Graphene functionalized field-effect transistors for ultrasensitive detection of Japanese encephalitis and Avian influenza virus

Graphene, a two-dimensional nanomaterial, has gained immense interest in biosensing applications due to its large surface-to-volume ratio, and excellent electrical properties. Herein, a compact and user-friendly graphene field effect transistor (GraFET) based ultrasensitive biosensor has been developed for detecting Japanese Encephalitis Virus (JEV) and Avian Influenza Virus (AIV). The novel sensing platform comprised of carboxy functionalized graphene on Si/SiO2 substrate for covalent immobilization of monoclonal antibodies of JEV and AIV. The bioconjugation and fabrication process of GraFET was characterized by various biophysical techniques such as Ultraviolet–Visible (UV–Vis), Raman, Fourier-Transform Infrared (FT-IR) spectroscopy, optical microscopy, Scanning Electron Microscopy (SEM) and Atomic Force Microscopy (AFM). The change in the resistance due to antigen–antibody interaction was monitored in real time to evaluate the electrical response of the sensors. The sensors were tested in the range of 1 fM to 1 μM for both JEV and AIV antigens, and showed a limit of detection (LOD) upto 1 fM and 10 fM for JEV and AIV respectively under optimised conditions. Along with ease of fabrication, the GraFET devices were highly sensitive, specific, reproducible, and capable of detecting ultralow levels of JEV and AIV antigen. Moreover, these devices can be easily integrated into miniaturized FET-based real-time sensors for the rapid, cost-effective, and early Point of Care (PoC) diagnosis of JEV and AIV.


Results and discussion
Characterization of graphene-bioconjugate. Figure 1a shows the steps involved in the fabrication of GraFET device, activation process, antibody attachment and detection of antigen by monitoring change in electrical characteristics whereas Fig. 1b depicts the steps involved in the binding of graphene-Ab conjugate.
The binding of graphene with JEV/AIV antibody was ensured by observing various parameters. The UV-Vis spectra (Fig. 2a) which showed a peak for graphene at 270 nm, displayed a blue shift of 5 nm that confirmed the binding of graphene-Ab complex (inset of Fig. 2a). Further addition of Ag led to a red shift of 10 nm (from 265 to 275 nm), as the graphene-Ab bound with specific Ag that confirmed the presence of graphene-Ab-Ag complex. The Raman spectra (Fig. 2b) showed typical single-layer graphene device, with characteristics peaks at 2,600 cm −1 and 1,600 cm −1 . FT-IR spectra in Fig. 2c showed one peak for graphene as well as all graphenebioconjugates at 1644 cm −1 corresponding to C=C bond of graphene molecules which states that structure of graphene does not undergo change during conjugation and further experimentation. On addition of Ab, two additional peaks were observed at 2,128 cm −1 (N=C=N) (carbodiimide bond) that confirmed EDC-NHS carbodiimide reaction for activation of carboxyl groups on graphene, and another peak at 1,076 cm −1 (C-N) that confirmed binding of amine group of Ab to activated carboxyl group of graphene. Moreover, all three peaks could be observed on addition of BSA and Ag which proved no further changes in the graphene-Ab conjugation. The morphological characteristics were observed by SEM (Fig. 2d) at different stages of conjugation i.e. (i) only graphene, (ii) graphene-Ab, (iii) graphene-Ab-Ag complex, that showed changes in the surface of graphene at each step of immobilization.
Electrical characterisation of GraFET biosensor. The optical micrograph of a typical GraFET device before measurement was showed in Fig. 3a. The scanning electron and atomic force micrograph of a typical device was depicted in Fig. 3b, c respectively, with the height of the graphene channel to be 1.22 nm. The inset in Fig. 3d showed the image of the packaged device mounted on a ceramic chip-carrier. To check the device characteristics before the measurements, resistance ( R ) vs gate-voltage ( V g ) measurements were performed, where V g was swept at a rate of 600 V/h. The R-V g curve of a prototypical device shown in Fig. 3d, displayed an ambipolar behaviour, where the peak in resistance corresponds to the Dirac point or charge neutrality point. Before the actual measurements were performed, we checked the response of the antibody-antigen combination Scientific RepoRtS | (2020) 10:14546 | https://doi.org/10.1038/s41598-020-71591-w www.nature.com/scientificreports/ on a standard Au-FET, which showed negligible change in channel resistance, implying that the Au electrodes used in the FET do not play any role 29 . Additionally, to ensure the stability of our GraFET sensors, we monitored the resistance of the GraFET channel up to four weeks. As can be seen from Fig. 3e, a negligible change in the channel resistance was observed which proved that the FETs were stable against environmental degradation for at least a month.
Binding, competitive and specificity assay for JEV and AIV. Binding assay for JEV was performed in the range of 1 μg/mL to 0.0019 μg/mL while for AIV from 0.0625 to 0.000031 μg/mL (Fig. 4a, c). The optimum binding was observed from 0.03 to 0.1 μg/mL for JEV, and 0.0003 to 0.01 μg/mL for AIV. Along with the binding assay, specificity assay for binding of JEV Ag with AIV Ab and vice-versa was also carried out (Fig. 4a, c respectively) and it was observed that no cross-reactivity takes place. This shows that using these sets of Ag and Ab will ensure specificity of the developed immunosensor as there are negligible chances of cross-reactivity. For competitive immunoassay, we used 0.1 μg/mL for JEV Ab and 0.01 μg/mL for AIV Ab as fixed concentration, that was used for competitive reaction with a range of antigen from 1 μg/mL to 0.0019 μg/mL. The LOD was found to be 0.25 μg/mL for JEV and 0.031 μg/mL for AIV as shown in Fig. 4b, d. Along with the competitive assay, specificity assay was once again carried out to check for non-specific binding of JEV Ag with AIV Ab and vice-versa (Fig. 4b, d respectively) and it was observed that no cross-reactivity took place. This confirmed  www.nature.com/scientificreports/ BSA to ensure that all the free-sites present in graphene channels were blocked and any further change in resistance that would appear was only due to antibody-antigen interaction. As observed in Fig. 5a,b, there was a clear change in the channel resistance due to a change in the doping profile of graphene with subsequent addition of antigen (JEV and AIV) ranging from 1 fM to 1 µM. As shown in Fig. 5a, after 1 fM addition of antigen, the resistance of graphene channel decreases and at 100 nM of JEV it saturated. However, as shown in Fig. 5b we observed a drop in resistance at 10 nM AIV antigen concentration and saturation at 1 µM. The antibody-antigen binding modifies the local electrostatic environment 48 which leads to a change in the number density of the graphene channel, manifested as a change in the resistance/conductance during the measurements. The eventual saturation in kinetic response on subsequent addition of antigens beyond a device specific threshold value could be an indicator that all the active sites on the device are occupied. This prevents any further change transfer, and hence any change in the channel resistance/conductance is also absent. The increase or decrease of resistance (or conductance) depends on the type of doping of both the channel and that due to the target-analyte interactions 49 .
For a quantitative analysis of the GraFET response towards sensing after addition of the specific antigen to the antibody bound to graphene, the percentage change in channel resistance was calculated, where resistance in buffer solution (PB) (added just before the antigen addition) was taken as the baseline as shown in Fig. 5a,b. The maximum change in channel resistance calculated for JEV was ~ 10% (Fig. 5a) and for AIV it was ~ 20% (Fig. 5b), for Ag concentrations ranging from 1 fM to 1 µM, which made our GraFET biosensors one of the most sensitive detectors till date in identifying JEV and AIV (as depicted in Table 1). It is more sensitive than the ELISA tests showed in Fig. 4 (a,b for JEV and c,d for AIV). Further improvement in sensitivity is possible by increasing the number of active sites, possibly by using a larger area of graphene. Reproducibility of these devices were carried out in our earlier research work 29 where different antigen were tested on three separate sensors each and the consistent results proved high reproducibility of this immunosensor. Similarly specificity studies were also carried out in the same research paper 29 using different antigen. Additionally, these devices are highly stable, and can be integrated into electronic chips for cost effective, on-field detection along with ease of handling.

Materials and methods
Materials and reagents. Single crystals of Kish graphite were obtained from Covalent Materials Corp.

Bioconjugation of graphene-antibody and biophysical characterization.
To immobilize antibodies on graphene, the graphene must first be activated using carbodiimide chemistry. 0.5 mg of graphene was added to 75 μm EDC and 75 μm sulfo-NHS (total volume 1 mL) and gently stirred for 2 h at RT. Here EDC in conjugation with NHS allows a two-step coupling reaction by activating the carboxyl groups for conjugation. The activated graphene complex was then centrifuged for 15 min at 10,000 rpm at 4 °C following which the pellet was resuspended in 1 mL phosphate buffer (PB) (50 mM, pH 7.4). Anti-JEV and anti-AIV monoclonal www.nature.com/scientificreports/ antibodies were drop-wise added to activated graphene complex, separately, for 30 min at RT and then left overnight for incubation at 4 °C. The unbound sites of the graphene-Ab complex were blocked using 1% BSA in 50 mM PB (pH 7.4) for 2 h at RT. Antigen (JEV and AIV separately) was added to the graphene-Ab complex after blocking, and incubated for 2 h at RT. To confirm the above binding steps, characterisation of each step was carried out using Single-beam UV-Vis spectrometer in the range of 190-800 nm, Fourier Transformed Infra-Red (500-4,000 cm −1 ), Raman Spectroscopy (1,300-3,000 cm −1 ) and Scanning Electron Microscope (morphological analysis). Quorum SC7620 sputter coater was used to electro-activate the samples for SEM imaging by applying a coat of gold.
Binding, competitive and specificity ELISA for JEV and AIV. Using the antigen and antibodies procured for JEV and AIV, both binding and competitive ELISA were carried out to check the sensitivity of immunoassay and later compared to the sensitivity of the developed GraFET sensor. For standardisation of the antibody concentration to be used in competitive assay, indirect binding ELISA was carried out. 96-well NUNC ELISA plates were coated with 100 μL of JEV and AIV antigen separately [0.25 μg/mL in carbonate buffer (pH 9.6)] (optimal standardised concentration) and incubated O/N at 4 °C. The plates were washed thrice with 0.02%  www.nature.com/scientificreports/ experimental steps were followed the same as the above binding assay. This cross reactivity test was carried out to check the specificity of the Ag-Ab interaction and reading were taken at 450 nm in Perkin Elmer Lambda 25 Multi-scan Spectrophotometer. From the above binding assay, the parameters and concentrations were standardised for competitive ELISA for both JEV and AIV. The initial steps remained the same as the above indirect ELISA. After blocking and washing, different Ag-1°Ab (JEV/AIV) were added to the plates (100 μL/well) where the 1°Ab concentration remained constant (0.1 μg/mL for JEV and 0.01 μg/mL for AIV as standardised in the indirect ELISA) and the Ag concentrations were diluted twofold in 0.1% PBS-M ranging from 1 μg/mL to 0.0019 μg/mL. After 2 h incubation at 37 °C the remaining steps i.e. 2°Ab, washing, TMB and stop solution remain the same as above. Readings were taken at 450 nm in Perkin Elmer Lambda 25 Multi-scan Spectrophotometer. After carrying out the competitive assay, the same experiment was carried out as above using AIV Ab-JEV Ag conjugate on JEV Ag coated plate and JEV Ab-AIV Ag conjugate on AIV Ag coated plate. All the concentrations and experimental steps were followed the same as the above competitive assay. This cross reactivity test was carried out to check the specificity of the Ag-Ab interaction and reading were taken at 450 nm in Perkin Elmer Lambda 25 Multi-scan Spectrophotometer.
Fabrication of GraFET sensor. Graphene was mechanically exfoliated from bulk graphite using the conventional scotch-tape method 85 with 3 M magic tape, on RCA-cleaned hot SiO 2 /Si substrate, where the 285 nm SiO 2 on Si functions as the back-gate dielectric. A high-resolution (1000×) optical microscope (Olympus BX51) was used to locate and select the desired graphene sheets. The substrates were spin-coated with PMMA 495 and PMMA 950 positive-resist, and baked at 150 °C. Electron beam lithography (Raith Pioneer) was performed to define the electrode patterns, which were developed using 1:3 mixture of methyl isobutyl ketone (MIBK) and isopropyl alcohol (IPA). After development, 5 nm thick Cr, and 50 nm thick Au layers were thermally evaporated to form the source-drain contacts followed by the lift-off process. Substrates were fixed onto a ceramic chipcarrier using silver-epoxy paste, and wire-bonded (TPT HB05 wire bonder) to the active contact pads.
Electrochemical performance of GraFET sensor. The JEV and AIV antibodies were pre-activated separately by EDC/NHS and immobilised on the surface of FET by being allowed to bind to the graphene for 30 min in 50 mM Phosphate Buffer (PB) at pH 7.4. Excess unbound Abs were washed off using 50 mM PB (pH 7.4). 1% BSA in 50 mM PB (pH 7.4), was used to block any non-specific sites. Different concentrations (1 fM to 1 μM) of JEV/AIV Ag were prepared in 50 mM PB (pH 7.4) to check the sensitivity of the fabricated FET biosensor. Measurements were performed and the change in the resistance was recorded at each stage by passing a constant current circuit of 100 nA through the FET device.

Conclusions
From our study, it can be concluded that GraFETs are an excellent tool for early detection of JEV and AIV. The developed biosensor proved to be highly sensitive with a detection range of 1 fM to 1 μM and limit of detection of 1 fM for JEV and 10 fM for AIV. This proved that graphene is an excellent choice for real-time sensing of low concentration of JEV/AIV antigen in the initial stages of the infection itself resulting in rapid diagnosis. The major advantage of this device over conventional JEV/AIV diagnosis methods, besides sensitivity and specificity, is quick response time and high POC potential. However, the miniaturization of the device for on-site applicability in the field would be expensive especially since each device can be used only once and can not be repeated. Alternate cost-effective fabrication processes need to be developed so that this device can be used as a Point of Care diagnostic platform for JEV/AIV as well as other diseases.