Molecularly imprinted polymer-based protein assay

Molecularly imprinted polymers for biomacromolecule recognition should have binding sites easily accessible for these large molecules, e.g. confined to their surface (see chapter 4.2.3.1./MIPs). The aim of this work was to create surface-imprinted polymeric nanostructures (surface-imprinted polymers, SIPs) by nanosphere lithography. The idea was to apply protein-modified nanoparticles on the transducer surface and electropolymerize a layer of conducting polymer around them with thicknesses on the order of the bead radius. After removal of the particles not only a 2D arrays of periodic complementary cavities remain in the polymer but the imprints of the protein as well. These protein imprints are selective recognition sites for the target protein on the enlarged surface of the polymer film.

As proof of principle, to imprint proteins via nanosphere lithography for the first time, avidin was used as target molecule. A layer of 750 nm diameter avidin-modified polystyrene particles were deposited on the planar gold surface of a quartz crystal resonator. The protein imprint was created in two dimensionally ordered PEDOT(PSS) film, electropolymerized around the nanoparticle conjugates, after the dissolution of the beads in toluene. The rebinding of avidin into the selective recognition sites of the surface-imprinted polymer surface was measured by QCM.

To determine the effect of the surface imprinting on the binding properties of the polymer, specific binding was studied as well by preparing non-imprinted polymer films (NIPs). The NIP films were synthesized in the exact same way as the SIPs but unmodified (not modified with avidin) nanoparticles were used for patterning the PEDOT(PSS). A compact layer of PEDOT(PSS) was prepared as well with the same polymerization charge, but untemplated.

The optimization steps were carried out on the NIP films as well.

My part of the work mainly contained the optimization of the polymer layer thickness, as well as preliminary non-specific binding studies including the optimization of the blocking process.

Similarly to the previously described nanosphere lithography work, the polystyrene nanoparticles were simply drop casted from an aqueous suspension, but this time onto the gold surface of a quartz crystal of a QCM resonator. Generating compact hexagonal monolayers of unmodified nanoparticles 750 nm in diameter on a 7.07 mm² glassy carbon electrode surface proved to be straightforward, as discussed before (see images on page 77). Approaching the same result, however, on a 3 times larger (20.5 mm²) gold

89 surface proved to be more challenging. (Especially later on for the protein-modified particles, the array of the avidin-protein-modified beads was less uniform and compact than those assembled from unmodified particles on the same surface.) The goal was to create a monolayer as compact as possible on the surface.

Obviously, a compact layer is desirable as it maximizes the imprinted to non-imprinted ratio and hence reduce the non-specific binding.

The voids between the deposited nanoparticles were filled by potentiostatically growing a 2D film of PEDOT(PSS) from aqueous solution. As such, the electropolymerization could be performed in mild conditions compatible with the protein target, unlike most of the conventional MIP generating polymerizations carried out in aprotic organic solvents. The choice of PEDOT(PSS) was furthermore motivated by its inherently high biocompatibility. The polymerization process was easily controllable by monitoring the current. The removal of the beads was carried out by dissolving the polystyrene particles in toluene.

Figure 43 Microscope image at the same magnification of (A) a monolayer of avidin-modified polystyrene beads (ø = 750 nm) drop casted onto the gold surface of the quartz crystal resonator and (B) partial removal of the beads from the imprinted polymer layer revealing both the beads and their imprints.

To obtain maximum imprinted to non-imprinted surface ratio, based on simple geometrical calculations and assuming uniform growth of the film, a polymer thickness of 395 nm, slightly higher than the half-height of the beads (375 nm) would be ideal. The optimal experimental conditions for the polymerization were determined empirically. Films of different thicknesses were prepared by controlling the electrical charge during the polymerization and examining the patterned polymer layers with atomic force microscopy (AFM) after the dissolution of the beads, as shown in Figure 44. It must be noted, that the depth of the cavities in case of the thicker films cannot be measured directly, due to the geometry of the gradually closing cavities and the tip.

90

Figure 44 AFM images and relevant line scans showing the surface topography of PEDOT(PSS) films prepared by nanosphere lithography using 750 nm diameter beads. Scans (A-D) are representative of 2.2 µm × 2.2 µm areas of patterned PEDOT(PSS) films prepared using 7.5, 19, 30, and 41 mC/cm² surface charge densities, respectively.

5 10 15 20 25 30 35 40 45

0 100 200 300 400 500 600 700 800

average polymer thickness / nm

polymerization charge / mC/cm²

Figure 45 Theoretical correspondence between the average polymer thickness and the applied polymerization charge based on the AFM measurements, assuming uniform growth of the film.

91 AFM measurements revealed preferential growth of the PEDOT(PSS) film alongside the particles, so the average polymer thicknesses (Figure 45) were calculated based on the exposed diameter of the cavity, measured on several AFM images of different parts of the corresponding layers.

At a charge density of 7.5 mC/cm² small cavities with a depth of ca. 57 nm were formed, that increased at 19 mC/cm² to 195 nm. Higher charge densities of 30 and 41 mC/cm² resulted in a gradual enfolding of the nanoparticles by the grown polymer film, as shown by the decrease in the exposed diameter of the cavity (550 and 410, respectively), and by the polymer thicknesses, 600 and 689 nm, respectively. As Figure 45 shows, even assuming uniform growth of the film the control over the polymer thickness between 250 and 500 nm is the most critical, and due to the preferential growth of the polymer around the particles it becomes more complex. To ensure a better control over the polymer growth and be less affected by preferential growth around the beads, a somewhat lower charge density, 17 mC/cm², was chosen.

The dissolution of polymer beads exposed parts of bare gold on the substrate corresponding to areas where the beads previously touched the surface, causing a significantly lower non-specific binding of avidin to the compact film than to the NIP. To reduce non-specific binding to these spots, the surface was treated with 1 mM HS-TEG prior to the QCM measurements. As Figure 46 shows, surface blocking indeed reduced the avidin binding of the NIP by 28.5±0.8 %.

0 1000 2000 3000 4000 5000

0.0 0.1 0.2 0.3

surface conc. of Av / g/cm2

c_Av / nM

Figure 46Effect of blocking the gold surface, exposed after removal of the beads, by HS-TEG.

The non-specific adsorption of avidin on (red) patterned PEDOT(PSS) films blocked with HS-TEG reduces by 28 % compared to the (black) same surface but without blocking.

The hereunder measurements were done by Júlia Erdőssy (nee Bognár). The SIP film-modified 10 MHz quartz crystal chips were mounted into a flow cell.

92

After stabilization of the frequency, increasing concentrations of avidin were injected in the carrier buffer and the frequency change was monitored in real-time. The avidin binding of NIPs was recorded in similar conditions but in a separate experiment. Under the optimized experimental conditions, at the highest avidin concentration studied the amount of avidin bound on the imprinted surface was 1.34 µg/cm² while on the NIP surface 0.21 µg/cm² (Figure 47). This accounts for an imprinting factor of 6.5. Although this value is higher than the vast majority of imprinting factors reported, cannot reach that of the surface-imprinted microbands^[192]. This is most likely due to the relatively large imprinted fraction of the surface as result of the non-compact avidin-modified nanoparticle layer.

0 1000 2000 3000 4000 5000

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

surface conc. of Av /g/cm2

c_Av / nM

Figure 47 Binding isotherms of avidin to the (red) imprinted and (black) non-imprinted PEDOT(PSS) patterns, measured by Júlia Erdőssy. The data points are the average of measurements performed on three different SIP- and NIP-modified quartz crystals.

Outlook

The interest for surface confined molecularly imprinted polymer films is growing recently. The template size-matched polymer film thickness was shown^[269] to be optimal for these sensors to achieve the best surface-imprinting performance in terms of binding capacity, binding constant, imprinting factor, and binding selectivity.

Lately, the molecular imprinting approach seemed to be carried forward from bulk imprinting to nanoparticle imprinting^[180], core-shell type nanocomposites seem to dominate the field. The magnetic core, either magnetic nanoparticle^[269]

93

[270][271] or magnetic carbon nanotube^[272] is modified with the target molecule and the imprint is created in the thin porous shell on the surface.

Most of protein binding MIPs, however, are based on chemical polymerization.

The advantages of electropolymerization, such as its ability to fine-tune the film thickness by controlling the charge consumed during deposition, and to grow the film directly and rapidly at a precise area on the transducer surface is more rarely used^[273].

An interesting approach is the surface imprinted hybrid nanofilms, where the natural binding receptor of the target molecule immobilized on a transducer surface is combined with molecularly imprinted thin polymer films electrosynthesized around the receptor-target conjugate^{[274] [275]}. When rebinding the target protein to these hybrid MIP nanolayers, the strength of the specific receptor binding is coupled with the shape-specific selectivity of the molecular imprints, resulting in one of the highest imprinting factors reported^[275].

6.3.1. Conclusions

Nanosphere lithography seem to offer a versatile technique to create surface-imprinted polymers. Due to the inverse nature of the method, the whole macromolecule is imprinted in the polymer film, it simply can’t be too thin for the recognition or too thick for the protein to access the binding sites. The nanosphere carrier, furthermore, offers utmost flexibility in terms of adjusting the local chemical environment of the macromolecular template. The electrochemical deposition of the conducting polymer from aqueous solution offers an easily controllable, rapid fabrication technique and mild conditions highly suitable for protein imprinting.

94