PEG Functionalization of Whispering Gallery Mode Optical Microresonator Biosensors to Minimize Non-Specific Adsorption during Targeted, Label-Free Sensing

Abstract

Whispering Gallery Mode (WGM) optical microresonator biosensors are a powerful tool for targeted detection of analytes at extremely low concentrations. However, in complex environments, non-specific adsorption can significantly reduce their signal to noise ratio, limiting their accuracy. To overcome this, poly(ethylene glycol) (PEG) can be employed in conjunction with appropriate recognition elements to create a nonfouling surface capable of detecting targeted analytes. This paper investigates a general route for the addition of nonfouling elements to WGM optical biosensors to reduce non-specific adsorption, while also retaining high sensitivity. We use the avidin-biotin analyte-recognition element system, in conjunction with PEG nonfouling elements, as a proof-of-concept, and explore the extent of non-specific adsorption of lysozyme and fibrinogen at multiple concentrations, as well as the ability to detect avidin in a concentration-dependent fashion. Ellipsometry, contact angle measurement, fluorescence microscopy, and optical resonator characterization methods were used to study non-specific adsorption, the quality of the functionalized surface, and the biosensor's performance. Using a recognition element ratio to nonfouling element ratio of 1:1, we showed that non-specific adsorption could be significantly reduced over the controls, and that high sensitivity could be maintained. Due to the frequent use of biotin-avidin-biotin sandwich complexes in functionalizing sensor surfaces with biotin-labeled recognition elements, this chemistry could provide a common basis for creating a non-fouling surface capable of targeted detection. This should improve the ability of WGM optical biosensors to operate in complex environments, extending their application towards real-world detection.

Keywords: PEG; non-specific adsorption; nonfouling surfaces; optical microresonator; surface characterization; surface functionalization.

Figure 1

Illustration of WGM resonator, based on [17] (adapted with permission). Light (orange) enters a WGM resonator, where it experiences total internal reflection (TIR) and generates an evanescent field. When an analyte binds or adsorbs onto the surface of the microsphere, it changes the effective refractive index of the circulating optical field resonator, and it pulls part of the evanescent field to the outside of the resonator (blue for pre-binding, red for post-binding). The expansion of the optical field’s boundary causes the round-trip wavelength of light to increase about 2πΔl. The increase in the light wavelength results in a frequency shift in the transmission spectrum. The evanescent field is an optical field extending to the surrounding environment and decreasing exponentially with the distance away from the resonator’s interface.

Figure 2

Goal: the PEG-biotin:PEG coating of the microsphere should improve the specificity by repelling non-specific protein adsortion. Avidin has four binding sites for biotin (two on each side); it is possible for avidin to bind 1–4 of these sites when it interacts with the surface, providing the biotin sites are within an appropriate distance. We expect avidin to bind to 1–2 biotin molecules on the surface, and present the other two binding sites to the environment; however, it is possible for a planar conformation to occur where all four sites bind the biotin tethered to the surface.

Figure 3

Two-dimensional schematic of the PEG-biotin:PEG ratios of 1:1, 1:2, and 1:3. The grey surface indicates the (100) silica-on-silicon wafer surfaces. The curved lines on the surface indicates the PEG coating, while the dark blue dots indicate the biotin molecules. Note that the amount of PEG-biotin in solution was kept constant, while the amount of PEG was varied.

Figure 4

Functionalization process of silica microsphere’s surface via a two-step covalent process. Adapted with permission from [35]. Copyright (2012) American Chemical Society. The silica surface was first populated with hydroxyl groups by exposing the device to piranha solution. The hydroxylated surface was then PEG-terminated and PEG-biotin terminated using silane coupling agents attached to PEG molecules via solvent-based, covalent grafting techniques.

Figure 5

Overall reaction scheme for the biotinylation of silica surfaces (based on reference [36]). (1) Hydroxylation of the silica surface; (2) Amination of the hydroxylated surface via the silane coupling agent; (3) Biotinylation of the aminated surface via NHS ester chemistry; (4) Stripping of the surface via oxygen plasma, resulting in a hydroxylated surface.

Figure 6

The experimental setup of device characterization. A—Laser B—Laser Controller C—Stage Controller D—Syringe Pump E—Photo Detector F—Nano-Positioning Stage G—Taper Holder H—Taper I—Computer. Light generated by a tunable diode laser propagates along an optical fiber. Once the Whispering Gallery Mode is excited upon coupling, the output signal is transferred to the detector and computer, and Q factor can be obtained automatically.

Figure 7

A model of the open-flow flow cell used for sensing experiments. This cell has injection ports for both the test molecule (avidin, lysozyme, or fibrinogen) and the PBS buffer solutions.

Figure 8

Mean thickness change (±standard deviation, nm) of functionalized wafers after adsorption. Three spots were measured on each wafer with ellipsometry, and three wafers were examined for each group. The values correspond to the amount of non-specific protein adsorbed onto the functionalized wafers.

Figure 9

Mean Roughness Parameters (±standard deviation) of functionalized wafers after lysozyme adsorption. Three spots were measured on each wafer with optical profilometry, and three wafers were examined for each group. The addition of PEG to the surface appears to slightly reduce the overall roughness, although the initial surfaces are very smooth.

Figure 10

Mean fluorescence intensity (±standard deviation) of functionalized wafers. Five regions of 256 by 256 pixels were measured on each fluorescence image with an Olympus IX 70, and three images were taken for each group of wafers.

Figure 11

Representative fluorescence microscopy image of the microsphere functionalized by PEG-biotin:PEG from a ratio of 1:1. The functionalization process resulted in a uniform and smooth coverage on the surface.

Figure 12

Quality factors of each microresonator used in the sensing study before (solid squares) and after (hollow squares) coating. Inset: a representative resonance (black line—data, red line—Lorentzian fit), showing a high quality factor device.

Figure 13

Average resonant wavelength shift of the PEG-biotin:PEG coated microresonators at concentrations of 1000, 500, 100, and 10 μg/mL for fibrinogen, lysozyme, and avidin.

Figure 14

A representative example of the resonant wavelength shift of one of the PEG-biotin:PEG coated microresonators used in the study for different concentrations of fibrinogen, lysozyme, and biotin at the concentration of 1000 μg/mL. Fibrinogen and lysozyme show a minimal response, while avidin has a clearly defined wavelength shift that is approximately 20 pm.

Figure 15

Representative response of one of the PEG-biotin:PEG coated microresonators for different concentrations of avidin. The wavelength shift declines as the concentration of avidin declines. Minimal response is seen at the concentration of 10 μg/mL.