F6 film preparation and Scanning Electron Microscopy (SEM)
F6 peptides were synthetized by solid phase peptide synthesis using Wang or Rink amide resin, which allow to achieve carboxylated and amidated variants, respectively.30 The capping of one or both the termini allows to obtain peptides with different charge states. Due to their aromatic and hydrophobic nature, all peptides are scarcely soluble in water (~0.5 mg/mL). On the contrary, they are highly soluble (up to 100 mg/mL) in HFIP (1,1,1,3,3,3-hexafluoro-2-propanol), generally used as solvent/co-solvent for highly hydrophobic Phe-containing peptides.32,33 Only the zwitterionic form, H+-F6-O-, showed a reduced solubility in HFIP (50 mg/mL). PL properties of these F6-peptides were investigated only for the samples at solid state, prepared by drop-casting peptide solutions at two different concentrations: 5 mg/mL and 100 mg/mL (50 mg/mL for H+-F6-O-). Prior to PL characterization, we evaluated the morphology of the peptide nanostructures at both the concentrations by Scanning Electron Microscopy (SEM). SEM microphotos in Fig. 1 allow to observe a dependence of the morphology from the concentration. At low concentration, the four variants assemble in twisted structures with a length and a thickness between 300-1000 μm and 5-20 μm, respectively. Instead, at high concentration, SEM microphotos show a sort of film formed by very short fibrillary structures aligned along all the directions, thus suggesting a lower degree of order with respect to samples at 5 mg/mL. This decrease of order is probably due to the high volatility and rapid evaporation of the organic solvent. Although our SEM analysis are only qualitative, there seems to be a somewhat higher degree of order in the two acetylated peptides with respect to the not acetylated ones. The optoelectronic properties of the air-dried F6 samples were preliminarily studied via fluorescence microscopy. The images of fibers and films in the dark field, in the blue and green regions are reported in Fig. S1 and Fig. 2, respectively. In clear accordance with the SEM characterization, the images showed µm-long fibrillary structures and films formed by ~100 µm-wide cracked plaque structures.
Successively, the optical properties of the F6 nanostructures were extensively investigated from a quantitative point of view: from the spectral emissive viewpoint, all peptides exhibit a characteristic PL in the visible spectrum, having two main emission bands evidenced at ~400 and ~450 nm (Fig. 3a). This finding was independent of the concentration of the precipitating solution. This trace, not ascribable to the π-π stacking between the phenyl ring of Phe residues,28,34 can rather be attributed to the presence of β sheet-rich nanostructures.15,22
A deeper investigation of this fluorescence emission clearly indicates that all four peptides present a highly similar multicolour emission profile as evidenced by the normalized PL spectra versus the excitation wavelength in the range from 330 and 430 nm (Fig. 3b, S2). This PL behaviour was previously reported for thermally treated triphenylalanine nanodots by Rosenman et al.32 Moreover, all F6 peptides show a fluorescence emission that is almost linearly dependent from the excitation wavelength in the whole investigated range (Fig. 3c, d and Fig. S2). The variation of the emission wavelength upon increasing the excitation wavelength exhibited by these peptides represents an interesting violation of the Kasha’s rule.31 Indeed, a linear behaviour could be clearly evidenced for all the four F6 compounds, with estimated slopes and intercepts demonstrated as fully compatible (Table 1). In order to gain further insights into this intricate process, we measured the excitation spectra corresponding to the main PL components (λem = 400 nm and 450 nm). As shown in Fig. 3d, the spectra exhibit a single well-defined peak, centred at ~325 nm for all four peptides. Optical transitions at such low energy in self-assembling β-rich polypeptides/proteins have been initially correlated to a putative long-range charge delocalization along their backbone due to hydrogen bonding occurring between peptide units.13 This hypothesis has been supported by the evidence that hydrophobic conditions (low humidity and pressure) occurring in amyloid fibrillary systems hamper both PL and charge transport.16 As a whole, the present data clearly demonstrate that this excitation-dependent emissive behaviour in the visible range as well as the low-energy excitation peak are invariant for all samples regardless their terminal charge. Therefore, in the examined systems, the photochemical explanation of this phenomenon seems not to be attributable to proton transfer or other mechanisms involving the termini of the peptide like few studies suggest.23,24 Nevertheless, the results are coherent with very recent theoretical findings reported in the literature,27 according to which the emission is promoted to the suppression of non-radiative paths due to the decreasing of the excitation energy and stabilization of n → π* transitions (i.e., electrons belonging to non-bonding pair jumping up to an antibonding π* orbital). This is shown to be the consequence of a deplanarization of amide groups, which may lead to the peculiar supramolecular organization of the examined compounds.27 Moreover, other studies demonstrate that PL from H-bonds rich, non-aromatic systems arises consequently to interactions between amide groups: the abundancy of hydrogen bonding was demonstrated to bring these functionalities into close proximity.35 Indeed, the results reported in the current work are in good agreement with these proposed models as they do not invoke any role for the charged ends.
The photoemissive behaviour of the F6 films is slightly modified upon increasing the peptide concentrations in the mother solution since another excitation band centred at around ~370 nm arises differently, in addition to the component at 325 nm. In this case a small but significant dependence on the terminal charge states is observed as the peak is observed only for the two peptides with the amidated C-terminal end (Fig. S3a). This experimental evidence allows to conclude that, when the peptide is prepared from highly concentrated solutions (100 mg/mL), the charge state of the terminal ends also plays a role in the fluorescence emission, attributed to Phe π-π stacking and induced by exciting the samples at 257 nm (Fig. S3b). Indeed, in these conditions, an increased emission of the peptides with the C-terminal capped ends compared to those with a charged -COO- moiety is observed. This aggregative phenomenon arises at high concentrations of deposited material and it can still be contextualized in terms of “preferential direction” in the self-assembling. It is reasonable to conclude that the main contribution to the optical response of the F6 films, in the higher concentration regime, comes from the stacking of the huge number of aromatic residues. This analysis is also supported by the SEM images above discussed for the samples at high concentrations. Finally, measurements of fluorescence quantum yield (PLQY) at the two excitation wavelengths of 325 and 370 nm (Table 2) provide further insights into the PL emission of these systems. At 325 nm an increase of the average PLQY from ~5(1)% to ~10(2)% is evidenced when going from 5 to 100 mg/mL. No specific effect of the charged/uncharged ends is observed for this PL increase. Although it is important to point out that PLQY only provides information about the integrated photoemissive behaviour of the samples and not about the exact origin and attribution of the spectral features of the samples. This finding points out that the concentration of the peptide in the stock solution has an impact on the resulting structure of the solid and on the related PL properties. The inspection of the PLQY obtained upon excitation at 370 nm, which has only been evaluated in the samples at higher concentrations, indicates some role of the charge state of the terminal as C-terminal capped variants present quantum yields higher than those exhibited by peptide containing –COO- groups.
Photobleaching study
Having shed light upon the intriguing spectroscopic properties of the nanostructures formed by F6 peptides, we evaluated their robustness against photodegradation and bleaching, which represent essential features to pursue any nanophotonic application. For this experiment, fluorescence spectra were collected and monitored after continuous exposure to the 330 nm excitation line, keeping the sample mounted in the integrating sphere and the excitation shutter open, for up to 180 mins at an incident power of ~820 μW. Assuming the interchangeability of the four variants in view of the previous results, the measurements were solely carried out on the Ac-F6-Am peptide film at 100 mg/mL concentration. After 3 h continuous exposure to UV light, a 64% retaining of the integrated PL intensity of the sample could be evidenced, revealing a remarkable bleaching resistance (Figure 4a). A rough estimation of the bleach rate, performing a single-exponential decay fit, led to a value of 7.2 (1.1) h, which is much higher than that of any conventional protein dye (whose characteristic range is ~101-102 s).36 Furthermore, the two deconvolved components of the PL spectrum (peaked at ~400 and ~450 nm) interestingly exhibit a different bleaching kinetics, highlighting the different origin of their photoemissive mechanisms (Fig. 4b). Moreover, no recovery was evidenced after 40 mins in dark (data not shown).
Photoemissive properties of Aβ16-21 peptides
In order to demonstrate that the PL results collected for F6 nanostructures can be generalized to other classes of β-sheet-assembling peptides, we investigated another amyloidogenic system. Specifically, for this study we choose the sequence 16-21 “extracted” from the Aβ1-42 peptide (Aβ16-21), in its fully neutral (Ac-Aβ16-21-Am) and zwitterionic state (H+-Aβ16-21-O-)(see Figure S4a). Analogously to their parental peptide Ac-Aβ16-20-Am,37 the structural characterization on Aβ16-21 variants (SEM and ThT assay) confirmed their capability to self-organize in β-sheet-rich nanostructures (see Figure S4b and S4c). The PL characterization of Aβ-peptide films at 5 and 100 mg/mL was carried out using the same setting chosen for the F6 samples. Due to the increased solubility in water of Aβ variants respect to F6 ones, samples at 5 mg/mL were prepared from their aqueous solutions. Analogously to the F6-peptides, PL microscopy images of Aβ (drop-casted from a solution at 5 mg/mL in H2O) show emissive fibrils in the blue spectral region (Fig. 5a). From the spectral viewpoint, both the Aβ variants show an emission peak around 400 nm (Fig. 5b), a multicolour PL emission profile in the range 330-400 nm (Fig. 5c) and a non-trivial dependence of the emission maximum on the excitation wavelength (Fig. 5d). Moreover, both the samples at 5 and at 100 mg/mL show a low-energy excitation peak at 325 nm (λem = 400 nm), which is compatible with what observed in F6 samples at the same concentration (Fig. 5e). The measured PLQYs at 100 mg/mL were (4.5 ± 0.2)% and (3.00 ± 0.15)% for H+-Aβ16-21-O- and Ac-Aβ16-21-Am, respectively. Such values are comparable to those measured for the F6 variants in the same experimental conditions, as this peculiar emission turned out to exhibit a quite high efficiency. Altogether, such analyses support and are coherent with the assumption made for the F6, i.e., that the excitation-dependent PL from an amyloidogenic system may be a direct consequence of a peculiar tridimensional-spatial disposition of the amide groups consequent to the β-sheet arrangement.