Variability of the microcystin synthetase gene cluster in the genus Planktothrix (Oscillatoriales, Cyanobacteria)


  • Edited by K. Forchhammer

*Corresponding author. Tel.: +49 33082 69922; fax: +49 33082 69917, E-mail address:


In populations of Planktothrix, microcystin-producers and non-producers, which are morphologically identical, coexist. In order to develop a basis for the reliable detection of microcystin producers in field samples with polymerase chain reaction (PCR) based methods, we studied the presence and variability of eight regions of the mcy gene cluster in 46 Planktothrix strains, including both microcystin-producing and non-producing ones. PCR-amplification products for two mcy gene regions were also found in non-microcystin-producing strains, indicating the existence of natural mutants. PCR-products of the other regions studied were only detected in microcystin-producing strains. Two of these mcy-amplicons were variable in sequence and length. Four gene regions remained that were conserved and specific for microcystin-producing Planktothrix strains, and thus qualified to detect the respective chemotypes in environmental samples.


Microcystins are hepatotoxic cyanobacterial metabolites, which have been implicated in livestock and human poisoning [1,2]. Microcystins can be synthesized by cyanobacteria of different taxa such as Microcystis[3], Anabaena[4] and Planktothrix[5]. Microcystin-producing and non-producing chemotypes can coexist in populations of cyanobacteria, as it was originally shown for strains of different taxa [5,6]. In Planktothrix these two chemotypes are morphologically identical, and thus microscopically indistinguishable. Therefore, information about the spatial and temporal distribution of both chemotypes is restricted. Knowledge about the occurrence and dynamics of these chemotypes is essential to explain the variability of microcystin concentrations in natural waters as well as to explore the function of microcystin that is still not known.

New possibilities to detect microcystin-producers and to study their distribution emerged since the genes that encode the microcystin synthetase complex have been discovered. These genes are arranged in the mcy gene cluster, encoding non-ribosomal polypeptide synthetases and polyketide synthetases, as well as tailoring enzymes. Complete mcy-sequences are available for M. aeruginosa PCC7806 [7], M. aeruginosa K-139 [8], P. agardhii NIVA-CYA 126/8 [9] and Anabaena sp. strain 90 [10]. Based on that information, primers were designed to amplify parts of mcy genes by PCR as a means of detecting microcystin-producing chemotypes. Microcystin-producing and non-producing strains of the genera Microcystis, Anabaena and Planktothrix could be distinguished with general mcyA-primers [11]. First results of the distribution of microcystin-producing chemotypes in natural populations were obtained, that indicate a strong variability in the relative abundance of these strains. A study about the presence of a mcyB fragment in Microcystis colonies showed that the proportion of microcystin-producing colonies is different in morphospecies: a PCR amplification product of mcyB was detected in 73% of M. aeruginosa, in 16% of M. ichthyoblabe, but not at all in M. wesenbergii[12]. Analysis of mcyE gene copy number with real-time quantitative PCR in samples from Finish lakes revealed that the relative abundance of mcyE-genotypes of Microcystis and Anabaena differed between lakes [14].

These examples demonstrate that PCR based methods are a promising tool to obtain further insights into the distribution of microcystin-producers in cyanobacteria populations. However, such a detection in field samples still runs the risk of false positive or false negative detection for two reasons: (i) DNA polymorphism in mcy genes, as it has been shown for the mcyB gene in Microcystis, where the region coding for the adenylation domain of the first module of mcyB was found to be highly variable [12,15]. (ii) The presence of mcy genes in non-microcystin-producers, as it has been found in M. aeruginosa[8,15–17] and Planktothrix[13]. Therefore, comprehensive information about the variability and presence of mcy genes is of particular importance to design primers for an unambiguous detection of microcystin-producing chemotypes in field samples.

In comparison to Microcystis, our knowledge about the mcy-sequence variability in Planktothrix is still limited. Two species of the genus Planktothrix, formerly Oscillatoria[18], are among the most common microcystin-producing cyanobacteria in temperate latitudes. A lake screening in Germany revealed that cyanobacterial blooms dominated by P. rubescens or P. agardhii have the highest microcystin content per dry weight [19]. The green coloured P. agardhii is primarily distributed in eutrophic polymictic shallow lakes, where mass developments of this species are frequently observed during late summer (e.g. [20,21]). The phycoerythrin containing, red coloured P. rubescens occurs in oligo- to mesotrophic dimictic lakes throughout the year and often stratifies in the metalimnic layer during summer stagnation (e.g. [22]).

The aim of this study was to find a region in the mcy-gene cluster that is suitable for an unambiguous detection of microcystin-producing chemotypes of the genus Planktothrix. Eight regions, specific for Planktothrix, in the mcy-gene cluster were selected. The presence and variability of these gene fragments were investigated in 46 strains, including both microcystin-producing and non-producing strains. The most adequate primer pair was verified using environmental samples.

2Materials and methods

2.1Strains and cultivation

A total of 46 Planktothrix strains, mainly from European freshwaters, were analyzed. They were either isolated during this study or obtained from culture collections (Table 1). The taxonomical classification was made according to Anagnostidis et al. [23] and Suda et al. [24]. Thirty two strains of P. agardhii were investigated, all green coloured, nine were microcystin-producers and 23 were not. Twelve red coloured microcystin-producing strains of P. rubescens were studied. Additionally, one strain of P. mougeotii and one strain of P. pseudagardhii were analyzed, both green coloured and non-microcystin-producing. P. rubescens were grown in Z8-medium [25], P. agardhii, P. mougeotii and P. pseudagardhii in BG11 [26]. The cultures were shaken in Erlenmeyer flasks and illuminated in a 16 h/8 h light–dark cycle with an average light intensity of 20 μmol m−2 s−1 at 20 °C.

Table 1.  Studied strains and results of PCR- and microcystin analysis
StrainGeographic originMicrocystinmcyTmcyTDmcyAmcyEGmcyBmcyEmcyCJmcyHA
  1. Strains provided by (a) NIVA-CYA: Norwegian Institute for Water Research, Oslo/Norway or (b) Andreas Nicklisch, Humboldt-University of Berlin/Germany. Geographical origin (D) Germany, (F) France, (FIN) Finland, (N) Norway, (RUS) Russia, (THA) Thailand. (+) microcystin/PCR-fragment detected. (−) microcystin/PCR-fragment not detected. (L)/(s) PCR-fragment longer/shorter than calculated according to P. agardhii NIVA-CYA 126/8-sequence. (n.d.) not determined.

P. agardhii NIVA-CY A126/8Lake Langsjoen (FIN)a+++++L++++
P. agardhii HUB076/A1River Spree (D)b+++++L++++
P. agardhii LAN2Lake Langer See (D)
P. agardhii MA 9810 G113Lake Maxsee (D)b+++++L++++
P. agardhii Max01Lake Maxsee (D)++
P. agardhii Max02Lake Maxsee (D)+
P. agardhii Max04Lake Maxsee (D)
P. agardhii Max05Lake Maxsee (D)+
P. agardhii Max06Lake Maxsee (D)+++++s++++
P. agardhii Max07Lake Maxsee (D)
P. agardhii Max08Lake Maxsee (D)++
P. agardhii Max09Lake Maxsee (D)++
P. agardhii Max11Lake Maxsee (D)++
P. agardhii Max12Lake Maxsee (D)+++++L++++
P. agardhii Max13Lake Maxsee (D)+++++L++++
P. agardhii Max14Lake Maxsee (D)+++++L++++
P. agardhii Max15Lake Maxsee (D)++
P. agardhii Max19Lake Maxsee (D)+++++L++++
P. agardhii Max21Lake Maxsee (D)++
P. agardhii Max22Lake Maxsee (D)+++++L++++
P. agardhii NIVA-CYA 34Lake Kolbotnvatnet, (N)a+++++s++++
P. agardhii NIVA-CYA12Akerhus (N)a+
P. pseudagardhii NIVA-CYA153River Chao Phya (THA)a
P. agardhii NIVA-CYA21Gulf of Finland (FIN)a++
P. agardhii NIVA-CYA232Lake Rybinsk, Borok (R)a++
P. agardhii NIVA-CYA64/1Lake Helgetjernet, (N)a+++++s++++
P. agardhii PaE1River Erdre (F)+
P. agardhii PaE2River Erdre (F)+
P. agardhii PaE3River Erdre (F)+
P. agardhii PaE4River Erdre (F)+
P. agardhii PaE5River Erdre (F)+
P. agardhii PaE6River Erdre (F)+
P. agardhii PaE7River Erdre (F)+
P. agardhii UH 9810 G105River Unter-Havel (D)
P. mougeotii Max16Lake Maxsee (D)++
P. rubescens BL 9810 G42Lake Breiter Luzin (D)b+++L++s++++
P. rubescens BL 9810 G54Lake Breiter Luzin (D)b++++L++++
P. rubescens BL1Lake Breiter Luzin (D)++++L++++
P. rubescens BL11Lake Breiter Luzin (D)+++++L++++
P. rubescens BL3Lake Breiter Luzin (D)++++L++++
P. rubescens BL9Lake Breiter Luzin (D)+++++s++++
P. rubescens HUB148Lake Arendsee (D)b++++s++++
P. rubescens HUB149Lake Arendsee (D)b++++s++++
P. rubescens HUB150Lake Arendsee (D)b++++L++++
P. rubescens HUB151Lake Stechlinsee (D)b+++n.d.++n.d.n.d.
P. rubescens SL 9810 G34Lake Schmaler Luzin (D)b++++L++++
P. rubescens WAHNDam Wahnbach (D)2+++L+L++++

2.2Environmental samples

Samples were collected from the deep, dimictic, mesotrophic lake Breiter Luzin on October 21st and from the shallow, polymictic, eutrophic lake Stolpsee on September 28th in 2004. For the phytoplankton analysis samples were taken with a plankton net (25 μm mesh size), fixed with formaldehyde (final concentration 4%) and the cyanobacteria composition was determined microscopically. For microcystin and DNA analysis water samples were collected with a 2-l Friedinger water sampler (Limnos) from the metalimnion of lake Breiter Luzin and from the whole water column of lake Stolpsee.

2.3Microcystin determination

Aliquots of 12 or 100 ml from cultures or environmental samples, respectively, were filtered (Whatman GF/C, 25 mm) and loaded filters were placed in 2 ml vials, immediately frozen in liquid nitrogen, lyophilized and stored at −20 °C. Microcystins were extracted 4 times with 1.5 ml 75% methanol [27]. The combined extracts were dried in nitrogen flow and redissolved in 50% methanol prior to analyses by reverse-phase high performance liquid chromatography (HPLC). The HPLC system (all components from Kontron) consisted of two 422 pumps, a M 491 mixer, a 560 auto sampler and a 440 photodiode array detector. Extracts were separated on a LiChrospher 100 ODS column (LiChorCART 250-4, Merck) applying a gradient of aqueous acetonitrile (0.05% trifluoroacetic acid) with a flow of 1 ml min−1[28]. Microcystins were identified by their characteristic UV-spectrum and quantified using Microcystin-LR (University of Dundee) as an external standard. For structure confirmation and elucidation of microcystins, peaks were collected manually and analyzed by matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS) as described previously [29,30]. The detection limit for microcystins with HPLC-analysis was estimated as 10 ng on column. This corresponds to some 5–10 ng per dry weight, depending on the coeluting matrix. Strains that did not contain detectable microcystins were classified as non-producers.

2.4DNA extraction

For extraction of genomic DNA, 50 ml culture of OD750 nm= 0.6 was centrifuged or 100 ml environmental samples were filtered (cellulose acetate filters, pore size 0.45 μm) and DNA was isolated as described elsewhere [31]. DNA was quantified using Eppendorf BioPhotometer (Eppendorf, Hamburg).

2.5Detection of mcy gene fragments

Eight regions of the mcy gene cluster were studied: gene fragments of mcyT, mcyE, mcyA and mcyB and intergenic regions mcyTD, mcyEG, mcyHA and mcyCJ (Fig. 1). Primers were designed based on the mcy gene cluster of P. agardhii NIVA-CYA 126/8 (AJ541056) [9] using Primer3 software ( The primers were synthesized by Metabion (Martinsried, Germany); the sequences are listed in Table 2. PCR was carried out in a final volume of 20 μl containing 1 × buffer (Qiagen, Hilden/Germany), variable MgCl2 concentrations (Table 2), 0.5 μM of each primer, 200 μM of each dNTP, 1 U Qiataq DNA Polymerase (Qiagen, Hilden/Germany) and approximately 3 ng of genomic DNA. Thermal cycling was done in a PTC-200 Peletier thermal cycler (MJ Research, Watertown/USA) using the following program: 10 min 96 °C initial denaturation; 35 cycles of 96 °C for 10 s, variable annealing temperatures for 10 s (Table 2), 72 °C for 30 s; and a final extension step at 72 °C for 10 min. The verification of Planktothrix-specific mcy gene fragments in environmental samples was done with mcyE-PCR analysis.

Figure 1.

mcy-Gene cluster of P. agardhii NIVA-CYA 126/8 [4]. Gene fragments studied here are marked with gray bars. Lengths of genes and bars are proportional to their true length.

Table 2.  Primers used in this study
GenePrimer namePrimer sequence 5′? 3′Tannealing (°C)C(MgCl2) (mM)Length of amplified gene fragment (bp)a
  1. aAccording to mcy-gene sequence of P. agardhii NIVA-CYA 126/8.

  2. bDeviant sequences of some strains in brackets.



Ten microlitres of each PCR reaction were analyzed on 1.5% agarose gels in 0.5 × Tris-acetate-buffer [32]. A low range DNA ladder (Peqlab, Erlangen, Germany) was used as size marker. The gels were stained with ethidium bromide and photographed under UV transillumination.


Sequence analysis was done for gene fragments mcyB (20 strains), mcyE (18 strains and two environmental samples) and mcyEG (16 strains), for both strands in each case. PCR products for sequencing were purified using a Qiaquick PCR purification kit (Qiagen) according to the manufacturer's protocol. DNA was redissolved in 50 μl elution buffer (Qiagen). Four microlitres of this DNA solution were sequenced with BigDye Terminator Cycle Sequencing Ready Reaction Kit 3.1 (Applied Biosystems, Darmstadt) as described in the user guide of the kit. Sequencing was carried out with an ABI Prism 3100-Avant Genetic Analyzer (Applied Biosystems, Darmstadt). The EMBL Accession Nos. for the sequences reported here are AJ854002–854039 and AJ854234–AJ854249.


Results of the occurrence and variability of the mcy-amplicons are presented in Table 1. Amplification products of two of the eight mcy-genes studied were obtained from all microcystin-producing strains but also from non-microcystin-producing strains. These were amplicons of mcyT, which were detected in 19 of 25 non-microcystin producing strains (examples are given in Fig. 2(a)), and of mcyTD, which were detected in nine of 25 non-producing strains. Moreover, elongated PCR products of the studied mcyTD fragment were found for two microcystin-producing strains: P. rubescens BL 9810 G42 and P. rubescens Wahn. These PCR products were approximately 1500 bp instead of 763 bp, according to the sequence information of strain P. agardhii CYA 126/8, indicating a 750 bp insertion within the mcyTD intergenic region (data not shown).

Figure 2.

Ethidium bromide-stained agarose electrophoresis gels showing PCR amplification products for selected strains: (a) mcyT amplification products: Lane 1, P. agardhii HUB076/A1; Lane 2, P. agardhii G105; Lane 3, P. agardhii Max01; Lane 4, P. rubescens BL3. (b) mcyA amplification products: Lane 1, P. agardhii HUB076/A1; Lane 2, P. agardhii G105; Lane 3, P. rubescens BL 9810 G42; Lane 4, P. rubescens BL3. (c) mcyEG amplification products: Lane 1, P. agardhii HUB076/A1; Lane 2, P. agardhii G105; Lane 3, P. agardhii Max12; Lane 4, P. rubescens BL3; Lane 5, P. rubescens BL9. (d) mcyE Amplification products: Lane 1, P. agardhii HUB076/A1; Lane 2, P. agardhii G105; Lane 3, P. rubescens BL3; Lane 4, P. rubescens BL9. Abbreviations: (PA) P. agardhii, (PR) P. rubescens, (NTC) no template control, (M) DNA length marker, (bp) basepairs, (+) microcystin-producing, (−) non-microcystin-producing.

PCR-amplification products of mcyA were not found in any of the non-microcystin-producing strains. However, this gene fragment was not detectable in nine of 12 microcystin-producing P. rubescens strains and the remaining three P. rubescens strains revealed only a very weak mcyA-PCR signal. In contrast to P. rubescens, all microcystin-producing P. agardhii showed PCR amplification products of mcyA (Fig. 2 (b)).

Amplicons of all other investigated regions mcyE, mcyB, mcyEG, mcyHA and mcyCJ were not detected in any of the non-microcystin-producing strains, but detected in all microcystin-producing strains. However, amplicons of this group were variable in their sequence as well as in their length.

PCR amplification products of two different lengths were found for mcyEG. A 775 bp product was obtained from 14 strains, and a 642 bp product was obtained from 7 strains (Fig. 2 (c)). Both, the short and the long fragment were equally distributed in strains of P. agardhii and P. rubescens. Sequencing of the mcyEG amplification products showed that the shorter products had an identical deletion of a 133 bp fragment within the non-coding region between the mcyE and mcyG genes compared to the DNA sequence of strain CYA 126/8. In strains with the longer mcyEG-sequence this 133 bp gene fragment is conserved except for two variable base pairs. (EMBL-database: AJ854234–AJ854249).

The remaining amplicons, mcyE (Fig. 2 (d)), mcyB, mcyHA and mcyCJ were detected in all microcystin-producing Planktothrix-strains with the same length. Nonetheless, sequencing of the mcyB and mcyE PCR products revealed that these amplicons are conserved to a differing degree. There were 19 variable bases pairs detected in the 555 bp mcyB-amplicon, while only one variable base pair was found in the 589 bp long mcyE-amplicon (EMBL-database: AJ854002–AJ854039).

The mcyE-primers were found to be the most suitable for the detection of microcystin-producing Planktothrix strains. Therefore, they were tested on environmental samples of lake Breiter Luzin and lake Stolpsee. The most common cyanobacterial species in Lake Breiter Luzin was P. rubescens, less abundant were Limnothrix spp. and Pseudanabaena spp. The most common cyanobacterial species in Lake Stolpsee was P. agardhii accompanied by Microcystis spp., Anabaena spp., Pseudanabaena spp. and Aphanizomenon spp. Microcystin was detected in both lakes with concentrations of 2.31 μg L−1 in lake Breiter Luzin and 1.81 μg L−1 in lake Stolpsee. The mcyE-amplification product was detectable with PCR analysis in both lakes (Fig. 3). The lengths and sequences of these products were identical with those of Planktothrix.

Figure 3.

mcyE-Amplification products from environmental samples. Lane 1, Lake Stolpsee, Lane 2, Lake Breiter Luzin, Lane 3, no template control, Lane 4, positive control: P. agardhii MA9810 G113, Lane 5, DNA-length marker.


The findings demonstrate that mcy-genes do not occur exclusively in microcystin-producing Planktothrix and that mcy-genes in microcystin-producers can be highly variable.

The occurrence of the mcyT- and mcyTD-amplicons in non-microcystin-producing strains can be explained by two hypotheses. Firstly, these amplicons may belong to other peptide synthesis gene clusters. Different groups of non-ribosomally synthesized peptides are known from Planktothrix (e.g. [33]). There is no final evidence that mcyT is required for microcystin biosynthesis. Its arrangement next to the mcy-genes in Planktothrix[9] could be random, because it is missed at that position in Microcystis[7,8] and Anabaena[10]. Secondly, the presence of mcyT- and mcyTD-amplicons in non-producers may indicate the existence of natural mutants. Indeed the mcy-genes are very ancient genes, and non-microcystin-producing strains probably evolved from microcystin-producing strains by repeated loss of mcy-genes [34]. Consequently, the existence of natural mutants can be assumed. In our work and the study of Kurmayer et al. [13], different kinds and degrees of deletions of mcy-genes in non-microcystin-producing Planktothrix strains were found. Kurmayer et al. [13] detected one P. agardhii that contained a mcyA-gene fragment, but did not produce microcystin. They detected three non-producing strains of P. rubescens that possess the intergenic regions mcyTD, mcyEG, mcyHA and mcyCJ and three further non-microcystin-producing strains with a deletion in mcyH and mcyA. In contrast, we did not find non-microcystin-producing P. rubescens at all. These discrepancies may be due to regional differences in the distribution of certain genotypes.

According to our data, mcyT and mcyTD are inadequate regions for the detection of microcystin-producing Planktothrix in field samples, since they also occur in non-producers. Using these regions would risk overestimating the portion of microcystin-producers.

The absence of the mcyA-amplicon in microcystin-producing P. rubescens is probably caused by sequence variability between P. agardhii and P. rubescens. The mcyA-gene is most likely also present in P. rubescens, because McyA is essential and unique for microcystin biosynthesis due to its involvement in the incorporation of N-methyl-dehydro-alanin at position seven and l-alanin at position one of the cyclic peptide [7]. A different mcyA-fragment is detectable in other microcystin-producing P. rubescens strains. [13]. However, mcyA could be composed of variable and conserved parts. The variability of mcyA may thus reflect the diversity of microcystin isoforms; Planktothrix produces at least 25 different isoforms of microcystin (e.g. [21,33,35]), of which some were preferentially produced by P. agardhii or P. rubescens, respectively. The microcystin patterns of the strains studied here were not analyzed, thus we cannot link the variability in mcyA to a certain microcystin isoform. However, for Microcystis it has been shown that the variability of mcyB is correlated to the microcystin-isoforms produced [15].

Other variations were detected in the intergenic regions mcyEG and mcyTD. Frequent variations in intergenic non-coding regions are not unusual. These regions are also highly variable in the intergenic spacer region (IGS) between the phycocyanin genes cpcA and cpcB[16,36] or the internal transcribed spacer sequence (ITS) between the 16SrDNA and 23SrDNA (e.g. [37]). All short mcyEG-sequences were similar, as well as all long mcyEG-sequences, and both amplicon lengths were found in strains of P. agardhii and of P. rubescens. This suggests that these regions could have a function.

The amplified regions of mcyA, mcyTD and mcyEG are too variable for the identification of microcystin-producing Planktothrix, because they might lead to an underestimation of microcystin-producers.

A reliable identification of microcystin-producing Planktothrix is only possible with those gene regions, whose amplicons were found in all microcystin-producing strains with the same length: mcyB, mcyE, mcyCJ or mcyHA. Though even here, sequencing of mcyB and mcyE showed, that these two amplicons are differently conserved. The mcyB-amplicon showed higher variability compared to the mcyE-amplicon. McyB catalyzes the incorporation of an l-amino acid at position two and of aspartic acid at position three of the microcystin molecule [7]. With the mcyB-primers used here, the respective gene fragment of the first adenylation domain of mcyB was amplified. This domain is responsible for the activation of the variable l-amino acid, which may be reflected in the variable sequence of mcyB. The low variability of the mcyE-amplicon could be explained by the function of McyE in microcystin-biosynthesis. This enzyme, a mixed polyketide/non-ribosomal peptide synthetase, is involved in the assembly of Adda (3-amino-9-methoxy-2,6,8,-trimethyl-10-phenyl-4,6-decadienoic acid), which can be found in all isoforms of microcystin. This qualifies mcyE as very solid molecular marker for microcystin-producing Planktothrix. The verification of the mcyE-primer with environmental samples supports this suggestion. Even though in lake Breiter Luzin and lake Stolpsee other cyanobacteria than Planktothrix occurred, neither a mcyE-fragment of Microcystis or Anabaena was detected, nor an amplification product of other than the microcystin synthetase gene cluster.

The results of our study show that molecular markers for the detection of microcystin-producing cyanobacteria in environmental samples must be checked carefully on a multitude of microcystin-producing as well as non-producing strains. Only thereby can a possible overestimation of microcystin-producers, due to a detection of potential natural mutants or of non-producers with mcy-similar genes, and a possible underestimation, due to the non-detection of microcystin-producers with variable mcy-genes, be avoided to the greatest possible extent. In our study we designed primers that are suitable for the detection of microcystin-producing Planktothrix in environmental samples. We provide a basis for the development of sophisticated methods for the detection of microcystin-producing Planktothrix sp.


Thanks to Andreas Nicklisch for providing strains from his collection, to Brigitte Nixdorf and her group for sampling lake Langer See and to Monika Degebrodt for her technical support. We are grateful to Elke Dittmann and Thomas Börner for productive discussions and critically reading the manuscript. We thank Ellen Roberts for correcting the English version of the script and two anonymous reviewers for helpful comments.