Multiplex biosensing based on highly sensitive ... - ACS Publications

34 downloads 172227 Views 3MB Size Report
detection limit and 7-order linear dynamic range for volumetric registration of mag- ... BoNT detection in milk, apple and orange juices. ... portable device.
Article pubs.acs.org/ac

Multiplex Biosensing Based on Highly Sensitive Magnetic Nanolabel Quantification: Rapid Detection of Botulinum Neurotoxins A, B, and E in Liquids Alexey V. Orlov,† Sergey L. Znoyko,† Vladimir R. Cherkasov,‡ Maxim P. Nikitin,†,‡ and Petr I. Nikitin*,†,§ †

Prokhorov General Physics Institute, Russian Academy of Sciences, 38 Vavilov Street, Moscow, 119991, Russia Moscow Institute of Physics and Technology, 9 Institutskii per., Dolgoprudny, Moscow Region, 141700, Russia § National Research Nuclear University MEPhI (Moscow Engineering Physics Institute), 31 Kashirskoe Shosse, Moscow, 115409, Russia ‡

S Supporting Information *

ABSTRACT: We present a multiplex quantitative lateral flow (LF) assay for simultaneous on-site detection of botulinum neurotoxin (BoNT) types A, B, and E in complex matrixes, which is innovative by virtually no sacrifice in performance while transition from the single-plex assays and by characteristics on the level of laboratory quantitative methods. The novel approach to easy multiplexing is realized via joining an on-demand set of single-plex LF strips, which employ magnetic nanolabels, into a miniature cylinder cartridge that mimics LF strip during all assay stages. The cartridge is read out by an original portable multichannel reader based on the magnetic particle quantification technique. The developed reader offers the unmatched 60 zmol detection limit and 7-order linear dynamic range for volumetric registration of magnetic labels inside a cartridge of several millimeters in diameter regardless of its optical transparency. Each of the test strips, developed here as building blocks for the multiplex assay, can be used “as is” for autonomous quantitative single-plex detection with the same measuring setup, exhibiting the limits of detection (LOD) of 0.22, 0.11, and 0.32 ng/mL for BoNT-A, -B, and -E, respectively. The proposed multiplex assay has demonstrated the remarkably similar LOD values of 0.20, 0.12, 0.35 ng/mL under the same conditions. The multiplex assay performance was successfully validated by BoNT detection in milk and apple and orange juices. The developed methods can be extended to other proteins and used for rapid multianalyte tests for point-of-care in vitro diagnostics, food analysis, biosafety and environmental monitoring, forensics, and security, etc.

B

The gold standard for BoNT detection is the mouse bioassay (MBA) that detects the doses ∼10−20 pg. 10 BoNT quantification with the MBA requires multiple animals, and the results depend on their age and strain. The duration of this lethal assay (1−4 days), automation difficulties, as well as the ethical issues restrict its use by specialized laboratories. The promising in vitro alternatives to the MBA include various immunoassays such as the enzyme-linked immunosorbent assay (ELISA)11,12 and its modifications employing signal amplification,13 fluorescence14 and luminescence immunochemical methods,15,16 and the methods based on specific immune- and enzyme activity of the BoNT fragments.17 Some of these methods offer close or even better limit of detection (LOD) than the MBA at relatively short assay times of 2−4 h. These assays, though, require sophisticated equipment and highly qualified personnel available predominantly in laboratories.4

otulinum neurotoxin (BoNT) is the most acutely lethal toxin known, with an estimated human median lethal dose (LD50) of 1.3−2.1 ng/kg intravenously or intramuscularly and 10−13 ng/kg when inhaled.1 The toxin is a protein produced by anaerobic bacteria of Clostridium botulinum widely present in soil and water. Besides, BoNT is increasingly used in medicine, e.g., in BOTOX and Dysport cosmetics.2,3 BoNT intoxication causes a life-threatening illness of botulism in humans and animals and can occur following ingestion of contaminated food, from a wound infection,4 overdose of cosmetological or pharmaceutical products, or from a bioterrorism attack.5 In foodborne botulism, symptoms can appear as early as 2 h.6 If untreated, they may progress to cause respiratory failure and paralysis.7 BoNT comprises several antigenically and serologically distinct neurotoxins (serotypes A−G);8 the serotypes A, B, E and, rarely, F produce botulism in humans while C and D affect animals only.9 The high toxicity, wide availability, and easy preparation of these substances as potential warfare agents dictate a growing need for rapid, highly sensitive, and simple tests for on-site screening in various complex matrixes. © 2016 American Chemical Society

Received: May 26, 2016 Accepted: October 6, 2016 Published: October 6, 2016 10419

DOI: 10.1021/acs.analchem.6b02066 Anal. Chem. 2016, 88, 10419−10426

Article

Analytical Chemistry

to signal readout. The magnetic particles employed as labels are quickly registered over the entire cartridge volume by nonlinear magnetization38−40 with a reader based on the magnetic particle quantification (MPQ) method which was efficiently used for logic gated biocomputing biosensing on LF strips and 3D filters.41 A new ultrasensitive three-channel portable reader has been developed that offers simultaneous detection of as low as 60 zmol or 0.4 ng of magnetic nanolabels in each 0.2-mL channel and features extremely wide 7-order linear dynamic range. Remarkably, the analytical characteristics demonstrated while multiplex detection of the neurotoxins in various complex matrixes are on the level of laboratory quantitative methods. The proposed assay can be easily extended to multiplex on-site quantitative detection of other proteins and can be used in ecological monitoring, food control, health and safety protection, etc.

The current trend toward detecting hazardous agents on-site is imperative for many fields such as ecological monitoring, food control, health and safety protection, etc. The on-site techniques are expected to take minimal time, be simple enough for nonqualified personnel, and implemented with a portable device. Another priority is elevated stability of the assay reagents and minimal matrix effect to have reliable results in complex matrixes and food samples without device recalibrating.18 Adding the multiplex feature, i.e., capability of concurrent identification of several agents in one sample, may substantially enrich the diagnostic value of the assay. The simultaneously obtained data are mutually consistent and provide more comprehensive information, e.g., about the whole repertoire of specific markers of a disease, and in the course of identification of hazardous agents such as different BoNT serotypes may help to quickly trace the source of contamination.19 Lateral flow (LF) methods meet the majority of the mentioned requirements. They are rapid, are easy to handle, employ stable dry-chemistry reagents, and some of them can be implemented multiplex.20,21 The development of multiplex LF assays is generally complicated. The great majority of the LF multiplex tests underperform the respective single-plex peers because one should optimize not only every participating single-plex test but also take into account cross-reactivity issues. The LF methods based on various optical labels (e.g., gold, latex, quantum dots, etc.)18,21−23 exhibit the sensitivity of 1−50 ng/mL BoNT at the total assay duration about 30 min.21 The complexity of realization of high-precision quantitative LF measurements, especially in opaque mediums, is intrinsic to optical registration, but it can be easily overcome with nonoptical labels, in particular, magnetic nanoparticles (MP). The MP suggest potentially high analytical sensitivity along with high stability and low background signal regardless of optical transparency of the analyzed media.24−26 A wide variety of methods based on magnetic immuno-chromatography testing (MICT),27,28 nonlinear magnetization,29 giant magnetic resistive biosensors,30,31 etc. have been developed for registration of the magnetic nanoparticles (MP) in the LF assays. However, multiplex quantitative techniques based on these approaches are still under investigation. The multiplexing LF tests are commonly realized by immobilization of different antibodies in various detection areas within a single LF strip as several test lines, microdots, etc.32,33 Alternatively, several monospecific LF strips may share one sample pad.34,35 The first realization is generally complicated because of cross-reactivity issues.36,37 The second approach has been implemented, to the best of our knowledge, solely with the optical labels. In spite of the considerable number of analytical techniques, high-precision quantitative multiplex assays are still to be developed for easy, inexpensive, and rapid on-site detection in complex matrixes, which use stable reagents and provide the analytical parameters comparable to those of quantitative laboratory tests. Here, we report the development of the first multiplex quantitative lateral flow assay for simultaneous on-site detection of botulinum neurotoxin types A, B, and E, which is notable for virtually no sacrifice in characteristics compared with highly sensitive single-plex tests and has the ability to analyze even opaque mediums. The novel concept of easy multiplexing is realized via joining several LF strips of different specificity into a miniature cylinder cartridge to be used as an integrated consumable during all assay stages from incubation with sample



EXPERIMENTAL SECTION Materials and Reagents. Formaldehyde inactivated toxoids of botulinum neurotoxins types A, B, and E were obtained from MicroGen (Moscow, Russia). The neurotoxins were produced and purified from the respective Clostridium botulinum cultures. Monoclonal anti-BoNT-A antibody clone BTA-232, clone KBA468; anti-BoNT-B antibody clone BTB224 and clone KBB36 were purchased from the Russian Research Center of Molecular Diagnostic and Therapy (Moscow, Russia). Polyclonal rabbit anti BoNT-E antibodies were kindly provided by Dr. I. D. Vinogradova (Gamaleya Federal Research Center for Epidemiology and Microbiology, Moscow, Russia). The specificity and sensitivity of antibodies to native neurotoxins was carried out by the producer according to the procedure described by Abbasova et al.42 2-Morpholinoethanesulfonic acid (MES) was purchased from AppliChem GmbH (Germany). Bovine serum albumin (BSA), phosphate buffered saline (PBS) tablets, N-(3dimethylaminopropiyl)-N-ethylcarbodiimide hydrochloride (EDC), Tween 20, and N-hydroxysulfosuccinimide sodium salt, sulfo-NHS were purchased from Sigma-Aldrich (Germany). Boric acid was obtained from Panreac Quimica (Spain). Deionized water, Milli-Q grade, was from Millipore Corp. (USA). All other chemicals were of analytical grade. Superparamagnetic 198 nm carboxyl-modified (COOH−) Bio-Estapor Microspheres were purchased from Estapor− Merck Millipore (Germany). Nitrocellulose (NC) membrane UniSart CN140 (260 μm thick and 100 μm backing) was kindly donated by Sartorius AG (Germany). Sample pads, conjugate pads and absorbent sinks/wicking pads were purchased from Ahlstrom CytoSep (Finland). Backing cards were obtained from Lohmann (USA). Preparation of “Antibody−Magnetic Nanoparticle” Conjugates. All steps were carried out at room temperature unless indicated otherwise. Before conjugation, 300 μg of MP was washed with deionized water and then MES buffer with occasional sonication to prevent MP aggregation. After that, the particles were incubated for 20 min in activation buffer (80 mg of EDC and 40 mg of sulfo-NHS in 0.8 mL of 100 mM MES buffer, pH 5.0) followed by 1-h conjugation in a solution of antibody in 80 μL of conjugation buffer (0.1 M borate buffer, pH-8.6). The antibodies (Ab) used during the latter step determined the MP-Ab conjugate type and were as follows: (i) 10 μL of anti-BoNT-A clone KBA468 (3.5 mg/mL), (ii) 10 μL of anti-BoNT-B clone BTB224 (3.0 mg/mL), and (iii) 25 μL of polyclonal rabbit anti BoNT-E antibodies (1.8 mg/mL). The 10420

DOI: 10.1021/acs.analchem.6b02066 Anal. Chem. 2016, 88, 10419−10426

Article

Analytical Chemistry

6.0 mg/mL anti-BoNT-B (clone KBB36), and 1.8 mg/mL polyclonal rabbit anti BoNT-E antibodies for the A-strips, Bstrips, and E-strips, respectively. The corresponding MP-Ab conjugate was deposited onto the conjugation pad at a jetting rate of 12 μL/cm. The NC membrane was air-dried at room temperature for 24 h and stored in dry place before use. A magnetic control line, similar to that of conventional lateral flow strips, can be easily added to speed up the procedure and for convenience of on-site applications.43 In this work, it has not been used because the mapping of MP distribution along test strips offers good discrimination between a truly negative result and failure of the assay.43 Multiplex Cartridges. Design and Fabrication. The cartridge for the multiplex assay represents a hollow polyethylene stick (95 mm long, 4.5 mm inner diameter) that comprises three LF test strips, namely, A-strip, B-strip, and Estrip. The strips are spatially separated by a porous spacer, which provides equal distribution of the sample. The strips were prepared as described above, and the test lines were positioned at 14 mm, 22 mm, and 30 mm from the front edge of nitrocellulose membrane for E-strip, B-strip, and A-strip, respectively, to avoid overlapping the magnetic signals. Measuring Procedure. After the assay, the test strip or multiplex cartridge is readout by inserting into the measuring coils of the portable multiplex MPQ reader. The magnetic labels are quantified from the entire volume within each measuring coil. The same measuring setup can be equally used for identification of a single analyte with the LF test strip (single-plex assay) or several analytes simultaneously with the multiplex cartridge (multiplex assay). While assay optimization, quantitative MP distribution along the test strip/cartridge was recorded. The sample volumes were 60 and 180 μL in singleplex and multiplex assays, respectively. The assay time was 25 min. Each test was repeated at least 3 times under the same conditions. The result was calculated as average and error was estimated as standard deviation of recorded values. The limit of detection was determined as the analyte concentration that produced an average signal greater than that of antigen-free negative control by double standard deviation. As a negative control, we used a sample of the same matrix but with no toxoid. Sample Preparation. Milk (1.8% fat) and apple and orange juices from concentrate were purchased from a local grocery store. No additional sample preparation was used for milk. Apple and orange juice were neutralized to pH 7.0 using 10% sodium hydroxide and centrifuged for 5 min at 7000 rpm. The resulting supernatant was stored at +4 °C. Toxoids were spiked into milk and juice samples before the sample preparation to get final concentrations 0.1, 0.33, 1, 33, 10, 33, and 100 ng/mL.

unreacted intermediate active groups were blocked for 1 h by 10 μL of 10% BSA in PBS (pH 7.4). Finally, the MP-Ab conjugates were twice washed with PBS (pH 7.4). Magnetic Particle Quantification Method. Magnetic particle quantification (MPQ) method, described in details in refs 38, 41, and 43 permits highly sensitive detection of magnetic nanoparticles at relatively large depths of several millimeters inside a volume regardless of its optical transparency. Briefly, MP subjected to an alternating magnetic field of two frequencies are registered by their nonlinear response at combinatorial frequencies. Linear dia- and paramagnetic materials do not contribute to the signal, and the method sensitivity to magneto-radioactive 59Fe-based MP is on the level of the γ-radioactive technique.44 The MPQ method was successfully used for noninvasive in vivo MP mapping in organs of small animals, as well as for long-term study of MP evolution, clearance, and redistribution in the animal’s body.44,45 For multiplex measurements in this research, our MPQ readers, which offered the record sensitivity from 60 zmol (36 000 pieces) of MP in up to 0.2 mL reading zone volume and 7-order linear dynamic range,43,46 were essentially improved by incorporation of three measuring inductive coils interrogated by a single processor unit. The advanced electronics allows readout of these coils separately and/or successively at adjustable time intervals. The multiplex readers retain the above-mentioned MP detection parameters that now have no analogues. Assay Principle. The proposed assays are based on the sandwich lateral flow assay principle47 with MP labels. A sample is deposited onto the sample pad (Figure 1). While migration

Figure 1. Test strip design based on sandwich-lateral flow assay with antibody-conjugated magnetic particles as labels.

along the test strip under capillary action, the sample interacts with dry MP-Ab conjugate deposited at the conjugation pad. Then, if the target antigen is contained in the sample, it binds to MP-Ab and to the capture Ab on the test line (TL). Test Strips. Design and Fabrication. Three types of the LF test strips were prepared: A-strips, B-strips, and E-strips intended for detection of BoNT-A, B, or E, respectively. These strips differed in the used MP-Ab conjugates and corresponding capture antibody. Each test strip (Figure 1) is 3 mm wide and is composed of overlapping sample pad, conjugation pad, nitrocellulose, and absorbent membranes assembled on an adhesive plastic backing sheet. The capture Ab deposited onto the 4 cm-long NC membrane at a jetting rate of 1 μL/cm with Biodot XYZ3060 Dispense Platform to form a test line at 22 mm from the NC front edge were 4.5 mg/mL anti-BoNT-A (clone BTA-232),



RESULTS AND DISCUSSION

The development of the multiplex LF assay for simultaneous quantitative detection of BoNT-A, BoNT-B, and BoNT-E included the following major steps: (i) development and optimization of three single-plex assays, each being specific for only one of the mentioned serotypes; (ii) multiplexing the obtained single-plex assays via joining the A-strip, B-strip, and E-strip into a miniature cartridge; (iii) comparison of characteristics of multiplex and single-plex assays; (iv) validation of the multiplex assay by analysis of food samples. 10421

DOI: 10.1021/acs.analchem.6b02066 Anal. Chem. 2016, 88, 10419−10426

Article

Analytical Chemistry

in buffer (Figure 3). The Y-axis exhibits “magnetic signal” calculated as the difference between the signals at the test line and 3 mm after it, the latter being considered as the background signal. The achieved limits of detection (LOD) were 0.20 ng/mL (1.33 pM or 12 pg or 80 amol per reaction), 0.12 ng/mL (0.80 pM or 7.2 pg or 48 amol per reaction), and 0.30 ng/mL (2 pM or 18 pg or 120 amol per reaction) for BoNT-A, BoNT-B, and BoNT-E, respectively, with the dynamic range exceeding 3 orders. Remarkably, these characteristics are highly competitive with the most up-to-date quantitative techniques having similar duration but executed in laboratory conditions only by qualified personnel.4,10 Multiplexing the Assays. The proposed novel approach to multiplexing is primarily based on the original multichannel magnetic particle quantification technique at relatively large depths of several millimeters inside a 3D volume regardless of its optical transparency properties. To realize this approach, three single-plex test strips (A-strip, B-strip, and E-strip) are combined in a miniature cartridge (Figure 4). The assay procedure and conditions as well as the measuring setup are the same as in the single-plex assays, with the cartridge used instead of the single-plex strip. A sample is deposited onto the front end of the cartridge and after ∼25 min is inserted into the advanced three-channel MPQ reader for measurements. The reader records MP at any point within the measuring coils, somewhat resembling “X-ray examination” for the cartridge. The results, therefore, do not depend on the strip position inside the cartridge, and the cartridge optical transparency is indifferent. Further, we have experimentally confirmed that no adaptation procedure is required for our method of multiplexing except provision of (i) spatial separation of the singleplex test strips inside the cartridge to avoid cross-reactivity, (ii) sample supply to the strips either in equal volumes or volumes sufficient to saturate the strips, and (iii) dissimilar positions of the test lines on the participating test strips to distinguish the magnetic signals from different strips. The first two requirements were easily met by proper positioning and fixation of the test strips inside the cartridge. The TL position on each of the strips was determined as follows. In the single-plex assays, we found that the signal slightly decreased with distance from the strip front end probably because not all MP reached TL within the limited assay time to contribute the signal. Therefore, the TL on the strip, which exhibited the lowest single-plex signal (E-strip), was shifted closer to the strip front end (at 14 mm) while the TL on the strip, which showed the highest single-plex signal (Astrip), was moved further from the strip front end at 30 mm. The B-strip TL with its in-between single-plex response kept the 22-mm position. According to the single-plex calibration curves measured with the new TL positions at the same optimal conditions, the effect of the TL relocation on the calibration curves and LOD was insignificant (Figure S1 in the Supporting Information). Notably, the used 8 mm separation of the TLs is not the MPQ method limitation as the reader can be easily adapted for multiplex measurements with less separated TLs. Multiplex Assay Characterization. Cross-reactivity is the common problem of multiplex assays that not only substantially deteriorates but may even compromise the assay performance.37 One of the advantages of the proposed multiplexing is that the cartridge design completely eliminates the need of the

Development of Single-Plex Assays. The single-plex LF assays were developed in this work as building blocks for the multiplex biosensing. These assays can be used “as is” for autonomous quantitative detection of the respective BoNT serotype. First, we optimized the assay conditions with quantitative mapping of magnetic nanolabels along the LF strip, which was volumetrically recorded with the MPQ-reader. For each BoNT serotype, we analyzed the following analyte concentrations: (0, 0.1, 0.33, 1, 3.3, 10, 33, and 100) ng/mL (see Figure 2a for

Figure 2. Quantitative single-plex detection of BoNT-A in buffer. (a) Typical distributions of magnetic nanolabels along the lateral flow test strip for different concentrations of BoNT-A. (b) SEM image of MP specifically captured at the test line on the membrane (FEI Quanta 200, accelerating voltage −1.00 kV, 10 000× magnification, and 3.7 mm working distance).

BoNT-A). The MP distributions exhibit three peaks corresponding to the sites of major MP localization and can be used to evaluate the MP portion of the whole MP amount at any length from the strip front end. MP specifically captured at the test line area of the NC membrane can be seen in the scanning electron microscope image (Figure 2b). The assay parameters were optimized toward the maximal MP portion at the test line and, concurrently, minimal MP portion remained at the conjugation pad. The optimized conditions are shown in the Experimental Section and include the protocols for MP-Ab conjugation, for deposition of the MP-Ab conjugates and antibodies onto the corresponding membranes, conditions of drying and storage of the prepared test strips. These optimal conditions were used to obtain calibration curves of the single-plex detection of the three BoNT serotypes 10422

DOI: 10.1021/acs.analchem.6b02066 Anal. Chem. 2016, 88, 10419−10426

Article

Analytical Chemistry

Figure 3. Calibration curves in log−log scale of quantitative single-plex detection of the studied BoNT serotypes in buffer. (a) BoNT-A, (b) BoNTB, and (c) BoNT-E. Dotted lines show the negative control signal plus double standard deviation.

0.36 ng/mL for BoNT-E, 0.12 ng/mL and 0.13 ng/mL for BoNT-B, and 0.20 ng/mL and 0.21 ng/mL for BoNT-A in the absence and presence of both other toxins, respectively. Such insignificant differences in LOD confirm the reliable specificity of the proposed multiplex quantitative biosensing. The pairwise comparison of top and middle rows in Figure 5 shows an increase in the total background signal in multiplex vs single-plex assay due to combining the signals from all the strips. Besides, as mentioned above, the slight (less than 15%) amplification and reduction of the signals compared with those obtained in single-plex assay were observed respective to the TL relocation toward the front end (E-strips) or back end (Astrips) of the strip. These variations produced minor differences in the calibration curves of the multiplex and single-plex assays (Figures S1 and S2 in the Supporting Information) with virtually no effect on the assay analytical characteristics. Thus, in contrast to vast majority of traditional approaches, the proposed concept of multiplexing offers high-specificity and practically the same LOD and dynamic range as those of respective single-plex assays. Test of the Multiplex Assay Robustness in Food Samples. Food is the most probable source of botulism, so the performance of the developed multiplex assay was verified by simultaneous detection of the three botulinum toxin serotypes in complex liquid food matrixes such as whole milk and juices commonly applied for verification of BoNT tests.13,49 All three BoNT serotypes were detected in milk with virtually the same LOD as in buffer (Table 1). In these experiments, no sample preparation was used. Apple and orange juices were more challenging due to their nonsoluble pulp suspension. Migrating with the sample flow, the pulp quickly clogged the NC membrane pores, thus blocking further MP flow and distorting the results. Our assay easily recognized this problem by an abnormally high first peak in the MP distribution along the strip caused by nonspecific MP deposition.43 Besides, as expected,22 the acid low pH values (3−5) inherent to pure juices affected both the colloidal stability of MP and the specificity of binding of the immobilized antibodies with antigen, producing a nonspecific signal even in the absence of toxins. The same effects, typical for any lateral flow analysis, were observed during single-plex detection. In our case, both issues were mostly resolved by juice clarification by 5 min centrifuging at 7000 rpm and neutralizing their pH. These procedures are not specific to our assay; they are standard for lateral flow analyses of these food products. If required, the process can be simplified (e.g., by filtration with a syringe

Figure 4. Multiplex assay setup: several single-plex test strips with dissimilar positions of the test lines are combined in a miniature cartridge (top); the cartridge with a sample deposited onto its front end is inserted into the portable MPQ reader (middle); and simultaneous readout of magnetic signals from all participating test strips (bottom).

time-consuming studies of cross-reactivity between the components of participating strips. The specificity of detection with the multiplex cartridge was tested for every toxin in two setups: in the absence and presence of both other toxins in concentration of 100 ng/mL (Figure 5). The pairwise comparison of the graphs in the middle and lower rows of the figure shows that the recorded specific signal is virtually independent of the presence of nontarget toxins. This indicates the absence of cross-reactivity between the antibodies on the strips inside the cartridge and nonspecific antigens in the sample as well as the nonexistent interference of the signals from the neighboring TLs. To check this consideration, we plotted the calibration curves for each toxin in both setups. The curves proved to be nearly identical (Figure S2 in the Supporting Information). The LOD values determined by these curves were as follows: 0.35 ng/mL and 10423

DOI: 10.1021/acs.analchem.6b02066 Anal. Chem. 2016, 88, 10419−10426

Article

Analytical Chemistry

Figure 5. Comparison of MP distributions along the test strips while multiplex and single-plex detection of BoNT-E, BoNT-B, and BoNT-A. Top row, single-plex detection of 100 ng/mL (solid line) and 0 ng/mL (dotted line) for each toxin (TL at central position of 22 mm from the strip front end, the samples contain target toxin only). Middle row, detection with the multiplex cartridge of 100 ng/mL (solid line) and 0 ng/mL (dotted line) for each toxin (TL positions from the front end, 14 mm, 22 mm, 30 mm for E-strips, B-strips, and A-strips, respectively; samples contain target toxin only). Lower row, detection with the multiplex cartridge of 100 ng/mL (solid line) and 0 ng/mL (dotted line) for each toxin (TL positions from the front end, 14 mm, 22 mm, 30 mm for E-strips, B-strips, and A-strips, respectively; samples contain all three toxins at a concentration of 100 ng/mL).

Table 1. Results of BoNT Detection with the Proposed Multiplex Assay in Food Matrixes PBS buffer

BoNT-A BoNT-B BoNT-E

milk

apple juice

orange juice

LOD, pg/mL

LOD, pg/mL

specific signal compared to buffer, %

LOD, pg/mL

specific signal compared to buffer, %

LOD, pg/mL

specific signal compared to buffer, %

185 140 350

197 143 254

97 ± 4 97 ± 2 127 ± 10

307 142 465

73 ± 6 99 ± 6 81 ± 1

287 139 410

75 ± 2 100 ± 5 91 ± 7

Since milk is also very complex matrix for lateral flow assays, these data suggest that the observed deviations from the experiments in buffer are mainly due to the chosen detecting bioreagents rather than to the properties or limitations of the developed assay. The LODs achieved with the developed assay of the toxoids in PBS buffer and milk were 11.1−11.8 pg/ reaction and 8.4−8.6 pg/reaction for BoNT-A and BoNT-B, respectively. That is on the same level as LODs obtained for toxoids with the immuno-quantitative PCR48 that offers 4.5−9

instead of centrifuging), and pH neutralization can be omitted with proper corrections of calibration curves. The effect of juice matrix on signal and LOD was minor (did not exceed 1 standard deviation) for BoNT-E. As can be seen from Table 1, both characteristics, though, were deteriorated (up to 25%) while BoNT-A detection in juices. In contrast, all the studied matrixes only slightly affected the calibration curve and LOD for BoNT-B. Moreover, LOD of the multiplex detection of BoNT-E in milk was even better than in buffer. 10424

DOI: 10.1021/acs.analchem.6b02066 Anal. Chem. 2016, 88, 10419−10426

Analytical Chemistry



pg/reaction for BoNT-A and 18.5−37 pg/reaction for BoNT-B in PBS + 1% BSA and semifat milk. The proposed method, though, features a wider dynamic range and higher usability for on-site detection. Considering frequent reports of LOD enhancement by 1−2 orders while transition from toxoids to native toxins in immunoanalytical methods,48,50 the proposed platform is promising for highly sensitive on-site detection of BoNTs.

CONCLUSION The present research delivers an effective tool for constructing multiplex quantitative LF assays, which eliminates the timeconsuming yet vital steps of cross-reactivity investigation. The developed multiplex assay can analyze media regardless of their optical properties, offers sensitivity on the level of the laboratory quantitative methods, but can be implemented onsite as well. The demonstrated triplex performance is not a method limitation due to the applied registration technique of magnetic particle quantification. The assay can be easily extended to detection of large number of other bioagents. The proposed approach is promising for rapid module design of the multiplex LF cartridges for an on-demand set of analytes for easy-to-use point-of-care in vitro diagnostics, for food analysis, environmental monitoring, national security, etc. ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.6b02066.



REFERENCES

(1) Arnon, S. S.; et al. JAMA 2001, 285 (8), 1059. (2) Benedetto, A. V. Botulinum Toxin in Clinical Dermatology; Taylor & Francis: Oxfordshire, U.K., 2006. (3) Aoki, K. R. Eur. J. Neurol. 1999, 6 (S4), s3−s10. (4) Saravanan, P.; Rajaseger, G.; Eric, Y. P.; Moochhala, S. In Toxinology; Gopalakrishnakone, P., Ed.; Springer: Dordrecht, The Netherlands, 2014; pp 1−13. (5) Bigalke, H.; Rummel, A. Toxicology 2005, 214 (3), 210−220. (6) Arnon, S. S.; Schechter, R.; Inglesby, T. V.; Henderson, D. A.; Bartlett, J. G.; Ascher, M. S.; Lillibridge, S. JAMA 2001, 285 (8), 1059−1070. (7) Sobel, J. Clin. Infect. Dis. 2005, 41 (8), 1167−1173. (8) Hill, K. K.; Smith, T. J. Genetic diversity within Clostridium botulinum serotypes, botulinum neurotoxin gene clusters and toxin subtypes. In Botulinum Neurotoxins; Springer: Berlin, Heidelberg, Germany, 2012; pp 1−20. (9) Woudstra, C.; Skarin, H.; Anniballi, F.; Fenicia, L.; Bano, L.; Drigo, I.; Fach, P. Appl. Environ. Microbiol. 2012, 78, 3120. (10) Singh, A. K.; Stanker, L. H.; Sharma, S. K. Crit. Rev. Microbiol. 2013, 39, 43. (11) Scotcher, M. C.; Cheng, L. W.; Stanker, L. H. PLoS One 2010, 5 (6), e11047. (12) Dressler, D.; Gessler, F.; Tacik, P.; Bigalke, H. Mov. Disord. 2014, 29 (10), 1322−1324. (13) Sharma, S. K.; Ferreira, J. L.; Eblen, B. S.; Whiting, R. C. Appl. Environ. Microbiol. 2006, 72 (2), 1231−1238. (14) Lee, J.; Brennan, M. B.; Wilton, R.; Rowland, C. E.; Rozhkova, E. A.; Forrester, S.; Hannah, D. C.; Carlson, J.; Shevchenko, E. V.; Schabacker, D. S.; Schaller, R. D. Nano Lett. 2015, 15 (10), 7161− 7167. (15) Frisk, M. L.; Tepp, W. H.; Johnson, E. A.; Beebe, D. J. Anal. Chem. 2009, 81 (7), 2760−2767. (16) Rivera, V. R.; Gamez, F. J.; Keener, W. K.; White, J. A.; Poli, M. A. Anal. Biochem. 2006, 353 (2), 248−256. (17) Barr, J. R.; Moura, H.; Boyer, A. E.; Woolfitt, A. R.; Kalb, S. R.; Pavlopoulos, A.; McWilliams, L. G.; Schmidt, J. G.; Martinez, R. A.; Ashley, D. L. Emerging Infect. Dis. 2005, 11 (10), 1578−1583. (18) Chiao, D.-J.; Shyu, R.-H.; Hu, C.-S.; Chiang, H.-Y.; Tang, S.-S. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2004, 809 (1), 37−41. (19) Ling, M. M.; Ricks, C.; Lea, P. Expert Rev. Mol. Diagn. 2007, 7 (1), 87−98. (20) Sharma, S. K.; Eblen, B. S.; Bull, R. L.; Burr, D. H.; Whiting, R. C. Appl. Environ. Microbiol. 2005, 71 (7), 3935−3941. (21) Ching, K. H.; Lin, A.; McGarvey, J. A.; Stanker, L. H.; Hnasko, R. J. Immunol. Methods 2012, 380 (1−2), 23−29. (22) Attrée, O.; Guglielmo-Viret, V.; Gros, V.; Thullier, P. J. Immunol. Methods 2007, 325 (1−2), 78−87. (23) Bruno, J. Pathogens 2014, 3 (2), 341−355. (24) Osterfeld, S. J.; Yu, H.; Gaster, R. S.; Caramuta, S.; Xu, L.; Han, S.-J.; Hall, D. A.; Wilson, R. J.; Sun, S.; White, R. L. Proc. Natl. Acad. Sci. U. S. A. 2008, 105 (52), 20637−20640. (25) Orlov, A. V.; Khodakova, J. A.; Nikitin, M. P.; Shepelyakovskaya, A. O.; Brovko, F. A.; Laman, A. G.; Grishin, E. V.; Nikitin, P. I. Anal. Chem. 2013, 85 (2), 1154−1163. (26) Orlov, A. V.; Nikitin, M. P.; Bragina, V. A.; Znoyko, S. L.; Zaikina, M. N.; Ksenevich, T. I.; Gorshkov, B. G.; Nikitin, P. I. J. Magn. Magn. Mater. 2015, 380, 231−235. (27) LaBorde, R. T.; O’Farrell, B. IVD Technol. 2002, 8, 36−41. (28) Workman, S.; Wells, S. K.; Pau, C.-P.; Owen, S. M.; Dong, X. F.; LaBorde, R.; Granade, T. C. J. Virol. Methods 2009, 160 (1−2), 14−21. (29) Nikitin, P. I.; Vetoshko, P. M.; Ksenevich, T. I. Sens. Lett. 2007, 5 (1), 296−299. (30) Taton, K.; Johnson, D.; Guire, P.; Lange, E.; Tondra, M. J. Magn. Magn. Mater. 2009, 321 (10), 1679−1682. (31) Marquina, C.; De Teresa, J. M.; Serrate, D.; Marzo, J.; Cardoso, F. A.; Saurel, D.; Cardoso, S.; Freitas, P. P.; Ibarra, M. R. J. Magn. Magn. Mater. 2012, 324 (21), 3495−3498.





Article

Detailed calibration curves at different conditions (Figures S1 and S2) (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions

A.V.O. and S.L.Z. contributed equally to this work. P.I.N. designed and directed the experiments. A.V.O., S.L.Z., V.R.C., and M.P.N. conducted the experiments. All authors discussed the results and cowrote the manuscript. Notes

The authors declare the following competing financial interest(s): A patent application on the multiplex assay has been filed by P.I.N.; P.I.N. is also an inventor on patents on MPQ.



ACKNOWLEDGMENTS The authors thank Dr. I. D. Vinogradova (Gamaleya Federal Research Center for Epidemiology and Microbiology, Moscow, Russia) for providing polyclonal rabbit anti BoNT-E antibodies and Sartorius AG (Germany) for the UniSart CN140 membrane. Different aspects and parts of this multidisciplinary research were partially supported by the Russian Science Foundation Grant No. 16-12-10543 (development of ultrasensitive multichannel detection of MP) and by the Russian Foundation for Basic Research Grant No. 16-33-60228 mol_a_dk (optimization and testing of the immunoassays robustness in food samples). 10425

DOI: 10.1021/acs.analchem.6b02066 Anal. Chem. 2016, 88, 10419−10426

Article

Analytical Chemistry (32) Rivas, L.; de la Escosura-Muñiz, A.; Serrano, L.; Altet, L.; Francino, O.; Sánchez, A.; Merkoçi, A. Nano Res. 2015, 8 (11), 3704− 3714. (33) Gantelius, J.; Hamsten, C.; Neiman, M.; Schwenk, J. M.; Persson, A.; Andersson-Svahn, H. J. Microbiol. Methods 2010, 82 (1), 11−18. (34) Fenton, E. M.; Mascarenas, M. R.; López, G. P.; Sibbett, S. S. ACS Appl. Mater. Interfaces 2009, 1 (1), 124−129. (35) Sajid, M.; Kawde, A.-N.; Daud, M. J. Saudi Chem. Soc. 2015, 19 (6), 689−705. (36) Ellington, A. A.; Kullo, I. J.; Bailey, K. R.; Klee, G. G. Clin. Chem. 2010, 56 (2), 186−193. (37) Juncker, D.; Bergeron, S.; Laforte, V.; Li, H. Curr. Opin. Chem. Biol. 2014, 18, 29−37. (38) Nikitin, P. I.; Vetoshko, P. M. Meter of magnetic susceptibility. Russian Patent RU 2177611, December 27, 2001. (39) Nikitin, P. I.; Vetoshko, P. M. Process of analysis of mixture of biological and/or chemical components with use of magnetic particles and device for its implementation. Russian Patent RU 2166751, May 10, 2001. (40) Nikitin, P. I.; Vetoshko, P. M.; Ksenevich, T. I. J. Magn. Magn. Mater. 2007, 311 (1), 445−449. (41) Nikitin, M. P.; Shipunova, V. O.; Deyev, S. M.; Nikitin, P. I. Nat. Nanotechnol. 2014, 9 (9), 716−722. (42) Abbasova, S. G.; Rudenko, N. V.; Gorokhovatskii, A. Y.; Kapralova, M. V.; Vinogradova, I. D.; Vertiev, Y. V.; Nesmeyanov, V. A.; Grishin, E. V. Russ. J. Bioorg. Chem. 2011, 37 (3), 307−315. (43) Orlov, A. V.; Bragina, V. A.; Nikitin, M. P.; Nikitin, P. I. Biosens. Bioelectron. 2016, 79, 423−429. (44) Nikitin, M. P.; Torno, M.; Chen, H.; Rosengart, A.; Nikitin, P. I. J. Appl. Phys. 2008, 103 (7), 07A304. (45) Nikitin, M. P.; Vetoshko, P. M.; Brusentsov, N. A.; Nikitin, P. I. J. Magn. Magn. Mater. 2009, 321 (10), 1658−1661. (46) Shipunova, V. O.; Nikitin, M. P.; Nikitin, P. I.; Deyev, S. M. Nanoscale 2016, 8, 12764−12772. (47) Posthuma-Trumpie, G. A.; Korf, J.; van Amerongen, A. Anal. Bioanal. Chem. 2009, 393 (2), 569−582. (48) Rajkovic, A.; El Moualij, B.; Fikri, Y.; Dierick, K.; Zorzi, W.; Heinen, E.; Uner, A.; Uyttendaele, M. Food Analytical Methods 2012, 5 (3), 319−326. (49) Wein, L. M.; Liu, Y. Proc. Natl. Acad. Sci. U. S. A. 2005, 102 (28), 9984−9989. (50) Keller, J. E. Clinical and Vaccine Immunology. 2008, 15 (9), 1374−1379.

10426

DOI: 10.1021/acs.analchem.6b02066 Anal. Chem. 2016, 88, 10419−10426