Harmful Algae 10 (2010) 88–97

Evidence for saxitoxins production by the cyanobacterium Aphanizomenon gracilein a French recreational water body

A. Ledreux a, S. Thomazeau a, A. Catherine a, C. Duval a, C. Yepremian a, A. Marie b, C. Bernard a,*a FRE 3206 CNRS/MNHN Molecules de Communication et Adaptation des Micro-organismes, USM 505 Cyanobacteries, Cyanotoxines et Environnement,

Museum National d’Histoire Naturelle, 57 rue Cuvier, CP 39, 75231 Paris Cedex 05, Franceb Plateforme de Spectrometrie de masse, FRE 3206 CNRS/MNHN Molecules de Communication et Adaptation des Micro-organismes,

Museum National d’Histoire Naturelle, 57 rue Cuvier, CP 54, 75231 Paris Cedex 05, France

A R T I C L E I N F O

Article history:

Received 24 September 2009

Received in revised form 23 May 2010

Accepted 30 July 2010

Keywords:

Aphanizomenon

Cyanobacteria

Freshwater

Saxitoxins

A B S T R A C T

In the last few years, a scientific consensus has emerged regarding the increasing frequency and severity

of cyanobacterial blooms in freshwater environments. Consequently, recreational and drinking water

bodies are now increasingly monitored by local authorities to prevent animal and human poisoning

related to cyanobacteria and their toxins. A survey of a water body used for recreational activities (at

Champs-sur-Marne, Paris suburb) was conducted from 2005 to 2008. The chlorophyll a concentrations,

the occurrence of toxic cyanobacteria and of the cyanotoxins microcystins, anatoxin-a and saxitoxins

were recorded. The potentially toxic genera Aphanizomenon, Planktothrix and Microcystis were observed

throughout the monitoring period. Microcystis was the dominant genus during the first two years (2005–

2007), whereas Aphanizomenon became the dominant genus towards the end of the survey (2007–2008).

As a consequence, microcystins were initially detected at high concentrations (max. = 89 mg equiv MC-

LR L�1 in raw water), and were later superceded by saxitoxins (STXs) in 2008. Using LC–MS/MS and the

Neuro-2a cell-based assay, we confirmed that the STX and neoSTX variants were present in the natural

samples at concentrations higher than 3 mg equiv STX L�1 in raw water. To attribute STXs production to

the appropriate cyanobacterial taxa, monoclonal cultures were obtained after an isolation procedure

then screened for the presence of the sxtA gene. SxtA-positive strains were analyzed for their STX content

using the Neuro-2a cell-based assay and LC–MS/MS. The results led to consider Aphanizomenon gracile as

the most likely producer of STXs. The first description of STXs in a French recreational water body lead to

discuss the health risks associated with the presence of STXs alone or with its co-occurrence with other

known cyanotoxins in recreational water bodies.

� 2010 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Harmful Algae

journal homepage: www.e lsev ier .com/ locate /ha l

1. Introduction

In freshwater ecosystems, potentially toxic cyanobacteria are asource of increasing concern as they may be associated to healthrisks for animals and human beings (Hitzfeld et al., 2000). A scientificconsensus has emerged regarding the impact of global changes(increasing human pressures and climate change) on the frequencyand severity of cyanobacterial blooms (Paerl and Huisman, 2008).The recent concerns regarding the monitoring of cyanobacteriaproliferation and their potential toxicity have led to the identifica-tion of a large number of areas contaminated with knowncyanotoxins microcystins (MCs), anatoxins (ANTX), saxitoxins(STXs) and cylindrospermopsin (CYN). To date, 67 species havebeen reported as producers of these cyanotoxins, which are also

* Corresponding author. Tel.: +33 1 40 79 31 83; fax: +33 1 40 79 35 94.

E-mail address: cbernard@mnhn.fr (C. Bernard).

1568-9883/$ – see front matter � 2010 Elsevier B.V. All rights reserved.

doi:10.1016/j.hal.2010.07.004

broadly classified into hepatotoxins, neurotoxins, and dermatotox-ins (Sivonen and Jones, 1999; Fristachi and Sinclair, 2008).

Recreational and drinking water bodies are monitored by therelevant authorities to prevent animal and human intoxicationsrelated to cyanobacteria and their toxins. In France, safetyguidelines for monitoring water resources fall under the purviewof different recommendations (Administrative Orders DGS/SD7A2003/270, 2004/364, 2005/304) that are implemented by Govern-mental safety institutions. The recommended monitoring pro-grams include evaluating the total phytoplankton biomassestimated from the chlorophyll a concentrations, identifying thecyanobacterial genera or species present, and measuring the in situ

MCs concentrations. The growing concern of the scientificcommunity and of health agencies has led to increased effortsto study the occurrence of cyanobacteria and identify additionalcyanotoxin and cyanotoxin producers.

In France, hepatotoxic microcystins associated with Plankto-

thrix or Microcystis blooms have been commonly reported (e.g.

A. Ledreux et al. / Harmful Algae 10 (2010) 88–97 89

Catherine et al., 2008a; Briand et al., 2009). A few cases of incidentsinvolving neurotoxins have also been reported. For example,anatoxin-a and homoanatoxin-a were detected in French rivers(Gugger et al., 2005; Cadel-Six et al., 2007), and benthiccyanobacteria were identified as the neurotoxins producers.Canine neurotoxicosis has also been directly linked to the presenceof these microorganisms (Gugger et al., 2005; Cadel-Six et al.,2007). However, to date, no cases of neurotoxic cyanobacterialbloom associated to saxitoxins have been reported.

The STXs are a group of neurotoxins consisting of at least 30carbamate alkaloid compounds (Araoz et al., 2009). This large familyof toxin is referred to as the Paralytic Shellfish Poisoning (PSP) toxinsin the marine environment. They can be divided into 3 groupsaccording to their degree of sulfatation, which is, in turn, related totheir specific toxic effects in the mouse (Llewellyn, 2006). Numerouscases of human poisoning due to STXs have been reported in marineareas, where these toxins are bioaccumulated by filter-feedingmolluscs, such as mussels and oysters. In freshwater environments,no cases of poisoning related to the consumption of these organismshave been reported so far, but there is some evidence that STXs canbioaccumulate in the freshwater mussel Alathyria condola (Negri andJones, 1995), in snails (Pomacea patula catemacensis) (Berry and Lind,2009) and in the freshwater fish Tilapia (Oreochromis niloticus)consumed by humans (Galvao et al., 2009). Nevertheless, exposureto STXs in freshwater recreational areas should also be consideredcarefully, since adverse health effects on children (e.g. fever, eyeirritation, abdominal pain, skin rash) have been reported in Finnishlakes (Rapala et al., 2005).

The methods used to detect and quantify STXs are based ontheir chemical structure, and include (i) immunoassays (Jellettet al., 2002) and (ii) chromatographic methods using availablechemical certified standards (n = 9 among more than 30 variantsalready reported).

Alternative methods based on the mechanisms of toxicity of theSTXs have also been proposed. STXs act on sodium voltage-gatedchannels of nerve cells. After binding to the site 1 of the sodiumchannels (Cestele and Catterall, 2000), sodium transport is blocked,resulting in the inhibition of axonal depolarization, and thusleading to muscle paralysis. These toxicological effects have led tothe widespread use of mouse bioassays to monitor STXs inshellfish. However, for ethical reasons and due to its poorreproducibility, this approach should be replaced by alternativemethods. The Neuro-2a cell line contains high levels of sodiumvoltage-gated channels (Kogure et al., 1988), and has been

[(Fig._1)TD$FIG]

Fig. 1. Geographic location of the Champs-sur-Marne water body within the Ile-de-Franc

Champs-sur-Marne water body) and location of sampling sites 1 and 2 (the dashed lin

employed for several years to detect STXs. The Neuro-2a cell-based bioassay has been shown to be an effective method fordetecting and quantifying cyanobacterial STXs (Humpage et al.,2007).

Finally, STXs can also been monitored by detecting STXs-producing algal blooms using molecular methods. The saxitoxinbiosynthesis pathway has recently been described in cyanobac-teria (Kellmann et al., 2008). The use of molecular probes targetingSTXs-producing cyanobacteria may provide useful early warningsystems of neurotoxic blooms.

In this study, the first evidence of an STX-contaminatedfreshwater body in France is presented. This new finding emergedfrom a regional survey of cyanobacteria and cyanotoxins in Ile-de-France area. Fifty water bodies were sampled and screened for thepresence of toxic cyanobacteria and MCs, ANTX, CYN and STXs(Catherine et al., 2008b). Special attention was focused on thewater body of the Champs-sur-Marne recreational area becausecyanobacteria were known to form summer blooms since 2005,which has aroused the interest of the local authorities in assessingthe risk linked to cyanotoxins. This paper focuses on the detectionof STXs in environmental samples collected from this water body,and on the identification of the causative cyanobacterial species.The implications of STXs occurrence for the health risk assessmentof recreational water bodies are also discussed.

2. Materials and methods

2.1. Sampling site

The recreational area of Champs-sur-Marne (4885105000N,283505300E) is in the outer suburb of Paris (France). This reservoiris a small (10.3 ha) and shallow (mean depth equal to 2.4 m) waterbody, which is isolated from the river system (Fig. 1), andoriginated from a disused sand pit. It is now a recreational areaused for bathing, kayaking, sailing and windsurfing.

The management committee of the Champs-sur-Marne recrea-tional area defined the sampling schedule according to the usesmade of the water body. It is used for sailing from April to Octoberand for bathing from late June to late August. As a consequence,sampling was performed weekly during the summer season,fortnightly in spring and summer and monthly during winter.

Site 1 corresponding to a bathing area separated from the rest ofthe water body by a mesh barrier (Fig. 1) was monitored from June2005 to December 2008 (n = 97 samples). Site 2, used for bathing

e region (open circles: water bodies included in the Cyanotox program; black circle:

e shows the mesh separating the bathing area from the rest of the water body).

A. Ledreux et al. / Harmful Algae 10 (2010) 88–9790

activities, was sampled from April 2006 to December 2008 (n = 38samples).

2.2. Sample collection and preparation

Four liters of sub-surface raw water collected at each stationwere placed in plastic bottles, kept refrigerated and protected fromlight. Phytoplankton biomass was concentrated by filtering 300 mLof raw water on GF-C filters (Whatman, 47 mm). The filters werethen stored frozen at �24 8C until being used for chlorophyll a

determination and toxin analysis. The remaining raw water wasused for identification and isolation procedures.

On 8 September 2008, surface scum was also collected, placedin plastic bottles, kept refrigerated and protected from light duringtransportation.

2.3. Study of the phytoplankton community

The chlorophyll a (Chl a) concentration was measured aftermethanol/water extraction (90/10, v/v) of frozen filters (Tallingand Driver, 1963), and analyzed using a Cary 50 Scan spectropho-tometer (Varian Inc., Palo Alto, USA).

The cyanobacterial genera or species (>20 mm) were deter-mined using samples fixed in buffered formaldehyde (5%, v/v) asdescribed previously (Yepremian et al., 2007). The identification ofthe phytoplankton was based on current literature (Komarek andAnagnostidis, 1989, 1998, 2005). For each sample, the number oftaxa present was assessed, and relative taxon abundances weredetermined by counting 400 individuals using a Malassez countingchamber. Cyanobacteria were considered as the dominantsubpopulation if they accounted for more 30% of the totalphytoplankton.

2.4. Isolation and obtaining of monoclonal cyanobacterial cultures

The water samples or scum were inoculated on a Z8 minusnitrogen (Z8-X) (Kotai, 1972) solid medium (7 g L�1 of agar).Isolations were carried out by repeated transfers of singlefilaments onto solid media (at least 10 steps) under an invertedmicroscope (Olympus CK2, Japan).

Growing clones were later cultured in 25 cm2 culture flasks(Nunc, Denmark) containing 10 mL of Z8-X medium. Strains weremaintained in the Paris Museum Collection (PMC) at 25 8C, usingdaylight fluorescent tubes providing an irradiance level of16 mmol photons m�2 s�1, with a photoperiod of 16L:8D. Strainsand cultures were monoclonal, non-axenic and named in the ParisMuseum Collection. Photographs and measurements were takenusing a microscope coupled to a Digital Sight DS-L1 imageacquisition system (Nikon Inc., Japan).

2.5. Molecular biology analysis

2.5.1. PCR and sequencing

The 16S rDNA genes were amplified with primers 8f, 920r, 861fand 1380r (Lane, 1991; Gugger et al., 2002; Gugger and Hoffmann,2004). Amplifications of the partial sxtA gene were performed withprimer pair sxtAF (50-AGG-TCT-TGA-CTT-GCA-TCC-AA-30) andsxtAR (50-AAC-CGG-CGA-CAT-AGA-TGA-TA-30), which weredesigned for this study with support from sxtA sequences availablein GenBank (Kellmann et al., 2008). The strain Anabaena circinalis

AWQC ANA311E maintained as PMC 00.14 in the Paris MuseumCollection since 2000 and known as an STXs producer (Teste et al.,2002) has been used as a positive PCR control. PCR amplificationwas performed in 20 mL reaction mixture containing 5 mL ofconcentrated culture (initial optical density at 750 nm was about 2absorbance units) or 1 mL of genomic DNA extracted following

procedure described in Briand et al. (2008), 250 mM of dNTPs,25 mM of each primer, 2.5 mL of 10X PCR Buffer with 2.5 mL ofMgCl2 (25 mM) and 1 U of GoldStar DNA polymerase (Eurogentec,San Diego, USA). The PCR reactions were performed with a PeltierThermal Cycler PTC-200 (MJ Research Inc., Waltham, MA, USA),and consisted, after a preliminary 5-min denaturing step at 94 8C,of 35 cycles at 94 8C for 1 min, at 58 8C for 1 min, and at 72 8C for2 min. An extension period was performed for 5 min at 72 8C,followed by a final 10-min period at 10 8C. The amplicons wereresolved on 1.5% agarose gels.

PCR products were directly sequenced by Genoscreen (Lille,France). Two sequence readings were performed for each amplifiedfragment, giving partial sequences of 1280 bp for 16S rDNA and602 bp for sxtA. Accession numbers for the sequences of the strainsare from HQ157685 to HQ157700 (16S rDNA) and from HQ157701to HQ157716 (sxtA).

2.5.2. Alignment and data set

The sequences included in the alignments were either newlygenerated for the present study or originated from previouslypublished data. In the latter case, only complete sequences wereselected from GenBank. Sequence alignment for each gene wasperformed by a multiple alignment process (ClustalW, version 1.7),and refined manually with Mega software version 3.1 (Kumar et al.,2004).

The 16S rDNA data set was intended to confirm morphologicalidentifications of Aphanizomenon gracile and Aphanizomenon

aphanizomenoides strains. Twenty Nostocaceae strains withpublished and complete 16S rDNA sequences were selected fromGenBank (Supplementary data S1). The sxtA data set wasincremented with four sxtA sequences from GenBank (Supple-mentary data S1). For both alignments, Lyngbya wollei sequencewas chosen as outgroup (Thomazeau et al., 2010).

2.5.3. Phylogenetic analysis

The phylogenetic analyses were run with MrBayes version 3.1.1(Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck,2003), based on Bayesian inference. The Bayesian inferences wereperformed with the models of DNA evolution chosen with the helpof the Modeltest 3.06 software (Posada and Crandall, 1998). Themodels chosen were calculated using the AIC (Akaike InformationCriterion). Metropolis-Coupled Markov Chain Monte Carlo(MCMCMC) analyses were conducted over 2,000,000 generations,with four incrementally chains. The default parameters were usedfor temperature and swapping. Parameter values and trees weresampled every 100 generations. The burn-in was set to 2000 treesor 200,000 generations for each run, and the remaining trees wereused to compute the consensus tree. The trees were edited usingTreeView version 1.6.1 (Page, 1996).

2.6. Screening for cyanotoxins

During the monitoring of Champs-sur-Marne recreationalwater body which started in 2005, a number of samples wereanalyzed for their cyanotoxin content. Samples were selected forthis analysis in accordance with the WHO guidelines (WHO, 2003),i.e. (i) when Chl a concentrations exceeded 10 mg L�1, (ii) whencyanobacteria were dominant (>30% of the total abundance) in thephytoplankton and (iii) when cyanobacteria genera known to bepotentially toxic were observed in the samples. For example, whenMicrocystis spp. were dominant in the phytoplankton, the MCscontent was determined.

2.6.1. Screening for MCs

MCs contents were estimated after methanolic extraction offrozen filters using the protein phosphatase 2A inhibition assay

Ta

ble

1C

hlo

rop

hy

lla

con

cen

tra

tio

ns,

nu

mb

er

of

cya

no

ba

cte

ria

lta

xa

,occ

urr

en

cea

nd

rela

tiv

efr

eq

ue

ncy

of

Ap

ha

niz

om

eno

nsp

p.a

nd

Mic

rocy

stis

spp

.,a

nd

ma

xim

um

lev

els

of

MC

sa

nd

ST

Xs

for

Sit

e1

(a)

an

dS

ite

2(b

)fo

re

ach

ye

ar

du

rin

gth

e

mo

nit

ori

ng

of

the

Ch

am

ps-

sur

Ma

rne

wa

ter

bo

dy

.

Ye

ar

Nu

mb

er

of

sam

ple

s

Ch

la

( mg

L�1)

Cy

an

ob

act

eri

aA

ph

an

izo

men

on

spp

.M

icro

cyst

issp

p.

MC

s

( mg

eq

uiv

MC

-LR

L�1)

ST

Xs

(mg

eq

uiv

ST

XL�

1)

Me

an

Ma

xC

V%

To

tal

nu

mb

er

of

tax

aO

ccu

rre

nce

Re

lati

ve

fre

qu

en

cy(%

)O

ccu

rre

nce

Re

lati

ve

fre

qu

en

cy(%

)

Ma

xM

ax

(a)

Sit

e1

20

05

22

20

01

59

01

95

13

94

11

35

98

9<

LOD

20

06

27

40

17

51

10

18

51

91

86

75

7<

LOD

20

07

24

29

86

82

25

16

67

21

88

<LO

D<

LOD

20

08

24

59

15

98

12

81

66

71

87

5<

LOD

5

(b)

Sit

e2

20

05

0–

––

––

––

––

<LO

D

20

06

10

41

21

41

58

11

22

05

50

<LO

D<

LOD

20

07

13

21

62

91

16

43

11

07

70

.1<

LOD

20

08

15

67

20

81

03

21

96

07

47

<LO

D7

LOD

:li

mit

of

de

tect

ion

.

A. Ledreux et al. / Harmful Algae 10 (2010) 88–97 91

(PP2A), as described previously (Briand et al., 2002). The MCsconcentrations were calculated from a standard MC-LR calibrationcurve (Alexis Biochemicals, Switzerland). Under these conditions,the detection limit was 0.01 mg equivalent MC-LR per liter of rawwater. A maximum of 21 samples were analyzed for their MCscontent each year during the monitoring program.

2.6.2. Screening for ANTX

Screening for ANTX was performed by mass spectrometryaccording to Gugger et al. (2005).

2.6.3. Screening for STXs

2.6.3.1. Neuro-2a cell-based assay. The saxitoxin content of frozenfilters was extracted in 6 mL of ultra pure water acidified withacetic acid to pH 2.4, sonicated for 3 min with an ultrasonic probein pulse mode (Sonics Vibra Cell 130W, Granuloshop, France), andfiltered on 0.22-mm filters (Analypore Labosi). The filtrates wereanalyzed by a Neuro-2a cell-based assay versus a STX standard(Institute of Marine Biosciences, Halifax, Canada). The STXs weredetected as described previously (Humpage et al., 2007). Thedetection limit was 0.5 mg equivalent STX per liter of raw water.

2.6.3.2. LC–MS/MS analyses. LC–MS/MS analyses were performedby a modification of the method published by Dell’Aversano et al.(2005). Hydrophilic interaction liquid chromatography (HILIC) wasemployed to separate the polar neurotoxin molecules, and wascarried out using a 3.5 mm Sequant ZIC HILIC column(150 mm � 1 mm i.d.). The HPLC system was from Dionex, USA(Ultimate 3000). The mobile phases consisted of (A) aqueousammonium acetate 10 mM adjusted to pH 5 with acetic acid and(B) ACN acidified with 0.05% HCOOH. Separation was carried outunder isocratic condition (40% A/60% B) at a flow rate of50 mL min�1. The separated molecules were analyzed on-lineusing a quadripole-time of flight (Q-TOF) mass spectrometer(Pulsar i, Applied Biosystems) equipped with an electrosprayionization (ESI) source. The information dependent acquisition(IDA) mode allowed both MS and MS/MS spectra to be acquired. Allexperiments were performed in positive-ion mode. Data wereacquired and processed using Analyst Qs software (version 1.1).Typical acquisition parameters were as follows: capillary volta-ge = 4500 V, declustering potential = 20 V, scan range ranged fromm/z = 50 to m/z = 1000, exclusion time = 60 s. The collision energyrequired for fragmentation (MS/MS experiments) was automati-cally calculated by the software.

A set of nine standards purchased from the Institute of MarineBiosciences (Halifax, Canada) was used and consisted of STX,neoSTX, GTX1/4, GTX2/3, GTX5, C1/C2, dcSTX, dcneoSTX anddcGTX2/3. The STXs variants in the cyanobacterial extracts wereidentified from the chromatographic retention time, molecularmass and MS/MS ion fragments.

3. Results

3.1. Chlorophyll a concentrations

The data obtained from the survey (2005–2008) allowedcharacterizing the water body as hypereutrophic on the basis ofthe chlorophyll a concentrations (Ryding and Rast, 1994) (Table 1and Fig. 2). For Site 1, the mean annual Chl a concentration rangedfrom a minimum value of 29 mg L�1 in 2007 to a maximum of200 mg L�1 in 2005. The highest Chl a levels (1590 mg L�1, Table 1)measured in September 2005 at Site 1 (Fig. 2) corresponded tosurface scum. For Site 2, the mean annual Chl a value reached itslowest level during the year 2007 (21 mg L�1), and peaked in 2008(67 mg L�1).

[(Fig._2)TD$FIG]

Fig. 2. Dynamics from June 2005 to November 2008 of chlorophyll a (—) (left y-axis), dominance of cyanobacteria (x), relative abundance of Microcystis spp. (- - -) and relative

abundance of Aphanizomenon spp. (� � �) at Champs-sur-Marne (Site 1) (right y-axis).

A. Ledreux et al. / Harmful Algae 10 (2010) 88–9792

3.2. Cyanobacterial species richness

During the monitoring, the number of cyanobacterial taxadoubled at both sampling sites (e.g. for Site 1: 13 taxa in 2005versus 28 taxa in 2008) (Table 1). Several potentially toxiccyanobacterial genera such as Anabaena, Aphanizomenon, Micro-

cystis, and Planktothrix were observed with variable intensity andfrequency of occurrence during the monitoring period (data notshown). Among these potentially toxic genera, the most abundantduring the survey were Aphanizomenon and Microcystis.

Occurrences of Aphanizomenon spp. were of particularinterest, since species belonging to this genus have beenidentified as potential producers of a range of toxins, includingSTXs. The relative frequency (number of occurrences/number ofsamples analyzed) of Aphanizomenon spp. increased during the

monitoring period (from 41% in 2005 to 67% in 2008 for Site 1,and from 20% in 2006 to 60% in 2008 for Site 2; Table 1),indicating that Aphanizomenon spp. had progressively settled inthe water body between 2005 and 2008. Throughout the survey,Aphanizomenon spp. was usually observed in summer, but therewas no apparent relationship between its presence and thedominance of cyanobacteria in the phytoplankton community(Fig. 2).

The relative frequency of the occurrence of the potential MCsproducers Microcystis spp. increased (from 59% in 2005 to 75% in2008) at Site 1, whereas it remained unchanged at Site 2 (from 50%in 2006 to 47% in 2008). The highest relative frequency wasrecorded in 2007 for both sites (88% and 77%, respectively); thesevalues indicate that Microcystis spp. was observed in most of thesamples collected in 2007.

[(Fig._3)TD$FIG]

Fig. 3. LC–MS/MS analysis of an extract from Champs-sur Marne recreational area

(Site 1). (A) ESI–MS/MS fragmentation pattern of the [M+H]+peak at m/z 300.1

corresponding to STX (Inset) Retention time of 25.6 min corresponding to STX. (B)

ESI–MS/MS fragmentation pattern of the [M+H]+peak at m/z 316.1 corresponding to

neoSTX (Inset) Retention time of 24.2 min corresponding to neoSTX.

A. Ledreux et al. / Harmful Algae 10 (2010) 88–97 93

3.3. Detection of cyanotoxins during the monitoring period

The highest concentrations of MCs at Site 1 (89 mg equiv MC-LR L�1 and 57 mg equiv MC-LR L�1) were observed in September2005 and August 2006, respectively. These concentrationscorresponded to the presence of Microcystis aeruginosa and M.

wesenbergii associated with high levels of Chl a (931 mg L�1 and147 mg L�1, respectively). At Site 2, MCs levels were close to orbelow the limit of detection.

Screening for ANTX by mass spectrometry was performed onselected samples when Chl a concentrations were above 30 mg L�1,and when Anabaena flos-aquae was observed in the phytoplankton.No ANTX was detected on those occasions.

3.4. Detection of STXs in natural samples

STXs were detected in September 2008 by the Neuro-2a cell-based assay at concentrations of 4.8 � 0.5 mg equiv STX L�1 (n = 3)and 6.7 � 3.5 mg equiv STX L�1 (n = 3) for Sites 1 and 2, respectively(Table 2). The occurrence of STX and neoSTX was confirmed in bothSites 1 and 2 by LC–MS/MS analysis. In MS experiments, STX wasdetected both as the protonated ion (m/z 300) and as a fragment ion(m/z 282) corresponding to the dehydrated ion at a retention time inthe range of 25.4–25.8 min compared to the 26.1 min of the STXstandard. In MS/MS mode, fragment ions of STX were observed (m/z282; 221; 204; 186; 179 and 138) (Fig. 3A). NeoSTX was detected asthe parent ion (m/z 316) with a retention time in the range 23.9–24.5 min, versus 24.8 min for the neoSTX standard. Fragment ionswith m/z 298; 237; 220; 177 and 138 were detected for neoSTX in MS/MS mode (Fig. 3B).

The high STXs concentrations were correlated to high Chl a

concentrations (153 mg L�1 and 208 mg L�1, for Sites 1 and 2,respectively), and to the dominance of Aphanizomenon flos-aquae

and Aph. gracile (Table 2).

3.5. Characterization of monoclonal cultures

From the STXs natural samples and after isolation procedures,27 monoclonal cultures were obtained: 2 isolates of Anabaena

planctonica Brunnthaler (1903) (PMC 623.10 and PMC 624.10), 23isolates of Aph. gracile Lemmermann (1910) (PMC 626.10 to PMC640.10; PMC 642.10 to PMC 649.10) and 2 isolates of Aph.

aphanizomenoides (Forti) Hortobagyi and Komarek (1979) (PMC625.10 and 641.10). Aph. aphanizomenoides has recently beenproposed to be re-evaluated as Sphaerospermopsis aphanizome-

noides by Zapomelova et al. (2009, 2010).

3.5.1. Description of Aphanizomenon isolates

Cyanobacterial species were identified on the basis ofmorphological and molecular (16S sequencing) criteria.

Table 2Chlorophyll a concentrations, cyanobacterial species observed and their relative abunda

and 2 of the Champs-sur-Marne water body.

Chl a (mg L�1) (n = 3) Cyanobacterial species

Site 1 153�1 Aphanizomenon flos-aquae

Aphanizomenon gracile

Aphanizomenon aphanizomenoides

Microcystis aeruginosa

Microcystis wesenbergii

Site 2 208�4 Aphanizomenon flos-aquae

Aphanizomenon gracile

Aphanizomenon aphanizomenoides

Microcystis wesenbergii

LOD: limit of detection.

Isolates have first been identified on morphological criteriaaccording to the literature (Komarek and Anagnostidis, 1989). Aph.

aphanizomenoides has a metameric structure of trichomes (straightor slightly arcuated), solitary, constricted at cross-walls and alwaysslightly narrowed to the ends. The cells are cylindrical to barrel-shaped with gas vesicles (2.5–3.5 mm long; 3.5–4.5 mm wide). Theterminal cells appeared round (3–4 mm long; 2.5–3.5 mm wide).The heterocytes are intercalar (spherical to cylindrical: 4–6 mm

nce and toxicity in the raw water samples harvested on 8 September 2008 at Sites 1

Relative

abundance (%)

MCs

content

ANTX

content

STXs content

mg equiv STX L�1 (n = 3)

36 <LOD <LOD 4.8� 0.5

6.5

1

<1

2

71.5 <LOD <LOD 6.7�3.5

7.5

<1

1.5

[(Fig._5)TD$FIG]

Fig. 5. Molecular characterization of Aphanizomenon strains isolated from Champs-

sur-Marne freshwater recreational area.

[(Fig._4)TD$FIG]

Fig. 4. Micrographs of Aphanizomenon species investigated in this study. (A and B)

Aphanizomenon aphanizomenoides PMC 641.10. (C and D) Aphanizomenon gracile

PMC 627.10. Scale bar correspond to 10 mm.

A. Ledreux et al. / Harmful Algae 10 (2010) 88–9794

long; 3–4 mm wide). The akinetes are adjacent to heterocytes,solitary or in pairs (spherical to cylindrical: 6–7 mm long; 4–5 mmwide).

Aph. gracile has a metameric structure of trichomes (straight orslightly arcuated). The trichomes are solitary with very finediffluent slime, constricted at cross-walls and always slightly ordistinctly narrowed to the ends. The cells are cylindrical to barrel-shaped with gas vesicles (5–7 mm long; 3–5 mm wide). Theterminal cells are round or conical (6–8 mm long; 3–4 mm wide).The heterocytes are intercalar, cylindrical or barrel-shaped (5–7 mm long; 4–5 mm wide). The akinetes are slightly distant ofheterocytes. They are intercalar, cylindrical, solitary, and alwayslonger (16–20 mm) than wide (4.5–6.5 mm) (Fig. 4).

A Bayesian phylogenetic tree based on 16S rDNA sequencesincluding GenBank references sequences of Aph. gracile and Aph.

aphanizomenoides confirmed these species identification (Fig. 5).

3.5.2. Molecular detection of putative STX synthetase gene fragments

All 25 isolates of Aphanizomenon were investigated for thepresence of the sxtA gene by PCR, and a 600-bp product was clearlyamplified in 14 isolates of Aph. gracile (PMC 627.10, PMC 628.10,PMC 631.10 to PMC 640.10, PMC 646.10 and PMC 649.10) and in

Table 3Presence of the sxtA gene and detection of STXs in 4 monoclonal cultures isolated from

Cyanobacterial species Strain number Presence of the sxtA gen

Anabaena circinalis AWQC ANA311E +

Aphanizomenon aphanizomenoıdes PMC 641.10 +

Aphanizomenon gracile PMC 627.10 +

PMC 638.10 +

PMC 649.10 +

+, detected; �, not detected; LOD: limit of detection.

one isolate of Aph. aphanizomenoides (PMC 641.10) (Supplemen-tary data S2). A sxtA Bayesian phylogenetic was constructed forthese 15 positive sxtA strains including strain ANA311E and 4GenBank reference sequences. The tree showed 96% homologies onsxtA sequences of Aphanizomenon and Anabaena (Supplementarydata S3).

3.5.3. Saxitoxin content of the isolates

Production of STXs was confirmed in sxtA-positive Aph. gracile

PMC 627.10, PMC 638.10 and PMC 649.10 cultures by the Neuro-2acell-based assay (Table 3). The STXs content ranged from0.9 � 0.2 pg equiv STX per cell to 3.5 � 1.2 pg equiv STX per cell.LC–MS/MS analyses identified STX and neoSTX among the 9 variantsscreened in PMC 627.10 and PMC 638.10 (data not shown). In Aph.

aphanizomenoides PMC 641.10, despite a positive sxtA response inPCR, STXs were not detected either by the Neuro-2a cell-based assayor the LC–MS/MS analyses (Table 3).

4. Discussion

The phytoplanktonic community in the Champs-sur-Marnerecreational water body was clearly dominated by cyanobacteriain summer, and displayed increasing numbers of cyanobacteriaspecies from 2005 to 2008. Despite the methodological constraintsinherent in monitoring cyanobacteria (i.e. the greater frequency ofsampling in summer because of the recreational use), a trend canbe observed in the dominant cyanobacterial genera during thesurvey. When monitoring began, Microcystis was the dominant

Champs-sur-Marne recreational area.

e Presence of STXs

Neuro-2a cell-based assay LC–MS/MS

pg equiv STX cell�1 mg equiv STX L�1 Qualitative detection of STXs

<LOD �<LOD �

1.8�0.7 396�142 +

3.5�1.2 766�257 +

0.9� 0.2 191�42 �

A. Ledreux et al. / Harmful Algae 10 (2010) 88–97 95

genus, whereas since 2007 the Nostocales have increased tobecome the dominant population. This change in the cyanobacter-ial community has had a clear impact on the cyanotoxins present(MCs in 2005 versus STXs in 2008). Because of the lack of datapertaining to the environmental variables during this period, it isimpossible to identify the causes of this shift. Teubner et al. (1999)demonstrated that the critical TN:TP ratio has an influence on thedevelopment of Aph. flos-aquae and other cyanobacterial species.However, the relevance of this ratio must be considered in relationto other environmental variables. Time series analyses ofenvironmental parameters and their correlation with cyanobac-terial biomass would be required to identify the main factorscontrolling Aphanizomenon and Microcystis dynamics.

During the last few decades, the increase in frequency ofcyanobacterial blooms in freshwater bodies has led watermanagement authorities to pay more attention to the risksassociated with toxic cyanobacteria. The diagnosis of toxicoccurrences still focuses mainly on microcystins, although reportsof neurotoxins are becoming more frequent (e.g. Gugger et al.,2005; Liu et al., 2006; Carrasco et al., 2007; Wood et al., 2007).

As far as we are aware, this is the first published report of thepresence of STXs in French freshwaters. The causative organismhas been identified as Aph. gracile, and monoclonal culturesproducing STXs have been successfully obtained.

On the basis of the strains cultured, nine cyanobacterial specieshave so far been identified as producing STXs in freshwater. Aph.

flos-aquae was first reported to be an STXs producer in the late1960s, first in Canada (Jackim and Gentile, 1968), and then in NewHampshire in the United States (Mahmood and Carmichael,1986). Australian blooms dominated by A. circinalis producingSTXs have caused extensive stock deaths in Australia (Humpageet al., 1994; Negri et al., 1995). STXs from L. wollei have since beenreported in Alabama (Carmichael et al., 1997), those fromCylindrospermopsis raciborskii and Raphidiopsis brookii in Brazil(Lagos et al., 1999; Molica et al., 2002; Yunes et al., 2009), fromAph. flos-aquae, and Aph. gracile in Portugal and Germany (Pereiraet al., 2000, 2004; Ballot et al., 2010), from Anabaena lemmer-

mannii in Denmark and Finland (Kaas and Henriksen, 2000; Rapalaet al., 2005) and from Limnothrix redekei in Italy (Pomati et al.,2000; Gkelis et al., 2005). However, two STXs-producing strainsAph. flos-aquae (LMECYA31 and NH-5) were reclassified as Aph.

issatschenkoi and Aphanizomenon sp., respectively, based onmorphological and genetic reevaluation (Li et al., 2000, 2003).Aph. flos-aquae STXs-producing strains have also been isolatedfrom a Chinese lake (Liu et al., 2006) and a Portuguese reservoir(Ferreira et al., 2001) but morphological description did not allowthe unequivocal identification of Aph. flos-aquae. In our study,isolations were attempted to provide isolates of Aph. flos-aquae

from Champs-sur-Marne recreational area but none of theisolations was successful. Thus, it remains unclear if Aph. flos-

aquae has the ability to produce STXs.In previous studies conducted in Portugal, Brazil and Germany,

the presence of STXs had already been attributed to Aph. gracile

(Pereira et al., 2004; Costa et al., 2006; Ballot et al., 2010).Moreover, there is some evidence that this species may alsoproduce toxins such as CYN (Rucker et al., 2007), and a not yetidentified hepatoxin (Dos S. Vieira et al., 2005).

In Champs-sur-Marne, Aph. gracile was part of a complexphytoplanktonic community in which other potentially toxicspecies, such as Anabaena flos-aquae, Aph. flos-aquae, M. aeruginosa,M. wesenbergii, and Planktothrix agardhii, were also encountered.Nevertheless, there was no evidence of the co-occurrence ofdifferent types of cyanotoxins at any time during the surveillanceof the Champs-sur-Marne water body, and there was no clearevidence of a time-dependent succession in the potentially toxicMicrocystis and Aphanizomenon populations.

In natural water samples analyzed for their STXs content, Aph.

gracile species accounted for only 6.5% and 7.5% of the phyto-planktonic biomass at Sites 1 and 2, respectively. As Aph. flos-aquae

was the dominant species, and has previously been reported to bean STXs producer (e.g. Ferreira et al., 2001; Liu et al., 2006), thepossibility that Aph. flos-aquae did in fact contribute to theproduction of STXs in the water body cannot be excluded.

In the natural samples analyzed, the concentrations of STXswere 4.8 � 0.5 mg equiv STX L�1 and 6.7 � 3.5 equiv STX L�1 for Sites1 and 2, respectively. These values were above the limit indicated inthe guideline (3 mg equiv STX L�1) for STXs in drinking watersproposed by Australia and Brazil. However, they are lower thanthe values reported by Rapala et al. (2005) (ranging from33 mg equiv STX L�1 to 1000 mg equiv STX L�1), which have beenlinked to possible associated effects on human health.

In the two monoclonal cultures of Aph. gracile and in the naturalsamples, only the two variants STX and neoSTX were detected byLC–MS/MS. In the past, the toxin profiles of other strains ofAphanizomenon have been studied (see Araoz et al., 2009 forreview). It is interesting to note that the Aph. gracile strainLMECYA40 isolated by Pereira et al. (2004) in Portugal alsoproduced only STX and neoSTX. One strain of Aphanizomenon sp.(NH-5) isolated by Mahmood and Carmichael (1986) in NewHampshire was reported to produce STX and neoSTX, whereasFerreira et al. (2001) detected a more complex toxin profile (GTX1,GTX3, GTX4 and C1) in cultures of Aph. flos-aquae isolated fromnorthern Portugal.

Other STXs variants may be found in natural samples,depending on the heterogenicity of the cyanobacterial toxin-producer subpopulations within the blooms, as reported infreshwater for microcystins producers (e.g. Yepremian et al.,2007) and marine environments for the STXs producers Alexan-

drium and Gymnodinium (e.g. Cho et al., 2008; Touzet et al., 2008).During our study, isolations procedures have resulted in the

obtaining of 15 positive sxtA Aph. gracile and Aph. aphanizomenoides

strains and 10 negative ones for both species. Interestingly, STXsproduction has not been evidenced in the strain sxtA-positive Aph.

aphanizomenoides (PMC 641.10). It can be hypothesized that (i) theSTXs synthesis rate of this strain may be so low that variants couldnot be detected either by the Neuro2a cell-based assay or by massspectrometry or/and, (ii) as proposed by Ballot et al. (2010), a lossor an inactivation of part of the saxitoxin gene cluster. Such geneinactivation has been described in cyanobacteria to due tomutations such as gene deletion events and the insertion oftransposable elements (e.g. Ostermaier and Kurmayer, 2009).

Moreover, the isolation of sxtA-positive and sxtA-negativeAphanizomenon strains within Champs-sur-Marne bloom con-firmed the heterogenicity of cyanobaterial populations. This hasalready been demonstrated for microcystins producers such as P.

agardhii (e.g. Briand et al., 2008) or M. aeruginosa (e.g. Rinta-Kantoet al., 2005). Such heterogenicity (chemotypes, genotypes, species)has to be considered for the monitoring of toxicity and/or toxins.

In the case of STXs monitoring in water bodies, several tools areavailable: (i) rapid microscopic observation and morphologicalcomparison with known cyanobacterial STXs producers, (ii)molecular tools to screen for the presence of genes implicatedin the synthesis of STXs, (iii) methods such as the Neuro-2a cell-based assay or easy-to-use ELISAs, which assimilate the toxicityfrom multiple toxin analogs into a single averaged result and (iv)analytical methods for identifying known STXs variants. However,depending on the manufacturer of the ELISA kit used, cross-reactivity issues may occur for some of the STXs variants, whichcould lead to an underestimation of the STX levels or to false-negatives. Analytical methods, such as LC–MS/MS, may not bereliable when natural samples contain unknown STXs variants(Humpage et al., 2007).

A. Ledreux et al. / Harmful Algae 10 (2010) 88–9796

In this study, Aph. gracile and Aph. flos-aquae were the dominantspecies when STX production was observed. These two speciesfrequently occur in freshwater environments of the Northernhemisphere, probably as a consequence of their ability to surviveunfavorable growth conditions by forming akinetes. They areknown to produce MCs, CYN or STXs (e.g. Willame et al., 2005;Preussel et al., 2006; Ballot et al., 2010), and it can be hypothesizedthat several types of toxins may occur simultaneously (Fastneret al., 2009). Such co-occurrence demands careful consideration, aslittle is known about the impact of mixtures of cyanotoxins onhuman and animal health. To date, the health risk associated tocyanobacteria and cyanotoxins is mainly related to the presence ofMCs in the particular fraction. Nevertheless, it has been shown thata significant proportion of other toxins, such as CYN and STXs, canbe found in the dissolved fraction of raw waters (Fastner et al.,2009), and that these toxins may persist in the environment (Jonesand Negri, 1997). This possible co-occurrence of several cyano-bacterial toxins, and the persistence of dissolved toxins in waterbodies highlights the need to re-evaluate the guidelines forrecreational waters so as to evaluate the risk linked to cyano-bacteria and to their toxins more accurately.

Acknowledgements

Sampling was done in cooperation by the technical team of theChamps-sur-Marne area and the MNHN during a research programinvolving the study of biodiversity in the Seine-Saint-DenisDepartment. Fellowships were awarded to A. Ledreux by AFSSA(Agence Francaise de Securite Sanitaire des Aliments) and to A.Catherine by the MNHN (Legs Prevost, Conseil General de Seine-Saint-Denis). The authors are grateful to S. Hamlaoui for assistancewith cell cultures, to L. Dubost for MS analyses and to Dr. KatiaComte for helpful discussion. The comments of anonymousreviewers were greatly appreciated. Monica Ghosh is acknowl-edged for improving the English version of the manuscript. Thiswork was funded by the program ANR SEST 07 CYANOTOX.[SS]

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in

the online version, at doi:10.1016/j.hal.2010.07.004.

References

Araoz, R., Molgo, J., Tandeau de Marsac, N., 2009. Neurotoxic cyanobacterial toxins.Toxicon, doi:10.1016/j.toxicon.2009.07.036.

Ballot, A., Fastner, J., Wiedner, C., 2010. Paralytic shellfish poisoning toxin-produc-ing cyanobacterium Aphanizomenon gracile in Northeast Germany. Appl. Envi-ron. Microbiol. 76, 1173–1180.

Berry, J.P., Lind, O., 2009. First evidence of ‘‘paralytic shellfish toxins’’ and cylin-drospermopsin in a Mexican freshwater system, Lago Catemaco, and apparentbioaccumulation of the toxins in ‘‘tegogolo’’ snails (Pomacea patula catemacen-sis). Toxicon, doi:10.1016/j.toxicon.2009.07.035.

Briand, J.F., Robillot, C., Quiblier, C., Bernard, C., 2002. A perennial bloom ofPlanktothrix agardhii (cyanobacteria) in a shallow eutrophic French lake: lim-nological and microcystin production studies. Arch. Hydrobiol. 153, 605–622.

Briand, E., Gugger, M., Francois, J.C., Bernard, C., Humbert, J.-F., Quiblier, C., 2008.Temporal variations in the dynamics of potentially microcystin-producingstrains in a bloom-forming Planktothrix agardhii (Cyanobacterium) population.Appl. Environ. Microbiol. 74, 3839–3848.

Briand, E., Escoffier, N., Straub, C., Sabart, M., Quiblier, C., Humbert, J.-F., 2009.Spatiotemporal changes in the genetic diversity of a bloom-forming Microcystisaeruginosa (cyanobacteria) population. ISME J. 3, 419–429.

Cadel-Six, S., Peyraud-Thomas, C., Brient, L., Tandeau de Marsac, N., Rippka, R.,Mejean, A., 2007. Different genotypes of anatoxin-producing cyanobacteriacoexist in the Tarn River, France. Appl. Environ. Microbiol. 73, 7605–7614.

Carmichael, W.W., Eans, W.R., Yin, Q.Q., Bell, P., Moczydlowsky, E., 1997. Evidencefor paralytic shellfish poisons in the freshwater cyanobacterium Lyngbya wollei(Farlow ex Gomont) comb.nov. Appl. Environ. Microbiol. 63, 3104–3110.

Carrasco, D., Moreno, E., Paniagua, T., de Hoyos, C., Wormer, L., Sanchis, D., Cires, S.,Martıin-del-Pozo, D., Codd, G.A., Quesada, A., 2007. Anatoxin-a occurrence and

potential cyanobacterial anatoxin-a producers in Spanish reservoirs. J. Phycol.43, 1120–1125.

Catherine, A., Quiblier, C., Yepremian, C., Got, P., Groleau, A., Vincon-Leite, B.,Bernard, C., Troussellier, M., 2008a. Collapse of a Planktothrix agardhii perennialbloom and microcystin dynamics in response to reduced phosphate concentra-tions in a temperate lake. FEMS Microbiol. Ecol. 65, 61–73.

Catherine, A., Troussellier, M., Bernard, C., 2008b. Design and application of astratified sampling strategy to study the regional distribution of cyanobacteria(Ile-de-France, France). Water Res. 42, 4989–5001.

Cestele, S., Catterall, W.A., 2000. Molecular mechanisms of neurotoxin action onvoltage-gated sodium channels. Biochimie 82, 883–892.

Cho, Y., Hiramatsu, K., Ogawa, M., Omura, T., Ishimaru, T., Oshima, Y., 2008. Non-toxic and toxic subclones obtained from a toxic clonal culture of Alexandriumtamarense (Dinophyceae): toxicity and molecular biological feature. HarmfulAlgae 7, 740–751.

Costa, I.A.S., Azevedo, S.M.F.O., Senna, P.A.C., Bernardo, R.R., Costa, S.M., Chellappa,N.T., 2006. Occurrence of toxin-producing cyanobacteria blooms in a Braziliansemiarid reservoir. Braz. J. Biol. 66, 211–219.

Dell’Aversano, C., Hess, P., Quilliam, M.A., 2005. Hydrophilic interaction liquidchromatography-mass spectrometry for the analysis of paralytic shellfishpoisoning (PSP) toxins. J. Chromatogr. A 1081, 190–201.

Dos S. Vieira, J.M., Azevedo, M.T., Oliveira Azevedo, S.M.F., Honda, R.Y., Correa, B.,2005. Toxic cyanobacteria and microcystin concentrations in a public watersupply reservoir in the Brazilian Amazonia region. Toxicon 45, 901–909.

Fastner, J., Rucker, J., Wiedner, C., Nixdorf, B., Chorus, I., 2009. Long-term trendsand seasonal patterns of cyanobacterial hepato- and neurotoxins in Germanwater bodies. In: International Symposium on Toxicity Assessment, Metz,France, p. 57.

Ferreira, F.M.B., Franco, J.M., Fidalgo, M.L., Fernandez-Vila, P., 2001. PSP toxins fromAphanizomenon flos-aquae (cyanobacteria) collected in the Crestuma-Leverreservoir (Douro river, Northern Portugal). Toxicon 39, 757–761.

Fristachi, A., Sinclair, J.L., 2008. Occurrence of cyanobacterial harmful algal blooms:workgroup report. Adv. Exp. Med. Biol. 619, 45–103.

Galvao, J.A., Oetterer, M., Bittencourt-Oliveira, M.C., Gouvea-Barros, S., Hiller, S.,Erler, K., Luckas, B., Pinto, E., Kujbida, P., 2009. Saxitoxins accumulation byfreshwater tilapia (Oreochromis niloticus) for human consumption. Toxicon 54,891–894.

Gkelis, S., Rajaniemi, P., Vardaka, E., Moustaka-Gouni, M., Lanaras, T., Sivonen, K.,2005. Limnothrix redekei (Van Goor) Meffert (Cyanobacteria) strains fromLake Kastoria, Greece form a separate phylogenetic group. Microb. Ecol. 49(1), 176–182.

Gugger, M., Lyra, C., Suominen, I., Tsitko, I., Humbert, J.F., Salkinoja-Salonen, M.S.,Sivonen, K., 2002. Cellular fatty acids as chemotaxonomic markers of the generaAnabaena, Aphanizomenon, Microcystis, Nostoc and Planktothrix (cyanobac-teria). Int. J. Syst. Evol. Microbiol. 52, 1007–1015.

Gugger, M., Hoffmann, L., 2004. Polyphyly of true branching cyanobacteria (Stigo-nematales). Int. J. Syst. Evol. Microbiol. 54, 349–357.

Gugger, M., Lenoir, S., Berger, C., Ledreux, A., Druart, J.-C., Humbert, J.-F., Guette, C.,Bernard, C., 2005. First report in a river in France of the benthic cyanobacteriumPhormidium favosum producing anatoxin-a associated with dog neurotoxicosis.Toxicon 45, 919–928.

Hitzfeld, B.C., Hoger, S.J., Dietrich, D.R., 2000. Cyanobacterial toxins: removal duringdrinking water treatment, and human risk assessment. Environ. Health Per-spect. 108 (S1), 113–122.

Huelsenbeck, J.P., Ronquist, F., 2001. MrBayes: Bayesian inference of phylogenetictrees. Bioinformatics 17, 754–755.

Humpage, A.R., Rositano, J., Bretag, A.H., Brown, R., Baker, P.D., Nicholson, B.C.,Steffensen, D.A., 1994. Paralytic shellfish poisons from Australian cyanobacter-ial blooms. Aust. J. Mar. Freshwater Res. 45, 761–771.

Humpage, A.R., Ledreux, A., Fanok, S., Bernard, C., Briand, J.-F., Eaglesham, G.,Papageorgiou, J., Nicholson, B., Steffensen, D., 2007. Application of the neuro-blastoma assay for paralytic shellfish poisons to neurotoxic freshwater cyano-bacteria: interlaboratory calibration and comparison with other methods ofanalysis. Environ. Toxicol. Chem. 26, 1512–1519.

Jackim, E., Gentile, J., 1968. Toxins of a blue-green alga: similarity to saxitoxin.Science 162, 915–916.

Jellett, J.F., Roberts, R.L., Laycock, M.V., Quilliam, M.A., Barrett, R.E., 2002. Detectionof paralytic shellfish poisoning (PSP) toxins in shellfish tissue using MIST Alert, anew rapid test in parallel with the regulatory AOAC mouse bioassay. Toxicon 40,1407–1425.

Jones, G.J., Negri, A.P., 1997. Persistence and degradation of cyanobacterial paralyticshellfish poisons (PSPs) in freshwaters. Water Res. 31, 525–533.

Kaas, H., Henriksen, P., 2000. Saxitoxins (PSP toxins) in Danish lakes. Water Res. 34,2089–2097.

Kellmann, R., Mihali, T.K., Jeon, Y.J., Pickford, R., Pomati, F., Neilan, B.A., 2008.Biosynthetic intermediate analysis and functional homology reveal a saxitoxingene cluster in cyanobacteria. Appl. Environ. Microbiol. 74 (13), 4044–4053.

Kogure, K., Tamplin, M.L., Simidu, U., Colwell, R.R., 1988. A tissue culture assay fortetrodotoxin, saxitoxin and related toxins. Toxicon 26, 191–197.

Komarek, J., Anagnostidis, K., 1989. Modern approach to the classification system ofCyanophytes. 4-Nostocales. Arch. Hydrobiol. Suppl. 82–83, 247–345.

Komarek, J., Anagnostidis, K., 1998. Cyanoprokaryota. 1. Teil: Chroococcales. In: Ettl,H., Gartner, G., Heynigh, H., Mollenhauer, D. (Eds.), Sußwasserflora von Mitte-leuropa 19/1. Gustav Fischer, Jena-Stuttgart-Lubeck-Ulm, p. 548.

A. Ledreux et al. / Harmful Algae 10 (2010) 88–97 97

Komarek, J., Anagnostidis, K., 2005. Cyanoprokaryota. 2. Teil: Oscillatoriales. In:Budel, B., Krienitz, L., Gartner, G., Schagerl, M. (Eds.), Sußwasserflora vonMitteleuropa 19/2. Elsevier GmbH, Munchen, p. 759.

Kotai, J., 1972. Instructions for Preparation of Modified Nutrient Solution Z8 forAlgae. Publication B-11/69. Norwegian Institute for Water Research, Oslo, pp.1–5.

Kumar, S., Tamura, K., Nei, M., 2004. MEGA3: integrated software for molecularevolutionary genetics analysis and sequence alignment. Brief. Bioinform. 5,150–163.

Lagos, N., Onodera, H., Zagatto, P.A., Andrinolo, D., Azevedo, S.M., Oshima, Y., 1999.The first evidence of paralytic shellfish toxins in the freshwater cyanobacteriumCylindrospermopsis raciborskii, isolated from Brazil. Toxicon 37, 1359–1373.

Lane, D.J., 1991. 16S/23S rDNA sequencing. In: Stackebrandt, E., Goodfellow, M.(Eds.), Nucleic Acid Techniques in Bacterial Systematics. Wiley, Chichester,Sussex, UK, pp. 115–174.

Li, R., Carmichael, W.W., Liu, Y., Watanabe, M.M., 2000. Taxonomic re-evaluation ofAphanizomenon flos-aquae NH-5 based on morphology and 16S rRNA genesequences. Hydrobiologia 438, 99–105.

Li, R., Carmichael, W.W., Pereira, P., 2003. Morphological and 16S rRNA geneevidence for reclassification of the paralytic shellfish toxin producing Aphani-zomenon flos-aquae LMECYA31 as Aphanizomenon issatschenkoi (Cyanophy-caea). J. Phycol. 39, 814–818.

Liu, Y., Chen, W., Li, D., Shen, Y., Liu, Y., Song, L., 2006. Analysis of paralytic shellfishtoxins in Aphanizomenon DC-1 from Lake Dianchi. China. Environ. Toxicol. 21,289–295.

Llewellyn, L.E., 2006. Saxitoxin, a toxic marine natural product that targets amultitude of receptors. Nat. Prod. Rep. 23, 200–222.

Mahmood, N.A., Carmichael, W.W., 1986. Paralytic shellfish poisons produced bythe freshwater cyanobacterium Aphanizomenon flos-aquae NH-5. Toxicon 24,175–186.

Molica, R., Onodera, H., Garcia, G., Rivas, M., Andrinolo, D., Nascimento, S., Meguro,H., Oshima, Y., Azevedo, S., Lagos, N., 2002. Toxins in the freshwater cyanobac-terium Cylindrospermopsis raciborskii (Cyanophyceae) isolated from Tabocasreservoir in Caruaru, Brazil, including demonstration of a new saxitoxin ana-logue. Phycologia 41, 606–611.

Negri, A.P., Jones, G.J., 1995. Bioaccumulation of paralytic shellfish poisoning (PSP)toxins from the cyanobacterium Anabaena circinalis by the freshwater musselAlathyria condola. Toxicon 33, 667–678.

Negri, A.P., Jones, G.J., Hindmarsh, M., 1995. Sheep mortality associated withparalytic shellfish poisons from the cyanobacterium Anabaena circinalis. Tox-icon 33, 1321–1329.

Ostermaier, V., Kurmayer, R., 2009. Distribution and abundance of nontoxicmutants of cyanobacteria in lakes of the Alps. Microb. Ecol. 58, 323–333.

Paerl, H.W., Huisman, J., 2008. Blooms like it hot. Science 320, 57–58.Page, R.D.M., 1996. Treeview: an application to display phylogenetic trees on

personal computers. Comput. Appl. Biosci. 12, 357–358.Pereira, P., Onodera, H., Andrinolo, D., Franca, S., Araujo, F., Lagos, N., Oshima, Y., 2000.

Paralytic shellfish toxins in the freshwater cyanobacterium Aphanizomenon flos-aquae, isolated from Montargil reservoir, Portugal. Toxicon 38, 1689–1702.

Pereira, P., Li, R.H., Carmichael, W.W., Dias, E., Franca, S., 2004. Taxonomy andproduction of paralytic shellfish toxins by the freshwater cyanobacteriumAphanizomenon gracile LMECYA40. Eur. J. Phycol. 39, 361–368.

Pomati, F., Sacchi, S., Rossetti, C., Giovannardi, S., Onodera, H., Oshima, Y., Neilan,B.A., 2000. The freshwater cyanobacterium Planktothrix sp. FP1: molecularidentification and detection of paralytic shellfish poisoning toxins. J. Phycol.36, 553–562.

Posada, D., Crandall, K.A., 1998. ModelTest: testing the model of DNA substitution.Bioinformatics (Oxf.) 149, 817–818.

Preussel, K., Stuken, A., Wiedner, C., Chorus, I., Fastner, J., 2006. First report oncylindrospermopsin producing Aphanizomenon flos-aquae (Cyanobacteria) iso-lated from two German lakes. Toxicon 47, 156–162.

Rapala, J., Robertson, A., Negri, A.P., Berg, K.A., Tuomi, P., Lyra, C., Erkomaa, K.,Lahti, K., Hoppu, K., Lepisto, L., 2005. First report of saxitoxin in Finnish lakesand possible associated effects on human health. Environ. Toxicol. 20, 331–340.

Rinta-Kanto, J.M., Ouellette, A.J., Boyer, G.L., Twiss, M.R., Bridgeman, T.B., Wilhelm,S.W., 2005. Quantification of toxic Microcystis spp. during the 2003 and 2004blooms in western Lake Erie using quantitative real-time PCR. Environ. Sci.Technol. 39, 4198–4205.

Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inferenceunder mixed models. Bioinformatics 19, 1572–1574.

Rucker, J., Stuken, A., Nixdorf, B., Fastner, J., Chorus, I., Wiedner, C., 2007. Concen-trations of particulate and dissolved cylindrospermopsin in 21 Aphanizome-non-dominated temperate lakes. Toxicon 50, 800–809.

Ryding, S.-O., Rast, W., 1994. Le controle de l’eutrophisation des lacs et reservoirs.UNESCO, Editions MASSON, Paris, France, 294 pp.

Sivonen, K., Jones, G., 1999. Cyanobacterial toxins. In: Chorus, I., Bartram, J. (Eds.),Toxic Cyanobacteria in Water: A Guide to Public Health. Significance, Monitor-ing and Management. Spon/Chapman & Hall, London, pp. 41–112.

Talling, J.F., Driver, D., 1963. Some problems in the estimation of chlorophyll a inphytoplankton. In: Doty, M. (Ed.), Proceedings on Primary Production Measure-ment, Marine and Freshwater. US Atomic Energy Engineering Commission,Hawaii, pp. 146–152.

Teste, V., Briand, J.-F., Nicholson, B.C., Puiseux-Dao, S., 2002. Comparison of changesin toxicity during growth of Anabaena circinalis (cyanobacteria) determinedby mouse neuroblastoma bioassay and HPLC. J. Appl. Phycol. 14, 399–407.

Teubner, K., Feyerabend, R., Henning, M., Nicklisch, A., Woitke, P., Kohl, J.-G., 1999.Alternative blooming of Aphanizomenon flos-aquae and Planktothrix agardhiiinduced by the timing of critical nitrogen: phosphorus ratio in hypertrophicriverine lakes. Arch. Hydrobiol. Spec. Issues Adv. Limnol. 54, 325–344.

Thomazeau, S., Houdan-Fourmont, A., Coute, A., Duval, C., Couloux, A., Rousseau, F.,Bernard, C., 2010. The contribution of sub-Saharan African strains to thephylogeny of cyanobacteria: focusing on the Nostocaceae family (Nostocalesorder, Cyanobacteria). J. Phycol. 46, 564–579.

Touzet, N., Franco, J.M., Raine, R., 2008. Morphogenetic diversity and biotoxincomposition of Alexandrium (Dinophyceae) in Irish coastal waters. HarmfulAlgae 7, 782–797.

WHO, 2003. Algae and cyanobacteria in fresh water. In: Guidelines for SafeRecreational Water Environments. World Health Organization, Geneva, pp.136–158.

Willame, R., Jurczak, T., Iffly, J.F., Kull, T., Meriluoto, J., Hoffmann, L., 2005. Distribu-tion of hepatotoxic cyanobacterial blooms in Belgium and Luxembourg. Hydro-biologia 551, 99–117.

Wood, S.A., Selwood, A.I., Rueckert, A., Holland, P.T., Milne, J.R., Smith, K.F., Smits, B.,Watts, L.F., Cary, C.S., 2007. First report of homoanatoxin-a and associated dogneurotoxicosis in New Zealand. Toxicon 50, 292–301.

Yepremian, C., Gugger, M.F., Briand, E., Catherine, A., Berger, C., Quiblier, C., Bernard,C., 2007. Microcystin ecotypes in a perennial Planktothrix agardhii bloom. WaterRes. 41, 4446–4456.

Yunes, J.S., De La Rocha, S., Giroldo, D., Bonoto da Silveira, S., Comin, R., da SilvaBicho, S., Melcher, S.S., Sant’anna, C.L., Vieira, A.A., 2009. Release of carbohy-drated and proteins by a subtropical strain of Raphidiopsis brookii (Cyanobac-teria) able to produce saxitoxin at three nitrate concentrations. J. Phycol. 45,585–591.

Zapomelova, E., Jezberova, J., Hrouzek, P., Hisem, D., Rehakova, K., Komarkova, J.,2009. Polyphasic characterization of three strains of Anabaena reniformis andAphanizomenon aphanizomenoides (Cyanobacteria) and their reclassification toSphaerospermum Gen. Nov. (Incl. Anabaena kisseleviana). J. Phycol. 45, 1363–1373.

Zapomelova, E., Jezberova, J., Hrouzek, P., Hisem, D., Rehakova, K., Komarkova, J.,2010. Nomenclature note. J. Phycol. 46, 415.