1. Introduction

Single protein studies typically bind a protein to a surface and then look at protein-protein interactions through additional binding events [1–5]. Impressively, recent studies have shown single protein binding sensitivity using plasmonic nanoparticle by monitoring resonance shifts [6] and by photo-thermal transduction to achieve increased sensitivity [7]. Those works, however, have the disadvantage of using one of the binding sites of the protein for the surface-attachment and obscuring/blocking a specific side of the protein by the surface. They also restrict the motion of the protein, and so the binding event is not in its native state.

Optical trapping of single proteins can, in principle, allow for studying protein interactions without the need for surface binding. Recently, our group has developed an optical trapping approach using nano-apertures [8,9], which allows for trapping of single proteins. Here we show single protein binding, using the biotin-streptavidin model system. Since biotin is a small molecule it requires tethering to a larger particle to observe binding with streptavidin, however, we might be able in the future to observe single molecule binding of bigger molecules without the need of tethering because molecules can be held in place by the optical trap. Protein binding is confirmed by performing two appropriate control experiments: 1) we trapped biotin-coated PS particles and then flowed in streptavidin with the binding sites being blocked off by mixing it with excess biotin, and 2) we trapped non-functionalized PS particles and then flowed in streptavidin. Both of the control experiments consistently did not show any binding. We believe that this is the first definitive measurement of single protein binding using an optical trapping system, without the need for tethering the protein to a larger particle.

2. Experimental setup

Figure 1 shows a schematic of the DNH optical trapping system which is based on Thorlabs optical tweezer kit (OTKB). We used an 830 nm continuous laser (Sacher Lasertechnik Group, Model TEC 120) which has the advantage of better detection efficiency for the photodetector than a 980 nm laser and helps with trapping smaller objects due to the favorable wavelength-dependent scaling. The trapping beam was focused into the sample using a 100 × oil immersion microscope objective with a 1.25 numerical aperture. We used an optical density filter (ODF) to limit the optical power at the output of the objective lens. A half-wave plate (HWP) was used to rotate the polarization of the incident beam so that the electric field of the beam is aligned along the cusps of the DNH giving a large local field enhancement and hence creating a strong trapping point.. Transmitted light through the DNH was collected using a 10 × condenser microscope objective with 0.25 numerical aperture, and measured by a silicon-based avalanche photodetector (APD) (Thorlabs APD110A). We used a data acquisition board to record the voltage values generated by the APD at a sampling frequency of 2 kHz. The setup is modified so as to allow for microfluidic delivery to the trapping site where we used a dual Fusion Syringe Pump (Chemyx Inc., Model Fusion 200) to sequentially flow biotin-coated PS particles and streptavidin into the microfluidic channel. Here, we followed the same microfluidic chip fabrication process outlined in the Supporting Information of our previous work on flow dependant trapping [10], with the resultant microfluidic channel dimensions being 65 µm by 800 µm.

 

Fig. 1 A schematic of the DNH optical trap with dual microfluidic input. Abbreviations used: ODF = optical density filter; HWP = half-wave plate; BE = beam expander; MR = mirror; MO = microscope objective; OI MO = oil immersion objective; APD = avalanche photodetector.

Download Full Size | PDF

A focused ion beam (Hitachi FB-2100 FIB) was used to mill a DNH in a 100 nm thick Au film on a glass substrate with a 5 nm titanium adhesion layer (EMF Corp.). Figure 2 shows a schematic of the protein binding experiments and ascanning electron microscope image (Hitachi S-4800 FESEM) of the DNH used in trapping experiments. The separation between the cusps was measured to be 30 nm which is close to the optimal 25 nm separation found to be the best for trapping 20 nm PS particles according to our previous studies [11]. As with our previous work on trapping a protein, prior to conducting trapping experiments we formed a monolayer of mPEG thiol on the Au surface by immersing the Au sample in a 5 mM aqueous solution of mPEG thiol (molecular weight, 5 kDa) at room temperature overnight and then rinsing it with DI water. This helps prevent streptavidin from adsorbing on the Au surface.

 

Fig. 2 A schematic showing the protein binding experiments .(a) 20 nm biotin-coated PS particle approaches the DNH. (b) Introduction of streptavidin to the trapping site once a successful trapping event of 20 nm biotin-coated PS particle is achieved. (c) Streptavidin is bound to biotin between the two sharp cusps of the DNH. (d) A scanning electron microscope image of the DNH used in the protein binding and control experiments.

Download Full Size | PDF

Here we experimentally demonstrate protein binding by trapping 20 nm biotin-coated PS particles (NANOCS PS20-BN-2) and then bind it with streptavidin with a molecular weight 60 kDa. The streptavidin was purchased in powder form (SIGMA-ALDRICH, Product No. 85878) and the desired concentration was obtained by dissolving it in phosphate buffered saline (PBS). Biotin, which is also known as vitamin B7 and vitamin H, has an extraordinary binding affinity to streptavidin, with a dissociation constant in the order of about 10⁻¹⁴ mol/L, making it the strongest non covalent bond found in nature [12].

3. Protein binding experiments

3.1 Protein binding detection

Figure 3 shows the time evolution of the optical power transmitted through the DNH and the schematics the corresponding steps of the protein binding experiments, where (a) shows flowing 20 nm biotin-coated PS particles (0.01% w/v in DI water), (b) trapping event of 20 nm biotin-coated PS particle between the two sharp tips of the DNH and subsequently flowing streptavidin (0.01% w/v in PBS), (c) binding of streptavidin with the trapped biotin-coated PS particle. We repeated all of the experiments and controls at least 4 times. When a biotin-coated PS particle gets trapped between the cusps of the DNH it dielectrically loads the region and causes an enhancement of the local optical field intensity and increased DNH transmission [13–18]. Once a stable trapping event of a biotin-coated PS particle is achieved, streptavidin is flowed into the channel where it gets bound to the trapped particle, causing further dielectric loading to the DNH region and consequently a higher optical transmission is obtained. We used a moderate flow rate of 5 µL/min which enables fast enough delivery of streptavidin to the trapping site and at the same time does not wash away the trapped particle, as was investigated in one of our previous studies [10].

 

Fig. 3 Time trace of optical transmission through the DNH where (a) shows flowing 20 nm biotin-coated PS particles through the microfluidic channel, (b) trapping of 20 nm biotin-coated PS particle between the two sharp tips formed by two overlapping DNHs and subsequently flowing streptavidin, and (c) binding between 20 nm biotin-coated PS particle and streptavidin.

Download Full Size | PDF

The setup is designed such that delivery of streptavidin to the trapping site takes about 150 seconds from the time a successful trapping of biotin-coated PS particle is achieved. The delivery time of streptavidin to the trapping site of 150 seconds was related to the time for it to flow from the syringe to the trapping site at a flow rate of 5 µL/min. This indicates that a binding event occurs within a few seconds after the introduction of streptavidin to the trapping site, which is expected due to the high binding affinity of the streptavidin-biotin pair. Similar binding results were obtained on different days and with freshly made solutions. After each set of experiments, we flowed deionized water through the microfluidic channel for 30 minutes to make sure that the channel is clean for the next set of experiments.

4. Control experiments

To show clearly that we are actually observing specific streptavidin-biotin binding, and not just co-trapping, we performed two separate control experiments using: 1) saturated streptavidin, 2) non-functionalized PS particles.

4.1 Saturated streptavidin

Figure 4 shows the time trace of the optical transmission through the DNH with the schematic of the corresponding steps for the saturated streptavidin control experiment where, (a) shows flowing 20 nm biotin-coated PS particles, (b) trapping event of 20 nm biotin-coated PS particle and subsequently flowing saturated streptavidin, and (c) saturated streptavidin does not bind to the trapped 20 nm biotin-coated PS particle. In this set of control experiments, we trapped a biotin-coated PS particle and then flowed in saturated streptavidin in which all the binding sites were blocked off by mixing it with excess biotin (SIGMA-ALDRICH), molecular weight 344.31 Da. As the figure shows, there is an increase in the optical transmission through the DNH observed after flowing the 20 nm biotin-coated PS particles indicating a successful trapping event of biotin-coated PS particle. However, we have not observed any further increase in the optical transmission for the subsequent 25 minutes after flowing saturated streptavidin.

 

Fig. 4 Time trace of optical transmission through the DNH for the saturated streptavidin control experiment where (a) shows flowing 20 nm biotin-coated PS particles through the microfluidic channel, (b) trapping of 20 nm biotin-coated PS particle between the cusps of the DNH and subsequently flowing saturated streptavidin, and (c) saturated streptavidin does not bind to the trapped 20 nm biotin-coated PS particle.

Download Full Size | PDF

4.2 Non-functionalized PS particles

Figure 5 shows the time evolution of optical transmission through the DNH with the schematic of the corresponding steps for the non-functionalized 20 nm PS particles where (a) shows flowing 20 nm non-functionalized PS particles (0.01% in DI water), (b) trapping event of 20 nm PS particle and then flowing streptavidin, and (c) streptavidin does not bind to the trapped 20 nm PS particle. Here, we flowed in streptavidin (with the same concentration used in the main binding experiments) after successfully achieving a trapping event of a 20 nm non-functionalized PS particle. Introduction of streptavidin to the trapping site did not cause any change in the optical transmission through the DNH for the subsequent 25 minutes after achieving a stable trapping event of a non-functionalized PS particle.

 

Fig. 5 Time trace of optical transmission through the DNH for the non-functionalized PS particle control experiment where (a) shows flowing 20 nm PS particles through the microfluidic channel, (b) trapping of 20 nm PS particle between the cusps of the DNH and subsequently flowing streptavidin, and (c) streptavidin does not bind to the trapped 20 nm PS particle.

Download Full Size | PDF

5. Discussion

5.1 DNH trap for single protein binding detection

Our objective in this current study has been to observe single protein binding using the DNH trap system. Our experimental results on single protein binding as well as the control experiments explicitly demonstrate that single protein binding detection is achievable by using the DNH optical trap system. Unlike other detection systems which require tethering and/or surface functionalizeation of the particles of interest [19–24], our system uses the trapping laser beam to hold the trapped particle in a well defined trapping region between the two cusps of the DNH. Furthermore, our DNH trap allows for single protein binding detection by simply measuring optical transmission of the same trapping laser beam through the DNH aperture without the need of measuring wavelength or having additional laser sources [25–28] or fluorescently labeling the target particle [29]. This straightforward intensity measurement of optical transmission could yield a much faster detection of protein binding and therefore, allow the study of single protein binding dynamics at a nanosecond scale. Besides, the distinct and abrupt jump in the optical transmission through the DNH serves as a potential protein binding sensor with a signal-to-noise ratio (SNR) of 18. Indeed a recent study used a streptavidin–R-phycoerythrin conjugate with a molecular weight of 300 kDa (50 times bigger molecule than the one we used in our study) to clearly detect protein binding using surface plasmon resonance shift of Au nanorods [7].

Protein-protein interactions are generally temperature-sensitive and require systems which operate below the denaturation temperature of the proteins under study. For instance, recent works on protein studies, like the resonant-based plasmonic particle trapping [30–32] and photonic crystal trapping [33,34] systems suffer from heating effects and might need to apply one or more thermal management strategies to overcome heating issues; for example, the use of adjacent metal films as a natural heat sink in nanopillar plasmonic trapping [35,36]. These resonant trapping systems do not only generate heat under resonant absorption but also add experimental complexity in terms of applying thermal management strategies. On the other hand, the DNH trap system is non-resonant and the Au film plays a role in reducing the heating effects due to its good thermal conductivity. Considering other works on optical trapping where there is a metal membrane [36], we expect that the temperature increase is only a few degrees celcius and so this is not expected to hinder protein binding, protein folding, or other molecular investigations. We are presently working on integrating Raman into the setup so that we could potentially measure the temperature distribution by looking at the ratio between Stokes and anti-Stokes peaks. Here, we expect the temperature at the trapping point to be well below 70 °C as this is the temperature required to break the streptavidin-biotin bond without denaturing streptavidin [37] and such thermal effects on streptavidin-biotin pair have not been observed in our experiments. This feature makes our DNH optical aperture trapping system thermally suitable for protein-protein studies, for example, studying the dynamics of protein folding and unfolding of BSA [8].

It is possible to envision schemes whereby the optical polarization may be used to rotate the trapped molecules, as has been achieved for other works [38]. The double nanohole constrains the polarization of the maximum trapping field along the cusps between the two nanoholes, and therefore rotation is not possible.

In our previous studies we found that steric hindrance prevents co-trapping of two similar particles in the region between the two sharp tips of DNH [10]. Therefore, it stands reasonable here to assume that co-trapping of two 20 nm biotin-coated PS particles in a 30 nm gap is subject to steric hindrance too. Indeed, in our current study we have neither observed co-trapping of a biotin-coated PS particle with another biotin-coated PS particle nor co-trapping of streptavidin with streptavidin. In addition, we have not observed any co-trapping behavior in our control experiments. This further proves that the results obtained in this study are as a consequence of streptavidin-biotin binding and not co-trapping of two similar particles.

5.2 Future directions

It is interesting that we observed protein binding at a single molecule level using the DNH aperture trap. We aim to extend our study to further test the binding sensitivity of our system by using binding affinity pairs which have lower disassociation constants. In addition, there is also the potential of analyzing time traces of binding events of different binding affinity pairs to investigate the possibility of identifying the nature of trapped/bound particle(s). It is also possible to modify our existing setup to analyze the Raman signal of the trapped/bound particle(s) for which we have experience with Raman spectroscopy [39]. We are presently attempting to simultaneously measure the Raman signal of the trapped object by using the trapping laser as the Raman excitation source, as well as appropriate filters and a spectrometer after the double nanohole. Moreover, it is interesting to study isolated nanoparticles within a heterogeneous mixture. Furthermore, we are interested in studying the dynamics of folding and unfolding of single proteins. Moreover, we are interested in fluorescence studies of trapped particles provided that quenching effects at the Au surface can be avoided.

6. Conclusion

In conclusion, we have used a DNH-based optical trap to detect streptavidin-biotin binding at the single molecule level. This was achieved by measuring the optical transmission through the DNH. To further confirm that the observed increase in optical transmission through the DNH is due to the specific binding of streptavidin to biotin, we preformed two appropriate control experiments that did not show binding, i.e. using saturated streptavidin and non-functionalized PS particles. The control experiment results were consistent with the main experimental results. These results are promising for biomolecular interaction studies at the single molecule/particle level; for example studying single protein binding interactions and dynamics at single molecule level.