www.elsevier.com/locate/ybbrc

Biochemical and Biophysical Research Communications 326 (2005) 687–694

BBRC

A novel rhythm of microcystin biosynthesis is describedin the cyanobacterium Microcystis panniformis Komarek et al.

Maria do Carmo Bittencourt-Oliveiraa, Paula Kujbidab, Karina Helena Morais Cardozoc,Valdemir Melechco Carvalhod, Ariadne do Nascimento Mourae,

Pio Colepicoloc, Ernani Pintob,*

a Departamento de Ciencias Biologicas, Escola Superior de Agricultura Luiz de Queiroz, Universidade de Sao Paulo, Piracicaba, SP, Brazilb Departamento de Analises Clınicas e Toxicologicas, Faculdade de Ciencias Farmaceuticas, CEP 05508-900, Universidade de Sao Paulo, SP, Brazil

c Departamento de Bioquımica, Instituto de Quımica, Universidade de Sao Paulo, Sao Paulo, SP, Brazild Instituto Fleury, Sao Paulo, SP, Brazil

e Departamento de Biologia, Universidade Federal Rural de Pernambuco, Recife, PE, Brazil

Received 12 November 2004Available online 2 December 2004

Abstract

The presence of microcystins (MCY) in the cyanobacteria Microcystis panniformis Komarek et al. is reported for the first time.This strain of cyanobacterium has been isolated from Barra Bonita, an eutrophicated water reservoir in Sao Paulo state, Brazil. Theidentification of M. panniformis was confirmed by both traditional morphological analysis and the phycocyanin intergenic spacersequences. MCY-LR and [Asp3]-MCY-LR were identified in this strain after HPLC purification and extensive ESI-MS/MS analysis.Their levels in this strain were determined by HPLC and ranged from 0.25 to 2.75 and 0.08 to 0.75 fmol/cell, respectively. Analyzingthe levels of MCY-LR and [Asp3]-MCY-LR in different times during the light:dark (L:D) cycle, it was found that levels of MCYsper cell were at least threefold as high during the day-phase than during the night-phase. This may be associated to the biologicalclock since prokaryotic cyanobacteria express robust circadian (daily) rhythms under the control of a timing mechanism that is inde-pendent of the cell division cycle. Our findings also showed the same pattern under light:light (L:L) cycle.� 2004 Elsevier Inc. All rights reserved.

Keywords: Diurnal rhythm; Microcystin; Cyanobacteria; Microcystis panniformis; HPLC; ESI-MS/MS

Some cyanobacterial blooms of the genus Microcystis

(Chroococcales, Cyanobacteria) can be a serious ecolog-ical and public health concern due to their ability todominate the planktonic environment and produce cyc-lic heptapeptide toxins, named microcystins (MCYs) [1].MCYs inhibit protein phosphatases, especially types 1and 2A, in a similar way to the action of okadaic acid[2].

Seasonal changes of Microcystis species and produc-tion of MCYs and aeruginopeptins have been reported

0006-291X/$ - see front matter � 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.bbrc.2004.11.091

* Corresponding author. Fax: +55-11-3815-6593.E-mail address: ernani@usp.br (E. Pinto).

in the Lake Suwa, Japan [3]. Photosynthetically activeradiation (PAR) can also interfere in some biochemicalprocess in cyanobacteria. The psbA2 gene has exhibitedlight-dependent and rhythmic expression in Microcystis

aeruginosa K-81 [4] and Microcystis strain PCC 7806showed a positive effect of PAR on microcystin produc-tion and content [5].

It has been shown that several physiological andbiochemical processes are controlled or at least areinfluenced by the biological clock in dinoflagellatesand cyanobacteria [6–11]. For example, the cyanobacte-ria Cyanothece strain CGD temporally separatesnitrogen fixation and photosynthetic activity to protect

688 M.C. Bittencourt-Oliveira et al. / Biochemical and Biophysical Research Communications 326 (2005) 687–694

oxygen-sensitive nitrogenase [12], three cyanobacteria,Synechocystis sp. PCC 6803, Synechococcus sp. PCC7942, and Cyanothece RF-1, clearly exhibited circadianphotosynthetic rhythm by employing a dissolved-oxygenmeter to monitor the photosynthesis [13] and the cyano-bacterial KaiB and KaiC proteins (encoded by the genecluster kaiABC) are robustly rhythmical, whereas theKaiA abundance undergoes little if any circadian oscil-lation in constant light [11,14].

These findings strongly implicate a circadian regula-tory mechanism operating on these metabolic processesare evidence for the importance of circadian rhythms inglobal metabolic regulation in some cyanobacteria spe-cies. For this reason, some natural peptides, includingMCYs and other bioactive oligopeptides, synthetizedby cyanobacteria may be influenced by the biologicalclock [15].

Microcystis panniformis strain BCCUSP 100 pre-sented the gene mcyB that encodes a MCY synthetase[16]; however, the toxin itself has not previously been de-tected in this cyanobacterium species.

In this report, we isolated and identified for the firsttime by ESI-MS/MS two MCYs in this species. Theidentification of M. panniformis strain BCCUSP 100was confirmed by traditional morphological criteriaand the phyletic relationships and genetic diversity ofthe Brazilian strains ofM. panniformis with other closelyrelated species of Microcystis were investigated. HPLCanalyses with M. panniformis samples collected duringL:D and L:L cycles indicated that the MCY-LR and[Asp3]-MCY-LR levels were higher during the day. Assupported by our data, the biosynthesis of MCYs iscontrolled by the biological clock. These findings areimportant to elucidate the mechanisms involved inMCYs biosynthesis as well as their isolation andcharacterization.

Materials and methods

Reagents. All reagents were obtained as ultrapure grade. Peptidetoxins MCY-LR and [Asp3]-MCY-LR were purified from the cyano-bacterium M. panniforms strain BCCUSP 100 grown in our laboratory[16] as described below.

Study area and field sampling. Water samples were collected fromblooms in Barra Bonita reservoir (22� 32 0 34.500 S, 48� 29 0 26.400 W) inApril 2000 using a 25 m plankton net. This reservoir is located at BarraBonita city, Sao Paulo state, Brazil (total area: 308 km2, water volume:2.566 · 106 m3, that has been used as hydroelectric plant and for rec-reation). Microcystis, Cylindrospermopsis, and Raphidiopsis bloomshave been occurring often in this reservoir.

Morphology. Field colonies were identified according to the mor-phological criteria [17,18], as: colony shape, cell diameter, and muci-lage aspect. Photomicrographs of colonies were taken with amicroscope (Nikon E200, Melville, NY, USA) equipped with a videocamera system Samsung SCC833 using the software Imagelab (Sof-tium, Brazil). One drop of diluted nankeen was used on the lamina tomake the mucilage evident. During the cultivation in the laboratory,the strain underwent morphological variation and became unicellular.

Measurements of cell diameter from M. panniformis strains were per-formed at random during the logarithmic phase of growth (n = 50) foreach strain. The averages of cell diameter and standard deviation werestatistically analyzed using SAS software (version 8.0) [19].

Strains and growth conditions. The non-axenic M. panniformis

BCCUSP 100 strain was isolated in modified BG-11 medium [20] inglass tubes containing 10 mL culture medium. One individual colonywas removed by micromanipulation techniques using a microscope(Nikon E200, Melville, NY, USA) at magnification of 100· to 400·.The isolated colony was washed by transferring it to several drops ofwater until all other microorganisms were removed and subsequentlytransferred to liquid medium. Some Brazilian Microcystis strains wereisolated in our previous study [16] and maintained at the BrazilianCyanobacteria Collection of University of Sao Paulo, Brazil (exFCLA).

Daily variation. Prior to the beginning of the experiments, a pre-culture of 150 mL was grown on a 12:12 h (light:dark) photoperiod, at23 ± 0.5 �C, light intensity was 75 ± 2 lmolphotonsm�2 s�1 to obtaininocula in an appropriate physiological condition. Light intensity wasmeasured using a spherical quantum photometer (LI-COR mod. LI-250 Lincoln, NE, USA) placed inside a culture flask with the medium.The pre-culture (15 days, exponential growth phase) was divided intoamounts of 50 mL and inoculated in 3 identical flasks with 2.2 L ofnew medium without aeration. The cultures were initiated with2.6 · 104 cellsmL�1 and maintained as previously described. Thesamples were collected in the middle of the exponential growth phase(around 106 cellsmL�1). Two experiments were carried out: (i) light:-dark cycle (12 h in the light and 12 h dark): samples were collected at2 h intervals, 50 mL to toxin analysis and cell counting (1 mL) and (ii)light:light cycle (24 h with light): the same procedure was used as in (i).In both experiments, cells were grown as previously described. Fiftymilliliters of culture was placed into Falcon tubes and centrifuged at4000 rpm for 15 min at 23 �C. Cell pellets were immediately frozen inliquid nitrogen and stored at �86 �C until analysis. Cell densities wereestimated by means of microscopic counts of cell samples, stained withLugol�s 4% solution in Fuchs Rosenthal hemocytometer. It wasestablished that a minimum of 400 cells needed to be counted to obtainan error of approximately 10% to a confidence level of 95%.

DNA extraction. Total genomic DNA was prepared using thecommercial kit Gnome DNA (BIO 101, Vista, CA, USA) and DNAsample was cleaned and purified through QiaQuick columns (Qiagen,Valencia, CA, USA) according to the manufacturer�s instructions.DNA concentrations were estimated directly from ethidium bromidefluorescence in agarose gel images against standard quantities of DNA(Low DNA mass, Invitrogen, Carlsbad, CA, USA) either by using aKodak gel documentation system (Kodak, Melville, NY, USA) andassociated software Kodak Digital Science 1D.

Amplification DNA and cpcB-cpcA intergenic spacer sequencing.The cpcB-cpcA phycocyanin intergenic spacer (PC-IGS) and flankingcoding region were amplified by PCR as described by Bolch et al. [21].The amplified fragments were directly sequenced using the forwardand reverse primers (PCb-F: 5 0-GGCTGCTTGTTTACGCGACA-3 0;PCa-R: 5 0-CCAGTACCACCAGCAACTAA-3 0) with ABI Prism BigDye Terminator Cycle Sequencing Ready reaction Kit (Perkin–ElmerApplied Biosystems, Foster City, CA, USA) and 3100 ABI sequencer(Perkin–Elmer Applied Biosystems, Foster City, CA, USA) accordingto the manufacturer�s instructions. To avoid errors by PCR, at least 3separate amplification reactions were pooled for sequencing. The PCRproducts were sequenced on both strands at least four times. ThecpcBA-IGS nucleotide sequence from strain BCCUSP 100 was com-pared to entries deposited in the GenBank database (http://www.ncbi.nml.nih.gov) to verify taxonomic accuracy. Automated basecalls for both strands were checked by manual inspection andambiguous calls and conflicts were resolved by alignment and com-parison using Sequencer program (version 3.0) to establish a consensussequence for the strain. The BCCUSP 100 consensus sequence andreference Microcystis sequences were assembled manually, using ESEE

M.C. Bittencourt-Oliveira et al. / Biochemical and Biophysical Research Communications 326 (2005) 687–694 689

3.2 [22]. Phylogenetic inferences were made by maximum-likelihood(ML), maximum-parsimony (MP), and neighbor-joining (NJ) methodsfor phylogenetic inferences with the computer program PAUP* 4.0[23]. Trees were supported by the bootstrap approach [23,24] using1000 replicates of the heuristic search algorithm.

Hip1-CA PCR. Template DNA (1–10 ng) was amplified using5 lmol of the Hip1-CA primer (5 0-GCGATCGCCA-3 0) [25] and thePCR commercial kit (puReTaq Ready-to-Go PCR beads, AmershamBiosciences). The following cycling parameter conditions were used:95 �C for 5 min, followed by 40 cycles of 95 �C for 30 s; 36 �C for1 min; and 72 �C for 2 min; followed by a final extension at 72 �C for5 min. Amplification products were visualized by electrophoresis in1.5% agarose gel and detected by staining with ethidium bromide by aKodak gel documentation system and associated software KodakDigital Science 1D. The amplified fragments were used to generate abinary data matrix from which similar matrices were obtained usingJaccard�s coefficient. The UPGMA algorithm [25] was used for thephenogram construction using the software NTSYS version 1.70.Doubtful bands of low resolution were disregarded.

Microcystin extraction and HPLC semipreparative analyses. MCYswere extracted with MeOH/H2O 3:1 from freeze-dried samples (around500 mg) and submitted to an ultrasound bath for 10 min. The extractwas centrifuged (10,000 rpm, 15 min) and the supernatant collected.The pellet obtained was re-extracted according to the proceduredescribed above. The supernatants were combined and dried in arotaevaporator (bath at 40 �C). The dried material was resuspended in3 mL DCM and applied to a silica column (20 · 5 cm, Silica GelKeiselgel 60, Merck). The column was equilibrated with DCM and theelution steps were: 30 mL DCM, 30 mL DCM:MeOH, 30 mL MeOH,and 30 mLMeOH:H2O. Fractionswere dried in a rotaevaporator (sameas described above). Toxins were found to be in the last elution step. Thelast dried fraction was resuspended with 1 mL of mobile phase and thenrepurified in a HPLC system equipped with a pump LC-10AD, a PDAdetector (SPD10AV), and a SCL-10Avp System Controller (Shimadzu,Kyoto, Japan). About 500 lL of the sample was injected in the systemand chromatographed in a semipreparative HPLC column (Phenome-nex, Luna C18, 5 lm, 250 · 10 mm) eluted with a mixture ofMeCN and0.05 mol/L, pH 3, NH4CH3COO (1:3) at a flow rate of 4.7 mLmin�1.The detection was set at 238 nm. Peaks 1 and 2 (Fig. 1) were furtheridentified as MCY-LR and [Asp3]-MCY-LR.

Fig. 1. Typical chromatogram obtained with the extracts ofM. panniformis strain BCCUSP 100 (Phenomenex Luna column C18,256 · 4,6 mm, 5 lm, ACN:NH4CH3COO buffer, pH 3, flow rate =1 mLmin�1 and k = 238 nm). 1, microcystin-LR and 2, [Asp3]-micro-cysitin-LR.

Mass spectrometric analysis. ESI-MS/MS spectra were obtained ona Quattro Micro tandem mass spectrometer (Waters Micromass,Manchester, UK). Samples (peaks 1 and 2) in MeCN:formic acid 0.1%(1:1, v/v) were introduced into the electrospray ion source by directinfusion using the integrated syringe pump at flow rates ranging from 2to 10 lLmin�1. The mass spectra were acquired in the positive ionmode with the source and analyzer parameters were optimized for theprotonated molecular ion. Tandem MS spectra (daughter scans) wereacquired using Ar as a collision gas (4 · 10�3 mbar) at different colli-sion energies (5–60 eV).

HPLC analysis for the daily variation experiments. MCYs wereextracted with 3 mL MeOH/H2O 3:1 from freeze-dried samples,around 10 mg, and submitted to an ultrasound bath for 10 min. Theextract was centrifuged (10,000 rpm, 15 min) and the supernatant wasevaporated (SpeedVac, Savant, City, US state). The precipitate wasresuspended in 1 mL MeOH and injected into a Sep-Pak cartridge(C18, Waters, Milford, MA). The preconditioning step includedwashing with MeOH (2 mL) and H2O (2 mL) and the elution stepswere: (i) 1 mL H2O, (ii) 1 mL MeOH/H2O 1:1, and (iii) 1 mL MeOH.MCYs were found to be in the last eluate. This fraction was alsoevaporated (SpeedVac) and the precipitate resuspended with 200 lLHPLC mobile phase and analyzed in the HPLC system equipped witha C18 Luna column (5 lm, 0.46 · 25 cm) eluted with MeCN/0.05 mol/L NH4CH3COO buffer (pH 3) 1:3 at 1 mLmin�1. Detection was per-formed at 238 nm with a SPD10AV PDA detector. HPLC analyseswere performed with a Shimadzu HPLC system. The calibration curvewas obtained with the MCY LR and [Asp3]-MCY-LR isolated fromM. panniformis.

Statistical analyses. The data were expressed as mean val-ues ± standard deviation (SD). The data for each experimental vari-able were tested for the basic premises of analysis of variance(ANOVA) model, using the Bartlett test for homogeneity of variancesand the v2 test for normality [26]. A single-factor ANOVA was appliedto the mean values obtained for each experimental time in order todetect significant differences. Whenever the null hypothesis of ANOVAwas rejected, the Tukey test of multiple comparisons [27,28] was em-ployed and the statistically significant differences (P < 0.05) betweeneach pair of mean values were discriminated.

Results and discussion

Microcystis panniformis identification

Previously, the strains BCCUSP 03, BCCUSP 30,BCCUSP 158, BCCUSP 200, and BCCUSP 310 weredenominated as belonging to the ‘‘Microcystis aerugin-

osa complex’’ [29]. In this paper, it was denominatedas M. panniformis following the suggestion of Dr. JiriKomarek ([30] see editorial remark). However, thediameter of the cells of the same strain was larger thanthe original descriptions, 3.9–6.4 lmdiam. (Table 1), re-sults that corroborated with White et al. [31].

The morphospecies M. panniformis showed broaddiversification (Fig. 2). Description of population andtheir life cycle were also described in the literature[16,29,30]. The M. panniformis populations found in theenvironmental samples from Barra Bonita reservoirshowed elongate and flat colonies with densely arrangedcells in (see Figs. 2A–C) agree withKomarek et al. [16,18].

Similar groupings emerged in the parsimony, neigh-bor-joining, and maximum-likelihood. Thus, only the

Table 1Mean, standard deviation,and maximum and minimum values of cell diameter for M. panniformis

Strain Sample location Mean (lm) Standard deviation Min–max (lm)

BCCUSP 100 Barra Bonita reservoir-SP 4.67 0.60 3.5–5.8BCCUSP 003 Garcas reservoir-SP 4.96 0.323 4–5.9BCCUSP 030 Garcas reservoir-SP 3.00 0.37 2.5–4BCCUSP 158 Garcas reservoir-SP 5.10 0.640 3.9–6.1BCCUSP 199 Garcas reservoir-SP 5.52 0.43 4.9–6.4BCCUSP 200 Americana reservoir-SP 3.81 0.49 3–4.9BCCUSP 310 Garcas reservoir-SP 3.20 0.39 2.9–4.0

Fig. 2. The morphospecies M. panniformis Komarek et al. (A,B) Ribbon-like and flat colonies with cells arranged in a single plane under cultureconditions, not unialgal; Barra Bonita reservoir. (C) Mature and juvenile colonies in environmental samples from Barra Bonita reservoir.

690 M.C. Bittencourt-Oliveira et al. / Biochemical and Biophysical Research Communications 326 (2005) 687–694

maximum-likelihood tree is shown (Fig. 3). The phylo-genetic tree was constructed from the 465 nucleotidesof the cpcBA-IGS and flanking regions. The 21 strainsanalyzed were separated in five major clusters (I–IV)

Fig. 3. Majority rule consensus maximum-likelihood tree of M. panniformis

alignment of 465 nucleotides of the cpcBA-IGS and flanking regions after excto the left of each node. Only bootstrap values of >50% are shown. Numbreplicates. The scale bar shows the branch length corresponding to 0.005 nutopology; the ML tree is shown. Accession numbers in the GenBank databindicated by BR, Brazil; JP, Japan, SP, Spain; CN, Canada; and N, not citedto BCCUSP. Black triangles, microcystin-producing strains. White triangles

divisible into taxonomically relevant groups. Two Bra-zilian M. aeruginosa strains were composed of the Clus-ter I and they did not group together with Cluster IIcomposed of eight M. aeruginosa, Microcystis viridis,

from Barra Bonita, Brazil (boldface), and other Microcystis spp. Anluding positions with gaps was used. Bootstrap percentages are showners above the branches are bootstrap support as a percentage of 1000cleotide substitution per site. NJ and MP methods gave the same treeases are underlined. The geographical origin of Microcystis strains is. Strains in each rectangle have identical sequence. FCLA is equivalent, not microcystin-producing strains.

M.C. Bittencourt-Oliveira et al. / Biochemical and Biophysical Research Communications 326 (2005) 687–694 691

and Microcystis flos-aquae strains from Brazil, Japan,Canada, and Spain. Three Spanish strains formed theCluster III: two Microcystis novacekii and M. aerugin-

osa. Finally, the Cluster IV showed three Microcystis

wesenbergii strains from Japan and Brazil shared 99.98to 100% similarity. The M. panniformis BCCUSP 100was isolated from the others.

On the other hand, the Hip1CA primer generatedPCR products and agarose gel electrophoresis bandingpatterns (Fig. 4) that did not distinguish the M. panni-

formis, M. wesenbergii, and Microcystis spp. Three dis-tinguished clusters and one strain that did not groupwere observed (Fig. 5). The three clusters had low simi-larity (A = 0.38, B = 0.49, and C = 0.40). M. pannifor-

Fig. 4. Ethidium bromide-stained 1.5% agarose electrophoresis gel showingMicrocystis spp. maintained at Brazilian Cyanobacteria Collection of Univpanniformis: BCCUSP 03, BCCUSP 30, BCCUSP 100, BCCUSP 158, BCaeruginosa: BCCUSP 09, BCCUSP 225, BCCUSP 232, BCCUSP 235, BCCBCCUSP 299, and BCCUSP 450.

Fig. 5. Clustering dendrogram ofMicrocystis spp. strains by UPGMA analysand Jaccard�s coefficient. The numerical scale indicates the level of similaritpanniformis: BCCUSP 3, BCCUSP 30, BCCUSP 100, BCCUSP 158, BCCUSsp.: BCCUSP 9, BCCUSP 225, BCCUSP 232, BCCUSP 235, BCCUSP 255,299, and BCCUSP 450. A = 0.38, B = 0.49, and C = 0.40.

mis strains were present among clusters A (BCCUSP158, BCCUSP 03, BCCUSP 30, and BCCUSP 310), B(BCCUSP 200), and C (BCCUSP 100).

Microcystin isolation and identification by ESI-MS/MS

We found a very effective and simple method formicrocystin purification combining extraction byMeOH/H2O [3:1], silica column chromatography, andone-step semipreparative HPLC. Due to microcystin�shigh polarity, it could be dissolved in DCM and elutedfrom silica gel using MeOH/H2O [1:1]. Two suspectedcompounds were found in the HPLC profile, named asPeaks 1 and 2. A typical chromatogram can be seen in

profiles of PCRs employing Hip1-CA primer of strains of Brazilianersity of Sao Paulo. M, molecular weight marker, in base pairs. M.

CUSP 200, and BCCUSP 310. M. wesenbergii: BCCUSPNinf. M.

USP 255, BCCUSP 258, BCCUSP 262, BCCUSP 236, BCCUSP298,

is of similarity matrix data using 92 OTU (operational units taxonomic)y at which clusters are formed, according to Jaccard�s coefficient. M.

P 200, and BCCUSP 310. M. wesenbergii: BCCUSP Ninf. Microcystis

BCCUSP 258, BCCUSP 262, BCCUSP 236, BCCUSP 298, BCCUSP

692 M.C. Bittencourt-Oliveira et al. / Biochemical and Biophysical Research Communications 326 (2005) 687–694

Fig. 1. These compounds were analyzed by ESI-MS;producing protonated molecular ions at m/z 995.4 andm/z 981.4, respectively. As suggested by their molecularweights and chromatographic behavior, we suspectedthey could be toxins [Asp3]-MCY-LR and MCY-LR.In order to confirm this hypothesis we analyzed bothcompounds by tandem MS.

Product ion spectra of 995.4 and 981.4 (Figs. 6A andB) showed typical fragments from MCYs such as m/z

135 that had been attributed to a fragment originatedfrom a-cleavage of the methoxy group of the Adda[32]. Ion m/z 213 had been assigned to the dipeptideGlu-Mdha and ion m/z was proposed to be producedby the loss of the fragment 135 from the tetrapeptideAdda-Glu-Mdha-Ala [32,33]. The ion m/z 599 is alsopresent in both spectra and could be assigned to the se-quence Arg-Adda-Glu.

Fig. 6. Positive mode ESI-MS/MS s

The identities of 1 and 2 as being MCY-LR and[Asp3]-MCY-LR are supported by a series of ions thatinclude MeAsp or Asp and consequently presented amass difference of 14 Da between spectra. Thus, itwas found m/z = 967 [cyclo(Ala-Leu-MeAsp-Arg-Adda-Glu-Mdha)-CO] and m/z = 953 [cyclo(Ala-Leu-Asp-Arg-Adda-Glu-Mdha)-CO]; m/z = 553 [Mdha-Ala-Leu-MeAsp-Arg] and m/z = 539 [Mdha-Ala-Leu-MeAsp-Arg]; m/z = 470 [Ala-Leu-MeAsp-Arg] andm/z = 456 [Ala-Leu-Asp-Arg]; m/z = 397 [Mdha-Ala-Leu-MeAsp] and m/z = 383 [Mdha-Ala-Leu-Asp]; andm/z = 286 [MeAsp-Arg] and m/z = 272 [Asp-Arg].

Daily variation of microcystins

It has been demonstrated that PAR, pH, and nutri-ents (iron, phosphorus, and nitrogen) influence the

pectra: (A) peak 1; (B) peak 2.

Fig. 7. Levels of toxins (fmolcell�1) produced by M. panniformis in 24 h (samples had been collected each 2 h), where (A) variation of MCY-LR atL:D condition, (B) variation of MCY-LR at L:L condition, (C) variations of [Asp3]-MCY-LR at L:D condition, and (D) variation of [Asp3]-MCY-LR at L:L condition.

M.C. Bittencourt-Oliveira et al. / Biochemical and Biophysical Research Communications 326 (2005) 687–694 693

growth and MCY content ofMicrocystis spp. ([5,34–39],respectively). However, to our knowledge, the circadianvariation of MCYs levels in Microcystis spp. has neverbeen reported.

In the studies presented, the levels of MCY-LR and[Asp3]-MCY-LR in M. panniformis under L:D andL:L cycles showed a peak at the middle of the day,around 12–14 h (Fig. 7). Also, in both L:D and L:L cy-cles, MCY-LR is almost fourfold more abundantaround 12 h than during the dark phase. [Asp3]-MCY-LR, in the L:D experiment, shows the same pattern asthat of MCY-LR. Nevertheless, [Asp3]-MCY-LR con-tents with the L:L treatment are twice as high in com-parison with the L:D experiment.

These results may be associated to the biologicalclock since in cyanobacteria, circadian rhythms havebeen found for photosynthesis, nitrogen fixation, someprotein synthesis, and cell division [6–10].

Conclusion

The identification of the M. panniformis strainBCCUSP 100 was confirmed by morphological profilesand based on sequence homology of the intergenicspacer region between the cpcA and cpcB-phycocyaninsubunits with database records and two MCYs were iso-lated and identified in this species, MYC-LR and [Asp3]-MYC-LR. We conclude from our results that the bio-logical clock controls the production of MYCs and their

content in M. panniformis, since their production wasobserved to peak at the middle of the day phase, in bothL:D and L:L experiments. This shows the importance ofcircadian regulation in this cyanobacterium.

Also, this strain can be readily utilized as a referencefor MYC-LR and [Asp3]-MYC-LR production and tofollow MCYs biosynthesis under many cultureconditions.

Acknowledgments

The authors thank Dr. Paul Gates (Organic and Bio-logical Chemistry Section, School of Chemistry, Univer-sity of Bristol, UK) for helpful comments on the massspectrometry analysis and also for the English languagerevision. This research was supported by grants fromFAPESP (2003/06443-0 and 2003/05773-6), CNPq(302439/2002-1), and CAPES.

References

[1] G.A. Codd, S.G. Bell, W.P. Brooks, Cyanobacterial toxins inwater, Water Sci. Technol. 21 (1989) 1–13.

[2] R.M. Dawson, Toxicology of microcystins, Toxicon 7 (1998) 953–962.

[3] K. Harada, T. Mayumi, T. Shimada, K. Fujii, F. Kondo, H.Park, M.F. Watanabe, Co-production of microcystins and aerug-inopeptins by natural cyanobacterial bloom, Environ. Toxicol. 16(2001) 298–305.

694 M.C. Bittencourt-Oliveira et al. / Biochemical and Biophysical Research Communications 326 (2005) 687–694

[4] G.K. Agrawal, M. Asayama, M. Shirai, Light-dependent andrhythmic psb transcripts in homologous/heterologous cyanobac-terial cells, Biochem. Biophys. Res. Commun. 255 (1999) 47–53.

[5] C. Wiedner, P.M. Visser, J. Fastner, J.S. Metcalf, G.A. Codd,L.R. Mur, Effects of light on the microcystin content ofMicrocystis strain PCC 7806, Appl. Environ. Microbiol. 69(2003) 1475–1481.

[6] O.K. Okamoto, D.L. Robertson, T.F. Fagan, J.W. Hastings, P.Colepicolo, Different regulatory mechanisms modulate the expres-sion of a dinoflagellate iron-superoxide dismutase, J. Biol. Chem.276 (2001) 19989–19993.

[7] E. Pinto, L.H. Catalani, N.P. Lopes, P. Di Mascio, P. Colepicolo,Peridinin as the major biological carotenoid quencher of singletoxygen in marine algae Gonyaulax polyedra, Biochem. Biophys.Res. Commun. 268 (2000) 496–500.

[8] J. Lorne, J. Scheffer, A. Lee, M. Painter, V.P.W. Miao, Genescontrolling circadian rhythm are widely distributed in cyanobac-teria, FEMS Microbiol. Lett. 189 (2000) 129–133.

[9] T. Kondo, M. Ishiura, The circadian clock of cyanobacteria,BioEssays 22 (2000) 10–15.

[10] T. Mori, C.H. Johnson, Circadian programming in cyanobacteria,Semin. Cell Dev. Biol. 12 (2001) 271–278.

[11] F. Hayashi, H. Ito, M. Fujita, R. Iwase, T. Uzumaki, M. Ishiura,Stoichiometric interactions between cyanobacterial clock proteinsKaiA and KaiC, Biochem. Biophys. Res. Commun. 316 (2004)195–202.

[12] M.A. Schneegurt, D.L. Tucker, J.K. Ondr, D.M. Sherman, L.A.Sherman, Metabolic rhythms of a diazotrophic cyanobacterium,Cyanothece sp. strain ATCC 51142, heterotrophically grown incontinuous dark, J. Phycol. 36 (2000) 107–117.

[13] U.C. Yen, T.C. Huang, T.C. Yen, Observation of the circadianphotosynthetic rhythm in cyanobacteria with a dissolved-oxygenmeter, Plant Sci. 166 (2004) 949–952.

[14] Y. Xu, T. Mori, C.H. Johnson, Circadian clock-protein expres-sion in cyanobacteria: rhythms and phase setting, EMBO J. 19(2000) 3349–3357.

[15] K. Harada, Production of secondary metabolites by freshwatercyanobacteria, Chem. Pharm. Bull. 52 (2004) 889–899.

[16] M.C. Bittencourt-Oliveira, Detection of potential microcystin-producing cyanobacteria in Brazilian reservoirs with a mcyBmolecular marker, Harmful Algae 2 (2003) 51–60.

[17] J. Komarek, J. Komarkova-legnerova, C.L. Sant�anna, M.T.P.Azevedo, P.A.C. Senna, Two common Microcystis species(Chroococcales, Cyanobacteria) from tropical America, includingM. panniformis sp. nov., Cryptogam. Algol. 23 (2002) 159–177.

[18] J. Komarek, J. Komarkova, Review of European Microcystis-morphospecies (Cyanoprokaryotes) from nature, Czech Phycol. 2(2002) 2–14.

[19] SAS Institute Inc. SAS Version 8, Cary, NC: SAS Institute Inc,1999.

[20] R. Rippka, J. Deruelles, J.B. Waterbury, M. Herdman, R.Y.Stanier, Generic assignments, strain histories and properties ofpure cultures of cyanobacteria, J. Gen. Microbiol. 111 (1979) 1–61.

[21] C.J.S. Bolch, S.I. Blackburn, B.A. Neilan, P.M. Grewe, Geneticcharacterization of strains of cyanobacteria using PCR-RFLP ofthe phycocyanin intergenic spacer and flanking regions, J. Phycol.32 (1996) 445–451.

[22] E.L. Cabot, A.T. Beckenbach, Simultaneous editing of multiplenucleic acid and protein sequences with ESEE, Comput. Appl.Biosci. 5 (1989) 233–234.

[23] D.L. Swofford, PAUP*: Phylogenetic Analysis Using Parsimony(* and Other Methods). Version 4.0. Sinauer Associates, Sunder-land, MA, 1998.

[24] J. Felsenstein, Confidence limits on phylogenies: An approachusing the bootstrap, Evolution 39 (1985) 783–791.

[25] J.K. Smith, J.D. Parry, J.G. Day, R.J. Smith, A PCR techniquebased on HIP1 interspersed repetitive sequence distinguishescyanobacterial species and strains, Microbiology 144 (1998) 2791–2801.

[26] P.H.A. Sneath, R.R. Sokal, Numerical Taxonomy, W.H. Free-man, San Francisco, 1973.

[27] R.R. Sokal, F.J. Rohlf, Biometry: The Principles and Practice ofStatistics in Biological Research, second ed., WH Freeman, NewYork, 1981.

[28] J. Neter, W. Wasserman, M.H. Kutner, Applied Linear StatisticalModels—Regression, Analysis of Variance and ExperimentalDesigns, second ed., R.D. Irwin, Homewood, 1985.

[29] M.C. Bittencourt-Oliveira, M.C. Oliveira, C.J.S. Bolch, Geneticvariability of some Brazilian strains of Microcystis aeruginosa

complex (Cyanophyceae/Cyanobacteria) using the nucleotidesequence analysis of the intergenic spacer and flanking regionsfrom cpcBA-phycocyanin operon, J. Phycol. 37 (2001) 810–818.

[30] M.C. Bittencourt-Oliveira, Development of Microcystis aerugin-osa (Kutzing) Kutzing (Cyanophyceae/Cyanobacteria) undercultivation and its taxonomic implications, Algolog. Stud. 99(2000) 27–37.

[31] S.H. White, L.D. Fabbro, L.J. Duivenvoorden, Changes inCyanoprokaryote populations, Microcystis morphology, andmicrocystin concentrations in lake Elphinstone (Central Queens-land, Australia), Environ. Toxicol. 18 (2003) 403–412.

[32] M. Namikoshi, K.L. Rinehart, R. Sakai, R.R. Stotts, A.M.Dahlem, V.R. Beasley, W.W. Carmichael, E.R. Evans, Identifi-cation of 12 hepatotoxins from a Homer bloom of the cyanobac-teria Microcystis aeruginosa, Microcystis viridis, and Microcystis

wesenbergii: nine new microcystins, J. Org. Chem. 57 (1992) 866–872.

[33] C. Edwards, L.A. Lawton, K.A. Beattie, G.A. Codd, Analysis ofmicrocystins from cyanobacteria by liquid chromatography withmass spectrometry using atmospheric pressure ionization, RapidCommun. Mass Spectrom. 7 (1993) 714–721.

[34] C. Hummert, M. Reichelt, J. Weiss, H.P. Liebert, B. Luckas,Identification of microcystins in cyanobacteria from the Bleilochformer drinking-water reservoir (Thuringia, Germany), Chemo-sphere 44 (2001) 1581–1588.

[35] C. Vezie, J. Rapala, J. Vaitomaa, J. Seitsonen, K. Sivonen, Effectof nitrogen and phosphorus on growth of toxic and nontoxicMicrocystis strains and on intracellular microcystin concentra-tions, Microb. Ecol. 43 (2002) 443–454.

[36] S. Jahnichen, T. Petzoldt, J. Benndorf, Evidence for control ofmicrocystin dynamics in Bautzen Reservoir (Germany) by cyano-bacterial population growth rates and dissolved inorganic carbon,Arch. Hydrobiol. 150 (2001) 177–196.

[37] H. Utkilen, N. Gjølme, Iron-stimulated toxin production inMicrocystis aeruginosa, Appl. Environ. Microbiol. 61 (1995) 797–800.

[38] H.M. Oh, S.J. Lee, M.H. Jang, B.D. Yoon, Microcystin produc-tion by Microcystis aeruginosa in a phosphorous-limited chemo-stat, Appl. Environ. Microbiol. 66 (2000) 176–179.

[39] P.T Orr, G.J. Jones, Relationship between microcystin productionand cell division rates in nitrogen-limited Microcystis aeruginosa

cultures, Limnol. Oceanogr. 43 (1998) 1604–1614.