Ultrabright Föster Resonance Energy Transfer Nanovesicles: The Role of Dye Diffusion

The development of contrast agents based on fluorescent nanoparticles with high brightness and stability is a key factor to improve the resolution and signal-to-noise ratio of current fluorescence imaging techniques. However, the design of bright fluorescent nanoparticles remains challenging due to fluorescence self-quenching at high concentrations. Developing bright nanoparticles showing FRET emission adds several advantages to the system, including an amplified Stokes shift, the possibility of ratiometric measurements, and of verifying the nanoparticle stability. Herein, we have developed Förster resonance energy transfer (FRET)-based nanovesicles at different dye loadings and investigated them through complementary experimental techniques, including conventional fluorescence spectroscopy and super-resolution microscopy supported by molecular dynamics calculations. We show that the optical properties can be modulated by dye loading at the nanoscopic level due to the dye’s molecular diffusion in fluid-like membranes. This work shows the first proof of a FRET pair dye’s dynamism in liquid-like membranes, resulting in optimized nanoprobes that are 120-fold brighter than QDot 605 and exhibit >80% FRET efficiency with vesicle-to-vesicle variations that are mostly below 10%.


Methods for Molecular Dynamics
. Process parameters of the Delos-susp methodology for QS preparation 19 Table S2. Characteristics of the dye-loaded nanoformulations 20 Table S3. Geometric parameters of the fluorescent QS 21   Table S4. Volume of QS membrane in relation to the total volume 22 Table S5. Calculation of the dye concentration at the QS membrane 23 Table S6. Brightness per particle 24 Table S7. Optical characteristics of fluorescent inorganic and organic nanoparticles reported in bibliography 25 Table S8. Estimation of donor-acceptor averaged distance in the Quatsome bilayer 27

Methods for Molecular Dynamics Simulations
Protocols for MD simulation. All MD simulations were performed using the NAMD program. 5 The equations of motion were solved using a 2 fs time step. Electrostatic interactions were computed using the PME method (PME) with usual settings in NAMD (1 Å resolution, updated each 2-time steps). Lennard-Jones interactions were truncated at 1.2 nm employing a switching function starting at 1.0 nm. Periodic boundary conditions were employed in all directions. The temperature was kept constant at 298 K using a Langevin thermostat with a relaxation time of 1 ps. The pressure of 1 atm and zero surface tension were imposed using the anisotropic Nosé-Hoover-Langevin piston implemented in NAMD (oscillation period of 100 fs and decay time of 50 fs). In all cases, we performed an initial energy minimization and an equilibration run until temperature, pressure, and membrane area reach stable values (these runs were short since initial configurations were built from equilibrated ones). The production runs were 40 ns for each case.
Results for S1 and S2 simulations. For both the DiI-DiI and DiD-DiD dye cases (S1 and S2 in Table S7), both pairs are perfectly incorporated and interdigitated between QS components, as illustrated in the snapshots of Figure S10 and S11. As observed there, both dyes are located and oriented as expected, being the head groups at the water-surface interface and the tails completely immerse in the hydrophobic region of the membrane. Results from the DiI-DiI pair (S1 simulation) show that both dyes remain inside the QS membrane stable along simulation time, not deforming the membrane nor aggregating. The calculated thickness of the bilayer (measured from the peaks of the nitrogen-nitrogen atom distance from CTAB distribution) is 4.2 nm for S1 and 4.3 for S2. In our previous simulations with a single dye, 4 we obtained 4.2 nm for both dyes. It is also close to the value 4.3 nm for the bilayer thickness obtained in our previous work in absence of dyes. 2 Note that MD simulations of DiI dyes in DPPC phospholipid bilayer 6 also predict the absence of aggregation.
We have computed the distance between both dyes present in S1 and S2 simulations as a function of time. As seen in Figure S10, the DiI-DiI distance (measured between the central carbon atoms of both dyes) presents a wide distribution, roughly uniform, indicating an absence of DiI-DiI interaction. In the case of DiD-DiD ( Figure S11), we do not observe aggregation but there is a larger tendency of both dyes to be one near the other instead of being fully separated, as seen in Figure S11 (right). As shown in that figure, there is a peak at a separation of about 3 nm between the two central carbons of each dye.
Additional results for S3 simulations. The main results for the S3 simulation are reported in the main paper, Figure 5. Here, as additional material, we report the DiD-DiI separation shown in Figure 5c calculated differently. In Figure 5c, the separation between dyes is calculated as the separation between the N atoms of each dye. Here, in Figure S12 we also report the separation between DiI and DiD during the S3 simulation computed as the donor-acceptor separation. The separation is computed as the distance between each nitrogen atom of the DiI (donor) and the centre of mass of the chromophore of the DiD dye (acceptor).  The configuration comprises a 7.3 mL high-pressure vessel (HPV), whose temperature is controlled by an external thermostatic bath; a syringe pump (model 260D, ISCO Inc, USA) (P) to introduce CO2 inside the HPV through valve V-4; a depressurization valve (V-7) from which the expanded liquid solution is depressurized into the aqueous phase placed in a collector (C) located after V-7, N2 is pumped into the vessel through V-6. A one-way valve is located after V-6 to prevent contamination of CO2 in the N2 line. V-2, V-3, and V-5 are dividing the CO2 and N2 pipelines. There is also a pressure indicator (PI) and another one-way valve before the vessel to prevent the backflow from HPV, which could lead to contamination of the gas lines.
The specific protocol for the preparation of the four QS-I,D formulations is detailed in Table S1.    were prepared (see Table S1)   [c] The dye encapsulation efficiency is defined as the ratio between the amount of dye present in the final formulation of QS-I,D and the initial amount of dye loaded into the reactor.
final dye concentration in bulk and the initial dye concentration in bulk.
[d] QS concentration and size distribution were measured by Nanoparticle Tracking Analysis (NTA). The averaged results are obtained from n ≥ 3 (error ± 7%)    [b] Brightness of the nanoparticle excited at DiD and emission recorded at DiD (λex = 600 nm, λem = 670 nm) Brightness per particle was determined as the product of the molar extinction coefficient per particle (εp) and the fluorescence quantum yield (φ). 17 The εp was determined by the Lambert-Beer Law (εp = Abs650 / CQS (M)), where Abs650 was obtained by UV-Vis spectroscopy and the molar concentration of QS (CQS) from the NTA values (Table S3). The φ measurements were carried out using a Quantaurus-QY Plus (UV-NIR absolute PL quantum yield spectrometer C13534-11), Hamamatsu Photonics. The samples were diluted until absorbance values OD ≈ 0.1 were obtained, the excitation wavelength was 520nm and 600nm, for FRET and direct DiD emission determination, respectively, (the direct excitation of DiD at 520nm can be considered negligible). 15 Illumination time was 0.9 seconds and the final φ value come from an average of 20 repetitions.  [a] Dyes are constantly diffusing through the QSs membrane, at higher loadings dye molecules are closer to each other due to the higher amount of molecules per volume, leading to shorter average distances. Composition of the systems considered in the MD simulations, including the total number of atoms, number of molecules of each component, simulation time, and box size (for each dye molecule a Clanion is also present and for each CTA + surfactant a Brcounterion is also present).