Single-molecule calorimeter and free energy landscape

Significance

Measuring the complete thermodynamic profile of a single molecule reveals unparalleled advantages in the biological realm. We develop a single-molecule calorimeter with plasmonic imaging and optical tweezer by probing the thermal fluctuations of single nanoparticles linked to the substrate via the single-molecule binding pair. This uncovers the complete thermodynamic information for single binding events and identifies the equilibrium state and energetic components of the free energy landscape, suggesting possibilities for offering high-throughout analysis of single-molecule interactions.

Abstract

The precise measurement of thermodynamic and kinetic properties for biomolecules provides the detailed information for a multitude of applications in biochemistry, biosensing, and health care. However, sensitivity in characterizing the thermodynamic binding affinity down to a single molecule, such as the Gibbs free energy (${\mathit{G}}_{\mathit{b}}$), enthalpy (${\mathit{H}}_{\mathit{b}}$), and entropy (${\mathit{S}}_{\mathit{b}}$), has not materialized. Here, we develop a nanoparticle-based technique to probe the energetic contributions of single-molecule binding events, which introduces a focused laser of optical tweezer to an optical path of plasmonic imaging to accumulate and monitor the transient local heating. This single-molecule calorimeter uncovers the complex nature of molecular interactions and binding characterizations, which can be employed to identify the thermodynamic equilibrium state and determine the energetic components and complete thermodynamic profile of the free energy landscape. This sensing platform promises a breakthrough in measuring thermal effect at the single-molecule level and provides a thorough description of biomolecular specific interactions.

Single-biomolecule optical detection has become indispensable in molecular biology and life science over the past decades and provides insight into the individual distinctive binding affinities, which may be hidden in ensemble-averaged measurements for bulk materials (1–5). The associated interactive and thermodynamic heterogeneity is the key to molecular function and poses a fundamental challenge to fully capturing the diversity of binding structures and mechanisms (6–9). To date, isothermal titration calorimetry is a well-defined technique to characterize the complete thermodynamic profiles in a biological system (10–12). It provides possibilities for studying the interplay of a wide range of biological systems including protein–protein interactions and antibody–antigen and enzymatic kinetic analysis. However, there are very few studies of heating effect and driving forces involved in a single-molecule binding event.

Recently, some force-driven technologies have attracted a large amount of attention to study thermodynamics and biomolecular interactions by measuring the mechanical responses at the single-molecule level (13), such as atomic force microscopy (14, 15) and optical (16–18) and magnetic tweezers (19, 20). The thermodynamic and interrogated structural properties including free energy profile can be extracted from molecular pulling experiments, which break the equilibrium state of interaction (16, 19). In addition to the above methods, fluorescence labels also paint a picture of intermolecular interactions by lighting up the specific biomolecules. These provide detailed information about dynamics, interactions, and structures with single-molecule resolution (21–24). Recently, we developed an optical platform to uncover the fluctuation of a single nanoparticle to probe the vertical positions and binding kinetics of molecular interactions with a plasmonic imaging method of subnanometer accuracy along z direction (25).

The thermal force of a biomolecular system alters the shaken motion of molecules, which can be optically probed by a linked nanoparticle (26–28). Tethered nanoparticle motion describes the conformational changes of linked soft molecule and dynamic interactions, which is achieved by tracking the motion behavior of a nanoparticle probe tethered to a substrate with a single biomolecule. It has been utilized to investigate biomolecular mechanisms and driving forces at the single-molecule level, and explore the equilibrium statistical analysis in the absence of external forces and specific labels. However, it is only sensitive to the transient variations along the x–y plane, which neglects the motions in the vertical direction.

Herein, we experimentally demonstrated the temperature dependence in single biomolecule system by utilizing a submicrometer laser spot to generate tiny temperature variation of a single binding pair linked between a nanoparticle and substrate. Through taking the rabbit IgG/goat anti-rabbit IgG binding pair as a model, the anti-IgG–coated gold nanoparticles (AuNPs) are trapped by the IgG immobilized gold substrate. By imaging the scattering of plasmonic wave propagating on the surface of AuNPs, the precise tracking of transient thermal fluctuations along three directions is allowed, and the complete thermodynamic profile (e.g., Δ$G_b$, Δ$H_b$, and Δ$S_b$) for single-molecule interplay can be determined. We also apply this single-molecule calorimetry platform to further study other binding pairs with different binding affinities to demonstrate its universality in biosystems. This single-molecule calorimeter provides a molecular-level understanding of the detailed dynamic information and binding characterizations, and elucidates thermal effect in kinetic equilibrium.

Results and Discussion

The optical apparatus is built to study the tiny thermal fluctuations of individual nanoparticles linked to substrate via single molecular binding pair by combining surface plasmon resonance microscopy (SPRM) with optical tweezer system (29, 30). Fig. 1A shows the schematic illustration of experimental setup, and more detailed description is provided in Materials and Methods. Briefly, a gold film is placed on an invert microscope equipped with 60× oil-immersed objective (numerical aperture [N.A.] = 1.49), and a polydimethylsiloxane cell is attached onto the gold film to hold sample solution (25, 31–35). A low-power red beam (wavelength = 680 nm) from a superluminescent emitting diode (SLED) is directed onto the gold film with an appropriate angle via the objective to excite surface plasmons at the surface of gold film. The scattering light is collected with the same objective and the plasmonic images are recorded by a charge-coupled device (CCD) camera with 100 frames per second dynamically. An optical tweezers system is introduced in the SPRM to focus a near-infrared beam (λ = 1,064 nm) on the gold slide, which influences the temperature of a single nanoparticle on the substrate surface by local heating. To obtain the thermodynamic profile of a single molecule, the rabbit IgG/goat anti-rabbit IgG (IgG/anti-IgG) binding pair is selected as a model system, as shown in Fig. 1B. The individual AuNPs are functionalized with the anti-IgG molecules, and captured by IgG molecules-immobilized gold substrate via single-molecule binding pair separately. The individual nanoparticles disturb the surface plasmon polariton and result in a parabolic pattern for each nanoparticle in the plasmonic image. The intensity of plasmonic pattern for single nanoparticle ($I$) decays exponentially with the distance ($z$) between nanoparticle and substrate according to the following:

$$I=I_0{e}^{-z/L},$$ [1]

where $I_0$ is the plasmonic intensity for single nanoparticle at $z=0$ and $L$ is the decay constant (36–40). By analyzing the time-lapsed plasmonic images, the thermal fluctuation along the z direction of a single nanoparticle is determined from the intensity.

Fig. 1.

Experimental setup of single-molecule calorimeter. (A) Schematic illustration of wide-field optical imaging and local heating configurations. A near-infrared laser (wavelength = 1,064 nm) is focused on a gold film to induce a localized heating spot within a few microns for triggering different thermodynamics of single-molecule binding events. To remove the optical interference of the heating laser beam, a short-pass filter (wavelength < 720 nm) is introduced. The wide-field plasmonic imaging utilizes p-polarized SLED light source to excite surface plasmons at the surface of gold film, typically deposited on a cover glass, and the reflected light is collected by a CCD camera. The scattering intensity is sensitive to the distance between nanoparticles and surface and the dielectric constant at the interfaces, providing precise tracking of the thermal fluctuations and temperature variations in the binding pair. (B) Illustration of single-molecule binding measurement by plasmonic imaging of nanoparticle fluctuations. Inset shows that the binding of goat anti-rabbit IgG-coated single AuNPs to rabbit IgG immobilized on a gold surface. The BSA molecules blocks the redundant active sites.

To validate the ability of optical technique for mapping the transient thermal variations, the plasmonic intensity changes of single AuNP on the gold surface tuned by hot water and focused laser for bulk and local heating are deliberated, respectively. The local temperature is determined from the SPR intensity with high sensitivity, which is based on the quantitative relationship between local refractive index and temperature (for more calculation details, see SI Appendix, S1). Adding hot water into the sample chamber and recording the cooling process, synchronous variation of plasmonic intensity and bulk temperature are measured by SPRM and thermocouple (SI Appendix, S2). The plasmonic intensity increases linearly with temperature gradually declines, where the temperature decrease of 10 K would induce an increase in 359 intensity units (IU) as shown in Fig. 2 A and B. From the calibration of temperature and plasmonic intensity, the conversion factor is determined to be −35.9 IU/K, allowing us to extract local temperature from the optical images. The local thermal distribution on the bare gold film generated by the focused heating laser is imaged with the above calibration of conversion factor in SI Appendix, Fig. S2. The heating laser focuses on the gold film to a small spot (diameter = 0.87 µm) and produces the thermal distribution surrounding the heating spot (SI Appendix, Fig. S2A). The thermal diffusion increases the local temperature of gold film, which is manifested by a decrease in refractive index of medium surrounding the heating spot. In our experiment, the location of heating laser is fixed, and the strength is set to increase periodically with altering the power (SI Appendix, Fig. S2 B and C). We introduce an absorptive filter to selectively remove the light of heating laser for investigating pure thermal perturbation. The temperature at the central position is elevated with increasing the heating strength, and decreases rapidly from the center to periphery (SI Appendix, Fig. S2D). The detection limit of this optical technique is 0.5 K estimated by three times of the noise level at “laser-off” state, which is originated from mechanical and thermal drift of the system.

Fig. 2.

Working principle and calibration of plasmonic intensity and local temperature. (A) The calibration curve of plasmonic intensity (ΔI) and local temperature (ΔT) showing linear dependence on bare gold substrate. The plasmonic intensity increases with the bulk temperature decrease of hot water in the sample chamber. (B) Plasmonic images of bare gold substrate at three temperature variations showing different plasmonic intensities. (Scale bar, 5 µm.) (C) The snapshots of local temperature distribution derived from plasmonic images for single AuNP (diameter, 150 nm) on gold substrate. The region of single AuNP is heated by the focused laser spot. (Scale bar, 5 µm.) (D) The time-lapsed plasmonic intensity change along the line profile in C (yellow dashed line, and the arrow reveals the direction for position) for different heating powers. The center for parabolic pattern shows the maximum value for intensity change owing to the highest temperature. (E) The line profiles of thermal distribution for single AuNP (yellow dashed line shown in C) with different powers of heating laser.

Next, we image the temperature distribution of a single nanoparticle tethered to the gold surface via nonspecific adsorption. The temperature distribution shows elongation compared to the results on the gold surface without nanoparticle; it is caused by the propagation of surface plasmon wave along the elongation direction (41). The nanoparticle shows a parabolic pattern due to the scattering of the plasmonic wave by the nanoparticle, and the time-lapsed plasmonic snapshots of it and corresponding temperature distribution along the line profile across the characteristic tail of nanoparticle marked are shown in Fig. 2 C and D. The intensity of nanoparticle region is switched with the laser on or off, and the temperature around nanoparticle rises with increasing the strength of heating laser. By plotting the thermal distribution of a single nanoparticle, the local temperature of central region shows the maximum value and decreases rapidly from the center to periphery in Fig. 2E.

In the light of the above demonstration, we immobilize rabbit IgG molecules to the gold film modified with 11-mercaptoundecanoic acid (11-MUA), and the goat anti-rabbit IgG-coated AuNPs with the diameter of 150 nm are introduced to bind with the IgG molecule on the gold film. To map the thermal fluctuations of single IgG/anti-IgG binding pair, we regulate the local temperature around single AuNP by tuning the strength of heating laser. Based on the quantificational relationship between plasmonic intensity ($I$) and z displacement ($z$), the thermal fluctuations of IgG/anti-IgG binding pair along z direction at different temperatures are extracted from the plasmonic images (Fig. 3A and Movie S1). The temperature is calculated from the plasmonic intensity according to the conversion factor mentioned in the above section. From the thermal fluctuations along the z direction at different temperatures, the probability density, $P\left(z\right)$, vs. $z$ for IgG/anti-IgG binding pair is acquired conforming to the Gaussian distribution for single binding event in Fig. 3A (just one IgG molecule binds with one anti-IgG molecule). The peak of $P\left(z\right)$ distribution is moving to a higher value with heating up, which indicates that the fluctuation of single binding pair is activated by the temperature. We calculate the free energy profile of single binding pair based on the relation between free energy $G\left(z\right)$ and probability distribution $P\left(z\right)$ by the following:

$$G\left(z\right)=-k_{\mathrm{B}}T\log \left[AP\left(z\right)\right],$$ [2]

where $A$ is a constant calculated by normalization of $P\left(z\right)$, $k_{\mathrm{B}}$ is the Boltzmann constant, and $T$ is temperature (for more calculation details, see SI Appendix, S4) (42, 43). The free energy profile for single-molecule binding event shows a parabola patterns, and shifts with the temperature revealing tiny change of the system energy and thermodynamic equilibrium states (Fig. 3B). The dynamic tracking of single nanoparticle along x and y directions from the plasmonic images has been performed to comprehensively uncover the thermodynamic details accompanied by the temperature in Fig. 3C, which is uncorrelated to the z displacement (SI Appendix, Fig. S3). At the first four temperature points, the two-dimensional (2D) mappings display circular symmetric patterns as the disk-like shape indicating the single-molecule binding event and the diameter of the pattern increases with gradual heating up, which further verified that the temperature activated the thermodynamics of the single-molecule binding pair mentioned above. Finally, it converts to two clusters faintly at 304.0 K and further demonstrates that the motion range of linked nanoparticle via IgG/anti-IgG binding pair enhances with increasingly raised temperature in this experimental system.

Fig. 3.

The thermodynamics and free energy profiles of single binding pair at different temperatures. (A) The z displacements and corresponding probability distributions for an anti-IgG–conjugated nanoparticle bound to IgG molecule at different temperatures. The violet curves represent Gaussian fitting results. (B) Free energy profiles of single IgG/anti-IgG binding pair determined from the probability distributions at different temperatures, where the red lines represent a guide to the eye. (C) Corresponding 2D histogram of displacements along x and y directions at different temperatures. (D) The z displacements of single AuNPs over the entire heating process, where the blue trace shows the displacements of nanoparticle stuck on the surface via nonspecific adsorption, and the black trace shows the displacements of single IgG/anti-IgG–linked AuNP. The zoom of the displacements in the purple dashed box to show a sudden jump of displacements by turning on laser. (E) The linear relation between local temperature and free energy of single IgG and anti-IgG binding pair with the slope $\mathrm{\Delta }H$ of 5.14 × 10⁻²² J/K, and intercept $\mathrm{\Delta }S$ of −5.81 × 10⁻²⁰ J.

To further validate that the thermal fluctuations are dominated by single-molecule binding pair rather than the nanoparticle, we compare the z displacements of single AuNP linked on the surface via IgG/anti-IgG binding pair with fluctuation and single AuNP stuck on surface without any fluctuation in the entire heating process, separately. As shown in Fig. 3D, the z displacements of the IgG/anti-IgG–linked AuNP are sensitive and reversible to the temperature by repeatedly turning the laser on and off, and show larger amplitude immediately when the laser is on and return to the previous level when the laser is off. For the nanoparticle stuck on the surface due to the nonspecific adsorption, the variation of z displacement maintains ∼0.3 nm in the entire heating process, which is independent with temperature revealing the noise level arising from the optics and electronics of the experiment setup. Additionally, considering the force-driven information of single nanoparticle in the system, the spring force from the molecular binding is large compared with the damping force of AuNPs (more details are in SI Appendix, S17). Therefore, the thermal fluctuations are manipulated by molecular binding pair mainly and influenced by the local temperature. In our system, increasing temperature activated the thermodynamics of single-molecule binding pair, which is probed by the large z displacement. We also verify this relation at much higher temperatures, and the observed nanoparticle is still linked to the surface within the range of 7 K (SI Appendix, Fig. S4). From the definition of the Gibbs free energy ($\mathrm{\Delta }G$) by two components, enthalpy ($\mathrm{\Delta }H$) and entropy ($\mathrm{\Delta }S$), the energetic information can be determined by the standard thermodynamic expression:

$$\mathrm{\Delta }G\left(z\right)=\mathrm{\Delta }H-T\mathrm{\Delta }S.$$ [3]

By fitting the free energy profiles of the nanoparticle near the energy minimum at different temperatures, we obtain the linear correlation between the free energy and temperature with the slope $\mathrm{\Delta }H$ of 5.14 × 10⁻²² J/K, and intercept $\mathrm{\Delta }S$ of −5.81 × 10⁻²⁰ J for single IgG/anti-IgG binding event (Fig. 3E). Almost all types of interaction that occur at molecular interfaces are accompanied by the specific set of energetic components making them distinguishable from other types of phenomena referred as “thermodynamic signatures.” It is the fundamental information in interpreting the results obtained from single molecule calorimeter, and speculating about the bound strength and driving forces of single-molecule interactions (for more details, see SI Appendix, S17 and S18).

The single-molecule binding event of IgG/anti-IgG shown in Fig. 3 is about 77% of all conditions, and the remaining 23% of the nanoparticles displays the multipeak fluctuation behaviors (SI Appendix, Table S1). The dynamic z displacements reveal a stochastic transition of single nanoparticle between two distinct equilibrium states indicating that one anti-IgG–coated AuNP can switch between two individual IgG molecules on the substrate (Fig. 4A). The plasmonic snapshots of different switching states in Fig. 4B show distinct features of the scattering tails, which is attributed to the amplitude of fluctuation. The probability distribution reveals two peaks with similar proportion, and the free energy profile probes two minima of equilibrium states (Fig. 4C). The 2D map among x–y plane brings out two separate clusters owing to the switching of single nanoparticle back and forth between two IgG molecules telling the same story as the above observations (Fig. 4D). We further measure the thermal fluctuations of this double-molecule binding event at different temperatures, and the z displacements maintain two states in SI Appendix, Fig. S5. The two-peak distribution of thermal fluctuations is moving to larger amplitude with heating up and the free energy profiles show two minima, corresponding to the two equilibrium states. However, it is surprising that a decrease of probability ratio of state 1 to state 2 for counting with increasing temperature indicates that the temperature activates the transient equilibrium of molecule binding dynamics and promotes the transition from state 1 to state 2.

Fig. 4.

Thermodynamics of double- and triple-molecule binding events at 300 K. (A) The z displacements for switching events of single anti-IgG–coated AuNP between two IgG molecules. The right curve indicates zoomed view of fluctuation trace showing two distinct transition states, and the bottom curve shows the corresponding stepwise switching process between two distinct equilibrium states. (B) The plasmonic images of two distinct transition states. (Scale bar, 5 µm.) (C) Free energy profile of IgG/anti-IgG binding pair in double-molecule binding events, where the red dashed line is a guide to the eye. (D) Lateral positions (x and y directions) reveal two clusters corresponded to the two distinct states. (E) The z displacements of single AuNP in triple-molecule binding events. The right curve indicates zoomed view of fluctuation trace showing three distinct transition states, and the bottom curve shows the corresponding stepwise switching process among three states. (F) The plasmonic images of three distinct transition states; state 1 is chosen as the reference. (Scale bar, 5 µm.) (G) Free energy profile of triple-molecule binding events. (H) Lateral positions (x and y directions) reveal three clusters corresponding to the three distinct states.

To disclose the thermodynamics of the multimolecule binding events thoroughly, we further study the triple-molecule binding event showing the three equilibrium states. The state 1 of individual nanoparticle is defined as the initial state at 300 K as shown in the stepwise transition process clearly. From the probability distribution of this event, it shows a state with large proportion (state 1) and two states with small proportions (states 2 and 3), indicating that state 1 is the equilibrium state among three states (Fig. 4E). The corresponding plasmonic images represent the three distinct states during the dynamic transition, and the corresponding free energy profile of the binding event reveals the three minima associated with three states (Fig. 4 F and G). In the 2D map along the x–y plane, there are three distinct clusters attributed to the transition between different IgG binding sites, denoted as the three states in Fig. 4H. We study the transition process of triple-molecule binding event at the different temperatures, and the intensity traces show the three states in Fig. 5A. The free energy profiles and the probability ratio of three states at different temperatures are plotted, and state 1 maintains the main equilibrium state as shown in Fig. 5 B and C. The proportion of state 1 increases with heating up and that of states 2 and 3 are decreasing attributed to the acceleration effect of high temperature in switching events (Fig. 5D). This method provides a direct version to determine the thermodynamic equilibrium state of single molecule at different temperatures, and uncovers the “fingerprint” details of transition process and each equilibrium state.

Fig. 5.

The thermodynamics and free energy profiles of IgG/anti-IgG binding pair in triple-molecule binding events at different temperatures. (A) Time traces of measured z displacements for single nanoparticle linked to the surface via IgG/anti-IgG binding pair at different temperatures. (B) Free energy profiles of IgG/anti-IgG in triple-molecule binding events, where the red dashed line is a guide to the eye. (C) The relative ratio of three transition states at different temperatures, and state 1 maintains the main equilibrium state. (D) The statistical analysis for switching events among different states at different temperatures.

To demonstrate that the above method can be generalized to different biomolecule binding systems, we investigate the IgG/protein G and biotin/streptavidin binding pairs. For IgG/protein G binding pair, the probability distribution of thermal fluctuations displays single (77%) and double peak (23%) (SI Appendix, Table S1), and the temperature dependence of the system is similar with IgG/anti-IgG binding pair discussed above. We probe the free energy profiles of IgG/protein G binding pair at different temperatures showing the linear relationship between the free energy and temperature of system, and extract the energetic components for IgG/protein G binding pair (SI Appendix, Figs. S7–S10), indicating an order of magnitude smaller than that of IgG/anti-IgG. This result is consistent with the different extents of binding affinity for two systems. From the multipeak thermal fluctuations, it shows three states precisely when the measured temperature is lower than 300.7 K, and it converts to two states with heating up due to the equilibrium transition. We further performed the same experiment for streptavidin-conjugated AuNPs to the biotin functionalized gold film with strong interaction, and only single-peak distribution of thermal fluctuations is observed and shows similar temperature dependence (SI Appendix, Figs. S11 and S12).

According to the above conclusions, we predict that the nanoparticles would detach from the surface accompanied with the fracture of binding site when the temperature is high enough. To validate, we build the platform for bulk heating all nanoparticles linked to the surface via single binding pair combined with SPRM and heating plate (SI Appendix, Fig. S13A). In the time-lapsed plasmonic images, some of the AuNPs detached from the surface during the heating process, and the corresponding z displacements further demonstrate the departure (SI Appendix, Fig. S13 B and C). The nanoparticle number linked to the surface via single binding pair is counted at different temperature points. For the binding pairs of IgG/anti-IgG and biotin/streptavidin, the numbers of nanoparticles remained on the surface decrease during the heating process (Fig. 6A). It is related to the thermal stability of binding interactions and the binding affinity of different systems. Based on the relationship between ln P and −1/k_BT as described in Eq. 2, the slopes reveal the free energy of different single-molecule binding systems (Fig. 6B). The corresponding free energy values are shown in the Fig. 6C, which is (8.55 ± 1.1) × 10⁻²⁰ J for the IgG/anti-IgG binding pair and (6.16 ± 0.32) × 10⁻²⁰ J for the biotin/streptavidin binding pair.

Fig. 6.

The fracture of single-molecule binding pair linked between individual nanoparticle and gold substrate by continuously heating up. (A) The statistical analysis for nanoparticles remained on the surface via single IgG/anti-IgG and biotin/streptavidin binding pair during heating process. (B) The linear relations between −1/(k_BT) and ln P for IgG/anti-IgG and biotin/streptavidin binding pair, where k_B is the Boltzmann constant, T is temperature, and P is the ratio of final count for nanoparticles stayed on the substrate by heating up to the initial count at 305 K. (C) The calculated free energy profiles for IgG/anti-IgG and biotin/streptavidin binding pairs of the detachment process.

Conclusion

In summary, we demonstrated the plasmonic imaging and tracking of single gold nanoparticle linked to substrate via single-molecule binding pair with monitoring the transient local heating effect by a focused laser. The rabbit IgG/anti-goat rabbit IgG binding pair was analyzed as a model of single-molecule calorimeter to uncover the thermodynamics of single, double, and triple binding events and determine the equilibrium state and energetic components of free energy landscape. The intermediate states of different binding events have been real-time tracked during local heating processes and reveal the temperature dependence of single-nanoparticle motion linked with biomolecular binding pair. To further prove the feasibility of this method, we studied different specific binding pairs and probed the free energy profiles at different temperatures to get valuable fingerprint information with single-molecule resolution. This strategy offers a molecular-level understanding of thermodynamic equilibrium state and energetic components, and provides a complete thermodynamic profile, such as the Gibbs free energy change ($\mathrm{\Delta }{G}_b$), enthalpy change ($\mathrm{\Delta }{H}_b$), and entropy change ($\mathrm{\Delta }{S}_b$). It also shows fingerprint details of the driving forces and bound strength down to the single-molecule level. Combined with other characterization techniques, this single-molecule calorimeter can be applied to reveal the conformational changes of the molecules upon interactions and biophysical kinetics for a wide range of soft biomaterials.

Materials and Methods

Chemicals.

Gold nanoparticle (diameter, 150 nm) coated with goat anti-rabbit IgG, protein G, and streptavidin molecules were purchased from Nanopartz and diluted in deionized water (18 MΩ⋅cm; Milli-Q; Millipore Corporation). Phosphate buffer saline buffer (1× PBS, 154 mM NaCl, 5.6 mM Na₂HPO₄, and 1.1 mM KH₂PO₄) was obtained from Corning. 11-MUA was purchased from Adamas Reagent Company. N-Ethyl-N′-(3-(dimethylamino)-propyl) carbodiimide hydrochloride (EDC), N-hydroxysuccinimide (NHS), Tween 20, rabbit IgG, and bovine serum albumin (BSA) samples were purchased from Sigma-Aldrich. d-Biotin hydrazide was purchased from Tokyo Chemical Industry Company.

Surface Preparation.

The gold films were prepared by coating BK-7 glass slides with a thickness of 2-nm Cr and 47-nm Au using the magnetron sputtering technique and treated with hydrogen flame to remove contaminations. After that, the clean gold film was immersed in the 1 mM 11-MUA/ethanol solution overnight. The modified gold film was rinsed with ethanol and dried under nitrogen gas. The Teflon solution cell was attached onto the 11-MUA modified gold substrate to hold sample solution. A volume of 200 µL of EDC/NHS mixture solution (M:M = 4:1) was added to activate the 11-MUA surface for 20 min. The rabbit IgG or d-biotin hydrazide solution was then introduced to the cell and incubated for 10 min to bind with the active sites of 11-MUA. After washing with PBST (PBS solution with 0.05% Tween) for three times, 200 µL of 3% (wt/vol) BSA in PBS solution was used to block the residual activated sites. Then the gold surface was rinsed with PBST three times to remove the unbound BSA molecules. A volume of 200 µL of PBST was added into the cell to further avoid nonspecific binding.

The Imaging Setup Combined with Optical Tweezer System.

The entire system was built on an inverted microscope (Nikon; Ti-U) with two decks of optical paths. The top deck was used for optical tweezer and the bottom deck was for plasmonic imaging. The system was equipped with a 60× microscope objective (N.A. = 1.49) to collimate, focus, and collect the light. A 680-nm superluminescent light-emitting diode (Q-photonics) was introduced as the light source for plasmonic imaging. The polarizer was placed in the microscope filter cube of the bottom deck, which was used to generate the p-polarized light and excite the surface plasmonic wave at the surface of gold film. The measurements were performed at a fixed incident angle, and the plasmonic intensity changed with the refractive index variation at the surface of gold film at different temperatures.

Here, the optical tweezers system (Aresis; Tweez 250si) was introduced to generate local heating of single nanoparticle by focusing the continuous-wave near-infrared laser beam (Nd:YAG; wavelength, 1,064 nm) onto the gold film through the same microscope objective. The size of heating spot is 0.87 μm. The absorptive filter (OD, 1.3) was added in the microscope filter cube on the top deck to decrease the intensity of heating spot. To remove the optical interference of near-infrared laser beam, a short-pass filter (wavelength < 720 nm) was placed in the same cube. A CCD camera (Pike F-032; Allied Vision Technology) was used to obtain the plasmonic images with the frame rate of 99.7 fps.

Data Analysis.

The data processing was performed with ImageJ and Matlab. The plasmonic images of single gold nanoparticle were obtained by subtracting the plasmonic image without heating up from those with local heating. By recording the time-lapsed plasmonic images, the thermal fluctuations in z direction were determined from the intensity.

We also developed an algorithm for tracking x and y positions of single nanoparticles from the plasmonic images. It allows us to identify the parabolic pattern of nanoparticles and then determine the x and y positions in two steps. First, the first image is chosen as the reference and 2D spatial correlation between the reference and the subsequent image sequence is calculated. The rough positions of nanoparticles are mapped based on the highest value in the 2D correlation. Second, the intensity profiles along the x and y directions of single nanoparticles are calculated from the rough positions and fitted with the polynomial function, which is used to determine the precise positions of individual nanoparticles.

Gold Nanoparticle Characterization.

Gold nanoparticles conjugated with goat anti-rabbit IgG, protein G, and streptavidin molecules deposited on gold film were imaged by scanning electron microscope (JSM-7800F) (shown in SI Appendix, Fig. S14, respectively).

Data Availability

All study data are included in the article and/or supporting information. SI Appendix and Movie S1 provide experimental data and parameter estimates.