Finescale spatial variability in anatoxina and homoanatoxina ...

Journal of Applied Microbiology ISSN 1364-5072

ORIGINAL ARTICLE

Fine-scale spatial variability in anatoxin-a and homoanatoxin-a concentrations in benthic cyanobacterial mats: implication for monitoring and management S.A. Wood1, M.W. Heath2, J. Kuhajek1 and K.G. Ryan2 1 Cawthron Institute, Nelson, New Zealand 2 School of Biological Sciences, Victoria University of Wellington, Wellington, New Zealand

Keywords anatoxin-a, benthic mats, cyanobacteria, homoanatoxin-a, liquid chromatography–mass spectrometry, polymerase chain reaction. Correspondence Susanna A. Wood, Cawthron Institute, Private Bag 2, Nelson 7042, New Zealand. E-mail: [email protected]

2010 ⁄ 0994: received 11 June 2010, revised 15 July 2010 and accepted 16 July 2010 doi:10.1111/j.1365-2672.2010.04831.x

Abstract Aims: The purpose of this study was to determine the variability in anatoxin-a (ATX) and homoanatoxin-a (HTX) concentrations in benthic cyanobacterial mats within sampling sites and to assess the applicability of using a PCR-based approach to determine ATX- and HTX-production potential. Methods and Results: ATX and HTX variability was investigated by collecting 15 samples from 10 · 10 m grids in seven rivers. ATX and HTX concentrations were determined using liquid chromatography–mass spectrometry (LC–MS). Samples from two sites contained no ATX or HTX and at one site ATX and HTX were detected in all samples. At four sites, both toxic and nontoxic samples co-occurred and these samples were sometimes spaced less than 1 m apart. PCR amplification of a region of a polyketide synthase (ks2, putatively involved in the biosynthetic pathway of ATX and HTX) successfully distinguished ATX-and-HTX- and non-ATX-and-HTX-producing cultured Phormidium strains. Results from environmental samples were more variable, and the results were in congruence with the LC–MS data in only 58% of samples. Conclusions: Fine-scale spatial variability in ATX and HTX concentrations occurs among benthic cyanobacterial mats. Significance and Impact of the Study: Multiple benthic cyanobacterial mat samples must be collected at a sampling site to provide an accurate assessment of ATX and HTX concentrations at that location. The PCR-based technique offers the potential to be a useful early warning technique.

Introduction An increasing number of cyanobacterial species are now known to produce natural toxins, known as cyanotoxins (Sivonen and Jones 1999). These toxins are a health threat to humans and animals when consumed or when there is contact with contaminated water. Water users have a growing awareness of the health risks associated with toxic planktonic cyanobacteria; however, the dangers of benthic cyanobacteria are less widely acknowledged. Although benthic taxa have received less attention than their planktonic counterparts, they are increasingly recognized as problematic and produce a range of cyanotoxins

(Edwards et al. 1992; Carmichael et al. 1997; Mez et al. 1997; Seifert et al. 2007). Animal poisonings associated with the consumption of benthic cyanobacteria and the presence of the neurotoxins anatoxin-a (ATX) and ⁄ or homoanatoxin-a (HTX) have been reported with increasing frequency worldwide (Hamill 2001; Wood et al. 2007; Cadel-Six et al. 2007). Benthic, mat-forming cyanobacteria are widespread throughout New Zealand rivers (Biggs and Kilroy 2000); the most common genus is Phormidium. During stable flow conditions, Phormidium can proliferate, forming expansive black ⁄ brown leathery mats across large areas of river substrate. Routine sampling of mats from around

ª 2010 The Authors Journal of Applied Microbiology 109, 2011–2018 ª 2010 The Society for Applied Microbiology

2011

Spatial variability in anatoxin-a and homoanatoxin-a in cyanobacterial mats

New Zealand has shown marked variation in the presence and concentrations of ATX and HTX. This has led to uncertainty regarding variables regulating toxin production and best sampling practices. Previous studies have demonstrated that both toxic and nontoxic genotypes co-occur in benthic cyanobacterial proliferations and that there is no correlation between ATX and HTX concentrations and the percentage of the river substrate that the cyanobacterial mats occupy (Heath et al. 2010; Cadel-Six et al. 2007). One of the aims of this study was to assess variability in the presence of ATX and HTX within sampling sites with the goal of establishing a methodology for collecting representative benthic cyanobacterial samples for use in risk assessments. Our previous research identified differences in 16S rRNA gene sequences, and sequence and length variability in the intergenic spacer (ITS) region between toxic and nontoxic Phormidium autumnale strains (Heath et al. 2010). This suggested the potential for toxic strains to be distinguished from nontoxic strains using molecular techniques. Automated rRNA intergenic spacer analysis (ARISA) is a recently developed DNA finger-printing method (Fisher and Triplett 1999) that exploits the length heterogeneity of the ITS region. In this study, cyanobacterial-specific ARISA primers (Wood et al. 2008b) were used to determine whether differences in ARISA fragment lengths (AFL) profiles could be used to predict which samples contained toxic Phormidium strains and to assess whether the cyanobacterial community structure of each mat influenced toxin production. Cadel-Six et al. (2009) identified a sequence (ks2) coding for a polyketide synthase that was present only in ATX or HTX-producing Oscillatoria sp., indicating its putative involvement in the biosynthesis of these toxins. We also assessed the applicability of using a PCR-based approach to screen environmental samples for ks2 and thus ATX- and HTX-production potential. Materials and Methods Cyanobacterial cultures Twenty-six cyanobacterial strains (Table 1), isolated as part of previous projects (Wood et al. 2008a; Heath et al. 2010) were screened to determine whether the presence of the ks2 gene fragment correlated with the detection of ATX and HTX. ATX and HTX were initially detected in Ph. autumnale strain CYN49; however, these toxins were undetectable in subsequent testing of subcultured samples of the strain (Wood et al. 2008a). Frozen material from an early frozen sample of CYN49, which tested positive for HTX and ATX, as well as the subsequent subcultured strain, which tested negative for HTX and ATX, were included in this study. 2012

S.A. Wood et al.

Table 1 Culture code, species identification, location of sample used for isolation, presence or absence of ATX and HTX and the ks2 fragment Code

Species

Location

Toxin

ks2

VUW3

)

)

CYN53 CYN55

) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ATX + and HTX ATX + ) )

Phormidium Wainuiomata River autumnale VUW4 Ph. autumnale Hutt River VUW5 Ph. autumnale Waingongoro River VUW6 Pseudanabaena sp. Wainuiomata River VUW7 Ph. autumnale Wainuiomata river VUW8 Ph. autumnale Akatarawa River VUW9 Ph. autumnale Hutt River VUW10 Ph. autumnale Wainuiomata River VUW11 Ph. autumnale Pembroke Road VUW12 Ph. autumnale Wainuiomata River VUW14 Ph. autumnale Hutt River VUW15 Pseudanabaena sp. Whakatiki River VUW16 Ph. autumnale Pelorous River VUW17 Ph. autumnale Mangatinoka Stream VUW18 Ph. autumnale Makarewa River VUW19 Ph. autumnale Mangaroa River VUW20 Ph. autumnale Rangataiki River VUW22 Ph. autumnale Waimana River CYN38 Ph. murrayi Red Hills Tarn CYN39 Ph. murrayi Red Hills Tarn CYN47 Ph. autumnale Ashley River CYN48 Ph. autumnale Ashley River CYN49 Ph. autumnale Hutt River CYN49* Ph. autumnale Hutt River Ph. autumnale Ph. autumnale

Rangataiki River Roding River

ATX, anatoxin-a; HTX, homoanatoxin-a. *Frozen archived material from when this culture was positive for ATX ⁄ HTX.

Environmental samples site description and sample collection Benthic cyanobacterial mats were collected from seven New Zealand rivers between January and March 2009 (Table 2). At each river, a 10 · 10 m grid was set up in a riffle (a shallow region of a river where the surface is broken into ripples or waves by totally or partially submerged obstructions). At each site, cyanobacterial mat coverage was measured in five 1 m2 quadrants that were randomly positioned within the grid. Random numbers were used to select fifteen sampling points within grids (Fig. 1). At each sampling point, cyanobacterial samples were collected by scraping a mat from one rock into sterile Falcon tubes (15 ml). Where no mats were present at the sampling point, the closest upstream mat was selected for sampling. Samples were frozen ()20C) for later molecular and toxin analysis. Subsamples (5 ml) were preserved using Lugol’s iodine for morphological identification.


S.A. Wood et al.


Table 2 Sampling sites, sampling dates and percentage of river substrate covered with benthic cyanobacterial mat Percentage Name

Location

Sampling date

coverage

Hutt River

4111¢29¢S, 17455¢55¢E

23 January 2009

30

Mangatarere

4103¢20¢S, 17529¢51¢E

16 March 2009

40

4005¢38¢S, 17507¢35¢E

17 March 2009

80

4055¢21¢S, 17538¢26¢E

15 March 2009

30

Ashley River

4117¢00¢S, 17232¢46¢E

24 March 2009

20

Maitai River

4116¢16¢S, 17318¢86¢E

20 March 2009

15

Waimea

4118¢41¢S, 17307¢41¢E

4 April 2009

20

North Island

Stream Managroa River Waipoua River South Island

River

Morphological identification The dominant cyanobacterium in each mat was identified by microscopy (BX51, Olympus, Wellington, New Zealand). Detection of ATX and HTX and statistical analysis Subsamples of all cyanobacterial mats were lyophilized and resuspended (100 mg) in 10 ml of MilliQ water containing 0Æ1% formic acid. Samples were sonicated (15 min) and centrifuged (4000 g, 10 min). The supernatant was collected and a second extraction undertaken on the pellet. The supernatants were combined and an aliquot analysed directly for ATX, HTX and their degradation products, dihydroanatoxin-a (dhATX), dihydrohomoanatoxin-a (dhHTX), epoxyanatoxin-a (epoxyATX) and epoxyhomoanatoxin-a (epoxyHTX) using liquid chromatography–mass spectrometry (LC–MS) as described in Heath et al. (2010). Toxin data from sites that tested positive for ATX and ⁄ or HTX were pooled and the percentage of ATX and ⁄ or HTX occurrence in these samples calculated. Using binomial probabilities, the number of samples required to detect ATX and ⁄ or HTX with 95, 99 and 100% confidence was determined. Additionally, the numbers of samples required to detect ATX and ⁄ or HTX with 95 and 99% confidence at a site with a prior determined percentage of ATX and HTX was modelled. Isolation of DNA, ARISA fingerprinting, ks2 PCR and analysis Six samples from each site were selected for PCR-based ks2 analysis and ARISA (Table 3). These samples were selected to provide a cross-section of toxic and nontoxic

(where available) samples from each site. DNA from each selected environmental sample and from all 26 laboratory cultures was extracted from approximately 0Æ05 g of lyophilized material using the Power Soil kit (Mo Bio Laboratories, Inc., Carlsbad, CA, USA) according to the manufacturer’s protocol. ARISA PCRs and analysis were carried out using cyanobacterial specific primers as described previously (Wood et al. 2008b). Nonmetric multidimensional scaling (MDS) based on Bray–Curtis similarities was undertaken using the primer 6 software package (Primer-E Ltd, Plymouth, UK). Nonmetric MDS was undertaken with 100 random restarts, and results were plotted in two dimensions. Agglomerative, hierarchical clustering of the Bray– Curtis similarities was carried out using the Cluster function of primer 6 and plotted onto the two-dimensional MDS at a similarity level of 40 and 60%. Cultures and environmental samples were screened for the presence of the ks2 fragment using a 25 -ll reaction mixture containing approximately 30 ng of DNA, 480 nmol l)1 of the primers given in Cadel-Six et al. (2009) (Geneworks, Auckland, New Zealand), 200 lM dNTPs (Roche Diagnostics, Auckland, New Zealand), 1· Taq PCR buffer (Invitrogen), 1 U of Platinum Taq DNA polymerase (Invitrogen), 2Æ5 mol l)1 MgCl2 (Invitrogen) and 2Æ0 lg nonacetylated bovine serum albumin (Sigma, Australia). PCRs were run on an iCycler thermal cycler (Bio-Rad, Germany) with the following conditions: 94C for 2 min followed by 94C for 30 s, 50C for 30 s, 72C for 1 min, repeated for 35 cycles with a final extension of 72C for 7 min. Amplicons of the expected size were purified using a High Pure PCR product purification kit (Roche Diagnostics) and sequenced bi-directionally using the BigDye Terminator ver. 3.1 Cycle Sequencing Kit (Applied Biosystems, CA, USA). Results The highest cyanobacterial percentage cover was observed at the Managaroa River site (80%, Table 2). This was markedly higher than other sites where coverage ranged from 15 to 40% (Table 2). The dominant cyanobacteria (identified by microscopy) in all samples was Ph. autumnale. ATX and ⁄ or HTX were detected at five of the seven sites (Fig. 1 and Table 3). The Ashley River was the only site in which all fifteen samples contained toxins. In contrast, only one sample in the Waipoua River site contained toxins (Fig. 1 and Table 3). At sites where both toxic and nontoxic samples were present (Hutt, Waipoua, Maitai and Waimea), no obvious correlation between sample location and ATX and ⁄ or HTX detection was found (Fig. 1).


2013


(a) 10

(b)

8 6 4 2 0 (c) 10

(d)

8 6 4 2 0 (e) 10

(f)

8 6 4 2 0

0

(g) 10

2

4

6

8

10

8 6 4 2 0

0

2

4

6

8

10

Figure 1 Location of samples within 10 · 10 m grids. (a), Hutt River; (b), Mangatarere Stream; (c), Managroa River; (d), Waipoua River; (e), Ashley River; (f), Maitai River; and (g), Waimea River. ( ) anatoxin-a (ATX) and ⁄ or homoanatoxin-a (HTX) detected and ( ) no ATX and ⁄ or HTX detected.

The composition and concentrations of ATX and HTX varied between sites. Twelve of the 15 Hutt River samples contained dhATX with maximal concentrations of 5Æ39 mg kg)1. Low levels of HTX and dhATX were also detected in three of these samples (Table 3). In the Ashley River, despite no ATX being detected, all samples contained dhATX (0Æ4–8Æ89 mg kg)1) and lesser amounts of HTX and dhHTX. The Maitai River samples contained low levels of dhHTX (