CoVSense: Ultrasensitive Nucleocapsid Antigen Immunosensor for Rapid Clinical Detection of Wildtype and Variant SARS‐CoV‐2

Abstract The widespread accessibility of commercial/clinically‐viable electrochemical diagnostic systems for rapid quantification of viral proteins demands translational/preclinical investigations. Here, Covid‐Sense (CoVSense) antigen testing platform; an all‐in‐one electrochemical nano‐immunosensor for sample‐to‐result, self‐validated, and accurate quantification of the severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) nucleocapsid (N)‐proteins in clinical examinations is developed. The platform's sensing strips benefit from a highly‐sensitive, nanostructured surface, created through the incorporation of carboxyl‐functionalized graphene nanosheets, and poly(3,4‐ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) conductive polymers, enhancing the overall conductivity of the system. The nanoengineered surface chemistry allows for compatible direct assembly of bioreceptor molecules. CoVSense offers an inexpensive (<$2 kit) and fast/digital response (<10 min), measured using a customized hand‐held reader (<$25), enabling data‐driven outbreak management. The sensor shows 95% clinical sensitivity and 100% specificity (Ct<25), and overall sensitivity of 91% for combined symptomatic/asymptomatic cohort with wildtype SARS‐CoV‐2 or B.1.1.7 variant (N = 105, nasal/throat samples). The sensor correlates the N‐protein levels to viral load, detecting high Ct values of ≈35, with no sample preparation steps, while outperforming the commercial rapid antigen tests. The current translational technology fills the gap in the workflow of rapid, point‐of‐care, and accurate diagnosis of COVID‐19.


S6-1. Principle of impedance measurement
In the low complexity potentiostat readout system developed and used in this work, the impedance is measured based on equations (1) and (2).
where α represents the gain related to the implemented circuit, P is the period of the excitation or response signals seen in Figure S8. A and B are amplitudes of excitation and response signals. The required sinusoidal excitation signal is created by the microcontroller using a 10 bits DAC channel. The response signal is the signal recorded by the microcontroller using a 10 bits ADC channel. Figure S8. Illustration of excitation and response signals.

S6-2. Potentiostat readout system
The handheld potentiostat readout system developed and used in this work consists of a microcontroller unit, an analog front-end circuitry, and a graphical user interface (GUI) unit as described below in the block diagram shown in Figure S9. Figure S9. Overview of the block diagram of the handheld potentiostat system.

S6-2-1. Microcontroller Unit
The microcontroller unit (MCU) breakout is the Arduino DUE with an ARM Cortex-M3 32-bit processor (64 MHz, two 256 kbytes embedded Flash, 128-bit wide access, memory accelerator, dual bank 100 (64+32) Kbytes embedded SRAM with dual banks). The USB interface allows for programming the MCU, powering the unit, and communicating through UART. The breakout can also be powered by a Lithium-Polymer (LiPo) battery and charged using the same USB interface. The firmware was developed using Arduino IDE.
The firmware of the MC was developed in C program to control the digitally controlled resistors in the analog front-end circuitry and generate excitation and record response sinusoidal signals. To decrease the error, the excitation and output sinusoidal signals were smoothened using a moving average operator of the buffer size of 11 samples. To further suppress the effect of noise, each measurement was repeated 20 times and the output signal was obtained by averaging the 20 measurements.

S5-2-2. Analog Front End circuitry
The Analog Front End circuitry consists of Low Pass Filter (LPF), Voltage Attenuator, and Potentiostat as described and demonstrated below.
Low-pass filter. A Sallen-Key LPF with variable cut-off frequency [1] was developed with two digitally controlled resistors via SPI serial interface. The microcontroller enables tuning the filter at seven different frequencies of 1, 3, 11, 37, 125, 412, and 1390 Hz.
Voltage attenuator. The excitation signal is a sinusoidal signal with an amplitude of 10 mV peak-to-peak. Therefore, to generate a small signal voltage from the output signal of LPF, a voltage attenuator (1/20) was developed using an inverting operational amplifier (OPAMP).

Potentiostat.
A standard potentiostat circuitry consisting of three OPAMPs was employed as shown in Figure 10. The digitally controlled feedback resistor (FB) via SPI serial interface enables the operation of the impedance measurement system in seven different ranges. A switch (SW) is used to select the first and second working electrodes (WE1 or WE2). Figure S10. The circuit diagram of the implemented three-electrode potentiostat.

S21
Printed circuit board design and implementation. The analog front was developed in a PCB using Altium Designer software with the layout shown in Figure S11. In this layout, the PCB ground plane was split into an analog plane (AGND) and a digital plane (DGND) to reduce the noise coupled from digital circuits into the analog signals. Analog and digital components were separated and their corresponding signal traces were placed right above their respective ground planes. In this way, the signals return to their ground planes by vias. This PCB was manufactured by www.seeedstudio.com ( Figure S12). A list of elements used for fabricating this PCB board is presented in Table S5.

S6-2-3. Graphical user interface
A Raspberry pi board connected to a 3.5-inch touchscreen LCD was used as an easy and friendly user interface, with the capability of wireless internet and Cloud storage. The Raspberry pi computer communicates with the EIS readout system through a serial port and controls the LCD via GPIO pins. A python-based App was developed and run under a customized Raspbian OS to make data collection, storage, and analysis easier, while the GUI provides instructions to collect the sample, run the test, and show the results.

S6-2-4. Full readout system
The readout system consists of a developed PCB board incorporated with Arduino and Raspberry pi boards and LCD and as seen in Figure S13. Figure S13. The integration of developed PCB and other boards and the fully packaged readout system.

S6-3-1. Testing the handheld impedimetric reader with known Resistor and capacitor values
For the detection of COVID-19 patients, measurement of the impedance in five different frequencies ranging logarithmically from 1 Hz to 1.4 kHz was proven to be sufficient based on the tests of the immunosensor with Metrohm Autolab measurement device. The results were then validated against the Autolab PGSTAT204 (Metrohm), with FRA32M electrochemical impedance spectroscopy (EIS) module and NOVA software. These frequencies were observed to be sufficient to uncover the impedance difference between the SARS-CoV-2 positive and negative clinical samples. Functionality assessment of the handheld reader was performed by testing them in the selected logarithmically-spaced frequencies. The functionality assessment was performed using the known resistors and capacitors combination ( Figure S14) by repeating each experiment 5 times. The repeatability was significant with only 1% variation among all the runs. As a result, the worst-case error between the measurements using the proposed readout system and the AutoLab device as an industrial and research-grade tool was below 20% in some frequencies. Figure S14. Test comparison between the proposed handheld reader and AutoLab for the known resistor and capacitor combinations; (A) a resistor (10 kΩ) in parallel with a capacitor (10 µF), (B) a resistor (10 kΩ) in parallel with a capacitor (1µF). Measured frequencies are 1, 3, 5, 10, 10, 30, and 1400 Hz.

S6-3-2. Redox testing of the proposed reader
The tests with the redox solution were performed first by sweeping the frequencies in the left electrode (L), followed by the right electrode (R). The results show similarity in high frequencies with more errors in low frequencies. The higher errors at lower frequencies are attributed to the limited phase measure resolution to the number of ADC values and the time-based phase measurement ( Figure S15).

S6-4. Comparison of the handheld potentiostat reader and the state-of-the-art potentiostat
We addressed the challenge of developing a low-cost but highly accurate impedimetric-based readout system using low complexity, yet programmable system suitable for performing an efficient calibration. The electrical characterization and biochemical testing results proved the functionality, sensitivity, and selectivity of this reader. The proposed low-cost readout system was successful in detecting COVID-19-positive patients. Table S6 compares the performance of our potentiostat reader once used in combination with our biosensor and other state-of-the-art electrochemical biosensors used for the detection of different SARS-CoV-2 biomarkers. The key benefits of our reader are its cost-effectiveness, fast measurement time, and the capability of transmitting the data to the Cloud.