pubs.acs.org/ac Article Paper-Based Electrophoretic Bioassay: Biosensing in Whole Blood Operating via Smartphone Amadeo Sena-Torralba,§ Ruslan Alvarez-Diduk,§ Claudio Parolo, Helena Torné-Morató, Alexander Müller, and Arben Merkoçi* Cite This: Anal. Chem. 2021, 93, 3112−3121 Read Online ACCESS Metrics & More Article Recommendations * sı Supporting Information See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. Downloaded via UNIV AUTONOMA DE BARCELONA on May 27, 2021 at 08:12:14 (UTC). ABSTRACT: Point-of-care (PoC) tests are practical and effective diagnostic solutions for major clinical problems, ranging from the monitoring of a pandemic to recurrent or simple measurements. Although, in recent years, a great improvement in the analytical performance of such sensors has been observed, there is still a major issue that has not been properly solved: the ability to perform adequate sample treatments. The main reason is that normally sample treatments require complicated or long procedures not adequate for deployment at the PoC. In response, a sensing platform, called paper- based electrophoretic bioassay (PEB), that combines the key characteristics of a lateral flow assay (LFA) with the sample treatment capabilities of electrophoresis is developed. In particular, the ability of PEB to separate different types of particles and to detect human antibodies in untreated spiked whole blood is demonstrated. Finally, to make the platform suitable for PoC, PEB is coupled with a smartphone that controls the electrophoresis and reads the optical signal generated. It is believed that the PEB platform represents a much-needed solution for the detection of low target concentrations in complex media, solving one of the major limitations of LFA and opening opportunities for point-of-care sensors. T he ability to measure a specific biomarker within minutes and in the easiest possible way is crucial for a variety of healthcare-related scenarios, ranging from the daily measure- The only sample treatment included in the former type of assays is the modification of the physical−chemical properties of the sample, thanks to the use of a prestored buffer (i.e., a ment of glucose level to the monitoring of a pandemic.1,2 In buffer dried into the sample pad, which is the region of the test this context, laboratory-based technologies are not useful, where the sample is applied). This allows for the stabilization considering that they require expensive/bulky equipment, of the pH, ionic strength, and viscosity of the sample, which long/multistep protocols, and qualified personnel.3,4 Instead, may be enough for clear samples with low viscosity (i.e., point-of-care (POC) tests represent a valid solution due to urine), but it is not sufficient for more viscous and complex their ease of use, speed, low cost, and portability.5−8 In the matrixes (i.e., serum/blood).12 Regarding the latter type of realm of POC technologies, lateral flow assays (LFAs) stand LFAs, they can include either a dilution step (prior to the out as the most widely used POC sensors across many addition of the sample onto the sample pad) or a washing step disciplines, from environmental analysis to safety and (the addition of a washing buffer after the application of the diagnostics. Their success lies in their single-step procedure sample).13,14 The washing buffer is usually based on a low (the user has to just apply the sample and in some cases a ionic strength and neutral buffer doped with anionic washing buffer) and their ability to provide results within 10− surfactants such as Tween-2015,16 or sodium dodecyl sulfate 15 min.9−11 The combination of their practicality with a low (SDS).12,17 These are frequently combined with external fabrication cost makes them a particularly attractive sensing filtration units or integrated external filtration membranes, platform for both developers and users. Indeed, the simple such as the commercial Cytosep and Vivid plasma separation operation of LFAs is one of their major advantages over other sensing technologies, but it also comes with a major drawback: the inability to perform efficient sample treatments. This Received: October 14, 2020 limitation of LFAs precludes their use for more challenging Accepted: January 21, 2021 applications, for example, the detection of biomarkers at low Published: February 3, 2021 concentrations in complex media. Most commercially available LFAs are either single-step (i.e., pregnancy tests) or double-step (i.e., serological tests) assays. © 2021 American Chemical Society https://dx.doi.org/10.1021/acs.analchem.0c04330 3112 Anal. Chem. 2021, 93, 3112−3121 Analytical Chemistry pubs.acs.org/ac Article membranes.18−20 There are also publications reporting novel bioreagent dispenser. The antibodies were dried for 2 h at integrated plasma separation strategies.21,22 In this case, LFAs 37 °C. can also work with blood and serum samples, but it Evaluation of the Electrophoresis Performance. compromises their ease of use (i.e., the user has to dilute the Commercial streptavidin-functionalized CdSe@ZnS QDs and sample, add a washing buffer or remove the filtration CdTe QDs were used. The graphene quantum dots (G-QDs) membrane at a fixed time) and possibly their sensitivity (the were synthesized using a homemade method following the dilution of the sample implies a dilution of the target analyte). procedure of Dong et al.25 The nitrogen- and sulfur-doped G- LFAs can also include reagents that remove specific QDS (N−S-doped G-QDs) were synthesized following the contaminants (for example, antibodies for the removal of procedure reported by Dong et al.26 The evaluation of the pH albumin), but this increases the complexity and cost of buffer effect on the electrophoretic mobility of the QDs was fabrication.23,24 carried out by drop-casting 1 mL of sodium borate (SB) buffer Here, we propose a sensing platform that allows the use of on the cellulose pads. Buffer solutions at different pH values whole blood while retaining all of the characteristics that make (8, 9, and 10) were prepared using boric acid with sodium LFAs the most popular POC diagnostic technology. hydroxide and tested on different PEB strips working Particularly, we developed a paper-based electrophoretic simultaneously. The gold nanoparticles (AuNPs) were bioassay (PEB) that uses the main components of LFAs, but synthesized by the Turkevich method27 using 50 mL of instead of relying on the capillary force to move the sample, it HAuCl4 (0.25 mM in Milli-Q water) and 1.25 mL of sodium uses electrophoresis. The use of a continuous and tunable flow citrate (1% in Milli-Q water). They were conjugated with anti- to move the sample easily overcomes the main issues observed H-IgG following the experimental procedure already reported when using blood samples in LFAs, like nonspecific binding or by the group,28 detailed in the Supporting Information. the clogging of the strip. We validated the PEB using as a test Detection of H-IgG in SB Buffer and Whole Blood. bed the detection of Human Immunoglobulin G (H-IgG) in One milliliter of SB buffer (10 mM, pH 8) was drop-casted untreated spiked whole blood. Finally, we demonstrated the onto the cellulose pads. Once the entire strip was embedded applicability of PEB at the point-of-care using a smartphone- with the electrophoresis buffer, the electrodes corresponding to based system that controls the electrophoresis and quantifies the positive and negative charged poles were connected to the the optical signal. power source. Next, 3 μL of the conjugated AuNPs was ■ EXPERIMENTAL SECTION Materials, Reagents, and Instruments. Materials, incubated with 3 μL of chicken whole blood for 30 s and drop- casted onto the NC pad at 1 cm from the TL. Then, the power source was connected, and the assay was performed for 15 min. reagents, and instruments, are detailed in the Supporting Evaluation of the PEB. Pictures of the strips were taken by Information (SI). adjusting the camera parameters of the smartphone (ISO 160, PEB Device Design and Fabrication. A nitrocellulose A 1/10, Manual focus) at a height distance of 9 cm from the (NC) membrane with a size of 10 cm × 0.49 cm was used as a strip and under constant light conditions, provided by the platform for running the electrophoresis. Two adhesive copper white LEDs. The pictures were evaluated using Image J electrodes were stacked through their adhesive at 0.5 cm from software, following the procedure reported by our group.29 both ends of the NC membrane. As stipulated by the The signal intensity was normalized following the equation: manufacturer, both the copper and the adhesive provide low (TL − BG)/(CL − BG), where TL, CL, and BG were the resistance to current and thus are suitable for their application signal intensities of the test line, control line, and background. as electrical connectors. Moreover, two cellulose pads (6 cm × The limit of detection (LoD) was calculated as optical 1.5 cm) were used for the storage and gradual dispensing of intensity (LoD) = blank + 3 σ blank (i.e., the corresponding the electrophoresis buffer and were stacked over the adhesive value of blank sample plus 3 times its standard deviation). The copper electrodes toward the extremes of the strip. The buffer limit of quantification (LoQ) was calculated as optical intensity is required to act as an electrolyte and maintain the pH at a (LoQ) = blank + 10 σ blank.30 relatively constant value. Next, the copper electrodes were Detection of H-IgG in Buffer and Whole Blood Using connected to the power source system that contained all of the LFA. The CytoSep membrane was used as the sample pad and electronics inside a three-dimensional (3D)-printed polylactic CN95 as the detection pad. The anti-H-IgG functionalized acid (PLA) cartridge (see the electronic diagram in Figure S1 AuNP solution used in the conjugate pad was the same as the and the Results and Discussion section for more details). For one used for the PEB experiments. The calibration curve for the evaluation of the quantum dot (QD) mobility, a black H-IgG was performed by spiking H-IgG (0−10 000 ng/mL) in cover was fabricated to measure the fluorescence emitted by PBS buffer (10 mM, pH 7.4). Then, 100 μL of each solution the QDs. The cover was made of PLA and fabricated using a were drop-casted onto the sample pad. For the calibration 3D printer. Two light-emitting diodes (LEDs) (365 nm curve of H-IgG in whole blood, a solution of chicken whole excitation wavelength) and two white LEDs were placed inside blood spiked with H-IgG (1 μg/mL) was prepared. The the device cover and connected by an on-the-go (OTG) wire solution was diluted 2.4- and 5-fold using PBS buffer (10 mM, to a Smartphone (Figure S2). pH 7.4). Then, 100 μL of the undiluted and diluted whole Nitrocellulose Membrane Strip Fabrication. The NC blood were drop-casted onto the sample pad. Pictures of the membranes used for the evaluation of the electrophoresis strips were taken after 15 min using the smartphone and performance were CN95, CN110, and CN150. For the adjusting the same camera parameters as for the PEB evaluation. ■ detection of H-IgG, the CN95 NC membrane was used. Polyclonal goat to H-IgG antibodies were deposited in the test line (TL) and polyclonal chicken to goat IgG antibodies were RESULTS AND DISCUSSION deposited in the control line (CL) (both at 1 mg/mL in 10 Concept. Electrophoresis is the migration of charged mM phosphate-buffered saline, PBS, pH 7.4) using a particles under the influence of an electric field. The system 3113 https://dx.doi.org/10.1021/acs.analchem.0c04330 Anal. Chem. 2021, 93, 3112−3121 Analytical Chemistry pubs.acs.org/ac Article is based on two electrodes of opposite charges (cathode and membrane has 109-fold higher resistivity than agarose gel and, anode) connected by a conductive medium (electrolyte). The therefore, it provides higher resistance to current.40,41 charged particles are able to separate due to differences in their The core of PEB is based on paper electrophoresis, which velocity (v), which depends on the particle’s mobility (m) and since 1950s has been developing strategies for the examination the applied field strength (E) in such a way of a variety of different samples, ranging from water to human v = mE (1) serum.42−44 Most of these works aimed at the separation of target analytes (including inorganic salts, amino acids, Moreover, the particle’s mobility (m) is determined by their enzymes, and proteins) along a filter paper and their size, shape, and charge and by the temperature during the subsequent analysis, such as the measurement of their separation. The mobility is constant under defined electro- concentration and/or specific activity.45 Nonetheless, these phoretic conditions, which are characterized both by electrical works relied on complex and tedious techniques for the parameters (current, voltage, power) and factors related to the identification of the separated components, which are based on electrophoresis medium (ionic strength, pH value, viscosity, staining or immunoblotting. The former serves for an and pore size). In this sense, the electrophoretic mobility (μ) estimation of the molecular size of the components upon can be expressed as comparison with a ladder/marker. The latter provides a more q specific determination of the separated components, but it still μ= requires several complex steps (membrane transferring, 6πrη (2) incubation with a primary antibody, washing, incubation with where q is the charge on the particle, η is the viscosity of the a secondary antibody, washing, signal revealing) and high media, and r is the hydrodynamic radius related to the size of volumes of antibodies. the particle. This means that under defined electrophoretic For the development of PEB, we took advantage of the conditions, small and highly charged molecules will move sample treatment capabilities of paper electrophoresis and toward the electrode with the opposite charge faster than bulky merged them with the convenience of LFAs. The result is a and neutral ones. From the perspective of sample treatment, diagnostic platform that can receive an untreated drop of blood this allows for an active and precise separation of contaminants and produce an easy-to-read optical signal in less than 15 min. from the target molecule (given the unique electrophoretic Design and Operation of the PEB Device. We designed properties of the different biomolecules present in blood/ the PEB device to assure its application at the point-of-care, serum samples), dramatically reducing nonspecific interac- retaining the portability, usability, and low cost of the LFA. For tions.31−33 this reason, we used the standard LFA materials: cellulose and Electrophoresis can be carried out either in free solution NC membranes. The cellulose pads were used as buffer (capillary electrophoresis) or in a support medium (paper or reservoirs to maintain the system constantly wet and the NC gel). The former offers fast and high-resolution separation but membrane was used as the main platform for carrying out both requires strong electric fields, which limits its application at the the electrophoretic separation and the signal generation. We POC (i.e., expensive and bulky equipment is needed to also used two sticky copper electrodes instead of applying the generate such electric fields, which may also be dangerous for voltage directly on the NC strip. Compared to crocodile clips the final, untrained user). The latter is simpler, cheaper, and or similar electrical connectors, the copper sticky tape provides enables facile visualization of the separation zone by direct a higher contact surface on the NC strip. In addition, it is staining. Within the types of support media, the gel provides cheap and can be disposed together with the paper strip once better resolution than paper but is more affected by heat oxidized (the copper electrode is corroded due to the oxidation generation, which causes variations in the particle’s migration reaction when used as an anode during electrophoresis). The rate and band’s distortion.32,34 Contrary to free electro- paper strip containing the copper electrodes is intended to be phoresis, the particle’s movement in support-medium-electro- disposable once the assay is performed. Regarding the reagents, phoresis is a sum of electrophoretic and electroendosmotic we used antibodies as bioreceptors (printed on the NC mobility. Electroendosmosis is the movement of liquid toward membrane like standard LFA) and analytes, while we an anode or a cathode depending on the type of support employed AuNPs and QDs as labels. To perform the medium, potential gradient, electrolyte ionic strength, and pH. electrophoresis separation, we designed a miniaturized and As a consequence, depending on the specific conditions, we portable power source as an alternative to the bulky and could observe charged molecules moving toward the electrode expensive power supply used in conventional electrophoresis with the same charge when the electroendosmosis forces are systems. We connected the smartphone battery to a Joule thief greater than the electrophoretic ones.35−37 inverter and a rectifier (Figure S1) to obtain a stable 200 v DC Unexpected particle’s mobility or band’s distortion is (0.1 mA, 20 mW). It is worth mentioning that although the commonly observed upon heat generation during gel electro- low intensity of the output current may suggest its inadequacy phoresis.38 This phenomenon is explained by Joule’s law and to perform electrophoresis, the high resistance of the system the generalized power equation (several megaohms) allows keeping a constant voltage in the V2 device. The paper strip size is 21 cm × 4 cm (Figure S3), P= weighs 2.1 grams, and costs around 1 euro (excluding the R (3) power source, which costs approximately 10 euros and weighs where (P) is the power converted from electrical energy to 10 grams). The assembly procedure of the different pads is thermal energy, (V) is the voltage applied, and (R) is the similar to that of LFA.18 A detailed description of the resistance to current.39 Considering this equation, electrical fabrication of the PEB device is in the Materials and Methods systems that exert higher resistance to current produce less section of this manuscript. heat. In this sense, paper electrophoresis is less prone to heat The operation of the sensor consists of few simple steps: (1) generation than agarose gel electrophoresis because the NC the two cellulose pads at the two ends of the sensor are wet 3114 https://dx.doi.org/10.1021/acs.analchem.0c04330 Anal. Chem. 2021, 93, 3112−3121 Analytical Chemistry pubs.acs.org/ac Article Figure 1. Schematic representation of the PEB strip and assay operation. The PEB strip is composed of an NC membrane strip functionalized with capture antibodies in TL and CL. (A) Two pieces of adhesive copper tape are stacked on both sides of the strip and serve as cathode and anode electrodes. Two pieces of cellulose membrane are stacked on both ends of the strip and serve as electrophoresis buffer reservoirs. The electrophoresis buffer is drop-casted on both cellulose pads and is gradually dispensed to the NC membrane. (B) Conjugate solution is mixed with the whole blood sample and it is drop-casted on the strip. (C) Sample moves by electrophoresis toward the detection zone. (D) Target analyte and the labeled antibodies are captured by the antibodies in TL and CL. The nonspecific interfering components of the whole blood sample flow out of the detection zone, enabling a clear visualization of the signal in TL and CL. Figure 2. Electrophoresis performance evaluation. (A) Effect of the NC membrane pore size on the electrophoretic mobility of different QDs. NC membranes, CN95, CN110, and CN150, were selected due to their big, medium, and small pore size, respectively. (B) Evaluation of the voltage effect applied on the electrophoresis mobility using the same QDs. (C) Effect of the pH value on the electrophoretic mobility. Data are shown as mean ± standard error of mean (SEM) of triplicate experiments. (D) G-QDs image at different pH values before applying the voltage 0 min (i) and after 10 min of voltage (ii). with the electrophoresis buffer, which contains the electrolytes for the whole duration of the assay (15 min). (2) Once the NC enabling the electric bridge between the cathode and the membrane is fully wet (approximately in 2 min from the anode. The high void volume of the two cellulose pads ensures addition of the electrophoresis buffer), the labeled bioreceptors the complete and homogenous wetting of the NC membrane and sample (6 μL) are drop-casted onto the membrane. (3) 3115 https://dx.doi.org/10.1021/acs.analchem.0c04330 Anal. Chem. 2021, 93, 3112−3121 Analytical Chemistry pubs.acs.org/ac Article The electrophoretic potential is applied to move the sample CdTe QDs compared to the G-QDs, their higher negative along the membrane, which will generate the optical readout. surface charge provided by the carboxyl functional groups The full operation is schematized in Figure 1. Additionally, a (−56.9 mV) makes them move faster (0.72 cm/min). demonstration video using a blood sample where the control The second parameter that we studied was the effect of line appears after running the electrophoresis assay can be different NC membranes over the migration speed. For this observed in Movie S1. Conversely, Movie S2 shows that there study, we used three different NC membranes (CN95, CN110, is no movement of the sample when the voltage is not applied and CN150) commonly used for the development of LFAs. (after the paper strip is totally wet with the buffer). Meaning Particularly, they provide three different capillary flow rates of that the sample does not move by capillarity in PEB and only 95 s/4 cm, 110 s/4 cm, and 150 s/4 cm for CN95, CN110, moves when the voltage is applied. Other authors have and CN150, respectively. The capillary flow rate represents the previously reported the inability of compounds to flow by required time for the sample to cover 4 cm of the NC capillarity in paper when it is soaked and in equilibrium.46 It is membrane, which is inversely related to the membrane pore important to note that, while here we demonstrate the proof of size. For example, CN150 with a capillary flow rate of 150 s/4 concept of the technology, the entire operation could be cm has the smallest pore size of the three membranes tested in condensed into a single-step procedure by keeping the sensor this work. The fabricants of NC membranes do not give wet in a sealed container and employing a conjugate pad on information on the pore size as the polymeric structure does top of the NC membrane. not actually create pores but rather a sponge-like pathway.49 Characterization of the PEB as an Electrophoretic We refer the reader to Figure S7 containing the scanning Platform. Given the complexity of paper-based electro- electron microscopy (SEM) images of the three different phoresis, we first characterized how using different materials membranes and the estimation of their relative pore size. and reagents during the fabrication of the PEB would affect its Again, the PEB device responded accordingly to the electro- electrophoretic behavior. We studied four main components: phoresis model. For this experiment, we used 10 mM SB buffer the particle mobility, the pore size of the NC membrane, the at pH 8, an operating voltage of 200 V (0.1 mA, 20 mW), and applied voltage, and the pH of the electrophoresis buffer the same QDs. As expected, the CN95 membrane produced (Figure 2). We evaluated how the size, shape, and charge of 2.67-, 1.37-, and 2.30-fold faster migration speed of G-QDs, nanoparticles affect their migration speed. According to the CdTe, and CdSe@ZnS QDs, respectively, compared to the concept of electrophoresis, we expect that bigger and less CN150 membrane (Figure 2A). charged particles would move slower than smaller and more The third parameter we tested was the applied voltage. In charged ones, given a fixed pore size, electrophoretic buffer, electrophoresis, under a uniform electric field, the field and applied potential. To validate this hypothesis, we used strength is directly proportional to the voltage applied and three different QDs: CdSe@ZnS QDs functionalized with hence to the particle’s velocity. Therefore, the higher the streptavidin (Ø 20 nm, considering the streptavidin), CdTe voltage, the faster the migration speed when the other QDs (Ø 5 nm), and G-QDs (Ø 2.5 nm). We chose these three parameters are fixed. From the engineering point of view of QDs because they cover the size range of most proteins and the sensor fabrication and design, this is a crucial parameter their charge has been previously studied. G-QDs have a high since we need to be able to achieve a quick migration applying negative charge due to the oxygen functional groups (C-OH, a relatively low voltage to have a sensor deployable at the CO, and COOH);47 CdTe QDs also have a negative charge point-of-care. In Figure 2B, we show the behavior of the due to their functionalization with carboxyl groups.48 The voltage on the electrophoretic mobility at 100 (0.05 mA, 5 streptavidin (pI ≈ 5) functionalization of the CdSe@ZnS QDs mW), 200 (0.1 mA, 20 mW), and 300 V (0.15 mA, 45 mW). also confers them a negative surface charge at pH 8. Their For this experiment, we used CN95 NC membrane because detailed characterization can be found in the Supporting this type of NC membrane provided similar speeds for the Information (emission spectra, excitation spectra, and trans- three QDs (Figure 2A). As expected, a higher voltage caused a mission electron microscopy (TEM) images are shown, faster movement of the QDs to the negative electrode. G-QDs, respectively, in Figures S4−S6). For this experiment, we CdTe, and CdSe QDs moved 2.86, 1.50, and 1.54-times faster fixed the field strength using 10 mM SB buffer at pH 8 and an when applying 300 V rather than 200 V. However, higher operating voltage of 200 V (0.1 mA, 20 mW) and the CN150 voltages resulted in less reproducible velocities. The standard NC membrane. Despite their negative surface charge, the QDs deviation of the velocities of G-QDs, CdTe, and CdSe QDs moved to the cathode (negative electrode) and this was were 2.56, 3.55, and 6.60-times higher when applying 300 V attributed to electroendosmosis, which is the flow of the rather than 100 V. The CdTe QDs were the fastest, particularly electrolyte solution. Since the buffer solution has a pH value they moved 1.3 times faster than the G-QDs when applying lower than the pKa value of the boric acid (pKa = 9.24), the 300 V. Instead, the CdSe QDs followed the same trend as in predominant concentration of the positively charged ions Figure 2A and moved slowest to the negative electrode. makes the electrolyte solution move toward the cathode, Finally, to evaluate the effect of the pH on the particle dragging the QDs with it. As expected, we found a correlation mobility direction and speed, we tested SB buffer (10 mM) at between the size/charge and the migration speed. In particular, three different pH values (8, 9, and 10). The particle’s rate and the CdSe@ZnS QDs moved at an average speed of 0.32 cm/ direction of movement are influenced by electrophoretic and min toward the negative electrode due to their bigger size (20 electroendosmotic mobilities. Variations in the pH have a nm) and the negative surface charge provided by the strong effect on both forces. In the former, the pH determines streptavidin (−60.3 mV). However, despite its z-potential the magnitude and sign of particle’s surface charge, while in the value is 3 times higher than the G-QDs one (−20 mV), the latter the pH determines the rate and direction of movement of bigger size of the CdSe@ZnS QDs (20 nm) compared to the the electrolyte solution. Being able to tune not only the speed G-QDs (2.5 nm) allows the G-QDs to move slightly faster but also the direction of the flow is paramount to achieve a (0.37 cm/min). Likewise, despite the 2-fold bigger size of complete sample treatment. In fact, just by changing the pH of 3116 https://dx.doi.org/10.1021/acs.analchem.0c04330 Anal. Chem. 2021, 93, 3112−3121 Analytical Chemistry pubs.acs.org/ac Article the electrophoresis buffer, we could separate the target analyte from other compounds. For this experiment, we challenged the PEB device with G-QDs since their surface functional groups are deprotonated when the pH value is higher than 7.50 This means that the G-QDs have a higher negative surface charge upon pH values higher than 7. However, we observed that at pH 8, the G-QDs moved toward the negative electrode, at pH 9 they slightly moved, and at pH 10 they moved toward the positive electrode (Figure 2C,D). This means that the mobility of the G-QDs was mostly influenced by the electroendosmotic mobility: at pH 8, the pH value of the buffer was lower than the pKa value of the boric acid (pKa = 9.24) and, therefore, the predominant ions in the electrolyte solution were positively charged. The buffer dragged the G-QDs to the negative electrode. At pH 9, the pH value of the buffer was nearly equal to the pKa value of the boric acid (pKa = 9.24); therefore, there were no predominantly charged ions in the electrolyte solution Figure 3. Separation of QDs by electrophoresis. (A) N−S-doped G- and the G-QDs had enough negative surface charge to slightly QDs are mixed with CdSe@ZnS QDs and drop-casted in the center move toward the positive electrode. Conversely, at pH 10, the of the NC membrane strip that is immersed in electrophoresis buffer, predominant ions in the electrolyte were negatively charged; pH 8. After running the electrophoresis for 10 min, the CdSe@ZnS thus, the buffer dragged the G-QDs toward the positive QDs (pink) moves to the left side of the strip (negative electrode), electrode. Once again, the PEB device was found to adhere to separating from the N−S-doped G-QDs (blue), which have slightly the laws of paper-based electrophoresis. moved to the left side. (B) CdTe QDs are mixed with CdSe@ZnS After demonstrating that the PEB device was responding QDs and drop-casted onto the center of the NC membrane strip that accordingly using simple homogenous buffer solutions, we is immersed in electrophoresis buffer, pH 8. After running the verified that our platform is able to separate different particles electrophoresis for 10 min, CdSe@ZnS QDs (orange) have moved from a heterogeneous sample. To achieve this, we used two faster to the left side (negative electrode), separating from the CdSe@ ZnS QDs (red), which have slightly moved to the left side. different solutions: one containing N−S-doped G-QDs and CdSe@ZnS QDs (at 25 nM), while the other containing CdTe QDs and CdSe@ZnS QDs (both at 25 nM). The choice of these two couples of QDs gives relevant information on the PEB Optimization for Gold Nanoparticles. Supported influence of the surface charge and size on the particle’s by the ability of the PEB device to separate different types of mobility. For instance, in the previous section, we observed QDs, we decided to test its ability to move also AuNPs. that the G-QDs moved faster toward the cathode than the Aiming at the development of a POC platform, the use of CdSe@ZnS QDs, mainly due to their big difference in size (2.5 AuNPs (the most used label in LFA) is fundamental to and 20 nm, respectively). However, we wanted to check if this translate the device from the benchtop to the market since they phenomenon could be conversed by manipulating the surface provide cost-effective synthesis, straightforward functionaliza- charge of the G-QDs. In this sense, we doped the G-QDs with tion, and naked-eye visualization.52−54 nitrogen and sulfur functional groups to increase their negative The AuNPs have a spherical shape and a diameter of around surface charge, which provides repulsion to the cathode. It is 15 ± 2 and 34.8 nm when measured in TEM and DLS, reported that the N−S-doped G-QDs have a more negative ζ- respectively. Their ζ-potential value is around −49.9 mV (the potential value than the G-QDs (−31.6 and −20 mV, reader can find a complete characterization of the AuNPs in respectively).47,51 The system set up included the use of the the SI (Figures S8 and S9)). In particular, we evaluated the SB buffer at pH 8, a potential of 200 V (0.1 mA, 20 mW), and mobility of both citrate-capped AuNPs and AuNPs function- the CN150 NC membrane. This membrane provides the alized with bovine serum albumin (BSA) and goat IgG (which slowest speeds and, therefore, the mobility progression of the are the detection reagents used in the final application). The QDs can be evaluated with more detail. Again, the system system set up consisted of the use of a CN150 NC membrane, responded as expected. Both QDs in the mixture moved 10 mM SB buffer, pH 8, and a potential of 200 V (0.1 mA, 20 toward the cathode due to electroendosmosis. However, the mW). The CN150 NC membrane runs at a slower speed N−S-doped G-QDs moved slower than the CdSe@ZnS QDs compared to CN95 and CN110 and thus more time to toward the negative electrode despite their smaller size, accurately determine the distance traveled by the function- indicating that the CdSe@ZnS QDs have less negative surface alized AuNPs. As expected, the citrate-capped AuNPs had charges than the N−S-doped G-QDs (Figure 3A). Instead, for higher mobility compared to the biofunctionalized AuNPs. the second mixture, the smaller CdTe QDs (diameter 5 nm) The electrophoretic velocity decreased 1.25-, 1.38-, and 1.51- moved faster than the bigger CdSe@ZnS QDs (Figure 3B), fold upon the functionalization with BSA (−44.1 mV, 38.0 which is similar to what we observed in Figure 3A. Crucially, in nm), BSA + anti-H-IgG (−35.2 mV, 56.3 nm), and BSA + both cases, we achieved the complete separation of different anti-H-IgG + H-IgG (target molecule) (−27.5 mV, 58.8 nm), QDs in the respective solutions in 10 min. This separation was respectively. This behavior is mainly due to the increase of the promoted by the difference in electrophoretic mobility of the hydrodynamic diameter of the AuNPs (BSA has an approx. QDs and not by electroendosmosis. This result supports the diameter of 3 nm, while IgG of 10 nm).55 As expected, we ability of the PEB device to carry out sample treatment for the observed that the AuNPs moved to the negative electrode due separation of both particles with different sizes and different to the movement of the buffer toward the cathode (electro- charges. endosmotic force) (Figure S10).35−37 3117 https://dx.doi.org/10.1021/acs.analchem.0c04330 Anal. Chem. 2021, 93, 3112−3121 Analytical Chemistry pubs.acs.org/ac Article Figure 4. Calibration curve for H-IgG detection using the smartphone camera. (A) Calibration curve with PEB in buffer media (black) and whole blood (red), covering the range from 0 to 1000 ng/mL. The curve equations with PEB in buffer and whole blood were (optical intensity = 2 × 10−7 [H-IgG (ng/mL)]2.0248 r2 = 0.98) and (optical intensity = 1 × 10−6 [H-IgG (ng/mL)]1.7807 r2 = 0.98), respectively. (B) Correlation between the signals obtained in PEB when performing the calibration curve (0−1000 ng/mL) in buffer and whole blood (optical intensity PEB Buffer = 0.9187, optical intensity PEB Whole blood + 0.0124 r2 = 0.99). Data are shown as mean ± SEM of triplicate experiments. (C) Pictures of PEB strips after performing the calibration curve (0−1000 ng/mL) in buffer and whole blood. (D) Pictures of LFA strips after the detection of H-IgG (1000 ng/ mL) in (i) 5-fold diluted whole blood, (ii) 2.4-fold diluted, and (iii) undiluted whole blood. The inset picture showing a zoom-in on the detection area. PEB Optimization for Whole Blood Analysis. As we LFAs developed in our lab and (2) the detection of H-IgG in previously described, the type of NC membrane can blood represents a real clinical scenario, as shown by the recent dramatically affect the performance of the PEB device. For pandemic.1,56 To facilitate the interpretation of the results and this reason, we tested the ability of CN95, CN110, and CN150 limiting the sample-to-sample variation, we spiked chicken membranes to move whole blood while applying a potential of whole blood with H-IgG, in this way, we removed the 200 V (0.1 mA, 20 mW) in the SB buffer at pH 8. Similar to variability lead by the endogenous concentration of human IgG what is reported in other works, we observed that the flow rate in human blood, which is in the range of 3−16 mg/mL.57 A was greater at the anode and decreased toward the cathode.45 detailed experimental procedure can be found in the SI section, This effect is related to the equilibrium reached between the while the optimization of the preincubation step can be found electroendosmotic force and the osmotic pressure, two in Figure S12. opposite forces leading to the accumulation of liquid nearby Before going to the analysis of whole blood, we the cathode. This means that while the applied potential characterized the analytical performance of the PEB device induces the movement of the electrolyte solution toward the and compared it with the one of LFAs using buffered solutions cathode, the accumulation of the electrolyte solution at the spiked with H-IgG. We performed a calibration curve in LFA cathode creates an osmotic pressure toward the anode until covering the range from 0 to 1000 ng/mL for the H-IgG reaching the equilibrium of forces. The result is that the sample concentration (values above 1000 ng/mL usually produces the stops moving after 15 min. We observed that CN95 was the hook effect, by which we observe a decrease in the signal due only NC membrane capable of washing out the blood from the to the saturation of the binding domain of the bioreceptors) detection zone in 15 min (see Figure S11). This is according to (Figure S13).58 We then performed a calibration curve in PEB the electrokinetic theory, stating that bigger pores produce covering the same range of H-IgG (Figure 4A). The LoD and faster mobility.53 While a lengthier strip (the equilibrium of LoQ achieved in LFA were 265.56 and 1876.94 ng/mL, while forces would be reached later) together with more volume of the ones for PEB were 489.78 and 856.09 ng/mL, respectively. electrophoresis buffer (to avoid fast evaporation) would enable These are similar to the publications recently reported by our a longer assay time, we preferred to use a bigger pore (CN95) group.10 Besides showing a higher LoD, PEB proved a 2.2-fold membrane in the next experiments since it allows the assay to lower LoQ and a 4.5-fold higher sensitivity than LFA when be performed in just 15 min. comparing the curve slopes (0.44 and 2.02 for LFA and PEB, Detection of H-IgG in Whole Blood. As the final step respectively). It is noteworthy that the strong affinity of the toward demonstrating the applicability of the PEB device as a bioreceptors to the target analyte created a stable signal in TL point-of-care diagnostic platform, we employed it for the and avoided the dragging of the labeled bioreceptors from the detection of a biomarker in spiked whole blood. As the test TL by the strong electric field. bed, we used a model system previously optimized in our lab Finally, supported by the positive results obtained with for the detection of human IgG. The choice of this analyte has buffer solutions, we used the same range of concentrations but two main reasons: (1) it provides a convenient comparison to using undiluted whole blood (Figure 4A,B). The PEB device 3118 https://dx.doi.org/10.1021/acs.analchem.0c04330 Anal. Chem. 2021, 93, 3112−3121 Analytical Chemistry pubs.acs.org/ac Article produced a specific and sensitive response (LoD = 284.52 ng/ min and the flow of blood along NC membranes. Regarding mL and LoQ = 441.07 ng/mL). It is noteworthy that PEB the former, the results we obtained indicate that using NC showed lower LoD and LoQ when detecting H-IgG in whole membranes, whose porosity is orders of magnitude higher than blood than in buffer medium. The reason behind this is that the diameter of the nanoparticles generally employed in LFA, the materials and electrophoretic conditions used in PEB have the main factor governing the particle movement is their been specifically optimized to achieve the highest sensitivity surface charge, rather than their size. We believe that the ability when analyzing spiked whole blood (details in Materials and of PEB to purify particles opens a myriad of opportunities for Methods). Conversely, buffer medium flow 1.83-fold faster the treatment and detection of different types of biomarkers. than whole blood when using the optimized conditions for Instead, the latter shows how PEB can handle complex and whole blood (see Figures S11 and S12). This has an impact on highly viscous samples without the need for extensive dilutions the assay’s sensitivity, which is highly affected by the flow rate (which would also decrease the amount of target) or extra (i.e., a faster flow decreases the interaction time between target washing steps (forcing the user to actuate at different times on and bioreceptors).59 Although PEB enabled the detection of the same test). In our opinion, this is a game changer within H-IgG in whole blood, avoiding, on the one hand, nonspecific the field of point-of-care sensors, allowing the use of media interactions in TL for blank samples and, on the other hand, such as blood, serum, and other viscous solutions in a simpler responding accordingly to the increase in the H-IgG and more effective way. In addition, although PEB operation is concentration (Figure 4C), CN95-based LFA failed in based on electrophoresis, it can deal with biological samples producing a reading signal due to the lack of movement by with different pH values or salt content. The reason behind the blood sample (Figure 4D). this is that the electrophoresis buffer (not the sample contents) The LFA strips were fabricated following the procedure is the one leading the flow rate and direction of the sample. reported in the materials and methods section, and a The almost 3 orders of magnitude difference in volumes used commercial cellulose membrane specifically designed for the for the electrophoresis buffer (2000 μL) and sample (6 μL) analysis of whole blood samples was used as the sample pad. makes the electrolytes of the buffer prevail over whole blood The whole blood sample did not flow properly until it was salts. Moreover, since there is not a significant diffusion diluted 5-fold in buffer medium, but still the background signal between the sample and the buffer during the electrophoresis, remained high and favored the masking of the red colorimetric the buffer is leading all the forces, and hence, it is the one signal in TL. In addition, the dilution of the whole blood responsible for the flow rate and direction modulation. In sample also implied the dilution of the spiked concentration of addition, it is reported that the whole blood samples have H-IgG up to 200 ng/mL, which was lower than the LoD. generally an ionic strength of 0.15 M60 and it is also reported Therefore, the extra dilution of the whole blood sample that the factors determining protein’s mobility in electro- improved its flow along the strip but was not appropriate from phoresis (retardation coefficients (KR), free mobilities (Yo), an analytical viewpoint. and Rf values) are not affected by NaCl in the samples up to ■ FINAL REMARKS AND CONCLUSIONS We observed three main forces affecting the movement of 0.8 M. In fact, the main issue with high-concentrated salt samples in electrophoresis is that SDS cannot bind to proteins.61 But this is not our case since SDS is not used in particles within the paper-based electrophoretic platform: PEB. Regarding the pH of electrophoresis buffer, this should electroendosmosis, electrophoresis, and osmosis. First, we be optimized for any given application. Similarly, blood saw that the mobility of QDs was influenced by electro- samples with different hematocrit levels would have a greater endosmosis since all QDs moved toward the cathode (in influence on analytical variability in LFA than in PEB since in electrolytic cells, the cathode is negatively charged), despite PEB all of the whole blood samples are removed from the their negative surface charge. We verified this hypothesis by detection zone by electrophoresis. changing the pH of the electrophoresis buffer: when the pH Finally, we validated the PEB for the detection of H-IgG in was lower than the pKa value of the boric acid (pKa = 9.24), buffered medium and spiked whole blood and compared its the movement of the electrolytes toward the cathode analytical performance with LFA. PEB proved a 2.2-fold lower represented the major force of the system, also dragging the LoQ and a 4.5-fold higher sensitivity (when comparing the negative particles toward the cathode. Instead, for pH close to curve slopes) than LFA for H-IgG detection in buffer media. the boric acid pKa value, the absence of the electrolyte’s Moreover, while LFAs were incapable of directly handling movement made the electrophoresis the main force, directing undiluted blood, the signals obtained with PEB when detecting the negatively charged nanoparticles toward the anode. H-IgG in spiked whole blood showed a correlation of 0.99 with Obviously, for pH higher than the pKa value, both the the ones for the detection in buffer samples. Therefore, PEB electrophoretic and electroendosmosis forces were moving the proved to be capable of maintaining its outstanding analytical particles toward the anode. Then, we observed the influence of sensitivity when evaluating nondiluted and nonpretreated electrophoresis in the PEB function since factors such as the spiked whole blood. While it is noteworthy that the current particle size, particle surface charge, membrane pore size, and approach consists of four main steps, these are easy to be applied voltage were enhancing or hindering the particle condensed into two steps (current switch-on and sample mobility, following the general principle of electrophoresis. addition). The relying of the PEB on the connection of the Additionally, we observed the effect of the osmosis limiting the electrodes and the power source can be easily re-engineered by amount of volume that could be displaced. the insertion of a power button, whereas the labeling of the The PEB demonstrated to perform efficient sample sample and the addition of the electrophoresis buffer can be treatment keeping the main characteristics of point-of-care solved by the storage of the buffer in burstable containers, such devices. In fact, employing either a portable battery or a as the ones reported by Prof. Mc Neil’s group,62 and the smartphone, we managed to perform two key activities: the incorporation of a dedicated conjugate pad using dissolvable separation of different types of nanoparticles in less than 10 wax barriers, such as the ones reported by our group.11 Most of 3119 https://dx.doi.org/10.1021/acs.analchem.0c04330 Anal. Chem. 2021, 93, 3112−3121 Analytical Chemistry pubs.acs.org/ac Article the commercially available LFA kits are based on a two-step manuscript. All authors reviewed and approved the final operation; thus, we believe that the proposed optimizations of version of the manuscript. the prototype are compatible with the application of the Notes concept at the point-of-care. ■ The authors declare no competing financial interest. * ASSOCIATED CONTENT sı Supporting Information The Supporting Information is available free of charge at ■ ACKNOWLEDGMENTS We acknowledge financial support from the NACANCEL https://pubs.acs.org/doi/10.1021/acs.analchem.0c04330. project PCIN-2016-066 (program Euronanomed 2). This Materials, reagents, and instruments; conjugation work is also funded by the CERCA Program/Generalitat de protocol of AuNPs with anti-H-IgG antibodies; Catalunya. The ICN2 is funded by the CERCA program/ detection protocol of H-IgG in PEB (Figures S1−S13) Generalitat de Catalunya. ICN2 acknowledges the support of (PDF) the Spanish MINECO for the Project MAT2017-87202-P and Demo video of the PEB device (Movie S1) (AVI) through the Severo Ochoa Centers of Excellence Program under Grant SEV2201320295. R.A.-D. acknowledges the Demo video when voltage is not applied (Movie S2) financial support from the EU Graphene Flagship Core 3 (AVI) ■ Project (No. 881603). A.S.-T. acknowledges the Autonomous University of Barcelona (UAB) for the possibility of perform- AUTHOR INFORMATION ing this work inside the framework of Biotechnology Ph.D. Corresponding Author Program. C.P. acknowledges Marie Skłodowska-Curie Actions Arben Merkoçi − Nanobioelectronics & Biosensors Group, Individual Fellowship, and this project has received funding Institut Català de Nanociència I Nanotecnologia (ICN2), from the European Union’s Horizon 2020 research and CSIC and The Barcelona Institute of Science and Technology innovation program under the Marie Skłodowska-Curie grant (BIST), 08193 Barcelona, Spain; Catalan Institution for agreement No. 795635. We also acknowledge Prof. Artur Research and Advanced Studies (ICREA), 08010 Barcelona, Xavier Roig from the Veterinary faculty in UAB for providing Spain; orcid.org/0000-0003-2486-8085; whole blood. ■ Email: arben.merkoci@icn2.cat Authors REFERENCES Amadeo Sena-Torralba − Nanobioelectronics & Biosensors (1) Udugama, B.; Kadhiresan, P.; Kozlowski, H. N.; Malekjahani, A.; Group, Institut Català de Nanociència I Nanotecnologia Osborne, M.; Li, V. Y. C.; Chen, H.; Mubareka, S.; Gubbay, J. B.; (ICN2), CSIC and The Barcelona Institute of Science and Chan, W. C. W. ACS Nano 2020, 14, 3822−3835. Technology (BIST), 08193 Barcelona, Spain (2) Waggoner, J. J.; Pinsky, A. J. Clin. Microbiol. 2016, 54, 860−867. (3) Yager, P.; Domingo, G. J.; Gerdes, J. Annu. Rev. Biomed. Eng. Ruslan Alvarez-Diduk − Nanobioelectronics & Biosensors 2008, 10, 107−144. Group, Institut Català de Nanociència I Nanotecnologia (4) Price, C. P. Br. Med. J. 2001, 322, 1285−1288. (ICN2), CSIC and The Barcelona Institute of Science and (5) Chin, C. D.; Linder, V.; Sia, S. K. Lab Chip 2012, 12, 2118− Technology (BIST), 08193 Barcelona, Spain; orcid.org/ 2134. 0000-0002-9876-1574 (6) Yetisen, A. K.; Akram, M. S.; Lowe, C. R. Lab Chip 2013, 13, Claudio Parolo − Nanobioelectronics & Biosensors Group, 2210−2251. Institut Català de Nanociència I Nanotecnologia (ICN2), (7) Quesada-González, D.; Merkoçi, A. Biosens. Bioelectron. 2015, 73, CSIC and The Barcelona Institute of Science and Technology 47−63. (BIST), 08193 Barcelona, Spain; orcid.org/0000-0001- (8) Parolo, C.; Greenwood, A. S.; Ogden, N. E.; Kang, D.; Hawes, 9481-4408 C.; Ortega, G.; Arroyo-Currás, N.; Plaxco, K. W. Microsyst. Nanoeng. 2020, 6, 4−11. Helena Torné-Morató − Nanobioelectronics & Biosensors (9) Posthuma-Trumpie, G. A.; Korf, J.; Van Amerongen, A. Anal. Group, Institut Català de Nanociència I Nanotecnologia Bioanal. Chem. 2009, 393, 569−582. (ICN2), CSIC and The Barcelona Institute of Science and (10) Quesada-González, D.; Sena-Torralba, A.; Wicaksono, W. P.; Technology (BIST), 08193 Barcelona, Spain de la Escosura-Muñiz, A.; Ivandini, T. A.; Merkoçi, A. Biosens. Alexander Müller − Nanobioelectronics & Biosensors Group, Bioelectron. 2019, 132, 132−135. Institut Català de Nanociència I Nanotecnologia (ICN2), (11) Sena-Torralba, A.; Ngo, D. B.; Parolo, C.; Hu, L.; Á lvarez- CSIC and The Barcelona Institute of Science and Technology Diduk, R.; Bergua, J. F.; Rosati, G.; Surareungchai, W.; Merkoçi, A. (BIST), 08193 Barcelona, Spain Biosens. Bioelectron. 2020, 168, No. 112559. (12) Millipore, M. Rapid Lateral Flow Test Strips Considerations for Complete contact information is available at: Product Development; EMD Millipore Corporation: Billerica, MA, https://pubs.acs.org/10.1021/acs.analchem.0c04330 USA, 2013. (13) Brangel, P.; Sobarzo, A.; Parolo, C.; Miller, B. S.; Howes, P. D.; Author Contributions Gelkop, S.; Lutwama, J. J.; Dye, J. M.; McKendry, R. A.; Lobel, L.; § A.S.-T. and R.A.-D. contributed equally to this work. Stevens, M. M. ACS Nano 2018, 12, 63−73. Author Contributions (14) Li, H.; Han, D.; Hegener, M. A.; Pauletti, G. M.; Steckl, A. J. Biomicrofluidics 2017, 11, 1−15. A.S.-T. and R.A.-D. contributed equally to this work. A.S.-T., (15) Gwyn, S.; Mitchell, A.; Dean, D.; Mkocha, H.; Handali, S.; R.A.-D., and C.P. designed the experiments. R.A.-D. designed Martin, D. L. J. Immunol. Methods 2016, 435, 27−31. and developed the power source. A.S.-T., R.A.-D., and A.M. (16) Kim, J.; Cao, X. E.; Finkelstein, J. L.; Cárdenas, W. B.; Erickson, designed and fabricated the device. A.S.-T., H.T.-M., and A.M. D.; Mehta, S. Malar. J. 2019, 18, 1−10. performed the experiments under the supervision of R.A.-D., (17) Zhang, Y.; Liu, X.; Wang, L.; Yang, H.; Zhang, X.; Zhu, C.; C.P., and A.M., A.S.-T., R.A.-D., and C.P. wrote the main Wang, W.; Yan, L.; Li, B. Sci. Rep. 2020, 10, No. 9604. 3120 https://dx.doi.org/10.1021/acs.analchem.0c04330 Anal. Chem. 2021, 93, 3112−3121 Analytical Chemistry pubs.acs.org/ac Article (18) Parolo, C.; Sena-Torralba, A.; Bergua, J. F.; Calucho, E.; (50) Kumawat, M. K.; Thakur, M.; Gurung, R. B.; Srivastava, R. Sci. Fuentes-Chust, C.; Hu, L.; Rivas, L.; Á lvarez-Diduk, R.; Nguyen, E. Rep. 2017, 7, No. 15858. P.; Cinti, S.; Quesada-González, D.; Merkoçi, A. Nat. Protoc. 2020, 15, (51) Zou, S.; Hou, C.; Fa, H.; Zhang, L.; Ma, Y.; Dong, L.; Li, D.; 3788−3816. Huo, D.; Yang, M. Sens. Actuators, B 2017, 239, 1033−1041. (19) Vivid. Plasma Separation Membrane. https://shop.pall.com/ (52) Aldewachi, H.; Chalati, T.; Woodroofe, M. N.; Bricklebank, N.; us/en/medical/advanced-materials/diagnostics/vivid-plasma- Sharrack, B.; Gardiner, P. Nanoscale 2018, 10, 18−33. separation-membrane-zidgri78lls (accessed Feb 27, 2020). (53) Cordeiro, M.; Carlos, F. F.; Pedrosa, P.; Lopez, A.; Baptista, P. (20) Ahlstrom-Munksjö. Lateral flow plasma separation pads. V. Diagnostics 2016, 6, 43. https://www.ahlstrom-munksjo.com/products/medical-life-sciences- (54) Quesada-González, D.; Merkoçi, A. Chem. Soc. Rev. 2018, 47, and-laboratory/lateral-flow-test-pads/plasma-separation-pads/ (ac- 4697−4709. cessed Nov 22, 2020). (55) González Flecha, F. L. G.; Levi, V. Biochem. Mol. Biol. Educ. (21) Lu, Z.; Rey, E.; Vemulapati, S.; Srinivasan, B.; Mehta, S.; 2003, 31, 319−322. Erickson, D.; Health, G. Lab Chip 2018, 18, 3865−3871. (56) Bedford, J.; Enria, D.; Giesecke, J.; Heymann, D. L.; Ihekweazu, (22) Gong, M. M.; MacDonald, B. D.; Nguyen, T. V.; Nguyen, K.; C.; Kobinger, G.; Lane, H. C.; Memish, Z.; don Oh, M.; Sall, A. A.; Van; Sinton, D. Biomicrofluidics 2013, 7, No. 044111. Schuchat, A.; Ungchusak, K.; Wieler, L. H. Lancet 2020, 395, 1015− (23) Tate, J.; Ward, G. Clin. Biochem. Rev. 2004, 25, 105−120. 1018. (24) Nash, M. A.; Waitumbi, J. N.; Hoffman, A. S.; Yager, P.; (57) Cassidy, J. T.; Nordby, G. L. J. Allergy Clin. Immunol. 1975, 55, 35−48. Stayton, P. S. ACS Nano 2012, 6, 6776−6785. (58) Rey, E.; O’Dell, D.; Mehta, S.; David Erickson, P. Anal. Chem. (25) Dong, Y.; Shao, J.; Chen, C.; Li, H.; Wang, R.; Chi, Y.; Lin, X.; 2017, 89, 5095−5100. Chen, G. Carbon 2012, 50, 4738−4743. (59) Bishop, J. D.; Hsieh, H. V.; Gasperino, D. J.; Weigl, B. H. Lab (26) Dong, Y.; Pang, H.; Yang, H.; Bin; Guo, C.; Shao, J.; Chi, Y.; Li, Chip 2019, 19, 2486−2499. C. M.; Yu, T. Angew. Chem., Int. Ed. 2013, 52, 7800−7804. (60) Covington, A. K.; Robinson, R. A. Anal. Chim. Acta 1975, 78, (27) Turkevich, J.; Stevenson, P. C.; Hillier, J. Discuss. Faraday Soc. 219−223. 1951, 11, 55−75. (61) See, Y. P.; Olley, P. M.; Jackowski, G. Electrophoresis 1985, 6, (28) Ambrosi, A.; Castañeda, M. T.; Killard, A. J.; Smyth, M. R.; 382−387. Alegret, S.; Merkoçi, A. Anal. Chem. 2007, 79, 5232−5240. (62) Fernández-Sánchez, C.; McNeil, C. J.; Rawson, K.; Nilsson, O. (29) Á lvarez-Diduk, R.; Orozco, J.; Merkoçi, A. Sci. Rep. 2017, 7, Anal. Chem. 2004, 76, 5649−5656. No. 976. (30) Armbruster, D. A.; Pry, T. Clin. Biochem. Rev. 2008, 29, S49− S52. (31) Adamson, N. J.; Reynolds, E. C. J. Chromatogr. B: Biomed. Sci. Appl. 1997, 699, 133−147. (32) Fritsch, R. J.; Electrophoresis, I. K. Encyclopedia of Food Sciences and Nutrition; Academic Press, 2003; pp 2055−2062. (33) Stalcup, A. M. Chiral Separations by Capillary Electrophoresis. In Chiral Analysis; Elsevier, 2006; pp 241−275. (34) Perrett, D. Electrophoresis. In Encyclopedia of Separation Science; Academic Press, 2000; pp 103−118. (35) Ferencik, M. Preparation of Pure Immunoglobulins. In Handbook of Immunohistochemistry; Springer, 2012; pp 191−196. (36) Bikerman, J. J. Electric Surface Phenomena. In Surface Chemistry: Theory and Applications; Academic Press, 1958; pp 405− 408. (37) Guo, Y.; Li, X.; Fang, Y. Prog. Biochem. Biophys. 1997, 24, 268− 272. (38) Zarei, M.; Goharshadi, E. K.; Ahmadzadeh, H.; Samiee, S. RSC Adv. 2015, 5, 88655−88665. (39) Experts, D. Electricity, Magnetism and Light. In The General Science Compendium; Disha Publications, 2018; pp 39−50. (40) Takahashi, M.; Takenaka, H. Polym. J. 1983, 15, 625−629. (41) Kandadai, M. A.; Raymond, J. L.; Shaw, G. J. Mater. Sci. Eng., C 2012, 32, 2664−2667. (42) Dangerfield, W. G.; Smith, E. B. J. Clin. Pathol. 1955, 8, 132− 139. (43) Mabry, C. C.; Karam, E. A. Am. J. Clin. Pathol. 1964, 34, 143− 152. (44) Nanthasurasak, P.; Cabot, J. M.; See, H. H.; Guijt, R. M.; Breadmore, M. C. Anal. Chim. Acta 2017, 985, 7−23. (45) Woods, E. F.; Gillespie, J. M. J. Biol. Sci. 1953, 6, 130−141. (46) Lan, W. J.; Maxwell, E. J.; Parolo, C.; Bwambok, D. K.; Subramaniam, A. B.; Whitesides, G. M. Lab Chip 2013, 13, 4103− 4108. (47) Lee, K. H.; Park, H.; Eom, W.; Kang, D. J.; Noh, S. H.; Han, T. H. J. Mater. Chem. A 2019, 7, 23727−23732. (48) Quevedo, I. R.; Tufenkji, N. Environ. Sci. Technol. 2012, 46, 4449−4457. (49) Wong, R.l.; Tse, H. Lateral Flow Immunoassay, 1st ed.; Springer, 2009. 3121 https://dx.doi.org/10.1021/acs.analchem.0c04330 Anal. Chem. 2021, 93, 3112−3121