• Toxic cyanobacteria;
  • Cylindrospermopsis raciborskii;
  • STRR sequences;
  • Thailand;
  • Japan;
  • Low-temperature adapted group;
  • Genotypic clusters


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

Cylindrospermopsis raciborskii is a planktonic, nostocalean cyanobacterium, which produces an alkaloid heptatoxin, cylindrospermopsin. We performed morphological observations, 16S rDNA sequence analysis, PCR fingerprint analysis of short tandemly repeated repetitive (STRR) sequences, temperature tolerances and toxin analysis to characterize 24 strains of this toxic cyanobacterium isolated from Thailand and Japan. All strains shared common morphological traits characteristic of C. raciborskii and showed high 16S rDNA sequence similarity, forming a defined cluster together with the reference strains from Australia. In particular, some of the Thai strains shared 99.9% to 100% similarity with the Australian strains. Various combinations of STRR primers revealed different and unique DNA band patterns among strains of C. raciborskii. The phylogenetic tree revealed two main clusters of C. raciborskii strains, the Thai/Japan-Shinobazugaike cluster (cluster I) and the Japan-Gonoike cluster (cluster II). Cluster I was further divided into two subclusters, A (only Thai strains) and B (one Thai strain and the Japan-Shinobazugaike strains). Thus, the results from 16S rDNA and STRR analyses showed no clear geographical distinction between Japanese and Thai strains and between Thai and Australian strains. Thai strains were separated into adaptive and non-adaptive groups to low temperature (15 and 17.5 °C) and Japanese strains were composed of only low-temperature-adaptive ones. The toxin cylindrospermopsin was detected in some strains of cluster I-A and in one strain of cluster II. We conclude that C. raciborskii is a species that has recently begun to invade, and a species with different physiological strains or ecotypes in temperature tolerance; the toxin is synthesized without any relation to phylogenetic or genetic clusters and to geography.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

Cylindrospermopsis raciborskii is a toxic and water-bloom-forming cyanobacterium. It produces an alkaloid hepatotoxin called cylindrospermopsin [1] and neurotoxic saxitoxin analogs [2]. In 1979, an outbreak of hepatoenteritis at Solomon Dam, Palm Island, Queensland, led to the hospitalization of 148 people [3]. Subsequent surveys suggested that C. raciborskii was the most likely causative agent of this outbreak [4,5]. Since this incident, this species has been studied by many investigators working in the fields of environmental and health sciences and water management and has been increasingly reported from water bodies in tropical, subtropical and temperate regions (cf. [6,7]).

The first intensive studies on C. raciborskii were performed on Australian strains. The strain C. raciborskii at Solomon Dam contained two distinct morphotypes of trichome: straight and coiled. Taxonomic and phylogenetic studies on the culture strains isolated in Australia revealed that these morphotypes displayed clear and consistent morphological and physiological differences [8], but all isolates had a high level of similarity of 16S rDNA gene sequences [8] and of rpoC1 gene sequences [9]. It has been suggested that all C. raciborskii isolates from Australia belonged to the same species, including the straight and coiled forms [9]. However, the highly iterative palindrome (HIP1) repeated sequence [10] and short tandemly repeated repetitive (STRR) sequences [9] showed wide variations among Australian isolates.

Padisák [7] summarized the existing knowledge of the taxonomic and morphological aspects, geographical distribution, population dynamics and physiological ecology of C. raciborskii and hypothesized its tropical origin and that its naturally expanding distribution was due to its invasive behavior. He raised the question of whether C. raciborskii is a species that has recently begun to invade and has excellent, multiple, competitive abilities, or whether it simply exists in different physiological strains or ecotypes. A phylogenetic tree constructed from analysis of both 16S rDNA sequences and HIP1 PCR reaction sequences using strains of C. raciborskii isolated from Australia, Brazil, Germany, Hungary, Portugal and the USA revealed three geographically distinct groups: Australian, European, and North/South American phylotypes in which only Australian strains produced cylindrospermopsin [6]. However, in that study, only a few strains were chosen from each country and it is unclear if those were representative of each country. Thus, the poor understanding of genetic and ecological variations of C. raciborskii within a country makes it difficult to formulate an answer to Padisák's hypothesis. In particular, only few biochemical, ecophysiological, or phylogenetic studies have been carried out on Asian strains of C. raciborskii in Asia.

Recently, we found C. raciborskii in freshwater reservoirs, dams and ponds in several parts of Thailand [11,12] and we were able to isolate a number of strains from these places. Cylindrospermopsin and deoxycylindrospermopsin were detected in some strains of C. raciborskii isolated from a reservoir in Thailand [13]. In addition, several strains were isolated from two ponds in Japan, Gonoike pond, in Ibaraki Prefecture, where C. raciborskii was first recorded in 1935 [14], and Shinobazugaike pond, in Ueno Park, Tokyo. Here we conducted morphological, phylogenic, toxin and temperature tolerance studies on 24 strains from Thailand and Japan and compared them with the characteristics of the two Australian strains. Our aim was to characterize these Asian strains of C. raciborskii and to determine what genetic or ecological variations can be recognized among them. Evolutionary relationships among Japanese, Thai and Australian strains are discussed.

2Materials and methods

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

2.1Strains and maintenance conditions

Twenty-four strains of C. raciborskii were newly isolated from lakes, reservoirs and ponds in Thailand and Japan or obtained from the NIES-Collection (Table 1). Strains that were not axenic and/or clonal were purified and cloned by the micropipette-washing method [15]. The purity of the strains was checked with bacterial check media (cf. [15]) and by microscopy. Axenic cultures were maintained in screw-capped test tubes (18 × 150 mm) with 10 ml CT medium [16] and incubated at 25C, under a 12:12 h light/dark cycle at a photon flux density of 40 μmol m−2 s−1, provided by daylight fluorescent lamps (National Co. Ltd., Osaka, Japan).

Table 1.  Strains of Cylindrospermopsis used in this study
StrainaTrichome shapeSampling placeProvince/countryAccession Nos.b
  1. aStrains were deposited in NIES – Microbial Culture Collection, Japan.

  2. bThe nucleotide sequence data reported in this paper appear in the DDBJ/EMBL/GenBank nucleotide sequence databases with the accession number(s) shown in this table.

DMKU51001StraightHuai SuppradooNakhon Ratchasima, Thailandgenbank:AB115466
DMKU51002StraightPond in Public parkTak, Thailandgenbank:AB115467
DMKU51003StraightHuai Tung ThaoChiang Mai, Thailandgenbank:AB115468
DMKU51004CoiledPa Chan ReservoirChiang Rai, Thailandgenbank:AB115469
DMKU51005CoiledShrimp pondChachoengsao, Thailandgenbank:AB115470
DMKU51006StraightShrimp pondChachoengsao, Thailandgenbank:AB115471
DMKU51007CoiledNong Phai KhaewChonburi, Thailandgenbank:AB115472
DMKU51008StraightNong Phai KhaewChonburi, Thailandgenbank:AB115473
DMKU51009CoiledThap Salao DamUthai Thani, Thailandgenbank:AB115474
DMKU51010StraightThap Salao DamUthai Thani, Thailandgenbank:AB115475
DMKU51011StraightKang Kra Chan DamPhetchaburi, Thailandgenbank:AB115476
DMKU51012StraightChong Khao Khat DamUttaradit, Thailandgenbank:AB115477
DMKU51013StraightHuai Pa Daeng ReservoirPhetchabun, Thailandgenbank:AB115478
DMKU51014StraightUbolrat DamKhonkaen, Thailandgenbank:AB115479
DMKU51015StraightNong MekUdon Thani, Thailandgenbank:AB115480
DMKU51016StraightBan Kok KangNong Khai, Thailandgenbank:AB115481
DMKU51017CoiledFish pond at SuwannaphumRoi Et, Thailandgenbank:AB115482
DMKU51018StraightFish pond at SuwannaphumRoi Et, Thailandgenbank:AB115483
DMKU51019CoiledFish pond at SuwannaphumRoi Et, Thailandgenbank:AB115484
CRJ1CoiledShinobazugaike pond in Ueno parkTokyo, Japangenbank:AB115485
CRJ2StraightShinobazugaike pond in Ueno parkTokyo, Japangenbank:AB115486
NIES-991StraightGonoike pondIbaraki, Japangenbank:AB115487
NIES-992StraightGonoike pondIbaraki, Japangenbank:AB115488
NIES-993StraightGonoike pondIbaraki, Japangenbank:AB115489


Each strain of Cylindrospermopsis was cultured in a 125 ml Erlenmeyer flask containing 25 ml of CT medium and incubated under the above-mentioned temperature and light conditions. After four weeks, the morphological characteristics of the strains were observed using a light microscope. Strains were identified by their cell size and shape, and heterocyst and akinete morphologies on the basis of the works by Saker et al. [8], Branco and Senna [17], Komárek and Kling [18] and Watanabe [19]. For the determination of cell dimensions, 50 filaments were randomly selected and the largest and smallest cells of each filament were measured. The mean cell size and standard deviation of 100 vegetative cells, heterocysts and akinetes were measured for each strain.

2.3DNA extraction

The culture conditions and medium employed for DNA extraction were the same as those used for morphological observations. Ten milliliters of Cylindrospermopsis culture was harvested by centrifugation at 5000g at 5 °C for 10 min. Five hundred microliters of 2% cetyl trimethyl ammonium bromide (CTAB) and 5 μl of mercaptoethanol were added, and the mixture incubated at 50 °C for 1 h. Cells were homogenized by adding glass beads (0.1 mm in diameter) and beaten 5 times with a Mini-BeadbeaterTM (3110BX, Biospec Products, Bartlesville, Oklahoma, USA) for 30 s. DNA was extracted with 500 μl of a chloroform: isoamyl alcohol mixture (24:1) followed by centrifugation at 5000g at 5 °C for 2 min. DNA was precipitated and purified with a QIAEX II kit (QIAGEN Sciences, MD, USA).

2.416S rDNA sequence analyses

Amplifications of 16S rDNA were performed using primers 107F and 1497R (Table 2). The PCR mixture (100 μl) consisted of 10–100 ng genomic DNA, 50 pmol primers, 10 × Ex TaqTM (Takara Biomedicals, Otsu, Japan) buffer, 2.5 mM of each deoxynucleoside triphosphate, and 5 units of Takara Ex TaqTM. The reaction was carried out in a thermal cycler (Takara PCR Thermal Cycler Personal, Takara Biomedicals, Otsu, Japan), which was initialized for 1 cycle at 95 °C for 3 min followed by 30 cycles at 94 °C for 10 s, 53 °C for 10 s and 72 °C for 1 min 30 s, terminated by 1 cycle at 72 °C for 8 min, and finally held at 4 °C.

Table 2.  Sequences of PCR primers for 16S rDNA
PrimerSequence (5′–3′)

For 16S rDNA sequence determination, about 0.1 pmol of PCR product, 1.5–2.5 pmol of the 16S rRNA gene-sequencing primers, 107F, 522F, 959R, 882F, 517R, 1185F, 1387R, and 1497R (Table 2) and a DYEnamic ET terminator cycle sequencing kit (Amersham Biosciences, Tokyo, Japan) were used to determine the nucleotide sequence of Cylindrospermopsis. The reaction conditions consisted of 30 cycles at 95 °C for 20 s, 50 °C for 15 s and 60 °C for 1 min, with final holding at 4 °C. Nucleotide sequencing was performed on an automated DNA sequencer (ABI PRISMTM 310 Genetic Analyzer, Applied Biosystems, Perkin–Elmer Corporation, Tokyo, Japan).

2.5Phylogenetic analysis

DNA sequences were aligned with the multiple sequence alignment software, ClustalW1.8 [20]. Ambiguous characters in which a deletion or insertion was recorded from any strains were removed from the aligned data. Phylogenetic and molecular evolutionary analyses were conducted with MEGA version 2.1 [21] and a tree was constructed by the neighbor-joining method [22]. Bootstrap analyses were performed with 1000 replicates. Accession numbers are given in Table 1.

2.6STRR sequence analyses

Primers of STRR sequence were developed from repeated sequences in the genome of the cyanobacteria [23]. The sequences of primers were STRR1F: (5′-CCCCARTCCCCART-3′); STRR1R: (5′-GGGGAYTGGGGAYT-3′); STRR3F: (5′-CAACAGTCAACAGT-3′); and STRR3R: (5′-ACTGTTGACTGTTG-3′). Each 50 μl of PCR reaction mixture consisted of 10–100 ng of genomic DNA, 50 pmol of primers, 10 × Ex TaqTM buffer, 2.5 mM of each deoxynucleoside triphosphate, and 5 units of Takara Ex TaqTM. PCR was performed in a Takara PCR Thermal Cycler programmed for 1 cycle at 94 °C for 10 min; 30 cycles at 94 °C for 30 s, 42 °C for 1 min and 65 °C for 4 min, 1 cycle at 65 °C for 7 min and final holding at 4 °C.

Fingerprint patterns of the STRR sequence were analyzed with Image Master 2D Elite Ver 2.00 software (Amersham Bioscience, Tokyo, Japan). The fingerprint patterns were converted to binary data by scoring the presence or absence of bands for each strain as 1 or 0. These data were used to calculate total character differences, which were used to construct a neighbor-joining tree with PAUP 4.0 β version [24], running on a Macintosh computer.

2.7Temperature experiment

Each strain of Cylindropermopsis reciborskii was cultured in triplicate in screw-capped test tubes (18×150 mm) containing 10 ml of CT medium. The media were inoculated with different (calculated) amounts of cultures in exponential growth phase to obtain similar initial cell concentrations. Incubation was carried out in a growth chamber (Advantec, PEL-2160, Toyo Seisakusho Co., Ltd., Chiba, Japan) controlled at 10, 15, 17.5, 20, 25, 30 and 35 °C with the light intensity of 40 μmol m−2 s−1 and 12:12 h light-dark cycle. Growth was checked at 3-day intervals using a spectrophotometer (Spectronic 20A, Shimadzu, Kyoto, Japan) at the wavelength of 680 nm until 12 days after inoculation.

2.8Toxin analysis

Each strain was cultured in 10 ml of Jaworski's medium [25] for a month. Toxin was extracted with methanol. After removal of the solvent, the residue was re-dispersed in water and filtered with a membrane filter. The solution was then passed through an ODS cartridge (Sep-Pak C18, Waters Corporation, Bedford, MA, USA). The pass-through fraction was collected and concentrated under reduced pressure. Concentrated samples were analyzed by HPLC-MS (Shimadzu, Tokyo, Japan) for cylindrospermopsin using a TSK-gel Amide-80 column (100 × 2 mm) (Tosoh Corporation, Tokyo, Japan) with a linear gradient 100 to 60% acetronitrile/water for 30 min at 0.2 ml min−1. A photodiode array (wavelength 262 nm, SIM) and MS (M + 416 m/z and M − 414 m/z) were used as detectors. Only presence or absence of toxin was determined because we did not exactly measure the cell mass of the culture strains.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References


The strains were separated into two distinct morphotypes, one with straight trichomes and another with coiled trichomes (Fig. 1). Apart from these variations in trichome form, all strains shared common morphological characteristics outlined below, although there were some small variations. Solitary trichomes, both coiled and straight ones, were not constricted or even slightly constricted at the cross walls. Cells with gas vesicles were cylindrical and 1.9–6.3 μm wide and 3.0–18.5 μm long. The ratio of cell length to cell width was 1:1 to 7:1. Terminal cells were cylindrical, conical or rounded. Heterocysts were developed from a terminal cell, being single or in pairs in some strains and at one or both ends of the trichome. Some Thai and Japanese strains rarely developed intercalary heterocysts (Fig. 1C). Terminal heterocysts were never curved, conical, cylindrical or rounded. They were 1.9–7.0 μm wide and 3.3–11.9 μm long, and the ratio of length to width was 1:1 to 5:1. Akinetes were observed in three Thai strains (DMKU51013, DMKU51014, DMKU51016) and in all Japanese strains. They were formed as neighbors to the heterocysts, sometimes two cells apart from the hetereocysts, solitary, in pairs, or, rarely, in series up to three. They were cylindrical, 2.8–7.0 μm wide, and 8.5–17.8 μm long. The ratio of length to width was 3:2 to 6:1. Because the ranges of standard deviation of the cell, heterocyst and akinete sizes of different strains overlapped continuously (data not shown), clear divisions cannot be made among the strains on the basis of their quantitative characters. In conclusion, there were no clear differences in morphology among the strains and the morphological characteristics observed were identical to those of C. raciborskii.


Figure 1. Morphology of C. raciborskii. (A) Strain DMKU51006 with straight trichome; (B) strain DMKU51007 with coiled trichome; and (C) intercalary heterocyst seen in strain DMKU51001.

Download figure to PowerPoint

3.2Phylogenetic analysis of the conserved regions of 16S rDNA

The corrected sequence alignment, which formed the basis of the phylogenetic analyses, corresponded to positions 112–1488 according to the Escherichia coli numbering and was 1376 bases long after gaps, inserts, and ambiguous positions had been removed. All strains of C. raciborskii, including the additional two strains, SDC (genbank:AF067818) and SDS (genbank:AF067819) showed high DNA sequence similarity exceeding 99.5%. A phylogenetic tree was constructed from an alignment of 33 sequences, including 24 strains of C. raciborskii, and two Australian strains of C. raciborskii, SDC (coiled: genbank:AF067818) and SDS (straight: genbank:AF067819) sequenced previously [8] with Anabaena flos-aquae as the outgroup using neighbor-joining method [22] (Fig. 2). All C. raciborskii strains formed a defined cluster, similarly as in the work of Saker and Neilan [10]. Although the topology within the C. raciborskii cluster may not be reliable because the 16S rDNA base differences were so small, it seems that the two Thai strains (DMKU51004 and 51009) and the Australian strains form a single defined cluster. In particular, 16S rDNA sequences of the Thai strain DMKU51004 and of the Australian strain SDS (genbank:AF067819) were identical. The same results were obtained when the neighbor-joining tree was constructed by excluding invariant sites.


Figure 2. Phylogenetic tree of C. raciborskii constructed from 16S rDNA sequences using neighbor-joining method. The scale bar represents five base substitutions per 1000 nucleotide position. Bootstrap percentages calculated from 1000 resamplings are indicated at nodes.

Download figure to PowerPoint

3.3Phylogenetic analysis of STRR sequences

The 24 strains of C. raciborskii were analyzed by the combination of three pairs of primers derived from cyanobacterial repeat sequences, namely STRR1F and STRR3F, STRR1F and STRR3R, and STRR1R and STRR3R. These combinations of primers established the DNA profiles for each isolate (Fig. 3). The sizes of STRR PCR products were approximately 0.3–5.6 kb in Thai and Japan-Shinobazugaike strains and 0.2–6.7 kb in Japan-Gonoike strains (Table 3); these were larger than previously reported in 13 isolates of C. raciborskii from Australia, which had ranged in size from approximately 0.1–2.5 kb [9] (Table 3).


Figure 3. DNA fingerprint of STRR sequences of C. raciborskii. A: combination of primer STRR1F and STRR3F; B: combination of primer STRR1F and STRR3R; C: combination of primer STRR1R and STRR3R. Lanes 1, Marker; 2, DMKU51001; 3, DMKU51002; 4, DMKU51003; 5, DMKU51006; 6, DMKU51008; 7, DMKU51010; 8, DMKU51011; 9, DMKU51012; 10, DMKU51013; 11, DMKU51014; 12, DMKU51015; 13, DMKU51016; 14, DMKU51018; 15, DMKU51004; 16, DMKU51005; 17, DMKU51007; 18, DMKU51009; 19, DMKU51017; 20, DMKU51019; 21, CRJ1; 22, CRJ2; 23, NIES991; 24, NIES992; 25, NIES993.

Download figure to PowerPoint

Table 3.  Size of the PCR products from STRR sequences of C. raciborskii isolated from different countries
Combination of primersApproximate size range (kb)
STRR1F and STRR3F0.3–5.60.3–5.60.2–5.6
STRR1F and STRR3R0.3–5.60.3–5.60.2–5.6
STRR1R and STRR3R0.6–4.80.6–4.80.3–6.7

The various primer combinations presented various DNA band patterns that were distinguishable among the strains of C. raciborskii. Clear differences in banding patterns were observed even between isolates from the same ponds. Furthermore, different banding patterns were obtained among all six strains with coiled trichomes, strains DMKU51004, DMKU51005, DMKU51007, DMKU51009, DMKU51017 and DMKU51019, but this primer combination grouped neither strains with coiled trichomes nor strains with straight trichomes (Figs. 3 and 4). Each primer combination showed that the DNA band patterns of Japan-Gonoike strains (NIES991, NIES992 and NIES993) were significantly different from those of the Thai and Japan-Shinobazugaike strains.


Figure 4. Phylogenetic tree of C. raciborskii constructed from band pattern difference of primer combination of STRR1R and STRR3R using Image Master 2D Elite Ver 2.00 software. Similar phylogenetic trees were obtained when primer combinations of STRR1F-STRR3F and STRR1F-STRR3R were used.

Download figure to PowerPoint

The phylogenetic tree constructed from the binary data of band patterns of STRR sequences demonstrated that the strains of C. raciborskii that we studied were separated into two major clusters, the Thai/Japan-Shinobazugaike strain cluster (cluster I) and the Japan-Gonoike strain cluster (cluster II) (Fig. 5). The Thai/Japan-Shinobazugaike cluster was further separated into two subclusters; subcluster A including only Thai strains and subcluster B including the Japan-Shinobazugaike strains and one Thai strain, DMKU51014.


Figure 5. Phylogenetic tree of C. raciborskii constructed from total character differences of band pattern analysis of STRR sequence by the neighbor-joining method. *, Toxin detected.

Download figure to PowerPoint

3.4Growth at different temperatures

None of the strains could survive at 10 °C. All of the Thai strains grew in the range of 20–35 °C with optimal growth at 30 and 35 °C. They were separated into a low-temperature adapted group (DNKU5006, 5008–50019) showing a moderate or good growth at 15 and/or 17.5 °C and a low-temperature non-adapted group (DMKU50001–50005, 50007) showing no or a little growth at the above temperatures. The low-temperature adapted group had a 3-day lag phase at 15 and 17.5 °C. All of the Japanese strains grew at a wide range of temperatures of 15–35 °C with optimal growth at 30–35 °C, showing the characteristics of the low-temperature adapted group, but they never showed a lag phase even at 15 and 17.5 °C. These results showed that some of the Thai strains could be adapted to lower temperatures at 15–20 °C with acclimatization periods, and the Japanese strains are completely adapted to these low temperatures. Interestingly, a Thai strain, DMKU51014, which forms cluster I-B with Japan-Shinobazugaike strains, was assigned to the low-temperature adapted group like the Japanese strains (Table 4). We found that cluster I-A was composed of both low-temperature adapted and non-adapted strains and cluster I-B and cluster II were composed of only low-temperature adapted strains.

Table 4.  Growth of C. raciborskii isolated from Thailand and Japan after 12 days of incubation at different temperatures
Strain No.Temperature (°C)
  1. −, No growth (OD < 0.01); ±, little growth (OD 0.01–0.05); +, moderate growth (OD 0.051–0.1); ++, good growth (OD 0.101–0.2); +++, excellent growth (OD > 0.2).

DMKU50005++ ++++

3.5Toxin production

The alkaloid hepatotoxin, cylindrospermopsin, was detected in some Thai strains of cluster I-A but not in strains of cluster I-B and only in one strain of cluster II (Fig. 5). From the trees reconstructed from the binary data of band patterns of STRR (Fig. 5) we found that toxin production was polyphyletic in distribution.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

All studied strains shared common morphological traits mostly corresponding to the description of C. raciborskii and formed a defined phylogenetic cluster together with two Australian strains of C. raciborskii SDC (genbank:AF067818) and SDS (genbank:AF067819), the morphological and growth characteristics of which had already been reported. Two morphologies, one with straight trichomes and the other with coiled trichomes have been observed in C. raciborskii[8]. Taxonomic and phylogenetic studies of the culture strains of these morphotypes have suggested that all of the C. raciborskii isolates from Australia belonged to the same species, which includes both straight and coiled forms [8–10]. The C. raciborskii strains studied here also formed a defined cluster together with two strains of C. raciborskii used by Saker et al. [8], in agreement with the result of Saker and Neilan [10]. Although STRR sequence analyses of 13 Australian strains indicated that Australian isolates with coiled trichomes could be grouped together, suggesting some linkage between trichome coiling and genotype [9], our STRR sequence analyses of Thai and Japanese strains clearly showed that the strains with coiled trichomes were not clustered together (Fig. 4) and no clear distinction was recognized between coiled and straight forms.

The phylogenetic tree constructed from the binary data of band patterns of STRR showed that C. raciborskii is composed of at least two genotype clusters. Cluster I is composed of strains from tropical or subtropical regions of Thailand and from Shinobazugaike pond in Japan, whose PCR products of STRR sequences are also similar in size. Cluster II is composed of strains from Gonoike pond in Japan that have wider size range of PCR products. From 16S rDNA sequence analysis, it seemed that two Thai strains (DMKU51004 and 51009) of cluster I-A formed a single cluster with two Australian strains (SDC and SDS). These four strains shared 99.9%–100% similarity of 16S rDNA sequences. It is likely that cluster I-A is more closely related to the Australian strains than cluster I-B and cluster II. These results showed that there was no clear geographical difference in the DNA banding patterns of strains between Japan and Thailand and between Thailand and Australia. Neilan et al. [6] proposed that C. raciborskii strains could be placed into different geographic clusters with higher between-country variation than within-country variation. However, in their study, the numbers of strains chosen from each country were too low to compare and evaluate the degrees of the variations. Also, it is unclear if the chosen strains were representative of each country. Our phylogenetic study using 19 Thai strains and five Japanese strains demonstrated that the recognized clusters did not coincide with geography. To obtain clear relationship between genetic variation and geography, much more strains should be isolated from various lakes and ponds in various countries and investigated on their genetic variations.

Usually, C. raciborskii populations grow and increase only in warm (>25 °C) water [7]. The culture experiments using Australian strains, SDC (genbank:AF067818) and SDS (genbank:AF067819), revealed that growth of C. raciborskii was most rapid for both forms at 28 °C but diminished at 21 °C [8]. All Japanese strains and some Thai strains, despite exhibiting optimal growth at high temperatures of 30 and 35 °C, like the Australian strains, showed moderate or good growth at low temperatures of 15 and/or 17.5 °C. The Thai strains can be separated into low-temperature adapted and non-adapted groups, but this grouping did not coincide with ecological or geographical differences in the habitats where the strains have been isolated. Even the strains isolated from the same pond, such as DMKU 51005 and 51006 from Shrimp pond, Chachoengsao, and DMKU 51007 and 51008 from Nong Phai Khaew, Chonburi, were separated into the two groups, respectively (Tables 1 and 4). This suggests that any of the Thai population of C. raciborskii could be adapted to temperature fluctuation by having physiologically different strains. On the other hand, the Japanese population of C. raciborskii would potentially be adapted to seasonal temperature fluctuations peculiar to temperate regions. This was demonstrated by seasonal dynamics of C. raciborskii in Gonoike pond in Japan as mentioned below. C. raciborskii was first recorded in Japan from Gonoike pond in 1935, only 23 years after the species was first described in the world. In this pond, C. raciborskii usually appears from May (water temperature of ca. 20 °C), increases its biomass from July (water temperature of more than 25 °C) to the end of October (water temperatures of ca. 15–17 °C) and decreases its biomass in December (temperatures of 8–10 °C) (data not shown). The seasonal dynamics of C. raciborskii in Gonoike pond coincide well with the results of temperature experiments on the growth of Japan-Gonoike strains. Regarding the relationship between genetic clusters and temperature tolerances, cluster I-A is composed of low-temperature adapted and non-adapted strains from Thailand with some strains closely related to the Australian strains, cluster I-B composed of Japan-Shinobazugaike strains and one Thai strains, both adaptive to low temperatures, and cluster II composed of the low-temperature adapted Japan-Gonoike strains. Padisák [7] summarized the existing knowledge of the taxonomy, morphology, geography, and ecology of C. raciborskii and hypothesized its tropical origin and that the naturally expanding distribution of C. raciborskii was caused by its invasive behavior. However, he has never excluded the gradual evolutionary adaptation to colder temperatures, because the maximum biomass development of C. raciborskii coincided with the temperature of only 15–18 °C in Austria [26]. The existence of strains or population of C. raciborskii adapted to low temperature strongly suggested that C. raciborskii was not only a species that has recently begun to invade but also a species with different physiological strains or ecotypes in temperature tolerance.

Padisák [7] also speculated that, because of the high-temperature demand of C. raciborskii and its inability to adapt to temperature fluctuations, it evolved in tropical lakes, with Africa as possible primary evolutionary center where the highest diversity of Cylindrospermopsis genus was found. In addition, to explain the ongoing invasion of C. raciborskii both in tropical, subtropical and temperate region, he hypothesized the existence of a secondary evolutionary center because superior shade tolerance of C. raciborskii is not likely to be selected in tropical region. The supposed secondary evolutionary center could be placed in Australia because there are not only the early epidemiological records suggesting the originally widespread occurrence of this species but as well the hydrological circumstances and its primarily subtropical position [7]. Our findings do not allow for any conclusion about the evolutionary dynamics of C. raciborskii as speculated by Padisák [7], however, his view is worthy to be adopted as a working hypothesis of evolutionary dynamics of C. raciborskii. Considering our results based on the working hypothesis, it is possible that Thai strains were derived from the Australian strains (high similarity of 16S rDNA sequences), the Japan-Shinobazugaike strains were further originated from the Thai strains with low-temperature tolerance, and Japan-Gonoike strains were differentiated from cluster I-B by gaining evolutionary adaptation to particular environments of Gonoike pond in Japan and formed an independent genetic cluster, cluster II.

Neilan et al. [6] found geographic segregation of toxin production with cylindrospermopsins being produced only by Australian strains. However, the present study showed that the toxic compound cylindrospermopsin was detected in some Thai strains of cluster I-A, and in one Japanese strain of cluster II (Fig. 5). A low number of toxic Japanese strains may have resulted from the limited number of studied Japanese strains. In fact, large amounts of cylindrospermopsin were detected in some other strains isolated from Gonoike pond and Shinobazugaike pond in a different year (Sano, personal communication). As shown in the trees reconstructed from the binary data of band patterns of STRR in Fig. 5, toxin production had polyphyletic distribution, which is suggesting horizontal transfer of toxin gene. We suggest that there are toxic and non-toxic strains of C. raciborskii that have no relation to phylogenetic or genetic clusters recognized on the basis of sequences of STRR and to geographic origin of the strains that these and the toxin gene may experience.

In conclusion, C. raciborskii is not only an ongoing invasive species but also a species with different physiological strains or ecotypes in temperature tolerance. Synthesis of the toxic compound, cylindrospermopsin, is possibly a result of dynamic genetic and evolutionary processes, such as a horizontal transfer. To obtain evidence to support these speculations, we will need to perform further genetic, biochemical and ecophysiological investigations on culture strains isolated from various areas of the world.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

We thank the Department of Microbiology, Faculty of Science, Kasetsart University, Thailand, and the Environmental Biology Division, National Institute for Environmental Studies, Japan, for providing research facilities. Financial support was supplied by the Thailand Research Fund for the Royal Golden Jubilee PhD research assistant fellowship.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References
  • [1]
    Ohtani, I., Moore, R.E., Runnegar, M.T.C. (1992) Cylindrospermopsin: a potent hepatotoxin from the blue-green Cylindrospermopsis raciborskii. J. Am. Chem. Soc. 114, 79417942.
  • [2]
    Lagos, N., Onodera, H., Zagatto, P.A., Andrinolo, D., Azevedo, S.M.F.Q., Oshima, Y. (1999) The first evidence of paralytic shellfish toxins in the freshwater cyanobacterium Cylindrospermopsis raciborskii, isolated from Brazil. Toxicon 37, 13591373.
  • [3]
    Byth, S. (1980) Palm Island mystery disease. Med. J. Aust. 2, 4042.
  • [4]
    Bourke, A.T.C., Hawes, R.B., Neilson, A., Stallman, N.D. (1983) An outbreak of hepatoenteritis (the Palm Island mystery disease) possibly caused by algal intoxication. Toxicon 3, 4548.
  • [5]
    Hawkins, P.R., Runnegar, M.T.C., Jackson, A.R.B., Falconer, I.R. (1985) Severe hepatotoxicity caused by the tropical cyanobacterium (blue-green algae) Cylindrospermopsis raciborskii (Woloszynska) Seenaya and Subba Raju isolated from a domestic water supply reservoir. Appl. Environ. Microbiol. 50, 12921295.
  • [6]
    Neilan, B.A., Saker, M.L., Fastner, J., Törökné§, A., Burns, B.P. (2003) Phylogeography of the invasive cyanobacterium Cylindrospermopsis raciborskii. Mol. Ecol. 12, 133140.
  • [7]
    Padisák, J. (1997) Cylindrospermopsis raciborskii (Wolozynska) Seenaya et Subba Raju, an expanding, highly adaptive cyanobacterium: worldwide distribution and review of its ecology. Arch. Hydrobiol./Suppl. 107, 563593.
  • [8]
    Saker, M.L., Neilan, B.A., Griffiths, D.J. (1999) Two morphological forms of Cylindrospermopsis raciborskii (cyanobacteria) isolated from Solomon Dam, Palm Island, Queensland. J. Phycol. 35, 599606.
  • [9]
    Wilson, K.M., Schembri, A., Baker, P.D., Saint, C.P. (2000) Molecular characterization of the toxic cyanobacterium Cylindrospermopsis raciborskii and design of a species-specific PCR. Appl. Environ. Microbiol. 66, 332338.
  • [10]
    Saker, M.L., Neilan, B.A. (2001) Varied diazotrophies, morphologies and toxicities of genetically similar isolates of Cylindrospermopsis raciborskii (Nostocales, Cyanophyceae) from Northern Australia. Appl. Environ. Microbiol. 67, 18391845.
  • [11]
    Chonudomkul, D., Yongmanitchai, W., Theeragool, G. and Watanabe, M.M. (2000). The occurrence of a toxic cyanobacteria Cylindrospermopsis spp. in Thailand. In: RGJ-PhD Congress I. May 2–4, 2000, (p. 129). Felix Hotel, Kanchana Buri, Thailand.
  • [12]
    Yongmanitchai, W., Chonudomkul, D., Intachote, P. and Noparatraraporn, N. (1999). Diversity of toxic cyanobacteria in water reservoirs of Thailand. In: International Conference on Asian Network on Microbial Research. 29 November–1 December, 1999, (pp. 689–696). Chiang Mai, Thailand.
  • [13]
    Li, R., Carmichael, W.W., Brittain, S., Eaglesham, G.K., Shaw, G.R., Mahakhant, A., Noparatnaraporn, N., Yongmanitchai, W., Kaya, K., Watanabe, M.M. (2001) Isolation and identification of the cyanotoxin cylindrospermopsin and deoxy-cylindrospermopsin from a Thailand strain of Cylindrospermopsis raciborskii (cyanobacteria). Toxicon 39, 937980.
  • [14]
    Negoro, K. Observations on the water-blooms, caused by Anabaenopsis raciborskii Woloszinska, in Lake Go-no-ike, Ibaraki Prefecture, Japan. Jpn. J. Limnol. 5. (1935) 148–162(in Japanese).
  • [15]
    Hoshaw, R., Rosowski, J.R. (1973) Methods for microscopic algae. In: Handbook of Phycological methods. Culture Methods and Growth Measurements (Stein, R., Ed.), pp.53–68 Cambridge University Press, London..
  • [16]
    Watanabe, M.M., Kawachi, M., Hiroki, M., Kasai, F. NIES-Collection, List of strains Microalgae and Protozoa. 6th ed. (2000) National Institute for Environmental Studies, Tsukuba, Japan. 159 pp.
  • [17]
    Branco, C.W.C., Senna, P.A.C. (1991) The taxonomic elucidation of the Pananoa Lake (Brasilia, Brazil) problem: Cylindrospermopsis raciborskii. Bull. Jard. Bot. Belg. 61, 8591.
  • [18]
    Komárek, J., Kling, H. (1991) Variation in six planktonic cyanophyte genera in Lake Victoria (East Africa). Algol. Stud. 61, 2145.
  • [19]
    Watanabe, M. (1995) Studies on planktonic blue-green algae 5. A new species of Cylindrospermopsis (Nostocaceae) from Japan. Bull. Natl. Sci. Mus., Tokyo, Ser. B. 21, 4548.
  • [20]
    Thompson, J.D., Higgins, D.G., Gibson, T.J. (1994) Clustal W: improving the sensitivity of progressive multiple sequence alignment through sequence weighing, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 46734680.
  • [21]
    Kumar, S., Tamura, K., Jakobsen, I.B., Nei, M. (2001) MEGA2: Molecular Evolutionary Genetics Analysis Software. Arizona State University, Tempe, AZ..
  • [22]
    Saitou, N., Nei, M. (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406425.
  • [23]
    Mazel, D., Houmard, J., Casters, A.M., Tandeau de Marsac, N. (1990) Highly repetitive sequences in cyanobacterial genomes. J. Bacteriol. 172, 27552761.
  • [24]
    Swofford, D.L. (2002) PAUP*. Phylogenetic Analysis Using Parsimony (* and Other Methods). Version4. Sinauer Associates, Sunderland, MA..
  • [25]
    Thompson, A.S., Rhodes, J.C., Pettman, I. Natural Environmental Research Council Culture Collection of Algae and Protozoa-Catalogue of Strains. (1988) Freshwater Biology Association, Ambleside, UK. p. 22.
  • [26]
    Dokulii, M.T., Myer, J. (1996) Population dynamics and photosynthetic rates of a CylindrospermopsisLimnothrix association in a highly eutrophic urban lake, Alte Donau, Austria. Algol. Stud. 83, 179195.