Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jun 15.
Published in final edited form as: Biosens Bioelectron. 2022 Mar 3;206:114142. doi: 10.1016/j.bios.2022.114142

Nano gold-doped molecularly imprinted electrochemical sensor for rapid and ultrasensitive cortisol detection

Sanjida Yeasmin 1, Bo Wu 1, Ye Liu 1, Ahasan Ullah 1, Li-Jing Cheng 1,*
PMCID: PMC9010368  NIHMSID: NIHMS1787057  PMID: 35259605

Abstract

Rapid and sensitive detection of steroid hormone cortisol can benefit the diagnosis of diseases related to adrenal gland disorders and chronic stress. We report a molecularly imprinted polymer (MIP)-based electrochemical sensor that utilized nano gold-doped poly o-phenylenediamine (poly-o-PD) film to selectively determine trace level cortisol with enhanced sensitivity. The sensor detected cortisol levels by measuring the current change of the redox-active probes in response to the binding of target cortisol to the imprinted sites in the polymer. The gold-doped MIP (Au@MIP) sensor was prepared using a facile one-step in situ gold reduction and electropolymerization method to distribute high-density gold nanoparticles in the vicinity of the binding cavities. The in situ gold reduction promote the polymerization reaction, enlarging the effective surface area of the sensor. The nano gold doping also facilitated charge transfer when exposed to redox reagents. It enabled efficient blocking of the charge transfer upon the occupation of the cavities by cortisol, resulting in enhanced detection response and sensitivity. The Au@MIP sensor exhibited a high affinity toward cortisol binding with a dissociation constant Kd of ~0.47 nM, a linear detection range from 1 pM to 500 nM with a detection limit of ~200 fM, and satisfied specificity over other steroid hormones with highly similar structures. The sensor was successfully demonstrated to determine cortisol levels in spiked saliva in normal and elevated ranges. The facile antibody-free cortisol detection method was proved to be highly sensitive and selective, suitable for point-of-care testing applications.

Keywords: Molecularly imprinted polymer, electropolymerization, steroid hormones, stress marker, salivary cortisol, electrochemical sensor

Graphical Abstract

graphic file with name nihms-1787057-f0001.jpg

1. Introduction

Cortisol is a steroid hormone produced by adrenal gland in response to stress or agitate states. It plays an essential role in regulating physiological processes, including blood pressure, glucose level, and metabolism (Gatti et al. 2009; Zea et al. 2020). Deficient or excess cortisol triggered by physical or psychological stress is linked to diseases, including post-traumatic stress disorders, Cushing’s disease, and chronic fatigue (Chobanian et al. 2003; Edwards et al. 1974; Raff and Findling 2003; Whitworth et al. 2000). High cortisol levels in patients with COVID-19 were also associated with increased mortality and reduced median survival, indicating that cortisol could be a biomarker for evaluating the severity of SARS-CoV-2 disease (Tan et al. 2020). Salivary cortisol level is an excellent index of the unbound plasma-free cortisol. Therefore, measuring salivary cortisol is an accurate and non-invasive method to assess the biologically active form of cortisol (El-Farhan et al. 2017; Zea et al. 2020). Cortisol concentration in the body fluctuates in a circadian rhythm, being the highest in the early morning (5–30 nM), moderate during the day, and the lowest at night (< 2 nM). The late-night salivary cortisol concentration in the range of 2.76–166 nM was determined for clinically confirmed Cushing’s patients (Papanicolaou et al. 2002; Raff et al. 1998).

The existing clinical cortisol measurement methods are high-cost, laborious tests, with urine or saliva samples taken from a patient and sent to a lab to await results. A standard method utilizes sophisticated, time-consuming liquid chromatography-tandem mass spectrometry to achieve sub-ng/mL detection limit (Baid et al. 2007). Competitive enzyme-linked immunosorbent assay (ELISA) detects cortisol in sub-pg/mL to ng/mL levels but with the drawbacks of tedious assay processes (~90 min to 2 hr) (Steckl and Ray 2018) and limited detection selectivity due to its cross-reactivity with other steroid hormones of similar structures (Krasowski et al. 2014; Zea et al. 2020). Besides ELISA, lateral flow assays and other immunosensors emerge with various detection techniques (Dalirirad and Steckl 2019; Liu et al. 2020; Madhu et al. 2020; Ray and Steckl 2019; Tu et al. 2020). Such immunosensors rely on antibodies or aptamers for specific cortisol recognition. These biological receptors are costly and tend to be unstable in ambient conditions, leading to a short shelf-life (Manickam et al. 2016). Alternatively, molecularly imprinted polymer (MIP), a synthetic polymer with biorecognition sites for target-specific binding, can substitute antibodies with numerous advantages, including ease of preparation, excellent stability, low cost, longer lifetime, high selectivity, and sensitivity (Beluomini et al. 2019; Reddy and Gobi 2013; Vlatakis et al. 1993). Combining MIP with electrochemical sensing techniques will enable sensitive detection with easy and rapid measurements, making it possible for point-of-care applications (Steckl and Ray 2018; Wu et al. 2022; Zea et al. 2020).

Limited reports have demonstrated MIP-based electrochemical cortisol sensors. A cortisol sensor used ethylene glycol-based MIP thermally polymerized on the electrode to achieve a LOD of 2 ng/mL (~5.5 nM) (Mugo and Alberkant 2020). A wearable MIP cortisol transistor fabricated by UV polymerization of methyl methacrylate-based copolymer successfully detected sweat cortisol in a range of 0.01 μM to 5 mM (Parlak et al. 2018). These cortisol MIP sensors have unsatisfied detection limits and require complicated fabrication processes. In contrast, electropolymerization allows MIP to directly grow on predefined electrodes with controllable film thickness and morphology using a one-step process at ambient temperature. (Crapnell et al. 2019). However, such planar MIP-coated electrodes often have recognition sites with limited accessibility, and slow charge transfer (Beluomini et al. 2017). Furthermore, the imprinted cavities in a non-conductive MIP tend to be isolated from each other, resulting in poor conductivity and low electrochemical signals. Incorporating nanomaterials with MIP can significantly improve the sensor’s recognition efficiency by offering a large surface area and more accessible imprinted sites to bind target molecules efficiently (Beluomini et al. 2017; Sanati et al. 2021).

Here, we demonstrate a gold nanoparticle (AuNP)-doped molecularly imprinted polymer (Au@MIP) for rapid and ultrasensitive cortisol detection. The MIP formed by polymerization of o-phenylenediamine (o-PD) created abundant amine groups to associate with cortisol molecules through hydrogen bonding. Also, poly-o-PD is biocompatible, mechanically stable, and can be managed to create a thin and compact polymer matrix required to deliver a rapid and stable sensing response (Buffon and Stradiotto 2019). The AuNP doping was introduced to promote the electron transfer between the redox probes and the electrode across the insulating MIP layer. The binding of target molecules in the imprinted cavities suppressed the electron transfer, yielding the change in current response. The presence of AuNPs nearby the imprinted cavities allowed the captured targets to more effectively block the electron transfer across the AuNPs, and, therefore, enhanced the detection sensitivity. A few reports demonstrated AuNP embedded MIPs by performing polymerization in the presence of pre-synthesized AuNPs (Motia et al. 2021; Motia et al. 2020; Sehit et al. 2020). Unlike the conventional approach, the co-polymerization process created an increased effective surface area of the MIP and provided a superior detection sensitivity of the MIP sensor. To the best of our knowledge, this type of MIP for cortisol detection has not been reported. The sensor was successfully validated to determine cortisol concentrations spiked in the artificial saliva in normal and elevated ranges, demonstrating a great potential to facilitate the rapid non-invasive assessment of steroid hormones for clinical applications.

2. Material and methods

2.1. Chemicals

Chemicals, including o-phenylenediamine (o-PD), cortisol, progesterone, β-estradiol, estriol, estrone, gold (III) chloride trihydrate (HAuCl4 ·3H2O), potassium ferricyanide (K3[Fe(CN)6], 99%), potassium ferrocyanide (K4[Fe(CN)6], 99%), and potassium chloride (KCl) were obtained from Sigma-Aldrich. Dehydroepiandrosterone (DHEA) was purchased from Chem-Impex International. Phosphate-buffered saline (PBS, 10X, pH 7.4), sodium acetate, and acetic acid (glacial, 99.7%) were obtained from Fisher Scientific. Artificial saliva was obtained from Pickering Laboratories (Mountain View, CA). Ferro-/ferricyanide redox couple solution was prepared from a mixture of 5 mM K3[Fe(CN)6])/K4[Fe(CN)6] (1:1 molar ratio) in 0.1 M KCl. All chemicals were of analytical grade and used without further purification. Ultrapure Milli-Q water with a resistivity of 18.2 MΩ-cm was used.

2.2. Preparation of the molecularly imprinted polymer (MIP) biosensors

A 3 mm diameter glassy carbon electrode (GCE) was polished, washed thoroughly with DI water, and then sonicated in ethanol for 5 min. The GCE was then electrochemically cleaned in a 0.5 M H2SO4 solution using 10 cycles from −0.5 to 1 V at 50 mVs−1 scan rate, followed by rinsing with deionized water. After drying, a gold layer was first electrodeposited onto the cleaned GCE in 0.5 mM HAuCl4 in 0.1 M KCl solution using cyclic scanning from 0.2 to −1 V at 50 mVs−1 scan rate for 10 consecutive cycles (Fig. S1). A Au@MIP layer was then electropolymerized on the gold-coated GCE in a 0.1 M, pH 4 acetate buffer solution containing a mixture of 5 mM o-PD, 0.5 mM cortisol templates, and 0.1 mM HAuCl4. The electropolymerization was carried out by applying a cyclic potential between 0 to 1 V at a scan rate of 50 mVs−1 for 30 cycles. The cortisol templates were then removed by soaking the electrode in an ethanol solution for 20 min under magnetic stirring, leaving cortisol-selective binding sites in the Au@MIP matrices ready for cortisol detection. As a control, a non-imprinted polymer (Au@NIP) modified electrode was also prepared under the same procedures except for the absence of template cortisol. For comparison, an undoped MIP electrode was also prepared in a similar procedure without HAuCl4 during electropolymerization.

2.3. Material and sensor characterizations

The sensors were analyzed using cyclic voltammetry (CV), square wave voltammetry (SWV), and electrochemical impedance spectroscopy (EIS) in the presence of redox probes using a Gamry Reference 600 potentiostat with a three-electrode system, including a platinum counter electrode and an Ag/AgCl (saturated KCl) reference electrode. The material properties of the MIP were characterized using scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) with a QUANTA 600F SEM (FEI, USA) and a Fourier-transform infrared (FTIR) spectroscopy (Thermo Fisher Nicolet IS50).

2.4. Sensor evaluations

To characterize the sensor performance, the MIP sensors were first incubated in the cortisol solutions of various concentrations in 1× PBS for 8 min, followed by rinsing with 1× PBS. CV measurements were performed between 0 to 0.5 V at a scan rate of 100 mVs−1. EIS measurements were conducted by applying 10 mV with the frequency ranging from 0.2 Hz to 100 kHz. SWV was applied from 0 to 0.5 V with 25 mV pulse size and 10 Hz frequency. To determine the cortisol level in artificial saliva, the spiked saliva samples were diluted 10 times with 1× PBS.

3. Results and discussion

3.1. Preparation and characterization of cortisol imprinted sensor

The CV responses during the electropolymerizations of undoped MIP and Au@MIP show distinct and irreversible anodic peaks at 0.62 V and 0.46 V, respectively, in the first polymerization cycle. The lower oxidation peak voltage and a greater oxidation peak current for the electropolymerization of Au@MIP can be attributed to a faster polymerization catalyzed by the gold ions (Han et al. 2011; Mallick et al. 2007; Sun et al. 2004). During electropolymerization, the protonated o-PD monomer served as electron donors and interacted with strong oxidizing agent AuCl4− ions. The formation of o-PD oligomers reduced the AuCl4− ions to AuNPs, which subsequently functioned as active catalysts, promoting the oxidative polymerization of additional o-PD monomers. The highly dispersed AuNPs were anchored to the polymeric network, possibly due to the strong affinity between the electron-rich N atoms on poly-o-PD and electron-deficient AuNPs (Sen and Sarkar 2015). With the increased scan cycles, the oxidation peak current reduced significantly, implying the formation of an insulating poly-o-PD layer on the electrode surface. The cortisol templates were embedded within the polymer network through the H-bondings between the hydroxyl groups on cortisol and the protonated amine groups on poly-o-PD in the pH 4 acetate buffer (Azadmehr and Zarei 2019). Fig. 2(c) shows that the electropolymerization of a gold-doped non-imprinted polymer (Au@NIP) yielded a similar CV to the Au@MIP. Cortisol is not an electroactive molecule and does not affect the current response in the electropolymerization process.

Fig. 2.

Fig. 2.

Open in a new tab

CV responses during electropolymerization of (a) undoped MIP, (b) Au@MIP, and (c) Au@NIP. (d-f) CVs of the three electrodes after each fabrication and sensing steps – (1) bare electrode, (2) gold-coated electrode, (3) after polymerization, (4) template removal, and (5) after binding of 100 nM cortisol. EIS of (g) MIP and (h) Au@MIP obtained after (A) electropolymerization, (B) template removal, and (C) 100 nM cortisol sensing. The inset represents an equivalent circuit of the electrode, where Rct is the electron transfer resistance. (i) Rct of MIP and Au@MIP electrodes. Both CV and EIS were measured in 0.1 M KCl solutions containing 5 mM redox probes.

Fig. 2 (d–f) shows the CVs of the undoped MIP, Au@MIP, and Au@NIP electrodes recorded after each step of device preparation and cortisol sensing in the presence of redox probes. The three electrodes displayed similar CV responses up to the electropolymerization step. Electrodeposition of gold on the bare GCE raised the redox current from curve (1) to curve (2) due to the increased electrical conductivity and surface area. The results agree with the EIS measurements in Fig. S2. Electropolymerization significantly suppressed the peak currents by more than 75 times to curve (3) due to the formation of the insulating MIP and NIP layers, blocking the entry of the redox probes to the electrode surface or the embedded AuNPs. After cortisol templates were removed from MIP using ethanol, the anodic peak of the undoped MIP electrode increased from 2.5 μA to 51.88 μA (curve (4) in Fig. 2(d)) due to its increased permeability to the redox probes through the cortisol specific cavities. Rebinding 100 nM cortisol on the MIP obstructed the access of redox probes to the electrode and thus decreased the anodic peak current to 43.13 μA at 0.38 V (curve (5), Fig. 2 (d)). Compared to the undoped MIP, the Au@MIP electrode showed a higher peak current of 70.17 μA after template removal, i.e., curve (4) in Fig. 2(e). The current dropped to an even lower current level of 37.52 μA after rebinding 100 nM cortisol. The enhanced sensitivity in the Au@MIP electrode is associated with the improved charge transfer rate between the redox probes and the electrode. For the control device Au@NIP, ethanol incubation moderately increases the current level, probably due to the swelling of the polymer matrix in the organic solvent (curve (4), Fig. 2 (f)). With the absence of imprinted cavities, the CV of Au@NIP shows negligible change after cortisol rebinding (curve (5)). The results proved the target binding capability of the undoped MIP and Au@MIP electrodes. Significantly, the Au@MIP electrode exhibited the most considerable current change in response to the rebinding of target cortisol and will be utilized to validate cortisol sensing.

Fig. 2(i) summarizes the electrode charge transfer resistances (Rct) extracted from the EIS in Fig. 2(g) and 2(h) agree with the CVs in Fig. 2(d) and 2(e). The template removal process rendered the MIPs more permeable to the redox probes and thus reduced Rct. The permeability declined after rebinding with 100 nM cortisol and, therefore, increased Rct. The presence of AuNPs in the polymer layer lowered the Rct and increased the cortisol sensing response compared with the undoped MIP, consistent with the results from CV measurements.

The EDX spectra of MIP and Au@MIP in Fig. 3(a) and 3(b) show that the significant peaks corresponding to Au were only observed on Au@MIP. The SEM images indicate that Au@MIP film was decorated by AuNPs with diameters varying from about 10 to 80 nm and had a rough polymer surface with granular morphology compared to the undoped MIP. The rough surface can be attributed to the fast polymerization of o-PD catalyzed by gold ions, which agrees with the lower oxidation peak voltages and increased oxidation peak current observed in Fig. 2 (a) and 2(b). The distinguishable surface morphologies were also evidenced by their different active surface areas. The active surface areas of MIP and Au@MIP were extracted from the relationship between the anodic peak current and the scan rate of their CV responses using a Randles-Sevcik equation, as shown in Fig. S3. The active surface area of Au@MIP was estimated to be 5.58 mm2, which is 51% larger than that of MIP. The increased active surface area of Au@MIP can be associated with the rough polymer surface due to the fast polymerization catalyzed by gold ions and the AuNPs in the resulting polymer.

Fig. 3.

Fig. 3.

Open in a new tab

Energy dispersive X-ray (EDX) spectra and scanning electron microscopy (SEM) images of (a) undoped MIP and (b) Au@MIP thin films. The scale bars represent 500 nm. (c) FTIR spectra of MIP, Au@MIP, and Au@NIP.

The FTIR spectra in Fig. 3(c) show that both MIP and Au@MIP contain peaks around 3410 cm−1 and 1710 cm−1, which can be assigned to the O–H and C=O stretching vibration, respectively, and a weak peak at 1239 cm−1 associated with C–O stretching (Manivannan et al.). These peaks were contributed by cortisol in the polymer matrix, not observable in Au@NIP. The primary amine peak in the polymerized o-PD appeared at 3350 cm−1 for Au@NIP and at the shoulder of O–H stretching peak (3410 cm−1) for MIP and Au@MIP. The peaks at 1643, 1572, 1408 cm−1 are associated with C=N, C=C, and C–N stretching vibration of phenazine structures in the poly-o-PD matrix, respectively (Olgun and Gülfen 2014; Wang et al. 2008). MIP and Au@MIP share similar FTIR spectral features, indicating the same chemical composition of their polymer matrices.

AuNPs-embedded MIPs have been demonstrated by forming MIP in the presence of pre-synthesized AuNPs (Motia et al. 2021; Motia et al. 2020; Sehit et al. 2020). In this approach, the AuNPs were trapped in the polymer matrix without a robust bonding mechanism, leading to leakage of AuNPs as the polymer swelled (Ahmad et al. 2015). We experimentally compared the gold doped MIPs created by in situ gold reduction and with pre-synthesized AuNPs. The CV response in Fig. S4(a) shows that the presence of pre-synthesized AuNPs did not increase the reaction rate of electropolymerization. Although the embedded AuNPs created by the conventional way improved the conductivity of the MIP (S4(b)), the Au@MIP sensor prepared by in situ gold reduction yielded a much larger sensitive response toward cortisol binding. The different sensitivity could be because the embedded AuNPs produced by the conventional method were well covered by the polymer, hindering the electron transfer. On the contrary, the in situ gold reduction promoted the polymerization reaction that increased the surface area of MIP. It allowed the binding cavities to form in the vicinity of high-density AuNPs, resulting in efficient modulation of electron transfer in response to the capture of cortisol.

3.2. Optimization of Au@MIP sensors

Multiple process conditions were investigated to optimize the performance of the Au@MIP sensor, including eluents and elution time for template removal, and polymerization conditions (e.g., compositions, pH, and gold ion concentration in electropolymerization). The study is detailed in supplementary information. Among these parameters, we observed that the nano gold doping enabled more effective transduction of cortisol binding events into current change. Fig. 4(a) summarizes the effect of HAuCl4 concentrations in the electropolymerization process on the sensing responses of the Au@MIP. We characterized the relative peak current change ΔI/I0 of the sensors in response to 100 nM cortisol, where I0 is the initial peak current in SWV in the absence of cortisol and ΔI is the reduction in current after cortisol binding. The result indicates that with the increase in HAuCl4 concentrations, I0 kept increasing and ΔI increased until HAuCl4 exceeded 0.1 mM, yielding a maximal ΔI/I0 at 0.1 mM HAuCl4. The result may be explained by a simple equivalent circuit model composed of a series connection of two resistances contributed by the polymer matrix and captured molecules, i.e., RMIP = Rpoly + Rmol, as illustrated in Fig. 4(b). Under a fixed electrode potential V, the current passing through the Au@MIP layer in the absence of target molecules (Rmol = 0) can be presented by I0 = V/Rpoly. The binding of target molecules yields a current change ΔI = I0 − V/(Rpoly + Rmol) = V[1/Rpoly − 1/(Rpoly + Rmol)], which leads to a relative sensing response given by ΔI/I0 = Rmol/(Rpoly + Rmol). The equation implies that the sensing response ΔI/I0 can be increased by reducing the polymer matrix resistance Rpoly. The trend matches the experimental observation. The increased HAuCl4 concentration in the electropolymerization process raised the AuNP density in the polymer matrix, improving the charge transfer across the polymer matrix and thus reducing Rpoly. Moreover, the AuNPs distributed in the vicinity of the binding cavities allow the charge transfer pathway to be efficiently blocked upon the occupation of the cavities by the target cortisol, promoting the current change ΔI. The equations also imply that a more conductive polymer matrix (smaller Rpoly) promotes both I0 and ΔI. However, with the HAuCl4 concentrations increased beyond 0.1 mM, plenty of large Au clusters were observed to form on the MIP surface due to the fast gold ions reduction reaction. The formation of large Au clusters can increase I0 by providing a conductive coating on the MIP surface, which is not considered in the proposed simple model. Also, the AuNPs density near the cavities may not continue to increase, hindering the further growth of ΔI. With the optimized HAuCl4 concentration, we compared the sensing responses ΔI/I0 of the undoped MIP and Au@MIP sensors for (n=3) towards three cortisol concentrations (1, 10, and 100 nM), as shown in Fig. 4(c). The result indicates that the presence of gold doping significantly enhanced the sensitivity of the sensor, in agreement with the ΔI/I0 derived from the equivalent circuit model and the sensing response observed in Fig. 2(i).

Fig. 4.

Fig. 4.

Open in a new tab

(a) Effects of gold ion concentration in electropolymerization on the current responses and relative sensing signal ΔI/I0 in SWV in response to 100 nM cortisol. (b) Schematics and the simple equivalent circuit model describing the charge transfer in the MIP and Au@MIP films. (c) ΔI/I0 of MIP and Au@MIP sensors in response to various cortisol concentrations (n=3). SWVs were measured in 0.1 M KCl solution containing 5 mM redox probes.

3.3. Analytical performance

The adsorption kinetics of 10 nM cortisol to Au@MIP can be measured by the current change in SWV ΔI(t) as a function of cortisol binding time t. Fig 5(a) shows that ΔI increased rapidly in the first 2 min, and reached a steady state in 8 min. A 8 min incubation time was used for all the sensor characterization. The fast adsorption kinetics can be attributed to the effective capture of target cortisol by the thin Au@MIP film with a large active surface area. The kinetics curve is fitted well to a second-order adsorption model, t/ΔI(t)=(k ΔIe2)−1+t/ΔIe with k and ΔIe being the equivalent adsorption rate constant and the current change at equilibrium, respectively (Ho and McKay 1999). The second-order kinetics is also evidenced by a linear relationship between t/ΔI and binding time t, as shown in the inset of Fig. 5(a). The analytical performance of the Au@MIP sensor was investigated by characterizing the change in SWV current peaks in the presence of redox probes in response to various cortisol concentrations, as shown in Fig. 5(b). The SWV peak currents occurring at ~0.23 V vs. Ag/AgCl decreased with the increase of cortisol concentration due to the elevated number of occupied cavities in the polymer film. The sensor offered a linear response to cortisol concentration, ranging between 1 pM and 500 nM. On the contrary, Au@NIP exhibited negligible current change with various cortisol concentrations because of the lack of imprinted cavities (Fig. 5(c)). The calibration curves in Fig. 5(d) summarize the current change ΔI in response to cortisol concentration c in nM which can be described by ΔI = 2.061 log c + 7.798 (R2=0.992, n=3) for Au@MIP and ΔI = 0.252 log c + 0.646, (R2=0.797, n=3) for the control sensor, Au@NIP. The Au@MIP sensor exhibited a broad linear dynamic range between 1 pM and 500 nM, with a LOD of ~200 fM, defined as three times the standard deviation σ of blank measurements divided by the slope of the calibration curve. The sensor yielded σ ~34 nA for n=3.

Fig. 5.

Fig. 5.

Open in a new tab

(a) Adsorption kinetics of Au@MIP towards 10 nM cortisol presented with ΔI in SWV against binding time t, (n=3). The blue line is the fitting line to the data based on a second-order kinetic model. Inset represents the data and fitting line in a t/ΔI vs. t plot. SWVs of (b) Au@MIP and (c) Au@NIP in response to various cortisol concentrations ranging from 0 to 500 nM. (d) Calibration plots of the ΔI in SWV vs. cortisol concentrations for Au@MIP and Au@NIP (n=3). (e) Binding isotherm of cortisol on Au@MIP (n=3) measured by the peak current change ΔI in SWV as a function of cortisol concentration. The red line represents the fitting curve of the Langmuir-Freundlich (LF) adsorption model. The dashed line indicates the Kd value. (f) Selectivity study of Au@MIP and Au@NIP sensors in responses to 10 nM steroid hormones, including cortisol and other analogues, and 100 nM interferents, including glucose, serotonin, dopamine, and ascorbic acid (n=3) with their chemical structures shown in (g). SWVs were measured in 0.1 M KCl solution containing 5 mM redox probes.

To evaluate the binding affinity of cortisol to the Au@MIP, we analyzed the SWV peak current change ΔI(c) for a broader range of free cortisol concentrations c (n=3), including the responses outside the linear range, as shown in Fig. 5(e). The current response can be described by a heterogeneous Langmuir–Freundlich (LF) isotherm model (Umpleby et al. 2001) given by

ΔI(c)=A cmKdm+cm.

Here, the current change ΔI(c) is directly related to the equilibrium concentration of the cortisol bound to the Au@MIP layer. Constant A is proportional to the total number of binding sites in Au@MIP. Heterogeneous parameter m, a number between 0 and 1, measures the heterogeneity of the binding sites. A small m implies a broad number distribution of binding affinity. Kd is the dissociation constant accounting for the binding strength between cortisol and the imprinted cavities in the Au@MIP film. The fitting yields m = 0.36 and Kd = 0.47 nM with R2 = 0.993. The low Kd value suggests that the Au@MIP sensor presented a high binding affinity towards cortisol.

3.4. Selectivity, reproducibility, and regeneration

The detection selectivity was evaluated by comparing the peak current changes ΔI in SWV of the Au@MIP and Au@NIP electrodes in response to 10 nM structurally similar steroid hormones, including cortisol, DHEA, progesterone, β-estradiol, estriol, and estrone, as well as 100 nM glucose, dopamine, serotonin, and ascorbic acid, as shown in Fig. 5(f) (n=3). The ratio α of ΔI obtained from Au@MIP to that from Au@NIP, i.e., α = ΔIAu@MIP/ΔIAu@NIP, was employed to compare the sensing signals contributed by the molecular affinity of the MIP cavities and the non-specific binding. It is worth noting that rinsing the sensor with 1× PBS after binding was found to significantly reduce the non-specific adsorption of analytes on MIP or NIP surface (Fig. S7) and was applied throughout the tests. Cortisol detection was observed to give the highest α value of 15.2, much larger than the other non-target molecules, indicating that the Au@MIP sensor exhibited specific recognition sites to distinguish cortisol from other analogues depicted in Fig. 5(e).

Fig. S8(a) shows that seven duplicate sensors yielded very similar sensing responses to 10 nM cortisol with an RSD value of 4.4%, proving the high reproducibility of the sensor. Furthermore, the Au@MIP sensor can be regenerated by rinsing with an ethanol solution to allow multiple measurements. Fig. S8(b) shows the reproducible current change observed from the same sensor under four repeating cycles of cortisol detection and rinsing steps. The RSD of the current responses for cortisol detection was observed to be 3.8% after three successive measurements and increased to 6.1% after the fourth regeneration, probably due to the slight degradation of the imprinted cavities.

We compare the analytical performance, synthesis method, and operation methods of the proposed Au@MIP cortisol sensor compared with the reported counterparts in Table S2, S3, and S4. The comparison suggests that this work provided superior performance in several aspects, including a broad dynamic range and low detection limit (~200 fM) with high selectivity against structurally similar hormones and interfering analytes. The in situ gold deposition in the MIP provided the Au@MIP sensor with greater accessibility of recognition cavities by (1) distributing high-density AuNPs in the vicinity of the cavities, (2) increasing active surface area, (3) and effective transduction of the target binding. Besides, the sensor can be realized by a facile and low-cost MIP synthesis process with high reproducibility and regeneration.

3.5. Artificial Saliva Sample analysis

To validate the clinically relevant application, we spiked cortisol of different concentrations varying from 1 nM to 1 μM in artificial saliva covering both normal and diseased conditions. Cortisol spiked saliva samples were diluted ten times with 1× PBS before detection. Table 1 shows that the sensor yields satisfied recovery rates of 91% - 105%, verifying the feasibility and accuracy of the Au@MIP sensor for real sample testing.

Table 1.

Determination of cortisol in artificial saliva sample (n=3).

Spiked cortisol (nM) Cortisol after dilution (nM) Found (nM) RSD (%) Recovery (%)
1 0.1 0.101 7.562 101.12
10 1 0.917 5.981 91.666
20 2 1.846 2.279 92.301
50 5 5.256 4.870 105.122
100 10 9.537 2.167 95.370
200 20 20.159 1.876 100.794
500 50 52.082 3.849 104.164
1000 100 99.595 1.803 99.595
Open in a new tab

4. Conclusion

We demonstrated a sensitive and selective electrochemical cortisol sensor based on AuNPs-embedded cortisol imprinted polymer. The sensor was prepared using a facile one-step electropolymerization method with in situ gold reduction. The incorporation of AuNPs in the MIP layer enhanced the detection sensitivity by enlarging the active surface area of the sensor, improving electron transfer across the MIP, and providing an efficient modulation of current in response to the capture of target cortisol. The Au@MIP cortisol sensor displayed a linear range from 1 pM to 500 nM with a detection limit of 200 fM, significantly improved compared with the existing counterparts. The sensor also demonstrated satisfying recovery rates for cortisol detection in artificial saliva samples. The high-performance and robust cortisol sensor was currently demonstrated on a discrete GCE, which can hardly be miniaturized. The Au@MIP technology and sensing method can be applied to screen-printed electrodes (SPE) in the future to facilitate device integration for point-of-care applications. Moreover, the Au@MIP technique can be easily modified to selectively capture different targets to achieve multiplexed detection. The sensing performance and fabrication technique of the sensor verify that the Au@MIP has considerable potential for low-cost, rapid detection of stress levels from non-invasive body fluids.

Supplementary Material

1

Fig. 1.

Fig. 1.

Open in a new tab

Fabrication process and sensing mechanism of a cortisol sensor based on a cortisol-selective gold-doped molecularly imprinted polymer (Au@MIP).

Highlights.

  • Nano gold-doped molecularly imprinted polymer (Au@MIP) improved the transduction of cortisol binding into current change, enhancing the detection sensitivity.

  • Au@MIP fabricated by electropolymerization and in situ gold reduction increased the active surface area and conductivity of MIP.

  • The new polymerization method led to a superior detection sensitivity to the conventional AuNP embedded MIP polymerized in the presence of pre-synthesized AuNPs.

  • The effects of the MIP synthesis process, gold doping, and assay approach on sensing performance were elucidated.

  • The sensor exhibited a 200 fM detection limit and selective detection against other steroid hormones.

  • The sensor was successfully applied to determine cortisol in spiked saliva at normal and elevated levels.

Acknowledgments

The authors acknowledge partial financial support from the National Science Foundation (nos. 1810067) and National Institutes of Health (no. 1R21DE027170-01). The authors acknowledge Dr. Changqing for the help on FTIR analysis. Part of this research was conducted at Oregon Process Innovation Center, part of the Northwest Nanotechnology Infrastructure, a National Nanotechnology Coordinated Infrastructure site at Oregon State University which is supported in part by the National Science Foundation (grant NNCI-2025489) and Oregon State University.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  1. Ahmad R, Griffete N, Lamouri A, Felidj N, Chehimi MM, Mangeney C, 2015. Nanocomposites of gold nanoparticles@ molecularly imprinted polymers: chemistry, processing, and applications in sensors. Chemistry of Materials 27(16), 5464–5478. [Google Scholar]
  2. Azadmehr F, Zarei K, 2019. An imprinted polymeric matrix containing DNA for electrochemical sensing of 2, 4–dichlorophenoxyacetic acid. Microchimica Acta 186(12), 1–8. [DOI] [PubMed] [Google Scholar]
  3. Baid SK, Sinaii N, Wade M, Rubino D, Nieman LK, 2007. Radioimmunoassay and tandem mass spectrometry measurement of bedtime salivary cortisol levels: a comparison of assays to establish hypercortisolism. The Journal of Clinical Endocrinology & Metabolism 92(8), 3102–3107. [DOI] [PubMed] [Google Scholar]
  4. Beluomini MA, da Silva JL, de Sá AC, Buffon E, Pereira TC, Stradiotto NR, 2019. Electrochemical sensors based on molecularly imprinted polymer on nanostructured carbon materials: A review. Journal of Electroanalytical Chemistry 840, 343–366. [Google Scholar]
  5. Beluomini MA, da Silva JL, Sedenho GC, Stradiotto NR, 2017. D-mannitol sensor based on molecularly imprinted polymer on electrode modified with reduced graphene oxide decorated with gold nanoparticles. Talanta 165, 231–239. [DOI] [PubMed] [Google Scholar]
  6. Buffon E, Stradiotto NR, 2019. Electrochemical sensor based on molecularly imprinted poly (ortho-phenylenediamine) for determination of hexahydrofarnesol in aviation biokerosene. Sensors and Actuators B: Chemical 287, 371–379. [Google Scholar]
  7. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, Jones DW, Materson BJ, Oparil S, Wright JT Jr, 2003. The seventh report of the joint national committee on prevention, detection, evaluation, and treatment of high blood pressure: the JNC 7 report. Jama 289(19), 2560–2571. [DOI] [PubMed] [Google Scholar]
  8. Crapnell RD, Hudson A, Foster CW, Eersels K, Grinsven B.v., Cleij TJ, Banks CE, Peeters M, 2019. Recent advances in electrosynthesized molecularly imprinted polymer sensing platforms for bioanalyte detection. Sensors 19(5), 1204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dalirirad S, Steckl AJ, 2019. Aptamer-based lateral flow assay for point of care cortisol detection in sweat. Sensors and Actuators B: Chemical 283, 79–86. [Google Scholar]
  10. Edwards O, Galley J, Courtenay-Evans R, Hunter J, Tait A, 1974. Changes in cortisol metabolism following rifampicin therapy. The Lancet 304(7880), 549–551. [PubMed] [Google Scholar]
  11. El-Farhan N, Rees DA, Evans C, 2017. Measuring cortisol in serum, urine and saliva–are our assays good enough? Annals of clinical biochemistry 54(3), 308–322. [DOI] [PubMed] [Google Scholar]
  12. Gatti R, Antonelli G, Prearo M, Spinella P, Cappellin E, Elio F, 2009. Cortisol assays and diagnostic laboratory procedures in human biological fluids. Clinical biochemistry 42(12), 1205–1217. [DOI] [PubMed] [Google Scholar]
  13. Han J, Dai J, Li L, Fang P, Guo R, 2011. Highly uniform self-assembled conducting polymer/gold fibrous nanocomposites: additive-free controllable synthesis and application as efficient recyclable catalysts. Langmuir 27(6), 2181–2187. [DOI] [PubMed] [Google Scholar]
  14. Ho Y-S, McKay G, 1999. Pseudo-second order model for sorption processes. Process biochemistry 34(5), 451–465. [Google Scholar]
  15. Krasowski MD, Drees D, Morris CS, Maakestad J, Blau JL, Ekins S, 2014. Cross-reactivity of steroid hormone immunoassays: clinical significance and two-dimensional molecular similarity prediction. BMC clinical pathology 14(1), 1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Liu Y, Wu B, Tanyi EK, Yeasmin S, Cheng L-J, 2020. Label-Free Sensitive Detection of Steroid Hormone Cortisol Based on Target-Induced Fluorescence Quenching of Quantum Dots. Langmuir 36(27), 7781–7788. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Madhu S, Anthuuvan AJ, Ramasamy S, Manickam P, Bhansali S, Nagamony P, Chinnuswamy V, 2020. ZnO nanorod integrated flexible carbon fibers for sweat cortisol detection. ACS Applied Electronic Materials 2(2), 499–509. [Google Scholar]
  18. Mallick K, Witcomb MJ, Erasmus R, Scurrell MS, 2007. Hydrophilic behaviour of gold-poly (o-phenylenediamine) hybrid nanocomposite. Materials Science and Engineering: B 140(3), 166–171. [Google Scholar]
  19. Manickam P, Pasha SK, Snipes SA, Bhansali S, 2016. A reusable electrochemical biosensor for monitoring of small molecules (cortisol) using molecularly imprinted polymers. Journal of the Electrochemical Society 164(2), B54. [Google Scholar]
  20. Manivannan P, Kumar RT, Periyanayagasamy V, SPECTRAL ANALYSIS OF HYDROCORTISONE.
  21. Motia S, Bouchikhi B, El Bari N, 2021. An electrochemical molecularly imprinted sensor based on chitosan capped with gold nanoparticles and its application for highly sensitive butylated hydroxyanisole analysis in foodstuff products. Talanta 223, 121689. [DOI] [PubMed] [Google Scholar]
  22. Motia S, Bouchikhi B, Llobet E, El Bari N, 2020. Synthesis and characterization of a highly sensitive and selective electrochemical sensor based on molecularly imprinted polymer with gold nanoparticles modified screen-printed electrode for glycerol determination in wastewater. Talanta 216, 120953. [DOI] [PubMed] [Google Scholar]
  23. Mugo SM, Alberkant J, 2020. Flexible molecularly imprinted electrochemical sensor for cortisol monitoring in sweat. Analytical and bioanalytical chemistry 412(8), 1825–1833. [DOI] [PubMed] [Google Scholar]
  24. Olgun U, Gülfen M, 2014. Doping of poly (o-phenylenediamine): spectroscopy, voltammetry, conductivity and band gap energy. Reactive and Functional Polymers 77, 23–29. [Google Scholar]
  25. Papanicolaou DA, Mullen N, Kyrou I, Nieman LK, 2002. Nighttime salivary cortisol: a useful test for the diagnosis of Cushing’s syndrome. The Journal of Clinical Endocrinology & Metabolism 87(10), 4515–4521. [DOI] [PubMed] [Google Scholar]
  26. Parlak O, Keene ST, Marais A, Curto VF, Salleo A, 2018. Molecularly selective nanoporous membrane-based wearable organic electrochemical device for noninvasive cortisol sensing. Science advances 4(7), eaar2904. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Raff H, Findling JW, 2003. A physiologic approach to diagnosis of the Cushing syndrome. Annals of Internal Medicine 138(12), 980–991. [DOI] [PubMed] [Google Scholar]
  28. Raff H, Raff JL, Findling JW, 1998. Late-night salivary cortisol as a screening test for Cushing’s syndrome. The Journal of Clinical Endocrinology & Metabolism 83(8), 2681–2686. [DOI] [PubMed] [Google Scholar]
  29. Ray P, Steckl AJ, 2019. Label-free optical detection of multiple biomarkers in sweat, plasma, urine, and saliva. ACS sensors 4(5), 1346–1357. [DOI] [PubMed] [Google Scholar]
  30. Reddy KK, Gobi KV, 2013. Artificial molecular recognition material based biosensor for creatinine by electrochemical impedance analysis. Sensors and actuators B: chemical 183, 356–363. [Google Scholar]
  31. Sanati A, Siavash Moakhar R, Hosseini II, Raeissi K, Karimzadeh F, Jalali M, Kharaziha M, Sheibani S, Shariati L, Presley JF, 2021. Gold nano/micro-islands overcome the molecularly imprinted polymer limitations to achieve ultrasensitive protein detection. ACS sensors 6(3), 797–807. [DOI] [PubMed] [Google Scholar]
  32. Sehit E, Drzazgowska J, Buchenau D, Yesildag C, Lensen M, Altintas Z, 2020. Ultrasensitive nonenzymatic electrochemical glucose sensor based on gold nanoparticles and molecularly imprinted polymers. Biosensors and Bioelectronics 165, 112432. [DOI] [PubMed] [Google Scholar]
  33. Sen S, Sarkar P, 2015. A novel third-generation xanthine biosensor with enzyme modified glassy carbon electrode using electrodeposited MWCNT and nanogold polymer composite film. RSC advances 5(116), 95911–95925. [Google Scholar]
  34. Steckl AJ, Ray P, 2018. Stress biomarkers in biological fluids and their point-of-use detection. ACS sensors 3(10), 2025–2044. [DOI] [PubMed] [Google Scholar]
  35. Sun X, Dong S, Wang E, 2004. Large scale, templateless, surfactantless route to rapid synthesis of uniform poly (o-phenylenediamine) nanobelts. Chemical Communications(10), 1182–1183. [Google Scholar]
  36. Tan T, Khoo B, Mills EG, Phylactou M, Patel B, Eng PC, Thurston L, Muzi B, Meeran K, Prevost AT, 2020. Association between high serum total cortisol concentrations and mortality from COVID-19. The Lancet Diabetes & Endocrinology 8(8), 659–660. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Tu E, Pearlmutter P, Tiangco M, Derose G, Begdache L, Koh A, 2020. Comparison of colorimetric analyses to determine cortisol in human sweat. ACS omega 5(14), 8211–8218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Umpleby RJ, Baxter SC, Chen Y, Shah RN, Shimizu KD, 2001. Characterization of molecularly imprinted polymers with the Langmuir− Freundlich isotherm. Analytical chemistry 73(19), 4584–4591. [DOI] [PubMed] [Google Scholar]
  39. Vlatakis G, Andersson LI, Müller R, Mosbach K, 1993. Drug assay using antibody mimics made by molecular imprinting. Nature 361(6413), 645–647. [DOI] [PubMed] [Google Scholar]
  40. Wang L, Guo S, Dong S, 2008. Facile synthesis of poly (o-phenylenediamine) microfibrils using cupric sulfate as the oxidant. Materials Letters 62(17–18), 3240–3242. [Google Scholar]
  41. Whitworth JA, Mangos GJ, Kelly JJ, 2000. Cushing, cortisol, and cardiovascular disease. Hypertension 36(5), 912–916. [DOI] [PubMed] [Google Scholar]
  42. Wu B, Yeasmin S, Liu Y, Cheng L-J, 2022. Sensitive and selective electrochemical sensor for serotonin detection based on ferrocene-gold nanoparticles decorated multiwall carbon nanotubes. Sensors and Actuators B: Chemical 354, 131216. [Google Scholar]
  43. Zea M, Bellagambi FG, Halima HB, Zine N, Jaffrezic-Renault N, Villa R, Gabriel G, Errachid A, 2020. Electrochemical sensors for cortisol detections: Almost there. TrAC Trends in Analytical Chemistry, 116058. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1787057-supplement-1.pdf (492.3KB, pdf)

RESOURCES