Characterization of paralytic shellfish toxins from Lyngbya wollei dominated mats collected from two Florida springs

Harmful Algae 16 (2012) 98–107

Contents lists available at SciVerse ScienceDirect

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

Characterization of paralytic shellfish toxins from Lyngbya wollei dominated mats collected from two Florida springs Amanda J. Foss a,b,*, Edward J. Phlips a, Mete Yilmaz a, Andrew Chapman b a The Fisheries and Aquatic Sciences Program in the School of Forest Resources and Conservation, Institute of Food and Agricultural Sciences (IFAS), University of Florida, 7922 NW 71st Street, Gainesville, FL 32653, USA b GreenWater Laboratories/CyanoLab, 205 Zeagler Drive, Palatka, FL 32177, USA

A R T I C L E I N F O

A B S T R A C T

Article history: Received 2 August 2011 Received in revised form 8 February 2012 Accepted 8 February 2012 Available online 17 February 2012

Lyngbya wollei, a commonly observed cyanobacterium in Florida’s spring fed systems, is considered a nuisance organism due to its formation of large benthic and floating mats. Standing crops and mats of Lyngbya from two Florida springs, Silver Glen Springs (Ocala National Forest) and Blue Hole Spring (Ichetucknee Springs State Park), were sampled and characterized via microscopy. A near full-length 16S rRNA gene sequence recovered from genomic DNA preparation of a filament collected from Silver Glen Natural Well was 99% identical to another L. wollei sequence. Paralytic shellfish toxin (PST) biosynthesis genes sxtA and sxtG were also detected in the filament DNA and were 97% and 98% identical in sequence, respectively, to those of L. wollei. PSTs were characterized utilizing High Performance Liquid Chromatography (HPLC) coupled with Mass Spectrometry (MS). Analysis of extracted algal material with LC/MS/MS verified that PSTs decarbamoylgonyautoxin 2&3 (dcGTX2&3) and decarbamoylsaxitoxin (dcSTX) were present in L. wollei mats in Florida springs and provided evidence supporting the presence of all L. wollei toxins (LWT 1-6). Levels of quantifiable toxins (dcGTX2&3 & dcSTX) ranged from 19 to 73 mg STX-eq (g dry weight)1. Although L. wollei toxins 1–6 could not be quantified due to a lack of available standards, their presence indicates samples may be higher in toxicity. This is the first detailed study confirming PST presence in L. wollei dominated mats in Florida spring systems. ß 2012 Elsevier B.V. All rights reserved.

Keywords: Cyanobacteria Lyngbya wollei Saxitoxin Paralytic shellfish toxin Florida springs 16S rRNA sxtA sxtG Mass spectrometry

1. Introduction Cyanobacteria produce many active metabolites, some of which induce toxic responses, including harmful effects on human and animal populations (Sivonen and Jones, 1999). In Florida’s spring fed rivers, one of the dominant cyanobacteria is Lyngbya wollei (Farlow ex Gomont) Speziale, which forms large benthic and floating mats (Cowell and Botts, 1994). A survey submitted to the Florida Department of Environmental Protection (FDEP) evaluating algal growth and nutrients in 21 different springs, found that L. wollei is one of the most common ‘‘macroalgae’’ present and has become a management concern (Stevenson et al., 2007). L. wollei has also been shown to produce toxins (Seifert et al., 2007; Berry et al., 2004; Teneva et al., 2003; Carmichael et al., 1997). Toxins of concern include dermatoxins (toxins that affect the skin), hepatotoxins (toxins that affect the liver and other internal organs), and neurotoxins (toxins that affect nerve cells) (Landsberg, 2002). Although anecdotal reports of adverse skin reactions

* Corresponding author at: GreenWater Laboratories, 205 Zeagler Drive Suite 302, Palatka, FL 32177, USA. Tel.: +1 386 328 0882. E-mail address: [email protected] (A.J. Foss). 1568-9883/$ – see front matter ß 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.hal.2012.02.004

(i.e. rashes, hives, and blisters), gastrointestinal disorders, respiratory illness, and even temporary loss of consciousness following potential exposure to cyanobacteria in Florida waterways have been reported to the Florida Department of Health (FDOH) (personal communication), detailed studies relating cyanobacterial toxins to health effects in Florida are limited. The FDOH initiated an evaluation of freshwater Lyngbya and its toxins in Florida springs in 2004 (PBS&J [Post, Buckley, Schuh, and Jernigan] 2007). The study reported the presence of ‘‘saxitoxin-like’’ compounds, which were designated unknown paralytic shellfish toxins (PSTs), since methods used at the time (i.e. ELISA) were not sufficiently selective to specify variants. Paralytic shellfish toxins, commonly referred to as saxitoxins or paralytic shellfish poisons, are a group of neurotoxic alkaloids produced primarily by marine dinoflagellates and freshwater cyanobacteria, with over 57 structural variants reported to date (Wiese et al., 2010). Because PSTs mainly act by blocking voltagegated sodium and calcium channels (Kao and Levinson, 1986; Su et al., 2004), exposure to these toxins may result in illness, paralysis, and even death (Kao, 1993). PSTs are potentially harmful to aquatic and terrestrial animals (Landsberg, 2002) in addition to humans. Analysis of L. wollei samples from Lake Guntersville in Alabama (USA) have yielded the PSTs decarbamoylgonyautoxins

A.J. Foss et al. / Harmful Algae 16 (2012) 98–107

99

2&3, decarbamoylsaxitoxin, and six novel PSTs, (L. wollei toxins 16) (Carmichael et al., 1997; Yin et al., 1997, Onodera et al., 1997). No other detailed reports of L. wollei toxins have since come out in the peer-reviewed literature. The goal of this study was to conduct detailed analyses of the PSTs associated with L. wollei dominated mats in Florida. Since the morphologic and molecular identification of L. wollei is under some scrutiny (Joyner et al., 2008) and toxin production cannot be determined by simply determining the presence of Lyngbya, the characterization of PST toxins in Florida’s L. wollei mats was an important goal. 2. Materials and methods 2.1. Sampling site description Three sites were selected for Lyngbya collection, one in Blue Hole Spring (Ichetucknee State Park) and two in the Silver Glen Springs Recreation Area (Ocala National forest). These springs were chosen because they contain extensive mats dominated by the cyanobacterium L. wollei. Silver Glen Springs is located in the Ocala National Forest in Marion County Florida and is utilized as a recreation area maintained by the United States Department of Agriculture (USDA) Forest Service. The spring run has a large combined pool emanating from two vents (east and southwest vents). The main pool is sectioned off with ropes into three regions, two of which contain large Lyngbya dominated mats (Fig. 1). The eastern spring vent is a 1st magnitude spring (flow rate > 100 cubic feet per second, [cfs]) and is frequented by swimmers. Very little Lyngbya is present in the latter region. The southwest vent (also known as the ‘‘Natural Well’’) is a 2nd magnitude spring (flow rates 10–100 cfs) and contains extensive Lyngbya-dominated mats. Both regions with mats were sampled over the study period. An underwater view representing the expansive nature of the L. wollei mat in Silver Glen Springs is shown in Fig. 2(A). Blue Hole Spring (aka Jug Spring) is a 1st magnitude spring (flow rate >100 cfs) in the Ichetucknee Springs State Park located in Columbia County, FL. Blue Hole Spring is one of many springs that feed into the Ichetucknee River system, which empties into the Santa Fe River. Although recreational activities such as tubing and canoeing are allowed in the main spring run, the activities in Blue Hole are limited to swimming, cave diving and snorkeling. Lyngbya-dominated mats in this spring are located along the bottom as well as vertically distributed along the northern face of the spring hole, attached to the roots of bald cypress trees (Taxodium distichum). Mat samples were collected from the vertically distributed mat (Fig. 2(B)).

Fig. 1. Silver Glen Springs Recreational Park photo; white lines represent ropes restricting general access; Natural Well and Main Stem were sampled in this study. Photo Courtesy of St. Johns River Water Management District.

Fig. 2. Underwater images of L. wollei mats at the study sites (A): Silver Glen Natural Well (southeastern vent) standing mat (August 2009), (B) Blue Hole (Jug) Spring vertically distributed mat loosely attached to tree roots (left). Mat is in close proximity to stairs utilized recreationally to access to spring head.

2.2. Collection A single line transect was used to collect Lyngbya-dominated mat samples. A 37 m transect was used in the main stem area of Silver Glen Spring. A surveyor tape was run across the length of the benthic mat from east to west. The Natural Well transect was 9 m in length, running along the bottom across the mat from north to south. Because the mat at Blue Hole Spring was not horizontally distributed, a 5 m transect was set up at a depth of 1 m (half the maximum depth) along the root line of bald cypress trees. Algal mat samples were collected as grab samples at five randomly selected locations along each transect. Each grab sample consisted of approximately 20 g wet weight, which were deposited in 10 in.  10 in. plastic bags for transport and analysis (approximately 100 g wet weight total). Large invertebrates and debris were lightly rinsed from the composited mat samples using spring water. There were six collection events during which all three sites were visited. One composited collection per site was taken during a sampling event, with a field replicate collected during each event. A total of 24 samples were collected for this study (Table 1). The samples were maintained below 10 8C after collection and for transport. All sample preparation for lyophilization, DNA isolation, and algal analysis was conducted within 4 h of collection. Additional data collected included tree canopy coverage (densiometer, Wildlife Supply Company, Yulee, FL), site depth,

A.J. Foss et al. / Harmful Algae 16 (2012) 98–107

100 Table 1 List of sample collections. Site

Sampling dates

Silver Glen Main Stem Silver Glen Natural Well Blue Hole Spring

8/25/09a 8/25/09 8/27/09

a

9/29/09 9/29/09 9/24/09a

10/23/09 10/23/09a 10/26/09

11/24/09a 11/24/09 12/4/09

1/27/10 1/27/10a 2/3/10

4/5/10 4/5/10 4/2/10a

Duplicate field sample collected at this time.

sampling depth, pH, dissolved oxygen (DO), specific conductivity, and temperature (YSI MP6600 sonde, YSI Inc., Yellow Springs, OH). Densiometer readings for overstory density (canopy coverage) were taken at every sample grab collection (5 readings per composited collection) and averaged for each sample site. Overstory density was determined from raw data using the following formula: 100%  (# unfilled squares  4.17) = average overstory density (according to manufacturer instructions). DO, pH, and conductivity were measured at the center of each transect at half the maximum depth. 2.3. Standards and reagents The Certified Reference Materials Program of the National Research Council of Canada provided certified calibration solutions of saxitoxin (STX), decarbamoylsaxitoxin (dcSTX), neosaxitoxin (NEO), decarbamoylneosaxitoxin (dcNEO), gonyautoxin 1&4 (GTX1&4), gonyautoxin 2&3 (GTX2&3), gonyautoxin 5 (GTX5), decarbamoylgonyautoxin 2&3 (dcGTX2&3) and N-sulfocarbamoylgonyautoxin-2 and -3 (C1&2) (Institute for Marine Biosciences, Halifax, Canada). RM 8642 FDA saxitoxin dihydrochloride solution (STX) was also purchased from the National Institute of Standards and Technology Standard Reference Materials (Gaithersberg, MD). All standards were further diluted with either 0.003 M hydrochloric acid or 0.01 M acetic acid when required. HPLC grade acetonitrile, sodium hydroxide, glacial acetic acid, hydrochloric acid (certified ACS), (>99%), formic acid, and ammonium formate were purchased from Thermo Fisher Scientific (Waltham MA) as was glutaraldehyde (25% in water) for algal preservation. All mobile phases for HPLC were filtered through 0.45-mm PVDF Millipore filters (Thermo Fisher Scientific) prior to utilization. Deionized water (18 MV cm) was provided in-house by a Pure Lab Ultra Filtration System (Siemens Water Technologies Corp., Warrendale PA). 2.4. Sample preparation 2.4.1. Qualitative analysis of mat samples Algal mats were placed in a 40 cm  30 cm  3.75 cm tray within 4 h of collection, spread out and rinsed with DI water to remove non-algal debris. Five subsamples (approximately 1 g wet weight each) were removed from each sample and preserved with gluteraldehyde (0.25%). The preserved samples were stored at 4 8C until species identification/characterization was conducted using microscopy. Microscopy subsamples (approximately 1 g wet weight) of preserved samples were placed in plastic Petri dishes with spring or distilled water. A minimum of two wet mounts were made from each subsample. For each wet mount, the entire area under the cover slip was scanned at 100 using a Nikon Eclipse TE200 inverted microscope equipped with phase contrast optics and epifluorescence. Higher magnifications were used as necessary for species identifications. Identifications were taken to species level when possible. A list of observed species was constructed, organized from most to least abundant based on the empirical judgment of the analyst. Epi-fluorescence was utilized to aid in characterization of epiphytes attached to Lyngbya filaments.

2.4.2. DNA isolation, polymerase chain reaction (PCR), and sequencing A single filament of L. wollei was isolated via microscopy from the sample collected in Silver Glen Springs Natural Well (9/29/09). An effort to cut a clean filament absent of epiphytic algae was made by observing the filament at 100 and 400 with phase contrast optics prior to DNA isolation. The filament was stored at 20 8C until DNA isolation could be conducted. Genomic DNA isolation, from a single filament, was performed according to Tillett and Neilan (2000), and this preparation was used for all PCR amplifications. Amplification and sequencing of the 16S rRNA gene were performed as described in Yilmaz et al. (2008). The presence of PST biosynthesis genes was checked by partial amplification of sxtA and sxtG genes, both of which are present in all PST-producing cyanobacteria examined (Kellmann et al., 2008). Polyketide synthase-like SxtA is formed from 4 catalytic domains (SxtA1– SxtA4) and is proposed to catalyze the first step in PST biosynthesis (Kellmann et al., 2008). Primer pair sxtAF (50 -AGCTGGACTCGGCTTGTTGCTTC) and sxtAR (50 -CACTTGCCAAACTCGCAACAGG) was used to amplify an approximately 657 bp fragment within the sxtA4 domain (Yilmaz and Phlips, 2011). The product of this domain is homologous to class II aminotransferases and proposed to perform the Claisen condensation between propionyl–Acyl carrier protein (ACP) and arginine. Primer pair sxtGF (50 -ATTGAAGCACCAATGGCAGATCG) and sxtGR (50 -AGAGTTCCGCGTCGGCGAAC), designed in this work, was utilized to amplify an approximately 700 bp fragment within the amidinotransferase gene, sxtG. SxtG is proposed to transfer an amidino group from arginine onto the product of SxtA (Kellmann et al., 2008). Primer pairs for sxtA and sxtG were designed by aligning all available cyanobacterial sequences present in the National Center for Biotechnology Information (NCBI) (http:// www.ncbi.nlm.nih.gov/) database to amplify the corresponding genes from all saxitoxin producing cyanobacteria. Both PCR reactions contained 5 mL of genomic DNA, 20 pmol of each primer (Eurofins, MWG operon, Huntsville, AL), 200 mM of each deoxynucleoside triphosphate (Thermo Fisher Scientific), 1.5 mM MgCl2, 10 mL of 5 green buffer, and 2 units of GoTaq1 Flexi DNA polymerase (Promega, Madison, WI) in a total volume of 50 mL. Amplification was initiated with denaturation of the genomic DNA at 95 8C for 3 min, followed by 33 cycles of 95 8C for 30 s, 58 8C for 30 s, 72 8C for 1 min, and ended with an extension step at 72 8C for 5 min. PCR products were purified from agarose gels (1.5%, w/v) with the QIAquick Gel extraction kit (Qiagen, Valencia, CA). Sequencing of the 16S rRNA gene was performed with sequencing primers reported in Yilmaz et al. (2008) and PST biosynthesis genes were sequenced on both strands using the same primers used in amplifications at the University of Florida’s Interdisciplinary Center for Biotechnology Research core sequencing facility. Sequences were manually checked and corrected using Mega version 4.1 (Tamura et al., 2007). New gene sequences obtained in this work are deposited in GenBank under accession numbers JQ282906 for 16S rRNA gene, JQ282907 for sxtA gene, and JQ282908 for sxtG gene. 2.5. Toxin extraction In preparation for toxin extraction excess water was lightly squeezed from the mats and filaments were cut into 1-cm lengths

A.J. Foss et al. / Harmful Algae 16 (2012) 98–107 Table 2 Monitored molecular and fragmentation ions. Toxin

[M + H]+

SIM ion

Product ions

STX dcSTX dcNEO dcGTX2 dcGTX3 GTX1 GTX4 NEO a LWT 1 a LWT 2 a LWT 3 a LWT 4 a LWT 5 a LWT 6

300 257 273 353 353 412 412 316 379 395 395 241 299 283

300 257 273 273 353 332 412 316 379 395 395 241 299 283

282 239 255 273 335 332 332

266 222 225 255 273 314 314

221

265

221

204

255

315 315 281

101

Method detection limits (MDLs) for the PSTs were calculated using commercially available standards. For toxins that were not detected in the samples, the limits were determined using standard addition techniques and a signal-to-noise ratio of 3:1. For toxins that were positively identified in the matrix, MDLs were estimated using a conservative standard addition approach. Detection limits for the following toxins are reported in parentheses with single ion monitoring (SIM) followed by selective reaction monitoring (SRM) MDLs: STX (5 mg g1, 10 mg g1), dcSTX (5 mg g1,13 mg g1), 1), dcNEO (20 mg g1, 25 mg g1), dcGTX2 (19 mg g1, 40 mg g1), dcGTX3 (6 mg g1, 40 mg g1), GTX1 (10 mg g1, 20 mg g1), GTX4 (12 mg g1, 20 mg g1), GTX5 (5 mg g1,25 mg g1) and NEO (9 mg g1, not calculated). 3. Results

Note: Data adopted from Dell’Aversano et al. (2005) except product ions (determined from autotune, XcaliburTM software). a No known standards available; data adopted and modified from Onodera et al. (1997).

to maximize homogenization and decrease lyophilization time. Samples were frozen (20 8C) in freeze flasks and lyophilized at 50 8C (Thermo Savant Modulyo Freeze Dryer System, Thermo Fisher Scientific). Lyophilized 1-g subsets were extracted in 50 mL of 0.1 M acetic acid at 100 8C  5 8C with constant stirring for 5 min. Once cooled to room temperature, the samples were brought back to the original volume with the addition of 0.1 M acetic acid. All samples were filtered with Whatman PuradiscTM 0.45 mm PVDF syringe filters (Whatman Inc. Piscataway, NJ) and stored at 4 8C prior to analysis. 2.6. LC/MS analysis A Thermo FinniganTM Surveyor HPLC system coupled with a Thermo FinniganTM LCQ Advantage MSn ion trap tandem mass spectrometer was utilized for analysis of PSTs in samples. A TSKgel1 Amide 80 250 mm  2 mm HPLC Column (5 mm particle size) was employed for chromatographic separation (Tosoh Bioscience LLC, Grove City, OH). Two mobile phases were used, solvent A: 100% DI with 3.6 mM formic acid and 2 mM ammonium formate and solvent B: 95% (v/v) acetonitrile with 3.6 mM formic acid and 2 mM ammonium formate. The elution gradient was as follows; hold at 35% A for 2 min, 35–70% A over 10 min, 70% held for 5 min, return to 35% A in 10 min and hold for ten additional minutes for re-equilibration. PSTs utilized in MS/MS analyses were directly infused into the mass spectrometer and autotuned with XcaliburTM software. Once MS/MS parameters and collision energies were optimized, scans (150–500 m/z), single ion monitoring (SIM), and selective reaction monitoring (SRM) were conducted on standards and samples from 20 mL injections. Monitored ions are given in Table 2. Concentrations of decarbamoyl toxins detected in the matrix were determined from standard addition techniques and quantification was conducted utilizing SIM. SRM was conducted only for purposes of validation of toxins present in samples above the detection limits.

3.1. Field measurements The ranges of values for field-measured parameters are shown in Table 3. Temperature and pH values showed very little spatial or temporal variation, which is common in Florida spring systems. Specific conductance varied between springs, ranging from near 2000 mS cm1 at Silver Glen Spring, to near 300 mS cm1 at Blue Hole. Dissolved oxygen levels were below saturation at both springs throughout the sampling period, as commonly observed near vents in spring systems (Harrington et al., 2010). All samples were collected at approximately 1 m in depth. The overstory densities were different from site to site. In Silver Glen Springs, the Main Stem collection area was free from canopy coverage (0% canopy coverage), while Natural Well was surrounded by an embankment with overstory density averaging 14%. Blue Hole had the highest canopy coverage averaging 81% for the study period. 3.2. Taxonomic analysis All 24 collected samples were dominated by the filamentous cyanobacterium L. wollei Farlow ex Gomont (syn. Plectonema wollei), as determined via microscopy and by meeting criteria published in Speziale and Dyck (1992). In some samples, other Lyngbya species were identified (Lyngbya cf. aestuarii & Lyngbya cf. major), but comprised only a small fraction of the entire sample. Although other potential cyanobacteria and cyanotoxin producers were present in the samples, the dominant and universally observed cyanobacterial species in every sample was L. wollei (Fig. 3). Silver Glen Main Stem samples yielded the longest algal species list, followed by Blue Hole Spring and then Silver Glen Natural Well (Table 4). Fine oscillatorialean filaments reported in Table 4 were primarily Heteroleibleinia sp. with some Leibleinia sp. (epiphytic) and Leptolyngbya sp. (metaphytic). The primary epiphytes on L. wollei were the previously mentioned oscillatorialean filaments, the pennate diatom Cocconeis sp. and the chroococcalean cyanophyte Chamaesiphon sp. Relative abundances of mat algae varied from month to month but L. wollei remained the dominant species.

Table 3 Range of field measurements from all sample sites (August 2009–April 2010) (n = 6). Parameter

Blue Hole

SGS Main Stem

SGS Natural Well

Temperature (8C) pH (s.u.) Salinity (psu) Specific conductance (mS cm1) Dissolved oxygen (mg L1) Average canopy coverage (%) Average collection depth (m)

21.4–21.7 7.5–7.6 0.14–0.15 289–310 1.8–2.2 68–89% 1.0–1.0

23.1–23.3 7.8–8.0 0.92–0.97 1814–1885 3.5–4.1 0–0% 0.8–1.4

23.0–23.1 7.7–7.9 1.01–1.02 1949–2008 3.1–3.5 6–21% 0.7–1.2

102

A.J. Foss et al. / Harmful Algae 16 (2012) 98–107

EU586735) to another Blennothrix sequence (Accession: EU253968) in the NCBI database was around 89%. Although the 16S rRNA gene sequence of the filament in the current study showed a high identity (98%) to a sequence presumably obtained from Blennothrix, this is probably a misidentified species. Additionally, there was not a published article on this sequence and there are no available morphological descriptions for this species from which the sequence was obtained. The partial sxtA gene sequence (584 bp) obtained from the L. wollei filament (Accession: JQ282907) showed 97% identity to the corresponding region in L. wollei strain Carmichael/Alabama (Accession: EU629174), while identities to other sxtA genes was lower from 90% to 94% from saxitoxin producing species such as Cylindrospermopsis raciborskii, Aphanizomenon sp., Aphanizomenon flos-aquae, Aphanizomenon gracile, Aphanizomenon issatschenkoi, Anabaena circinalis, Anabaena planktonica, Anabaenopsis elenkinii, and Scytonema sp. A 672 bp sxtG sequence (Accession: JQ282908) was recovered from the L. wollei filament, which showed 98% identity to that of L. wollei strain Carmichael/Alabama (Accession EU629180). Identities to other sxtG sequences from other species were lower, ranging from 93% to 96%. 3.4. LC/MS

Fig. 3. Micrographs of Lyngbya wollei, scale bars are 50 mm: (A) exhibiting false branching with phase contrast optics (Blue Hole Spring), (B) bright field optics at 400 (Silver Glen Main Stem).

3.3. 16S rRNA and PST biosynthesis gene sequences A 1366 bp near full-length16S rRNA gene fragment (Accession: JQ282906) was recovered from the single L. wollei filament. A Basic Local Alignment Search Tool (BLAST) search against nucleotide sequences in the NCBI database revealed highest similarity to a Plectonema (Lyngbya) wollei 16S rRNA gene sequence (Accession: HQ419203) with 99% identity, followed by two other Plectonema (Lyngbya) wollei sequences (Accession: HQ419202 and HQ419204) with 98% identity. The 16S rRNA gene sequence obtained in this work was only 96% identical to that of L. wollei strain Carmichael/ Alabama (Accession: EU439567). Comparison of our single filament 16S rRNA gene sequence to others obtained from L. wollei mats in Florida springs (Joyner et al., 2008), revealed sequence identities ranging from 93% to 98% over approximately 380 bp. Interestingly, similarities to sequences obtained from Silver Glen springs were between 94% and 97%. The blast search also showed a Blennothrix sp. strain (Accession: EU586735) having 98% identity to our sequence over 1107 bp. However the identity of 16S rRNA gene sequence of this Blennothrix strain (Accession:

All samples extracted with 0.1 M acetic acid and analyzed with LC/MS contained decarbamoylgonyautoxin(s), decarbamoylsaxitoxin, and LWTs 1-6. All other toxins, including STX, dcNEO, GTX1, GTX4, and NEO, were either absent or present below the detection limits established in this study. Some samples collected from Blue Hole Spring and Natural Well Spring did indicate a presence of dcNEO via SIM, but the concentrations were below the MDL and could not be verified with SRM. There was no indication of NEO in any of the samples collected. Other PSTs commonly screened for, GTX 2, GTX 3, GTX 5, C1, and C2, were not detected when analyzed utilizing LC/fluorescence (unpublished data). Matrix effects were observed in all spiked samples when compared to standards. The effects included retention time shifts and reduced MS response. Efforts to minimize matrix effects utilizing SPE (C18), dilution, or by replacing the sample solution with 0.003 M HCl did not prevent retention time shifts or enhance MS response. These shifts were reproducible in all samples and care was taken to avoid misidentification of peaks. Retention times for peaks positively identified in the matrix, dcGTX2, dcGTX3, and dcSTX, shifted 0.4, 0.4, and 3.7 min (respectively). This was validated utilizing both pre- and post-extraction spikes with LC/ MS/MS. Toxins not detected in the samples, STX, dcSTX, dcNEO, NEO, GTX 1, GTX4 and GTX5, exhibited RT shifts of 0.4–5.7 min when compared to standards without the matrix. This was also confirmed utilizing standard addition techniques. Table 5 highlights shifts in retention time. Decarbamoylgonyautoxin 2&3 concentrations are shown in Figs. 4 and 5. The levels of dcGTX2 ranged from non-detectable to 60 mg g1 and dcGTX3 concentrations ranged from 9 to 28 mg g1 for all sites. Concentrations of decarbamoylsaxitoxin (dcSTX) at the three sampling sites are shown in Fig. 6, with a range of concentrations from 16 to 33 mg g1. All sites showed temporal variability in concentrations of dcGTX2, dcGTX3, and dcSTX, but there was insufficient data to clearly define temporal or spatial trends in toxin concentrations. Representative SCAN data showing SIM ions for L. wollei toxins from Blue Hole Spring can be seen in Fig. 7. L. wollei toxins LWT 2 and LWT 3 are isomers and resolution of the two toxins was not obtainable under the conditions of this study, so they were combined as LWT 2&3. As represented in Fig. 8, L. wollei toxin profiles were similar for each site. This data is somewhat limited

A.J. Foss et al. / Harmful Algae 16 (2012) 98–107

103

Table 4 Qualitative algae list from all samples, determined through microscopic observation. DATE

Silver Glen Main Stem

Silver Glen Natural Well

Blue Hole Spring

August 2009

Lyngbya wollei Fine oscillatorialean filaments spp. Spirogyra sp. 1 Microchaete sp. 1 Oedogonium sp. 1 Stigeoclonium sp. 1 Oedogonium sp. 2 Ulva flexuosa Cladophora glomerata Anabaena sp. 1 Oscillatorialean filament sp. 1 Pseudanabaena sp. 1

Lyngbya wollei Fine oscillatorialean filaments spp.

Lyngbya wollei Fine oscillatorialean filaments spp. Phormidium sp. 2 Phormidium sp. 3

September 2009

Lyngbya wollei Fine oscillatorialean filaments spp. Stigeoclonium sp. 1 Microchaete sp. 1 Oedogonium sp. 1 Anabaena sp. 1 Microchaete sp. 2

Lyngbya wollei Fine oscillatorialean filaments spp.

Lyngbya wollei Fine oscillatorialean filaments spp. Spirogyra sp. 1 Phormidium sp. 3 Phormidium sp. 2 Batrachospermum sp. 1 Mougeotia sp. 1

October 2009

Lyngbya wollei Fine oscillatorialean filaments spp. Anabaena sp. 1 Spirogyra sp. 1 Oedogonium sp. 2 Microchaete sp. 1 Stigeoclonium sp. 1 Oedogonium sp. 1

Lyngbya wollei Fine oscillatorialean filaments spp. Homoeothrix sp.

Lyngbya wollei Phormidium sp. 3 Fine oscillatorialean filaments spp. Phormidium sp. 2

November/December 2009

Lyngbya wollei Fine oscillatorialean filaments spp. a Lyngbya cf. aestuarii Lyngbya cf. major Phormidium sp. 1 Anabaena sp. 1 Phormidium sp. 2 Microchaete sp. 1 Rhizoclonium hieroglyphicum Stigeoclonium sp. 1 Oedogonium sp. 1 Spirogyra sp. 1

Lyngbya wollei Fine oscillatorialean filaments spp.

Lyngbya wollei Phormidium sp. 2 Fine oscillatorialean filaments spp. Homoeothrix sp. 1 Phormidium sp. 2 Phormidium sp. 4

January/February 2010

Lyngbya wollei Fine oscillatorialean filaments spp. Lyngbya cf. aestuarii Oscillatorialean filament sp. 2 Microchaete sp. 1 Mougeotia sp. 1

Lyngbya wollei Fine oscillatorialean filaments spp.

Lyngbya wollei Fine oscillatorialean filaments spp. Phormidium sp. 2 Chlorophyte filament sp. 1

April 2010

Lyngbya wollei Fine oscillatorialean filaments spp. Stigeoclonium sp. 1

Lyngbya wollei Fine oscillatorialean filaments spp.

Lyngbya wollei Fine oscillatorialean filaments spp. Phormidium sp. 3

a

L. aestuarii (Mertens) Liebmann sensu Prescott 1962; unrevised in Komarek and Anagnostidis.

however as it is derived only from MS responses; the standard materials necessary for quantification are not available. This lack of available toxin standards limits determinations as to which PSTs dominate the profile and comparisons of the relative presence of

Table 5 Retention time shifts observed in matrix spiked with standards. Toxin

Retention time (min) standard

Retention time shift (min)

STX dcSTX dcNEO dcGTX2 dcGTX3 GTX1 GTX4 GTX5 NEO

16.5 16.7 18.8 8.0 8.3 8.1 8.5 9.1 16.9

3.4 3.7 5.7 0.4 0.4 4.5 4.9 0.4 3.8

Fig. 4. dcGTX 2 concentrations for all sites over sampling period (reported in mg g1 dry weight of lyophilized L. wollei). Some samples fell below the MDL.

104

A.J. Foss et al. / Harmful Algae 16 (2012) 98–107

finding would change the relative toxicity of samples from each site. 4. Discussion 4.1. PSTs associated with Lyngbya dominated mats

Fig. 5. dcGTX 3 concentrations for all sites over sampling period (reported in mg g1 dry weight of lyophilized L. wollei). All samples were detected above the MDL.

Fig. 6. dcSTX concentrations for all sites over sampling period (mg g1-dry weight of lyophilized L. wollei). All samples were detected above the MDL for dcSTX.

toxins to estimates. An analysis of variance showed that the concentrations of L. wollei toxins 1, 4, 5, and 6 were not statistically significant between sites (P = 0.729, 0.098, 0.741 and 0.153, respectively). However, LWT 2&3 concentrations were significantly different between sites (P = 0.0034). Since LWT 2&3 are toxic constituents of the L. wollei toxin group, confirmation of this

The results of this study show that Lyngbya dominated mats in two spring fed rivers in Florida are a potential source of PSTs, with toxin profiles similar to Lyngbya mat samples collected from Guntersville Reservoir on the Tennessee River in Alabama (Carmichael et al., 1997; Onodera et al., 1997). PCR and sequencing also confirmed the presence of sxtA and sxtG genes involved in PST biosynthesis. Morphological examination verified the presence of L. wollei in all samples and the 16S rRNA gene sequence obtained from an isolated filament of Lyngbya showed highest identities to three L. wollei sequences in the NCBI database (98–99% identity). On the contrary, identity of our sequence to that of the L. wollei collected from the Tennessee River in Alabama, which also produces PSTs (Carmichael et al., 1997; Kellmann et al., 2008), was only 96%. Bacterial strains with less than 97.5% 16S rRNA gene sequence identity are generally regarded as different species (Stackebrandt and Goebel, 1994), therefore the L. wollei filament isolated in our study and the Alabama L. wollei strains are likely different species. Joyner et al. (2008) noted similar discrepancies and it is suggested that L. wollei identified in Florida’s spring systems might belong to two or more species. The single L. wollei filament 16S rRNA gene sequence determined in this study showed between 93% and 98% identity to sequences obtained from Joyner et al. (2008) Lyngbya mats. This raises some important questions with regard to the significance of morphological and molecular identification of L. wollei and the relationship to toxin production. Cultured isolates or genomic DNA from single filaments would be required for further clarification of this issue. The presence of PST biosynthesis genes in conjunction with the 16S rRNA gene sequence resulting from a single isolated filament of Lyngbya, provided additional evidence that L. wollei is the source of the PSTs observed in Silver Glen Springs Lyngbya mats. Toxins detected in the Florida Lyngbya dominated samples via LC/MS included dcGTX2, dcGTX3, dcSTX and LWTs 1-6. Concentrations of toxins (nmol g1-dry weight) in this study and that of Onodera et al. (1997) are shown in Table 6. The levels of dcSTX, dcGTX2, and dcGTX3 in the Florida samples were comparable to those reported by Onodera et al. (1997) from Alabama. Comparisons of concentrations of LWTs 1-6 in the Florida and Alabama samples were not possible due to the unavailability of appropriate

Fig. 7. HPLC/MS SCAN of Blue Hole Spring (collected 1/27/10) with SIM ions for L. wollei Toxins 1–6. All samples exhibited peaks for all LWT 1-6 SIM ions.

A.J. Foss et al. / Harmful Algae 16 (2012) 98–107

105

Fig. 8. Mass spectrometry response to monitored SIM ions of LWT 1-6 for all sites. Standard error bars represent the standard deviation of MS response data over the sampling period. There was a significant difference in LWT 2&3 levels between sites, a toxic variant of LWTs.

standards. Levels of decarbamoyl toxins appeared to correlate with temporal and spatial variability indicating that seasonal and regional differences may be potential driving factors, but further work would be needed to define these trends and relationships. Although the two spring systems chosen in this study represented both high and low specific conductance waters, as well as high and low overstory density areas, toxin profiles remained similar. Kellmann et al. (2008) proposed that sxtX, a portion of the saxitoxin biosynthesis gene cluster, coded for a cephalosporin hydroxylase homolog, which performs N-1 hydroxylation of the parent STX molecule forming NEO. Mihali et al. (2011) reported the presence of sxtX in the PST biosynthesis gene cluster of L. wollei. However, they could not verify the presence of NEO in their freeze-dried bloom material. Similar to the samples analyzed in Onodera et al. (1997), NEO was also not detected in any of our samples.

4.2. The toxic threat associated with Lyngbya dominated mats A central question with regard to PSTs is the nature of potential ecosystem and human health threats posed by L. wollei dominated mats, which are not limited to the two springs sampled in this study, but are prevalent in many Florida springs (Stevenson et al., 2007; Quinlan et al., 2008; Joyner et al., 2008). PSTs have been detected using immunological methods (i.e. ELISA) in L. wolleidominated mats collected from a number of Florida springs in conjunction with work conducted by the FDOH (Florida Department of Health), including Wakulla Springs, Fanning Springs, Homosassa Springs, Alexander Springs, Juniper Springs and multiple springs in the Ichetucknee Spring System (PBSJ [Post, Buckley, Schuh, and Jernigan] 2007, FDOH-personal communication). It is therefore clear that the potential for exposure to PSTs associated with L. wollei-dominated mats in Florida is widespread.

A.J. Foss et al. / Harmful Algae 16 (2012) 98–107

106

Table 6 Average and range of toxin concentrations in nmol per gram-dried material of mats collected from Alabama (Onodera et al., 1997) compared with toxin concentrations measured in Florida samples. Toxin

Alabama Lyngbyaa (nmol g1)

SGS main stem (nmol 1)

SGS natural well (nmol g1)

Blue Hole Spring (nmol g1)

dcSTX dcGTX2 dcGTX3 b LWT 1 LWT 2 LWT 3 b LWT 4 LWT 5 b LWT 6

40 110 41 690 390 920 20 500 97

98 (68–129) 30 (ND-121) 37 (21–48) – – – – – –

102 (64–113) 86 (ND-170) 59 (32-80) – – – – – –

89 (65–118) 113 (82–146) 64 (45–79) – – – – – –

–Could not quantify without available standards ND = not detected above the detection limit a Data from Onodera et al. (1997). b Non-toxic derivatives.

The extent, specific nature, and seriousness of that potential remain uncertain. The relative toxicity of the levels of PST observed in this study can be viewed from the context of previous studies of the toxins involving mouse bioassays. Table 7 shows the toxicity of PSTs from L. wollei mats as reported by Oshima (1995) and Onodera et al. (1997) in Mouse Units (MU) mmol1 of toxin, using the AOAC Method for Paralytic Shellfish Poisons, i.e. 20 g ddY strain male mice. Saxitoxin, exhibits the highest toxicity at 2483 MU mmol1 of toxin, based on research conducted by Oshima (1995). Although other toxicity levels have been reported for saxitoxin, Oshima’s values are used in this paper due to the similarity in the methods and source materials used for analyses. Lethal doses of toxic samples of L. wollei collected in Alabama ranged from 150 mg of lyophilized L. wollei to nontoxic quantities of 1500 mg lyophilized L. wollei per kg mouse, with reported saxitoxin equivalents from 0 to 58 mg STX-eq g(dry weight)1. In this study, total STX-eqs could not be determined for LWTs 1-6, but STX-eqs based on dcSTX & dcGTX2&3 concentrations represent a toxicity range of 19–73 mg STX-eq g(dry weight)1. Based on these findings, L. wollei mats collected from Florida springs are comparable in toxicity, or are more toxic, than L. wollei mats from Alabama. There is considerable uncertainty regarding the actual risk of human exposure to the PSTs produced by L. wollei. Direct ingestion of Lyngbya by humans is highly unlikely; however, there are other potential means of exposure. One avenue of exposure may be consumption of aquatic animals that have bioaccumulated PSTs trophically. PSTs are known to bioaccumulate in marine and freshwater bivalves through filter-feeding of phytoplankton, and

Table 7 Toxicity of PST variants reported in MU mmol1 and relative toxicity to STX (STX-eq) of variants confirmed in L. wollei dominated mat samples from Florida. Toxin

a

Mouse units (MU) (mmol1)

Relative toxicity

a

2483 1274 1617 1872 <10 178 52 <10 326 <10

1.00 0.51 0.65 0.75 – 0.07 0.02 – 0.13 –

STX dcSTX a dcGTX2 a dcGTX3 b LWT 1 b LWT 2 b LWT 3 b LWT 4 b LWT 5 b LWT 6 a

– Non toxic. a Oshima (1995) data. b Onodera et al. (1997) data.

even in freshwater fish through exposure to cyanobacteria blooms (Negri and Jones, 1995; Smith et al., 2001, Galva˜o et al., 2009), but little is known about what organisms graze on L. wollei or whether PSTs from L. wollei can transfer to other organisms. Another means of exposure may be through the release of PSTs into the water column through leakage or senescence of L. wollei filaments. PSTs do not degrade quickly in buffered freshwater systems, such as the freshwater springs of Florida, and may remain intact until photolytic or bacterial degradation acts on the toxins (Jones and Negri, 1997). As levels of accessible drinking water in Florida and other regions decline, sources of surface water with L. wollei present may be tapped to support growing needs, thereby increasing the risk of human exposure to PSTs. On the other hand, the high flow rates in spring systems may rapidly dilute released toxins. One of the most common current means of human exposure to PSTs from L. wollei is likely direct contact with mats. For example, physical removal of L. wollei mats by water managers may involve occupational exposure related to handling, treating, removing, or disposing of mat material. Recreational exposure is also of concern since direct contact with L. wollei dominated mats in spring runs is likely to occur. PSTs may be limited in their ability to cross the dermal layer due to hydrophilic properties, but individuals with lesions or cuts in the skin may be at greater risk. To date, there are no confirmed cases directly linking PSTs from L. wollei-dominated mats to symptomatic responses. It is unknown if this is due to actual risks or if this is because epidemiological data is limited. Currently, data collected by the FDOH in areas dominated by L. wollei is focused on dermatoxic and gastrointestinal responses, with a lack of focus on neurotoxic responses, which may be mild. Questionnaires and reports would have to be extended to include neurotoxic responses associated with PST exposure, such as numbness, tingling sensations, headaches, nausea, blurred vision, and muscle incoordination. Epidemiological data of this nature might help to provide a better understanding of risks related to PSTs derived from L. wollei in Florida and other southeastern US spring systems. Acknowledgments The authors would like to acknowledge the assistance of Dr. Mark Aubel (GreenWater Laboratories), Dr. Nancy Szabo (Analytical Toxicology Core Laboratory, University of Florida), Dr. Karl Havens (chair, Fisheries and Aquatic Sciences, University of Florida) and Andrew Reich (Florida Department of Health). The authors also thank Alicia Plakotaris and Johnny May for their help in the field. The work was supported by GreenWater Laboratories/ CyanoLab, Palatka, FL.[SS]

A.J. Foss et al. / Harmful Algae 16 (2012) 98–107

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.hal.2012.02.004. References Berry, J.P., Gantar, M., Gawley, R.E., Wang, M., Rein, K.S., 2004. Pharmacology and toxicology of pahayokolide A, a bioactive metabolite from a freshwater species of Lyngbya isolated from the Florida Everglades. Comp. Biochem. Phys. C: Toxicol. Pharmacol. 139, 231–238. Carmichael, W.W., Evans, W.R., Yin, Q.Q., Bell, P., Moczydlowski, E., 1997. Evidence for paralytic shellfish poisons in the freshwater cyanobacterium Lyngbya wollei (Farlow ex Gomont) comb. nov. Appl. Environ. Microbiol. 63, 3104. Cowell, B.C., Botts, P.S., 1994. Factors influencing the distribution, abundance and growth of Lyngbya wollei in central Florida. Aquat. Bot. 49, 1–17. Dell’Aversano, C., Hess, P., Quilliam, M.A., 2005. Hydrophilic interaction liquid chromatography–mass spectrometry for the analysis of paralytic shellfish poisoning (PSP) toxins. J. Chromatogr. A. 1081, 190–201. Galva˜o, J.A., Oetterer, M., Bittencourt-Oliveira, M.C., Barros, S.G., Hiller, S., Erler, K., Luckas, B., Pinto, E., Kujbida, P., 2009. Saxitoxins accumulation by freshwater tilapia (Oreochromis niloticus) for human consumption. Toxicon 54 (6), 891– 894. Harrington D., Maddox G., Hicks R., 2010. Florida springs initiative monitoring network report and recognized sources of nitrate. Florida Department of Environmental Protection; Division of Environmental Assessment and Restoration Bureau of Watershed Restoration, pp. 1–113. Jones, G.J., Negri, A.P., 1997. Persistence and degradation of cyanobacterial paralytic shellfish poisons (PSPs) in freshwaters. Water Res. 31, 525–533. Joyner, J.J., Litaker, R.W., Paerl, H.W., 2008. Morphological and genetic evidence that the cyanobacterium Lyngbya wollei (Farlow ex Gomont) Speziale and Dyck encompasses at least two species. Appl. Environ. Microbiol. 74, 3710–3717. Kao, C.Y., 1993. Paralytic shellfish poisoning. In: Falconer, E.R. (Ed.), Algal Toxins in Seafood and Drinking Water, Academic, London, UK, pp. 75. Kao, C.Y., Levinson, S.R., 1986. Tetrodotoxin, saxitoxin, and the molecular biology of the sodium channel. In: Boland, B., Cullinan, J., Cohn, T. (Eds.), Annals of the New York Academy of Science. New York Academy of Science, New York, pp. 1–445. Kellmann, R., Michali, T., Neilan, B., 2008. Identification of a saxitoxin biosynthesis gene with a history of frequent horizontal gene transfers. J. Mol. Evol. 67, 526–538. Landsberg, J.H., 2002. The effects of harmful algal blooms on aquatic organisms. Rev. Fish Sci. 10 (2), 113–390. Mihali, K.T., Carmichael, W.W., Neilan, B., 2011. A putative gene cluster from a Lyngbya wollei bloom that encodes paralytic shellfish toxin biosynthesis. PLoS One 6 (2), e14657, doi:10.1371/journal.pone.0014657. Negri, A.P., Jones, G.J., 1995. Bioaccumulation of paralytic shellfish poisoning (PSP) toxins from the cyanobacterium Anabaena circinalis by the freshwater mussel Alathyria condola. Toxicon 33, 667–678.

107

Onodera, H., Satake, M., Oshima, Y., Yasimoto, T., Carmichael, W.W., 1997. New saxitoxin analogues from the freshwater filamentous. Cyanobacterium Lyngbya wollei. Nat. Toxins 5 (4), 146–151. Oshima, Y., 1995. Post-column derivatization liquid chromatographic method for paralytic shellfish toxins. J. AOAC Int. 78, 528–532. PBS&J [Post, Buckley, Schuh and Jernigan, Inc.]. 2007. Ecological and Toxicological Assessment of Lyngbya in Florida Springs. Florida Department of Health, Tallahassee, Florida, USA. http://www.myfloridaeh.com/medicine/aquatic/ index.html. Quinlan, E.L., Phlips, E.J., Donnelly, K.A., Jett, C.H., Sleszynski, P., Keller, S., 2008. Primary producer biomass and nutrient loading in Silver Springs, Florida, USA. Aquat. Bot. 88, 247–255. Seifert, M., McGregor, G., Eaglesham, G., Wickramasinghe, W., Shaw, G., 2007. First evidence for the production of cylindrospermopsin and deoxy-cylindrospermopsin by the freshwater benthic cyanobacterium, Lyngbya wollei (Farlow ex Gomont) Speziale and Dyck. Harmful Algae 6, 73–80. Sivonen, K., Jones, G., 1999. Cyanobacterial toxins. In: Chorus, I., Bertram, J. (Eds.), Toxic Cyanobacteria in Water: A Guide to Public Health, Significance, and Monitoring and Management, The World Health Organization, E & FN Spon, London, UK, ISBN: 0-419-23930-8pp. 41–111. Smith, E.A., Grant, F., Ferguson, C.M., Gallacher, S., 2001. Biotransformations of paralytic shellfish toxins by bacteria isolated from bivalve molluscs. Appl. Environ. Microbiol. 67, 2345–2353. Speziale, B.J., Dyck, L.A., 1992. Lyngbya infestations: Comparative taxonomy of Lyngbya wollei Comb. Nov. (cyanobacteria). J. Phycol. 28, 693–706. Stackebrandt, E., Goebel, B.M., 1994. Taxonomic note: a place for DNA–DNA reassociation and 16S rRNA sequence analysis in the present species definition in bacteriology. Int. J. Syst. Bacteriol. 44, 846–849. Stevenson, R.J., Pinowska A., Albertin A., Sickman J., 2007. Ecological Condition of Algae and Nutrients in Florida Springs: The Synthesis Report. Environmental Assessment Section, Florida Department of Environmental Protection, pp. 1–58. Su, Z., Sheets, M., Ishida, H., Li, F., Barry, W.H., 2004. Saxitoxin blocks L-type ICa. J. Pharmacol. Exp. Ther. 308 (1), 324–329. Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: molecular evolutionary genetics analysis (MEGA) Software Version 4.0. Mol. Biol. Evol. 24, 1596– 1599. Teneva, I., Asparuhova, D., Dzhambazov, B., Mladenov, R., Schirmer, K., 2003. The freshwater cyanobacterium Lyngbya aerugineo-coerulea produces compounds toxic to mice and to mammalian and fish cells. Environ. Toxicol. 18, 9. Tillett, D., Neilan, B.A., 2000. Xanthogenate nucleic acid isolation from cultured and environmental cyanobacteria. J. Phycol. 36, 251–258. Wiese, M., D’Agostino, P.M., Mihali, T.K., Moffitt, M.C., Neilan, B.A., 2010. Neurotoxic alkaloids: saxitoxin and its analogs. Mar. Drugs 8 (7), 2185–2211. Yin, Q., Carmichael, W.W., Evans, W.R., 1997. Factors influencing growth and toxin production by cultures of the freshwater cyanobacterium Lyngbya wollei Farlow ex Gomont. J. Appl. Phycol 9, 55. Yilmaz, M., Phlips, E.J., Szabo, N.J., Badylak, S., 2008. A comparative study of Florida strains of Cylindrospermopsis and Aphanizomenon for cylindrospermopsin production. Toxicon 51, 130–139. Yilmaz, M., Phlips, E.J., 2011. Toxicity and genetic diversity of Cylindrospermopsis raciborskii in Florida, USA. Lake Reserv. Manage. 27 (3), 235–244.