*2.2. Optical Characterization*

The probe 3HF18 was analyzed by UV-Vis spectrophotometry and spectrofluorimetry in solvents with different polarity (Table 1, Figures 2 and 3). The UV-Vis spectra show absorption maxima between 400 nm and 450 nm, with a redshift of more than 50 nm and an increase in absorbance as the polarity of the solvent increases. Emission spectra confirm the solvatochromic property of the probe in emission rather than in absorption. The probe has been shown to detect even slight changes in the polarity of the environment. In polar solvents both N\* and T\* are energy stabilized and this stabilization is more pronounced for N\* due to the higher electrical dipole moment. For very polar protic solvents, N\* is at energy lower than T\*, and therefore only one transition is observed (Figure 3). Moreover, as reported in the literature for analogous compounds, the probe's core is poorly soluble and nonfluorescent in water [23]. The solubility increases in water/methanol mixtures. In these cases, the fluorescence spectra show a single fluorescence peak around 520 nm (Figure S5 in Supplementary Materials). As can be observed in Figure 3, the maximum emission ranges from 475 nm in dioxane to 551 nm ethylene glycolfor the N\* band. While the T\* band, located at longer wavelengths, shows almost no influence with the solvent polarity because the stabilization effect of polar solvents acts on both T and T\* states. The IN\*/I T\* ratio represents the dependence of the fluorescence intensity on the solvent polarity. It is possible to observe an increase in the intensity of fluorescence, which accompanies the redshift, changing from apolar solvents such as dioxane to a single band (N\* state) in protic polar solvents such as methanol, propan-2-ol and ethylene glycol.

**Table 1.** Absorption and emission data of 3HF18 in solvents with different polarities.


<sup>1</sup> Maximum absorption wavelength; <sup>2</sup> Molar extinction coefficients, from Lambert-Beer Law; <sup>3</sup> A single N\* peak is observed; <sup>4</sup> The T\* band, poorly resolved, appears as a shoulder; <sup>5</sup> The probe is poorly soluble in water. For the emission spectra, all the solutions were excited at 398 nm.

**Figure 2.** Absorption spectra of 3HF18 in solvents of different polarity.

**Figure 3.** Emission spectra of 3HF18 in solvents of different polarity, excited at 398 nm.

In Figure 4, a picture of different solutions of the probe 3HF18, in representative solvents with increasing polarity is reported, under white (a) and UV (b) light.

**Figure 4.** Picture of 3HF18 dissolved in different representative solvents: (1) Dioxane, (2) DCM, (3) MeOH, (4) H2O: (**a**) under white light and (**b**) under UV light (365 nm).

From the image in Figure 4, the solvatochromism of the 3HF18 molecule is visible to the naked eye. The emission color of the solutions ranges from orange to yellow-green, as the polarity of the solvent increases. The aqueous solution appears colorless due to the poor solubility of the compound.

#### *2.3. Analysis of Fluorescence in Lipid Membranes*

The heterogeneity and dynamism of cell membranes make it challenging to study their structure because they are incredibly complex entities. For this reason, the simplest model for membranes is represented by unilamellar vesicles. In particular, giant unilamellar vesicles, GUVs, with dimensions comparable to the sizes of a cell (in an interval of about 1–100 µm), are particularly useful as model membranes. Over the years, several methods have been developed for the design of GUVs. Many methods have been proposed to control the size and composition of the vesicles [24,25]. One of the most recently developed methods is the "droplet transfer" or "reverse emulsion" [26], characterized by the stratification of an aqueous phase and a water-in-oil emulsion with the formation of the bilayer at the interface of the two phases. This method permits to easily encapsulate hydrophilic compounds in the aqueous lumen [27,28] and blend hydrophobic species in the membrane, such as fatty acids [29,30] or polymers [31,32]. In this work, the "droplet transfer" method was used for the formation of vesicles of POPC and a ternary mixture of POPC:DPPC:Chol to simulate the two phases

Lo and Ld of a membrane. The solvatochromism of the probe 3HF18 can be used to characterize cell membranes or their simplified models consisting of lipid vesicles. The probe has been successfully encapsulated both in pure POPC vesicles in the Ld phase, Figure 5A,C and in mixed POPC:DPPC:Chol membranes (Lo phases, Figure 5B,D). Fluorescence has been recorded at two different wavelength intervals, where the emission maxima do not overlap: in (Figure 5A,B) λexc = 405 nm, the emission range is λ = 477–527 nm; in Figure 5C,D λexc = 405 nm and the emission range is λ = 550–600 nm. Cholesterol is reported to increase polarity along the membrane surface, while in the hydrophobic core it decreases polarity [33]. In the case of mixed vesicles, therefore, due to the effect of Chol, the surface polarity may be higher than in vesicles of POPC alone. In a previous work [21], we have shown that membrane probes with a zwitterionic head can be easily placed with the chromophore beneath the membrane surface and with the hydrophobic tail inserted between the tails of phospholipids. In this case, a higher IN\*/I T\* emission ratio in GUVs of POPC/DPPC/Chol is observed (Figure 5E,F) which supports the notion that the chromophore is positioned below the surface of the membrane.

**Figure 5.** Fluorescence confocal images (**A**–**D**) and intensity ratio (**E**,**F**) of vesicles made of POPC (**A**,**C**,**E**) and POPC/DPPC/Chol 1/0.5/0.7 (**B**,**D**,**F**). Excitation at λ = 405 nm, Emission at λ = 477–527 nm (**A**,**B**) and λ = 550–600 nm (**C**,**D**). Bars scale 5 µm. Contrast of the images has been adjusted to improve the quality of the pictures.

The fluorescence of the probe is limited to the surface of the vesicles, thus confirming the anchorage of the probe to the double layer. Molecular dynamics studies indicate that the flip flop mechanism of the probe is significantly lower than that of fluorophore alone and also than that of POPC and DPPC. In fact, the zwitterionic head of the probe has a dipole inverted compared to that of phospholipids and increases its membrane cohesion.

#### *2.4. Molecular Dynamics in Model Membranes. Thickness and Area Per Lipid*

Through molecular dynamics (MD) analysis, double layer thickness and area per lipid values were calculated. The experimental values of double layer thickness and area per lipid (A<sup>l</sup> ) are commonly used for the validation of simulations [34,35]. The average lipid bilayer thickness, calculated as the average distance between the phosphorus atoms of the two double-layer sheets, showed that the insertion of the probe, in both excited N\* and T\* forms, does not cause membrane perturbation (Table 2).


**Table 2.** Thickness values (in Å) of the two membranes in the presence and absence of the probe.

<sup>1</sup> Thickness of simulated POPC and DPPC-Chol membranes; <sup>2</sup> Experimental thickness of POPC and DPPC-Chol membranes; <sup>3</sup> Thickness of POPC membrane with N\* 3HF18; <sup>4</sup> Thickness of POPC membrane with the T\* 3HF18; <sup>5</sup> Thickness of DPPC-Chol membrane with N\* 3HF18; <sup>6</sup> Thickness of DPPC-Chol membrane with T\* 3HF18.

The thickness values obtained from pure membrane simulations are 37.3 Å for POPC and 45.9 Å for DPPC-Chol. The calculated thickness values are in good agreement with the experimental values of 39.8 ± 0.8 Å [35] for POPC and 44.8 Å for DPPC-Chol [34], respectively. From the values shown in Table 2, it can be seen that the inclusion of the 3HF18 molecule does not cause a change in the thickness of the analyzed membranes. Table 3 shows the area per lipid (A<sup>l</sup> ) data obtained from simulations on different systems compared with the experimental data obtained from CHARMM-GUI [36].


68.3 43.2-48.6 69.9 ± 0.5 69.2 ± 0.5 43.2 ± 0.2 43.6 ± 0.2

**Table 3.** Area per Lipid (A<sup>l</sup> ) values of the two membranes with and without the probe 3HF18.

<sup>1</sup> Experimental A<sup>l</sup> of POPC e DPPC-Chol membranes; <sup>2</sup> A<sup>l</sup> of POPC membrane with N\* 3HF18; <sup>3</sup> A<sup>l</sup> of POPC membrane with T\* 3HF18; <sup>4</sup> A<sup>l</sup> of DPPC-Chol membrane with N\* 3HF18; <sup>5</sup> A<sup>l</sup> of DPPC-Chol membrane with T\* 3HF18.

When the cholesterol concentration inside the membrane is high (>20%), it is necessary to decrease the area per lipid values by 10–20%. Therefore DPPC-Chol displays a range of values. Also, for this analysis, the areas per lipid in the presence and absence of the probe are comparable, so we can conclude that the presence of the probe does not cause perturbation in the membranes.

### *2.5. Deuterium Order Parameter SCD*

To assess whether the probe causes a change in membrane fluidity, we performed an analysis of the deuterium order parameter (SCD. The SCD parameter is a measure of the motor anisotropy of the C-D bond analyzed and provides its orientation over time. It is derived from NMR experiments. SCD analysis provides comprehensive information on membrane fluidity and allows us to predict changes in the fluidity of the lipophilic chains of phospholipids near a guest molecule (in this case, the 3HF18 probe). The fluidity of the POPC and DPPC-Chol chains was evaluated in the presence and absence of the 3HF18 probe in the forms N, N\* and T\* (Figures 6 and 7). The results of the N-form calculations are reported in the Supplementary Materials section. In Figure 6, the SCDs for POPC phospholipid chains around 5 Å from the guest molecule is shown. In Figure 6a, the unsaturated chain is shown, and in Figure 6b, the saturated one.

**Figure 6.** Order parameter SCD for (**a**) the unsaturated oleic and (**b**) the saturated palmitoyl acyl chains of POPC phospholipid in POPC pure membrane (orange curve) used as control and in POPC/3HF18 membrane (N\*, yellow curve; T\* green curve). Notes: On the *Y*-axis, the SCD is indicated; on the *X*-axis, the carbon atom position is reported, starting from the first (1) alpha carbon atom in the chains.

**Figure 7.** Order parameter SCD for (**a**) the saturated Sn-1 palmitoyl chain and (**b**) the Sn-2 saturated palmitoyl chains of DPPC phospholipid in DPPC pure membrane (green curve) used as control and in DPPC/3HF18 membrane (N\*, blue curve; T\* yellow curve). Notes: On the *Y*-axis, the SCD is indicated; on the *X*-axis, the carbon atom position is reported, starting from the first (1) alpha carbon atom in the chains.

The SCD decreases sharply in the region near the double bond (carbon atom 9) and approaches 0 in the tail region, indicating the highest disorder within the double layer (Figure 6). Similar SCD profiles were observed experimentally using NMR studies [37]. Figure 7 shows the SCD values of the two saturated chains of DPPC. Even for the DPPC, in the final region of the tail, the SCD decreases towards 0, thus showing a greater disorder.

From Figures 6 and 7, we can see that the probe (both tautomeric forms of 3HF18) interacts with the aliphatic chains of phospholipids and does not cause a significant change in the order of the membranes.
