OUP user menu

Phylogenetic and functional diversity of the cultivable bacterial community associated with the paralytic shellfish poisoning dinoflagellate Gymnodinium catenatum

David H. Green, Lyndon E. Llewellyn, Andrew P. Negri, Susan I. Blackburn, Christopher J.S. Bolch
DOI: http://dx.doi.org/10.1016/S0168-6496(03)00298-8 345-357 First published online: 1 March 2004


Gymnodinium catenatum is one of several dinoflagellates that produce a suite of neurotoxins called the paralytic shellfish toxins (PST), responsible for outbreaks of paralytic shellfish poisoning in temperate and tropical waters. Previous research suggested that the bacteria associated with the surface of the sexual resting stages (cyst) were important to the production of PST by G. catenatum. This study sought to characterise the cultivable bacterial diversity of seven different strains of G. catenatum that produce both high and abnormally low amounts of PST, with the long-term aim of understanding the role the bacterial flora has in bloom development and toxicity of this alga. Sixty-one bacterial isolates were cultured and phylogenetically identified as belonging to the Proteobacteria (70%), Bacteroidetes (26%) or Actinobacteria (3%). The Alphaproteobacteria were the most numerous both in terms of the number of isolates cultured (49%) and were also the most abundant type of bacteria in each G. catenatum culture. Two phenotypic (functional) traits inferred from the phylogenetic data were shown to be a common feature of the bacteria present in each G. catenatum culture: firstly, Alphaproteobacteria capable of aerobic anoxygenic photosynthesis, and secondly, Gammaproteobacteria capable of hydrocarbon utilisation and oligotrophic growth. In relation to reports of autonomous production of PST by dinoflagellate-associated bacteria, PST production by bacterial isolates was investigated, but none were shown to produce any PST-like toxins. Overall, this study has identified a number of emergent trends in the bacterial community of G. catenatum which are mirrored in the bacterial flora of other dinoflagellates, and that are likely to be of especial relevance to the population dynamics of natural and harmful algal blooms.

  • Dinoflagellate
  • Paralytic shellfish poisoning
  • Small subunit rDNA
  • Cultivable bacterial diversity
  • Rhodobacteraceae
  • Alteromonadaceae
  • Aerobic anoxygenic photosynthesis
  • Hydrocarbon utilisation
  • Gymnodinium catenatum

1 Introduction

Gymnodinium catenatum (Graham) is an unarmoured (naked) marine dinoflagellate responsible for paralytic shellfish poisoning (PSP) in temperate and tropical waters off all major continents [1]. It is the only naked dinoflagellate known to produce a suite of potent neurotoxins, the saxitoxins (STX) and related derivatives, collectively termed paralytic shellfish toxins (PSTs). Outbreaks of PSP in the human population result from blooms of G. catenatum being ingested by filter-feeding bivalves that accumulate the PSTs. Human consumers of the PST-contaminated shellfish suffer symptoms ranging from tingling of the lips through to respiratory failure in extreme cases. The worldwide distribution and incidence of G. catenatum and other harmful algal blooms (HABs) has increased markedly in the past few decades [2,3] fuelling significant interest in the factors that influence the growth and toxicity of HAB species.

There is considerable debate surrounding the ‘true’ source of the PST produced by G. catenatum, Alexandrium spp. and Pyrodinium bahamense var. compressum, particularly the role that bacteria may play in bloom toxicity. A bacterial source of PST production was first put forward by Silva et al. [4], who described the presence of intracellular bacteria in PST-producing dinoflagellates. Later, bacteria associated with these PST-producing dinoflagellates were also shown to produce PST-like toxins [5]. Subsequent studies indicate that bacteria may potentially influence PST content or production of the algal culture in several ways: autonomous production of PST-like toxins (e.g. [6]); modulating the toxicity of their host [79]; biotransformation of the PST derivatives [10]; or possibly a combination of these factors. The specific mechanisms of interaction are currently unknown, although the relationship between toxin production and the bacterial flora may well be manifested at a much broader community level, through the production of stimulatory or inhibitory factors that may induce or repress toxin production [11].

Bacteria have a potentially more profound role in the development of HAB events, beyond just their potential effects on toxin production (for review see [12]). They are able to exert considerable influence on bloom population dynamics through ‘bacterial–algal’ interactions, for example positive stimulation of growth [1315], promotion of sexuality [16], antagonism mediated via the production of algicidal factors (e.g. [17]) and inhibition of cyst formation [18], and protective effects associated with an existing microflora insulating the algal host from antagonistic bacterial activity [19].

Our interest in the bacteria associated with G. catenatum stems from the observation that as part of its sexual cycle, G. catenatum produces sexual resting cysts [20] that, when germinated in the laboratory setting, often produce ‘atypical’ cultures which express abnormally low amounts of PST per cell (Table 1). We postulated that the process of sonicating and sterile washing during isolation was likely to remove much of the cysts’ surface-associated bacteria, and that after laboratory germination the reduced bacterial diversity may affect the cells ability to produce ‘normal’ levels of PSTs [21]. Preliminary evidence showed that these atypical PST-producing cultures had a reduced bacterial diversity compared to their toxic counterparts [21].

View this table:

The geographic origin and the PST content of the G. catenatum strains examined in this study

Strain nameTotal PST content (fmol cell−1)Isolation sourceCulture collectiona
Typical PST content:
CAWD101n.d.bKaitaia, New Zealand (2000)Cawthron Institute (NZ)
GC21V244Ria de Vigo, Spain (1986)CCMP (USA)
GCDE08189Derwent Estuary, Tasmania (1987)CSIRO (Australia)
YC499B15316Yellow Sea, Korea (1998)T.G. Park (Korea)
Atypical PST content:
GCHU114Huon Estuary, Tasmania (1988)CSIRO (Australia)
GCJP01b.d.cSeto Inland Sea, Japan (1985)CSIRO (Australia)
GCTRA143Spring Bay, Tasmania (1993)CSIRO (Australia)
  • aAlgal culture collection from which strains analysed in this study were acquired.

  • bNot determined.

  • cBelow detection limit.

To examine the potential link between the bacterial flora on the cyst surface and resulting G. catenatum culture toxicity, we compared the cultivable bacterial flora from a range of G. catenatum strains – representative of both typical and atypical PST-producing cultures from a variety of geographic populations (Table 1). Our main objectives were to: (1) document the typical bacterial flora of this dinoflagellate; (2) assess the phylogenetic, and by inference from this, the functional diversity of the bacteria associated with typical and atypical PST-producing cultures; and (3) examine the function that specific bacteria might have in both the lifecycle and toxicity of G. catenatum. To address these questions, small subunit rDNA (SSU rDNA) sequencing was used to establish the phylogenetic affiliations of all cultivable bacteria, and subsequently, we examined the ability of bacteria to grow photoheterotrophically, to utilise hydrocarbons and to autonomously produce PST.

2 Materials and methods

2.1 Algal culture

The G. catenatum strains used in this study (Table 1) were grown at 18°C in 25 cm2 tissue culture flasks (Nunc, Norway) in GSe medium [22] with a photon flux density of 50–70 μmol m−2 s−1 from cool-white fluorescent lighting (Phillips, The Netherlands) with a 12:12 h light:dark photoperiod. All cultures were handled aseptically to prevent bacterial contamination and cross-contamination between cultures.

2.2 Bacterial culture

Bacteria were isolated and maintained on a modified marine agar (ZM/10) prepared with 75% aged filtered (1.0 μm pore size) natural seawater, 0.05% bacto-peptone (Difco, USA), 0.01% yeast extract (Difco) and 1.5% bacto-agar (Difco), supplemented following autoclaving with sterile trace elements and vitamins at the same concentrations as used in GSe medium [22]. Zobells 2216E marine agar and broth supplemented with GSe trace elements and vitamins (ZM/1) were used as necessary.

Bacterial isolation was performed by harvesting 1 ml of a late-logarithmic phase G. catenatum culture by brief centrifugation (12 000×g for 10 s). The spent medium was removed and the algal cell pellet resuspended in 100 μl of sterile seawater and vortexed. The cell suspension was diluted 10-fold and 100 μl of each dilution spread onto ZM/10 agar plates and incubated in the dark at 18°C for 3 weeks. Bacterial colonies with distinct colony morphology were picked (and the number of that colony morphology counted) and serially passaged on ZM/10 agar until the purity of the isolate was assured. Bacterial isolates were grown in ZM/10 or ZM/1 broth, glycerol was added (20% v/v), and the cells were stored at −80°C.

The ability of bacterial strains to utilise aliphatic hydrocarbons as a sole carbon source was demonstrated following the growth of bacteria on a synthetic seawater agar SM1 [23] prepared with deionised distilled water and supplemented with sodium nitrate or ammonium nitrate as the nitrogen source. Hydrocarbons, n-tetradecane or n-hexadecane (Sigma-Aldrich, UK), were supplied in the vapour phase by adding 200 μl to a sterile filter pad in the lid of the Petri dish. Growth was examined after 3 weeks of incubation at 25°C.

2.3 Toxin analysis

Analysis of the ability of bacterial strains to autonomously produce STX and PST-like toxins was conducted on bacterial cell pellets harvested from 100 ml of ZM/1 broth grown in the dark for 48 h at 25°C on an orbital shaker (160 rpm). Cell pellets were resuspended in 0.05 N acetic acid, sonicated, and cell debris removed by centrifugation and 0.2 μm membrane filtration prior to being assayed with the saxiphilin and rat brain sodium channel assays [24].

Analysis of the PST content of G. catenatum cultures was by liquid chromatography fluorescence detection of the C-toxins, gonyautoxins and STX as described by Negri and Llewellyn [25], and detection of the GC-toxins as described by Negri et al. [26].

2.4 DNA manipulation

Bacterial genomic DNA was extracted using a method based on cetyltrimethylammonium bromide purification (CTAB) [27]. The polymerase chain reaction (PCR) was used to amplify the SSU rDNA gene from chromosomal DNA using the primer pair 27F (AGAGTTTGATCMTGGCTCAG) and 1492R (ACGGCTACCTTGTTACGACTT) [28]. PCR was carried out on an MJ Research PTC200 DNA Engine thermocycler and used 1 U of Taq polymerase (ABgene, UK) in a 50 μl reaction containing a final concentration of 1.8 mM Mg2+, 20 mM NH4SO4, 75 mM Tris–HCl (pH 8.8) and 0.01% Tween 20. Cycling parameters were as follows: 94°C for 2 min; 26 cycles of 55°C for 30 s, 72°C for 2.5 min and 94°C for 10 s; followed by 72°C for 10 min. The PCR products were purified through Centricon-PCR ultrafilters (Millipore, UK) according to the manufacturer's instructions and sequenced in both directions using 27F and 1492R primers and ABI-Prism ‘Big-Dye’ terminator chemistry (Applied Biosystems, USA) according to the standard protocols. Sequence reactions were electrophoresed on an ABI 377 DNA sequencer (Applied Biosystems), and resulting sequences aligned and manually checked for consistent base-calling, using Sequence Navigator (Ver. 1.0.1, Applied Biosystems, USA).

Detection of pufLM photosynthetic reaction centre genes in bacteria was carried out by PCR amplification with the forward (5′-CTKTTCGACTTCTGGGTSGG-3′) and reverse (5′-CCATSGTCCAGCGCCAGAA-3′) primers as described by Béja et al. [29]. PCR reactions contained 1.0 U/50 μl of a 7:1 (U/U) mixture of Taq and Pfu polymerase (Promega, UK) in a reaction containing a final concentration of 2 mM Mg2+, 20 mM NH4SO4, 75 mM Tris–HCl (pH 8.8) and 0.01% Tween 20. Cycling was carried out using an MJ PTC200 DNA Engine thermocycler as follows: 94°C for 2 min; 30 cycles of annealing between 48 and 58°C for 30 s, 72°C for 2.5 min, 94°C for 10 s; followed by 72°C for 10 min. PCR products of ca. 1500 bp were purified through Centricon-PCR filters, A-tailed, ligated into pGEM-T Easy (Promega, UK) and transformed into Escherichia coli XL1-Blue (Stratagene, The Netherlands). Selected clones were sequenced in both directions using the primers ARD-F (GCCATGGCGGCCGCGGGAATT) and ARD-R (AGGCGGCCGCGAATTCACTAG) that bind immediately adjacent to the pGEM-T Easy cloning site. DNA sequencing and sequence analysis were carried out as outlined above.

2.5 Genetic identity and phylogenetic inference

SSU rDNA sequences were aligned with representative SSU rDNA sequences of other bacteria using the software programme Clustal X [30] and were again checked for any potential sequencing errors. Genetic identity was determined using Ribosomal Database Project II (RDP II) Sequence Match [31]. Genetic identity was in most instances assigned to the closest taxonomically described bacterial sequence in the RDP II and listed in Table 2 according to Bergey's taxonomic outline of the prokaryotes [32].

View this table:

Phylogenetic affiliation of the bacteria isolated from G. catenatum and their percentage dominance in the culture from which they were isolated

Strain no. and taxonomic affiliationAccession no.CFUa (%)Closest relative in the RDP IIbAccession no.ID (%)G. catenatum strain
Proteobacteria (Alpha)
DG943AY2580897.2Mesorhizobium amorphaeAF04144296.4GCJP01
DG1023AY25809620.7Mesorhizobium chacoenseAJ27824994.2CAWD101
DG948AY2580913.6Rhodobium orientisD3079290.8GCJP01
DG869AY25807461.2Roseobacter gallaeciensisY1324493.8GCHU11
DG874AY2580752.0Roseobacter sp.AF09849592.6GCDE08
DG877AY25807632.9unidentified alphaproteobacteriumAB01868999.8GCDE08
DG878AY2580774.6Roseobacter sp.AF09849592.5GCDE08
DG882AY25807839.4Ruegeria gelatinovoransD8852395.0GCDE08
DG885AY2580790.8Sulfitobacter mediterraneusY1738794.5GCDE08
DG888AY25808065.1Roseobacter sp.AF09849592.0YC499B15
DG889AY25808113.1Roseobacter litoralisX7831294.4YC499B15
DG891AY2580820.7Stappia aggregataD8852098.1YC499B15
DG895AY2580840.1Hyphomonas johnsoniiAF08279198.4YC499B15
DG897AY2580856.5Roseobacter sp.AF09849592.4YC499B15
DG898AY2580860.4Ruegeria algicolaX7831597.8YC499B15
DG941AY2580877.2Roseobacter sp.AF09849592.6GCJP01
DG942AY25808846.9R. litoralisX7831294.5GCJP01
DG944AY25809018.1Roseovarius toleransY1155194.2GCJP01
DG981AY25809439.3unidentified alphaproteobacteriumAJ00256594.2GCTRA14
DG1020AY25809541.4S. mediterraneusY1738798.8CAWD101
DG1127AY25809932.2Roseobacter sp.AF09849596.0GC21V
DG1128AY25810048.3Ruegeria atlanticaAF12452192.4GC21V
DG1132AY2581025.6R. litoralisX7831293.1GC21V
DG1133AY2581030.8unidentified alphaproteobacteriumAB01868999.8GC21V
DG949AY25809210.8Azospirillum doebereineraeAJ23856786.3GCJP01
DG1131AY2581012.4Thalassospira lucentensisAF35866489.0GC21V
DG892AY2580830.1Sphingopyxis alaskensisAF14575498.4YC499B15
DG978AY25809332.7Blastomonas ursincolaAB02428992.3GCTRA14
DG1024AY2580978.3B. ursincolaAB02428992.9CAWD101
Proteobacteria (Gamma)
DG870AY25810620.4M. hydrocarbonoclasticusAB01914895.5GCHU11
DG879AY25810713.1M. hydrocarbonoclasticusAB01914895.5GCDE08
DG893AY2581100.1M. hydrocarbonoclasticusAB01914895.7YC499B15
DG979AY25811214.7M. hydrocarbonoclasticusAB01914895.1GCTRA14
DG980AY2581138.2M. hydrocarbonoclasticusAB01914895.2GCTRA14
DG1022AY2581148.3Alteromonas macleodiiY1822892.2CAWD101
DG1135AY2581150.8Pseudoalteromonas luteoviolaceaX8214496.3GC21V
DG1136AY2581160.1M. hydrocarbonoclasticusAB01914895.2GC21V
DG940AY2581113.6Oceanobacter kriegiiAB00676790.4GCJP01
DG812AY2581040.2Alcanivorax borkumensisY1257999.4CAWD101
DG813AY2581050.3A. borkumensisY1257999.4YC499B15
DG881AY2581090.7A. borkumensisY1257999.5GCDE08
DG880AY2581080.7Acinetobacter lwoffiiX8166599.3GCDE08
DG868AY2581177.1Zobellia sp.AF53013796.3GCHU11
DG886AY2581200.7Zobellia sp.AF53013796.3GCDE08
DG945AY2581231.1Aequorivita antarcticaAY02780489.1GCJP01
DG975AY2581251.6Zobellia sp.AF53013793.7GCTRA14
DG976AY2581260.2marine psychrophileU8588296.8GCTRA14
DG977AY2581273.3Zobellia uliginosaM6279985.8GCTRA14
DG1025AY2581294.1Polaribacter franzmanniiU1458692.3CAWD101
DG1027AY2581308.3Arenibacter latericiusAF05274291.8CAWD101
DG1030AY2581320.2Zobellia sp.AF53013792.3CAWD101
DG1134AY2581341.6Cellulophaga fucicolaAJ00597393.2GC21V
DG873AY25811811.2Microscilla furvescensAB07807991.3GCHU11
DG887AY2581210.7Cyclobacterium sp.AJ24468991.8YC499B15
DG890AY25812213.1unidentified bacteriumAJ22494284.7YC499B15
DG946AY2581241.4M. furvescensAB07807996.5GCJP01
DG1021AY2581284.1unidentified bacteriumAF08704385.7CAWD101
DG1129AY2581338.1Flexibacter aggregansAB07803888.0GC21V
DG876AY2581195.3Micrococcus luteusAF05728999.2GCDE08
DG1029AY2581310.4Nocardioides sp.U6129891.4CAWD101
  • aRelative abundance (percentage) of each bacterial isolate in the G. catenatum culture from which it was isolated.

  • bNamed species are based on the most similar named SSU rDNA sequence listed on the RDP II [31]. The percentage identity and accession number of the closest species is shown alongside.

Phylogenetic inference was performed on visually corrected alignments using PAUP 4.0*[33]. All ambiguous alignment positions were masked from the analysis. The method of maximum likelihood (ML) using the estimated rates of transition/transversion and α-shape parameters (estimated from the data set following heuristic searching) were used to infer each phylogeny. Bootstrap support for each inferred tree was established following resampling of 1000 data sets based on neighbour-joining analysis [34].

3 Results

3.1 Strain characterisation

A total of 61 distinct bacteria spanning three phyla were cultured from the seven strains of G. catenatum (Tables 2 and 3). Thirty (49%) of the bacterial strains were affiliated with the Alphaproteobacteria, of which 21 (34%) were affiliated within the Rhodobacteraceae, with one or more isolates of this family identified in all seven G. catenatum cultures. Thirteen (21%) isolates were affiliated with the Gammaproteobacteria, with each G. catenatum culture having one or more Gammaproteobacteria. Of this group, eight isolates were affiliated to the Alteromonadaceae, cultured from six of the seven G. catenatum cultures. The remaining isolates came from two phyla, the Bacteroidetes (26%) and the high G+C% Gram-positive Actinobacteria (3%) (Table 3).

View this table:

Summary of the cultivable bacteria identified in each G. catenatum culture

G. catenatum strainTotal (n)Number of isolates per lineage
Typical PST content:
Atypical PST content:
Total (n)613013162

From cross-referencing of colony counts and SSU rDNA data, the cultivable bacterial flora of the G. catenatum cultures examined was shown to be dominated by the Alphaproteobacteria (Table 2). In all G. catenatum cultures examined, a single member of the Rhodobacteraceae was the most abundant bacterium, accounting for between 39 and 65% of the cultivable flora. The relative abundance of any one individual Alphaproteobacteria ranged considerably (0.1–65%), averaging 19%, while the average culture abundance of the Rhodobacteraceae was 22%. The abundance of Gammaproteobacteria and Bacteroidetes in G. catenatum cultures averaged 5 and 4% respectively.

The phylogenetic relationship of the cultivable strains was compared to their closest SSU rDNA affiliates and selected bacterial type species (Figs. 1, 2 and 3). Bootstrap support for higher order branching between Alphaproteobacteria lineages (Fig. 1) was typically low. While this implies uncertainty, maximum parsimony analysis did nevertheless infer a similar tree to the ML analysis. Bootstrap support for the branching order of both the Gammaproteobacteria and Bacteroidetes/Actinobacteria ML was high. A significant number of the bacterial strains (39%) had less than 93% SSU rDNA identity and only a relative few (18%) exceeded an identity of 97% to their closest relative. Phenotypic characteristics such as cell morphology, pigmentation, motility, catalase and oxidase activity were recorded (data not shown) and, for most novel genera and species, these tests were consistent with the family or genus to which they showed highest phylogenetic affiliation.


Phylogenetic inference of the cultivable Alphaproteobacteria and AAP growth potential. ML neighbour-joining analysis of 1254 bp of SSU rDNA was used in the analysis (bootstrap support ≥70% is shown next to its respective branch). The scale bar indicates the number of nucleotide substitutions per site. Asterisks next to accession numbers denote that this bacterial sequence was isolated from a dinoflagellate other than G. catenatum. Right of the tree, illustrates which of the Alphaproteobacteria were potentially capable of AAP growth based on colony pigmentation following aerobic, dark growth and the PCR detection of the conserved photosynthetic reaction centre genes, pufLM. +, pufLM detected; −, pufLM not detected; B, beige; C, cream; G, grey; P, pink; PP, pale pink; PR, pale red; W, white; Y, yellow.


Phylogenetic inference of the cultivable Gammaproteobacteria and hydrocarbon utilisation. ML neighbour-joining analysis was conducted on 1347 bp of aligned SSU rDNA (bootstrap support ≥70% is shown next to its respective branch). Right of the tree, shows qualitative assessment of each bacterial strain's ability to utilise hydrocarbon as the sole carbon source. TET, n-tetradecane as carbon source; HEX, n-hexadecane as carbon source; NEG, no carbon source; +, ++, +++ indicate the degree of growth; ± possible growth; − no growth.


Phylogenetic inference of the cultivable Bacteroidetes and Actinobacteria. ML neighbour-joining analysis was conducted on 1203 bp of aligned SSU rDNA (bootstrap support ≥70% is shown next to its respective branch).

A number of the bacterial strains isolated were phylogenetically closely related to one another, while having originated from G. catenatum cultures from different parts of the world (see Figs. 1 and 2). For example, strains of Alcanivorax sp. were isolated from G. catenatum cultures originating from New Zealand, Australia and Korea (Table 2). All three of these G. catenatum cultures had been isolated and maintained in exclusion of each other (Table 1) up to the point of this study. Another distinct group was a specific clade of Roseobacter–Roseovarius-like strains (DG874, 878, 888, 897, 941, 944 and 1127; Fig. 1) that originated from G. catenatum cultures isolated from the sea areas as separate as Australia, Korea, Japan and Spain. In addition, many strains were also closely related to bacteria identified in association with other dinoflagellates (denoted with an asterisk; Figs. 1, 2 and 3) such as the PST-producing Alexandrium tamarense, Alexandrium lusitanicum and Alexandrium affine, the diarrhetic shellfish poisoning (DSP) Prorocentrum lima, and the non-toxic dinoflagellate Scrippsiella trochoidea. These similarities were especially evident among the dinoflagellate-derived strains belonging to the Rhodobacteraceae (Fig. 1) and Alteromonadaceae (Fig. 2) families.

3.2 Bacterial diversity of atypical PST-producing G. catenatum

One of the specific aims of this work was to understand if atypical PST-producing G. catenatum cultures had a microbial flora distinct from that of toxic G. catenatum cultures. The data presented in Table 3 show that of the three atypical PST-producing G. catenatum cultures analysed here, there were in two instances a reduction in the number of strains cultivable. Bacterial isolation from other atypical PST-producing G. catenatum cultures showed that lower numbers of cultivable bacteria were typical (data not shown). The number of bacteria isolated from two of the three atypical PST-producing G. catenatum cultures showed a reduction in the Alpha- and Gammaproteobacteria diversity of these cultures, as compared to typical PST-producing G. catenatum (Table 3). The numbers of Bacteroidetes isolated remained approximately the same (Table 3), but while this appeared to make them a relatively more dominant feature of these cultures, their relative abundance (colony-forming units (CFU)) in the atypical PST-producing G. catenatum cultures (5%) was not obviously different to that in the typical PST-producing cultures (3%).

3.3 Bacterial phenotypes

Extrapolation from the phylogenetic data showed that a number of the strains isolated may have belonged to one or two phenotypic groups: the aerobic anoxygenic photosynthetic (AAP) Alphaproteobacteria, and Gammaproteobacteria capable of hydrocarbon utilisation. Phenotypic characterisation showed that the 13 (67%) of the 21 Rhodobacteraceae isolates were potentially capable of AAP as evidenced by the presence of the conserved photosynthetic reaction centre genes pufLM, and in most cases pink-red colony pigmentation following aerobic incubation in the dark, which is indicative of the production of bacteriochlorophyll a (Fig. 1). Atypical responses were demonstrated by two isolates (DG874 and 878) that only produced a pink-red pigmentation when grown for at least 4 weeks on a rich medium in the dark. Three isolates were not observed to have any pigmentation (DG869, 882 and 1128), but had pufLM genes that could be detected by the PCR. The bacterium DG943 (Mesorhizobium sp.) was the only isolate not phylogenetically affiliated to the Rhodobacteraceae, that was capable of AAP, as shown by its pink colony pigmentation following dark, aerobic incubation, and PCR detection of pufLM (Fig. 1). Although no AAP Alphaproteobacteria was cultured from G. catenatum CAWD101, the PCR was able to detect the pufLM genes from the total DNA extracted from G. catenatum CAWD101 culture (data not shown). The PCR was used to screen all Gammaproteobacteria isolates for the gene coding for the light-harvesting proteorhodopsin protein [35]. No PCR product was detected in any of the isolates tested (data not shown).

Bacteria with the potential to utilise aliphatic hydrocarbons as a carbon source were identified in the majority of the G. catenatum strains examined. Members of the Gammaproteobacteria, with close phylogenetic affiliations to either Alcanivorax borkumensis SK2 or Marinobacter hydrocarbonoclasticus ATCC 27132, were isolated from six of the seven G. catenatum cultures. All three Alcanivorax isolates and four of the six Marinobacter isolates were shown to be able to utilise n-tetradecane and n-hexadecane as the sole carbon source. While the remaining Marinobacter isolates grew poorly or insufficiently different to the control (no carbon source added) to enable accurate assessment of their growth potential (Fig. 2). All of the Alphaproteobacteria were unable to utilise hydrocarbon as a sole carbon source (data not shown). Unexpectedly, all Alteromonadaceae and Alcanivorax were capable of growth on the synthetic seawater agar without any additional carbon source (Fig. 2). Agarolytic activity (colony-sized, shallow depressions ≤1 mm) and subsurface growth were observed in all of Alcanivorax and Marinobacter isolates. Both Alteromonas sp. DG1022 and Pseudoalteromonas sp. DG1135 grew in the absence of a carbon source but with no evidence of agarolytic activity or subsurface growth.

The bacteria cultured from two strains of G. catenatum (GCDE08 and GCHU11) were analysed for the ability to autonomously produce PST-like toxins. All of the nine bacterial strains (DG869, 870, 874, 877, 878, 880, 881, 882 and 885) tested for STX/PST showed no evidence for sodium channel-blocking activity in any of the cell pellets examined (data not shown).

4 Discussion

This study aimed to comprehensively document the bacterial flora associated with the PST-producing marine dinoflagellate G. catenatum as a first step to understanding the influence of bacteria on algal growth, physiology and toxicity, and their role in the development of HABs. We have initially focussed on the cultivable bacterial flora so that future research investigating the biochemistry, physiology and genetics of the bacterial–algal interactions can utilise characterised, cultivable organisms [36]. Our previous work had suggested that the bacterial flora may influence the expression of PST production in G. catenatum, we therefore chose to examine the bacterial diversity of G. catenatum cultures that produce abnormally low amounts of PST.

4.1 Bacterial community composition

The phylogenetic characterisation of the cultivable bacteria isolated from the seven G. catenatum cultures demonstrated that there were a number of emergent trends across the G. catenatum examined. In summary, the Alphaproteobacteria dominated the strains isolated, and an individual Alphaproteobacteria (Rhodobacteraceae) was always the most numerically abundant bacterium present in each culture (Table 2). Half of all of the Alphaproteobacteria isolated were capable of a mode of photosynthetic growth, termed AAP ([37] and references therein). The second trend was for there to be cultivable oligotrophic and/or hydrocarbon-degrading Gammaproteobacteria present in almost all of the cultures. And thirdly, one or more cultivable isolates belonging within the Flexibacteraceae or Flavobacteriaceae families of the Bacteroidetes were always present in each culture.

The bacterial flora of G. catenatum generally mirrors that found associated with other dinoflagellates, being dominated by the Alphaproteobacteria (principally the Rhodobacteraceae – frequently referred to as Roseobacter clade; e.g. [38]). For example, 50% of all phylotypes identified in four Pfiesteria sp. cultures were affiliated with the Alphaproteobacteria, with the Rhodobacteraceae Rg. algicola and Hyphomonas jannaschiana-like bacteria among the most numerous of these phylotypes [15]. Rhodobacteraceae were also a dominant feature of the bacterial flora associated with the DSP-producing dinoflagellate P. lima[39], and from which the association Rg. algicola was originally described [40]. The bacterial flora of Alexandrium spp. and S. trochoidea cultures were also dominated by Alphaproteobacteria, with the Roseobacter clade dominating both the cultivable species and ribotype clones identified [38]. Like G. catenatum, members of the Alteromonadaceae (Marinobacter and Alteromonas) were consistently identified in other dinoflagellate cultures [14,15,38,39,41].

The high incidence of Alphaproteobacteria associated with algae does not appear to be restricted to the dinoflagellates, as Alphaproteobacteria, primarily Rhodobacteraceae, were always identified in association with each of six different species of diatom culture [42]. Bacterial culture from the domoic acid-producing pennate diatoms, Pseudo-nitzschia multiseries, Pseudo-nitzschia seriata and non-toxic Pseudo-nitzschia delicatissima, consistently identified one or more Alphaproteobacteria associated with each of these cultures (D. Green, J. Fehling and S. Bates, unpublished data).

A striking feature of the bacterial flora of G. catenatum was the high degree of genetic similarity of members of the Alpha- and Gammaproteobacteria (Rhodobacteraceae and the Alteromonadaceae, respectively) compared to other dinoflagellates, particularly the PST-producing genus Alexandrium[38]. For example, G. catenatum bacterial isolate DG898 and R. algicola from P. lima[40] shared a sequence identity of 97.6% (Fig. 1) and DG893/DG1136 and Marinobacter sp. 407-13 from A. tamarense[38] all shared a sequence identity ≥99.5% (Fig. 2). Similarly, these groups also showed consistently high degree of genetic similarity between the G. catenatum cultures examined, even though the G. catenatum cultures had originated from separate geographic regions around the world.

The similarities of bacterial flora across different dinoflagellates (compare this study, [15,38,41]) have two potential explanations. Firstly, there are selective mechanisms operating in laboratory cultures [15,42] that favour genera from within the Rhodobacteraceae and Gammaproteobacteria, such as Marinobacter and Alcanivorax. The selection for specific populations of bacteria adapted to the utilisation of algal extracellular products [43] is one mechanism. Algal extracellular products such as dimethylsulphoniopropionate (DMSP) may well select for specific bacterial taxa such as the Roseobacter clade [44], of which the latter are recognised to utilise as a source of carbon and sulphur ([45] and references therein). G. catenatum is recognised, like many dinoflagellates, to be rich in fatty acids, sterols, lipids and oils [46], and this may explain the high frequency with which Marinobacter and Alcanivorax were identified in G. catenatum cultures. Isolates of both genera were observed to be capable of utilising complex carbohydrates and hydrocarbons (Fig. 2), and Alcanivorax is regarded as a highly fastidious organism that can only utilise aliphatic hydrocarbons and a few fatty acids as a carbon source [23]. The high N and P concentrations used in algal media, or abiotic forces exerted by laboratory glassware and plastics used in the isolation and culture of algae, represent obvious departures from the natural environment, and thus may also contribute to the in vitro development of communities dominated by specific bacterial taxa.

The second explanation for the similarities in the bacterial community across G. catenatum cultures and with other dinoflagellates is that the bacteria from these groups may be of specific importance to the growth and physiology of dinoflagellate cells. Bacterial mineralisation of the algal extracellular products and phytodetritus is recognised as being an important part of the ‘microbial loop’, re-supplying algal cells with readily utilisable forms of C, N and P [47]. The supply of vitamins [48], chelated iron by bacterially produced siderophores [49], or the production of cytokinins [50] are examples where bacterially produced factors have been shown to stimulate algal growth. It may also be that the aerobic photoheterotrophs (AAP) identified in this study, which dominated the cultivable bacterial flora of G. catenatum cultures, may have a role in contributing energy to G. catenatum growth. This premise extends from the work which has shown that the aerobic photoheterotrophs are an abundant [29] and important component of the marine bacterioplankton community through their ability to contribute to carbon cycling in the oceans [51].

Three reports have identified specific bacteria as key components of the bacterial flora associated with the stimulation of dinoflagellate growth. Sakami et al. [13] demonstrated that several bacteria, one of which was identified as an Alteromonas sp., isolated from the cell surface of the ciguatera-producing dinoflagellate Gambierdiscus toxicus, stimulated growth of this dinoflagellate. A bacterium related to Rg. algicola was identified as especially important to enhancing the growth rate of axenic cultures of Pfiesteria sp. [15], and Alteromonas spp. were observed to be a fundamental part of the bacterial community responsible for growth enhancement of Alexandrium fundyense[14]. Importantly, these bacteria belong to the two bacterial families consistently encountered in G. catenatum and other dinoflagellate bacterial communities.

While it is recognised that selective effects may influence the bacterial community of the algal cultures we observed, similar patterns of bacterial association have been noted from field populations. Using fluorescent in situ hybridisation probes, an association was observed between the Roseobacter and Alteromonas clade's abundance and PST-producing Alexandrium spp. cell numbers in natural bloom populations [52]. Analysis of the associated bacterial population with a bloom of the dinoflagellate Lingulodinium polyedrum also demonstrated the presence of Roseobacter and Marinobacter in the free-living fraction and Roseobacter in attached bacterial population [53]. Several other field studies have reported the Rhodobacteraceae as a notable feature of the bacterioplankton community associated with mixed phytoplankton blooms [5456]. These data support the idea that the specific bacterial–algal associations that we and others have observed, are not culture-induced artefacts, but are likely to be relevant and important in natural field populations.

4.2 Role of bacteria in PST production

The role bacteria have in dinoflagellate PST production and bloom toxicity is still an open question. There are several reports of autonomous PST production by bacteria isolated from toxic dinoflagellate cultures (e.g. [5,6,57]) and some have presented convincing chromatographic evidence of their similarity to PST compounds [6], but convincing structural data are still lacking. Recent studies have shown that some suspected PST-like compounds produced by bacteria are structurally unrelated and have been labelled as ‘imposter’ toxins [58,59].

Bacteria isolated from G. catenatum have previously been shown to autonomously produce compounds with activity similar to PSTs [57]. However, our analysis of nine bacterial strains isolated from typical (GCDE08) and atypical (GCHU11) PST-producing G. catenatum cultures, demonstrated no bacterium produced any compound with PST-like activity. This has also been confirmed during parallel studies of the bacteria isolated from GCDE08 and another G. catenatum strain, GCDE09 (S. Geier and A. Negri, unpublished data). While it is possible that we failed to induce these bacteria to produce PST-like toxins during growth, we did however use growth media and conditions known to elicit autonomous PST production in taxonomically related bacteria isolated from Alexandrium spp. [60]. The absence of autonomous PST production by any of our bacterial isolates, together with reported Mendelian inheritance of PST profiles by Alexandrium spp. [61,62], and G. catenatum ([63]; D. Green, C. Bolch and A. Negri, unpublished data), suggests that PST biosynthesis is associated with the dinoflagellate cell, not its associated bacteria.

While a direct role for bacteria in PST production by G. catenatum seems highly unlikely, previous work shows a strong correlation between G. catenatum cultures that produce low amounts of PST with cultures generated from cysts germinated under laboratory conditions. We hypothesised that specific bacterial types may be removed during sterile washing of the cysts, and that the resulting low PST production levels were mediated through an altered physiological or nutrient status of the algal cells, in the absence of particular bacterial types [21].

Our observations suggest that the bacterial flora may influence the biosynthesis of the PSTs. A number of explanations are possible, for example: a threshold number or critical consortium of Proteobacteria may be required to induce/promote toxin production by G. catenatum, through for example, the supply of sufficient cofactors or precursors necessary for toxin production [11]; or a specific Proteobacteria may be required to trigger post-germination resumption of ‘normal’ PST production of G. catenatum; and alternatively, a particular bacterium (or community mix) may ‘switch-off’ PST biosynthesis in some of our G. catenatum strains.

4.3 Conclusions

This study has shown that the cultivable bacterial flora of each G. catenatum strain is unique at the level of SSU rDNA sequence, yet each of the bacterial communities bares a number of similarities at both the phylogenetic and phenotypic levels, supporting reported studies with other dinoflagellates. Bacteria affiliated to the Rhodobacteraceae and Alteromonadaceae were a consistent feature of G. catenatum and other dinoflagellates such as Alexandrium spp., and these bacteria may have an important role in the growth of dinoflagellates in laboratory culture and natural bloom populations. The lack of PST/PST-like toxin production by bacteria supports the idea that they do not have a direct role in PST production by G. catenatum, however, the reduced bacterial diversity associated with atypical PST-producing G. catenatum strains suggests that key bacteria (or consortia of bacteria) are important for the induction of PST production by G. catenatum following excystment. Overall, we believe that this study points to a number of bacterial groups that are key to the growth and physiology of G. catenatum and that their activity may be fundamental in the development and maintenance of natural blooms. Further work is now underway to characterise their individual and combined influence on the growth and toxicity of G. catenatum, and elucidate the mechanisms by with these effects are mediated.


We are grateful for G. catenatum strain donations from Lincoln MacKenzie, Cawthron Institute (CAWD101) and Mr T.G. Park (YC499B15). D.H.G. is very grateful to the New Zealand Foundation for Research, Science and Technology Postdoctoral Fellowship Scheme for funding this research.


  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]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
View Abstract