A monomer bearing a pendant mannose (p-acrylamidophenyl-α-D-mannoside [α-man]) was prepared via procedures reported previously 9. N, N-dimethylacrylamide [DMAAm] (Wako Pure Chemical Industries, Ltd., 3-1-2, Doshomachi, Chuo-Ku, Osaka, 540-8605, Japan) was purified under reduced pressure (13 mmHg) in a nitrogen atmosphere before use. N, N'-methylenebisacrylamide [MBAAm], 2,2'-Azobis(2-methylpropionamidine) dihydrochloride (V50), acryloyl chloride, dimethyl sulfoxide [DMSO], anhydrous ethanol, manganese (II) chloride tetrahydrate, calcium chloride (Wako Pure Chemical Industries, Ltd.), 1-[4-(2'-hydroxyethoxy)-phenyl]-2'-hidroxy-2"-methyl-1-propane-1-one (Irgacure2959) (Ciba Specialty Chemicals, Inc., Klybeckstrasse, 141 CH-4002 Basel, Switzerland), Con A (lectin from Canavalia ensiformis) (J-Oil Mills, 1-11-1, Marunouchi, Chiyoda-Ku, Tokyo, 100-6226, Japan), 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid [HEPES] and 11-amino-1-undecanethiol, hydrochloride [AUT] (Dojindo Molecular Technologies, Inc., Kumamoto Techno Research Park, 2025-5 Tabaru, Mashikimachi, Kamimashiki gun, Kumamoto 861-2202, Japan) were all used without further purification. Deionized ultra-pure Milli-Q water (Millipore, 290 Concord Road, Billerica, MA, 01821, USA) (18.2 MΩ cm^-1) was used throughout the experiments. Hydrofluoric acid [HF] (23%) was prepared by diluting 46% HF (Morita Chemical Industries, Co., Ltd., 4-1-3, Kyutaromachi, Chuo-Ku, Osaka, 541-0056, Japan) with water.

HEPES buffers (20 mM) were prepared using recipes from buffer calculator 1011 developed by R. Beynon at the University of Liverpool, UK. The buffers were titrated to a specific pH with KOH solution, and the overall ionic strengths were fixed with KCl.

DMAAm, α-Man, MBAAm (as a crosslinker) and V50 (as an initiator) were dissolved in 50:50 (v/v) DMSO/water-mixed solvent as shown in Table 1. The total monomer concentration (DMAAm and α-Man) was either 2 or 4 M. The radical co-polymerization was carried out in a glass capillary (internal diameter [ID] = 0.93 mm; d_t) (Drummond Scientific Company, 500 Parkway, Box 700, Broomall, PA, 19008, USA), of which, both ends were sealed with an epoxy resin at 50°C for 24 h. After the reaction, the glass capillary was dissolved by immersing it in a 23% HF solution for 6 h. The obtained α-man gel was immersed in water and dialyzed in HEPES buffer (pH = 6.9, 25°C, I = 0.3 M) for 48 h to remove any remaining unreacted components. The purified gel was cut into pieces of 1-cm length, transferred to and equilibrated in a HEPES buffer (pH = 6.9, 4°C, I = 0.3 M) containing CaCl₂ (1 mM) and MnCl₂ (1 mM).

After equilibration at 4°C, the surface water of a piece of the gel was removed by gentle blotting with a laboratory tissue and was weighted (W₀). The gel was transferred to a HEPES buffer containing Con A (10 μM) in the presence of both CaCl₂ (1 mM) and MnCl₂ (1 mM). The gel weight at given duration (W) was measured and the swelling ratio of the gel was defined as in Equation 1:

Swelling ratio  [ % ] = { ( W - W 0 ) / W 0 } × 100

A gold electrode (5 × 5 mm²) was fabricated by the sputter deposition of an adhesion layer of titanium (10 nm) and then of a gold layer (90 nm, 99.99% purity) on a silicon substrate (Ferrotec silicon; Ferrotec Corporation, 1-4-14 Kyobashi, Chuo-ku, Tokyo, 104-0031, Japan). The gold electrode was cleaned before use with acetone, ethanol, water and, finally, piranha solution (H₂O₂/H₂SO₄ = 25/75 v/v) (extreme caution must be exercised when using piranha etch; an explosion-proof hood should be used). The surface was rinsed thoroughly with water and ethanol, and was dried with nitrogen gas. An AUT self-assembled mono-layer [AUT SAM] was formed on a clean gold electrode by immersing the electrode in a 2-mM AUT ethanol solution under nitrogen atmosphere at room temperature for 24 h. After repeated rinses and sonication in pure ethanol for 5 min, the electrode was dried with nitrogen gas. The AUT SAM, thus, obtained was then treated with acryloyl chloride under nitrogen atmosphere at room temperature for 6 h to modify the SAM terminus amino groups with acryloyl groups. After the reaction, the electrode was washed by sonication in ethanol for 5 min and rinsed with water. The electrode was covered with a teflon tape with a round hole (thickness = 50 μm, φ = 6 mm), in a way, exposing the central gold surface to the air.

A pregel solution was prepared by dissolving DMAAm, α-man, MBAAm as a crosslinker and Irgacure2959 as an initiator in DMSO solvent as shown in Table 2. α-man gel micro-film was synthesized by pipetting 1- μl droplets of the pregel solution onto the gold surface, which was then covered with a cover slip and placed under UV light (500 mW/cm², Aicure UJ20; Panasonic Electric Works Co., Ltd., 1048 Kadoma, Kadoma-Shi, Osaka, 571-8686, Japan) to allow radical co-polymerization.

A reaction chamber (200 μL volume) was made on the electrode. A glass tube (ID = 5 mm) was immobilized on the gold surface with a thermosetting-insulating epoxy resin (XA-1295/HQ-1) (Pelnox, Ltd., 8-7 Bodai, Kanagawa, 259-1302, Japan) by heating at 60°C for 12 h. The rest of the substrate, including bonding wires, had to be completely protected against water penetration from the reaction solution so the periphery of the chamber was completely covered with the epoxy resin. HEPES buffer (pH = 6.9, 4°C, I = 0.3 M) containing CaCl₂ (1 mM) and MnCl₂ (1 mM) was pipetted into the reaction chamber. The source and drain of a commercially available FET (N-Channel Depletion-Mode MOSFET, LND150) (Supertex Inc., 1235 Bordeaux Drive, Sunnyvale, CA, 94089, USA) and Ag/AgCl reference electrode were connected to a real-time FET analyzer (Optogenesis, 2-1-8 Kenpuku, Honjo-Shi, Saitama, 367-0044, Japan), and the gate of the FET was connected to the wire, which was connected to separated α-man gel-modified gold electrode.

The α-man gel-modified gold electrode was equilibrated in HEPES buffer (pH = 6.9, 4°C, I = 0.3 M) containing CaCl₂ (1 mM) and MnCl₂ (1 mM) prior to use. Real-time change in the gate potential was monitored and recorded at source current (I_S) of 1,800 μA and gate voltage (V_G) of 0 V (versus Ag/AgCl) when fixing source-drain voltage (V_D) at 1 V. Various concentrations of Con A solution were prepared in the HEPES buffer. Con A detection was carried out by adding 5 μM Con A into the reaction chamber and incubated at 25°C.

Several different types of α-man gels were synthesized by radical co-polymerization in shape of capillary (Table 1). As shown in Figure 2, molar ratio of the α-man was found to have significant impact on the diameter of the obtained gels (d/d_t). When the gels were prepared from 2 M total monomer concentration (2D, 2DS3, 2DS20 and 2DS80), a maximal equilibrium state diameter was found at 3 mol% of the α-man content. It is interesting to note that, at this critical value of the α-man content, the gel yielded in slightly white and opaque color which was not observed for those with 0 or 20 mol% α-man (2D and 2DS20), whereas a gel with 80 mol% α-man (2DS80) showed more profoundly white and opaque color. Transparency of a gel is an indicator of its own structural uniformity in the scale of visible light wavelength (several hundred nano-meters). Namely, with increased heterogeneity in this scale of topological structure of the gel, a fraction of scattered visible light (through the gel) increases, making the gel appear white and opaque. Therefore, observations in Figure 2 indicate that the composition of the α-man significantly affects the nano-structure of the resulting gel. We postulate that a strong hydrophobicity of the α-man monomer plays a role to induce micro- or nano-phase separation during the polymerization. The magnitude is dependent on both the concentration of itself as well as the polarity of the reaction solvent used. Preparation and characterization of such nano- and micro-structured hydrogels have been reported previously 12131415. Advantageously enough, it has been demonstrated that such heterogeneous gels are able to swell (hydrate) to a greater extent compared to homogeneous gels. This is also consistent with the results obtained in Figure 2. As will be discussed later, when introduced onto the gate surface of the FET, the swelling (hydration) degree of the gel is a determinant for both the signal direction and amplitude. This means that controlled nano-structure of the gel provides a means to modulate electrical signals that we actually observe.

Lectin binding to a single saccharide ligand is typically a low-affinity interaction. However, the multi-valent nature of lectin-saccharide interactions allows many low-affinity binding events to occur, resulting in high overall avidity 16. Multimeric lectins can cross-link multi-valent carbohydrate ligands, and they selectively cross-link with a single species of glycoprotein to form uniform lectin-carbohydrate lattices 17. Since tetrameric Con A can bind with four independent carbohydrates at once, one can anticipate that Con A-α-man complexation would result in inter- or intra-chain cross-linkages and, thus, deswelling of the gel (Figure 1a).

Figure 3 shows time courses of the swelling degree of the gels with the addition of Con A. In both cases, the gels swelled immediately after the addition of Con A; the extent of which is more prominent for 4DS3. At early stage of the α-man-Con A binding, it may occur in a 1 to 1 fashion rather than multi-valent complexation (Figure 4). Then a weakly (anionically) charged Con A should contribute to increase osmotic pressure of the gel, which would help the gel to swell rather than shrink (Figures 3 and 4). At later stage, as the thermal motion of the polymer chain matures, chances may increase for the multi-valent α-man-Con A complexation, which makes the gel deswell as observed in 4DS20 (Figures 3 and 4). 4DS3 gel bearing much less amount of α-man compared to 4DS20 did not show any significant deswelling even after 24 h. This can be explained by an insufficient number of α-man units available for multimeric recognition in the network, preventing the inter- or intra-chain cross-linkages. These results demonstrate feasibility to control the direction (swell or deswell) of the gel response simply by modulating the content of α-man.

As proposed in our previous work 6, the expression of threshold voltage [V_T] of the gel-modified FET is given by Equations 2 and 3:

V T = E ref - ψ 0 + χ sol - Φ Si q - Q OX + Q SS + Q B + Q gel C Com + 2 Φ f

with

C Com = C OX × C gel C OX + C gel = C OX 1 + C OX C gel

where E_ref is the constant potential of the reference electrode, ψ₀ is the potential drop in the electrolyte at the insulator-electrolyte interface, χ^sol is the surface dipole potential of the solution, Φ si q is the silicon electron work function, the fifth term is due to accumulated charge in the oxide(Q_OX) at the oxide-silicon interface (Q_SS) and the depletion charge in the silicon (Q_B), Φ_f is the potential difference between the Fermi levels of doped and intrinsic silicon, Q_gel is the charge in the gel layer and C_Com is the combined capacitance of the gate oxide (C_OX) and the α-man gel layer (C_gel). All other variables than Q_gel and C_gel are regarded constant throughout the chemical stimulation to the gel. On the basis of the operation function of FET (Equation 2), an increased anionic density on the gate surface gives rise to a positive directional shift of V_T, whereas a decreased anionic density gives a negative directional shift of V_T. Likewise, an increase in the gel permittivity on the gate surface leads to a negative direction shift of V_T and vice versa.

Figure 5a shows time course response of V_T of 4DS20-modified FET upon addition of 5 μM Con A. V_T undergoes a three-stage response: the first was an immediate and abrupt increase (stage 1) and the second response was a gradual decrease (stage 2) followed by slow re-increase (stage 3). The shrinking and swelling processes of the polymer gels take place in order of (1) diffusion of the solute molecules in the polymer network, (2) relaxation of the polymer chains due to (de) solvation and (3) diffusion of the polymer chains into (or out of) the solvent 18. From the directions of the V_T shifts observed in Figure 5a and in light of the above discussion, the stage 1 response can be attributed to the charge effect due to both diffusion of Con A (pI = ca.5-7) 1920212223 into the hydrogel and the complexation with α-man (in a 1:1 fashion) corresponding to step 1 of the gel response. On the other hand, the secondary and third changes can be attributed to the change in permittivity of the gel. It should be noted that the volume change of a hydrogel is practically equivalent to the change in water content. That is to say, as the gel deswells over time (stages 2 and 3), the fraction of high permittivity water decreases, decreasing the capacitance factor of the gel/gate interface. This safely explains the observed recovery of V_T at later stages of the response. The V_T shifts of 4DS50-modified FET as a function of Con A concentration were shown in Figure 5b. A data point at 1 h after the addition of the solution was defined as the equilibrium V_T shift for buffer without Con A, and a data point at the end of the second stage was defined as the equilibrium V_T shift for Con A-containing buffer. When the Con A-containing buffer was added to the gel-modified FET, positive shifts of the V_T were observed and increased as the Con A concentration increased. This result shows the potential of the α-man gel-modified FET for detecting saccharide-protein interaction. These results show that the gel-modified FET is not only able to electrically detect weakly charged proteins of large molecular weights on the basis of protein-saccharide interactions but also reveal kinetics of the signal transduction.