*For correspondence. E-mail: firstname.lastname@example.org; Tel.: +55 19 3429-4128; Fax:
+55 19 3447-8649
§Supplemental material for this article may be found at
Copyright G2019, The Microbiological Society of Korea
Adriana Sturion Lorenzi1,2, Mathias Ahii Chia1,3,
Fabyano Alvares Cardoso Lopes2,4,
Genivaldo Gueiros Z. Silva2, Robert A. Edwards2,5,
and Maria do Carmo Bittencourt-Oliveir
1Laboratory of Cyanobacteria, Department of Biological Sciences, Luiz
de Queiroz College of Agriculture, University of São Paulo (USP),
Piracicaba, SP, Brazil
2Computational Science Research Center, San Diego State University,
San Diego, California, USA
3School of Marine and Atmospheric Sciences, Stony Brook University,
Southampton Campus, New York, USA
4Laboratory of Enzymology, Department of Cell Biology, University of
Brasília (UNB), Brasília, DF, Brazil
5Department of Computer Science, San Diego State University, San
Diego, California, USA
(Received Jun 27, 2018 / Revised Nov 13, 2018 / Accepted Dec 13, 2018)
Journal of Microbiology (2019) Vol. 57
Cyanobacterial biodiversity of semiarid public drinking water supply
reservoirs assessed via next-generation DNA sequencing technology§
Next-generation DNA sequencing technology was applied
to generate molecular data from semiarid reservoirs during
well-defined seasons. Target sequences of 16S-23S rRNA ITS
and cpcBA-IGS were used to reveal the taxonomic groups of
cyanobacteria present in the samples, and genes coding for
cyanotoxins such as microcystins (mcyE), saxitoxins (sxtA),
and cylindrospermopsins (cyrJ) were investigated. The pre-
sence of saxitoxins in the environmental samples was eval-
uated using ELISA kit. Taxonomic analyses of high-through-
put DNA sequencing data showed the dominance of the ge-
nus Microcystis in Mundaú reservoir. Furthermore, it was the
most abundant genus in the dry season in Ingazeira reservoir.
In the rainy season, 16S-23S rRNA ITS analysis revealed that
Cylindrospermopsis raciborskii comprised 46.8% of the cyano-
bacterial community in Ingazeira reservoir, while the cpcBA-
IGS region revealed that C. raciborskii (31.8%) was the most
abundant taxon followed by Sphaerospermopsis aphanizo-
menoides (17.3%) and Planktothrix zahidii (16.6%). Despite
the presence of other potential toxin-producing genera, the
detected sxtA gene belonged to C. raciborskii, while the mcyE
gene belonged to Microcystis in both reservoirs. The detected
mcyE gene had good correlation with MC content, while the
amplification of the sxtA gene was related to the presence of
STX. The cyrJ gene was not detected in these samples. Using
DNA analyses, our results showed that the cyanobacterial
composition of Mundaú reservoir was similar in successive
dry seasons, and it varied between seasons in Ingazeira re-
servoir. In addition, our data suggest that some biases of an-
alysis influenced the cyanobacterial communities seen in
the NGS output of Ingazeira reservoir.
Keywords:cyanotoxins, genotypic composition, NGS, pub-
lic water supply, water quality
Northeastern Brazil has well defined rainy and dry seasons
(Almeida et al., 2009), and water scarcity has worsened in
this semiarid region due to increased frequency and duration
of drought caused by changing global climatic conditions
(World Bank, 2013). The construction of reservoirs has re-
solved this problem by ensuring water is available during the
dry season and prolonged drought. However, the discharge
of domestic and industrial sewage eutrophies these water
bodies due to the nutrient-enriched conditions of the sewage.
Eutrophic conditions coupled with high water temperatures
and long water residence times have led to the excessive pro-
liferation of cyanobacteria in semiarid reservoirs over the
years (Chellappa and Costa, 2003; Costa et al., 2006; Bitten-
court-Oliveira et al., 2014). In Pernambuco (PE) State, 90%
of the reservoirs are eutrophic (Bouvy et al., 2000), and per-
ennially have blooms of Cylindrospermopsis raciborskii (Wol-
oszynska) Seenayya and Subba Raju, Microcystis aeruginosa
(Kützing) Kützing, Microcystis panniformis Komárek et al.,
Sphaerospermopsis torques-reginae (Komárek) Werner, Laug-
hinghouse IV, Fiore and Sant’Anna Klebahn (formerly de-
scribed as Anabaena spiroides), Sphaerospermopsis aphani-
zomenoides (Forti) Zapomělová, Jezberová, Hrouzek, Hizem,
Reháková and Komárková, and Planktothrix agardhii (Go-
mont) Anagnostidis and Komárek (Bouvy et al., 2001; Molica
et al., 2005; Bittencourt-Oliveira et al., 2014).
Among the secondary metabolites produced by cyanobac-
teria, cyanotoxins such as microcystins (MCs), cylindrosper-
mopsins (CYNs), and saxitoxins (STXs) have been detected
in Brazilian semiarid water supply reservoirs (Bittencourt-
Oliveira et al., 2011a, 2014) or in cyanobacterial strains iso-
lated from these aquatic systems (Borges et al., 2015). MCs
are assembled by multifunctional enzyme complexes con-
sisting of nonribosomal peptide synthetases (NRPSs) and
polyketide synthases (PKSs) (Christiansen et al., 2003; Rou-
hiainen et al., 2004). The MC gene cluster in the model cy-
anobacterium Microcystis aeruginosa PCC7806 spans 55 kb
of DNA composed of 10 (mcyABCDEFGHIJ) bidirectionally
transcribed open reading frames (ORFs) arranged in two pu-
tative operons, mcyA-C and mcyD-J (Tillet et al., 2000; Ras-
togi et al., 2015). Similarly, STXs and CYNs are synthesized
2Lorenzi et al.
by modular NRPSs and PKSs enzyme complexes (Dittmann
et al., 2013). In cyanobacteria such as Anabaena circinalis,
Aphanizomenon, Cylindrospermopsis raciborskii, Lyngbya
wollei, and Raphidiopsis brookii the STX gene cluster (25.7–
36.0 kb) consists of up to 33 genes encoding biosynthetic en-
zymes, transporters and regulatory proteins (Neilan et al.,
2013). The CYN gene cluster (cyr) spans 43 kb of DNA and
is comprised of 15 open reading frames containing genes re-
quired for the biosynthesis, regulation, and export of the
toxin (Mihali et al., 2008).
In response to the tragic incident in Caruaru (PE) – Brazil,
in 1996, where the death of several dialysis patients occurred
after exposure to MCs contaminated water (Jochimsen et al.,
1998; Azevedo et al., 2002), the Brazilian government pro-
posed a specific regulation (Brazil, 2011) that stipulates man-
datory monitoring of water reservoirs used for public supply.
This resolution establishes maximum permissible limits of
1 μg/L for MCs and 3 μg/L for SXTs, and weekly monitoring
when the cyanobacterial cell density is higher than 10,000
cells/ml (Brazil, 2011).
The objective of several studies has been to determine the
composition and ecology of cyanobacteria in the Brazilian
semiarid water bodies (Bittencourt-Oliveira et al., 2011b,
2012a; Lira et al., 2011; Dantas et al., 2012; Moura et al., 2012;
Fonseca et al., 2015). However, the molecular ecology of
toxin-producing cyanobacteria has been poorly addressed
(Bittencourt-Oliveira, 2003; Bittencourt-Oliverira et al., 2010,
2011a, 2012b; Borges et al., 2015; Lorenzi et al., 2015).
Recent advances in sequencing technologies allied with bio-
informatics tools have increased the possibility of large-scale
studies of cyanobacterial communities (Steffen et al., 2012;
D’Agostino et al., 2016). Next-Generation Sequencing (NGS)
technologies employed in metagenomics, comparative ge-
nomics, and metatranscriptomics have significantly contri-
buted to the understanding of the ecology and control of
harmful blooms of cyanobacteria (Li et al., 2011; Steffen et
al., 2012, 2015; Penn et al., 2014; Voorhies et al., 2016; Wood-
house et al., 2016; Walter et al., 2018). Therefore, the objec-
tives of the present study were to: (1) to characterize the di-
versity of cyanobacteria and cyanotoxin genes in Ingazeira
and Mundaú reservoirs (Brazil) during the rainy and/or dry
seasons using NGS, and (2) determine whether the detected
cyanotoxin genes were associated with the presence of toxins
in the investigated environmental samples. The Ingazeira and
Mundaú reservoirs are very important semiarid water bodies
that provide potable water to more than 300,000 people (SRH,
2000), and are perennially plagued with noxious bloom-for-
ming and potential toxin-producing cyanobacteria (Bitten-
court-Oliveira et al., 2014). This work represents an effort to
evaluate the Brazilian semiarid cyanobacterial composition
using NGS, and the results provide valuable data on the di-
versity and abundance of cyanobacteria in public water sup-
ply reservoirs in unique ecosystems worldwide. The public
availability of these data will serve as a basis for comparison
with other distinct environments, which will aid the compre-
hensive understanding of the cyanobacterial community of
Materials and Methods
Site description and sample collection
In addition to supplying drinking water to over 300,000 in-
habitants during periods of drought, which are characteristic
of the semiarid climate, Ingazeira and Mundaú reservoirs in
northeastern Brazil are used for irrigation, fishing, ranching
and bathing (Bouvy et al., 2000). The reservoirs are located
in the “Agreste” phytogeographic zone of Pernambuco State
(Supplementary data Fig. S1), which is characterized by re-
gular rainy and dry regimes, and an average yearly temper-
ature of 26°C. Cyanotoxins have been frequently detected in
these reservoirs (Lorenzi et al., 2018).
Surface water samples were previously collected preferen-
tially from the middle of Ingazeira (08°3641.2S; 36°54G
23.7W) and Mundaú (08°5721.1S; 36°3007.3W) reser-
voirs by Bittencourt-Oliveira et al. (2014). Water samples were
collected twice from Ingazeira (I1 - Apr 14 and I18 - Oct 13,
2009) and Mundaú (M1 - Mar 17 and M27 - Nov 09, 2009)
reservoirs using 20-μm mesh plankton net. Sample I1 was
collected in the rainy season, while samples I18, M1, and M27
were obtained in the dry season. Due to problems associated
with the sampling difficulties and sample availability, inter-
seasonal comparison was only possible for the Ingazeira re-
servoir but not the Mundaú reservoir. Further, the compari-
son was made only for samples from the same reservoir be-
cause there is a 58.57 km-straight line distance between Inga-
zeira and Mundaú reservoirs, and the NGS technology was
employed for site discovery to investigate related samples.
Metagenomic DNA extraction, PCR amplification, and next-
Environmental water samples were subjected to DNA ex-
traction immediately after samplings using a cetyltrimethyl-
ammonium bromide (CTAB)-based extraction method (Ro-
gers and Bendich, 1985) as described previously (Lorenzi et
al., 2015), and at least three biological replicates were car-
ried out per sample. The samples were stored at -20°C until
analysed. In the present study, three technical replicates of
PCR reactions targeting the 16S rRNA gene with the 16S–23S
intergenetic segment (Strunecký et al., 2011) and cpcBA-IGS
(Neilan et al., 1995) region were performed on pooled DNA
samples. In order to investigate the presence of microcystins,
cylindropermopsins, and saxitoxins genes, the mcyE (Rantala
et al., 2004), cyrJ (Mihali et al., 2008), and sxtA (Smith et al.,
2011) genes were targeted, respectively. PCR reactions were
performed using 10 ng of total DNA and pureTaq Ready-
To-Go PCR Beads kit (GE Healthcare) in a GeneAmp PCR
System 9700 Thermal Cycler (Applied Biosystems). PCR ther-
mocycling conditions were in accordance with those speci-
fied by the authors indicated above. Negative controls (with-
out DNA) were prepared using the same reaction conditions
and primers. Subsequently, amplicons stained with SYBR®
Gold Nucleic Acid Gel Stain (Invitrogen) were analyzed on
1.0% agarose gels after electrophoresis in 0.5 × TBE run-
ning buffer. PCR products were subjected to shotgun sequen-
cing. Paired-end libraries were prepared with the Nextera
XT DNA Library Prep Kit part# FC-131-1096 (Illumina) ac-
cording to the manufacturer’s instructions. The libraries were
Next-generation DNA sequencing of cyanobacterial blooms 3
sequenced on a MiSeq Personal Sequencing System (Illumina)
using the 500-cycle MiSeq reagent kit v2 (2 × 250) (Illumina).
An average of 60,238.9 reads per sample was obtained from
a subset of the original 12 samples.
Data preprocessing, annotation, and analyses
Paired-end reads were merged using PEAR software (Zhang
et al., 2013) to produce consensus sequences and increase the
annotation accuracy. Low-quality bases (quality score < 20)
from merged and unmerged sequences were trimmed from
both ends using the Phred algorithm with the python script.
Merged and unmerged trimmed sequences from the same
sample were concatenated into a single file for each sequence
category (16S-23S rRNA ITS, cpcBA-IGS, mcyE, and sxtA),
and the replicates were analyzed separately. These files were
uploaded to the Metagenomics Rapid Annotation (MG-
RAST) server (Meyer et al., 2008), and made publicly acces-
sible under the project accession ‘Metacommunity of cya-
nobacterial blooms – ESALQ USP / SDSU’ and accession
numbers 4676155.3 to 4676165.3 and 4676167.3.
Unassembled DNA sequences (250-bp length) of each sam-
ple dataset were annotated with the BLASTN 2.2.32+ soft-
ware (Altschul et al., 1997) using a cut-off E-value of 1e-5 and
97% minimum sequence identity against the NCBI-NT da-
tabase for downstream analyses. Tables of frequencies of the
hits to each target sequence for each sample dataset were ge-
nerated and normalized by dividing by the total number of
hits to remove bias difference in read lengths and sequencing
efforts. Heat maps, which are graphical representations of
individual values contained in a matrix in colors, were con-
structed from the predicted relative abundance of the taxa
using a homemade python script, and Euclidean distance em-
ployed as the distance method. Only the best and relevant
number of hits for each query sequence was used in the count.
The abundances of DNA sequences from the water samples
were compared using the Statistical Analysis of Metagenomic
Profiles (STAMP) software version 2.1.3 (Parks et al., 2014),
to test the hypothesis that there were no significant differ-
ences between the samples and provide exploratory plots for
analyzing the profile of samples from the same researvoir.
P values were calculated using the Welch’s t-test (P < 0.05),
and correction was applied using Bonferroni (16S-23S rRNA
ITS and cpcBA-IGS) and Benjamini-Hochberg (mcyE and
sxtA) methods. Taxa with a small effect size were removed
by filtering (effect size < 1.0), and asymptotic confidence
intervals (95%) were calculated. Correlation matrix based
principal component analysis (PCA), a process that employs
orthogonal transformation, was performed on abundance
measures using the Factoextra 1.0.3 package with the R ver-
sion 3.2.2 software (R Development Core Team, 2015). With
an ellipse level of 0.95, the PCA was used to convert the set
of observations of possibly correlated variables into a set of
values of linearly uncorrelated variables. Cross-assembly of
cyanobacterial sequences was conducted using MIRA (Che-
vreux et al., 2004) by concatenating all the sequenced reads
from all the samples into one unique file. Individual sequ-
ences compared to the cross-assembly were visualized using
distance and cladogram from the crAss (Dutilh et al., 2012)
MCs and CYNs results were retrieved from Bittencourt-
Oliveira et al. (2014), and STXs results from Lorenzi et al.
(2018). These data were used for comparisons with the de-
tection of potential toxin-producing cyanobacterial geno-
types in the samples. The ELISA method is widely used for
cyanotoxin detection and quantification in field and labo-
ratory samples, and many studies have demonstrated a strong
correlation between ELISA and LC-MS results (Babica et
al., 2006; Wood et al., 2008; Bláhová et al., 2009; Ballot et al.,
2010), and HPLC-DAD (R = 0.96, P < 1 × 10−10) (Metcalf et
Heat maps generated from sequencing data on the basis of
the relative abundance of cyanobacteria are shown in Fig. 1.
The cyanobacterial profiles were variable between seasons
in the Ingazeira reservoir. The 16S-23S rRNA ITS sequence
results showed that Cylindrospermopsis raciborskii (46.8%)
was dominant in the rainy season. Furthermore, C. raci-
borskii (31.8%) was the most dominant species followed by
Sphaerospermopsis aphanizomenoides (17.26%) and Plank-
tothrix zahidii (16.4%) on the basis of cpcBA-IGS sequences.
Microcystis aeruginosa was the most abundant cyanobacte-
rial species in Ingazeira reservoir in the dry season with re-
gard to 16S-23S rRNA ITS sequences (Fig. 1), while M. pan-
niformis was dominant on the basis of cpcBA-IGS sequences.
The 16S-23S rRNA ITS sequences of uncultured cyanobac-
terium were as abundant as those of M. aeruginosa in the dry
season. Moreover, the sequences of uncultured Microcystis
sp. comprised 18.18% of total cpcBA-IGS sequences in the
Microcystis was the most abundant genus in the two con-
secutive dry seasons in Mundaú reservoir (Fig. 1). With re-
gard to the 16S-23S rRNA ITS sequences, the proportion
(57.97% for M1, and 56.10% for M27) of M. aeruginosa was
similar in both dry seasons. Despite the occurrence of M.
aeruginosa in M1, M. flos-aquae significantly contributed to
the abundance of the Microcystis genus on the basis of cpcBA-
IGS sequences. On the other hand, M. flos-aquae was domi-
nant in M27, followed by M. aeruginosa and M. panniformis
according to cpcBA-IGS sequence analysis. In addition, the
proportion of M. aeruginosa population was almost unaltered
in M1 and M27 based on cpcBA-IGS sequences. Analysis of
cpcBA-IGS sequences revealed a similar relative abundance
of Radiocystis fernandoii in M1 and M27.
Characterization of cyanotoxin genes
The detected mcyE gene belonged to Microcystis in Ingazeira
reservoir, despite the presence of other potential MC-pro-
ducing genera in samples I1 and I18 of the rainy and dry
seasons, respectively (Fig. 1). The proportion of Microcystis
sequences changed during the seasons, leading to the domin-
ance of Microcystis sp. CYN10 (accession number FJ393328)
in the dry season (Fig. 1). The detected sxtA gene in Ingazeira
reservoir was strongly associated with C. raciborskii in the
4Lorenzi et al.
Fig. 1. Relative abundance of cya-
nobacteria and cyanotoxins genes
in Ingazeira (I) and Mundaú (M)
reservoirs. Lowercase letters a, b,
and c represent triplicate samples.
White color (-) represents zero
rainy and dry seasons. The presence of potential CYN-pro-
ducing cyanobacteria was not detected by PCR amplifica-
tion of the cyrJ gene with the oligonucleotide primers cyn-
sulfF and cylnamR (Mihali et al., 2008).
In Mundaú reservoir, the detected mcyE gene belonged to
M. aeruginosa and Microcystis sp. RST 9501 (JQ771642) in
both dry seasons. However, the proportion of mcyE gene be-
longing to M. aeruginosa and Microcystis sp. RST 9501 (ac-
cession number JQ771642) varied between the two dry sea-
sons (Fig. 1). The detected sxtA gene in samples M1 and M27
belonged to C. raciborskii. The cyrJ gene was not detected in
Seasonal comparison of cyanobacterial diversity and cyano-
toxin genes in the studied reservoirs
The profiles of cyanobacterial diversity in the samples col-
lected from the same reservoir were compared with the
STAMP software, and the detected differences are shown
in Fig. 2. Pairwise comparisons of 16S-23S rRNA ITS sequ-
ences of samples I1 (rainy) and I18 (dry) from Ingazeira
reservoir indicated that C. raciborskii population was the
highest in I1 (Fig. 2A, powder gray), which was indicative
of significant positive differences. On the other hand, the
population size of M. aeruginosa was the highest in I18 (Fig.
2A, gray), corresponding to negative differences between
Next-generation DNA sequencing of cyanobacterial blooms 5
Fig. 3. Principal component analysis of the most abundant taxa (A) and cyanotoxins genes (B) in Ingazeira (I) and Mundaú (M) reservoirs. Light gray col-
ored taxa show low contribution to the analysis. ITS – 16S-23S rRNA ITS; IGS – cpcBA-IGS; I1 – rainy season; I18, M1 and M27 – dry season samples.
Lowercase letters a, b, and c represent triplicate samples.
6Lorenzi et al.
Next-generation DNA sequencing of cyanobacterial blooms 7
the populations. As visualized by STAMP, highest abun-
dance of C. raciborskii and S. aphanizomenoides popula-
tions in I1 (Fig. 2C, powder gray) corresponded to positive
differences, while that of M. panniformis in I18 (Fig. 2C, gray)
corresponded to negative differences between proportions
on the basis of cpcBA-IGS sequences. In the case of mcyE
gene analysis, the overrepresentation of M. aeruginosa po-
pulation in I1 (Fig. 2E, powder gray) showed positive differ-
ences between the analyzed samples.
For Mundaú reservoir, pairwise comparisons of cpcBA-IGS
sequences of samples M1 and M27 (dry seasons) indicated
that the population of M. flos-aquae was higher in M1 (Fig.
2D, light gray), which corresponded to positive differences.
The population of M. panniformis was overrepresented in
M27 (Fig. 2D, dark gray), indicating negative differences be-
tween proportions of this species. In addition, positive and
negative differences between mcyE proportions were shown
for M. aeruginosa population, which was highest in M1 (Fig.
2F, light gray), and Microcystis sp. RST 9501 that was over-
represented in M27 (Fig. 2F, dark gray), respectively. No sig-
nificant differences were found with regard to the relative
abundance of sxtA gene sequences in Ingazeira and Mundaú
Principal component analysis (PCA) showed a significant
contribution of the 16S-23S rRNA ITS sequences to the cor-
relation between the collected samples and reservoirs (Fig.
3A). PCA indicated a strong positive correlation between
the Mundaú samples M1 and M27 (dry seasons) and M. ae-
ruginosa based on 16S-23S rRNA ITS sequences. The rela-
tive abundance of the 16S-23S rRNA ITS sequences of Plank-
tothrix pseudagardhii and uncultured proteobacterium sig-
nificantly distinguished the Ingazeira I18 (dry season) sam-
ple from the I1 (rainy season) sample (Fig. 3A). The uncul-
tured proteobacterium sequences were generated using the
359F and 23S30R (Strunecký et al., 2011) primers, and were
also annotated with a cut-off E-value of 1e-5 and 97% mini-
mum sequence identity against the NCBI-NT database. Due
to the significant number of the hits obtained with the un-
cultured proteobacterium sequences, they were not omitted
from the results. The first 2 principal components accounted
for over 96% of the total variation (Fig. 3A). PCA results for
toxin-related genes are shown in Fig. 3B. The detected mcyE
gene was significantly associated with the different climatic
seasons and reservoirs (Fig. 3B), while the sxtA gene was sig-
nificantly correlated with C. raciborskii on the second prin-
cipal component. The cyanobacterial taxa related to mcyE se-
quences strongly separated the water samples and reservoirs
into four distinct quadrants. This ensured that even samp-
lings performed in consecutive dry seasons were not grouped
together. The first 2 principal components of the PCA were
responsible for over 74% of the total variation (Fig. 3B).
To better evaluate whether each environmental sample har-
bored a specific cyanobacterial community per climatic sea-
son, their taxonomic and cyanotoxin genes compositions
were assessed with the crAss program. The crAss cladogram
showed that cyanobacterial community profiles were clearly
separated according to the sampling times and reservoirs
(Supplementary data Fig. S2). However, an exception was one
of the replicates of M27 (M27c) that was similar to those of
M1, showing that the technical reproducibility of the PCR
reactions should be taken into consideration when DNA
sequences are analyzed.
MC and STX were detected in all the samples analyzed (Sup-
plementary data Table S1). We are aware that the ELISA
kit crossreacts with MC-LR, MC-RR, and MC-YR iso-
forms, which made it impossible to define the type of MC
detected in the different samples. The highest MC concen-
tration was recorded in Mundaú reservoir during the first
dry season, while the highest concentrations of STX were
found in Ingazeira reservoir in the rainy and dry seasons.
CYN was not detected in the samples analyzed.
Cyanobacteria are an important group because of the roles
they play in the carbon and nitrogen cycling, environmental
and human health risk, and their biotechnological and phar-
macological applications. Historically, these organisms have
been studied on the basis of microscopic inspection, cell
counts, and biochemistry.
In the present study, the molecular analysis of environmental
samples using different target sequences aided the compar-
ison of taxonomically dominant groups in the Ingazeira and
Mundaú reservoirs per climatic season. Since these reservoirs
are located in the Brazilian semiarid region that is a particular
ecosystem worldwide, the public availability of these data will
facilitate comparisons with other environments, and provide
comprehensive information on the community of cyanobac-
teria of unique ecosystems. In addition, the molecular pro-
files obtained have valuable applications for future studies
on biochip constructions for the detection of potential cya-
notoxin-producing species in specific environments.
Several studies conducted on Brazilian semiarid reservoirs
have shown that these environments harbor similar cyano-
bacterial communities (Bouvy et al., 2003; Bittencourt-Oli-
veira et al., 2011a, 2011b, 2012a, 2014; Dantas et al., 2011;
Lira et al., 2011; Moura et al., 2011). According to Bittencourt-
Oliveira et al. (2014), cyanobacterial diversity was mainly
constituted by the toxin-producing species Cylindrospermopsis
raciborskii, Planktothrix agardhii, Sphaerospermopsis apha-
nizomenoides, and Geitlerinema amphibium in 10 reservoirs
analyzed. The only exception was the Venturosa reservoir,
where the dominant species was Merismopedia tenuissima.
Using a DNA sequence based-approach, the dominant ge-
nus found in Mundaú reservoir was Microcystis (consecu-
tive dry seasons), which is in accordance with the report by
Bittencourt-Oliveira et al. (2014) that Microcystis panniformis
was the dominant species in the reservoir. M. panniformis
is a very common species in Southern America, especially in
the eastern part of Brazil (Sant’Anna et al., 2004). However,
using morphological features for taxonomic identification
of different Microcystis species is limited by the quality of the
samples and microscopic equipment, and the competence
of the taxonomist. As the current morphological classifica-
8Lorenzi et al.
tion of Microcystis species is not supported by 16S rRNA gene
phylogenetic analysis (Willame et al., 2006) and the average
sequence divergence has been generally less than 1% (Rudi
et al., 1997; Otsuka et al., 1998), we used the 16S-23S rRNA
ITS and cpcBA-IGS sequences. Studies of the ITS region of
ribosomal RNA genes (16S-23S rRNA ITS) have distinguished
closely related prokaryotic species or populations (Bolch et
al., 1996), whereas the restriction enzyme digestion of 16S-23S
rRNA ITS has been used to resolve closely related cyanobac-
terial strains (Janse et al., 2003). Despite the inclusion of the
ITS region, the molecular marker did not effectively assess
the different Microcystis species when compared to the cpcBA-
IGS results (Fig. 1). In fact, the unification of some Micro-
cystis species, including M. aeruginosa, M. ichthyoblabe, M.
novackekii, M. viridis, and M. wesenbergii, into a single M.
aeruginosa species has been proposed under the Rules of the
Bacteriological Code (Otsuka et al., 2001; Harke et al., 2016).
In the present study, the best results were obtained using the
cpcBA-IGS region as the target molecule (Fig. 1). Compared
to the 16S-23S rRNA ITS sequences, Microcystis species were
better differentiated using the IGS sequence between the β-
and α-subunit ORFs (cpcBA-IGS), which encode for β-phy-
cocyanin and α-phycocyanin, respectively (Neilan et al., 1995).
The intergenic spacer (IGS) between the two bilin subunit
genes of the phycocyanin operon (PC) has been shown to
be a variable region of DNA sequence that is useful for the
identification of cyanobacteria to the strain level (Neilan et
In Ingazeira reservoir, C. raciborskii was the dominant spe-
cies in the rainy season, while M. aeruginosa was dominant
in the dry season (Fig. 1). There were significant differences
(P < 0.05) between these taxonomic groups in the rainy and
dry seasons. The relative abundance of C. raciborskii was
higher in the rainy season, followed by a decrease in the dry
season (Fig. 2A and C). The opposite was observed for M.
aeruginosa in the same seasons. Recently, metagenomics de-
monstrated the presence of Microcystis and Cylindrosper-
mopsis in all ponds sampled in a severely drought-impacted
semiarid region of Paraíba State, Brazil (Walter et al., 2018).
Similar to the results obtained in Mundaú reservoir, the
16S-23S rRNA ITS sequence was not as discriminatory as
cpcBA-IGS in assessing the cyanobacterial diversity of Inga-
zeira reservoir (Fig. 1). Species such as P. zahidii and S. apha-
ninomenoides were only relevant in the rainy season when
the cpcBA-IGS sequence was applied as a molecular mar-
ker (Fig. 1). In the same way, Microcystis species were well
characterized by the cpcBA-IGS sequence in the dry season.
On the basis of the DNA sequencing approach, our data
showed that the composition of cyanobacterial communi-
ties changed during the rainy and dry seasons in Ingazeira
reservoir. However, using traditional microscopic techni-
ques, Bittencourt-Oliveira et al. (2014) reported multi-spe-
cies blooms of cyanobacteria, with Planktothrix agardhii as
the most abundant (69.70%) species in the rainy season in
Ingazeira reservoir. Furthermore, the authors reported that
Geitlerinema amphibium (52.76%) and Planktothrix agardhii
(32.67%) were dominant in the reservoir in the dry season.
Although DNA sequencingGtechniques can efficiently assess
previously undetectable organisms in an environment, some
bias introduced during the processing steps of the study such
as DNA extraction protocol, sequencing artifacts, DNA copy
numbers, and primer design may produce different results
when compared to traditional methods (Brooks et al., 2015).
Large discrepancies in the proportion of Gram-negative bac-
teria have already been observed using next-generation se-
quencing, microscopy, and culture-based methods (Lagier
et al., 2012). Despite performing at least 3 DNA extractions
and 3 PCR amplifications per sample, our results demon-
strated that some bias remained, which probably led to the
differences between the results reported by Bittencourt-Oli-
veira et al. (2014) and those of the present study. It is possible
that the primer sets we used are more specific to coccoid than
filamentous cyanobacteria. This may lead to a preferential
amplification of multi-template PCR and/or primer mis-
match, which probably caused the discrepancies observed.
Furthermore, the dominant genus found in Mundaú reser-
voir was Microcystis (consecutive dry seasons), which is in ac-
cordance with the report by Bittencourt-Oliveira et al. (2014).
From the results obtained with environmental samples, Sipos
et al. (2007) recommended the use of a low annealing tem-
perature to reduce preferential amplification and maintain
PCR stability. In addition, the BLAST hits obtained by sear-
ching against reference databases are generally used for taxo-
nomic identification of query sequences (Kim et al., 2013),
but are limited by the sequences available in public databases.
Sequences of P. agardhii and G. amphibium isolated from
Brazilian ecosystems are extremely scarce in public data-
bases, and may contain genetic diversities that do not permit
the annealing of currently designed primers. Furthermore,
the genetic diversity of uncultured cyanobacteria should be
taken into account when designing PCR based-methods. As
previously stated above, it is important to note that the use
of morphological features for taxonomic classification of cy-
anobacteria has limitations, which also may have been re-
sponsible for the differences recorded between the results
reported in Bittencourt-Oliveira et al. (2014) and those of
the present study.
Characterization of cyanotoxin genes
The toxicity potential of water bodies from semiarid eco-
systems was recently shown by Walter et al. (2018) using
NGS. COG functional annotation of metagenomic sequences
revealed gene sequences related to toxin production in drin-
king water from a semiarid region of Paraíba State (Brazil),
including cyanopeptolin synthetase (mcn), microcystin syn-
thetase (mcy), and non-ribosomal peptide synthase genes.
The PCR-based sequencing method employed in this study
generated sequences of mcyE and sxtA genes, which are in-
volved in MC and STX biosynthesis, respectively. The mcyE
gene codes for a partial adenylation domain and a phospho-
pantetheine-binding site, the region that activates glutamic
acid in the MC biosynthesis pathway (Rantala et al., 2004).
The sxtA gene encodes a polyketide synthase (PKS)-like struc-
ture that is uncommon to STX biosynthesis in cyanobacteria
(Kellmann et al., 2008). Despite the presence of other poten-
tial toxin-producing genera, the detection of the sxtA gene
was mainly associated with C. raciborskii, and the presence
of the mcyE gene was related to Microcystis in both reser-
voirs. As previously reported, the effects of bias during sam-
ple processing, sequencing, and sequence annotation may
Next-generation DNA sequencing of cyanobacterial blooms 9
have interfered with the affiliation of the sequences. Never-
theless, by integrating molecular taxonomy and functional
features, it was possible to separate different samples from
the reservoirs analyzed (Supplementary data Fig. S2), which
demonstrated the sensitivity of DNA-based methods.
Cyanotoxins and related genes
The detected mcyE gene had good correlation with the pre-
sence of MC, while the amplification of the sxtA gene was
related to STX content (Supplementary data Table S1). The
C. raciborskii populations of Brazilian aquatic systems are
potential STX producers (Lagos et al., 1999), and the pre-
sence of the sxt biosynthetic gene cluster has been confir-
med in the T3 strain of the cyanobacterium (Kellmann et al.,
2008). Unlike the Southern American isolates, the C. raci-
borskii strains isolated from Australian and Asian aquatic
ecosystems synthesize CYN. This provided a basis for the
characterization of the cyr gene cluster in C. raciborskii AWT-
205 (Mihali et al., 2008). Although the cyrJ gene is a suitable
probe for the detection of potential CYN producers (Mihali
et al., 2008), its absence in samples collected from Ingazeira
and Mundaú reservoirs correlates with the absence of the
cyanotoxin in both reservoirs. It is important to note that CYN
was recently reported in some Brazilian semiarid water sup-
ply reservoirs containing C. raciborskii (Bittencourt-Oliveira
et al., 2011a). Despite the lack of PCR amplification of the cyrJ
gene and the absence of CYN in the present study, careful
consideration should be given to non-annealing of the oli-
gonucleotide primers and non-expression of the cyr gene
The PCR-based methodology applied in this study gives an
overview of the genetic basis of the cyanobacterial profiles
present in two Brazilian semiarid public water supply reser-
voirs. Taxonomic analysis of high-throughput DNA sequen-
cing data showed that the cyanobacterial diversity did not
significantly change in Mundaú reservoir during the two con-
secutive dry seasons, and it varied over the seasons in Inga-
zeira reservoir. The detected mcyE and sxtA genes were as-
sociated with the presence of the cyanotoxins MC and STX
in both reservoirs, respectively. Using the molecular profiles
generated, biochip constructions for the detection of poten-
tial cyanotoxin-producing species in specific environments
may be obtained in future works. However, our results sug-
gested some biases of analysis on the basis of DNA sequ-
encing such as preferential amplification of multi-template
PCR, primer mismatch, and/or sequence availability in pub-
lic databases, which probably influenced the cyanobacterial
communities seen in NGS output. Based on the findings of
the present study, we recommend the integration of tradi-
tional approaches with NGS methods for the comprehensive
understanding of the community structure and function of
cyanobacterial populations in natural aquatic ecosystems.
This study was sponsored by grants from the São Paulo
Research Foundation (FAPESP – 2013/15296-2) and the Bra-
zilian National Research Council (CNPq – 442083/2014-9).
A.S.L. and M.A.C were supported by FAPESP post-doctoral
fellowships (Grant 2014/01913-2 and 2013/11306-3, respec-
tively). G.G.Z.S. was supported by NSF Grants (CNS-1305112,
MCB-1330800, and DUE-132809 to R.A.E). F.A.C.L. was
supported by CAPES graduate scholarship.
Conflict of Interest
The authors declare that there are no conflicts of interest.
Almeida, V.L.S., Dantas, E.W., Melo-Júnior, M., Bittencourt-Oliveira,
M.C., and Moura, A.N. 2009. Zooplanktonic community of six
reservoirs in northeast Brazil. Braz. J. Biol. 69, 57–65.
Altschul, S.F., Madden, T.L., Schäffer, A.A., Zhang, J., Zhang, Z.,
Miller, W., and Lipman, D.J. 1997. Gapped BLAST and PSI-
BLAST: a new generation of protein database search programs.
Nucleic Acids Res. 25, 3389–3402.
Azevedo, S.M.F.O., Carmichael, W.W., Jochimsen, E.M., Rinehart,
K.L., Lau, S., Shaw, G.R., and Eaglesham, G.K. 2002. Human in-
toxication by microcystin during renal dialysis treatment in
Caruaru. Toxicology 181-182, 441–446.
Babica, P., Kohoutek, J., Bláha, L., Adamovsky, O., and Marsálek,
B. 2006. Evaluation of extraction approaches linked to ELISA and
HPLC for analyses of microcystin-LR, -RR and -YR in fresh-
water sediments with different organic material contents. Anal.
Bioanal. Chem. 85, 1545–1551.
Ballot, A., Fastner, J., and Wiedner, C. 2010. Paralytic shellfish poi-
soning toxin-producing cyanobacterium Aphanizomenon gracile
in northeast Germany. Appl. Environ. Microbiol. 76, 1173–1180.
Bittencourt-Oliveira, M.C. 2003. Detection of potentially micro-
cystin producing cyanobacteria in Brazilian reservoirs with a
mcyB molecular marker. Harmful Algae 2, 51–60.
Bittencourt-Oliveira, M.C., Piccin-Santos, V., Kujbida, P., and Moura,
A.N. 2011a. Cylindrospermopsin in water supply reservoirs in
Brazil determined by immunochemical and molecular methods.
J. Water Resource Prot. 3, 349–355.
Bittencourt-Oliveira, M.C., Moura, A.N., Hereman, T.C., and Dantas,
E.W. 2011b. Increase in straight and coiled Cylindrospermopsis
raciborskii (cyanobacteria) populations under conditions of ther-
mal de-stratification in a shallow tropical reservoir. J. Water Re-
source Prot. 3, 245–252.
Bittencourt-Oliveira, M.C., Dias, S.N., Moura, A.N., Cordeiro-Araújo,
M.K., and Dantas, E.W. 2012a. Seasonal dynamics of cyanobac-
teria in a eutrophic reservoir (Arcoverde) in the semi-arid region
of Brazil. Braz. J. Biol. 72, 533–544.
Bittencourt-Oliveira, M.C., Piccin-Santos, V., and Barros, S.G. 2012b.
Microcystin-producing genotypes from cyanobacteria in Brazilian
reservoirs. Environ. Toxicol. 27, 461–471.
Bittencourt-Oliveira, M.C., Piccin-Santos, V., Moura, A.N., Aragão-
Tavares, N.K.C., and Cordeiro-Araújo, M.K. 2014. Cyanobacteria,
microcystins and cylindrospermopsin in public drinking supply
reservoirs of Brazil. An. Acad. Bras. Ciênc. 86, 297–310.
Bittencourt-Oliveira, M.C., Santos, D.M.S., and Moura, A.N. 2010.
Toxic cyanobacteria in reservoirs in northeastern Brazil: detec-
tion using molecular method. Braz. J. Biol. 70, 1005–1010.
Bláhová, L., Oravec, M., Marsálek, B., Sejnohová, L., Simek, Z., and
10 Lorenzi et al.
Bláha, L. 2009. The first occurrence of the cyanobacterial alkaloi
toxin cylindrospermopsin in the Czech Republic as determined
by immunochemical and LC/MS methods. Toxicon 53, 519–524.
Bolch, C.J.S., Blaekburn, S.I., Neilan, B.A., and Grewe, P.M. 1996.
Genetic characterization of strains of cyanobacteria using PCR-
RFLP of the cpcBA intergenic spacer and flanking regions. J.
Phycol. 32, 445–451.
Borges, H.L.F., Branco, L.H.Z., Martins, M.D., Lima, C.S., Barbosa,
P.T., Lira, G.A.S.T., Bittencourt-Oliveira, M.C., and Molica, R.J.R.
2015. Cyanotoxin production and phylogeny of benthic cyano-
bacterial strains isolated from the northeast of Brazil. Harmful
Algae 43, 46–57.
Bouvy, M., Falcão, D., Marinho, M., Pagano, M., and Moura, A. 2000.
Occurrence of Cylindrospermopsis (cyanobacteria) in 39 Brazilian
tropical reservoirs during the 1998 drought. Aquat. Microb. Ecol.
Bouvy, M., Molica, R., Oliveira, S., Marinho, M., and Beker, B. 2001.
Effects of a cyanobacterial bloom (Cylindrospermopsis raciborskii)
on bacterial and zooplankton communities in Ingazeira reser-
voir (Northeast, Brazil). Aquat. Microb. Ecol. 25, 215–227.
Bouvy, M., Nascimento, S.M., Molica, R.J.R., Ferreira, A., Huszar,
V., and Azevedo, S.M.F.O. 2003. Limnological features in Tapa-
curá reservoir (northeast Brazil) during a severe drought. Hydro-
biologia 493, 115–130.
Brazil. 2011. Ministry of Health. Regulation No. 2914, pp. 39–46.
Brazilian official gazette, Executive Branch, Brasília, DF, Dec.
12, 2011. Section 1.
Brooks, J.P., Edwards, D.J., Harwich, M.D.Jr., Rivera, M.C., Fettweis,
J.M., Serrano, M.G., Reris, R.A., Sheth, N.U., Huang, B., Girerd,
P., et al. 2015. The truth about metagenomics: quantifying and
counteracting bias in 16S rRNA studies. BMC Microbiol. 15, 66.
Chellappa, N.T. and Costa, M.A.M. 2003. Dominant and co-exist-
ing species of cyanobacteria from a eutrophicated reservoir of
Rio Grande do Norte State, Brazil. Acta Oecol. 24, S3–S10.
Chevreux, B., Pfisterer, T., Drescher, B., Driesel, A.J., Müller, W.E.,
Wetter, T., and Suhai, S. 2004. Using the miraEST assembler
for reliable and automated mRNA transcript assembly and SNP
detection in sequenced ESTs. Genome Res. 14, 1147–1159.
Christiansen, G., Fastner, J., Erhard, M., Börner, T., and Dittmann,
E. 2003. Microcystin biosynthesis in Planktothrix: genes, evolu-
tion, and manipulation. J. Bacteriol. 185, 564–572.
Costa, I.A.S., Azevedo, S.M.F.O., Senna, P.A.C., Bernardo, R.R.,
Costa, S.M., and Chellappa, N.T. 2006. Occurrence of toxin-
producing cyanobacteria blooms in a Brazilian semiarid reser-
voir. Braz. J. Biol. 66, 211–219.
D’Agostino, P.M., Woodhouse, J.N., Katharina Makower, A.,
Yeung, A.C.Y., Ongley, S.E., Micallef, M.L., Moffitt, M.C., and
Neilan, B.A. 2016. Advances in genomics, transcriptomics and
proteomics of toxin-producing cyanobacteria. Environ. Micro-
biol. Rep. 8, 3–13.
Dantas, E.W., Bittencourt-Oliveira, M.C., and Moura, A.N. 2012.
Dynamics of phytoplankton associations in three reservoirs in
northeastern Brazil assessed using Reynolds’ theory. Limnologica
Dantas, E.W., Moura, A.N., and Bittencourt-Oliveira, M.C. 2011.
Cyanobacterial blooms in stratified and destratified eutrophic
reservoirs in semi-arid region of Brazil. An. Acad. Bras. Cienc.
Dittmann, E., Fewer, D.P., and Neilan, B.A. 2013. Cyanobacterial
toxins: biosynthetic routes and evolutionary roots. FEMS Mic-
robiol. Rev. 37, 23–43.
Dutilh, B.E., Schmieder, R., Nulton, J., Felts, B., Salamon, P., Ed-
wards, R.A., and Mokili, J.L. 2012. Reference-independent com-
parative metagenomics using cross-assembly: crAss. Bioinfor-
matics 28, 3225–3231.
Fonseca, J.R., Vieira, P.C.S., Kujbida, P., and da Costa, I.A.S. 2015.
Cyanobacterial occurrence and detection of microcystins and
saxitoxins in reservoirs of the Brazilian semi-arid. Acta Limnol.
Bras. 27, 78–92.
Harke, M.J., Steffen, M.M., Gobler, C.J., Otten, T.G., Wilhelm, S.W.,
Wood, S.A., and Paerl, H.W. 2016. A review of the global eco-
logy, genomics, and biogeography of the toxic cyanobacterium,
Microcystis spp. Harmful Algae 54, 4–20.
Janse, I., Meima, M., Edwin, W., Kardinaal, A., and Zwart, G. 2003.
High-resolution differentiation of cyanobacteria by using rRNA-
internal transcribed spacer denaturing gradient gel electropho-
resis. Appl. Environ. Microbiol. 69, 6634–6643.
Jochimsen, E.M., Carmichael, W.W., An, J., Cardo, D.M., Cookson,
S.T., Holmes, C.E., Antunes, B.C., Melo Filho, D.A., Lyra, T.M.,
Barreto, V.S.T., et al. 1998. Liver failure and death after exposure
to microcystins at a hemodialysis center in Brazil. N. Engl. J.
Med. 338, 873–878.
Kellmann, R., Mihali, T.K., Jeon, Y.J., Pickford, R., Pomati, F., and
Neilan, B.A. 2008. Biosynthetic intermediate analysis and func-
tional homology reveal a saxitoxin gene cluster in cyanobacteria.
Appl. Environ. Microbiol. 74, 4044–4053.
Kim, M., Lee, K.H., Yoon, S.W., Kim, B.S., Chun, J., and Yi, H. 2013.
Analytical tools and databases for metagenomics in the next-
generation sequencing era. Genomics Inform. 11, 102–113.
Lagier, J.C., Million, M., Hugon, P., Armougom, F., and Raoult, D.
2012. Human gut microbiota: Repertoire and variations. Front.
Cell. Infect. Microbiol. 2, 136.
Lagos, N., Onodera, H., Zagatto, P.A., Andrinolo, D., Azevedo, S.M.,
and Oshima, Y. 1999. The first evidence of paralytic shellfish
toxins in the water cyanobacterium Cylindrospermopsis racibor-
skii, isolated from Brazil. Toxicon 37, 1359–1373.
Li, N., Zhang, L., Li, F., Wang, Y., Zhu, Y., Kang, H., Wang, S., and
Qin, S. 2011. Metagenome of microorganisms associated with the
toxic cyanobacteria Microcystis aeruginosa analyzed using the 454
sequencing platform. Chinese J. Oceanol. Limnol. 29, 505–513.
Lira, G.A.S.T., Araújo, E.L., Bittencourt-Oliveira, M.C., and Moura,
A.N. 2011. Phytoplankton abundance, dominance and coexis-
tence in an eutrophic reservoir in the state of Pernambuco, Nor-
theast Brazil. An. Acad. Bras. Cienc. 83, 1313–1326.
Lorenzi, A.S., Chia, M.A., Piccin-Santos, V., and Bittencourt-Oli-
veira, M.C. 2015. Microcystins and cylindrospermopsins mole-
cular markers for the detection of toxic cyanobacteria: a case
study of northeastern Brazilian reservoirs. Limnetica 34, 269–282.
Lorenzi, A.S., Cordeiro
Araujo, M.K., Chia, M.A., and Bittencourt-
Oliveira, M.C. 2018. Cyanotoxin contamination of semiarid drin-
king water supply reservoirs. Environ. Earth Sci. 77, 595.
Metcalf, J.S., Bell, S.G., and Codd, G.A. 2000. Production of novel
polyclonal antibodies against the cyanobacterial toxin micro-
cystin-LR and their application for the detection and quantifi-
cation of microcystins and nodularin. Water Res. 34, 2761–2769.
Meyer, F., Paarmann, D., D’Souza, M., Olson, R., Glass, E.M., Kubal,
M., Paczian, T., Rodriguez, A., Stevens, R., Wilke, A., et al. 2008.
The metagenomics RAST server – a public resource for the auto-
matic phylogenetic and functional analysis of metagenomes. BMC
Bioinformatics 9, 386.
Mihali, T.K., Kellmann, R., Muenchhoff, J., Barrow, K.D., and Neilan,
B.A. 2008. Characterization of the gene cluster responsible for
cylindrospermopsin biosynthesis. Appl. Environ. Microbiol. 74,
Molica, R.J.R., Oliveira, E.J.A., Carvalho, P.V.V.C., Costa, A.N.S.F.,
Cunha, M.C.C., Melo, G.L., and Azevedo, S.M.F.O. 2005. Occur-
rence of saxitoxins and an anatoxin-a(s)-like anticholinesterase
in a Brazilian drinking water supply. Harmful Algae 4, 743–753.
Moura, A.N., Dantas, E.W., Oliveira, H.S.B., and Bittencourt-Oliveira,
M.C. 2011.Vertical and temporal dynamics of cyanobacteria in
the Carpina potable water reservoir in Northeastern Brazilian.
J. Biol. 71, 451–459.
Moura, A.N., Nascimento, E.C., and Dantas, Ê.W. 2012. Temporal
and spatial dynamics of phytoplankton near farm fish in eutrophic
Next-generation DNA sequencing of cyanobacterial blooms 11
reservoir in Pernambuco, Brazil. Rev. Biol. Trop. 60, 581–597.
Neilan, B.A., Jacobs, D., and Goodman, A.E. 1995. Genetic diver-
sity and phylogeny of toxic cyanobacteria determined by DNA
polymorphisms within the phycocyanin locus. Appl. Environ.
Microbiol. 61, 3875–3883.
Neilan, B.A., Pearson, L.A., Muenchhoff, J., Moffitt, M.C., and Ditt-
mann, E. 2013. Environmental conditions that influence toxin
biosynthesis in cyanobacteria. Environ. Microbiol. 15, 1239–1253.
Otsuka, S., Suda, S., Li, R., Watanabe, M., Oyaizu, H., Matsumoto,
S., and Watanabe, M.M. 1998. 16S rDNA sequences and phylo-
genetic analysis of Microcystis strains with and without phyco-
erythrin. FEMS Microbiol. Lett. 164, 119–124.
Otsuka, S., Suda, S., Shibata, S., Oyaizu, H., Matsumoto, S., and
Watanabe, M.M. 2001. A proposal for the unification of five spe-
cies of the cyanobacterial genus Microcystis Kützing ex Lem-
mermann 1907 under the rules of the bacteriological code. Int.
J. Syst. Evol. Microbiol. 51, 873–879.
Parks, D.H., Tyson, G.W., Hugenholtz, P., and Beiko, R.G. 2014.
STAMP: Statistical analysis of taxonomic and functional profiles.
Bioinformatics 30, 3123–3124.
Penn, K., Wang, J., Fernando, S.C., and Thompson, J.R. 2014. Se-
condary metabolite gene expression and interplay of bacterial
functions in a tropical freshwater cyanobacterial bloom. ISME
J. 8, 1866–1878.
R Development Core Team. 2015. R: A language and environment
for statistical computing. R foundation for statistical comput-
ing, Vienna, Austria. URL https://www.R-project.org/.
Rantala, A., Fewer, D.P., Hisbergues, M., Rouhiainen, L., Vaitomaa,
J., Börner, T., and Sivonen, K. 2004. Phylogenetic evidence for
the early evolution of microcystin synthesis. Proc. Natl. Acad. Sci.
USA 101, 568–573.
Rastogi, R.P., Madamwar, D., and Incharoensakdi, A. 2015. Bloom
dynamics of cyanobacteria and their toxins: environmental health
impacts and mitigation strategies. Front. Microbiol. 6, 1254.
Rogers, S.O. and Bendich, A.J. 1985. Extraction of DNA from milli-
gram amounts of fresh, herbarium, and mummified plant tissues.
Plant Mol. Biol. 5, 69–76.
Rouhiainen, L., Vakkilainen, T., Siemer, B.L., Buikema, W., Hasel-
korn, R., and Sivonen, K. 2004. Genes coding for hepatotoxic
heptapeptides (Microcystins) in the cyanobacterium Anabaena
strain 90. Appl. Environ. Microbiol. 70, 686–692.
Rudi, K., Skulberg, O.M., Larsen, F., and Jakobsen, K.S. 1997. Strain
characterisation and classification of oxyphotobacteria in clone
cultures on the basis of 16S rRNA sequences from variable re-
gions V6, V7 and V8. Appl. Environ. Microbiol. 63, 2593–2599.
Sant’Anna, C.L., Azevedo, M.T.P., Senna, P.A.C., Komárek, J., and
Komárková, J. 2004. Planktic cyanobacteria from São Paulo State,
Brazil: Chroococcales. Rev. Bras. Bot. 27, 213–227.
Sipos, R., Székely, A.J., Palatinszky, M., Révész, S., Márialigeti, K.,
and Nikolausz, M. 2007. Effect of primer mismatch, annealing
temperature and PCR cycle number on 16S rRNA gene-targeting
bacterial community analysis. FEMS Microbiol. Ecol. 60, 341–350.
Smith, F.M.J., Wood, S.A., van Ginkel, R., Broady, P.A., and Gaw,
S. 2011. First report of saxitoxin production by a species of the
freshwater benthic cyanobacterium, Scytonema Agardh. Toxicon
SRH (Secretaria de Recursos Hídricos de Pernambuco). 2000. Plano
Estadual de Recursos Hídricos do Estado de Pernambuco –
Documento Síntese, Recife.
Steffen, M.M., Belisle, B.S., Watson, S.B., Boyer, G.L., Bourbonniere,
R.A., and Wilhelm, S.W. 2015. Metatranscriptomic evidence for
co-occurring top-down and bottom-up controls on toxic cyano-
bacterial communities. Appl. Environ. Microbiol. 81, 3268–3276.
Steffen, M.M., Li, Z., Effler, T.C., Hauser, L.J., Boyer, G.L., and Wil-
helm, S.W. 2012. Comparative metagenomics of toxic freshwater
cyanobacteria bloom communities on two continents. PLoS One
Strunecký, O., Elster, J., and Komárek, J. 2011. Taxonomic revision
of the freshwater cyanobacterium Phormidium murrayi = Wil-
mottia murrayi. Fottea 11, 57–71.
Tillett, D., Dittmann, E., Erhard, M., von Döhren, H., Börner, T.,
and Neilan, B.A. 2000. Structural organization of microcystin
biosynthesis in Microcystis aeruginosa PCC7806: an integrated
peptide-polyketide synthetase system. Chem. Biol. 7, 753–764.
Voorhies, A.A., Eisenlord, S.D., Marcus, D.N., Duhaime, M.B.,
Biddanda, B.A., Cavalcoli, J.D., and Dick, G.J. 2016. Ecological
and genetic interactions between cyanobacteria and viruses in a
low-oxygen mat community inferred through metagenomics and
metatranscriptomics. Environ. Microbiol. 18, 358–371.
Walter, J.M., Lopes, F.A.C., Lopes-Ferreira, M., Vidal, L.M., Leomil,
L., Melo, F., de Azevedo, G.S., Oliveira, R.M.S., Medeiros, A.J.,
Melo, A.S.O., et al. 2018. Occurrence of harmful cyanobacteria in
drinking water from a severely drought-impacted semi-arid re-
gion. Front. Microbiol. 9, 176.
Willame, R., Boutte, C., Grubisic, S., Willmotte, A., Komárek, J., and
Hoffmann, L. 2006. Morphological and molecular characteriza-
tion of planktonic cyanobacteria from Belgium and Luxemburg.
J. Phycol. 42, 1312–1332.
Wood, S.A., Mountfort, D., Selwood, A.I., Holland, P.T., Puddick,
J., and Cary, S.C. 2008. Widespread distribution and identifica-
tion of eight novel microcystins in antarctic cyanobacterial mats.
Appl. Environ. Microbiol. 74, 7243–7251.
Woodhouse, J.N., Kinsela, A.S., Collins, R.N., Bowling, L.C., Honey-
man, G.L., Holliday, J.K., and Neilan, B.A. 2016. Microbial com-
munities reflect temporal changes in cyanobacterial composition
in a shallow ephemeral freshwater lake. ISME J. 10, 1337–1351.
World Bank. 2013. Water resources planning and adaptation to cli-
mate variability and climate change in selected river Basins in
Northeast Brazil. Final report on a non lending technical assis-
tance program (P123869). World Bank, Washington DC, USA.
Zhang, J., Kobert, K., Flouri, T., and Stamatakis, A. 2013. PEAR: a
fast and accurate Illumina Paired-End reAd mergeR. Bioinfor-
matics 30, 614–620.