Sandwich Electrochemical Immunosensor for Early Detection of Tuberculosis Based on Graphene/Polyaniline-Modified Screen-Printed Gold Electrode

A rapid and sensitive sandwich electrochemical immunosensor was developed based on the fabrication of the graphene/polyaniline (GP/PANI) nanocomposite onto screen-printed gold electrode (SPGE) for detection of tuberculosis biomarker 10-kDa culture filtrate protein (CFP10). The prepared GP/PANI nanocomposite was characterized using Fourier transform infrared spectroscopy (FTIR) and field emission scanning electron microscopy (FESEM). The chemical bonding and morphology of GP/PANI-modified SPGE were studied by Raman spectroscopy and FESEM coupled with energy dispersive X-ray spectroscopy, respectively. From both studies, it clearly showed that GP/PANI was successfully coated onto SPGE through drop cast technique. Cyclic voltammetry was used to study the electrochemical properties of the modified electrode. The effective surface area for GP/PANI-modified SPGE was enhanced about five times compared with bare SPGE. Differential pulse voltammetry was used to detect the CFP10 antigen. The GP/PANI-modified SPGE that was fortified with sandwich type immunosensor exhibited a wide linear range (20–100 ng/mL) with a low detection limit of 15 ng/mL. This proposed electrochemical immunosensor is sensitive, low sample volume, rapid and disposable, which is suitable for tuberculosis detection in real samples.


Introduction
Globally, one of the most critical infectious diseases is tuberculosis (TB), which can be fatal [1]. Mycobacterium tuberculosis (M. tuberculosis) is the causative agent of TB which can be transmitted via minute aerosol droplets such as coughing, sneezing, or even talking by the infected TB person [2]. This'airborne contagious disease poses a particular infection-control challenge since the healthy person developed by immobilizing the capture anti-CFP10 antibodies (CapAb) onto GP/PANI-modified SPGE to capture CFP10 in the sample and iron oxide-gold magnetic nanoparticles (Fe 3 O 4 -Au MNPs) conjugated with primary anti-CFP10 antibodies (Ab) served as a signal probe. We also demonstrated the effectiveness of developed immunoassay for the sensitive and quantitative detection of CFP10 in sputum, collected from human samples of TB infected persons. These studies suggested that the blood test-free method through sputum detection of biomarkers such as CFP10 might be of value in the early detection of TB disease, especially in HIV-positive cases and thus could provide wide potential applications in clinical analysis.

Apparatus
Electrochemical measurements were performed using potentiostat Metrohm @ µAutolab type III (Eco Chemie, Utrecht, The Netherlands) integrated with screen printed junction cable controlled by NOVA 1.11 software. The SPGE was purchased from DropSens (Oviedo, Spain). The diameter of the disk-shaped working electrode was 4 mm. The working electrode and auxiliary electrode were made of gold, whereas the reference electrode was made of silver. The electrodes were all printed on a ceramic support (L 33 × W 10 × H 0.5 mm). All electrochemical measurements were performed at room temperature. Field emission scanning electron microscopy (FESEM) analysis was carried out using an FEI Nova Nanosem 230 microscope equipped with an energy dispersive X-ray (EDX) system. Raman spectroscopy studies were performed with a WITec Alpha 300R microscope while Fourier transform infrared spectroscopy (FTIR) coupled with attenuated total reflectance analysis were executed using Shimadzu.

Synthesis of Iron Oxide Magnetic Nanoparticles Coated with Gold (Fe 3 O 4 /Au MNPs)
Magnetic nanoparticles composed of Fe 3 O 4 and gold nanohybrid material (Fe 3 O 4 /Au MNPs), in particular, were synthesized via co-precipitation method according to previous reported methods with slight modification [25,26] The color of the mixture changed from black to reddish brown gradually. The mixture was heated for another 1 h and then cooled to room temperature. The formed Fe 3 O 4 /Au MNPs were collected by a magnet, followed by centrifuged twice with deionized water and ethanol. Finally, the Fe 3 O 4 /Au MNPs was dried at 70 • C in a vacuum oven for 1 h.
For carboxylation steps, 2-mercaptoethanol and 12-mercaptododecanoic acid (ME-MDDA) was self-assembled on Fe 3 O 4 /Au MNPs surface by the well-known gold-thiol (Au-S) chemistry. Then, 10 mg of the Fe 3 O 4 /Au MNPs were suspended in 1 mM ethanolic ME-MDDA and incubated for 24 h at room temperature. The carboxylated nanoparticles were washed several times with water and dispersed in 0.01 M PBS solution with pH 7.4.

Immobilization of Primary Anti-CFP10 Antibody on Fe 3 O 4 /Au MNPs
Briefly, 1 mL functionalized Fe 3 O 4 /Au MNPs was mixed with 1 mL 25 µg/mL antibody (Ab) solution followed by incubation for 2 h. Subsequently, Fe 3 O 4 /Au-Ab was collected by a magnet and washed three times with 0.01 M PBS solution. Then, Fe 3 O 4 /Au-Ab was re-dispersed in 1 mL of 1% BSA for 2 h and separated again using a magnet. Finally, Fe 3 O 4 /Au-Ab was washed with PBS again and re-dispersed the particles in PBS solution. The particles were stored at 4 • C for further use.

Preparation of GP/PANI Nanocomposite
The method to prepare GP/PANI nanocomposite was done according to an in-situ polymerazation procedure previously reported by Mohamad and group [18]. Briefly, 0.2 M aniline and 3 wt % PMVEA were dissolved in deionized water. Then, the solution was cooled in a refrigerator at 4 • C for 60 min and followed by adding a pre-cooled aqueous solution of 0.2 M APS. Next, 50% weight ratio of GP to aniline was added into the mixture and placed at room temperature at least 6 h in order to complete the whole reaction for polymerization process. The black-green precipitate was filtered and rinsed with methanol, deionized water and acetone for several times. Finally, the GP/PANI nanocomposite product was dried overnight at 40 • C.

Modification of SPGE-Based on Sandwich Immunoassay Format
The development of sandwich immunosensor for CFP10 detection in this study is depicted in Figure 1. Firstly, the electrode was activated with 0.5 M H 2 SO 4 solution using CV between 0.0-1.6 V for 40 cycles, scan rate of 100 mV/s and dried at room temperature. Then, 1 mg of GP/PANI powder was dispersed in 1 mL of 2% APTES solution to prepare GP/PANI solution. Various amount of the GP/PANI solution (0, 4, 5, 6 and 7 µL) was drop casted onto electrode surface and dried overnight at room temperature in order to optimize the GP/PANI thickness on the electrode surface. After washing with ethanol, the electrode was then dried at 70 • C in the oven for 30 min. The CapAb was immobilized onto GP/PANI surface through a cross-linker EDC and NHS [27]. The electrode was then washed with PBS and immersed in 4 mL of 0.25% BSA solution for 1 h to avoid non-specific binding. Next, the modified electrode was incubated with 4 µL of M. tuberculosis CFP10 antigen solution for 1 h. After washing the electrode, 4 µL of the prepared Fe 3 O 4 /Au-Ab buffer solution was drop casted onto the surface and incubated for 40 min. After washing, the electrode was ready for measurement.
All CV measurements were carried out in 1 mM K 3 Fe(CN) 6 (Figure 2b) shows that the GP sheets are mostly covered with PANI nanofibers and formed cauliflower-like structures. It indicates that the presence of GP promotes the formation of agglomerate PANI nanofibers. Figure 2c presents IR spectra of GP nanosheet and GP/PANI nanocomposite. The peaks at 1655 cm −1 correspond to the C-C bond in GP. Compared to spectra of GP/PANI nanocomposite, the broad peaks at 1477 cm −1 indicates the N-H bond was overlapped with C=C aromatic. Besides that, the C-N peak appeared at 1274 cm −1 and 1084 cm −1 . Therefore, we can conclude that the GP was successfully bound with PANI to form GP/PANI nanocomposite.    Figure 2c presents IR spectra of GP nanosheet and GP/PANI nanocomposite. The peaks at 1655 cm −1 correspond to the C-C bond in GP. Compared to spectra of GP/PANI nanocomposite, the broad peaks at 1477 cm −1 indicates the N-H bond was overlapped with C=C aromatic. Besides that, the C-N peak appeared at 1274 cm −1 and 1084 cm −1 . Therefore, we can conclude that the GP was successfully bound with PANI to form GP/PANI nanocomposite.  It indicates that the presence of GP promotes the formation of agglomerate PANI nanofibers. Figure 2c presents IR spectra of GP nanosheet and GP/PANI nanocomposite. The peaks at 1655 cm −1 correspond to the C-C bond in GP. Compared to spectra of GP/PANI nanocomposite, the broad peaks at 1477 cm −1 indicates the N-H bond was overlapped with C=C aromatic. Besides that, the C-N peak appeared at 1274 cm −1 and 1084 cm −1 . Therefore, we can conclude that the GP was successfully bound with PANI to form GP/PANI nanocomposite.

Characterization of GP/PANI-Modified SPGE
GP/PANI nanocomposite material was used to modify bare SPGE in order to enhance the sensor performance. The morphology of GP/PANI-modified SPGE surface was studied by FESEM-EDX. It was observed in Figure 3a,b that the surface structure of GP/PANI-modified electrode turned relatively rough and agglomerate compared to bare SPGE due to the formation of nanoparticle clusters on the electrode surface, which resulted in an increase of active surface area [28,29]. The presence of GP/PANI nanocomposite was also confirmed using EDX analysis. The presence of Si in GP/PANI-modified SPGE as depicted in Figure 3b was due to the APTES used as a dispersion agent. Meanwhile, the existence of N, O and a significant increase of C in GP/PANI-modified SPGE clearly indicated that the GP/PANI was successfully bound to the surface of SPGE. In addition, Raman spectroscopy was used to identify the presence of PANI in GP/PANI-modified electrode. As shown in Figure 3c, the spectrum of the GP-modified SPGE clearly shows three bands at 1350, 1580 and 2700 cm −1 which represent D, G and 2D bands, respectively. The G band represents the in-plane bondstretching motion of the pairs of carbon sp 2 atoms, while the D band is related to the conversion of a sp 2 -hybridized carbon to a sp 3 -hybridized carbon [18]. Compared with graphene, the D band of PANI showed a slight shift to higher frequencies, probably due to the strong long-range π-π and electrostatic interaction between PANI and GP-modified SPGE [30].

Characterization of GP/PANI-Modified SPGE
GP/PANI nanocomposite material was used to modify bare SPGE in order to enhance the sensor performance. The morphology of GP/PANI-modified SPGE surface was studied by FESEM-EDX. It was observed in Figure 3a,b that the surface structure of GP/PANI-modified electrode turned relatively rough and agglomerate compared to bare SPGE due to the formation of nanoparticle clusters on the electrode surface, which resulted in an increase of active surface area [28,29]. The presence of GP/PANI nanocomposite was also confirmed using EDX analysis. The presence of Si in GP/PANI-modified SPGE as depicted in Figure 3b was due to the APTES used as a dispersion agent. Meanwhile, the existence of N, O and a significant increase of C in GP/PANI-modified SPGE clearly indicated that the GP/PANI was successfully bound to the surface of SPGE. In addition, Raman spectroscopy was used to identify the presence of PANI in GP/PANI-modified electrode. As shown in Figure 3c, the spectrum of the GP-modified SPGE clearly shows three bands at 1350, 1580 and 2700 cm −1 which represent D, G and 2D bands, respectively. The G band represents the in-plane bond-stretching motion of the pairs of carbon sp 2 atoms, while the D band is related to the conversion of a sp 2 -hybridized carbon to a sp 3 -hybridized carbon [18]. Compared with graphene, the D band of PANI showed a slight shift to higher frequencies, probably due to the strong long-range π-π and electrostatic interaction between PANI and GP-modified SPGE [30].

Characterization of GP/PANI-Modified SPGE
GP/PANI nanocomposite material was used to modify bare SPGE in order to enhance the sensor performance. The morphology of GP/PANI-modified SPGE surface was studied by FESEM-EDX. It was observed in Figure 3a,b that the surface structure of GP/PANI-modified electrode turned relatively rough and agglomerate compared to bare SPGE due to the formation of nanoparticle clusters on the electrode surface, which resulted in an increase of active surface area [28,29]. The presence of GP/PANI nanocomposite was also confirmed using EDX analysis. The presence of Si in GP/PANI-modified SPGE as depicted in Figure 3b was due to the APTES used as a dispersion agent. Meanwhile, the existence of N, O and a significant increase of C in GP/PANI-modified SPGE clearly indicated that the GP/PANI was successfully bound to the surface of SPGE. In addition, Raman spectroscopy was used to identify the presence of PANI in GP/PANI-modified electrode. As shown in Figure 3c, the spectrum of the GP-modified SPGE clearly shows three bands at 1350, 1580 and 2700 cm −1 which represent D, G and 2D bands, respectively. The G band represents the in-plane bondstretching motion of the pairs of carbon sp 2 atoms, while the D band is related to the conversion of a sp 2 -hybridized carbon to a sp 3 -hybridized carbon [18]. Compared with graphene, the D band of PANI showed a slight shift to higher frequencies, probably due to the strong long-range π-π and electrostatic interaction between PANI and GP-modified SPGE [30]. Next, the electrochemical behavior of GP/PANI-modified SPGE in potassium ferricyanide (K3Fe(CN)6) solution was investigated. K3Fe(CN)6 solution was used to study the electrocatalytic response of GP/PANI nanocomposite coated SPGE surface using the CV technique. It is notable that Ferrocyanide ion has been a good redox indicator for the analysis of the redox properties of modified electrodes due to its reversibility and rapid electrochemical reactions [31]. Moreover, it is believed that the thickness of GP/PANI nanocomposite on surface modification of SPGE played a crucial role for sensor performance. In this regard, we have studied different amounts of GP/PANI solution coated onto SPGE.
As depicted in Figure 4a, the sensor electrode with modification of 4 µL GP/PANI solution shows the highest current value approximately 50 µA. As the volume of GP/PANI solution was increased to 5 and 6 µL, the value of peak current decreased to 33 and 15 µA, respectively. However, it is noted that the peak current was slightly increased to 20 µA as the volume of GP/PANI solution increased to 7 µL. This could happen to the optimized value of 4 µL GP/PANI solution that gives highest surface electrostatic interactions, as compared to the GP/PANI volume solutions of 5, 6 and 7 µL. On the other hand, 4 µL of GP/PANI solution provides more stable of current observation (in terms of standard deviation) and higher peak current than bare SPGE. Thus, the thickness of GP/PANI particles on a SPGE electrode is crucial as it will affect the electrochemical performance. In this work, 4 µL GP/PANI solution was used to modify the SPGE surface.
To monitor the performance of the GP/PANI-modified SPGE for immunosensor development, the effective surface area for the electrode was studied and determined. To study the effective surface area of the electrode, the CV was performed in 1 mM K3Fe(CN)6 containing 50 mM KCl solution at different scan rates (10-100 mV/s). The effective surface area can be estimated according to the Randles-Sevcik equation [7].  Next, the electrochemical behavior of GP/PANI-modified SPGE in potassium ferricyanide (K 3 Fe(CN) 6 ) solution was investigated. K 3 Fe(CN) 6 solution was used to study the electrocatalytic response of GP/PANI nanocomposite coated SPGE surface using the CV technique. It is notable that Ferrocyanide ion has been a good redox indicator for the analysis of the redox properties of modified electrodes due to its reversibility and rapid electrochemical reactions [31]. Moreover, it is believed that the thickness of GP/PANI nanocomposite on surface modification of SPGE played a crucial role for sensor performance. In this regard, we have studied different amounts of GP/PANI solution coated onto SPGE.
As depicted in Figure 4a, the sensor electrode with modification of 4 µL GP/PANI solution shows the highest current value approximately 50 µA. As the volume of GP/PANI solution was increased to 5 and 6 µL, the value of peak current decreased to 33 and 15 µA, respectively. However, it is noted that the peak current was slightly increased to 20 µA as the volume of GP/PANI solution increased to 7 µL. This could happen to the optimized value of 4 µL GP/PANI solution that gives highest surface electrostatic interactions, as compared to the GP/PANI volume solutions of 5, 6 and 7 µL. On the other hand, 4 µL of GP/PANI solution provides more stable of current observation (in terms of standard deviation) and higher peak current than bare SPGE. Thus, the thickness of GP/PANI particles on a SPGE electrode is crucial as it will affect the electrochemical performance. In this work, 4 µL GP/PANI solution was used to modify the SPGE surface.
To monitor the performance of the GP/PANI-modified SPGE for immunosensor development, the effective surface area for the electrode was studied and determined. To study the effective surface area of the electrode, the CV was performed in 1 mM K 3 Fe(CN) 6 containing 50 mM KCl solution at different scan rates (10-100 mV/s). The effective surface area can be estimated according to the Randles-Sevcik equation [7]. I p = 2.69 × 10 5 A × D 1/2 n 2/3 Cv 1/2 , where I p represents the oxidation peak current (µA), n = 1 which is the number of electrons transferred, D is the diffusion coefficient of ferricyanide solution (7.6 × 10 −6 cm 2 /s) [29], A is the electrode surface area (A) (cm 2 ), C is the concentration of ferricyanide (mol/cm 3 ) and v is the scan rate (V/s). Based on the graph of I p versus v 1/2 as shown in Figure 4b, the effective surface area values of bare SPGE and GP/PANI-modified SPGE were calculated as 0.101 and 0.535 cm 2 , respectively. The utilization of GP/PANI nanocomposite as electrochemical enhancer has led to five times greater sensor performance with respect to the bare SPGE. In addition, the linear relationship between redox peak current of GP/PANI-modified SPGE versus square root of scan rate (v 1/2 ) reveals that the electrochemical reaction of the ferricyanide surface is a diffusion-controlled process [29].
where Ip represents the oxidation peak current (µA), n = 1 which is the number of electrons transferred, D is the diffusion coefficient of ferricyanide solution (7.6 × 10 −6 cm 2 /s) [29], A is the electrode surface area (A) (cm 2 ), C is the concentration of ferricyanide (mol/cm 3 ) and v is the scan rate (V/s). Based on the graph of Ip versus v 1/2 as shown in Figure 4b, the effective surface area values of bare SPGE and GP/PANI-modified SPGE were calculated as 0.101 and 0.535 cm 2 , respectively. The utilization of GP/PANI nanocomposite as electrochemical enhancer has led to five times greater sensor performance with respect to the bare SPGE. In addition, the linear relationship between redox peak current of GP/PANI-modified SPGE versus square root of scan rate (v 1/2 ) reveals that the electrochemical reaction of the ferricyanide surface is a diffusion-controlled process [29].

Sandwich Electrochemical Immunosensor for CFP10 Detection Using GP/PANI-Modified SPGE
A differential pulse voltammetry (DPV) technique was used in this study to measure the immunoreaction in detection of CFP10. Figure 5

Sandwich Electrochemical Immunosensor for CFP10 Detection Using GP/PANI-Modified SPGE
A differential pulse voltammetry (DPV) technique was used in this study to measure the immunoreaction in detection of CFP10. Figure 5 shows the electrochemical responses of bare SGPE, SPGE/GP/PANI, SPGE/GP/PANI/CapAb, SPGE/GP/PANI/CapAb/CFP10/Ab-Fe@Au, SPGE/GP/PANI/CapAb/BSA/Ab-Fe@Au and SPGE/Fe@Au that were characterized using DPV in 0.01 M PBS (pH 7.4) containing 2.7 mM KCl. It is clearly observed that the bare SPGE (peak a) produce a small oxidation peak current compared to GP/PANI-modified SPGE (peak b). Besides that, the oxidation peak current of bare SPGE was observed at 0.9 V but after modification of electrode with GP/PANI, the oxidation peak current was shifted to a lower positive potential, which indicates that the developed immunosensor have improved the electronic transport capacity [13]. The enhanced electrochemical response in the oxidation processes indicated the electroconducting properties of the GP/PANI nanocomposite in facilitating the electron transfer between the electrode surface and chloride ion [13]. After stepwise immobilization the modified electrode with capture antibody (CapAb) (peak c) and Ab-Fe@Au with CFP10 antigen (peak d), the peak currents gradually increase might be due to the signal enhancement of Fe@Au MNPs [32]. Meanwhile, for the SPGE/GP/PANI/CapAb/BSA/Ab-Fe@Au (peak e), the peak current decrease, showing that the Fe@Au MNPs did not attach on the surface of the electrode.
Sensors 2018, 18, x FOR PEER REVIEW 9 of 14 small oxidation peak current compared to GP/PANI-modified SPGE (peak b). Besides that, the oxidation peak current of bare SPGE was observed at 0.9 V but after modification of electrode with GP/PANI, the oxidation peak current was shifted to a lower positive potential, which indicates that the developed immunosensor have improved the electronic transport capacity [13]. The enhanced electrochemical response in the oxidation processes indicated the electroconducting properties of the GP/PANI nanocomposite in facilitating the electron transfer between the electrode surface and chloride ion [13]. After stepwise immobilization the modified electrode with capture antibody (CapAb) (peak c) and Ab-Fe@Au with CFP10 antigen (peak d), the peak currents gradually increase might be due to the signal enhancement of Fe@Au MNPs [32]. Meanwhile, for the SPGE/GP/PANI/CapAb/BSA/Ab-Fe@Au (peak e), the peak current decrease, showing that the Fe@Au MNPs did not attach on the surface of the electrode. The sensitivity of DPV using the developed sensor towards different concentrations of M. tuberculosis CFP10 antigen was studied in the range of 20 to 100 ng/mL. The correlation between the DPV current and the target concentrations as obtained in Figure 6a indicates that the detection responses of the immunosensor was linear with the value of the complementary target with the linear regression of R 2 = 0.99. An increase in peak current was observed as the concentration of CFP10 was increased. The device sensitivity was approximately 3.41 × 10 −7 A/ng·mL −1 calculated from the curved plotted in the Figure 6a. The detection response of the developed immunosensor can be expressed as the relative change in current (ΔI/I0), where ΔI is the change in peak current after the target detection and I0 is the peak current corresponding to the probe (capture antibody) immobilization response [33]. The relative change of peak current towards different concentrations of CFP10 is plotted in Figure 6b. The significant current change around 5% was obtained as 100 ng/mL of CFP10 was employed on the developed immunoassay. As the concentration of CFP10 was decreased to 60, 50 and 40 ng/mL, a descending trend of detection response was observed approximately 3.5%, 3.2% and 2.4%, respectively. Meanwhile, merely 0.7% current change was recorded when 20 ng/mL of CFP10 was introduced. The limit of detection (LOD) can be defined as YLOD = Yprobe + 3σ; where Yprobe is the mean of the probe measurement and σ is the standard deviation of the calibration curve [34,35]. The linear regression equation of the plots shown in Figure 6b is YLOD = 5.7 log10 x + 44.65, where x is the concentration of target CFP10 antigen and YLOD is calculated from the equation. The calculated LOD obtained in this work was approximately 15 ng/mL. This was estimated as the lowest concentration of target CFP10 that can be provided a detectable electronic signal. The developed immunosensor assay can meet the demands of detecting the CFP10 in real sample, as the physiological range of interest of CFP10 is around 100 ng/mL [36]. Remarkably, this approach compared reasonably in term The sensitivity of DPV using the developed sensor towards different concentrations of M. tuberculosis CFP10 antigen was studied in the range of 20 to 100 ng/mL. The correlation between the DPV current and the target concentrations as obtained in Figure 6a indicates that the detection responses of the immunosensor was linear with the value of the complementary target with the linear regression of R 2 = 0.99. An increase in peak current was observed as the concentration of CFP10 was increased. The device sensitivity was approximately 3.41 × 10 −7 A/ng·mL −1 calculated from the curved plotted in the Figure 6a. The detection response of the developed immunosensor can be expressed as the relative change in current (∆I/I 0 ), where ∆I is the change in peak current after the target detection and I 0 is the peak current corresponding to the probe (capture antibody) immobilization response [33]. The relative change of peak current towards different concentrations of CFP10 is plotted in Figure 6b. The significant current change around 5% was obtained as 100 ng/mL of CFP10 was employed on the developed immunoassay. As the concentration of CFP10 was decreased to 60, 50 and 40 ng/mL, a descending trend of detection response was observed approximately 3.5%, 3.2% and 2.4%, respectively. Meanwhile, merely 0.7% current change was recorded when 20 ng/mL of CFP10 was introduced. The limit of detection (LOD) can be defined as Y LOD = Y probe + 3σ; where Y probe is the mean of the probe measurement and σ is the standard deviation of the calibration curve [34,35]. The linear regression equation of the plots shown in Figure 6b is Y LOD = 5.7 log 10 x + 44.65, where x is the concentration of target CFP10 antigen and Y LOD is calculated from the equation. The calculated LOD obtained in this work was approximately 15 ng/mL. This was estimated as the lowest concentration of target CFP10 that can be provided a detectable electronic signal. The developed immunosensor assay can meet the demands of detecting the CFP10 in real sample, as the physiological range of interest of CFP10 is around 100 ng/mL [36]. Remarkably, this approach compared reasonably in term of linearity range and LOD with other previously reported CFP10-based detections as summarized in Table 1.   In order to test the validity of this method in real sample application, the developed immunosensor was then applied for the determination of M. tuberculosis CFP10 antigen in sputum samples. Five sputum samples consisted of four sputum samples from positive TB patients and one from a healthy person as the control experiment were collected from Hospital Universiti Sains Malaysia (HUSM). The fresh sputum samples were supplied to the Science laboratory at School of Medical Science (USM) and smear microscope examination was implemented prior to the culture  In order to test the validity of this method in real sample application, the developed immunosensor was then applied for the determination of M. tuberculosis CFP10 antigen in sputum samples. Five sputum samples consisted of four sputum samples from positive TB patients and one from a healthy person as the control experiment were collected from Hospital Universiti Sains Malaysia (HUSM). The fresh sputum samples were supplied to the Science laboratory at School of Medical Science (USM) and smear microscope examination was implemented prior to the culture method. The electrochemical analysis was performed simultaneously in the same workspace. We found that this electrochemical immunosensor showed the same trend of signals for all positive TB patients' samples (Sample 1-Sample 4) as depicted in Figure 7, which obviously tended to give the highest signal change as compare to the healthy person sample (negative sample as control). From this viewpoint, we have verified that our developed GP/PANI modified SPGE electrode in a sandwich immunoassay format is feasible for early TB diagnosis through detection of the CFP10-specific antigen using sputum samples. method. The electrochemical analysis was performed simultaneously in the same workspace. We found that this electrochemical immunosensor showed the same trend of signals for all positive TB patients' samples (Sample 1-Sample 4) as depicted in Figure 7, which obviously tended to give the highest signal change as compare to the healthy person sample (negative sample as control). From this viewpoint, we have verified that our developed GP/PANI modified SPGE electrode in a sandwich immunoassay format is feasible for early TB diagnosis through detection of the CFP10specific antigen using sputum samples. Figure 7. DPV response of sandwich electrochemical immunosensor for CFP10 detection using GP/PANI-modified SPGE in real sample application.

Reproducibility of the Developed Immunosensor
In biosensor performance assessment, reproducibility is a very significant feature in order to identify the reliability in working activity. It can be defined as the capability of a developed immunosensor of producing equivalent feedbacks for a repeated experimental setup [41]. Therefore, it is compulsory to test the reproducibility of the fabricated immunosensor for reliability verification of this method since it is a very significant parameter for immunosensor. For detection of 60 ng/mL CFP10, a series of five electrodes was setup. Peak currents within the range of 6.094-6.168 µA were measured from the developed immunosensors as tabulated in Table 2. Based on the measured peak currents for the five electrodes, the relative standard deviation (RSD) can be calculated from the following equation: where σ is standard deviation and µ is the mean for the measurements. Acceptable RSD of 0.55% was obtained, which signify the developed immunosensor has an excellent reproducibility for the detection of low concentration CFP10 antigen.

Reproducibility of the Developed Immunosensor
In biosensor performance assessment, reproducibility is a very significant feature in order to identify the reliability in working activity. It can be defined as the capability of a developed immunosensor of producing equivalent feedbacks for a repeated experimental setup [41]. Therefore, it is compulsory to test the reproducibility of the fabricated immunosensor for reliability verification of this method since it is a very significant parameter for immunosensor. For detection of 60 ng/mL CFP10, a series of five electrodes was setup. Peak currents within the range of 6.094-6.168 µA were measured from the developed immunosensors as tabulated in Table 2. Based on the measured peak currents for the five electrodes, the relative standard deviation (RSD) can be calculated from the following equation: RSD = (σ/µ) × 100 (2) where σ is standard deviation and µ is the mean for the measurements. Acceptable RSD of 0.55% was obtained, which signify the developed immunosensor has an excellent reproducibility for the detection of low concentration CFP10 antigen.

Conclusions
In this study, we presented a novel detection strategy that uses GP/PANI-modified SPGE and sandwich immunoassay of TB marker protein CFP10 for early detection of TB infection. This platform employs the CapAb immobilized with GP/PANI as an electrochemical transducer and the Ab was conjugated by Fe 3 O 4 /Au MNPs allowing for the antigen-specific detection of probes from the assay solution. The morphological and surface analyses of GP/PANI modified on the SPGE surface were studied using FESEM equipped with EDX, which confirmed the existence of N, O and C elements of this nanocomposite on the surface electrode. The Raman spectra showed that the D band changes significantly with the presence of PANI in GP/PANI modified surface. The presence of the amine groups (NH 2 ) in GP/PANI nanocomposite allowed the direct anchoring of the biolayer, which was identified by FTIR spectra at a peak of 1477 cm −1 . It was observed that the integration of GP/PANI on SPGE represents an effective strategy in improving about five times sensor performance compared to unmodified SPGE. This method was capable to detect the CFP10 within 3 h with linearity in the range of 20-100 ng/mL (R 2 = 0.99). The estimated LOD obtained in this work was 15 ng/mL. Moreover, the developed sandwich assay showed exceptional detection response towards CFP10 in sputum specimens for a real sample application and demonstrated high reproducibility (RSD of 0.55%) as a disposable-based immunoassay. Our findings signify that TB can be detected in a simple, sensitive and economic way through sputum-based observation, indicating that this assay analysis might be applied in the development of rapid-sensing tools for early TB monitoring.