The mosaic structure of the mcyABC operon in
Ave Tooming-Klunderud,1,2Bjørg Mikalsen,23 Tom Kristensen1,3
and Kjetill S. Jakobsen2
Kjetill S. Jakobsen
1University of Oslo, Department of Molecular Biosciences, 0316 Oslo, Norway
2University of Oslo, Department of Biology, Centre for Ecological and Evolutionary Synthesis
(CEES), 0316 Oslo, Norway
3University of Oslo, Microbial Evolution Research Group (MERG), 0316 Oslo, Norway
Received 14 December 2007
Revised16 April 2008
Accepted 17 April 2008
An extensive study of the mcyABC genes and regions flanking the mcy gene cluster was
performed in naturally occurring Microcystis strains. Lack of methylation in strains producing only
desmethyl7-microcystin was found to be associated with point mutations in substrate-binding
sequence motifs of the N-methyltransferase (NMT) domain in McyA. Multiple recombination
events giving rise to ‘phylogenetic mosaics’ were detected within the NMT-domain-encoding
mcyA sequences and the adenylation (A) domain sequences of mcyB and mcyC. Recombination
leading to exchanges between the mcyB and mcyC regions encoding A domains in modules
McyB1 and McyC was also detected. A previously reported replacement of the A domain in
McyB1 was found to involve the region between the conserved motifs A3 and A8/A9. In all
microcystin-producing strains the mcy gene cluster was flanked by the genes uma1 and dnaN.
Clear indications of recombination, an insertion element and footprints of IS elements were found
in the dnaN–mcyJ intergenic region. Among the non-microcystin producers, uma1 and dnaN
were linked in some, but not all strains. Most non-producing strains lacked all mcy genes, while
one strain possessed a partially deleted mcy operon. Our results show that frequent horizontal
gene transfer events in addition to point mutations and insertions/deletions contribute to variation
in the mcy gene cluster.
Blooms of cyanobacteria are common in freshwater lakes,
ponds and water reservoirs worldwide. Cyanobacteria of
the genus Microcystis are widespread in such blooms and
constitute a health risk for animals and humans through
the production of hepatotoxic microcystins – a family of
cyclic heptapeptides. Microcystins share a common
structure of cyclo (D-Ala1–L-X2–D-MeAsp3–L-Z4–Adda5–
D-Glu6–Mdha7), where Mdha is N-methyldehydroalanine,
D-MeAsp is 3-methylaspartic acid and Adda is 3-amino-9-
X and Z indicate variable L-amino acids. At least 89 different
microcystin isoforms, differing in modifications of the
peptide backbone or the type of amino acids incorporated,
have been identified (Welker & von Do ¨hren, 2006). Many
strains produce several microcystin isoforms, but only one
or two isoforms dominate in any single strain.
Microcystins are synthesized nonribosomally by a thio-
template mechanism catalysed by microcystin synthetase
(Marahiel et al., 1997). Microcystin synthetase gene clusters
(mcy) include genes for peptide synthetases, polyketide
synthases, mixed peptide/polyketide synthases and tailor-
ing enzymes, and have been characterized in detail in
Microcystis (Nishizawa et al., 1999, 2000; Tillett et al.,
2000), Planktothrix (Christiansen et al., 2003) and
Anabaena (Rouhiainen et al., 2004). Insertional gene
knockout experiments have demonstrated that all micro-
cystin variants produced by a strain are synthesized by a
single enzyme complex encoded by a 55 kb gene cluster
(Christiansen et al., 2003; Dittmann et al., 1997; Nishizawa
et al., 1999).
The phylogenetic concurrence between a few housekeeping
genes and microcystin synthetase genes from the same
strains representing a wide selection of microcystin-
producing genera suggests that mcy genes represent an
3Present address: Division of Pathology, Rikshospitalet University
Hospital, 0027 Oslo, Norway.
Abbreviations: A, adenylation (domain); C, condensation (domain), ML,
maximum likelihood; NJ, neighbour joining; NMT, N-methyltransferase
(domain); SAM, S-adenosylmethionine.
The GenBank/EMBL/DDBJ accession numbers for the sequences
determined in this work are shown in Table 1.
A supplementary table of primers and three supplementary figures are
available with the online version of this paper.
Microbiology (2008), 154, 1886–1899
18862007/015875G2008 SGMPrinted in Great Britain
ancestral state of many lineages (Rantala et al., 2004). Thus,
within lineages producing microcystin, non-producing
strains would be the result of inactivation either by point
mutation, or by partial or complete deletion of mcy genes
(Christiansen et al., 2006; Dittmann et al., 1997; Kurmayer
et al., 2004; Mikalsen et al., 2003).
One of the most intriguing features of microcystin is the
high number of isoforms, often associated with changes in
the synthetases as a result of mutations in the gene cluster
(Kurmayer et al., 2005; Mikalsen et al., 2003). Several
independent investigations on Microcystis and Planktothrix
strains have concluded that the natural variation in the gene
cluster within each strain is often caused by recombination.
The regions with highest site variation seem to be mcyA
(Kurmayer et al., 2005; Tanabe et al., 2004) and mcyB
(Kurmayer & Gumpenberger, 2006; Mikalsen et al., 2003).
For example within Planktothrix spp., the typical N-
domain of the first module of McyA has been replaced by
an A domain without NMT, leading to production of
desmethyl7-microcystin isoforms (Kurmayer et al., 2005).
For Microcystis Mikalsen et al. (2003) showed that a
recombination event between segments encoding two
different adenylation domains has led to the presence of a
second type of A domain in McyB1. Strains containing the
Arg-activating A domain in McyB1 (hereafter denoted as C-
like due to properties similar to the A domain of McyC)
produce mainly microcystin-RR (indicating a microcystin
with Arg in both the X and Z position), while strains with
theLeu-activating A domain in McyB1 (hereafter denoted as
B-like) produce mainly microcystin-LR.
In the previous study of Mikalsen et al. (2003) only a part
of mcyB (A1) was analysed in detail. We therefore
expanded the investigation using the same 17 field-
collected strains to include the NMT domain of McyA
and the A domain of McyC, as well as the flanking regions
on both sides of the gene cluster. We employed a
combination of alignments, a phylogenetic approach using
split decomposition analysis and various statistical software
(GENECONV, RPD and MaxChi) to address the genomic
processes leading to new microcystin gene cluster variants
(and thus to new Microcystis chemotypes, or inactivation of
the cluster), the genetic basis for desmethyl7-microcystin
variants and the recombination events involving mcyB1
Bacterial strains and DNA isolation. The Microcystis strains used
in this study (Table 1) have previously been genetically typed using
RAPD and REP fingerprinting (Mikalsen et al. 2003). Unialgal
cultures were grown at the Norwegian Institute of Water Research
(NIVA) as previously described (Skulberg & Skulberg, 1990) except
for strain PCC 7806, which was kindly provided by H. Utkilen
(National Institute of Public Health, Norway). Genomic DNA for
Southern blotting and PCR was isolated as described earlier (Mikalsen
et al., 2003).
Southern hybridization. Southern hybridization analyses were
performed using probes derived from dnaN, uma1 and mcyA
sequences in Microcystis aeruginosa PCC 7806 (AF183408). Probe
locations are shown in Fig. 1. Approximately 1 mg genomic DNA was
digested overnight with 15 U HindIII (37 uC). Fragments were
separated on a 1% agarose gel and transferred to an Amersham
Hybond-N membrane bySouthern
Hybridization was performed at 68 uC by standard procedures
(Galau et al., 1986). Radiolabelled probes were generated from
NMT-F/NMT-R, dnaN-F/dnaN-R and uma1 F/uma1-R PCR pro-
ducts from strain PCC 7806 by standard MSPL (magnetic solid-phase
labelling) random-primed synthesis (Espelund et al., 1990) using the
Random Primed Labelling kit (Roche Diagnostics).
PCR amplification and sequencing. Primers used for amplification
of different regions of the mcy gene cluster are listed in
Supplementary Table S1 (available with the online version of this
paper) and their relative positions in the gene cluster are shown in
Fig. 1. Primers were designed based on the publicly available mcy gene
sequence of M. aeruginosa PCC 7806, accession no. AF183408 (Tillett
et al., 2000). The primer hypX (Nishizawa et al., 2007) was used to
investigate the genetic organization downstream of uma1 in two non-
toxic strains. BD Advantage 2 polymerase (BD Biosciences) was used
in PCRs. Amplicons were purified using the E.Z.N.A Gel Extraction
kit (Omega Biotek) and sequenced directly. DNA sequencing was
performed on both strands with the BigDye Terminator v3.1 Cycle
Sequencing kit (Applied Biosystems) with a capillary electrophoresis
sequencer (ABI 3730). Sequences have been submitted to GenBank;
accession numbers are shown in Table 1.
Phylogenetic and recombination analyses. Sequences were
aligned with the publicly available mcy sequence of M. aeruginosa
PCC7806 (accessionno.AF183408)using CLUSTAL W (Thompsonetal.,
1997) and manual editing. The sequences were used to infer the
phylogeny in a Bayesian framework applying the program MrBayes
v3.1 (Ronquist & Huelsenbeck, 2003). Analysis was performed with the
following parameters: GTR model, gamma distribution, running 2
million generations and sampling trees every 100 generation, burn-in
3000 trees. The maximum-likelihood (ML) tree was estimated for
mcyB1 and mcyC A domain sequences using PhyML (Guindon &
Gascuel, 2003) under the GTR model, gamma distribution and with
parameter values indicated by MrmodelTest (Nylander, 2004). The
neighbour joining (NJ) tree was obtained under the default nucleotide
substitution model using MEGA version 3.1 (Kumar et al., 2004).
Mrmodelltest, PhyML and MrBayes analyses were performed on the
Bioportal computer platform resources (http://www.bioportal.uio.no).
Recombination was detected by split decomposition analysis using
SplitsTree version 4.8 (Huson & Bryant, 2006) with default settings
(uncorrectedP method) and 1000 bootstrap replicas, and Phi test for
recombination (implemented to test split decomposition analysis
reliability)(Bruen et al., 2006). In addition, statistical tests for detecting
recombination were used: GENECONV (Padidam et al., 1999), RDP and
MaxChi (Martin et al., 2005) analyses in the RDP version 2 b08
program package. For GENECONV, G-scale values 0 and 1 were used, for
detecting recent and older recombination events, respectively.
Recombination was also investigated by visual analysis of informative
sites (variable sites where each variant occurs in at least two strains).
Sequence variation in the NMT domain encoded
Southern hybridization analysis with a probe derived from
the NMT region of mcyA (Fig. 1) was used to investigate
Genetic mosaics in the microcystin cluster
Table 1. Microcystis strains investigated
Microcystis strain*Geographical origin Accession no. (PCR product, bp)
N-C 31 M. aeruginosaLittle Rideau Lake,
Wisconsin, USA (ATCC
L. Gjersjøen, Norway
L. Finjasjo ¨n, Sweden
L. Ma ¨laren, Sweden
L. Akersvatnet, Norway
L. Borrevatnet, Norway
L. Mosvatnet, Norway
L. Arresø, Denmark
L. Arresø, Denmark
EF115379 (1009)EF115388 (1491)EU009866 (1406)EF115396 (1063) EF115405 (515)–
N-C 43 M. aeruginosa–D–D–D–D–D EU475874d (668)
N-C 57 M. aeruginosa EF115380 (1011)EF115389 (1491) EU009867 (1406) EF115397 (1063)EF115406 (515)–
N-C 118/2 M. sp.
N-C 122/2 M. viridis
N-C 123/1 M. aeruginosa
N-C 143 M. aeruginosa
N-C 144 M. cf. flos-aquae
N-C 161/1 M. botrys
N-C 166 M. aeruginosa
N-C 169/7 M. viridis
N-C 172/5 M. cf.
N-C 228/1 M. aeruginosa
N-C 264 M. botrys
L. Akersvatnet, Norway
L. Tøra ˚ssjøen, Norway
N-C 279 M. cf.
N-C 324/1 M. sp.
PCC 7806 M. aeruginosa
–D–D–D–D–D EF115426§ (756)
*N-C, NIVA-CYA, Norwegian Institute for Water Research Cyanobacterial Culture Collection; PCC, Pasteur Culture Collection. The NIVA-CYA strains have been genetically typed using RAPD
and REP fingerprinting by Mikalsen et al. (2003). There is no evidence for clonality among the strains.
DThe strain does not possess mcy genes.
dPCR product contains hypX and uma1 segments.
§PCR product contains dnaN and uma1 segments.
||PCR product dnaN–mcyA due to partial loss of mcy operon in this strain.
The strain does not possess the NMT-domain-encoding part of mcyA.
#PCR product contains ftsH and uma1 segments.
A. Tooming-Klunderud and others1888
the presence of this domain in 17 Microcystis strains (see
Table 1 for description of the strains). Positive hybridiza-
tion signals were obtained for all nine toxic strains
(Fig. 2A), including the strains producing only des-
methyl7-microcystin isoforms (N-C 57, N-C 228/1 and
N-C 264) (Mikalsen et al., 2003). The same HindIII
restriction pattern was obtained for all strains except N-C
118/2. No hybridization signal was observed for the non-
toxic strain N-C 143, which possesses mcy genes, but is
deficient in microcystin production (Mikalsen et al., 2003).
Using an NMT-specific primer pair (see Fig. 1), PCR
products were obtained from all toxic Microcystis strains,
but not from strain N-C 143 (Table 1), thus confirming the
Southern hybridization results. Sequencing of the NMT
region (see Fig. 1) showed that strain N-C 118/2 lacked the
HindIII restriction site present in the other hybridizing
strains, thus explaining the polymorphism seen for this
The amino acid residues constituting the S-adenosyl-
methionine (SAM)-binding site of cyclosporin synthetase
NMT domains (Velkov & Lawen, 2003) were identified in
the deduced amino acid sequences of the NMT domains.
The NMT domain sequence from Microcystis strain K-139
(AB019578), which exclusively produces [Dha7]-MC, was
also included in the NMT dataset. The analysis revealed
that invariant amino acid residues in the SAM-binding site
(indicated in Fig. 3A) were intact in all strains producing
only [Mdha7]-MC. In all methylation-defective strains one
or more of these amino acid residues were altered (Fig.
3A). Interestingly, in strain N-C 324/1, which produces
both [Mdha7]-MC and [Dha7]-MC (Mikalsen et al., 2003),
the invariant Gln residue in the SAM-binding site was
replaced by Arg (Fig. 3A).
Apart from these functionally relevant differences, the
NMT sequences showed relatively little variation (1–7%)
between strains (as might be anticipated from the almost
identical Southern blots). However, an alignment of the 55
informative sites present showed a block-like structure with
respect to phylogenetic affiliation. For all sequences one or
more blocks showed similarity to different subsets of the
strains examined, i.e. a mosaic pattern (Supplementary
Fig. S1A). Such block differences were originally demon-
strated for the rbcLX operon in cyanobacteria; they were
attributed to horizontal (lateral) gene transfer and denoted
mosaic structures (Rudi et al., 1998) – a term we will use in
the following. The split decomposition analysis revealed a
reticulate phylogeny (Fig. 3B) and the Phi test found
(P,0.01). All three recombination detection programs
used suggested recombination events in this dataset
Adenylation domains in modules McyB1 and
Using the mcyB-specific primer pair, a 1406/1409 bp
segment coding for almost the entire A domain of
McyB1 was amplified and sequenced from all toxin-
producing strains and from strain N-C 143 (Table 1). An
alignment of the sequences was used to identify the 39
breakpoint of the recombination event involving the A-
domain-encoding region of mcyB reported by Mikalsen et
al. (2003). The 59 breakpoint of this recombination event is
located near the conserved motif A3 (Mikalsen et al., 2003).
Recombination detection programs suggested a 39 break-
point near the conserved motif A8 (at position 1209 in
nucleotide alignment) (Table 2) in strain N-C 31. Visual
inspection of amino acid alignment revealed a putative
recombination breakpoint near the conserved motif A9 in
strains N-C 118/2, N-C 161/1 and PCC 7806 (Fig. 4). This
breakpoint could not be detected by Mikalsen et al. (2003)
due to the shorter McyB1 sequences studied in that work.
Aligning the four B-like nucleotide sequences revealed 3–
6% sequence variation. The 16 informative sites found in
the alignment revealed a mosaic structure (Supplementary
Fig. S1B). A reticulate phylogeny was obtained by split
decomposition analysis (Fig. 5) and the Phi test found
(P,0.01). However, no recombination events received
statistical support by the recombination detection pro-
grams (Table 2). The C-like McyB1 A domain nucleotide
sequences showed 2–5% sequence variation. The split
decomposition analysis revealed a reticulate phylogeny
Fig. 1. The microcystin synthetase gene cluster in M. aeruginosa PCC 7806. The relative positions of PCR primers used to
amplify flanking regions of mcy gene clusters and parts of mcyA, mcyB and mcyC are shown. The positions of the probes used
for Southern hybridization are indicated.
Genetic mosaics in the microcystin cluster
Fig. 2. Southern blot analyses of the mcy
gene cluster and flanking region. The relative
positions of primers, Southern probes and
Genomic DNA was digested with HindIII and
hybridized with the dnaN and uma1 probes.
Strains containing mcy genes (Mikalsen et al.,
2003) are indicated in bold. (A) Organization in
strain PCC 7806 of the region in mcyA that
contains the NMT domain and Southern blot
analysis of the corresponding region in various
strains. The A, NMT and C domains are
microcystin isoforms are indicated by black
dots. (B) Organization of the upstream region
flanking the mcy gene cluster in strain PCC
7806 and Southern blot analysis of this region
in various strains. (C) Organization of the
downstream region flanking the mcy gene
cluster in strain PCC 7806 and Southern blot
analysis of this region in various strains.
A. Tooming-Klunderud and others
1890 Microbiology 154
(Fig. 5) and statistically significant evidence for recom-
bination wasfound by
Recombination was also detected by the recombination
detection programs (Table 2) and suggested by the mosaic
structure of informative sites (Supplementary Fig. S1C).
The A-domain-encoding region of mcyC was successfully
amplified from all toxic strains as well as the non-toxic
strain N-C 143 (Table 1). The sequences showed 0–5%
sequence variation. Multiple recombination events were
detected in this dataset by the mosaic structure of
informative sites (Supplementary Fig. S1C); the reticulate
phylogeny revealed by split decomposition (Fig. 5) was
supported by the Phi test (P,0.01) and all recombination
detection programs (Table 2).
A comparison of mcyB and mcyC A-domain-encoding
regions from the same strain revealed relatively low
sequence variation (9–13%) in strains possessing the C-
like A domain in McyB1. In strains with a B-like A domain
in McyB1, these sequences were rather divergent, with 39–
40% sequence variation. Split decomposition analysis
including all mcyB and mcyC sequences revealed a
reticulate phylogeny within, but not between the clusters
(P,0.01) (Fig. 5). The 161 informative sites found in the
1063 bp alignment of mcyB (C-like) and mcyC sequences
showed a mosaic structure, indicating recombination also
between these regions (Supplementary Fig. S1C). Putative
recombination events involving the A domain regions of
mcyB (C-like) and mcyC identified by the recombination
detection programs are listed in Table 2. The phylogenetic
analyses of all A domain amino acid sequences revealed
three clades: McyB1 (B-like), McyB1 (C-like) and McyC
(Supplementary Fig. S2).
Regions flanking the mcy gene cluster
Probes derived from the genes flanking the mcy gene
cluster in strain PCC 7806, dnaN and uma1 (Fig. 1) (Tillett
et al., 2000), were used for Southern hybridization analysis
of flanking regions. All 17 strains gave positive hybridiza-
tion signals with both probes (Fig. 2B, C). For the uma1
probe, all mcy-containing strains (Mikalsen et al., 2003)
gave a 4.5 kb HindIII fragment similar to that from strain
PCC 7806. The dnaN probe displayed a far more variable
restriction pattern, with few shared bands between the
The primer pairs dnaN-mcyJ and mcyC-uma1 (Fig. 1 and
Supplementary Table S1) amplified flanking regions from
all toxic strains. The mcy-containing non-toxic strain N-C
143 gave a PCR product with the mcyC-uma1 combination
but not with the dnaN-mcyJ primer pair. The region
between dnaN and mcyB could however be amplified using
the dnaN-676R (mcyB-located) primer pair. Sequencing
revealed a deletion of the whole mcyD–J gene cluster and a
large part of mcyA (everything before position 43852 in
AF183408). Comparison of DNA sequences from the toxic
strains showed a high degree of conservation of all coding
regions. A 90 bp intergenic region between mcyC and
Fig. 3. Region of mcyA encoding the NMT
domain in Microcystis strains. Strains pro-
ducing desmethyl7-microcystin isoforms are
indicated by black dots. (A) Amino acid
sequences of NMT domains. Amino acid
residues constituting the SAM-binding-site
according to Velkov & Lawen (2003) are
indicated. Function of residues altered in
non-methylating strains: Gly (G) residues form
hydrogen bonds with the Met moiety of SAM.
Acidic residue (D/E) forms hydrogen bonds
with the adenine ribose. Gln (Q) forms
hydrogen bonds with Met (Velkov & Lawen,
2003). (B) Split decomposition analysis of
NMT-domain sequences constructed using
SplitsTree4 at default setting, showing 1000
bootstrap replicas above 50%.
Genetic mosaics in the microcystin cluster
Table 2. Putative recombination events detected by RPD, GENECONV and MaxChi2 (events detected by two or more programs are
FragmentD, P-valueFragmentD, P-value FragmentD, P-valueFragmentD, P-value
Putative recombination events within NMT-domain-encoding sequences of mcyA
1–399, 0.001 1–440, ,0.001N-C 228/1
1–440, 0.00215–411, ,0.001
406–1129, 0.006–– 492–1142, ,0.001
295–1129, 0.045–– 15–1128, ,0.001
–– 1130–1490, ,0.001825–1470, ,0.001
– 517–1082, ,0.001– 517–1082, ,0.001
295–479, 0.016–– 295–479, 0.008
Putative recombination events within A-domain-encoding sequences of mcyB1
1–254, ,0.0011–302, ,0.001N-C 143
N-C 264, 324/1, 169/7, 57
N-C 31, 118/2, 161/1, PCC 7806
1–486, ,0.001169–491, 0.006––
1209–1405, ,0.001 1243–1405, ,0.001 1243–1405, ,0.0011209–1404, ,0.001
Putative recombination events within mcyB1 sequences encoding C-like adenylation domaind
33–953, 0.001N-C 143
–– 33–953, ,0.001
N-C 57 (143, 228/1)1062–1362, 0.0041066–1338, 0.004–1068–1404, ,0.001
33–633, 0.002––33–633, ,0.001
Putative recombination events within A-domain-encoding mcyC sequences
36–169, 0.009N-C 118/2
1–203, ,0.0011–246, 0.002 ––
––43–454, 0.004111–417, ,0.001
Putative recombination events between sequences encoding A domains of McyB1 (C-like) and McyC
297–996, ,0.001262–1062, ,0.001
McyC N-C 31 (57, 143, 161/1,
228/1, 264, 324/1)
McyB1 N-C 264 –249–1014, ,0.001
McyB1 N-C 169/7
McyC N-C 31 (57, 161/1, 169/7)
876–1029, ,0.001–887–1013, ,0.001707–1056, ,0.001
McyC N-C 161/1
948–1062, 0.006–934–1055, 0.006–
#Pairs of strains analysed for recombination.
*Reference to programs (see Methods).
DRelative position of the recombining regions in our alignments. The alignment of NMT-domain-encoding mcyA sequences covered positions
38722–40212 in accession no. AF183408 from strain PCC 7806. The alignments of A-domain-encoding sequences in mcyB1 covered positions
46789–47834 in accession no. AF183408 from strain PCC 7806. The alignment of A-domain-encoding sequences in mcyC covered positions
53181–54242 in accession no. AF183408 from strain PCC 7806.
dNo recombination events were detected within sequences encoding the B-like A domain of McyB1.
A. Tooming-Klunderud and others
uma1 was also conserved in all strains except strain N-C
118/2, which has a shorter (80 bp) region with no apparent
similarity to the intergenic regions from the other strains
(Fig. 6A). In contrast, the dnaN–mcyJ intergenic regions
were highly different among the strains, with regard to
both length (from 360 to 2192 bp) and sequence. Based on
the alignments, the region was divided into segments A–D
(Fig. 6A and Supplementary Fig. S3) that were assigned to
subgroups containing similar sequences (colour coded in
Fig. 6A). The dnaN–mcyJ sequence from strain N-C 118/2
Fig. 4. Alignment of the A domain amino acid sequences in the first module of McyB. Conserved motifs are underlined and
genetic variants (B-like and C-like) indicated to the right of the alignment.
Genetic mosaics in the microcystin cluster
was at the dnaN end highly similar to the corresponding
sequences from strains N-C 161/1, N-C 169/7, N-C 264
and N-C 324/1, while it was most similar to the sequences
from strains N-C 228/1, N-C 31, N-C 57 and PCC 7806 at
the mcyJ end, indicating recombination.
A 1585 bp insertion containing an IS element with
similarity to Tn5-like transposases from Gloeobacter
violaceus PCC 7421 (NP923092) and bacteriophage WO
(BAA89624) was found in strain N-C 228/1. This IS
element (named ISMae7 and submitted to the ISfinder
database) belongs to the IS4 family and consists of 7 bp
direct repeats, 29 bp terminal inverted repeats (Fig. 6B)
and a single 1407 bp ORF transcribed in the same direction
as mcyJ. A GATC methylation site known to play a
modulating role in transposition activity (Mahillon &
Chandler, 1998) was found in the left inverted repeat. The
transposase encoded by ISMae7 contains a DDE motif
known to be necessary for efficient DNA transposition
(Mahillon & Chandler, 1998) and has probably retained
transposition activity, since no stop codons were present
within the ORF. Direct repeats associated with ISMae7
were also found in the dnaN–mcyJ intergenic spacer in
strains N-C 118/2 and N-C 143. Interestingly, another
16 bp direct repeat not associated with ISMae7 was present
in dnaN–mcyJ intergenic sequences from several strains
Phylogenetic analysis of the mcy flanking regions gave fairly
similar trees, except for strain N-C 324/1, which clustered
with different clades at the two flanks, and the clade
including strains N-C 31, N-C 57, N-C 228/1 and PCC
7806, which showed incongruent phylogenies (Fig. 7). Split
decomposition analysis indicated recombination events for
both flanks (Fig. 7), while the Phi test found statistically
significant evidence for recombination only in the dnaN–
mcyJ alignment (P,0.01).
PCR with the dnaN-uma1 primer pair gave PCR products
from five of seven non-toxic strains investigated (Table 1).
PCR with the hypX (Nishizawa et al., 2007) and uma1
primer pair gave a PCR product for one of the two
remaining non-toxic strains, N-C 43 (Table 1). In three
strains (N-C 123/1, N-C 144 and N-C 279), dnaN was
found to be situated close to uma1. The intergenic region
between these genes showed no similarity to the intergenic
regions in the toxic strains, except for a 80 bp segment
close to uma1 that was similar to the region between mcyC
and uma1 in strain N-C 118/2 (Fig. 7A). In three other
non-toxic Microcystis strains, the genetic organization
downstream of uma1 was different: in strains N-C 166
and N-C 172/5, an ORF encoding a protein homologous to
FtsH (Kaneko et al., 1995) and in strain N-C 43, a
hypothetical protein (HypX) (Nishizawa et al., 2007) was
found (Fig. 6A). The flanking regions of strain N-C 122/2
could not be amplified using these primer pairs, indicating
differences in the flanking regions of its mcy gene cluster,
and therefore this strain is not included in Fig. 6(A).
Loss of N-methyltransferase activity is associated
with specific point mutations in mcyA in
Deletion of the entire NMT domain in McyA has been
reported to cause production of desmethyl7-microcystin in
Fig. 5. Split decomposition analysis of gene sequences encoding A domains in modules McyB1 and McyC constructed using
SplitsTree4 at default setting, showing 1000 bootstrap replicas above 50%.
A. Tooming-Klunderud and others
some strains of Planktothrix (Kurmayer et al., 2005) and
Anabaena (Fewer et al., 2008). Our results indicate that the
production of desmethyl7-microcystin by five of our
Microcystis strains is not due to deletion of the NMT
domain, but rather to specific point mutations altering
amino acid residues in the cofactor SAM-binding site. Not
all mutations were associated with total inactivation of the
NMT domain, as shown for strain N-C 324/1, where
replacement of invariant Gln with Arg (Fig. 3A) was
associated with co-production of methylated and non-
methylated microcystin isoforms (Mikalsen et al., 2003).
The replacement of the second conserved Gly residue (as in
strain N-C 57, Fig. 3A) has previously been shown to
inactivate the NMT domain of pyochelin synthetase in
Pseudomonas aeruginosa (Patel & Walsh, 2001). The effect
of the remaining amino acid residues in the SAM-binding
site has not been verified by mutations. However, residues
which form the SAM-binding site (Velkov & Lawen, 2003)
were found to be replaced only in Microcystis strains
producing desmethyl7-microcystin, and always with resi-
dues with different physicochemical properties (Fig. 3A).
Although the effects of these replacements on the
methyltransferase activity have not yet been demonstrated,
the results suggest that the NMT domain may be
inactivated by point mutations, in addition to the complete
deletion of this domain described in other nonmethylating
strains (Fewer et al., 2008; Kurmayer et al., 2005).
Inter- and intragenomic recombinations within
and between domains
Previously, we have shown recombination involving the A
domain of McyB1 resulting in altered substrate specificity
(Mikalsen et al., 2003). Here we show that this recombina-
tion included the A domain segment between the conserved
motifs A3 and A8/A9 (Table 2, Fig. 4). Notably, Fewer et al.
(2007) have described a similar recombination in Anabaena
spp. and Hapalosiphon hibernicus, involving only the A
domain regions of mcyB1 and mcyC and not affecting the
condensation (C) domains in these modules. In both the
Hapalosiphon) case the recombinations involve larger,
non-identical functional units (different A domains) and
affect essentially an entire domain. For Microcystis the 59
recombination breakpoint (near the conserved motif A3) is
the same in all strains, but the downstream breakpoint has
two different locations (near conserved motifs A8 in strain
N-C 31 and A9 in strains N-C 118/2, N-C 161/1 and PCC
7806), separated by about 120 bp (Fig. 4). This might
indicate two independent recombination events. However,
as seen from the high similarity between the conserved A8
and A9 regions of McyB1 and McyC in strains PCC 7806, K-
139 and UV027 (AP183408, AB019578 and AP458094,
respectively), the variable location of the 39 breakpoint may
alternatively be the result of a recombination event between
two different A domain sequences followed by intragenomic
or intergenomic recombinations spanning the regions
encoding the conserved motifs A8 and A9.
Our data also indicate several recombination events
between the A domain regions of mcyB (C-like) and
mcyC, covering almost the whole segment between the
conserved motifs A3 and A8 (as in strain N-C 264) or
shorter segments of the A domain (Supplementary Fig. S1C
and Table 2) and mainly resulting in the replacement of
McyB1 A domain segments with corresponding McyC
segments. Recombination between these two A-domain-
encoding gene regions has been shown also for cyanobac-
teria from other genera (Fewer et al., 2007; A. Tooming-
Klunderud, D. P. Fewer, T. Rohrlack, J. Jokela, L.
Rouhiainen, K. Sivonen, T. Kristensen & K. S. Jakobsen,
Our data suggest multiple recombination events involving
smaller segments within all examined domains of mcyABC
operon. These recombinations lead to a mosaic pattern of
phylogenetic affinities in the alignments covering nearly
the entire analysed sequence of the NMT domain and the
McyB1A domain. As in other studies of the NMT domain
of Microcystis (Tanabe et al., 2004) and the McyB1A
domain of Planktothrix (Kurmayer et al., 2005; Kurmayer
& Gumpenberger, 2006) many of these recombinations do
not change the amino acid sequence. Within the A
domain of McyC, however, such a pattern was mainly
found upstream of the conserved motif A3 (Table 2). The
apparent lack of recombination in the region coding for
the substrate-binding part of this domain may be
explained by a more restricted function of the Arg-
activating McyC A domain compared to the A domain of
McyB1, which will activate several amino acids (Mikalsen
et al., 2003).
Evidence indicates that all microcystin variants are
synthesized by a single enzyme complex. Given a single
mcy operon in each strain, the observed recombinations
(Kurmayer & Gumpenberger, 2006; Tanabe et al., 2004;
present study) most likely represent intergenomic processes
and are thus manifestations of horizontal gene transfer
(HGT) between strains. The frequent HGT between closely
related cyanobacterial strains was originally suggested by
Rudi et al. (1998) and recently reinforced by whole-
genome analysis of cyanobacteria (Zhaxybayeva et al.,
2006) and in a general bacterial context (Papke et al.,
2007). HGT may be the result of uptake of free DNA
through natural transformation, since Microcystis is
naturally competent (Dittmann et al., 1997), or DNA
shuttling catalysed by cyanophages through transduction,
or possibly due to other transposable elements.
The mcy flanking sequences do not indicate
frequent transfer of complete mcy gene clusters
between Microcystis strains
In all microcystin-producing strains investigated here the
same two genes (dnaN and uma1) flank the mcy operon, in
agreement with Tillett et al. (2001), who previously have
shown that uma1 is located upstream of mcyC in toxic
Microcystis strains. Thus, it seems likely that the genomic
Genetic mosaics in the microcystin cluster
A. Tooming-Klunderud and others
1896 Microbiology 154
location of the mcy gene cluster in Microcystis spp. is the
same in all strains examined, in agreement with another
recent study (Nishizawa et al., 2007). These findings do not
support the idea of frequent transfer of complete mcy gene
clusters between Microcystis strains. The same two genes
(dnaN and uma1) were also found to be neighbours in
three of the non-toxic Microcystis strains, as also reported
by Nishizawa et al. (2007), indicating that loss of the mcy
gene cluster was not accompanied by further rearrange-
ments in this genomic region. Partial loss of the mcy gene
cluster in strain N-C 143 clearly illustrates that loss of
microcystin production can be caused by deletion (Fig. 6A).
A gene encoding a putative transposase (uma4) has
Microcystis strain PCC 7806 (Tillett et al., 2000). Here we
in themcy region of
report the presence of an IS element (ISMae7) in the
intergenic region between mcyJ and dnaN in strain N-C
228/1 (Fig. 6). Recently, two other genes coding for
different types of transposases have been identified between
dnaN and mcyJ in several Microcystis strains (Nishizawa
et al., 2007), but none of these reported transposases are
similar to the one encoded by uma4. The additional direct
repeats detected by us in this intergenic region in several
strains may also indicate an additional, now lost IS
element. According to these results, the intergenic region
between dnaN and mcyJ seems to be a recombinatorial ‘hot
spot’ while the mcyC–uma1 region is more conserved.
The frequent recombinations – representing both intrage-
nomic and intergenomic rearrangements – give rise to
novel mcy gene cluster variants, most of which encode
Fig. 6. Regions flanking the mcy operon in Microcystis strains. (A) A schematic overview of both flanks. In strains containing
mcy genes, the intergenic region between dnaN and mcyJ was divided into segments based on DNA sequence similarity
(Supplementary Fig. S2). For each segment, sequences marked with the same colours are highly similar, while unique colours
indicate different sequences. Strains producing microcystins are indicated by red diamonds. *For strains N-C 166 and N-C
172/5, the PCR product contained the uma1-R primer in both ends. (B) Parts of the sequences of the dnaN–mcyJ intergenic
region in strains containing mcy genes. Inverted repeats in the IS element present in strain N-C 228/1 (ISMae7) and direct
repeated sequences (DR) generated on insertion are indicated. Directly repeated sequences not associated with ISMae7,
present in some strains, are also indicated.
Fig. 7. Phylogenetic and split-decomposition analysis of regions flanking the mcy gene cluster in Microcystis spp. Both NJ and
Bayesian trees revealed the same topology (NJ tree shown). Bootstrap values (1000 replicates) and posterior probabilities are
indicated at the nodes. Strain N-C 324/1, for which the two flanks cluster with different clades, is marked by a grey box. (A)
Intergenic region between dnaN and mcyJ. (B) Intergenic region between uma1 and mcyC.
Genetic mosaics in the microcystin cluster
peptide synthetases with the same biosynthetic properties,
due to a constant selective pressure. Changes in selection
will, however, favour some variants within a particular
ecosystem. The results presented here along with some
previous studies (Kurmayer et al., 2005; Kurmayer &
Gumpenberger, 2006; Mikalsen et al., 2003; Tanabe et al.,
2004) suggest that the function of microcystins needs to be
understood within the frame of the natural ecosystems of
the bacterial strains. It seems likely that the processes
promoting variation within the microcystin synthetase
genes are crucial for adaptivity of a given population within
a particular habitat.
We thank Randi Skulberg for providing the N-C strains and Hans
Utkilen for providing the PCC 7806 strain. We also thank Trine B.
Rounge and Thomas Rohrlack for fruitful discussions. This work was
supported by the Research Council of Norway by a grant 157338/140
Bruen, T. C., Philippe, H. & Bryant, D. (2006). A simple and robust
statistical test for detecting the presence of recombination. Genetics
Christiansen, G., Fastner, J., Erhard, M., Bo ¨rner, T. & Dittmann, E.
(2003). Microcystin biosynthesis in Planktothrix: genes, evolution,
and manipulation. J Bacteriol 185, 564–572.
Christiansen, G., Kurmayer, R., Liu, Q. & Bo ¨rner, T. (2006).
Transposons inactivate biosynthesis of the nonribosomal peptide
microcystin in naturally occurring Planktothrix spp. Appl Environ
Microbiol 72, 117–123.
Dittmann, E., Neilan, B. A., Erhard, M., von Do ¨hren, H. & Bo ¨rner, T.
(1997). Insertional mutagenesis of a peptide synthetase gene that is
responsible for hepatotoxin production in the cyanobacterium
Microcystis aeruginosa PCC 7806. Mol Microbiol 26, 779–787.
Espelund, M., Stacy, R. A. & Jakobsen, K. S. (1990). A simple method
for generating single-stranded DNA probes labeled to high activities.
Nucleic Acids Res 18, 6157–6158.
Fewer, D. P., Rouhiainen, L., Jokela, J., Wahlsten, M., Laakso, K.,
Wang, H. & Sivonen, K. (2007). Recurrent adenylation domain
replacement in the microcystin synthetase gene cluster. BMC Evol Biol
Fewer, D. P., Tooming-Klunderud, A., Jokela, J., Wahlsten, M.,
Rouhiainen, L., Kristensen, T., Rohrlack, T., Jakobsen, K. S. &
Sivonen, K. (2008). Natural occurrence of microcystin synthetase
deletion mutants capable of producing microcystins in strains of the
genus Anabaena (Cyanobacteria). Microbiology 154, 1007–1014.
Galau, G. A., Hughes, D. W. & Dure, L., III (1986). Abscisic acid
induction of cloned cotton late embryogenesis-abundant (Lea)
mRNAs. Plant Mol Biol 7, 155–177.
Guindon, S. & Gascuel, O. (2003). A simple, fast, and accurate
algorithm to estimate large phylogenies by maximum likelihood. Syst
Biol 52, 696–704.
Huson, D. H. & Bryant, D. (2006). Application of phylogenetic
networks in evolutionary studies. Mol Biol Evol 23, 254–267.
Kaneko, T., Tanaka, A., Sato, S., Kotani, H., Sazuka, T., Miyajima, N.,
Sugiura, M. & Tabata, S. (1995). Sequence analysis of the genome of
the unicellular cyanobacterium Synechocystis sp. strain PCC6803. I.
Sequence features in the 1 Mb region from map positions 64% to
92% of the genome. DNA Res 2, 153–166.
Kumar, S., Tamura, K. & Nei, M. (2004). MEGA3: Integrated software
for molecular evolutionary genetics analysis and sequence alignment.
Brief Bioinform 5, 150–163.
Kurmayer, R. & Gumpenberger, M. (2006). Diversity of microcystin
genotypes among populations of the filamentous cyanobacteria
Planktothrix rubescens and Planktothrix agardhii. Mol Ecol 15,
Kurmayer, R., Christiansen, G., Fastner, J. & Bo ¨rner, T. (2004).
Abundance of active and inactive microcystin genotypes in popula-
tions of the toxic cyanobacterium Planktothrix spp. Environ Microbiol
Kurmayer, R., Christiansen, G., Gumpenberger, M. & Fastner, J.
(2005). Genetic identification of microcystin ecotypes in toxic
cyanobacteria of the genus Planktothrix. Microbiology 151, 1525–1533.
Mahillon, J. & Chandler, M. (1998). Insertion sequences. Microbiol Mol
Biol Rev 62, 725–774.
Marahiel, M. A., Stachelhaus, T. & Mootz, H. D. (1997). Modular
peptide synthetases involved in nonribosomal peptide synthesis.
Chem Rev 97, 2651–2674.
Martin, D. P., Williamson, C. & Posada, D. (2005). RDP2:
recombination detection and analysis from sequence alignments.
Bioinformatics 21, 260–262.
Mikalsen, B., Boison, G., Skulberg, O. M., Fastner, J., Davies, W.,
Gabrielsen, T. M., Rudi, K. & Jakobsen, K. S. (2003). Natural
variation in the microcystin synthetase operon mcyABC and impact
on microcystin production in Microcystis strains. J Bacteriol 185,
Nishizawa, T., Asayama, M., Fujii, K., Harada, K. & Shirai, M. (1999).
Genetic analysis of the peptide synthetase genes for a cyclic
heptapeptide microcystin in Microcystis spp. J Biochem 126,
Nishizawa, T., Ueda, A., Asayama, M., Fujii, K., Harada, K., Ochi, K. &
Shirai, M. (2000). Polyketide synthase gene coupled to the peptide
synthetase module involved in the biosynthesis of the cyclic
heptapeptide microcystin. J Biochem 127, 779–789.
Nishizawa, T., Nishizawa, A., Asayama, M., Harada, K. & Shirai, M.
(2007). Diversity within the microcystin biosynthetic gene clusters
among the genus Microcystis. Microbes Environ 22, 380–390.
Nylander, J. A. A. (2004). MrModeltest v2. Program distributed by the
author. Evolutionary Biology Centre, Uppsala University.
Padidam, M., Sawyer, S. & Fauquet, C. M. (1999). Possible emergence
of new geminiviruses by frequent recombination. Virology 265,
Papke, R. T., Zhaxybayeva, O., Feil, E. J., Sommerfeld, K., Muise, D. &
Doolittle, W. F. (2007). Searching for species in haloarchaea. Proc Natl
Acad Sci U S A 104, 14092–14097.
Patel, H. M. & Walsh, C. T. (2001). In vitro reconstitution of the
Pseudomonas aeruginosa nonribosomal peptide synthesis of pyochelin:
characterization of backbone tailoring thiazoline reductase and N-
methyltransferase activities. Biochemistry 40, 9023–9031.
Vaitomaa, J., Bo ¨rner, T. & Sivonen, K. (2004). Phylogenetic evidence
for the early evolution of microcystin synthesis. Proc Natl Acad Sci U
S A 101, 568–573.
Ronquist, F. & Huelsenbeck, J. (2003). MrBayes 3: Bayesian
phylogenetic inference under mixed models. Bioinformatics 19,
Rouhiainen, L., Vakkilainen, T., Siemer, B. L., Buikema, W.,
Haselkorn, R. & Sivonen, K. (2004). Genes coding for hepatotoxic
A. Tooming-Klunderud and others
heptapeptides (microcystins) in the cyanobacterium Anabaena strain
90. Appl Environ Microbiol 70, 686–692.
Rudi, K., Skulberg, O. M. & Jakobsen, K. S. (1998). Evolution of
cyanobacteria by exchange of genetic material among phyletically
related strains. J Bacteriol 180, 3453–3461.
Skulberg, R. & Skulberg, O. M. (1990). Research with Algal Cultures.
NIVA’s Culture Collection of Algae. Oslo: Norway.
Tanabe, Y., Kaya, K. & Watanabe, M. M. (2004). Evidence for
recombination in the microcystin synthetase (mcy) genes of toxic
cyanobacteria Microcystis spp. J Mol Evol 58, 633–641.
Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F. &
Higgins, D. G. (1997). The CLUSTAL_X Windows interface: flexible
strategies for multiple sequence alignment aided by quality analysis
tools. Nucleic Acids Res 25, 4876–4882.
Tillett, D., Dittmann, E., Erhard, M., von Do ¨hren, H., Bo ¨rner, T. &
Neilan, B.A.(2000). Structural organization of microcystin
biosynthesis in Microcystis aeruginosa PCC7806: an integrated
peptide-polyketide synthetase system. Chem Biol 7, 753–764.
Tillett, D., Parker, D. L. & Neilan, B. A. (2001). Detection of toxigenicity
by a probe for the microcystin synthetase A gene (mcyA) of the
cyanobacterial genus Microcystis: comparison of toxicities with 16S
rRNA and phycocyanin operon (Phycocyanin Intergenic Spacer)
phylogenies. Appl Environ Microbiol 67, 2810–2818.
Velkov, T. & Lawen, A. (2003). Mapping and molecular modeling of
domains of the multifunctional polypeptide cyclosporin synthetase.
J Biol Chem 278, 1137–1148.
Welker, M. & von Do ¨hren, H. (2006). Cyanobacterial peptides –
nature’s own combinatorial biosynthesis. FEMS Microbiol Rev 30,
Zhaxybayeva, O., Gogarten, J. P., Charlebois, R. L., Doolittle, W. F. &
Papke, R. T. (2006). Phylogenetic analyses of cyanobacterial genomes:
quantification of horizontal gene transfer events. Genome Res 16,
Edited by: K. Forchhammer
Genetic mosaics in the microcystin cluster