Evaluation of taxa-specific real-time PCR, whole-cell FISH and morphotaxonomy analyses for the detection and quantification of the toxic microalgae Alexandrium minutum (Dinophyceae), Global Clade ribotype

Authors


  • Editor: Riks Laanbroek

Correspondence: Nicolas Touzet, The Martin Ryan Institute (r.304), National University of Ireland, Galway, Ireland. Tel.: +353 91 493 231; fax: +353 91 525 005; e-mail: nicolas.touzet@nuigalway.ie

Abstract

The dinoflagellate genus Alexandrium contains neurotoxin-producing species that have adversely affected the aquaculture industry in many countries. The morphological similarity between Alexandrium species has led to the development of molecular methods for the discrimination, enumeration and monitoring of toxic and nontoxic species. A quantitative real-time PCR assay (qRT-PCR) targeting the internal transcribed spacer 1-5.8S rRNA gene using hybridization probe technology was developed for the potentially toxic species Alexandrium minutum (Global Clade) (GC). The assay was specific with a detection limit of less than one cell equivalent. The assay was used to detect and quantify A. minutum (GC) in seawater samples collected during summer 2007 in Cork Harbour, Ireland. The results were compared with those obtained using whole-cell FISH (WC-FISH) and morphotaxonomy analyses. Alexandrium minutum did not reach high bloom concentrations over the sampling period (maximum of c. 6 × 104 cells L−1), and the average concentrations determined using qRT-PCR, WC-FISH and morphotaxonomy did not significantly differ in eight of nine comparisons. Regression curves showed positive relationships between the methods; WC-FISH and qRT-PCR slightly under- and overestimated, respectively, the A. minutum concentrations compared with the morphotaxonomy method. The qRT-PCR assay for A. minutum (GC) offers high-throughput sample analysis and may prove suitable for implementation in microalgae monitoring programmes and assist in population dynamics studies of the species.

Introduction

Occurrences of harmful algal blooms (HABs) have increased in frequency over the past decades in many coastal environments (Hallegraeff, 1993; Glibert et al., 2005). HABs have had a variety of detrimental economical, environmental and public health effects such as cage-reared finfish stock losses, biofouling of recreational areas and seafood poisoning in humans (Shumway, 1990; Hallegraeff & Hara, 1995). Microalgae and biotoxin monitoring programmes have been implemented in many countries to ensure public safety and to protect the aquaculture industry through the provision of early detection systems for toxic blooms and the development of coupled biophysical models of HAB occurrences (Andersen et al., 2003; McGillicuddy et al., 2005).

The cosmopolitan dinoflagellate genus Alexandrium includes a number of toxic species associated with HABs (Anderson, 1998). Several species have been recognized as being responsible for paralytic shellfish poisoning (PSP), an alimentary intoxication syndrome arising from the contamination of shellfish with potent alkaloid neurotoxins collectively known as saxitoxins. In the environment, Alexandrium populations are subject to various biophysical pressures that influence their distribution and abundance. A better understanding of the link between Alexandrium spp. population dynamics and their biophysical environment is therefore necessary to facilitate the management of harmful blooms and enable the prediction of their occurrence and toxicity.

In Ireland, blooms of Alexandrium spp. have mostly been observed along the west and south coasts (Moran et al., 2005). PSP toxicity in shellfish has only been recorded in Cork Harbour, an inlet located on the south coast, where the most likely causative organism has been identified as Alexandrium minutum. A population of this species co-occurs here with a nontoxic form of Alexandrium tamarense (West European ribotype) (Touzet et al., 2007). Previous studies have shown that blooms mainly develop in the period June–July in the North Channel area of Cork Harbour, often reaching cell concentrations >1 × 105 cells L−1 (Ní Rathaille et al., 2009).

Species discrimination within the Alexandrium genus is based on the examination of morphological characteristics (Balech, 1995). Identification is time consuming, requires a high degree of taxonomic expertise and is not reliable when performed by conventional light microscopy, the standard procedure used for monitoring. The high degree of morphological similarity between closely related species and occurrences of toxic and nontoxic populations within a single species also complicate the monitoring of Alexandrium spp. (Lilly et al., 2005; Touzet et al., 2008a). Several molecular techniques have been successfully adapted to enable the rapid detection and quantification of individual HAB species in environmental samples. These methods have relied mainly on the use of taxa-specific molecular probes such as antibodies, lectins and oligonucleotide sequences (Costas et al., 1993; Vrieling et al., 1994; Tyrrell et al., 2001). The development of taxa-specific detection approaches based on the variability of genetic signatures, particularly those of the ribosomal genes, between distantly or closely related microalgal taxa has been widely published (Scholin & Anderson, 1996; Penna & Magnani, 2000; Godhe et al., 2001; Edvardsen et al., 2003). Whole-cell FISH (WC-FISH) assays, sandwich hybridization assays (SHA) and a quartz crystal microbalance biosensor are available for Alexandrium spp., and have allowed the detection of target species through the binding of taxa-specific oligonucleotide probes to the rRNA component of ribosomes present in actively growing cells (Anderson et al., 2005; John et al., 2005; Kim et al., 2005; Lazerges et al., 2006; Touzet & Raine, 2007; Diercks et al., 2008). Several quantitative real-time PCR assays (qRT-PCR) based mostly on SYBR Green, Taqman and molecular beacon technologies have also been developed for a number of HAB species including Alexandrium (Bowers et al., 2000; Galluzzi et al., 2004; Coyne et al., 2005; Hosoi-Tanabe & Sako, 2005; Dyhrman et al., 2006; Moorthi et al., 2006). Galluzzi et al. (2004) designed a genus-specific qRT-PCR assay for Alexandrium spp. and demonstrated its application for quantification of A. minutum in field samples. As the assay was not designed to discriminate to the species level, its value may have limitations when analysing samples collected in areas where several Alexandrium species co-occur.

This study reports the development of a qRT-PCR assay based on the internal transcribed spacer (ITS)1-5.8S rRNA gene region of A. minutum Global Clade (GC) using hybridization probe technology on the LightCycler (Anon, 2000). The assay performance for the detection and quantification of A. minutum in seawater samples collected in Cork Harbour in the period May–August 2007 was assessed by comparing the results with those obtained by morphotaxonomy and WC-FISH analyses.

Materials and methods

Culturing of dinoflagellates

Clonal cultures of A. minutum derived from resting cysts and/or vegetative cells taken along the south and west coasts of Ireland were selected for the design and testing of PCR primers and taxa-specific hybridization probes. A selection of strains of A. minutum (n=15), A. tamarense (n=9), Alexandrium tamutum (n=2), Alexandrium ostenfeldii, Alexandrium peruvianum and Alexandrium andersoni, as well as some strains of dinoflagellate species found in Irish coastal waters, were used for cross-reactivity analysis (Table 1). All cultures were maintained in borosilicate glass tubes containing f/2 medium minus silicate (Guillard, 1975) in an illuminated culture chamber at 15 °C and under a 14 : 10 h light : dark photoperiod cycle (photon flux density of 100 μmol m−2 s−1).

Table 1.   Taxa list, strain designations and origins of the panel of species used for the specificity assessment of the Alexandrium minutum (GC) RT-PCR assay
SpeciesStrain designationsOriginsRibotype*Assay
reactivity
  • The assay reactivity obtained after testing DNA extracts in RT-PCR is indicated.

  • *

    GC and PC refer to the Global and Pacific clades of Alexandrium minutum (Lilly et al., 2005) whereas NA and WE refer to the West European and North American ribotypes of Alexandrium tamarense (Scholin et al., 1995), respectively.

A. minutum (n=11)CK.A02, CK.A14, CK.A17, CK.D05, Kill.C6, Kill.E4, Kill.G3, Kill.H3, SHA.A12, SHA.B11, SHA.C4Cork Harbour, Shannon Estuary, Killary Harbour (Ireland)GC+
A. minutum (n=1)W07/001/01ScotlandGC+
A. minutum (n=1)AL2VGalicia (Spain)GC+
A. minutum (n=1)AL8CCatalunia (Spain)GC+
A. minutum (n=1)AMBOP006New ZealandPC
A. tamutum (n=2)Kill.D5, SHA.A11Shannon Estuary, Killary Harbour (Ireland)
A. insuetum (n=1)ICMB218Catalunia (Spain)
A. ostenfeldii (n=1)BY.K04Bantry Bay (Ireland)
A. peruvianum (n=1)LS.D05Lough Swilly (Ireland)
A. andersoni (n=1)BY.KF4Bantry Bay (Ireland)
A. tamarense (n=7)BL.C10, CB.A02, CK.B01, CK.C01, CK.D01, CK.D02, KY.A0αBelfast Lough, Clew Bay, Cork Harbour, Killary Harbour (Ireland)WE
A. tamarense (n=2)BL.B10, BL.F10Belfast Lough (Ireland)NA
Glenodinium foliaceum (n=1)CK.A08Cork Harbour (Ireland)
Akashiwo sanguinea (n=1)H10Shannon Estuary (Ireland)
Karenia mikimotoi (n=1)CCMP429Plymouth (UK)
Lingulodinium polyedrum (n=1)BY.AB2Bantry Bay (Ireland)
Prorocentrum lima (n=1)CCMP1370Florida (USA)
Prorocentrum micans (n=1)Kill.A04Killary Harbour (Ireland)
Scrippsiella sp. (n=1)KY.A06Killary Harbour (Ireland)

Field sample collection, fixation and processing

Forty-two stations were sampled onboard a vessel throughout the summer of 2007 in Cork Harbour, Ireland (Fig. 1, Table 2). Three surface seawater samples were collected independently at each station. For analysis by WC-FISH, 2-L seawater samples were concentrated by filtration through a 150-μm-mesh sieve and collected on a 5-μm sieve onboard ship. The planktonic material was carefully backwashed into 50-mL tubes with 0.22 μm filtered seawater preserved in formalin (1% final concentration v/v) and stored in darkness for 2–10 h. On land, samples were centrifuged immediately (4500 g, 5 min), the supernatants were carefully discarded by aspiration and the remaining pellets were treated with 15 mL of 100% ice-cold methanol to remove pigments and stabilize nucleic acids before storage at −20 °C. For qRT-PCR and morphology-based analysis with the cellulose-staining dye calcofluor, independent 50-mL samples collected at each station were fixed with Lugol's iodine and stored in darkness.

Figure 1.

 Map of Cork Harbour indicating the stations sampled between May and August 2007.

Table 2.   Details of the stations sampled during the surveys carried out in Cork Harbour
Survey no.StationSampling locationTime (local)Latitude (N)Longitude (W)Sounding (m)
  1. ND, not determined.

1-(23/05/07)5303R16:4051°52.943′8°12.094′7.7
5304Q17:0551°52.905′8°13.054′5.5
5305P17:2751°52.881′8°13.992′2.7
5306O17:5951°52.817′8°15.015′2.6
5307NO18:1551°52.765′8°15.594′5.2
2-(05/06/07)5402O13:3851°52.824′8°15.019′2.0
5403P14:1351°52.866′8°13.958′2.0
5404Q14:4751°52.890′8°12.974′6.4
5405R15:1251°52.932′8°12.071′7.0
5406T15:5251°52.181′8°12.385′7.7
5407U16:1851°51.123′8°12.794′9.5
5410Belvelly Bridge18:0551°53.341′8°18.256′ND
3-(14/06/07)5503NO10:5751°52.752′8°15.495′4.3
5504O11:0551°52.832′8°14.999′3.7
5506P11:4151°52.893′8°13.991′5.0
5507Q12:1951°52.902′8°12.984′6.8
4-(21/06/07)5602NO16:3551°52.760′8°15.493′5.7
5603O16:5351°52.834′8°14.994′3.8
5604P17:0951°52.907′8°13.990′6.1
5605Q17:2051°52.897′8°12.985′7.1
5606R17:4551°52.907′8°12.107′8.7
5-(27/06/07)5728R09:1951°52.912′8°12.043′8.1
5729Q09:4551°52.896′8°12.996′7.0
5730P10:0651°52.906′8°13.995′6.0
5732O10:4251°52.826′8°14.997′2.1
5733NO10:5851°52.752′8°15.407′4.6
6-(04/07/07)5806R14:0451°52.948′8°12.089′6.8
5807Q14:4651°52.882′8°13.004′5.5
5808PQ15:3351°52.884′8°13.492′1.5
7-(10/07/07)5902NO07:4451°52.761′8°15.463′4.9
5903O08:1051°52.829′8°15.010′2.1
5905P08:3151°52.899′8°13.969′5.6
5906Q08:4751°52.891′8°12.985′6.8
5907R09:0351°52.894′8°12.192′8.8
8-(13/07/07)6004NO10:2751°52.756′8°15.482′4.5
6005O10:5451°52.828′8°14.989′2.0
6007P11:1351°52.906′8°13.995′5.8
6008Q11:3651°52.893′8°12.997′6.7
6009R11:5751°52.906′8°12.098′8.4
9-(26/07/07)6101NO09:1051°52.776′-8°15.465′5.2
6102O09:2651°52.841′8°14.979′5.0
6104P09:5051°52.915′8°13.981′6.1
6105Q10:1851°52.905′8°12.988′7.4
6106R10:3251°52.913′8°12.088′8.8
10-(10/08/07)6203P11:2051°52.930′8°14.015′5.3

Morphology-based quantification of A. minutum

Alexandrium minutum identification and quantification was performed using samples settled in an Utermöhl sedimentation chamber (Utermöhl, 1958). Armoured dinoflagellates were stained with calcofluor white (Fritz & Triemer, 1985) and examined with an inverted microscope (Olympus CKX-41) fitted with epifluorescence optics. Identification at the species level was based on Balech's (1995) description of the genus and performed by rotating each Alexandrium cell with a dissecting needle to observe the morphological characteristics of the thecal plates. Species were differentiated by the presence or absence of a ventral pore on the first apical plate (1′) and the shapes of the sixth precingular (6″) and posterior sulcal plates (sp) (Kofoid notation). Observations were made at × 200 and × 400 magnifications.

WC-FISH assay and epifluorescence microscopy

The design and testing of the A. minutum (GC)-specific oligonucleotide probe MinA was performed as described in Touzet & Raine (2007). Cross-reactivity assessments were carried out using Alexandrium spp. and dinoflagellate species found in Irish coastal waters. The probe MinA (5′-TTATATGGTTGATGTGGGTGC-3′), synthesized with the 5′- end labelled with CY.3, was purchased from MWG Biotech, Germany. Upon receipt, concentrated probe stocks were diluted in double-distilled water (final concentration 250 ng μL−1), divided into 100-μL aliquots and stored at −20 °C as working solutions. WC-FISH was carried out with the probe MinA according to a modified version of the protocol described in Miller & Scholin (1998). Aliquots of field samples preserved in methanol were carefully placed onto 13-mm-diameter (1.2 μm pore size) polycarbonate membranes secured in a custom vacuum manifold (Miller & Scholin, 1998), and gentle vacuum was applied to remove liquid. The membranes were then incubated at room temperature for 2 min with 400 μL hybridization buffer (5X SET, 20% formamide, 0.1% IGEPAL and 25 μg mL−1 polyadenylic acid). After filtration, 400 μL hybridization buffer containing 500 ng of probe was added to the filters and incubated in a dark chamber at 55 °C for 60 min. After filtration, membranes were washed once for 1 min with 400 μL 0.2 × SET at 55 °C in the dark to remove excess unbound probes. Filters were carefully removed from the manifold and placed onto a slide. Before mounting the coverslips, 5 μL of a mix of calcofluor (100 μg mL−1) and 4′,6-diamidino-2-phenylindole (DAPI) (3 μg mL−1) was added to the filters as well as 10 μL of SlowFade® Light Antifade reagent to prevent the fading of the fluorescence signals.

Membranes were examined with a microscope (Olympus CKX-41) fitted with a U-RFL-T epifluorescence attachment and a 100 W Mercury lamp. The following filter combinations were used for the detection of the fluorescent signals: calcofluor and DAPI (355DF25 excitation filter, 400DRLP dichroic mirror and 420 long-pass barrier filter) and CY.3 (525AF45 excitation filter, 560DRLP dichroic mirror and 595AF60 band-pass barrier filter). Observations were made at × 200 magnification by scanning the entire filter and counting all positive signals. Probe specificity was confirmed with calcofluor by inspecting the general organization of the Alexandrium plate tabulation for each positive signal recorded.

Sequence alignment, PCR primer and fluorescence resonance energy transfer (FRET) probe design

The ITS1-5.8S-ITS2 rRNA gene sequences of Alexandrium spp. (n=8) isolated from Irish coastal waters obtained in previous studies (Touzet et al., 2007) were compiled with those of other Alexandrium species imported from GenBank (http://www.ncbi.nlm.nih.gov), and aligned with the pairwise alignment function of the Genedoc package (http://www.psc.edu/biomed/genedoc). The GenBank accession numbers for the Alexandrium spp. sequences used for the construction of the partial rRNA gene alignment are as follows: A. minutum CNR AMI-A4, AJ318460; Alexandrium lusitanicum LAC27, ALAJ5050; A. minutum AM, DQ176668; Alexandrium insuetum S1, AB006996; A. tamarense FK-788B, AB006994; Alexandrium catenella ACC01, AJ272120; A. tamarense WKS-1, AB006991; A. catenella M17, AB006990; Alexandrium affine H1, AB006995; A. tamarense CU-15, AB006992; Alexandrium cohorticula ACMS01, AF145224; Alexandrium fraterculus AF-1, AF208242; Alexandrium taylori AV-8, AJ251654; Alexandrium margalefi AM-1, AJ251208; and Alexandrium pseudogonyaulax H1, AB006997 (Fig. 2).

Figure 2.

 Multiple nucleotide sequence alignment of the ITS1-58S rRNA gene region of Alexandrium spp. strains. The black areas indicate the target sites of ITSmin-f1 and ITSmin-r3 primers as well as of the hybridization probes ITSminB-Fl and ITSminB-LC. The taxon designations (species strain) are as follows: 1, Alexandrium lusitanicum LAC27 (GC) (con-specific to Alexandrium minutum; Lilly et al., 2005); 2, A. minutum CNR AMI-A4 (GC); 3, A. minutum CK.A14 (GC); 4, A. minutum SHA.C4 (GC); 5, A. minutum Kill.E4 (GC); 6, A. minutum AM (PC); 7, Alexandrium tamutum SHA.A11; 8, Alexandrium insuetum S1; 9, Alexandrium ostenfeldii BY.K04; 10, Alexandrium peruvianum LS.D05; 11, Alexandrium andersoni BY.KF4; 12, Alexandrium tamarense FK,788B; 13, A. tamarense BL.D10; 14, Alexandrium catenella ACC01; 15, A. tamarense WKS-1; 16, A. catenella M17; 17, Alexandrium affine H1; 18, A. tamarense CU-15; 19, Alexandrium cohorticula ACMS01; 20, Alexandrium fraterculus AF-1; 21, Alexandrium taylori AV-8; 22, Alexandrium margalefi AM-1; 23, Alexandrium pseudogonyaulax H1.

Primers ITSmin-f1 and ITSmin-r3 were 100% homologous to A. minutum (GC) and designed to amplify a PCR product of c. 250 bp within the ITS1-5.8S domain (Table 3). Sensor and anchor probes (ITSminB-Fl and ITSminB-LC, respectively) for the hybridization probe pair were selected based on in silico analysis of the sequence alignment. The sensor probe was designed to contain at least five mis-matches with closely related nontarget species and to have a lower melting temperature (Tm) than the anchor probe. The probes were designed to hybridize to the PCR amplicon in a head-to-tail arrangement with one nucleotide gap so as to facilitate the energy transfer from one fluorescent dye to the other during the polymerization reaction, a process referred to as FRET. Following PCR amplification, the lower Tm of the sensor probe confirms the hybridization probe pair specificity during melt peak analysis, whereby FRET comes to an end as the sensor probe melts off its target and the dyes are separated (Wilhelm & Pingoud, 2003). Primers and hybridization probes were screened using the Basic Local Alignment Search Tool (blast, Altschul et al., 1990) to assess their specificity against the NCBI sequence database. Primers and probes were purchased from Tib-MolBiol, Germany.

Table 3.   Oligonucleotide primers and probes used in the Alexandrium minutum (GC) qRT-PCR assay
DesignationSequence (5′→3′)Tm (°C)
  • *

    Fl and LC640 refer to the dyes fluorescein and LC640, respectively, conjugated to the probes.

  • Tm melting temperature.

ITSmin-f1TTCTGCATCATTTGCT51.7
ITSmin-r3CCGAAGGTGCAAATTA52.3
ITSminB-FlTGCATTGAACTTGAATGTGA – Fl*59.0
ITSminB-LC*LC640-ATGTGTTTTTGCAATGAATGTCT – PH61.2

Sample processing and DNA extraction for RT-PCR

The initial testing of the specificity of the RT-PCR assay was carried out using DNA extracted from cultured Alexandrium spp. and dinoflagellate vegetative cells fixed with Lugol's iodine. Known numbers of A. minutum (GC) cells were used for the assessment of the sensitivity of the RT-PCR assay. Estimations of A. minutum concentrations in cultures were carried out by harvesting cells in triplicate (100 μL) into 1.5-mL tubes containing 1390 μL of filtered seawater and 10 μL of Lugol's iodine. After thorough homogenization, 50-μL aliquots were pipetted into wells of a 96-well plate and cells were enumerated using an inverted microscope. Culture and 25-mL field samples from Cork Harbour were filtered through 25-mm-diameter Whatman Purabind filters (1.2 μm pore size) and the filters were stored in microcentrifuge tubes at −20 °C until nucleic acid extraction.

Before DNA extraction, filters were thawed and 500 μL lysis buffer (1% CTAB, 100 mM Tris-HCl, 20 mM EDTA, and 1 M NaCl) containing silica beads was added. Cell disruption was achieved with two cycles of bead beating (20 s, 6.5 intensity) in a ribolyser (ThermoHybaid-Supplier). DNA extraction was performed on the lysates using the DNeasy® Plant Mini Kit according to the manufacturer's instructions. DNA preparations were quantified by spectrophotometry and stored at –20 °C.

RT-PCR assay for A. minutum (GC)

RT-PCR amplification reactions were performed in 20 μL volumes in glass capillaries in the LighCycler 1.2 (Roche Diagnostics, Germany). The PCR reaction comprised 1X LightCycler FastStart DNA Master Hybridization Probe mix (Roche Molecular Biochemicals), 2 mM MgCl2, 0.5 μM of primers, 0.2 μM of hybridization probes and 2 μL of template DNA. The thermocycling conditions included an initial denaturation step at 95 °C for 10 min, followed by 45 cycles of amplification [denaturation (95 °C, 15 s), annealing (55 °C, 15 s) and extension (72 °C, 15 s)] with a temperature transition rate of 20 °C s−1. Melting curve analysis profile was performed post-PCR as follows: 95 °C for 60 s, 40 °C for 60 s and then 40–80 °C at a transition rate of 0.1 °C s−1. One final cooling step (40 °C) was also included.

Standard curve construction

Taking into account the potential variability of rRNA operon target copies among strains, equal numbers of cells from ten A. minutum (GC) cultured strains that originated from Irish coastal waters were pooled together and fixed with Lugol's iodine as standards for generating the calibration curve. DNA was extracted from triplicate samples of filtered cells spanning five orders of magnitude (from 1 to 105 cells) and included in the RT-PCR assay. A calibration curve was then constructed from the average Ct values, where Ct is the threshold cycle parameter, defined as the cycle number at which the fluorescence crosses an automatically determined value above the baseline during the amplification reaction. The quantification of A. minutum cells in field samples was accomplished using the lightcycler software version 3.5 through the interpretation of the linear regression log(A. minutum cell number)=(Ct−intercept)/slope. For quantification, an A. minutum (GC) extract standard (100 cells) was included in triplicate in each batch of samples analysed to enable auto-adjustment by the lightcycler software to the pregenerated calibration curve. Ct values greater than 35 were not considered for quantification.

Statistical analysis

Statistical analyses were performed using spss version 15.0 for Windows. One-way anova, followed by Tukey's post hoc test (Tukey, 1949), was performed (1) to identify significant differences between the average A. minutum (GC) concentrations found in the North Channel area of Cork Harbour using morphotaxonomy, WC-FISH and qRT-PCR methods and (2) to ascertain whether the A. minutum concentrations recorded by morphotaxonomy over the sampling period differed significantly. The data were tested for normality and homogeneity of variance before performing the post hoc test. Natural logarithm was used to transform the data to meet these assumptions. Independent-sample Student's t-tests (Student, 1908) were used to compare the slope of the linear regression curve derived from the morphotaxonomy count data with that of WC-FISH and qRT-PCR.

Results

DNA sequence analysis and assay design

Analysis of a sequence alignment of the ITS1-5.8S rRNA gene region from several species of Alexandrium showed genetic divergence between the A. tamarense species complex and the clade supporting A. minutum, A. tamutum, A. insuetum, A. ostenfeldii and A. peruvianum (data not shown). PCR primers ITSmin-f1 and ITSmin-r3 contained one to two mismatches with closely related members within the A. minutum supporting clade and up to six mismatches with the A. tamarense species complex and some species of the Gessnerium subgenus. The sensor hybridization probe ITSminB-Fl was 100% homologous to A. minutum (GC), showing five mismatches with A. insuetum and more than seven with other members of the clade, including the Pacific Clade type A. minutum. The anchor probe ITSminB-LC designed from the 5′ end of the 5.8S rRNA gene was 100% homologous to A. minutum, A. tamutum, A. insuetum, A. ostenfeldii, A. peruvianum and A. andersoni 5.8S sequences but contained up to three mismatches with A. margalefi and A. taylori, and at least five or more mismatches with members of the A. tamarense species complex. The blast-n program (National Center for Biotechnology Information; http://www.ncbi.nlm.nih.gov) was used to confirm that the PCR primers and FRET hybridization probes did not recognize any other dinoflagellate DNA sequence in GenBank.

Specificity assessment of the RT-PCR assay

The specificity of the A. minutum (GC) assay was evaluated using DNA extracted from a panel of Alexandrium spp. and other dinoflagellate strains (Table 1). The hybridization probe pair appeared specific for A. minutum (GC) and generated a species-specific melt peak of c. 57 °C only for the A. minutum (GC) strains (n=14). Figure 3 shows an example of the quantification and melt curves obtained with three A. minutum (GC) strains tested. No melt peaks were obtained for the eight other Alexandrium spp. and seven dinoflagellate strains tested. After analysis of the RT-PCR products by agarose gel-electrophoresis, single-band 220-bp PCR products were visible for A. minutum strains and other closely related Alexandrium species including A. tamutum, A. insuetum, A. ostenfeldii and A. peruvianum. Multiple PCR products were visible for other Alexandrium and dinoflagellate strains (data not shown).

Figure 3.

 Real-time PCR (a) and melt curve (b) outputs showing the specificity of the Alexandrium minutum (GC) assay tested against a panel of Alexandrium and dinoflagellate species. Only the DNA extracts of A. minutum (GC) are detected (strains CK.A14, SHA.C4 and Kill.E4 were selected for this specific run). The nontarget taxa are the same as those listed in Table 1.

Sensitivity assessment of the RT-PCR assay

DNA was extracted from a cultured A. minutum strain (CK.A17), serially diluted over eight orders of magnitude (10 ng to 1 fg DNA) and included in the A. minutum (GC) RT-PCR (data not shown). Three independent experimental assessments of the sensitivity of the A. minutum RT-PCR assay were performed. The detection limit of the assay was reproducibly 0.1 pg of A. minutum DNA, which is less than one cell equivalent.

Standard curve construction

Ten Irish A. minutum (GC) strains were mixed together (7 × 104 cells each) to generate a concentrated stock. From this mix, A. minutum cells spanning five orders of magnitude (1–105 cells) were filtered in triplicate and genomic DNA was extracted for inclusion in the RT-PCR assay. A negative linear relationship (r2=0.9818) between Ct values and the log of cell numbers extracted was obtained (Fig. 4). The efficiency of the RT-PCR reaction E was derived using the relation E=10(−1/μ), where μ is the slope of the standard curve, and had an average value of 1.94 (range 1.89–1.99) from three runs.

Figure 4.

 Calibration curve of the cycle threshold (Ct) generated for known concentrations of mixed Alexandrium minutum strains. Error bars denote the SD from triplicate PCR amplifications.

Field sample analysis and method comparison

Cell concentration estimates of A. minutum (GC) were determined in 42 seawater samples from 10 surveys (1–10) in Cork Harbour by morphotaxonomy, WC-FISH and qRT-PCR analyses. Because of logistical constraints onboard the vessel during sampling, the three independent seawater samples taken at each station and used for comparison were processed in different ways and cannot be considered as real subsamples. The theoretical limits of quantification of A. minutum calculated for each method according to the amount of seawater collected and processed were c. 8, 20, and 40 cells L−1 for WC-FISH, morphotaxonomy and qRT-PCR, respectively. The morphotaxonomy-based method using calcofluor was considered as the reference method on the basis that it required the least sample processing and that individual target cells were identified and enumerated. Spatial and temporal variability was observed in the results among stations in the North Channel of Cork Harbour (Fig. 5). Average A. minutum concentrations were the highest for all methods during survey 5. The anova of the A. minutum concentrations in the North Channel determined by morphotaxonomy, WC-FISH and qRT-PCR, and performed individually on the natural logarithm values of samples from surveys 1–9 (sample series 5300s–6100s), showed that only the results obtained for survey 5 (5700s) for qRT-PCR and WC-FISH differed significantly (anova, F=5.5, P=0.024). Considering the morphotaxonomy-derived results, the anova for the data set including samples from surveys 3 to 9 (5500s–6100s) showed that the average A. minutum concentrations for samples from surveys 5, 7 and 8 (5700s, 5900s and 6000s) did not differ significantly from each other, but were significantly higher than those from samples from surveys 3, 4, 6 and 9 (5500s, 5600s, 5800s and 6100s) (anova, F=8.8, P<0.001).

Figure 5.

 Average Alexandrium minutum concentrations in the North Channel area of Cork Harbour derived for each survey carried out during summer 2007 by morphotaxonomy (▪), qRT-PCR (inline image) and WC-FISH analyses (□).

The morphology-based analysis of samples using calcofluor and an Utermöhl counting chamber showed the co-occurrence of A. minutum with A. tamarense, and to a lesser extent with A. ostenfeldii. Other morphologically similar armoured dinoflagellate species such as Glenodinium foliaceum, Protoperidinium minimum or Scrippsiella spp. were also observed at high numbers in some samples. The concentrations of A. minutum in the water column over the sampling period ranged from 0 to c. 5 × 104 cells L−1 as determined with the calcofluor-based method (for some samples from survey 2 and 6008-Q, respectively). Four samples (5405, 5406, 5407 and 5410) collected on 5 June (survey 2) had no A. minutum cells. The range of A. minutum (GC) concentrations determined in the WC-FISH assays using the probe Min.A was between 0 and c. 3.4 × 104 cells L−1, with the maximum value obtained at station 6008-Q (survey 8), consistent with the result derived with the calcofluor analysis. No A. minutum cells were detected in samples 5406 and 5410 only. However, cell concentrations in samples 5405 and 5407 were, respectively, 15 and 8 cells L−1, values below the detection limit of the morphotaxonomy method. Alexandrium minutum cells were clearly labelled and identifiable in bright orange, nontarget species and detritus showing only weak auto-fluorescence. The simultaneous use of DAPI and calcofluor along with Min.A also proved particularly useful to confirm the absence of matrix effect or cross-reactivity with other phytoplankton species. Based on melting curve analysis, the RT-PCR assay allowed the detection of A. minutum (GC) DNA in 38 of the 42 samples that showed a single melt peak at c. 57 °C. The standard curve generated for the assay allowed the quantification of A. minutum in 34 samples, with concentrations ranging from 94 to c. 6 × 104 cells L−1 (at stations 5405-R and 6008-Q, respectively). No A. minutum was detected in samples 5303-5305-5406-5410 and the presence of the species was observed through melt peak analysis in samples 5304-5403-5404-5407, but at levels below the limit of quantification (Ct>35).

Linear regression analyses were carried out to examine the distributions of the A. minutum concentrations derived using the three quantification methods. Positive relationships were observed between the methods (Fig. 6). The best coefficient of determination (r2=0.81) was obtained for the morphotaxonomy and qRT-PCR pairings. However, the corresponding linear regression curve showed a slight overestimate of A. minutum concentrations in the field compared with the 1 : 1 correlation line. The regression curve derived from the morphotaxonomy and WC-FISH results indicated an underestimation of A. minutum by the latter method by a factor of c. 1.2. The plotting of the WC-FISH and qRT-PCR derived results showed the highest dispersion of data around the regression curve (r2=0.57), whose slope was, however, not significantly different (Student-t test, P<0.05) from that obtained for the morphotaxonomy and qRT-PCR pairings.

Figure 6.

 Comparison of the Alexandrium minutum cell concentrations derived in the seawater samples collected in Cork Harbour in summer 2007 by morphotaxonomy, qRT-PCR and WC-FISH analyses. Solid lines represent the regression curve fits forced through the origin and the dotted lines correspond to the 1 : 1 correlation.

Discussion

A number of rRNA- and rDNA-targeting WC-FISH and qRT-PCR assays are available for the discrimination and quantification of Alexandrium spp. in natural populations (John et al., 2003; Galluzzi et al., 2004; Gribble et al., 2005; Touzet et al., 2008b). A taxon-specific RT-PCR assay based on the ITS1-5.8S region of the rRNA gene with a detection limit equivalent to one cell or less was developed in this study to detect and quantify A. minutum (GC) in environmental samples. The FRET-based hybridization probe chemistry allowed the identification of the target through melt curve analysis with one taxon-specific melt peak only (c. 57 °C) for 11 Irish A. minutum (GC) strains and three from elsewhere, with no detection of 23 other Alexandrium spp. and dinoflagellate strains. Additionally, A. minutum (GC) was successfully identified in 38 seawater samples that contained the target taxon as established by morphotaxonomy and/or WC-FISH analyses.

The performance of the qRT-PCR assay was compared with results obtained after analysis by morphotaxonomy and WC-FISH of seawater samples collected in Cork Harbour. Positive relationships were obtained between the three methods, and the discrepancies observed among the data sets highlighted that distinct entities were measured, whereby morphotaxonomy allowed the counting of cells, WC-FISH allowed the enumeration of cells with ribosomal activity and qPCR allowed the quantification of double-stranded nucleic acids. The apparent underestimation of A. minutum by WC-FISH probably reflects cell losses that may have occurred during sample processing. High heterogeneity in A. minutum concentrations derived by qRT-PCR was observed for samples from survey 8 (CH6000s) compared with analyses by WC-FISH and morphotaxonomy. This may be due to the high degree of patchiness in the distribution of the A. minutum population across the North Channel (1 × 104 cells L−1 at the eastern end of the channel and up to c. 5 × 104 cells L−1 elsewhere) and the fact that the cell concentration range obtained by qRT-PCR was wider than those obtained with the other two methods. Additionally, A. minutum strains within developing populations may have varying cellular copy numbers of the rRNA gene operon, and nucleic acids contained in phycopellets or the variation of frequency of diploid sexual stages in populations may also be contributing factors (Moorthi et al., 2006). The quality of the nucleic acids extracted from field samples may be affected by the sample matrix, which can impact the efficiency of qPCR. Despite providing purified DNA of suitable quality for analysis, commercial extraction kits have been criticized for the variability in the amount of nucleic acids recovered from samples (Galluzzi et al., 2004). The use of sample crude lysates is possible. However, the stability of nucleic acids is not guaranteed for long-term storage of extracts and the requirement for dilutions before PCR in order to reduce interferences from coextracted inhibitors such as humic acids may impair the sensitivity of assays and result in the underestimation of target taxa (Miller, 2001).

Only two samples that contained A. minutum as determined by morphotaxonomy or WC-FISH were negative by qRT-PCR. Based on the detection limit of the assay, it should have been possible to detect A. minutum (GC) in these samples. However, cell concentrations were close to the limit of quantification (40 cells L−1) and nondetection may be due to the combination of low cell concentration, competitive PCR amplification of nontarget gene fragments and heterogeneity among the three samples collected at each station for independent analysis with the different methods. The underestimation of A. minutum cells in the Cork Harbour samples obtained using WC-FISH, as well as the efficiency (E) of the qRT-PCR assay, were both comparable to those recorded in previous studies carried out in the Gulf of Maine on natural A. fundyense populations (Anderson et al., 2005; Dyhrman et al., 2006).

Morphotaxonomy-based analysis relies on the visual identification of structurally intact ‘real’ phytoplankton cells and allows the quantitative assessments of many taxa in environmental samples. However, the occurrence of morphologically similar species, as observed in many of the Cork Harbour samples, can make analyses time consuming. WC-FISH combines the advantages of both phenotypic and genetic analyses by allowing monitoring of multiple species and the reliable discrimination of toxic and nontoxic taxa that may be morphologically identical but genetically distinct (e.g. A. tamarense species complex) (John et al., 2005). The examination of specimens still requires c. 20 min per membrane, and the number of sample processing steps involved can also lead to some losses of target cells (Anderson et al., 2005). While qRT-PCR does not provide information on the global taxonomic diversity in samples, high-throughput analysis of samples can be achieved, making the method ideal for rapid screening and quantification of a small number of taxa. However, assays do not provide information on the global taxonomic diversity in samples. Quantification of HAB species by qRT-PCR is relatively new. Despite the requirement of high capital equipment and consumable budgets compared with WC-FISH, RT-PCR shows promise as a monitoring tool because it can be easily standardized and is suitable for high sample throughput.

Monitoring of and studies on Alexandrium population dynamics at the species level cannot be carried out reliably by light microscopy. Simultaneous multiple species detection and enumeration through multiplexing in RT-PCR has recently been demonstrated for HABs (Handy et al., 2006). The development of such approaches using hybridization probes may facilitate the monitoring of closely related co-occurring Alexandrium species (Ciervo & Ciceroni, 2004; Keaveney et al., 2006; O'Grady et al., 2008). The use of qRT-PCR may also have applications for Alexandrium lifecycle studies for facilitating the tedious analysis of cyst-containing sediment samples (Anderson et al., 2003). Kamikawa et al. (2007) recently demonstrated the potential of qRT-PCR for the quantification of A. catenella and A. tamarense resting cysts in marine sediments. In the context of Cork Harbour, this may assist in high-resolution cyst mapping and parameterization of models of cyst bed development in the North Channel area, where blooms of Alexandrium spp. occur on a regular basis (Ní Rathaille et al., 2009).

In conclusion, an ITS-5.8S rRNA gene-targeted qRT-PCR assay for A. minutum (GC) was developed for the rapid detection and quantification of the species in water column samples. Comparable A. minutum concentration estimates were obtained after analysis by morphotaxonomy, WC-FISH and qRT-PCR of samples collected during surveys carried out in Cork Harbour. The assay shows great potential for monitoring and population dynamics studies.

Acknowledgements

The authors acknowledge D. Hugh-Jones, B. Byrne and D. Geary for the use of their onsite facilities and the provision of boats. Thanks are due to A. Ní Rathaille for assistance in field work. The A. minutum AMBOP006, A. minutum W07/001/01, A. minutum AL2V, A. minutum AL8C and A. insuetum ICMB 218 strains were generously provided by D. Anderson, C. Collins, S. Fraga and E. Garces, respectively. This work was supported by the EC 6th Framework Programme through the Interreg NWE IIII FINAL (G142) and SPIES-DETOX (Coll CT 2006, Contract no. 030270) and SEED (GOCE-CT-2005-003375) projects.

Ancillary