Studying Protein and Gold Nanoparticle Interaction Using ...dli9/Publications/Studying Protein and Gold... · 1 Studying Protein and Gold Nanoparticle Interaction Using ... layer

1 Studying Protein and Gold Nanoparticle Interaction Using2 Organothiols as Molecular Probes3 Karthikeshwar Vangala,† Fathimar Ameer,† George Salomon,† Vu Le,† Edwin Lewis,† Leyuan Yu,‡

4 Dong Liu,‡ and Dongmao Zhang*,†

5†Department of Chemistry, Mississippi State University, Mississippi 39762, United States

6‡Department of Mechanical Engineering, University of Houston, Texas 77204, United States

7 *S Supporting Information

8 ABSTRACT: The protein and gold nanoparticle (AuNP) interfacial interaction has broad9 implications for biological and biomedical applications of AuNPs. In situ characterization of10 the morphology and structural evolution of protein on AuNPs is difficult. We have found11 that the protein coating layer formed by bovine serum albumin (BSA) on AuNP is highly12 permeable to further organothiol adsorption. Using mercaptobenzimidazole (MBI) as a13 molecular probe, it is found that BSA interaction with AuNP is an exceedingly lengthy14 process. Structural modification of BSA coating layer on AuNP continues even after 2 days’15 aging of the (BSA/AuNP) mixture. While BSA is in a near full monolayer packing on the16 AuNPs, it passivates only up to 30% of the AuNP surfaces against MBI adsorption. Aging17 reduces the kinetics of the MBI adsorption. However, even in the most aged BSA-coated18 AuNP (3 days), 80% of the MBI adsorption occurs within the first 5 min of the MBI19 addition to the (AuNP/BSA) mixture. The possibility of MBI displacing the adsorbed BSA was excluded with quantitative BSA20 adsorption studies. Besides MBI, other organothiols including endogenous amino acid thiols (cysteine, homocysteine, and21 glutathione) were also shown to penetrate through the protein coating layer and be adsorbed onto AuNPs. In addition to22 providing critical new understanding of the morphology and structural evolution of protein on AuNPs, this work also provides a23 new venue for preparation of multicomponent composite nanoparticle with applications in drug delivery, cancer imaging and24 therapy, and material sciences.

25 ■ INTRODUCTION26 Recent experimental evidence shows that when AuNPs are27 exposed to protein or serum plasma, protein spontaneously28 accumulates onto the AuNP surface, forming a protein coating29 layer that is commonly referred to as a protein “corona”.1−3

30 Multiple works have probed the fundamental mechanism31 governing the protein interfacial interaction with AuNPs.3−8

32 However, there are many unsettling questions regarding the33 protein structure on the AuNP surfaces. For example, to date,34 binding constants that differ by ∼4 orders of magnitude35 (∼105−1011 M−1) have been reported for BSA binding with36 AuNPs of 10−40 nm in diameter.4,8−10 Recently, Casals et al.37 reported that aging enhances the stability of the protein38 adsorbed on AuNPs, and the protein coating layer evolves from39 a soft corona to a hard corona. However, the fundamental40 mechanism of this aging effect and the structural characteristics41 of the “soft” and “hard” coronas are currently unclear.42 Important questions such as the possible mobility and the43 interprotein spacing of the adsorbed protein have not been44 addressed. Answering these questions is important for our45 understanding of protein/AuNP interfacial interaction, which46 will aid future design of protein-functionalized AuNPs for47 biological and biomedical applications.11

48 One difficulty in studying the morphology of the protein49 coating layer on AuNPs is the lack of reliable spectroscopic and50 imaging techniques for in situ investigation of protein

51immobilized on AuNPs in solution. While circular dichroism52and protein tryptophan fluorescence have been used for53understanding the protein/AuNP interactions,2,9,10 background54signal from the protein that remained free in solution and the55inner filtration effect of the AuNP compromise sensitivity of the56techniques and complicate the interpretation of the exper-57imental results. Dynamic light scattering (DLS) techniques58measure the change in the hydrodynamic radii of the AuNP59induced by protein adsorption. It is unlikely to be sensitive to60protein conformational change on the AuNP surfaces. Trans-61mission electron microscopic (TEM), atomic force microscopic62(AFM), and optical imaging techniques were applied to study63AuNP interfacial interaction with protein; however, the samples64have to be dried12 or frozen (cryogenic TEM),13 which likely65perturbs the configuration of the protein/AuNP complex.66In this work, we employ mercaptobenzimidazole (MBI), an67organothiol, as the molecular probe to investigate the structural68characteristics of the protein coating layer on AuNPs. Recent69work suggests that organothiols such as glutathione cannot70displace BSA adsorbed onto the AuNPs.14 Whether or not an71organothiol can be adsorbed onto BSA-covered AuNPs is an72open question. Indeed, as it will be shown later, our

Received: November 8, 2011Revised: January 10, 2012

Article

pubs.acs.org/JPCC

© XXXX American Chemical Society A dx.doi.org/10.1021/jp2107318 | J. Phys. Chem. C XXXX, XXX, XXX−XXX

dmadmin | ACSJCA | JCA10.0.1408/W Unicode | research.3f (R2.1.i5 HF03:2984 | 2.0 alpha 39) 2011/12/22 17:29 | PROD-JCAVA | rq_1327205 | 1/24/2012 15:05:54 | 8

73 quantitative MBI adsorption data revealed that MBI can be74 readily adsorbed onto the BSA-covered AuNPs without causing75 BSA displacement. By studying the kinetics and binding76 capacity of the MBI adsorption onto the AuNP/BSA complex,77 we are able to obtain critical information regarding the78 morphology of the protein coating layer and the binding79 characteristics between BSA and AuNPs. We also investigated80 the possible binding of the amino acid thiols (AAT) [cysteine81 (Cys), homocysteine (Hcy), and glutathione (GSH)] onto the82 BSA-stabilized AuNPs. For simplicity’s sake, we refer to GSH as83 an AAT even though it is a tripeptide.

84 ■ EXPERIMENTAL SECTION85 Materials and Equipment. All the chemicals used were86 purchased from Sigma−Aldrich. BSA with a purity of 97% (lot87 064K1251) was used as received. Nanopure water (Thermo88 Scientific) was used in all our measurements. The SERS spectra89 were obtained with a LabRam HR confocal Raman microscopy90 system (Horiba Jobin Yvon, Edison, NJ) using a 632.8 nm91 HeNe laser for Raman excitation. The RamChipTM slides (Z-S92 Tech LLC) were used for SERS spectral acquisitions. The UV−93 visible measurements were measured using an Evolution 30094 spectrophotometer (Thermo Scientific, Waltham, MA) or an95 Olis HP 8452 diode array spectrophotometer (for the time-96 resolved localized surface plasmon resonance (LSPR) and97 kinetics of MBI adsorption measurements). DLS measurements98 were performed on a DynaProTM NanoStar system (Wyatt99 Technology, Santa Barbara, CA). Centrifugations were100 conducted with a Marathon 21000R Fisher Scientific instru-101 ment (Pittsburgh, PA).102 Silver and Gold Nanoparticle Synthesis. Unless103 specified otherwise, all SERS spectra were acquired using silver104 nanoparticles (AgNPs) synthesized by the Lee and Meisel105 method.15 The gold nanoparticles were synthesized using the106 citrate reduction method.16,17 In brief, 0.0415 g of gold(III)107 chloride trihydrate was dissolved in 100 mL of distilled water108 and was heated to reflux with vigorous stirring. Then, 10 mL of109 1.14% (w/v) sodium citrate dihydrate aqueous solution was110 added to the solution right after boiling commenced, and111 boiling continued for 20 min. The average particle size of112 prepared AuNPs was ∼15 nm in diameter with peak UV−vis113 absorption centered at 520 nm (Figure S1). The concentration114 of AuNPs was calculated as 9.2 nM by assuming that all115 gold(III) ions are reduced to gold(0), which is also consistent116 with the concentration of the AuNP estimated using the UV−117 vis absorbance of the as-synthesized AuNPs.118 DLS Measurements. DLS measurements were performed119 with a DynaProTM NanoStar system equipped with a HeNe120 laser at 658 nm and an Avalanche photodiode detector. A121 Wyatt quartz cuvette of path length 1 cm and an active volume122 of 10 μL was used as a sample holder. After its loading, the123 sample in the cuvette was left to sit for 5−10 min to allow any124 turbulence to dissipate before spectral acquisition. The125 DYNAMICS software package (v.7.1.0) was used to analyze126 the data. A detection angle of 90° was chosen for all size127 measurements.128 Ratiometric SERS Quantification of MBI Adsorption129 on AuNP. The amount of MBI adsorbed onto the BSA-130 stabilized AuNP was determined using the ratiometric SERS131 method we recently developed.18 After centrifugation removal132 of the AuNPs together with their surface adsorbates, 100 μL of133 supernatant was transferred into a vial and then mixed with an134 equal volume of known concentration of isotope-substituted

135MBI (MBId4) where the four hydrogen atoms in the benzene136ring of normal MBI (MBId0) are substituted with deuterium137atoms (Figure S2, Supporting Information). After vortexing for138∼1 min, the MBId4 spiked supernatant was transferred to an139ultrafiltration tube (MWCO 3500, Millipore (Bedford, MA)) to140separate the BSA and AuNP from the isotope-substituted MBI141pair. The filtrate was then subjected to SERS spectral142acquisition. The removal of BSA is important as it interferes143with the SERS acquisition of the MBId4 and MBId0 mixture.19

144The concentration of the free MBI (MBId0) in the ligand145binding solution was deduced from the SERS intensity ratio of146MBId4 and MBId0 using the ratiometric SERS calibration curve147we derived before.18

148Kinetics of MBI Adsorption onto BSA-Covered AuNPs.149MBI adsorption kinetic measurements were carried out with an150Olis HP 8452 diode array spectrophotometer. To probe the151effect of aging the BSA-covered AuNP on the MBI adsorption,152the AuNP was mixed with BSA and the mixture was left to sit at153room temperature for a predefined time period. Time-resolved154UV−vis spectra were taken immediately after the addition of155MBI into the (AuNP/BSA) mixture. The time interval between156each consecutive spectral acquisition was 0.5 s. The dead (lead)157time of this instrument was about 4 s, and the integration time158of each spectrum was 1.1 s. The adsorption kinetics of MBI159onto the (AuNP/BSA) complex was determined on quenching160of the MBI UV−vis absorption upon its adsorption onto AuNP,161an effect that will be discussed in the Results and Discussion162section.163Quantification of BSA Adsorbed on AuNP. The as-164synthesized AuNPs were concentrated using centrifugation165precipitation followed by sonication redispersion to a final166concentration of 200 nM. The concentrated AuNPs are then167mixed with equal volume of 20 μM BSA solution. The amount168of BSA adsorbed onto the AuNPs was determined with169(AuNP/BSA) mixtures that were aged for 10 min, 12 h, and 48170h, respectively. Three independent samples were prepared for171each time point. After incubation, the samples were centrifuged172at 9000 g for 90 min (Marathon 21000R, Fisher Scientific), and173the amount of BSA that remained free in the supernatant was174quantified using the BSA UV−vis absorbance in the 280 nm175region.

176■ RESULTS AND DISCUSSION177LSPR Measurement of BSA Adsorption onto AuNPs.178BSA adsorption onto AuNP was monitored using a179combination of LSPR, DLS, and quantitative BSA adsorption180studies. LSPR is one of the very few techniques appropriate for181in situ study of the AuNP interfacial interactions in solution,182and it has been used extensively for probing the stability of183BSA-covered AuNP.2,5,14,20−22 To our knowledge, however,184application of LSPR for monitoring the BSA adsorption kinetics185 f1has not been reported. Figure 1 shows the time-resolved LSPR186spectra of the AuNP/BSA solutions, which demonstrated that187immediately following the BSA addition into as-synthesized188AuNP, the peak absorbance of the AuNP increased189significantly, accompanied by a red shift of the LSPR peak.190After the first ∼10 min or so of the sample preparation, there is191essentially no spectral modification in the LSPR feature,192indicating that the BSA adsorption has reached a steady state193on the AuNPs.194The changes in the AuNP LSPR features may be induced by195AuNP aggregation and/or BSA adsorption. Subsequent DLS196measurement reveals that the maximum increment of the

The Journal of Physical Chemistry C Article

dx.doi.org/10.1021/jp2107318 | J. Phys. Chem. C XXXX, XXX, XXX−XXXB

197 AuNPs exposed to BSA is about ∼3.9 nm in radius, which is the198 significantly smaller than would be expected from AuNP199 aggregation. As a result we concluded that the LSPR spectral200 change is induced by the binding of BSA onto AuNPs.201 To investigate the possibility of using LSPR to estimate the202 binding affinity of BSA onto AuNP, we studied the correlation203 between the BSA concentration and the AuNP LSPR feature in204 the BSA/AuNP mixtures (Figure 1B). Compared to the LSPR205 peak positions, the LSPR peak absorbance correlates much206 better with the BSA concentration. The LSPR data in Figure 1B207 showed that the AuNP peak absorbance reached a plateau208 when the concentration of BSA is equal to or larger than 4 μM,209 suggesting that BSA reaches saturation adsorption at this210 concentration.211 The kinetic curve and the adsorption isotherm shown in212 Figure 1 are remarkably similar to the previously reported data213 obtained with the quartz crystal microbalance (QCM) study of214 BSA adsorption onto citrate-coated flat gold film.23 This result215 suggests that BSA binding onto the citrate-reduced AuNP used216 in this work is similar to BSA binding onto the citrate-coated217 gold film. It also indicates that the AuNP LSPR peak218 absorbance has a strong correlation with the amount of BSA219 adsorbed, even though the quantitative relationship between220 these two variables is not currently known. This observation is221 important as it validates the use of AuNP LSPR absorbance for222 in situ monitoring of BSA binding onto collodial AuNP in223 solution.224 Assuming that the BSA adsorption onto AuNP follows a225 Langmuir isotherm,23 the results in Figure 1B strongly226 suggested that the Langmuir binding constant between AuNP227 and BSA should be between 1 × 106 and 5 × 107 M−1. It can be228 shown that if the binding concentration is equal to or smaller229 than 1 × 106 M−1, the BSA packing density on AuNP in the 4230 μM BSA sample will be about 20% lower than that in the 50231 μM BSA solution. Such a difference would most likely induce a232 notable difference in the AuNP LSPR feature between these233 two samples. On the other hand, if the binding constant of BSA234 is higher than 5 × 107 M−1, the packing density of BSA on235 AuNP should be largely the same between the 1 and 4 μM BSA

236samples as the amounts of BSA in both solutions are sufficient237for a full monolayer BSA packing on the AuNPs,23 which would238make the LSPR feature of these two samples identical.239Dynamic Light Scattering and Quantitative BSA240Adsorption Study. The time-dependent DLS measurements241conducted with equal volume mixtures of 10 μM BSA and as-242synthesized AuNP showed that immediately following the243sample preparation (the lead time for the DLS measurement is244∼5 min for sample handling), the particle size of the AuNP245increases from 7.8 ± 0.6 to 11.7 ± 0.5 nm in radius. This 3.9246nm increment in the AuNP radius is consistent with recently247reported data.20 Prolonged incubation (up to 2 days) has no248detectable effect on the AuNP size (Supporting Information,249Figure S3). This result suggests that BSA reaches a steady-state250adsorption onto AuNP within the first 10 min of the251preparation of (AuNP/BSA) mixture under our experimental252conditions, which is consistent with the time-resolved LSPR253result shown Figure 1. However, quantitative BSA adsorption254measurements conducted by centrifugation precipitation of255AuNP with the adsorbed BSA reveal that aging increases the256amount of BSA adsorbed on the AuNP surfaces. The amount of257BSA on centrifugation-precipitated AuNP changes from ∼21 ±2583 BSA per AuNP in the sample aged for 10 min to ∼28 ± 3259BSA per AuNP in the overnight aged (10 h) sample. Further260aging of the AuNP/BSA sample has no significant impact on261the amount of BSA adsorbed (Figure S4, Supporting262Information). The difference between the time courses of263BSA adsorption onto AuNP revealed by the LSPR and DLS264and quantitative BSA measurements is likely due to the nature265of the characterization techniques. While DLS and LSPR are in266situ techniques that do not perturb the sample, the quantitative267BSA adsorption measurement requires prolonged centrifuga-268tion precipitation (90 min) with relatively high centrifuge269forces (9000 g) for AuNP removal. It is possible that the270binding of some of the freshly adsorbed BSA with AuNPs is not271strong enough to sustain the lengthy centrifugation process. As272a result they are dissociated from AuNP during the273centrifugation. In contrast, the strength of the binding between274BSA and the AuNP likely has no detectable effect on the DLS275and LSPR measurements. We hypothesize that the higher BSA276adsorption onto AuNP in the aged samples is due to the aging-277enhanced BSA stability on AuNPs, which is consistent with the278recent report that aging enhances the stability of the protein279corona on AuNPs formed in serum plasma.20

280The quantitative ligand adsorption allows us to determine281packing density of BSA on AuNPs. The nominal footprint of282BSA on AuNPs, calculated using the amount of the BSA283adsorbed in the overnight aged (AuNP/BSA) sample, is about28425 nm2/BSA, which is significantly larger than 17.5 nm2/BSA285reported previously by De Roe et al.24 but in agreement with286the 27 nm2/BSA determined for BSA adsorbed on citrated-287adsorbed gold film.23 Since native human serum albumin (an288analogue of BSA) can be approximated as an equilateral289triangular prism with a thickness of 3 nm and side length of 8290nm,7 our DLS particle size and BSA footprint analysis suggest291that BSA most likely lies flat on the AuNPs with a near full292monolayer BSA coverage in the overnight aged AuNP/BSA293sample.294Organothiol Adsorption onto BSA-Stabilized AuNP.295Regardless of the age of the BSA/AuNP solutions, the BSA-296covered AuNPs exhibit excellent stability in the organothiol297 f2and/or the electrolyte containing solutions (Figure 2), while as-298synthesized AuNPs aggregate rapidly in the organothiol and/or

Figure 1. (A) Time course of the peak UV−vis absorbance of LSPRspectra obtained with (AuNP/BSA) mixture where the concentrationsof AuNP and BSA were 4.6 nM and 5 μM, respectively. The first datapoint (0 min) is acquired with (AuNP/H2O) where the concentrationof AuNP is also 4.6 nM. Inset: Representative time-resolved LSPRspectra of (AuNP/BSA). The spectrum in blue is taken with (AuNP/H2O) (0 min). The arrow indicates increasing incubation time. (B)Correlation between the peak UV−vis absorbance of (AuNP/BSA)with BSA concentration. Inset: Representative steady-state UV−visspectra of (AuNP/BSA) solutions where the concentration of BSA waschanged from 10 nM to 50 μM. The concentration of the AuNP was4.6 nM. All the steady-state LSPR spectra were acquired afterovernight incubation of the (AuNP/BSA). Arrow indicates spectrawith increasing BSA concentration.

The Journal of Physical Chemistry C Article

dx.doi.org/10.1021/jp2107318 | J. Phys. Chem. C XXXX, XXX, XXX−XXXC

299 electrolyte containing solutions. Quantitative comparison of the300 amount of BSA adsorbed onto the overnight aged (AuNP/301 BSA) mixture with and without subsequent MBI addition (that302 is subsequently aged for another night) showed that the303 addition of MBI has no significant effect on the amounts of304 BSA adsorbed (Figure S5, Supporting Information). This result305 is consistent with the literature report that GSH is incapable of306 displacing dye-labeled BSA from AuNPs.14

307 Importantly, the inability of organothiols to displace BSA on308 AuNPs does not exclude the possibility of organothiol309 adsorption onto the BSA-covered AuNPs. Such a possibility310 has not, to our knowledge, been explored before. Our311 quantitative MBI adsorption experiments, conducted with the312 ratiometric SERS method we recently reported for MBI313 quantification,18 reveal that AuNPs retain ∼80% of their MBI

t1 314 binding capacity after the BSA stabilization (Table 1). This

315result, combined with our quantitative BSA adsorption onto316AuNP with and without subsequent MBI adsorption, indicates317that even though BSA maintains a near full packing on the318AuNP, it passivates only ∼20% of the AuNP surface, and the319rest of the surface is or can be made available for organothiol320adsorption. The uptake of AATs by the BSA-stabilized AuNPs321was indirectly confirmed from the reduced MBI uptake by322AuNPs that were sequentially mixed with BSA and competing323AAT (Table 1). The sequential AAT and MBI binding324experiments indicate that some of the binding sites that325would be available for MBI binding in the BSA-stabilized326AuNPs are occupied by the competing AATs, which reduces327the MBI binding capacity of the AuNP.328Several lines of evidence support that the MBI adsorbed is329directly attached to the AuNP, not on the BSA bonded to the330AuNPs. If MBI were bonded through BSA, the amount of MBI331adsorbed would be similar to the amount of BSA adsorbed. It is332deduced from our quantitative BSA and MBI adsorption results333that the amount of MBI adsorbed onto the BSA-covered AuNP334is ∼70 times higher than the BSA adsorbed (Table S1,335Supporting Information). Such a large difference in the amount336of BSA and MBI adsorbed is too high to be attributed solely to337the possible MBI/BSA interactions. Further evidence of direct338MBI/AuNP interactions is the complete quenching of the MBI339UV−vis absorbance upon its adsorption onto the BSA-340 f3stabilized AuNP (Figure 3), while the UV−vis spectrum of a341BSA and MBI mixture is additive of the BSA and MBI UV−vis342spectra. The AuNP quenching of the UV−vis transition of343surface adsorbates has been reported before,25 and it is likely344due to the charge transfer between the surface adsorbate and345AuNPs.26

346Kinetics of the MBI Adsorption onto the BSA-347Stabilized AuNPs. Taking advantage of AuNP adsorption348quenching of the UV−vis absorption of MBI and induced349spectral changes in the AuNP LSPR features, we studied the350kinetics of the MBI adsorption onto the BSA-covered AuNP351using a time-resolved UV−vis spectrum obtained with352 f4((AuNP/BSA)/MBI) (Figure 4). Importantly, the time course353of the UV−vis change in the MBI adsorption region correlates354very well with that of the AuNP LSPR features, indicating that355MBI adsorption further modified the AuNP LSPR features.356The possible effects of BSA concentration on the kinetics357and amount of MBI adsorption were investigated with358((AuNP/BSA)/MBI) solutions where the concentration of359AuNP and MBI were kept constant (3.1 nM and 15 μM,360respectively), but the concentration of BSA was varied from 3.3361 f5t2t3to 10 to 16.5 μM, respectively (Figure 5 and Tables 2 and 3).362All the kinetic data of MBI adsorption were fitted empirically363with two pseudo-first-order reaction functions (eq 1) where k1364and Γ1 refer to the rate constant and the amount of MBI365adsorbed, respectively, for the faster, which is also the major366process responsible for the MBI adsorption.

= Γ − − + Γ − −M k t k t(1 exp( )) (1 exp( ))1 1 2 2 (1)

367Aging of the AuNP/BSA complex reduces both the rate and368the amount of MBI adsorption (Figure 5, Tables 2 and 3).369However, the time scale and the magnitude of the aging effect370on these two parameters are significantly different. While more371than a 10- and a 2-fold reduction in the MBI adsorption time372constant of k1 and k2 were observed within the first 5 min of373sample aging for all the samples, it takes ∼10 h to produce a374detectable aging effect on the total amount of MBI adsorbed375(Γ1 + Γ2) .

Figure 2. Photographs of solution of (1) (((AuNP/H2O)/AAT)/H2O), (2) (((AuNP/H2O)/AAT)/PBS), (3) (((AuNP/BSA)/AAT)/H2O), and (4) (((AuNP/BSA)/AAT)/PBS). The AAT is (A) MBI,(B) Cys, (C) Hcy, and (D) GSH. For samples (3) and (4), the(AuNP/BSA) solution was aged for 24 h before the addition of AAT.Pictures were taken after the final component was added for 24 h.Concentration of the AuNP, BSA, and AAT are 2.3 nM, 2.5 μM, and 8μM, respectively. The composition of PBS used was 11.9 mMNa2HPO4, 137 mM NaCl, and 2.7 mM KCl.

Table 1. Quantitative MBI Adsorption and DLS Particle Sizeof AuNPa

sample ΓMBIb (μM) particle size in radiusc (nm)

(((AuNP/H2O)/H2O)/H2O) NA 7.7 ± 0.6(((AuNP/BSA)/H2O)/H2O) NA 11.4 ± 0.5(((AuNP/H2O)/H2O)/MBI) 5.7 ± 0.5 aggregated(((AuNP/BSA)/H2O)/MBI) 4.8 ± 0.5 12.2 ± 0.5(((AuNP/BSA)/Cys)/MBI) 1.8 ± 0.4 11.8 ± 0.4(((AuNP/BSA)/Hcy)/MBI) 1.2 ± 0.3 11.7 ± 0.5(((AuNP/BSA)/GSH)/MBI) 3.5 ± 0.3 11.6 ± 0.4

aThe components in the inner parentheses of (((A/B)/C)/D) weremixed first and aged 2 h before the addition of the third and fourthcomponents. Concentrations of the AuNP, BSA, MBI, and AAT in thesolutions are 2.3 nM, 2.5 μM, 8 μM, and 8 μM, respectively.bAdsorption of MBI was measured using the ratiometric SERS methodafter the samples were incubated overnight at room temperature(Supporting Information). cParticle size measured by DLS.

The Journal of Physical Chemistry C Article

dx.doi.org/10.1021/jp2107318 | J. Phys. Chem. C XXXX, XXX, XXX−XXXD

376 The independence of the rate and amount of MBI adsorption

377 from the BSA concentrations tested in the AuNP/BSA mixtures

378 further supports a relatively high binding constant (>106 M−1)

379 between BSA and AuNP .24 If the binding constant between

380 BSA and AuNP is smaller than 106 M−1, the Langmuir

381 equilibrium packing density of BSA on the AuNP had to be382 significantly different among the three (AuNP/BSA) samples

383tested (Tables 2 and 3), which would inevitably induce

384significant differences in the rates and amounts of MBI385adsorption.

386The near-perfect fitting of the MBI kinetic data with the two

387pseudo-first-order reactions indicates that the MBI adsorption

388can be approximated by two parallel, first-order processes of389different rate constants. Conceivably, MBI binding onto the

Figure 3. (A) (blue) Additive spectrum of ((AuNP/BSA)/H2O) and ((MBI/H2O)/H2O); (green) experimental UV−vis spectrum ((AuNP/BSA)/MBI). Inset: (black) UV−vis spectra of ((MBI/H2O)/H2O), (red) the difference spectra obtained by subtracting the experimental spectrum of((AuNP/BSA)/MBI) from the additive spectra of ((AuNP/BSA)/H2O) and ((MBI/H2O)/H2O). (B) (blue) Additive spectrum of ((BSA/H2O)/H2O) and ((MBI/H2O)/H2O), (green) experimental UV−vis spectrum ((BSA)/MBI) /H2O). Inset: (black) UV−vis spectra of ((MBI/H2O)/H2O), (red) the difference spectra obtained by subtracting the experimental spectrum of ((BSA)/MBI) /H2O) from the additive spectra of ((BSA/H2O)/H2O) and ((MBI/H2O)/H2O). The concentrations of AuNP, BSA, and MBI in (A) are 3.1 nM, 3.3 μM, and 5 μM, respectively.

Figure 4. (A) Representative time-resolved UV−vis spectra of ((AuNP/BSA)/MBI) complex. Inset: (a) Zoom-in of the UV−vis spectra in theregion from 290 to 310 nm where the MBI absorbs. (b) Zoom-in of the region of the AuNP peak absorbance. (B) Time courses of amount of MBIadsorbed (black squares) and the change in the UV−vis peak absorbance of AuNP (red squares); both are calculated on the basis of the differencespectra shown in the inset. Inset: Representative UV−vis difference spectra obtained by subtracting the time-resolved UV−vis spectrum of ((AuNP/BSA)/MBI) from the control that is the additive spectra of ((AuNP/BSA)/H2O) and ((MBI/H2O)/H2O). The arrow indicates the differencespectra of increasing time. Concentrations of AuNP, BSA, and MBI are 3.1 nM, 3.3 μM, and 15 μM, respectively.

Figure 5. Kinetics of MBI adsorption onto (AuNP/BSA) complexes that are aged (a−e) for 5 s, 10 min, 1 h, 48 h. and 72 h, respectively. Red solidcurves were obtained by fitting the kinetics data (in dots) with two pseudo-first-order reaction equations (eq 1 in the text). The concentrations ofAuNP and MBI were kept constant at 3.1 nM and 15 μM, respectively, but the BSA concentrations were (A) 3.3, (B) 10, and (C) 16.5 μM.

The Journal of Physical Chemistry C Article

dx.doi.org/10.1021/jp2107318 | J. Phys. Chem. C XXXX, XXX, XXX−XXXE

390 BSA-covered AuNP is a two-step process, the diffusion of MBI391 in bulk aqueous solution (Mb) onto the AuNP surfaces (Ms)392 and the MBI/AuNP binding (eqs 3−5). While the average rate393 of the MBI diffusion to the AuNP surfaces is dictated by the394 morphological features (inter- and intraprotein spacing) of the395 BSA coating layer on AuNP, the rate of MBI binding onto396 AuNP is much more complicated. Theoretically the surface of397 the BSA-covered AuNP can be divided into three categories398 (eqs 3−5): (i) void area SV, i.e., the AuNP surface on which399 there are no BSA amino acid residues in direct contact, (ii)400 loosely bonded area SL, i.e., the area on which BSA amino acid401 residues are in direct contact, but the binding between these402 amine acid residues and gold is too weak to prevent MBI403 displacement, and (iii) tightly bonded area ST where binding404 between the BSA amino acid residues with AuNP is strong405 enough to preclude MBI displacement. Since the binding406 between MBI and ST is impossible (kT = 0), MBI binding onto407 a BSA-covered surface can be simplified as a combination of408 two parallel processes, i.e, MBI adsorption onto SV and SL.409 Step 1:

→M Mk

B Sd

(2)

410 Step 2:

+ →M S M Sk

S V S VV

(3)

+ →M S M Sk

S L S LL

(4)

+ →M S M Sk

S T S TT

(5)

411Control experiment showed that MBI binding to a naked412AuNP is an exceedingly rapid process (90% of the ligand413adsorption is completed within seconds) (Figure S6,414Supporting Information). As a result the rate limiting step for415MBI binding to SV is kd, the rate constant of MBI diffusion416through the BSA coating layer onto AuNPs. However, for MBI417binding to SL, the overall rate constant is likely dictated by kL,418the rate constant of MBI displacing the loosely bonded BSA419amino acids.420Assignment of the two MBI adsorption processes charac-421terized by (k1, Γ1) and (k2, Γ2) were made on the basis of422differential aging effect on k1 and k2 (Table 2). Conceivably,423aging should have a significant effect on the kd value as any424aging-induced additional BSA adsorption and/or BSA spread-425ing on the AuNPs would greatly reduce the kd value. In426contrast, the effect of aging on the kL value is likely small427because the strength of the interaction between the loosely428bonded BSA amino acids and AuNP is unlikely to have big429change with aging. On the basis of these considerations, and the430observation that aging has much larger effect on k1 than on k2,431we hypothesize that the rate constants and the adsorption432capacities of (k1, Γ1) and (k2, Γ2) described in Tables 2 and 3

Table 2. Time Constants Obtained by Fitting the Kinetics Data of MBI Adsorption on (AuNP/BSA) Complex

[BSA] = 3.3 μM [BSA] = 10 μM [BSA] = 16.5 μM

ta (min) k1b (s−1) k2

b (s−1) k1b (s−1) k2

b (s−1) k1b (s−1) k2

b (s−1)

0.08 4.6 ± 1.4 0.02 ± 0.01 10 ± 3.0 0.06 ± 0.02 10 ± 3.6 0.06 ± 0.020.25 3.8 ± 1.8 0.03 ± 0.02 5.9 ± 2.0 0.06 ± 0.01 5.6 ± 2.0 0.07 ± 0.0011.00 4.0 ± 1.4 0.08 ± 0.02 3.8 ± 0.8 0.05 ± 0.02 4.4 ± 0.8 0.06 ± 0.0015.00 0.3 ± 0.03 0.01 ± 0.001 0.3 ± 0.07 0.02 ± 0.01 0.3 ± 0.07 0.03 ± 0.0210.00 0.2 ± 0.03 0.01 ± 0.001 0.2 ± 0.03 0.01 ± 0.001 0.2 ± 0.03 0.01 ± 0.00130.00 0.2 ± 0.02 0.01 ± 0.001 0.2 ± 0.02 0.01 ± 0.001 0.2 ± 0.02 0.01 ± 0.00160.00 0.1 ± 0.02 0.01 ± 0.001 0.1 ± 0.02 0.01 ± 0.001 0.2 ± 0.02 0.01 ± 0.001720.00 0.1 ± 0.01 0.01 ± 0.001 0.1 ± 0.01 0.01 ± 0.001 0.1 ± 0.02 0.01 ± 0.0012900.00 0.06 ± 0.01 0.01 ± 0.001 0.06 ± 0.01 0.01 ± 0.001 0.06 ± 0.01 0.008 ± 0.0014300.00 0.05 ± 0.01 0.01 ± 0.001 0.05 ± 0.01 0.01 ± 0.001 0.05 ± 0.01 0.008 ± 0.001

aAging time of (AuNP/BSA) complex before mixing with MBI. bRate constants of MBI adsorption onto (AuNP/BSA) complex extracted by fittingthe kinetics data. The concentrations of AuNP and MBI are 3.1 nM and 15 μM, respectively. Measurement results calculated from three independentmeasurements.

Table 3. Amount of the MBI Adsorption Obtained by Fitting the Kinetic Data of MBI Adsorption onto (AuNP/BSA)

[BSA] = 3.3 μM [BSA] = 10 μM [BSA] = 16.5 μM

ta (min) Γ1b (μM) Γ2

b (μM) Γ1b (μM) Γ2

b (μM) Γ1b (μM) Γ2

b (μM)

0.08 4.9 ± 0.4 1.2 ± 0.2 3.5 ± 0.2 2.8 ± 0.3 3.5 ± 0.2 2.9 ± 0.30.25 4.7 ± 0.3 1.7 ± 0.5 3.5 ± 0.3 2.7 ± 0.1 3.3 ± 0.2 2.9 ± 0.31.00 4.5 ± 0.3 1.6 ± 0.2 4.1 ± 0.4 2.4 ± 0.2 4.1 ± 0.5 2.4 ± 0.35.00 4.6 ± 0.3 1.2 ± 0.2 4.2 ± 0.5 2.4 ± 0.2 4.6 ± 0.4 1.8 ± 0.310.00 4.5 ± 0.4 1.2 ± 0.2 4.9 ± 0.1 1.4 ± 0.2 4.8 ± 0.1 1.5 ± 0.330.00 4.5 ± 0.1 1.2 ± 0.1 4.8 ± 0.2 1.5 ± 0.2 4.7 ± 0.1 1.6 ± 0.360.00 4.2 ± 0.1 1.2 ± 0.1 4.4 ± 0.2 1.7 ± 0.2 4.5 ± 0.2 1.8 ± 0.3720.00 3.4 ± 0.2 1.7 ± 0.2 3.7 ± 0.1 2.3 ± 0.1 3.5 ± 0.3 2.3 ± 0.32900.00 3.2 ± 0.2 1.7 ± 0.1 3.1 ± 0.1 2.4 ± 0.1 3.4 ± 0.2 2.3 ± 0.34300.00 3.4 ± 0.1 1.7 ± 0.1 3.2 ± 0.1 2.3 ± 0.1 3.2 ± 0.2 2.4 ± 0.3

aAging time of (AuNP/BSA) complex before mixing with MBI. bThe amount of MBI adsorbed onto (AuNP/BSA) complex by fitting the kineticdata. The concentrations of AuNP and MBI are 3.1 nM and 15 μM, respectively. Measurement results calculated from three independentmeasurements.

The Journal of Physical Chemistry C Article

dx.doi.org/10.1021/jp2107318 | J. Phys. Chem. C XXXX, XXX, XXX−XXXF

433 correspond to the processes of MBI adsorption to SV and SL,434 respectively.435 Using the theoretical saturation packing density of 600436 pmol/cm2 for MBI on naked AuNPs,27 it is deduced from the437 data in Table 3 that the fraction of the AuNP surface areas (ST)438 passivated by the BSA coating layer increases from ∼20% in the439 freshly prepared sample to ∼30% in the fully aged samples (70440 h or more). Since both the DLS and LSPR data indicate that441 BSA adsorption is largely completed within the first 10 min of442 the BSA mixing with AuNP, the only sensible explanation of443 the slow, but significant (50%) increment in the fraction of the444 BSA-passivated AuNP surface is that, with aging, the adsorbed445 BSA spreads and forms more anchoring points on the AuNP446 surface that are stable against MBI displacement. It is important447 to note, however, even in the most aged (AuNP/BSA) sample,448 there is still a large fraction of the AuNP surface that remains449 void or loosely bonded to BSA amino acid residues that allows450 subsequent, high-capacity MBI adsorption.451 The fact that there is still a notable aging effect on MBI452 adsorption in the 2-day-aged BSA/AuNP mixtures indicates453 that BSA deformation on the AuNP is a lengthy process.454 Otherwise, the MBI binding kinetics and capacity would be the455 same for the AuNP/BSA samples that are aged 2 days or more.456 The ability for BSA to change structure on the AuNP has a457 significant implication for the BSA structure on the AuNP458 surface. Although BSA on AuNP is “hardened” in terms of its459 stability after 1 day of the (AuNP/BSA) aging, it remains “soft”460 enough to undergo additional structural modification as461 evidenced by the reduced MBI adsorption rate constants and462 capacity in 3-day-aged AuNP.463 Additional important insight into the binding characteristics464 between BSA and AuNP is garnered on the basis of the amount465 of MBI adsorption onto the BSA-stabilized AuNPs (Table 3).466 Although BSA has a nominal footprint of 25 nm2/protein on467 the AuNPs in the fully aged (AuNP/BSA) mixtures (aged 1 day468 or more), BSA passivates only ∼9 nm2 of the AuNP surface469 against MBI adsorption in the 72-h-aged sample. The rest of470 the surface is either void (∼10.6 nm2) or linked to loosely471 bonded BSA amino acids that can be replaced by MBI (∼5.3472 nm2), as estimated from the amount of MBI adsorbed in Table473 3. Taking into consideration that the average van der Waals474 radius for the amino acids is 3.3 Å,28 and the necessary spacing475 between different protein anchoring points, this 9 nm2

476 passivated surface indicates that for each adsorbed BSA, the477 number of amino acid residues that can be in direct contact478 with the AuNP should be smaller than 26. This conclusion is479 consistent with the recent NMR study of ubiquitin bonded480 onto AuNPs which shows there are only five amino acids481 residues involved in the AuNP/protein interaction.29

482 Taking all the experimental results into consideration, we483 propose the following mechanism explaining structural484 evolution of BSA on AuNPs and its impact on subsequent

f6 485 organothiol adsorption (Figure 6). When BSA is initially486 adsorbed onto the AuNPs, the interprotein spacing between487 the surface-adsorbed BSAs is likely large. As a result MBI488 penetrates rapidly through the BSA coating layer and is489 adsorbed onto the AuNPs. With aging, however, the adsorbed490 BSA spreads on AuNP, which on one hand increases the BSA491 stability on the AuNP while on the other hand reduces the492 speed and quantity of the subsequent organothiol adsorption493 onto the BSA-covered AuNPs. It is important to note, however,494 even for the most aged (3 days) AuNP/BSA sample, there is495 still a large amount of void space between the AuNP and BSA

496coating layer, and the protein coating layer remains exceedingly497permeable. Indeed, the data in Figure 5 showed that for all the498samples tested, over 80% of MBI adsorption occurs within the499first 5 min of MBI mixing with (AuNP/BSA) complex (Table5003).

501■ CONCLUSION

502Organothiol is a powerful molecular probe for studying the503structural evolution, morphology, and binding stability of504protein on AuNPs. With this technique, we found that BSA505interaction with AuNP is an exceedingly lengthy process. While506BSA adsorption per se is rapid and is largely completed within50710 min of the sample preparation, the conformational508modification continues for at least 2 days. Even though BSA509forms a nearly full monolayer packing on AuNP, it passivates510only a small fraction of the AuNP surfaces against further511organothiol adsorption. In addition to shedding critical new512insights on the morphological feature and structural evolution513of protein on AuNPs, the results from our organothiol514adsorption study also have significant implications for AuNP515applications. For example, the ability of the protein-covered516AuNP to uptake organothiols with high binding capacity opens517a convenient avenue for fabrication of multicomponent (nano,518bio, and organo) composite materials that can be exploited in519AuNP applications including nanoparticle drug delivery, nano/520biomaterial fabrication, and biosensor development. This521finding also raises new nanotoxicity concerns of AuNPs.522Given the biological significance of the amino acid thiols to523human health30,31 and their relative high abundance in biofluids524including serum plasma, the possibility for inclusion of525organothiol into protein-covered AuNPs and its potential526impact on the toxicity and functionality of the protein-covered527AuNP should not be overlooked in biological/biomedical528applications of AuNPs.

529■ ASSOCIATED CONTENT

530*S Supporting Information531Detailed experimental procedures which include AgNP and532AuNP synthesis, ratiometric SERS quantification of MBI,533LSPR, and DLS measurements, and quenching of MBI

Figure 6. Proposed mechanism of the effect of aging on BSA structureon AuNP.

The Journal of Physical Chemistry C Article

dx.doi.org/10.1021/jp2107318 | J. Phys. Chem. C XXXX, XXX, XXX−XXXG

534 absorbance on AuNP. This material is available free of charge535 via the Internet at http://pubs.acs.org.

536 ■ AUTHOR INFORMATION537 Corresponding Author538 *E-mail: [email protected] Notes540 The authors declare no competing financial interest.

541 ■ ACKNOWLEDGMENTS542 This work was supported in part by a NSF fund (EPS-543 0903787) provided to D.Z. The authors thank Dr. Willard544 Collier and Ms. Laura Lewis for their editorial assistance.

545 ■ REFERENCES(1)546 Monopoli, M. P.; Walczyk, D.; Campbell, A.; Elia, G.; Lynch, I.;

547 Baldelli Bombelli, F.; Dawson, K. A. J. Am. Chem. Soc. 2011, 133, 2525.(2)548 Lacerda, S. H. D. P.; Park, J. J.; Meuse, C.; Pristinski, D.; Becker,

549 M. L.; Karim, A.; Douglas, J. F. ACS Nano 2009, 4, 365.(3)550 Mahmoudi, M.; Lynch, I.; Ejtehadi, M. R.; Monopoli, M. P.;

551 Bombelli, F. B.; Laurent, S. Chem. Rev. 2011, 5610.(4)552 Li, N.; Zeng, S.; He, L.; Zhong, W. Anal. Chem. 2010, 82, 7460.(5)553 Maiorano, G.; Sabella, S.; Sorce, B.; Brunetti, V.; Malvindi, M. A.;

554 Cingolani, R.; Pompa, P. P. ACS Nano 2010, 4, 7481.(6)555 Cedervall, T.; Lynch, I.; Lindman, S.; Berggard, T.; Thulin, E.;

556 Nilsson, H.; Dawson, K. A.; Linse, S. Proc. Natl. Acad. Sci. U.S.A 2007,557 104, 2050.

(7)558 Rocker, C.; Potzl, M.; Zhang, F.; Parak, W. J.; Nienhaus, G. U.559 Nat. Nano 2009, 4, 577.

(8)560 Tsai, D.-H.; DelRio, F. W.; Keene, A. M.; Tyner, K. M.;561 MacCuspie, R. I.; Cho, T. J.; Zachariah, M. R.; Hackley, V. A.562 Langmuir 2011, 27, 2464.

(9)563 Wangoo, N.; Suri, C. R.; Shekhawat, G. Appl. Phys. Lett. 2008, 92,564 133104.

(10)565 Treuel, L.; Malissek, M.; Gebauer, J. S.; Zellner, R.566 ChemPhysChem 2010, 11, 3093.

(11)567 Rozanova, N.; Zhang, J. Z. J. Biomed. Nanotechnol. 2008, 4, 377.(12)568 Gole, A.; Murphy, C. J. Langmuir 2005, 21, 10756.(13)569 Zhang, D.; Neumann, O.; Wang, H.; Yuwono Virany, M.;

570 Barhoumi, A.; Perham, M.; Hartgerink Jeffrey, D.; Wittung-Stafshede,571 P.; Halas Naomi, J. Nano Lett. 2009, 9, 666.

(14)572 Xie, H.; Tkachenko, A. G.; Glomm, W. R.; Ryan, J. A.;573 Brennaman, M. K.; Papanikolas, J. M.; Franzen, S.; Feldheim, D. L.574 Anal. Chem. 2003, 75, 5797.

(15)575 Lee, P. C.; Meisel, D. J. Phys. Chem. 1982, 86, 3391.(16)576 Frens, G. Nature 1973, 241, 20.(17)577 Freeman, R. G.; Hommer, M. B.; Grabar, K. C.; Jackson, M. A.;

578 Natan, M. J. J. Phys. Chem. 1996, 100, 718.(18)579 Zhang, D.; Ansar, S. M. Anal. Chem. 2010, 82, 5910.(19)580 Zhang, D.; Ansar, S. M.; Vangala, K.; Jiang, D. J. Raman

581 Spectrosc. 2010, 41, 952.(20)582 Casals, E.; Pfaller, T.; Duschl, A.; Oostingh, G. J.; Puntes, V.

583 ACS Nano 2010, 4, 3623.(21)584 Delfino, I.; Cannistraro, S. Biophys. Chem. 2009, 139, 1.(22)585 Deka, J.; Paul, A.; Chattopadhyay, A. J. Phys. Chem. C 2009, 113,

586 6936.(23)587 Brewer, S. H.; Glomm, W. R.; Johnson, M. C.; Knag, M. K.;

588 Franzen, S. Langmuir 2005, 21, 9303.(24)589 De Roe, C.; Courtoy, P. J.; Baudhuin, P. J. Histochem. Cytochem.

590 1987, 35, 1191.(25)591 Reyes, E.; Madueno, R.; Blazquez, M.; Pineda, T. J. Phys. Chem.

592 C 2010, 114, 15955.(26)593 Franzen, S.; Folmer, J. C. W.; Glomm, W. R.; O’Neal, R. J. Phys.

594 Chem. A 2002, 106, 6533.(27)595 Raj, C. R.; Kitamura, F.; Ohsaka, T. Analyst 2002, 127, 1155.(28)596 Thomas, C. E. Proteins: Structrures and Molecular Properties, 2nd

597 ed.; W. H. Freeman and Company: New York, 2002.

(29) 598Calzolai, L.; Franchini, F.; Gilliland, D.; Rossi, F. Nano Lett.5992010, 10, 3101.

(30) 600Alkilany, A.; Murphy, C. J. Nanopart. Res. 2010, 12, 2313.(31) 601Murphy, C. J.; Gole, A. M.; Stone, J. W.; Sisco, P. N.; Alkilany,

602A. M.; Goldsmith, E. C.; Baxter, S. C. Acc. Chem. Res. 2008, 41, 1721.

The Journal of Physical Chemistry C Article

dx.doi.org/10.1021/jp2107318 | J. Phys. Chem. C XXXX, XXX, XXX−XXXH