Abstract Microcystins are small hepatotoxic peptides pro-duced by a number of cyanobacteria. They are synthesizednon-ribosomally by multifunctional enzyme complex syn-thetases encoded by the mcy genes. Primers deduced frommcy genes were designed to discriminate between toxicmicrocystin-producing strains and non-toxic strains. Thus,PCR-mediated detection of mcy genes could be a simpleand efficient means to identify potentially harmful geno-types among cyanobacterial populations in bodies of wa-ter. We surveyed the distribution of the mcyB gene in dif-ferent Microcystis strains isolated from Chinese bodies ofwater and confirmed that PCR can be reliably used toidentify toxic strains. By omitting any DNA purificationsteps, the modified PCR protocol can greatly simplify theprocess. Cyanobacterial cells enriched from cultures, fieldsamples, or even sediment samples could be used in thePCR assay. This method proved sensitive enough to detectmcyB genes in samples with less than 2,000 Microcystiscells per ml. Its accuracy, specificity and applicabilitywere confirmed by sequencing selected DNA amplicons,as well as by HPLC, ELISA and mouse bioassay as con-trols for toxin production of every strain used.

Keywords Cyanobacteria · Microcystis · Microcystin ·Toxin · Peptide synthetase · Whole-cell PCR

Introduction

In eutrophic water habitats, the proliferation of cyanobac-teria results in the formation of water blooms, a world-wide phenomenon that causes the poisoning of animals.Most bloom-forming cyanobacteria are known to synthe-size a variety of toxins. Among the cyanotoxins is a fam-ily of cyclic heptapeptides called microcystins which arepotent hepatotoxins and liver tumor promoters (Car-michael 1996; Sivonen, 1996). The most common produc-ers of microcystins are various Microcystis species (Car-michael 1996; Sivonen, 1996). These species often form aheavy bloom on water surface of lakes or reservoirs. Theyare also abundant in the sediment of lakes and reservoirsthroughout the winter. These overwintering colonies mayprovide the inoculum for the next bloom (Noriko et al.1984; Preston and Stewart 1980). Bodies of water that areused for recreational purposes and/or as the source ofdrinking water should thus be regularly controlled for thepresence of toxic cyanobacteria.

The majority of Microcystis aeruginosa isolates aretoxic and synthesize microcystins. Only toxic isolatescontain the genes for the enzymes involved in microcystinbiosynthesis (Dittmann et al. 1999; Meissner et al. 1996;Neilan et al. 1999). Microcystins are synthesized nonribo-somally by the large modular multifunctional enzymecomplexes known as peptide synthetases encoded by themcy (microcystin synthetase) gene cluster (Dittmann et al.1999; Tillett et al. 2000; Nishizawa et al. 1999, 2000).Amplification of mcy genes by PCR from DNA isolatedfrom axenic cultures and field samples has proven to be asensitive means to differentiate between microcystin-pro-ducing hepatotoxic and non-hepatotoxic Microcystis strains(Dittmann et al. 1999; Meissner et al. 1996; Neilan et al.1999). Two pairs of oligonucleotide primers, TOX1P/TOX1 M and TOX2P/TOX2 M, were designed to detectand characterize microcystin-producing and non-toxiccyanobacteria species by specifically amplifying frag-ments of the mcyB gene (Dittmann et al. 1999). ThesePCR-based gene detection procedures established a corre-

Hui Pan · Lirong Song · Yongding Liu · Thomas Börner

Detection of hepatotoxic Microcystis strains by PCR with intact cells from both culture and environmental samples

Received: 21 September 2001 / Revised: 25 June 2002 / Accepted: 1 July 2002 / Published online: 27 August 2002

ORIGINAL PAPER

H. Pan · L. Song (✉) · Y. LiuDepartment of Phycology, Institute of Hydrobiology, The Chinese Academy of Sciences, Luojiashan, Wuhan 430072, P. R. Chinae-mail: lrsong@ihb.ac.cn, Tel.: +86-27-87647715, Fax: +86-27-87647715

T. BörnerInstitute for Biology, Humboldt University, Chausseestrasse 117, 10115 Berlin, Germany

Arch Microbiol (2002) 178 :421–427DOI 10.1007/s00203-002-0464-9

© Springer-Verlag 2002

lation between the existence of mcyB gene and the pro-duction of microcystins.

Cyanobacteria are gram-negative bacteria with a thickpeptidoglycan layer that makes them resistant to routinecell disruption methods. Thus, the available cell disrup-tion protocols may decrease the efficiency of cyanobacte-rial DNA extraction and inhibit PCR (Howitt 1996; Sid-ney et al. 1988). Here we describe a method using intactcyanobacteria cells for PCR amplification in order toidentify toxic and non-toxic strains of Microcystis. Thismethod is a simple, efficient, and reliable way to identifythe toxic cells from axenic strains, isolates, and lyo-philized samples as well as from water and sediment sam-ples.

Materials and methods

Cyanobacterial strains, isolates, cultivation and lyophilization

Axenic cyanobacteria strains were obtained from the Culture Col-lections of the Freshwater Algae of the Institute of Hydrobiology(FACHB; Wuhan, China), the National Institute for Environmen-tal Studies (NIES; Tsukuba, Japan), the Culture Collection of Al-gae at the University of Texas (UTEX; Austin, USA), and the In-stitute Pasteur (PCC; Paris, France). Cyanobacteria were culturedin liquid MA (Ichimura 1979) medium at 25±2°C under continu-ous illumination of 25 µmol photons m–2 s–1. In addition, 16 Mi-crocystis isolates including M. aeruginosa, M. viridis, and Micro-cystis sp. were sampled from Dianchi Lake (Kunming, China) in2000 and cultured under conditions as for the axenic strains. Thecells of the following axenic Microcystis strains were also investi-gated after lyophilization and storage at –20°C: M. aeruginosa8641, M. wesenbergii 574, M. aeruginosa PCC 7820, M. viridis(Dian-2), M. aeruginosa (Dian-1), and Oscillatoria raciborskii(Daye-1).

Environmental samples

Cyanobacterial cells from water blooms were collected from TaihuLake, Wuxi, in December 1999; from East Lake, Wuhan, in Sep-tember 2000; and from 20 sampling spots of Big Bay and SmallBay of Dianchi Lake, Kunming, monthly from March to December2000. The top-layer (about 10 cm) sediment samples were takenfrom Dianchi Lake, Kunming, and from Guanqiao pool, Wuhan, inDecember 2000. These samples contained single cells and smallcolonies of Microcystis, mainly M. wesenbergii and M. aerugi-nosa, as observed with an inverted light microscope (OPTONIM35, Germany). All environmental samples were stored at 4 °Cfor no more than 1 week.

Toxicity measurement and analysis of microcystins

To detect microcystins, mouse bioassay, high-pressure liquid chro-matography (HPLC), and an indirect competitive enzyme-linkedimmunosorbent assay (ELISA) were used according to publishedprotocols. In the mouse bioassay, aqueous cell suspensions wereinjected intraperitonealy, the minimum lethal dose was measured,and hepatic hemorrhage was observed histologically (Song et al.1999). In HPLC analyses, samples were extracted with 5% aceticacid while being emulsified by ultrasonication After centrifuga-tion(1,250×g, 15 min), the pellets were extracted three times with100% methanol and the methanolic extracts were dried in a vac-uum evaporator. The residues were dissolved in 5% acetic acid andpassed through conditioned (10 ml 100% methanol, 50 ml 100%distilled water) Sep-Pak C18 cartridges (Waters). The cartridgeswere then rinsed with 20% methanol and the samples eluted with

90% methanol. After drying, the eluates were dissolved in 20%methanol, and the fractions were subjected to reversed-phase HPLCcolumn (Shimadzu LC-10A) analysis. ELISA was conducted withmonoclonal antibodies kindly provided by Y. Ueno (Yashio Insti-tute of Environmental Sciences). To estimate the total microcystincontent, samples were treated twice by freeze-thawing followed byfiltration over glass filters (Whatman GF/C, 25 mm in diameter).Microcystin-LR-bovine serum albumin conjugate was coated on amicrotiter plate, which was subsequently incubated with standardmicrocystin-LR or samples and anti-microcystin-LR monoclonalantibody. The amount of antibody bound on the surface of thewells was determined by the reaction of peroxidase-labeled goatanti-mouse IgG (TAGO 4550) with its substrate (0.1 mg 3,3,5,5,-tetramethylbenzidine ml–1, 0.005% H2 O2 in 0.1 M acetate buffer,pH 5). Microcystin concentrations were determined from the stan-dard curve of microcystin-LR (also see Ueno et al. 1996).

DNA extraction

Total genomic DNA was prepared by a commonly used methodfor cyanobacteria (Sidney and Kaplan 1988).

Pretreatments of samples for whole-cell PCR

Cyanobacteria for the whole-cell PCR were collected from the en-vironmental water samples by centrifugation. Before resuspensionin distilled water to a defined volume, the cells were washed one tothree times with distilled water. Sediment sample (about 100 g)was mixed with 500 ml distilled water in a beaker. Cells floatingon the surface were collected and treated as described for watersamples. The floating cells were also prepared by adding 40 mlsterilized MA medium slowly to a glass tube containing about 10 gsediment sample. Both water and sediment samples were also di-rectly used in PCR without any pretreatment.

PCR amplification

PCR was performed in a GeneAmp2400 thermocycler (Perkin-Elmer Cetus, Emeryville, Calif., USA). The thermal cycling proto-col included an initial denaturation at 94°C for 2 min, followed by35 cycles. Each cycle began with 10 s at 93°C followed by 20 s atthe annealing temperature at 50, 55 or 52°C for primers MTF2/MTR2, TOX1P/TOX1 M or TOX2P/TOX2M, respectively, andended with 1 min at 72°C (Neilan et al. 1999). When extractedDNA was used, the amplification reactions contained a 10×ampli-fication buffer with 1.5 mm MgCl2, 0.2 mm dNTPs, 20 pmol ofeach primer and 1 U Taq DNA polymerase, and 3–5 ng purifiedDNA in a final volume of 50 µl (Dittmann et al. 1999). The PCRamplification with whole cells started with 6 µl of crude sample,pretreated subsample with an approximate cell density of 8×106 cells/ml, or 0.1 µg lyophilized cyanobacterial cells. The samplewas added directly to a 20-µl reaction solution containing bovineserum albumin (0.1 mg/ml) or skim milk (0.1 –100 mg/ml, w/v),and a 10×amplification buffer, which contained 1.5 mM MgCl2,0.2 mM dNTPs, 20 pmol of each primer, and 0.5 U Taq DNApolymerase (Howitt 1996). The PCR amplifications conditionswere identical to those for the samples described above. An extraramp rate of 3 s/°C between the denaturing and annealing stepswas set when a GeneAmp9600 cycler instead of GeneAmp2400was used for PCR amplification. The dosage for the skim milkranging from 1 to 100 mg/ml was determined to be appropriatebased on the results of PCR.

DNA sequencing

DNA was sequenced automatically on a PE ABI 377 sequencer(Perkin-Elmer, Foster City, Calif.) using cyclic sequencing with adye terminator. PCR fragments were sequenced using 300 pmolprimers and automated protocols. The similarities between the

422

mcyB genes in GenBank and the PCR products were analyzed withNCBI sequence similarity search tool (BLAST 2.1) and with mul-tiple-sequence cluster alignment software (DNATools 5.1, S.W.Rasmussen, Carlsbad Laboratory, Copenhagen).

Results and discussion

Comparison of PCR using extracted DNA and whole cells from axenic strains

The primer pair of MTF2/MTR2 was previously used toamplify nonspecifically fragments of all the genes encod-ing the peptide synthetases in cyanobacteria, allowing the

detection of strains synthesizing small peptides nonribo-somally (Dittmann et al. 1999; Meissner et al. 1996; Neilanet al. 1999). By using the primer pairs TOX1P/TOX1 Mand TOX2P/TOX2 M (Dittmann et al. 1999), it is possibleto simply but reliably detect mcyB genes, which are in-volved in the biosynthesis of microcystins in the Micro-cystis strains and, to an unknown extent, other cyanobac-terial genera . Using one of the three primer pairs men-tioned above, both the extracted DNA and whole cellsfrom 38 axenic cyanobacterial strains belonging to differ-ent species were analyzed by PCR (Table 1). The samestrains were checked with all three of the methods de-scribed in Materials and methods for the presence of toxinand toxicity. In most cases, all three methods delivered

423

Strains Whole-cell PCR HPLC Bioassay ELISA

Primers Primers Primers MTF/MTR TOX1P/1M TOX2P/2M

Microcystis aeruginosa FACHB (Wuda’erchi) + + + + + +Microcystis aeruginosa FACHB (Dianshanhu) + + + + + +Microcystis aeruginosa FACHB (Bao’anhu) + + + + + +Microcystis aeruginosa NIES-101 + – – – / –Microcystis aeruginosa NIES- 90 + + + + + +Microcystis aeruginosa FACHB8641 + + + + + +Microcystis sp. FACHB10 + + + + + +Microcystis aeruginosa FACHB86 + + + + + +Microcystis aeruginosa PCC 7806 + + + + / +Microcystis viridis FACHB (Dianchi-1) + + + + + +Microcystis aeruginosa UTEX2061 + – – – – –Microcystis sp. FACHB20 + – – – – –Microcystis wesenbergii FACHB574 + + + – + –Microcystis aeruginosa FACHB (Dianchi-1) + + + + + +Microcystis aeruginosa FACHB (Dianchi-3) + / + + + +Microcystis viridis FACHB (Dianchi-2) + / + + + +Microcystis aeruginosa FACHB (Wuda’erquan) + / + + + +Microcystis aeruginosa PCC 7820 + + + / / +Microcystis aeruginosa NIES-98 + – – – / –Microcystis sp. FACHB434 + / + / / +Microcystis sp. FACHB573 + / – / / –Microcystis sp. FACHB572 + / – / / –Microcystis sp. FACHB525 + / + / / +Microcystis sp. FACHB526 + / – / / –Microcystis sp. FACHB569 + / + / / +Microcystis sp. FACHB469 + / – / / –Microcystis sp. FACHB502 + / – / / –Microcystis sp. FACHB575 + / – / / –Microcystis elabens NIES-42 + – – / / –Microcystis holsatica NIES-43 + – – / / –Oscillatoria rubescens NIES-610 / / – + / +Oscillatoria agardhii NIES- 595 – – – + / +Oscillatoria raciborskii FACHB (Daye-1) – – – / – –Oscillatoria planctonica UTEX708 – – – / – –Anabaena flos-aquae FACHB1444 + – – / – –Anabaena sp. PCC 7120 – – – / – –Aphanizomeon flos-aquae FACHB 44–1 + – – / – –

Table 1 Comparison of PCR, HPLC, bioassay and ELISA in de-termination of toxicity of different cyanobacteria. + Toxic (forHPLC, bioassay and ELISA) or specific band amplified (for PCR);

– non-toxic (for HPLC, bioassay and ELISA) or no specific bandamplified (for PCR); / not assayed

consistent results. In a few cases, however, ELISA wasthe most sensitive of the three methods (although it occa-sionally gave false-positive results). The PCR results agreedwell with those obtained by HPLC, ELISA and mousebioassay, except for M. wesenbergii 574, O. rubescens610 and O. agardhii 595, in which the PCR results differedwith those obtained using HPLC and ELISA (Table 1).

The MTF2/MTR2-PCR results indicated that all 30Microcystis strains examined, whether toxic or not, pos-sessed at least one peptide synthetase gene. The preva-lence of the peptide synthetase genes among Microcystisstrains was also reported in other studies (Christiansen etal. 2001). An expected fragment of about 1,000 bp couldbe amplified from nontoxic Anabaena flos-aquae 1444and Aphanizomenon flos-aquae 44–1, but not from theOscillatoria strains or another Anabaena strain ( Table 1).

In this investigation, two pairs of primers, TOX1P/TOX1 M and TOX2P/TOX2 M, were used to amplify thespecific fragments of mcyB. Among the 30 tested Micro-cystis strains, 18 hepatotoxic Microcystis strains gave pos-itive PCR signals. As illustrated in Fig.1 and Table 1, aproduct of 1,300 or 350 bp was measured, depending onwhich primer pair was used. Conversely, the non-toxicstrains yielded no detectable signals. The results with theTOX1P/TOX1 M pair were consistent with the data fromthe TOX2P/TOX2 M pair in most cases. The PCR data

with the latter primer pair detected more toxic strains andwere consistently in agreement with results from the de-termination of toxin and toxicity (Table 1). Therefore, theTOX2P/TOX2 M primer pair appears to be more reliableand was thus used in subsequent experiments.

Among the eight filamentous cyanobacteria, no PCR-amplified product was measured from the non-hepato-toxic strains, which included A. flos-aquae 1444, An-abaena sp. 7120, Ap. flos-aquae 44–1, and O. planctonica708. Table 1 also shows that neither O. agardhii NIES-595nor O. rubescens NIES-610 gave a detectable signal whenthe primer pair TOX2P/TOX2 M was used.

In the context of the present study, the most important ob-servation is that the PCR with extracted DNA and withwhole cells without pretreatment gave identical results(Table 1, Fig.1). Therefore, for quick identification of micro-cystin-producing strains, whole cells may be used directly.

Whole-cell PCR used in samples other than axenic cultures

With the positive results from axenically cultivated cells,the same simple protocol was applied to lyophilized

424

Fig.1 Electrophoresis of the PCR products with primers TOX2P/TOX2 M from various Microcystis strains cultured axenically. A DNA-template PCR. Lanes: 1 Microcystis sp. FACHB20, 2 Mi-crocystis viridis FACHB (Dianchi-1), 3 Microcystis aeruginosaNIES-90, 4 Microcystis wesenbergii FACHB574, 5 Microcystisaeruginosa FACHB (Dianshanhu), 6 Microcystis aeruginosaUTEX2061, 7 Microcystis aeruginosa FACHB (Wuda’erchi), 8 Microcystisa FACHB8641, 9 Microcystis aeruginosa FACHB86,10 Microcystis sp. FACHB10, 11 Microcystis aeruginosa FACHB(Dianchi-1), 12 Microcystis aeruginosa FACHB (Bao’anhu). B Whole-cell PCR. Lanes: 1 Microcystis sp. FACHB20, 2 Micro-cystis aeruginosa UTEX2061, 3 Microcystis aeruginosa NIES-90,4 Microcystis wesenbergii FACHB574, 5 Microcystisa FACHB(Dianshanhu), 6 Microcystis viridis FACHB (Dianchi-1), 7 Micro-cystis aeruginosa FACHB (Wuda’erchi), 8 Microcystis aerugi-nosa FACHB8641, 9 Microcystis aeruginosa FACHB86, 10 Mi-crocystis sp. FACHB10, 11 Microcystis aeruginosa FACHB (Dianchi-1), 12 Microcystis aeruginosa FACHB (Bao’anhu), 13 Genomic DNA of Microcystis aeruginosa PCC 7806 (control),14 Genomic DNA of Microcystis aeruginosa PCC 7820 (control).A and B M DNA marker. Note that lane 2 and lane 6 in A do notcorrespond to lane 2 and lane 6 in B

Fig.2 Discrimination of the hepatotoxic Microcystis by whole-cell PCR. A Water samples, B lyophilized cells. Lanes 1–4: Watersamples from Taihu Lake, December, 1999. 5 Genomic DNA ofMicrocystis aeruginosa PCC 7806 (control), 6 Microcystis aerug-inosa FACHB8641, 7 Microcystis wesenbergii FACHB574, 8 Mi-crocystis aeruginosa PCC 7820, 9 Microcystis viridis FACHB(Dianchi-2), 10 Microcystis aeruginosa FACHB (Dianchi-1), 11 Oscillatoria raciborskii FACHB (Daye-1). M DNA marker

cyanobacterial cells, uncultured environmental samples,such as Microcystist bloom samples, and sediments with amixed background of other cyanobacterial species andbacteria besides Microcystis. Six samples of the lyophiliz-ed cyanobacteria cells were directly examined by PCRwithout any pretreatment or DNA extraction. M. aerugi-nosa 8641, M. wesenbergii 574, M. aeruginosa PCC 7820,M. viridis (Dian-2) and M. aeruginos (Dian-1) were provento produce toxins by HPLC, ELISA and mouse bioassay.As a negative control, lyophilized O. raciborskii (Daye-1)cells generated no detectable PCR signals (Fig.2B).

More than 200 water samples collected from theblooms in Dianchi Lake, East Lake, and Taihu Lake werealso checked . Whole-cell PCR amplified the mcyB frag-ments in all 200 samples that were also found positive byELISA. More than 95% of the samples were positive inmicrocystin assays by ELISA. Every month from Marchto December, 2000, the mixed water samples were testedby HPLC, and the results were consistent with those of whole-cell PCR (Fig.2A.). A similar consistency wasalso observed in the tests of the sediment samples from

Dianchi Lake and Guanqiao pool. As shown in Table 2,positive results of these samples were confirmed by theexpected amplification products in the corresponding wa-ter samples.

However, the crude sediment samples without any pre-treatments yielded no whole-cell PCR products. Severalreports have suggested that humic acids or other materialsin sediments might remain associated with the cyanobac-teria cells, inhibiting DNA polymerases and thus prevent-ing PCR (Carol 1996; Higuchi 1989; Picard et al. 1992;Tsai and Olson 1992a, b). After extensive treatment steps,such as adding skim milk, bovine serum albumin, orphage-T4 gene 32 protein to the PCR reaction mixture(Carol 1996), this PCR inhibition could be alleviated. Theaddition of BSA or skim milk was found to be essentialfor the whole-cell PCR; without this step, no DNA frag-ment could be amplified even from the whole cyanobacte-rial cells.

The whole-cell PCR protocol can be optimized by rins-ing the cyanobacterial cells one to three times with dis-tilled water and by adjusting the density of the primers

425

Table 2 Identification of toxicMicrocystis strains from sedi-ment samples by whole-cellPCR

Samples Sampling spots Results of TOX-PCR

TOX1P/1F TOX2P/2F

Haigeng + +No. 13 + +

Sediment samples

No. 17 + +

Haigeng + +No. 13 + +

Floating cells collected from sediment samples after being stirred and

being suspended in distilled waterNo. 17 + +

Haigeng + +No. 13 + +

Cells collected from and cultured in medium MA for 1 week

No. 17 + +

Haigeng + +No. 13 + +

Cells collected from sediment samples and cultured in MA media for 1 week

No. 17 + +

Haigeng + +No. 13 + +

Corresponding water samples above sediment samples

No. 17 + +

No. 1 + +No. 2 + +No. 3 + +No. 4 + +

Microsystis isolates from sediment samples of Dian Chi Lake

No. 5 + +

Sterilized sediment samples

Microcystis.aeruginosaPCC 7806

+

Microcystis.aeruginosaUTEX2061

–

Microcystis.aeruginosaFACHB -(Dianchi-4)

+ +

Microcystis strains inoculated in sterilized sediments

Microcystis.aeruginosaFACHB -(Bao’anhu)

+ +

+

–

and cells in the reactions. Following such steps, amplifi-cation of non-specific products could be minimized.Hence, we suggest that the non-specific products do notcome from cyanobacteria themselves, but perhaps fromother bacteria associated with cyanobacteria.

Sensitivity of the whole-cell PCR

The sensitivity of the whole-cell PCR was investigatedwith axenically cultured toxin-producing Microcystis cells.A concentration as low as about 2,000 cells/ml was foundsufficient to give the expected PCR product (Fig.3). Forlyophilized cyanobacterial cells, 1×10–4mg (dry wt) toxiccells were necessary to amplify the mcyB fragments. Bythis means, the toxic Microcystis cells could be roughlyestimated by the intensity of the PCR signals.

Sequences and comparisons

To confirm that the amplified DNA fragments were in-deed parts of the mcyB genes, 11 PCR products were se-lected for sequencing. These fragments were derived ei-ther from DNA templates or from the whole-cell PCR.All 11 sequenced PCR products exhibited 98–99% iden-tity with the corresponding mcyB from M. aeruginosa(GenBank accession no. U97078). For the axenic strainM. aeruginosa PCC 7806 and a fresh water sample collectedfrom Dianchi Lake, the products of both types of PCRwere sequenced; no difference was found between the twomethods. Among seven other strains, including M. aeru-ginosa PCC 7820, M. aeruginosa 90, M. aeruginosa(Wuda’erchi), M. aeruginosa (Dianchi-3), M. viridis(Dianchi-1), M. viridis (Dianchi-2) and M. wesenbergii574, the alignments revealed that their sequences had a

high level of identity, ranging from 97 to 99% (data notshown).

The present study presents a simple and rapid PCRprotocol which, by virtue of using intact cells, permits theassessment of toxicity in large number of samples simul-taneously without the tedious DNA extraction and purifi-cation steps. This method could be applied to routinelymonitor the temporal and spatial changes of the toxic Mi-crocystis populations in bodies of water, thus, supportingtheir long-term evaluation. Whole-cell PCR was demon-strated to be applicable not only to the Microcystis strainsbut also to other cyanobacteria, such as An. flos-aquae1444 or Ap. flos-aquae 44–1. In order to develop a reli-able detection of all the mcy genes in Microcystis as wellas other microcystin-producing genera, further efforts areessential to select more optimum primers and PCR condi-tions..

Acknowledgements The authors would like to thank two anony-mous reviewers for their critical comments to the paper. Thanksare also due to Lian Min and Shen Qiang for their kind assistancein the ELISA and HPLC analysis in this study. This research wassupported by the National 863 Program (no. 2001AA641030) andthe Chinese Academy of Sciences Project (STZ-01–31) to L. Song,and the National Key Project on “Dianchi Lake Bloom Control”(K99–05–35–01) to Y. Liu. This research was also financially sup-ported by the Frontier Science Projects Programme of the Instituteof Hydrobiology, CAS

References

Carmichael WW (1996) Cyanobacterial toxins. IOC Manuals andGuides 33:163–175

Carol AK (1996) Relief of amplification inhibition in PCR withbovine serum albumin or T4 gene 32 protein. Appl EnvironMicrobiol 62:1102–1106

Cenis JL (1992) Rapid extraction of fungal DNA for PCR amplifi-cation. Nucleic Acids Res 20:2380

Christiansen G, Dittmann E, Ordorika LV, Rippka R, Herdman M,Börner T (2001) Nonribosomal peptide synthetase genes occurin most cyanobacterial genera as evidenced by their distribu-tion in axenic strains of the PCC. Arch Microbiol 176: 452–458

Dittmann E, Neilan BA, Börner T (1999) Peptide synthetase genesoccur in various species of cyanobacteria. In: Peschek GA,Loeffelhardt W, Schemetterer G (eds) The phototrophic pro-karyotes. Kluwer Academic/Plenum, New York, pp 615–621

Higuchi R (1989) Simple and rapid preparation of samples forPCR. In: Erlich HA (ed) PCR technology: principles and appli-cations for DNA amplification. Stockton, New York, pp 31–38

Howitt CA (1996) Amplification of DNA from whole cells ofcyanobacteria using PCR. Biotechniques 21: 32–33

Ichimura T (1979) Media for blue-green algae. In: Nishizawa K,Chihara M(eds) Methods in algological studies. Kyoritsu, Tokyo,pp 294–305 (in Japanese)

Meissner K, Dittmann E, Börner T (1996) Toxic and non-toxicstrains of the cyanobacterium Microcystis aeruginosa sequenceshomologous to peptide synthetase genes. FEMS Microbiol Lett135: 295–303

Neilan BA, Dittmann E, Rouhiainen L, Bass RA, Schaub V, Sivo-nen K, Börner T (1999) Nonribosomal peptide synthesis andtoxigenicity of cyanobacteria. J Bacteriol 13: 4089–4097

Nishizawa T, Asayama M, Fujii K, Harada K, Shirai M (1999) Ge-netic analysis of the peptide synthetase genes for a cyclic hep-tapeptide microcystin in Microcystis spp. J Biochem (Tokyo)126(3):520–9

426

Fig.3 Correlation between the cell concentration and the intensityof the PCR signals in toxic strains of Microcystis. Lanes: 1 Ge-nomic DNA of Microcystis aeruginosa PCC 7806 (control), 2 crude water sample, 3 cells from the water sample rinsed withdistilled water three times, 4–10 dilutions of cell suspensions ofMicrocystis aeruginosa PCC 7806 (5-, 10-, 50-,100-, 500-, 1,000-,and 2,000-fold dilutions). The starting cell concentration was4×106 cells/m l. M DNA marker

Nishizawa T, Ueda A, Asayama M, Fujii K, Harada K, Ochi K,Shirai M. (2000) Polyketide synthase gene coupled to the pep-tide synthetase module involved in the biosynthesis of thecyclic heptapeptide microcystin. J Biochem (Tokyo) 127:779–89

Noriko T, Masayuki Y, Kiyoshi S (1984) Overwintering of Micro-cystis aeruginosa Kuetz. in a shallow lake. J Plankton Res 6:1019–1029

Picard C, Ponsonnet C, Paget E, Nesme X, Simonet P (1992) De-tection and eumeration of bacteria in soil by direct DNA ex-traction and polymerase chain reaction. Appl Environ Micro-biol 58:2717–2722

Preston T, Stewart WDP (1980) Bloom-forming cyanobacteriumMicrocystis aeruginosa overwinters on sediment surface. Na-ture 288:365–367

Sidney PC, Kaplan NO (1988) Methods in enzymology, vol. 167.Academic, New York

Sivonen, K (1996) Cyanobacterial toxins and toxin production.Phycologia. 35:12–24(Suppl)

Song LR, Lei LM, He ZR, Liu YD (1999) Toxicity analysis of Mi-crocystis aeruginosa and Microcystis viridis in Dianchi Lake.China Acta Hydrobiologica Sinica 23:402–407

Tillett D, Dittmann E, Erhard M, von Dohren H, Börner T, NeilanBA (2000) Structural organization of microcystin biosynthesisin Microcystis aeruginosa PCC7806: an integrated peptide-polyketide synthetase system. Chem Biol 7:753–64

Tsai YL Olson BH (1992a) Rapid method for separation of bacte-rial DNA from humic substances in sediments for polymerasechain reaction. Appl Environ Microbiol 58:2292–2295

Tsai YL, Olson BH (1992b) Detection of low numbers of bacterialcells in soil and sediments by polymerase chain reaction. ApplEnviron Microbiol 58:754–757

Ueno Y, Nagata S, Tsutsumi T, Hasegawa A, Yoshida F, SuttajitM, Mebs D, Pütsch M, Vasconcelos V (1996) Survey of mi-crocystins in environmental water by a highly sensitive im-munoassay based on monoclonal antibody. Nat Toxins 4:271–276

427