Quantitative and Selective Surface Plasmon Resonance Response Based on a Reduced Graphene Oxide–Polyamidoamine Nanocomposite for Detection of Dengue Virus E-Proteins

Dengue viral infection is one of the most common deadliest diseases and has become a recurrent issue for public health in tropical countries. Although the spectrum of clinical diagnosis and treatment have recently been established, the efficient and rapid detection of dengue virus (DENV) during viremia and the early febrile phase is still a great challenge. In this study, a dithiobis (succinimidyl undecanoate, DSU)/amine-functionalized reduced graphene oxide-–polyamidoamine dendrimer (DSU/amine-functionalized rGO–PAMAM) thin film-based surface plasmon resonance (SPR) sensor was developed for the detection of DENV 2 E-proteins. Different concentrations of DENV 2 E-proteins were successfully tested by the developed SPR sensor-based system. The performance of the developed sensor showed increased shift in the SPR angle, narrow full-width–half-maximum of the SPR curve, high detection accuracy, excellent figure of merit and signal-to-noise ratio, good sensitivity values in the range of 0.08–0.5 pM (S = 0.2576°/pM, R2 = 0.92), and a high equilibrium association constant (KA) of 7.6452 TM−1. The developed sensor also showed a sensitive and selective response towards DENV 2 E-proteins compared to DENV 1 E-proteins and ZIKV (Zika virus) E-proteins. Overall, it was concluded that the Au/DSU/amine-functionalized rGO–PAMAM thin film-based SPR sensor has potential to serve as a rapid clinical diagnostic tool for DENV infection.


Introduction
The demand for rapid, sensitive, and quantifiable methods for the detection of dengue virus is significant. Optical sensors have emerged as a promising technology to potentially identify biological phenomena. They have the advantages of being highly sensitivity, allowing fast detection and real-time measurements, and having a simple configuration [1][2][3][4][5][6][7]. The main advantage of optical sensors is the possibility of performing label-free quantitative detection, which allows direct binding without fluorescent labels or isotope labelling [8][9][10][11][12][13][14]. Recently, two well-known optical sensors have been

Reagents
DSU (MW = 628.84 g/mol) was purchased from Dojindo Japan. Graphene oxide was purchased from Graphanea, Spain. PAMAM dendrimer (ethylenediamine core, generation 4.0 solution in methanol), ethylenediamine (EDA), N-hydroxysuccinimide (NHS), bovine serum albumin (BSA) were purchased from Sigma Aldrich, Germany. N-Ethyl-N-(3-(dimethylaminopropyl) carbodiimide (EDC) was bought from Fluka, Switzerland. The standard powder of recombinant DENV 2 E-proteins and antibodies (IgM) against dengue type 2 envelope proteins were purchased from Meridian Life Science. The preparation of amine-functionalized rGO was begun by amalgamating the graphene oxide with EDC for 5 min, followed by the addition of EDA. After the solution was vigorously stirred, the suspension was slowly dissolved and started to change its color from brown to dark black. Then, the suspension was washed for 5 times with ethanol and centrifuged to remove excessive EDA and EDC and then was allowed to dry in an oven at 60 • C for at least 1 h. Afterward, the PAMAM solution with the desired concentration was mixed to the amine-functionalized rGO to obtain a composite. Unless otherwise indicated, all antibodies and antigen solutions were diluted in 10 mM phosphate-buffered saline (PBS) at pH 7.4. All chemicals were of reagent or higher grade, and deionized water was used throughout the experiments.

Sensor Surface Fabrication
To produce the best SPR peak for sensor application, a metal deposition technique to obtain a thin Au film (~50 nm) with high thickness tolerance is required. The generally used methods for metal thin-film deposition are electroplating, sputtering, and evaporation (such as vacuum thermal, electron beam, laser beam, and ion-plating evaporation) [54][55][56]. Sputter coating was chosen in this study because it achieves uniform thickness and high film quality, meeting the thickness tolerance requirement for the SPR experiment. It is also an easy-to-use, environment-friendly, and relatively low-cost technique. Figure 1 shows the preparation procedure of the proposed sensor film. To begin, a glass substrate (Menzel glass, 24 mm × 24 mm) was coated with an Au film (~50 nm) using a sputter coater, with a deposition time of 67 s and a current of 20 mA. The prepared Au film was then rinsed thoroughly using deionized water and then ethanol and dried in a nitrogen flow. Afterward, the Au film was self-assembled in 2 mM DSU solution for 24 h to allow chemisorption of biomolecules through amide linkages. After rinsing with acetone and PBS, the gold substrate was incubated with amine-functionalized rGO-PAMAM for 30 min, followed by spinning. The amine-functionalized rGO-PAMAM-modified film was cross-linked with EDC/NHS for 30 min and spun at 6000 rpm for 30 s. After cross-linking, 0.01 µM antibodies specific to DENV 2 E-proteins was incubated over the surface for 30 min. This sensor film will be referred to as DSU/amine-functionalized rGO-PAMAM/IgM in the following discussion.

DENV E-Proteins Detection
For the detection of DENV 2 E-proteins, the sensor film was mounted on a prism with a refractiveindex-matching liquid and loaded onto the SPR rotating stage for SPR measurements. Figure 2 shows the schematic diagram of our custom-made SPR experimental set up. In the beginning, 1 mL of PBS solution was injected into the measuring cell to obtain the baseline signal. Afterward, 1 mL of 0.08 pM DENV 2 E-proteins solution was injected into the measuring cell while monitoring the SPR response for 30 min. A triplicate SPR response was recorded for each concentration of DENV 2 E-proteins solution (0.08-1 pM) by using new sensor films. The changes of the SPR shift for each concentration of DENV 2 E-proteins were calculated by measuring the difference between the resonance angle of the sample and the PBS solution (as shown in Figure 2). The selectivity of the proposed sensor was also evaluated by comparing its SPR response to different concentrations of DENV 2 E-proteins with that to other proteins.

DENV E-Proteins Detection
For the detection of DENV 2 E-proteins, the sensor film was mounted on a prism with a refractive-index-matching liquid and loaded onto the SPR rotating stage for SPR measurements. Figure 2 shows the schematic diagram of our custom-made SPR experimental set up. In the beginning, 1 mL of PBS solution was injected into the measuring cell to obtain the baseline signal. Afterward, 1 mL of 0.08 pM DENV 2 E-proteins solution was injected into the measuring cell while monitoring the SPR response for 30 min. A triplicate SPR response was recorded for each concentration of DENV 2 E-proteins solution (0.08-1 pM) by using new sensor films. The changes of the SPR shift for each concentration of DENV 2 E-proteins were calculated by measuring the difference between the resonance angle of the sample and the PBS solution (as shown in Figure 2). The selectivity of the proposed sensor was also evaluated by comparing its SPR response to different concentrations of DENV 2 E-proteins with that to other proteins.

DENV E-Proteins Detection
For the detection of DENV 2 E-proteins, the sensor film was mounted on a prism with a refractive-index-matching liquid and loaded onto the SPR rotating stage for SPR measurements. Figure 2 shows the schematic diagram of our custom-made SPR experimental set up. In the beginning, 1 mL of PBS solution was injected into the measuring cell to obtain the baseline signal. Afterward, 1 mL of 0.08 pM DENV 2 E-proteins solution was injected into the measuring cell while monitoring the SPR response for 30 min. A triplicate SPR response was recorded for each concentration of DENV 2 E-proteins solution (0.08-1 pM) by using new sensor films. The changes of the SPR shift for each concentration of DENV 2 E-proteins were calculated by measuring the difference between the resonance angle of the sample and the PBS solution (as shown in Figure 2). The selectivity of the proposed sensor was also evaluated by comparing its SPR response to different concentrations of DENV 2 E-proteins with that to other proteins.

SPR Performances
There are several parameters, i.e., sensitivity, full width at half minimum (FWHM), detection accuracy (DA), signal-to-noise ratio (SNR), figure of merit (FOM), selectivity, and spike recovery that are used to analyze the performance of SPR sensors. The sensitivity of an SPR sensor can be defined by calculating the slope in the graph of the resonance angle shift, ∆θ, versus DENV 2 E-protein concentrations, as follows: A second parameter is the angular width of the SPR curve for the half value of the maximum reflectance, which is known as the FWH). The value of FWHM should be as small as possible in order to get an accurate reading on the dip of the SPR curve [57]. Therefore, the DA of the SPR sensor is y inversely proportional to the value of FWHM.
To comprehensively evaluate the performance of the SPR signal, the SNR can be determined according to Finally, the FOM can be defined as the ratio between the spectral sensitivity and the width of the SPR curve and can be written as

Sensor Surface Characterization
The surface morphology and roughness of the sensor films were imaged using an atomic force microscope (AFM, Bruker AFM multimode 8, Santa Barbara, CA, USA) in Scan Asyst mode. This measurement was performed in ambient atmosphere at room temperature. Figure 3 shows the SPR responses of the proposed sensor film for the detection of DENV 2 E-proteins. To verify the viability of the sensor film, various concentrations of DENV 2 E-proteins in the range of 0.08-1 pM were injected into the cell. The results showed that the resonance angle for the reference solution (0 pM), i.e., the PBS solution, was 54.2138 • . When the proposed sensor was exposed to 0.08 pM DENV 2 E-proteins, the resonance angle of the reflected light increased to 54.3052 • . Subsequently, for DENV 2 E-proteins concentrations of 0.1 pM, 0.3 pM, 0.5 pM, and 1 pM, the resonance angles from the SPR curves were found to be 54.3137 • , 54.3925 • , 54.4004 • , and 54.4083 • , respectively. To measure the amount of antigen bound to the sensor surface, the resonance angle shift (∆θ) was determined from the difference between the resonance angle of the antigen and the resonance angle of the reference solution. It was found that increases of ∆θ of 0.0914 • , 0.0999 • , 0.1708 • , 0.1866 • , and 0.1945 • were obtained when detecting 0.08 pM, 0.1 pM, 0.3 pM, 0.5 pM, and 1 pM DENV 2 E-proteins, respectively. These ∆θ can be attributed to the changes in the refractive index of the sensor surface which in turn changed the real part of the dielectric constant of the gold film caused by the binding of DENV 2 E-proteins. It was inferred that a change of the thickness of the sensing layer would also result in a slight angle shift of SPR, as the evanescent wave possesses longer penetration depths [58,59]. The inset of Figure 3 shows how ∆θ varied with time when detecting different concentrations of DENV 2 E-proteins. It can be seen that the proposed sensor achieved a stable detection time of 8 min for the lowest concentration of DENV 2 E-proteins. Therefore, all DENV E-proteins concentrations were left for 8 min before the SPR curve was recorded. Figure 4 reports the evolution of FWHM and DA as a function of the concentration of DENV 2 E-proteins. As expected, the most extensive broadening of the SPR curves was observed when the proposed sensor surface was exposed to the highest concentration of DENV 2 E-proteins. This phenomenon was due to the increasing refractive index on the sensor surface caused by the kinetics of the binding between ligand and analyte [60]. The increase in surface roughness also led to broadening of the SPR curves as the coupling rate of SPR became poor. On the other hand, the resolution of the SPR signal decreased due to the increase in electron energy loss in the dense sensing layer. The minimum value of FWHM was achieved with the first introduction of DENV 2 E-proteins at the concentration of 0.08 pM. A more distinct and narrower SPR curve allows for higher accuracy of SPR sensors throughout measurements. Thus, the highest DA of 0.3964 degree −1 was achieved in the detection of 0.00008 nM DENV E-proteins. Nanomaterials 2020, 10, x FOR PEER REVIEW 6 of 14  Figure 4 reports the evolution of FWHM and DA as a function of the concentration of DENV 2 E-proteins. As expected, the most extensive broadening of the SPR curves was observed when the proposed sensor surface was exposed to the highest concentration of DENV 2 E-proteins. This phenomenon was due to the increasing refractive index on the sensor surface caused by the kinetics of the binding between ligand and analyte [60]. The increase in surface roughness also led to broadening of the SPR curves as the coupling rate of SPR became poor. On the other hand, the resolution of the SPR signal decreased due to the increase in electron energy loss in the dense sensing layer. The minimum value of FWHM was achieved with the first introduction of DENV 2 E-proteins at the concentration of 0.08 pM. A more distinct and narrower SPR curve allows for higher accuracy of SPR sensors throughout measurements. Thus, the highest DA of 0.3964 degree −1 was achieved in the detection of 0.00008 nM DENV E-proteins. Figure 5 depicts the variation of SNR and FOM with the concentration of DENV 2 E-proteins. Expectedly, the SNR increased with the increase of the concentration of DENV 2 E-proteins. This indicated that a high concentration of DENV 2 E-proteins led to a lower noise of the signal, thus providing the maximum values of SNR. This effect was due to the stronger coupling of surface plasmon, which led to a stronger radiation damping of the surface [61,62]. In addition, the variation in SNR was almost similar to the binding kinetic analysis, as it is dependent on the resonance angle shift. Beyond that, a high FOM is required for a better quality of the SPR sensor. Apparently, a higher FOM was observed when a 0.08 pM DENV 2 E-proteins solution was introduced. The reason for the higher value of FOM at this concentration could be the smaller dissipation in metal upon the excitation of surface plasmon at a shorter wavelength, resulting in a narrow full width at half maximum of the SPR curve. Consequently, the FOM decreased with a further increase of the concentration of DENV 2 E-proteins . The decreasing of the FOM could be due to a faster decrease in the sensitivity compared to FWHM, which resulted in a notable uncertainty in determining the position of the resonance angle.     Figure 5 depicts the variation of SNR and FOM with the concentration of DENV 2 E-proteins. Expectedly, the SNR increased with the increase of the concentration of DENV 2 E-proteins. This indicated that a high concentration of DENV 2 E-proteins led to a lower noise of the signal, thus providing the maximum values of SNR. This effect was due to the stronger coupling of surface plasmon, which led to a stronger radiation damping of the surface [61,62]. In addition, the variation in SNR was almost similar to the binding kinetic analysis, as it is dependent on the resonance angle shift. Beyond that, a high FOM is required for a better quality of the SPR sensor. Apparently, a higher FOM was observed when a 0.08 pM DENV 2 E-proteins solution was introduced. The reason for the higher value of FOM at this concentration could be the smaller dissipation in metal upon the excitation of surface plasmon at a shorter wavelength, resulting in a narrow full width at half maximum of the SPR curve. Consequently, the FOM decreased with a further increase of the concentration of DENV 2 E-proteins . The decreasing of the FOM could be due to a faster decrease in the sensitivity compared to FWHM, which resulted in a notable uncertainty in determining the position of the resonance angle.  To investigate the binding between Au/DSU/amine-functionalized rGO-PAMAM/IgM-based SPR sensor and target, the changes in the resonance angle as a function of the concentration of DENV 2 E-proteins from 0.08 to 1 pM were plotted and fitted according to the Langmuir adsorption model (Figure 6), represented by the following equation [63,64]: To investigate the binding between Au/DSU/amine-functionalized rGO-PAMAM/IgM-based SPR sensor and target, the changes in the resonance angle as a function of the concentration of DENV 2 E-proteins from 0.08 to 1 pM were plotted and fitted according to the Langmuir adsorption model (Figure 6), represented by the following equation [63,64]:

Results
where ∆θ max is the maximum SPR shift at saturation, C is the concentration of DENV, and K D is the equilibrium dissociation constant. Typically, the strength of biomolecular interactions is quantified through K D (reciprocal of K A ). The fitted graph gave a K D of 0.1308 pM with R 2 of 0.97, which was found to be consistent with the standard K D value for protein interactions (K D < 10 nM) [65,66]. A smaller K D value led to higher binding affinity of the ligand for its target. The apparent K A values for the binding of DENV 2 E-proteins to the sensor surface was 7.6452 TM −1 , indicating the high potential for the use of the proposed SPR sensor in detecting DENV 2 E-proteins. In order to determine the sensitivity of the proposed sensor film for DENV 2 E-proteins, a linear correlation curve was plotted and is presented in Figure 7. The results were fitted linearly with a correlation coefficient, R 2 , of 0.92. The slope of the linear fit was considered to be the sensitivity value of the proposed SPR sensor, thus yielding a sensitivity of 0.2576 • /pM. It appears that DENV 2 E-proteins can be sensitively detected at the lowest concentration of 0.08 pM. This is possible thanks to the great penetration depth of the SPR evanescent field along the DSU/amine-functionalized rGO-PAMAM/IgM sensor layer. correlation curve was plotted and is presented in Figure 7. The results were fitted linearly with a correlation coefficient, R 2 , of 0.92. The slope of the linear fit was considered to be the sensitivity value of the proposed SPR sensor, thus yielding a sensitivity of 0.2576°/pM. It appears that DENV 2 Eproteins can be sensitively detected at the lowest concentration of 0.08 pM. This is possible thanks to the great penetration depth of the SPR evanescent field along the DSU/amine-functionalized rGO-PAMAM/IgM sensor layer.  Table 1 presents the performance of other techniques developed for DENV detection, in terms of sensitivity and detection limit [67][68][69][70][71]. From the table, it can be seen that the sensitivity of these techniques has not been extensively evaluated compared to their detection limit. Therefore, with respect to the detection limit, the lowest DENV concentration measured by our SPR sensor was 0.08 pM. The performance of our SPR sensor is highly dependent on its surface modification, which can cause significant changes in surface mass due to binding events, thus improving its sensing performance.   Table 1 presents the performance of other techniques developed for DENV detection, in terms of sensitivity and detection limit [67][68][69][70][71]. From the table, it can be seen that the sensitivity of these techniques has not been extensively evaluated compared to their detection limit. Therefore, with respect to the detection limit, the lowest DENV concentration measured by our SPR sensor was 0.08 pM. The performance of our SPR sensor is highly dependent on its surface modification, which can cause significant changes in surface mass due to binding events, thus improving its sensing performance. To determine the selectivity of the Au/DSU/amine-functionalized rGO-PAMAM/IgM sensor film towards DENV 2 E-proteins, other envelope proteins, such as DENV 1 E-proteins and ZIKV E-proteins, at the concentration of 10 pM were tested. An excellent sensor should be highly selective in the detection of DENV 2 E-proteins. The results are shown in Figure 8. Clearly, the highest SPR signal was obtained for the detection of DENV 2 E-proteins, even of their concentration was much lower than those of DENV 1 E-proteins and ZIKV E-proteins. This demonstrated that the proposed sensor is highly selective for the detection of DENV 2 E-proteins. AFM images were taken before ( Figure 9a) and after the introduction of DENV 2 E-proteins (Figure 9b) onto the surface of the Au/DSU/amine-functionalized rGO-PAMAM/IgM sensor. A crumpled and wrinkled morphology of the rGO-PAMAM structure is observed in both images. It can be observed that when DENV 2 E-proteins were present onto the sensor surface, the surface structure became less crumpled and wrinkled. The surface roughness of the sensor films increased from 1.56 to 3.19 nm upon the introduction of DENV 2 E-proteins. The change in surface roughness proves that the attachment of DENV 2 E-proteins changed the surface morphology of the sensing layer. AFM images were taken before ( Figure 9a) and after the introduction of DENV 2 E-proteins (Figure 9b) onto the surface of the Au/DSU/amine-functionalized rGO-PAMAM/IgM sensor. A crumpled and wrinkled morphology of the rGO-PAMAM structure is observed in both images. It can be observed that when DENV 2 E-proteins were present onto the sensor surface, the surface structure became less crumpled and wrinkled. The surface roughness of the sensor films increased from 1.56 to 3.19 nm upon the introduction of DENV 2 E-proteins. The change in surface roughness proves that the attachment of DENV 2 E-proteins changed the surface morphology of the sensing layer.
crumpled and wrinkled morphology of the rGO-PAMAM structure is observed in both images. It can be observed that when DENV 2 E-proteins were present onto the sensor surface, the surface structure became less crumpled and wrinkled. The surface roughness of the sensor films increased from 1.56 to 3.19 nm upon the introduction of DENV 2 E-proteins. The change in surface roughness proves that the attachment of DENV 2 E-proteins changed the surface morphology of the sensing layer.

Conclusions
In this study, a DSU/amine-functionalized rGO-PAMAM thin film-based SPR sensor was developed for the detection and quantification of DENV 2 E-proteins. The results demonstrate that

Conclusions
In this study, a DSU/amine-functionalized rGO-PAMAM thin film-based SPR sensor was developed for the detection and quantification of DENV 2 E-proteins. The results demonstrate that the developed sensor could successfully monitor the changes in the SPR angle when detecting the lowest concentration of DENV 2 E-proteins, i.e., 0.08 pM, within 8 min. The sensitivity and binding affinity constant of the developed sensor were 0.2576 • /pM and 7.6452 TM −1. The highest FOM and DA values were obtained when detecting 0.08 pM of DENV 2 E-proteins and corresponded to the minimum values of SNR. The selectivity of the proposed sensor was demonstrated, since the SPR response to other proteins was reduced compared to that to DENV 2 E-proteins. It can be concluded that the developed SPR sensor has a great potentiality for rapid DENV diagnostic analyses or point-of-care tests.