Simultaneous Detection of Local Polarizability and Viscosity by a Single Fluorescent Probe in Cells

Abstract

Many intracellular reactions are dependent on the dielectric (“polarity”) and viscosity properties of their milieu. Fluorescence imaging offers a convenient strategy to report on such environmental properties. Yet, concomitant and independent monitoring of polarity and viscosity in cells at submicron scale is currently hampered by the lack of fluorescence probes characterized by unmixed responses to both parameters. Here, the peculiar photophysics of a green fluorescent protein chromophore analog is exploited for quantifying and imaging polarity and viscosity independently in living cells. We show that the polarity and viscosity profile around a novel hybrid drug-delivery peptide changes dramatically upon cell internalization via endosomes, shedding light on the spatiotemporal features of the release mechanism. Accordingly, our fluorescent probe opens the way to monitor the environmental effects on several processes relevant to cell biochemistry and nanomedicine.

Introduction

The cell milieu hosts thousands of biochemical processes that differ in location, number, and kind of molecules. Yet, some physicochemical features play regulative roles irrespective of the specific nature of any biochemical reaction. Among these, electrostatic and viscosity properties of the local environments stand out as the most general ones (1). The cell machinery arranges biochemical processes in cell compartments characterized by strong differences of local polarity. Examples are the hydrophobic lipid bilayer of membranes and the hydrophilic cytoplasmic regions (2, 3). As for viscosity, diffusion-mediated cellular processes are related to the hydrodynamic properties of local media. Several experiments demonstrated that, depending on the nature of the intracellular compartment, local viscosity may increase by 3–4 orders of magnitude with respect to pure water (4, 5, 6, 7). The key roles of polarity and viscosity in determining the reactive complexity of the cell have recently prompted intensive research on bioanalytical methods capable of reporting on these physicochemical properties at the nanoscale. In this context, the use of fluorescent probes coupled with high-resolution fluorescent microscopy of living cells stands out as one of the most promising approaches. Indeed, the characteristic long lifetime of electronic excited states (from a few picoseconds to nanoseconds) makes fluorescence exquisitely sensitive to changes in the local (nano) environments. Accordingly, many fluorescent polarity and viscosity probes have been employed to image the cell environment and have contributed to the elucidation of biochemical mechanisms (8, 9, 10, 11, 12, 13, 14, 15). Most probes, however, are characterized by photophysical processes of their excited states (e.g., twisted intramolecular charge transfer) that lead to complex and intertwined responses to polarity and viscosity (16, 17, 18). In most cases, this allows for detection of only one parameter at a time, usually by monitoring it in restricted cellular media where the other is rather constant (6, 19, 20). Nonetheless, the simultaneous and independent detection of both properties by a “dual” environmental probe, i.e., a fluorophore characterized by fully decoupled optical responses to polarity and viscosity, has so far represented a long-term goal in the field of fluorescent molecular imaging (21). In this work, we describe the first, at least to our knowledge, dual sensor of polarity and viscosity for intracellular use, Ge1 (Fig. 1).

Figure 1.

Molecular structure of Ge1 and the bioconjugable analog Ge1a.

Ge1 responds to polarity through emission red-shift and to viscosity by increasing its fluorescence lifetime. The independent responses to these physicochemical properties stem from a peculiar photophysical mechanism that decouples the energy of emission from the evolution kinetics of excited state(s). Notably, Ge1 is obtainable in a functional form, Ge1a (Fig. 1), that is easily conjugable to biochemical moieties, such as lipids or peptides.

To demonstrate its broad range of application in the cell context, Ge1 has been targeted to intracellular endosomes by means of engineered peptides, and there it reported environmental changes related to endosomal degradation. By this approach, we have been able to monitor in real time the molecular processes of a novel, to our knowledge, drug-delivery strategy recently proposed by some of us.

Materials and Methods

Peptide synthesis, purification, and labeling

All peptides were prepared by solid-phase synthesis using fluorenylmethyloxycarbonyl chemistry on an automatic Liberty-12-Channel automated peptide synthesizer with an integrated microwave system (CEM Corporation, Matthews, NC). The crude peptides were purified by reversed-phase high-performance liquid chromatography (HPLC) on a Jupiter 4 m Proteo 90 A column (250 × 10 mm) (Phenomenex, Torrance, CA). The HPLC analysis and purification were performed on a Dionex Ultimate 3000 HPLC system with autosampler. The purity of the product was confirmed by electrospray mass spectroscopy. The cysteine residue added to the C-terminus of each peptide provided a sulfhydryl group for further ligation to the Ge1-maleimide fluorophore. The labeling of purified peptides was performed by incubating for 3 h with a threefold molar excess of Ge1-maleimide, 150 mM PBS buffer, tris(2-carboxyethyl)phosphine, at pH 7.4. Finally, Ge1 peptides were purified by HPLC and then lyophilized overnight. The molecular weight of all conjugated peptides was confirmed by electrospray mass spectroscopy, and the concentration of each peptide stock solution was verified by ultraviolet-visible absorbance. The electrospray ionization mass spectra of the peptides were obtained with an API3200QTRAP hybrid triple quadrupole/linear ion trap (AB SCIEX, Foster City, CA). Peptides were stored at −80°C.

Time resolved measurements

Time-resolved absorption spectra (TRAS) measurements were performed using a home-built femtosecond pump-probe spectrometer. A titanium-sapphire regenerative amplifier (Coherent, Santa Clara, CA) was used as laser source, delivering 100 fs pulses at a central wavelength of 800 nm with 4 mJ pulse energy at a repetition rate of 1 kHz. The 400 nm pump pulses were obtained by second-harmonic generation in a β-barium borate crystal. A white light continuum generated in a 2-mm-thick sapphire plate was used for probing in the visible from 450 to 750 nm. The probe transmission through the sample is measured with an optical multichannel analyzer working at the full repetition rate of the laser source. The acquisition of the pump-perturbed and pump-unperturbed probe spectra allowed extraction of the differential absorbance (ΔA) spectrum. At the specific pump-probe delay τ, it is calculated as: ΔA(λ,τ) ≈ −[I_ON(λ,τ)-I_OFF(λ)]/I_OFF(λ), where λ is the probe wavelength and I_ON(I_OFF) is the transmitted probe intensity with(without) the pump pulse. The overall temporal resolution of the system is ∼150 fs (22). The pump-probe measurements were analyzed with a home-built software based on the model by van Stokkum et al (23). A sequential model with three components was selected to fit the data and disentangle the different components. The analysis fitted the instrumental response function of the experiment and probe pulse chirp.

Fluorescence imaging and lifetime measurements

Fluorescence imaging and lifetime measurements were performed by means of a Leica TCS SP5 SMD inverted confocal microscope (Leica Microsystems AG, Wetzlar, Germany) equipped with an external pulsed diode laser for excitation at 405 and 470 nm and a time-correlated single-photon counting acquisition card (PicoHarp 300; PicoQuant, Berlin, Germany) connected to internal spectral detectors. The images were collected using low excitation power at the sample (10–20 μW), and the laser repetition rate was set to 40 MHz. The pinhole aperture was set to 1.0 Airy. Solutions of Ge1 (1–5 μM) in different solvent mixtures were placed into glass bottom cuvettes while Chinese hamster ovary (CHO) K1 cells were treated with Ge1 and Ge1-peptides in glass-bottom WillCo dishes (WillCo Wells, Amsterdam, the Netherlands) and observed after 15 min and 1 h incubation, respectively, without washing. All the sample were mounted in a thermostated chamber at 37°C and viewed with a 43× or 100×1.5 NA oil immersion objective (Leica Microsystems). Image size was 256 × 256 (solutions) or 512 × 512 (cells) pixels. The emission was monitored in the 480–525 nm and 540–580 nm ranges using the built-in acousto-optical beam splitter detection system of the confocal microscope. The two acquired ranges allowed us to evaluate the generalized polarization (GP) while the lifetime analysis was performed on the joined channels.

Results and Discussion

Ge1 dual-sensing properties

Representative steady-state absorption and fluorescence spectra of Ge1 (structure in Fig. 1) in different solvents are reported in Fig. 2 a. Regardless of solvent polarity, Ge1 absorption peaks in the narrow 420–430 nm range. On the contrary, Ge1 fluorescence displays significant solvatochromism from 490 up to 560 nm. The emission band is structured and blue-shifted in apolar media, whereas increasing medium polarity leads to a smoother and red-shifted band. The emission shoulder visible at high energies in apolar media is assigned to a vibrational subband because absorption and excitation spectra almost overlap (Fig. S1). Low temperature prevents solvent dipolar relaxation around Ge1 and therefore prevents red-shift of the emission, as visible by comparing the spectra in polar dimethylformamide (DMF) at 298 and 77 K (full light gray and dotted black spectra in Fig. 2 a).

Figure 2.

Steady-state optical properties of Ge1. (a) Absorption (dashed lines) and emission (full lines) spectra of Ge1 in cyclohexane (black), dichloromethane (gray), and dimethylformamide (light gray) at 298 K are shown. The dotted black spectrum refers to the emission in dimethylformamide at 77 K. (b) Fluorescence emission time decay at different viscosities is shown.

We previously demonstrated (3) that the solvatochromic behavior of Ge1 emission is associated with a linear relationship (Fig. 3 a) between the logarithm of the local dielectric constant and the GP calculated from the fluorescence intensities that were collected in the “hydrophobic” range 480–525 nm (F_L) and the “hydrophilic” range 540–580 nm (F_H), as follows:

$$GP=\frac{F_L-F_H}{F_L+F_H}\text{.}$$ (1)

Significantly, this linear trend holds regardless of the viscosity properties of the medium, representing a general property of the dye (Fig. 3 a).

Figure 3.

Polarity and viscosity dependence of the fluorescence response of Ge1. (a) Polarity plot: the generalized polarization (GP) versus dielectric constant of the medium is shown: Pearson’s correlation coefficient ρ = cov(X,Y)/(σ_{X ×}σ_Y), where σ_i = standard deviation of all the solutions ρ = −0.97. (b) Rigidochromic log-log (Förster-Hoffmann) plot: the logarithm of average fluorescence lifetime versus logarithm of viscosity is shown: ρ_{pure solvent} = −0.24 and ρ_{high viscosity} = 0.95. (c) GP versus logarithm of viscosity is shown: ρ_{pure solvent} = −0.22, ρ_{high viscosity} = 0.74. (d) Logarithm of average fluorescence lifetime versus dielectric constant is shown: ρ_{pure solvent} = −0.88 and ρ_{high viscosity} = −0.83. Black full circles: pure solvents are shown. Black full triangles, squares, and diamond: isoviscous mixtures are shown (corresponding to those enclosed in the black empty square of (b) with the same symbol code).

Because GP is a concentration-independent (“ratiometric”) optical response, its dependence on dielectric constant can be used to monitor local polarity in systems (e.g., living cells) in which the control over the dye concentration is unfeasible.

Time-resolved fluorescence analysis of Ge1 emission highlights monoexponential decays in both highly apolar cyclohexane and strongly H-bonding methanol, whereas a biexponential decay is observed in all other media (Fig. 2 b;Table S1). In low-viscosity solvents (η < 3.5 cP; see Table S1), the average lifetime (‹τ›) ranges from 0.49 to 2.58 ns. The rigidochromic properties of Ge1 are witnessed by the Förster-Hoffmann (F-H) double logarithmic plot reporting the dependence of ‹τ› on the solvent viscosity η (Fig. 3 b). Remarkably, the F-H plot displays a strict linear behavior for viscosity above 3.5 cP with a negligible dependence on the medium polarity with the exception of strongly H-bonding solvents (as identified by the Kamlet-Taft H-bond donor parameter (24), α > 0.8 (Table S2)). Analogously to GP, ‹τ› does not depend on the dye concentration, and the observed general F-H dependence can be exploited to monitor local viscosity in systems where the amount of Ge1 cannot be controlled, such as in living cells.

The determination of polarity by GP is fully independent of local viscosity (Fig. 3 c). Instead, ‹τ› displays negative correlation with the dielectric constant. Yet the strength of ‹τ› versus Log(ε), as measured by the slope m obtained by linear fitting, is dramatically related to the viscosity range. At very low viscosity, m = −0.135 (Fig. 3 d, black dashed line), whereas for η ≥ 3.5, the slope reduces to m = −0.026 (Fig. 3 d, gray dashed line). Thus, changes of ε in the meaningful interval 1–50 [Log(ε) = 0–1.7] affect to a very low extent the lifetimes of viscous mixtures, supporting the strong decoupling of polarity from viscosity highlighted in Fig. 3 b. This effect makes this dye the first example—at least to our knowledge—of a dual indicator of local polarity and viscosity, i.e., a probe whose calibrations from these physicochemical parameters are universal and mutually independent from each other.

Photophysical origins of the dual-sensing property

We studied the photophysical origin of this unique dual-sensing behavior by time-resolved spectroscopy in media of different polarity and viscosity. At first, time-resolved emission spectra were acquired from 100 ps to 25 ns (Fig. S2, a–c), and data were globally fit to a sum of exponential functions to obtain the relevant decay-associated spectra. In low-viscosity solvents, two states S_1f and S_1s were identified. S_1f decays faster to the ground state, is more intense, and slightly blue-shifts with respect to its longer-lived counterpart S_1s (Fig. S2, d and e). Across the whole wavelength range, S_1f and S_1s never display negative values, thereby ruling out the nanosecond-scale evolution of S_1f into S_1s or vice versa. Significantly, the Stokes’s shifts of S_1f and S_1s are linearly related to the orientational polarizability of the solvent, following a strict Lippert-Mataga behavior (Fig. S3). Linear fitting of the Lippert-Mataga reveals that S_1s possesses an electric dipole moment 16% larger than that of S_1f and that for low polarity solvents, such as cyclohexane, S_1s and S_1f merge into a single emitting state. In viscous solvents, the Stokes’s shifts of S_1f and S_1s fully comply with the Lippert-Mataga trend found in low-viscosity media. Yet, decay-associated spectrum analysis reveals an additional very fast (200–400 ps) component (S_1vf), peaking at 490–500 nm (similarly to Ge1 in cyclohexane) and displaying negative amplitudes at longer wavelengths with an overall intensity integral of about zero (Fig. S2 f). The latter property explains the undetectability of S_1vf when the whole 480–700 nm emission is collected. Notably, the relative amplitude and lifetime of S_1s grow linearly with log(η) (Fig. S4, a and b), whereas the lifetime of S_1f is weakly affected by viscosity (Fig. S4 b).

To provide a further insight into the emission mechanism of Ge1, TRAS were collected in dichloromethane (DCM), DMF, and PEG400 across the 150 fs – 1 ns time window (Fig. S5). Consistently with Ge1 solvatochromism, the significant red-shift of the stimulated emission band in polar DMF as compared with that of apolar DCM (Fig. S5, a and b) indicates that in nonviscous media, the locally excited (LE) state of Ge1 undergoes a classical solvent dipolar relaxation in the first 2–3 ps of excited-state lifetime. As expected, in the more viscous polar PEG400, dipolar relaxation occurs on a longer timescale (20 ps) (Document S1. Supporting Materials and Methods, Figs. S1–S8, and Tables S1 and S2, Document S2. Article plus Supporting Material). After this first kinetic step, TRAS highlights a further evolution of the excited state. Global fitting (23) of TRAS identifies two additional components: one with a characteristic lifetime of 60–120 ps (the longest lifetime was observed in PEG400) and the other lasting longer than 1 ns and identifiable with a mixture of S_1s and S_1f, nonresoluble due to the limited timespan of the TRAS measurements (Fig. S5, d–f).

Overall, these findings may be rationalized by a simple photophysical model (Fig. 4). In nonviscous solvents, the Franck-Condon excited state undergoes intramolecular vibrational relaxation followed by dipolar relaxation until a solvent-relaxed LE state is reached (2–3 ps). Then, the relaxed LE bifurcates into two solvent-relaxed, fluorescence-emitting excited states (S_1f and S_1s) that possess slightly different electronic and conformational characteristics. The kinetic process favors S_1f, which displays larger abundance than does S_1s. Increasing viscosity has a number of effects. First, it inverts the relative populations of S_1f and S_1s. Under the reasonable assumption that viscosity always slows down excited-state molecular processes (negligibly affecting the radiative rate), this finding indicates that viscosity hampers more the formation of S_1f than S_1s in Ge1, suggesting S_1f to be characterized by some internal twisting with respect to the LE state and S_1s. Second, it slows down the decay of S_1s to ground state according to a classical F-H behavior. This suggests that intramolecular twisting of S_1s promotes efficient internal conversion to the ground state. Third, the dipolar relaxation of solvent around LE does not take place in a few picoseconds, yielding a further fast-emitting state S_1vf that can be identified with partially unrelaxed LE on account of its emission spectral fingerprint. Yet, S_1vf has almost no effect on the global Ge1 emission.

Figure 4.

Schematic representation of the photophysical behavior of Ge1.

The viscosity-dependent bifurcation of LE into two emissive states is reminiscent of the photophysical performance of some boron-dipyrromethene rotors (25) and accounts for the polarity independence of the rigidochromic behavior of Ge1. At odds with the boron-dipyrromethene rotors, however, the energies of S_1f and S_1s depend on medium polarity, and the global emission spectrum can be independently used to probe the dielectric properties of the environment.

Dual sensing of polarity and viscosity in the cell milieu

As a preliminary test of Ge1 ability to report on polarity and viscosity in an independent way in a cellular context, we administered Ge1 to CHO cells (Fig. S6). As already reported (3), Ge1 diffuses into cells and stains the plasma membrane (PM) and the endoplasmic reticulum (ER). From the polarity calibration, we obtained ‹ε›(PM) = 4.9 ± 0.1 and ‹ε›(ER) = 7.4 ± 0.2 (Document S1. Supporting Materials and Methods, Figs. S1–S8, and Tables S1 and S2, Document S2. Article plus Supporting Material). These dielectric properties indicate that Ge1 localizes well within the lipid bilayers of the PM and ER. Notably, these two compartments are strongly different in terms of local viscosity sensed by Ge1. We found ‹η›(PM) = 7000 ± 2500 cP and ‹η›(ER) = 200 ± 60 cP, respectively (Fig. S6 f). This dissimilarity could reflect the larger fraction of rigid lipids present in the PM as compared to the ER (26). We should also note that ‹η›(ER) is in good agreement with other literature data (27, 28, 29, 30), whereas our measurements would indicate that the PM is more rigid than what is reported by some authors (4).

Next, we applied a bioconjugated form of Ge1 to follow a biological process of relevant interest for nanomedicine applications. Recent approaches to intracellular delivery of drugs explored their bioconjugation with cell-penetrating peptides (CPPs) on account of the ability of CPPs to cross the PM and improve drug availability at the intracellular target. Among CPPs, the arginine-rich motif of Tat protein from human immunodeficiency virus 1 (sequence: YGRKKRRQRRR) is particularly effective in promoting PM penetration (31). Yet, Tat₁₁-linked cargo molecules typically end up sequestered into endosomal vesicles conveyed to metabolic degradation, thus hampering the application of Tat₁₁ for delivery purposes. To tackle this issue, some of us recently proposed (32, 33) to combine the well-known CPP uptake property of Tat₁₁ motif to the membrane-disruptive (concentration-dependent) ability of a linear cationic α-helical antimicrobial peptide belonging to the “CM” series (cecropin A and melittin hybrids). Indeed, several studies showed that CM hybrids efficiently disrupt the bacterial outer membrane (34). CM hybrids appear to trigger a detergent-like “carpet” mechanism (35, 36, 37) or the formation of discrete pores that dissipate ion gradients (38, 39, 40).

Starting from existing data on shorter variants (41), the CM₁₈-Tat₁₁ hybrid (CM₁₈ sequence: KWKLFKKIGAVLKVLTTG, C(1–7)M(2–12)) has been developed as a good means of disrupting endosomal membranes when linked with Tat₁₁. Previous data on fluorescently labeled CM₁₈-Tat₁₁ highlighted the effective release of the fluorescent cargos from the endosomes into the cell cytoplasm (32). Accordingly, the chimera CM₁₈-Tat₁₁ was conjugated to the functional analog of Ge1, which is Ge1a (Fig. 1), yielding CM₁₈-Tat₁₁-Ge1 (Fig. S7). As benchmarks of the chimera behavior, we also synthesized Tat₁₁-Ge1 and CM₁₈-Ge1 derivatives. All these peptides were then tested in CHO cells. Polarity and viscosity were monitored after 1 h of incubation when the peptide accumulation on the PM and in the endosomal vesicles reached a steady state. Fig. 5 shows the transmission (Nomarski) and fluorescence images as well as the polarity and viscosity maps for each Ge1-linked peptide.

Figure 5.

Administration of Ge1 peptides to CHO cells. (a) Transmission, fluorescence, log(ε) and log(η) maps of cells treated with CM₁₈-Ge1 (upper panels), CM₁₈-Tat₁₁-Ge1 (medium panels), and Tat₁₁-Ge1 (lower panels) are shown. (b) A zoomed region (white squares in (a)) is shown of cells treated with CM₁₈-Tat₁₁-Ge1. To see this figure in color, go online.

CM₁₈-Ge1 sensed the lowest value of the dielectric constant (‹ε› = 6.55 ± 0.10, 34 cells) and a very high local viscosity (‹η› = 1740 ± 770 cP, 34 cells). These data suggest that CM₁₈-Ge1 binds to the PM and endosomes leading to the intercalation of the fluorophore in the inner region of their bilayers (Fig. 6 a). On the contrary, Tat₁₁-Ge1 reported a significantly higher polarity (‹ε› = 10.02 ± 0.07, 43 cells) and a much lower viscosity (‹η› = 185 ± 20 cP, 43 cells). These findings indicate that Tat₁₁-Ge1 interacts mostly with the surface of the bilayer (Fig. 6 b). This model was further confirmed by molecular dynamics simulations of Tat₁₁-Ge1 and CM₁₈-Ge1 inside the bilayer (Fig. S8). Molecular dynamics data show that the carboxymethyl tethers of Tat₁₁-Ge1 and CM₁₈-Ge1 adopt considerably different positions within the bilayer. For Tat₁₁-Ge1, two stable positions of the carboxymethyl function minimize the free energy: one near the cholesterol hydroxyls and the other deep within the bilayer. On the contrary, CM₁₈-Ge1 displayed a preferential positioning of the carboxymethyl tether in the central part of the bilayer. Note that Tat₁₁ peptide located itself preferentially on top of the hydroxyl groups of the membrane, whereas CM₁₈—in agreement with its more hydrophobic character—inserted within the bilayer.

Figure 6.

Schematic representation of the action mechanism of Ge1 peptides. (a) Tat₁₁-Ge1, (b) CM₁₈-Ge1, (c) CM₁₈-Tat₁₁-Ge1 near the plasma membrane, and (d) CM₁₈-Tat₁₁-Ge1 in the internal vesicles are shown. The average of 80–100 normalized-distance profiles of (e) log(ε) and (f) log(η) radially collected (as shown in the graphic scheme of the cell) between the plasma membrane (PM - 0) and the nucleus (N - 1) for cells treated with Tat₁₁-Ge1 (gray squares), CM₁₈-Ge1 (black circles), and CM₁₈-Tat₁₁-Ge1 (light gray triangles) is shown.

Remarkably, the chimera Tat₁₁-CM₁₈-Ge1 detected a different polarity for the PM or endosomes close to it (‹ε› = 7.31 ± 0.04, 44 cells) with respect to endosomes located in the inner part of the cell (‹ε› = 9.15 ± 0.07). The average viscosity sensed by Tat₁₁-CM₁₈-Ge1 in the PM or the endosomes near the PM was similar to that reported by CM₁₈-Ge1 (‹η› = 2040 ± 100 cP), whereas the value detected in the internal vesicles was significantly much lower (‹η› = 90 ± 14 cP). Comparison of these real-time measurements with the behavior of benchmark peptides added an intracellular “spatial” coordinate to the “detergent-like” model of membrane disruption by Tat₁₁-CM₁₈ previously advanced by some of us (32, 33, 42). In the initial stage of the interaction of the chimera peptide with the membrane, i.e., when still embedded in the PM or in the endosomes spatially close to the PM, the low polarity and the high viscosity sensed by Ge1 were similar to the membrane intercalation of CM₁₈-Ge1 (Fig. 6 c). Note that the absolute values of ‹ε› and ‹η› calculated using our peptide-conjugated Ge1 cannot be considered, at this stage, to be true representations of the actual values found in cells; indeed, the solubility properties of peptide-conjugated probes did not allow the obtaining of accurate calibration curves in organic solvents of different polarity and viscosity. However, we want to emphasize that the observed variations of ‹ε› and ‹η› clearly demonstrate that a specific disruption mechanism is taking place. The higher polarity and lower viscosity sensed in late endosomes, which can be found in the inner cell, comply with the fluidification of the membrane and the dissolution of its structural integrity by a concentration effect of the chimera (Fig. 6 d). The observed spatial dependence of both polarity and viscosity along a generic line starting from the PM and ending at the nuclear periphery explains well the effect described above (Fig. 6, e and f). In short, our findings locate the onset of CM₁₈ detergent-like activity at a delayed stage of the uptake process, probably involving late endosomes of the inner part of the cell. Of note, this observation corroborates for the first time—to our knowledge—the hypothesis proposed by some of us based on indirect evidence (32, 33) that CM₁₈ is able to reach its critical membrane-perturbing concentration during its intracellular vesicular trafficking and not at the level of the PM, a key requisite for the successful application of such a system for delivery purposes.

Conclusion

To our knowledge, we have reported here a novel fluorescent molecular probe suitable for quantifying and imaging polarity and viscosity at the same time in living cells. The responses to the two parameters as well as the spectroscopic readouts recorded during the measurement (fluorescence intensity and lifetime) are fully decoupled from each other. A set of time-resolved measurements helped to outline the photophysical model underlying the peculiar behavior of the probe. The efficacy of the totally decoupled responses has been illustrated by quantitative measurements of local properties at the interface between probe-functionalized CPPs and the endosomal vesicle membrane. The accurate knowledge of polarity and viscosity adds significant information to better understand the dynamical process occurring during the entrance of such useful drug delivery vectors.

The peculiar behavior makes our probe suitable for a number of applications. Besides its use in monitoring the endocytic internalization of drug-delivery systems, as presented here, we may envisage Ge1 to monitor the transition properties of self-assembling materials, such as micelles and sol-gel systems, based on macromolecular binding (e.g., formation of stereocomplexes; see, for instance, (43). In the cell context, Ge1 could be applied to monitor changes in the viscosity properties of the endothelial cell membrane when it is subjected to the variation of shear forces (9). Experiments in both fields are currently under course.

Author Contributions

G.A., G.S., and R.B. designed the research. G.A., D.P., B.S., F.C., F.S., R.N., G.S., and R.B. performed the research. G.A., D.P., D.V., F.C., F.S., R.N., G.S., and R.B. analyzed the data. G.A., D.P., G.C., F.C., G.S., and R.B. wrote the manuscript.

Editor: Jason Swedlow.