J. Phys. Chem. A 2010, 114, 1521–1528
1521
Temperature Dependence of Conjugation of Amyloid Beta Protein on the Surfaces of Gold Colloidal Nanoparticles† Kazushige Yokoyama,* Nicole B. Gaulin, Hyunah Cho, and Nicole M. Briglio‡ Department of Chemistry, The State UniVersity of New York, Geneseo College, One College Circle, Geneseo, New York 14454 ReceiVed: August 14, 2009; ReVised Manuscript ReceiVed: December 2, 2009
The absorption spectrum of the amyloid beta 1-40 peptide (Aβ1-40) conjugated to gold colloidal suspension of 15, 20, 30, and 40 nm size were examined under temperature ranging from 5 to 50 °C. As the pH was externally altered repetitively between pH 4 and 10, Aβ1-40-coated 20 nm gold colloid nanoparticles exhibited a reversible color change at the entire temperature range tested in this study except for 5 ( 0.2 °C. This reversible change may be due to the fact that hydrophilic Aβ1-40 evolves between a three-dimensional network containing mainly β-sheet and R-helices, and an intermediate of this process implies a reversible step reported as initiation of the fibrillogenesis in Alzheimer’s disease. When other nanosize particles were investigated, Aβ1-40-coated 30 and 40 nm colloids exhibited the reversible color change when temperature was lowered to 18 ( 0.2 and 6 ( 0.2 °C, respectively. This specific and unique size and temperature dependence in reversible color change strongly suggests that the noncovalent intrinsic intermolecular potential formed between the nanocolloidal surface and each Aβ1-40 monomer conjugated at the surface drives the process. Introduction Pathologically, a key hallmark of the neuritic and cerebrovascular amyloid in Alzheimer’s disease is the formation of insoluble fibrillar deposits of the amyloid beta peptide (Aβ) as both diffuse and senile amyloid plaque that invades the brain’s seat of memory and cognition (brain parenchyma and vasculature) before spreading to other areas.1–4 Although a terminal process of misfolding producing fiber-like insoluble aggregates of Aβ has been extensively studied,5–7 an initial stage involving a soluble Aβ complex has been regarded as an important step and a key onset for the subsequent aggregation.6,8,9 The precursor of the fibril, nonfibrillar soluble oligomer forms of Aβ, possess neurotoxic properties and may, therefore, play a role in the molecular pathogenesis of Alzheimer’s disease.6 As a result, the oligomers have been regarded as the most dangerous entities for brain cells, whereas their fully formed fibril are not.8,9 Although still relatively little is known about the structure of this soluble oligomer Aβ, a selective detection of the oligomer form of Aβ aggregate will be an important step in conducting a study of determining its neurotoxicity and clarifying its role in fibrillogenesis. A direct detection of the reversible process associated with conformation in an oligomer has not yet been fully achieved. The conformation found on the membrane surface is a critical stage for initial fibrillogenesis, so the direct investigation of Aβ placed under the interfacial environment must realistically project the conformation associated with intermediates in fibrillogenesis. Generally speaking, the proteins immobilized at an interface are expected to behave differently from their counterparts in bulk solutions,10–13 and understanding the interactions of the proteins on the surface is challenging but crucial to fully understand the mechanism of fibrillogenesis †
Part of the “W. Carl Lineberger Festschrift”. * To whom correspondence should be addressed. E-mail: yokoyama@ geneseo.edu. ‡ Current address: Department of Chemistry, University of Rochester, Rochester, NY, 14627.
involving the membrane surface. Perutz and Windle considered the epidemiological data of Huntington’s disease and modeled the aggregation process with a homogeneous nucleation theory.14 Amyloid aggregates are believed to grow through a nucleation mediated pathway, whereas the nucleation radius remains unclear. To investigate the fibril precursor at the interfacial environment, our group took an approach to prepare the Aβ on an interfacial boundary of the surface of gold colloidal nanoparticles and investigated the associated structural information of conjugated Aβ.15–17 Our group intended to create a boundary environment for Aβ to reproduce a structure similar to Aβ located at brain cell or membrane interfacial surfaces. Although metal colloidal surfaces do not create the same environment as physiological conditions, this approach enabled us to systematically examine the structure of Aβ on the size-controlled interfacial environment by externally varying the pH. When the conjugation of various sequences of Aβ (Aβ1-11, Aβ12-28, Aβ31-35, Aβ1-40, and Aβ1-42) with gold colloidal suspension of 20 nm size was examined by absorption spectroscopy, color changes were seen for all tested proteins at a higher pH than what bare gold colloid exhibits, under pH ranging from pH 2 to pH 10. The color changes observed at a pH lower than 5 are attributed to the unfolded Aβ monomer units around the gold colloidal surface. Interestingly, only Aβ1-40-coated gold colloidal nanoparticles exhibited a reversible color change as the pH was externally altered between pH 4 and 10. This reversibility is interpreted to be an important implication of the observation of a reversible step reported for the fibrillogenesis as mentioned above.15 We further investigated the conjugation of Aβ1-40 and chicken egg albumin (ovalbumin) with various sizes of gold colloidal nanoparticles under various pH values, ranging from pH 2 to pH 10. The pH value that indicates the color change, pHo, exhibited colloidal size dependence for both Aβ1-40 and ovalbumin-coated particles. In particular, Aβ1-40-coated gold colloidal particles exhibited noncontinuous size dependence peaking at 40 and 80 nm, implying that their corresponding
10.1021/jp907880f 2010 American Chemical Society Published on Web 12/29/2009
1522
J. Phys. Chem. A, Vol. 114, No. 3, 2010
cage-like structures provide efficient net-charge cancellation at these core sizes. Remarkably, only the pHo value for ovalbumincoated 80 nm gold colloid was pH > 7, and a specific cage-like structure is speculated to have a positive net charge facing outward when ovalbumin self-assembles over this particular gold colloid. For ovalbumin-coated gold colloids, all tested sizes of gold colloidal nanoparticles showed a quasi-reversible color change whereas only 20 nm gold colloids exhibited reversible color change as Aβ1-40 is conjugated at room temperature.16 This reversible color change was also examined by microscopic investigations. AFM images on graphite surfaces revealed the morphology of Aβ aggregates with gold colloids. TEM images clearly demonstrate the correspondence between spectroscopic features and conformational changes of the gold colloid.17 The conformational change uniquely observed on the surface of the gold colloidal nanoparticles in our studies has opened new opportunities for investigation of protein associated nanomaterials.18 In this study, the reversible color change was examined at various temperature conditions for Aβ1-40conjugated 20 nm gold colloids. We aim to understand the role of the thermal condition in the reversible process as well as the corresponding conformation of the Aβ aggregates formed over the gold colloid surface. The variation of the thermal condition must change the surface potential enough to overcome the physical restriction (particularly size of the gold colloids) preventing the reversible process from taking place at room temperature otherwise. If this assumption is true, the reversible process may be observed in sizes of the gold colloids other than 20 nm diameter at temperatures other than room temperature, and the condition that supports the reversible process may describe the possible structure or conformation of the intermediates associated with the fibrillogenesis. Experimental Section The detailed contents on the experimental procedure utilized in this study were already given,15–17 and only a brief description is provided here. The ultrapure Aβ1-40 (MW: 4329.9 Da), in the form of lyophilized powder (97% by HPLC) was purchased from American Peptide (Sunnyvale, California, USA) and was stored at -12 °C. The stock solution of 100 µM Aβ1-40 was prepared at approximately 18 °C using double-distilled deionized and filtered water. The amount of Aβ was determined by UV absorption at 280 nm (absorbance of Tyrosine at 275 nm, 275 ) 1390 cm-1 M-1).19 Gold colloidal nanoparticles were purchased from Ted Pella Inc. (Redding, California, USA) with the following diameters (and particles/mL in O.D. ) 0.2 at 528 nm): 15.2 ( 1.5 nm (2.8 × 1011 particles/mL), 19.7 ( 1.1 nm (1.4 × 1011 particles/mL), 30.7 ( 1.3 nm (4.0 × 1010 particles/ mL), and 40.6 ( 1.1 nm (1.8 × 1010 particles/mL). In the procedure, any water used was purified to more than 18 MΩ using a Milli-Q water system (Millipore) in order to prevent salt-induced aggregation of the gold particles and to prevent possible interference in the adsorption of the protein to the gold particle surface.20 The concentration of protein was adjusted to a factor of 1000 more than the gold colloids (e.g., 20 nm colloid particles in this experiment were 2.30 nM and the concentration of the protein was 2.30 µM), which was found to optimize the conjugation of Aβ over the gold surface. A series of absorption spectra were collected by a UV-vis spectrophotometer (Varian Carry Bio 300 Varian Inc. Palo Alto, California USA) with a UV-enhanced quartz cuvette cell. The temperature control of the absorption cell was conducted by a feedback loop temperature control system with the Peltier device. The temperature of the solution was monitored by a
Yokoyama et al. temperature sensor installed in the cuvette holder and was confirmed by a digital thermometer inserted in a sample cuvette for the temperature range between 5 and 50 °C taking 10, 20, 30, and 40 °C as middle data points with an accuracy of (0.2 °C. A step of 1 °C was taken when more fine temperature steps were required to clearly observe the transition in spectral features. The pH of the solution was repeatedly altered between pH 4 and pH 10 for 10 cycles, and the corresponding absorption spectrum was monitored at each pH condition. Here, one cycle indicates the transition where pH changes from pH 4 to pH 10 to pH 4. The pH value of each sample was directly measured in a cuvette cell while stirring, using a micro pH electrode. For each operation of acid/base insertion, the pH value was managed to reproduce pH 4 or pH 10 with an accuracy of (0.1. The original pH of freshly prepared sample was around pH 7. The pH change to pH 4 from pH 7 was completed by dropwise addition of hydrochloric acid (HCl), and the change to pH 10 from pH 4 was by addition of sodium hydroxide (NaOH). The utilized acids (HCl) and bases (NaOH) of various concentration were prepared in the temperature control cell holder with targeted temperature in order to minimize the temperature variation during the pH adjustment. Results The average peak position of the surface plasmon resonance (SPR) band of the regions between 400 and 800 nm was monitored as the pH was repetitively changed between pH 4 and pH 10. We used the identification of operation to alter pH condition by “n”, which varies from 1 to 21 for the process of 10 cycles. Here, n ) 1 corresponds to the starting pH condition of solution before acid is inserted and its pH was around pH 7. After n ) 1, odd numbers of n (nodd ) 3, 5, 7,..., 21) indicate an operation of acid addition leading the pH of solution to be pH 4, whereas even numbers of n (neven ) 2, 4, 6,..., 20) show an operation of base addition to increase the pH of solution to be pH 10. The average peak position at given temperature (T) and operation (n), λpeak(n, T), was calculated based on eq 1,
λpeak(n, T) )
∑ ai(n, T)λi(n, T)
(1)
i
where λi(n, T) and ai(n, T) represent the peak position and fraction of the ith component band. In this study, the fraction ai was determined by the fraction of the area (Ai) of the band to the area of the total sum of the entire bands, for example, a1(n, T) ) A1/(A1 + A2) for the case of two bands. Most of the bands observed in our study were analyzed with two components or one component with a large background band with a maximum at 350 ( 50 nm. All the absorption bands were fully explained by the fit with Gaussian profile using the peak-fitmodule of ORIGIN (Version 7.0) in the range of 400-800 nm. After the background band was excluded, the average peak shift for two components band is given by λpeak(n, T) ) a1(n, T)λ1(n, T) + a2(n, T)λ2(n, T). The color change is considered to be caused by electric dipole-dipole interactions and coupling between surface plasmons of neighboring particles inside the aggregates, provided that the interparticle distance inside the aggregate is smaller than the gold colloidal particle diameter. The initial color of the bare gold nanoparticle solutions or Aβ1-40-coated gold particles had a pH around neutral with reddish color; thus, the SPR band is maximum at 528 nm. The color of the solution that changed into bluish color had bimodal components with absorption band maximum around 600 nm as
Temperature Dependence of Nanoscale Conjugation
J. Phys. Chem. A, Vol. 114, No. 3, 2010 1523
Figure 2. Collections of peak shift, λpeak(n, T), as a function of pH changes (n) for Aβ1-40-coated gold colloid particles of (a) 15 nm, (b) 20 nm, (c) 30 nm, and (d) 40 nm at selected temperatures between 5 and 50 °C. The dashed lines are provided to indicate the feature of the peak shift at a given pH.
Figure 1. (a) A comparison of peak shift (λpeak) as a function of pH changes between 20 nm gold colloid (white circles) and Aβ1-40coated 20 nm gold colloid particles (black circles) at 20 °C. The upward and downward arrows indicate an injection of HCl and NaOH, adjusting the pH of the solution to pH 4 and pH 10, respectively. The dashed lines indicate the values predicted by eq 2. The thin line is provided to indicate the trend of the peak shift observed for 20 nm gold colloid particles. (b) A peak shift (λpeak) as a function of pH changes for Aβ1-40-coated 20 nm gold colloid particles at 20 °C is shown along with the TEM images of Aβ1-40coated 20 nm gold colloids: n ) 1 (pH 7), n ) 2 (pH 4), n ) 3 (pH 10), n ) 7 (pH 10), n ) 8 (pH 4), n ) 13 (pH 10), n ) 14 (pH 10), n ) 20 (pH 4), and n ) 21 (pH 10).
the pH was adjusted to pH 4. These bimodal absorption peaks consisted of a band with maximum around 528 ( 1 nm (λ1(n, T)) indicating the existence of free gold colloidal nanoparticles and a band with a maximum between 590-630 nm (λ2(n, T)). The contribution of the λ1(n, T) component was larger at higher pH values (pH > 7), whereas the λ2(n, T) component shifts to red and gains more amplitude at pH 4; this component became more prominent than the λ1(n, T) component at this pH. When the reversible color change took place, the color of the solution returned to reddish color as the pH condition was shifted back to pH 10 from pH 4 as observed in the Aβ1-40coated 20 nm gold colloids (Figure 1a). To show the dispersed or aggregated form of gold colloids at basic or acidic condition, the TEM images are shown for representative n in Figure 1b. It should be noted that the color change in the reversible process was not between pure blue and red; rather, it was between purple and red. The peak at pH 10 shifts gradually to 580 ( 1 nm
from 528 nm as the repetition number of the pH change increased, while the absorption band at pH 4 appears around 580 nm. The bare gold nanoparticles (15, 20, 30, and 40 nm) indicated no visual colorimetric change between pH 4 and pH 10, by maintaining the bluish color created at n ) 2 (i.e., the first insertion of the acid for pH 4) as shown in Figure 1. When the reversible process was examined for Aβ1-40-coated gold colloid with the sizes of 15, 20, 30, and 40 nm at the temperature ranges between 5 and 50 °C, all Aβ1-40-coated gold colloids exhibited an undulating feature in λpeak(n, T) as the pH changed between pH 4 and pH 10. However, in most cases the amount of shift at the basic condition is not significant enough to provide reddish color preventing visual observation of repetitive color change between reddish and bluish color. For example, in Figure 2 the reversible process was visually observed for Aβ1-40-coated 20 nm gold at all shown temperatures except for the case when temperature is at 5 °C. For Aβ1-40-coated 30 nm the amplitudes of λpeak(n) were significantly large enough to provide the reversible color change at 10 °C (Figure 2c). This was true at 5 °C for Aβ1-40-coated 40 nm gold colloids (Figure 2d). To clarify the transition point where the reversible change took place, the pH-induced peak shift was investigated with a step of 1 °C as shown in Figure 3, panels a-c. (In order to avoid congestions, some representative peaks are selected here.) For 20 nm, the peak shift was investigated from 10 to 5 °C for every 1 °C, since 5 °C is the lowest temperature available in this experiment. The amplitude of peak shift was significantly reduced at 5 °C when reversible color change became visually unidentifiable, as shown in Figure 3a. For 30 nm gold coated with Aβ1-40, the peak shift was investigated between 20 and 10 °C in 1 °C increments in order to find the temperature where the reversible color change began. A significant increase in amplitude was found at 18.0 ( 0.2 °C and lower, and reversible color change was visually identified. (Figure 3b) As for 40 nm gold covered with Aβ1-40, we investigated the peak shift by scanning the temperature between 10 and 5 °C by 1 °C step. (Figure 3c) We found a sign of the reversible color change when temperature was lowered to 6.0 ( 0.2 °C. In our study, the undulation of wave peak of each acid or base addition operation, λpeak(n, T), is represented by the following analytical formula:
1524
J. Phys. Chem. A, Vol. 114, No. 3, 2010 d
λpeak(n, T) ) ATd + BTd(n - 1)C T + DTd exp(-(n - 1)ETd)cos(nπ) (2) The parameters ATd, BTd, CTd, DTd, and ETd were extracted for each temperature, T (°C), of gold colloids with diameter of d nm and plotted in Figure 4. Since the meaning and implication of each parameter was already explained in detail,18 a brief description of the parameters is given here. Equation 2 was derived to reproduce the spectral shift and repetition change. In this equation, an initial peak position at neutral pH (i.e., λpeak(1, T)) is given by ATd - DTd, and the parameters BTd and CTd show the average wave peak position shift as pH varies between pH 4 and pH 10. The parameters DTd and ETd imply amplitude and damping factor for the repetitive event, and the cosine function was used to indicate increase and decrease of λpeak(n, T) upon change between pH 4 and pH 10. The values calculated by eq 2 are effective only for each n value, not the values between each n. Thus, the dotted line shown in Figures 2 and 3 is given only for the purpose of clarifying the repetitive trend. Among listed parameters, the parameter DTd is critical; it indicates whether the reversible process can be recognized visually or not. The half amount of peak shift between λpeak(nodd, T) and λpeak(neven, T) is represented by parameter DTd. Our observation concludes that the visual confirmation of the reversible process was possible when parameter DTd was over 5.5 nm, which was indicated in Figure 4d with dashed lines in order to clarify that the reversible change was observed at the temperatures when DTd resides above this threshold. The major feature of temperature dependence of each parameter is described for each size of gold colloid (15, 20, 30, and 40 nm) conjugated with Aβ1-40. 15 nm Gold Colloids Coated with Aβ1-40. The reversible color change was not colorimetrically observed at all temperatures for 15 nm gold colloid conjugated with Aβ1-40. The peak shift, λpeak(n, T), shows the undulating feature with a small amplitude as a function of pH as shown in Figure 2a, that is, DT15 < 5.5 nm (Figure 4d-i), where DT15 indicates a critical threshold required to support a visual observation of the reversible color change. As the overall position of the peak shift shows in the figure, the color of the solution was bluish (or dark purple) regardless of pH except for that at the initial condition (n ) 1).
Figure 3. The demonstration of color reversibility of (a) Aβ1-40-coated 20 nm gold particles between 5 and 8 °C, (b) Aβ1-40-coated 30 nm gold particles between 17 and 20 °C, and (c) Aβ1-40-coated 40 nm gold particles between 5 and 8 °C.
Yokoyama et al. All parameters of eq 2 extracted for Aβ1-40-coated 15 nm gold colloids exhibited a monotonic, or nondiscontinuous, trend as shown in Figure 4a-i, b-i, c-i, d-i, and e-i. The parameters 15 15 A15 T , DT , and ET showed almost constant values under all tested 15 temperature ranges, whereas the parameters B15 T and CT showed slight monotonic increases or decreases as the temperature increased. The trend observed for Aβ1-40-coated 15 nm gold colloids was used as the standard feature for identifying the case where no visual confirmation of the reversibility was made. 20 nm Gold Colloids Coated with Aβ1-40. The reversible color change was confirmed for the entire temperature range visually except for that at 5 °C, and it was confirmed by the fact that D520 did not go over the threshold DTd ) 5.5 nm. The DdT values had a local maximum at 10 °C and an abrupt increase at 50 °C among the tested temperature range. The high value of DT20 at the high temperature is an especially unique feature found only for 20 nm gold colloid conjugated with Aβ1-40. The trend seen in parameter C20 T somewhat correlates with parameters AT20 andDT20. As seen in Figure 4e-ii, all ET20 values stayed above zero, indicating that the amplitude of the undulating feature damped down as n increased. 30 nm Gold Colloids Coated with Aβ1-40. When Aβ1-40 was coated over 30 nm gold colloid, the reversible color change was visually observed at 18 °C ( 0.02 and lower. Although the feature at 18 °C possessed a high damping mode, a clear reversible change with relatively high amplitude and lower damping mode was observed from 20 to 17 °C as shown in Figure 3b. The induction of the visually observable reversible change was clearly confirmed in the plot of parameter DT30 (Figure 4d-iii) as a stepwise increase at 18 °C. Since all other parameters indicated the noncontinuous spot at 18 °C as shown in Figures 4a-iii, b-iii, c-iii, and e-iii, 18 °C is a critical temperature for Aβ1-40-coated 30 nm gold colloid to cause a significant change in the structure of Aβ1-40 constructed over 30 the surface. The parameters A30 T and CT showed almost the same trend as that seen in parameter DT30, where a stepwise increase began at 18 °C. Especially, the relatively higher values of CT30 at T < 20 °C reflect the high nonlinear or non-monotonic increase of the average peak shift as a function of n. This nonlinearity is considered to be highly contributed by the feature seen at smaller n, where the peak shift shows very small changes between neven and nodd. Quite interestingly, the damping factor represented in parameter ET30 showed a cusp at both 18 and 19 °C, whereas the overall trend indicated a smooth decrease as temperature decreased. The high damping factor at these temperatures can be easily confirmed from the features seen in the peak shift plot shown in Figure 3b, where amplitude of λpeak(n, T ) 18 and 19 °C) decreased drastically as n increased. Since the damping factor E530 is negative, the amplitude was slightly larger at 5 °C. 40 nm Gold Colloids Coated with Aβ1-40. The reversible color change was visually observed at 6 °C and lower where a clear stepwise increase in parameter D640 was seen as shown in Figure 4d-iv. However, λpeak(n ) 1-3, T ) 5 and 6 °C) exhibited a peculiar feature; an initial peak shift at n ) 2 and 3 did not indicate a clear trend of the reversible change as shown in Figure 3c. This was reflected in C40 T at T ) 5 and 6 °C (Figure 4c-iv) as non-monotonic change in peak shift between neven and nodd. The parameters BT40, DT40, and ET40indicated the noncontinuous spot around 6 °C as shown in Figures 4b-iv, d-iv, and e-iv. The parameters C40 T highly correlated with parameter DT40, where a stepwise change seen at 6 °C. The negative ET 40 e 8 indicates that the amplitude was slightly larger at these temperatures as n increased.
Temperature Dependence of Nanoscale Conjugation
J. Phys. Chem. A, Vol. 114, No. 3, 2010 1525
Figure 4. (a) Plotting of the parameter ATd in eq 2 as a function of gold colloidal sizes for Aβ1-40-conjugated (i) 15 nm, (ii) 20 nm, (iii) 30 nm, and (iv) 40 nm gold colloid particles. (b) Plotting of the parameter BTd in eq 2 as a function of gold colloidal sizes for Aβ1-40-conjugated (i) 15 nm, (ii) 20 nm, (iii) 30 nm, and (iv) 40 nm gold colloid particles. (c) Plotting of the parameter CTd in eq 2 as a function of gold colloidal sizes for Aβ1-40-conjugated (i) 15 nm, (ii) 20 nm, (iii) 30 nm, and (iv) 40 nm gold colloid particles. (d) Plot of the parameter DTd in eq 2 as a function of gold colloidal sizes for Aβ1-40-conjugated (i) 15 nm, (ii) 20 nm, (iii) 30 nm, and (iv) 40 nm gold colloid particles. The horizontal dotted line at 5.5 nm indicates the threshold and is the minimum amount required to support visual observation of the reversible color change. (e) Plot of the parameter ETd in eq 2 as a function of gold colloidal sizes for Aβ1-40-conjugated (i) 15 nm, (ii) 20 nm, (iii) 30 nm, and (iv) 40 nm gold colloid particles.
Discussion Prior to this study, only the 20 nm gold colloid conjugated with Aβ1-40 was considered to exhibit the reversible color change. While the colloidal size dependence was suspected to be due to chemicals, such as a stabilizer, used to prepare each gold colloidal size, we confirmed with Ted Pella Inc. that the stabilizer used for all gold colloids were the same. All sizes of gold colloids were formed by a Frens-derived citrate reduction method possessing traces of citrate
