Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Detection of the cyanobacterial toxin, microcystin-LR, using a novel recombinant antibody-based optical-planar waveguide platform Caroline Murphy a,n, Edwina Stack a, Svetlana Krivelo a, Daniel A. McPartlin a, Barry Byrne a, Charles Greef b, Michael J. Lochhead b, Greg Husar b, Shauna Devlin c, Christopher T. Elliott c, Richard J. O’Kennedy a a
School of Biotechnology, National Centre for Sensor Research and Biomedical Diagnostics Institute, Dublin City University, Dublin 9, Ireland MBio Diagnostics Inc., Boulder, CO, 80301, USA c Institute for Global Food Security (IGFS), School of Biological Sciences, Queen’s University, Malone Road, 18-30 Belfast, BT9 5BN, Northern Ireland, United Kingdom b
art ic l e i nf o
a b s t r a c t
Article history: Received 13 June 2014 Received in revised form 2 October 2014 Accepted 13 October 2014
Microcystins are a major group of cyanobacterial heptapeptide toxins found in freshwater and brackish environments. There is currently an urgent requirement for highly-sensitive, rapid and in-expensive detection methodologies for these toxins. A novel single chain fragment variable (scFv) fragment was generated and is the first known report of a recombinant anti-microcystin avian antibody. In a surface plasmon resonance-based immunoassay, the antibody fragment displayed cross-reactivity with seven microcystin congeners (microcystin–leucine–arginine (MC-LR) 100%, microcystin–tyrosine–arginine (MC-YR) 79.7%, microcystin–leucine–alanine (MC-LA) 74.8%, microcystin–leucine–phenylalanine (MC-LF) 67.5%, microcystin–leucine–tryptophan (MC-LW) 63.7%, microcystin–arginine–arginine (MC-RR) 60.1% and nodularin (Nod) 69.3%, % cross reactivity). Following directed molecular evolution of the parental clone the resultant affinity-enhanced antibody fragment was applied in an optimized fluorescence immunoassay on a planar waveguide detection system. This novel immuno-sensing format can detect free microcystin-LR with a functional limit of detection of 0.19 ng mL 1and a detection range of 0.21–5.9 ng mL 1. The assay is highly reproducible (displaying percentage coefficients of variance below 8% for intra-day assays and below 11% for inter-day assays), utilizes an inexpensive cartridge system with low reagent volumes and can be completed in less than twenty minutes. & 2014 Published by Elsevier B.V.
Keywords: Cyanobacteria Microcystin Recombinant Antibody Immunoassay Fluorescent planar wave-guide
Introduction Cyanobacteria are ‘blue-green’ prokaryotes found globally in fresh-water bodies such as lakes, ponds, reservoirs, slow moving rivers and brackish environments (Vareli et al., 2013; WHO, 2008). During periods of excessive nutrient enrichment, cyanobacterial ‘blooms’ flourish and, depending on the predominance of certain cyanobacteria, toxins known as cyanotoxins, are generated. Species of cyanobacteria such as Anabaena spp. and Microcystis spp. produce the most commonly found cyanotoxin, microcystin-LR (Vareli et al., 2013), a cyclic-heptapeptide (7 amino acids) with over 80 structurally related congeners (Humpage, 2008). A similar structurally related pentapeptide (5 amino acids), nodularin, is produced by Nodularia spp. Microcystin’s mode of toxicity is n Correspondence to: School of Biotechnology, Dublin City University, Glasnevin, Dublin 9, Dublin, Ireland. E-mail address: [email protected] (C. Murphy).
associated with its ability to inhibit serine/threonine protein phosphatases 1 and 2 A (MacKintosh et al., 1990) leading to hepatocytic necrosis and haemorrhaging (Bhattacharya et al., 1997). Cases of acute poisoning have resulted in severe nausea, vomiting, diarrhoea, abdominal pains, sore throats, headaches, blisters around the mouth and dry coughs (Turner et al., 1990). Chronic exposure to microcystin is a worrying phenomenon as microcystin over an extended period can cause cancer through the inhibition of protein phosphatases PP1 and PP2A (Hernandez et al., 2009). Bursts of cyanobacterial growth have significant socioeconomic impacts. In 2007, a cyanobacterial bloom of Microcystis aeruginosa contaminated the Chinese city of Wuxi's only drinking water supply, and the city's two million residents were without water for one week (Qin et al., 2010). The ubiquitous nature of microcystins illustrates their global prevalence and underpins the requirement for regular and accurate monitoring (Cheung et al., 2013). Current EU legislation relies on the World Health Organization's (WHO) guidelines for drinking water (WHO, 2008). In 1998,
http://dx.doi.org/10.1016/j.bios.2014.10.039 0956-5663/& 2014 Published by Elsevier B.V.
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
the WHO defined the regulatory limit for the presence of microcystin-LR in drinking water as 1 μg L 1 (or 1 ng mL 1) (WHO, 1998) and since then has maintained this guideline following two subsequent reviews (WHO, 2008). Analytical methodologies set out by the International Office for Standardization (ISO) for microcystin detection, such as HPLC coupled to UV, require extensive sample pre-treatment procedures, are highly susceptible to interferences by matrix effects and are often fundamentally incompatible with ‘in-situ’ measurement. This article describes the application of an evanescent opticalplanar waveguide sensor that utilizes a recombinant antibodysensing element. The use of optical-sensors is advantageous in that they can be miniaturized for portable ‘in-situ’ use, exhibit low signal-to-noise ratios and require low reagent volumes and function well in complex matrices (Conroy et al., 2009). Planar waveguide arrays were developed by Long et al. and Herranz et al. using monoclonal antibodies to detect microcystin. The system developed by Long utilizes a fibre optic probe covalently immobilized with microcystin-conjugate onto a self-assembled thiol-silane monolayer, while Herranz's system uses a microscope slide covalently coated with MC-LR toxin, has very low detection limits and a total analysis time of 60 min (Herranz et al., 2012; Long et al., 2009). The evanescent waveguide used in this article was initially developed by Lochhead et al. (2011) as an inexpensive, disposable, single-use system next-generation lateral-flow device (LFD) (Devlin et al., 2013). In this study bespoke toxin-conjugates were printed onto the plastic planar-waveguide which was bonded to a plastic upper component defining the flow of the channel. The toxin conjugates are housed within the cartridge, thus mitigating exposure risk to the user (Lochhead et al., 2011). A novel scFv fragment was generated during this study and is the first documentation of a microcystin-specific recombinant antibody from an immunized avian source (Andris-Widhopf et al., 2000). A number of recombinant antibodies were previously generated to microcystin (McElhiney et al., 2002; McElhiney et al., 2000), however, these antibodies were from non-immune sources. A major advantage of using immune libraries, over conventional naïve antibody libraries, is that the host animal immune system can refine and enhance antigen affinity and specificity. It was reported that the most sensitive microcystin-specific scAb isolated, was capable of detecting microcystin-LR at levels below the World Health Organisation limit in drinking water (1 μg L 1). However, the antibody fragment only cross-reacted with three microcystin variants (MC-RR, -LW, -LF and nodularin). The incorporation of a highly engineered antibody fragment provides a system with greater detection capabilities and greater assay specificity.
Materials and methods
Host immunization and antibody library generation Immune library generation was carried out according to Andris-Widhopf et al. (2000). MC-LR-BSA (500 μg) was mixed in a 1:1 ratio with Freund's complete adjuvant (FCA) in a 1 mL final volume. The emulsified mixture was administered sub-cutaneously over four sites in 250 mL injection volumes. Subsequent booster immunizations were carried out in the same manner using Freund's incomplete adjuvant (FIA) over a twelve-week period. RNA isolated from the spleen and bone marrow was reverse transcribed to complementary DNA (Superscript III kit, Invitrogen). Antibody variable heavy (VH) and variable light (VL) chain genes were amplified from the cDNA. The scFv gene products were amplified using splice by overlap extension (SOE) PCR, and cloned into the pComb3XSS vector using restriction enzyme SfiI (New England Biolabs). The antibody library was transformed into electro-competent Escherichia coli XL1-Blue cells (Agilent Technologies) by electroporation. The library titre was approximately 3.5 107 CFU mL 1. Phage selection and screening Immune library selection was carried out over five rounds using M13K07 helper phage (New England Biolabs). Phage were incubated and enriched by decreasing the concentration of MC-LROVA (100–5 μg mL 1) coated on immunotubes (Nalgenes). Further information on the selection of the most sensitive anti-microcystin-LR clones is available in supplementary information. The 10 best clones, with minimal cross-reactivity to BSA, OVA and KLH (2 μg mL 1 coated direct ELISA) were selected. Concomitantly, DNA sequencing analysis was carried out for identification of unique sequences (Eurofins). Antibody expression and purification by immobilized metal affinity and size exclusion chromatography The presence of a polyhistidine tag on the recombinant antibodies facilitated purification by immobilized metal affinity chromatography (IMAC). For more information please see the accompanying supplementary documentation. Size exclusion chromatography (SEC) was carried out using an AKTA FPLC (GE Healthcare) utilizing UNICORN software. A HiLoad Superdex 200 16/60 column (GE Healthcare) was equilibrated with 3 volumes of PBS. A 1 mL sample of IMAC-purified scFv was filtered through a 0.2 μm filter and injected into a 2 mL loop. The flow rate was fixed at 1 mL min 1. Fractions of interest were collected in 5 mL volumes and were concentrated to 500 μL volumes using 10,000 MW cutoff vivaspin columns. The purity of each fraction was analysed by SDS-PAGE, the fractions were combined and the concentration was established using a Nanodrop NC-1000 system and a modified Lowry protein quantification assay (Thermo Scientific).
Preparation of microcystin-LR conjugates Microcystin-LR (2 mg) derivatisation was carried out by Enzo Life Sciences (UK), as described by Moorhead et al. (1994). The reaction mixture was analysed by reverse-phase high-performance liquid chromatography (RP-HPLC) and was found to contain 495% derivatised aminoethanethiol-microcystin-LR (1 mg) which was subsequently conjugated to bovine serum albumin (BSA) (3 mg) and ovalbumin (OVA) (3 mg) using a glutaraldehyde coupling procedure (Metcalf et al., 2000). The two bespoke conjugates were dialysed against 0.9% (w/v) saline. The resulting microcystin-LR conjugates were used for subsequent immune library generation, selection and screening.
Assessment of matrix interferences in competitive MC-LR determination assay Three different lake water samples from three different sites were taken from two different lakes in Ireland, Lough Ennell and Lough Derg, and were compared to deionized water, tap water and PBS to examine the effect of the matrix on microcystin determination. The water samples were filtered through a 0.4 μm filter to remove any large debris. The analytical performance of the antiMC-LR recombinant antibody was assessed using a competitive ELISA (please see Supplementary information for further experimental detail).
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
MC-LR (Enzo Life Sciences) was directly conjugated to Biacore CM5 sensor chips (GE Healthcare) according to Vinogradova et al. (2011). The SPR-based Biacore inhibition assay was developed using the purified 2G1 recombinant anti-microcystin-LR scFv (optimized at 1/7000 in HBS buffer). The scFv was mixed with a range of concentrations of microcystin-LR, -YR -LA, -LF, -LW, -RR, and nodularin (0.69–300 ng mL 1). The assay was carried out at a flow rate of 5 μL min 1 with 2 min injections (n ¼ 3) followed by surface regeneration (two pulses of 200 mM NaOH for 1 min). The IC50 values for each toxin were directly compared to that of MC-LR to determine the degree of cross reactivity. Inter-day and intra-day assay studies were performed to verify the accuracy and precision. The standard curve was generated (n ¼3) on a single day (intra-day) and on three consecutive days (inter-day) and the mean and standard error were calculated. The %CV were determined for each concentration and calibration curves were generated from the data with a four-parameter equation, using BIAevaluation™ 3.0 software. Site-directed mutagenesis of 2G1 antibody fragment using molecular docking approach Maximum Entropy based Docking (MEDock) (Chang et al., 2005) predicted a structurally important amino acid change from arginine to serine at position 66 in the framework region in the heavy chain of the scFv may impact on antibody:toxin binding. Quikchange II XL site directed mutagenesis kit from Stratagene was used to mutate the amino acid using ‘in-house’ designed primers (synthesized by Eurofins Genomics). The novel antibody against microcystin 2G1-R66S was used in MBio SnapESIs opticalplanar waveguide immunoassay studies.
Microarrays were then assembled into an injection-moulded cartridge, which contained a 5 mm wide fluidic channel (max volume 30 mL) with a single inlet port for the addition of sample and reagents.
Results Isolation of anti-microcystin scFv fragments from an immunized phage library Recombinant antibody fragments targeting MC-LR were developed from an immunized phage library. The enrichment of microcystin-specific phage binders over five rounds of biopanning was confirmed by polyclonal phage ELISA (Fig. 1A) (for more information see Supplementary information). Following infection of phage outputs from round 4 and 5 into E. coli Top10F’, preliminary screening was carried out using a direct monoclonal ELISA for identification of MC-LR-specific clones (Fig. 1B). All scFv fragments displayed binding to the MC-LR-OVA
A Absorbance (450 nm)
Development of a surface plasmon resonance (SPR)-based inhibition assay
3
2.5 2.0 1.5 1.0 0.5
Biotinylation of 2G1-R66S variant
Microarray printing and cartridge development The generation of portable optical-planar waveguide cartridges for the determination of microcystin was carried out in conjunction with MBio Diagnostics, Colorado, USA. Briefly, the planar waveguide substrate is an injection moulded plastic component activated with a proprietary silane-based surface chemistry. Biomolecule binding to the waveguide surface is through a combination of covalent reactions (e.g. via surface lysines on the printed proteins) and non-specific interactions (e.g., hydrophobic, electrostatic, van der Waals). MC-LR-conjugates were printed using a Bio-Dot AD3200 robotic arrayer onto microarrays at varying concentrations (microcystin-LR-OVA at 10 and 5 μg mL 1). Spots were produced, with a diameter of 0.5 mm using a Bio-Jet print head that dispensed 20 nL. Two replicates of the toxin-conjugates were printed onto a grid with 1 mm centres, blocked with a proteinbased blocking agent (0.5% (w/v) casein in PBS (150 mM, pH 7.4) with Proclin300 antimicrobial agent) prior to spin-drying.
0.0
B
Prepan Pan 1 Pan 2 Pan 3 Pan 4 MC-LR-OVA OVA
Pan 5
1.6 1.4
Absorbance (450 nm)
ScFv 2G1-R66S used in the development of the MBio SnapESI immunoassay was purified as described. 2G1-R66S (anti-MC-LR) was biotinylated using 100-fold molar excess of sulfo-NHS-LCBiotin (Pierce). The 2G1-R66S anti-MCLR-biotin mixture was buffer exchanged (PBS 150 mM, pH 7.4) to remove unreacted biotin. Using a biotin quantification kit (Pierce), a conjugation ratio of 1:1.7 mmol 2G1-R66S to biotin was determined. StreptavidinAlexa-647 (Molecular Probes, Invitrogen) was used to detect the presence of biotinylated-anti-MC-LR 2G1-R66S in the MBio SnapESIs MC-LR detection assay.
1.2 1.0 0.8 0.6 0.4 0.2 0.0
Clones 1-96
Fig. 1. Polyclonal phage ELISA and soluble scFv monoclonal ELISA: (A) polyclonal phage ELISA for the assessment of anti-microcystin-LR phage-scFv binding over five rounds of bio-panning. The experiments were performed in triplicate and error bars represent the standard error of the mean. (B) Direct binding to MC-LR-OVA was observed using monoclonal, soluble ELISA, using an anti-HA-HRP conjugate as the secondary detection reagent. The background binding (assessed as three times the blank measurement) is indicated by a dashed red line. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
conjugate above the background signal. No cross-reactivity to protein carriers including OVA, BSA or KLH was observed. The positively identified candidate clones were analyzed by inhibition ELISA (data not shown). The best scFv, 2G1, was expressed and purified by immobilized metal affinity and size exclusion chromatography, as described.
Development of biacore inhibition immunoassay and assessment of cross-reactivity profile A Biacore inhibition assay for the detection of MC-LR was developed using the specific scFv fragment 2G1. The 2G1 scFv displayed an IC50 of 4.8 ng mL 1. The antibody response values (R) for each concentration were normalized by the transformation to % R/R0, according to the equation: %R/R0 ¼(R Rexcess)/(R0 Rexcess), where R is the relative response, R0 is the response at a zero concentration of analyte, and Rexcess is the response at an excess of analyte. The %R/R0 value was plotted against the logarithm of the MC-LR concentration (ng mL 1). The inter-day calibration curve for the 2G1 scFv fragment is presented in Fig. 2A. The limit of detection was calculated from the mean of the zero analyte concentration, minus three standard deviations, as 1.7 ng mL 1. To determine the degree of cross-reactivity of the scFv fragment 2G1 to the six structurally-related variants of microcystin, MC-LW, MCRR, MC-YR, MC-LA, MC-LF and nodularin, an inhibition analysis was performed. Percentage cross-reactivities (%CR50) were extrapolated by comparing the relevant IC50 of the microcystin variants to the IC50 of the MC-LR toxin. The %CR50 was calculated using the equation: %cross reactivity 50 (CR50)¼ (IC50 MC variant/IC50 MCLR) 100. The cross-reactivity profiles for the 2G1 scFv clone are illustrated in Fig. 2B. The degree of cross-reactivity of the 2G1 scFv to the MC variants are comparable with MC-LR, with MC-YR and MC-LA showing the highest cross-reactivity relative to MC-LR.
A 100 % R/Ro
80 60 40 20 0 0.1
1
10
100
Microcystin-LR (ng mL-1)
B Variant MC-LR MC-YR MC-LA Nodularin MC-LF MC-LW MC-RR
IC50 (ng mL-1) 4.80 8.06 6.42 9.28 7.11 10.08 7.99
% CR 100.00 79.73 74.75 69.25 67.49 63.75 60.07
Fig. 2. Biacore inhibition assay curve and microcystin cross reactivity profiles. (A) Biacore inhibition assay curve for the determination of MC-LR in PBS, for the 2G1 (wild-type) scFv fragment. MC-LR concentrations, ranging from 40 to 0.69 ng mL 1 were pre-incubated with 2G1 anti-MC-LR scFv (1/7000) and injected over the MC-LR-immobilized surface. The error bars represent the standard deviation for each concentration. (B) Cross-reactivity profile of MC toxins with the 2G1 scFv fragment in HBS-EP buffer.
Identification of important structural scFv regions Increased MC-LR binding affinity was achieved by identifying important structural aspects of the scFv-2G1 polypeptide. To evaluate possible molecular docking permutations between MC-LR and anti-microcystin scFv, MEDock was used. Fig. 3 displays ligand (MC-LR) docked to 2G1-scFv. The programme predicted a number of possible amino acid mutations to increase antibody/toxin affinity (http://medock.csie.ntu.edu.tw/ and http://bioinfo.mc.ntu. edu.tw/medock/) (Chang et al., 2005). An increase in antibody affinity for toxin microcystin-LR was determined by inhibition ELISA (data not shown). The novel antibody against microcystin 2G1R66S was used in subsequent MBio SnapESIs optical-planar waveguide immunoassay validation studies. Testing the performance of 2G1-R66S in various sample matrices To assess the influence of sample matrix on the determination of MC-LR in freshwater water compared to PBS, tap water and deionized water a competitive ELISA was carried out (Fig. 4). Lake water from three different locations in Ireland (Lough Ennell, and two different locations in Lough Derg (Dromineer and Luska) were assessed. The ability of the antibody to bind to free toxin in the presence of the different matrices (A) was compared to the binding of each antibody to toxin conjugate in the absence of free toxin (A0). The results were normalized by subtracting the absorbance value at excess analyte (Aexcess). The resultant normalized absorbance (A Aexcess/A0 Aexcess) was plotted as A/A0 versus MC-LR concentration (ng mL 1) as shown in Fig. 4. The experiment was repeated twice and each sample determination was carried out in triplicate. The IC50 values for each sample were obtained (IC50 values: PBS, 5.0 ng mL 1; tap water, 3.8 ng mL 1; deionized H2O, 6.7 ng mL 1; L. Derg-Dromineer, 5.7 ng mL 1; L. Derg -Luska, 4.4 ng mL 1 and L. Ennell, 4.4 ng mL 1. Development of optical-planar waveguide assay for the determination of MC-LR in cyanobacterial cultures using recombinant antibody 2G1-R66S Assay format The ability of free microcystin in solution to inhibit the binding of antibody to immobilized microcystin-conjugate forms the basis of the assay. Different conjugate coating concentrations were tested, including MC-LR-BSA at 200, 100 and 50 μg mL 1 and MCLR-OVA at 100, 50, 10 and 5 μg mL 1 and it was determined that MC-LR-OVA at 10 μg mL 1 achieved the most reproducible and sensitive results. Antibody stock was stored at 20 °C in 50 μL aliquots of a concentration 1 mg mL 1 in PBS (150 mM, pH 7.4). A range of antibody dilutions were tested i.e. 1:102, 1:103, 1:104 and 1:105, it was found that scFv 2G1-R66S worked optimally on the MBio SnapESIs system at 1 in 104 dilution. The format of the assay relies on the addition of Streptavidin-Alexa-647s to identify the presence of biotinylated anti-microcystin scFv. All steps were carried out at room temperature and are as follows:
Addition of antibody: toxin sample to cartridge – 10 min incubation at room temperature.
1 Wash in PBS for 2 min. Detection molecule (streptavidin-Alexa-647) addition – 5 min incubation at room temperature.
1 Wash in PBS 2 min. Insert cartridge into Mbio system and obtain fluorescent reading.
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Fig. 3. Molecular docking model of microcystin-LR and scFv-2G1. The ligand (MC-LR) is displayed in ‘ball and stick’ mode and is represented in red, the white area shows the binding cavity, the scFv light and heavy chain are displayed in green and yellow respectively, and the glycine-serine linker is displayed in blue (scFv is displayed in ‘space filling’ mode). The structure was visualized using ICM-Browser tool. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
1.0
PBS Deionized H20 Tap water L. Derg Luska L. Derg Dromineer L. Ennell
A/A0
0.8 0.6 0.4 0.2 0.0
0.5
1
2
4
8
Microcystin-LR
16
32
64
(ng mL-1)
Fig. 4. Competitive ELISA comparing the effects of sample matrices on the determination of MC-LR using recombinant scFv 2G1-R66S. PBS, deionized H2O and tap water were compared to three fresh water Irish lake samples (Lough Derg Luska, Lough Derg Dromineer and Lough Ennell). The results were normalized by subtracting the absorbance value at excess analyte (Aexcess). The resultant normalized absorbances (A Aexcess/A0 Aexcess) were plotted as A/A0 versus MC-LR concentration (ng mL 1).
Assay validation Outline of assay format and generation of standard curve. To assess assay reproducibility, analytical standards were used to generate a standard curve and the ability of anti-MC-LR 2G1-R66S-biotin (1:104 diluted in 1% BSA-PBS (150 mM, pH 7.4) (w/v)) to accurately detect the standards was assessed. MC-LR standards were prepared in M. aeruginosa sample matrix (CCAP 1450/03) (described by Devlin et al.) to account for matrix effects. The calibration curve contained toxin concentrations from 25 ng mL 1 to 48.8 pg mL 1 (at final working concentrations). Anti-MC-LR 2G1-R66S-biotin was incubated with the calibrants for 5 minutes at room temperature prior to testing. Each MC-LR-calibrant and anti-MC-LR mixture was applied to the MBio SnapESIs cartridge at a volume of 175 μL and incubated for at room temperature (10 min). Filter sterilized PBS (175 μL, 150 mM, pH 7.4) was added to the well (2 min). Streptavidin-alexa-647 (Molecular Probes, Invitrogen) diluted in PBS (150 mM, pH 7.4) (1:2,000, 175 μL) was added to the cartridge (5 min). The cartridge was washed with 175 μL filter sterile PBS (2 min) and the signal was measured using MBio SnapESIs system and a calibration curve was generated.
Determination of limit of detection (LOD). The LOD was established by measuring the limit of blank plus three standard deviations (Armbruster and Pry, 2008). A calibration curve was generated using inter-assay fluorescence results generated over three days using scFv 2G1-R66S-biotin diluted in sample matrix. The mean fluorescence value of twenty ‘microcystin-free’ matrix samples plus the value of three standard deviations was determined and interpolated onto the curve. An analytical LOD of 0.28 ng mL 1 was achieved Table 1. The IC50 value and the dynamic range (IC10 to IC90) were established using two batches of cartridges on three separate days. The values are displayed in Table 1. The assay displayed an IC50 value of 1.1 ng mL 1 and a dynamic range of 0.21–5.94 ng mL 1. To establish the detection capability of the assay (CCβ) (also known as the functional limit of detection), low concentrations of MC-LR (0.1 ng mL 1, 0.19 ng mL 1 and 0.5 ng mL 1) were assayed measuring 20 toxin samples at each concentration and 20 blank samples with all tests carried out in the complex sample matrix. Samples containing 0.19 and 0.5 ng mL 1 (analysed by un-paired two-tailed t-test) were found to be significantly different, each displaying a p value o 0.0001. Therefore, as the lowest value, 0.19 ng mL 1 was identified as the functional limit of detection (CCβ) (Fig. 5). Repeatability studies were carried out over three days using four different batches of cartridges. Matrix samples were spiked with 3.3, 1.19 and 0.5 ng mL 1 (IC15, IC50 and half the WHO recommended detection limit, respectively). The results are shown in Table 2. Data were fit using a four-parameter non-linear regression calibration curve. Inter-assay results displayed percentage accuracies between 88 and 99%, with intra-day assay displaying results between 82 and 103%. The percentage CV was determined using equation CV%¼ (SD/Mean) 100. All percentage coefficients of variance were below 8% for intra-day assays and below 11% for between day assays. The assay was assessed for its ability to detect MC-LR in lake water (Lough Derg, Ireland) and deionized water (diH2O). Two of Table 1 Analytical limit of detection, IC50 values and dynamic range of assay.
10 μg mL 1 MC-LROVA-coated assay spot
Analytical LOD (ng mL 1)
IC50 (ng mL 1)
Dynamic range (IC10 to IC90, ng mL 1)
0.28
1.11
0.21–5.94
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
Fig. 5. Functional LOD determination (CCβ) of microcystin specific scFv 2G1-R66S. The dotted line represents the point at which over 95% of spiked samples are distinguishable over blank samples and displayed p values o0.0001. Table 2 MBio SnapESI repeatability measurements.
Intra-day Day 1
Day 2
Day 3
Inter-day Days 1–3
Spiking level (ng mL 1)
Back calculated concentration (ng mL 1)
3.30 1.19 0.50 3.30 1.19 0.50 3.30 1.19 0.50
3.25 1.21 0.46 3.37 1.14 0.41 3.16 1.08 0.45
4.9 5.1 7.8 6.3 4.8 4.5 5.1 5.2 5.8
98.5 101.7 92.0 102.1 95.8 82.0 95.8 90.8 90.0
3.30
3.26
7.35
98.8
1.19 0.50
1.14 0.44
10.44 1.40
95.8 88.0
CV (%) Specific accuracy (%)
the samples were supplemented with M. aerguinosa extract (preparation previously described in paper). The presence of M. aerguinosa extract or lake water contaminants had minimal impact on the assay performance, and diH2O, M. aerguinosa in diH2O and M. aerguinosa in lake water displaying IC50 values of 1.62, 1.66 and 1.57 ng mL 1, respectively (see Supplementary information).
Discussion The planar waveguide immunoassay presented is of exquisite sensitivity, and exploits the in-built ability to genetically alter recombinant antibodies to increase their affinity for target antigen, thus enhancing their capacity in every-day immunoassays. Recombinant antibodies have been developed previously to microcystin. However, none have been developed using an immune host. In the present system, the single chain fragment variable (scFv) molecule recognizes the presence of microcystin-LR and its congeners displaying cross-reactivites to a range of MC-congeners (MC-YR 79%, MC-LA 74.8%, MC-LR 67.5%, MC-LW 63.7%, 60.1%, Nod 69.3%). The ability of 2G1 scFv to cross-react with other members of the MC family increases the utility value of the scFv in immunoassays, facilitating accurate toxicity determination in an environmental, complex sample which could potentially contain a
number of microcystin congeners. The ability of the antibody to determine MC-LR in the presence of various sample matrices was assessed by competitive ELISA. Different samples of fresh water were spiked with MC-LR and the resultant IC50 values displayed the relatively low influence that sample matrix has on the ability of the recombinant antibody to measure MC-LR. To further enhance the binding avidity of scFv-2G1 for MC-LR, docking permutations between scFv 2G1 and MC-LR were evaluated using a molecular docking simulation. The MEDock server predicted a possible ligand-binding site at position 66 of the heavy chain that could be mutated to increase affinity. This amino acid was also predicted by the (CASTp, 2011) server to play a role in MC-LR binding, and interestingly is highly conserved as a serine residue in human antibodies (ClustalW analysis). Using the genetically altered and biotinylated scFv (2G1-R66Sbiotin) an evanescent optical-planar waveguide immunoassay using the MBio SnapESIs system was developed for the detection of microcystin and a conservative analytical LOD of 0.28 ng mL 1 was determined. The functional LOD (CCβ) of 0.19 ng mL 1 toxin in matrix sample was identified (Fig. 3). These findings improve upon a similar study for the detection of microcystin (Devlin et al., 2013), by lowering the limit of detection 5 times from 1 ng mL 1 to 0.19 ng mL 1. It is important to note that using this system, the presence of toxin can be detected following a simple mechanical toxin extraction procedure that only takes 10 min (Devlin et al., 2013). The presence of M. aeruginosa cellular debris impacts on the sensitivity of assays; however, in the presence of cellular debris this system can still detect the presence of toxin below the WHO recommended limit. Intra-day analyses generated data that displayed percentage accuracies between 82 to 102.2% (Table 2) while the inter-day study returned values of between 88 to 99%, thus displaying better values than previous reports. The resultant percentage CVs produced very tight (low) values with the highest recorded CV at 7.8%. Intermediate repeatability was quantified by inter-day studies, whereupon the highest exhibited % CV value was 10.44. The results thus highlight the high precision afforded to the detection of MCLR by the use of this assay.
Conclusion This is the first report, to our knowledge, of the generation of a recombinant antibody fragment against MC-LR, from an immunized avian host. The antibody fragment displayed cross-reactivity values towards six MC congeners in an SPR-based immunoassay. The antibody fragment was subjected to structural scrutiny and an amino acid at position 66 in the variable heavy chain was genetically modified from arginine to serine. Subsequently, a recombinant antibody-based evanescent optical-planar waveguide system capable of reliably detecting the presence of MC-LR was developed which displayed sensitivities comparable to those achievable with reverse phase HPLC (Pyo et al., 2005; WHO, 2008). The inhibition assay has a complete run time of 19 min and can detect the presence of microcystin with a functional LOD of 0.19 ng mL 1 in complex sample matrix, a five times greater sensitivity over previous reports.
Acknowledgements This work was supported by Science Foundation Ireland (SFI) as part of a US-Ireland research and development partnership programme BEACONS (Biosafety for EnvironmentAL Contaminants
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
using Novel Sensors) Grant code 08/US/I1512 and the Beaufort Marine Research Award under the Sea Change Strategy and the Strategy for Science Technology and Innovation (2006–2013) and with the support of the Marine Institute, funded under the Marine Research sub-programme of the National Development plan 2007–2013.
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at. http://dx.doi.org/10.1016/j.bios.2014.10.039
References Andris-Widhopf, J., Rader, C., Steinberger, P., Fuller, R., Barbas 3rd, C.F., 2000. Methods for the generation of chicken monoclonal antibody fragments by phage display. J. Immunol. Methods 242, 159–181. Armbruster, D.A., Pry, T., 2008. Limit of blank, limit of detection and limit of quantitation. The Clinical biochemist. Rev. Aust. Assoc. Clin. Biochem. 29 (Suppl. 1), S49–S52. Bhattacharya, R., Sugendran, K., Dangi, R.S., Rao, P.V., 1997. Toxicity evaluation of freshwater cyanobacterium microcystis aeruginosa PCC 7806: II. Nephrotoxicity in rats. Biomed. Environ. Sci. 10 (1), 93–101. CASTp, 2011. 〈http://sts-fw.bioengr.uic.edu/castp/calculation.php〉. Chang, D.T., Oyang, Y.J., Lin, J.H., 2005. MEDock: a web server for efficient prediction of ligand binding sites based on a novel optimization algorithm. Nucl. Acids Res. 33, W233–238. Cheung, M.Y., Liang, S., Lee, J., 2013. Toxin-producing cyanobacteria in freshwater: a review of the problems, impact on drinking water safety, and efforts for protecting public health. J. Microbiol. 51 (1), 1–10. Conroy, P.J., Hearty, S., Leonard, P., O'Kennedy, R.J., 2009. Antibody production, design and use for biosensor-based applications. Semin. Cell Dev. Biol. 20 (1), 10–26. Devlin, S., Meneely, J.P., Greer, B., Greef, C., Lochhead, M.J., Elliott, C.T., 2013. Next generation planar waveguide detection of microcystins in freshwater and cyanobacterial extracts, utilising a novel lysis method for portable sample preparation and analysis. Anal. Chim. Acta 769, 108–113. Hernandez, J.M., Lopez-Rodas, V., Costas, E., 2009. Microcystins from tap water could be a risk factor for liver and colorectal cancer: a risk intensified by global change. Med. Hypotheses 72 (5), 539–540. Herranz, S., Marazuela, M.D., Moreno-Bondi, M.C., 2012. Automated portable array biosensor for multisample microcystin analysis in freshwater samples. Biosens. Bioelectron. 33 (1), 50–55.
7
Humpage, A., 2008. Toxin types, toxicokinetics and toxicodynamics. Adv. Exp. Med. Biol. 619, 383–415. Lochhead, M.J., Todorof, K., Delaney, M., Ives, J.T., Greef, C., Moll, K., Rowley, K., Vogel, K., Myatt, C., Zhang, X.Q., Logan, C., Benson, C., Reed, S., Schooley, R.T., 2011. Rapid multiplexed immunoassay for simultaneous serodiagnosis of HIV-1 and coinfections. J. Clin. Microbiol. 49 (10), 3584–3590. Long, F., He, M., Zhu, A.N., Shi, H.C., 2009. Portable optical immunosensor for highly sensitive detection of microcystin-LR in water samples. Biosens. Bioelectron. 24 (8), 2346–2351. MacKintosh, C., Beattie, K.A., Klumpp, S., Cohen, P., Codd, G.A., 1990. Cyanobacterial microcystin-LR is a potent and specific inhibitor of protein phosphatases 1 and 2 A from both mammals and higher plants. FEBS Lett. 264 (2), 187–192. McElhiney, J., Drever, M., Lawton, L.A., Porter, A.J., 2002. Rapid isolation of a singlechain antibody against the cyanobacterial toxin microcystin-LR by phage display and its use in the immunoaffinity concentration of microcystins from water. Appl. Environ. Microbiol. 68 (11), 5288–5295. McElhiney, J., Lawton, L.A., Porter, A.J., 2000. Detection and quantification of microcystins (cyanobacterial hepatotoxins) with recombinant antibody fragments isolated from a naive human phage display library. FEMS Microbiol. Lett. 193 (1), 83–88. Metcalf, J.S., Bell, S.G., Codd, G.A., 2000. Production of novel polyclonal antibodies against the cyanobacterial toxin microcystin-LR and their application for the detection and quantification of microcystins and nodularin. Water Res. 34 (10), 2761–2769. Moorhead, G., MacKintosh, R.W., Morrice, N., Gallagher, T., MacKintosh, C., 1994. Purification of type 1 protein (serine/threonine) phosphatases by microcystinsepharose affinity chromatography. FEBS Lett. 356 (1), 46–50. Pyo, D., Park, G., Choi, J., Oh, C., 2005. Microcystin detection characteristics of fluorescence immunochromatography and high performance liquid chromatography. Bull. Korean Chem. Soc. 26 (2), 268–272. Qin, B., Zhu, G., Gao, G., Zhang, Y., Li, W., Paerl, H.W., Carmichael, W.W., 2010. A Drinking Water Crisis in Lake Taihu, China: linkage to climatic variability and lake management. Environm. Manag. 45 (1), 105–112. Turner, P.C., Gammie, A.J., Hollinrake, K., Codd, G.A., 1990. Pneumonia associated with contact with cyanobacteria. Br. Med. J. 300 (6737), 1440–1441. Vareli, K., Jaeger, W., Touka, A., Frillingos, S., Briasoulis, E., Sainis, I., 2013. Hepatotoxic seafood poisoning (HSP) due to microcystins: a threat from the ocean? Mar. Drugs 11 (8), 2751–2768. Vinogradova, T., Danaher, M., Baxter, A., Moloney, M., Victory, D., Haughey, S.A., 2011. Rapid surface plasmon resonance immunobiosensor assay for microcystin toxins in blue-green algae food supplements. Talanta 84 (3), 638–643. WHO, 1998. Guidelines for Drinking Water Quality, 2nd ed. World Health Organisation, Geneva, pp. 95–110. WHO, 2008. Guidelines for Drinking Water Quality (incorporating the first and second addenda). Third Edition World Health Organization, Geneva, Switzerland, ISBN: 978 924 154 761 1 (web version).
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Detection of the cyanobacterial toxin, microcystin-LR, using a novel recombinant antibody-based optical-planar waveguide platform Caroline Murphy a,n, Edwina Stack a, Svetlana Krivelo a, Daniel A. McPartlin a, Barry Byrne a, Charles Greef b, Michael J. Lochhead b, Greg Husar b, Shauna Devlin c, Christopher T. Elliott c, Richard J. O’Kennedy a a
School of Biotechnology, National Centre for Sensor Research and Biomedical Diagnostics Institute, Dublin City University, Dublin 9, Ireland MBio Diagnostics Inc., Boulder, CO, 80301, USA c Institute for Global Food Security (IGFS), School of Biological Sciences, Queen’s University, Malone Road, 18-30 Belfast, BT9 5BN, Northern Ireland, United Kingdom b
art ic l e i nf o
a b s t r a c t
Article history: Received 13 June 2014 Received in revised form 2 October 2014 Accepted 13 October 2014
Microcystins are a major group of cyanobacterial heptapeptide toxins found in freshwater and brackish environments. There is currently an urgent requirement for highly-sensitive, rapid and in-expensive detection methodologies for these toxins. A novel single chain fragment variable (scFv) fragment was generated and is the first known report of a recombinant anti-microcystin avian antibody. In a surface plasmon resonance-based immunoassay, the antibody fragment displayed cross-reactivity with seven microcystin congeners (microcystin–leucine–arginine (MC-LR) 100%, microcystin–tyrosine–arginine (MC-YR) 79.7%, microcystin–leucine–alanine (MC-LA) 74.8%, microcystin–leucine–phenylalanine (MC-LF) 67.5%, microcystin–leucine–tryptophan (MC-LW) 63.7%, microcystin–arginine–arginine (MC-RR) 60.1% and nodularin (Nod) 69.3%, % cross reactivity). Following directed molecular evolution of the parental clone the resultant affinity-enhanced antibody fragment was applied in an optimized fluorescence immunoassay on a planar waveguide detection system. This novel immuno-sensing format can detect free microcystin-LR with a functional limit of detection of 0.19 ng mL 1and a detection range of 0.21–5.9 ng mL 1. The assay is highly reproducible (displaying percentage coefficients of variance below 8% for intra-day assays and below 11% for inter-day assays), utilizes an inexpensive cartridge system with low reagent volumes and can be completed in less than twenty minutes. & 2014 Published by Elsevier B.V.
Keywords: Cyanobacteria Microcystin Recombinant Antibody Immunoassay Fluorescent planar wave-guide
Introduction Cyanobacteria are ‘blue-green’ prokaryotes found globally in fresh-water bodies such as lakes, ponds, reservoirs, slow moving rivers and brackish environments (Vareli et al., 2013; WHO, 2008). During periods of excessive nutrient enrichment, cyanobacterial ‘blooms’ flourish and, depending on the predominance of certain cyanobacteria, toxins known as cyanotoxins, are generated. Species of cyanobacteria such as Anabaena spp. and Microcystis spp. produce the most commonly found cyanotoxin, microcystin-LR (Vareli et al., 2013), a cyclic-heptapeptide (7 amino acids) with over 80 structurally related congeners (Humpage, 2008). A similar structurally related pentapeptide (5 amino acids), nodularin, is produced by Nodularia spp. Microcystin’s mode of toxicity is n Correspondence to: School of Biotechnology, Dublin City University, Glasnevin, Dublin 9, Dublin, Ireland. E-mail address: [email protected] (C. Murphy).
associated with its ability to inhibit serine/threonine protein phosphatases 1 and 2 A (MacKintosh et al., 1990) leading to hepatocytic necrosis and haemorrhaging (Bhattacharya et al., 1997). Cases of acute poisoning have resulted in severe nausea, vomiting, diarrhoea, abdominal pains, sore throats, headaches, blisters around the mouth and dry coughs (Turner et al., 1990). Chronic exposure to microcystin is a worrying phenomenon as microcystin over an extended period can cause cancer through the inhibition of protein phosphatases PP1 and PP2A (Hernandez et al., 2009). Bursts of cyanobacterial growth have significant socioeconomic impacts. In 2007, a cyanobacterial bloom of Microcystis aeruginosa contaminated the Chinese city of Wuxi's only drinking water supply, and the city's two million residents were without water for one week (Qin et al., 2010). The ubiquitous nature of microcystins illustrates their global prevalence and underpins the requirement for regular and accurate monitoring (Cheung et al., 2013). Current EU legislation relies on the World Health Organization's (WHO) guidelines for drinking water (WHO, 2008). In 1998,
http://dx.doi.org/10.1016/j.bios.2014.10.039 0956-5663/& 2014 Published by Elsevier B.V.
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
the WHO defined the regulatory limit for the presence of microcystin-LR in drinking water as 1 μg L 1 (or 1 ng mL 1) (WHO, 1998) and since then has maintained this guideline following two subsequent reviews (WHO, 2008). Analytical methodologies set out by the International Office for Standardization (ISO) for microcystin detection, such as HPLC coupled to UV, require extensive sample pre-treatment procedures, are highly susceptible to interferences by matrix effects and are often fundamentally incompatible with ‘in-situ’ measurement. This article describes the application of an evanescent opticalplanar waveguide sensor that utilizes a recombinant antibodysensing element. The use of optical-sensors is advantageous in that they can be miniaturized for portable ‘in-situ’ use, exhibit low signal-to-noise ratios and require low reagent volumes and function well in complex matrices (Conroy et al., 2009). Planar waveguide arrays were developed by Long et al. and Herranz et al. using monoclonal antibodies to detect microcystin. The system developed by Long utilizes a fibre optic probe covalently immobilized with microcystin-conjugate onto a self-assembled thiol-silane monolayer, while Herranz's system uses a microscope slide covalently coated with MC-LR toxin, has very low detection limits and a total analysis time of 60 min (Herranz et al., 2012; Long et al., 2009). The evanescent waveguide used in this article was initially developed by Lochhead et al. (2011) as an inexpensive, disposable, single-use system next-generation lateral-flow device (LFD) (Devlin et al., 2013). In this study bespoke toxin-conjugates were printed onto the plastic planar-waveguide which was bonded to a plastic upper component defining the flow of the channel. The toxin conjugates are housed within the cartridge, thus mitigating exposure risk to the user (Lochhead et al., 2011). A novel scFv fragment was generated during this study and is the first documentation of a microcystin-specific recombinant antibody from an immunized avian source (Andris-Widhopf et al., 2000). A number of recombinant antibodies were previously generated to microcystin (McElhiney et al., 2002; McElhiney et al., 2000), however, these antibodies were from non-immune sources. A major advantage of using immune libraries, over conventional naïve antibody libraries, is that the host animal immune system can refine and enhance antigen affinity and specificity. It was reported that the most sensitive microcystin-specific scAb isolated, was capable of detecting microcystin-LR at levels below the World Health Organisation limit in drinking water (1 μg L 1). However, the antibody fragment only cross-reacted with three microcystin variants (MC-RR, -LW, -LF and nodularin). The incorporation of a highly engineered antibody fragment provides a system with greater detection capabilities and greater assay specificity.
Materials and methods
Host immunization and antibody library generation Immune library generation was carried out according to Andris-Widhopf et al. (2000). MC-LR-BSA (500 μg) was mixed in a 1:1 ratio with Freund's complete adjuvant (FCA) in a 1 mL final volume. The emulsified mixture was administered sub-cutaneously over four sites in 250 mL injection volumes. Subsequent booster immunizations were carried out in the same manner using Freund's incomplete adjuvant (FIA) over a twelve-week period. RNA isolated from the spleen and bone marrow was reverse transcribed to complementary DNA (Superscript III kit, Invitrogen). Antibody variable heavy (VH) and variable light (VL) chain genes were amplified from the cDNA. The scFv gene products were amplified using splice by overlap extension (SOE) PCR, and cloned into the pComb3XSS vector using restriction enzyme SfiI (New England Biolabs). The antibody library was transformed into electro-competent Escherichia coli XL1-Blue cells (Agilent Technologies) by electroporation. The library titre was approximately 3.5 107 CFU mL 1. Phage selection and screening Immune library selection was carried out over five rounds using M13K07 helper phage (New England Biolabs). Phage were incubated and enriched by decreasing the concentration of MC-LROVA (100–5 μg mL 1) coated on immunotubes (Nalgenes). Further information on the selection of the most sensitive anti-microcystin-LR clones is available in supplementary information. The 10 best clones, with minimal cross-reactivity to BSA, OVA and KLH (2 μg mL 1 coated direct ELISA) were selected. Concomitantly, DNA sequencing analysis was carried out for identification of unique sequences (Eurofins). Antibody expression and purification by immobilized metal affinity and size exclusion chromatography The presence of a polyhistidine tag on the recombinant antibodies facilitated purification by immobilized metal affinity chromatography (IMAC). For more information please see the accompanying supplementary documentation. Size exclusion chromatography (SEC) was carried out using an AKTA FPLC (GE Healthcare) utilizing UNICORN software. A HiLoad Superdex 200 16/60 column (GE Healthcare) was equilibrated with 3 volumes of PBS. A 1 mL sample of IMAC-purified scFv was filtered through a 0.2 μm filter and injected into a 2 mL loop. The flow rate was fixed at 1 mL min 1. Fractions of interest were collected in 5 mL volumes and were concentrated to 500 μL volumes using 10,000 MW cutoff vivaspin columns. The purity of each fraction was analysed by SDS-PAGE, the fractions were combined and the concentration was established using a Nanodrop NC-1000 system and a modified Lowry protein quantification assay (Thermo Scientific).
Preparation of microcystin-LR conjugates Microcystin-LR (2 mg) derivatisation was carried out by Enzo Life Sciences (UK), as described by Moorhead et al. (1994). The reaction mixture was analysed by reverse-phase high-performance liquid chromatography (RP-HPLC) and was found to contain 495% derivatised aminoethanethiol-microcystin-LR (1 mg) which was subsequently conjugated to bovine serum albumin (BSA) (3 mg) and ovalbumin (OVA) (3 mg) using a glutaraldehyde coupling procedure (Metcalf et al., 2000). The two bespoke conjugates were dialysed against 0.9% (w/v) saline. The resulting microcystin-LR conjugates were used for subsequent immune library generation, selection and screening.
Assessment of matrix interferences in competitive MC-LR determination assay Three different lake water samples from three different sites were taken from two different lakes in Ireland, Lough Ennell and Lough Derg, and were compared to deionized water, tap water and PBS to examine the effect of the matrix on microcystin determination. The water samples were filtered through a 0.4 μm filter to remove any large debris. The analytical performance of the antiMC-LR recombinant antibody was assessed using a competitive ELISA (please see Supplementary information for further experimental detail).
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
MC-LR (Enzo Life Sciences) was directly conjugated to Biacore CM5 sensor chips (GE Healthcare) according to Vinogradova et al. (2011). The SPR-based Biacore inhibition assay was developed using the purified 2G1 recombinant anti-microcystin-LR scFv (optimized at 1/7000 in HBS buffer). The scFv was mixed with a range of concentrations of microcystin-LR, -YR -LA, -LF, -LW, -RR, and nodularin (0.69–300 ng mL 1). The assay was carried out at a flow rate of 5 μL min 1 with 2 min injections (n ¼ 3) followed by surface regeneration (two pulses of 200 mM NaOH for 1 min). The IC50 values for each toxin were directly compared to that of MC-LR to determine the degree of cross reactivity. Inter-day and intra-day assay studies were performed to verify the accuracy and precision. The standard curve was generated (n ¼3) on a single day (intra-day) and on three consecutive days (inter-day) and the mean and standard error were calculated. The %CV were determined for each concentration and calibration curves were generated from the data with a four-parameter equation, using BIAevaluation™ 3.0 software. Site-directed mutagenesis of 2G1 antibody fragment using molecular docking approach Maximum Entropy based Docking (MEDock) (Chang et al., 2005) predicted a structurally important amino acid change from arginine to serine at position 66 in the framework region in the heavy chain of the scFv may impact on antibody:toxin binding. Quikchange II XL site directed mutagenesis kit from Stratagene was used to mutate the amino acid using ‘in-house’ designed primers (synthesized by Eurofins Genomics). The novel antibody against microcystin 2G1-R66S was used in MBio SnapESIs opticalplanar waveguide immunoassay studies.
Microarrays were then assembled into an injection-moulded cartridge, which contained a 5 mm wide fluidic channel (max volume 30 mL) with a single inlet port for the addition of sample and reagents.
Results Isolation of anti-microcystin scFv fragments from an immunized phage library Recombinant antibody fragments targeting MC-LR were developed from an immunized phage library. The enrichment of microcystin-specific phage binders over five rounds of biopanning was confirmed by polyclonal phage ELISA (Fig. 1A) (for more information see Supplementary information). Following infection of phage outputs from round 4 and 5 into E. coli Top10F’, preliminary screening was carried out using a direct monoclonal ELISA for identification of MC-LR-specific clones (Fig. 1B). All scFv fragments displayed binding to the MC-LR-OVA
A Absorbance (450 nm)
Development of a surface plasmon resonance (SPR)-based inhibition assay
3
2.5 2.0 1.5 1.0 0.5
Biotinylation of 2G1-R66S variant
Microarray printing and cartridge development The generation of portable optical-planar waveguide cartridges for the determination of microcystin was carried out in conjunction with MBio Diagnostics, Colorado, USA. Briefly, the planar waveguide substrate is an injection moulded plastic component activated with a proprietary silane-based surface chemistry. Biomolecule binding to the waveguide surface is through a combination of covalent reactions (e.g. via surface lysines on the printed proteins) and non-specific interactions (e.g., hydrophobic, electrostatic, van der Waals). MC-LR-conjugates were printed using a Bio-Dot AD3200 robotic arrayer onto microarrays at varying concentrations (microcystin-LR-OVA at 10 and 5 μg mL 1). Spots were produced, with a diameter of 0.5 mm using a Bio-Jet print head that dispensed 20 nL. Two replicates of the toxin-conjugates were printed onto a grid with 1 mm centres, blocked with a proteinbased blocking agent (0.5% (w/v) casein in PBS (150 mM, pH 7.4) with Proclin300 antimicrobial agent) prior to spin-drying.
0.0
B
Prepan Pan 1 Pan 2 Pan 3 Pan 4 MC-LR-OVA OVA
Pan 5
1.6 1.4
Absorbance (450 nm)
ScFv 2G1-R66S used in the development of the MBio SnapESI immunoassay was purified as described. 2G1-R66S (anti-MC-LR) was biotinylated using 100-fold molar excess of sulfo-NHS-LCBiotin (Pierce). The 2G1-R66S anti-MCLR-biotin mixture was buffer exchanged (PBS 150 mM, pH 7.4) to remove unreacted biotin. Using a biotin quantification kit (Pierce), a conjugation ratio of 1:1.7 mmol 2G1-R66S to biotin was determined. StreptavidinAlexa-647 (Molecular Probes, Invitrogen) was used to detect the presence of biotinylated-anti-MC-LR 2G1-R66S in the MBio SnapESIs MC-LR detection assay.
1.2 1.0 0.8 0.6 0.4 0.2 0.0
Clones 1-96
Fig. 1. Polyclonal phage ELISA and soluble scFv monoclonal ELISA: (A) polyclonal phage ELISA for the assessment of anti-microcystin-LR phage-scFv binding over five rounds of bio-panning. The experiments were performed in triplicate and error bars represent the standard error of the mean. (B) Direct binding to MC-LR-OVA was observed using monoclonal, soluble ELISA, using an anti-HA-HRP conjugate as the secondary detection reagent. The background binding (assessed as three times the blank measurement) is indicated by a dashed red line. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
conjugate above the background signal. No cross-reactivity to protein carriers including OVA, BSA or KLH was observed. The positively identified candidate clones were analyzed by inhibition ELISA (data not shown). The best scFv, 2G1, was expressed and purified by immobilized metal affinity and size exclusion chromatography, as described.
Development of biacore inhibition immunoassay and assessment of cross-reactivity profile A Biacore inhibition assay for the detection of MC-LR was developed using the specific scFv fragment 2G1. The 2G1 scFv displayed an IC50 of 4.8 ng mL 1. The antibody response values (R) for each concentration were normalized by the transformation to % R/R0, according to the equation: %R/R0 ¼(R Rexcess)/(R0 Rexcess), where R is the relative response, R0 is the response at a zero concentration of analyte, and Rexcess is the response at an excess of analyte. The %R/R0 value was plotted against the logarithm of the MC-LR concentration (ng mL 1). The inter-day calibration curve for the 2G1 scFv fragment is presented in Fig. 2A. The limit of detection was calculated from the mean of the zero analyte concentration, minus three standard deviations, as 1.7 ng mL 1. To determine the degree of cross-reactivity of the scFv fragment 2G1 to the six structurally-related variants of microcystin, MC-LW, MCRR, MC-YR, MC-LA, MC-LF and nodularin, an inhibition analysis was performed. Percentage cross-reactivities (%CR50) were extrapolated by comparing the relevant IC50 of the microcystin variants to the IC50 of the MC-LR toxin. The %CR50 was calculated using the equation: %cross reactivity 50 (CR50)¼ (IC50 MC variant/IC50 MCLR) 100. The cross-reactivity profiles for the 2G1 scFv clone are illustrated in Fig. 2B. The degree of cross-reactivity of the 2G1 scFv to the MC variants are comparable with MC-LR, with MC-YR and MC-LA showing the highest cross-reactivity relative to MC-LR.
A 100 % R/Ro
80 60 40 20 0 0.1
1
10
100
Microcystin-LR (ng mL-1)
B Variant MC-LR MC-YR MC-LA Nodularin MC-LF MC-LW MC-RR
IC50 (ng mL-1) 4.80 8.06 6.42 9.28 7.11 10.08 7.99
% CR 100.00 79.73 74.75 69.25 67.49 63.75 60.07
Fig. 2. Biacore inhibition assay curve and microcystin cross reactivity profiles. (A) Biacore inhibition assay curve for the determination of MC-LR in PBS, for the 2G1 (wild-type) scFv fragment. MC-LR concentrations, ranging from 40 to 0.69 ng mL 1 were pre-incubated with 2G1 anti-MC-LR scFv (1/7000) and injected over the MC-LR-immobilized surface. The error bars represent the standard deviation for each concentration. (B) Cross-reactivity profile of MC toxins with the 2G1 scFv fragment in HBS-EP buffer.
Identification of important structural scFv regions Increased MC-LR binding affinity was achieved by identifying important structural aspects of the scFv-2G1 polypeptide. To evaluate possible molecular docking permutations between MC-LR and anti-microcystin scFv, MEDock was used. Fig. 3 displays ligand (MC-LR) docked to 2G1-scFv. The programme predicted a number of possible amino acid mutations to increase antibody/toxin affinity (http://medock.csie.ntu.edu.tw/ and http://bioinfo.mc.ntu. edu.tw/medock/) (Chang et al., 2005). An increase in antibody affinity for toxin microcystin-LR was determined by inhibition ELISA (data not shown). The novel antibody against microcystin 2G1R66S was used in subsequent MBio SnapESIs optical-planar waveguide immunoassay validation studies. Testing the performance of 2G1-R66S in various sample matrices To assess the influence of sample matrix on the determination of MC-LR in freshwater water compared to PBS, tap water and deionized water a competitive ELISA was carried out (Fig. 4). Lake water from three different locations in Ireland (Lough Ennell, and two different locations in Lough Derg (Dromineer and Luska) were assessed. The ability of the antibody to bind to free toxin in the presence of the different matrices (A) was compared to the binding of each antibody to toxin conjugate in the absence of free toxin (A0). The results were normalized by subtracting the absorbance value at excess analyte (Aexcess). The resultant normalized absorbance (A Aexcess/A0 Aexcess) was plotted as A/A0 versus MC-LR concentration (ng mL 1) as shown in Fig. 4. The experiment was repeated twice and each sample determination was carried out in triplicate. The IC50 values for each sample were obtained (IC50 values: PBS, 5.0 ng mL 1; tap water, 3.8 ng mL 1; deionized H2O, 6.7 ng mL 1; L. Derg-Dromineer, 5.7 ng mL 1; L. Derg -Luska, 4.4 ng mL 1 and L. Ennell, 4.4 ng mL 1. Development of optical-planar waveguide assay for the determination of MC-LR in cyanobacterial cultures using recombinant antibody 2G1-R66S Assay format The ability of free microcystin in solution to inhibit the binding of antibody to immobilized microcystin-conjugate forms the basis of the assay. Different conjugate coating concentrations were tested, including MC-LR-BSA at 200, 100 and 50 μg mL 1 and MCLR-OVA at 100, 50, 10 and 5 μg mL 1 and it was determined that MC-LR-OVA at 10 μg mL 1 achieved the most reproducible and sensitive results. Antibody stock was stored at 20 °C in 50 μL aliquots of a concentration 1 mg mL 1 in PBS (150 mM, pH 7.4). A range of antibody dilutions were tested i.e. 1:102, 1:103, 1:104 and 1:105, it was found that scFv 2G1-R66S worked optimally on the MBio SnapESIs system at 1 in 104 dilution. The format of the assay relies on the addition of Streptavidin-Alexa-647s to identify the presence of biotinylated anti-microcystin scFv. All steps were carried out at room temperature and are as follows:
Addition of antibody: toxin sample to cartridge – 10 min incubation at room temperature.
1 Wash in PBS for 2 min. Detection molecule (streptavidin-Alexa-647) addition – 5 min incubation at room temperature.
1 Wash in PBS 2 min. Insert cartridge into Mbio system and obtain fluorescent reading.
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Fig. 3. Molecular docking model of microcystin-LR and scFv-2G1. The ligand (MC-LR) is displayed in ‘ball and stick’ mode and is represented in red, the white area shows the binding cavity, the scFv light and heavy chain are displayed in green and yellow respectively, and the glycine-serine linker is displayed in blue (scFv is displayed in ‘space filling’ mode). The structure was visualized using ICM-Browser tool. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
1.0
PBS Deionized H20 Tap water L. Derg Luska L. Derg Dromineer L. Ennell
A/A0
0.8 0.6 0.4 0.2 0.0
0.5
1
2
4
8
Microcystin-LR
16
32
64
(ng mL-1)
Fig. 4. Competitive ELISA comparing the effects of sample matrices on the determination of MC-LR using recombinant scFv 2G1-R66S. PBS, deionized H2O and tap water were compared to three fresh water Irish lake samples (Lough Derg Luska, Lough Derg Dromineer and Lough Ennell). The results were normalized by subtracting the absorbance value at excess analyte (Aexcess). The resultant normalized absorbances (A Aexcess/A0 Aexcess) were plotted as A/A0 versus MC-LR concentration (ng mL 1).
Assay validation Outline of assay format and generation of standard curve. To assess assay reproducibility, analytical standards were used to generate a standard curve and the ability of anti-MC-LR 2G1-R66S-biotin (1:104 diluted in 1% BSA-PBS (150 mM, pH 7.4) (w/v)) to accurately detect the standards was assessed. MC-LR standards were prepared in M. aeruginosa sample matrix (CCAP 1450/03) (described by Devlin et al.) to account for matrix effects. The calibration curve contained toxin concentrations from 25 ng mL 1 to 48.8 pg mL 1 (at final working concentrations). Anti-MC-LR 2G1-R66S-biotin was incubated with the calibrants for 5 minutes at room temperature prior to testing. Each MC-LR-calibrant and anti-MC-LR mixture was applied to the MBio SnapESIs cartridge at a volume of 175 μL and incubated for at room temperature (10 min). Filter sterilized PBS (175 μL, 150 mM, pH 7.4) was added to the well (2 min). Streptavidin-alexa-647 (Molecular Probes, Invitrogen) diluted in PBS (150 mM, pH 7.4) (1:2,000, 175 μL) was added to the cartridge (5 min). The cartridge was washed with 175 μL filter sterile PBS (2 min) and the signal was measured using MBio SnapESIs system and a calibration curve was generated.
Determination of limit of detection (LOD). The LOD was established by measuring the limit of blank plus three standard deviations (Armbruster and Pry, 2008). A calibration curve was generated using inter-assay fluorescence results generated over three days using scFv 2G1-R66S-biotin diluted in sample matrix. The mean fluorescence value of twenty ‘microcystin-free’ matrix samples plus the value of three standard deviations was determined and interpolated onto the curve. An analytical LOD of 0.28 ng mL 1 was achieved Table 1. The IC50 value and the dynamic range (IC10 to IC90) were established using two batches of cartridges on three separate days. The values are displayed in Table 1. The assay displayed an IC50 value of 1.1 ng mL 1 and a dynamic range of 0.21–5.94 ng mL 1. To establish the detection capability of the assay (CCβ) (also known as the functional limit of detection), low concentrations of MC-LR (0.1 ng mL 1, 0.19 ng mL 1 and 0.5 ng mL 1) were assayed measuring 20 toxin samples at each concentration and 20 blank samples with all tests carried out in the complex sample matrix. Samples containing 0.19 and 0.5 ng mL 1 (analysed by un-paired two-tailed t-test) were found to be significantly different, each displaying a p value o 0.0001. Therefore, as the lowest value, 0.19 ng mL 1 was identified as the functional limit of detection (CCβ) (Fig. 5). Repeatability studies were carried out over three days using four different batches of cartridges. Matrix samples were spiked with 3.3, 1.19 and 0.5 ng mL 1 (IC15, IC50 and half the WHO recommended detection limit, respectively). The results are shown in Table 2. Data were fit using a four-parameter non-linear regression calibration curve. Inter-assay results displayed percentage accuracies between 88 and 99%, with intra-day assay displaying results between 82 and 103%. The percentage CV was determined using equation CV%¼ (SD/Mean) 100. All percentage coefficients of variance were below 8% for intra-day assays and below 11% for between day assays. The assay was assessed for its ability to detect MC-LR in lake water (Lough Derg, Ireland) and deionized water (diH2O). Two of Table 1 Analytical limit of detection, IC50 values and dynamic range of assay.
10 μg mL 1 MC-LROVA-coated assay spot
Analytical LOD (ng mL 1)
IC50 (ng mL 1)
Dynamic range (IC10 to IC90, ng mL 1)
0.28
1.11
0.21–5.94
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
Fig. 5. Functional LOD determination (CCβ) of microcystin specific scFv 2G1-R66S. The dotted line represents the point at which over 95% of spiked samples are distinguishable over blank samples and displayed p values o0.0001. Table 2 MBio SnapESI repeatability measurements.
Intra-day Day 1
Day 2
Day 3
Inter-day Days 1–3
Spiking level (ng mL 1)
Back calculated concentration (ng mL 1)
3.30 1.19 0.50 3.30 1.19 0.50 3.30 1.19 0.50
3.25 1.21 0.46 3.37 1.14 0.41 3.16 1.08 0.45
4.9 5.1 7.8 6.3 4.8 4.5 5.1 5.2 5.8
98.5 101.7 92.0 102.1 95.8 82.0 95.8 90.8 90.0
3.30
3.26
7.35
98.8
1.19 0.50
1.14 0.44
10.44 1.40
95.8 88.0
CV (%) Specific accuracy (%)
the samples were supplemented with M. aerguinosa extract (preparation previously described in paper). The presence of M. aerguinosa extract or lake water contaminants had minimal impact on the assay performance, and diH2O, M. aerguinosa in diH2O and M. aerguinosa in lake water displaying IC50 values of 1.62, 1.66 and 1.57 ng mL 1, respectively (see Supplementary information).
Discussion The planar waveguide immunoassay presented is of exquisite sensitivity, and exploits the in-built ability to genetically alter recombinant antibodies to increase their affinity for target antigen, thus enhancing their capacity in every-day immunoassays. Recombinant antibodies have been developed previously to microcystin. However, none have been developed using an immune host. In the present system, the single chain fragment variable (scFv) molecule recognizes the presence of microcystin-LR and its congeners displaying cross-reactivites to a range of MC-congeners (MC-YR 79%, MC-LA 74.8%, MC-LR 67.5%, MC-LW 63.7%, 60.1%, Nod 69.3%). The ability of 2G1 scFv to cross-react with other members of the MC family increases the utility value of the scFv in immunoassays, facilitating accurate toxicity determination in an environmental, complex sample which could potentially contain a
number of microcystin congeners. The ability of the antibody to determine MC-LR in the presence of various sample matrices was assessed by competitive ELISA. Different samples of fresh water were spiked with MC-LR and the resultant IC50 values displayed the relatively low influence that sample matrix has on the ability of the recombinant antibody to measure MC-LR. To further enhance the binding avidity of scFv-2G1 for MC-LR, docking permutations between scFv 2G1 and MC-LR were evaluated using a molecular docking simulation. The MEDock server predicted a possible ligand-binding site at position 66 of the heavy chain that could be mutated to increase affinity. This amino acid was also predicted by the (CASTp, 2011) server to play a role in MC-LR binding, and interestingly is highly conserved as a serine residue in human antibodies (ClustalW analysis). Using the genetically altered and biotinylated scFv (2G1-R66Sbiotin) an evanescent optical-planar waveguide immunoassay using the MBio SnapESIs system was developed for the detection of microcystin and a conservative analytical LOD of 0.28 ng mL 1 was determined. The functional LOD (CCβ) of 0.19 ng mL 1 toxin in matrix sample was identified (Fig. 3). These findings improve upon a similar study for the detection of microcystin (Devlin et al., 2013), by lowering the limit of detection 5 times from 1 ng mL 1 to 0.19 ng mL 1. It is important to note that using this system, the presence of toxin can be detected following a simple mechanical toxin extraction procedure that only takes 10 min (Devlin et al., 2013). The presence of M. aeruginosa cellular debris impacts on the sensitivity of assays; however, in the presence of cellular debris this system can still detect the presence of toxin below the WHO recommended limit. Intra-day analyses generated data that displayed percentage accuracies between 82 to 102.2% (Table 2) while the inter-day study returned values of between 88 to 99%, thus displaying better values than previous reports. The resultant percentage CVs produced very tight (low) values with the highest recorded CV at 7.8%. Intermediate repeatability was quantified by inter-day studies, whereupon the highest exhibited % CV value was 10.44. The results thus highlight the high precision afforded to the detection of MCLR by the use of this assay.
Conclusion This is the first report, to our knowledge, of the generation of a recombinant antibody fragment against MC-LR, from an immunized avian host. The antibody fragment displayed cross-reactivity values towards six MC congeners in an SPR-based immunoassay. The antibody fragment was subjected to structural scrutiny and an amino acid at position 66 in the variable heavy chain was genetically modified from arginine to serine. Subsequently, a recombinant antibody-based evanescent optical-planar waveguide system capable of reliably detecting the presence of MC-LR was developed which displayed sensitivities comparable to those achievable with reverse phase HPLC (Pyo et al., 2005; WHO, 2008). The inhibition assay has a complete run time of 19 min and can detect the presence of microcystin with a functional LOD of 0.19 ng mL 1 in complex sample matrix, a five times greater sensitivity over previous reports.
Acknowledgements This work was supported by Science Foundation Ireland (SFI) as part of a US-Ireland research and development partnership programme BEACONS (Biosafety for EnvironmentAL Contaminants
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
C. Murphy et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
using Novel Sensors) Grant code 08/US/I1512 and the Beaufort Marine Research Award under the Sea Change Strategy and the Strategy for Science Technology and Innovation (2006–2013) and with the support of the Marine Institute, funded under the Marine Research sub-programme of the National Development plan 2007–2013.
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at. http://dx.doi.org/10.1016/j.bios.2014.10.039
References Andris-Widhopf, J., Rader, C., Steinberger, P., Fuller, R., Barbas 3rd, C.F., 2000. Methods for the generation of chicken monoclonal antibody fragments by phage display. J. Immunol. Methods 242, 159–181. Armbruster, D.A., Pry, T., 2008. Limit of blank, limit of detection and limit of quantitation. The Clinical biochemist. Rev. Aust. Assoc. Clin. Biochem. 29 (Suppl. 1), S49–S52. Bhattacharya, R., Sugendran, K., Dangi, R.S., Rao, P.V., 1997. Toxicity evaluation of freshwater cyanobacterium microcystis aeruginosa PCC 7806: II. Nephrotoxicity in rats. Biomed. Environ. Sci. 10 (1), 93–101. CASTp, 2011. 〈http://sts-fw.bioengr.uic.edu/castp/calculation.php〉. Chang, D.T., Oyang, Y.J., Lin, J.H., 2005. MEDock: a web server for efficient prediction of ligand binding sites based on a novel optimization algorithm. Nucl. Acids Res. 33, W233–238. Cheung, M.Y., Liang, S., Lee, J., 2013. Toxin-producing cyanobacteria in freshwater: a review of the problems, impact on drinking water safety, and efforts for protecting public health. J. Microbiol. 51 (1), 1–10. Conroy, P.J., Hearty, S., Leonard, P., O'Kennedy, R.J., 2009. Antibody production, design and use for biosensor-based applications. Semin. Cell Dev. Biol. 20 (1), 10–26. Devlin, S., Meneely, J.P., Greer, B., Greef, C., Lochhead, M.J., Elliott, C.T., 2013. Next generation planar waveguide detection of microcystins in freshwater and cyanobacterial extracts, utilising a novel lysis method for portable sample preparation and analysis. Anal. Chim. Acta 769, 108–113. Hernandez, J.M., Lopez-Rodas, V., Costas, E., 2009. Microcystins from tap water could be a risk factor for liver and colorectal cancer: a risk intensified by global change. Med. Hypotheses 72 (5), 539–540. Herranz, S., Marazuela, M.D., Moreno-Bondi, M.C., 2012. Automated portable array biosensor for multisample microcystin analysis in freshwater samples. Biosens. Bioelectron. 33 (1), 50–55.
7
Humpage, A., 2008. Toxin types, toxicokinetics and toxicodynamics. Adv. Exp. Med. Biol. 619, 383–415. Lochhead, M.J., Todorof, K., Delaney, M., Ives, J.T., Greef, C., Moll, K., Rowley, K., Vogel, K., Myatt, C., Zhang, X.Q., Logan, C., Benson, C., Reed, S., Schooley, R.T., 2011. Rapid multiplexed immunoassay for simultaneous serodiagnosis of HIV-1 and coinfections. J. Clin. Microbiol. 49 (10), 3584–3590. Long, F., He, M., Zhu, A.N., Shi, H.C., 2009. Portable optical immunosensor for highly sensitive detection of microcystin-LR in water samples. Biosens. Bioelectron. 24 (8), 2346–2351. MacKintosh, C., Beattie, K.A., Klumpp, S., Cohen, P., Codd, G.A., 1990. Cyanobacterial microcystin-LR is a potent and specific inhibitor of protein phosphatases 1 and 2 A from both mammals and higher plants. FEBS Lett. 264 (2), 187–192. McElhiney, J., Drever, M., Lawton, L.A., Porter, A.J., 2002. Rapid isolation of a singlechain antibody against the cyanobacterial toxin microcystin-LR by phage display and its use in the immunoaffinity concentration of microcystins from water. Appl. Environ. Microbiol. 68 (11), 5288–5295. McElhiney, J., Lawton, L.A., Porter, A.J., 2000. Detection and quantification of microcystins (cyanobacterial hepatotoxins) with recombinant antibody fragments isolated from a naive human phage display library. FEMS Microbiol. Lett. 193 (1), 83–88. Metcalf, J.S., Bell, S.G., Codd, G.A., 2000. Production of novel polyclonal antibodies against the cyanobacterial toxin microcystin-LR and their application for the detection and quantification of microcystins and nodularin. Water Res. 34 (10), 2761–2769. Moorhead, G., MacKintosh, R.W., Morrice, N., Gallagher, T., MacKintosh, C., 1994. Purification of type 1 protein (serine/threonine) phosphatases by microcystinsepharose affinity chromatography. FEBS Lett. 356 (1), 46–50. Pyo, D., Park, G., Choi, J., Oh, C., 2005. Microcystin detection characteristics of fluorescence immunochromatography and high performance liquid chromatography. Bull. Korean Chem. Soc. 26 (2), 268–272. Qin, B., Zhu, G., Gao, G., Zhang, Y., Li, W., Paerl, H.W., Carmichael, W.W., 2010. A Drinking Water Crisis in Lake Taihu, China: linkage to climatic variability and lake management. Environm. Manag. 45 (1), 105–112. Turner, P.C., Gammie, A.J., Hollinrake, K., Codd, G.A., 1990. Pneumonia associated with contact with cyanobacteria. Br. Med. J. 300 (6737), 1440–1441. Vareli, K., Jaeger, W., Touka, A., Frillingos, S., Briasoulis, E., Sainis, I., 2013. Hepatotoxic seafood poisoning (HSP) due to microcystins: a threat from the ocean? Mar. Drugs 11 (8), 2751–2768. Vinogradova, T., Danaher, M., Baxter, A., Moloney, M., Victory, D., Haughey, S.A., 2011. Rapid surface plasmon resonance immunobiosensor assay for microcystin toxins in blue-green algae food supplements. Talanta 84 (3), 638–643. WHO, 1998. Guidelines for Drinking Water Quality, 2nd ed. World Health Organisation, Geneva, pp. 95–110. WHO, 2008. Guidelines for Drinking Water Quality (incorporating the first and second addenda). Third Edition World Health Organization, Geneva, Switzerland, ISBN: 978 924 154 761 1 (web version).
Please cite this article as: Murphy, C., et al., Biosensors and Bioelectronics (2014), http://dx.doi.org/10.1016/j.bios.2014.10.039i
