Zwitterionic polymer electroplating facilitates the preparation of electrode surface for biosensing

Abstract

Surface chemistry critically affects the diagnostic performance of biosensors. An ideal sensor surface should be resistant to nonspecific protein adsorption, yet be conducive to analytical responses. Here we report a new polymeric material, zwitterionic polypyrrole (ZiPPy), to produce optimal surface condition for biosensing electrodes. ZiPPy combines merits from two unique precursors: the zwitterionic function that efficiently hydrates electrode surface, hindering nonspecific binding of hydrophobic proteins; and the pyrrole backbone which enables rapid (<7 min), controlled deposition of ZiPPy through electropolymerization. ZiPPy-coated electrodes showed lower electrochemical impedance and less non-specific protein adsorption (low fouling), outperforming bare and polypyrrole-coated electrodes. Moreover, affinity ligands for target biomarkers could be immobilized together with ZiPPy in a single-step electropolymerization. We developed ZiPPy-coated electrodes with specificity for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The prepared sensor detected SARS-CoV-2 antibodies in human saliva down to 50 ng mL⁻¹, without need for sample purification or secondary labeling.

Graphical Abstract

Zwitterionic polypyrrole (ZiPPy) is applied to produce the optimal surface for biosensing electrodes. ZiPPy combines merits of two components: the zwitterionic function efficiently hydrates electrode surface to minimize non-specific protein adsorption, and the pyrrole backbone enables rapid, controlled deposition of ZiPPy through electropolymerization. ZiPPy-coated electrodes are used to detect neutralizing antibodies against SARS-CoV-2 in human saliva.

1. Introduction

Electrochemical sensing and electrical assays are powerful methods for on-site point-of care diagnostics.^[1-3] Electrochemical sensing is generally fast, and the signal can be read out by compact devices.^[4-7] Electrodes play a critical role in this type of assay, converting molecular information into electrical signals. However, as electrodes come in contact with physiological fluids, they often suffer from fouling artifacts — the interference of specific signaling by nonspecific protein absorption. Commonly abundant proteins (e.g., human serum albumin) flood the electrode surface and interfere with specific protein-protein binding, causing false positives and a lower signal-to-background ratio.^[8] The issue can be mitigated by purifying samples before measurements, although adding pre-analytical steps can complicate assays, require more sample amounts, and often results in loss of target biomarkers. An alternative workaround is to surface-treat electrodes with non-fouling polymers.^[9] A list of organic polymers are available to minimize nonspecific molecular adsorption, including poly(ethylene glycol) and its derivatives,^[10] polyamides, polyurethanes, and naturally occurring polysaccharides such as dextran or chitosan.^[11] For biosensing applications, however, polymers need to meet other requirements as well: they should not impede electrochemical measurements, and the deposition process should be controllable for reproducibility. Charged polymers are non-ideal, as they can hinder ion migrations towards electrodes. Zwitterionic molecules such as sulfo- or carboxybetaines can be used to treat electrodes for antifouling with minimal impact on electrochemical reactions,^[12] but growing them into thin layers on sensor surfaces remains chemically challenging.^[13]

Electrochemical deposition of conducting polymers is a widely used method to modify sensor electrodes.^[14,15] Electropolymerization can simplify the incorporation of biomolecules on electrodes. For example, biomolecules can be mixed with monomer solutions for co-deposition, or they can be bound to polymers through various surface chemistry strategies.^[16]. Among conducting polymer candidates, polypyrrole (PPy) has been extensively used for its biocompatibility and ease of functionalization.^[16,17]. Unfortunately, PPy is positively charged and susceptible to nonspecific protein adsorption.^[18] Additional chemical modifications are often necessary to minimize biofouling on Pay surface.^[18,19]

Here we present a one-step strategy to prepare an anti-fouling biosensing electrode, termed ZiPPy (zwitterionic polypyrrole). We hypothesized dual benefits of zwitterionic PPy: i) the polymer is highly hydrated, thereby minimizing nonspecific adsorption of hydrophobic proteins;^[20] and ii) the hydration also makes it easier for ions to migrate through the polymer layer (i.e., high ion conductivity),^[21,22] which benefits electrochemical detection. To prove the concept, we thus synthesized a new zwitterionic-pyrrole derivative that combines the advantages of these two precursor classes. The developed ZiPPy-coated electrodes indeed showed superior hydrophilicity and considerably lower protein adsorption when compared to non-coated or PPy-coated electrodes. Preparing ZiPPy-coated electrodes for biosensing was also streamlined: i) the coating process (electropolymerization) was fast (7 min) and easy to control; and ii) affinity ligands, such as antibodies or other proteins, could be co-immobilized during the polymer coating.

As a biosensing application, we tuned the ZiPPy system to detect neutralizing antibodies against SARS-CoV-2 in human saliva samples. SARS-CoV-2 infection or vaccination induces the production of neutralizing antibodies whose presence is a good indicator of the reduced risk of future infection.^[23,24] Neutralizing antibodies are typically measured in serum, which involves blood draw and sample transport to specialized laboratories. We reasoned that saliva could be an appealing test medium: i) samples can be non-invasively collected; and ii) saliva has been shown to carry neutralizing antibodies against SARS-CoV-2, with the salivary antibody concentrations correlating with those in serum.^[25,26] To detect neutralizing antibodies, we functionalized a set of electrodes, each containing spike (S) proteins of SARS-CoV-2 wild type, B.1.7 variant, or B.1.351 variant. We then used the ZiPPy-coated electrodes to analyze native saliva samples from individuals who have received (n = 24) or have not received (n = 7) coronavirus disease 2019 (COVID-19) vaccines.

2. Results and Discussion

2.1. ZiPPy strategy

Figure 1 summarizes the assay workflow. We first synthesized zwitterionic pyrrole (ZiPy) monomers in a two-step reaction (Figure 1a) (see Experimental Section for details). The synthesis of ZiPy was confirmed via nuclear magnetic resonance (NMR) measurements and mass spectrometry (Figure S1), which also demonstrated the high purity of ZiPy monomers (Figure S2). We next coated bare electrodes (carbon or gold) with ZiPy via electropolymerization (Figure 1b). ZiPy monomer solution was drop-casted on electrodes, and electrical potential was applied to polymerize ZiPy into ZiPPy (Experimental Section). The experimental condition for ZiPY polymerization (e.g., electrolyte, potential range) was identical to that of Py, suggesting that ZiPY shared the similar polymerization mechanism as Py: i) oxidation of monomers on the anode; ii) polymer growth; and ii) the incorporation of anions from electrolyte.^[27-29] At this coating step, we mixed affinity ligands in the ZiPy monomer solution, which allowed us to entrap the ligands in the growing polymer brush. Compared to other bioconjugation methods (e.g., covalent coupling, physisorption), this approach enabled a rapid, one-step functionalization of biosensors.

Figure 1.

Preparing biosensor electrodes using zwitterionic polypyrrole (ZiPPy). a) Synthesis of zwitterionic pyrrole (ZiPy) monomers. Pyrrole is reacted with 3-dimethylaminopropylchloride hydrochloride to produce N,N-dimethyl-3-(1H-pyrrol-1-yl)propan-1-amine (1). The compound is then converted to ZiPy (2) upon reaction with β-propiolactone in THF. b) Electrode functionalization. A mixture of affinity ligands and ZiPy monomers is drop-cast on electrodes. Antifouling coating embedded with affinity ligands is created by electropolymerizing ZiPy into ZiPPy through cyclic voltammetry. c) Illustrative example showing how ZiPPy-coated electrodes can be used for label-free biosensing. Antibodies (target biomarker) present in saliva are captured by different antigens immobilized on ZiPPy-coated electrodes. The binding events change electrochemical impedance. A ferri- and ferrocyanide couple is used as a redox label.

The prepared sensors can detect molecular targets directly from biological samples (Figure 1c). Samples are simply incubated on the electrodes. Binding of targets to affinity ligands results in changes in electrochemical impedance, and the net impedance difference before and after target binding can be used as an analytical metric. Overall, the ZiPPy approach offers practical advantages: i) sensors can be rapidly prepared (<7 min) without requiring lengthy chemical modifications; ii) the antifouling surface improves detection specificity; and iii) the assay requires no additional labeling for signal generation (i.e., label-free).

2.2. Surface characterization

We first characterized the surface properties of ZiPPy electrodes. As a comparative control, a separate set of electrodes were coated with PPy. Atomic force microscopy (AFM) of a large area (10 × 10 μm²) confirmed extended polymer coating on flat gold electrodes (Figure 2a). The surface roughness was similar between PPy and ZiPPy coatings (Figure S3). The root mean square roughness values were 78.1 ± 17.6 nm (ZiPPy) and 67.3 ± 2.4 nm (PPy), and were statistically non-different (P = 0.57, unpaired t test). The maximum peak-to-valley values were also statistically identical (P = 0.94, unpaired t test) with 465 ± 93 nm for ZiPPy and 458 ± 39 nm for PPy. Carbon electrodes showed similar morphological features between PPy and ZiPPy coatings as well (Figure S4).

Figure 2.

Surface characterization of PPy and ZiPPy coating. a) The surface morphology of bare, PPy-coated, and ZiPPy-coated gold electrode was measured by atomic force microscopy. PPy and ZiPPy coatings had a similar roughness, with the root mean square roughness values of 67.3 ± 2.4 nm (PPy) and 78.1 ± 17.6 nm (ZiPPy). b) Fourier transform infrared spectra of bare, PPy-coated, and ZiPPy-coated gold electrodes. Both PPy and ZiPPy coatings showed vibration characteristics of oxidized PPy. The ZiPPy coating displayed additional peaks coming from the terminal carboxylic acid. c) Photos of water drops on bare, PPy-coated, and ZiPPy-coated electrodes. The contact angle between an electrode and a water-air interface was measured. d) Comparison of contact angles with different coatings. Bare electrodes had high contact angles, and the values were significantly different (**P = 0.0014) between gold and carbon electrodes. With the polymer coating, contact angles decreased and became similar regardless of the electrode type (P = 0.64, PPy coating; P = 0.28, ZiPPy coating). Electrodes with the ZiPPy coating had the smallest contact angle and thereby the highest hydrophilicity. Data is displayed as mean ± s.d. from technical triplicates. Two-sided unpaired t-test was used for statistics comparison.

Fourier transform infrared (FTIR) spectroscopy further confirmed the presence of polymer matrices on PPy- and ZiPPy-coated electrodes. Both electrodes showed vibration characteristics of oxidized PPy (Figure 2b):^[30] peaks at 817 cm⁻¹ (C─H out of plane ring deformation), 943 cm⁻¹ (C═C bending), 1122 cm⁻¹ (breathing vibration of the pyrrole ring), 1130-1162 cm⁻¹ (C─N stretching), 1314-1379 cm⁻¹ (C─H and/or C─N in plane deformation, ring stretching), 1469 cm⁻¹ (pyrrole ring vibration). The ZiPPy coating showed additional C─O (1277 cm⁻¹) and C═O (1740 cm⁻¹) stretching bands that came from the terminal carboxylic acid.^[31] Surface adsorbed water molecules were also visible at 1610 cm⁻¹, which supported the high hydrophilicity of ZiPPy.^[32]

We next evaluated the surface hydrophilicity of coated electrodes by measuring water contact angles. Increasing hydrophilicity is desirable to enhance the antifouling effect, as it reduces the protein adsorption by hindering hydrophobic interactions.^[33] We imaged water drops on carbon and gold electrodes with PPy or ZiPPy coatings or without coating (Figure 2c). The polymer-coated electrodes had contact angles (between the electrode surface and the water-air interface) smaller than bare electrodes (Figure 2d).^[34] For a given coating type, contact angles were statistically identical regardless of the substate type, which suggested effective coverage of bare electrodes with polymers. The ZiPPy coating was most efficient in reducing the contact angle (i.e., the highest hydrophilicity), indicating its potential for better protein repulsion among three surface types tested.^[35]

2.3. Electrochemical properties

To characterize the electrochemical performance of bare, PPy-, and ZiPPy-coated electrodes, we recorded cyclic voltammograms (scan rate, 100 mV s⁻¹) in the presence of K₃(FeCN)₆ (Figure 3a). The ratio between anodic (I_PA) and cathodic (I_PC) peak currents was close to unity, indicating that all electrodes supported quasi-reversible redox conversion between [Fe(CN)₆]³⁻ and [Fe(CN)₆]⁴⁻. However, ZiPPy coating had further desirable properties. The PPy-coated electrode showed the highest peak current, which was due the capacitive charging effect of anionic PPy.^[28] The ZiPPy-coated electrode, on the other hand, displayed cyclic voltammograms similar to those of bare electrodes, indicating a minimal capacitive effect. Minimizing the charging layer would facilitate the charge transport between electrolyte and electrodes, thereby improving the sensitivity of electrochemical sensors. We further measured the peak potential separation (i.e., potential difference between anodic and cathodic peaks): 0.41 V (bare electrode), 0.39 V (PPy-coated electrode), and 0.16 V (ZiPPy-coated electrode). The ZiPPy-coated electrode had the smallest peak separation, which also confirmed that it was most efficient in electron transfer.^[36]

Figure 3.

Electrochemical properties of PPy-coated and ZiPPy-coated electrodes. a) Cyclic voltammograms of [Fe(CN)₆]^3−/4− were recorded using bare, PPy-coated, and ZiPPy-coated carbon electrodes. The ZiPPy-coated electrode had less capacitive charging effect and was more efficient in electron transfer. b) Electrochemical impedances (Z_Bare, Z_PPy, Z_ZiPPy) of bare, PPy-coated, and ZiPPy-coated carbon electrodes were measured over the frequency range of 10⁻¹ to 10⁵ Hz, and the real (Z_Re) and the imaginary (Z_Im) values were plotted. The charge transfer resistance (R_ct), estimated from the Randles circuit model (Figure S5), was the smallest with the ZiPPy-coated electrode. c) Bode-impedance diagrams. The ZiPPy-coated electrode had the lowest impedance over a wide range of frequency. d) The relative differences of Z_PPy and Z_ZiPPy against Z_Bare were plotted. The differences were stable around 10 Hz (the green rectangle), which was set as the measurement frequency.

We next measured electrochemical impedance of bare, PPy-, and ZiPPy-coated carbon electrodes in the presence of K₃(FeCN)₆, and generated Nyquist plots (Figure 3b). The bare electrode showed a typical Randles-circuit behavior characterized by a charge transfer resistance (R_ct = 982 Ω), solution resistance, double layer capacitance, and diffusion-related Warburg impedance (see Figure S5 for the circuit model). Upon PPy or ZiPPy coating, however, R_ct values decreased to 72 Ω (PPy) and 43 Ω (ZiPPy), with ZiPPy coating showing the lowest charge transfer resistance. This observation was further supported by Bode-impedance plots which showed a significant change in total impedance upon polymer modification (Figure 3c). Both PPy and ZiPPy matrices had a porous structure that likely trapped electrolyte ions and thereby improved electrical conductivity. The higher conductivity of ZiPPy can be attributed to its zwitterionic nature: charge neutral ZiPPy minimized the formation of the double layer that could slower ion migrations.^[21,37] Gold electrodes, coated with PPy or ZiPPy, displayed the similar results (Figure S6). For the impedance assay, we set the measurement frequency at 10 Hz wherein the impedance differences between coated and bare electrodes were stable (Figure 3d).

2.4. Performance in biological media

We next tested the antifouling performance of bare, PPy, and ZiPPy-coated electrodes. We first used a solution of fluorescent streptavidin (1 mg mL⁻¹) as a fouling media. Electrodes were incubated with the solution (1 h, 20 °C), washed with buffer, and imaged for adsorbed fluorescent streptavidin (Figure 4a). Polymer-coated electrodes all displayed antifouling capacity: the observed fluorescent intensities were lower than to those of bare electrodes, with ZiPPy (54% reduction) outperforming PPy coating (64% reduction). Protein adsorption assays showed a similar same trend (Figure 4b). We incubated electrodes (1 h, 20 °C) with a 10% bovine serum albumin (BSA) solution (125 μg mL⁻¹), and then quantified proteins remaining in supernatant. The ZiPPy coating had the least amount of proteins adsorbed on electrodes (38% of the initial amount, 0.189 μg mm⁻²), followed by the PPy coating (61% of the initial loading, 0.303 μg mm⁻²). With bare gold electrodes, the adsorption ratio reached up to 86% (0.428 μg mm⁻²).

Figure 4.

Antifouling study. a) Bare, PPy-coated, and ZiPPy-coated electrodes were incubated with fluorescently labeled streptavidin solution, washed, and imaged. The ZiPPy-coated electrode had the lowest fluorescent intensity from adsorbed streptavidin. The inset shows raw images. The graph displays mean ± s.d. of intensity values from 15 different field-of-views. b) Electrodes were incubated with 10% BSA, and the remaining BSA in supernatant was quantified. The ZiPPy-coated electrode had the lowest BSA adsorption. Data is displayed as mean ± s.d. from quadruplicate measurements. c) Raw data showing electrochemical impedance of a ZiPPy-coated electrode before and after incubation with human saliva. d) Relative changes of electrochemical impedance before and after incubation with human serum, saliva, or 10% BSA. Impedance values were measured at 10 Hz. ZiPPy-coated electrodes had the lowest fouling. Data is shown as mean ± s.d. from triplicate measurements.

We further challenged electrodes with unaltered human serum and saliva samples, monitoring the change of electrochemical impedance before and after sample incubation (Figure 4c, Figure S7). The ZiPPy-coated electrodes had the best performance against all types of media (Figure 4d). The impedance change was minimal (6 – 11%), compared to PPy-coated (8 – 32%) and bare electrodes (16 – 84%).

2.5. Profiling of saliva samples for SARS-CoV-2 antibodies

We finally applied the ZiPPy system to detect neutralizing antibodies against SARS-CoV-2 in human saliva samples. We functionalized ZiPPy-coated electrodes with spike (S) viral proteins. Three sensing electrodes were prepared (Figure 5a), with each electrode having S proteins specific to a different SARS-CoV-2 type: wild type, B1.1.7 (alpha) variant, and B.1.351 (beta) variant. We mixed ZiPy monomers with spike proteins and co-deposited the mixture onto electrodes via electropolymerization. About 330 ng (≈ 4.3 pmol) of S proteins were estimated to be bound per working electrode (4 mm in diameter; see Experimental Section), and the protein binding was stable with negligible leaching of bound S proteins into buffer (Figure S8).

Figure 5.

Detection of SARS-CoV-2 antibodies in saliva. a) A set of carbon electrodes were prepared to detect antibodies against SARS-CoV-2. Each electrode was coated with ZiPPy along with spike (S) proteins of either SARS-CoV-2 wild type, B.1.17 variant, or B.1.351 variant. Saliva samples were directly applied to ZiPPy-coated electrodes for electrochemical impedance measurements. b) Saliva samples spiked with varying amounts of SARS-CoV-2 IgG antibodies were measured by ZiPPy-coated electrodes specific to SARS-CoV-2 wild type. The limit of detection was about 50 ng mL⁻¹ (IgG). Data is shown as mean ± s.d. from technical triplicates. c) Saliva samples from individuals who were vaccinated (n = 24) or non-vaccinated (n = 7) against SARS-CoV-2 were analyzed. The effective antibody concentrations were estimated from titration curves. d) The amount of anti-SARS-CoV-2 neutralizing antibodies was significantly higher (p < 0.0001, unpaired two-sided t-test) in vaccinated than non-vaccinated cohorts. e) Measurement data (n = 24) were divided into three groups according to the elapsed days after the second vaccination dose. For a given SARS-CoV-2 type, the average antibody concentrations showed no significant temporal changes (one-way ANOVA).

We first assessed the analytical sensitivity by generating a titration curve. Samples were prepared by spiking varying amounts of SARS-CoV-2 IgG antibodies in saliva from a donor with no history of SARS-CoV-2 infection nor SARS-CoV-2 vaccination. Electrochemical impedances were measured before (Z₀) and after (Z₁) incubating test samples on electrodes, and the net change (ΔZ = Z₁ − Z₀) was used as an analytical metric. We observed the increase of ΔZ as more target antibodies bound to the sensor (Figure 5b). With the sensor for SARS-CoV-2 wild type, the limit of detection was about 50 ng mL⁻¹ (IgG) and the dynamic range spanned three orders of magnitude. This sensing range encompassed the reported IgG range in saliva (total IgG concentration ~15 μg mL⁻¹).^[38] Experiments performed with control IgG samples showed negligible impedance changes, and ZiPPy sensors without S protein also showed negligible signal; these results confirmed the SARS-CoV-2 selectivity of prepared electrodes (Figure S9).

We next analyzed human saliva samples collected from vaccinated (n = 24) and non-vaccinated (n = 7) individuals (see Table S1 for demographic data). All subjects in the vaccinated cohort finished their two-dose regimen of either AZD1222 (AstraZeneca; n = 1), BNT162b2 (Pfizer; n = 12), or mRNA-1273 (Moderna; n = 11). The non-vaccinated, asymptomatic and COVID-negative participants served as a comparative negative control group to set the signal baseline. For a given saliva sample, we split it into three aliquots and tested them for the reactivity against SARS-CoV-2 wild type and two other variants (B.1.17, B.1.351). Anti-SARS-CoV-2 neutralizing antibodies were present in saliva of all vaccinated cohort (Figure 5c). The vaccination, regardless of the vaccine type, was found to be effective against SARS-CoV-2 wild type and variants (Figure 5d, Figure S10). We further divided the vaccinated cohort into three groups according to relapsed days after the second vaccine shot: ≤50 days; 51 – 100 days; 101 – 150 days. The antibody levels showed no significant difference among these groups (Figure 5e). The results confirmed immune protection of vaccines against SARS-CoV-2 and the presence of SARS-CoV-2 neutralizing antibodies in saliva.^[38,39]

3. Conclusions

By combining the merits of two precursors, zwitterionic molecules and conducting polymers, the ZiPPy approach provides an efficient way to rapidly produce high-performance biosensors. The zwitterionic nature of ZiPPy hydrates the sensor surface through strong electrostatic interactions, thereby rendering the surface highly resistant to protein adsorption. Using the conducting polymer, pyrrole, as a backbone markedly simplifies the surface treatment – ZiPPy can be coated on a sensing device through electropolymerization in a controlled manner, and affinity ligands for biosensing can be deposited together. To prove the concept, we designed a zwitterionic pyrrole derivative and applied the material to prepare electrodes for electrochemical impedance measurements. We could coat electrodes with ZiPPy within 7 min. The coated electrodes not only suppressed non-specific protein binding, but also minimized the double-layer effect to lower the overall electrochemical impedance; these advantages enabled the ZiPPy system to achieve high selectivity and sensitivity. Using the ZiPPy sensor, we could detect SARS-CoV-2 antibodies in saliva down to 50 ng mL⁻¹, without need for sample purification or a secondary labeling. A capacity to detect antibodies in saliva on-site would facilitate wide-spread and frequent monitoring of SARS-CoV-2 immunity with minimal hesitancy.

The ZiPPy approach could be further developed to achieve even higher antifouling properties. Since the non-specific protein adsorption is inversely related to the hydrophilicity of the material surface, one may consider incorporating branched or dendritic structures to enhance the hydrophilicity of monomers. Such structures could be synthesized by employing a cross linker or incorporating multiple zwitterionic units into monomers. Keeping the pyrrole backbone would be desirable to keep using electrodeposition as a one-step strategy for electrode-coating and bioconjugation.

Experimental Section

Unless otherwise mentioned, all materials were used as received. Tetrabutylammonium bromide (TBAB, 99.0 %), sodium hydroxide (NaOH, 50wt%), anhydrous tetrahydrofuran (THF, ≥99.9%), potassium ferrocyanide (K₄[Fe(CN)₆], ≥98.5%), potassium ferricyanide (K₃[Fe(CN)₆], ≥99%), and potassium chloride (KCl, ≥99.0%) were purchased from Sigma-Aldrich. β-propiolactone (95%) and anhydrous magnesium sulfate (MgSO₄, ≥99.5%) were received from Alfa Aesar and dimethyl sulfoxide (DMSO, >99.0%), pyrrole (99%) and 3-dimethylaminopropylchloride hydrochloride (98%) were purchased from TCI. Methanol (MeOH, ≥99.8%) and diethyl ether (99.0%) were ordered from VWR. D₂O and CDCl₃ were purchased from Cambridge Isotope Laboratories (USA). Carbon (Cat. #C110) and gold screen printed (Cat. #220 AT) electrodes were purchased from Metrohm DropSens (Spain). Both electrode types have the same dimensions: 3.4 × 1.0 × 0.05 cm³ (length × width × height). Reaction mixtures were purified using a Biotage SNAP Bio C18 25 μm 60 g on a Biotage Isolera with a gradient composed of water and methanol for reversed-phase chromatography. In all cases, water refers to MilliQ water with a resistivity of 18.2 MΩ cm⁻¹ at 25 °C. ¹H and ¹³C NMR spectra were recorded on a Bruker AC-400 MHz spectrometer. Electrospray ionization mass spectra were recorded using a Waters 3100 Mass Detector.

Zwitterionic pyrrole synthesis

Zwitterionic pyrrole (2) was synthesized in a two-step process by first preparing a pyrrole amine (1) derivative which was transferred into a zwitterionic form through a ring-opening reaction with β-propiolactone (Figure 1a).

Synthesis of N,N-dimethyl-3-(1H-pyrrol-1-yl)propan-1-amine (1)

Pyrrole (1 mL, 14 mmol) was dissolved in dimethyl sulfoxide (20 mL) under argon. TBAB (35 mg, 0.1mmol) was then added, followed by the addition of 50wt% NaOH (6.5 mL). Water (0.5 mL) was added to allow for a better mixing before dropwise addition of 3-dimethylaminopropylchloride hydrochloride (5 g, 32 mmol) dissolved in DMSO (7 mL). This mixture was stirred for 24 h and water (20 mL) was added to dissolve the precipitate. The product was extracted with diethyl ether (3 × 25 mL) and the organic phase was washed with water (2 × 25 mL) and brine (1 × 25 mL). The organic phase was dried over MgSO₄ and the product was obtained in 70% yield as a colorless liquid upon removal of the solvent under vacuum. This product was used without further purification. ¹H NMR (CDCl₃, 400 MHz): δ 6.66 (t, J = 2.2 Hz, 2H), 6.13 (t, J = 1.5 Hz, 2H), 3.97 (t, J = 6.9 Hz, 2H), 2.41-2.33 (m, 8H), 2.06-1.97 (m, 2H) (Figure S1a). MS: m z⁻¹: calculated for C₉H₁₇N₂ [M + H]⁺ 153.14, found 153.08 (Figure S1b).

Synthesis of 3-((3-(1H-pyrrol-1-yl)propyl)dimethylammonio)propanoate (ZiPy) (2)

β-propiolactone (0.76 mL, 12 mmol) dissolved in anhydrous THF (10 mL) was added dropwise to a solution of compound 1 (1.4 g, 8 mmol) in anhydrous THF (50 mL) under argon at 0 °C. The mixture was continued stirring at 15 °C for a further 6 h. Then, diethyl ether (50 mL) was added. The product was extracted with water (2 × 25 mL), the aqueous phases were combined, and the solvent was removed under vacuum. Compound 2 was purified via reverse-phase column chromatography (5% MeOH in water) using Biotage SNAP C18 cartridges (60 g), which provided the product as a hygroscopic white solid at 37% yield and high purity. ¹H NMR (D₂O, 400 MHz): δ 6.83 (t, J = 2.2 Hz, 2H), 6.20 (t, J = 2.2 Hz, 2H), 4.06 (t, J = 6.4 Hz, 2H), 3.52 (t, J = 8.6 Hz, 2H), 3.22-3.18 (m, 2H), 3.01 (s, 6H), 2.59 (t, J = 8.6 Hz, 2H), 2.29-2.21 (m, 2H) (Figure S2a). ¹³C NMR (D₂O, 100 MHz): δ 176.5, 121.4, 108.2, 61.4 (t, 1:1:1), 61.1 (t, 1:1:1), 50.5 (t, 1:1:1), 45.5, 30.5, 24.1 (Figure S2b). MS: m z⁻¹: calculated for C₁₂H₂₁N₂O₂ [M + H]⁺ 225.16, found 224.98 (Figure S2c). The estimated molecular weight of ZiPy monomer was 224.30 g mol⁻¹. The ZiPy monomer solution (0.1 m with 0.1 m NaCl) maintained high solubility without precipitation (Figure S11).

Preparation of ZiPPy electrodes for antibody detection

Spike protein of interest (Table S2) was diluted to a final concentration of 20 μg mL⁻¹ in ZiPy monomer solution (0.1 m with 0.1 m NaCl). Thirty microliter of the mixture solution was dropped on a working electrode (4 mm in diameter). Electropolymerization of ZiPy to ZiPPy was performed via cyclic voltammetry by scanning the potentials between 0.0 and +0.95 V with respect to Ag/AgCl reference electrode at the 50 mV s⁻¹ scan rate. After the polymerization step, the amount of S protein left in the solution was quantified via Qubit protein assay (Thermo Fisher Scientific). Two microliter of the remaining solution was added to 198 μL of Qubit protein reagent (200-fold diluted with Qubit protein buffer solution). The mixture was incubated for 15 min and the protein concentration was read out using Qubit 4 fluorometer (Thermo Fisher Scientific). For the leaching test, S-protein coated electrodes were incubated with buffer and the amount of S protein in the buffer was measured via Qubit protein assay.

Surface characterization

Both carbon and gold electrodes were coated with polymers, and the coated surface was characterized via AFM (NX-10, Park Systems equipped with CONTSCR cantilevers, Nano world). AFM images were analyzed using XEI software (Park Systems). FTIR spectra were obtained as average of 32 scans, and OMNIC software was used for analysis. Carbon electrodes were excluded in the peak analysis, as they displayed numerous peaks associated with the substrate itself.

Contact angle measurement

Distilled water (DIW) contact angles on bare, PPy-, ZiPPy- coated carbon and gold surfaces were recorded with home-built contact angle measuring system (Center for Nanosystems, Harvard) equipped with ImageJ plug-in to analyze the contact angle. A 5 μL drop of DIW was added on samples by a micropipette for the static contact angle measurement. After adjusting the illumination and the focus, the images were taken and later on analyzed with ImageJ contact angle plug-in. The electrode-drop-air interface was specified by replacing five crosses around the drop and Both BestFit model was selected to analyze the contact angles.

Fouling study

Different fouling media were used: human saliva collected from unvaccinated individuals, 10% bovine serum albumin (BSA), human serum and fluorescent streptavidine solution (1 mg mL⁻¹). Fouling analyses were made via Qubit protein quantitation assay (ThermoFisher Scientific, Q33212), fluorescence imaging, and electrochemical measurements. Bare, PPy and ZiPPy electrodes were drop-casted with each fouling medium and incubated for 1 h at 20 °C. For qubit protein quantitation assay, after the incubation period, the fouling medium was collected and analyzed according to the manufacturer’s protocol. Fluorescence imaging and electrochemical characterizations were made at the electrode surfaces after washing electrodes with PBS (1×). Fluorescence images were taken by an inverted fluorescence microscope (Ti-E, Nikon) and recorded by a CMOS camera (Zyla 5.5, Andor). Fluorescent intensity was obtained using Image-J software. Two types of electrochemical signals were measured to compare the antifouling characteristics of different electrode surfaces. Both cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were performed in the mixture of K₄[Fe(CN)₆] (10 mm), K₃[Fe(CN)₆] (10 mm), and KCl (100 mm; pH 7.0). CV studies were performed within the potential range from – 0.6 to 0.6 V at a fixed scan rate of 100 mV s⁻¹. Impedance was measured between 100 mHz and 100 kHz with a 10 mV sinusoidal voltage superimposed on a constant bias of −0.195 V. EIS data were analyzed using ZView software (version 3.2, Scribner Associates).

Saliva sample collection and preparation

This study was approved by the Institutional Review Board of Massachusetts General Hospital (IRB number 2019P003472, PI: Hakho Lee), and the overall procedures followed institutional guidelines. The sample source and test results were blinded until the statistics analyses. No data were excluded. SalivaBio Oral Swab (SOS) Device (5001.02) (Salimetrics, USA) was used to collect saliva samples. SOS was kept in the mouth for 3-5 min and replaced into a collection tube. To extract saliva, the tube was centrifuged for 15 min at 1500 ×g. Samples were kept at 4 °C until use.

Sars-CoV-2 antibody detection in saliva

Saliva samples (50 μL) were drop-casted on S-protein functionalized ZiPPy electrodes and incubated for 1 h at 20 °C. The electrodes were then washed with PBS (1×) and an electrolyte solution (10 mm K₄[Fe(CN)₆] and 10 mm K₃[Fe(CN)₆] in 100 mm KCl) was added to the electrodes. Bio-logic SP-200 potentiostat with impedance analyzer (Bio-Logic Science Instruments SAS, Claix, France) was used to measure EIS.