Bithiophene‐Cored, mono‐, bis‐, and tris‐(Trimethylammonium)‐Substituted, bis‐Triarylborane Chromophores: Effect of the Number and Position of Charges on Cell Imaging and DNA/RNA Sensing

Abstract The synthesis, photophysical, and electrochemical properties of selectively mono‐, bis‐ and tris‐dimethylamino‐ and trimethylammonium‐substituted bis‐triarylborane bithiophene chromophores are presented along with the water solubility and singlet oxygen sensitizing efficiency of the cationic compounds Cat1+ , Cat2+ , Cat(i)2+ , and Cat3+ . Comparison with the mono‐triarylboranes reveals the large influence of the bridging unit on the properties of the bis‐triarylboranes, especially those of the cationic compounds. Based on these preliminary investigations, the interactions of Cat1+ , Cat2+ , Cat(i)2+ , and Cat3+ with DNA, RNA, and DNApore were investigated in buffered solutions. The same compounds were investigated for their ability to enter and localize within organelles of human lung carcinoma (A549) and normal lung (WI38) cells showing that not only the number of charges but also their distribution over the chromophore influences interactions and staining properties.

The synthesis of Neut1, Neut(i)2, and Neut3 was made possible via a synthetic route to unsymmetrically-substituted triarylboranes which was previously not well developed, [39] but has been reported very recently by our group. [40] Results and Discussion
To obtain Neut1, Neut2, and Neut3, 5-bromo-2,2'-bithiophene 2 was cross-coupled with 4b, 5b, and 6b, respectively, in Suzuki-Miyaura reactions using the conditions described above to give 7a, 8a, and 9a in 71 %, 93 %, and 80 % yields, respectively (Scheme 2). Prior to the second Suzuki-Miyaura cross-coupling, these intermediates were halogenated at the 5'position of the bithiophene. Compound 7a was brominated at this position in 93 % yield using NBS in DMF. For the aminated compounds 8a and 9a, it was not possible to affect bromination with NBS at the desired position due to the more activated positions at the dimethylamino-substituted arene. Therefore, the bithiophene moiety was lithiated with nBuLi and then treated with I 2 . This gave inseparable mixtures containing 80 % of the iodinated compounds 8b and 9b and their respective starting materials as determined by 1 H NMR spectroscopy. As the starting material cannot undergo Suzuki-Miyaura crosscoupling, these mixtures were used directly, after minor purification, for the synthesis of Neut3. It was possible to isolate Neut3 from both reaction sequences (8b + 6b and 9b + 5b). However, the first reaction sequence gave higher yields in all steps, especially the last one (60 % vs. 6 % yield).
With all uncharged compounds Neut1-3 in hand, their cationic counterparts were synthesized using methyl triflate in CH 2 Cl 2 similar to previous reports [26][27]29,35,40] giving Cat 1 + , Cat 2 + , Cat(i) 2 + , and Cat 3 + in 88 %, 69 %, 49 %, and 57 % yields, respectively. The neutral triarylboranes 5a and 6a were methylated similarly giving 5c and 6c in 89 % and 93 % yields, respectively. With the neutral and cationic triarylboranes and bis-triarylboranes in hand, their photophysical, and electrochemical properties were investigated as well as the singlet oxygen sensitizing efficiency and water-solubility of Cat 1 + -Cat 3 + and their DNA/RNA binding affinities in buffered solutions. Additionally, the behavior of the most promising compounds Cat(i) 2 + and Cat 3 + was investigated in two different cell lines.

Solubility in water
While Cat 4 + is soluble in pure water with a maximum solubility of 1.0 mM, [26] none of the new, mono-, di-or tri-cationic bistriarylboranes are soluble in pure water at concentrations of 2.6-3.4 μM ( Figure S64, Table S2). However, concentrated solutions of Cat 1 + -Cat 3 + in acetonitrile can be diluted with water to less than 1 % without precipitation or aggregation. The concentrations of the cationic compounds in these diluted

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 solutions were as high as 15 to 34 μM and are suitable for photophysical studies. Thus, aqueous solutions of the cationic bis-triarylboranes were prepared from concentrated acetonitrile solutions which were diluted with water until the solution contained � 1 % acetonitrile. For studies of interactions with DNA/RNA in buffered solutions and cells, stock solutions were prepared in DMSO (c = 1 × 10 À 2 mol L À 1 ) and diluted prior to use with the buffered solutions indicated.

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 and Figures S65-S78 in the Supporting Information). The photophysical properties of the triarylboranes and of the neutral bistriarylboranes are discussed briefly here (Figures 4-6) and a more detailed discussion is given in the Supporting Information.
The absorption spectra of the triarylboranes display no solvatochromism. The introduction of a strong electron-donating functionality, in this case dimethylamino group(s), leads to bathochromically shifted absorption and emission maxima. Upon methylation, the positive solvatochromism of the emission is almost lost due to the loss of the charge-transfer process which was supported by DFT and TD-DFT calculations using the orbital overlap parameter Λ. This parameter is defined as with f a and f i representing the occupied and unoccupied one-electron wavefunctions, respectively, j < f a j j j f i > j representing the norm of the one-electron wavefunction centroid and c i,a representing the weight of the one-electron excitation. Λ takes values between 0 and 1, with 0 corresponding to no overlap and 1 to complete overlap. [45] For all mono-triarylboranes, these values are smaller than 0.6, thus suggesting localized orbitals and CT character of the transition. [45] Hence, the photophysical properties of 4a, 5c, and 6c are very similar, as these compounds are electronically similar due to the conversion of the electron-donating amine to trimethylammonium groups upon methylation. The molar extinction coefficient increases almost linearly with the number of dimethylamino groups in the neutral triarylboranes. For 5c and 6c, these values are lower and independent of the number of trimethylammonium groups. The properties of the neutral bis-triarylboranes are very similar when dissolved in the same solvent, except for the molar

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 extinction coefficients, which increase non-linearly with the number of amino groups. The emission spectra display chargetransfer character for Neut1-Neut3, but not for Neut0 as the latter bears no amino groups. DFT calculations show the HOMOs and LUMOs of the neutral bis-triarylboranes to be mainly localized at the bithiophene bridge. Thus, transitions are of π-π* nature. This is supported by TD-DFT calculations and Λ values resulting therefrom, which are between 0.66 and 0.72 thus reflecting delocalized orbitals over the bridge and locally excited transitions. [45] In contrast, for the cationic bis-triarylboranes, Λ values are between 0.30 and 0.53 reflecting chargetransfer character of the transitions, and rather localized orbitals. Thus, charge transfer character of the transition of lowest energy is lost upon methylation, which has been shown previously for very similar compounds. [26] The orbitals obtained from DFT calculations suggest that the HOMOs of the cationic bis-triarylboranes are mainly localized at the bithiophene bridge as it is for the neutral compounds. However, the LUMOs of the cationic compounds are mainly localized at the most electron deficient boron center (5c for Cat 1 + ; 6c for Cat 2 + , Cat 3 + ). Therefore, the transition for most of the cationic bis-triarylboranes is not of π-π* nature but of π-B (p) nature. In the case of Cat(i) 2 + and Cat 4 + , [26] the LUMO is delocalized from one boron center to the other over the bithiophene bridge. As a result, the transition is of π-π* nature. Due to the different origins of the main transitions of the cationic bis-triarylboranes, the photophysical properties are not the same for these compounds as they depend to some extent on the number and distribution of the trimethylammonium groups. For example, in acetonitrile, the absorption maxima shift bathochromically with increasing number of trimethylammonium groups. The emission maxima shift bathochromically with increasing dipole moment of the cationic bis-triarylboranes in the order Cat 1 + � Cat(i) 2 + ! Cat 3 + < Cat 2 + . Note that the dipole moment of charged compounds is defined relative to its origin. Thus, it is not an observable quantity. However, the term dipole moment will be used herein to describe the distribution of the electron density over the molecules for convenience. The molar extinction coefficient does not change significantly with increasing number of trimethylammonium groups. Similar, but less pronounced trends are found in 1 % MeCN in water and dilutions of DMSO stock solutions with sodium cacodylate buffer solutions (see Supporting Information).
In summary, the absorption and emission maxima are solvent dependent as expected for donor-acceptor (D-A) compounds in which the boron center is the electron-acceptor. The electron-donor for the neutral compounds is the dimethylamine moiety. In the case of 4a and 5c, the mesityl moiety acts as an electron-donor, whereas for 6c, only the 2,6-dimethylphenyl motif serves as the donor, as supported by DFT and TD-DFT calculations. Upon methylation of the neutral bistriarylboranes, the bithiophene becomes the electron-donating motif. The absorption and emission maxima of the cationic bistriarylboranes Cat 1 + , Cat 2 + , Cat(i) 2 + , and Cat 3 + are bathochromically shifted compared to their respective monotriarylboranes. However, no additive effect from combining the triarylboranes with the bithiophene bridge was observed for the molar extinction coefficient. Thus, it can be concluded that the photophysical properties mainly result from the interaction of the most electron-poor boron center with the most electronrich donating moiety for all compounds investigated.

Electrochemistry
Electrochemical measurements were performed in the respective solvent (THF or acetonitrile) with [nBu 4 N][PF 6 ] as the electrolyte and a scan rate of 250 mV s -1 . Each measurement is referenced to the ferrocene/ferrocenium (Fc/Fc + ) couple. The redox potentials of uncharged and cationic triarylboranes 4a-6a, 5c, and 6c were examined as well as those of the neutral and cationic bithiophene bis-triarylboranes Neut0-4 and Cat 1 + -

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 Cat 2 + . Potentials of the neutral compounds were measured in THF, and those of the cationic compounds in MeCN. All cyclic voltammograms are shown in the Supporting Information ( Figures S82-S99) together with an extended discussion thereof. The electrochemical data are summarized in Table 2. No oxidation potentials are expected for 4a, 5c and 6c, within the electrochemical window of the solvent used. In the case of 5c and 6c, the presence of ca. 0.5 % starting material was confirmed by 1 H NMR spectra. Thus, the minor oxidation waves observed for both compounds can be attributed to the presence of non-or partially methylated starting materials.
For the triarylboranes 4a, 5a, 6a, 5c, and 6c one partially reversible 1e À reduction potential for the boron center and one 1e À oxidation potential per amino group was found. The reduction potentials decrease with increasing electron density at the boron center which increases in the order 6c < 5c < 4a < 5a < 6a. The electrochemical data are in good agreement with values reported by Gabbaï and coworkers for very similar trimethylammonium-substituted triarylboranes. However, the trend that each trimethylammonium group leads to a decrease of 0.26 V in MeCN [35] was only observed for the addition of the second cationic group (5c vs. 6c) but not for the addition of the first one (4a vs. 5c).

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 However, for Cat 1 + , Cat 2 + , and Cat 3 + , two partially reversible 1e À reductions were observed for each compound whereas Cat(i) 2 + and Cat 4 + display only one reduction potential. These result from two simultaneous 1e À reductions (Table S20), as the two boron centers in the latter are the same on both sides of the molecule. It was not possible to improve the resolution of the quasi-reversible reduction waves of Cat 1 + by pulse voltammetry experiments ( Figure S94D).

HOMO-LUMO gaps
HOMO-LUMO gaps (ΔE calc , ΔE opt , ΔE CV ) were obtained from DFT calculations and experimental values. The absorption spectra were used to determine the optical bandgap ΔE opt and ΔE CV was calculated from the HOMO and LUMO energies according to a literature procedure (see Supporting Information). [46][47][48][49] The results are summarized in Table 3. ΔE opt does not change significantly within a series of neutral or cationic bis-triarylboranes but decreases by ca. 0.5 eV upon introduction of the strong dimethylamino electron donor in the mono-triarylboranes, due to destabilization of the HOMO more than the LUMO, as supported by DFT calculations. The values of ΔE opt and ΔE CV are of the same magnitude and display the same trends. For the cationic compounds, only the cationic moiety was used in the calculations, so ion-ion-interactions were not considered.

Singlet oxygen sensitization
The ability to sensitize singlet oxygen in O 2 -saturated acetonitrile solutions was investigated for the triarylboranes 4a, 6c, and 5c as well as for the charged bis-triarylboranes Cat 1 + -Cat 4 + by monitoring the weak 1 O 2 phosphorescence emission at 1275 nm vs. that obtained from a solution of the known sensitizer perinaphthenone [50] with an estimated error of � 0.1. Triarylboranes 4a, 6c, and 5c sensitize 1 O 2 with efficiencies of 0.3, 0.5, and 0.6, respectively, (Figures S79) which increases with increasing number of trimethylammonium substituents (Table S17). The corresponding fluorescence quantum yields in acetonitrile decrease in the same order from 0.34 to 0.11 and 0.10. Thus, the sum of singlet oxygen efficiency Φ Δ and fluorescence quantum Φ f yield for these compounds is ca. 0.6 ( Table S17).
The charged bis-triarylboranes sensitize 1 O 2 with efficiencies of 0.6, 0.8, 0.7, 0.6, and 0.6 for Cat 1 + , Cat 2 + , Cat(i) 2 + , Cat 3 + , and Cat 4 + , respectively ( Figure 9, Table 4). The values for Cat 2 + and Cat 3 + may be overestimated due to tailing of the compound's  [a] No data were reported for this compound. [26] [b] A different basis set was used. [26] [c] Calculated from reported absorption spectra. [26] Chemistry-A European Journal Full Paper doi.org/10.1002/chem.202102308 emission into the NIR region of the singlet oxygen emission ( Figure S80), which is the reason for the raised baseline between 1230 and 1260 nm in Figures 9B and 9D. Thus, Φ Δ of ca. 0.6 is estimated for all charged bis-triarylboranes.
The S 1 excited states of Cat 1 + -Cat 4 + thus decay ca. 40 % by fluorescence and ca. 60 % by intersystem crossing to the triplet state, which very efficiently sensitizes 1 O 2 . Attaching two triarylboranes to a bithiophene motif leads to a more efficient

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 energy transfer to triplet oxygen than for the mono-triarylboranes. Thus, interactions with DNA, RNA and DNApore and cell imaging studies were only examined for the cationic bistriarylboranes.

Studies in buffered solutions
Interactions of the cationic bis-triarylboranes Cat 1 + -Cat 3 + and naturally occurring calf thymus DNA (ctDNA; typical B-helical structure with a balanced ratio of GC-(48 %) and AT-(52 %) base pairs), double stranded (ds) poly rA-poly rU (pApU; A-helical structure characterized by its major groove being available for binding of bulky small molecules) and DNApore ( Figure 10A) were investigated in aqueous buffered solutions. Experimental details and all spectra are given in the Supporting Information. DNApore was chosen as Cat 1 + -Cat 4 + should fit in the cavity of the structure ( Figure 10B).
The stability of Cat 1 + -Cat 3 + in buffered solutions was confirmed by temperature dependent absorption and emission spectra (Figures S101-S108).
Changes in the fluorescence of Cat 1 + -Cat 3 + upon addition of increasing amounts of DNA, or RNA were studied at pH 7 and/or pH 8. In combination with the results of thermal denaturation (T m ) and circular dichroism (CD) experiments of the polynucleotides, which are known to be unique for the respective molecule, conclusions about the mode of interaction (intercalation, major or minor groove binding, or external binding) can be drawn. [51][52][53][54] For example, increasing denaturation temperatures of polynucleotides in the presence of small molecules, as observed for our bis-triarylboranes Cat 2 + -Cat 3 + , suggest non-covalent binding of the latter, characterized by the type of interactions which can increase the stability of DNA/ RNA (e. g. intercalation, H-bonding, and/or cation-anion interactions). However, the emission of Cat 1 + showed negligible changes in the presence of any polynucleotide at any pH ( Figure 11, Figure 13, Table 6). In some cases, precipitation of Cat 1 + was observed. In addition, the denaturation temperatures ( Figure 12, Figure S109, Figure S125, Table 5) or the CD spectra (Figures S121, Figure S141, Figure 14) of the polynucleotides were not influenced by the presence of the mono-cation. This behavior is similar to that of neutral analogues investigated previously. [37] Thus, one positive charge is not sufficient for the bis-triarylboranes to provide efficient interaction with polynucleotides.

Interaction of Cat 2 + -Cat 3 + with ctDNA, and pApU at pH 7
In the presence of Cat 2 + , Cat(i) 2 + , and Cat 3 + , T m is increased compared to pure ctDNA or pApU at pH 7. Thus, these three compounds efficiently bind to ctDNA and pApU and stabilize these polynucleotides (Table 5; Figure S110-S112). However, the stabilizing effects of Cat 2 + , Cat(i) 2 + , and Cat 3 + are lower than that reported for Cat 4 + . [38] [26] [a] Values obtained from singlet oxygen sensitization measurements overestimate the actual value due to tailing of the emission spectrum of the compound between 1230 nm and 1330 nm.

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 Upon addition of ctDNA or pApU, the emission of Cat 2 + -Cat 3 + increases depending on the structure of the polynucleotide and the number of charges. Using the Scatchard equation [55] and the McGhee, von Hippel formalism, [56] binding constants were obtained, which are very similar for all compounds and polynucleotides investigated ( Table 6). Such non-selective affinity with respect to the secondary structure of the ds-polynucleotide and the composition of its respective base pairs was reported previously for structurally related compounds, [36][37][38] and suggests that Cat 2 + , Cat(i) 2 + , and Cat 3 + bind in the minor groove of DNA or the major groove of RNA.
Upon addition of Cat 2 + , Cat(i) 2 + , and Cat 3 + to ctDNA or pApU, a small decrease of the intensity of the CD spectra between 230 nm and 300 nm (Figures S122-S124) was ob-served. This can be attributed to an unwinding of the double helix of the polynucleotides upon insertion of the bis-triarylboranes. In contrast to the previously reported tetra-cationic compound Cat 4 + , [38] no ICD bands were observed at 400-500 nm. Thus, Cat 2 + , Cat(i) 2 + , and Cat 3 + might be nonuniformly oriented with respect to the chiral axis of the polynucleotide, but it was not possible to derive direct structural information. Nevertheless, in analogy with our previous binding studies, [36][37][38] groove binding of Cat 2 + , Cat(i) 2 + , and Cat 3 + can be assumed.

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 Interaction of Cat 1 + -Cat 3 + with ctDNA, and DNApore at pH 8 Similarly, as described above for sodium cacodylate solutions, thermal denaturation temperatures of ctDNA and DNApore in Tris-buffer at pH 8 (15 mmol L -1 Tris-HCl, 300 mmol L À 1 KCl) were increased in the presence of Cat 2 + , Cat(i) 2 + , Cat 3 + , and Cat 4 + (Table 5, Figure 12). The denaturation curve of DNApore is slightly biphasic, most likely due to the presence of two different conformers. Both transitions were stabilized by interaction with the bis-triarylboranes, with the effect on the melting temperature of DNApore being proportional to the number of cationic charges in the order Cat 1 + < Cat 2 + � Cat(i) 2 + < Cat 3 + < Cat 4 + . The unusual drop of absorbance observed for Cat(i) 2 + upon addition of DNA is likely due to precipitation of a Cat 2 + /DNA complex upon complete unwinding.
In contrast to the emission increase of Cat 2 + and Cat(i) 2 + upon addition of ctDNA at pH 7, at pH 8, the emission of these compounds decreases ( Figure 13A), while the emission of Cat 3 + increases in both solutions. However, upon addition of DNApore, the emission of Cat 2 + , Cat(i) 2 + , Cat 3 + , and Cat 4 + decreases ( Figure 13B). Binding constants are very similar for the different bis-triarylboranes at pH 8 (Table 6). However, upon changing the pH from 7 to 8, the binding constants of ctDNA are lower, especially for Cat 3 + . In contrast, for all compounds, binding constants to the DNApore are larger than those for ctDNA at pH 8. Thus, the chromophores bind more strongly to DNApore than to ctDNA.
Addition of Cat 1 + to ctDNA at pH 8 resulted in changes of the CD spectrum at < 300 nm which were not observed at pH 7. Thus, a change of the interaction between this bistriarylborane and ctDNA takes place upon increasing the pH from 7 to 8. For Cat(i) 2 + and Cat 3 + , a negative band at ca. 425 nm was observed, indicating a perpendicular orientation of the transition moments to the chiral axis of ctDNA at pH 8.
The chiral axis of the DNApore cannot be directly correlated to the usually well-defined chiral axes of typical ds-DNAs due to the six intertwined 50-mer oligonucleotides and, consequently, the tilted angles of the double helices with respect to the central axis of the DNApore. Thus, the following ICD signal assignments do not necessarily match those described in the literature. [54,[57][58] However, due to the similar structures of the bis-triarylboranes studied herein, and the same experimental procedures applied, the results of these measurements can be used for comparisons within this series of compounds. For Cat 2 + , weak, negative ICD bands were observed which is Table 6. Binding constants [a] (log Ks) of Cat 1 + , Cat 2 + , Cat(i) 2 + , Cat 3 + , and Cat 4 + with polynucleotides calculated from fluorimetric titrations according to literature procedures. [55][56]  [a] Analyses of titration data by means of the Scatchard equation [55] with von Hippel formalism [56]

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 characteristic for a perpendicular orientation of the transition moment of the chromophore with respect to the chiral axis of the DNApore. In contrast to the CD measurements at pH 7, which did not change at wavelengths > 300 nm, addition of Cat(i) 2 + to DNApore at pH 8 induced strong ICD bands at 400-450 nm ( Figure 14). These changes fit the absorption spectra of the corresponding bis-triarylboranes perfectly, which shows that the ICD signals are derived from our chromophores. The positive ICD bands suggest a tilted orientation by ca. 60°to the chiral axis of the DNApore. For Cat 3 + , complete disintegration of the CD bands of DNApore suggests strong changes of the secondary structure of the DNApore upon binding. Thus, Cat 2 + , Cat(i) 2 + , Cat 3 + , and Cat 4 + interact with all polynucleotides investigated. Although there are some similarities, e. g. interaction with DNApore leads to emission decrease of the compounds, the number and distribution of charges affects the interaction of bis-triarylboranes with different polynucleotides. In particular, the induced CD bands of various derivatives ( Figure 14) can be used for fine sensing of the DNApore shape. Thus, Cat 2 + , Cat(i) 2 + , Cat 3 + , and Cat 4 + are useful probes for further studies of the DNApore structure and interactions with novel ligands.

Cell toxicity and imaging studies
Studies were conducted on the cationic bis-triarylboranes to see whether they penetrate cell membranes, stain specific organelles therein, or if they have anti-proliferative effects to A549 and WI38 cells. Confocal microscopy studies showed that the mono-cationic compound Cat 1 + does not penetrate the cells, in contrast to Cat 2 + -Cat 3 + . Thus, this compound was not further investigated in biological studies. For Cat 2 + , only preliminary results were obtained, suggesting localization at the cellular membrane and certain compartments in the cell nucleus. However, due to the low stability of this compound in Dulbecco Modified Eagle's Medium (DMEM) containing 10 % fetal bovine serum (FBS), only the staining patterns of Cat(i) 2 + and Cat 3 + were further investigated.
The cytotoxicity of Cat(i) 2 + and Cat 3 + was tested using the MTT assay against human lung carcinoma (A549) and normal lung (WI38) cell lines ( Figure S146). In the dark, Cat(i) 2 + is not cytotoxic in all concentrations tested (0.1 μM, 1 μM, 10 μM). However, when irradiating the incubated cells of both cell lines in the presence of 1 μM Cat(i) 2 + at 400-700 nm for 5 min, the viability of the cells was reduced drastically with increasing concentration of Cat(i) 2 + .
In the dark, Cat 3 + is not cytotoxic to both cell lines in concentrations of 0.1 μM and 1 μM. However, at 10 μM concentrations, Cat 3 + is cytotoxic to A549 cells. This toxicity is even more pronounced for WI38 cells. For both cell lines, the toxicity increases under irradiation (400-700 nm) at 1 μM and 10 μM concentrations.
The increased toxicity of these compounds when irradiated with UV light results most likely from their ability to sensitize singlet oxygen (see above) as the latter is a highly reactive compound leading to cell death. [59] This assumption was confirmed by monitoring cells incubated with Cat 2 + , Cat(i) 2 + , or Cat 3 + by confocal microscopy while irradiating at 405 nm at the full power of the light source (LED). Within 2 min, significant changes in the cell morphology were observed (cell blebbing, cell contraction; Figures S148, S150, S152) suggesting strong cellular damage. Simultaneously, the emission of the dyes bleached rapidly within 30 s to 1 min (Figures S147, S149, S151). The stability of the emission increases with the number of cationic charges. After the emission is quenched completely, no additional cell damage was observed. Control experiments on untreated cells did not show any sensitivity to light exposure. This is in good agreement with the cell toxicity of Cat(i) 2 + and Cat 3 + under UV irradiation found during the MTT assay.
However, when incubating A549 cells for 90 min (37°C, 5 % CO 2 ) with Cat(i) 2 + , or Cat 3 + and exciting with light (405 nm; LED source), the cells show strong green emission due to the accumulation of the compounds in some organelles in the cytoplasm. The staining patterns of both compounds were further analyzed by co-localization experiments ( Figure 15; Figures S153-S154). The degree of co-localization was quantified using the Pearson correlation coefficient (R r ) [60] and corrected by replicate-based noise correction correlation. [61] For Cat(i) 2 + (0.54) and Cat 3 + (0.46), the best overlap was obtained for co-localization with LAMP1, a staining antibody selective for lysosomes. However, the staining pattern of the bis-triarylboranes suggests further interactions with early endosomes and golgi apparati (Figure S153-S154; Tables S23-S24). Thus, these new compounds are not specific for one organelle but are distributed between the organelles mentioned with some selectivity for lysosomes.

Chemistry-A European Journal
Full Paper doi.org/10.1002/chem.202102308 Cat 2 + , Cat(i) 2 + , Cat 3 + ) and their neutral precursors (Neut1, Neut2, Neut(i)2, Neut3) are described. The compounds were isolated in good yields and the influence of different numbers and locations of charges on electrochemical and photophysical properties and cellular imaging was investigated. Photophysical investigations reveal expected behavior for D-A compounds wherein the boron center is the electron-acceptor in all compounds with differing electron-donating moieties, as supported by DFT and TD-DFT calculations. For the cationic bistriarylboranes, an increasing hypsochromic shift of the emission spectra with increasing dipole moment was observed in acetonitrile. CV measurements of the neutral and cationic bistriarylboranes show that the π-conjugation, and therefore the communication between the boron centers, is improved upon methylation of the dimethylamino groups. Efficient singlet oxygen sensitization was demonstrated for the cationic bistriarylboranes Cat 1 + , Cat 2 + , Cat(i) 2 + , and Cat 3 + . For comparison with the properties of the bis-triarylboranes, selected properties were also investigated for the mono-triarylboranes 4a-6a, 5c, and 6c.
Due to the moderate solubilities of Cat 1 + , Cat 2 + , Cat(i) 2 + , and Cat 3 + in pure water, photophysical investigations and biological studies were performed in aqueous environments containing less than 1 % acetonitrile and 0.1 % DMSO, respectively. Studies with DNA and RNA in buffered solutions show negligible binding for the mono-cationic compound Cat 1 + but strong binding for Cat 2 + , Cat(i) 2 + , and Cat 3 + within the minor groove of DNA or the major groove of RNA exhibiting a strong fluorescence increase under physiological conditions. The compounds also bind strongly to DNApore, most likely inside the pore. Thus, they could potentially act as a "stopper" of ubiquitous material flow, which DNApore enables when inserted in, e. g., bacterial membranes. [62] Due to their induced CD response, Cat 2 + , Cat(i) 2 + , and Cat 3 + might be useful to study the fine structure of DNApore and the interaction with other ligands. Studies with A549 and WI38 cells show all compounds to be cell permeable except for the mono-cation Cat 1 + . Under intense irradiation (400-700 nm), the emission of Cat 2 + , Cat(i) 2 + , and Cat 3 + bleaches rapidly with simultaneous changes of cell morphology, most likely due to in situ formation of singlet oxygen. Thus, the cytotoxicity of the compounds is strongly increased upon irradiation with light (400-700 nm) as demonstrated using the MTT assay. Without irradiation, Cat(i) 2 + and Cat 3 + are not cytotoxic up to concentrations of 10 and 1 μM, respectively, for both cell lines. Co-localization experiments of both compounds in A549 cells demonstrate the lysosome as the main but not only accumulation site, since Cat(i) 2 + and Cat 3 + also localize in other organelles. Thus, the cationic bis-triarylboranes are promising theranostic agents [63] as they combine the potential for photodynamic therapy due to their relatively high singlet-oxygen sensitizing efficiency with high fluorescence quantum yields for simultaneous imaging of the location of the chromophores in cells or tissues. Furthermore, this study demonstrates that the number of cationic Figure 15. Intracellular localization of Cat(i) 2 + and Cat 3 + in A549 cells. Co-localization with lysosomes (LAMP1) was monitored by confocal microscopy. Cells were treated with the bis-triarylboranes at concentrations of 10 μM for 90 min at 37°C. Nuclei were stained with DAPI. Co-localization was assessed by determining the Pearson correlation coefficient. trimethylammonium groups and their distribution influences all properties investigated, including intracellular localization. Thus, when designing new bis-triarylborane chromophores for biological applications, the number and distribution of charges should be considered as well as the bridging unit.

Crystal structures
"Deposition Number(s) 2072401 (for 4a) and 2072402 (for 4b) contain(s) the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structures service."