The peptides (Table 1) were prepared by solid phase peptide synthesis (SPPS) using the Fmoc/t-Bu strategy. All the peptides were synthesized manually on Rink-Amide resin (0.217 g, loading 0.69 mmol/g), according to the procedure described in the SI. The coupling systems used was 3:5:5:8 equivalents of Fmoc-amino acid/HOBt/HBTU/DIEA for the first two amino acids and for the third ones of peptides 1, 4, 6, 9 and 10, whereas the other couplings were performed using 3:5:5 equivalents of Fmoc-amino acid/HBTU/DIC. The coupling with biotin was performed directly on resin (see SI), using as coupling system 10:10:10 equivalents of biotin/DIC/Oxyma Pure. After each coupling, a Kaiser test was performed to check the completeness of the reaction. The peptidyl bound resins were cleaved with a mixture of TFA/phenol/H₂O/thioanisole/TIPS (84:5:5:5:1 v/v/v/v) for 2 h under magnetic stirring. After the work-up, the crude peptides were purified by semi-preparative RP-HPLC (see SI for the details) and lyophilized. Finally, the pure peptides were analyzed by ESI-MS, RP-HPLC, ¹H and ¹³C-NMR spectroscopy.

AuNPs with 20 nm diameter are prepared by a modified Turkevich-Frens method (Turkevich et al., 1951; Giljohann et al., 2010). A boiling solution of HAuCl₄ trihydrate (15 mg, 40 μmol) in 150 ml of Milli-Q water is quickly mixed with 4.5 ml of warm (60–80°C) 1% (wt) aqueous solution of trisodium citrate, followed by reflux for 60 min. The resulting nanoparticle suspension was then allowed to cool down to room temperature and stirred overnight.

The concentration c of AuNPs in solution (0.9 nM) was calculated from absorption (A) at 450 nm for a standard path length l of 1 cm according to Lambert-Beer Law (Eq.2.1). c = A 450/ ε 450 ⋅ l (2.1)

ε₄₅₀ is the molar extinction coefficient of AuNPs at 450 nm in water and it depends on the size of nanoparticles. The theoretical value for AuNPs with diameter of 20 nm is 5.41 × 10⁸ Mcm according to the literature data (Haiss et al., 2007). From the concentration, it is possible to calculate the number of particles in solution. All experiments were performed on 2 ml of AuNPs dispersion ( V ), which contains 1.08 × 10¹³ particles according to the following equation: N AuNPs = c . N A V ⋅ V (2.2) where N _AV is Avagadro constant (6.02 × 10²³ mol⁻¹).

The functionalization of AuNPs with peptides is carried out following a procedure described by (Levy et al., 2004).

All peptides were dissolved in DMSO to give 0.5 mM stock solution. Since 2 ml of AuNPs solution contains 1.08 × 10¹³ nanoparticles, 3.5 × 10¹⁶ peptide molecules (N_p) (117 μl of 0.5 mM solution) were calculated to exchange all citrate ions and completely cover nanoparticles. An excess of peptide (175 μl, 1.5-fold excess) has been used to facilitate the exchange reaction and to be sure to create the self-assembled monolayer.

For the AuNPs functionalization, the proper peptide was added to 2 ml of 0.9 nM AuNPs solution under vigorous stirring, and the mixture was left under stirring overnight. Then peptide and citrate excess was washed out by centrifugation at 14,000 rpm for 15 min.

AuNPs stability was studied by UV-vis spectroscopy and DLS by incubating nanoparticles in different environments: water, phosphate buffer saline (PBS, pH 7.4), Human Serum (HS) 10% in water or HS 100%. An aliquot of concentrated nanoparticles was dispersed in the choosen medium and the stability was evaluated over the time. In the case of 100% HS, an aliquot of concentrated nanoparticles was dispersed in the medium and incubated for 24 h at 37°C, followed by centrifugation at 14,000 rpm for 15 min and redespersion in water. Subsequently, the absorption spectra of nanoparticles in the 400–700 nm range were measured.

An aqueous solution of avidin (0.125 mg/ml or 0.0125 mg/ml, 20 μL) was added to 2 ml (0.9 nM) of biotinylated peptides AuNPs suspension. The reaction mixture was mixed thoroughly for 2 h at room temperature. Subsequently, functionalized AuNPs were washed by centrifugation (14,000 rpm for 15 min) and the supernatant was analyzed by BCA protein assay (Sigma Aldrich) and HABA test.

Agarose gel electrophoresis of AuNPs was carried out using a Biorad electrophoresis gel apparatus. Small aliquots (16 μl) of the conjugates were supplemented with 30% glycerol (4 μL) and loaded on 0.75% agarose gel in 1X Tris-acetic acid-EDTA (TAE) buffer, and ran for 1–2 h at 80 V. After migration, nanoparticles were visible with the unaided eye as a red band.

The rational design of the peptides considered the need of a strong affinity for gold and a good balance between hydrophobicity and hydrophilicity to ensure both self-assembly and colloidal stabilization. To this purpose, Ala-Aib containing sequences seemed particularly profitable. Indeed, they possess helical conformation, can form stable self-assemblies in water (Locarno et al., 2020), and can be easily functionalized with a thiol group allowing gold binding. We thus prepared a small peptide library (Table 1) by modulating: 1) the sulphur anchor point (at C- or N-terminus, using respectively Fmoc-Cys (Trt)-OH or 3-(Tritylthio)propionic acid); 2) the peptide length; 3) the amino acid composition. In this last case, we investigated the effect of Ac5c insertion. Ac5c is a more constrained C^α-tetrasubstituted AA with respect to Aib that, although its steric hindrance, it seems not decreasing the peptide self-assembly behavior (Locarno et al., 2020). Finally, peptides sequence one and six were functionalized with biotin at N-terminus allowing avidin protein binding via molecular recognition.

All the peptides (Table 1) were prepared by solid phase peptide synthesis (SPPS) (see SI for the details) using the Fmoc/t-Bu strategyon Rink-Amide resin (0.217 g, loading 0.69 mmol/g). Different coupling systems (HOBt/HBTU and DIPEA oror HOBt/DIC or DIC/Oxyma Pure) were used depending on the type, the position and the steric hinderance offered by the single functionality (natural AA, Cα-tetrasubstituted AA, thiol gold binding moiety and biotin). The purity of the synthetized peptides was calculated by analytical RP-HPLC, while their identity was confirmed by NMR and mass spectrometry Helical propensity was finally assessed by FT-IR on selected peptides (see SI).

Citrate-stabilized AuNPs were prepared by modified Turkevich-Frens method (Turkevich et al., 1951; Giljohann et al., 2010) and were characterized by UV-spectroscopy. The absorption spectrum in water of synthesized AuNPs are reported in Figure 1A. SPR peak occurred at 518 nm and the calculated A_spr/A₄₅₀ value is 1.73 which corresponds to an average size of 20 nm (Haiss et al., 2007; Rahman, 2016). This value was also confirmed by DLS (Figure 1B, 22.7 ± 0.2 nm, PdI 0.2 ± 0.02).

The most common approach to functionalize AuNPs is the ligand exchange method involving the displacement of one ligand for another. In the presence of molecules bearing a thiol group, citrate-capped AuNPs undergo ligand exchange giving functionalized nanoparticles (Zong et al., 2017).

In order to establish the amount of peptide needed for the functionalization, the circular peptide axial footprint, which is defined as the average area each peptide occupies on the nanoparticle surface, was measured. The peptide area (A_p = 0.38 nm²) has been approximated according to Eq. 3.1, measuring the peptide axial average diameter 4) from the crystal structure of the pentapeptide analogue Ac-Ala-Aib-Ala-Aib-Ala-NH₂, which has been found of 7 Å (Locarno et al., 2020): A p = π ( d / 2 ) 2 (3.1)

The ratio between the AuNPs and the peptide areas (K = A_AuNPs/A_p) gives the average number of peptides per particle with a 20 nm radius. Some assumptions regarding surface area were made: first, nanoparticles were modelled as a perfect sphere; second, the numbers of particles larger and smaller than the average particle size were assumed to be equal; third, peptides were assumed to be evenly distributed on the nanoparticle surface. According to these assumptions, K is 3,265 peptides/AuNP. Conversely, in the ligand exchange experiments we used 175ul of a 0.5 mM peptide DMSO solution (see Gold Nanoparticles Functionalization).

The effect of peptides functionalization on AuNPs was evaluated by UV-Vis and DLS. The functionalization of AuNPs modifies the position and width of SPR absorption since this reaction changes the chemical and physical environment of the particle and consequently, its optical properties. The monitoring of the SPR signal is considered an easy method to control the functionalization reaction and the stability of AuNPs (Krpetic et al., 2011). In the case of Ala-Aib peptides 1-5, the redshift of UV−Vis plasmon absorption from 518 to 525 nm, is the first confirmation of the AuNPs functionalization (Figure 2A). Moreover, changes in size-distribution features of nanoparticles determined by DLS were observed (increased diameters from 22 to 26–30 nm, Figure 2B). Only in the case of tripeptide four aggregation was observed (Figure 2B). This is a consequence of the presence of Cysteine at C-terminus. In fact, as demonstrated below, functionalization of AuNPs by peptides with Cysteine instead of thiopropionic acid, gives less stable AuNPs. Figure 2 shows also the effect of peptide length on AuNPs stability. For peptide 4, which is the shortest peptide, the aggregation parameters increased rapidly, as shown in Figure 2A,B As the length of the peptide increased, the aggregation was less evident, suggesting a direct correlation between peptide length and stability of the peptide-capped AuNPs. The colloidal stability of capped AuNPs after freeze-drying was then studied to evaluate the possibility to store them as a powder. This is an important parameter since very few AuNPs preparations beside CALNN-capped AuNPs can be stored in dry state and redispersed in water (Levy et al., 2004). As shown in Figure 2 C-D, only 3@AuNPs maintained the same hydrodynamic diameter after redispersion in water, while for 5@AuNPs aggregation started to occur. However, aggregation was more evident in case of 1@AuNPs and 2@AuNPs, both containing Cysteine residue. These data suggested that in freeze-drying process, the gold anchor point plays a more crucial role than the peptide length.

Taken together all the data we can thus postulate that pentapeptides and thiopropionic acid are the best combination for inhibiting AuNPs aggregation.

Peptides 6, 7 and 8 containing the hindered Ac5c AA were then evaluated by UV-Vis and DLS (Figure 3). Also in this case, when Cysteine is at C-terminus (peptide 6), AuNPs showed evident sign of aggregation: in addition to the color change of suspension, the SPR peak resulted to be broader. Moreover, the hydrodynamic diameter increased more than 20 nm (Figures 3A,B). Peptides 7 and 8, instead, were able to stabilize the particles. However, after freeze-drying 8@AuNPs were completely destabilized and big aggregates were observed in DLS profiles (Figure 3D). On the other hand, 7@AuNPs were found more stable, although a partial aggregation started to occur, since a small second population was observed in DLS experiments (Figures 3C,D). These results suggested that the steric hindrance of Ac5c could have negative influence on the AuNPs stability.

From UV-Vis and DLS data, peptides 3 and 7, both containing thiopropionic acid at N-terminus, represented the best candidates for inhibiting AuNPs aggregation propensity before and after freeze-drying. For this reason, 3@AuNPs and 7@AuNPs were selected for a further functionalization with biotinylated peptides 9 and 10 (Table 1), respectively. Also in this case, the ligand exchange method was employed. Three different ratios of non-biotinylated 3 and 7 peptides and biotinylated 9 and 10 ones were considered (80:20, 90:10, 95:5). All the analyzed ratios showed the ability to stabilize AuNPs in solution (see Supplementary Figure S2 in SI), so the 80:20 one, having the highest percentage of biotinylated peptide (20%), was selected for further evaluations.

Also in this case, the colloidal stability after freeze-drying of functionalized AuNPs was evaluated. As shown in Figure 5, only 3 + 9@AuNPs show good colloidal stability while 7 + 10@AuNPs aggregated. To overcome colloidal instability of 7 + 10@AuNPs, Ac5c peptide 10 was replaced Aib peptide 9 (7 + 9@AuNPs). This substitution increased colloidal stabilization, as demostrated by the SPR peak that remained unchanged. This finding confirmed the higher efficacy of Aib containing peptides in inhibiting AuNPs aggregation than Ac5c ones. However, the DLS size distribution analysis showed, beyond the expected peak around 20–30 nm, a second peak of more than 200 nm, suggesting a partial aggregation of the particles. (Figure 4).

TEM analysis on biotinylated 3 + 9@AuNPs, 7 + 10@AuNPs, and 7 + 9@AuNPs was then undertaken. As reported in Figure 5, a better NPs dispersion was observed for 3 + 9@AuNPs (Figure 5A), while in the case of 7 + 10@AuNPs the presence of aggregates was observed, confirming the UV-Vis and DLS data.

The colloidal stability of AuNPs in a physiological environment is a fundamental prerequisite for biological applications. Three different mediums (water, PBS, 10% HS in water and 100% HS) were then considered and the peptide stability of functionalized-AuNPs was evaluated against naked AuNPs. Apart from the 7 + 10@AuNPs in PBS, in all other cases, the functionalization with peptides ensured the dispersion stability of AuNPs and the DLS profile remained the same for weeks (see Supplementary Figure S3 in SI). As expected, in water, all functionalized-AuNPs showed a greater stability than naked AuNPs (Figure 6A). The major differences in dispersion stability were found in PBS.

PBS is a buffer solution frequently utilized as stimulated human body fluid in biological research. As already known from the literature (Du et al., 2012), when naked AuNPs were dispersed in PBS, the relatively high ionic strength could cause particle destabilization leading both to aggregation and adhesion to surrounding substrates through van der Waals forces. However, the functionalization of AuNPs with peptides 3 and 7 and with the mixture of 3 + 9 induced a great improvement of particle stability in PBS (Figure 6B), since no detectable changes in the UV-Vis absorbance and DLS spectra were observed. This is due to the local change of dielectric permittivity, resulting from the formation of a peptide layer on the surface of the particles. Thus, the peptide provides immediate protection to the particles from the aggregation. On the other hand, when the AuNPs were functionalized with the mixture of 7 + 10, the destabilization occurred. Also in this case, the substitution of peptide 10 with Aib peptide 9 (mixture 7 + 9) increased the particles stability in PBS, confirming the detrimental effect of Ac5c in the peptide sequence. In HS, the abundance of serum proteins hampered DLS analysis. Spectroscopic analyses, instead, showed interesting results. In fact, when uncapped AuNPs were dispersed in HS, both 10% in water and 100%, (Figures 6C,D) a shift of UV-Vis plasmon absorption of 10 nm (from 518 to 529 nm) was observed suggesting an increment of particle size due to the great amount of serum proteins conjugation to the particle surface. The shift was, instead, absent in all the peptide capped-AuNPs, that showed a better stability both in 10 and 100% HS. However, the broader SPR peak suggested the possibility of unspecific binding of proteins, also in presence of functionalized-AuNPs.

The “aging” of AuNPs over the time was recorded in all the considered mediums (see Supplementary Figure S3 in SI). The stability study over 2 months in water showed a gradual degradation of functionalized AuNPs since the SPR peak decreased gradually as well as the count rate measured by DLS analysis. However, AuNPs aggregation was prevented: DLS analyses showed at any time points the same peak at around 20 nm. The degradation rate was faster in biotinylated than in non-biotinylated AuNPs. Same results were obtained also in 10% aqueous HS solution. In PBS, instead, particle destabilization occurred after 15 days, when adhesion to surrounding substrates was observed.

Finally, agarose gel electrophoresis (Figure 7) was conducted on both peptide capped- and uncapped-AuNPs in order to confirm the previous data. All AuNPs were dispersed in water or incubated in HS 100% for 24 h at 37°C, followed by centrifugation in order to remove the incubation medium. Then, the particles were supplemented with 30% glycerol and loaded on agarose gel. Since the electrophoretic motilities are dependent on charge, size and isoelectric point of the particle, there was a distinct change in mobility of unstable and stable AuNPs. It is not surprising that AuNPs washed from the excess of citrate did not migrate over the gel, since they lost the negative charge of citrate, almost reaching a neutral charge. The dark color of AuNPs corroborated the formation of big aggregates and the loss of stability. All the peptide-functionalized AuNPs migrated to the cathode but, also in this case,7 + 10@AuNPs showed only a partial stability since a dark band was visible in the well of the gel and the red band was consequently less intensive. The substitution of 10 with 9 (7 + 9@AuNPs), once again, overcame the instability, confirming the data previous reported. Finally, when capped-AuNPs were dispersed in HS 100%, the gel migration decreased due to the high concentration of proteins that masked the particles’ charge. The capped-AuNPs appeared stable but, probably, a partial unspecific binding of proteins occurred.

First, the interaction of biotinylated peptides with avidin was assessed by the HABA (4′-hydroxyazobenzene-2-carboxylic acid) test that allowed also to evaluate the number of peptides per particle (see Supplementary Figure S2 and text in SI for details). HABA is a dye that enables spectrophotometric estimation of biotinylation levels of labeled molecules. A total of peptides was calculated to be 4.0 × 10¹⁵ biotinylated molecules for 3 + 9@AuNPs and 3.6 × 10¹⁵ biotinylated molecules for 7 + 10@AuNPs. This translates to 370 molecules of peptide 9 per nanoparticle and to 334 molecules of peptide 10 per nanoparticle.

The specific binding of avidin to biotinylated AuNPs was further evaluated with BCA (Bicinchoninic Acid) Protein assay. The principle of this method is the biuret reaction by which proteins can reduce Cu⁺² to Cu⁺¹ in an alkaline solution and results in a purple color formation by bicinchoninic acid. Avidin aqueous solution (0.125 mg/ml) was added to biotinylated AuNPs (3 + 9@AuNPs, 7 + 10@AuNPs or 7 + 9@AuNPs) and to non-biotinylated 3@AuNPs or 7@AuNPs. The mixtures were shacked for 2 h. Then, the mixture was centrifuged and the supernatant was analyzed calorimetrically by BCA assay. As shown in Figure 8, the presence of avidin in blank supernatant was evident both optically (the supernatant solutions analyzed with BCA assay acquired the typical purple color) and spectrophotometrically, while the supernatants of biotinylated AuNPs were completely free from avidin. These results suggested that nonspecific interaction of avidin with peptides capped-AuNPs was prevented and the avidin specific interaction occurred only in presence of biotin.

The addition of avidin to the biotinylated AuNPs resulted in particle aggregation and the redispersion in water was no possible, thus confirming the accessibility and recognition of the biotin moieties. This aggregation happened because avidin can bind up four biotin molecules since it has two pairs of biotin binding sites on opposite sides of the molecule (Aslan et al., 2004). The aggregate grows only when different biotinylated AuNPs cross-linked with the same avidin molecule. Therefore, the concentration of avidin solution was decreased up to 0.0125 mg/ml to reduce the possibility of cross-linking between the particles. The avidin-biotinylated AuNPs interaction was analyzed by UV-Vis and DLS (Figure 9). The UV-Vis measurements (Figure 9A) gave a single sharp SPR peak signifying that the integrity of the biotinylated AuNPs was maintained. On the other hand, DLS analysis (Figure 9B) showed two peaks: one at ∼ 20 nm corresponding to the peak of AuNPs, the other at ∼ 200 nm suggesting a partial aggregation probably induced by biotin-avidin interaction. The aggregation is more evident with 7 + 9@AuNPs and 3 + 9@AuNPs, both with the biotinylated Aib peptide 9.

Finally, TEM analysis was performed on avidin bound biotinylated 3 + 9@AuNPs, 7 + 10@AuNPs, and 7 + 9@AuNPs confirming the formation of large NPs aggregates (Figure 10) and thus avidin binding. As previously observed in DLS experiments, the aggregates were particularly relevant in the case of 7 + 9@AuNPs.