Biosensor for the Detection of Cyanobacterial Toxin Microcystin-LR

Abstract

Cyanobacteria are found everywhere in the environment, and their growth accelerates significantly with rising amounts of sunlight and temperatures. The proliferation of cyanobacteria begins when the average temperatures rise above 15°C. The proliferation can lead to high amounts of secondary metabolites, such as cyanotoxins, in surrounding waters. The most common cyanotoxin is microcystin-LR (MC-LR). MC-LR can cause rashes, abdominal cramps, and liver damage in humans and animals, so continuous monitoring of its content in water is of great importance. MC-LR is commonly detected with high-performance liquid chromatography, but phosphatase inhibition-based bioassays and enzyme-linked immunosorbent tests are also available. However, these are all lab-based methods and require sample transport and preparation for analytical procedures, not allowing for obtaining quick results. Therefore, there is a need for a rapid and field-based analysis method, and one promising option is to use biosensors. The present study aimed to design and construct an aptamer/antibody-based biosensor to detect MC-LR and test its applicability to detect MC-LR in cyanobacteria culture (Microcystis aeruginosa).

1. Introduction

Cyanotoxins are metabolites produced by cyanobacteria, a group of photosynthetic prokaryotes found in freshwater. The intake of contaminated water, skin contact, or swallowing water during swimming are among the most common reasons for poisoning caused by cyanotoxins [1]. An increase in temperature causes the cyanobacteria to grow faster. In addition, the spread of cyanobacteria is also affected by the pH of the environment, salinity, the presence of necessary nutrients (e.g., nitrogen and phosphorus), and light. It has been observed that global warming may increase the frequency and extent of cyanobacterial proliferation [2, 3]. Cyanobacteria can release toxins into the environment during the mass spread of microorganisms, that is, the water blooming.

1.1 Cyanotoxins

Microcystins are the most widespread cyanobacterial toxins produced by Microcystis aeruginosa in freshwater lakes and rivers worldwide [4]. Microcystins are hepatotoxins that substantially affect serine/threonine protein phosphatases (PPs), which can remove phosphate from the protein in many biochemical pathways [5]. They are cyclic heptapeptides with a molecular weight of 800–1100 Da, and more than 250 different microcystins have been described [6]. The general structure of microcystins is cyclo-D-Ala¹-X²-D-MeAsp³-Z⁴-Adda⁵-D-Glu⁶-Mdha⁷ (superscript number indicates the position number, Figure 1), where X and Z are variable L-amino acids, D-MeAsp is D-erythro-β-methylaspartic acid, Adda is 3-amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid, and Mdha is N-methyldehydroalanine [5, 7]. Structure variations occur in all seven amino acid residues; most common are the replacement of L-amino acids in positions 2 and 4, replacement of Mdha by dehydrobutyrine (Dhb) or by serine in position 7, and a lack of methylation of amino acids in positions 3 and/or 7 [7]. The variations in Adda are essential because they may affect analytical test results, which use Adda as a marker, and in addition, Adda moiety is critical to microcystin activity [5]. The hydrophobicity of the amino acids at positions two and four influences the overall hydrophobicity. Hydrophobicity of the microcystin congener determines how the toxin interacts with cell membranes, and therefore, affects its specific toxicity [8].

Figure 1.

Generic structure of microcystin-LR: Superscript numbers indicate the position numbers, and X and Z are the variable L-amino acids in different microcystins. The two specific L-amino acids of MC-LR are shown in black (leucine, L) and blue (arginine, R). Abbreviations: Ala is alanine; Leu is leucine; MeAsp is erythro-β-methylaspartic acid; Arg is arginine; Adda is 3-amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid; Glu is glutamic acid, and Mdha is N-methyldehydroalanine.

The most common and most toxic is microcystin-LR (MC-LR), with leucine (Leu = L) in the second position and arginine (Arg = R) in the fourth position (Figure 1). Modeling of the MC-LR molecule has shown that the alanine and leucine residues in positions 1 and 2 extend beyond the ring plane; thus, allow selective binding to receptor molecules, which causes high toxicity of MC-LR and another metabolite, a cyclic non-ribosomal pentapeptide nodularin [7]. Also, three-dimensional structure studies of the MC-LR have shown that Adda and Arg side chains protrude from the ring distal from one another caused by the repulsion between the guanidino function of Arg and the hydrophobic Adda [9]. Microcystin-LR inhibits protein phosphatase type 1 and type 2A (PP1 and PP2A) activities in the cytoplasm of liver cells [10, 11], which leads to an increase in the phosphorylation of proteins in liver cells. The Adda side-chain is accommodated to the hydrophobic channel [7]. The carboxylic D-Glu site makes hydrogen bonds to metal-bound water molecules [7]. The carboxyl group of the MeAsp site makes hydrogen bonds to conserved arginine and tyrosine residues in the PPP enzyme [7]. Finally, the methylene group at the Mdha site binds to an S-atom of a cysteine residue, and the leucine residue folds closely to another well-maintained tyrosine residue [12].

The degradation of microcystins is slow in most water environments [13]. Most mycotoxins are heat-resistant [14], and the water treatment process cannot altogether remove them. Still, they can be degraded, when using UV treatment close to their absorption peak (UV lambda max for microcystin-LR is 238 nm) [4, 15]. Due to carboxyl, amino, and acylamino groups in the structure, mycotoxins have different ionization propensities at different pH values.

Limits for MC-LR in natural waters have been set in only a few countries. In Hungary and some US states, such as Indiana and New York, the MC-LR limit in water is 4 μg/l [16]. The World Health Organization (WHO) has set a limit of 1 μg/l for MC-LR in drinking water [6, 17].

1.2 Methods to detect cyanotoxins

For the detection of cyanotoxins in water, the following methodologies are used: high-performance liquid chromatography (HPLC) combined with mass spectrometry (MS, MS/MS) or ultraviolet/photodiode array detectors (UV/PDA), enzyme-linked immunosorbent assays (ELISA), and protein phosphatase inhibition assay (PPIA).

HPLC is a selective and sensitive method that allows the simultaneous determination of different microcystins at very low concentrations (0.02 μg/l). Still, the determination is technically complex and time-consuming, and the cost of the apparatus and analysis is high [18]. In addition, pretreatment of samples is required [19]. Chromatographic methods do not allow on-site monitoring, and given the need to transport samples, results can be obtained in a minimum of 4–6 hours [19]. It is also important to consider matrix effects in chromatographic analysis, and prior calibration with the matrix is required [20].

ELISA and PPIA are the other technologies often used to detect microcystins. ELISA assays are based on antigen-antibody interactions, and analytes are detected by the color change resulting from the reaction. The ELISA assay is highly sensitive and relatively straightforward [19]. For commercial ELISA rapid tests, the limit of determination for microcystins is 0.06 μg/l, and the test time varies between 4 and 6 h [21]. A significant disadvantage of many commercial ELISAs is that they are based on anti-Adda antibodies and do not measure the specific microcystin, but the total microcystin and nodularin content [22] and there is cross-reactivity [20]. ELISA tests based on a monoclonal antibody against arginine at position 4 limit detection of as low as 0.002–0.006 μg/l [18, 23].

PPIA allows to perform assays efficiently and quickly (approx. 2 h), and is based on a protein phosphatase-catalyzed protein dephosphorylation reaction in which the presence of a chromogenic substrate (e.g., p-nitrophenol phosphate) releases p-nitrophenol, which is detected at 410 nm [24]. The enzymes used, such as protein phosphatase 1 (PP1), are readily available, and this method has a medium sensitivity of 0.1 μg/l for MC-LR [25]. The main disadvantage of PPIA is the low selectivity because cyanobacteria contain phosphatases, and it is impossible to identify different microcystins [26]. The interaction of microcystins with PP1 is thought to be related to non-coding amino acid residues: Adda, D-Glu, and Mdha at positions 5, 6, and 7 of the microcystin molecules, respectively [27]. The essential analytical parameters characterizing the above-described microcystin determination methods are summarized in Table 1.

Methods	LOD (μg/l)	Analysis time (h)	Pretreatment of the sample	On-site possibility	Approximate cost	Ref.
Chromatography methods (HPLC-UV ja LC-MS/MS)	0.02	4–6	+	—	100–200€	[18, 20]
ELISA	0.002	4	+/−	+		[18]
Commercial ELISA tests	0.06	5–6	+/−	+	50€	[21]
PPIA	0.1	2	+	—	20–50€	[25]

Table 1.

Overview of MC-LR determination methods.

LOD: limit of detection; HPLC-UV: high-performance liquid chromatography with UV detector; LC-MS/MS: liquid chromatography combined with mass spectrometry; ELISA: enzyme-linked immunosorbent assay; and PPIA: protein phosphatase inhibitors.

Each technique has some limitations in sensitivity, reliability, detection limit, or speed and cost. The selection of a suitable method is based on the information they provide and the technical expertise needed. The cost of analytical equipment, long-lasting measurements, and the need for qualified personnel to perform the analysis are a challenge for routine monitoring. Nowadays, methods suitable for the end-user that can be validated and accepted worldwide continue to be an objective for regulators and the industry. The variety of commercially available assays or testing kits for marine toxin analysis remains limited. The list of currently available point-on-site marine toxin end-product testing technologies is provided in ref. [28].

1.3 MC-LR biosensors

Cyanotoxins are not monitored regularly in most countries due to technical complications in the detection and quantification. Biosensors for freshwater monitoring and safety applications are prospective alternatives to traditional methods. Biosensors are analytical devices that include a bio-recognition element linked to a transducer that transforms the chemical information produced into a readable signal followed by a detector. Most biosensors used to detect MC-LR are immunosensors that use antibodies or aptamers to recognize the analyte. The detection limit of biosensors ranges from 0.00003 to 0.37 μg/l, and the assay time is from 0.8 to 2.3 hours [29, 30, 31, 32, 33, 34]. The most important parameters characterizing the biosensors used to determine microcystins are summarized in Table 2. High selectivity of bio-recognition is assured by using specific antibodies or aptamers.

Methods	LOD (μg/l)	Analysis time (h)	Pretreatment of the sample	On-site possibility	Ref.
Aptasensor	0.37	0.7	Yes	+	[29]
Aptasensor (fluorescence)	0.002	~1	Not needed	+/−	[30]
Aptasensor (Raman)	0.002	2.3	Not needed	+/−	[31]
Aptasensor (square wave voltammetry)	0.002	~1	Yes	—	[32]
Sensor (differential cyclic voltammetry)	0.0003	~1	Not needed	—	[34]

Table 2.

An overview of MC-LR biosensors.

LOD: limit of detection.

1.3.1 Bio-recognition elements

Antibodies or immunoglobulins (Ig) are glycoproteins used in nature to detect and neutralize foreign objects. They have a characteristic basic structure consisting of a protein chain linked by a disulfide bridge and a very high affinity for the antigen detected, described by the dissociation constant of the antigen/antibody complex. The values of this constant are usually between 10⁻¹² and 10⁻⁸ M [35]. The MC-LR monoclonal antibody used in the present work has an affinity toward mycotoxins, which have arginine in the 4th position, with the dissociation constant of 1.4 · 10⁻¹¹ M [36].

Aptamers are synthetic single-stranded oligonucleotides capable of binding various molecules with high affinity and specificity. Aptamers are considered artificial antibodies and can adapt through intermolecular interactions [37, 38]. Nevertheless, compared with antibodies, they are more stable. When interacting with its target, the “lock key” is formed by matching the spatial conformation with the aptamer molecules [39]. Aptamers are produced using SELEX (systematic evolution of ligands by exponential enrichment), and once the aptamer sequence is developed, it can be reproduced with high precision. The characteristics of the aptamers selected for microcystin, identifying modifications at 3′ or 5′ ends to label or link the aptamers to the sensor platform, and their affinity to the target toxin are summarized in refs. [40, 41]. The MC-LR aptamer used in the present work, AN6 (5′ GGC GCC AAA CAG GAC CAC CAT GAC AAT TAC CCA TAC CAC CTC ATT ATG CCC CAT CTC CGC 3′), is a microcystin-LR specific aptamer with the affinity (K_d) of 5 x 10⁻⁸ M. It can also bind to microcystin-LA but with 3-fold reduced affinity (approx. 15.8 · 10⁻⁸ M), and no binding to microcystin-YR has been observed [40, 41]. AN6 is a synthetic 60-base DNA aptamer with a molecular weight of 18167.79 Da.

1.4 Bead injection analysis

It is promising to use measurements in analyte-containing flows for continuous monitoring of toxins for continuous monitoring of on-site analyses. One option for designing in-flow sensor systems is to use the principle of BIA (Bead Injection Analysis) [42, 43]. This microgranule insertion assay uses microgranule transport in a flowing solution to form microcolumns required for the assay. A selective component recognizing an analyte is immobilized on the surface of the microcolumn-forming granules allowing it to pre-concentrate and bind the targeted compound. After removing the sample matrix, selective detection of the bound analyte occurs, for example, using an antigen/antibody interaction. The signal of the recognition reaction is detected spectrometrically or by measuring the fluorescence signal. The scheme of BIA operation is shown in Figure 2 [44].

Figure 2.

Working principle of bead insertion analysis (BIA) . [44]: (A) injection of bio-activated granules into a flow channel to form a micro-column. (B) Sample injection (the sample binds to the activated granules). (C) Washing of the column to remove unbound sample components and matrix. (D) Labeled bio-component injection, incubation, and washing out of the unbound components. (E) Signal detection. (F) System regeneration.

The amount of activated granules required to form a microcolumn is small (approx. 20 μg), which allows the assay to be a single-use one. This technique eliminates the need to regenerate the bio-recognition system, the risk of contamination, and the risk of denaturation of the bio-component bound to the granules. It allows operation in a continuous flow system. To ensure the reliability and accuracy of the results obtained, it is also essential to ensure a consistently high quality of the bio-activation of the granules [45].

2. Experimental part

2.1 Materials

Solutions of aptamer (5’ GGC GCC AAA CAG GAC CAC CAT GAC AAT TAC CCA TAC CAC CTC ATT ATG CCC CAT CTC CGC 3′ (AN6, Integrated DNA Technologies)) and microcystin (Enzo Life Sciences, ALX-350-012-C100) were prepared in 0.01 M phosphate buffer saline (0.15 M NaCl, pH 7.2, PBS). Monoclonal MC-LR antibody (Enzo Life Sciences, MC10E7) solution was prepared in 0.5 M carbonate buffer (pH 9.5). The solutions were stored at 4°C. Epichlorohydrin was from Acros Organics (A0386058) and Coomassie brilliant blue R-250 was from Fluka AG (99%, CH-9470). Sephadex G-50 medium granules were from Pharmacia Fine Chemicals (FB-14567). All other chemicals used were at analytical grade. Buffer solutions were prepared using ultrapure MilliQ water (specific resistance 18.2 MΩ·cm).

2.2 Preparation of activated microgranules

Epichlorohydrin (Acros Organics, A0386058) was used to activate the Sephadex G-50 medium granules (Pharmacia Fine Chemicals, FB-14567). The antibody was covalently attached to the epoxy carbon of the epichlorohydrin via the amino group, using a previously published protocol with minor modifications was used [46]. First, 47 mg of granules were allowed to swell overnight at 4°C in 1 ml of water. After swelling, 400 μl of NaOH solution (concentration 0.1 M - 1 M) and 100 μl of epichlorohydrin to activate the granules were added and left on a shaker for 3 hours at room temperature. The beads were then washed twice with water and once with 0.5 M carbonate buffer (pH 9.5). The mixture was centrifuged at 2450 x g for 5 min after each washing step to separate beads. 0.5 to 2 ml of MC-LR antibody in 0.5 M carbonate buffer (pH 9.5) was added, with the antibody concentration varying from 10 to 350 μg/ml (the total amount). The mixture was incubated on a shaker for 24 h at room temperature. After incubation, the suspension was centrifuged (5 min at 2450 x g) and washed once with 0.5 M carbonate buffer (pH 9.5). To block free binding sites on the surface of the granules, ethanolamine solution (85 μl/ml in 0.5 M carbonate buffer) was added. The mixture was incubated on a shaker for 2 hours at room temperature. The suspension was centrifuged (5 min at 2450 x g), washed twice with water and several times with PBS buffer, and stored in PBS buffer at 4°C.

The yield of the attached antibody on the granules was evaluated with two different methods. First, it was visually inspected by adding 0.1% Coomassie brilliant blue R-250 (99% (Fluka AG, CH-9470)) to 30 μl of granules before adding ethanolamine. It was assessed by whether the granules turned blue, indicating the presence of bound protein on the granules. In addition, the protein content in the antibody solution was spectrophotometrically evaluated before and after the antibody attachment process. The protein content of the samples was determined at 280 nm, and the concentration was calculated using an absorption coefficient of 1.37 for IgG (ε^1%).

2.3 Carrying out measurements with biosensor

The outflow channel of the BIA system was partially sealed with a moving cap, and 20 μl of bio-activated microgranules were injected into the measuring cell at a flow rate of 1 μl/sec to form a microcolumn. 30 μl of PBS buffer was added at a flow rate of 2 μl/sec to ensure the column’s packing. A sample containing 150 μl of MC-LR was added at a flow rate of 1 μl/sec, the flow was stopped, and the system was incubated for 30 min. The measuring cell was washed with 150 μl of PBS at a flow rate of 2 μl/sec to remove the unbound toxin. Then 30 μl of MC-LR aptamer labeled with a fluorescence marker (Alexa Flour 647) was added at a flow rate of 1 μl/sec and incubated for 30 min. The concentration of the marker varied from 0.5 to 5 μg/ml. The unbound aptamer was removed from the microcolumn by adding 350 μl of PBS at a flow rate of 1 μl/sec. After each measurement, the cap was opened, and the system was washed at least four times with PBS buffer. PBS buffer with no added MC-LR was used for experimental determination of the system’s background signal (all other measurement steps were left unchanged).

All measurements were performed in triplicate. Measurements were performed at room temperature. The fluorescence intensity was measured at 670 nm (excitation wavelength 650 nm) of the Alexa Flour 647 emission peak perpendicular to the excitation light. To calculate the signal change, the signal after washing off the unbound MC-LR was subtracted from the final signal (signal after washing off unbound aptamer, recording started 5 min after completion of aptamer wash). The signal was recorded at 1-sec intervals. After stabilization, the mean signal was calculated as an average of 100 points to reduce experimental noise.

2.4 Cultivation and preparation of cyanobacteria sample

To cultivate Microcystis aeruginosa (Norwegian Culture Collection of Algae, K-0540) cells, 1 ml of culture was inoculated into 50 ml of liquid sterilized BG11+ medium, and grown under artificial light for 14 days at 16°C [47]. A LED lamp (16 h white/8 h dark, 6 W 3000 K) kept at a distance of 20 cm from the culture vessel was used as a light source. The cells were stored at -20°C.

After thawing, the samples were concentrated. Repeated centrifugation (5 min at 10000 x g) reduced the sample volume five times. An ultrasonic probe sonicator (Bandelin HD 2020 Sonopuls, horn

3 mm) was used to disrupt the cyanobacterial cells for 1 min at a cycle intensity of 7/10 and a power of 75%.

2.5 The characterization of the formation of MC-LR complexes with size-exclusion chromatography (SEC)

The formation of MC-LR complexes with antibody/aptamer was studied with an ÄKTA Purifier 10 liquid chromatography system (GE Healthcare) equipped with a UPC detector (280 nm) and a conductivity detector. The column (height 29 cm and diameter 1 cm) was loaded with Sephacryl S-200 HR (GE Healthcare, product ID: 10090795) gel pre-expanded overnight at room temperature in PBS buffer (pH 7.2) and packed under pressure to ensure the high quality of packing. For analyses, we optimized the flow rate (0.18 to 0.39 ml/min), sample volume (50 and 100 μl), and sample concentration (0.05 to 1 mg/ml). The column was calibrated with different proteins with molar weights ranging from 20 to 240 kDa. Dextran blue (2000 kDa) and potassium dichromate (294 Da) were used to determine the column void volume and total volume. All optimizations, calibrations, and measurements were performed at 8°C.

The experiments were performed at an optimum flow rate of 0.18 ml/min, a sample volume of 50 μl, and the column was flushed with 70 ml of PBS buffer (pH 7.2). The aptamer/microcystin mixture was prepared, the aptamer was incubated with MC-LR for 30 min (1:1 molar ratio). To prepare the aptamer/microcystin/antibody mixture, the antibody was incubated with microcystin for 30 minutes and re-incubated for another 30 minutes with the aptamer (1:2:43 molar ratio).

3. Results and discussion

3.1 Aptasensor design

In the MC-LR aptasensor, a sandwich system consisting of an antibody, MC-LR, and an aptamer was used. MC-LR molecule is relatively small compared to the antibody and aptamer molecules (molecular weights 0.995, 150, and 18.17 kDa). To assure effective binding, the binding sites of the antibody and the aptamer to the MC-LR molecule should be different. According to the manufacturer, the MC-LR monoclonal antibody binds to the MC-LR molecule in position two [23]. The AN6 aptamer is used to detect the bound MC-LR molecules and was selected to bind to another characteristic amino acid residue in position four [41]. In addition to the amino acid residues in positions two and four in the MC-LR molecule, the ADDA residue in position five can be used for interactions. Still, its use would reduce the selectivity of the biosensor because ADDA is present in all microcystins and nodularin. The MC-LR determination scheme is shown in Figure 3.

Figure 3.

MC-LR aptasensor.

The MC-LR Sephadex G-50 M granules were activated with MC10E7 monoclonal antibody for selective binding. The granules were chosen according to their size, to enable the collection of granules into a closed measuring channel, and the diameter of the granules must be >80 μm (Sephadex G-50 M is approximately 100–300 μm in a buffer solution) [45].

During immobilization, 44 ± 0.1% of the antibody in solution adhered to the surface of the microgranules. Visual inspection revealed that the granules turned light blue after mixing them with Coomassie brilliant blue, indicating the presence of bound antibodies on the granules. Coomassie brilliant blue stain is a widely used method for routine visualization of proteins because it makes complexes with essential amino acids, such as lysine, histidine, tyrosine, and arginine [48]. The efficiency of the immobilization process did not depend on the concentration of antibodies in the immobilization solution.

3.2 The formation of a detectable antibody/MC-LR/aptamer complex

The complex components have a significant molecular weight difference: MC-LR molecule 995 Da; the antibody and the aptamer of 150 kDa and 18.17 kDa, respectively. Considering these significant differences in molecular size, the potential formation of an antibody/MC-LR/aptamer triple complex was studied using SEC. The optimization was needed to achieve a sufficient resolution: flow rate, sample volume, and analyte concentration in a sample were modified. The best resolution over a significant range of molecular weights (100–200 kDa) was achieved at a flow rate of 0.18 ml/min, and a sample concentration of 0.5 mg/ml. Changing the sample volume did not significantly affect the resolution of the peaks, so 50 μl was chosen. These optimal conditions were used for all experiments. Dextran blue (M = 2000 kDa) was used to determine column void volume, and the total volume of the column was obtained with potassium dichromate. Individual compounds were analyzed and compared to investigate the formation of possible complexes. The chromatograms of MC-LR monoclonal antibody, aptamer AN6, and microcystin alone were compared to chromatograms of component mixtures, which were incubated before analysis in different modes:

1. Aptamer solution was incubated with microcystin for 30 min before injection;

2. Antibody was incubated with microcystin for 30 min, and then the aptamer was added and incubated again for 30 min before injecting.

As SEC separates particles according to their size, several peaks were obtained from the various spatial structures of the aptamer AN6 [49], which moved through the column significantly faster than expected. The chromatogram of the aptamer showed two clear peaks at flow volumes of 8.10 ml and 11.50 ml (± 4%). For the antibody, it was also possible to identify two characteristic peaks at flow rates of 8.40 ml and 9.99 ml (± 4%).

Due to the aptamer AN6, it is impossible to characterize the chromatograms of mixtures by molecular weights; instead, the shape, intensity, and area of the peaks were compared. There were no differences in incubating the aptamer with MC-LR (mode 1) compared to the chromatogram of aptamer alone. The chromatogram of the mixture of the antibody, MC-LR, and the aptamer prepared according to mode 2 showed that there was no peak with an elution volume of 8.1 ml, and the intensity of the peak with an elution volume of 11.5 ml was increased (Figure 4), which may indicate interactions between different components. Comparing the areas under peaks for aptamer, antibody, and antibody/MC-LR/aptamer solutions, where the amount of material injected into the column was similar, the difference was less than one unit (approximately 6%), indicating that all substances injected into the column had passed through the column, and nothing was stuck into the column.

Figure 4.

Aptamer (red line), antibody (green line), and antibody, MC-LR and aptamer mixture (blue line) chromatograms (Sephacryl S-200 HR (1/29) column). Aptamer and antibody concentrations were 0.06 mg/ml and 0.24 mg/ml, respectively. In the mixture of antibody/MC-LR/aptamer the toxins concentration was 0.07 mg/ml (aptamers and antibody concentrations 0.06 mg/ml and 0.24 mg/ml, respectively). The flow rate was 0.18 ml/min, sample size of 50 μl.

3.3 Determination of MC-LR

3.3.1 Optimization of the protocol

Measurements were performed using a protocol for detecting pathogens with a BIA-based sensor with some modifications [45]. To achieve a low limit of quantification, both antibody/MC-LR (MC-LR binding to activated microgranules) and MC-LR/aptamer (MC-LR binding to aptamer) incubation times were 30 minutes as the incubation at 15 minutes was not sufficient to obtain a reliable signal below the established WHO limit of 1 μg/l [6, 16]. The minimum volume of PBS for the efficient removal of unbound aptamer was 350 μl, as with lower PBS amounts, some of the aptamer remained in the flow channel and caused unstable signals (signal increased by hundreds of units in 5 minutes). We also optimized the aptamer concentration from 0.5 μg/ml to 5.0 μg/ml. With higher aptamer concentration, no stable end signal was achieved within 5 minutes. These results show aptamer concentration of 0.5 μg/ml was used in further experiments. The optimal protocol used for the determination of MC-LR was as follows:

• 20 μl of bio-activated gel was injected at a flow rate of 1 μl/s;

• 30 μl of PBS buffer was added at a flow rate of 2 μl/s;

• 150 μl of MC-LR sample was injected at a flow rate of 1 μl/s;

• MC-LR sample was incubated for 30 minutes to secure the attachment of MC-LR to the granules;

• The flow cell was washed with 150 μl PBS buffer at a flow rate of 2 μl/s;

• 30 μl of Alexa Flour 647-labeled aptamer AN6 was injected at a flow rate of 1 μl/s;

• The aptamer was incubated for 30 minutes to secure the attachment of the aptamer to the bound MC-LR;

• The flow cell was washed with 350 μl PBS buffer at a flow rate of 1 μl/s;

• The aptasensor signal was measured after its stabilization in 5 minutes.

3.3.2 The calibration of MC-LR biosensor

A calibration graph was plotted to characterize the sensitivity and operating range of the aptasensor (Figure 5).

Figure 5.

The dependence of aptasensors signal on MC-LR concentration.

The results showed that the signal of the aptasensor was linearly dependent on the concentration of MC-LR over a relatively wide concentration range from 1.3·10⁻⁷ to 8.0·10⁻⁴ mg/ml, and the experimental errors in this range were relatively minor from 0.8 to 4.3 AU. The coefficient of determination (R²) of the graph was 0.97. The sensitivity of the aptasensor was characterized by the slope of the graph being 0.151 ± 0.006 logAU/log(mg/ml). The background signal was measured using PBS without added MC-LR, and the background value was 4.6 ± 0.9 AU. The theoretical detection limit (LOD) of the MC-LR aptasensor was calculated as the background signal + three standard deviations of the background signal, and the limit of quantification (LOQ) as the background signal +10 standard deviations of the background signal. The LOD and LOQ values for the aptasensor were 1.7·10⁻⁸ mg/ml and 3.4·10⁻⁸ mg/ml, respectively.

Comparing the LOD of the MC-LR aptasensor with the allowed limit of MC-LR in drinking water, established by the WHO (1 μg/l = 10⁻⁶ mg/ml) [50], the LOD value of the proposed aptasensor is significantly lower, serving as a good precondition for the application of this aptasensor for the determination of MC-LR content and monitoring quality of natural water bodies.

The analysis took approximately 75 minutes, of which 60 minutes were for the analyte binding and the formation of a detectable complex. Compared to laboratory-based methods for detecting MC-LR, typically taking 4–6 hours [18, 19, 20], this method allows the determination of cyanotoxins much faster. However, compared to other aptasensors, the results can be obtained within a longer time. The time of analysis can be reduced by reducing the incubation time. It is also interesting to mention that the average material cost for one measurement was estimated to be 4.4 €.

3.3.3 Testing of the aptasensor

A cyanobacterial culture was used to test the performance of the designed aptasensor. The sample of the cultured cyanobacteria treated with ultrasound to break up the bacterial cells was diluted 50 times as it contained broken blue-green algae cells, and the solution had green color. The MC-LR concentration in the diluted culture sample was 3.0·10⁻⁷ mg/ml, indicating that the designed aptasensor was sensitive enough in the matrix, assumingly more complex than the one of natural water.

4. Conclusions

An aptasensor was designed and constructed to detect the cyanobacterial toxin MC-LR. The aptasensor was integrated with a bead injection system, where bio-activated micro-granules formed a disposable microcolumn in a partially closed flow channel. It took about 75 minutes to determine MC-LR. The aptasensor’s detection limit (LOD) was 1.7·10⁻⁸ mg/ml, and the limit of quantitation (LOQ) was 3.4·10⁻⁸ mg/ml. The LOD and LOQ values of the aptasensor were below the allowed MC-LR limit of 1 μg/l in drinking water set by WHO. The MC-LR aptasensor was used for testing the MC-LR concentration in a cyanobacterial culture. The sensitivity of the aptasensor is sufficient to determine MC-LR in samples containing algae, which creates good conditions for using the constructed aptasensor in natural water bodies for water quality monitoring.

Conflict of interest

The authors have no conflicts of interest to declare. We certify that the submission is original work and is not under review at any other publication.