pH-Dependent Synthesis of Anisotropic Gold Nanostructures by Bioinspired Cysteine-Containing Peptides.

In the present study, alkaline peptides AAAXCX (X = lysine or arginine residues) were designed based on the conserved motif of the enzyme thioredoxin and used for the synthesis of gold nanoparticles (GNPs) in the pH range of 2-11. These peptides were compared with free cysteine, the counterpart acidic peptides AAAECE and γ-ECG (glutathione), and the neutral peptide AAAACA. The objective was to investigate the effect of the amino acids neighboring a cysteine residue on the pH-dependent synthesis of gold nanocrystals. Kohn-Sham density functional theory (KS-DFT) calculations indicated an increase in the reducing capacity of AAAKCK favored by the successive deprotonation of their ionizable groups at increasing pH values. Experimentally, it was observed that gold speciation and the peptide structure also have a strong influence on the synthesis and stabilization of GNPs. AAAKCK produced GNPs at room temperature, in the whole investigated pH range. By contrast, alkaline pH was the best condition for the synthesis of GNP assisted by the AAARCR peptide. The acidic peptides produced GNPs only in the presence of polyethylene glycol, and the synthesis using AAAECE and γ-ECG also required heating. The ionization state of AAAKCK had a strong influence on the preferential growth of the GNPs. Therefore, pH had a remarkable effect on the synthesis, kinetics, size, shape, and polydispersity of GNPs produced using AAAKCK. The AAAKCK peptide produced anisotropic decahedral and platelike nanocrystals at acidic pH values and spherical GNPs at alkaline pH values. Both alkaline peptides were also efficient capping agents for GNPs, but they produced a significant difference in the zeta potential, probably because of different orientations on the gold surface.

Introduction

There is a growing interest in exploiting the ability to synthesize gold nanoparticles (GNPs) of controlled size and shape mediated by amino acids and peptides.[1−7] For many years, numerous synthetic strategies developed for the synthesis of GNPs have frequently required the use of harsh reagents for controlling the particle morphology. The search for eco-friendly conditions showed that peptides are an interesting alternative for reducing the environmental impacts during the synthesis of GNPs. Also, the capping of GNPs provided by peptides assures efficient delivery and biocompatibility that can contribute to the use of theranostic properties of these materials.[8] Moreover, the functional groups available in the peptides, such as −SH and −NH2, allow a fine control of the synthesis for achieving the desired functionality for the material. Peptides provide templates that direct the growth and shape of the metal nanoparticles as well as their capping.[4,5] Particularly, the presence of a cysteine residue in a peptide provides a mild reducing agent for biomineralization and an efficient anchor for the peptide on the GNP surface.[9] Therefore, a peptide containing a cysteine residue is a promising compound for the controlled synthesis and capping of GNPs. The manipulation of the reaction conditions and a rational choice of the amino acids arrangement (bioinspired sequences) make feasible the specific directing of the GNP spatial growth assisted by peptides. Bioinspired peptides have been shown to be useful for the assembly and synthesis of gold nanostructures.

Hwang and co-workers developed the strategy of using peptides conjugated with nonpeptide organic molecules to provide the characteristics of a programmable self-assembly and inorganic recognition.[9] These motifs conduct the one-pot synthesis, arrangements of regular structure, and the formation of superstructures of high complexity. The authors demonstrated that some members of this class of molecules used for the synthesis and assembly of nanoparticles could be used to prepare 1D nanoparticle superstructures and GNP superstructures at the scale of 100 nm formed by the assembly of spherical structures. One particularly useful feature of this methodology is that small changes in the composition and sequence of the peptide conjugate can dramatically influence the structure of the resulting nanoparticle assembly.[9] Brown et al. created a genetic system in the bacterium Escherichia coli to produce nonbiological proteins.[10] The genetic system allowed the production of proteins tailored to control the crystal growth. Si and Mandal used synthetic oligopeptides with a tryptophan residue at the C-terminal for the synthesis of gold and silver nanoparticles.[11] The authors corroborated tryptophan as the reducing agent for the gold and silver ions by the identification of the corresponding products of the peptide oxidation: ditryptophan, kynurenine, and some cross-linked products. These results are consistent with the formation of free radical intermediates resulting from the oxidation promoted by the metal ions. Kim et al. showed that TGTSVLIATPYV peptide, named Midas-2, formed gold nanostructures of diverse shapes (nanoparticles, nanowires, nanoribbons, kite and tail structures) and sizes (nanometer-thick platelets) by changing the reaction conditions, including pH and the concentration of gold salt precursor.[7] For a more extensive understanding of the state of the art of the use of peptides for the GNP synthesis, we recommend the reviews by Dickerson et al. and Sarikaya et al.[4,5]

In the present study, we designed two peptides inspired on the conserved motif of a protein involved in cellular redox processes, thioredoxin (Trx), with the objective to ensure a highly reducing moiety for gold ions. Thioredoxins are characterized by the presence of positively charged amino acids flanking the nucleophilic cysteine (CXXC motif) (Figure ), which leads to lowering its pKa and increasing their reactivity.[12,13] The peptides were designed with the AAAXCX sequence, X being lysine residues (K) or arginine residues (R), and compared it with the counterpart acidic AAAECE peptide and the biological γ-ECG peptide, which is known as glutathione (GSH). The neutral peptide AAAACA and free cysteine were also used. Only the alkaline AAAKCK multifunctional peptide was efficient to assist and stabilize GNPs at the pH range of 2–11, and the results will be discussed herein.

Figure 1

Cartoon representation of the crystal structure of thioredoxin (PDB: 1ERT) and the zoomed-in view of the catalytic motif (lysine residue in blue and cysteine residue in red).

Experimental Section

Chemicals

Hydrogen tetrachloroaurate (III) (HAuCl4·3H2O) and potassium phosphate monobasic and dibasic were purchased from Sigma-Aldrich. All of the chemicals were used as received. Finally, pure water (18.2 MU cm; Millipore) was used for all experiments.

Synthesis of Peptides

An automated bench-top simultaneous multiple solid-phase peptide synthesizer (PSSM 8 system from Shimadzu) was used for the solid-phase synthesis of all peptides by the Fmoc procedure.[14] The final peptides were deprotected in TFA and purified by semipreparative high-performance liquid chromatography (HPLC) using an Econosil C-18 column (10 μ, 22.5 × 250 mm2) and a two-solvent system consisting of (A) trifluoroacetic acid (TFA)/H2O (1:1000) and (B) TFA/acetonitrile (ACN)/H2O (1:900:100). The column was eluted at a flow rate of 5 mL/min with a 10%–50% gradient of solvent B over 30 min or with a 30%–60% gradient over 45 min. Analytical HPLC was performed using a binary HPLC system from Shimadzu with an SPD-10AV Shimadzu UV–vis detector coupled to an Ultrasphere C-18 column (5 μ, 4.6 × 150 mm), which was eluted with solvent systems A1 (H3PO4/H2O, 1:1000) and B1 (ACN/H2O/H3PO4, 900:100:1) at a flow rate of 1.0 mL/min and a 10%–80% gradient of B1 over 20 min. The HPLC column elutes were monitored for their absorbance at 220 nm. The molecular weight and purity of the synthesized peptides were checked using a MALDI-TOF mass spectrometry (Bruker Daltons) or electron spray ionization source LC/MS-2010 (Shimadzu).[14]

Synthesis of Gold Nanoparticles

An aqueous solution of GNPs was prepared by reducing tetrachloroauric acid (HAuCl4) with a peptide sequence. For this, 30 μL of 10 mM HAuCl4 was added to 270 μL of phosphate buffer solution (20 mM) and was reduced by 30 μL of each peptide solution (7 mM). To evaluate the effect of pH on the synthesis, the buffer solutions were adjusted using HCl or 0.1 M NaOH to 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, and 11.0. The system was kept at room temperature without the influence of stirring. The colloidal gold solution was subjected to centrifugation, and the conjugate was washed with deionized water to remove any uncoordinated peptides. The peptide–GNP systems were collected by centrifugation (5000 rpm, 25 °C, and 7 min) and subsequently washed three times to remove free peptides or other compounds present in the reaction solution.

Characterization

The formation of GNPs was further confirmed by UV–vis spectroscopic measurement of the reaction mixture solution, which was performed on a Thermo Scientific Evolution Array UV–visible spectrophotometer, employing a 1 cm quartz cuvette with ultrapure water as the reference. The zeta potential (ζ) measurements were carried out using a Zetasizer Nano ZS, Malvern Instruments, at the temperature of 25 °C to investigate the superficial charge. Fourier transform Raman (FT-Raman) spectroscopic measurement was carried out to identify the potential functional groups responsible for the reduction of gold ions and stabilization of the synthesized GNPs, which was recorded on a FT-Raman MultiRAM spectrometer (Bruker Optics). Transmission electron microscopy (TEM) analysis was conducted using a Jeol JEM-100CX microscope operating at an accelerating voltage of 80 kV to determine the morphology and dimension of GNPs. The samples for TEM study were prepared onto a carbon-coated Cu grid, which was allowed to dry under ambient conditions. To understand the crystalline characteristics, X-ray powder diffraction data were recorded at room temperature on an STADI-P powder diffractometer in transmission geometry by using a Mo Kα1 (λ = 0.7093 Å) wavelength selected using a curved Ge (111) crystal, with a tube voltage of 50 kV and a current of 40 mA.

Results and Discussion

The sequence AAAXCX was designed to investigate the effect of cysteine vicinity on the capacity of these peptides for assisting and stabilizing GNPs. Literature data show that the alanine residues are not directly involved in the synthesis of GNPs and play structural and stabilizing roles in the peptide sequence.[15] The cysteine thiol group of the peptide AAAXCX is expected to have a lowered pKa value when X = K or R that was corroborated using peptide titration. The AAAKCK peptide exhibited pKa values of 2.2, 7.7, and 9.7. These pKa values can be assigned, respectively, to the terminal C, the cysteine thiol group, and the amino groups of the peptide (Figure ). The pKa value of alanine α-amino group is 2.18, and a different value is unexpected for this group in the peptide structure.[16] The pKa value of 9.7 is probably a mean value that resulted from the contribution of the three amino groups of the peptide, that is, the terminal N and the lateral chains of the two lysine residues.

Figure 2

Changes in the net charge of AAAKCK peptide sequence during the titration using HCl and NaOH. The pKa values of the ionizable groups are indicated beside the arrows.

The AAAKCK peptide was used as a reducing and stabilizing agent for the synthesis of GNP, at room temperature within a pH range of 2–11 with 1 unit of interval among the values. The formation of the GNPs was formerly evidenced by the color change of the solution from pale yellow to purple, pink, or red under different reaction conditions. The presence of GNPs was then corroborated by the appearance of the surface plasmon resonance (SPR) at the wavelength range of 500–600 nm, which is typical of GNP plasmon absorbance.[17]Figure A–D shows the spectra obtained during the synthesis of GNPs using the AAAKCK peptide at pH 4, 6, 9, and 11, respectively. These spectra are representative of the GNP SPR obtained at the acidic and alkaline pH range. The SPR of GNPs produced using AAAKCK at acidic pH is featured as narrow bands, peaking at the wavelength range of 530–550 nm. The GNPs obtained at pH 6 also exhibit a broad band that peaks at 800 nm. This result is consistent with the presence of anisotropic particles. The SPR bands of GNPs obtained at alkaline pH are broad bands peaking at the wavelength range of 550–600 nm.

Figure 3

Time-resolved UV–vis spectra obtained during the synthesis of GNPs using AAAKCK at pH (a) 4.0, (b) 6.0, (c) 9.0, and (d) 11.0.

The kinetic analysis shows that the GNPs synthesized at the acidic pH using AAAKCK (Figure square symbols) exhibited SPR bands with the maximal intensity around 0.2. However, an increasing lag time to the maximal intensity occurred concomitant with the increase in the pH (white, light gray, and gray symbols for pH 4, 5, and 6, respectively). The maximal intensity of the GNP SPR bands synthesized at the pH range of 7–11 was around 0.3. In the alkaline medium, an increasing lag time for attaining the maximal intensity of the SPR band was observed at increasing pH values.

Figure 4

pH-dependent kinetics of the GNP synthesis using AAAKCK in the pH range of 4–10 as indicated in the figure.

Figure shows the pH-dependent intensity of SPR bands at 560 nm after 10 and 40 min from the starting of the synthesis reactions (blue and red lines and symbols, respectively). Also, in Figure , the black line and symbols show the pH-dependent time necessary for the maximal intensity of SPR band to be attained (lag time). The maximal SPR intensity was attained before 10 min of incubation at acidic and neutral pH values and in all conditions after 40 min of incubation. According to Figure , the influence of three ionizable groups on the synthesis of GNPs using AAAKCK is evident. The turning point at pH 3.6 reflects, more probably, gold speciation, that is, the hydrolysis of HAuCl4– (eq ) rather than the deprotonation of terminal COOH group (pKa 2.2).According to Moreau et al.,[18] equimolar concentrations of AuCl4 and the water-coordinated species AuCl3(OH–) + AuCl3(OH2) are expected at pH 3. The authors stated that 20% of AuCl3(X) species at pH 3.0 are AuCl3(OH–) and that 30% of AuCl3(X) species are AuCl3(OH2). The pKa value of gold speciation can suffer a slight deviation influenced by the presence of peptide and phosphate in the medium. In fact, recently a significant influence of gold speciation on the synthesis of GNPs using albumins was reported.[19] The pKa value of 6.2 probably reflects the deprotonation of the lateral chains of the amino acid involved in the reduction of gold ions. Two groups might be responsible for gold reduction: the thiol group of cysteine and the amino group of lysines. The value of 6.2 is significantly lower than the pKa value of these groups in the free amino acids: 8.3 and 10.5, respectively. However, the protonated amino groups of lysine residues that are neighboring cysteine could lower the pKa value of the thiol group. It is also probable that one of the two lysine residues put each other in proximity also has a lower pKa value of the lateral chain. The titration of the peptide showed pKa values of 7.7 and 9.7, which are consistent with −SH and −NH3+ with lowered pKa values. The coordination of these ionizable groups with gold ions can respond to the more exacerbated lowering of their pKa values to 6.2 and 9.0. It is important to note that the value of 9.7 obtained from the peptide titration is an average value of the ε and terminal α-amino groups. However, the pKa value of around 9.0 obtained from the rate of synthesis refers to the ionizable alkaline groups that participate in the synthesis and stabilization processes.

Figure 5

pH-dependent maximal intensity of the SPR at 560 nm after 10 min (blue line and symbols) and 40 min (red line and symbols). The black line and symbols indicate the pH-dependent delay time to the maximal SPR intensity.

The pH of the medium has a significant influence not only on the net charge and structure of the peptides but also on the speciation of aqueous HAuCl4. The speciation of gold influences the oxidizing capacity of the species. A previous study demonstrated that the reduction potential of the aqueous HAuCl4 solution decreased with the increase in pH.[20] The voltammograms of the aqueous HAuCl4 solution measured at pH 2.91, 6.16, and 8.01 exhibited reduction potentials of 0.66, 0.59, and 0.53 V, respectively, and no reduction peak at pH 10.35.[20] It is important to note that the values of the reduction potential of gold ion species at different pH values are not extremely different. Thus, although the reactivity of gold species decreases with the increasing pH, the combination of fast rate of synthesis and high intensity of SPR observed at pH 7 cannot be explained by gold reactivity. Probably, the displacement of water from AuCl3(OH–) is more favorable than that of one Cl– from AuCl4– for the coordination with the peptide.

To better understand the factors influencing the synthesis of GNP using AAAKCK, ab initio calculations based on total Kohn–Sham density functional theory (KS-DFT)[21,22] with local density approximation (LDA)[23] were performed to determine the charge reduction of the peptide complexed with AuCl4–. Our results were obtained using the projector augmented wave (PAW) method,[24] with an energy cutoff of 465.0 eV as implemented in the VASP code.[25,26] The relaxation was allowed until the forces were smaller than 0.02 eV/Å in a 15 × 25 × 25 Å supercell dimension for x, y, and z directions, respectively. For this calculation, three configurations of the peptide–AuCl4– complex were considered: protonated thiol (−SH) with a neighboring positively charged amino group (acidic form), deprotonated thiol (−S–) with a neighboring positively charged amino group (neutral form), and deprotonated thiol with a neighboring amino group (basic form). The first configuration was obtained by abstracting one electron of the peptide–AuCl4–complex, and the last configuration was obtained by abstracting one proton. The peptide–AuCl4– complex was studied using Bader analysis, in which the space was divided into regions limited by surfaces where the gradient of electron density had no normal component.[27] These configurations allowed the decomposition of the electron density of each atom. Using the Bader analysis, the total approximated electronic charge was obtained by the sum of the corresponding points of the grid.[28] The sum of the Bader electronic charges for the peptide–AuCl4–complexes was 0.350, 0.391, and 0.479, for the acidic, neutral, and basic configurations, respectively. These values corroborated that the deprotonation of the thiol group at pH values above its pKa value—lowered by a neighboring positively charged amino group—increases the reducing capacity of the peptide. However, at highly alkaline pH values, the deprotonation of the neighboring amino group increased the reducing capacity of the peptide once again. Solution pH also exhibited a great influence on the characteristics and particle size distributions of the GNPs. Figure shows the representative electronic microscopy images of GNPs obtained under different conditions, which illustrates the formation of gold nanostructures with various shapes. The histograms shown in panels D and E shows the shape distributions of the GNPs synthesized at pH 4.0 and 6.0, respectively. Panel F shows the size distribution of GNPs synthesized at pH 9.0. The histograms D and E evidenced the shape polydispersity of the GNPs synthesized at pH 4.0 and 6.0. GNPs synthesized at pH 4.0 are predominantly icosahedron and decahedron (71.4%). Hexagonal plates were 6.1%, and other shapes constituted 22.5% of the GNPs synthesized at pH 4.0. GNPs synthesized at pH 6.0 exhibited a more pronounced polydispersity of shapes and sizes (Figure B); similar morphologies for biosynthetic GNPs were observed previously.[29−32] This result is consistent with the presence of two SPR bands of GNPs produced at pH 6.0 as shown in Figure B. In this condition, that is, at pH 6.0, icosahedron and decahedron remained as the predominant shapes (∼68%), the contribution of other irregular shapes decreased to ∼11%, and the bipyramidal and spherical GNPs—absent in the sample produced at pH 4.0—constituted ∼13%. Icosahedron, decahedron, and plates were produced at pH 4.0 and 6.0 but with higher size polydispersity in the latter pH. The respective size ranges of the icosahedron, decahedron, and plates produced at pH 4.0 were 55–90, 50–100, and 65–125 nm, respectively. For the synthesis performed at pH 6.0, the size ranges of the icosahedron, decahedron, and plates were 15–45, 15–55, and 20–210 nm, respectively. The polydispersity in shapes and sizes probably results from twinning, a very common process for face-centered cubic (fcc) metals.[33,34] Our analysis suggests that random twinning is likely a product of fast and nonuniform reduction, predominant in acid reaction media, as shown in Figure . GNPs produced at pH 9.0 (Figure C) did not present shape polydispersity, and the size distribution histogram revealed the predominance of 18 nm GNPs.

Figure 6

Representative electronic microscopy images of GNP synthesis using AAAKCK sequence at the pH values of (A) 4.0 (TEM), (B) 6.0 (TEM), and (C) 9.0 (FEG-SEM). Panel A shows a zoomed-in view of hexagonal plates and a decahedral NP. Panel B shows a zoomed-in view of plates with incomplete growing and decahedral NPs. (D,E) Corresponding shape distributions of the nanocrystals produced at pH 4.0 and 6.0, respectively. (F) Size distribution histogram of GNPs synthesized at pH 9.0. (ICO = icosahedron; DEC = decahedron; OT = others; PL = plate; BP = bipyramid; and SPH = sphere).

The crystalline nature of GNPs synthesized with the AAAKCK sequence at pH 4.0 was further corroborated using X-ray diffraction (XRD) analysis. The X-ray diffraction pattern is shown in Figure . The Bragg reflection peaks are corresponding to (111), (200), (220), (311), and (222) lattice planes. The 2θ degree values are consistent with the face-centered cubic (fcc) gold crystal structure as determined by the Joint Committee on Powder Diffraction Standards (JCPDS) file no: 04-0784. The electron diffraction pattern of gold nanostructures (inset of Figure ) exhibits the characteristics of polycrystalline gold.

Figure 7

XRD pattern of GNP synthesized with AAAKCK at pH = 4.0 and the corresponding electron diffraction analysis (inset).

The UV–visible spectra in Figure showed another interesting characteristic of the GNP synthesis with the AAAKCK sequence. The spectra obtained immediately after mixing the reagents (dotted lines) did not present bands at the UV region. However, at the subsequent times of synthesis, the appearance of bands was observed in the range of 325–340 nm that are assigned to two unresolved ligands (π) to metal (σ*) charge transfer (LMCT) bands.[18,20]Figure shows the pH profile of the LMCT band. The LMCT bands are assigned to the coordination of the peptide—probably SH—with the GNP surface that is consistent with the capping of GNPs by the peptides.

Figure 8

pH-dependent maximal absorbance of LMCT band resulted from the capping of GNPs synthesized by the AAAKCK peptide.

The GNPs produced by the AAAKCK peptide were also characterized by the zeta potential (Figure ) and Raman spectroscopy (Figure ). The results obtained with these techniques corroborated the capping of GNPs by the peptides. The samples remained unchanged even after some months at room temperature, indicating that the synthesized GNPs were very stable, and no aggregation occurred. It is known that peptides can bind to metal nanoparticles through free amine group or cysteine side chain and via electrostatic attraction of negatively charged carboxyl group, forming a coat on the particle surface to prevent agglomeration and ensuring the stabilization of GNPs.[35−37] The ζ potential values of these GNPs varied from −7.6 to −24.9 mV. These values are consistent with a partial covering of gold surface by AAAKCK, considering that the ζ value of bare GNPs is −38 mV.[38] For our samples, the values of zeta potential decreased by increasing the pH, which is consistent with the deprotonation of the lysine ε-amino groups of the peptide covering the GNPs and decreasing their tendency for aggregation (Figure ). In Figure , the decrease in the ζ potential (−8.6 to −17.4 mV) observed by changing the pH from 3 to 4 and (−17.4 to −24.9 mV) from pH from 7 to 10 is consistent with the deprotonation of carboxylic and amino groups at the gold surface, respectively.

Figure 9

pH-dependent zeta potential of GNPs synthesized using AAAKCK peptide.

Figure 10

FT-Raman spectra of AAAKCK peptide at pH 4.0 (red line) and the corresponding SERS effect promoted by the association of the peptide with GNPs. The inset shows the differential Raman spectra obtained by subtracting the Raman spectra at pH 4.0 from those at pH 9.0.

Raman spectra were recorded out to identify the possible functional groups present in the peptide sequence involved in the reduction of gold ions and the type of interaction in the capping and stabilization of synthesized nanostructures.[39]Figure shows the Raman spectra of AAAKCK solution (red line) and GNPs synthesized and stabilized by AAAKCK at pH 4.0 (black line). In Figure , the more pronounced surface-enhanced Raman scattering (SERS) effect for some groups is consistent with a predominant orientation of the peptide on the GNP surface. The SERS effects were the most intense for bands peaking at 3207, 1640, 1191, 1065, 804, 601 (shoulder), and 428 cm–1. The band at 3207 cm–1 is assigned to NH stretch modes. The band at 1640 cm–1 is amide I band assigned to C=O stretch of the α-helix structure. In the Raman spectrum of AAAKCK solution, this band is weak and peaks at 1660 cm–1.[40] The SERS effect associated with the shift of the amide I band to a lower frequency, consistent with the α-helix structure, corroborated the association of the peptide with the GNP surface. The bands at 1191 and 804 cm–1 are assigned to CH2 twist and CH2 rock modes. The spectral region of 1065–1071 cm–1 involves the respective contribution of Cα–Cβ stretch and C–S asymmetric stretch. SERS effect was also observed at 601 cm–1 (shoulder) that could be assigned to the blue-shifted contribution of δCSS and νCSS modes. The increased signals in the spectral range from 500 to 200 cm–1 result from the overlapped contribution of CCC and CCS bending vibration (ω), C–S stretching mode (νCS), and CSH bending mode (βCSH).[41,42] The inset of Figure shows the differential Raman spectra obtained by subtracting the Raman spectra obtained at pH 4.0 from those obtained at pH 9.0. The more pronounced differences in the differential spectra were observed at 601 cm–1 (δCSS and νCSS), in the spectral region from 500 to 200 cm–1 (overlapped CCC and CCS ω, νCS, and βCSH), and in the band at 3207 cm–1 (NH stretch modes). Additionally, a band at 3406 cm–1 region appeared in the differential spectrum (region highlighted in gray). Considering that the spectral region of 3450–3250 cm–1 contains the contribution of amide A bands that are assigned to the stretching vibration of the peptide N–H, the additional red-shifted band could correspond to another orientation of the peptide on GNP.

For the AAAXCX structure, the replacement of lysine residues by arginine residues (AAARCR) changed the properties of the peptide for GNP synthesis. The GNP synthesis with AAARCR at acidic pH values resulted in low-intensity plasmon resonance bands at pH 2, 3, 4, and 5. Plasmon resonance bands were not detected at pH 6, 7, and 8. At acidic pH, the GNPs presented broad plasmon resonance bands with a high contribution of turbidity. GNPs synthesized with AAARCR at pH 9 and 10 presented narrower plasmon resonance bands peaking at 536 nm, consistent with a low polydispersity and mean diameter of 79 ± 6 nm, as determined by scanning electron microscopy (SEM) (Figure ). GNPs synthesized with AAARCR at pH 3, 9, and 10 presented better stability and were more intense and featured the plasmon resonance band. The zeta potential values determined for the GNPs synthesized at the pH values cited were +39.7 ± 3; +30.4 ± 1.7, and +33.2 ± 1.4 mV, respectively. These values are consistent with a highly efficient covering in which the guanidine group of arginine residues is exposed on the GNP surface. Probably the orientation and packing of the AAARCR peptide on the GNP surface shielded the carboxylic groups, leading to a positively charged particle.

Figure 11

pH-dependent SPR of GNPs synthesized using AAARCR peptide. Panel A shows the SPR bands of the synthesis of pH 2.0, 3.0, and 4.0 as indicated. Panel B shows the SPR bands of the synthesis of pH 5.0, 6.0, and 7.0 as indicated. Panel C shows the SPR bands of the synthesis of pH 8.0, 9.0, and 10.0 as indicated. Panel D shows the SEM image of GNPs obtained at pH 10.0.

The present results demonstrated that the reducing capacity of AAAXCX is favored by the deprotonation of the ionizable groups, but the molecular structure of the peptides has a strong influence on the growth of the GNPs. Therefore, very similar peptide structures resulted in the different pH-dependent synthesis of GNPs. Considering that the replacement of lysine by arginine drastically changed the pH profile of the GNP synthesis, a neutral sequence and two acidic sequences were also tested and compared with free cysteine: AAAACA, AAAECE, and GSH (γ-ECG). Under the same conditions as for the synthesis used for the alkaline peptides, all of these peptides were unable to assist the GNP synthesis. This result shows that the acidic peptides with different structures were unfavorable for the synthesis of GNP. The AAAACA peptide was able to assist the GNP synthesis only in the presence of polyethylene glycol (PEG10) at pH 8 and 11 (Figure S1B). AAAECE peptide and GSH required PEG10 and heating at 80 °C to assist GNP synthesis. In these conditions, AAAECE assisted the GNP synthesis at pH 7 and 11 (Figure S2A) and GSH assisted the GNP synthesis only at pH 11. AAAECE-assisted and AAAECE-stabilized GNPs were characterized by electronic absorption spectroscopy and by TEM (Figure S2B). The XRD analysis of GNPs synthesized with AAAECE at pH 7.0 is shown in Figure S3. The major diffraction peaks were indexed as the face-centered cubic (fcc) gold phase based on the data of JCPDS file (JCPDS no. 04-0784). The dependence of PEG for the GNP synthesis assisted by acidic peptides suggests that the interaction with gold species is not favorable for these structures. PEG created a template and a microenvironment that favored the gold ions–peptide interactions. Differently, cysteine-assisted GNP synthesis occurred at the same pH range (2–11) that was effective for AAAKCK (Figure S1A). However, cysteine, when used as a reducing agent, did not efficiently stabilize GNPs, and the aggregation was also observed 24 h after the synthesis (not shown). The peptide structure also influenced the crystal growth and GNP capping with repercussions for the nanocrystal properties. The routes of metallic nanoparticle synthesis that provide the control for nanocrystal size and shape are desired because they dictate the properties of these materials. Bulk metallic materials exhibit typical properties such as magnetism and conductivity that are related to the quasicontinuous density of states.[43] At nanoscale level, the continuum of orbital energy is replaced by a discrete-energy-level structure. At the quantum dot scale, the material can exhibit well-defined bonding and antibonding molecular orbitals. Otherwise, anisotropy has dramatic effects on the optical properties of metal nanoparticles such as increased absorption, fluorescence, and Rayleigh (Mie) scattering.[44] These effects are extensive to the molecules adsorbed on their surface as the well-known SERS technique. Therefore, controlling the nanostructure shape is an efficient and sophisticated strategy for optical tuning. Anisotropy of metal nanoparticles results from the particle growing in a preferred direction. Synthesis of anisotropic metal nanoparticles can be achieved by the use of hard templates or the control of thermodynamic and kinetic parameters involved in the crystallographic control.[45] Under the conditions of the present study, GNPs were synthesized by the addition of gold salt in the peptide solution. In the solution-based method, the steps for the synthesis of fcc gold nanocrystal involve the reduction of Au3+ to Au0, nucleation, and growth to form seeds with a diversity of structures dictated by thermodynamic and kinetic factors.[46−48] TEM images of GNPs produced with the AAAKCK peptide at acidic pH revealed the predominance of the icosahedron and decahedron structures with the presence of some plates and other irregular nanostructures. The increase in pH from 4.0 to 6.0 decreased the size range of the nanostructures. GNPs produced at pH 6.0 exhibited higher shape polydispersity with the additional contribution of bipyramidal and spherical nanoparticles as well as complete and incomplete triangular plates. The icosahedra, decahedra, bipyramids, and plates with different sizes result from multiple twinned seeds and plates with stacking faults.[49] More details on the factors that govern the growth of seeds toward the diversity of nanostructures can be found in the refs (50) and (51). The pH of the medium is described as a key growth parameter in the synthesis of GNPs using biomolecules. The pH of the medium determines the ionization state of the biomolecules with repercussions on their reducing capacity and binding affinity for the gold surface. The effect of pH on the shape, size, and polydispersity of GNPs produced by the AAAKCK peptide was similar to the results presented by other studies using the biomolecules present in plant and microorganism extracts.[52,53] The production of larger nanoparticles with a diversity of shapes at lower pH has been attributed to the weakened repulsion between the negatively charged AuCl4– ions and the predominantly protonated carboxylic groups of biomolecules resulting in uncontrolled nucleation of seeds. Also, at lower pH values, the coagulation (coalescence) of smaller nuclei is favored, leading to the formation of larger GNPs.[54] The size range of icosahedra, decahedra, and plates produced, respectively, at pH 4.0 was consistently 55–90, 50–100 and 65–125 nm, whereas the size range of the same nanostructures produced at pH 6.0 decreased to 15–45, 15–55, and 20–210 nm, respectively. It is worth to remark that spherical GNPs were detected among the GNPs synthesized at pH 6.0. Higher pH values move the equilibrium between the ionization states of peptides toward the negatively charged forms. In this condition, the stronger repulsion between the negatively charged peptides and gold ions impairs the further growth of nuclei and favors the formation of small spherical nanoparticles. Under the conditions of the present study, it was also observed that the particle size and polydispersity decreased by increasing the pH. The pH also tailors the shape of GNPs by influencing the differentiated capping of crystal facets.[55−57] It is well known that the capping agents with a specific affinity to a crystal facet can direct the anisotropic growth.[43] An increase in LMCT band intensity was consistently observed during the synthesis of GNPs. Therefore, it is reasonable to assume that the preferential ligation of the peptide on a facet of the crystal seed contributed to the formation of the hexagonal and triangular nanoplates observed for the synthesis performed at acidic pH values.[58] Thus, a rationale for the pH-dependent formation of anisotropic nanocrystals in the synthesis using AAAKCK can be achieved. At acidic pH values, the equilibrium of the peptide is dislocated to the protonated thiol group of cysteine. In this condition, lysine could have a significant contribution as the reducing agent, and the availability of SH groups provide the capping, directing the anisotropic growth. At increasing pH values, the dislocation of the equilibrium toward deprotonated thiol group (−S–) could modify the capping properties of the peptide. A 10 nm red shift of the LMCT bands at alkaline pH and a significant loss of intensity of this band at pH 11 were consistently observed. The SERS effect on different vibrations involving the thiol group corroborated that cysteine lateral chains interact with the GNP surface. The significant SERS effect on the vibration of peptide N–H suggests that the polypeptide chain is also close to the gold surface. The differential Raman spectra shown in the inset of Figure suggests that N–H groups assume different interactions with the GNP surface synthesized at alkaline pH. The differentiated interaction could change the capping, leading to an impairment of the anisotropic growth. The negative values of the zeta potential of GNPs synthesized with the AAAKCK peptide and the strong pH-dependent SERS effect on N–H stretching suggest that the peptide is positioned horizontally along the peptide backbone on the gold surface. Interestingly, the LMCT band was not present in the UV spectra of the GNPs synthesized with the AAARCR peptide. In this condition, the GNPs exhibited positive values of zeta potential that is consistent with the guanidine group of arginine lateral chain on the surface of the GNPs. It is unlikely that the cysteine thiol group of AAARCR is not coordinated with the gold surface. However, the complete capping of the GNP surface by the AAARCR peptide could shield the gold-interacting thiol group, impairing the UV-light access and the occurrence of the ligand–metal transition. These results suggest that AAARCR binds to the gold surface by the thiol group and projects its polypeptide chain vertically from the surface. This positioning of AAARCR allows a better covering of the particle that filters the incident UV light. Figure summarizes the proposal of AAAKCK and AAARCR orientation on the gold surface.

Figure 12

Proposed orientation of AAAKCK and AAARCR on the GNP surface.

Conclusions

In the present study, we exploited the structural features of two alkaline bioinspired peptides for the synthesis of GNPs at the pH range of 2–11 and compared them with acidic peptides and cysteine. The objective was to investigate whether the lowering of the thiol group pKa of peptides by positively charged neighboring amino acid chains favors the GNP synthesis. The striking differences observed for the GNP synthesis assisted by AAAKCK and AAARCR, and the inefficiency of the acidic peptides demonstrated that the synthesis and stabilization of GNPs is a very complex process in which the ionization of cysteine is one of the participating factors. A synergistic combination of structure, net charge, reducing capacity, and accessibility of the thiol group operates in the synthesis of GNP assisted and stabilized by cysteine-containing peptides. In this scenario, the sequence array KCK is unique for a pH modulation of the synthesis of GNP with different sizes and shapes that are key characteristics for different applications.