A Plasma Membrane Polarity Sensor Harnessing Conjoined Twisted Intramolecular Charge Transfer Modulation and Charge Number Control Strategy

Open AccessCCS ChemistryRESEARCH ARTICLES1 Mar 2024A Plasma Membrane Polarity Sensor Harnessing Conjoined Twisted Intramolecular Charge Transfer Modulation and Charge Number Control Strategy Yu-Qiang Zhao†, Le Yu†, Liping Zhu, Jinsong Liang, Ying Zhou and Jong Seung Kim Yu-Qiang Zhao† College of Chemical Science and Technology, Yunnan University, Kunming 650091 , Le Yu† College of Chemical Science and Technology, Yunnan University, Kunming 650091 Department of Chemistry, Korea University, Seoul 02841 , Liping Zhu College of Chemical Science and Technology, Yunnan University, Kunming 650091 , Jinsong Liang College of Chemical Science and Technology, Yunnan University, Kunming 650091 , Ying Zhou *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] College of Chemical Science and Technology, Yunnan University, Kunming 650091 and Jong Seung Kim *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Department of Chemistry, Korea University, Seoul 02841 TheranoChem Incorporation, Seoul 02856 https://doi.org/10.31635/ccschem.023.202303309 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Polarity is a critical microenvironmental factor of the plasma membrane, which can offer valuable insights into various biological processes. Herein, we proposed a novel strategy for the construction of fluorescent agents to measure plasma membrane polarity by conjoining twisted intramolecular charge transfer (TICT) modulation and charge number control. It is shown that compounds with a stronger TICT tendency are more sensitive to polarity shifts due to the number of dialkylated amino groups present (from 1 to 3), and the molecules with two or more charged centers remain in the plasma membrane. Therefore, we developed two fluorescent agents with high polarity sensitivity, excellent turn-on ratios, and superior ability, to target the plasma membrane. In the wash-free fluorescence imaging and fluorescence lifetime tests, our designed agent could detect plasma membrane polarity with high precision, allowing effective distinction between cancer cells and normal cells based on their differences in plasma membrane polarity. Moreover, both fluorescence and fluorescence lifetime changes of the plasma membrane in the ferroptosis model established by Sorafenib confirmed an increase in plasma membrane polarity during cell ferroptosis. Download figure Download PowerPoint Introduction Plasma membranes, as fundamental and necessary subcellular structures, influence most biological processes.1,2 They act as a semipermeable barrier to defend the cell and allow bidirectional transport of specific molecules between the intracellular and extracellular environments. However, they also act as sophisticated organelles to organize and tune receptor molecules.3–5 Thus, the aberrant changes in the plasma membrane can cause cellular malfunctions, resulting in a wide range of human diseases such as cystic fibrosis, Alzheimer's disease, and in particular, malignancies.6,7 Recent studies have demonstrated that cancer cells possess higher levels of cholesterol for plasma membrane biogenesis due to their abnormal metabolism and rapid proliferation compared to normal cells.8,9 Since cholesterol with relatively low polarity is mainly enriched in the plasma membranes,10 it can be assumed that cell carcinogenesis can result in a variation in plasma membrane polarities. In addition, the structures and compositions of cancer cells' membranes frequently change upon treatment with anticarcinogens,11–13 accumulating in a shift in membrane polarity. Hence, membrane polarity monitoring has the potential to easily comprehend the behavior of plasma membranes during various biological processes. Currently, the majority of clinical imaging modalities, such as magnetic resonance imaging (MRI), computed tomography (CT), and positron emission tomography (PET), can only image at tissue level, however, plasma membrane polarity remains elusive with these techniques.14,15 Due to the superior sensitivity and selectivity, ease of manipulation, and high spatiotemporal resolution, fluorescence imaging techniques have shown great promise for the measurement of pathological microenvironments (e.g., pH, temperature, viscosity, and polarity).16–31 Unfortunately, very few small-molecule fluorescent agents that are polarity-sensitive and plasma membrane-specific have been reported ( Supporting Information Table S6).32–38 In 2022, Feng's group35,39 described the construction of fluorescence agents for plasma membrane polarity detection by associating plasma membrane-targeted moieties and polarity-sensitive fluorophores (Scheme 1a). However, it was difficult to further improve the performance of plasma membrane anchoring and the polarity-sensitivity of such agents due to the fixed structures of the targeted moieties and the sensitive fluorophores. Inspired by these works, we, therefore, wondered if there is a new strategy for the preparation of plasma membrane-specific and polarity-sensitive fluorescent scaffolds simultaneously. Scheme 1 | A general approach assessment of the designed plasma membrane-specific and polarity-sensitive fluorescent agents. Download figure Download PowerPoint Design Planning Here, we examined how to improve the polarity sensitivity and plasma membrane anchorage ability. Earlier reports have utilized the twisted intramolecular charge transfer (TICT) mechanism for the development of fluorescence agents for the purpose of detecting polarity in microenvironments.40–43 In general, nonemissive TICT states preferred to form in polar solvents, hence, polarity reduction suppressed the formation of TICT states and restored fluorescence emission (Scheme 1b).43 However, the majority of biocompatible fluorophores such as coumarin, naphthalimides, and rhodamine derivatives (cf. Scheme 1b) exhibited a poor polarity response owing to their weak TICT effects.42 Therefore, constructing fluorophore scaffolds that exhibit strong TICT effects remains a challenge, but is an effective approach to improve polarity sensitivity. (E)-4-(2-(5-(4-(diphenylamino)phenyl)thiophen-2-yl)vinyl)-1-methyl-pyridin-1-ium (TTPy) and (E)-4-(2-(5-(4-(diphenylamino)phenyl)-thiophen-2-yl)vinyl)- 1-(3-(trimethylammonio)propyl)pyridin-1-ium (TTVP) are well-known fluorescence dyes, which possess comparative structures but with unique organelle anchoring properties (Scheme 1c).44 Importantly, TTPy can target the mitochondria, whereas TTVP is able to anchor in the plasma membrane. However, the main contrast between these compounds is that TTVP contains an additional positive charge. Therefore, if anchoring to the plasma membrane with fluorescence agents can be enhanced with increasing charge-number, we hypothesized that the charge-number control could be a new avenue for the design of plasma membrane-specific fluorescent scaffolds. More importantly, multicharge control and TICT modulation could work in unison to create a plasma membrane-specific and polarity-sensitive fluorescence scaffold. The most crucial technical point of our approach is the design of a minimal fluorophore with both a positive charge and twisted electron donating group (EDG). This type of fluorophore can be changed readily via covalent bonding to increase both the charge number and TICT tendency. Our research group has previously employed anthocyanidin derivatives as fluorescence dyes to detect microenvironments.45–47 Importantly, as shown in Scheme 1d, these anthocyanidin-based dyes frequently contain positive charge and flexible modified structures that can link twisted EDG (such as alkylated amino groups). To achieve our goal, one anthocyanidin derivative with a dialkylated amino group was chosen as the twisted EDG for the construction of a multi-branched fluorescence scaffold and controlling both TICT tendency and the charge numbers (Scheme 1d). We successfully designed a series of anthocyanidin derivative-based branched fluorescent scaffolds (referred to as M1, M2, and M3) with benzene as the central core (Figure 1a). We found that the TICT tendency and charge number of the prepared fluorescent scaffold were enhanced by increased branching, which improved the plasma membrane anchorage ability and polarity sensitivity. Surprisingly, both M2 and M3 showed more than 60-fold increase in fluorescence response towards polarity, and the plasma membrane anchorage of M2 and M3 was significantly superior to that of the commercially available plasma membrane dye, 3,3′-dioctadecyloxacarbocyanine perchlorate (DiO). Moreover, additional benefits such as great biocompatibility and wash-free imaging ability were observed for M2 and M3 (Figure 1b). Figure 1 | (a) Molecular structures of newly designed fluorescent agents, referred to as M1, M2, and M3. (b) Key requirements of plasma membrane-specific and polarity-sensitive fluorescent agents for fluorescence imaging. (c) Schematic diagram of application of M2 for plasma membrane polarity monitoring in diverse biological processes through fluorescence and fluorescence lifetime imaging. Fl., fluorescence. Download figure Download PowerPoint M2 was employed to specifically label cancer cells among normal cells via enhanced fluorescence and fluorescence lifetime signals, indicating that the plasma membrane polarity of the cancer cells was a latent biomarker due to the relatively low polarity compared to normal cells (Figure 1c). Additional information was obtained on plasma membrane changes in the cancer cells during therapy using Sorafenib (a typical anticarcinogen) to treat cancer cells and monitor plasma membrane polarity changes using M2. The results showed that Sorafenib treatment enhanced plasma membrane polarity (Figure 1c). Further investigations suggested that the increase in plasma membrane polarity stemmed from lipid peroxidation during ferroptosis. Our findings demonstrate that in conjunction with high-performance plasma membrane-specific and polarity-sensitive fluorescent agents, the real-time monitoring of plasma membrane polarity via the combination of fluorescence and fluorescence lifetime imaging (FLIM) allows for a greater understanding of plasma membrane behavior in diverse biological processes. Experimental Methods Synthesis M1, M2, and M3 were synthesized according to the operations in the Supporting Information and characterized by NMR and high-resolution mass spectrometry (HRMS). Computational methods Density functional theory/time-dependent density functional theory calculations Molecular geometry optimization, vibration frequency analysis, and single point energy were carried out within the density functional theory (DFT) framework using the Gaussian 09 software.48 The computational details are presented in the Supporting Information. Molecular dynamics All molecular dynamics (MD) simulations were performed using the GROMACS 2019.6 suite of programs.49 The computational details are presented in the Supporting Information. Solution polarity titration Polarity solutions were prepared by mixing methanol and dichloromethane (DCM) or 1,4-dioxane and methanol in different volume proportions. The magnitude of polarity is uniformly represented by the orientation polarizability (Δf), where the orientation polarizability corresponding to each solvent volume ratio, together with experimental details, are shown in Supporting Information Tables S3 and S4. Cytotoxicity assay We created suspensions of single cells from human hepatoma (HepG-2), human lung cancer (A-549), human fetal hepatocyte (LO2), human non-tumorigenic lung epithelial (BEAS-2B), and rat adrenal medulla neuroendocrine tumor (PC12) cell lines. Each cell type was inoculated into 96-well plates at 3000 to 5000 cells/well in Dulbecco's modified eagle medium (DMEM) culture medium and incubated for 12–24 h in an incubator set at 37 °C and 5% CO2. M1, M2, and M3 compounds were dissolved in dimethyl sulfoxide (DMSO) to obtain a stock solution of specific concentration, then an appropriate volume of each was added to the cell-containing 96-well plates to achieve final concentrations of 0, 1.25, 5, 10, and 20 μM/compound, respectively. The final volume of each well was 200 μL. Following a further 48 h of incubation, the culture medium was removed and replaced with 20 μL of the cell viability tetrazolium agent, MTS solution, and 100 μL of fresh culture medium, before being left to incubate for 3 h. After the MTS had fully reacted, the absorbance at 492 nm/well was recorded using a MULTISKAN FC microplate reader (Thermo Fisher Scientific, Shanghai, China). Results and Discussion Theoretical validations of enhanced TICT effects We hypothesized that the introduction of additional dialkylated amino groups through the branched structure could effectively enhance the TICT tendency of the molecule. Upon light irradiation, the dialkylated amino groups rotated ∼90 deg with respect to the fluorophore backbone, leading to a perpendicular alignment in the excited state.50–52 This type of rotation was typically accompanied by charge separation between the amino group and the fluorophore backbone, resulting in a nonemissive TICT state. Since the polar environment favored the stabilization of the TICT state, we envisaged that a decrease in polarity could inhibit the TICT state and recover the fluorescence emission.53 Therefore, enhancement of TICT tendency could improve the fluorophore's sensitivity towards polarity detection. To verify our hypothesis, compounds M1, M2, and M3 were prepared containing different amounts of anthocyanidin branches. Notably, in addition to dialkylated amino groups, the prepared compounds possessed other twisted structures quantified by various dihedral angles ( Supporting Information Figure S10). Since all the twisted structures could lead to TICT effects, we first considered whether the dialkylated amino groups were the main contributors to the TICT effects. It is well-known that TICT only occurs between an electron donor and electron acceptor; thus, time-dependent DFT (TD-DFT) calculations were performed to display the donor and acceptor parts in M1, M2, and M3.48 Hole-charge analysis was used to examine the first excited state (S1) of the prepared compounds,54,55 where the electrons were mainly distributed on the dialkylated amino groups and transferred to the hole (central benzene) ( Supporting Information Figures S11–S13). This clearly demonstrated that the dialkylated amino groups were the electron donors. More importantly, through geometry optimization of S1 states of M1, M2, and M3, we found that as the energy of S1 decreased, the dihedral angles (Δθ > 4°, cf. Supporting Information Figure S14 and Figure 2a–c) between dialkylated amino groups and fluorophore backbone remained relatively constant for twisted formation. However, the remaining dihedral angles were prone to change in planarization (Δθ > 9°, cf. Supporting Information Figures S14 and S15). These results further suggested that the dialkylated amino groups in M1, M2, and M3 were the key factors in the formation of the TICT state. Figure 2 | Evaluation of TICT tendency among M1, M2, and M3 via TD-DFT calculations. Variation of S1 energy with excited state geometry optimization of (a) M1, (b) M2, and (c) M3 by TD-DFT calculation at the PBE0/def2-TZVP/LR-SMD level in DCM, the illustrations display the excited state geometry and the amount of change in the associated angle (θ) during the excited state optimization. Hole-charge analysis of the first excited state of (d) M1, (e) M2, and (f) M3 with Multiwfn-3.8(dev). Blue and red isosurfaces represent hole and electron distributions, respectively. The oscillator strength (f) of the ICT to TICT state transition is labeled in the inset. Calculated S1 PESs of (g) M1, (h) M2 and (i) M3 at PBE0/def2-TZVP/LR-SMD level in methanol. The driving energy (Ede) for TICT is labeled in the inset. Download figure Download PowerPoint We next examined if the influence of the TICT state by the twist of the dialkylated amino groups could act as a nonirradiative channel, resulting in weak fluorescence emission of M1, M2, and M3. The oscillator strength (f) changes of M1, M2, and M3 in different rotation angles (θ) of the dialkylated amino group on the S1 potential energy surface (PES) were calculated ( Supporting Information Figure S16). As shown in Figure 2d–f, in typical intramolecular charge transfer (ICT) states (i.e., θ ≈ 0°), M1, M2, and M3 exhibited large oscillator strengths (f > 0.50), which suggested a bright fluorescence emission. In contrast, negligible oscillator strengths (f ≈ 0) were observed in the typical TICT states (i.e., θ = 90°) of the studied compounds, indicative of the weak emission. Subsequently, we explored the variation of TICT tendency among M1, M2, and M3. As depicted in Figure 2g–i, the amount of dialkylated amino groups increased in the order of M1 > M2 > M3 based on the calculated TICT driving energy (Ede, defined as the energy change from the ICT state to the TICT state), which increased progressively from −0.72 eV to −0.88 eV and −1.10 eV, respectively. The data obtained showed that the spontaneous transition from ICT state to TICT state in polar solvents was much easier for M2 and M3 compared to M1, hence, M2 and M3 had a stronger tendency to TICT state. Therefore, we deduced that the fluorescence response towards the polarity change of M2 and M3 was more sensitive due to their effective TICT formation with an increasing number of dialkylated amino groups. Improved polarity sensitivity through TICT modulation M1, M2, and M3 were prepared to validate that the enhanced TICT tendency was capable of improving the fluorescence responses toward polarity sensitivity. The synthetic routes are illustrated in Supporting Information Scheme S1, and the chemical structures of M1, M2, and M3 were characterized by 1H and 13C NMRs and high-resolution mass spectrometry ( Supporting Information Figures S1–S9). The maximum UV–vis absorption peaks of the compounds in DCM were located at 542, 561, and 548 nm, respectively (Table 1 and Supporting Information Figure S17). Depending on the hole–electron analysis ( Supporting Information Figure S18), the maximum absorption peaks of M1 and M2 were attributed to the local excitation (LE) from the ground state (S0) to S1, and that of M3 was ascribed to the accumulation effect of LE from S0 to both S1 and S2 due to the centrosymmetric structure ( Supporting Information Figures S11–S13 and Tables S7–S9). Table 1 | Photophysical Properties of M1, M2, and M3 Compd. λabs (nm)a εmax (L·mol/cm)a λem (nm)a Δvb (cm−1) ΦF (%)c ΦF (%)d ΦF (%) HOMO (eV)g LUMO (eV)g Eg (eV)h M1 542 32,210 615 2190 5.66 1.80 24.01e −6.58 −3.86 2.72 M2 561 58,220 635 2077 3.25 0.94 15.26e −6.66 −4.06 2.60 M3 547 50,290 625 2282 1.42 0.83 4.11f −8.89 −6.18 2.71 a10 μM in DCM containing 0.25% DMSO (v/v). bStokes shift in DCM. cFluorescence quantum yield in MeOH with rhodamine 6G as the standard. dFluorescence quantum yield in water with rhodamine 6G as the standard. eFluorescence quantum yield in DCM. fFluorescence quantum yield in dioxane. gCalculations with the B3LYP/def2-TZVP level using Gaussian 09. hThe energy difference between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO). The shoulder peak near the maximum absorption peak of M1, M2, and M3 was caused by the partial TICT state. Due to the enlarged conjugate structure, M2 (εmax = 58220 M−1cm−1) and M3 (εmax = 50290 M−1cm−1) had larger molar extinction coefficient compared to M1 (εmax = 32210 M−1cm−1, cf. Table 1 and Supporting Information Figures S19–S21), which could improve the utilization of absorbed light for fluorescence imaging. Furthermore, the fluorescence spectra of the compounds had a maximum absorption at 615, 635, and 625 nm, respectively (cf. Table 1 and Supporting Information Figure S22), and the red-shifted order of the absorption and emission peak matched their energy gap (Eg) values (i.e., M1 > M3 > M2, cf. Table 1 and Supporting Information Figure S23). Moreover, in a polar solvent (e.g., methanol and water), M2 and M3 displayed much lower fluorescence quantum yield (ΦF) than that of M1 (Table 1), suggesting a more efficient formation of nonemissive TICT state in M2 and M3, consistent with the calculated Ede. To investigate the fluorescence response towards polarity, the fluorescence emission spectra of M1, M2, and M3 were measured in various mixed solutions with the orientation polarizability (Δf, a parameter used for polarity expressing) from ≈ 0.31 to ≈ 0.22. As shown in Figure 3a, ∼ 47-fold enhancement in the fluorescence was attributed to M1 with decreased polarity. However, the observed fluorescence increase produced by M1 was markedly below that of M2 (63-fold, Figure 3b) and M3 (78-fold, Figure which that due to the stronger TICT M2 and M3 possessed superior sensitivity for polarity Figure 3 | (a) spectra of M1 in with volume The the peak fluorescence 615 of M1 as a of orientation polarizability from 0.31 to (b) spectra of M2 in with volume The the peak fluorescence 635 of M2 as a of orientation polarizability from 0.31 to (c) spectra of M3 in with volume The the peak fluorescence 635 of M3 as a of orientation polarizability from 0.31 to (d) turn-on of M1, M2, and M3 in low polarity and high (e) of M3 at 625 in the of various Download figure Download PowerPoint studies have reported that the TICT mechanism can also be in hence, the fluorescence of M1, M2, and M3 in a diverse of solvent with from to was examined ( Supporting Information Table In with the TICT tendency among the compounds the fluorescence enhancement of M2 and M3 was higher than that of M1 ( Supporting Information Figure As shown in Figure compared to polarity, the fluorescence response of the studied compounds towards was much their poor sensitivity for detection. To our M1, M2, and M3 exhibited great ability in the of diverse such as and (Figure and Supporting Information Figure as shown in Supporting Information Figures and the studied compounds exhibited and in the Therefore, M1, M2, and M3 have the potential for microenvironmental polarity detection. Theoretical validations of enhanced plasma membrane-targeted ability by charge number control To due to the of imaging techniques and the high sensitivity of chemical structures towards the intracellular of the the specific of subcellular organelles by fluorescence agents has been Herein, by the chemical structure of TTPy and we that the increased charge numbers in M2 and M3 improve their plasma membrane To hypothesis, simulations were performed by GROMACS 2019.6 suite of to investigate the plasma membrane processes of M1, M2, and M3. As shown in Figure a typical (i.e., and plasma membrane was to in the energy of M1, M2, and M3. The on only the model due to the plasma and the of the membrane was used as the with the = = was from lipid which possessed the lowest density and polarity. = was the between the and the of the = was between the lipid and the > mainly of the environment (e.g., extracellular Figure | (a) of the lipid membrane model and the partial density of the lipid membrane. The lipid membrane is into and labeled as and (b) The energy of M1, M2, and M3 the lipid membrane were obtained by The energy barrier of M1, M2, and M3 is labeled in the inset. (c) The of the of M1, M2, and M3 on the plasma membrane. Download figure Download PowerPoint To investigate the between various agents with different numbers of and the plasma membrane, M1, M2, and M3 were into the membrane model using the and then Subsequently, dynamics simulations were performed on the studied compounds, which from the membrane = to = The energy of the were obtained using during the and the from the to the was for each agent ( Supporting Information Figures As depicted in Figure the energy barrier increased from M1 to M2 and M3 indicating that M2 and M3 were more difficult to the membrane compared with Moreover, as shown in Figure and the energy of M1 and M2 in was lower than in water suggesting that they could the of the membrane the energy barrier of the groups. However, the energy of M2 in the of the membrane was much lower than that of M1, leading to a higher energy barrier of M2 in hence, M2 had the of the membrane. In contrast, the energy of M3 enhanced from to and only decreased in (Figure red suggesting that it was to the membrane due to strong and low to the M1 was capable of the membrane, whereas M2 was able to target the of the membrane of the of one more positive and M3 had the ability due to positive (Figure Experimental validations of enhanced plasma membrane-targeted ability by charge number control to the of plasma membrane we performed subcellular imaging using M1, M2, and M3, respectively. fluorescent agents with a positive charge to in the with membrane Hence, we first the ability of the studied compounds to to the results were using as the As depicted in Figure the fluorescence of M1 in the cells was consistent with that of with a high coefficient = that M1 the plasma membrane and targeted the In of M2 = and M3 = with were relatively which that they could specifically target the To