Comparative summer dynamics of surface cyanobacterial communities in two connected lakes from the west of Ireland

Science of the Total Environment 553 (2016) 416–428

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Comparative summer dynamics of surface cyanobacterial communities in two connected lakes from the west of Ireland N. Touzet a,⁎, D. McCarthy b, A. Gill b, G.T.A. Fleming b a b

Centre for Environmental Research, Innovation and Sustainability, School of Science, Department of Environmental Science, Institute of Technology Sligo, Sligo, Ireland Microbiology, School of Natural Sciences, National University of Ireland, Galway, Galway, Ireland

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• DGGE highlighted cyanobacterial community differences in two connected lakes. • Microcystin-coding genes and microcystin-like activity were detected in both lakes. • Low cyanobacterial richness was associated with increased microcystin-like activity.

a r t i c l e

i n f o

Article history: Received 9 December 2015 Received in revised form 27 January 2016 Accepted 17 February 2016 Available online xxxx Editor: D. Barcelo Keywords: Cyanobacteria Eutrophication DGGE Molecular fingerprinting Community dynamics Microcystins

⁎ Corresponding author. E-mail address: [email protected] (N. Touzet).

http://dx.doi.org/10.1016/j.scitotenv.2016.02.117 0048-9697/© 2016 Elsevier B.V. All rights reserved.

a b s t r a c t The eutrophication of lakes is typically associated with high biomass proliferations of potentially toxic cyanobacteria. At a regional level, the sustainable management of water resources necessitates an approach that recognises the interconnectivity of multiple water systems within river catchments. This study examined the dynamics in summer diversity of planktonic cyanobacterial communities and microcystin toxin concentrations in two inter-connected lakes from the west of Ireland prone to nutrient enrichment. DGGE analysis of 16S rRNA gene amplicons of genotype-I cyanobacteria (typically spherical) showed changes in the communities of both Lough Corrib and Ballyquirke Lough throughout the summer, and identified cyanobacterial genotypes both unique and shared to both lakes. Microcystin concentrations, estimated via the protein phosphatase 2A inhibition assay, were greater in August than in July and June in both lakes. This was concomitant to the increased occurrence of Microcystis as evidenced by DGGE band excision and subsequent sequencing and BLAST analysis. RFLP analysis of PCR amplified mcy-A/E genes clustered together the August samples of both lakes, highlighting a potential change in microcystin producers across the two lakes. Finally, the multiple factor analysis of the combined environmental data set for the two lakes highlighted the expected pattern opposing greater water temperature and chlorophyll concentration against macronutrient concentrations, but also indicated a negative relationship between microcystin concentration and cyanobacterial diversity, possibly underlining allelopathic interactions. Despite some element of connectivity, the dissimilarity in the composition of the cyanobacterial

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

417

assemblages and the timing of community change in the two lakes likely were a reflexion of niche differences determined by meteorologically-forced variation in physico-chemical parameters in the two water bodies. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The management of water quality is an important endeavour worldwide relevant to biodiversity preservation, ecosystem balance and public health safety (Papastergiadou et al., 2010). Increasing anthropogenic pressures on aquatic environments have been met with the development of stringent policies, such as the EU Water Framework Directive in Europe, implemented to preserve the sustainability of aquatic resources (European Communities, 2000). Their efficient management requires a better understanding of the interconnections between the factors that determine quantitative and qualitative patterns in the biota components of aquatic ecosystems. The eutrophication of continental waters is a significant environmental problem, which has usually been associated with high biomass algal proliferations, in particular cyanobacteria (Carpenter et al., 1998; Schindler, 2006; Istvánovics, 2009). Those dominated by toxic species have caused a variety of ecological disruptions and present serious threats toward animal and human health (Codd et al., 2005). Cyanobacteria are ubiquitous components of phytoplankton assemblages in aquatic environments and although proliferations are typically seasonal, perennial blooms have been documented in several eutrophic lakes (Mur et al., 1999). The dynamics of cyanobacterial populations in lakes have been extensively studied due to the potential noxious effects some species can elicit. Animal death and human illness have been associated with the production by some species of a diverse array of biotoxins (Codd et al., 2005). The most common are the hepatotoxic cyclic heptapeptides microcystins, which have been non-exhaustively found in the Microcystis, Anabaena, Nostoc or Planktothrix genera (Vezie et al., 1998; Marsalek et al., 2000). Many microcystin structural forms of varying toxicity have been characterised with a mode of action associated with the inhibition of protein phosphatases. Cyanotoxins constitute a serious problem for the management of drinking water as toxin concentrations in lakes do not necessarily reflect the size of cyanobacterial communities. Succession in toxigenic cyanobacteria may also impact upon the overall amount of cyanotoxins present in water (Kardinnal and Visser, 2005). Moreover, variation in cellular toxin quotas may result from changes in environmental conditions (Dokulil and Teubner, 2000). It has hence been recommended to base the management of cyanobacterial blooms on both toxin concentration and biomass. For example, guideline values for microcystin concentration and cyanobacterial biomass in water, issued by the World Health Organization, have been set to 1 μg·l−1 and 10 μg·l−1, respectively (WHO, 2003). The determination of the taxonomic diversity of cyanobacterial communities has traditionally been carried out by light microscopy. Difficulties in the management of toxic blooms have lied in the common cooccurrence of multiple morphologically similar species, which may or not synthesise toxins. Difficulties have been further compounded by the fact that there exist in individual species both toxic and nontoxic forms (Kurmayer et al., 2004). Molecular techniques have permitted to better characterise microbial genetic diversity in various ecosystems. In particular for cyanobacteria, phylogenetic studies based on 16S rDNA polymorphism have demonstrated the existence of both polyphyletic and monophyletic groups interspersed in various clades (Willmotte and Herdman, 2001; Abed et al., 2002; Litvaitis, 2002). Other studies have led to the design of taxaspecific molecular markers suitable for the discrimination of morphologically similar species (Zehr et al., 1997; Runnegar et al., 1995; Beard et al., 1999; Iteman et al., 2000; Tillett et al., 2001; Zeidner et al., 2003). The

molecular fingerprinting of microbial communities using typing methods such as denaturant gradient gel electrophoresis (DGGE) of PCR-amplified genes have notably proved successful to determine patterns in microbial species succession and community structure, including cyanobacteria (Geiss et al., 2004; Song et al., 2005; Briand et al., 2009). The elucidation of the genetic basis of non-ribosomal microcystin biosynthesis has also resulted with the characterisation of the microcystin toxin gene cluster (mcy) in several cyanobacterial genera (Neilan et al., 1999; Christiansen et al., 2003; Rouhiainen et al., 2004). Primers have since been tested using both culture and field samples, for the specific amplification of different domains of the mcy gene cluster (Pan et al., 2002; Hisbergues et al., 2003). These methodological developments offer opportunities to better manage the risks associated with cyanobacterial blooms. As aquatic resources are increasingly being managed at the water catchment or regional level, there is a need to better understand the dynamics in space and time of biota communities in relation to the degree of connectivity of water bodies. In this context this study was carried out to compare the dynamics in summer diversity of planktonic cyanobacterial communities in two inter-connected lakes prone to nutrient enrichment. The main hypothesis focused on ascertaining whether or not there was a difference in cyanobacterial community composition in Ballyquirke Lough and Lough Corrib. To this end, cyanobacterial diversity was estimated by genotypic analysis via DGGE resolution of 16S rDNA amplicons. Along with the measurement of limnological variables, the potential existence of a relationship between variation in cyanobacterial community composition and the onset of microcystin production in the two lakes was also examined. 2. Material and methods 2.1. Study area Sampling was carried during summer 2010 in two connected lakes in the west of Ireland (Fig. 1). Western Irish lakes are managed as recreational fisheries resources and constitute important tourist attractions for the region. Changes in farming practices near some lakes have however generated concerns about their vulnerability to eutrophication (McCarthy et al., 2001). Lough Corrib is the largest of the western Irish lakes with a surface extending over 180 km2 and a catchment area of ~ 3000 km2. The lake spans two geological basins mainly constituted of granite and limestone and contains multiple islands and bays. Lough Corrib Lower has an average depth of 2.1 m (max. 5.5 m) while that of Lough Corrib Upper is 8.4 m (max. 42.0 m). Ballyquirke Lough lies southeast of Lough Corrib Lower and an ~3 km long river connects the two water bodies. The underlying limestone bedrock also provides ample subterranean drainage routes. Ballyquirke Lough is much smaller with a surface and catchment area of 79.2 ha and 7319 ha, respectively. 2.2. Sampling and data acquisition Sampling was carried out on a near weekly basis in the two lakes along transects of stations between June 6th and September 2nd 2010. Surface water temperature and depth profiles were recorded using a YSI MultiQuatroPro probe. Geographical coordinates were taken with a GPS and water transparency was determined with a Secchi disk. Solar irradiance, wind speed and daily rainfall data were retrieved from the meteorological station at the National University of Ireland Galway but were only available from mid-July for seven weeks. Surface water samples were immediately filtered onto Whatmann GF/F filters for subsequent analysis of chlorophyll-a and microcystin-like activity.

418

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

Fig. 1. Map of the west of Ireland indicating the location of Lough Corrib and Ballyquirke Lough and the stations sampled therein throughout summer 2010.

Additional water samples were filtered (150 ml) onto Whatmann cellulose nitrate filters (1 μm pore size) for subsequent nucleic acid extraction. The corresponding filtrates were retained in polypropylene tubes for analysis of inorganic nutrients. All samples were stored on ice in a cooling box during sampling and were placed at −20 °C back in the laboratory until further processing. Only one replicate was collected per station. The average values of the stations sampled in Lough Corrib and the stations sampled in Ballyquirke Lough were then compared on a weekly basis.

manufacturer's instructions for Gram-negative bacteria. Minor modifications to the protocol were carried out as follows. First, the membranes were mechanically disrupted by bead-beating in a Hybaid ribolyser (full power, 3 cycles of 20 s each) using 180 μm diameter glass beads. Cell lysis was then carried out with proteinase-K in a Thermomixer (56 °C, 180 min). Nucleic acids were eluted in a final volume of 100 μl with PCR grade water and the concentration and quality of the extracts determined using a Nanodrop 1000 spectrophotometer (Thermo Scientific). Samples were stored at −20 °C until use.

2.3. Phytoplankton biomass determination and macronutrient analysis

2.5. Cyanobacteria community fingerprinting by 16S rDNA DGGE analysis

Phytoplankton biomass was estimated via chlorophyll-a analysis, which was performed on the surface water filters. This was carried out by spectrophotometry after overnight extraction in 90% acetone (Lorenzen, 1967). Samples were homogenised using a vortex mixer and kept in darkness overnight at 5 °C. Samples were centrifuged (4000 rpm, 15 min) prior to carrying out the measurements using a spectrophotometer (Hitachi U-1100) at 665 nm and 750 nm wavelengths. Measurements were also taken after adding a drop of 0.1 N hydrochloric acid to each extract to compensate for the presence of phaeopigments. Dissolved inorganic nutrients, specifically nitrate + nitrite and phosphate, were measured by flow injection using a QC800 automated analyser (LaChat Instruments) with computer-controlled sample selection and peak processing using the methods recommended by the manufacturer (QuikChem® Method 31-107-04-1-E and QuikChem® Method 31-1115-01-1-I).

A semi-nested PCR approach was carried out with an initial partial amplification of the 16S rDNA gene using the cyanobacteria-specific primer combination CYA359f and 23S30r (Boutte et al., 2006) (Table 1). The PCR cocktail contained 1 × buffer with loading dye, 2 mM MgCl2, 200 μM dNTPs, 0.2 μM of both primers, 1 unit of Taq polymerase (Promega, GoTaq) and 1 μl of template DNA. The thermocycling conditions were as follows: initial denaturation step (94 °C, 5 min), 15 cycles of amplification [denaturation (94 °C, 0.5 min), annealing (58 °C, 1 min) and extension (72 °C, 2 min)] and a final extension step (72 °C, 30 min). The second master mix had a similar composition and was supplemented with 1 μl of the first PCR reaction prior to subjecting the samples to the second amplification carried out with the primers CYA359f and CYA781r(a) or CYA781r(b), which preferentially amplify filamentous and spherical morphotypes, respectively (Nübel et al., 1997; Boutte et al., 2006). In this study, these designations have been replaced by genotype groups I and II (GG-I and GG-II), respectively, as in Touzet et al. (2013). The GC clamp was placed on the forward primer. The cycling conditions were as follows: initial denaturation step (94 °C, 5 min), 40 cycles of amplification [denaturation (94 °C, 1 min), annealing (62 °C, 1 min) and extension (72 °C, 1 min)] and a final extension step (72 °C, 30 min). Visualisation of the PCR products was carried

2.4. Nucleic acid extraction from filtered water column samples Total genomic DNA extraction was carried out on filtered water samples using the Qiagen DNeasy® Blood & Tissue Kit according to the

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

419

Table 1 Sequences and characteristics of the oligonucleotide primers used in the study. Primer designation

Sequence (5′–3′)

GC content

Melting point

Reference

CYA359f 23S30r CYA781r(a) CYA781r(b) GC clamp CYAseq-f CYAseq-r mcy-A-cd1-f mcy-A-cd1-r HEPF-mcy-E HEPR-mcy-E

ggggaattttccgcaatggg cttcgcctctgtgtgcctaggt gactactggggtatctaatcccatt gactacaggggtatctaatcccttt cgcccgccgcgccccgcgcccgtcccgccgcccccgcccg gcgaaagcctgacggagc ggggtatctaatcccattcgct aaaagtgttttattagcggctcat aaaattaaaagccgtatcaaa tttggggttaacttttttgggcatagtc aattcttgaggctgtaaatcgggttt

55% 59% 44% 44% 98% 67% 50% 33% 24% 39% 38%

59.3 °C 64.0 °C 61.3 °C 61.3 °C N75 °C 60.5 °C 60.3 °C 55.9 °C 48.1 °C 62.2 °C 60.1 °C

Nübel et al. (1997) Taton et al. (2003) Nübel et al. (1997) Nübel et al. (1997) Nübel et al. (1997) This study This study Hisbergues et al. (2003) Hisbergues et al. (2003) Jungblut and Neilan (2006) Jungblut and Neilan (2006)

out by 1%-agarose gel electrophoresis using the nucleic acid stain Sybr®Safe (Invitrogen). Separation of individual amplicons was performed on vertical denaturing acrylamide gels using the DCode Mutation Detection Sytems (Bio-Rad Laboratories, USA). Gels were constituted of 8% acrylamide/ bis 40/1 (w/v) with a 45–55% denaturant gradient prepared from a 100% denaturation solution of 7 M urea and 40% formamide (v/v). Amplicons were loaded into the wells (35 μl) and separation allowed for 16.5 h at 60 °C at a constant voltage of 60 V. Prior to visualisation under UV illumination, gels were stained in 1 × TAE buffer containing Ethidium Bromide (10 mg·ml−1) for 40 min and destained in deionised water for 12 min. All gels were imaged using the G:Box and the GeneSnap® programme (Syngene). Several DGGE profiles were run to verify the consistency of band separation for random samples. Selected bands on each gel were excised with sterile blades and the DNA eluted overnight at 5 °C in 100 μl of PCR grade water. Reamplification was then performed using 1 μl of template with the internal primer combination CYAseq-f and CYAseq-r using the second round PCR (Table 1). Amplicons were purified using the QIAquick PCR purification kit as described in the manufacturer's instructions. Samples were eluted in 20 μl of PCR grade water and stored at 5 °C prior to external sequencing (GATC, Germany). Analysis of the DGGE amplicon migration patterns representing the cyanobacterial community composition was carried out using the TotalLab TL120 software (Nonlinear Dynamics, UK). A densitometric scan of the gels was created and background noise was subtracted using a rolling disk algorithm. A band-matching matrix was constructed using peak height values for each gel. 2.6. Phylogenetic inference Upon reception, the rDNA sequences of amplified cyanobacterial DGGE bands were initially screened using BLASTn (Altschul et al., 1990) to orientate the identification of the isolates. The sequences were then compiled with other partial 16S rDNA cyanobacteria sequences imported from GenBank, and aligned using the pairwise alignment function of GeneDoc and Clustal-X. PAUP version 4.0b10 (Swofford, 2002) was used to infer the phylogeny of cyanobacteria and determine the relative position of the sequenced bands. Modeltest 3.7 (Posada and Crandall, 1998) was used to determine the optimal base substitution models for the alignments relative to the partial 16S rDNA sequences. Maximum likelihood (ML) analyses were carried out using PAUP with the parameters derived from the Akaike Information Criterion (AIC) in Modeltest. Phylogenies were then reconstructed by performing heuristic searches with random addition of sequences (10 replicates) and a tri-bisection-reconnection branch swapping algorithm. Bootstrap analyses were finally performed on the tree topologies to evaluate the robustness of the sequence arrangements. Due to computation constraints, bootstrap values were derived using the ‘fast’ stepwise addition with 1000 replicates under the distance criterion. The

Gram-positive species Clostridium lavalense (strain CCRI-9929, GenBank entry: EF564278) was selected as out-group. 2.7. mcy PCR and RFLP analysis Partial mcy-A and mcy-E genes were amplified by PCR from the field sample extracts using appropriate primer combinations (Table 1). Both PCRs were carried out in 50 μl reaction volumes using a cocktail consisting of 1 × buffer, 3 mM MgCl2 , 200 μM dNTPs, 0.5 μM of both primers, 1 unit of Taq polymerase (Promega, GoTaq) and 1 μl of template. The thermocycling sequences were as follows: an initial denaturation step (95 °C, 5 min), 35 cycles of amplification [mcy-A: denaturation (94 °C, 45 s), annealing (50 °C, 45 s) and extension (72 °C, 45 s)]–[mcy-E: denaturation (95 °C, 1 min), annealing (58 °C, 1 min) and extension (72 °C, 1 min)] and a final extension step (72 °C, 5 min). For Restriction Fragment Length Polymorphism analysis (RFLP), the enzymes RsaI (GT^AC) and HhaI (G_CG^C) (Fermentas) were used according to the manufacturer's instructions to conduct the dual digestion (15 min, 37 °C) of the combined mcy-A and mcy-E amplicons. In silico analysis of mcy-A and mcy-E sequences in GenBank showed these enzymes to typically cut amplicons once at various sites for several microcystin-producing cyanobacteria. The digests were resolved on 2% agarose 1 × TAE buffer gels and the gel image analysed in Total-Lab TL120. The corresponding similarity matrix was then converted into a dendrogram using the PASWStatistic17.0 module in SPSS. 2.8. Microcystin analysis by protein phosphatase 2A assay (PP2A) The assay is based on the inhibition by microcystins of the dephosphorylation by the enzyme phosphatase 2A of the substrate paranitrophenyl-phosphate (nPP) into the product paranitrophenol (pNP), which is detectable by spectrophotometry (Simon and Vemoux, 1994; Rivasseau et al., 1999). Microcystins were extracted from filtered samples using 4 ml of ice-cold 100% methanol and mechanical disruption by bead-beating and sonication. The homogenates were passed through a 5 ml syringe fitted with a 0.45 μm pore size filter. The eluates were then evaporated to dryness, resuspended with 400 μl of 100% methanol and stored at −20 °C until analysis. The PP2A assay was carried out in a 96-well plate format. Each well was inoculated with 100 μl of buffered sample (field extract, MC-LR standard or buffer blank, in 40 mM Tris/HC1, 34 mM MgCl2, 4 mM EDTA and 4 mM DL-dithiothreitol, pH 8.0) and 50 μl of enzyme (0.05 units ml−1 final concentration). Each determination was done in duplicate and a set of three dilutions for each extract was tested. The reaction was started by the addition of 50 μl of the substrate p-NPP (28.2 mM final concentration) and carried out for 1 h at 37 °C. The hydrolysis of p-NPP to p-NP was recorded at 405 nm using a Microplate Autoreader (Biotek Instruments, USA). Microcystin concentration in samples, expressed in μg·l−1 of MC-LR eq., was determined by

420

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

interpolation of the absorbance values of the samples from the linear portion of the standard curve.

2.9. Data treatment and analysis Pearson correlation was used to determine the significance of the relationships between selected quantitative variables. Analyses of variance (ANOVA) and Student t-tests were carried out to identify significant differences in the distribution of quantitative variables in Lough Corrib and Ballyquirke Lough. Those included temperature, Secchi depth, chlorophyll-a concentration, dissolved inorganic nitrogen and phosphorus, PP2A-based microcystin concentrations and DGGE-based estimate of cyanobacterial abundance. Hierarchical clustering was performed on the RFLP band migration profiles obtained with the samples collected throughout the summer. Fisher's exact test was applied to selected bands to determine the presence of relationships between their presence/absence and environmental descriptors. Finally, multiple factor analysis (MFA) was applied to the Lough Corrib and Ballyquirke Lough combined data sets to assess the relationships between different variables. MFA explores the common structures present in several data sets in two steps (Escofier and Pages, 1990). Principal component analyses (PCA) are first performed on each individual data set, which are then weighted by dividing all their elements by the square root of the first eigenvalue obtained from their PCA. The normalised tables are juxtaposed into a single matrix and subjected to a PCA again. The analyses were run for the following variables: temperature, Secchi depth, chlorophyll-a concentration, dissolved inorganic nitrogen and phosphorus, DGGE-based genotype-II cyanobacteria richness and microcystin concentration estimates. The final PCA was run for a correlation matrix in a Varimax rotation mode. The data sets were autoscaled by column standardisation prior to analysis. All statistical treatments were performed using the PASWStatistic17.0 module in SPSS.

3. Results 3.1. Summer variation in meteorological and limnological variables Solar irradiance, daily rainfall and wind speed were variable between July 15th and the end of August (Fig. 2). Data for surface temperature, Secchi depth and chlorophyll-a in Lough Corrib and Ballyquirke Lough during summer 2010 are illustrated in Fig. 3. Secchi depth in Lough Corrib was significantly lower in July than in June and August (ANOVA, F = 3.66, p = 0.03, n = 62). In Lough Ballyquirke, the average surface temperature, Secchi depth and chlorophyll-a concentration were all significantly lower in August than in June and July (ANOVA, F N 5.82, p b 0.01, n N 43). Concentrations of nitrate and nitrite in Lough Corrib were in general superior to 150 μg·l−1 at station 4, except from mid July to early August when they dropped below 50 μg·l−1 (Fig. 4). At the southern end of Ballyquirke Lough, the concentrations progressively increased during the summer from ~50 μg·l−1 to 125 μg·l−1, with substantial variation between mid-July and mid-August. Phosphate concentration in Lough Corrib was on average 3.5 μg·l−1 (s.d. = 1.7, n = 12) throughout the summer. In Ballyquirke Lough, phosphate levels were similar except from mid-July to early August when a surge reaching ~18 μg·l−1 was detected. On a comparative basis, average nitrate and nitrite concentrations over the summer were significantly greater in Lough Corrib than in Lough Ballyquirke (Student t-test, p b 0.01, n = 25). No such relationship was found for phosphate (Student t-test, p = 0.18, n = 25). Vertical profiling of the water column was carried out during each sampling occasion. Periods of stratification and mixing were observed for both Lough Ballyquirke and Lough Corrib (S1).

Fig. 2. Daily variations in sunlight, wind speed and precipitations recorded at the weather station adjacent to Lough Corrib from July to August 2010.

3.2. Cyanobacteria community fingerprinting by 16S rDNA DGGE analysis The DGGE migration patterns showed variation in the communities of genotype group II cyanobacteria for both lakes during the summer. This was most noticeable for Lough Corrib after August 5th while in Ballyquirke Lough, the first apparent switch was observed after July 5th, followed by minor changes from July 20th (Fig. 5). Genotype

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

421

Fig. 4. Near-weekly variations of surface concentrations of nitrate–nitrite and phosphate in Lough Corrib and Ballyquirke Lough during summer 2010.

The frequency of occurrence of individual bands of genotype II cyanobacteria in the DGGE gels and their intensities were considered for both lakes and highlighted four groups (Fig. 6). There was an overall positive correlation between genotype II cyanobacterial richness and cumulated band intensity, which may be viewed as a proxy for abundance (Pearson's, p b 0.01, r = 0.76) (S2). Noticeably, cumulated band intensities were significantly greater in August than in June, and July than June, for Gr-A (rare but abundant when present) and Gr-C (frequent but not abundant when present) bands, respectively (ANOVA, F = 5.57, p = 0.01). Gr-C bands were also significantly more intense in Ballyquirke Lough than Lough Corrib (Student t-test, p b 0.01). The occurrence of prominent bands from the four groups was assessed in relation to environmental descriptors using Fisher's exact tests. In particular, bands 33, 37, 38, 24, 14 and 15, the first four corresponding to Microcystis sp., showed significant relationships with the descriptors pertaining to the month of sampling, surface chlorophyll-a levels, light penetration and nitrate + nitrite levels (Table 2).

3.3. Partial 16S rDNA phylogenetic inference Fig. 3. Near-weekly variations of surface temperature, Secchi depth and chlorophyll-a in Lough Corrib and Ballyquirke Lough during summer 2010.

group I cyanobacteria were only detected in some of the Ballyquirke samples during August, showing very low richness and band intensity (data not shown). On average there was a moderate significant difference in the richness of cyanobacterial genotype II between the two lakes (Student ttest, p = 0.06, n = 25). Only for Ballyquirke Lough was cyanobacterial richness significantly different during the summer, with the diversity in June being significantly lower than that recorded during July (ANOVA, F = 4.83, p = 0.03, n = 13).

Individual bands were successfully excised from the gels and reamplified prior to purification and sequencing. The BLASTn search returned matches to cyanobacterial sequences for all the bands excised (n = 18), which positions in the 16S rDNA phylogeny were ascertained by maximum likelihood (Fig. 7). The alignment contained 90 partial 16s rDNA sequences with 383 characters, 213 being constant, 43 variable parsimony-uninformative and 127 parsimony-informative. All the bands sequenced clustered in Chroococcales clades. Noticeably, some sequences corresponded to Worochinina sp. and Microcystis sp., species notorious for their potential production of mycrocystin toxins. About 50% of the bands sequenced grouped in a single Chroococcales clade containing species typically considered nontoxic.

422

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

Fig. 5. DGGE migration patterns of partial 16S rDNA amplicons for cyanobacterial GG-II in Lough Corrib and Ballyquirke Lough during summer 2010. The numbers above the lanes indicate the week during which samples were collected. The circles and numbers throughout the gel indicate the bands that were excised and subsequently sequenced.

3.4. Detection of microcystin-like activity in filtered water samples

3.5. RFLP analysis of microcystin genes

Analysis of methanolic extracts using the PP2A assay was carried out on samples collected from the two lakes and showed microcystin-like activity in most samples. Microcystin concentration estimates ranged from 0.01 to 1.15 μg eq. MC-LR per litre. Concentrations were N0.05 μg eq. MC-LR per litre in 33% of the samples from Lough Corrib and 11% in those from Ballyquirke Lough. Concentrations in Lough Corrib were significantly higher in August than in June and July (ANOVA, F = 13.64, p b 0.01, n = 11). No such relationship was found for Lough Ballyquirke (ANOVA, F = 0.92, p = 0.43, n = 13) (Fig. 8). This is largely due to the great standard deviation of the August data set, attributable to sample 469, for which the microcystin concentration estimate was the highest recorded.

PCR amplification of partial mcy genes was carried out on surface water extracts. Amplicons of the expected size (~400 bp) were detected in 10 out of 12 samples from Lough Corrib and in all those collected in Ballyquirke Lough. Double digestion was carried out using the restriction enzymes RsaI and HhaI. The restriction profiles were converted into a matrix and subsequent cluster analysis defined two groups (Fig. 9). Contingency analysis showed a significant relationship in the clustering of samples according to the time at which they were sampled (X = 10.65, p b 0.01, df = 1, n = 25) but none with respect to the lake of origin (X = 0.18, p N 0.10, df = 1, n = 25). Contingency analysis also showed a significant relationship in the distribution of samples in the RFLP-derived clusters A and B and corresponding levels of microcystin-like activity estimated by PP2A (X = 6.62, p b 0.05, df = 1, n = 25). Subsequent to sequencing, several 16S rRNA gene DGGE bands identified as Microcystis or Woronichinia (bands 18, 22, 33, 38) were prominent in samples of Lough Corrib and Ballyquirke Lough which grouped together in cluster B based on mcy RFLP analysis. Those corresponded to the samples collected at the end of August when the highest levels of microcystin-like activity were recorded. 3.6. Multivariate analysis of the combined lake datasets

Fig. 6. Segregation of the cyanobacterial GG-II DGGE bands according to their maximum intensity and frequency of occurrence in Lough Corrib and Ballyquirke Lough during summer 2010.

The multiple factor analysis applied on the Lough Corrib and Ballyquirke Lough autoscaled data sets resulted in three components that accounted for 68% of the total variance (Table 3). Component 1 (PC1) was negatively related to both nitrate and phosphate concentrations and positively to temperature and chlorophyll-a concentration. Component 2 (PC2) was positively related to genotype-II cyanobacterial richness and negatively to microcystin concentrations. Component 3 (PC3) was mostly related to Secchi depth. The sample score plot did reveal some grouping tendencies (Fig. 10). Notably, the samples collected during the first half of the summer, mostly positioned toward the positive end of PC1, separated along PC2 two clusters of samples collected in the second half of the summer. Noticeably, those corresponded to samples with greater cyanobacterial genotype-II richness and higher microcystin concentrations, respectively.

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

423

Table 2 Significant relationships (2-tailed Fisher's exact test) between the presence of selected individual bands and environmental descriptors. Band

14 15 25 33 37 38

Lake

Ballyquirke Ballyquirke Ballyquirke Ballyquirke Ballyquirke Corrib

Environmental descriptor Chl-a

NO− 3

Temp.

Secchi

Time

0.060 0.060 – – 0.060 –

0.080 0.080 – 0.045 – 0.050

0.090 0.090 – – – –

0.080 0.080 0.080 0.0450 – 0.050

0.060 0.060 – – – 0.002

BLAST ID

Match simil.

Access. number

Synechococcus Uncultured Uncultured Microcystis Microcystis Microcystis

98% 98% 99% 99% 91% 96%

NR125481 KF856235 JN032868 KF372571 HM449786 KJ746519

Fig. 7. Most likely tree inferred from the maximum likelihood analysis of 16S rDNA cyanobacterial sequences. The optimal base substitution model derived from the AIC criterion in Modeltest3.7 was a TVM+G model with the following constraining parameters for base substitution frequencies and gamma distribution shape parameter: A = 0.2511, C = 0.2319, G = 0.2921, T = 0.2250; A–C = 1.0475, A–G = 2.8377, A–T = 1.7277, C–G = 0.2615, C–T = 2.8377, G–T = 1.0000; γ = 0.3140. Numbers on the branches indicate branch frequency from 1000 bootstrap replicates (values b50% not included). The selected sequenced bands from Lough Corrib and Ballyquirke Lough are indicated in bold.

424

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

Fig. 8. Monthly average values of microcystin-like activity estimated by the PP2A assay in Lough Corrib (left) and Ballyquirke Lough (right) during summer 2010. For clarity, sample 469 (1.15 μg eq. MC-LR per litre) was excluded from the data shown for August in Ballyquirke Lough.

4. Discussion Even though the consequences of nutrient enrichment on aquatic communities have been studied at regional levels (Donohue et al., 2005; Donohue et al., 2009), little is known at smaller scales, in particular for microbial communities, when connectivity elements between water bodies exist. The dynamics and diversity of surface cyanobacteria communities were hence investigated in two connected lakes from the West Irish lake system. The main aspect of the study focused on ascertaining if there was a difference in cyanobacterial community composition in Ballyquirke Lough and Lough Corrib. Cyanobacteria were detected throughout summer 2010 in both lakes using 16S rDNA DGGE fingerprinting but only for genotype II cyanobacteria were profiles obtained. The suitability of amplicon concentration as a proxy for microbial abundance has previously been questioned owing to the biases associated with semi-nested DGGE PCR. In this regard, Park and Crowley (2010) highlighted the importance of a moderate number of cycles in the primary PCR to limit the loss of quantitative information.

On the other hand, Neilson et al. (2013) have severely cautioned against PCR-DGGE for accurate comparative diversity analysis; yet, they acknowledged the technique as an excellent high-throughput tool for comparative community structure analysis. Compared to New Generation Sequencing (NGS), which is a far more powerful method, PCRDGGE still remains an attractive method for microbial community analysis and the routine comparison of multiple environmental samples, which the costs of NGS still somehow preclude. The DGGE procedure employed in the present study was successfully applied for the analysis of lacustrian cyanobacteria in previous work (Touzet et al., 2013). Cumulated DGGE band intensities for weekly samples collected throughout the summer showed a positive correlation between Lough Corrib and Ballyquirke Lough, indicating that conditions promoting the development of genotype-II cyanobacteria populations, typically small-sized and spherical, were met in the two lakes. An overall correlation between cyanobacterial richness and band intensity was also observed. This observation echoes the multiple studies which have examined diversity– abundance relationships in microbial communities in relation to ecosystem functioning and indicated that more diverse communities tend to have higher contributions to ecosystem balance (Giovannoni, 2004; Bell et al., 2005). 4.1. Summer community change

Fig. 9. Clustering of the Lough Corrib and Ballyquirke Lough samples collected during summer 2010 according to their similarity in the restriction patterns obtained for partial mcyA + E gene amplicons.

DGGE profiling showed shifts during the summer in the surface cyanobacterial community structure, with the apparition and disappearing of genotypes in both Lough Corrib and Ballyquirke Lough. All the partial 16S rDNA bands excised from the DGGE gel returned highly similar Chroococcales sequences closely related to Worochinina, Aphanothece, Cyanobium and Microcystis. Several of those bands noticeably clustered as small branches into a single Chroococcales clade of the phylogenetic tree. Sequencing work performed on strains of the Chrococcales genus Microcystis has highlighted the high plasticity of its genome (Kaneko et al., 2007; Frangeul et al., 2008). It is tempting to extrapolate this microdiversity feature to other Chroococcales and to anticipate individual populations to contain multiple genotypes able to thrive under varying environmental conditions. The timing of the switch in cyanobacterial community structure in the two lakes could be linked to meteorological and physico-chemical forcing, which created niches suitable for the development of populations fit to compete in their respective environments. Physical changes in water column structure occurred during the summer with interspersed precipitations and sustained wind regimes, which were also associated with pulsed availability of nutrients, in particular in Ballyquirke Lough. Variations in abundance caused by competition for limiting resources are associated with the coexistence of multiple

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

425

Table 3 Eigen values of the principal components and variable scores along the corresponding axes following the multiple factor analysis in the summer 2010 data sets. Initial eigenvalues

Variable projections in the rotated component matrix

Component

Total

% of variance

Cumul. %

Variable

PC1

PC2

PC3

PC1 PC2 PC3

2.058 1.576 1.156

29.398 22.509 16.514

29.398 51.907 68.421

Temperature Chlorophyll-a Secchi depth Phosphate Nitrate + nitrite Cyanobacterial richness Microcystin-like activity

0.75 0.751 0.237 −0.596 −0.666 −0.088 −0.079

−0.045 0.284 0.068 0.014 0.109 0.822 −0.798

0.205 0.316 0.878 0.63 0.181 −0.084 −0.176

species whose responses to perturbation affect the outcome of competitive interactions (Armstrong and McGehee, 1980; Hastings, 1980; Gons et al., 1991; Huisman and Weissing, 1999). The importance of light and physical variables in general, as factors shaping phytoplankton assemblages, has been re-emphasised in recent years (Zohary et al., 2010). Owing to the decreases in Secchi depth and phytoplankton biomass recorded in Ballyquirke Lough at the end of July, which was concomitant to a sharp decrease of chlorophyll-a at the time, light penetration is likely to have substantially contributed to the shaping of the cyanobacterial community, probably favouring gas vesicles containing species, such as Microcystis spp. Wind-driven turbulence might have played a greater role in Lough Corrib, which is both deeper and of greater size. Vertical profiling carried out in Upper Lough Corrib showed periods of stratification and temperature homogeneity during June and August. This could not be examined during July due to instrument malfunction. 4.2. Water body connectivity From a metacommunity structuring perspective, it is difficult to separate dispersal-related effects from those caused by environmental forcing. Phytoplankton assembly at spatial scales in the order of kilometers involves complex processes with environmental gradients affecting taxonomic composition (Leibold et al., 2004). Even though they are linked

on the surface via a short river, the Ballyquirke Lough outlet connecting into the Lower Corrib is distant by ~ 20 km from station 4 where the samples used for community composition assessment were collected. Owing to the natural north to south flow of water through Lough Corrib, the rapid dispersion of cyanobacterial genotypes from Ballyquirke Lough to the Upper Corrib via Lower Corrib is questionable. However, this situation is not to be dismissed as under prevailing westerly winds, numerical model simulations have shown complex surface and nearbed water circulation, which could promote dispersion (Dabrowski and Berry, 2009). A delay in the onset of population development in both lakes would be anticipated under such scenario and would return differential detection by 16S rDNA DGGE analysis. However, in order to capture such events in greater detail, sampling frequency and location in both lakes should have been substantially increased. Dispersal depends on the connectivity between habitats, which is determined by geographic distances and the existence of dispersal barriers (Ricketts, 2001; Angeler et al., 2010; Horváth et al., 2015). The degree of connectivity and geographic proximity between Lough Corrib and Ballyquirke Lough warrants for dispersal and potential similarity in microbial community composition. However, the distinct limnology and contrasting topographies encountered in the two water bodies during the summer would probably force the selection of species more suited to those two different niches.

Fig. 10. Sample projections along the first two principal components highlighting the segregation of cyanobacteria rich samples from microcystin-containing samples in Lough Corrib and Ballyquirke Lough during summer 2010. The selected variables for the analysis included water temperature, chlorophyll-a, Secchi depth, phosphate and nitrate concentrations, cyanobacterial GG-II richness and PP2A-derived microcystin concentration.

426

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

4.3. Lake morphology and niche typology Lough Corrib and Ballyquirke Lough have previously been the centre of attention owing to their regional importance for commercial fisheries, leisure fishing activities or water abstraction (Irvine et al., 1998; McCarthy et al., 2001). Noticeably, the Secchi depths and chlorophylla and phosphate concentrations in the lakes recorded in the late 1990s were similar to those determined in the present study. This suggests, despite the land development that occurred in the catchments over the past 20 years, that the trophic status of both lakes has probably remained stable. Seven stations were sampled on a near weekly basis in the Upper Corrib during summer 2010 and analysis of limnological variables showed spatial variability in their distribution. In particular, were recorded on several occasions in domain-A, delineated by stations 7–5, greater temperatures, Secchi depths, chlorophyll-a and nutrient levels compared to domain-B, extending from stations 4–1. This was not surprising given the topography of the water basin in those areas. Domain-A is a larger and deeper area susceptible to wind exposure and water column mixing. On the other hand, domain-B is a shallow and narrow straight. This body of water is hence more amenable to rapid warm up and can be protected from wind-driven turbulence by topographical features, which can contribute to phytoplankton biomass development. The separation of water masses in domains A and B of the Upper Corrib was evidenced by the thermal stratification that was detected on occasions at station 4. Noticeably, the composition of genotype-II cyanobacterial community by 16S rDNA semi-nested DGGE analysis, although largely similar, showed minor variations in the two domains with the presence and absence of particular bands (data not shown). Bands were also of greater intensity in domain-B, probably highlighting the greater abundance of cyanobacteria and phytoplankton in the area, as reflected by greater chlorophyll-a levels. These differences need to be viewed in relation to the topographylinked temperature gradients previously identified in Lough Corrib and which constitute important pathways for the transport of material from the periphery to the deeper regions of the lake (Cannaby et al., 2007). Such circulation mechanisms are important in calm conditions as density differences and rapid cooling of shallow areas create cascading with potential biological implications, such as the directional dispersion of planktonic microbes. As highlighted by Dabrowski and Berry (2009), such pathways may have implications for representative sampling for lake monitoring and management.

It is likely for a toxic population that variation in the intensity of amplified mcy gene bands would reflect variation in biomass. However, the competitive replacement of toxic genotypes by nontoxic ones offers a plausible explanation for the gradual decrease in average toxicity per cell that has been observed during the development of dense Microcystis blooms (Kardinaal et al., 2007). This aspect needs to be considered further in cases where 1) potentially toxic populations do not reach high biomass bloom proportions, 2) the remainder of the phytoplankton community contributes substantial biomass and 3) where light penetration is naturally low due to the coloration of the lake. Ballyquirke Lough would provide a suitable water body to examine such aspects. 4.5. Detection of microcystin-like activity PP2A results from the samples collected in the two lakes returned concentration estimates well below the WHO recommended level of 1 μg·l−1, except on one occasion in August where visible agglomerations of spherical cyanobacterial colonies were sampled in Ballyquirke Lough. Noticeably, substantial PP2A inhibition coincided with the detection of mcy gene bands of greater intensity. Relating PP2A inhibition to the detection of mcy genes warrants caution, as gene presence does not necessarily translate into toxin synthesis (Beversdorf et al., 2015). There also exist many variants of microcystins displaying a wide range of toxicities or toxin profiles in single populations can be complex and vary according to the physiological state of cells (Kardinnal and Visser, 2005). Alternatively, PP2A inhibition could be caused by co-extracted compounds concomitant to the non-detection of mcy genes, or low amounts of weakly toxic variants could be synthesised, returning weak PP2A responses but allowing the detection of genes associated with their synthesis. Overall, the MFA for the two lakes returned on PC1 the expected pattern of positively correlated increasing temperature and chlorophyll-a concentration to be opposed to low macronutrient concentrations. Noticeably, the analysis also highlighted on PC2 the negative relationship between microcystin concentration and cyanobacterial diversity. This pattern, only observed at the scale of two lakes, would need to be extended to a regional scale to add substantial support to the hypothesis that microcystin toxins might display allelopathic effects onto competing phytoplankton species, a mechanism which has been verified via laboratory-based experiments (Bittencourt-Oliveira et al., 2015). 5. Conclusion

4.4. Detection of toxigenic populations Microcystins are cyanotoxins for which the World Health Organization (WHO) has issued recommendatory levels (Sivonen and Jones, 1999). Genes of the microcystin operon have proved useful to identify toxigenic cyanobacterial genera in culture and field samples (Christiansen et al., 2003; Hisbergues et al., 2003; Garneau et al., 2015). The presence of the mcy-A and -E genes was detected in surface water extracts from Lough Corrib and Ballyquirke Lough. In Lough Corrib, the intensity of mcy bands in August was consistent with the apparition from the end of July of band 38 in the DGGE gel, identified as Microcystis sp., one of the most important microcystin toxin-producing genera in lakes (Sivonen and Jones, 1999). In Ballyquirke Lough, the mcy-A + E RFLP profile was similar throughout July and August, albeit those in August were more intense, which was concomitant to the observation in the DGGE gel of band 18, identified as the potentially toxic Woronichinia sp. The difference observed in the RFLP profiles for the two lakes possibly reflected the presence of different microcystin-producing species. This variation must also be considered in relation to the presence and absence of genotype-I cyanobacteria in Lough Corrib and Ballyquirke Lough, respectively. Even though the DGGE patterns were not complex, some genotype-I bands were detected in Ballyquirke Lough, in particular in August, which could also have corresponded to potential microcystin-synthesisers.

Two connected lakes from the western Irish lake system were sampled during summer 2010 in order to ascertain if there were differences in surface cyanobacterial community dynamics. Community change was observed in both Upper Lough Corrib and Ballyquirke Lough, which also showed differences in nutrient regimes and lake morphology. Noticeably, community succession was concomitant to physico-chemical change in the water column, suggesting forcing influenced by meteorological conditions. In addition, the decrease in community richness observed in August was associated with greater microcystin-like activity, brighter mcy gene PCR bands and the presence of Microcystis sp. as determined by DGGE gel band excision and sequencing. Future work will focus on identifying trends in the dispersal of specific genotypes and will compare the succession dynamics of cyanobacterial genotypes in a regional system of connected lakes. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2016.02.117. Acknowledgments The authors wish to acknowledge R. Ham, K. Kilroy and A. Smith for laboratory assistance. This work was supported by the Irish Environmental Protection Agency through the STRIVE programme 2007–2013 (fellowship 2008-FS-EH-3-S5).

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

References Abed, R.M.M., Safi, N.M.D., Köster, J., de Beer, D., El-Nahhal, Y., Rullkötter, J., Garcia-Pichel, F., 2002. Microbial diversity of a heavily polluted microbial mat and its community changes following degradation of petroleum compounds. Appl. Environ. Microbiol. 68, 1674–1683. Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1990. Basic local alignment search tool. J. Mol. Biol. 215, 403–410. Angeler, D.G., Alvarez-Cobelas, M., Rojo, C., Sanchez-Carrillo, S., 2010. Phytoplankton community similarity in a semiarid floodplain under contrasting hydrological connectivity regimes. Ecol. Res. 25, 513–520. Armstrong, R.A., McGehee, R., 1980. Competitive exclusion. Am. Nat. 115, 151–170. Beard, S.J., Handley, B.A., Hayes, P.K., Walsby, A.E., 1999. The diversity of gas vesicle genes in Planktothrix rubescens from Lake Zürich. Microbiology 145, 2757–2768. Bell, T., Newman, J.A., Silverman, B.W., Turner, S.L., Lilley, A.K., 2005. The contribution of species richness and composition to bacterial services. Nature 436, 1157–1160. Beversdorf, L.J., Chaston, S.D., Miller, T.R., McMahon, K.D., 2015. Microcystin mcyA and mcyE gene abundances are not appropriate indicators of microcystin concentrations in lakes. PLoS One http://dx.doi.org/10.1371/journal.pone.0125353. Bittencourt-Oliveira, M.C., Chia, M.A., Bezerra de Oliveira, H.S., Cordeiro Araujo, M.K., Molica, R.J.R., Dias, C.T.S., 2015. Allelopathic interactions between microcystinproducing and non-microcystin-producing cyanobacteria and green microalgae: implications for microcystins production. J. Appl. Phycol. 27, 275–284. Boutte, C., Grubisic, S., Balthasart, P., Wilmotte, M., 2006. Testing of primers for the study of cyanobacterial molecular diversity by DGGE. J. Microbiol. Methods 65, 542–550. Briand, E., Escoffier, N., Straub, C., Sabart, M., Quiblier, C., Humbert, J.F., 2009. Spatiotemporal changes in the genetic diversity of a bloom-forming Microcystis aeruginosa (cyanobacteria) population. ISME J. 3, 419–429. Cannaby, H., White, M., Bowyer, P., 2007. Dynamical response to differential rates of temperature change over sloping topography: Lough Corrib, Ireland. J. Geophys. Res. Oceans Atmos. 112, C09009. http://dx.doi.org/10.1029/2006JC003890. Carpenter, S.R., Cole, J.J., Kitchell, J.F., Pace, M.L., 1998. Impact of dissolved organic carbon, phosphorus, and grazing on phytoplankton biomass and production in experimental lakes. Limnol. Oceanogr. 43, 73–80. Christiansen, G., Fastner, J., Erhard, M., Borner, T., Dittman, E., 2003. Microcystin biosynthesis in Planktothrix: genes, evolution, and manipulation. J. Bacteriol. 185, 564–572. Codd, G.A., Lindsay, J., Young, F.M., Morrison, L.F., Metcalf, J.S., 2005. Harmful cyanobacteria: from mass mortalities to management measures. In: Huisman, J., Matthijs, H.C.P., Visser, P.M. (Eds.), Harmful Cyanobacteria. Springer, The Netherlands, pp. 1–23. Dabrowski, T., Berry, A., 2009. Use of numerical models for determination of best sampling locations for monitoring of large lakes. Sci. Total Environ. 407, 4207–4219. Dokulil, M.T., Teubner, K., 2000. Cyanobacterial dominance in lakes. Hydrobiologia 438, 1–12. Donohue, I., Styles, D., Coxon, C., Irvine, K., 2005. Importance of spatial and temporal patterns for assessment of risk of diffuse nutrient emissions to surface waters. J. Hydrol. 304, 183–192. Donohue, I., Jackson, A.L., Pusch, M.T., Irvine, K., 2009. Nutrient enrichment homogenizes lake benthic assemblages at local and regional scales. Ecology 90, 3470–3477. EC Parliament and Council, 2000. Directive of the European Parliament and of the Council 2000/60/EC Establishing a Framework for Community Action in the Field of Water Policy. European Commission, Luxembourg (PE-CONS 3639/1/100 Rev 1). Escofier, B., Pages, J., 1990. Analyses factorielles simples et multiples: objectifs, methodes, interpretations. Dunod, Paris. Frangeul, L., Quillardet, P., Castets, A.M., Humbert, J.F., Matthijs, H.C.P., Cortez, D., Tolonen, A., Zhang, C.C., Gribaldo, S., Kehr, J.C., Zilliges, Y., Ziemert, N., Becker, S., Talla, E., Latifi, A., Billault, A., Lepelletier, A., Dittmann, E., Bouchier, C., Tandeau de Marsac, N., 2008. Highly plastic genome of Microcystis aeruginosa PCC 7806, a ubiquitous toxic freshwater cyanobacterium. BMC Genomics 9, 274. http://dx.doi.org/10.1186/14712164-9-274. Garneau, M.E., Posch, T., Pernthaler, J., 2015. Seasonal patterns of microcystin-producing and non-producing Planktothrix rubescens genotypes in a deep pre-alpine lake. Harmful Algae 50, 21–31. Geiss, U., Selig, U., Schumann, R., Steinbruch, R., Bastrop, R., Hagemann, M., Schoor, A., 2004. Investigations on cyanobacterial diversity in a shallow estuary (Southern Baltic Sea) including genes relevant to salinity resistance and iron starvation acclimation. Environ. Microbiol. 6, 377–387. Giovannoni, S., 2004. Oceans of bacteria. Nature 430, 515–516. Gons, H.J., Otten, J.H., Rijkeboer, M., 1991. The significance of wind resuspension for the predominance of filamentous cyanobacteria in a shallow, eutrophic lake. Mem. Ist. Ital. Idrobiol. 48, 233–249. Hastings, A., 1980. Disturbance, coexistence, history, and competition for space. Theor. Popul. Biol. 18, 363–373. Hisbergues, M., Christiansen, G., Rouhiainen, L., Sivonen, K., Borner, T., 2003. PCR-based identification of microcystin-producing genotypes of different cyanobacterial genera. Arch. Microbiol. 180, 402–410. Horváth, Z., Vad, C.F., Ptacnik, R., 2015. Wind dispersal results in a gradient of dispersal limitation and environmental match among discrete aquatic habitats. Ecography http://dx.doi.org/10.1111/ecog.01685. Huisman, J., Weissing, F.J., 1999. Biodiversity of plankton by species oscillations and chaos. Nature 402, 407–410. Irvine, K., Allott, N., Caroni, R., DeEyto, E., Free, G., White, J., 1998. The water quality and ecological status of Ballyquirke, County Galway. Report to Galway County Council. Galway County Council. Istvánovics, V., 2009. Eutrophication of lakes and reservoirs. In: Likens, G.E. (Ed.), Encyclopedia of Inland Waters 1. Elsevier, Oxford, pp. 157–165.

427

Iteman, I., Rippka, R., Tandeau de Marsac, N., Herdman, M., 2000. Comparison of conserved structural and regulatory domains within divergent 16S rRNA–23S rRNA spacer sequences of cyanobacteria. Microbiology 146, 1275–1286. Jungblut, A.D., Neilan, B.A., 2006. Molecular identification and evolution of the cyclic peptide hepatotoxins, microcystin and nodularin, synthetase genes in three orders of cyanobacteria. Arch. Microbiol. 185, 107–114. Kaneko, T., Nakajima, N., Okamoto, S., Suzuki, I., Tanabe, Y., Tamaoki, M., Nakamura, Y., Kasai, F., Watanabe, A., Kawashima, K., Kishida, Y., Ono, A., Shimizu, Y., Takahashi, C., Minami, C., Fujishiro, T., Kohara, M., Katoh, M., NakazakI, N., Nakayama, S., Yamada, M., Tabata, S., Watanabe, M.M., 2007. Complete genomic structure of the bloom-forming toxic cyanobacterium Microcystis aeruginosa NIES-843. DNA Res. 14, 247–256. Kardinaal, W.E.A., Janse, I., van Agterveld, M.K., Meima, M., Snoek, J., Mur, L.R., Huisman, J., Zwart, G., Visser, P.M., 2007. Microcystis genotype succession in relation to microcystin concentrations in freshwater lakes. Aquat. Microb. Ecol. 48, 1–12. Kardinnal, W.E.A., Visser, P.M., 2005. Dynamics of cyanobacterial toxins: Sources of variability in microcystin concentrations. In: Huisman, J., Matthijs, H.C.P., Visser, P.M. (Eds.), Harmful cyanobacteria. Springer, The Netherlands, pp. 41–63. Kurmayer, R., Christiansen, G., Fastner, J., Börner, T., 2004. Abundance of active and inactive microcystin genotypes in populations of the toxic cyanobacterium Planktothrix spp. Environ. Microbiol. 6, 831–841. Leibold, M.A., Holyoak, M., Mouquet, N., Amarasekare, P., Chase, J.M., Hoopes, M.F., Holt, R.D., Shurin, J.B., Law, R., Tilman, D., Loreau, M., Gonzalez, A., 2004. The metacommunity concept: a framework for multi-scale community ecology. Ecol. Lett. 7, 601–613. Litvaitis, M.K., 2002. A molecular test of cyanobacterial phylogeny: inferences from constraint analyses. Hydrobiologia 468, 135–145. Lorenzen, C.J., 1967. Determination of chlorophyll and pheo-pigments: spectrophotometric equations. Limnol. Oceanogr. 12, 343–346. Marsalek, B., Blaha, L., Hindak, F., 2000. Review of toxicity of cyanobacteria in Slovakia. Biologia 55, 645–652. McCarthy, T.K., Barbiero, R., Doherty, D., Cullen, P., Ashe, P., O'Connell, M., Guiry, M., Sheehy-Skeffington, M., King, M., O'Connor, B., 2001. Investigation of eutrophication processes in the littoral zones of western Irish lakes. Synthesis Report for Environmental Protection Agency. Environmental Protection. Mur, L.R., Skulberg, O.M., Utkilen, H., 1999. Cyanobacteria in the environment. In: Chorus, I., Bartram, J. (Eds.), Toxic Cyanobacteria in Water Monitoring and Management. E & FN Spon, London, pp. 15–40. Neilan, B.A., Dittman, E., Rouhiainen, L., Bass, R.A., Schaub, V., Sivonen, K., Borner, T., 1999. Nonribosomal peptide synthesis and toxigenicity of cyanobacteria. J. Bacteriol. 181, 4089–4097. Neilson, J.W., Jordan, F.L., Maier, R.M., 2013. Analysis of artifacts suggests DGGE should not be used for quantitative diversity analysis. J. Microbiol. Methods 92, 256–263. Nübel, U., Garcia-Pichel, F., Muyzer, G., 1997. PCR primers to amplify 16S rRNA genes from cyanobacteria. Appl. Environ. Microbiol. 63, 3327–3332. Pan, H., Song, L.R., Liu, Y.D., Borner, T., 2002. Detection of hepatotoxic Microcystis strains by PCR with intact cells from both culture and environmental samples. Arch. Microbiol. 178, 421–427. Papastergiadou, E., Kagalou, I., Stefanidis, K., Retalis, A., Leonardos, I., 2010. Effects of anthropogenic influences on the trophic state, land uses and aquatic vegetation in a shallow Mediterranean lake: implications for restoration. Water Resour. Manag. 24, 415–435. Park, J.W., Crowley, D.E., 2010. Nested PCR bias: a case study of Pseudomonas spp. in soil microcosms. J. Environ. Monit. 12, 985–988. Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14, 817–818. Ricketts, T.H., 2001. The matrix matters: effective isolation in fragmented landscapes. Am. Nat. 158, 87–99. Rivasseau, C., Racaud, P., Deguin, A., Hennion, M.C., 1999. Development of a bioanalytical phosphatase inhibition test for the monitoring of microcystins in environmental water samples. Anal. Chim. Acta 394, 243–257. Rouhiainen, L., Vakkilainen, T., Siemer, B.L., Buikema, W., Haselkorn, R., Sivonen, K., 2004. Genes coding for hepatotoxic heptapeptides (microcystins) in the cyanobacterium Anabaena strain 90. Appl. Environ. Microbiol. 70, 686–692. Runnegar, M., Berndt, N., Kong, S.M., Lee, E.Y., Zhang, L., 1995. In vivo and in vitro binding of microcystin to protein phosphatases 1 and 2A. Biochem. Biophys. Res. Commun. 216, 162–169. Schindler, D.W., 2006. Recent advances in the understanding and management of eutrophication. Limnol. Oceanogr. 51, 356–363. Simon, J.F., Vemoux, J.P., 1994. Highly sensitive assay of akadaic acid using protein phosphatase and paranitrophenyl phosphate. Nat. Toxins 2, 293–301. Sivonen, K., Jones, G., 1999. Cyanobacterial toxins. In: Chorus, I., Bartram, J. (Eds.), Toxic Cyanobacteria in Water: A Guide to Their Public Health Consequences, Monitoring and Management. E & FN Spon, London, pp. 41–111. Song, T.Y., Martensson, L., Eriksson, T., Zheng, W.W., Rasmussen, U., 2005. Biodiversity and seasonal variation of the cyanobacterial assemblage in a rice paddy field in Fujian, China. FEMS Microbiol. Ecol. 54, 131–140. Swofford, D., 2002. PAUP: Phylogenetic Analysis Using Parsimony. Sinauer Associates, Sunderland, MA. Taton, A., Grubisic, S., Brambilla, E., De Wit, R., Wilmotte, A., 2003. Cyanobacterial diversity in natural and artificial microbial mats of Lake Fryxell (McMurdo Dry Valleys, Antarctica): a morphological and molecular approach. Appl. Environ. Microbiol. 69, 5157–5169. Tillett, D., Parker, D.L., Neilan, B.A., 2001. Detection of toxigenicity by a probe for the microcystin synthetase A gene (mcyA) of the cyanobacterial genus Microcystis: comparison of toxicities with 16S rRNA and phycocyanin operon (phycocyanin intergenic spacer) phylogenies. Appl. Environ. Microbiol. 67, 2810–2818.

428

N. Touzet et al. / Science of the Total Environment 553 (2016) 416–428

Touzet, N., McCarthy, D., Fleming, G.T.A., 2013. Molecular fingerprinting of lacustrian cyanobacterial communities: mesoscale patterns in summer diversity. FEMS Microbiol. Ecol. 86, 444–457. Vezie, C., Brient, L., Sivonen, K., Bertru, G., Lefeuvre, J.C., Salkinojasalonen, M., 1998. Variation of microcystin content of cyanobacterial blooms and isolated strains in lake Gand-lieu (France). Microb. Ecol. 35, 126–135. WHO, 2003. Guidelines for Safe Recreational Water Environments, Vol 1, Coastal and Fresh Waters. World Health Organization, Switzerland, Geneva. Willmotte, A.Y., Herdman, M., 2001. Phylogenetic relationships among the cyanobacteria based on 16S rRNA sequences. In: Boone, D., Castenhol, R., Garrity, G. (Eds.), Bergey's Manual of Systematic Bacteriology. Springer, New York, pp. 487–514.

Zehr, J.P., Mellon, M.T., Hiorns, W.D., 1997. Phylogeny of cyanobacterial nifH genes: evolutionary implications and potential applications to natural assemblages. Microbiology 143, 1443–1450. Zeidner, G., Preston, C.M., Delong, E.F., Massana, R., Post, A.F., Scanlan, D.J., Beja, O., 2003. Molecular diversity among marine picophytoplankton as revealed by psbA analyses. Environ. Microbiol. 5, 212–216. Zohary, T., Padisak, J., Naselli-Flores, L., 2010. Phytoplankton in the physical environment: beyond nutrients, at the end, there is some light. Hydrobiologia 639, 261–269.