Surface Engineering of Gold Nanorods for Cytochrome c Bioconjugation: An Effective Strategy To Preserve the Protein Structure.

The surface of gold nanorods (Au NRs) has been appropriately engineered to achieve a suitable interface for bioconjugation with horse heart cytochrome c (HCc). HCc, an extensively studied and well-characterized protein, represents an ideal model for nanoparticle (NP)-protein conjugation studies because of its small size, high stability, and commercial availability. Here, the native state of the protein has been demonstrated for the first time, by means of Raman spectroscopy, to be retained upon conjugation with the anisotropic Au nanostructures, thus validating the proposed protocol as specifically suited to mostly preserve the plasmonic properties of the NRs and to retain the structure of the protein. The successful creation of such bioconjugates with the retention of the protein structure and function along with the preservation of the NP properties represents a challenging but essential task, as it provides the only way to access functional hybrid systems with potential applications in biotechnology, medicine, and catalysis. In this perspective, the organic capping surrounding the Au NRs plays a key role, as it represents the functional interface for the conjugation step. Cetyltrimethylammonium bromide-coated Au NRs, prepared by using a seed-mediated synthetic route, have been wrapped with polyacrylic acid (PAA) by means of electrostatic interactions following a layer-by-layer approach. The resulting water-dispersible negatively charged AuNRs@PAA NPs have then been electrostatically bound to the positively charged HCc. The bioconjugation procedure has been thoroughly monitored by the combined analysis of UV-vis absorption, resonance Raman and Fourier transform infrared spectroscopies, transmission electron microscopy microscopy, and ζ-potential, which verified the successful conjugation of the protein to the nanorods.

Introduction

In the recent years, new perspectives have been offered by anisotropic gold nanoparticles (NPs) such as nanorods [gold nanorods (Au NRs)] because of their interesting optical properties, which have been exploited especially in biomedical diagnostics, in photothermal therapy, and as dark field imaging contrast agents and surface enhanced Raman spectroscopic (SERS) substrates.[1−8] Au NRs are characterized by two surface plasmon bands, namely, a transverse surface plasmon mode located in the visible region around 520 nm and a longitudinal surface plasmon mode at longer wavelengths, which can be tuned from the visible to the near-infrared (NIR) region by varying the aspect ratio (length-to-width ratio) of the Au NRs.[9−11] Bioconjugates of Au NRs are exceptionally promising candidates for important biomedical applications, especially when the resonances of the NRs are tuned to the NIR, where biological tissues are relatively transparent. In particular, NP–protein hybrid systems have already demonstrated their potential for many applications in research and diagnostics, for example, optical biosensing,[12] imaging of tissues, cells, or macromolecules,[5,13,14] and physicochemical manipulation of biological systems.[15] In this perspective, it is fundamental that the functional properties of the bioconjugates are maintained during the preparation. To accomplish this bioconjugation task, the nature of the NP interface and the adsorption are crucial, as they can determine conformational changes, denaturation, or undesirable protein orientations with respect to the substrate[16] that endanger their activity. To overcome these problems, NPs have been modified with small molecules[17] or polymeric matrixes[18−20] and characterized by means of bioanalytical techniques involving transduction of protein-based signals for detection, including electrochemistry,[21] and SERS[22] to verify the native state of protein. The goal of this work is to bioconjugate water-dispersible Au NRs, coated with the positively charged surfactant stabilizer [cetyltrimethylammonium bromide (CTAB)], to horse heart cytochrome c (HCc), while preserving the nanostructure geometry-dependent properties and, at the same time, the protein structure. Generally, three approaches have been proposed in the literature to link proteins to Au:[23] (i) direct anchoring on the CTAB coating of Au NRs by exploiting electrostatic adsorption, provided that the protein is negatively charged; (ii) interposing a negatively charged layer as a “glue” between CTAB and the protein;[24] and (iii) using as intermediate linkers bifunctional molecules containing a functional group with strong or good affinity for the metal surface [i.e., cyano (−CN), mercapto (−SH), carboxylic acid (−COOH), and amino (−NH2) groups] to, at least partially, displace the native bilayer and, on the other side, a functional group that interacts with the protein by means of covalent or noncovalent bonds.[23] HCc is used here as a protein probe, being a small, single chain, mitochondrial protein that plays an important role both in electron transfer and in programmed cell death (apoptosis). The identification of a strategy suited to bioconjugate HCc on Au NRs can be of great interest for biological applications and can allow the refinement of new operating procedures for protein bioconjugation. A possible direct covalent interaction of horse heart HCc with the bare Au NRs surface is not viable because of the lack of functional groups able to covalently bind the Au surface. Furthermore, HCc exhibits a positively charged domain around the exposed heme edge[25,26] that hampers any direct electrostatic interaction with the positively charged CTAB-capped Au NRs. In addition, a surface functionalization step to replace the pristine CTAB coating with bifunctional molecules more suited for subsequent binding of HCc could cause irreversible aggregation of Au NRs or a continuous release of the residual CTAB, thus endangering the protein stability also.[27] A reliable strategy, typically used to modify the surface chemistry and simultaneously protect the native NR coating from possible detachment of CTAB molecules is based on polyelectrolyte wrapping,[27] that endows the nanostructure surface with functional groups that enable their conjugation with biomolecules, including DNA, peptides, antibodies, and proteins by means of either electrostatic or covalent interactions.[28] We herein investigate the viability of poly(allylamine hydrochloride) (PAA) which, being a weak polyelectrolyte, allows the interaction between HCc and Au NRs to be modulated by controlling ionic strength and pH, thus obtaining a suitable wrap for the Au NRs.

Recent studies have shown that the conformation of HCc adsorbed on Au NPs depends on surface coverage and NP size.[29] To date, preservation of the native structure of HCc has been achieved on electrodes by using appropriate self-assembled monolayers,[26,30] but bioconjugation of colloidal Au NPs has mostly resulted in protein denaturation.[22,31−35] In particular, the conjugation of HCc to Au NPs has been found to preserve the HCc structure only when the structure has been decorated with aggregated Ag nanospheres.[22] The bioconjugation of Au NRs with proteins is even more difficult. In fact, the functionalization of Au NRs becomes more crucial because of the geometrical features of the nanostructures, as NRs are particularly sensitive to self-assembly and aggregation problems[36] and to the possible toxicity issues arising from the capping molecules that may detach from the nanostructures.[27] Herein, “as prepared” CTAB-capped Au NRs have been functionalized by using the electrostatic layer-by-layer approach to proceed with their subsequent decoration with HCc. Namely, the NRs have been coated with a negatively charged polyelectrolyte, PAA, which effectively enables the interaction of the NRs with HCc, which is positively charged because of the presence of Lys residues in its structure. The bioconjugation of the resulting AuNRs@PAA NPs with HCc, thoroughly characterized by means of ζ-potential, UV–vis absorption spectroscopy, transmission electron microscopy, Fourier transform infrared (FTIR) and resonance Raman (RR) spectroscopies, has demonstrated, for the first time, that the native state of the protein is preserved. The overall results indicate that the proposed protocol is specifically suited to accomplish an effective interface engineering of the Au NRs to bioconjugate HCc, enabling the design of hybrid nanosystems with a precise control of the inorganic geometry-dependent properties and the final biological functionality.

Results and Discussion

The characterization of the surface interactions between Au NRs and proteins is of importance for many potential applications. However, it has been proven that it is very difficult to achieve protein binding without damaging the native structure or modifying the NR plasmonic properties. Furthermore, neither of the two moieties have specific anchor points that facilitate covalent binding; therefore, the task represents a significant challenge. To overcome such difficulties and attain effective bioconjugation of the probe protein HCc, the surface properties of the Au NRs have been modified by means of a layer-by-layer electrostatic functionalization procedure. Herein, CTAB-coated Au NRs with an aspect ratio of 4.2 ± 0.4 (diameter = 14.1 ± 1.6 nm; length = 58.5 ± 5.8 nm) have been synthesized by using the seed-mediated approach, which results in NRs that have the surface coated with a bilayer of CTAB molecules, which ensures a very good particle dispersion in water, thus minimizing the risk of aggregation phenomena. Au NRs were prepared by reducing the Au(III) precursor at room temperature to Au(0) in the presence of CTAB as a surfactant.

The UV–vis spectrum of colloidal Au NRs is reported in Figure (black line) and shows the typical two plasmon bands located at 516 and 800 nm, corresponding to transverse and longitudinal plasmon modes, respectively. Moreover, the longitudinal plasmon band is quite narrow and intense compared with the transverse band, suggesting a high yield of nearly monodisperse Au NRs.

Figure 1

Evolution of normalized UV–vis absorption spectra of “as prepared” Au NRs (black line), Au NRs coordinated by PAA (red line), and HCc-conjugated Au NRs (AuHCc, green line). The inset shows a zoom of the spectra within the 350–450 nm wavelength range. The calculated HCc concentration of the AuHCc solution is about 10–6 M (molar absorptivity 1.06 M–1 cm–1).[37] A sketch presenting the multiple functionalization process of Au NRs with HCc is also included.

Moreover, the ammonium-terminated polar heads of the surfactants interact with water molecules in the surrounding environment, conferring a positive charge to the Au NRs, as demonstrated by the ζ-potential measurements (Table , Figure S1).

Table 1

Particle Size (d), Size Distribution and Surface Charge of Au NRs

sample	d mean (nm) z-average	PDIa	ζ-potential (mV)
Au NRs	86.1 ± 0.5	0.223 ± 0.046	+36.4 ± 1.0
AuNRs@PAA	127.2 ± 3.0	0.203 ± 0.071	–37.9 ± 0.3
AuHCc	160.2 ± 1.6	0.115 ± 0.032	–18.4 ± 1.0

PDI = polydispersity index.

The positively charged quaternary ammonium head of the CTAB molecules coating the Au NRs promote the electrostatic binding of the positively charged HCc by interaction with the negatively charged PAA (see the sketch in Figure ). Such a polyelectrolyte wrapping of the NRs is, in principle, able to limit the possible release of the CTAB molecules of the bilayer coating the NR surface, thus representing a safe “glue” compatible with the presence of biomolecules. The most investigated polyelectrolyte to electrostatically link HCc has been polystyrene sulfonated (PSS). However, the strong electrostatic interaction with the negative charges of PSS has been found to cause extensive disruption of the native structure of HCc.[26,38] Conversely, PAA, being a weak polyelectrolyte, allows the interaction between HCc and Au NRs to be modulated by the tuning of ionic strength and pH, thus obtaining the most-suited wrap for the Au NRs. In fact, this process has been demonstrated to be affected not only by the polymer charge density and the initial charge distribution on the CTAB-capped Au[27] but also by the balance between polymeric unit–NP attraction and repulsion among polymeric units.[39] The charge density of PAA is a function of the solution pH and, therefore, affects the arrangement of the chains and their conformation upon surface adsorption.[40] In fact, for pH > pKa (PAA, pKa = 4.8), many of the PAA carboxylic groups are deprotonated, thus inducing stretching of the polymeric chains because of reciprocal electrostatic repulsion.[41] In this study, PAA has been incubated in solution with the Au NRs at pH 7, obtaining a complete surface coverage. This is demonstrated by the inversion of the ζ-potential value (Table and Figure S1) from the initial positive charge of the Au NRs (+36.4 mV) to the negative charge of AuNRs@PAA (−37.9 mV). The ionic strength of the PAA solution has been left unchanged although, in principle, tuning this parameter could help optimize the electrostatic screening of the polyelectrolyte in solution. Nevertheless, any addition of salt would be detrimental, as it is known to cause aggregation of Au NRs.[42] As can be seen in the absorption spectrum of the AuNRs@PAA sample, after the purification procedure, the coverage with the PAA layer does not induce any significant change in the position of the longitudinal plasmon band (Figure , red line). At this stage, the AuNRs@PAA solution has been incubated with an excess of positively charged HCc for 1 h to promote electrostatic anchoring of the protein on the Au NRs (AuHCc sample). The unbound HCc has been removed by centrifugation cycles, to precipitate and isolate only functionalized Au NRs. The experimental conditions are decisive not only to control the amount of protein anchored on the Au NRs surface by simply varying the incubation time of the protein in the AuNRs@PAA solution but also to modulate the strength of the interaction between HCc and AuNRs@PAA, that may affect the extent of protein denaturation at the NR surface.[31,43] In addition, it has been reported that a higher protein coverage could result in an overcrowded arrangement and thus enhance the protein–protein and protein–NRs interactions that, consequently, may affect protein folding and, ultimately, its activity.[44,45] Therefore, to maintain constant the number of free carboxylates of PAA coating the Au NRs, the AuNRs@PAA solution has been washed and redispersed in phosphate buffer solution at physiological pH (pH 7). The presence of HCc in the Au NRs solution is clearly shown by the absorption spectrum of the AuHCc sample. The appearance of the Soret band of native HCc (Fe3+) at 409 nm, concomitant with the slight red shift of the longitudinal plasmon band because of modification of the chemical environment, is clearly evident (Figure , green line). Remarkably, because of its low molecular weight, the free HCc in solution does not precipitate along with the bioconjugated Au NRs upon numerous washing cycles of the sample by centrifugation. In fact, no precipitate is observed when pure HCc solution was centrifugated in the same experimental conditions as the bioconjugated sample. The shoulder on the longitudinal plasmon band observed at higher wavelengths can be ascribed to minor aggregation phenomena arising from the purification/centrifugation cycles.

The ζ-potential value after incubation with HCc (Table , Figure S1) decreases because of the reduction in the number of free PAA carboxyl groups, which confirms that the protein is well-anchored and also indicates a good stability of the prepared NRs. Transmission electron microscopy (TEM) images of as-synthesized Au NRs (a), Au NRs coordinated by PAA (b), and HCc-conjugated Au NRs are reported in Figure . The TEM micrograph in Figure a shows that the shape and size distribution of the Au NRs is quite uniform with an aspect ratio of 4.2 ± 0.4 (Figure g). The TEM analysis of the aspect ratios of the particles for AuNRs@PAA and AuHCc (Figure h,i), performed on the TEM images (Figure b,c), shows that they are 4.2 ± 0.5 and 4.1 ± 0.5, respectively. This confirms that the Au NRs preserve their size and shape after the polyelectrolyte wrapping and HCc binding steps. Therefore, the slight modification of the longitudinal plasmon band position observed in Figure can be safely ascribed to the modification of the chemical environment of Au NRs. Interestingly, TEM images of the negatively stained samples (Figure d–f) highlight the modification of the Au NRs organic coating, that passes from a uniform and thin layer for Au NRs to an inhomogeneous coating of about 2.5 nm for AuNRs@PAA, consistent with the less compact structure of a polyelectrolyte layer. Such features are even more evident when the HCc is anchored to the Au NRs. Furthermore, the diameter analysis by dynamic light scattering (DLS) (Table , Figure S2) shows that the Au NRs are characterized by an equivalent spherical shape and rather homogeneous in size: 86.1 ± 0.5 nm for Au NRs, 127.2 ± 3.0 nm for AuNRs@PAA, and 160.2 ± 1.6 nm for AuHCc, whereas their diameters are higher than those observed by TEM. This discrepancy can be explained by taking into account that DLS probes the hydrodynamic diameter, which includes the hydration shell due to the PAA coating. In addition, the DLS analysis of anisotropic particles, such as Au NRs, has to be interpreted cautiously because of the possible occurrence of multiple diffusion modes and must be considered only qualitatively. Nevertheless, such values are reliable enough to provide a useful indication of the size evolution trend of the nanostructure along the overall functionalization process. The diffusion of rod-like NPs is more complicated to describe than that of nanospheres because the presence of two geometrical parameters (length l and radius R) requires that both the rotational and the translational diffusion coefficients be taken into account for the treatment of the rod-like particle hydrodynamics. In addition, the overall translational diffusion coefficient is determined by a suitable combination of the diffusion coefficients for motion parallel (D∥) and perpendicular (D⊥) to the long axis of the NR.[46,47]

Figure 2

TEM images of as-synthesized Au NRs (a), Au NRs coordinated by PAA (b) and HCc-conjugated Au NRs (c) (100 nm scale bar). The TEM images of the same samples captured in negative staining mode are shown in (d–f). (50 nm scale bar). The corresponding aspect ratio statistical analyses of the particles are presented in (g–i).

The bioconjugation process has also been monitored by FTIR investigation (Figure , Table S1). The intense signal at 1702 cm–1 and the broad band with medium intensity at 1248 cm–1 are typically associated with ν(C=O) and ν(C–O) stretching modes of carboxylic group dimers. The weak bending δ(O–H) mode at 814 cm–1 further confirms the presence of carboxylic group dimers in PAA. In addition, the asymmetric stretching ν(C–O) mode, associated with –COO– at 1558 cm−1, is coupled with the weaker symmetric ν(C–O) mode at 1408 cm–1, indicating the presence of carboxylate ions.[48] Therefore, this suggests that at the pH used (pH 7), the carboxylic groups of the PAA stock solution are not completely deprotonated and thus both dimers and carboxylate ions coexist. Remarkably, when PAA interacts with CTAB to wrap the Au NRs, the intense band at 1702 cm–1 significantly decreases, leaving the intense band at 1558 cm–1 of the carboxylate ions, which interacts with the ammonium heads of CTAB (Figure B). The binding of HCc to the PAA layer results in the appearance of the typical band at 1652 cm–1 (Figure D), assigned to the ν(C=O) stretching mode of amide I of the protein (Figure C) that is a very sensitive probe of protein linkages.[35] In fact, the intensity of this signal slightly decreases when HCc electrostatically interacts with PAA.

Figure 3

ATR–FTIR spectra of PAA (A), AuNRs@PAA (B), HCc (C), and AuHCc (D).

Therefore, the overlap of the HCc and PAA bands in the 1500–1700 cm–1 region does not permit any further considerations on the bioconjugation event. However, any denaturation of the protein can be ruled out, as indicated by the lack of a band at 1635 cm–1 in the FTIR spectrum of AuHCc, typically ascribed to denatured HCc (Figure D).[49]

The AuHCc RR spectra (Figure , red) show, for both 406.7 (b) and 514.5 (a) nm excitation, the coexistence of the oxidized and reduced HCc forms. For 406.7 nm excitation, in resonance with the HCc Soret bands of both the ferric (409 nm) and ferrous (417 nm) forms, the spectra are dominated by the totally symmetric modes of native HCc. In fact, for the Fe2+ form, ν4, ν3, ν2, and ν10 bands are observed at 1363, 1492, 1592, and 1622 cm–1, and for the Fe3+ form, ν4, ν3, ν2, and ν10 bands are observed at 1372, 1502, 1585, and 1637 cm–1, respectively. The full assignment is reported in Table S2. Complete HCc oxidation has been achieved by adding a minimum amount of K3[Fe(CN)6] solution to the AuHCc solution (Figure , violet), without any change in the position of the Au plasmon bands and the HCc Soret band (data not shown). This suggests that the reduced species represents only a small amount of the total HCc bound to the NRs, even though it is clearly visible in the Raman spectrum. This apparent discrepancy probably results from different resonance intensification factors and cross sections of the core-size marker bands of the reduced form compared with the oxidized species. Moreover, the treatment with K3[Fe(CN)6] did not induce any detachment of the protein from the NRs, as verified by subsequent centrifugation and washing of the sample (Figure S5). Nevertheless, after the washing, the reduced species appears again in the Raman spectrum. Upon 514.5 nm excitation (Figure , a), in resonance with the Q band transitions of HCc (β-band at 530 and 520 nm for the ferric and ferrous forms, respectively), the spectrum is characterized, as expected, by B1g, A2g, and B2g modes. The bands at 1313, 1402, and 1585 cm–1 are assigned to ν21, ν20, and ν19 of the ferrous form, respectively. However, the coexistence of the reduced and oxidized forms causes some changes in the relative intensity ratios of the bands compared with the spectra of the pure species[50,51] (Table S2).

Figure 4

RR spectra of AuHCc (red) and AuHCc after addition of K3[Fe(CN)6] (violet) obtained with the 514.5 (a) and 406.7 (b,c) nm excitation wavelengths. The bands assigned to (Fe2+) and (Fe3+) HCc are indicated in green and black, respectively. Experimental conditions: laser power at the sample 40 (a) and 10 mW (b,c); average of 15 spectra with 2 h 30 min integration time (a), 4 spectra with 40 min integration time (b), and 8 spectra with 80 min integration time (c). The spectra have been shifted along the ordinate axis to allow better visualization.

The HCc adsorbed on the NRs retains its native conformation (i.e., the heme iron is bound to the His18 and Met80 internal ligands), as demonstrated by the close correspondence of the difference spectra obtained by subtracting alternatively the native HCc (Fe2+) (Figure , b-c) and the native HCc (Fe3+) (Figure , b-a) from the spectrum of AuHCc (Figure b). In fact, the core-size marker band frequencies are identical to those of the native protein in solution, indicating that no changes in the heme distortion, and hence ligation,[52,53] have been induced by bioconjugation. This conclusion is confirmed by the “fingerprint” region between 300 and 450 cm–1 that shows that features are typical of a Met–Fe–His species (Figure ). Indeed, previous reports found that upon denaturation the disruption of the Met80–iron bond occurs concomitantly with the loss of the secondary structure and the increase in the molecular size under equilibrium unfolding conditions.[54] The replacement of the Met80 heme ligand with other residues should cause marked changes both in the positions and in the relative intensities of the Raman bands. In particular, specific iron–ligand stretching vibrations have been identified for His18–Fe–Lys and His18–Fe–His misligation as well as specific band patterns for His18–Fe–OH– and His18–Fe–H2O conformations.[55−99] Because the band frequencies and the intensity pattern of AuHCc (Figure b, red) are identical to those of the oxidized protein in solution (Figure a), with some minor changes due to the presence of a small percentage of the reduced species (Figure b), the HCc structure and heme ligation are fully conserved on the Au NR surface.

Figure 5

RR spectrum of AuHCc (b) together with the difference spectra (b-c, b-a) after subtraction of the pure ferric HCc (a) and ferrous HCc (c) solution spectra. Experimental conditions: (a) 406.7 nm excitation wavelength, laser power at the sample 5 mW, and average of 5 spectra with 25 min integration time; (b) 406.7 nm excitation wavelength, laser power at the sample 10 mW, and average of 48 spectra with 4 h integration time; and (c) 413.1 nm excitation wavelength, laser power at the sample 5 mW, and average of 4 spectra with 20 min integration time. The intensities are normalized to that of the ν4 band. The spectra have been shifted along the ordinate axis to allow better visualization.

Figure 6

RR spectra of AuHCc (b, red) compared with ferric (a) and ferrous (c) HCc in the low-frequency region, obtained with the 406.7 nm excitation wavelength. Experimental conditions: laser power at the sample 5 mW (a,c) and 10 mW (b); average of 5 spectra with 25 min integration time (a), 18 spectra with 3 h integration time (b), and 4 spectra with 20 min integration time (c). The spectra have been shifted along the ordinate axis to allow better visualization.

No SE(R)RS effect has been observed for both 406.7 (Figure S6) and 514.5 nm excitation because the intensity of the HCc signals is the same (or marginally lower) as that of a pure HCc buffer solution at 10–6 M (Figure S3), that is, the calculated HCc concentration from the spectrum in Figure (green line). Excitation with the 632.8, 647.1, 785, and 1064 nm laser lines has been also checked. However, no HCc signals have been observed (data not shown).

Conclusions

In conclusion, this work has demonstrated that water-dispersible CTAB-coated Au NRs, prepared by means of a seed-mediated synthetic route, can be wrapped with polyacrylic acid (PAA) by exploiting the electrostatic interactions between CTAB and PAA, without modifying the NRs optical properties. This layer-by-layer preparation scheme has been then used also to accomplish the electrostatic binding of positively charged HCc on the negatively charged AuNRs@PAA NPs. The thorough characterization of the bioconjugation procedure has established that the native state of the protein has been retained upon conjugation with Au NRs and that the plasmonic properties of the Au NRs have also been mostly maintained. This result, confirming the success of the proposed protocol, represents the first example of HCc bioconjugation with anisotropic Au NPs. Hence, it is now possible to envisage the design of similar systems for biomimetic applications and for integration in bioelectronics devices.

Experimental Section

Materials

Sodium borohydride (NaBH4, ∼99%), l-ascorbic acid (99%), potassium ferricyanide (K3[Fe(CN)6], analytical grade), sodium dihydrogen phosphate (NaH2PO4, analytical grade), sodium hydrogen phosphate (Na2HPO4, reagent grade), hydrogen tetrachloroaurate(III) trihydrate (HAuCl4·3H2O, 99.9%), silver nitrate (AgNO3, 99.9999%), PAA (Mn ≈ 130 kDa), potassium phosphate monobasic (≥98%), potassium phosphate dibasic (≥98%), potassium chloride (≥99%), and HCc ≥95% based on mol. wt. 12 384) were purchased from Sigma-Aldrich. Stock solutions were prepared using deionized water (Millipore Milli-Q gradient A-10 system).

General Protocol for the “Seed-Mediated” Synthesis of Au Nanorods

CTAB-capped Au NRs with a longitudinal plasmon band at 800 nm were prepared using a suitably modified literature procedure commonly known as the seed-mediated growth method.[61,62] Briefly, a “seed” solution containing CTAB-stabilized Au NPs (<3 nm) was prepared by reducing the Au precursor, HAuCl4·3H2O (5 mL, 5 × 10–4 M), with NaBH4 (0.01 M) in CTAB (5 mL, 0.2 M) and then was kept under vigorous stirring for 2 h. At the same time, the Au NR growth solution was prepared: 8.5 mg of AgNO3, 18.22 g of CTAB, and 100 mg of HAuCl4·3H2O were dissolved in 500 mL of deionized water (CTAB dissolution was assisted by sonication). This mixture was kept at room temperature under continuous stirring. After 3 min, 7 mL of a 0.0778 M aqueous solution of ascorbic acid was added dropwise to the above mixture.[61] The mixture became colorless (indicative of the reduction of Au(III) to Au(I),[63] and then, 0.8 mL of the “seed” solution was added to the entire growth solution. A dark-red color developed within 10 min. After preparation, cycles of centrifugation were performed at 7000 rpm for 20 min to wash the Au NRs eliminating uncoordinated CTAB.

Functionalization of Au NRs with HCc

The Au NRs were functionalized with HCc by exploiting electrostatic interactions in a layer-by-layer fashion (see the sketch in Figure ). In the first step, the “as prepared” Au NRs were coated with PAA by incubation of 1 mL of Au NRs (2 × 10–9 M) and 1 mL of PAA (1.6 mg/mL, pH 7) for 20 min under gentle stirring. Then, this mixture was centrifuged two times at 10 000 rpm for 10 min to wash the Au NRs and eliminate free PAA. The AuNRs@PAA that precipitated upon centrifugation were dispersed in a phosphate buffer (5 mM, pH 7) to ensure deprotonation of the PAA carboxylate groups and confer negative charges to the NR surface. Thus, in the second step, 200 μL of the HCc solution (0.5 mg/mL) in 5 mM phosphate buffer (pH 7.0) was added to the as prepared AuNRs@PAA solution (5 × 10–10 M), and this mixture was kept under gentle stirring for 1 h. Finally, the obtained HCc-conjugated Au NRs (AuHCc) solution was centrifuged at 7000 rpm for 10 min, and washed with phosphate buffer (5 mM, pH 7) to remove free HCc until no further decrease of the Soret band absorbance was observed. To ensure complete oxidization of the HCc adsorbed on the NRs, 2 μL of a K3[Fe(CN)6] solution (1 grain in 1 mL) was added. The resulting solution was characterized by optical and morphological measurements to monitor the functionalization process.

Sample Characterization

UV–Visible Spectroscopy

UV–vis absorption spectra of solutions of the as prepared NRs and layer-by-layer multilayer-modified substrate were recorded at room temperature using a Varian Cary 5000 UV–vis–NIR scanning spectrophotometer in the 300–1200 nm range or a Cary 60 (Agilent Technologies) with a 1 mm cuvette and a 600 nm/min scan rate. Absorption spectra using a 5 mm NMR tube with a 300 nm/min scan rate were measured both prior to and after RR measurements to ensure that no degradation had taken place under the experimental conditions used. The HCc concentration was determined using the molar absorptivity (ε) of 106 mM–1 cm–1 at 409 nm.[37]

IR Spectroscopy

Mid-infrared spectra were acquired with a PerkinElmer Spectrum One FTIR spectrometer equipped with a deuteratedtryglicinesulfate detector. The spectral resolution used for all experiments was 4 cm–1. For attenuated total reflection (ATR) measurements, the internal reflection element (IRE) used was a three-bounce 4 mm diameter diamond microprism. Cast films were prepared directly on the IRE by depositing the solution of interest (3–5 μL) onto the upper face of the diamond crystal and allowing the solvent to evaporate.

Transmission Electron Microscopy

TEM analyses were performed with a JEOL JEM-1011 microscope operating at 100 kV. The specimens were prepared by depositing a few drops of aqueous NP dispersions onto a carbon-coated copper grid and then allowing the solvent to evaporate. The samples were stained with a 2% (v/v) phosphotungstic acid solution for 30 s. A statistical analysis of NR size distributions was performed on the basis of low-magnification TEM images with the help of ImageJ software. At least 150 NPs were counted for each sample.

Particle Size, Size Distribution, and ζ-Potential Measurements

The average hydrodynamic diameters (z-average), size distribution (polydispersity index, PDI), and ζ-potential were determined using Zetasizer Nano ZS, Malvern Instruments Ltd., Worcestershire, UK. Size and size distribution were measured by DLS after suspension in demineralized water at a suitable concentration (≈0.1%, w/v NRs concentration). The ZP was determined by laser Doppler velocimetry after a suitable dilution of the samples with KCl 1 mM (≈0.01%, w/v NRs concentration).

RR Measurements

RR spectra were recorded using a 5 mm NMR tube and 406.7 (Kr+ laser, Innova 300 C, Coherent, Santa Clara, CA) and 514.5 nm (Ar+ laser, Innova 90 Coherent, Santa Clara, CA) excitation wavelengths. Backscattered light from a slowly rotating NMR tube was collected and focused into a triple spectrometer (consisting of two Acton Research SpectraPro 2300i and a SpectraPro 2500i in the final stage with a grating of 3600 or 1800 grooves/mm) working in the subtractive mode, equipped with a liquid nitrogen-cooled charge-coupled device detector. A spectral resolution of 1.2 cm–1 and spectral dispersion of 0.4 cm–1/pixel were calculated theoretically on the basis of the optical properties of the spectrometer for the 3600 grating; for the 1800 grating, used to collect the RR spectra with the 514.5 nm excitation wavelength, the spectral resolution was 4 cm–1 and spectral dispersion 1.2 cm–1/pixel.

The RR spectra were calibrated with indene and carbon tetrachloride as standards to an accuracy of 1 cm–1 for intense isolated bands. All the RR measurements were repeated several times under the same conditions to ensure reproducibility. To improve the signal-to-noise ratio, a number of spectra were accumulated and summed only if no spectral differences were noted. Because the spectra of the HCc-conjugated Au NRs (AuHCc) solutions were characterized by a strong fluorescence background and a broad intense water band in the region 1550–1700 cm–1 on which the weak HCc signals were observed, for both excitation at 406.7 and 514.5 nm, a blank sample spectrum (AuNRs@PAA in buffer solution, see Figure S7) was subtracted from all spectra. All spectra were baseline-corrected.