Ecological Dynamics of Toxic Microcystis spp. and Microcystin-Degrading Bacteria in Dianchi Lake, China

Lin Zhu,a,b Yanlong Wu,a,b Lirong Song,a Nanqin Gana

State Key Laboratory of Freshwater Ecology and Biotechnology, Institute of Hydrobiology, The Chinese Academy of Sciences, Wuhan, People’s Republic of Chinaa;University of Chinese Academy of Sciences, Beijing, People’s Republic of Chinab

Toxic cyanobacterial blooms directly threaten both human safety and the ecosystem of surface waters. The widespread occur-rence of these organisms, coupled with the tumor-promoting properties of the microcystin toxins that they produce, demandsaction to mitigate their potential impacts and, thus, a robust understanding of their ecological dynamics. In the present work,the abundance of toxic Microcystis spp. and microcystin (MC)-degrading bacteria in Dianchi Lake, located in Yunnan Province,China, was studied using quantitative PCR. Samples were taken at monthly intervals from June 2010 to December 2011 at threesampling stations within this freshwater lake. Results revealed that variation in the abundance of both total Microcystis spp. andtoxic Microcystis spp. exhibited similar trends during the period of the algal bloom, including the reinvasion, pelagic growth,sedimentation, and overwintering periods, and that the proportion of toxic Microcystis was highest during the bloom and lowestin winter. Importantly, we observed that peaks in mlrA gene copy numbers of MC-degrading bacteria occurred in the monthsfollowing observed peaks in MC concentrations. To understand this phenomenon, we added MCs to the MC-degrading bacteria(designated strains HW and SW in this study) and found that MCs significantly enhanced mlrA gene copy numbers over thenumber for the control by a factor of 5.2 for the microcystin-RR treatment and a factor of 3.7 for the microcystin-LR treatment.These results indicate that toxic Microcystis and MC-degrading bacteria exert both direct and indirect effects on each other andthat MC-degrading bacteria also mediate a shift from toxic to nontoxic populations of Microcystis.

Algal blooms are a frequent and problematic feature of manyfreshwater bodies around the world. Toxic cyanobacterial

blooms directly threaten both human safety and the ecologicalquality of surface waters (1). Hepatotoxic cyanobacterial bloomsoccur more frequently than neurotoxic blooms (2). In China,most hepatotoxic blooms are caused by cyanobacteria belongingto the genus Microcystis, which produce cyclic heptapeptide hepa-totoxin microcystins (MCs). The widespread occurrence of thesebacteria, coupled with the tumor-promoting properties of themicrocystin toxins that they produce, demands action to mitigatetheir potential impacts and, thus, a robust understanding of theecological dynamics of the toxic blooms.

The typing of single colonies by matrix-assisted laser desorp-tion ionization–time of flight mass spectrometry (MALDI-TOFMS) (3, 4) can identify toxic and nontoxic Microcystis organismsand is a valuable tool for ecological studies of the genus Microcys-tis. However, this method requires the use of very expensiveequipment that is not always available (3). An alternative ap-proach, using molecular genotyping via PCR techniques, has beensuccessfully applied to the detection of the microcystin biosynthe-sis mcy gene cluster and has revealed a direct relationship betweengene expression and toxin production (5, 6). Quantitative reversetranscription-PCR (qRT-PCR) methods for identifying bacterialentities have also been established for Microcystis and Anabaena(7–9). Furthermore, Janse et al. (10) characterized toxic and non-toxic Microcystis colonies in natural populations using rRNA-in-ternal transcribed spacer denaturing gradient gel electrophoresis(rRNA ITS-DGGE).

Microcystins are stable in water, and microcystin-LR is evenresistant to temperatures up to 300°C and to extremes in pH (11).MCs present in water, however, eventually undergo biodegrada-tion and photolysis. Jones et al. (12) isolated the first MC-degrad-ing bacterium, MJ-PV, from Australian water bodies. Since then,

large numbers of MC-degrading bacteria have also been isolatedfrom natural water (13–18). In addition, the mlr gene cluster(mlrA, mlrB, mlrC, and mlrD) of the MC-degrading bacteriumSphingomonas sp. strain ACM-3962 has been demonstrated to en-code proteins involved in the initial steps of microcystin biodeg-radation (19, 20). In particular, the mlrA gene has been shown toencode an enzyme responsible for the hydrolytic cleavage of thecyclic structure of MCs (20). Conventional PCR techniques areable to identify the mlrA gene and detect MC-degrading bacteriafrom field water and from the biofilms of biofilters (21, 22). However,these conventional PCR assays cannot reveal the abundance of themlrA gene that is present in a sample. Recently, a quantitative mlrAgene-directed TaqMan PCR assay that can rapidly detect MC-de-grading bacteria via the use of degenerate oligonucleotides that targetconserved DNA regions has been developed (22). In this study, theTaqMan PCR assay was used to quantitatively detect total Microcystisspp., toxic Microcystis spp., and MC-degrading bacteria using primersthat target the 16S rRNA gene, mcyD genes, and the microcystin deg-radation protein-encoding gene (mlrA).

Most previous studies of algal bloom dynamics have focusedon their relationship to physicochemical factors affecting growthin the aquatic environment or on quantifying changes in mcy genecopy numbers that occur during blooms (23). To date, little is

Received 5 September 2013 Accepted 3 January 2014

Published ahead of print 10 January 2014

Editor: C. R. Lovell

Address correspondence to Nanqin Gan, gannq@ihb.ac.cn, or Lirong Song,lrsong@ihb.ac.cn.

Copyright © 2014, American Society for Microbiology. All Rights Reserved.

doi:10.1128/AEM.02972-13

1874 aem.asm.org Applied and Environmental Microbiology p. 1874 –1881 March 2014 Volume 80 Number 6

on June 27, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

known about the effects of toxic Microcystis spp. and microcystinson the abundance of MC-degrading bacteria and vice versa.Therefore, this study aimed to characterize the quantitative andqualitative effects of toxic Microcystis and microcystins on MC-degrading bacterial abundance and mlrA gene expression.

MATERIALS AND METHODSStudy site and sampling. Sampling was conducted at three sampling sta-tions (station D13 [24°57=39.17�N, 102°38=45.40�E], station D22[24°49=48.00�N, 102°42=47.00�E], and station D24 [24°41=58.00�N,102°39=53.00�E]) in Dianchi Lake, the largest lake (300 km2) in China’sYunnan Province (Fig. 1). Dianchi Lake is a plateau lake that has a verysmall watershed for a relatively large water surface area. Sampling stationswere visited once a month from June 2010 to December 2011. Watersamples were collected from a depth of 0.5 m using a van Dorn bottle. Thevan Dorn bottle was oriented vertically for sampling. Water samples (50ml) were filtered through 0.22-�m-pore-size polycarbonate membranefilters (Xinya Factory, Shanghai, China), transported on ice, and stored at�20°C prior to DNA extraction. Samples used for MC detection werefiltered through a 0.45-�m-pore-size glass fiber membrane (Jingteng,China). Ten milliliters of this filtrate was stored at �20°C.

Microcystin analysis. Dissolved microcystins were measured induplicate using an enzyme-linked immunosorbent assay (ELISA) kit(manufactured by the Institute of Hydrobiology [IHB], Chinese Academyof Sciences [CAS], China). The ELISA kit used a microcystin monoclonalantibody that reacts with most microcystin variants as well as with micro-cystin-LR and had a detection limit of 0.1 �g/liter. Further details regard-ing the use of the ELISA are provided by Lei et al. (24).

DNA extraction. DNA extraction from Microcystis was performed ac-cording to the protocol of Rinta-Kanto et al. (8), with slight modification.Briefly, sterilized scissors were used to fragment (cut) the filtrated mem-

branes into sections that were suspended in 2 ml of lysis buffer (10 mMTris, 100 mM NaCl, 1 mM EDTA, pH 9) and stirred constantly with apipette tip to ensure complete separation of the filter. To this we thenadded 20 �l of lysozyme (100 mg/ml) and incubated the solution at 37°Cfor 20 min. After incubation, 6 �l of proteinase K (20 mg/ml) and 105 �lof 10% sodium dodecyl sulfate (SDS) were added, and the cell suspensionwas then incubated at 50°C for 2 h. DNA was extracted by adding a volumeof phenol-chloroform-isoamyl alcohol (25:24:1) equal to the volume ofthe aqueous phase, mixing the solution gently for 2 min, and then centri-fuging the solution at 11,000 rpm for 20 min. A subsequent extraction ofthe aqueous phase was then undertaken, adding an equal volume of chlo-roform-isoamyl alcohol (24:1), mixing the solution, and again centrifug-ing at 11,000 rpm for 20 min. The upper aqueous phase was transferred toa new tube to which absolute ethanol (2� the aqueous-phase volume) and10 M ammonium acetate (0.1� the aqueous-phase volume) were thenadded. This mixture was precipitated overnight (at �20°C). Followingcentrifuging at 11,000 rpm for 30 min, the supernatant was removed andthe precipitated DNA was washed with cold 70% ethanol and then cen-trifuged again at 11,000 rpm for 10 min. DNA pellets were air dried andsubsequently resuspended in 50 �l deionized, distilled water and stored at�20°C. The concentration and purity of the extracted DNA were mea-sured with a spectrophotometer (NanoDrop Technologies, Wilmington,DE). This process was repeated in triplicate for each sample.

DNA extraction from MC-degrading bacteria, meanwhile, was under-taken with a PowerWater DNA isolation kit (Mo Bio Laboratories Inc., Carls-bad, CA) according to the manufacturer’s instructions. Again, DNA extrac-tion was repeated in triplicate with three filters for each sample. Each DNAextract was used as the PCR template for quantitative PCR (qPCR).

Standard curve preparation. To quantify the abundance of totalMicrocystis spp., toxic Microcystis spp., and MC-degrading bacteria in Di-anchi Lake, a qPCR assay was performed using primers based on 16SrRNA genes, mcyD genes, and mlrA genes, according to previously de-scribed methods (6, 9, 22). Microcystis aeruginosa PCC7806 was used asthe standard strain for the quantification of both total Microcystis spp. andtoxic Microcystis spp. The Sphingomonas sp. of MC-degrading bacteriathat was isolated from field water was used to quantify the MC-degradingbacteria (25). The standard curve was established by correlating knownDNA concentrations (in cell equivalents) with the threshold cycle (CT)values of the diluted samples (26). Ten milliliters of the PCC7806 straincontaining 1.71 � 107 cells/ml (determined by a direct microscopiccount) was filtered through 0.22-�m-pore-size polycarbonate membranefilters (Xinya Factory, Shanghai, China). A series of 10-fold dilutions(ranging from 3.4 � 106 cells to 3.4 cells) of the PCC7806 DNA templatestandard solution was used as the external standard for the qPCR. Tenmilliliters of the Sphingomonas sp. containing 1.2 � 109 cells/ml (deter-mined by serial dilution of the count obtained by the coated plate method)was collected using the same method used for strain PCC7806. Again, a seriesof 10-fold dilutions (ranging from 1.2 � 108 cells to 1.2 � 102 cells) of theSphingomonas sp. was used as the external standard for the qPCR. CT

calculations were completed automatically for each qPCR assay using iQ5software (version 2.0; Bio-Rad, Hercules, CA) and the maximum corre-lation coefficient approach. In this approach, the threshold is automati-cally determined to obtain the highest possible correlation coefficient (r2)for the standard curve.

TaqMan quantitative PCR conditions. We used a Bio-Rad cyclerequipped with the iQ5 real-time fluorescence detection system and soft-ware (version 2.0; Bio-Rad) to amplify and quantify gene copies. All reac-tions were completed in a total volume of 20 �l comprising 0.5 mM eachprimer, 0.1 mM TaqMan probe (Invitrogen, CA), 10 �l Bestar real-timePCR master mix (DBI Bioscience, China), 1 �l bovine serum albumin (3mg/ml; Sigma), double-distilled H2O, and template DNA. The primersused for amplification of total Microcystis spp., toxic Microcystis spp., andMC-degrading bacteria are listed in Table 1. Three separate assays wereperformed to quantify the Microcystis 16S rRNA gene, mcyD, and qmlrAfor all samples. The qPCR program for Microcystis 16S rRNA and mcyD

FIG 1 Map showing sampling stations in Dianchi Lake, Yunnan Province,China: site D13, 24°57=39.17�N, 102°38=45.40�E; site D22, 24°49=48.00�N,102°42=47.00�E; site D24, 24°41=58.00�N, 102°39=53.00�E.

Dynamics of Toxic Freshwater Algal Blooms

March 2014 Volume 80 Number 6 aem.asm.org 1875

on June 27, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

(toxic Microcystis/total Microcystis) was as follows: 95°C for 2 min, fol-lowed by 45 cycles of 95°C for 30 s and 55°C for 1 min. The qPCR programfor qmlrA was as follows: 95°C for 2 min, followed by 45 cycles of 95°C for15 s and 62°C for 45 s. All PCRs were run in triplicate on 96-well plates(Bio-Rad) sealed with optical-quality sealing tape (Bio-Rad). Three neg-ative controls without DNA were included for each PCR run.

RNA extraction, cDNA synthesis, and qRT-PCR amplification. TwoMC-degrading bacteria (designated strains HW and SW) were isolatedfrom field water and identified as a Sphingomonas sp. and a Sphingopyxissp. HW and SW were cultured with physiological saline and treated with-out and with MC-LR and MC-RR each day. This was done at a finalconcentration of 100 �g/liter while the cultures were shaken at 150 rpm at28°C in a 50-ml glass flask with a volume of 25 ml. Two milliliters ofculture was taken from the flasks each day and centrifuged (11,000 rpmfor 3 min at 4°C).

Total RNA was extracted from 2 ml of the cultured cell suspensionusing an E.Z.N.A. bacterial RNA kit (Omega). The amount and purity ofthe extracted RNA were determined using comparison of the optical den-sity at 260 nm and the optical density at 280 nm and agarose gel electro-

phoresis to evaluate integrity. After digestion with DNase I, 2 �g of totalRNA was reverse transcribed using a RevertAid first-strand cDNA synthesiskit (Thermo Scientific). Two pairs of specific primers (Table 1) were used toquantify the number of copies of the mlrA and the 16S rRNA genes, respec-tively. The RT-PCR commenced with 95°C for 2 min, followed by 45 cycles of95°C for 15 s and 62°C for 45 s. The mRNA copy number was determinedusing the CT value. The gene expression ratio was calculated by the 2���CT

method according to the handbook for the Bio-Rad real-time PCRsystem, where ��CT � (CT, target gene � CT, 16S rRNA)stress � (CT, target gene �C

T, 16S rRNA)control. All assays were performed in triplicate, and analyses were

conducted on means � standard deviations (SDs). Analyses were conductedwith Origin software (version 8.0; OriginLab).

RESULTSTotal Microcystis abundance and environmental parameters.Several cyanobacterial blooms occurred in Dianchi Lake during thesampling period (Table 2). Total Microcystis abundance remainedbetween 2.84 � 107 copies/liter and 5.29 � 108 copies/liter for all

TABLE 1 Oligonucleotides used as primers for qPCR and qRT-PCR

DNA target Primer Sequencea (5=–3=) Reference

Microcystis 16S rRNA 184F GCCGCRAGGTGAAAMCTAA 40431R AATCCAAARACCTTCCTCCCProbe (TaqMan) FAM-AAGAGCTTGCGTCTGATTAGCTAGT-BHQ-1 8

mcyD F2 GGTTCGCCTGGTCAAAGTAA 41R2 CCTCGCTAAAGAAGGGTTGAProbe (TaqMan) FAM-ATGCTCTAATGCAGCAACGGCAAA-BHQ-1 8

mlrA qmlrAf AGCCCKGGCCCRCTGC 22qmlrAr ATGCCARGCCCACCACATProbe (TaqMan) FAM-TGCCSCAGCTSCTCAAGAAGTTTG-BHQ-1

Bacterial 16S rRNA BACT1369F CGGTGAATACGTTCYCGG 42PROK1492R GGWTACCTTGTTACGACTT

a FAM, 6-carboxyfluorescein; BHQ-1, black hole quencher 1.

TABLE 2 Microcystis abundance and environmental parameters in Dianchi Lake from June 2010 to December 2011a

Sampling date(yr/mo/day)

No. of Microcystiscopies liter�1

Chla concn(�g/liter) Temp (°C)

Concn (mg/liter)

TP TN SRP TDP

2010/06/19 1.12E08 80.881 21.233 0.467 5.682 0.025 0.0392010/07/16 1.04E08 187.649 23.267 0.323 4.766 0.011 0.0262010/08/17 1.45E08 96.600 23.500 0.227 3.479 0.022 0.0552010/09/15 4.74E08 169.61 21.967 0.311 3.245 0.021 0.0182010/10/16 2.28E08 103.92 18.833 0.418 3.613 0.025 0.0262010/11/19 3.27E08 83.577 14.933 0.205 1.833 0.033 0.0362010/12/16 2.40E08 78.823 12.867 0.350 4.416 0.035 0.0462011/01/14 1.10E08 51.08 10.100 0.257 4.935 0.064 0.0832011/02/19 2.84E07 29.595 13.100 0.196 4.349 0.041 0.0362011/03/18 4.22E07 65.567 13.833 0.218 4.517 0.032 0.0542011/04/20 4.52E07 122.553 18.000 0.207 4.294 0.011 0.0172011/05/18 5.92E07 65.804 20.933 0.335 6.475 0.000 0.0292011/06/23 1.93E08 162.403 22.600 0.225 3.942 0.004 0.0182011/07/22 1.68E08 249.067 23.167 0.313 4.628 0.000 0.0302011/08/17 1.47E08 277.841 22.100 0.332 5.105 0.019 0.0382011/09/20 5.29E08 139.965 20.067 0.237 3.580 0.010 0.0372011/10/14 2.30E08 147.229 19.367 0.238 3.813 0.000 0.0382011/11/21 3.49E08 100.312 15.167 0.234 3.587 0.016 0.0972011/12/15 5.29E08 91.315 11.367 0.244 4.137 0.004 0.039a Data are averages of three sampling sites. Chla, chlorophyll a; TP, total phosphorus; TN, total nitrogen; SRP, solubility-reactive phosphorus; TDP, total dissolved phosphorus.

Zhu et al.

1876 aem.asm.org Applied and Environmental Microbiology

on June 27, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

these blooms, while the water temperature remained between 10°Cand 24°C. The total nitrogen content recorded during blooms variedfrom 1.833 mg/liter to 6.475 mg/liter, while the total phosphorusrecorded varied from 0.196 mg/liter to 0.418 mg/liter.

Seasonal variation in abundance of total and toxic Microcys-tis spp. Our results showed that large Microcystis blooms started fromJune to December and a peak occurred in September, while smallerMicrocystis blooms occurred from February to April. Microcystis 16SrRNA gene copy numbers ranged from 1.54 � 106 copies/liter to2.59 � 108 copies/liter at site D13, 4.62 � 106 copies/liter to 1.21 �109 copies/liter at site D22, and 8.65 � 106 copies/liter to 1.27 �109

copies/liter at site D24 (Fig. 2). Copy numbers of the mcyD generevealed a trend similar to that seen for Microcystis spp. (5.34 � 105

copies/liter to 4.89 � 107 copies/liter at site D13, 2.74 � 106 copies/liter to 1.77 � 108 copies/liter at site D22, and 1.67 � 105 copies/literto 2.22 � 107 copies/liter at site D24; Fig. 2). Meanwhile, we foundthat the proportions of toxic Microcystis were low from Septemberthrough April but then increased from May to reach a maximum inJune and July at all three sites (Fig. 2). These results indicate that cellnumbers of both toxic Microcystis and total Microcystis exhibit similartrends during the entire period of the algal bloom, including the re-invasion, pelagic growth, sedimentation, and overwintering periods(27), and that the proportion of toxic Microcystis was highest duringblooms and lowest in winter.

Seasonal variation in the abundance of microcystins andMC-degrading bacteria. Our results showed that at all threesampling sites, the MC concentration exhibited two distinctpeaks in September 2010 and October 2011 (Fig. 3). Concen-trations of MCs were higher in October 2011 than in September2010 at two of the three sites (September 2010 concentrationswere 1.33 �g/liter at site D13, 1.3 �g/liter at site D22, and 1.63�g/liter at site D24; October 2011 concentrations were 1.421�g/liter at site D13, 1.39 �g/liter at site D22, and 1.53 �g/literat site D24) (Fig. 3).

We found that mlrA gene copy numbers at site D13 rangedfrom 4.2 � 104 copies/liter to 3.49 � 108 copies/liter and exhibitedtwo periods of increase: from June 2010 to October 2010 and fromApril 2011 to November 2011 (observations were not obtained forsamples from December 2010 to March 2011, as copy numbers atthis time were below the detection limit, i.e., 6.1 � 103 copies/liter) (Fig. 3). A similar pattern was observed at both sites D22 andD24, in which mlrA gene copy numbers ranged from 2.4 � 104

copies/liter to 4.44 � 107 copies/liter at site D22 and from 8.8 �104 copies/liter to 9.62 � 107 copies/liter at site D24, with twoperiods of increase from June 2010 to October 2010 and from May2011 to November 2011 (observations were not obtained for sam-ples collected between January 2011 and April 2011 at these twosites). Peaks in the microcystin concentration were apparent inSeptember 2010 and October 2011, while peaks in the mlrA genecopy numbers of MC-degrading bacteria appeared in the follow-ing months (October 2010 and November 2011).

Effects of MCs on mlrA gene expression. Our results showedthat the rate of MC degradation increased when MCs were addedtwice, suggesting that MC degradation activity is promoted by theintermittent addition of MC-LR and MC-RR to cultures ofstrain HW and strain SW (Fig. 4). We found that over 4 days,levels of expression of the strain HW mlrA gene increased 3.7times under MC-LR stimulation and 5.2 times under MC-RRstimulation (Fig. 4). In contrast, the levels of expression of thestrain SW mlrA gene increased marginally (Fig. 4).

FIG 2 Seasonal variation in the abundance of total and toxic Microcystis spp.at each of the three sampling stations, D13 (a), D22 (b), and D24 (c), inDianchi Lake from June 2010 to December 2011. The numbers of DNA copieswere determined using qPCR. Error bars represent standard deviations. Thetoxic proportion was determined by dividing the relative number of copies ofthe mcyD gene by the total number of copies in Microcystis determined with theMicrocystis 16S rRNA primer set.

Dynamics of Toxic Freshwater Algal Blooms

March 2014 Volume 80 Number 6 aem.asm.org 1877

on June 27, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

DISCUSSION

Our analysis of the variation in the abundance of toxic Microcystisand MC-degrading bacterial communities in Dianchi Lake re-vealed that the cell numbers of Microcystis spp. and the mcyD gene

FIG 3 Seasonal variation in the abundance of dissolved MCs and MC-degradingbacteria at each of the three sampling stations, D13 (a), D22 (b), and D24 (c), inDianchi Lake from June 2010 to December 2011. The numbers of mlrA genecopies were determined using qPCR. Error bars represent standard deviations.

FIG 4 Effect of MCs on mlrA gene expression. (a) Changes in MC concentra-tions with time (LR100 and RR100, MC-LR and MC-RR at 100 �g/liter, re-spectively); (b) changes in mlrA gene expression at different concentrations ofMC-LR; (c) changes in mlrA gene expression at different concentrations ofMC-RR.

Zhu et al.

1878 aem.asm.org Applied and Environmental Microbiology

on June 27, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

varied in parallel, while the abundance of MC-degrading bacteriaincreased immediately after MC concentrations had reached theirmaximum. These findings suggest that the abundance of toxicMicrocystis and microcystins in this lake affects the abundance ofMC-degrading bacteria.

To analyze the seasonal dynamics of toxic Microcystis spp.and the MC-degrading bacterial community in field-collectedwater, TaqMan qPCR was used to quantify total Microcystis cellnumbers, toxic Microcystis cell numbers, and MC-degradingbacterial abundance under the assumption that each M. aerugi-nosa PCC7806 cell has a constant copy number of the Micro-cystis 16S rRNA and mcyD genes. Previous studies have re-ported that Microcystis strains have one mcyD gene copy pergenome (28), whereas genome sequencing analyses have foundthat a single Microcystis cell contains two copies of the rRNAgene cluster (29). Hence, the percentage of toxic Microcystisorganisms is likely to be overestimated by qPCR. This problemhas been discussed in detail by Rinta-Kanto et al. (8). Despitethis possibility, our results nonetheless provide a snapshot ofthe relative dominance of the microcystin-producing strainswithin this lake community.

In Dianchi Lake, eutrophication has been evident since the late1980s as a result of the escalating input of untreated wastewaterand municipal sewage. This has resulted in dense Microcystisblooms (�108 cells liter�1 in summer) and microcystin contam-ination from April to November (30, 31). The results of this studyreveal that the proportion of toxic Microcystis cells in the lakevaried from 2.9% to 93.8% and was low from September to Apriland reached a maximum in June and July. We also were able toconfirm that the largest Microcystis blooms always occurred fromJune to December (108 copies/liter to 109 copies/liter), while thesmallest blooms occurred from February to April (106 copies/li-ter). This result is consistent with the findings of Davis et al. (32),who showed that eutrophication and climatic warming may actadditively to promote the growth of toxic, rather than nontoxic,populations of Microcystis, leading to blooms with higher MCcontents.

Several studies have shown that Alphaproteobacteria, Beta-proteobacteria, Gammaproteobacteria, Bacteroidetes, Firmicutes,Deinococcus-Thermus, and Gemmatimonadetes are involved in thebacterial community diversity associated with Microcystis blooms(33–35). A GenBank and Ribosomal Database Project (RDP)analysis showed that the two MC-degrading bacteria (designatedHW and SW) that we isolated were very similar to samples withthe accession numbers AB161684.1 and JF833116, which belongto the class Alphaproteobacteria, order Sphingomonadales, familySphingomonadaceae, genus Sphingopyxis. A number of molecularstudies of bacterial communities associated with Microcystisblooms performed using targeted, 16S rRNA-type and meta-genomic analyses have recently been reported. The use of targetedSphingomonas showed that the Sphingomonadales seem to be anintegral element of Microcystis blooms and can effectively degradeMCs as a result of their breakdown of the toxins (36). Comparisonof the metagenomes of free-living bacterial plankton assemblagesfrom Lake Erie in the United States showed that a diverse array ofbacterial phyla was responsive to an elevated supply of MCs. Thisalso included Sphingomonadales, although they were present atlower levels than other bacteria (37).

A key result of this study was that peaks in MC concentra-tions were followed (in the following months) by peaks in mlrA

gene copy numbers of MC-degrading bacteria, which suggeststhat MCs released in field water may affect the MC-degradingbacterial communities in Dianchi Lake. Due to a sampling fre-quency of once per month, MC-degrading bacteria could po-tentially appear and again disappear in this time span. In addi-tion, the concentrations of MCs in the lake also increased as theproportion of toxic Microcystis increased. These observationssuggest that there is a strong association between the propor-tion of toxic Microcystis in the lake and the concentration ofMCs and MC-degrading bacteria. Jones et al. (12) reportedthat, under experimental conditions, isolates of MC-degradingbacteria require a lag time to initiate degradation of MCs whenthey have not previously been exposed to MCs. Maruyama et al.(38), meanwhile, assumed that MC-degrading bacteria re-sponded to changes in the concentration of MCs and began todegrade MCs as they were released from Microcystis cells. Theyspeculated that MC-degrading bacteria in the mucilage re-mained on standby until the degradation of MCs occurred (i.e.,they could be directly exposed to MCs released from cells in thewater bloom). The findings of this study provide empiricalsupport for these previous field study-based assumptions,demonstrating that seasonal variation in the abundance ofMC-degrading bacteria in freshwater is related to the concen-trations of both toxic Microcystis and MCs in the water.

These findings further suggest that MCs may act as sourcesof nutrition for MC-degrading bacteria. Previous studies pro-posed that MCs are used as secondary substrates by MC-de-grading bacteria that come into contact with them (17, 22).Microorganisms accumulate in polysaccharide matrices andform structural and functional microbial assemblies on sur-faces submerged in water commonly known as biofilms. Li etal. (39) found that the indigenous MC-degrading bacteria inbiofilm can accumulate and become activated with MC-LRpresent, thereby increasing the efficiency with which MC deg-radation occurs. Furthermore, Li et al. (4) reported that mlrAgene abundance increased with increasing MC-LR concentra-tions, suggesting that biodegradation of MCs depends on thepopulation of indigenous MC-degrading bacteria which pri-marily use MCs over other nutrients. In this study, we foundthat mlrA gene abundances increased under high concentra-tions of MCs, indicating that MC-degrading bacteria can useMCs for growth, without requiring other nutrients.

In conclusion, the data presented here show that large Micro-cystis blooms are correlated with a high abundance of toxic Micro-cystis, with concentrations of both microorganisms influencingthe abundance of MC-degrading bacteria. These results suggestthat toxic Microcystis and MC-degrading bacteria exert both indi-rect and direct effects on one another.

ACKNOWLEDGMENTS

This work was supported by grants from the National Natural ScienceFoundation of China (31370418), the Chinese Academy of Sciences(KSCX2-EW-Z-3), and the Natural Science Foundation of China-Yun-nan Project (U0833604).

REFERENCES1. Jochimsen EM, Carmichael WW, An JS, Cardo DM, Cookson ST,

Holmes CE, Antunes MB, de Melo Filho DA, Lyra TM, Barreto VS,Azevedo SM, Jarvis WR. 1998. Liver failure and death after exposure tomicrocystins at a hemodialysis center in Brazil. N. Engl. J. Med. 338:873–878. http://dx.doi.org/10.1056/NEJM199803263381304.

Dynamics of Toxic Freshwater Algal Blooms

March 2014 Volume 80 Number 6 aem.asm.org 1879

on June 27, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

2. Sivonen K, Niemelä S, Niemi R, Lepistö L, Luoma T, Räsänenl L.1990. Toxic cyanobacteria (blue-green algae) in Finnish fresh andcoastal waters. Hydrobiologia 190:267–275. http://dx.doi.org/10.1007/BF00008195.

3. Kurmayer R, Dittmann E, Fastner J, Chorus I. 2002. Diversity of mi-crocystin genes within a population of the toxic cyanobacterium Micro-cystis spp. in Lake Wannsee (Berlin, Germany). Microb. Ecol. 43:107–118.http://dx.doi.org/10.1007/s00248-001-0039-3.

4. Li J, Shimizu K, Maseda H, Lu Z, Utsumi M, Zhang Z, Sugiura N. 2012.Investigations into the biodegradation of microcystin-LR mediated by thebiofilm in wintertime from a biological treatment facility in a drinkingwater treatment plant. Bioresour. Technol. 106:27–35. http://dx.doi.org/10.1016/j.biortech.2011.11.099.

5. Meissner K, Dittmann E, Börner T. 1996. Toxic and non-toxic strainsof the cyanobacterium Microcystis aeruginosa contain sequences ho-mologous to peptide synthetase genes. FEMS Microbiol. Lett. 135:295–303.

6. Pan H, Song LR, Liu YD, Börner T. 2002. Detection of hepatotoxicMicrocystis strains by PCR with intact cells from both culture and envi-ronmental samples. Arch. Microbiol. 178:421– 427. http://dx.doi.org/10.1007/s00203-002-0464-9.

7. Vaitomaa J, Rantala A, Halinen K, Rouhiainen L, Tallberg P, MokelkeL, Sivonen1 K. 2003. Quantitative real-time PCR for determination ofmicrocystin synthetase E copy numbers for Microcystis and Anabaena inlakes. Appl. Environ. Microbiol. 69:7289 –7297. http://dx.doi.org/10.1128/AEM.69.12.7289-7297.2003.

8. Rinta-Kanto JM, Ouellette AJA, Boyer GL, Twiss MR, Bridgeman TB,Wilhelm SW. 2005. Quantification of toxic Microcystis spp. during the2003 and 2004 blooms in western Lake Erie using quantitative real-timePCR. Environ. Sci. Technol. 39:4198 – 4205. http://dx.doi.org/10.1021/es048249u.

9. Ha JH, Hidaka T, Tsuno H. 2009. Quantification of toxic Microcystis andevaluation of its dominance ratio in blooms using real-time PCR. Environ.Sci. Technol. 43:812– 818. http://dx.doi.org/10.1021/es801265f.

10. Janse I, Kardinaal WE, Meima M, Fastner J, Visser PM, Zwart G. 2004.Toxic and nontoxic Microcystis colonies in natural populations can bedifferentiated on the basis of rRNA gene internal transcribed spacer diver-sity. Appl. Environ. Microbiol. 70:3979 –3987. http://dx.doi.org/10.1128/AEM.70.7.3979-3987.2004.

11. Watanabe MF, Harada KI, Matsuura K, Watanabe M, Suzuki M. 1989.Heptapeptide toxin production during the batch culture of two Microcys-tis species (cyanobacteria). J. Appl. Phycol. 1:161–165. http://dx.doi.org/10.1007/BF00003879.

12. Jones GJ, Bourne DG, Blakeley RL, Doelle H. 1994. Degradation of thecyanobacterial hepatotoxin microcystin by aquatic bacteria. Nat. Toxins2:228 –235. http://dx.doi.org/10.1002/nt.2620020412.

13. Ishii H, Nishijima M, Abe T. 2004. Characterization of degradationprocess of cyanobacterial hepatotoxins by a gram-negative aerobic bacte-rium. Water Res. 38:2667–2676. http://dx.doi.org/10.1016/j.watres.2004.03.014.

14. Tsuji K, Asakawa M, Anzai Y, Sumino T, Harada K. 2006. Degradation ofmicrocystins using immobilized microorganism isolated in an eutrophic lake.Chemosphere 65:117–124. http://dx.doi.org/10.1016/j.chemosphere.2006.02.018.

15. Ho L, Hoefel D, Saint CP, Newcombe G. 2007. Isolation and identifi-cation of a novel microcystin-degrading bacterium from a biological sandfilter. Water Res. 41:4685– 4695. http://dx.doi.org/10.1016/j.watres.2007.06.057.

16. Manage PM, Edwards C, Singh BK, Lawton LA. 2009. Isolation andidentification of novel microcystin-degrading bacteria. Appl. Environ.Microbiol. 75:6924 – 6928. http://dx.doi.org/10.1128/AEM.01928-09.

17. Eleuterio L, Batista JR. 2010. Biodegradation studies and sequencing ofmicrocystin-LR degrading bacteria isolated from a drinking-water biofil-ter and a fresh water lake. Toxicon 55:1434 –1442. http://dx.doi.org/10.1016/j.toxicon.2010.02.020.

18. Jiang Y, Shao J, Wu X, Xu Y, Li R. 2011. Active and silent members in themlr gene cluster of a microcystin-degrading bacterium isolated from LakeTaihu, China. FEMS Microbiol. Lett. 322:108 –114. http://dx.doi.org/10.1111/j.1574-6968.2011.02337.x.

19. Bourne DG, Jones GJ, Blakeley RL, Jones A, Negri AP, Riddles P. 1996.Enzymatic pathway for the bacterial degradation of the cyanobacterialcyclic peptide toxin microcystin LR. Appl. Environ. Microbiol. 62:4086 –4094.

20. Bourne DG, Riddles P, Jones GJ, Smith W, Blakeley RL. 2001. Charac-terisation of a gene cluster involved in bacterial degradation of the cyano-bacterial toxin microcystin LR. Environ. Toxicol. 16:523–534. http://dx.doi.org/10.1002/tox.10013.abs.

21. Saito T, Okano K, Park HD, Itayama T, Inamori Y, Neilan BA, BurnsBP, Sugiura N. 2003. Detection and sequencing of the microcystin LR-degrading gene, mlrA, from new bacteria isolated from Japanese lakes.FEMS Microbiol. Lett. 229:271–276. http://dx.doi.org/10.1016/S0378-1097(03)00847-4.

22. Hoefel D, Adriansen CM, Bouyssou MA, Saint CP, Newcombe G, Ho L.2009. Development of an mlrA gene-directed TaqMan PCR assay forquantitative assessment of microcystin-degrading bacteria within watertreatment plant sand filter biofilms. Appl. Environ. Microbiol. 75:5167–5169. http://dx.doi.org/10.1128/AEM.00036-09.

23. Yoshida M, Yoshida T, Kashima A, Takashima Y, Hosoda N, NagasakiK, Hiroishi S. 2008. Ecological dynamics of the toxic bloom-formingcyanobacterium Microcystis aeruginosa and its cyanophages in freshwater.Appl. Environ. Microbiol. 74:3269 –3273. http://dx.doi.org/10.1128/AEM.02240-07.

24. Lei LM, Wu YS, Gan NQ, Song LR. 2004. An ELISA-like time-resolvedfluorescence immunoassay for microcystin detection. Clin. Chim. Acta348:177–180. http://dx.doi.org/10.1016/j.cccn.2004.05.019.

25. Gan NQ, Xiao Y, Zhu L, Wu ZX, Liu J, Song LR. 2012. The role ofmicrocystins in maintaining colonies of bloom-forming Microcystis spp.Environ. Microbiol. 14:730 –742. http://dx.doi.org/10.1111/j.1462-2920.2011.02624.x.

26. Furukawa K, Noda N, Tsuneda S, Saito T, Itayama T, Inamori Y. 2006.Highly sensitive real-time PCR assay for quantification of toxic cyanobac-teria based on microcystin synthetase A gene. J. Biosci. Bioeng. 102:90 –96.http://dx.doi.org/10.1263/jbb.102.90.

27. Tilo I, Sabine J, Jürgen B. 2005. Wax and wane of Microcystis (Cyano-phyceae) and microcystins in lake sediments: a case study in QutzdorfReservoir (Germany). J. Phycol. 41:479 – 488. http://dx.doi.org/10.1111/j.1529-8817.2005.00071.x.

28. Kaebernick M, Dittmann E, Borner T, Neilan BA. 2002. Multiple alter-nate transcripts direct the biosynthesis of microcystin, a cyanobacterialnonribosomal peptide. Appl. Environ. Microbiol. 68:449 – 455. http://dx.doi.org/10.1128/AEM.68.2.449-455.2002.

29. Kaneko T, Nakajima N, Okamoto S, Suzuki I, Tanabe Y, Tamaoki M,Nakamura Y, Kasai F, Watanabe A, Kawashima K, Kishida Y, Ono A,Shimizu Y, Takahashi C, Minami C, Fujishiro T, Kohara M, Katoh M,Nakazaki N, Nakayama S, Yamada M, Tabata S, Watanabe MM. 2007.Complete genomic structure of the bloom-forming toxic cyanobacteriumMicrocystis aeruginosa NIES-843. DNA Res. 14:247–256. http://dx.doi.org/10.1093/dnares/dsm026.

30. Botes DP, Wessels PL, Kruger H, Runnegar MTC, Santikarn S, SmithRJ, Barna JCJ, Williams DH. 1985. Structural studies on cyanoginosin-LR, -YR, -YA, and -YM, peptide toxins from Microcystis aeruginosa. J.Chem. Soc. Perkin Trans. 1, pp 2747–2748.

31. Li H, Hou G, Dakui F, Xiao B, Song L, Liu Y. 2007. Prediction andelucidation of the population dynamics of Microcystis spp. in Lake Dianchi(China) by means of artificial neural networks. Ecol. Informatics 2:184 –192. http://dx.doi.org/10.1016/j.ecoinf.2007.03.007.

32. Davis TW, Berry DL, Boyer GL, Gobler CJ. 2009. The effects of temper-ature and nutrients on the growth and dynamics of toxic and non-toxicstrains of Microcystis during cyanobacteria blooms. Harmful Algae 8:715–725. http://dx.doi.org/10.1016/j.hal.2009.02.004.

33. Wu X, Xi W, Ye W, Yang H. 2007. Bacterial community composition ofa shallow hypertrophic freshwater lake in China, revealed by 16S rRNAgene sequences. FEMS Microbiol. Ecol. 61:85–96. http://dx.doi.org/10.1111/j.1574-6941.2007.00326.x.

34. Shi L, Cai Y, Wang X, Li P, Yu Y, Kong F. 2010. Community structureof bacteria associated with Microcystis colonies from cyanobacterialblooms. J. Freshwater Ecol. 25:193–203. http://dx.doi.org/10.1080/02705060.2010.9665068.

35. Shi L, Cai Y, Kong F, Yu Y. 2011. Changes in abundance and com-munity structure of bacteria associated with buoyant Microcystis colo-nies during the decline of cyanobacterial bloom (autumn-winter tran-sition). Ann. Limnol. Int. J. Limnol. 47:355–362. http://dx.doi.org/10.1051/limn/2011047.

36. Dziallas C, Grossart H. 2011. Temperature and biotic factors influencebacterial communities associated with the cyanobacterium Microcystis sp.

Zhu et al.

1880 aem.asm.org Applied and Environmental Microbiology

on June 27, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

Environ. Microbiol. 13:1632–1641. http://dx.doi.org/10.1111/j.1462-2920.2011.02479.x.

37. Mou X, Lu X, Jacob J, Sun S, Robert H. 2013. Metagenomic identifica-tion of bacterioplankton taxa and pathways involved in microcystin deg-radation in Lake Erie. PLoS One 8:e61890. http://dx.doi.org/10.1371/journal.pone.0061890.

38. Maruyama T, Kato K, Yokoyama A, Tanaka T, Hiraishi A, Park HD. 2003.Dynamics of microcystin-degrading bacteria in mucilage of Microcystis. Mi-crob. Ecol. 46:279–288. http://dx.doi.org/10.1007/s00248-002-3007-7.

39. Li J, Shimizu K, Utsumi M, Nakamoto T, Sakharkar MK, Zhang Z,Sugiura N. 2011. Dynamics of the functional gene copy number andoverall bacterial community during microcystin-LR degradation by a bi-

ological treatment facility in a drinking-water treatment plant. J. Biosci.Bioeng. 111:695–701. http://dx.doi.org/10.1016/j.jbiosc.2011.02.007.

40. Neilan BA, Jacobs D, Del Dot T, Blackall LL, Hawkins PR, Cox PT,Goodman AE. 1997. rRNA sequences and evolutionary relationshipsamong toxic and nontoxic cyanobacteria of the genus Microcystis. Int. J.Syst. Bacteriol. 47:693– 697.

41. Kaebernick M, Neilan BA, Börner T, Dittmann E. 2000. Light and thetranscriptional response of the microcystin biosynthesis gene cluster.Appl. Environ. Microbiol. 66:3387–3392.

42. Suzuki MT, Taylor LT, DeLong EF. 2000. Quantitative analysis of small-subunit rRNA genes in mixed microbial populations via 5=-nuclease as-says. Appl. Environ. Microbiol. 66:4605– 4614.

Dynamics of Toxic Freshwater Algal Blooms

March 2014 Volume 80 Number 6 aem.asm.org 1881

on June 27, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from