Enhancing Protein Capture Using a Combination of Nanoyeast Single-Chain Fragment Affinity Reagents and Alternating Current Electrohydrodynamic Forces

Abstract

New high-performance detection technologies and more robust protein capture agents can be combined to both rapidly and specifically capture and detect protein biomarkers associated with disease in complex biological samples. Here we demonstrate the use of recently developed recombinant affinity reagents, namely nanoyeast-scFv, in combination with alternating current electrohydrodynamic (ac-EHD)-induced shear forces, to enhance capture performance during protein biomarker analysis. The use of ac-EHD significantly improves fluid transport across the capture domain, resulting in enhanced sensor-target interaction and simultaneous displacement of nonspecific molecules from the electrode surface. We demonstrate this simple proof-of-concept approach for the capture and detection of Entamoeba histolytica antigens from disinfected stool, within a span of 5 min using an ac-EHD microfluidic device. Under an ac-EHD field, antigens were captured on a nanoyeast-scFv immobilized device and subsequently detected using a quantum dot conjugated antibody. This immunosensor specifically detected antigen in disinfected stool with low background noise at concentrations down to 58.8 fM with an interassay reproducibility (%RSD of n = 3) < 17.2%, and in buffer down to 5.88 fM with an interassay reproducibility (% RSD, n = 3) of 8.4%. Furthermore, antigen detection using this immunosensor was 10 times more sensitive than previously obtained with the same nanoyeast-scFv reagents in a microfluidic device employing surface-enhanced Raman scattering (SERS) detection in buffer and at least 200 times more sensitive than methods using screen printed gold electrodes in disinfected stool. We predict this rapid and sensitive approach using these stable affinity reagents may offer a new methodology to detect protein disease biomarkers from biological matrices.

Graphical Abstract

Recent advances in detection technologies have yielded new methods that improve the speed, sensitivity and accuracy of identifying disease biomarkers in complex samples.¹ Currently, most immunosensors capture and detect candidate biomarkers using whole antibodies that are produced via immunizing animals to generate polyclonal antibody preparations or alternatively monoclonal antibodies (mAbs) by hybridoma technology. However, full-length antibody production encounter problems with batch to batch variability in polyclonal antibodies and changes to quality over time for monoclonals,^2,3 which is a drain on research time and expenses,³ and relatively long, costly production times.^2,4 Due to these issues, focus over the past decade has shifted to recombinant antibody fragments such as single-chain variable fragment (scFv) reagents. scFvs are a promising class of protein capture reagents which can be used as alternatives to full-length mAbs in immunosensors.^2,5,6 These reagents show similar specificity toward target antigens as their parent, full-length monoclonal antibody. scFvs have the added flexibility to engineer in tags and customize the fragment antigen binding site, which allows custom production of reagents with the most sought-after affinity traits (i.e., orientation and positioning on immunosensor surface to optimize protein capture).^7–9 Furthermore, scFvs can be rapidly isolated from libraries of antibody fragment genes using various display technologies (e.g., yeast display), and can be produced in prokaryotic production systems followed by scale-up manufacture.^10,11 Despite these advantages, many solubly expressed scFvs suffer from stability problems,^12,13 which necessitates new designs and strategies to produce stable protein capture agents.^12,14–16

Protein capture agents composed of lyophilized whole yeast cells with displayed scFv (yeast-scFv) have shown to be potential alternatives to mAbs in immunosensors.¹⁷ Yeast-scFv can be selected from nonimmune yeast display libraries in 2–3 weeks¹⁸ and do not require special storage or cultivation conditions, allowing these reagents to be stored at room temperature in lyophilized form prior to use in an immunoassay. However, yeast-scFvs are insoluble and too large to adopt into many of the new detection technologies that are being developed each year. We recently developed an immunoaffinity reagent, called nanoyeast-scFv, that stabilizes soluble scFvs on biosensor surfaces by retaining scFvs anchorages to yeast-cell wall fragments, the environment in which the scFv were selected to function when they were culled from yeast-display libraries.^19–22 It has been hypothesized that the yeast cell wall fragments act as in-built stabilizing excipients, possibly due to the stabilizing sugars found in the yeast cell wall.²² The combination of yeast-display selection and the comparatively lower-cost production of using the yeast-display library with the added advantage of retaining the cell wall structure as in inbuilt stabilizing agents presents these reagents as alternative capture reagents to full-length mAbs.^19–22 Our prior work on nanoyeast-scFv demonstrated their utility in immunosensing using electrochemical²⁰ and surface enhanced Raman scattering (SERS) techniques.²¹ Herein, we present a simple approach to enhance the capture performance of nanoyeast-scFv systems in pretreated stool samples using alternating current electrohydrodynamic (ac-EHD)-induced surface shear forces.

Electrically driven fluid systems represent a promising approach for rapid and specific target capture with the ability to enhance fluid transport across the sensing domain. The use of electrical forces covered by electrohydrodynamics such as ac-electroosmosis (ACEO) and dielectrophoresis has previously been demonstrated for the concentration of biological particles such as pathogens,^23,24 cells,²⁵ or nucleic acid populations^26,27 and also the manipulation of colloidal objects on electrode surfaces.²⁸ Overall, these previous developments demonstrated high sensitivity and reduced overall assay times to typically a few minutes. Berggren et al. demonstrated the use of ac-field-induced forces for the manipulation of protein molecules on electrode surfaces to achieve detection limits as low as 10⁻¹⁵ M.²⁹ Similarly, Hart et al. utilized ac electro-osmosis to enhance fluid transport in microdevices during protein capture and demonstrated significant improvements in capture performance.³⁰ Under an applied ac field, ACEO force generates microflows that convey particles or targets from the bulk fluid onto the electrode surface. For instance, in systems comprising circular and/or ring shaped electrodes^31,32(inner circular and outer ring), the direction of slip velocity is toward the center of the inner electrode and away from it vertically. Because of mass conservation, a circulating flow is generated on the electrode surface to transport and cluster the molecules. ACEO flow generates two large vortices and two small vortices, and the counter-rotating vortices trap molecular targets onto the electrode surface. With a bias dc offset applied with an ac signal, more ions are induced in double layer due to Faradaic charging, leading to enhanced ACEO flow, and thus, more particles can be trapped. The special features of this fluid flow include the ability to (i) tune the force exerted on the particle or target molecules and simultaneously remove nonspecific molecules by simply adjusting the magnitude of the electrical body forces (via applied ac field) and (ii) modulate analyte transport without the need for external fluid handling systems. This is potentially difficult to achieve with alternative techniques using liquid flow, evaporation, sedimentation, or mechanical manipulation.³³

In this report, we for the first time demonstrate sensitive and rapid protein sensing using systems combining ac-EHD induced fluid flow with nanoyeast-scFv probes as capture agents. The utility of this approach is demonstrated as a proof-of-concept model system using a purpose built ac-EHD microfluidic device for the detection of recombinant Entamoeba histolytica antigens spiked into disinfected stool samples. E. histolytica infections are primarily found in developing countries and require low-cost molecular diagnostics for rapid detection of E. histolytica antigens.³⁴ By combining the facile production process of these nanoyeast-scFv reagents (e.g., due to the inbuilt stability and relatively lower-cost of recombinant production), with ac-EHD effects, this paper details a simple demonstration to reduce total assay time and improve the sensitivity of nanoyeast-scFv based assays.

EXPERIMENTAL SECTION

Reagents.

All reagents were of analytical grade and purchased from Sigma-Aldrich, Australia. Unless stated otherwise, 1 mM phosphate buffer saline (PBS) was used in all studies. The SiteClick Qdot 655 Antibody Labeling Kit was purchased from Life Technologies, Australia. Biotinylated anti-HA antibody (Abcam #ab26228) was purchased from Sapphire Bioscience, and streptavidin was procured from Invitrogen. Antigen “350”, a fragment of E. histolytica protein EHI_115350, was cloned and purified as previously described.^17,20 An anti-350 polyclonal rabbit antiserum was generated by Cocalico Biologicals, Inc., Reamstown, PA, USA.¹⁷

The yeast-scFv library was a generous gift from K. Dane Wittrup. For selections, antigen “350” was biotinylated with EZ-Link Sulfo-NHS-LC biotin (Thermo Pierce, Rockford, IL). Yeast-scFv selections were performed as described earlier, with two rounds of magnetic selection and two rounds of sorting on a Beckman Aria II cell sorter.^17,20 The final sort output was plated, and yeast-scFv clones were confirmed specifically by flow cytometry analysis and uniquely by scFv PCR amplification and BstNI digest (New England Biosciences, USA).

The stool matrix utilized in this study was prepared from human samples collected by the International Centre for Diarrheal Disease Research, Bangladesh (ICDDR,B) with institutional review board approval from the University of Washington and the ICDDR,B. Disinfected stool matrix was generated by pooling five E. histolytica-negative samples, diluting them to a 1:1 ratio in PBS, centrifuging for 5 min at high speed, and heating the subsequent supernatant for 20 min at 95 °C. Stool was considered disinfected when streaks of the stool produced no growth on nutrient and tryptic soy agar plates for ≥3 days at 37 °C.

Generation of Nanoyeast-scFv.

Yeast cells were lyophilized for long-term storage, and nanoyeast-scFv were created as described previously.²⁰ Briefly, lyophilized yeast was disrupted with a mortar and pestle into a fine powder and then reconstituted with 10 mL of PBS, 5% glycerol, and protease inhibitor EDTA-free cocktail. Samples were centrifuged at 500 RCF for 2 min at 15 °C to remove whole yeast. To prevent microbial growth, sodium azide 0.05% was added. The processed lysates (supernatant) were stored at 4 °C until use and filtered using a 0.1 μm filter.

Design of ac-EHD Microdevice.

Pyrex glass wafers (4″, 1 mm thick, double-side polished) were obtained from Bonda Technology, Singapore. Positive (AZ9260) and negative (AZnLOF 2070) photoresists were purchased from Micro-chemicals, Germany. Sodalime chrome masks (5″ × 5″) were obtained from Qingyi Precision Maskmaking (Shenzhen) Ltd., China.

We designed a simple microdevice (Figure 1a) with an asymmetric electrode pair containing an inner circular small electrode and a large outer ring electrode with an edge to edge distance of 1000 μm between the electrodes.²⁸ The diameter of the inner electrode and the width of the outer ring electrode were 250 and 30 μm, respectively. The microdevices were designed using Layout Editor (L-Edit V15, Tanner Research Inc., CA).

Figure 1.

Schematic of the electrode and ac-EHD mechanism. (a) Optical image of the asymmetric electrode pair containing an inner circular small electrode and a large outer ring electrode with an edge to edge distance of 1000 μm between the electrodes. The diameter of the inner electrode and the width of the outer ring electrode were 250 and 30 μm, respectively. Scale bar = 200 μm. (b) Schematic representation of the mechanism of ac-EHD induced surface shear forces for rapid capture and detection of E. histolytica antigen (not drawn to scale). Under an ac-EHD field, the charges induced within the electrical double layer of an electrode experience an electrical body force that drives the bulk fluid onto the inner circular electrode. This fluid flow transports target molecules or detection antibody in the bulk fluid and can continuously supply target molecules (i.e., increase sensor-target affinity interactions) onto the capture domain. Further, the fluid flow can be tuned using the applied ac field to achieve optimal fluid flow that can maximize device performance. This ability of ac-EHD flow can enable rapid capture and detection of target antigens.

Fabrication of Microdevices.

The asymmetric electrode pair based microdevices were fabricated using a previously described method.²⁸ Briefly, Pyrex glass wafers were cleaned by sonication in acetone and isopropanol (IPA) and dried under a flow of nitrogen gas. The cleaned wafers were then coated with a thin layer of positive photoresist (AZ9260) using a spin coater at 4000 rpm for 60 s, followed by a soft bake step for 7 min at 110 °C. These wafers were then UV exposed (1000 mJ cm⁻²) using a mask aligner (EVG620, EV Group GmbH, Austria) and developed in AZ 726 developer solution (Microchemicals, Germany) for 8 min. The developed wafers were then rinsed with deionized (DI) water and dried under the flow of nitrogen gas. Prior to gold deposition, any residual resist layers were removed with a brief oxygen plasma treatment (60 W for 45 s). Metallic layers of titanium (10 nm adhesion layer) and gold (200 nm) were deposited using an e-beam evaporator (Temescal BJD-2000 E-beam). Subsequently, acetone lift-off revealed the gold patterns with an inner circular electrode (diameter, d = 250 μm) and an outer ring electrode (width, w = 30 μm). The distance between the inner and outer electrode (edge to edge) was 1000 μm. The inner and outer circular electrodes were connected to large gold connecting pads. Following this, the gold electrodes were coated with a layer of negative photoresist (AZnLOF2070) at 3000 rpm for 30 s and soft baked at 110 °C for 3 min. The wafers were then UV exposed (1000 mJ cm⁻²) and developed in AZ 726 developer solution to expose the gold electrodes and connecting pads. Finally, the wafers were diced (ADT-7100 dicing saw) and cleaned using oxygen plasma treatment (100 W for 45 s) prior to functionalization.

Device Functionalization.

Gold microelectrodes were modified in a three-step process using standard avidin–biotin chemistry. Initially, the device was incubated with biotinylated BSA solution (500 μg mL⁻¹) for 2 h. The device was then incubated in streptavidin (500 μg mL⁻¹) for 1 h followed by incubation with biotinylated anti-HA antibody (100 μg mL⁻¹) solution for another hour. Finally, the nanoyeast-scFv reagent diluted 1:5 in PBS (pH 7.4, 1 mM) was allowed to bind (by virtue of the interaction between anti-HA and the HA affinity tag cloned into the yeast-display scFv) for 1 h and washed with PBS. This step tethered the nanoyeast-scFv to the electrode surface while at the same time removing yeast fragments that do not bear displayed scFv. The device was washed with PBS after each incubation step.

Quantum Dot Modified Detection Antibody.

The detection antibody was labeled with Qdots using SiteClick antibody labeling kits (Invitrogen, UK) as per manufacturer’s instructions. Briefly, the antibody concentrator provided with the kit was rinsed with distilled water, and a designated volume (100 μL) of the polyclonal rabbit anti-350 detection antibody (antibody concentration: 2 mg mL⁻¹) was added to the concentrator. The concentrator was then filled with preparation buffer to obtain a final volume of 500 μL. The antibody was concentrated through a series of centrifugation steps (5000 × g for 6 min followed by 1000 × g for 3 min) and subsequently incubated with β-galactosidase for 4 h at 37 °C. The modified antibody was then incubated with Tris-UDP-N-azidoacetylgalactosamine (UDP-GalNAz; 20× Tris; pH 7.0) solution overnight at 30 °C. The azide-modified antibody was then transferred to a new concentrator component, and repeated washes using Tris buffer (1× Tris; pH 7.0) removed any unmodified antibody. Subsequently, the concentrator was filled with the azide modification solution, and the antibody was purified through a series of centrifugation steps (1200 × g for 6 min followed by 1200 × g for 10 min and 1400 × g for 10 min). The flow through was discarded after each centrifugation step, and the concentrator was rinsed repeatedly with Tris buffer. Qdots were conjugated with the purified antibody upon the addition of Qdot DIBO label (50 μL) and overnight incubation at 25 °C. Finally, any unconjugated antibody was separated by repeated centrifugation (1500 × g for 5 min followed by 1500 × g for 10 min) and wash steps using PBS buffer. The conjugated antibody was stored at 4 °C until further use.

Antigen Capture and Detection.

The large and small electrodes of the microdevice were connected to a signal generator (Agilent 33510B waveform Generator, Agilent Technologies, Inc., CA) via gold connecting pads. Samples (20 μL) containing designated concentrations of E. histolytica “350” antigen spiked in 1 mM PBS or pretreated stool samples were manually added to the device and driven under the applied ac-EHD field. Subsequently, Qdot conjugated detection antibody (20 μL; 2 μg mL⁻¹ polyclonal rabbit anti-350) was manually added onto the device and driven under the applied ac-EHD field. The capture and subsequent detection steps in all the experiments were performed for designated time intervals (0.5 to 3 min) under the applied field strength of f = 100 Hz and V_pp = 4 V. Control experiments were performed under static conditions (e.g., incubation) for both capture (1 h) and detection (30 min) steps. The device was washed with PBS after the capture and detection steps. Finally, the device was imaged under a confocal microscope (Zeiss LSM 710 confocal microscope) using a ×63 oil immersion objective lens to obtain fluorescence images of the detected antigen (Qdots 655 with excitation at 405 nm). Subsequently, fluorescence intensity measurements were obtained upon analyzing the captured images using Zeiss LSM analysis software.

RESULTS AND DISCUSSION

To demonstrate the assay platform, we constructed a device (see Experimental Section and Figure 1a) containing an asymmetric electrode pair with a large (ring) and small (inner circular) electrode forming the cathode and anode or vice versa of an electrolytic cell. The design of the system employed in our study was adopted from our previous investigations on the use of ac-EHD induced fluid flow for the manipulation of colloidal particles.²⁸ Although the use of similar ring electrode geometries for fluid manipulation has been reported earlier,^31,32 we optimized the device geometry, associated design parameters, and ac-field conditions for fluid flow with fluorescent latex bead-based model systems in order to achieve enhanced sensitivity and specificity. In this study, we have adopted a similar design to establish an effective methodology for the detection of pathogen antigens based on in-house synthesized nanoyeast-scFv probes. Figure 1 illustrates the use of ac-EHD induced surface shear forces for the specific capture of E. histolytica antigen. The application of a potential difference across the inner circular and outer ring electrode results in charges being induced onto the electrodes. Since the surface charge is balanced by the redistribution of ions close to the surface, it leads to accumulation of counterions on the electrode surface. This charge accumulation results in the formation of a capacitive electrical double layer. The asymmetric geometry of the electrode results in a lateral variation in the number of accumulated charges on the respective electrodes and their spatial distribution on the electrode surface.³⁵ Subsequently, the resulting tangential electric field generates two nonuniform forces with the force on the ring electrode being greater than that on the circular electrode. The resultant force drives the bulk fluid flow at the interface across the functionalized capture domain and results in improved target–sensor interactions. These forces are generated within the electrical double layer of an electrode. This double layer is typically 3–4 nm thick for phosphate buffer saline (PBS) solution (derived from Debye–Hückle approximation described in ref.^36,37), thereby suggesting that all ac-EHD forces are engendered within molecular distances from the electrode surface. The ability of these forces to engender fluid flow vortices and concomitant fluid mixing (Figure 1b) increase the immunocapture efficiency as a result of increase in a number of sensor–target collisions and physical displacement of nonspecific molecules from the electrode surface.

To investigate the potential of ac-EHD-induced forces for E. histolytica antigen capture and detection, the devices were functionalized using nanoyeast-scFv affinity reagents (Figure 2). Nanoyeast-scFv was immobilized with the use of a human influenza hemagglutinin (HA) antigen tag cloned into the recombinant scFv construct.^18,22 Thus, this assay uses biotinylated anti-HA antibody to build immunoassays using any yeast-scFvs selected to bind antigens of interest. Samples (20 μL) containing E. histolytica antigen EHI_115350 (“350” antigen from here on) in PBS were driven through functionalized devices. Subsequently, the captured antigens were detected using a rabbit polyclonal anti-“350” detection antibody conjugated to quantum dots (QD) (Figure 2).

Figure 2.

Schematic representation of the capture and detection of E. histolytica antigen using nanoyeast-scFv as capture probes for pathogen antigen detection (not drawn to scale).

The assay performance in our device is determined by the magnitude of bulk fluid motion under ac-EHD field. This bulk fluid flow is characterized by the applied field strength, electrolyte concentration, and time intervals used in capture and detection steps. These parameters work synergistically, as they determine the conditions for effective target nanoyeast–scFv interactions. The optimal ac-EHD field conditions were obtained from our previous investigations²⁸ on the effect of different field strengths (i.e., different frequency and amplitude ranges) for the specific capture of streptavidin beads. It was found that, at the field strength of f = 100 Hz and V_pp = 4 V, the shear forces were strong enough to displace the nonspecific molecules and enhance target capture as compared to other tested electric field strengths. Since the reaction time for antigen–antibody binding is highly dependent on the fluid flow rate, a stronger flow (i.e., high electric field strength) can wash out not only nonspecific molecules (weakly interacted to surface-bound antibodies) but also specific target molecules that can significantly reduce the number of effective antibody–antigen collisions (i.e., a condition where shear forces > antibody–antigen affinity forces). Thus, the optimal field strength (i.e., f = 100 Hz and V_pp = 4 V) was utilized for investigations on the effect of ac-EHD on E. histolytica detection. At this field strength, the stimulation of fluid flow around the sensor surface was optimal to maximize sensor–target collisions. However, the time interval for capture and detection under ac-EHD also needs to be identified in order to determine the optimal shear forces that are strong enough to displace nonspecific molecules while facilitating target capture.

In order to determine the optimal time intervals (i.e., time of application of ac-EHD field) for capture and detection of E. histolytica antigen, the antigen capture and detection steps were performed at different time intervals (0.5–3 min) under the ac-EHD fluid flow conditions of f = 100 Hz and V_pp = 4 V. Samples containing “350” antigen (1 ng mL⁻¹) and/or detection antibody spiked in PBS were driven through the functionalized devices. Figure 3 demonstrates the capture of “350” antigen, with the antigen capture and detection steps performed at different time intervals (0.5–3 min) under ac-EHD fluid flow conditions. Fluorescence images and corresponding image analysis (see Experimental Section for details) suggest that the capture performance of the device was a function of ac-EHD time intervals. The affinity interactions during capture (antigen to nanoyeast-scFv) and detection steps (antigen to detection antibody) depend on the flow rate. Thus, the high capture efficiency at time intervals of 3 and 2 min for capture and detection steps probably reflects the effective manipulation of shear forces and subsequent stimulation of fluid flow around the sensors. This manipulation maximizes effective antigen-scFv and antigen–antibody collisions (a condition where shear force < antigen-scFv or antigen–antibody affinity force). A further increase in the time of fluid flow (>6 min for capture and detection) resulted in electrolysis causing electrode damage (data not shown). Thus, the capture and subsequent detection steps were performed for time intervals of 3 and 2 min respectively in subsequent experiments.

Figure 3.

Optimal conditions for capture and detection of E. histolytica antigen in buffer. Representative fluorescence images and corresponding image analysis of captured “350” antigen, with the antigen capture (0.5–3 min) and detection (1–2 min) performed at different time intervals under ac-EHD induced fluid flow conditions of f = 100 Hz; V_pp = 4 V. Scale bar is 50 μm. Each data point represents the average of three separate measurements (n = 3), and error bars represent standard error of measurements within each experiment.

To investigate the specificity of immunocapture, control experiments were performed using samples spiked with (1 ng mL⁻¹; Figure 4a) and without “350” antigen (Figure 4b) in PBS. Under the optimal experimental conditions, negligible nonspecific binding of the QD conjugated detection antibody was observed (Figure 4). To further assess the specificity and accuracy of immunocapture, additional control experiments were performed using devices functionalized without (i) anti-HA antibody (Figure 4c) and (ii) nanoyeast-scFv (Figure 4d). Samples containing “350” antigen (1 ng mL⁻¹) and/or detection antibody spiked in PBS were driven through the functionalized devices. These control results suggest that ac-EHD induced surface shear forces minimize any background signal from the detection antibody.

Figure 4.

Specificity of immunocapture in buffer samples. (a, b) Representative fluorescence images and corresponding image analysis of samples (a) spiked with (1000 pg mL⁻¹) and (b) without “350” antigen in PBS under ac-EHD field strength of f = 100 Hz and V_pp = 4 V. (c,d) Representative fluorescence images and corresponding image analysis of captured “350” antigen (1000 pg mL⁻¹) in PBS on devices without (c) anti-HA antibody and (d) nanoyeast-scFv under ac-EHD field strength of f = 100 Hz and V_pp = 4 V. Each data point represents the average of three separate measurements (n = 3), and error bars represent standard error of measurements within each experiment.

To investigate the dynamic range and limit of detection (LOD) of this device, concentrations of “350” antigen (1 ng mL⁻¹ to 100 fg mL⁻¹) were spiked in PBS, and the samples were run on nanoyeast-scFv functionalized devices. Fluorescence micrographs and image analysis suggest that the device was sufficiently sensitive to detect 100 fg mL⁻¹ (5.88 fM) “350” antigen spiked in PBS (Figures 4b and 5a). We believe the spatial distribution in capture trend observed in fluorescence images for each concentration can be attributed to the structure and orientation of scFv on the capture domain. Our recent investigations on the structural and orientation of nanoyeast-scFv fragments indicate that they are globular in structure and heterogeneous in size and were found to aggregate in solution.²² This could possibly result in an inhomogeneous surface coating on the sensor surface and subsequently influence the capture trend of the device. The dynamic range of detection was determined to be from 1 ng mL⁻¹ to 100 fg mL⁻¹. The correlation in signal enhancement with an increase in antigen concentration was found to be R² = 0.9849. Despite the inhomogeneous surface coating and aggregation of nanoyeast-scFv on the surface resulting in a different capture trend for each trial, we observed that the total resultant fluorescence signal was consistent for each of the tested concentrations. This is evident from the reproducibility (% RSD) of <8.4% (n = 3), determined across three individual trials (Figure 6). Antigen detection by this device in buffer was 10 times more sensitive than previously obtained with the same nanoyeast-scFv reagents in a microfluidic device employing surface-enhanced Raman scattering (SERS) detection,²¹ and at least 100 times more sensitive than methods using screen printed gold electrodes that require longer assay times.²⁰ This enhancement in sensitivity is possibly due to the effective manipulation of shear forces to create fluid flow vortices and subsequent fluid mixing that can enhance the number of sensor–target collision. The use of quantum dots as a signal readout with fluorescence microscopy may have also improved sensitivity. This level of sensitivity and reproducibility is comparable to existing technologies based on scFv based immunocapture,^7–9 SERS based immunoassays for pathogen detection from bacteria,²³ enzyme based signal amplification methods,³⁸ and ac electrokinetic enrichment^39,40 for the rapid detection of disease based biomarkers. Additionally, our nanoyeast-scFv assay utilizing ac-EHD surface sheer forces overcomes/addresses some limitations of these previously mentioned assays, including (i) nonspecific adsorption of nontarget species onto the capture domain, (ii) the use of full length mAbs that add production cost and are time-consuming to reliably generate, and (iii) the requirement for complex surface modification or detection procedures and longer analysis periods. Thus, we believe the use of ac-EHD surface shear forces in combination with nanoyeast-scFv and/or other affinity reagents could potentially be an effective alternative to conventional immunoassay devices.

Figure 5.

Rapid capture and detection of E. histolytica antigen. (a) Representative fluorescence images and corresponding fluorescence image analysis of the captured “350 antigen” spiked in PBS at designated concentrations (1 ng mL⁻¹ to 10 fg mL⁻¹) under the applied ac-EHD field strength of f = 100 Hz and V_pp = 4 V. Scale bar is 50 μm. Each data point represents the average of three separate measurements (n = 3), and error bars represent standard error of measurements within each experiment. (b) Representative fluorescence images and corresponding fluorescence image analysis of the captured “350 antigen” spiked in disinfected stool samples at concentrations of 1000 pg mL⁻¹ and 1 pg mL⁻¹ under the applied ac-EHD field strength of f = 100 Hz and V_pp = 4 V. Scale bar is 50 μm. Each data point represents the average of three separate measurements (n = 3), and error bars represent standard error of measurements within each experiment.

Figure 6.

Reproducibility of E. histolytica antigen immunocapture. Representative fluorescence images of samples spiked with (1000 pg mL⁻¹) “350” antigen in PBS under ac-EHD field strength of f = 100 Hz and V_pp = 4 V.

We then tested the assay performance in analyzing E. histolytica cyst antigens present in a complex biological matrix. Defined concentrations of “350” antigen (1 ng mL⁻¹ and 1 pg mL⁻¹) were spiked into disinfected, E. histolytica-negative stool samples (see Experimental Section). These simulated clinical samples were run on nanoyeast-scFv functionalized devices under the ac-EHD fluid flow conditions ( f = 100 Hz and V_pp = 4 V). It was observed that, under the applied ac-EHD field, the optimal capture (3 min) and detection (2 min) time intervals resulted in electrolysis causing electrode damage. This could possibly be attributed to the high electrolyte concentration of stool samples caused by their dilution in PBS during disinfection procedures. Thus, ac-EHD time intervals of 1.5 and 1 min were used for antigen capture and detection steps, respectively. Despite the reduction in antigen capture and detection time intervals, the device was sensitive enough to detect 1 pg mL⁻¹ of “350” antigen (Figure 5b) in a disinfected stool matrix. Reproducibility of concentrations from three separate trials resulted in a %RSD of <17.2% (n = 3) between the electrodes at each concentration. Generally, the concentrations of E. histolytica antigens detected in stool samples range from μM to pM levels.^41–44 Commercial kits and other detection platforms based on standard ELISA systems have reported the detection of antigens with a molecular weight (170 kDa) 10 times higher than that of the currently tested “350” antigen to as low as 1000 pgmL⁻¹ (5.88 pM) in stool samples.⁴² Thus, we believe that our device performance (e.g., 1 pg mL⁻¹ (58.8 fM)) in stool samples is comparable with existing methodologies for E. histolytica detection. Furthermore, the analytical sensitivity of the device used is at least 200 to 500 times more sensitive in disinfected stool than previously reported demonstrations on “350” antigen that also require longer assay times.^20,22 To our knowledge, this is the best reported sensitivity for the detection of E. histolytica antigens from a complex biological matrix. In addition to this, our approach also demonstrates a significant improvement in assay time with current methodologies and commercial kits for E. histolytica detection requiring an overall assay time of 30 min to 2 h.^34,42,45 Although the presented technology involves microscopy read-out systems to evaluate the device performance, we believe this method can potentially be adapted to other read-out platforms such as chemiluminescence, colorimetry, or even a simple fluorescence reader to align with the needs of routine diagnostic settings.

Finally, to investigate the specificity of immunocapture (Figure 7), disinfected stool samples spiked with and without “350” antigens were run on nanoyeast-scFv functionalized devices under the ac-EHD fluid flow conditions. Further, to investigate the potential of ac-EHD induced shear forces on minimizing background response during stool sample analysis, control experiments were performed under static conditions (e.g., without ac-EHD) with simple incubation of stool samples without any spiked antigen. Representative fluorescence images and corresponding intensity measurements suggested negligible nonspecific adsorption of stool biomolecules in the case of samples without any spiked antigen analyzed under ac-EHD (Figure 7b) in comparison to control experiments (Figure 7c) performed under static conditions (i.e., 2-fold decrease under ac-EHD vs static conditions). These data indicate that ac-EHD induced shear forces are effective in enhancing the detection capabilities of the device.

Figure 7.

Specificity of immunocapture in disinfected stool samples. (a–c) Representative fluorescence images and (d) corresponding image analysis for the capture and detection of stool samples spiked (a) with (1000 pg mL⁻¹) and (b) without “350” antigen under the applied ac-EHD field strength of f = 100 Hz and V_pp = 4 V. Additional control experiments were performed using stool samples (c) without any spiked antigen under static conditions (without ac-EHD) with simple incubation of target and detection antibody during capture and detection steps, respectively. Scale bar is 50 μm. Each data point represents the average of three separate measurements (n = 3), and error bars represent standard error of measurements within each experiment.

CONCLUSION

In conclusion, we have demonstrated a simple methodology that combines ac-EHD-induced shear forces and stable recombinant nanoyeast-scFv affinity reagents for rapid and sensitive detection of recombinant E. histolytica antigen in disinfected stool samples. This work is the first attempt at using ac-EHD-induced forces to enhance the capture performance of nanoyeast-scFv platforms in complex biological matrices. The use of ac-EHD enabled rapid (e.g., total assay time of 5 min for both capture and detection steps) and specific protein capture with the ability to minimize background noise particularly during detection from stool samples. Using this approach, we achieved reproducible (%RSD = 17.2 for n = 3) and sensitive (1 pg mL⁻¹) detection of E. histolytica antigen from pretreated stool samples. We believe the use of nanoyeast-scFv reagents can potentially provide simpler alternatives to full-length mAbs. Further, we believe this approach, combining the effectiveness of ac-EHD forces with nanoyeast-scFv, presents a simple method for protein biomarker detection.