First development and characterisation of polyclonal and monoclonal antibodies to the emerging fresh water toxin cylindrospermopsin

Harmful Algae 24 (2013) 10–19

Contents lists available at SciVerse ScienceDirect

Harmful Algae journal homepage: www.elsevier.com/locate/hal

First development and characterisation of polyclonal and monoclonal antibodies to the emerging fresh water toxin cylindrospermopsin Christopher T. Elliott a,*, Clare H. Redshaw b,c, Suja E. George a, Katrina Campbell a a

Institute for Global Food Security, School of Biological Sciences, Queen’s University, Belfast BT9 5AG, UK European Centre for Environment and Human Health, University of Exeter Medical School, Royal Cornwall Hospital, Truro, Cornwall TR1 3HD, UK c School of Geography, Earth and Environmental Science, University of Plymouth, Drake Circus, Plymouth, Devon PL4 8AA, UK b

A R T I C L E I N F O

A B S T R A C T

Article history: Received 10 July 2012 Received in revised form 30 December 2012 Accepted 30 December 2012

As increasing incidences in the occurrence of cylindrospermopsin (CYN) appear, in addition to further research on its toxicological nature, improved rapid methods to detect this toxin are required. Antibody based assays are renowned for their ability to provide rapid, portable, simple to use tests. As yet however there are no publications outlining how an antibody to CYN can be produced. A range of chemical approaches was investigated to synthesise CYN immunogens for antibody production but failed to generate a response. Finally, a modified Mannich reaction for immunogen synthesis was employed to couple the toxin to two carrier proteins. Both protein conjugates were successfully used to raise both polyclonal and monoclonal antibodies of high sensitivity to CYN. These antibodies were characterised employing competitive indirect ELISA and an optical biosensor assay. By ELISA the sensitivity achieved ranged from 27 to 131 pg/mL and by SPR 4.4 to 11.1 ng/mL thus demonstrating that the selection of immunoassay platform is important for the detection level required by the end user for their application. Low cross-reactivity to the much less toxic metabolite deoxyCYN was observed. This is the first reported production of antibodies to this toxin. ß 2013 Elsevier B.V. All rights reserved.

Keywords: Cylindrospermopsin Fresh water toxin Antibody Immunoassay

1. Introduction The appearance of a number of previously unreported marine biotoxins, such as tetrodotoxin, palytoxin and cyclic imines, within European waters, has led to concerns regarding risks to human health (Paredes et al., 2011), and is evidenced by the formation of a number of European Food Safety Authority opinion reports on these emerging toxins (EFSA, 2008, 2009, 2010). Similar emergence of cyanobacterial produced toxins has also been observed in temperate freshwater systems with the appearance of anatoxin, and their analogues (Cadel-Six et al., 2007; Codd et al., 2005; Rawn et al., 2007). This has resulted in concerns over both human health and environmental consequences due to contamination of drinking water, food stuffs and bioaccumulation in exposed aquatic organisms (Kinnear, 2010). The presence of these emerging toxins and the increased occurrence of blooms of established cyanobacteria (such as the microcystin producer Microcystis spp.) is largely attributed to increased water temperature due to climate change and the degradation of water quality; particularly eutrophication due to high agricultural fertiliser usage (Hudnell et al., 2008; Paerl and Huisman, 2008).

* Corresponding author. E-mail address: [email protected] (C.T. Elliott). 1568-9883/$ – see front matter ß 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.hal.2012.12.005

The best known cyanobacteria responsible for producing cylindrospermopsin (CYN) is Cylindrospermopsis raciborskii, although in recent years a number of other genera including Aphanizomenon, Anabaena, Raphidiopsis, Umezakia and Lyngbia have emerged as CYN producers in a number of geographical locations across Europe, the Americas, E and SE Asia, West Africa, the Middle East and Australasia (Banker et al., 2000; Bernard et al., 2003; Berry and Lind, 2010; Blahova et al., 2009; Brient et al., 2009; Carson, 2000; Cheng et al., 2009; Li et al., 2001; Messineo et al., 2009; Preussel et al., 2009). The structure and relative stereochemistry of CYN (C15H21N5O7S; molecular weight 415.43 g/mole; Fig. 1) was first elucidated by Ohtani and Moore (1992) as a highly water-soluble alkaloid with a zwitterionic tricyclic guanidine moiety and sulphate ester functionality, bridged with a 5-substituted-2,4dioxypyrimidine (uracil) moiety. 7-epi-CYN (Fig. 1), a diasteriomer of CYN, has been identified by Banker et al. (2000) and has a similar toxic potency in mouse bioassays (Banker et al., 2000, 2001; Griffiths and Saker, 2003). The toxicity of the slightly less polar deoxy-CYN (C15H21N5O6S; molecular weight 399; Fig. 1), which is produced as a major analogue by some cyanobacterial spp., is under debate as it is non-toxic by mouse bioassay, but is a potent protein synthesis inhibitor and can inhibit glutathione synthesis (Griffiths and Saker, 2003; Li et al., 2001; Looper et al., 2005; Norris et al., 1999).

C.T. Elliott et al. / Harmful Algae 24 (2013) 10–19

B

A

D

11

A

C

D

C

Cylindrospermopsin (CYN), favored enol tautomer

B

A

B

D

C

7-epi-Cylindrospermopsin (7-epi-CYN)

Cylindrospermopsin (CYN), keto tautomer

A

B

D

C

Deoxy-Cylindrospermopsin (deoxy-CYN)

Fig. 1. Cylindrospermopsin (CYN) and its natural cyanobacterial produced analogues 7-epi-CYN and deoxy-CYN. The positive charge is shared between the 3 nitrogen atoms, and is not upon the carbon atom as illustrated.

The first documented large scale human CYN poisoning event occurred in 1979 in Australia, although the aetiology of this event was not elucidated until several years later, thereby earning CYN poisoning the colloquial name of ‘‘Palm Island Mystery Disease’’ (Griffiths and Saker, 2003). High environmental CYN concentrations have been reported in farm dams and highly eutrophic waters (up to 800 mg/L; Griffiths and Saker, 2003). Exposure to these potent hepatotoxins from submerged summer blooms can have serious health effects ranging from gastroenteritis to pathological impacts upon the intestine, spleen, thymus, kidneys, liver and heart (Carson, 2000; Falconer et al., 1999; Griffiths and Saker, 2003). The primary mechanism of toxicity is by no means fully understood, however evidence suggests inhibition of protein synthesis (Carson, 2000; Froscio et al., 2001, 2003; Griffiths and Saker, 2003), glutathione synthesis inhibition (Froscio et al., 2003; Griffiths and Saker, 2003; Norris et al., 2002), involvement of cytochrome P450 enzymes in CYN activation/metabolite formation and hepatocytotoxicity, (Froscio et al., 2003; Norris et al., 2001, 2002), inhibition of pyrimidine nucleotide synthesis (Beyer et al., 2009), DNA strand breakage (Griffiths and Saker, 2003) and alteration of ratios of free cholesterol and phospholipids in red blood cell membranes (Reisner et al., 2004), all have a role to pay in the toxic effects observed. Due to the toxic effects the tolerable daily intake for CYN was initially calculated to be 0.02 g/kg body weight/day based on a no observable effect level in mice. It was estimated that guideline values for adults, children, and infants were 0.48, 0.16, and 0.11 g/L, respectively, based on drinking water consumption of 2 L for a 60-kg adult, 1 L for a 10-kg child, and 0.75 L for a 5-kg infant (Duy et al., 2000). The guideline values for the number of cyanobacterial cells (based on 0.026 pg CYN/cell) were 4231 in infants, 6154 in children, and 18,461 in adults (Duy et al., 2000). Concerns over the emergence of this potent toxin have led a number of countries to introduce drinking water guidelines for CYN, which range from 0.3 mg/L (France) to 1 mg/L (Brazil, Australia and New Zealand). Nevertheless, CYN poisoning has been reported in wild and domestic animals more than any other organisms, due to the higher probability of exposure directly from toxic cyanobacterial blooms in lakes and ponds. In reports of toxin exposure in cattle the incidence of mortality and morbidity was higher in calves than in adult cattle (Thomas et al., 1998). This implies that toxicity of CYN in livestock may be greater in young

animals but may be a consideration for livestock (e.g. cattle, goats) whereby it could present in milk. CYN accumulation in plants (Kittler et al., 2012), snails (Berry and Lind, 2010) and finfish (Gallo et al., 2009) has been reported. With the emergence of cyanotoxins, such as CYN, it is imperative that both qualitative and quantitative analytical methodologies are developed for the detection of these toxins in a range of matrices, including water and food, to ensure human and animal wellbeing. A number of advanced analytical techniques, e.g. HPLC–MS/MS have been developed for these purposes (Blahova et al., 2009; Eaglesham et al., 1999; Gallo et al., 2009; Nybom et al., 2008; Oehrle et al., 2010; Rawn et al., 2007), however there is a paucity of rapid, low-cost, high-throughput techniques available, which are essential for monitoring these emerging cyanotoxins. To meet this need, monoclonal and polyclonal antibodies were raised against CYN, which could be utilised in various immunological assays such as enzyme-linked immunoassays (ELISAs) and the optical biosensing technique of surface-plasmon resonance (SPR). The aim of this study was to assess different protein conjugation approaches for the synthesis of immunogens for antibody production and to characterise both antibody types by comparing both traditional ELISA and advanced surface plasmon resonance (SPR) methodologies. 2. Materials and methods 2.1. Chemicals and reagents Bovine thyroglobulin (BTG), ovalbumin (OVA), 1-ethyl-3-(3dimethylaminopropyl) carbodiimide hydrochloride (EDC), Nhydroxysuccinimide (NHS), sodium hypochlorite, succinic anhydride, sodium borohydride, butandiol-diglycidyl-ether, methyl morpholine, isobutylchloroformate, glacial acetic acid, acetic anhydride, perchloric acid, potassium carbonate, N-(t-Boc)-2bromoethylamine (IUPAC: tert-butyl N-(2-bromoethyl)carbamate, ethylenediamine, 2,2-(ethylenedioxy)-bis(ethylamine) (Jeffamine), formaldehyde and all other organic solvents used in the study (Analar grade) were obtained from Sigma–Aldrich Chemical Company (Dorset, UK). The Nunc Maxisorp plates were purchased from Fisher Scientific UK (Leicestershire, UK) and the horseradish peroxidase-linked anti-mouse immunoglobulin from DAKO

12

C.T. Elliott et al. / Harmful Algae 24 (2013) 10–19

(Cambridge, UK). TMB/E solution was obtained from Millipore Limited (Watford, UK). Slide-A-Lyzer dialysis cassettes (10K MWCO, 0.5–3 mL Capacity) and Imject Mariculture KLH (keyhole limpet hemocyanin) were supplied by Medical Supply Company Ltd. (Dublin, Republic of Ireland). PD10 columns were obtained from GE Healthcare, UK and the immunoassay stabiliser was purchased from Source BioScience Life Sciences (Nottingham, UK). Freund’s adjuvant (complete and incomplete) was obtained from Sigma–Aldrich (Bornem, Belgium), Quil-ATM was obtained from Brenntag Biosector A/S (Frederikssund, Denmark) and the adjuvant Pam3Cys-Ser-(Lys)4 (PCSL) was purchased from EMC microcollections (Tu¨bingen, Germany). Horseradish peroxidise (HRP) and mouse monoclonal antibody isotyping strips were obtained from Roche Diagnostics Limited (West Sussex, UK) and (CYN) from n’Tox (France). Ethanolamine, HBS–EP buffer (pH 7.4, 0.01 M HEPES, 0.15 M NaCl, 3 mM EDTA, 0.005% polysorbate), and an amine coupling kit were provided by GE Healthcare Biacore AB (Uppsala, Sweden). 2.2. Preparation of protein conjugates for immunogens 2.2.1. Immunogen 1: CYN–acid–BTG prepared via carbodiimide reaction CYN was derivatised to CYN–acid through the chlorination reaction in a similar manner as described by Banker et al. (2001). In brief, CYN (1.5 mg) was dissolved in 100 mL of deionised water. Sodium hypochlorite (1.1 mL of 1 mg/mL) was added and the mixture was left stirring in the dark for seven days at room temperature. The final mixture was dried and analysed by UPLC– MS (Quattropremier XE) for the presence of the CYN–acid product (see details later). The crude CYN–acid product was then coupled to BTG via the carbodiimide reaction with minor modifications to that reported for dinitrocarbanilide mimics (Connolly et al., 2002). Briefly, the CYN–acid (0.75 mg) was added to deionised water and activated by the addition of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) (1 mg) and N-hydroxysuccinimide (NHS) (0.5 mg) in 2(N-morpholino)ethanesulfonic acid (MES) buffer (0.05 M MES, 0.5 M NaCl, pH 4.7). The activated CYN–acid was added to BTG (2.5 mg) dissolved in phosphate buffered saline (pH 7.2), and the mixture was incubated overnight at room temperature. The BTG conjugate was purified by dialysis against 0.15 M saline solution. 2.2.2. Immunogen 2: CYN–acid to BTG prepared via acid anhydride reaction The CYN–acid derivative, prepared as described in Section 2.2.1, was coupled to BTG via an acid anhydride intermediate as described for steroids (Erlanger et al., 1957), monensin (Crooks et al., 1997) and chloramphenicol succinate (Fodey et al., 2007). CYN–acid (0.44 mg) was dissolved in 0.5 mL anhydrous dimethyl formamide (DMF). Methyl morpholine (20 mL) was added to the mixture which was immediately cooled at 20 8C for 10 min. Isobutylchloroformate (20 mL) was added and the mixture was stirred for 15 min at room temperature. The activated CYN–acid was added slowly to BTG (2.8 mg) dissolved in 1 mM sodium acetate solution and allowed to react for 1 h at 4 8C. The conjugate was purified by dialysis against 0.15 M saline solution. 2.2.3. Immunogen 3: CYN–succinate–BTG via carbodiimide reaction CYN was derivatised to CYN succinate in an analogous approach as presented for 30 -azido-30 -deoxythymidine (Tadayoni et al., 1993). CYN (1 mg) with succinic anhydride (1 mg) dissolved in anhydrous pyridine (1 mL) at room temperature for 48 h stirring in the dark. After the incubation period, pyridine was removed from the final mixture by solvent evaporation. This derivative was conjugated to BTG (2.5 mg) using the carbodiimide method as described in Section 2.2.1.

2.2.4. Immunogen 4: CYN–BTG via an epoxide linker CYN–BTG was prepared using the approach of coupling through an epoxide linker (Hermanson, 1996). Sodium borohydride (10 mg) was dissolved in 5 mL of deionised water that was adjusted to pH 11 with 0.1 M sodium hydroxide. To this 20 mL of butanediol-diglycidyl-ether was added and the epoxide solution was degassed with nitrogen. BTG (3 mg) was dissolved in 500 mL of deionised water and 180 mL of the epoxide solution was added. The mixture was incubated for 40 h under nitrogen at room temperature before purification by PD-10 gel filtration to remove excess epoxide linker. CYN was dissolved in degassed 0.1 M sodium hydroxide (0.5 mL). The hapten solution was added to the epoxide modified BTG under nitrogen and left to incubate in the dark at room temperature for 40 h. Finally, the BTG conjugate was purified by dialysis against saline solution. 2.2.5. Immunogens 5 and 6: CYN–KLH and CYN–OVA via modified Mannich reaction CYN was coupled to KLH and OVA using a modification of the Mannich reaction as described for saxitoxin (Campbell et al., 2007). CYN was added to KLH or OVA (3 mg) dissolved in 500 mL of phosphate buffer. Formaldehyde (13.5 mL) was added and the mixture was stirred in the dark at room temperature. The conjugate was purified by dialysis in 0.15 M saline solution 2.3. Preparation of CYN–HRP conjugates for competitive direct ELISA CYN–HRP conjugates were prepared as described for immunogens 1–6 as outlined in Section 2.2 to be used for the competitive direct ELISA format. 2.4. Coating antigen in competitive indirect ELISA Both CYN–KLH and CYN–OVA protein conjugates (immunogens 7 and 8) were also employed for ELISA coating antigens in the competitive indirect ELISA format. 2.5. Preparation of CYN biosensor chip surfaces The carboxymethylated surface of a CM5 certified grade biosensor chip (Biacore AB) was equilibrated to room temperature and activated using an amine coupling kit. Briefly, EDC and NHS from the kit were mixed (1:1, v/v) and applied to the chip surface for 30 min. Excess solution was removed, and Jeffamine (0.1 M) in borate buffer was added to the chip surface and allowed to react for 1 h. The chip surface was deactivated by exposure to ethanolamine (1 M) for 30 min followed by immobilisation of using three different surface chemistries. Chip 1: The carboxylic acid group of CYN-acid was activated using EDC/NHS and coupled to the free amine on the Jeffamine linker. Chip 2: CYN was also activated using the epoxide chemistry in the same manner as the conjugation to BTG and coupled to an ethylenediamine linker used instead of Jeffamine. Chip 3: CYN was activated using formaldehyde chemistry and coupled to the amine of the Jeffamine linker Each sensor chip surface was then washed with deionized water, dried using a stream of nitrogen gas, and stored desiccated at 4 8C when not in use. 2.6. Immunisation procedures for antibody production 2.6.1. Polyclonal antibody production Two rabbits were immunised with each of the CYN immunogens (1–6) prepared. Prior to these immunisations serum was

C.T. Elliott et al. / Harmful Algae 24 (2013) 10–19

collected from each rabbit to use as a baseline control for titre comparisons with blood samples taken post immunisations. For each rabbit 200 mg of the immunogen preparation (in sterile saline) was added slowly to an equal volume of Freund’s adjuvant with constant vortexing to produce an emulsion. For the initial injection Freund’s complete adjuvant was used while the incomplete form was used for all subsequent immunisations. The emulsion was delivered subcutaneously and divided equally between four injection sites on the animal. An interval of two weeks was used between the first three immunisations, while remaining injections were delivered on a monthly basis. A blood sample (5 mL) was collected ten days after each immunisation for monitoring for the presence of antibodies to CYN. 2.6.2. Monoclonal antibody production Prior to immunisation sera from a blood sample was collected from each mouse to use as a baseline control for titre comparisons with blood samples taken post immunisation. For the first two immunisations of mice, 40 mg of immunogen was mixed with 50 mg of the adjuvant Quil-A in a total volume of 200 mL of sterile saline. This was administered subcutaneously, equally divided between two injection sites, with an interval of two weeks between immunisations. Subsequent immunisations were delivered by the same route on a monthly basis using the same amount of immunogen (40 mg) and 50 mg of the adjuvant Pam3Cys-Ser(Lys)4 (PCSL). A blood sample (0.1 mL) was collected ten days after each immunisation to determine the presence of antibodies to CYN. When a sufficiently high antibody titre was observed a final immunisation was performed intraperitoneally, without adjuvant. Three days later the mouse was sacrificed and the spleen removed for fusion with an SP2 murine myeloma cell line to produce antibody producing hybridomas (Galfre and Milstein, 1981); a blood sample from the heart was also obtained to use as a positive control in the protocol for screening the clones for specific antibodies to CYN. All protocols carried out using mice were performed in accordance with a licence issued by the Department of Health, Social Services and Public Safety in the UK under The animals (scientific procedures) Act 1986. 2.7. Screening procedures for determination of antibody titre Sera samples from all the immunised animals were assessed for their antibody content by performing an ELISA chequerboard titration using both competitive direct and indirect assay formats. Each rabbit serum was tested by direct ELISA using all CYN–HRP conjugates produced. Test-bleed sera was also analysed by SPR using the biosensor chips produced. In theory this allowed for the assessment of the different chemistries employed with the varying orientations of the immobilised CYN on the protein conjugate because of the group reacted. During the monoclonal fusion the hybridomas formed were assessed for specific antibody production by both competitive indirect ELISA and a SPR biosensor (Biacore Q, GE Healthcare) screening assay. 2.7.1. Competitive direct ELISA protocol for antibody screening Serial dilutions (1/1K to 1/8K) of antibody (100 mL/well) in 1 mM sodium acetate buffer were coated onto Falcon plates and incubated overnight at room temperature. All sera from immunogen approaches 1–5 were tested with each CYN–HRP from approaches 1–5. The antibody was removed and the HRP serially diluted in 1% BSA solution were added (50 mL/well) to appropriate wells in a chequerboard design. Negative and positive (1000 ng/ mL) controls (50 mL/well) were incorporated into the assay to determine binding and inhibition. The CYN–HRP and standards were discarded and the plates washed with wash buffer. TMB/E

13

(100 mL/well) was added to each well and the reaction stopped after 12 min using 2.5 mM H2SO4 (25 mL/well). The plates were read at 450 nm using a TECAN Safire2. 2.7.2. Competitive indirect ELISA protocol for antibody screening Nunc Maxisorp plates were coated with either CYN–OVA or CYN–KLH conjugate (100 mL/well) in PBS (0.01 M, pH 7.4) and incubated at 4 8C overnight in the dark. The plates were blocked using a commercial immunoassay stabiliser solution (100 mL/well) and incubated in an incubator/shaker for 1 h at 37 8C. The blocking buffer was discarded and antisera or hybridoma supernatant (diluted 1:5 in PB buffer, pH 7.2) was added to the plate (100 mL/ well). For hybridoma screening only CYN–OVA was used as the coating antigen, the positive control was the sera obtained from the final heart bleed of the mouse and the negative control was cell culture medium diluted in PB buffer (pH 7.2). The plates were incubated at 4 8C overnight in the dark, the solution discarded and the plate washed four times using wash buffer. Secondary antibody (horseradish peroxidase-linked anti-species immunoglobulin) at 1:2000 dilution was added to each well (100 mL/well) and incubated in the incubator shaker for 1 h at 37 8C in the dark. The secondary antibody was discarded and the plates washed four times with wash buffer. TMB/E (100 mL/well) was added to each well and the reaction stopped after a few minutes using 2.5 mM H2SO4. The plates were read at 450 nm using a TECAN Safire2. 2.7.3. Biosensor protocol for antibody screening The screening of the test sera for CYN antibodies was assessed on each chip surface by injecting dilutions (1 in 100) of the serum over the chip surface at a flow rate of 12 mL/min for 2 min. Report points were recorded before (5 s) and after each injection (30 s), and the relative response units were determined. The chip surface was regenerated with a 25 mL injection of hydrochloric acid (10 mM) with 10% SDS at a flow rate of 25 mL/min. A typical screening cycle was 6–8 min. 2.8. Assessment of antibody sensitivity and specificity 2.8.1. Polyclonal antibodies The sensitivity and specificity of the final polyclonal antisera produced in rabbits was assessed using a competitive indirect ELISA and the Biacore Q biosensor assay with a CYN surface produced using formaldehyde chemistry. Various parameters were initially investigated when developing both the ELISA and biosensor assays to achieve optimal conditions. The conditions applied for each antibody characterised fully are summarised in Table 1. Stock solutions of CYN and deoxyCYN standard were used to prepare working standards. For ELISA a set of working standards, ranging in concentration from 0 to 100 ng/mL in pH 7.2 PB and for biosensor 0–100 ng/mL in pH 7.4 HBS–EP buffer, were prepared to produce calibration curves for each toxin based on dose–response by each method. The response units relative to the toxin concentration curves were evaluated using a non linear regression curve fit using Graphpad Prism for Windows (version 5) in addition to determining the % binding at each toxin concentration, the midpoint (IC50) and the dynamic range (IC20–IC80) of each standard curve and the specificity of the assay to other compounds. The specificity or % cross-reactivity for deoxyCYN relative to CYN was calculated using a % ratio of mid-points for each of the standard curves. 2.8.2. Monoclonal antibodies On selection of the best performing monoclonal antibody, the cell line was grown and antibody was concentrated in cell culture medium using Viva spin centrifuge tubes (30,000 MWCO) followed

C.T. Elliott et al. / Harmful Algae 24 (2013) 10–19

14 Table 1 ELISA and SPR assay parameters for each antibody. Antibody

Raised to

ELISA

SPR biosensor

Coating antigen

Antibody titre

Antibody titre

Flow rate (mL/min)

Contact time (min)

Regeneration solution(s)

Contact time (s)

10 mM HCl/1% SDS 250 mM NaOH 10 mM HCl/1% SDS 250 mM NaOH/20% acetonitrile 250 mM NaOH/20% acetonitrile 250 mM NaOH 10 mM HCl/1% SDS 75 mM NaOH 75 mM NaOH

60 60 30 40 40 60 30 30 30

Polyclonal 1 Polyclonal 2

CYN–KLH CYN–KLH

CYN–OVA (0.5 mg/mL) CYN–OVA (1.0 mg/mL)

1 in 8K 1 in 16K

1 in 100 1 in 100

12 12

2 2

Polyclonal 3

CYN–OVA

CYN–KLH (5.0 mg/mL)

1 in 8K

1 in 100

12

2

Polyclonal 4

CYN–OVA

CYN–KLH (5.0 mg/mL)

1 in 4K

1 in 100

12

2

Monoclonal 1 Monoclonal 2

CYN–KLH CYN–KLH

CYN–OVA (0.5 mg/mL) CYN–OVA (0.5 mg/mL)

1 in 5K 1 in 5K

1 in 250 1 in 250

12 12

2 2

by purification via affinity chromatography using a protein Gsepharose gel column (MAbTrap Kit). Dialysis of the antibody over 24 h in 0.15 M saline (3 4 L) was performed. The protein concentration and isotyping of each antibody was determined at A280 nm and a mouse monoclonal antibody isotyping kit, respectively. The monoclonal antibodies were then assessed by ELISA and SPR in the same way as the polyclonal antibodies using the parameters as outlined in Table 1. 3. Results and discussion 3.1. Preparation of CYN immunogens Six different immunogens were prepared comparing directly five different chemical approaches and two different carrier proteins. The rationale behind assessing different chemical approaches was to link the carrier protein to different sites of the CYN molecule or its derivatised analogue, as illustrated in Fig. 2, to determine the best orientation of the CYN molecule to induce an immune response. For immunogens 1 and 2 the production of CYN–acid (molecular weight: 349 g/mole) was confirmed by UPLC–mass spectrometry. Aliquots (20 mL) of the NaOCl–CYN reaction mixture were mixed with 80 mL of eluent (98:2% (v/v) water:ACN modified with 0.1% TFA) and analysed every 24 h by isocratic UPLC–MS with a C18 column (0.35 mL/min flow rate; 5 mL injection volume; 50 8C column temperature). Selective ion monitoring (SIM) in positive mode for pseudomolecular ions, sodium adducts and in the case of

CYN–acid a dimer, was performed so that the relative proportions of starting material (CYN; [M+H]+, m/z 416; [M+Na]+, m/z 438), intermediate product (5-chloro-CYN; [M+H]+, m/z 450; [M+Na]+, m/ z 472) and final product (CYN–acid; [M+H]+, m/z 350; [M+Na]+, m/z 372; [2M+H]+, m/z 699) could be evaluated. The reaction mixture analysed on day 7 contained no detectable levels of 5-chloro-CYN or CYN and was deemed ready for use in conjugation (Fig. 3). Although it can be confirmed that the CYN–acid was formed, the quantity or percentage conversion could not be determined due to the lack of available analytical standards. For immunogens 3 and 4 spacer molecules were incorporated into the conjugation process. Succinic anhydride was reacted with the hydroxyl reaction site (Fig. 2C) of the CYN molecule to form a hemisuccinate derivative which creates a terminal carboxylic acid functional group that is then conjugated to the amine of the protein via the carbodiimide reaction. Similarly, the bis-epoxide homobifunctional compound 1,4-butanediol diglycidyl ether was employed to cross-link the amine of the protein with the hydroxyl reaction site of the CYN molecule (Fig. 2C). For immunogens 5 and 6 two different proteins were both conjugated via a modified Mannich reaction which most probably reacted through the indole group of the guanidinium moiety of the CYN (Fig. 2C) as was previously reported for cyclopiazonic acid (Huang and Chu, 1993). 3.2. Antibody titres by ELISA and SPR The three screening methods were utilised to assess antibody titre for the sera produced. For the determination of antibody titre

Fig. 2. Conjugation sites on the CYN molecule and derivatised analogues.

C.T. Elliott et al. / Harmful Algae 24 (2013) 10–19

100

A

0.67

349.89

C

0 100

0.2

0.4

0.6

B

0.8

1.0

1.2

1.4

Percentage

0.67

Percentage

Percentage

100

15

699.30 371.86 0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0

300

400

500

Time

600

700

800

m/z

Fig. 3. UPLC–MS analysis of CYN–acid. (A) SIM chromatogram of CYN–acid [2M+H]+, m/z 699.30. (B) SIM chromatogram of CYN–acid [M+H]+, m/z 349.89. (C) Mass spectra indicating presence of CYN–acid pseudomolecular ion and dimer.

development of competitive indirect ELISAs and SPR methods. For immunogens 1–4, as a derivative was synthesised prior to the conjugation reaction to the protein, there is a high uncertainty as to the percentage conversion ratio from CYN to the derivatised product being conjugated. It may be as the derivatives were uncharacterised that noteably less was being conjugated to the protein compared to the assumed 100% conversion which may therefore mean that the dose that was then administered to the animal was too low in CYN to stimulate a specific immune response. Due to the chronological nature of the study the direct competitive ELISA and SPR analysis for immunogens 1–4 were performed prior to work for immunogens 5 and 6. As such new efforts in the antibody production then focussed on this chemical approach which conjugated the CYN directly to the protein. Coating antigens for competitive indirect ELISA were only prepared using this method due to the limited availability of the toxin. In general for marine and fresh water toxins, the expense of purchase 1000

Response Units

by competitive direct immunoassay, testing all sera using all HRPs, no appreciable antibody titre was observed. These results could have been due to either no antibodies being raised by each of the chemical approaches or to non-functional HRP enzyme conjugates due to the harsh nature of some of the chemistries applied denaturing the HRP particularly when high pHs were applied. For the determination of antibody titre by competitive indirect immunoassay using only CYN–KLH and CYN–OVA as coating antigens, suitable antibody titres and inhibition with CYN were observed for both rabbits and mice immunised with CYN–OVA and CYN–KLH (immunogens 5 and 6) respectively. For the determination of antibody titre by SPR, testing all sera using the three different sensor surfaces, no binding response was observed for animals immunised with immunogens 1, 2 or 3. A limited binding response was observed for the sera of animals immunised with immunogen 4 on chip 2 produced using the epoxide linker but this response was insufficient for further antibody characterisation. However, suitable antibody titres were observed for both mice and rabbits immunised with immunogens 5 and 6 on chip 3 produced using formaldehyde chemistry. By SPR the antibody titre was deemed suitable if a dilution of the antibody produced a specific binding response to the biosensor surface greater than 300 response units over a 1 or 2 min injection but which could also be inhibited using the free CYN in solution. The polyclonal binding response of rabbits immunised with immunogen 5 and 6 over the immunisation schedule is shown in Fig. 4. Polyclonals 1 and 2 were raised to CYN–KLH and Polyclonals 3 and 4 were raised to CYN–OVA. For each rabbit the binding response rises quickly over the first two months then slowly increases or plateaus out until month 8. At this point there was a dip in the binding response so it was deemed a suitable point to harvest a final yield of antisera. From the different chemical conjugation approaches applied to the study only immunogens 5 and 6 produced a suitable antibody titre in both mice and rabbits worth investigating further to the

800

600

400

200

0 0

2

4

6

8

Months Fig. 4. Polyclonal antibody binding response by SPR over the immunisation schedule (* Polyclonal 1; & Polyclonal 2; ~ Polyclonal 3; & Polyclonal 4).

C.T. Elliott et al. / Harmful Algae 24 (2013) 10–19

16

(A)

120

Normalised optical density

Normalised optical density

120 100 80 60 40 20

0 0.001

0.01

0.1

1

10

100

(B)

100 80 60 40 20 0 0.001

1000

CYN Concentration (ng/mL)

(C)

80 60 40 20 0 0.001

0.01

0.1

1

10

0.1

1

10

100

1000

DeoxyCYN Concentration (ng/mL)

Normalised response

Normalised response

100

0.01

100

1000

CYN Concentration (ng/mL)

100

(D)

80 60 40 20 0 0.001

0.01

0.1

1

10

100

1000

DeoxyCYN Concentration (ng/mL)

Fig. 5. Calibration curves for polyclonal antibodies determined by ELISA for (A) CYN and (B) deoxyCYN and by SPR for (C) CYN and (D) deoxyCYN (* Polyclonal 1; & Polyclonal 2; ~ Polyclonal 3; ! Polyclonal 4).

and limited availability makes the challenge of finding a suitable chemistry to produce effective immunogens a difficult task. 3.3. Polyclonal antibody characterisation For polyclonal antibody production, animals demonstrating a suitable titre received a total of nine immunisations before their sacrifice whereby a final bleed was taken. The assay parameters used for both indirect competitive ELISA and SPR for the final bleeds of the four polyclonal sera (Fig. 5) from immunogens 5 and 6 are summarised in Table 1. All four polyclonal antisera demonstrated high sensitivity, determined by the IC50, for CYN by ELISA ranging from 27 to 131 pg/mL (Table 2). The dynamic range (IC20–IC80) deemed to be the theoretical lower and upper detection limits for the assay was similar for all four polyclonal antibodies with Polyclonal 1 showing the lowest theoretical detection limit of 10 pg/mL. For all antibodies the dynamic range was in the pg/mL range. Currently deoxyCYN is not considered to be toxic compared to CYN therefore the low percentage cross-reactivity demonstrated for all 4 antibodies of up to 6.2% of this analogue of CYN in the SPR assay is beneficial so that an assay developed will not detect the deoxyCYN to the same extent the much more potent CYN. This finding strongly suggests that the hydroxyl molecule present in CYN but absent in deoxyCYN plays an important role in the ability of the antibody to bind to the toxin. These polyclonal antibodies showed no cross-reactivity towards microcystins but this was not unexpected due to the much greater differences in molecular structure. By SPR all four polyclonal antibodies showed a similar titre with those raised to KLH as the carrier protein displaying marginally

better sensitivity to CYN compared to those raised to OVA. The sensitivity of the polyclonal antibodies by SPR ranged between 4.4 and 9.6 ng/mL (Table 2). The dynamic range (IC20–IC80) was similar for all four polyclonal antibodies with Polyclonal 1 showing the lowest theoretical detection limit of 1.9 ng/mL. The SPR assay demonstrated slightly higher theoretical detection limits yet similar low % cross-reactivity for all 4 antibodies towards deoxyCYN of between 9.6 and 12.5% (Table 2). 3.4. Monoclonal antibody characterisation For the purposes of monoclonal antibody production the mice received a total of five immunisations and the best performing mouse in terms of titre detected was identified as belonging to a mouse immunised with CYN–KLH. This animal was sacrificed, the spleen harvested and a fusion performed. From the fusion over a two-week period 714 hybridomas were screened for antibodies binding to CYN conjugate. For those screened 10% (71) demonstrated antibody binding greater than the final heart bleed polyclonal control and 26% (18) of those tested positive for inhibition when challenged with CYN. The best 4 performing hybridomas in terms of intensity of reaction observed were cloned and following the cloning procedure two clones which demonstrated the best inhibition were grown, purified and characterised. The final protein concentration of the purified monoclonal antibodies Monoclonal 1 and Monoclonal 2 was 0.948 and 1.115 mg/mL determined by UV A280. Isotyping of the monoclonal antibodies demonstrated that both were IgG1 heavy chain with k light chains. By ELISA and SPR the sensitivity of the monoclonal antibodies, Monoclonal 1 and Monoclonal 2, was on average slightly higher

Normalised optical density

9.6 9.8 12.5 11.9 10.5 11.2

120

13.4–134 29.3–204.5 25.5–186.8 39.2–161.8 38.7–205.3 40.1–228.1

DeoxyCYN IC20–IC80 (ng/mL)

%CR to deoxyCYN

C.T. Elliott et al. / Harmful Algae 24 (2013) 10–19

(A)

100 80 60 40 20

0.01

0.1

1

10

100

100

Normalised response

1.9–10.0 4.0–14.7 3.9–19.6 4.5–19.9 4.4–20.6 5.7–21.2

45.8 77.2 71.3 80.9 92.0 98.9

Concentration (ng/mL)

4.4 7.6 8.9 9.6 9.7 11.1

Biosensor

CYN IC20–IC80 (ng/mL)

DeoxyCYN IC50 (ng/mL)

0 0.001

CYN IC50 (ng/mL)

17

(B)

80 60 40 20 0 0.001

0.01

0.1

1

10

100

1000

0.25–6.8 0.16–9.4 0.61–6.1 0.71–4.9 1.6–8.7 0.56–6.4 1.31 1.11 1.55 2.14 3.8 2.1

2.1 4.2 6.2 6.1 3.2 5.1

DeoxyCYN IC20–IC80 (ng/mL) DeoxyCYN IC50 (ng/mL)

0.01–0.08 0.01–0.19 0.06–0.17 0.07–0.22 0.04–0.36 0.06–0.18 CYN–KLH CYN–KLH CYN–OVA CYN–OVA CYN–KLH CYN–KLH Polyclonal 1 Polyclonal 2 Polyclonal 3 Polyclonal 4 Monoclonal 1 Monoclonal 2

0.027 0.047 0.097 0.131 0.121 0.108

CYN IC20–IC80 (ng/mL) CYN IC50 (ng/mL)

Raised to

ELISA

Fig. 6. Calibration curves for monoclonal antibodies determined by (A) ELISA and (B) SPR for CYN and deoxyCYN (* CYN monoclonal 1; ^ deoxyCYN monoclonal 1; * CYN monoclonal 2; ^ deoxyCYN monoclonal 2.

than that of the polyclonal antibodies (Fig. 6). By ELISA the sensitivities were 108 and 121 pg/mL with % cross-reactivities to deoxyCYN of 3.2% and 5.1% and by SPR the sensitivities were 9.8 and 11.1 ng/mL with % cross-reactivity for deoxyCYN at 10.5– 11.2% (Table 2). The benefit of monoclonal antibodies over their polyclonal counterparts is well documented such as their homogeneous immortal reproducible characteristics compared to polyclonal sera. There is no variability in specificity and affinity of monoclonal antibodies compared to polyclonal sera though they are more laborious to produce and have a limited stability in changes in pH or salt concentration. 3.5. Comparison of methods

Antibody

Table 2 Sensitivity and specificity of each antibody by ELISA and SPR.

%CR to deoxyCYN

Concentration (ng/ml)

This level of sensitivity and specificity for all antibodies would allow for the development of either fit for purpose ELISA or SPR assays to be constructed for food and environmental samples. The ELISA assay employed demonstrated a range in factor of 60–360 times more sensitivity compared to the corresponding SPR assay. However, the SPR assay employed was a real time assay whereby the sample being analysed could be monitored for CYN during the analysis which would present a rapid result. For SPR one important parameter to note is that the regeneration solutions employed had to be optimised for each animal with Polyclonal 3 proving to bind strongly to the chip surface thereby requiring harsher conditions to be removed for the automated regeneration of the next sample. It is important that these regeneration conditions are considered whenever any automated sensor system is being employed to ensure that the chip is fully conditioned to its original state for the next analysis to avoid any issues with carryover in the system. When food or environmental samples are being analysed or the antibodies are being employed in different sensor technologies

18

C.T. Elliott et al. / Harmful Algae 24 (2013) 10–19

these regeneration conditions may need re-optimised to the analysis conditions at that time. For the polyclonal sera a higher level of non-specific background response was observed for the antibodies raised to OVA as the carrier protein compared to those raised to KLH when evaluated by ELISA. This is most likely due to non-specific binding of the antibody to KLH as the coating antigen. This effect is reduced when free CYN is immobilised onto sensor surfaces which then eliminates non-specific protein–protein interactions. For both the polyclonal and the monoclonal antibodies generated in this study the selection of an appropriate platform for the assay development would be more dependent of the requirements of the end user. The ELISA based tests would offer greater sensitivity whereas the SPR based assay would allow for rapid analysis of a sample at that time. The sample preparation of the environmental or food sample would also need to be considered and the ELISA based test would offer more scope in the dilution of the sample to reduce interfering components based on its improved sensitivity. For rapid portable testing the antibodies could be incorporated or employed using other immunoassay-based platforms such as lateral flow devices, flow cytometry or electrochemical sensors. 4. Conclusion Antibodies of high sensitivity and specificity to CYN were successfully produced for a range of future applications in rapid, immunobased detection platforms. This is the first report of the successful production of antibodies to this toxin. One of the limitations which hampers all researchers in the field of natural toxin analysis is either the availability or expense of the toxin to purchase. Since no work has been published on the development of antibodies to this toxin it was important to find a suitable chemistry to use as an effective immunogen and to discuss all the approaches taken to achieve the success reported. Using a modified Mannich reaction for immunogen synthesis both polyclonal and monoclonal antibodies of high sensitivity were produced. It was demonstrated that these antibodies could be used in two assay formats of ELISA and SPR. Future work will be employed in the application of these antibodies on different immunoassay based platforms for food and environmental sample analysis to determine CYN presence. Acknowledgements This research was funded by The Department of Employment and Learning All Island Research programme (ASSET), the Interreg Programme through the project ‘‘Atlantox: Advanced Tests about New Toxins appeared in the Atlantic Area’’ and the EU FP7 project ‘‘mAQUA: Universal microarrays for the evaluation of fresh-water quality based on detection of pathogens and their toxins’’ of Grant agreement no: 265409. The authors would like to acknowledge both Brett Greer and Eliza Barszczewska-Lyner for their technical support.[SS] References Banker, R., Carmeli, S., Werman, M., Teltsch, B., Porat, R., Sukenik, A., 2001. Uracil moiety is required for toxicity of the cyanobacterial hepatotoxin. Journal of Toxicology and Environmental Health. Part A 62, 281–288. Banker, R., Teltsch, B., Sukenik, A., Carmeli, S., 2000. 7-Epi, a toxic minor metabolite of the cyanobacterium Aphanizomenon ovalisporum from Lake Kinneret, Israel. Journal of Natural Products 63, 387–389. Bernard, C., Harvey, M., Briand, J.F., Bire, R., Krys, S., Fontaine, J.J., 2003. Toxicological comparison of diverse Cylindrospermopsis raciborskii strains: evidence of liver damage caused by a French C. raciborskii strain. Environmental Toxicology 18, 176–186. Berry, J.P., Lind, O., 2010. First evidence of paralytic ‘‘ralytic shellfish toxins’’ and in a Mexican freshwater system, LagoCatemaco, and apparent bioaccumulation of

the toxins in ‘‘tegogolo’’ snails (Pomacea patula catemacensis). Toxicon 55, 930–938. Beyer, D., Sura´nyi, G., Vasas, G., Roszik, J., Erdodi, F., M-Hamvas, M., Ba´csi, I., Ba´tori, R., Serfozo, Z., Szigeti, Z.M., Vereb, G., Demeter, Z., Gonda, S., Ma´the´, C., 2009. Cylindrospermopsin induces alterations of root histology and microtubule organization in common reed (Phragmites australis) plantlets cultured in vitro. Toxicon 54 (4), 440–449. Blahova, L., Oravec, M., Marsalek, B., Sejnohova, L., Simek, Z., Blaha, L., 2009. The first occurrence of the cyanobacterial alkaloid toxin in the Czech Republic as determined by immunochemical and LC/MS methods. Toxicon 53 (5), 519–524. Brient, L., Lengronne, M., Bormans, M., Fastner, J., 2009. First occurrence of Cylindrospermopsin in freshwater in France. Environmental Toxicology 24 (4), 415–420. Cadel-Six, S., Peyraud-Thomas, C., Brient, L., de Marsac, N.T., Rippka, R., Mejean, A., 2007. Different genotypes of anatoxin-producing cyanobacteria coexist in the Tarn river, France. Applied and Environmental Microbiology 73 (23), 7605– 7614. Campbell, K., Stewart, L.D., Fodey, T.L., Haughey, S.A., Doucette, G.J., Kawatsu, K., Elliott, C.T., 2007. An assessment of specific binding proteins suitable for the detection of paralytic shellfish poisons (PSP) using optical biosensor technology. Analytical Chemistry 79 (15), 5906–5914. Carson, B., 2000. Cylindrospermopsin: Review of Toxicological Literature. Integrated Laboratory Systems, North Carolina. Cheng, X.L., Shi, H.L., Adams, C.D., Timmons, T., Ma, Y.F., 2009. Effects of oxidative and physical treatments on inactivation of Cylindrospermopsis raciborskii and removal of cylindrospermopsin. Water Science and Technology 60 (3), 689– 697. Codd, G., Lindsay, J., Young, F., Morrison, L., Metcalf, J., 2005. In: Huisman, J., Matthijs, H., Visser, P. (Eds.), Harmful Cyanobacteria. Springer, Netherlands, pp. 1–23. Crooks, S.R.H., Traynor, I.M., Elliott, C.T., McCaughey, W.J., 1997. Detection of monensin residues in poultry liver using an enzyme immunoassay. Analyst 122, 161–163. Duy, T.N., Lam, P.K.S., Shaw, G.R., Connell, D.W., 2000. Toxicology and risk assessment of freshwater cyanobacterial (blue-green algal) toxins in water. Reviews of Environment Contamination and Toxicology 163, 113–186. Eaglesham, G.K., Norris, R.L., Shaw, G.R., Smith, M.J., Chiswell, R.K., Davis, B.C., Neville, G.R., Seawright, A.A., Moore, M.R., 1999. Use of HPLC–MS/MS to monitor, a blue-green algal toxin, for public health purposes. Environmental Toxicology 14, 151–154. EFSA, 2008. Opinion of the scientific panel on contaminants in the food chain on a request from the European Commission on marine biotoxins in shellfish – azaspiracids. The EFSA Journal 723, 1–52. EFSA, 2009. Scientific opinion of the panel on contaminants in the food chain on a request from the European Commission on marine biotoxins in shellfish – influence of processing in the levels of lipophilic marine biotoxins in bivalve molluscs. The EFSA Journal 1016, 1–10. EFSA, 2010. Scientific opinion on marine biotoxins in shellfish – cyclic imines (spirolides, gymnodimines, pinnatoxins and pteriatoxins). The EFSA Journal 1628, 1–39. Erlanger, B.F., Borek, F., Beiser, S.M., Lieberman, S., 1957. Steroid-protein conjugates. I. Preparation and characterization of conjugates of bovine serum albumin with testosterone and with cortisone. Journal of Biological Chemistry 228, 713–727. Falconer, I.R., Hardy, S.J., Humpage, A.R., Froscio, S.M., Tozer, G.J., Hawkins, P.R., 1999. Hepatic and renal toxicity of the blue-green alga (cyanobacterium) Cylindrospermopsis raciborskii in male Swiss albino mice. Environmental Toxicology 14, 143–150. Fodey, T., Murilla, G., Cannavan, A., Elliott, C., 2007. Characterisation of antibodies to chloramphenicol, produced in different species by enzyme-linked immunosorbent assay and biosensor technologies. Analytica Chimica Acta 592, 51–57. Froscio, S.M., Humpage, A.R., Burcham, P.C., Falconer, I.R., 2001. Cell-free protein synthesis inhibition assay for the cyanobacterial toxin. Environmental Toxicology 16, 408–412. Froscio, S.M., Humpage, A.R., Burcham, P.C., Falconer, I.R., 2003. Cylindrospermopsin-induced protein synthesis inhibition and its dissociation from acute toxicity in mouse hepatocytes. Environmental Toxicology 18, 243–251. Galfre, G., Milstein, C., 1981. Preparation of monoclonal antibodies: strategies and procedures. Methods in Enzymology 72, 3–46. Gallo, P., Fabbrocino, S., Cerulo, M.G., Ferranti, P., Bruno, M., Serpe, L., 2009. Determination of Cylindrospermopsin in freshwaters and fish tissue by liquid chromatography coupled to electrospray ion trap mass spectrometry. Rapid Communications in Mass Spectrometry 23 (20), 3279–3284. Griffiths, D.J., Saker, M.L., 2003. The Palm island mystery disease 20 years on: a review of research on the cyanotoxin. Environmental Toxicology 18 (2), 78–93. Hermanson, G.T. (Ed.), 1996. Homobifunctional cross-linkers: bioconjugate techniques. Academic Press, London, pp. p.220–p.221. Huang, X., Chu, F.S., 1993. Production and characterization of monoclonal and polyclonal antibodies against the mycotoxin cyclopiazonic acid. Journal of Agricultural and Food Chemistry 41, 329–333. Hudnell, H.K., Dortch, Q., Zenick, H., 2008. An overview of the interagency. International Symposium on Cyanobacterial Harmful Algal Blooms (ISOC-HAB): Advancing the Scientific Understanding of Freshwater Harmful Algal Blooms. In: Hudnell, H.K. (Ed.), Cyanobacterial Harmful Algal Blooms: State of the Science and Research Needs. Springer, New York, pp. 1–16. Kinnear, S., 2010. A decade of progress on bioaccumulation research. Marine Drugs 8 (3), 542–564.

C.T. Elliott et al. / Harmful Algae 24 (2013) 10–19 Kittler, K., Schreiner, M., Krumbein, A., Manzei, S., Koch, M., Rohn, S., Maul, R., 2012. Uptake of the cyanobacterial toxin in Brassica vegetables. Food Chemistry, http://dx.doi.org/10.1016/j.foodchem.2012.01.107. Li, R., Carmichael, W.W., Brittain, S., Eaglesham, G.K., Shaw, G.R., Liu, Y., Watanabe, M.M., 2001. First report of the cyanotoxins Cylindrospermopsin and deoxy cylindrospermopsin from Raphidiopsis curvata (cyanobacteria). Journal of Phycology 37, 1121–1126. Looper, R.G., Runnegar, M.T.C., Williams, R.M., 2005. Synthesis of the putative structure of 7-deoxy: C7 oxygenation is not required for the inhibition of protein synthesis. Angewandte Chemie International Edition 44, 3879–3881. Messineo, V., Bogialli, S., Melchiorre, S., Sechi, N., Luglie, A., Casiddu, P., Marlani, M.A., Padedda, B.M., Di Corcia, A., Mazza, R., Carloni, E., Bruno, M., 2009. Cyanobacterial toxins in Italian freshwaters. Limnologica 39 (2), 95–106. Norris, R.L., Eaglesham, G.K., Pierens, G., Shaw, G.R., Smith, M.J., Chiswell, R.K., Seawright, A.A., Moore, M.R., 1999. Deoxy, cylindrospermopsin an analog of cylindrospermopsin from Cylindrospermopsis raciborskii. Environmental Toxicology 14, 163–165. Norris, R.L.G., Seawright, A.A., Shaw, G.R., Senogles, P., Eaglesham, G.K., Smith, M.J., Chiswell, R.K., Moore, M.R., 2002. Hepatic xenobiotic metabolism of cylindrospermopsin in vivo in the mouse. Toxicon 40, 471–476. Norris, R.L.G., Seawright, A.A., Shaw, G.R., Smith, M.J., Chiswell, R.K., Moore, M.R., 2001. Distribution of 14C cylindrospermopsin in vivo in the mouse. Environmental Toxicology 16, 498–505. Nybom, S.M.K., Salminen, S.J., Meriluoto, J.A.O., 2008. Specific strains of probiotic bacteria are efficient in removal of several different cyanobacterial toxins from solution. Toxicon 52 (2), 214–220.

19

Oehrle, S.A., Southwell, B., Westrick, J., 2010. Detection of various freshwater cyanobacterial toxins using ultra-performance liquid chromatography tandem mass spectrometry. Toxicon 55, 965–972. Ohtani, I., Moore, R.E., 1992. Cylindrospermopsin: A potent hepatotoxin from the blue-green alga Cylindrospermopsis raciborskii. Journal of the American Chemical Society 114, 7941–7942. Paerl, H.W., Huisman, J., 2008. Blooms like it hot. Science 320 (5872), 57–58. Paredes, I., Rietjens, I.M.C.M., Vieites, J.M., Cabado, A.G., 2011. Update of risk assessments of main marine biotoxins in the European Union. Toxicon 58 (4), 336–354. Preussel, K., Wessel, G., Fastner, J., Chorus, I., 2009. Response of cylindrospermopsin production and release in Aphanizomenon flos-aquae (Cyanobacteria) to varying light and temperature conditions. Harmful Algae 8 (5), 645–650. Rawn, D.F.K., Niedzwiadek, B., Lau, B.P.Y., Saker, M., 2007. Anatoxin-a and its metabolites in blue-green algae food supplements from Canada and Portugal. Journal of Food Protection 70 (3), 776–779. Reisner, M., Carmeli, S., Werman, M., Sukenik, A., 2004. The cyanobacterial toxin cylindrospermopsin inhibits pyrimidine nucleotide synthesis and alters cholesterol distribution in mice. Toxicological Sciences 82 (2), 620–627. Tadayoni, B.M., Friden, P.M., Walus, L.R., Musso, G.F., 1993. Synthesis in vitro kinetics, and in vivo studies on protein conjugates of AZT: evaluation as a transport system to increase brain delivery. Bioconjugate Chemistry 4, 139–145. Thomas, A.D., Saker, M.L., Norton, J.H., Olsen, R.D., 1998. Cyanobacterium Cylindrospermopsis raciborskii as a probable cause of death in cattle in Queensland. Australian Veterinary Journal 76, 592–594.