patterns (Figure 9). Indeed, Z-CGCP-4-OH and Z-CGCP-4-NO₂ appear to be completely trapped in vesicles and thus unable to escape, which in turn results in a very low toxicity profile. Z-CGCP-4-OMe and Z-CGCP-3-NO₂ on the other hand can be partially observed in the nucleus, which necessitates prior endosomal escape. Here, high levels of cytotoxicity were observed for both compounds. Notably, Z-CGCP-3,5-OMe is less toxic than Z-CGCP-4-OMe or Z-CGCP-3-NO₂ and is also mostly observed in vesicles, although it seems to be able to escape the endosomes to some degree. Hence, internalization is probably mediated by the GCP group that is shared by all compounds. However, their bioavailability is then determined by the corresponding functionalization. The effects on cell viability in turn seem to rely on the degree of bioavailability and possibly also by the distinct intracellular localization of the compounds: Z-CGCP-4-OMe exhibits the highest level of bioavailability and is the most toxic compound. Interestingly, it also most prominently accumulates at nucleoli-like structures inside the cell nucleus. In conclusion,he here investigated Z-CGCP derivatives exhibit EC₅₀ values in the range of ten to higher than 100 μM, which lies in the range of most commonly used luminophores (e. g. BODIPY and rhodamine derivatives), where concentrations around and above 100 μM arecytotoxic.³⁸

Experimental Section

General Procedure for the Synthesis of (Boc)CGCP-3,5-OMe, (Boc) CGCP-4-OMe and (Boc)CGCP-4-OMOM

To a solution of the corresponding GCP aldehyde (GCPA, see ESI) (1 eq.) and the corresponding phenyl acetonitrile (2 eq.) in dry THF, potassium tert-butanolate (2 eq.) was added under argon. Then, the mixture was refluxed for 18 hours. After cooling down to room temperature, the mixture was poured into water and extracted with ethyl acetate. The combined organic layers were dried over sodium sulfate, filtered and the solvent was removed in vacuo. The crude product was then purified by silica chromatography.

General Procedure for the Synthesis of (Boc)CGCP-4-NO₂ and (Boc)CGCP-3-NO₂

A mixture of GCP aldehyde (GCPA, see ESI) (1 eq.), the corresponding nitro-phenyl acetonitrile (2 eq.) and boric acid (2 eq.) in dry methanol was refluxed for 18 hours under an argon atmosphere. After cooling down to room temperature, the mixture was poured into water, filtered and washed with a 2 : 1 mixture of water/methanol. After purification by silica chromatography the pure products were obtained.

General Procedure for the Acidic Deprotection

6 M HCl was added to a solution of the protected CGCP derivative in THF and the mixture was refluxed for 4 hours. After cooling down to room temperature, ethyl acetate or a mixture of ethyl actetate/ethanol was added, the precipitate was filtered off followed by washing with a mixture of ethyl acetate/ethanol. After drying in vacuo, the pure deprotected Z-isomers were obtained.

Conclusions

In summary, it was possible to synthesize a series of luminescent GCP-derivatives featuring light-responsiveness in terms of Z-E isomerization as well as potential application in bioimaging. Depending on the electronic effect of the substituents attached, emission maxima in the solid state ranging from 389 to 595 nm and hence covering a broad range of the visible electromagnetic spectrum were enabled. This also applies for the excitation wavelengths, which were all found to be higher than 400 nm making these compounds suitable candidates for bioimaging. In addition, as a representative example, we were able to investigate the dynamic photoisomerization of Z-CGCP-4-OMe under visible light (Figure S29). The E-isomer was isolated and also investigated in terms of photophysical properties, although a high sensitivity towards light was observed due to rapid double-bond isomerization in the solid state. By comparing the molecular structures in crystals of the (Boc)CGCP-4-OMe, the light sensitivity of the compound could be explained by the out-of-plane rotation of the pyrrole and phenyl moieties, which interrupts the conjugation of the π-system. In addition to the photophysical studies, the compounds’ cellular uptake and its effect on cell viability were also investigated. Hereby, it suggested that small structural changes are suffice to drastically influence the compound's cytotoxic potential. Relying on the photophysical properties, the cellular distribution could be monitored by confocal laser scanning microscopy. This revealed a clear correlation between compound internalization into the nucleoplasm, its nucleolar accumulation and its perturbative effect on cell viability. We foresee that this merging approach of supramolecular bioactive ligands with luminophores such as cyanostilbenes opens a broad range of application avenues. These compounds overcome classic drawbacks of ligand labelling by using the merged product as active species. CGCP's will be investigated in the future with regard to anion recognition, as building block in supramolecular materials as well as potential gene transfection units, by enabling a fast tracking and an environmentally sensitive emission due to their intrinsic AIE properties. The use of Z or E isomers enables structure-property relationship studies in terms dipole and hence polarity as well as geometry changes in biomedical samples. In addition in situ photoisomerization in novel materials, leading to adaptive behavior will be of interest for future research projects.

Supporting Information

The authors have cited additional references within the Supporting Information.^{33a, 39}

Author Contributions

Jan Balszuweit: Data curation, Conceptualization, Writing – original draft, Investigation. Paul Stahl: Data curation, Writing – original draft, Investigation. Victoria Cappellari: Photophysical characterization, Rick Y. Lorberg: Data curation, Investigation, Writing – review and editing, Christoph Wölper: Data curation, Writing – original draft, Oleg Prymak – Data curation, Writing – original draft, Felix C. Niemeyer: Data curation, Investigation. Johannes Koch: Data curation, Investigation, Shirley K. Knauer: Funding acquisition. Cristian A. Strassert: Funding acquisition, Writing – original draft. Jens Voskuhl: Conceptualization, Funding acquisition, Writing – original draft, Supervision, Project administration.

Acknowledgments

J.V. S.K.K and J.B. thank the “Deutsche Forschungs-gemeinschaft“ (CRC 1093, project number: 229838028, project A10, B5) for financial support. We acknowledge the use of the imaging equipment and the support of the “Imaging Center Campus Essen” (ICCE). Instrument Leica TCS SP8X FALCON was obtained through DFG funding (Major Research Instrumentation Program as per Art. 91b GG, INST 20876/294-1 FUGG). This work was supported by the DFG/ Land NRW (INST 211/915–1 FUGG: “Integrated confocal luminescence spectrometer with spatiotemporal resolution and multiphoton excitation”, C.A.S.). Kevin Rudolph is acknowledged for his support during the synthesis of the compounds. Open Access funding enabled and organized by Projekt DEAL.

Conflict of Interests

There are no conflicts to declare.

Open Research

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.