Solid-state optical properties of self-assembling amyloid-like peptides with different charged states at the terminal ends

The self-assembling of small peptides not only leads to the formation of intriguing nanoarchitectures, but also generates materials with unexpected functional properties. Oligopeptides can form amyloid-like cross-β assemblies that are able to emit intrinsic photoluminescence (PL), over the whole near-UV/visible range, whose origin is still largely debated. As proton transfer between the peptide chain termini within the assembly is one of the invoked interpretations of this phenomenon, we here evaluated the solid state PL properties of a series of self-assembled hexaphenylalanine peptides characterized by a different terminal charge state. Overall, our data indicate that the charge state of these peptides has a marginal role in the PL emission as all systems exhibit very similar multicolour PL associated with a violation of the Kasha’s rule. On the other hand, charged/uncharged ends occasionally produce differences in the quantum yields. The generality of these observations has been proven by extending these analyses to the Aβ16–21 peptide. Collectively, the present findings provide useful information for deciphering the code that links the spectroscopic properties of these assemblies to their structural/electronic features.

www.nature.com/scientificreports/ nuclear quantum effect 26 ; (VI) extensive delocalization of peptide bond electrons in the β-structure 27 . In 2016, Pinotsi et al. reported an interesting study in which PL emission from amyloid-like fibrils was correlated to the proton transfer between the C-and the N-termini of two adjacent strands in the β-sheet structure 23,24 . Despite an alternative explanation, involving the deformation of the peptide bond planarity in the assembly, has been successively reported 27 , the contribution of the proton transfer in the origin of the PL properties has not been further highlighted from the experimental point of view 27 . In this framework, we have recently undertaken extensive structural and spectroscopic investigations on analogous amyloidogenic systems constituted by selfassembled phenylalanine-based homopeptides, specifically in form of hexa-Phe (F6). Due to their hydrophobic nature, they were frequently conjugated to PEG (polyethylene glycol) moieties of different length to increase their solubility and to study their properties in solution (PEG8-F6, PEG12-F6, PEG18-F6 and PEG24-F6) 21,28,29 . Moreover, by studying a series of F6 variants endowed with either free or capped N-/C-termini we have also investigated the effects that charged ends play on the aggregation mechanism and on the structural properties of the final fibers 30 . Since the charged state of the peptide end crucially affects the probability of proton transfer between the termini, we here perform a comparative analysis of the photoemissive properties of four peptide analogues having the same peptide sequence (F6) and different charged states of their termini (see Fig. 1). Together with the zwitterionic peptide H + -F6-O − , which has charged termini, we characterized the fully capped peptide Ac-F6-Am (in which the C-and the N-termini are amidated and acetylated, respectively), and the two intermediate variants in which the termini are alternatively capped or uncapped, Ac-F6-O − and H + -F6-Am. Collectively, these four variants provide all possible combinations of charged/uncharged states on the F6 peptide,  www.nature.com/scientificreports/ thus allowing an insightful comparison of the dependence of fluorescence behaviour as function of the terminal state. Indeed, our previous structural/morphological characterization pointed out the tendency of these peptidic variants to self-assemble in amyloid-like structures, with an organization in β-sheet nanostructures with an antiparallel orientation of the β-strands in the cross-β framework, regardless their terminal charge 30 . The structural resemblance of the supramolecular organization of these nanostructures indicates that potential differences in the PL properties among different peptides can only be ascribed to their different chemical properties. The solid-state spectroscopic characterization of these compounds reported in the current work unravel some interesting features such as (1) the independence of the PL emission from the charge states of the peptide termini, (2) a multicolour PL emission, and (3) the general violation of the Kasha's rule 31 . We also demonstrate that F6 films are stable against photodegradation and bleaching, an essential feature required for nanophotonic applications. The generality of the findings obtained for the F6 peptides are proven by studying the completely charged and uncharged variants of the Aβ 16-21 peptide (H + -Aβ 16-21 -O − and Ac-Aβ 16-21 -Am), which self-assemble in amyloid-like structures analogously to F6 variants.

Results and discussion
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). The effective capability of these peptides to give zwitterionic form in HFIP has been assessed by the Safranine T qualitative assay. Safranine T (reported also as Safranine O or basic red 2) is cationic azonium compounds of symmetrical 2,8-dimethyl-3,7-diaminophenazine able to interact with negative charged via electrostatic interactions 34 . This assay was carried out on samples in their solid state. The selected peptides were the completely protected peptide Ac-F6-Am-used as negative control-and, on the unprotected one, H + -F6-O − used as positive control. Peptides were drop-casted on a slide glass from 5.0 mg/mL HFIP stock solutions and air-dried at room temperature. The resulting films were stained using a pure ethanol Safranine T solution at 0.30 mg/mL and imaged by optical microscopy under bright and cross polarized light. From the inspection of images in Fig. S1, an effective interaction between the peptide film and the dye only for H + -F6-O − appears, thus confirming the existence of charges on the peptide under these experimental conditions. 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 micrographs 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 and 1000 μm and 5-20 μm, respectively. Instead, at high concentration, SEM micrographs 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 Figs. S2 and 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 28,35 , 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 to 430 nm (Figs. 3b, S3). 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 (Figs. 3c, d and S3). Also, very recent outcomes have shown the occurring of this same effect in solid-state polyethylene glycol-derivatized F6 films 36 . The variation of the emission wavelength upon increasing the excitation wavelength exhibited by these peptides represents an interesting violation of the Kasha's rule, according to which "the emitting level of a given multiplicity is the lowest excited level of that multiplicity" 31 . The violation of the rule implies the possibility that the photons can be emitted by a higher energy state thereby showing a dependence on the excitation wavelength; in particular, the shorter the excitation wavelength the shorter the emission wavelength. Indeed, a linear behaviour could be clearly evidenced for all the four F6 compounds, with slopes and intercepts fully compatible each other for all the compounds (Table 1). In order to gain further insights into this intricate process, we measured the excitation spectra corresponding to the main PL components (λ em = 400 nm-continuous linesand 450 nm-dash-dotted lines). As shown in Fig. 3d, the spectra exhibit a single well-defined peak, centred at ~ 325 nm for all four peptides, regardless the excitation wavelength. 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  www.nature.com/scientificreports/ 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 37 . 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. S4a). 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. S4b). 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)   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 min 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 (Fig. 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 ~ 10 1 -10 2 s) 38 . 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 min in dark (data not shown).
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 H 2 O) 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  www.nature.com/scientificreports/ 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 (λ em = 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 + -Aβ 16-21 -O − and Ac-Aβ 16-21 -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.

Conclusions
The PL results collected on two amyloidogenic systems, F6 and Aβ [16][17][18][19][20][21] , allow to conclude that all the generated peptide films are able to emit multicolor PL profiles regardless their terminal charge state. The observed optical properties provide further support to the notion that proton transfer between the C-and the N-terminal of two adjacent strands in a β-sheet structure is not essential for generation of the intrinsic PL exhibited by amyloid-like systems.These results are compatible with the hypothesis that the emission is a consequence of the decrease of the excitation energy and stabilization of the peptide bond due to a deplanarization of amide groups 27 , although this explanation has to demonstrate its validity in interpreting the fluorescence exhibited by helical peptides 40 as well as the its dependence from the physical state of the sample 36 . The efficient excitation-dependent PL of the examined compounds was investigated and demonstrated to be retained even at higher concentrations where other aggregation-induced emissive phenomena occur, reaching PLQY values of the order of ~ 10%. The origin of this visible PL trace seems to be a direct consequence of their peculiar 3D-spatial disposition at the very first-stage supramolecular level.
Peptide synthesis. F6-homosequences were synthesized according to the method previously described 30 .
100 µL of peptide solutions of H + -F6-O − and Ac-F6-Am were deposited from HFIP stocks at 5.0 mg/mL. After solvent evaporation at room temperature, peptide films were stained for 1 min with 50 µL of a 0.30 mg/mL Safranine T solution in anhydrous ethanol. After total dryness of the films, stained samples were inspected by optical microscope under bright-field illumination and between crossed polars by using a Nikon AZ100 microscope.
Thioflavin T (ThT) assay. Aggregate formation of H + -Aβ 16-21 -O − and Ac-Aβ 16-21 -Am was assessed by fluorescence spectroscopy using ThT. This dye associates rapidly with β-rich aggregates giving rise to an enhanced fluorescence emission at 482 nm 41 . Spectra of an aqueous solution of 25 µM ThT, before and after the addition of Ac-Aβ 16-21 -Am derivative (3.0 mg/mL), were recorded at room temperature. The spectrum of the peptide solution alone was also acquired as reference. Fluorescence measurements were recorded between 460 and 650 nm exciting the samples at 450 nm.
Scanning electron microscopy (SEM). Morphological analysis of the nanostructures was carried out using field emission scanning electron microscope (Nova NanoSem 450-FEI). Briefly, 10μL of sample solutions at 100.0 (50 mg/mL for H + -F6-O − ) and 5.0 mg/mL were dropped off on aluminum stub using a graphite adhesive tape. A thin coat of Au and Pd was sputtered at a current of 20 mA for 120 s. The sputter coated samples were then introduced into the specimen chamber and the images were acquired at an accelerating voltage of 2-5 kV, spot 3, through the Everhart Thornley Detector (ETD) and the Through the Lens Detector (TLD).
Fluorescence microscopy. 10  Fluorescence measurements. The as-prepared samples were spectrally characterized using a Jasco FP-8300 spectrofluorometer equipped with an ILF-835 integrating sphere, which allowed us to evaluate the photoluminescence quantum yield (PLQY) with the absolute method thereby overcoming the issue of unavoidable differences in the relative emission intensities arising from poorly reproducible drop-casted peptide films. Fluorescence spectra of the four F6 films were acquired in a window up to 650 nm, probing the 330-430 nm excitation range at 10 nm steps. All acquisitions were averaged over five measurements after spectral correction.
Photodegradation of peptide films. PL spectra were collected after continuous exposure to the 330 nm excitation line of the instrument, keeping the sample mounted in the integrating sphere and the excitation shutter open, for up to 180 min (at an incident power of ~ 820 μW).