OUP user menu

Comparison of paralytic shellfish toxin (PST) production by the dinoflagellates Alexandrium lusitanicum NEPCC 253 and Alexandrium tamarense NEPCC 407 in the presence and absence of bacteria

Georgina L. Hold, Elizabeth A. Smith, T. Harry Birkbeck, Susan Gallacher
DOI: http://dx.doi.org/10.1111/j.1574-6941.2001.tb00843.x 223-234 First published online: 1 July 2001


The ability of two Alexandrium species to produce paralytic shellfish toxins (PST) in laboratory culture following the generation of bacteria-free cultures was investigated. The dinoflagellates Alexandrium lusitanicum NEPCC 253 and Alexandrium tamarense NEPCC 407 were cultured in the presence of antibiotics and tested for residual bacteria. After treatment with a cocktail of streptomycin, ciprofloxacin, gentamicin and penicillin G, bacteria could not be detected in either of the treated Alexandrium cultures using 17 different solid and broth bacterial growth media, by epifluorescence microscopy with the dye Sybr green 1, or polymerase chain reaction amplification using universal eubacterial primers designed to target the 16S rRNA gene. Subsequent analysis of A. lusitanicum for PST using high performance liquid chromatography demonstrated that the growth rate and toxin profile remained similar in both bacteria-free and control cultures, although the quantity of toxins produced differed with the bacteria-free culture producing generally more of each compound and also having a greater toxin content in terms of saxitoxin equivalents. A. tamarense also retained similarities between the bacteria-free and control cultures in terms of growth rates and toxin profile, although in this instance, depending on the growth stage and the toxin, the control culture produced more of some toxins than the bacteria-free culture. The control culture was also more toxic in terms of saxitoxin equivalents than the axenic culture. These results suggest that bacteria can influence toxin production in laboratory cultures of Alexandrium species although the mechanisms remain unknown.

  • Dinoflagellate
  • Axenic culture
  • Paralytic shellfish toxin (PST)
  • Epifluorescence microscopy
  • High performance liquid chromatography (HPLC)
  • Polymerase chain reaction (PCR)

1 Introduction

The production of paralytic shellfish toxins (PST) has been attributed to dinoflagellates, with members of the Alexandrium, Gymnodinium and Pyrodinium genera reported to produce PST [1]. Cyanobacteria are also recognised as producers of PST [24], and there is some evidence for production of these toxins by heterotrophic bacteria, which is reviewed in [5].

PST consist of saxitoxin (STX) and at least 20 chemically related derivatives which vary in toxicity, hence the total toxicity of dinoflagellates may alter depending on the combination and concentration of toxins present [5]. It has been suggested that bacteria associated with dinoflagellates either in the natural environment or in laboratory cultures, may influence the production of PST by dinoflagellates [5,6]. Several authors have shown that bacteria-free dinoflagellate cultures produce PST [79], although conflicting evidence has been presented on the effect of bacteria on dinoflagellate toxicity. Some authors report higher toxicity in bacteria-free cultures [10,11] while others show the reverse [6]. However, there is no evidence for a species-specific difference.

The above studies are dependent on the production and maintenance of axenic cultures. Several methods have been used to produce these cultures including washing with sterile diluents, using dilution series, ultrasonic treatment and differential centrifugation [1015]. In recent years these methods have been used either in conjunction with, or have been superseded by treatment with antibiotics [16,17] or other anti-bacterial agents, e.g. lysozyme [6]. The advantage of bactericidal treatments over physical dissociation methods is the ease of application, coupled with their effectiveness in removing bacteria from mucilaginous algal species which cannot easily be physically separated from their accompanying microflora [16].

In many of the older published works assessment of the bacterial status of dinoflagellate cultures was limited to determining bacterial growth on one or two different bacteriological media and in some cases, observations with light microscopy [18,19]. Often, the media formulations used were unsuitable for detection of culturable marine bacterial species [2022] and did not take into account that 95% of marine strains are currently considered unculturable [23]. Therefore, more stringent methods for testing axenic cultures have been implemented, generally involving epifluorescence microscopy [6,8,14,15] using dyes such as 4′,6-diamidino-2-phenylindole (DAPI) and acridine orange [2426]. Recently, more effective fluorescent stains such as Sybr green 1 have become available [27].

The aim of this investigation was twofold, firstly to produce bacteria-free dinoflagellate cultures and rigorously assess their bacterial status using a combination of culture methods, epifluorescence microscopy and molecular techniques. Secondly, to determine if PST production in Alexandrium cultures was influenced by the presence of bacteria as determined by changes in toxin profile measured by high performance liquid chromatography (HPLC). The cultures examined were Alexandrium lusitanicum NEPCC 253 and Alexandrium tamarense NEPCC 407. A. lusitanicum was originally isolated from Portuguese waters in 1962 and reported to remain toxic with the same toxin composition during 30 years in culture [28]. A. tamarense was also reported to remain toxic in culture after its isolation from Canadian waters in 1981 [29].

2 Materials and methods

2.1 Dinoflagellate strains and maintenance

Dinoflagellate strains A. lusitanicum NEPCC 253, also known as AL 18 [30], and A. tamarense NEPCC 407 were supplied by the North East Pacific Culture Collection, Canada. Dinoflagellate cultures were aseptically subcultured into seawater enriched with f/2 media without silica (Sigma, Dorset, UK) as described in [17]. Prior to use, aged (3 months) seawater was autoclaved at 110°C for 30 min in borosilicate wide-neck flasks with non-absorbent cotton-wool bungs and cooled to room temperature before the addition of 20 ml l−1 f/2 nutrients. All cultures were maintained at 15°C with a 14:10 h light:dark cycle (irradiance level 0.5–1.5×1016 quanta s−1 cm−2). Nine identical flasks containing 1.25 l of f/2-enriched seawater were prepared for each dinoflagellate culture with approximately 1000 dinoflagellate cells from mid-exponential growth into each flask introduced using aseptic transfer. This allowed an adequate sample volume to continuously monitor growth and also to sample for HPLC and PCR analyses at each growth point.

2.2 Dinoflagellate growth curves

5 ml volumes of dinoflagellate culture were aseptically removed every second day following gentle swirling of flasks to generate a uniform suspension. Lugol's iodine (10 μl) prepared as described by Cowan and Steel [31] was added to each sample to fix the cells. Cells were counted using a Sedgewick-rafter slide following the procedure described by McAlice [32]. Twelve random fields of vision per ml of fixed sample were counted, with this procedure carried out in triplicate for each growth curve sample to give statistically valid results. Dense cultures were serially diluted in sterile seawater to give between 10 and 20 cells per μl before counting. Growth rates were calculated as described by Pirt [33].

2.3 Isolation of bacteria from dinoflagellate cultures

Samples (1 ml) of dinoflagellate cultures taken from early and late exponential phase, and stationary phase corresponding to days 4, 14 and 28 in A. lusitanicum and days 6, 21 and 28 in A. tamarense, and serially diluted (10-fold dilutions from neat – 10−9) in sterile seawater. Samples (100 μl) of each dilution were spread in triplicate onto marine agar plates (Difco 2216, MI, USA), and subsequently incubated for 14 days at 20°C. Bacteria from the dilution generating between 50 and 100 colonies were isolated from one of the triplicate plates, and replated individually onto marine agar to obtain pure cultures. Resultant pure bacterial isolates were categorised by colony morphology before being stored in marine broth plus 10% (v/v) glycerol at −70°C, prior to further investigation.

2.4 Antibiotic resistance profiling of bacteria isolated from dinoflagellate cultures

Bacteria isolated from dinoflagellate cultures were inoculated into 10 ml marine broth (Difco, MI, USA), and incubated for 24 h in a shaking incubator (20°C, 120 osc. min−1). Each bacterial suspension (200 μl), was spread evenly onto marine agar plates and allowed to air dry to form a bacterial lawn. Sterile individual antibiotic sensitivity discs (5 mm) containing: – kanamycin 30 and 5 μg; streptomycin 25 and 10 μg; penicillin G 10 unit; ciprofloxacin 5 μg; gentamicin 120 μg; novobiocin 30 μg (Oxoid, Basingstoke, UK), were pressed onto the agar surface. Plates were subsequently inverted and incubated at 20°C for 48 h. Antibiotic sensitivity was identified by production of a zone of inhibition of 5 mm or greater diameter around the disc.

2.5 Antibiotic treatment of dinoflagellate cultures

Antibiotic cocktails containing either ciprofloxacin (23 μg ml−1), gentamicin (120 μg ml−1) and streptomycin (12.5 μg ml−1) or ciprofloxacin (46 μg ml−1), gentamicin (240 μg ml−1) and streptomycin (25 μg ml−1) or ciprofloxacin (46 μg ml−1), gentamicin (240 μg ml−1), streptomycin (25 μg ml−1) and penicillin G (20 units ml−1), were added to mid-exponential phase dinoflagellate cultures and incubated for 12 days. Dinoflagellates were subcultured through three growth cycles to dilute out the antibiotics prior to assessment of their bacteriological status, growth rates and toxin profiles. Untreated cultures i.e. without addition of antibiotics, were maintained concurrently. All cultures were incubated under conditions as previously described.

2.6 Use of different marine culture media to assess the presence of bacteria in antibiotic-treated dinoflagellate cultures

To detect the presence of remaining culturable bacteria in antibiotic-treated cultures, 17 different media, including both general and selective formulations, were used (see Table 1). Samples of antibiotic-treated cultures were inoculated into broths and onto agar plates of the 17 media formulations. Five tubes containing 9 ml of broth from each of the 17 media formulations were inoculated with 1 ml of the antibiotic-treated culture, whilst three agar plates of each media formulation were inoculated with 100 μl of culture. These were incubated at 20°C for 1 month, with control, untreated cultures included for comparison.

View this table:

Bacterial growth media formulations

Medium nameReference
Peptone A[46]
Peptone Bc[46]
Malt extract[48]
Yeast extractc[20]
Peptone glucose[50]
Peptone yeastc[50]
  • a a Medium was also used at a 1/100 dilution.

  • b b Medium was also used at a 1/10 and 1/10 000 dilution.

  • c c Media were also prepared without the addition of iron.

2.7 Epifluorescence microscopy

Sybr green 1 nucleic acid stain (Molecular Probes Inc., The Netherlands), was prepared and stored following the method described by Marie et al. [27]. Samples (10 ml) from dinoflagellate cultures were taken aseptically and diluted in sterile seawater to give approximately 105 cells ml−1. Sybr green 1 stain at a final dilution of 10−4 of the commercial solution was added to samples, which were subsequently incubated in the dark at 20°C for 5 min before placing in a filter column and slowly drawing (<150 mm Hg vacuum) the cell suspension onto a black polycarbonate membrane (0.2 μm, Porvair Filtronics, Middlesex, UK). The membrane was fixed onto a microscope slide by addition of a drop of immersion oil above and below the membrane, and a coverslip added. Slides were immediately examined by epifluorescence microscopy (Zeiss, Axiovert 10) using oil immersion at an excitation wavelength of 460 nm.

2.8 Bacterial 16S rRNA PCR amplification of dinoflagellate cultures.

Antibiotic-treated dinoflagellate cultures of A. lusitanicum NEPCC 253 and A. tamarense NEPCC 407 from early, late exponential and stationary growth phases (1000 ml), were collected by centrifugation (10 000×g, 10 min) and DNA extracted following the method of Scholin et al. [34]. Each sample was subjected to two sets of PCR reactions: set 1 eubacterial primers described by Muyzer et al. [35] and set 2 by Suzuki et al. [36]. These universal primer sets amplified different length fragments of the 16S rRNA gene, with primer set 1 targeting the hypervariable V3 region corresponding to nucleotide positions 341–534 in Escherichia coli, and generating an approximately 200-bp PCR product. Primer set 2 eubacterial primers 27F and 1522R amplified almost the whole 16S rRNA gene (approximately 1500 bp). Amplifications were performed using a Techne Genius thermocycler, with each 100-μl reaction volume containing 100 ng genomic DNA, 2 mM MgCl2 and 0.3 μM of each primer. Thermocycling conditions for primer set 1 following initial denaturation of 94°C for 5 min were, 30 cycles of 94°C for 30 s, 55°C for 45 s and 72°C for 30 s, followed by a final extension for 10 min at 72°C. Thermocycling conditions for primer set 2 were as described by Suzuki et al. [36]. Prior to visualisation of PCR products, the 100-μl PCR reaction volumes were concentrated to 10 μl using Prep-a-Gene (Bio-Rad, Hemel Hempstead, UK). Concentrated PCR products were subjected to electrophoresis in a 2% (w/v) agarose gel in 1×TAE containing ethidium bromide (0.5 μg ml−1), with markers (100 bp and 1 kb; Gibco, Paisley, UK) included on the gel for reference.

The lower detection limit of the PCR amplification using the two primer sets was estimated by assessing their ability to generate amplification products with differing concentrations of various bacterial morphotypes isolated from A. lusitanicum. Dilutions of representatives of the previously isolated morphotypes were prepared to provide 10–107 colony forming units (cfu) per 100-μl PCR reaction. These amplification products were visualised following concentration as previously described.

2.9 PST extraction from dinoflagellates

Dinoflagellates from three points in the growth cycle were sampled for toxin analysis. Following determination of cell concentrations, cells from 1000 ml were collected by centrifugation (2000×g, 10 min), the supernate decanted and the resulting cell pellet resuspended in 1 ml acetic acid (0.05 M), before storage at −20°C overnight. Following thawing of samples, cells were disrupted by the addition of 25% w/v glass beads (150–212 μm diameter, Sigma, Dorset, UK), and vortex-mixed for 3 min. Microscopic examination of cell debris revealed that the cells had been completely disrupted. All samples were briefly centrifuged (13 000×g, 30 s), with the supernatant filtered through 0.45-μm syringe filters (Nalgene, Rochester, USA) prior to storage at −20°C for subsequent toxin analysis.

2.10 HPLC analysis of dinoflagellate samples

HPLC analysis of dinoflagellate samples followed the method described by Franco et al. [37], with the following amendments: a silica-based reverse-phase column was used for toxin separation (C18; 250×4 mm internal diameter, Purospher, Merck, Glasgow, UK), with a mobile phase flow rate of 1.0 ml min−1, held at constant temperature (35°C). External calibration standards gonyautoxins 1/4 and 2/3 (GTX 1, 2, 3 and 4), STX and neo-STX (NRC, Canada) were included before sample analysis and after every fourth injection to monitor the performance of the system. Toxin composition profiles were determined from triplicate analyses and expressed as toxicity per cell. Confirmation of retention times of sample peaks with those of known toxins was determined by inclusion of internal toxin standards within samples. Total toxicity values were determined as described by Parkhill and Cembella [37], with conversion figures used as detailed in Franco et al. [38].

HPLC was performed using an autosampler and Spectra System P4000 pumps (Thermo Separation Products, Hemel Hempstead, UK), and a RF-535 fluorescence detector (Shimadzu, Milton Keynes, UK) with computer integration (PC1000, V 3.5; Thermo Separation Products).

The signal to noise ratio allowed toxin peaks which were three times the size of the background noise level to be determined, with minimum limits for each toxin on the column determined as: GTX 1, 0.28 ng; GTX 2, 0.33 ng; GTX 3, 0.08 ng; GTX 4, 0.12 ng; STX, 0.57 ng and neo-STX, 3.77 ng.

3 Results

3.1 Assessment of the bacteriological status of antibiotic-treated Alexandrium cultures using a range of marine media

Bacteria isolated on marine agar from the A. lusitanicum culture were categorised into four distinct morphotypes, as were bacteria isolated from A. tamarense, with some similarities noted between the morphotypes from the two dinoflagellate cultures. The antibiotic sensitivity of the bacterial isolates was assessed using a range of antibiotics (Table 2), with all strains sensitive to a combination of ciprofloxacin, gentamicin and streptomycin (Table 2). This combination of antibiotics (combination 1 and 2; Table 2) was subsequently added to the two dinoflagellate cultures and incubated for 12 days, after which time the dinoflagellates were subcultured into fresh antibiotic-free media, and their bacterial status assessed by checking for bacterial growth on marine agar. Although antibiotic combination 2 was more effective at reducing bacterial numbers than combination 1, one morphotype still remained in each dinoflagellate culture (Table 2). Further antibiotic profiling of the resistant bacteria demonstrated that they were sensitive to penicillin G. This antibiotic was subsequently added to the antibiotic cocktail (combination 3) and proved bactericidal to the microflora of the two dinoflagellate cultures, as determined by lack of bacterial growth after 14 days at 20°C initially on marine agar and subsequently using 16 other different agar formulations detailed in Table 1. However, a fungus identified as a Botrytis sp. (Dr S. Moss, personal communication) was isolated from the antibiotic-treated A. tamarense culture from most of the agar preparations shown in Table 1, with the exception of the low nutrient formulations, namely ST10−4, f/2, seawater media and 1/100 strength marine media. The fungus was not isolated from the control A. tamarense culture. Efforts to remove the fungi using a number of fungicidal reagents were unsuccessful.

View this table:

Sensitivity of the different bacterial morphotypes to antibiotics and percentage removal of bacterial isolates following antibiotic treatment

Antibiotic sensitivityBacteria removed (%) with antibiotic combination
Colony description –A. lusitanicum NEPCC 253
Large cream+++2085100
Small rose++++45100100
Flat beige+75100100
Colony description –A. tamarense NEPCC 407
Small rose++++55100100
Large cream+++30100100
Small cream+++1580100
Rough beige+++35100100
  • a a Ciprofloxacin (23 μg ml−1), gentamicin (120 μg ml−1) and streptomycin (12.5 μg ml−1).

  • b b Ciprofloxacin (46 μg ml−1), gentamicin (240 μg ml−1) and streptomycin (25 μg ml−1).

  • c c Ciprofloxacin (46 μg ml−1), gentamicin (240 μg ml−1), streptomycin (25 μg ml−1) and penicillin G (20 units ml−1).

  • d d+=antibiotic sensitive.

  • e e−=antibiotic resistant.

As the detection limit for bacteria by the plating method was 10 cfu ml−1, a more sensitive test for the presence of bacteria, with a detection limit of 2 cfu ml−1[39] was carried out by inoculating 1 ml of each dinoflagellate culture into five replicate tubes containing 9 ml broth of each media listed in Table 1. After 30 days incubation, no turbidity was observed, although Botrytis was again detected in most of the broth cultures inoculated with the antibiotic-treated A. tamarense culture, with the exception of the low nutrient formulations mentioned above.

3.2 Epifluorescence microscopy of dinoflagellate cultures

The bacterial status of the dinoflagellate control and antibiotic-treated cultures was examined using epifluorescence microscopy with Sybr green 1 stain. Fig. 1a shows an A. lusitanicum NEPCC 253 cell prior to antibiotic treatment. The nucleus of the cell was visible as a large green fluorescing mass with the rest of the dinoflagellate cell faintly autofluorescing red due to the presence of chlorophyll. The remainder of the green fluorescence, in the form of small particles, indicated the presence of bacterial cells. Following antibiotic treatment (Fig. 1b), bacteria were not visible either attached to the dinoflagellate cell or free-living in the culture media.


Epifluorescence microscopy pictures of A. lusitanicum NEPCC 253 before and after antibiotic treatment Sybr green 1 (40 000× magnification). a: control culture. b: antibiotic-treated culture.

3.3 Use of PCR to confirm the axenic status of dinoflagellate cultures

Prior to determining the bacterial status of the dinoflagellate cultures with PCR, the spectrum of bacteria targeted by each of the two primer sets was investigated. Searches of databases, namely GenBank and the ribosomal database project (RDP), indicated that some β-Proteobacteria and the Planctomycetes were not amplified by primer set 1 and certain Vibrio species were not amplified by primer set 2. However, use of both primer sets covered the range of Eubacteria deposited in the databases.

The sensitivity of the method was also assessed. DNA was extracted from representatives of the four different bacterial morphotypes obtained from A. lusitanicum NEPCC 253 and PCR-amplified using the two primer sets. This indicated that the method could successfully amplify all bacterial morphotypes when 10 cfu were present in the PCR reaction. Although 1 cfu per reaction was enough to generate an amplification product on occasions.

PCR products were not detected in antibiotic-treated dinoflagellate cultures with either primer set but were detected, as expected, in control cultures. The bacteria-free cultures were subsequently passaged three times over a 90-day period in antibiotic-free media and their bacterial status again checked as described above. These cultures remained bacteria-free.

3.4 Dinoflagellate growth rates

The growth of the dinoflagellate cultures over a 30-day period was monitored using dinoflagellate cell counts, with Fig. 2A,B showing mean log10 values over the 30-day period. Differences were observed in the cell counts between the control and antibiotic-treated cultures of A. lusitanicum during the first eight days although growth rates of the cultures were identical (μ=0.1 day−1). As shown in Fig. 2A, the control culture initially contained a lower cell concentration than the axenic culture (Fig. 2A). However, a steep increase in cell concentration was observed in the control culture over the first 8 days compared to the axenic culture, but from day 8 onwards concentrations were very similar between the two cultures. With A. tamarense, the growth rates of the two cultures were also identical (μ=0.14 day−1), although differences were observed in cell concentrations between days 7 and 15 (Fig. 2B).


Growth curve (n=3±S.E.M.) over 30 days for (A) A. lusitanicum NEPCC 253 and (B) A. tamarense NEPCC 407 in the presence/absence of bacteria. Arrows indicate cell sampling points for HPLC analysis. △=control culture; ◼=axenic culture.

3.5 Production of PST by control and axenic cultures of A. lusitanicum and A. tamarense

Control and axenic cultures of A. lusitanicum and A. tamarense were sampled at three different points in the dinoflagellate growth cycle, namely early and late exponential phase, and stationary phase (as indicated by the sampling points on Fig. 2) and the bacteria-free status of the axenic cultures confirmed as described above. These samples were subsequently analysed by HPLC to assess production of PST. This was repeated twice more with results similar to those presented below. Fig. 3 is an example of the chromatograms obtained using toxin standards GTX 1–4 (Fig. 3a) and sample 1 of the control (Fig. 3b) and axenic (Fig. 3c) cultures of A. lusitanicum, before and after spiking with toxin standards. The major toxin peaks in the control dinoflagellate culture co-eluted with standards of GTX 1 and 4 with traces of GTX 2 and 3 also observed (Fig. 3b). A further peak at ∼3.5 min was also apparent on sample chromatograms, but could not be attributed to a PST toxin standard. In the axenic culture, peaks co-eluting with standards of GTX 1 and 4 were again observed although at different ratios to the control culture, and the peaks co-eluting with standards of GTX 2 and 3 were more pronounced (Fig. 3c). The peak at ∼3.5 min was also present in axenic samples.


HPLC chromatograms of (a) PST standards GTX 1–4, (b) A. lusitanicum NPECC 253 control culture – early exponential growth (sample 1) and (c) A. lusitanicum NEPCC 253 axenic culture – early exponential growth (sample 1). GTX standard spiked dinoflagellate samples overlaid as coloured traces (blue and red respectively) on chromatograms (b) and (c).

The highest proportion of the toxin load in A. lusitanicum cells from both the control and axenic cultures, at all three sample points, was due to GTX 4 followed by GTX 1 (Table 3). In the early and late exponential growth phase (samples 1 and 2), the control culture contained a higher proportion of GTX 4 than the axenic culture but the reverse was true in the stationary phase of growth. The axenic culture contained the highest proportion of GTX 1 in early and late exponential phase whereas the control culture contained the highest proportion in the stationary phase (sample 3; Table 3).

View this table:

Contribution of each toxin to the total toxin load and total toxicity, expressed as STX equivalents for A. lusitanicum NEPCC 253

SampleIndividual toxin contribution to total toxin load (%)Total toxicity (STX equivalents per cell (pg))
culture treatmentGTX 1GTX 2GTX 3GTX 4
1 – Early exponentialcontrol32.5TaT67.50.179
2 – Late exponentialcontrol22.3TT77.70.017
3 – Stationarycontrol50.0ndbnd50.00.011
  • a a T=toxin present at <0.33 ng (on the column) for GTX 2, and <0.08 ng (on the column) for GTX 3.

  • b b nd=toxin not detected.

The axenic culture produced a higher quantity of each toxin compared to control cultures, with the exception of GTX 4 at sample point 1 (early exponential phase) for which equivalent quantities were produced by both the axenic and control cultures (Fig. 4). Measurable quantities of GTX 2 and GTX 3 were only detectable in early exponential phase from the axenic culture.


Effect of bacteria on PST production by A. lusitanicum NEPCC 253 grown in batch culture over 30 days. Results calculated from HPLC data (n=3), with error bars indicating the maximum and minimum toxin concentration when minimum and maximum cell counts are considered. Sample points 1–3 refer to early exponential, late exponential and stationary phase respectively. ◻=control culture, ▩=axenic culture.

A. tamarense demonstrated a more diverse toxin profile than A. lusitanicum, with six toxin groups quantified including both carbamate and sulfocarbamoyl toxins (Table 4). The presence of a further toxin, the decarbamoyl toxin dcGTX 2, was suspected but could not be confirmed due to the lack of a calibrated standard. Again a peak at ∼3.5 min was observed in all sample chromatograms, but remained unidentified. GTX 4 was the dominant toxin in both the axenic and control cultures, with the exception of early exponential phase from the axenic culture in which C2 dominated (Table 4). It was also noted that the contribution of GTX 1 to the toxin loading of these cultures was much lower than that observed for A. lusitanicum (Tables 3 and 4).

View this table:

Contributions of each toxin detected to the total toxin load and total toxicity, expressed as STX equivalents for A. tamarense NEPCC 407

SampleIndividual toxin contribution to total toxin load (%)Total toxicity (STX equivalents per cell (pg))
culture treatmentGTX 1GTX 4neo-STXC1C2C4
1 – Early exponentialcontrolTa95.60.3nd4.1nd0.520
2 – Late exponentialcontrol1.
3 – Stationarycontrol5.
  • a a T=toxin present at <0.33 ng (on the column) for GTX 2, and <0.08 ng (on the column) for GTX 3.

  • b b nd=toxin not detected.

With A. tamarense, the axenic cultures produced a higher concentration of GTX 1 and C4 (when detected) and a greater quantity of GTX 4 in late exponential and stationary phase samples (samples 2 and 3). However, control cultures produced more GTX 4 in early exponential phase, C1 in late exponential and stationary phase, C2 in stationary phase, and neo-STX in all samples (Fig. 5). Additionally, different trends were noted in the concentration of GTX 4 between the axenic and control cultures, with the control culture decreasing in concentration of GTX 4 over the growth cycle, while the axenic culture showed an increase in GTX 4 as the culture approached the stationary phase of growth (Fig. 5).


Effect of bacteria on PST production by A. tamarense NEPCC 407 grown in batch culture over 30 days. Results calculated from HPLC data (n=3), with error bars indicating the maximum and minimum toxin concentration when minimum and maximum cell counts are considered. Sample points 1–3 refer to early exponential, late exponential and stationary phase respectively. ◻=control culture, ▩=axenic culture.

Tables 3 and 4 also summarise the data in terms of the total toxicity of each culture in STX equivalents derived from the concentration of each toxin and its reported toxicity in the mouse bioassay [38]. Comparing control and bacteria-free cultures, total toxicity was consistently 1.5–3-fold higher in axenic cultures of A. lusitanicum, while the reverse was true for A. tamarense with between 3- and 220-fold higher levels of toxicity observed in control cultures.

4 Discussion

Bacteria can be considered an integral part of the physical environment of toxic dinoflagellates, alternating between being free-living in the medium, attached to the dinoflagellate cell wall or existing internally within the dinoflagellate cell [5]. In all cases there is limited information available on the effect of such bacteria on dinoflagellate production of PST [5].

However, some studies have shown that PST are still produced in bacteria-free cultures of Alexandrium species [79]. Although, a few studies [6,11] have compared the toxicity of axenic dinoflagellate cultures to those of control dinoflagellates containing their normal bacterial flora, none have provided detailed information on the toxin profile and toxin concentrations in these cultures. This study aimed to provide such information by initially producing axenic cultures and subsequently assessing differences in the concentration of PST produced between dinoflagellates containing their natural microbial population and those in which the bacterial flora had been removed.

In this investigation axenic dinoflagellates were prepared after treatment with a bactericidal antibiotic cocktail which was shown to be effective against all the isolated bacterial morphotypes present. A more stringent approach was taken than reported in other studies in assessing the axenic nature of the resultant dinoflagellate cultures. This involved using 17 different bacterial growth media in both solid and broth form, epifluorescent microscopy and PCR.

By using a wide range of solid and broth media formulations, including both oligotrophic and nutrient rich formulas, the probability of detecting bacteria with differing nutrient requirements was greatly increased. The possibility that non-culturable bacteria remained in the dinoflagellate cultures was also assessed using epifluorescence microscopy. Other researchers have applied this technique in conjunction with axenic dinoflagellates, generally with DAPI or acridine orange dyes [6,15]. Substituting Sybr green 1 [27] for these dyes in this study had the advantage of reducing autofluorescence from background detritus and increasing the clarity of the images.

Although molecular techniques have been used routinely to detect the presence of bacteria within a community from a range of environments, such as marine picoplankton [40], hot springs [41,42] and seawater [36], there are no published reports of the use of such techniques to examine the bacterial status of dinoflagellate cultures. In this study, a PCR method with an amplification lower level of 10 cfu per PCR reaction was used to demonstrate that control dinoflagellate cultures contained bacterial DNA. Use of two primer sets capable of amplifying DNA from a wide range of Eubacteria provided strong evidence that lack of amplification of DNA from axenic cultures was due to the absence of Eubacteria from the cultures.

In obtaining bacteria-free dinoflagellate cultures doubts can always arise as to whether bacteria are still present but are not detected by the techniques used. However, by using a combination of the three methods described in this study, this uncertainty is reduced and the possibility that endosymbiotic bacteria remain is vastly reduced compared to other investigations. The treated A. lusitanicum and A. tamarense cultures could reasonably be considered bacteria-free as determined by the above methods.

The GTX 1, 2, 3 and 4, with the epimer pair GTX 1 and 4 dominant, were detected in both axenic and control cultures of A. lusitanicum which is consistent with data reported by other researchers for bacteria-containing cultures [28,29]. This implies that removal of bacteria did not inhibit toxin production by this dinoflagellate nor change the toxin profile.

In the A. lusitanicum culture differences were observed in cell concentrations at the beginning of the growth cycle and it is possible that the cultures were in asynchronous growth in the early part of the cycle, with convergence at day 8 representing the beginning of synchronous cell growth. The differences in cell concentration between days 1–8 could be due to several reasons, including the use of the antibiotic cocktail. As the antibiotic-treated dinoflagellate cultures had been subcultured several times into antibiotic-free media prior to the experiment, it is unlikely that this was responsible for the differences between the cultures. Other possible reasons to explain the differences in cell concentrations and toxicity between the cultures include differences in the handling of the control and axenic cultures. However, all cultures were handled and maintained using identical procedures and environmental parameters, including light exposure times and light intensities. It was, therefore, concluded that the changes in toxin production and dinoflagellate cell concentrations were the direct result of removal of the bacterial microflora.

The growth rate of the A. lusitanicum culture was unaffected by removal of its bacterial microflora, although the concentration of individual toxins and total toxicity generally was greater in axenic cultures compared to control cultures. However, these results conflict with those from the study of Doucette and Powell [6], in which the same dinoflagellate strain was found to be approximately 50% less toxic in axenic cultures compared to a control. The lack of growth rate information in that study makes direct comparison difficult.

Removal of bacterial had a different effect on the toxicity of A. tamarense NEPCC 407, compared to A. lusitanicum. In most instances the toxin profile, including the dominance of GTX 4, was similar in control and axenic cultures, although some differences were found, unlike the A. lusitanicum culture. The most notable difference between the two dinoflagellate species is that removal of the bacterial microflora from A. tamarense NEPCC 407 resulted in a decrease in overall toxicity of the culture. Depending on the growth stage and the toxin, the control culture produced more of some compounds than the axenic culture; this was particularly noticeable with neo-STX. However, the bacteria-free culture did contain a culturable fungus, Botrytis sp., which was not isolated from the control culture. Thus, it was difficult to conclude whether the differences observed between the A. tamarense cultures were due to bacteria alone.

It is possible that the fungus was present as a spore or in very low concentrations in the control culture. Bacteria have been shown to suppress fungal growth in other environments [43], which leads to the suggestion that removal of bacteria from the culture reduced competition for nutrients and enhanced growth of the fungus. Evidence to support this view, is the inability of the fungus to grow in low nutrient media including the dinoflagellate growth media, and the observation that no other axenic dinoflagellate cultures maintained in this laboratory, and subcultured using the same aseptic technique, contained this fungus.

The study of Singh et al. [10] demonstrated that removing bacteria from cultures of A. tamarense resulted in changes in growth rate and higher toxicity in the axenic culture, the reverse of that observed here with the A. tamarense culture in which growth rates were unaffected and toxicity was lower in axenic cultures. Also, Dantzer and Levin [11] showed higher toxicity in some axenic A. tamarense cultures on a per cell basis. However, as with the Doucette and Powell study [6], dinoflagellate growth rate information was omitted, although cell counts for the sampling points would suggest that the growth rates differed, indicating an indirect effect on toxicity due to the removal of bacteria.

This study is the first to provide detailed information on individual toxin profiles between axenic and control dinoflagellate cultures and shows that bacteria have a limited effect, if any on the dinoflagellate toxin profile but can affect the quantity of toxin produced. Summarising this data with that of previous studies [6,10,11] leads us to conclude that bacteria directly influence toxin production in some dinoflagellate cultures while in others they may exert an indirect effect through influencing the growth rate. The mechanisms by which this occurs are unknown but Gallacher and Smith [5] suggest that a number of factors could be involved including the production of unknown co-factors which inhibit or stimulate dinoflagellate toxin synthesis, signalling molecules, nutrient availability and metabolism of the toxins by bacteria.

Changes in nutrients and other parameters, e.g. salinity, influence dinoflagellate growth and toxin production in laboratory cultures [1]. This study suggests that bacteria may have a quantitatively similar effect to these other more easily measurable factors. Providing the above information on laboratory cultures is of interest but leads to questions on the role of bacteria in dinoflagellate toxicity in the much more complex and dynamic marine environment. Therefore, future studies should also investigate the role of viruses in dinoflagellate growth and toxin production given that phytoplankton are readily susceptible to these agents [44].

Defining the identity of bacteria which influence toxicity in dinoflagellate cultures will aid in field investigations by allowing the development of bacterial oligonucleotide probes. Bacteria isolated from dinoflagellates used in this study have subsequently been identified [45], with studies currently underway to develop oligonucleotide probes to isolates of interest for use in investigations into dinoflagellate toxicity in the environment.


This research was supported by the Fisheries Research Services, Marine Laboratory, Aberdeen and the EU FAIR programme CT96-1558. Thanks are extended to Jennifer Graham for technical assistance and Sylvia Duncan for advice on growth curve analysis.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
View Abstract