Sensitive Detection of Dengue Virus Type 2 E-Proteins Signals Using Self-Assembled Monolayers/Reduced Graphene Oxide-PAMAM Dendrimer Thin Film-SPR Optical Sensor

In this work, sensitive detection of dengue virus type 2 E-proteins (DENV-2 E-proteins) was performed in the range of 0.08 pM to 0.5 pM. The successful DENV detection at very low concentration is a matter of concern for targeting the early detection after the onset of dengue symptoms. Here, we developed a SPR sensor based on self-assembled monolayer/reduced graphene oxide-polyamidoamine dendrimer (SAM/NH2rGO/PAMAM) thin film to detect DENV-2 E-proteins. Surface characterizations involving X-ray diffraction (XRD) and Fourier-transform infrared spectroscopy (FTIR) confirms the incorporation of NH2rGO-PAMAM nanoparticles in the prepared sensor films. The specificity, sensitivity, binding affinity, and selectivity of the SPR sensor were then evaluated. Results indicated that the variation of the sensing layer due to different spin speed, time incubation, and concentration provided a better interaction between the analyte and sensing layer. The linear dependence of the SPR sensor showed good linearity (R2 = 0.92) with the lowest detection of 0.08 pM DENV-2 E-proteins. By using the Langmuir model, the equilibrium association constant was obtained at very high value of 6.6844 TM−1 (R2 = 0.99). High selectivity of the SPR sensor towards DENV-2 E-proteins was achieved in the presence of other competitors.

www.nature.com/scientificreports www.nature.com/scientificreports/ Materials characterization. XRD patterns of the synthesized NH 2 rGO, and NH 2 rGO-PAMAM composite were recorded on X-ray Diffractometer (Philips X'Pert X-ray) with the Cu Kα radiation. The surface functionalization of the proposed sensor film was then confirmed using Fourier transform infrared (FTIR) spectrometer (VERTEX 70) in the wavenumber range of 4000-400 cm −1 .
Surface plasmon resonance (SPR) sensing. SPR measurements were conducted based on Kretschmann configuration [65][66][67][68] by evaporating Au/DSU/NH 2 rGO-PAMAM/IgM sensor film onto the prism surface. The prism was placed on an optical stage driven by a stepper motor with a resolution of 0.001° (Newport MM 3000) to let the incident light from laser beam (632.8 nm, 5 mW) pass through the prism and hits a gold layer to generate the surface plasmon waves at the interface. At a specific angle of the incident light, the SPR response was induced when the evanescent wave is generated due to the change in the refractive index of the medium in close to the vicinity of a gold surface. The SPR response is the reflected light intensity, at which its angle at minimum intensity was recorded with time. A 100 µl flow cell was attached to the sensor film to be filled up by DENV-2 E-proteins solution for the detection system. All experiments were conducted at room temperature and replicated three times with a new set of sensor surface for each concentration of DENV-2 E-proteins (0.08 pM -0.5 pM). Figure 1 is a schematic illustration representing the proposed Au/DSU/NH 2 rGO-PAMAM/IgM sensor with the introduction of DENV-2 E-proteins.

Results and Discussion
Materials characterization. Figure 2 depicts the XRD patterns of the synthesized NH 2 rGO, and NH 2 rGO-PAMAM composite. The formation of the synthesized NH 2 rGO was confirmed by a broad diffraction peak at 2θ = 26.35 and a small hump at 2θ = 40.17°, which stand for the (002) and (100) crystal plane of NH 2 rGO [69][70][71][72] . After being composited with PAMAM, the amorphous nature of NH 2 rGO was restored and slightly shifted to the left. This shift can be associated to the changes in lattice parameters caused by the covalent binding between PAMAM and NH 2 rGO.
Characterization was continued using FTIR spectroscopy to examine the chemical structure of the Au/DSU/ NH 2 rGO-PAMAM/IgM thin film before and after antigen conjugation, as shown in Fig. 3. From the FTIR spectra, the absorption band of Au-S at 650-720 cm −1 confirms the adsorption of a self-assembled monolayer on a gold film. The peaks at 1570 cm −1 and 1631 cm −1 is related to the N-H bending (amide II) and C-O stretching (amide I) vibration of PAMAM dendrimer, respectively 73 . Three peaks at 3448, 2912, and ~1400 cm −1 are observed that can be indicated to the stretching of adsorbed water molecules and structural O-H groups in graphene oxide. In the case of rGO with amine functionalized, the bandwidth at 3200-3400 cm −1 belonged to the N-H stretching, which overlapped with O-H bonds 74 . Another overlapped peak observed at 1550 cm −1 and a small peak at 1150 cm −1 were attributed to N-H bending and C-N stretching, respectively, hence confirms the amine functionalization of GO followed by reduction to amino-rGO 59 . After covalent bonding with the antibodies via EDC/NHS, a small peak of C=O stretching could be observed at 1716 cm −1 , with regard to certifying the amide bonding 75,76 . This proved that the bioconjugation proceeded successfully. Furthermore, with the introduction of DENV-2 E-proteins into the sensor film, the reduction in the intensity of amide bands and O-H band can be observed quantitatively, indicate the structural changes of the sensor surface due to binding. This finding thus validates the immunoreaction between antibodies immobilized on the sensor surface with the DENV-2 E-proteins has taken place. www.nature.com/scientificreports www.nature.com/scientificreports/ Optimization of SPR sensing layers. The excellent performance of SPR sensor lies in its sensor surface functionalization, which offers a significant binding towards DENV-2 E-proteins. Significant binding of antigens results in a change in the refractive index of the medium in close to the vicinity of a gold surface, which in turn is measured as a shift in the SPR angle 77,78 . To achieve the best SPR enhancement performance, we plotted the shift in resonance angle as a function of various fabricating material-coated gold films as shown in Fig. 4(a,b), i.e. DSU/PAMAM/IgM, DSU/NH 2 rGO/IgM, NH 2 rGO-PAMAM/IgM, DSU(24 h)/NH 2 rGO-PAMAM/IgM, DSU(48 h)/NH 2 rGO-PAMAM/IgM, and DSU(72 h)/NH 2 rGO-PAMAM/IgM. As shown in Fig. 4(c), the binding between DENV-2 E-proteins and sensing surface of DSU(24 h)/NH 2 rGO-PAMAM/IgM resulted in an obvious shift in the SPR resonance angle. Furthermore, when comparing to the last three sensing surfaces which differ in self-assembly time optimization, the assembly time of 24 h gives the highest shift in the resonance angle. This is because the longer incubation time than 24 led to the desorption of SAM structures. This indicates that DSU(24 h)/NH 2 rGO-PAMAM/IgM sensing layers can serve as the best dengue sensing medium in SPR technique, thus offering a significant enhancement in the penetration depth of evanescent waves.
Effect of PAMAM concentrations. In order to establish the optimal condition for the sensing layer, we varied the concentrations of PAMAM dendrimer (Fig. 5). From the SPR experiments, the shift in the  www.nature.com/scientificreports www.nature.com/scientificreports/ resonance angle increased dramatically with PAMAM concentration of 10 mM. The higher value in Δθ obtained for 10 mM PAMAM dendrimer showed significant improvement in the detection of DENV-2 E-proteins. Perhaps, the effect of a higher concentration of PAMAM dendrimer can be attributed to the higher activity of dendrimer-encapsulated reduced graphene oxide. As a result, the presence of dense concentration and globular shape of dendrimer might have provided better interaction between antigen and sensing layer 79 . Therefore, 10 mM PAMAM dendrimer was established as optimal concentration, which is in agreement with those reported in the literature 80 .

Effect of sputtering time and spin speed.
In order to achieve a relatively high shift in resonance angle and low width of the resonance curve, we optimized the sputtering time for gold layer and spin speed of NH 2 rGO-PAMAM layer (Fig. 6). The important of width (i.e., FWHM) lies in determining the resonance angle accurately 81,82 . From Table 1, it is noted that as the sputtering time for gold was fixed at 67 s with the increasing spin speed of deposition, the reflectance curve is redshifted and narrower. The increases in resonance angle indicate that the sensor layer tends to absorb the biomolecules as they become thinner. In addition, the interaction between evanescent field and sensing medium has resulted in the deeper penetration depth of the field in the biomolecular analyte layer. In this case, the propagation constant (wavevector) of surface plasmons will be enhanced due to smaller electron energy loss. Comparing with the sputtering time of 65 s and 63 s, the smallest FWHM was achieved at 63 s. The reason behind this is might be due to the less scattering near surface plasmon, thus has selective detection towards DENV-2 E-proteins. Regardless of the narrower FWHM at sputtering time of 63 s, the optimum sputtering time for the gold layer and spin speed of NH 2 rGO-PAMAM layer were 67 s and 8000 rpm, www.nature.com/scientificreports www.nature.com/scientificreports/ respectively. A significant increase in FWHM could be due to the intensified internal loss as resulted from the higher binding of target antigen 83,84 . On the other hand, the average thickness of the gold layer at 67 s was found to be ~48 nm, confirmed by AFM analysis as shown in Fig. 7 (see vertical distance). It can also be mentioned the obtained thickness is nearer to the most typical configuration in the SPR system with a 50 nm gold film, which is responsive to the local gold-dielectric interface 85-87 . Comparison of SPR signals for DENV-2 E-proteins detection. Figure 8a-d display the variation of SPR signal on modified SPR sensor surfaces generated by the introduction of 100 pM DENV E-proteins. As observed in Fig. 8a, the first bare gold film shows the resonance angle shifts from 53.6554° to 53.6642° upon the introduction of target antigens. Further, the resonance angle was blue-shifted after the bare gold film was self-assembled with the amine-reactive sites of DSU (Fig. 8b). The remarkable blueshift of the resonance curve www.nature.com/scientificreports www.nature.com/scientificreports/ after injection of DENV solution can be explained by the dissociation of the target antigen towards the sensor surface. As the composite layer of NH 2 rGO-PAMAM was developed onto the sensor surface (Fig. 8c), a decreasing shift in resonance angle was observed. This leaves the sensor surface requires specific biomolecules for selective detection of DENV solution. When a solution containing antibodies specific to DENV E-proteins is immobilized on the sensor surface (Fig. 8d), the shift in resonance angle is greatly enhanced to 0.2577° which results in a significant change in SPR sensitivity. Therefore, complete development of sensor film, Au/DSU/NH 2 rGO-PAMAM/ IgM is responsible for selective and sensitive detection of target antigens as they have a great change in SPR angle. Figure 9 shows the SPR real-time measurement of the proposed Au/DSU/NH 2 rGO-PAMAM/IgM sensor film for detection of dengue virus. To verify the viability of the sensor film, different concentrations of DENV-2 E-proteins in the range of 0.08-0.5 pM were injected accordingly  www.nature.com/scientificreports www.nature.com/scientificreports/ into the cell. The proposed sensor had a complete response time at approximately 10 min for detection of higher DENV-2 E-proteins concentration, while 6-8 min for lowest concentration of DENV-2 E-proteins, 0.08 pM. It indicates that an increase in concentration caused an increase in real-time detection of DENV-2 E-proteins,  www.nature.com/scientificreports www.nature.com/scientificreports/ which might be due to saturation of all binding sites. Due to that, all DENV-2 E-proteins concentrations were left for 8 min before the SPR response was taken. Figure 10 shows the SPR responses of the proposed sensor film for detection of DENV-2 E-proteins. The results showed that the resonance angle for PBS solution was 54.2138°. When the proposed sensor was exposed to the lowest concentration of DENV-2 E-proteins (0.08 pM), the resonance angle of the reflected light increased to 54.3052°. Subsequently, the resonance angles from SPR curves were found to be 54.3137°, 54.3925°, and 54.4004° with further addition of DENV-2 E-proteins concentrations of 0.1 pM, 0.3 pM, and 0.5 pM, respectively. To measure the number of antigens bound to sensor surfaces, the resonance angle shift (Δθ) was taken from the difference in resonance angle of antigen and resonance angle of the reference solution. It was found that the rise in Δθ of 0.0914°, 0.0999°, 0.1708°, and 0.1866° were obtained for detection of 0.08 pM, 0.1 pM, 0.3 pM, and 0.5 pM, of DENV-2 E-proteins, respectively. These Δθ can be attributed to the changes in the refractive index of the sensor surface which in turn changes 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 also brings out a slight angle shift of SPR as the evanescent wave possesses longer penetration depths 88,89 . Sensitivity and binding affinity of DSU/NH 2 rGO-PAMAM/IgM sensor film. Prior to the sensitivity measurement of Au/DSU/NH 2 rGO-PAMAM/IgM sensor film, a control experiment was performed using  www.nature.com/scientificreports www.nature.com/scientificreports/ Au/IgM. For this purpose, different concentrations of DENV-2 E-proteins ranging from 0.08-0.5 pM were injected onto Au/IgM surface. Table 2 depicts the obtained resonance angle shift in a triplicate manner. It was observed that there were no variations in the SPR resonance angle with average standard deviation value of ±0.0004. Further, the linear regression analysis for Au/DSU/NH 2 rGO-PAMAM/IgM sensor film was plotted as shown in Fig. 11, yielded y = 0.25762x + 0.07492; R 2 = 0.92 with average standard deviation value of ±0.0044. From a comparison of the gradient for Au/IgM and Au/DSU/NH 2 rGO-PAMAM/IgM, the gradient of Au/DSU/ NH 2 rGO-PAMAM/IgM reveals higher sensitivity than Au/IgM. The results indicated that DENV-2 E-proteins can be sensitively detected at the lowest concentration of 0.08 pM using Au/DSU/NH 2 rGO-PAMAM/IgM based SPR sensor.

SPR reflectivity with real-time measurement.
To extract the binding strength involved in analyte-ligand interactions, the data of SPR angle shifts (Δθ) and DENV-2 E-proteins concentrations were fitted using Langmuir isotherm model (Fig. 12). The equation of this model is represented by 90,91 where Δθ max is the maximum SPR shift at the saturation, C is the concentration of DENV and K D is the equilibrium dissociation constant. The K D for the assessment of the DENV-2 E-proteins towards Au/IgM and Au/DSU/ NH 2 rGO-PAMAM/IgM were then calculated and found to be 1.1306 pM; R 2 = −0.5 and 0.1496 pM; R 2 = 0.99, respectively. The obtained K D values are found to be consistent with the standard K D value for protein interaction (K D < 10 nM) 92,93 . The smaller K D value revealed that the Au/DSU/NH 2 rGO-PAMAM/IgM sensor film has higher affinity (K A ) interaction towards DENV E-proteins compared with Au/IgM sensor film. The K A values for Au/IgM and Au/DSU/NH 2 rGO-PAMAM/IgM sensor films were then calculated to be 0.8844 TM −1 and 6.6844 TM −1 , respectively. The cause of these changes was significantly due to differences in ligand density and stability, which can potentially affect the evanescent field distribution at the gold interface, thus affect the quantitative detection  www.nature.com/scientificreports www.nature.com/scientificreports/ of the target analyte. It thus validated that the integration of DSU/NH 2 rGO-PAMAM/IgM sensor layer into SPR gold film is important to improve the detection of DENV-2 E-proteins. Table 3 presents the performance comparison of other reported SPR-based DENV sensor in terms of sensitivity (S), dissociation constant (K D ), and detection limit. The proposed SPR sensor in this work shows excellent DENV detection performance including the highest sensitivity and binding affinity, lowest detection limit, and faster response times within 8 minutes. It is hypothesized that modification of the SPR-gold thin film would have an influence on the surface mass associated with a binding event and consequently cause a significant change in the SPR signal.
Selectivity study. The selectivity of the proposed SPR sensor film toward DENV-2 E-proteins was also investigated relative to other potentially competitive proteins such as HSA and DENV-1 E-proteins (Fig. 13).
The tests revealed that the sensor response to 0.1 pM DENV-2 E-proteins was higher than of other proteins. One can conclude that the Au/DSU/NH 2 rGO-PAMAM/IgM-based SPR sensor film has good selectivity to DENV-2 E-proteins. This remarkable selectivity may be due to the high affinity between DENV-2 E-proteins and its specific antibodies immobilized on the sensor surface. It was later verified that the NH 2 rGO-PAMAM sensor layer is developed to increase the adsorption of antibodies to provide more active sites for attachment of DENV-2 E-proteins. In the case of HSA proteins and DENV-1 E-proteins, the results indicated a least SPR response due to the non-specific antibody binding. The least response from DENV-1 E-proteins likely reflects the successful interactions of similar genome, which shares 65% of single-stranded RNA genomes encoded by other DENV serotypes 94,95 . As for HSA proteins, high SPR response can be accounted as an excessive proteins in the blood with a molecular weight of 66.4 kDa when compared to 50 kDa DENV-1 E-proteins 96,97 . Figure 14 depicts the effect of diverse analyte solutions on the selective detection of DENV-2 E-proteins using SPR sensor. The concentrations of each analyte were fixed at 10 pM. The multiple solutions that having DENV-2 E-proteins showed larger SPR signal compared to other solutions containing no DENV-2 E-proteins. The results suggest that the interference by other analyte solutions does not affect the quantitative detection of DENV-2 E-proteins.  Table 3. Performance comparison of the previously reported using SPR-based DENV sensor.