How to unveil self-quenched fluorophores and subsequently map the subcellular distribution of exogenous peptides

Confocal laser scanning microscopy (CLSM) is the most popular technique for mapping the subcellular distribution of a fluorescent molecule and is widely used to investigate the penetration properties of exogenous macromolecules, such as cell-penetrating peptides (CPPs), within cells. Despite the membrane-association propensity of all these CPPs, the signal of the fluorescently labeled CPPs did not colocalize with the plasma membrane. We studied the origin of this fluorescence extinction and the overall consequence on the interpretation of intracellular localizations from CLSM pictures. We demonstrated that this discrepancy originated from fluorescence self-quenching. The fluorescence was unveiled by a “dilution” protocol, i.e. by varying the ratio fluorescent/non-fluorescent CPP. This strategy allowed us to rank with confidence the subcellular distribution of several CPPs, contributing to the elucidation of the penetration mechanism. More generally, this study proposes a broadly applicable and reliable method to study the subcellular distribution of any fluorescently labeled molecules.


S1.1.2. Procedure to follow progression and completion of the reactions
Progression of the in-solution reactions was monitored on analytical TLC commercially available precoated aluminium packed plates (Merck Kieselgel 60 F254). Chromogenic products were directly observed on TLC plates.
Completion of the amide bond formation on solid support was monitored with the Kaiser test.
Propargylglycine does not generate the apparition of the expected blue coloration of the beads upon deprotection but a light red coloration.

S1.1.3. General procedure for automated peptide synthesis
Peptides were partially synthesized using an ABI 433A automated peptide synthesizer from Applied Biosystem in a 0.1 mmol scale in Boc strategy. A 10-fold molar excess of amino acid was used (1 mmol) with the classical HOBt/DCC activation protocol.
Peptides were manually synthesized in various scales by Boc strategy using the following protocol: amino acids (5 eq.) were activated using HBTU (4.5 eq.) and DIPEA (10 eq.) in NMP (in an The derivatization of the peptides with fluorophores will be introduced below.
At the end of the synthesis, the peptidyl-resin was washed with NMP (3 x), CH 2 Cl 2 (3 x) and MeOH (3 x) and dried under vacuum. The weight of the peptidyl-resin was measured and compared with the expected yield to anticipate any eventual truncation of the sequence.
After cleavage, the peptide was precipitated into cold diethyl ether. The precipitate was filtrated and washed with cold ether and finally dissolved in 10 % AcOH in water. The crude peptide was lyophilized.
Peptides were lyophilized using an Alpha 2/4 Freeze dryer from Bioblock Scientific.

S1.3.2. Click reaction protocol to obtain rhodamine labeled peptides
The click reaction with rhodamine B azide ( Supplementary Fig. 2) was performed on solid support using the following protocol.

S1.4. Peptide sequences and characterization of the studied peptides
The peptide sequences and their analytical characterizations are reported in Supplementary Table 2 and Supplementary Table 3, respectively.

The multiple origins of fluorescence self-quenching
The goal of this note is to describe concisely the possible origins of fluorescence self-quenching in our system. In particular, we would like to interpret the evolution of the fluorescence of the CPPs in the presence of an excess of DOPG ( Fig. 1 and Supplementary Fig. 3). In such a case, the CPP is adsorbed and confined in the phospholipid phase and is getting progressively diluted upon addition of DOPG.

S2.1. Note on the inner-filter effect
The order of magnitude of the number of vesicle encountered by a photon emitted from a fluorescent peptide on a vesicle, noted n, is: L is the width of the cuvette (L ~1 cm), a is the radius of a vesicule (a ~100 nm) and d is the distance between two vesicles. Because d is defined as the distance between two vesicles, curve, whose slope is K SV .
If the fluorescence quencher is a fluorophore F: Fluorescence energy transfer between two identical fluorophores cannot be the physical origin of dynamic self-quenching since it is not accompanied by any energy loss. In consequence, fluorescence transfers with "energy losses" have to be envisioned.
Alternatively, when the concentration of quencher is high enough, so that the quencher is always in interaction with the fluorophore, static quenching may occur. Static quenching might be due to the formation of a ground-state complex between the fluorophore and the quencher and in our case the quencher can be a fluorophore F. The equilibrium constant of the ground state complex formation, K S , is: The dependence of fluorescence intensity on the concentration of the quencher, here the fluorophore F, can be derived and gives again a linear relationship between I 0 /I and [F]: Not surprisingly, static quenching decreases the number of fluorescent dyes inducing a global decrease of the quantum yield.
In conclusion, if classical dynamic or static quenching is responsible for the observed fluorescence decrease, linear Stern-Volmer plots will be observed.  Fig. 3). However, when the CPP is densely packed on the surface, an upward curvature is always observed (Supplementary Fig. 3). Such a curvature signifies that at high CPP:DOPG ratio, a very important self-quenching occurs. Because the simple dynamic or static models that we proposed do not satisfactorily describe the deviation, they have to be adapted.
The upward curvature of our plot, concave towards the y-axis may originate from three different phenomena: (i) dynamic and static quenching can occur simultaneously or (ii) the space between the fluorophore and the quencher (i.e. another fluorophore) is small enough for the quencher to be next to the fluorophore at the moment of the excitation -as a result, the fluorophore is immediately quenched without the requirement of the formation of a physically defined ground complex-or (iii) the static quenching yields "black dimers", which are themselves efficient quenchers. We will successively investigate these possibilities.

S2.4.1. Combined dynamic and static fluorescence quenching
We interpreted static quenching as the formation of a complex between the fluorophore and the quencher. If static and dynamic quenching occurs simultaneously, this means that dynamic quenching can decrease the fluorescence of the remaining non-complexed fluorophore.
Symmetrically, this means that the formation of a non-fluorescent complex can reduce the fraction of free fluorophore that were dynamically quenched:

S2.4.3. Formation of "dark dimers"
This last model is based on the static fluorescence quenching and assumes that fluorophore dimers (or aggregates) can absorb light at the emission wavelength of the fluorophores. These dimers thus become "black dimers". 2 In this model, a "transfer with losses" between two fluorophores is not required: the dimers are the fluorescence quenchers.
If dimers are fluorescence quenchers, at low CPP concentration we only observe a linear plot due to formation of dimers; this is the static quenching.