OPEN 3 ACCESS Freely available online © PLOSI o -

Insights into the Physiology and Ecology of the Brackish- Water-Adapted Cyanobacterium Nodularia spumigena CCY9414 Based on a Genome-Transcriptome AnalysisBjörn Voß1, Henk Bolhuis2, David P. Fewer3, Matthias Kopf1, Fred Moke4, Fabian Haas1, Rehab El-Shehawy5, Paul Hayes6, Birgitta Bergman7, Kaarina Sivonen3, Elke Dittmann8, Dave J. Scanlan9, Martin Hagemann4, Lucas J. Stal2'10, Wolfgang R. Hess1"1 Genetics and Experimental Bioinformatics Group, Faculty o f Biology, University o f Freiburg, Freiburg, Germany, 2 Department o f Marine Microbiology, Royal Netherlands Institute o f Sea Research, Yerseke, The Netherlands, 3 Food and Environmental Sciences, Division o f Microbiology, Viikki Biocenter, University o f Helsinki, Helsinki, Finland, 4 Plant Physiology, Institute Biosciences, University o f Rostock, Rostock, Germany, 5IMDEA-Agua, Alcalá de Henares, Madrid, Spain, 6 Faculty o f Science, University o f Portsmouth, Portsmouth, United Kingdom, 7 Department o f Botany, Stockholm University, Stockholm, Sweden, 8 Institute fo r Biochemistry and Biology, University o f Potsdam, Golm, Germany, 9 School o f Life Sciences, University o f Warwick, Coventry, United Kingdom, 10 Department o f Aquatic Microbiology, University o f Amsterdam, Amsterdam, The Netherlands

AbstractNodularia spumigena is a filamentous diazotrophic cyanobacterium that dominates the annual late summer cyanobacteria! blooms in the Baltic Sea. But N. spumigena also is common in brackish water bodies worldwide, suggesting special adaptation allowing it to thrive at moderate salinities. A draft genome analysis of N. spumigena sp. CCY9414 yielded a single scaffold of 5,462,271 nucleotides in length on which genes for 5,294 proteins were annotated. A subsequent strand-specific transcriptome analysis identified more than 6,000 putative transcriptional start sites (TSS). Orphan TSSs located in intergenic regions led us to predict 764 non-coding RNAs, among them 70 copies of a possible ret rot rans poso n and several potential RNA regulators, some of which are also present in other N2-fixing cyanobacteria. Approximately 4% of the total coding capacity is devoted to the production of secondary metabolites, among them the potent hepatotoxin nodularin, the linear spumigin and the cyclic nodulapeptin. The transcriptional complexity associated with genes involved in nitrogen fixation and heterocyst differentiation is considerably smaller compared to other Nostocales. In contrast, sophisticated systems exist for the uptake and assimilation of iron and phosphorus compounds, for the synthesis of compatible solutes, and for the formation of gas vesicles, required for the active control o f buoyancy. Hence, the annotation and interpretation of this sequence provides a vast array of clues into the genomic underpinnings of the physiology of this cyanobacterium and indicates in particular a competitive edge of N. spumigena in nutrient-limited brackish water ecosystems.

Citation: Voß B, Bolhuis H, Fewer DP, Kopf M, Moke F, et al. (2013) Insights into the Physiology and Ecology o f the Brackish-Water-Adapted Cyanobacterium Nodularia spumigena CCY9414 Based on a Genome-Transcriptome Analysis. PLoS ONE 8(3): e60224. doi:10.1371/journal.pone.0060224

Editor: Paul Jaak Janssen, Belgian Nuclear Research Centre SCK/CEN, Belgium

Received January 15, 2013; Accepted February 23, 2013; Published March 28, 2013

Copyright: © 2013 Voß, e t al. This is an open-access article distributed under the terms o f the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: The authors thank the Gordon and Betty Moore Foundation fo r launching and supporting the Marine Microbiology initiative aiming at a substantial increase in the number o f genome sequences o f ecológica I ly-relevant marine microorganisms and the EU-project MaCuMBA (Marine Microorganisms: Cultivation Methods for Improving their Biotechnological Applications) (Grant agreement no: 311975) for supporting the final project phase. Work on N. spumigena CCY9414 at Rostock University was supported by a research stipendium for F.M. from the INF (Interdisziplinäre Fakultät, Maritime Systeme) and DFG (Deutsche Forschungsgemeinschaft) grant to M.H. Financial support from the BalticSea2020 Foundation and the Baltic Ecosystem Adaptive Management Program (to BB) is acknowledged. The research at the University o f Helsinki was funded by the Academy o f Finland grant no. 118637 to K.S. The article processing charge was funded by the German Research Foundation (DFG) and the Albert Ludwigs University Freiburg in the funding programme Open Access Publishing. The funders had no role in study design, data collection and analysis, decision to publish, or preparation o f the manuscript.

Com peting Interests: The authors have declared that no competing interests exist.

* E-mail: wolfgang.hess@biologie.uni-freiburg.de

Introduction

T ox ic cyanobacteria! blooms in aquatic ecosystems are a w o rld ­w ide problem , w h ich are predicted to increase according to the present scenarios o f clim ate change [1]. Here, we report the results o f a d ra ft genome analysis targeting Nodularia spumigena sp. CC Y 9414 (from here on N spumigena CCY9414), a toxin- producing, N 2-fix ing, filamentous cyanobacterium isolated from the brackish waters o f the southern Baltic Sea. N spumigena as mem ber o f the Nostocales has a com plex lifestyle, capable o f cell d iffe rentia tion w ith in the ir long trichomes [2]. Th is cyanobacte­rium can differentiate vegetative cells in to akinetes, heterocysts or horm ogonia. Heterocysts are specialized cells fo r N 2-fixation,

w hich develop a th ick cell w a ll and have lost photosystem I I in order to decrease the in terna l oxygen concentration to a level that allows nitrogenase activ ity du ring the day tim e (for reviews see [3,4]). Heterocysts are usually on ly form ed when combined nitrogen is not available, bu t in N spumigena A V I heterocyst d iffe rentia tion appeared to be uncoupled from the n itrogen supply [5]. Akinetes are cell types that serve the long-term survival o f the organism under stress and non-grow th pe rm itting conditions. I t is thought that N spumigena forms akinetes in the Baltic Sea during autum n. The akinetes sink and overw inter in the bottom sediments from where they m ay be m ixed back in to the water co lum n during spring and as such serve as the inocu lum fo r a new population [6].

PLOS ONE I w w w .p loso n e .o rg 1 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ra n scrip to m e o f a Toxic C yanobacte rium

H orm ogon ia are short m otile trichomes consisting o f small-sized vegetative cells. They are form ed from akinetes o r from vegetative cells and serve the dispersal o f the organism.

Heterocystous cyanobacteria o f d ie group Nostocales can be d ivided in to two m ajor groups. There are several genome sequences available for d ie clade encompassing species such as Nostoc punctiforme A T G G 29133, Anabaena sp. PGG 7120 (from here Anabaena PGG 7120) and Anabaena variabilis A T G G 29413, whereas for the od ier clade, includ ing Nodularia (F ig . 1), genome- level studies have only recendy been started [7], The strain N spumigena GGY9414 was isolated from brackish surface waters o f d ie Baltic Sea (near Bornholm). This isolate is a typical representative o f the b loom -fo rm ing p lanktonic filamentous N 2- fix ing cyanobacteria and an im portan t component in an ecological context. These cyanobacteria release considerable amounts o f die ‘new ’ n itrogen fixed in to die n itrogen-poor surface waters, thereby feeding die rest o f the com m unity w ith a key nutrient. They contribute an estimated annual n itrogen inp u t almost as large as

the entire riverine load and twice die atmospheric load in to the Baltic Sea proper [8,9],

However, a m ajor concern is the tox ic ity o f these blooms, which may severely interfere w ith hum an activities [10,11] and regularly causes anim al poisonings in coastal regions o f d ie Baltic Sea (e.g. [12,13]). For instance, N. spumigena produces the potent hepato- tox in nodularin [10] bu t i t is still unclear to w ha t extent the toxic blooms im pact on related food chains. H ig h phosphorus combined w ith low to undetectable n itrogen concentrations du ring the summer season (hence low N :P ratios) are p rinc ipa l factors favouring grow th and bloom form ation o f Nodularia in the stratified Baltic Proper and G u lf o f F in land [14], Th is phenomenon is pa rticu la rly pronounced under periods o f stably stratified warm water conditions when its gas vesicles provide buoyancy leading to the form ation o f large surface scums in die absence o f m ixing. The decomposition o f such blooms causes depletion o f dissolved oxygen con tribu ting to anoxic bottom waters across large areas o f d ie Baltic.

A BNostocales

Nostocales I

A na ba e na 90a A n a b a e n a 90bAnabaena flos-aquae PCC 9302 Nodularia PCC 9350

N o d u la r ia s p u m ig e n a CCY9414a N o d u la r ia s p u m ig e n a CCY9414b

Nostoc punctiform e PCC 73102 Tolypothrix 1AM Anabaena PCC 7120 Anabaena variabilis ATCC29413 Fischerella sp. BB98

Fischerella m a jo r NIES 592 Calothrix PCC 7714

— Leptolyngbya fox Synechococcus PCC 7002

-Synechocystis PCC 6803

— Cyanothece PCC 7424Microcystis PCC 7806

io p i Trichodesmium X70767

Trichodesmium erythraeum — Oscillatoria sp. PCC 7112

Lyngbya aestuarii CCY9616a

¡¡¿I Lyngbya sp. PCC 7419

95 Lyngbya aestuarii CCY9616b Acaryochloris m arina MBIC11017

Synechococcus PCC 7942 Synechococcus MW6C694 j— Synechococcus WH 7803

Synechococcus WH 8102 ~ Prochlorococcus M IT 9313

si I Prochlorococcus marinus SS120 91 Prochlorococcus MED4

-Thermosynechococcus BPl - Gloeobacter violaceus PCC 7421

Chlorobium tepidum

F ig u re 1 . G e n e ra l fe a tu r e s o f N. s p u m ig e n a C C Y 9 4 1 4 . A . P h o to m ic ro g rap h o f N. spum igena CCY9414 trichom es. The a rrow s p o in t to heterocysts. The ve rtica l bar co rresponds to 40 pm . B. P hy logene tic p o s itio n o f N. spum igena CCY9414 (boxed) w ith in th e cyanobacte ria ! p hyu m , based on its tw o 16S rRNA sequences (labe led a and b). The tw o sub-clades w ith in th e N ostocales, c lade I and c lade II, are ind ica te d . Species fo r w h ich a to ta l g e n o m e sequence is p u b lic ly ava ilab le , are in b lue . The sequence o f C hlorob ium tep idum TLS served as o u tg ro u p . The n um bers a t nodes re fer to b o o ts tra p s u p p o rt va lues (1000 re pe tit io n s ) i f > 6 0 % . The p h y lo g e n e tic tre e was g en e ra te d us ing th e M in im u m E vo lu tion m e th o d w ith in MEGA5 [158]. The o p tim a l tre e w ith th e sum o f b ranch le n g th = 0 .85445647 is show n . The tre e is d raw n to scale, w ith b ranch len g th s in th e sam e u n its as th o se o f th e e vo lu tio n a ry d is tances used to in fe r th e p h y lo g e n e tic tre e and are g ive n in th e n u m b e r o f base s u b s titu tio n s pe r site. The m u lt ip le sequence a lig n m e n t w as sh o rte ne d to a to ta l o f 1407 p os itio n s in th e fin a l da taset to inc lu d e also 16S rRNA sequences fro m species w ith o u t a g e n o m e sequence.doi:10 .1371/jou rna l.pone .0060224 .g001

PLOS ONE I w w w .p loso n e .o rg 2 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

Thus, as a diazotroph, JV. spumigena has a selective advantage under the v irtua lly nitrogen-free, stably stratified w arm brackish water conditions o f the Baltic Sea w ith its salinity gradient from 28 practica l salinity units (PSU, equivalent to perm ille) to almost freshwater conditions in the surface waters above the pycnodine. In the central Baltic Sea, the preferred hab itat o f jV. spumigena, the salinity varies between 1-2 PSU. JV. spumigena is found at sim ilar locations throughout the w orld , where brackish water conditions prevail, fo r instance in the Peel-Harvey in le t (Western Australia [15,16]), o r the Neuse R iver estuary (USA, [17]). In Austra lian brackish waters, JV. spumigena blooms usually fo rm between spring and autum n. The p rim a ry m otiva tion fo r this study was to obtain genomic in fo rm ation from brackish-water-adapted, b loom -fo rm ­ing and toxic cyanobacteria, in order to gain insights in to adaptations p e rm itting it to dom inate in brackish water environ­ments. The d ra ft genome sequence o f JV. spumigena C C Y9414 allows a comparative genome analysis o f its physiological capabilities. The genome analysis was complemented by a transcriptome-w ide m apping o f transcrip tional start sites (TSS) to be able to set its regulatory com plexity in the context o f previously studied cyanobacteria Synechocystis 6803 and Anabaena 7120 [18,19] and to identify the suite o f putative non-coding RN As (ncRNAs) [20,21].

Results and Discussion

General Genomic PropertiesThe 16S rRNA-based phylogenetic tree o f cyanobacteria shows

two clades containing representatives from the Nostocales, clade I and clade I I (F ig . 1). JV. spumigena C C Y 9414 is located in clade I as opposed to clade I I conta in ing the m uch better studied Nostocales Anabaena PCC 7120 and JV. punctiforme A T C C 29133 (PCC 73102). As in the closely related Anabaena sp. 90 [7], and some other related cyanobacteria [22], there are two 16S rR N A genes, w h ich differ by 4 nt (99% identity), labelled .Nodularia C C Y 9414 a and b. These 16S rR N A genes are associated w ith two d istinct ribosom al R N A opérons characterized by the ir d ifferent intergenic transcribed spacer types, one also containing the tR N A -Ile G A T and tR N A - A la T G C genes, whereas the other is lacking these tR N A genes, as also previously described for the section V cyanobacterium Fischerella sp. R V 14 [22].

As summarized in Table 1, the JV. spumigena C C Y 9414 draft genome sequence is d istributed over 264 contigs. F rom these, one m ajor scaffold o f 5,462,271 n t length and several short scaffolds (< 2 kb) were assembled. W ith this length, the genome appears smaller than those o f several o ther Nostocales sequenced before (6.34, 6.41, 7.75 and 8.23 Ml:) fo r Anabaena variabilis A T C C 29413, Anabaena 7120, T richodesm ium erythraeum IM S101 and Nostoc punctiforme PC Cl 73102) bu t larger than the m in im a l Nostocales genomes o f Cylindrospermopsis raciborskii CIS-505 and R aph i­diopsis b rook ii D 9 (3.9 and 3.2 Ml:) [23]) and is comparable to the genome a í Anabaena sp. strain 90 [7]. The genomic G C content is 42% and 5,294 pro te in-coding genes were modeled. W e predicted 48 tR N A genes and one tm R N A gene. The tR N A -L e u UAA gene contains a group I in tron , w h ich has been suggested as being o f ancient o rig in [24], w h ile the gene fo r the in itia to r tR N A , tR N A - fM e tGAT, is in tron-free, d ifferent from its ortho log in some other cyanobacteria [25].

The annotated scaffold o f 5,462,271 n t length is available under GenBank accession num ber AO FE00000000, add itiona lly the file containing all in fo rm a tion on mapped transcrip tional start points (TSS) can be downloaded from h ttp ://w w w .cya n o la b .d e / suppdata/Nodularia_genom e/Nodularia_spum igena_CCY9414. gbk.

Table 1. General genome a n d a n n o ta t io n in fo r m a t io n .

N. spumigena CCY9414

Genome Length 5,462,271

Scaffolds 1

Avg. Contig Size 26,308

Genomic GC% 42.00%

Genes 5,294

Coding% 80.00%

tRNA Count 48

rRNA Count 8

Introns 1

sRNAs 4

Inteins 2

Sigma factors 8

Total number o f TSS 6,519

Number o f gTSS 1,628

Number o f aTSS 2,084

Number o f iTSS 2,043

Number o f nTSS 764

For information on non-ubiquitious sRNAs, see Table 3. doi:10.1371 /journal.pone.0060224.t001

Mobile Genetic ElementsJV. spumigena C C Y9414 possesses 164 genes encoding transpos-

ases. These transposases were identified by B L A S T p searches against the IS finder [26] requ iring a B LA S T p value o f < le l ( ) -5 and were assigned to 11 d ifferent families, each containing 1-32 identical copies, w ith highest copy numbers found fo r the IS 2 0 0 / IS605, IS607 and IS630 families o f IS elements (Table SI). The large num ber, the h igh sequence s im ila rity and the fact that active prom oters were detected fo r m any o f the transposases indicate that a large pa rt o f the m obile genetic elements associated w ith them are active. Nevertheless, when norm alized to the genome size, the num ber o f transposase genes is sim ilar to m any other cyanobac­teria, fo r instance, 70 transposase genes are present in Synechocystis PCC 6803 w ith 3.7 M b genome size. However, jV. spumigena C C Y 9414 has far fewer transposases than other m arine N o-fix ing cyanobacteria such as Crocosphaera sp. W H 8501, w h ich has as m any as 1,211 transposase genes [27].

A nother class o f m obile elements in the JV. spumigena CCY9414 genome is represented by at least two d ifferent D iversity G enerating Retroelements (DGR1 and D G R 2). D G R s introduce vast amounts o f sequence diversity in to the ir target genes [28], using a distinct type o f reverse transcriptase (genes nsp38130 fo r D G R 1 and mj/>13150 for D G R 2; 70% am ino acid identity). The very strong nTSS located 199 n t downstream o f nsp38130 may give rise to the ncR N A interm ediate, which, fo llow ing reverse transcription, is essential fo r homologous recom bination in to the target site fo r codon rew riting and pro te in diversification [28], Follow ing previously established protocols [29], we identified nsp38150, encoding a FGE-sulfatase superfam ily-dom ain contain­ing prote in, as the like ly target o f D G R 1. Closely related D G R systems, inc lud ing homologs o f the Nsp.?<?i.?0 reverse-transcriptase and JVsp38150 FGE-sulfatase superfam ily proteins, exist in JV. punctiforme PCC 73102 (Npun_F4892, Npun_F4890,Npun_F4889) and, in Anabaena PCC 7120 (Alr3497, Alt'3495). However, the JV. spumigena C C Y 9414 genome contains 70

PLOS ONE I w w w .p loso n e .o rg 3 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

copies(&98% sequence identity) o f this potentia l D G R 1 ncR N A element consisting o f the transcribed region, suggesting that D G R 1 is a h igh ly active retroelement that also inserts in to non­coding regions independently o f its codon rew riting capability.

M oreover, a free-standing rot dom ain conta in ing reverse transcriptase (nspl0420) was annotated, w h ich belongs to the R N A -d irected D N A po lym erase:H N H endonuclease type. Such rot dom ain proteins are no t components o f retrotransposons or viruses. These genes occur frequendy in syntenic regions, evolve under pu rify ing selection and are found in all m ajor taxonom ic groups inc lud ing bacteria, protists, fungi, animals and plants, bu t the ir function is unknow n [30]. These genes also exist in many other cyanobacterial genomes, exemplified by A lr7241 in Anabaena PCC 7120 and three paralogs in Anabaena sp. 90. A th ird type o f putative reverse transcriptase is encoded by nsp37000.

F ig . 2A shows a comparison o f the predicted proteome o f A i spumigena C C Y 9414 w ith those o f other well-studied Nostocales, Nostoc punctiforme PCC 73102, Anabaena variabilis A T C C 29413 and Anabaena PCC 7120. The core set o f proteins comprises 2,778 gene clusters com m on to all fou r strains. A subgroup o f these gene clusters represents m ulti-copy gene families o f functiona l rele­vance. For example, N spumigena CC Y 9414 harbors four identical copies o f the psbA gene encoding the D 1 pro te in o f photosystem I I , 9 copies o f genes encoding proteins o f the C A B /E L IP /H L IP superfam ily b u t 2 hetP-like genes proposed to be involved in heterocyst d iffe rentia tion [31], whereas Anabaena PCC 7120 possesses 5 D l - and 8 C A B /E L IP /H L IP -c o d in g genes b u t 4 d ifferent hetPAske genes.

There are 608 gene clusters com m on to the three other Nostocales w ith the exclusion o f A i spumigena CC Y 9414 (Table S2). These are like ly genes specific fo r the Nostocales clade I I . However, w ith 1,098 potentia lly unique coding sequences (1,047 gene clusters) there are also a substantial num ber o f proteins in A i spumigena CC Y 9414 fo r w h ich no homologs exists in the clade I I genomes o r only at low s im ilarity (F ig . 2A; Table S3). F ig . 2B shows the taxonom ic relationships o f these A i spumigena CC Y9414 genes. The largest fraction (719 genes) could no t be assigned to any phylogenetic group (i.e. have not been reported before in any other organism). A bou t 30% o f the rem ain ing 379 genes have a clear cyanobacterial orig in . A no ther quite large group o f genes were assigned to the taxon bacteria because they could no t be unambiguously assigned to a particu la r group.

A m ong the 1,098 potentia lly unique A i spumigena C C Y9414 genes are genes that m ight be expected to be more m obile, such as several restriction-m odifica tion cassettes, glycosyltransfeases (e.g. the three genes nspl3820—13840), bu t also m any genes w ith a surprising annotation or taxonom ic relation. N o tew orthy are the genes nsp5‘280, nsp5300 and nsp5310, w hich resemble the genes M X A N 3 8 8 5 -38 8 3 o f Myxococcus xanthus D K 1622 fo r fim bria ! biogenesis outer membrane proteins functional in spore coat biogenesis [32].

In accordance w ith the p lanktonic lifestyle o f A i spumigena CCY9414, ten genes gopAlA2CNJKLFGVW (nspl5380- nspl5470) fo r gas vesicle proteins are arranged in one consecutive stretch o f 6,372 n t that are critica l fo r the regulation o f buoyancy and are not found in benthic A i spumigena [33,34].

Organization o f the Primary TranscriptomeThe d ra ft genome sequencing o f A i spumigena CC Y 9414 was

combined w ith an analysis o f its transcriptome. Follow ing established approaches fo r a transcriptome-wide m apping o f TSS [18,19], we analyzed a cD N A population enriched for p rim a ry transcripts obtained from an R N A sample o f A i spumigena CC Y 9414 grown under standard conditions. In total, 41,519,905

sequence reads were obtained, from these 40,577,305 unique reads were mapped to the A i spumigena C C Y 9914 scaffold. The m a jo rity o f these, 28,214,827 (70%) unique reads, am ounting to 2,819,120,699 bases o f cD N A , represented n o n -rR N A sequences, ind icating a very h igh efficiency o f the rR N A depletion and cD N A preparation. A pp ly ing a m in im um threshold o f 280 reads o rig ina ting w ith in a 7-nt w indow , 6,519 putative TSS were identified. In the absence o f in fo rm a tion about the real lengths o f 5 ' U TR s, a ll TSS were classified based on the ir position and according to published criteria [19]. Hence, a ll TSS w ith in a distance o f ^ 2 0 0 n t upstream o f an annotated gene were categorized as gene TSS. TSSs w ith in a pro te in-coding region, w h ich frequendy also contribute to the generation o f m RNAs, were classified as in terna l TSS (iTSS). TSSs for non-coding R NAs were found on the reverse complem entary strand fo r antisense R N As (aTSS) or w ith in intergenic regions fo r non-coding sRNAs (nTSS) (Table 1). A ccord ing to this classification, only 25% (1,628 gTSS) o f all TSS were in the classical arrangement 5 ' o f an annotated gene. However, sim ilar observations have been made d u ring genome-wide TSS m apping in other bacteria, inc lud ing the cyanobacteria Synechocystis 6803 [18] and Anabaena 7120 [19]. The TSS associated w ith the by far highest num ber o f reads is located upstream o f one o f the psbA genes (psbA l, nsp5370). The 50 gTSS associated w ith the highest numbers o f reads (Table 2) comprise one additional mem ber o f the psbA gene fam ily (psbA4, nsp35290), together w ith seven fu rther photosynthesis-related genes {psbV, cpcG3, transport proteins fo r inorganic carbon and carbon concentration and C alv in Cycle proteins). One o f the genes in this category encodes a CP12 pro te in (Table 2). CP12 proteins are small regulators o f the C alv in cycle in response to changes in lig h t availab ility, b u t recent evidence suggests add itiona l functions o f C P I2 proteins in cyanobacteria [35]. A functional class o f sim ilar size w ith in this top-50 group o f gTSS drives the transcrip tion o f translation-related genes fo r ribosomal proteins (S I4, S16, L Í9 , L32 and L35), the DnaJ chaperone, or translation factor IF3. The fact tha t photosynthesis- and transla­tion-related gTSS are so dom inant in the top-50 group illustrates that photosynthetic energy metabolism and pro te in biosynthesis were h igh ly active in the culture taken fo r R N A analysis.

Some o f the mapped TSS gave rise to orthologs o f non-coding transcripts in other cyanobacteria. For instance, the Anabaena PCC 7120 gene a!13278, whose m uta tion leads to the in a b ility to fix N2 in the presence o f 0 2 [36], was associated w ith an asRNA [19]. This was also observed fo r the N spumigena CC Y 9914 hom olog nspl5990. A nother example is the conservation o f the nitrogen- stress-induced R N A 3 (NsiR3) first observed in Anabaena PCC 7120. NsiR3 is a 115 n t sR N A that is strongly induced upon rem oval o f am m onia and contro lled by an N tcA b ind ing site [19]. The hom olog in N spumigena CC Y 9914 is transcribed from an nTSS at position 2888943, structurally conserved and also associated w ith a putative N tcA b ind ing site (G T G -N 8-T A C ) centered at position -41. A n overview o f identified ncR N As and fu rther details are presented in Table 3.

The transcriptome analysis allowed insight in to the expression and p rom oter organization o f genes involved in h igh ly divergent physiological processes. This in fo rm a tion is available by dow n­load ing the annotated genbank file associated w ith this m anuscript under h t tp : / /w w w .cyano lab .de /suppdata /N odularia_genom e/ Nodularia_spum igena_CCY9414.gbk. In the fo llow ing, we ana­lyzed in more detail genes involved in the form ation o f heterocysts, the regulation o f n itrogen metabolism and N2 fixa tion that were transcribed from h igh ly active TSS. The global nitrogen regulatory p ro te in N tcA was transcribed from a single TSS located 45 n t upstream o f the start codon, associated w ith a perfect

PLOS ONE I w w w .p loso n e .o rg 4 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

N 7120 N_punct

A var N_spumi

B

Bacteria 143

cellular organisms 94

>•o

♦Actinobacteria <phylum> 3

Bacteroidetes/Chlorobi group 1

Chloroflexi <phylum> 2

i------------------ ‘Acaryochloris 1Chroococcales 4 |

o — •Nostocales 0

Nostocaceae

Cyanobacteria 66

llatoriales 3

-------------------Crocosphaera watsonii WH 0003 2

Synechococcus sp. PCC 7335 1

----------------- •Anabaena sp. 90 8

------------------- Anabaena sp. CA 1

-------------------Cylindrospermopsis raciborskii 1

1----------------- ♦Nostoc flagelliforme 4

-------------------Raphidiopsis brookii 1

------------------ ‘Trichormus azollae 2

Leptolyngbya sp. PCC 7375 1

Lyngbya sp. PCC 8106 1

Oscillatoria sp. PCC 6506 1

Oscillatoriales cyanobacterium JSC-12 1

’Fischerella sp. JSC-11 3

— Fibrobacteres/Acidobacteria group 1

-•P roteobacteria 8

— unclassified Bacteria 1

-Archaea 1

-Eukaryota 1

Not assigned 1

No hits 719

F ig u r e 2 . C la s s if ic a t io n o f t h e p r e d ic te d N . s p u m ig e n a C C Y 9 4 1 4 p r o te o m e . A . C om parison o f all p red ic te d p ro te in s o f N. spum igena (N _spum i) a ga ins t th e p ro te o m e s o f o th e r w e ll-s tu d ie d N ostocales, N ostoc pun c tifo rm e sp. PCC 73102 (N _punct), A nabaena va riab ilis sp. ATCC 29413 (A_var) and A nabaena PCC 7120 (N_7120) based o n MCL c lus te rin g o f BLASTp results (m in im u m e-value: 10"8). The n um b e rs re fer to th e n u m b e r o f p ro te in c lusters in each ca tegory , th e n um b e rs in b rackets to th e to ta l n u m b e r o f in d iv id u a l p ro te in s . B . Ta xo n om ic to p h its fo r th e 1,098 N. spum igena CCY9414 s ing le to ns fro m p a rt A (Table S3) v isua lized b y MEGAN. d o i:1 0.1371/jo u rn a l.pone.0060224.g002

(G T A -N 8 -T A C ) N tcA -b ind ing m o tif [37]. In comparison, six d ifferent TSS were reported fo r the n tcA gene in Anabaena PCC 7120 [19,38,39]. S im ilarly, JV. spumigena C CY94T4 hetR was transcribed from a single TSS 109 n t upstream o f the start codon, compared to four TSS d riv ing the transcrip tion o f hetR in Anabaena PCC 7120 [19,38,40,41], I t should be stressed that the m ultip le TSS va. Anabaena PCC 7120 were detected by the same approach in the absence o f com bined n itrogen [19] as used here fo r JV.

spumigena CCY9414. Therefore, these genes that code fo r proteins central fo r the d ifferentia tion o f heterocysts and N2 fixation appear to be contro lled from less complex p rom oter regions in JV. spumigena C C Y 9414 when compared to the well-studied Anabaena PCC 7120. A sim plified genom e/transcrip tom e arrangement was also detected for the genes encoding glutam ine synthetase and the glutam ine synthetase inactiva ting factor IF7, g lnA and g ifA (uspi6180 and usp i6190). In Anabaena PCC 7120, these genes

PLOS ONE I w w w .p loso n e .o rg 5 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

Table 2. The 50 gTSS of protein-coding genes associated with the highest number of reads.

Position S Reads ID Gene Annotation

512315 + 4058097 nsp5370 psbA1 photosystem II protein D1

1911567 + 420026 nsp19010 rbpA2 RNA-binding protein

5407423 404779 nsp53100 rbpA1 RNA-binding protein

818076 + 254679 nsp8200 rps16 SSU ribosomal protein S16p

963628 + 194603 nsp9560 hypothetical protein

4901855 + 186690 nsp48610 rp!32 LSU ribosomal protein L32p

3168087 + 168603 nsp31110 rps14 SSU ribosomal protein S14p (S29e)

429606 + 161747 nsp4480 - unknown protein

2033905 + 140335 nsp20320 infC Translation initiation factor 3, TSS2

1997768 + 135635 nsp19930 sbtA putative sodium-dependent bicarbonate transporter

25246 + 108746 nsp280 hypothetical protein

125437 + 74766 nsp1390 - unknown protein

818072 + 69294 nsp8200 rps16 SSU ribosomal protein S16p

3595360 + 67550 nsp35290 psbA4 photosystem II protein D1 (PsbA)

2100481 - 66076 nsp20970 ^B hypothetical protein

490455 + 61106 nsp5100 psbV Photosystem II protein PsbV, cytochrome c550

1301909 - 58894 nsp 12740 glyA Serine hydroxymethyltransferase (EC 2.1.2.1)

3327951 + 54540 nsp32680 - DnaJ-class molecular chaperone

4211981 - 54170 nsp41850 hliE CAB/E UP/HUP superfamily

4066960 + 52955 nsp40310 ccmK Possible carbon dioxide concentrating mechanism protein CcmK

3743395 - 51014 nsp36850 ^B Branched-chain amino acid permeases, DUF4079

2477247 + 48670 nsp24980 cpcG3 Phycobilisome rod-core linker polypeptide, phycocyanin-associated

2126456 + 46218 nsp21270 acpP Acyl carrier protein

3427650 - 45498 nsp33690 - hypothetical protein

4477663 - 45304 nsp44370 purS Phosphoribosylformylglycinamidine synthase, PurS subunit (EC 6.3.5.3)

4875128 - 44715 nsp48350 prx5 Peroxiredoxin

3279390 + 43459 nsp32180 chIL Light-independent protochlorophyllide reductase iron-sulfur ATP-binding protein ChIL

308174 - 39899 nsp3210 - hypothetical protein

4424601 + 36350 nsp43850 ndhN Putative subunit N o f NAD(P)H:quinone oxidoreductase

3660323 - 35393 nsp35880 - Peptidase M23B precursor

2945315 - 34180 nsp28800 ^B FIG00871618: hypothetical protein

2963323 + 33480 nsp29010 trxA Thioredoxin

2609225 - 33225 nsp26260 rbcL Ribulose bisphosphate carboxylase large chain (EC 4.1.1.39)

948959 - 32821 nsp9420 rpl19 LSU ribosomal protein L19p

2961668 - 32694 nsp28990 ^B Transposase, OrfB family

1827328 + 32133 nsp18160 fbpl Fructose-1,6-bisphosphatase, GIpX type (EC 3.1.3.11)/Sedoheptulose-1,7- bisphosphatase (EC 3.1.3.37)

1563165 - 32001 nsp15230 rpaC putative regulator o f phycobilisome association C

4656026 - 31933 nsp46210 chIP Geranylgeranyl reductase (EC 1.3.1.83)

1811000 + 30834 nsp17890 rpoZ, ycf61 DNA-directed RNA polymerase omega subunit

4154749 - 30790 nsp41280 - hypothetical protein

2033866 + 29710 nsp20320 infC Translation initiation factor 3, TSS1

2898416 + 29643 nsp28400 - hypothetical protein

3831169 - 29380 nsp37800 ^B hypothetical protein

4611031 + 29275 nsp45760 rpl35 LSU ribosomal protein L35p

175701 + 29188 nsp 1870 ^B hypothetical protein

4907404 + 29084 nsp48690 - Protein CP12, regulation o f Calvin cycle

4319669 29012 nsp42500 hypothetical protein

3648306 - 28409

PLOS ONE I w w w .p loso n e .o rg

nsp35770 metX S-adenosylmethionine synthetase (EC 2.5.1.6)

6 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

Table 2. Cont.

Position S Reads ID Gene Annotation

472034 + 27553 nsp4910 ubiB Ubiquinone biosynthesis monooxygenase UbiB

3264182 - 25978 nsp31980 - hypothetical protein

For each gTSS, the position w ith respect to the forward strand, the orientation (S), the number o f reads, the gene ID, gene name (if known) and gene annotation is given. The gTSS are ordered according to the number o f reads. doi:10.1371 /journal.pone.0060224.t002

have six [19,42-44] and one TSS [45] and are arranged ta il-to- ta il. Th is genomic arrangement is also conserved in „/Vi spumigena CC Y 9414 but on ly three TSS were detected fo r g lnA (Table 4). The phycobilisome degradation p ro te in N b lA is another example from this set: i t has five TSS mapped by dRNAseq and confirm ed by p rim er extension in Anabaena PCC 7120 [19], but only two TSS were detected upstream o f the JV. spumigena C C Y 9414 hom olog (nsp44910), at positions -102 and -340. In contrast to these examples, genes not involved in n itrogen assim ilation exhibited a conserved prom oter architecture. For example the rbcLXS operon is transcribed from two TSS at positions -25 and -504 in Anabaena PCC 7120 [19,46—48] and this is also the case in JV. spumigena CCY9414, at positions -31 and -512.

Genome-Based Prediction o f Compatible Solute Accumulation Capabilities

Analysis o f salt-induced compatible solute accum ulation in approxim ately 200 d ifferent cyanobacterial strains proposed that freshwater and brackish water strains (low salt resistance) accumulate the disaccharides sucrose a n d /o r trehalose, w h ile true m arine strains (moderate salt tolerance) contain the heteroside glucosylglycerol (GG), and ha lophilic and hypersaline strains accumulate betaines, m ain ly glycine betaine (reviewed in [49]). Salt-loaded cells o f Nostocales accumulated on ly disaccharides, in agreement w ith the ir low salt tolerance. JV. spumigena C C Y9414 occurs in the Baltic Sea m ain ly at salinities ranging from 2 - 1 0 PSU (equivalent to 0.2 - 1% N a C l and 5 - 33% o f fu ll seawater salinity). Its genome was searched using sucrose-phosphate synthase (SpsA) and sucrose-phosphate phosphatase (Spp) se­quences from Synechocystis sp. PCC 6803 (sps—sll0045, spp-slr0953 [50]). The JV. spumigena C C Y 9414 O R F nsp8740 shows significant sim ilarities to SpsA, w h ich clusters w ith sim ilar enzymes from unice llu lar cyanobacteria, while putative Sps proteins from Nostocales were found in a separate clade (F ig . S I). The nsp8740 gene was found to be h igh ly expressed under our

standard cu ltiva tion conditions (10 PSU) expla in ing the observed sucrose accum ulation in these cells (F. M oke, unpublished). The Sps o f .Y spumigena C C Y 9414 could be a com bined enzyme w ith Sps as w ell as Spp activity, because bo th domains are present in the sequence o f O R F nsp8740. A n apparently truncated Sps pro te in is encoded by nsp23670, w ith closely related proteins in o ther Nostocales (F ig . S I). S im ilar to Synechocystis sp. PCC 6803 and Anabaena PCC 7120, the JV. spumigena C C Y 9414 genome also harbors a single spp gene (nsp21420). Anabaena PCC 7120 also possesses sucrose synthase (susA - a!14985, susB - a l l i059) [51]. S im ilar proteins, SusA (nsp24720) and susB (nsp48150) were also found. Hence, sucrose metabolism in JV. spumigena C C Y 9414 is sim ilar to Anabaena PCC 7120, where Sps is used to synthesize sucrose to serve as a compatible solute as well as serving as a source o f energy and reducing power for N 2-fixa tion in the heterocyst [52]: SusA and SusB are p robably involved in sucrose breakdown to provide electrons and energy fo r No-fixation [51,53],

Low salt to lerant cyanobacteria often also accumulate trehalose, w h ich is usually synthesized by the m altooligosyl trehalose synthases (M ts l and Mts2) using glycogen as precursor. This trehalose synthesis pathway often includes also TreS as enzyme capable o f hydro lyzing trehalose in to glucose (e.g. [54]). In Anabaena PCC 7120, an operon was identified comprising the treS, mtsl and mts'2 genes (all0166, all0167, all0168\ [55]). Proteins very sim ilar to M ts l and Alts2 are encoded in JV. spumigena C C Y 9414 by two genes like ly fo rm ing an operon (nsp4l87Q/nsp41880). These genes are linked to nsp41890 (the first gene in a putative operon w ith mtsl and mts2) that encodes for a glycogen de-branching enzyme, m aking the precursor glycogen fo r trehalose synthesis accessible. Finally, O R F nsp39450 is a good candidate encoding TreS fo r degradation o f trehalose. H ithe rto , there is no experim ental verification fo r trehalose accum ulation in JV. spumigena CCY9414, i.e. cells grown in liq u id m edia at different salinities accumulated only sucrose (F. M oke, unpublished). The absence o f trehalose correlates well w ith the absence o f an active

Table 3. Selected non-coding RNA elements mentioned in the text or ubiquitious among bacteria.

Gene Product From To s Com ment Reference

ssrA tmRNA 3524428 3524830 + [159,160]

ssaA 6S RNA 3750385 3750114 single TSS located in purK (nsp26940) gene, tw o TSS identified in other cyanobacteria

[161]

Ffs scRNA, 10S RNA 2912155 2912257 +

rnpB RNAse P RNA subunit 5025615 5025994 +

yfr2 Yfr2 460195 460090 - [162]

nsiR3 NsiR3 2888944 2888814 NtcA binding site conserved [19]

The respective gene name is given together w ith the sRNA product, the location, orientation (S; +, forward strand; -, reverse strand), comments and references. doi:10.1371 /journal, pone.0060224.t003

PLOS ONE I w w w .p loso n e .o rg 7 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

Table 4. Selected proteins of heterocyst differentiation, pattern formation and nitrogen assimilation in N. spumigena CCY9414.

Category/Protein N. spumigena CCY9414 Anabaena PCC 7120

NsORF TSS Identity e-value

Early events

NtcAs nsp2630 254230R 99% e-121

HanA, HupB - - -

HetR§ nsp16830 1713443R 91% e-162

HetF nsp22100 2206991F 61% 0

HetC

HetL

HetP (alr2818)§ nsp7850 - 69% 2e-046

NrrA§* nsp18040 1819958R 92% 1e-134

HetZ* nsp39970 4037960F 90% 1 e-173

Pattern form ation

PatS nsp47965 - 82% 0.011

PatA* nsp24860 - 61% 1 e-130

PatB nsp41440 81% 0

PatN*§ nsp 12530 1277587F, 1277705F 71% 5e-084

HetN - - - -

Nitrogen assimilation

GlnA§* nsp 16180 1656268F, 1656364F, 1656438F 88% 0

GifA nsp16190 1658460R 79% 3e-025

NblA§ nsp44910 4524908F 92% 1 e-029

The protein names are given, followed by the ORF ID in N. spumigena CCY9414 (NsORF), the position o f the TSS (F, forward or R, reverse strand), the% ID and e-value In a pairwise alignment w ith the orthologs from Anabaena PCC 7120. Only amino acid Identities >60% were considered; (-) not detected; *gene Is associated w ith an antisense RNA In In N. spumigena CCY9414; §gene Is associated w ith multiple TSS In Anabaena PCC 7120. dol:10.1371 /journal.pone.0060224.t004

prom oter fo r the m ts l/2 genes under the grow th conditions tested here. In this respect, it is interesting to note tha t salt-stressed cells a í Anabaena PCC 7120 also on ly accumulate sucrose ([52]; own observations), while the trehalose biosynthesis genes were induced upon desiccation in this organism [56,57].

Besides de novo synthesis, compatible solutes are often seques­tered via specific transporters. JV. spumigena C C Y 9414 contains m ultip le genes fo r such transporters. A n AB C -type transporter for glycine beta ine /cho line uptake [58,59] appears to be encoded by nsp43160 to nsp43200. A no ther gene cluster seems to encode a p ro line /g lyc ine betaine A B C transporter (nsp6940/nsp6950). In contrast, an AB C -type transporter fo r compatible solutes sucrose, trehalose, and G G , such as G g tA B C D from Synechocystis sp. PCC 6803 [49], was not found in the JV. spumigena C C Y 9414 genome. The presence o f m ultip le compatible solute uptake systems m ight be favorable in com plex m icrob ia l communities, in w hich dissolved compatible solutes such as pro line and glycine betaine released from other m icrobes can be qu ick ly taken up and used in add ition to the de novo biosynthesis o f sucrose.

Acclimation Strategies to Low Iron Levels: a M ultitude of Transport Systems

Iro n is one o f the m ain factors determ in ing cyanobacterial p roductiv ity in the m arine pelagic environm ent includ ing cyanobacterial blooms in the Baltic Sea [60], because most inorganic iron in the oxygenated biosphere was converted in to v irtua lly insoluble ferric iron. Acclim ation o f cyanobacteria to iron starvation includes the induction o f specific transport systems [61], Synechocystis sp. PCC 6803 possesses at least three AB C -type iron-

specific transporters, w h ich seem to be specialized fo r uptake o f Fe2+ (feoB, s lrl392, etc.), Fe3+ IfutA, s lrl295/s lr0513, etc.), and Fe3+- d ic itrate (fecB, s i l l202 or s ir i491, etc.). S im ila r gene clusters are present in the genome a í Anabaena PCC 7120, w h ich encodes m ultip le copies o f the fee operon [62]. These genes were used to search the genome o f JV. spumigena C C Y 9414 (Table 5). C orre­sponding to the ecological niche, the genome o f JV. spumigena C C Y 9414 lacks a Fe2+ uptake system o f the Feo-type, w h ich is consistent w ith the nearly complete absence o f Fe2+ in the oxygenated seawater environm ent o f JV. spumigena. However, as expected fo r an organism that is exposed to iron lim ita tion , at least three alternative iron uptake systems were found. One operon contains four genes sim ilar to the fu t operon (nspl9100-nspl9130), w h ich encode an AB C -type Fe3+ uptake system. A dd itiona lly , two systems for the uptake o f Fe3+ bound to organic chelators (siderophores), such as d ic itrate o r hydroxam ate exist in JV. spumigena CCY9414. One o f these transporters is sim ilar to the Fee system from Synechocystis sp. PCC 6803 or Anabaena PCC 7120 (fecBEDQ uspi 193 0-nspl 1960). I t is linked to a TonB-dependent ferrichrom e-like receptor (uspi 1910) used fo r the uptake o f chelated Fe3+ [62]. This p ro te in also shows sim ilarities to A lr0397, w h ich was characterized as the receptor fo r the siderophore schizokinen in Anabaena PCC 7120 [63]. The genes putative ly involved in schizokinen synthesis in Anabaena PCC 7120 [63] were not found in the genome o f JV. spumigena CCY9414. However, the JV. spumigena C C Y 9414 genome contains an JhuCDB operon (nsp27490-nsp27510), annotated as a Ferric-hydroxam ate A B C transporter. This implies that JV. spumigena C C Y 9414 is able to accept m any forms o f chelated Fe3+ inc lud ing those bound to

PLOS ONE I w w w .p loso n e .o rg M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

Table 5. Proteins related to the uptake of iron in N. spumigena CCY9414 identified on the basis of gene clusters present in the genome of Anabaena PCC 7120.

ORF NsORF Annotation % ID Reference

Operon 1

AII2618 nsp 11930, nsp2720 Iron(lll) dicitrate transport system, periplasmatic Iron(lll) dicitrate transport system, periplasmatic

36, 32

AII2619 Ferrobactin Receptor nsp11910, nsp2710 TonB-dependent receptor; Outer membrane Ferrichrome receptor

50, 25

AII2620, Ferrobactin Receptor nsp 11910, nsp2710, nsp26750 TonB-dependent receptor; Outer membrane Ferrichrome-iron receptor, Ferrichrome-iron receptor

47, 38, 29

Operon II

Alr2209 Aerobactin receptor nsp 11910, nsp2710, nsp26750 TonB-dependent receptor; Outer membrane Ferrichrome-iron receptor, Ferrichrome-iron receptor

46, 27, 25

Air2210 nsp2720, nsp 11930 iron(lll) dicitrate-binding periplasmic protein 38, 38

Alr2211 ferrichrome-iron receptor nsp2710, nsp26750 Ferrichrome-iron receptor 45, 38

Alr2212 nsp 11930, nsp2720, nsp27500 Iron(lll) dicitrate transport system, periplasmic proteins, Ferric hydroxamate ABC transporter

37, 32, 37

Alr2213 nsp 11930 nsp2720 Iron(lll) dicitrate transport system, periplasmic proteins

Operon III

Alr0397, schizokinen receptor nspl 1910, nsp2710, nsp26750 TonB-dependent receptor; Outer membrane, Ferrichrome-iron receptor, Ferrichrome-iron receptor

47, 37, 24 [164]

AII0396, iaminobutyrate-pyruvate transaminase

nsp46650 Acetylornithine aminotransferase 29 [164]

AII0395, L-2,4-diaminobutyrate decarboxylase

nsp37340 Cysteine desulfurase 26 [164]

Operon IV

AII0390, rhbF nothing [62]

AII0389, fhuC nsp27490, nsp 11960 Ferric hydroxamate ABC transporter, ABC-type Fe3+-siderophore transport system, ATP-binding

89, 50 [62]

AII0388, fhuD nsp27500 nsp11930 Ferric hydroxamate ABC transporter Iron(lll) dicitrate transport system, periplasmic binding

83,30 [62]

AII0387, fhuB nsp27510, nsp 11950, nsp 11940 Ferric hydroxamate ABC transporter, ABC-type Fe3+-siderophore transport system, permeases, Nsp11950 and 11940 probably 1 ORF

85, 37, 48 [62]

Operon V

Alr2581 aerobactin receptor nspl 1910, nsp2710, nsp26750 TonB-dependent receptor; Outer membrane Ferrichrome-iron receptor, Ferrichrome-iron receptor

78, 27, 24

Alr2582; Hyp. Prot. nsp28410 hypothetical protein 47

Alr2583, fecBI nspl 1930, nsp2720 Iron(lll) dicitrate transport system, periplasmic binding protein

80, 36 [62]

AII2584, fecEl nsp 11960, nsp27490 ABC-type Fe3+-siderophore transport system, ATP-binding Ferric hydroxamate ABC transporter

71, 52 [62]

AÍÍ2585, fecDI nsp 11950, nsp 11940, nsp27510 ABC-type Fe3+-siderophore transport system, permeases, Ferric hydroxamate ABC transporter

67, 36, 43 [62]

AII2586, fecC! nsp 11950, nsp 11940, nsp27510 ABC-type Fe3+-siderophore transport system, permeases, Ferric hydroxamate ABC transporter

61, 37, 34 [62]

Operon VI

Alr2587, Transcription factor nsp2700 Transcriptional regulator, AraC family 38

Alr2588, fer rich rom e-Fe-receptor nsp2710, nsp26750, nsp 11910 Ferrichrome-iron receptor, Ferrichrome-iron receptor, TonB-dependent receptor

50, 40, 41

Air2589, Hyp. Prot. nsp28410 hypothetical protein 47

Alr2590, iron(lll) dicitrate-binding protein

nsp 11930 nsp2720, nsp27500 Iron(lll) dicitrate transport system, periplasmic binding proteins, Ferric hydroxamate ABC transporter

28, 27, 37

Alr2591, Transcription factor nsp2700 Transcriptional regulator, AraC family 38

Alr2592, fer rich rom e-Fe receptor nsp2710, nsp26750, nsp 11910 Ferrichrome-iron receptor, Ferrichrome-iron receptor, TonB-dependent receptor

46, 38,26

Alr2593 i ron (111) dicitrate-binding nspl 1930,nsp2720 Iron(lll) dicitrate transport system, binding periplasmic proteins

38, 35

PLOS ONE I w w w .p loso n e .o rg 9 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

Table 5. Cont.

ORF NsORF Annotation % ID Reference

Alr2594, hypothetical protein

Alr2595, Transcription Factor nsp2700 Transcriptional regulator, AraC family 40

Alr2596, fer rich rom e-Fe receptor nsp2710, nsp26750, nsp 11910 Ferrichrome-iron receptor, Ferrichrome-iron receptor, TonB- dependent receptor

49, 38, 27

Air2597, i ron (111) dicitrate-binding nspl 1930, nsp2720 Iron(lll) dicitrate transport system, periplasmic binding proteins 29, 26

Operon VII

Alr3240, FecD2 nsp 11950, nspl 1940 ABC-type Fe3+-siderophore transport system, permease proteins 42, 37 [62]

Alr3241, FecE2 nsp 11960, nsp27490 ABC-type Fe3+-siderophore transport system, ATP-binding, Ferric 46, 44 hydroxamate ABC transporter

[62]

Alr3242, hutA2 nspl 1910, nsp2710 TonB-dependent receptor, Ferrichrome-iron receptor 25, 22 [62]

Alr3243, fecB2 nsp11930 Iron(lll) dicitrate transport system, periplasmic binding protein 24 [62]

Operon VIII

Air4030, Hypot. Prot. nspl 4910 hypothetical protein 38 [62]

Alr4031, fecB3 nsp11930 Iron(lll) dicitrate transport system, periplasmic binding protein 23 [62]

Alr4032, fecD3 nsp 11950, nsp 11940, nsp27510 ABC-type Fe3+-siderophore transport system, permease proteins, 32, 28, 31 Ferric hydroxamate ABC transport

[62]

Alr4033fecE3

nsp 11960 ABC-type Fe3 +-siderophore transport system, ATP-binding 33 [62]

Operon IX

Air 1381, prcA nsp18110 Calcium-dependent protease precursor 32 [62]

Air 1382, futA nsp19110, nsp19100 Ferric iron-binding periplasmic proteins o f ABC transporter 57, 49 [62]

Air 1383, futB nspl 9120 Ferric iron ABC transporter, permease protein 64 [62]

Air 1384, futC nspl 9130 lron(lll)-transport ATP-binding protein 60 [62]

The first row contains the ORF ID, annotation and gene name (if available) o f the respective Anabaena protein (according to the published sequence [163] In Genbank file NC_003272.1), followed by the ORF ID In N. spumigena CCY9414 (NsORF), the detailed annotation, the% ID In a pairwise alignment and the reference, If available. dol:10.1371 /journal.pone.0060224.t005

siderophores produced by other bacteria present in the brackish water com m unity.

Acclimation Strategies to Low Iron Levels: a M ultitude ofpsbC/isiA/pcb Genes

O ne gene that becomes strongly expressed under iro n -lim itin g conditions in m any cyanobacteria is isiA, coding fo r the iron stress induced pro te in A [64 -66 ]. A dd itiona lly , Is iA participates in high ligh t acclim ation [67]. Is iA belongs, together w ith the CP43 (PsbC) and the Pcb’s from Prochloron, Prochlorothrix, Prochlorococcus and Acaryochloris, to a fam ily o f related antenna proteins that b ind chlorophylls. In JV. spumigena CCY9414, psbC (nsp52950) is located at one genomic location as part o f a psbDC d icistronic operon, w h ich is the typical gene organization o f these photosystem I I core antenna genes among cyanobacteria. However, four additional genes o f the isiA/psbC/pcb fam ily (nsp37450, nsp37460, nsp37500, nsp37510) are clustered w ith a flavodoxin gene (nsp37490, isiB) at another site in the genome (F ig . 3A). In between the flavodoxin and is iA genes a p ro te in o f unknow n function w ith an a lpha/beta hydrolase dom ain is encoded (nsp37480), homologs o f w h ich are associated w ith flavodoxin genes also in most other N 2-fix ing cyanobacteria. A sim ilar situation w ith several tigh tly clustered genes o f the Is iA /C P 4 3 fam ily exists in Anabaena PCC 7120 and other filamentous, N 2 -fix ing cyanobacteria such as Fischerella muscicola PCC 73103 [68], A phylogenetic analysis o f these proteins shows that one o f the proteins from this fam ily (nsp37460) clusters w ith several w ell characterized Is iA proteins and hence is a distinct Is iA homolog. In contrast, the other fou r proteins belong to a tigh t cluster also contain ing PsbC (F ig . 3B).

One o f the PsbC homologs (JVsp37500) possesses a considerable Cl te rm ina l extension (total length 477 am ino acids compared to 319-344 residues fo r the other PsbC homologs). A closer inspection revealed that JVsp37500 possesses a PsaL dom ain in this add itiona l segment and that nine transmembrane regions are predicted fo r the PsbC-PsaL hyb rid p ro te in (SI, F ig . S2). S im ilar genes have recently been identified in several more cyanobacterial genomes and the PsbC-PsaL hyb rid proteins have been classified as ch lorophyll b ind ing proteins type V (CBPV) [69]. Analysis o f a PsaL-less m utant o f Synechocystis sp. PCC 6803 indicated that PsaL is required fo r the fo rm ation PSI trimers. However, iron- starved cells o f this m utant were still able to fo rm Is iA rings around PSI monomers bu t to a lesser extent [66,70], The PsbC-PsaL fusion present in JVsp37500 suggests that this strain is hard-w ired fo r the add ition o f ch lorophyll-antenna to PSI monomers over and above the IsiA-rings associated w ith PSI trimers. This possibility is supported by the results o f a recent hom ology m odelling and insertion o f the PsaL-like dom ain in to the PSI structure [69]. Such an antenna com plex may be a particu la rly efficient fo rm o f ligh t- harvesting by PSI in the ecological niche o f JV. spumigena. The regulation o f these genes in JV. spumigena is not known, bu t at least fo r F. muscicola PCC 73103 the iron-stress-regulation o f a comparable large operon w ith Pcb/PsbC homologs was detected

[6 8 ]-The transcriptome data provides an in itia l snapshot on the

expression o f the different members o f the psbC!¿¿4-like gene fam ily in JV. spumigena CCY9414. W hile the classical psbDC operon is strongly expressed, we detected only a rather weak TSS associated w ith the genes nsp37450, nsp37460, nsp3 7500 and nsp37510, w h ich is

PLOS ONE I w w w .p loso n e .o rg 10 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ra n scrip to m e o f a Toxic C yanobacte rium

nsp37450 nsp37460 nsp37480nsp37490nsp37500nsp37510 \ nsp52940 nsp52950

psbC-lk1 isiA a/ß isiB PsaL PsbC psbC-lk3 hydrolase domain

psbC-lk2

psbD psbC

B Synechocystis 6803 SII0851 Crocosphaera WH8501 EAM47669|

Lyngbya AE26130 PsbC (CP43)Trichodesmium Tery0513

Anabaena Alr4291 Nodularia NSP52950 Trichodesmium Tery2484

Anabaena 7120 AII4003 Nodularia NSP37510

PsbC-lk3

Crocosphaera WH0003 EAM47499 Trichodesmium Tery2483 Trichodesmium Tery2485

Anabaena AII4002 Nodularia NSP37500

Fischerella JSC11 EHC15956 Anabaena 7120 AII4000

Nodularia NSP37450

PsbC_PsaLhybrid

PsbC-lk1

Anabaena AII4001 Nodularia NSP37460

Crocosphaera WH 0003 EHJ09590 Synechocystis 6803 SII0247 IsiA (CP43*) Trichodesmium Teryl 667- Anabaena 7120 CP47 PsbB Outgroup

1620 reads

i l m i i i i i i i i i i i m i i m i i 11 i

.......................................................................................................... ncf3802558 ......................... .......

I I I I II I I I I I I I I ’SP37S1° I III I I I I I I I I I I I= l I I II II II I I I II I I I

nsp37510 (psbC-lk3) ................................................................................ ' "

Figure 3. Analysis o f loci encod ing pro te in s o f th e C P 43/ls iA /P cb fa m ily . A. O rgan iza tion o f th e ch rom o so m a l re g io n h a rb o ring th e isiA and psbC-like genes (psbC -ik l-3 ) o f N. spum igena and th e separa te psbDC o pe ro n . The P saL-cod ing d o m a in in psbC-lk2 (nsp37500) is h ig h lig h te d in o range . B. P hy lo g e ne tic analysis o f CP43, IsiA and re la ted c h lo ro p h y ll-b in d in g p ro te in s fro m N. spum igena and o f se lected o th e r cyanobacte ria was In fe rred us ing th e M in im u m E vo lu tion m e th o d . The o p tim a l tre e w ith th e sum o f b ranch len g th = 3 ,97009738 is show n . The p e rce n ta g e o f rep lica te trees in w h ich th e associa ted taxa c lus tered In th e b o o ts tra p te s t (1000 rep licates) are show n n e x t to th e branches. The tre e is d raw n to scale, w ith branch len g th s in th e sam e u n its as th o se o f th e e v o lu tio n a ry d is tances used to in fe r th e p h y lo g e n e tic tree. The e v o lu tio n a ry d is tances w ere c o m p u te d using th e Poisson co rre c tio n m e th o d and are in th e u n its o f th e n u m b e r o f a m in o acid s u b s titu tio n s pe r site. A ll p o s itio n s co n ta in in g gaps and m iss ing da ta w ere e lim in a te d fro m th e d a tase t (co m p le te d e le tio n o p tio n ). There w ere a to ta l o f 279 p os itio n s in th e fin a l dataset. C. Transcrip tiona l o rga n iza tio n a ro u n d th e isiA, isiB and psbC-like g ene c luster. There are th re e m a pp e d TSS In th e re g io n d isp layed in Fig. 3A, all associa ted w ith o r c lose to th e 5 ' end o f nsp37510. TSS are ind ica te d by b lu e a rrow s and th e n u m b e r o f cD N A reads associa ted w ith th e m are g ive n as a p p ro x im a tio n fo r th e ir a c tiv ity . O ne gTSS g ives rise to th e 83 n t lon g 5 ' UTR upstream o f nsp37510 (b lue) and th e g ene o r o p e ro n mRNA. An antisense RNA o rig ina tes fro m a s ing le aTSS in th e o p p o s ite d ire c tio n (purp le ). The th ird TSS Is a p u ta tiv e nTSS d riv in g th e tra n s c rip tio n o f an ncRNA in th e nsp37510- nsp37520 in te rg e n ic spacer. Except fo r th e nsp37510 5 ' UTR, all TSS d isp layed are d raw n w ith a 100 n t- lo n g b o x th a t co rresp o n de d to th e m a x im u m read len g th in th e dRNAseq approach. d o i:1 0.1371 /jou rna l.pone .0060224 .g003

PLOS ONE I w w w .p loso n e .o rg 11 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

located 83 n t upstream o f mp3 7510 and could indicate the presence o f a long operon consisting o f these genes (Fig. 3C). However, its activ ity m ight be decreased by the activity o f an aTSS at position +426 (Fig. 3C) under some conditions. I f so, this antisense R N A could have an analogous contro l function as the Is rR antisense R N A to the isiA gene in Synechocystis sp. 6803 [20].

Dinitrogen Fixation: Nitrogenase and Hydrogenases in N. spumigena CCY9414

JV. spumigena CC Y 9414 has one complete set o f n if genes coding fo r a Mo-nitrogenase 1 and add itiona l N 2 fixa tion genes w ith in a region o f 26,173 base pairs (genes nsp40650-nsp40900), transcribed from a single b u t strong TSS 296 n t upstream o f nsp40650 (nifB). A nifHDK gene cluster is present in this region, inc lud ing a split nifH (nsp40720) and a split nifD (nsp40770) gene. A second copy o f nifH (nifH‘2; nsp34540), coding fo r dinitrogenase reductase, is present at an unrelated site in the genome. In JV. spumigena strain A V I expression o f nijH2 seems to be under n itrogen con tro l [71].

JV. spumigena C C Y 9414 encodes two [N iFe] hydrogenases as is the case in all o ther N 2-fix ing cyanobacteria investigated to date. The genes fo r catalytic subunits o f uptake hydrogenase, hupS (nsp41100) and hupL (nsp41090 and nsp41000, w h ich become fused fo llow ing heterocyst-specific recom bination), are separated by an intergenic stretch that m ight fo rm a ha irp in as has been described fo r o ther cyanobacteria [72]. The genome also contains hoxEFUYH (genes nsp28020-nsp28070), encoding the structural proteins o f the b id irectiona l hydrogenase, an enzyme com m on in m any diazotrophic and non-d iazotrophic cyanobacteria. Both uptake and b id irectiona l hydrogenase gene clusters possess genes fo r a putative endoprotease, H u p W (nsp40980) and H oxW (nsp28070), processing the large subunits HupL (genes nsp41090 and nsp41000) and H oxH (nsp28060), respectively. These genes are located downstream from hupL and hoxH, and in the case o f the uptake hydrogenase separated from hupL by a small O R F. In JV. spumigena C C Y 9414 the hox genes fo rm a contiguous cluster hoxEFUYHW (nsp28020-nsp28070) w ith o u t add itiona l O R F ’s. The hyp genes fo r m atura tion proteins are present as single copy genes in the genome o f A i spumigena CCY9414.

Dinitrogen Fixation: Regulation o f Fleterocyst Differentiation

JV. spumigena forms regularly spaced heterocysts along the filaments as other Nostocales. The structural genes fo r din itogen fixa tion and heterocyst fo rm ation are closely related to those from Anabaena PCC 7120 (see Table 5 fo r overview). Regulatory proteins, such as the transcrip tion factor N tcA (nsp2630), w hich senses the in trace llu lar accumulation o f 2-oxoglutarate as an ind ica to r o f n itrogen lim ita tio n [73] and then triggers the d iffe rentia tion process towards heterocysts via HetR [nspl 6830, [74]), are present in the JV. spumigena CC Y 9414 genome (nsp2630). In m any heterocystous cyanobacteria, such as Anabaena PCC 7120, hetR is expressed in a spatial pattern along the trichomes [75-78 ], triggered from a heterocyst-specific TSS. M any fu rther well- characterized genes encoding pro te in factors involved in hetero­cyst fo rm ation (reviewed in [3]), such as N rrA , PatN and the signalling peptide PatS (sequence here:M K T T M L V N F L D E R G S G R the m in im um pentapeptide re­qu ired fo r norm al heterocyst pattern fo rm ation underlined), H etF and H etP and hetP-Yke genes (2 copies) are also present in the JV. spumigena CC Y 9414 genome. H e tF is required for heterocyst fo rm ation and fo r the no rm a lly spaced expression o f hetR in JV. punctiforme [79] and Anabaena PCC 7120 [80].

In add ition to the know n regulatory proteins, associated regulatory R N A s fo r heterocyst d ifferentia tion are w ell conserved in JV. spumigena C C Y 9414 as they are in other Nostocales. A tandem array o f 12 short repeats was found upstream o f hetF (:nsp22100) [81]. A homologous tandem array in Anabaena PCC 7120 gives rise to the N s iR l sR N A that like ly plays a role in the regulatory cascade leading to heterocyst d ifferentia tion [19,81]. H e tZ is a pro te in involved in Anabaena PCC 7120 in early heterocyst d iffe rentia tion [82]. Recendy, contro l o f its transcrip­tion by a 40 n t HetR b ind ing site upstream o f a TSS was suggested, w h ich is located at position -425, antisense to gene asl0097 [83]. This arrangement is almost exactly conserved in JV. spumigena CCY9414: the only TSS upstream o f the hetjf hom olog nsp39970, was mapped at position -429 and in antisense orienta tion to nsp39970, the hom olog o f asl0097. M oreover, the sequence 5 '- A T T T G A G G G T C A A G C C C A G C A G G T G A A C T T A G G G A G A C A T -3 ', located 56-17 n t upstream o f this TSS is almost identical to the reported HetR b ind ing site in Anabaena PCC 7120 [83]. These facts, together w ith the conserved arrangement, inc lud ing a long 5 '-U T R o f hetif and aTSS located w ith in the gene upstream, suggest that the HetR b ind ing site is also functional in JV. spumigena CCY9414.

Genes fo r PatA and PatB, w h ich p lay essential roles in con tro lling the spacing o f heterocysts along a filam ent [84-86] are also present in the JV. spumigena C C Y 9414 genome. However, other proteins involved in heterocyst fo rm ation were not found. A m ong these are hetJV.] in Anabaena PCC 7120 involved in pa ttern ing o f heterocysts along the filaments, and hanA (hupB), encoding the histone-like H U p ro te in [87], w h ich is essential fo r heterocyst d ifferentia tion in Anabaena PCC 7120 [88]. Likewise, the hetC gene, proposed to be expressed in pro-heterocysts and to stimulate f t s i f expression [89,90], and hetL, w h ich simulates heterocyst development even in the presence o f com bined nitrogen [91], were not found. The lack o f genes fo r some o f the proteins involved in early events o f heterocyst fo rm ation indicates tha t JV. spumigena C C Y 9414 uses a mechanism fo r regulating early heterocyst d ifferentia tion d ifferent from that in Anabaena PCC 7120. These findings correspond w ith the less stringent regulation o f heterocyst fo rm ation by the n itrogen supply as reported fo r JV. spumigena A V 1 [5].

Dinitrogen fixation: DNA Rearrangements Involved in Fleterocyst Differentiation

D N A rearrangements as pa rt o f heterocyst developmental processes are know n from the heterocystous cyanobacteria Anabaena and Nostoc [92,93]. A D N A element, in te rrup ting a gene in the vegetative cell, is excised leading to recom bination and transcrip tion o f the genes in the heterocyst in order to perfo rm the function tha t is heterocyst specific. In Anabaena PCC 7120 three D N A elements have been identified and named after the genes they in te rrup t: nifD, fd x N and hupL element. The genome o f A i spumigena C C Y 9414 also is like ly to undergo three D N A rearrangements; i t contains a nifD and a hupL, bu t instead o f a

fd x N it has a nifH l element. However, the size o f the nifD and hupL elements is smaller than in most other Nostocales. The nifD element o f A i spumigena C C Y 9414 differs from those o f other cyanobacteria also in the num ber o f ORFs. In add ition to xisA, w hich encodes the site-specific recombinase, on ly a single other O R F (nsp40780) fo r a hypothetical p ro te in was identified on this element in A i spumigena CCY9414. The hupL element o f A i spumigena C C Y 9414 is 7.6 kb and also smaller than the 10.5 kb element a í Anabaena PCC 7120. Five out o f 7 ORFs found on the hupL element in A i spumigena CCY9414, inc lud ing the recombinase gene xisC, have sequence identities o f 86-97% at the D N A level to

PLOS ONE I w w w .p loso n e .o rg 12 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

Table 6. Complement of P- and arsenate-related gene orthologs in N. spumigena CCY9914.

NsORF, gene name Annotation com ment Reference

Inorganic P transport

nsp 1550 low affinity P permease

nsp 16870 low affinity P permease

nsp 15300, sphX freshwater sphX (P binding protein) similar to Synechocystis PCC6803 sll0540

nsp28900, pstS1 periplasmic P binding protein PstS similar to Synechocystis PCC6803 s ir i247 [100,106]

nsp28910, pstC1 PstC component o f high affinity ABC P transporter [100,106]

nsp28920, pstA1 PstA component o f high affinity ABC P transporter [100,106]

nsp28930, pstB1 PstB component o f high affin ity ABC P transporter ATP-binding protein component

[100,106]

nsp28940 pstB2 PstB component o f high affin ity ABC P transporter ATP-binding protein component

[100,106]

nsp52600, pstS2 periplasmic P binding protein PstS similar to Synechocystis PCC6803 SÜ0680 [100,106]

nsp52610, pstC2 PstC component o f high affinity ABC P transporter [100,106]

nsp52620, pstA2 PstA component o f high affinity ABC P transporter [100,106]

nsp52630, pstB2 PstB component o f high affin ity ABC P transporter ATP-binding protein component

[100,106]

Phosphonate transport

nsp7590, phnF PhnF component o f a C-P lyase

nsp7580, phnG PhnG component o f a C-P lyase

nsp7570, phnH2 PhnH component o f a C-P lyase

nsp7560, phnl Phnl component o f a C-P lyase

nsp7540, phnJ PhnJ component o f a C-P lyase

nsp7530, phnK PhnK component o f a C-P lyase

nsp7520, phnL PhnL component o f a C-P lyase

nsp7510, phnM PhnM component o f a C-P lyase

nsp7490 hypothetical protein in phn cluster

nsp7500 hypothetical protein in phn cluster

nsp7480, phnD2 PhnD component o f phosphonate ABC transporter phosphate-binding periplasmic component

nsp7470, phnC2 PhnC Phosphonate ABC transporter ATP-binding protein

nsp7460, phnE2 PhnE Phosphonate ABC transporter permease protein

nsp7450, phnE3 PhnE3 Phosphonate ABC transporter permease protein

nsp35120, phnC1 PhnO Phosphonate ABC transporter ATP-binding protein

nsp35130, phnD1 PhnD1 Phosphonate ABC transporter phosphate-binding periplasmic component

nsp35140, phnE1 PhnE1 Phosphonate ABC transporter permease protein

nsp35150, phnH1 PhnH truncated version, translationally coupled to nsp35160 - PhnM component o f a C-P lyase

nsp 18360, phnD2 PhnD2 Phosphonate ABC transporter phosphate-binding periplasmic component

nsp 18370, phnC2 PhnC2 Phosphonate ABC transporter ATP-binding protein

nsp 18380, phnE4 PhnE4 Phosphonate ABC transporter permease protein

Phosphite transport

nsp35050, ptxA PtxA Phosphite ABC transporter permease protein

nsp35060, ptxB PtxB Phosphite ABC transporter phosphate-binding periplasmic component

nsp35070, ptxC PtxC Phosphite Phosphite ABC transporter permease protein

nsp35080 phosphite dehydrogenase

nsp35090 LysR transcriptional regulator consistent w ith the operon structure o f the characterised[118] Pseudomonas stutzeri phosphite transporter

glycerol-3-phosphate transport

nsp7940, ugpC glycerol-3-phosphate ATP-binding protein component no other components o f the ugp operon appear present

PLOS ONE I w w w .p loso n e .o rg 13 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

Table 6. Cont.

NsORF, gene name Annotation com ment Reference

P stress inducible

nsp8220, phoH PhoH family protein

P storage and degradation of P polymers

nsp 10230, ppk polyphosphate kinase

nsp29750, ppa inorganic pyrophosphatase

nsp42550, ppx exopolyphosphatase

Degradation of organic P sources

nsp6490 glycerophophoryl diester phosphodiesterase contains also a potential phytase domain

nsp7010 atypical alkaline phosphatase akin to those present in several cyanobacteria

nsp12860 DedA-like phosphatase

nsp12920 alkaline phosphatase

nsp 12940 PhoX-like phosphatase

nsp18960 putative PhoX phosphatase

nsp20770 COG4246 superfamily sometimes annotated as a phytase

nsp29340 Metallophophoesterase

nsp29350 Metallophosphoesterase

nsp33000 predicted phosphatase

nsp31680 metal dependent PHP family phosphoesterase

nsp35720 acid phosphatase

nsp46480 metallophosphoesterase (GIpQ-like)

nsp53310 PhoD-like phosphatase

Arsenate-related gene orthologs/operons

nsp40 ArsA

nsp 1880 ArsA

nsp15480 ArsA

nsp33490, arsR regulator o f arsenate resistance

nsp33500 SphX periplasmic P binding component o f P ABC transporter

nsp33510 glycera Id ehyde-3-phosp hate dehydrogenase

nsp33520 major facilitator superfamily permease

nsp33540, acr3 Acr3 (ArsB) Arsenical-resistanee protein ACR3

nsp33550, arsH ArsH Arsenic-res¡stance protein

nsp41360 ArsC-family protein ArsC similarity not obvious

Haloacid dehalogenase-like hydrolases

nsp1610 HAD-superfamily hydrolase

nsp3740 HAD-superfamily hydrolase

nsp6980 HAD-superfamily hydrolase

nsp48160 glycoside hydrolase/HAD-superfamily hydrolase

P sensing and regulation

nsp 10800, phoB PhoB (SphR) Response regulator

nsp 10810, phoR PhoR (SphS) sensor kinase

nsp 10830, phoU PhoU putative negative regulator o f the Pi regulon

doi:10.1371 /journal.pone.0060224.t006

6 out o f 10 ORFs present on the element o f Anabaena PCC 7120. The two ORFs on the JV. spumigena C CY94T4 hupL element that do not have homologs on the Anabaena PCC 7120 element, are sim ilar to a DNA-cytosine methyltransferase and a F IN Fl-type endonu­clease (nsp41020 and nsp41030) and appear to be transcribed from a specific TSS 28 n t upstream o f nsp41020.

The d irectly repeated sequences flanking the nifD element d iffe r in JV. spumigena C C Y 9414 by one nucleotide from each other. The

repeat flanking the 5 ' part o f nifD is identical to the 11 bp sequence o f o ther strains (G G A T T A C T C C G ), w h ile the repeat flanking the 3 ' part o f nifD, close to xisA, differs by one nucleotide (G G A A T A C T C C G ). A sim ilar difference was observed in the element o f Anabaena sp. A TC C 33047 [94], bu t the d iffe ring nucleotides are not the same. Also the repeated sequences o f the hupL element d iffe r in JV. spumigena C C Y 9414 by a single nucleotide. The repeat at the 5 ' pa rt o f hupL is identica l to the

PLOS ONE I w w w .p loso n e .o rg 14 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

16 bp repeat from Anabaena PCC 7120 (C A C A G C A G T T A - T A T G G ) w hile the repeat close to xisC at the 3 ' pa rt o f hupL is d ifferent (C A T A G C A G T T A T A T G G ). The direct repeated se­quences from bo th nifD and hupL elements are present only once on the genome o f JV. spumigena CCY9414, thus, i t appears that these excisions are very specific.

In contrast to Anabaena PCC 7120 no rearrangement in fd x N seems to take place in JV. spumigena CCY9414. Instead a th ird rearrangem ent exists in nifH l in the nifH D Kcluster. The activ ity o f this previously unknow n D N A rearrangement mechanism was recendy demonstrated [71]. The nipHl element is 5.2 kb and encodes a X isA /X isC -typ e site-specific recombinase (nsp40750). In add ition to this recombinase only two fu rther ORFs are located on the nifH l element, coding fo r a hypothetical p ro te in and a putative D N A m odifica tion methylase. The recombinase encoded by nsp40750, w hich we term X isG , is 48% identical to X isA and 4% identica l to X isC o f JV. spumigena CCY9414. A ll three recombinases contain the h igh ly conserved tetrad R -H -R -Y o f the phage integrase fam ily w ith the catalytically active residue tyrosine, b u t as was described fo r X isA and X isC o f Anabaena PCC 7120 [95], the histidine is substituted by a tyrosine in JV. spumigena C C Y 9414 in X isA and X isC and the newly described X isG . The identical d irect repeats flanking the nifH l element are on ly 8 bp long (C C G TG A A G ). These repeats are overrepresented w ith 111 occurrences in the genome. Therefore, how the correct d irect repeats fo r recom bination are chosen by the recombinase is an open question. Interestingly, o ther strains o f JV. spumigena know n to develop heterocysts in the presence o f combined n itrogen [5,96], also have the nifH l element [71] found in JV. spumigena CCY9414.

Phosphate Acquisition: a Multitude o f Phosphatases and Transport Systems

The im portance o f phosphorus as a key lim itin g nu trien t in aquatic systems (see [97,98]) awoke m uch interest in defin ing P- scavenging mechanisms in cyanobacteria, pa rticu la rly at the genetic level (e.g. see [99-102 ]. This is especially relevant here since b io logica lly available dissolved inorganic and organic phosphorus forms appear critica l fo r JV. spumigena bloom form ation in the Baltic Sea [103,104], M oreover, expression o f the nodularin synthetase gene cluster increases during P-depletion [105]. Based on existing in fo rm ation , searches o f the JV. spumigena CC Y9414 genome fo r components o f inorganic phosphate transport and assim ilation were conducted (Table 6).

JV. spumigena possesses extensive P acquisition m achinery and strong TSS were mapped for most o f the genes involved. JV. spumigena C C Y9414 contains two copies o f a gene encoding a low a ffin ity permease fo r inorganic phosphate (Pi) transport akin to the E. coli P itA system (nspl550 and nspl 6870) unlike most m arine picocyanobacteria w h ich lack this capacity fo r P acquisition [102]. In addition, as is the case w ith several freshwater cyanobacteria [100,106], the genome o f JV. spumigena C C Y 9414 contains two gene clusters encoding components o f the h igh a ffin ity Pi transport system. This transport system is comprised o f components o f the membrane bound A B C transport system (PstABC) and the periplasm ic b ind ing p ro te in (PstS) (Table 6). These two high a ffin ity systems appear genetically sim ilar to those characterized biochem ically in the freshwater cyanobacterium Synechocystis sp. PCC 6803 and may equate to P; A B C transporters w ith significant differences in bo th k inetic and regulatory properties [106]. Together, these low and h igh a ffin ity P; acquisition systems m ight a llow JV. spumigena to acquire inorganic phosphate over a wide range o f concentrations. O the r potentia l h igh a ffin ity periplasm ic P i b ind ing proteins are also encoded in the JV. spumigena C C Y9414 genome sim ilar to sll0540 (nspl.5300) and sll0679 (nsp33500) from

Synechocystis sp. PCC 6803. The la tte r encodes a varian t o f the PstS b ind ing p ro te in term ed S phX [107,108], w h ich appears to be regulated d iffe rently from the other ‘classic’ PstS proteins at least in Synechocystis [106].

In JV. spumigena CCY9414, nsp33500 is located in a cluster o f genes, nsp33490-nsp33550 (Table 6) tha t includes one gene encoding glyceraldehyde-3-phosphate dehydrogenase, bu t also several others that are all involved in resistance to arsenic acid. Arsenate (As[V]), a toxic P; analog, has a nutrien t-like depth pro file in seawater [109] and competes w ith P; fo r uptake through the PstSCAB system. The gene nsp33490 encodes a potentia l A rsR regulator o f arsenate resistance and nsp33540 (A C R 3/A rsB ) encodes a putative arsenite efflux system (nsp33550 encodes a putative A rsH b u t the function o f this p ro te in is unknown). JV. spumigena also encodes three separate copies o f genes (nsp40, nsp l880 and nspl5480) potentia lly encoding ArsA, an arsenite- stimulated ATPase thought to allow more efficient arsenite efflux th rough ArsB [110,111]. However, ArsC encoding arsenate reductase appears to be lacking in the JV. spumigena CCY9414 genome, although an A rsC -fam ily pro te in is present (nsp41360) w hich may fu lfill the role o f arsenate reduction.

In add ition to transport systems fo r P; (i.e. phosphorus in its most oxidized form , +5 valence), JV. spumigena CC Y 9414 also contains transport systems fo r phosphonates and phosphite (i.e. +3 valence phosphorus compounds) (Table 6). T ransport capacity for these phosphorus sources has on ly been found in the genomes o f some cyanobacteria [99,100,112], hence, the presence o f trans­porters fo r phosphonates and phosphite in JV. spumigena is in trigu ing. Phosphonates, organic phosphorus compounds con­ta in ing a C-P linkage, require a specific C-P lyase enzyme to break this stable bond. In Pseudomonas stutzeri and E. coli, phosphonate u tiliza tion is mediated by a cluster o f 14 genes phnC to phnP) encoding a C -P lyase pathw ay [113,114], The JV. spumigena C C Y 9414 genome contains phnC-phnM (nsp7450-nsp7590), w ith phnCDE encoding potentia l components o f a high a ffin ity AB C transport system fo r phosphonates (there is another copy o fphnE in this cluster w h ich we have named phnES) and phnG-phnM encoding the putative m em brane-bound C-P lyase complex. In E. coli phnF and phnJV-0 are no t required fo r phosphonate u tiliza tion b u t may encode accessory proteins o f the C-P lyase o r be transcrip tional regulators [113], hence the ir absence in the JV. spumigena C CY9414 genome does not preclude the cluster encoding a functional C-P lyase and phosphonate transporter. The JV. spumigena C CY9414 genome also contains two other gene clusters (nspl8360-nspl8380 and nsp351‘20-nsp35160) potentia lly encoding phosphonate AB C transporter components (Table 6), although the la tte r cluster also contains a truncated phnH linked to the phnM com ponent o f the C — P lyase. The role o f these clusters in phosphonate u tilisation by JV. spumigena remains to be determ ined, a lthough i t is know n that other cyanobacteria can utilize this source o f phosphorus [115,116]. In add ition to C -P lyase cleavage enzymes bacteria may also possess other phosphonatases tha t cleave the C -P bond e.g. phosphonoacetaldehyde phosphonohydrolase [117] belonging to the haloacid dehalogenase (H A D ) superfam ily. Putative members o f this fam ily are also found in the JV. spumigena C C Y 9414 genome (Table 6).

The putative JV. spumigena C C Y 9414 phosphite transport system (genes nsp35050-nsp35090, (Table 6) is sim ilar to the well- characterized ptxABCDE system from Pseudomonas stutzeri [118], w ith am ino acid identities to the corresponding P. stutzeri proteins ranging from 40 -62% . In P. stutzeri, ptxABC encode components o f a h igh a ffin ity phosphite transport system, ptxD encodes a N A D - dependent phosphite dehydrogenase oxid iz ing phosphite to phosphate and ptxE is a lysR fam ily transcrip tional regulator.

PLOS ONE I w w w .p loso n e .o rg 15 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

The ptxABCD gene cluster was found in Prochlorococcus sp. M IT 9 3 0 1 (only 2 o f 18 Prochlorococcus genomes curren tly available possess this cluster) and this was concom itant w ith the ab ility o f this strain to utilise phosphite as sole phosphorus source [119]. A lthough the concentration o f phosphite in m arine waters is unknow n the potentia l obviously exists for jV. spumigena to supplement its phosphorus demand by u tiliz ing this +3 valence phosphorus form .

Further b io in fo rm a tic evidence suggestive o f the critica l nature o f phosphorus in the b io logy o f JV. spumigena is the p lethora o f genes coding fo r phosphatases tha t can be found in the genome encompassing over a dozen d ifferent gene products, presumably fo r degradation o f organic phosphorus sources (Table 6). These genes include an atypical alkaline phosphatase (nsp7010) found in several o ther cyanobacteria [102,120], putative P hoX phospha­tases (nspl2940 and nspl9860) (see [121]), an acid phosphatase (nsp35720), and several metallophosphoesterases (nsp‘29340, nsp‘29350, nsp46480). The product o f gene nsp6490 contains two G lp Q domains and a phytase domain. The form er corresponds to the glycerophosphodiester phosphodiesterase dom ain (GDPD) present in a group o f putative bacterial and eukaryotic glyceropho­sphodiester phosphodiesterases (G P-GDE, EC 3.1.4.46) sim ilar to E. coli periplasm ic phosphodiesterase G lp Q [122], as w ell as p lan t glycerophosphodiester phosphodiesterases (GP-PDEs), all o f w hich catalyze the Ca2+-dependent degradation o f periplasm ic glycer- ophosphodiesters to produce sn-glycerol-3-phosphate (G3P) and the corresponding alcohols. Phytase is a secreted enzyme w hich hydrolyses phytic acid (the dom inant source o f phosphorus in soils) to release inorganic phosphate, re inforc ing the idea tha t JV. spumigena is very w ell equipped to access an array o f potentia l organic, as well as inorganic, P sources in its environm ent.

Secondary Metabolites: a Multitude o f Biosynthetic Pathways

JV. spumigena C C Y 9414 produces nodularin , a potent hepato- tox in com prising a cyclic pentapetide contain ing unusual non- proteinogenic am ino acids [10] that is responsible fo r the deaths o f domestic and w ild animals th roughout the w orld [10,123]. N odu la rin is synthesized by a hyb rid nonribosom al peptide synthetase (NRPS)/polyketide synthase (PKS) enzyme complex [124], as are the heptapeptide hepatotoxic m icrocystins o f freshwater cyanobacteria [125]. The complete nodu larin synthe­tase (nda) gene cluster was elucidated from an Austra lian JV. spumigena strain [124], JV. spumigena C C Y 9414 contains the nodularin synthetase gene cluster (nsp42130-nsp42220), where the order o f the genes in the operon and its length, 48 kb, is identical to the Austra lian isolate (Fig. 4).

Investigation o f JV. spumigena strain A V I from the Baltic Sea led to the discovery o f cyclic nodulapeptin peptides and linear spumigin peptides in add ition to nodularin [126]. The m ajority o f isolated strains and o f trichomes analyzed from the pelagic Baltic Sea are identified as JV. spumigena [34,127,128] and contain nodu larin as well as spumigins and nodulapeptins [129]. Peptide synthetase gene clusters encoding the biosynthetic pathways for the p roduction o f spumigins (nsp49190-nsp49250) and nodulapep­tins (nsp49350-nsp49400) were identified in the genome o f JV. spumigena C C Y 9414 [130,131] (Fig. 4).

Surprisingly, analysis o f the genome identified gene clusters for one add itiona l NRPS, two add itiona l PKS and one additional hyb rid N R P S /P K S gene cluster encoding unknow n peptides (Fig. 4). A compact NRPS gene cluster (nsp50530-nsp50600) consisting o f 3 modules and proteins encoding the biosynthesis o f a 2-carboxy-6-hydroxyoctahydroindole m oiety (Choi) was identified suggesting tha t JV. spumigena C C Y 9414 m igh t produce an

aeruginosin (Fig. 4). Aeruginosins are linear tetrapeptide protease inh ib ito rs found in the genera Planktothrix and Microcystis [132,133] b u t w h ich have never been reported from JV. spumigena. A d d itio n ­ally, a large cryptic NRPS-PKS gene cluster (nsp26910-nsp27060) was found (Fig. 4). The product is no t known, b u t a very sim ilar gene cluster is present in Anabaena PCC 7120. I t is interesting to note that the gene clusters fo r nodularin , spumigin, nodulapeptin and the cryptic gene cluster tha t is supposed to make aeruginosin are not random ly distributed bu t cluster in a 0.8 M b region o f the genome.

In add ition to the PKS modules that were identified as pa rt o f non-ribosom al peptide synthetase (NRPS) gene clusters, two fu rther PK S gene clusters (nsp26710-nsp26730 and nsp l3 6 4 0 - nsp!3650) were discovered in the genome o f JV. spumigena C C Y 9414 (Fig. 4). U nlike the m odular PKS, these enzymes comprise up to three consecutive acyl carrie r p ro te in domains (ACPs, data no t shown), ind icating the ir involvem ent in an iterative fa tty acid like mode o f biosynthesis [134], The classification as iterative PKS is fu rther supported by their phylogenetic clustering in an overall phylogenetic tree o f PKS sequences ([135], data not shown). T w o closely related heterocyst g lyco lip id synthases (nsp!3650 and nsp46340) were also identified (Fig. 4). One o f the clusters shows close s im ilarity to the heterocyst g lyco lip id biosynthesis clusters o f Anabaena PCC 7120 [136] and is most like ly involved in the biosynthesis o f this im portan t heterocyst envelope compound. Structure-based models allow the pred iction o f the substrate for the acyltransferase (AT) dom ain o f PKS proteins (h ttp ://w w w .n ii.re s .in /n rp s -pks .h tm l). Using this speci­fic ity conferring software it was predicted that the two unchar­acterized PK S could be involved in the synthesis o f unusual (e.g. branched) fa tty acids. One o f the clusters is also present in Anabaena PCC 7120. The structure and role o f these unusual lip ids is unknown.

Cyanobacteria are increasingly recognized as a source o f a second class o f peptid ic natura l products that are produced through the post-translational m od ifica tion o f precursor proteins. Three different peptide families, cyanobactins [137,138], m icro- v irid ins [139,140] and lantipeptides (prochlorosins) [141] have been described and d iffe r substantially in the ir respective am ino acid functionalities and mode o f m acrocyclization. The genetic in fo rm a tion fo r the p roduction o f two o f these classes, cyanobactin (nsp33610-nsp33660) and m ic rov ir id in (nsp49400-nsp49480) is present in the JV. spumigena C C Y 9414 genome [139]. However, the PatA hom olog encoded in the cyanobactin cluster o f JV. spumigena C C Y 9414 is truncated and the cluster lacks a precursor gene, most like ly rendering the gene cluster non-functional. The JV. spumigena C C Y 9414 genome fu rther features 7 cryptic bacteriocin gene clusters although none encodes the L a n M enzyme, w hich characterizes the lantipeptide fam ily [142], and two gene clusters related to sunscreen biosynthesis.

Genom ic m in ing approaches and subsequent in vitro reconsti­tu tion studies have previously uncovered the biosynthetic pa th ­ways fo r two im portan t sunscreen compounds in cyanobacteria, mycosporic acids (M A A) and scytonemin [143-145]. Both compounds show a sporadic d is tribu tion in cyanobacteria and are predom inandy detected in terrestria l and m icrob ia l m at comm unities [146]. The fact tha t bo th biosynthesis gene clusters are present in the genome o f the brackish water JV. spumigena C C Y 9414 was therefore unexpected and may give some new im plications fo r the specific adaptation to the brackish water environm ent as well as the capability to fo rm surface scums.

Sum marizing, NRPS and PKS comprise at least 4% o f the genome o f JV. spumigena CCY9414. This num ber includes 9 gene clusters encoding 58 genes and occupying 222 kb o f the genome.

PLOS ONE I w w w .p loso n e .o rg 16 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ra n scrip to m e o f a T ox ic C yanobacte rium

ndaG ndaH ndal

Nodularin synthetase (Modular NRPS/PKSI)

NSOR10 ndaB ndaA ndaC ndaD ndaE

putativesubstrate Ar9 D-Me-Asp Thr Phenyl Malonyl- Malonyl- Malonyl- D-Glu

acetate? CoA CoA CoA

N9414 nsp42130 nsp42140 nsp42150 nsp42160 nsp42180 nsp42190-42220

HO 0

Heterocyst Glycolipid synthase (Iterative PKSI)

PCC7120

putativesubstrate

N9414

hglE

Arg

nsp46340

hgIC alr5356 hetM hetN

D-Me-Asp Thr Phenyl Malonyl- Malonyl- acetate? CoA CoA

Malonyl- D-Glu CoA

nsp46330 nsp46310 nsp46300 nsp46290 nsp46280 nsp46270

Unknown synthetase (Modular NRPS/PKSI)

PCC7120 8112649 all2648 all2647 all2646 all2645 all2643 all2642 all2641

putativesubstrate

N9414

Asn Val Gly Gly

nsp26910 nsp26920 nsp26930 nsp26940 nsp26950

Malonyl- Ala Cys Gly Ala CoA

nsp26960

Ser Tyr

nsp26980 nsp26990 nsp27000

Malonyl-CoA

nsp27060

A90

putativesubstrate

N9414

Nodulapeptin synthetase (Modular NRPS)

aptA apis aptC

Ile Lys ? ?

nsp49350 nsp49360 nsp49370

aptE aptF

nsp49380 nsp49390 nsp49400

L H PT xY^Vroi 991 S8=M»1 R- CH,SMe i917na=Mot|Q) R= CHjSI-OIMo 899 aa=Sor R=OAo

NIES843/NIVA126

Unknown synthetase (Modular NRPS)

aerM aerB aerD aerE aerF aerG

putativesubstrate Arg Tyr

N9414 nsp50530 nsp50540 nsp50550-nsp50570 nsp50580 nsp50590 nsp50600

Spumigin synthetase (Modular NRPS)

spuA spuB spuC spuD spuE spuF

I Iputativesubstrate Pia Htyr MePro Arg

N9414 nsp49250 nsp49230

Unknown synthase (Iterative PKSI)

PCC7120 alr2678 alr2679

nsp49220-nsp49190

putativesubstrate

N9414

?

nsp26730

Malonyl-CoA

Malonyl-CoA

nsp26720 nsp26710

Unknown synthase (Iterative PKSI)

putativesubstrate

N9414 nspl3650 nsp13640

PLOS ONE I w w w .p lo so n e .o rg 17 M arch 2013 | V o lu m e 8 | Issue 3 I e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

F igure 4 . G ene clusters fo r secondary m e ta b o lite biosyntheses In N. spumigena C C Y 9414 . The ass ig n m e nt o f th e gene p ro d u c ts to non - r ibosom a l p e p tid e syn thetases (NRPS) o r p o ly k e tid e synthases (PKS) is ind ica te d b y red and green co lo u r, respective ly . Genes e n co d in g p u ta tiv e ta ilo r in g p ro te in s are ind ica te d in b lack. The c lass ifica tion o f PKS in to ite ra tive and m o d u la r PKS is show n in th e s u b tit le o f each gene c luster. For characterized gene c lusters p ro d u c t nam es w ere inc lu d ed in th e su b title s . Related gene clusters i f p rese n t in th e da tabase are sh ow n w ith th e ir gene a n n o ta tio n s and s tra in nam es above each gene c luster. S ubstra te spec ific ities as p red ic te d by h ttp ://w w w .n ii.re s .in /n rp s -p k s .h tm l are show n u n d e rn ea th each NRPS o r PKS gene co n ta in in g subs tra te a c tiv a tin g dom a ins . Q uestion m arks ind ica te dom a in s w ith unc lea r substra te specific ities. The n um b e rs in th e second line b e lo w th e gene c lusters re la te to th e gene n um bers in N. spum igena CCY9414. d o i:1 0.1371/jo u rn a l.pone .0060224.g004

This is more than the 3% reported for Moorea producera (Lyngbya majuscula 3L), one o f the most p ro lific sources o f natura l metabolites among cyanobacteria [147]. Thus, the genetic in fo rm a tion required fo r the generation o f these secondary metabolites takes a substantial pa rt o f the genomic coding capacity. Even though JV. spumigena is the subject o f frequent chemical analysis, the only other secondary metabolites observed were nodularin , nodulapeptins and spumigins [126,130,148]. This genome analysis suggests tha t JV. spumigena has the potentia l to synthesise a wealth o f o ther peptides and polyketides. There is still enormous interest in new bioactive compounds from bacteria and the ir biosynthetic pathways. M any bacterial NRPS and PKS products have served as lead products fo r d rug development and the in fo rm a tion gained on NRPS and PKS can provide new insights fo r the generation o f “ unnatu ra l” com pound libraries by com binatoria l biosynthesis approaches (e.g. [149]). A nother, new class o f b ioactive compounds in cyanobacteria are ribosom ally produced and posttranslationally m odified peptides [150]. In order to use the potentia l o f this JV. spumigena strain in the future, genomic m in ing strategies have to be developed in order to iden tify the secondary metabolites guided by the substrate predictions fo r the synthesizing enzymes.

Strains and Methods

Ethic StatementThis research d id no t involve endangered o r protected species

and no w ork on vertebrates. The m icrob ia l sampling was done on board o f a Germ an research vessel (FS A lko r, Institute o f m arine Sciences, K ie l) tha t had all the permissions to sample in the Baltic Sea waters. The B ornholm Sea is neither a m arine park nor private property.

JV. spumigena CC Y 9414 was isolated from samples collected from the surface water in the B ornholm Sea by p ick ing single aggregates o f trichomes and p la ting on agar m edium o f a m ixture o f 1 part ASN3 and 2 parts B G 1 1, devoid o f com bined n itrogen [33]. The isolated strain JV. spumigena C C Y 9414 is a toxic p lanktonic, heterocyst-form ing, gas-vacuolate b loom -fo rm ing cyanobacterium and is representative o f those JV. spumigena tha t fo rm toxic surface blooms in brackish coastal seas.

Genome analysisThe genome was sequenced using a com bination o f Sanger and

454 sequencing platforms. For Sanger sequencing, two genomic libraries w ith insert sizes o f 4 and 40 kb were made. The prepared plasmid and fosm id clones were end-sequenced to provide paired- end reads at the J. C ra ig V ente r Science Foundation Jo in t Technology Center on A B I 3730X L D N A sequencers (Applied Biosystems, Foster C ity, CA). W hole-genome random shotgun sequencing produced 47,486 high qua lity reads averaging 811 bp in length, fo r a to ta l o f approxim ately 38.5 M b p o f D N A sequence, analysed as described [151] and leading to the 5.32 M b W hole Genome Shotgun Assembly deposited in GenBank under the accession num ber P iy N A 13447. For this assembly, 4,904 genes, among them 4,860 pro te in-coding genes were predicted.

Since it was not possible to get a single large scaffold from Sanger sequencing reads alone, and because several previously analysed genes were missing, add itiona l sequence data was obtained by pyrosequencing using the GS F L X system provided by Eurofins M W G G m bH Ebersberg, Germany. The GS F L X system delivered 109,881 sequence reads w ith an average read length o f 251 base pairs. A hyb rid 454/Sanger assembly was made using the M IR A assembler [152]. Resulting contigs were jo ined in to scaffolds using B A M B U S [153]. A ltogether, an average 13- fo ld coverage o f the genome was obtained. Gene calling and in itia l annotation was perform ed apply ing the R apid Annotations using Subsystems Technology (RAST) system [154], leading to the W hole Genome Shotgun Assembly deposited at D D B J /E M B L / GenBank under the accession AOFEOOOOOOOO. The version described in this paper is the first version, AOFEOIOOOOOO.

Cultivation and RNA Preparation for Transcriptome Analysis

JV. spumigena C C Y 9414 cells were grown in cell cu lture bottíes using a 2:1 m ixture o f nitrate-free B G 1 1 and - A S N -II I media [2] (salinity 10 PSU). Cells were incubated at am bient a ir in a temperature contro lled incubator at 20°C, 40 pm ol photons m -2 s“ 1. The photoperiod was set at 16 h ligh t and 8 h dark. Cells were m ixed by daily shaking o f the cell culture bottíes. 50 m l o f cells from the m iddle o f the lig h t period were harvested by quick filtra tio n through sterile glass fibre filters (W hatm an G F /F ). Filters and cells were im m ediate ly frozen in liq u id n itrogen and stored at — 80°C.

T o ta l R N A o f JV. spumigena CC Y 9414 was isolated using the T o ta l R N A Isolation K i t fo r plants (Macherey-Nagel). T o im prove R N A yield, ice-cold lysis buffer (buffer RAP, Macherey-Nagel) was added to the frozen cells on filters and the m ixture was shaken w ith steel beads (cell m ill M M 4 0 0 , Retsch) w ith m axim um speed, three times fo r 30 seconds. For sequence analysis, cD N A libraries were constructed (vertís B iotechnologie A G , Germany) and analysed on an Illu m in a sequencer as previously described [19]. In brief, total R N A was enriched fo r p rim a ry transcripts by treatm ent w ith T e rm ina to r 5’phosphate-dependent exonuclease (Epicentre). Then, 5’PPP R N A was cleaved enzymatically using tobacco acid pyrophosphatase (TAP), the ’de-capped’ R N A was ligated to an R N A linke r [19] and l st-strand c D N A synthesis in itia ted by random p rim ing . The 2nd strand c D N A synthesis was p rim ed w ith a b io tiny la ted antisense 5’-Solexa p rim er, after w h ich cD N A fragments were bound to streptavidin beads.

Bead-bound cD N A was b lunted and 3’ ligated to a Solexa adapter. The c D N A fragments were am plified by 22 cycles o f PCR. For Illu m in a H iSeq analysis (100 bp read length), the cD N A in the size range o f 200 - 500 bp was eluted from a preparative agarose gel. A total o f 41,519,905 reads was obtained. The data was deposited in the N C B I Short Read A rch ive under accession SRS392745.

Reads were mapped to the genome using segemehl [155] w ith default settings, resulting in 40,577,305 mapped reads. T ranscrip ­tiona l start sites (TSSs) were predicted fo r positions where &280 reads start and the num ber o f reads starting at the position is

PLOS ONE I w w w .p loso n e .o rg 18 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

> 5 0% larger than the num ber o f reads covering the position. C lassification o f TSSs in to gTSSs, iTSSs, aTSSs and nTSSs was carried out according as described [19].

Data InterpretationProtein sequences were compared w ith those from Anabaena

variabilis A T C C 29413, Anabaena PCC 7120, Nostoc punctiforme PCC 73102 and Synechocystis sp. PCC 6803 using B LA S T p w ith an e- value cu t-o ff o f le 8. H ig h scoring sequence pairs fo r the same sequences were merged and the per cent iden tity and alignm ent length values recomputed. M erged h igh scoring sequence pairs w ith alignm ent length coverage less than 10% o f the longer sequence were removed. Those sharing the same query o r subject sequence were filtered as follows: first, the best h it was kept together w ith hits whose per cent iden tity is at most ten percentage points smaller; second, we removed those hits whose alignm ent length coverage was more than 20 percentage points smaller than tha t o f the best h it. The rem ain ing hits were clustered using M C L w ith default parameters. Based on this clustering we defined unique and shared genes o f the genomes. Phylogenetic classifica­tion o f pro te in sequences was carried out using M E G A N . B LA S T p results against the N C B I n r database requ iring a m in im um e-value o f le -8 were used as input.

IS elements were identified and assigned to IS families based on the genes o r gene fragments encoding transposase by the IS finder a lgorithm [26] using default parameters and a B LA S T p threshold o f E < le “ 5.

Analysis o f secondary metabolite genesNRPS and PK S gene clusters gene clusters were identified using

met2db [156]. Adényla tion dom ain substrate specificity predic­tions fo r NRPS enzymes were made using NRPSpreditor2 [157]. Cata lytic dom ain annotations fo r NRPS and PK S proteins were refined m anually using CD-search, B LA S T P and InterProScan. Putative functions were assigned to proteins encoding ta ilo ring enzymes associated w ith these cluster were also identified using CD-search, B LA S T P and InterProScan searches. The cyanobac­tin gene cluster was identified using sequences from the patellam ide gene cluster as a query in B LA S T p searches.

Supporting Information

Figure SI C lu ster a n a ly s is o f p ro te in s p o ten tia lly involved in su cro se m eta b o lism in cyan ob acteria . Puta­tive proteins from N spumigena CC Y 9414 (labelled nsp and in boldface letters) are included. Sps - sucrosephosphate synthase, Spp - sucrosephosphate phosphatase, Sus - sucrose synthase. The evolutionary history was in ferred using the M in im u m Evolu tion m ethod w ith in M E G A 5 [158]. The optim a l tree w ith the sum o f

References1. Paerl H W , H u is m a n J (2008) C lim ate. Blooms like it hot. Science 320: 57-58.2. R ippka R , Deruelles J , W a te rb u ry JB , H e rd m an n M , Stanier R Y (1979)

G eneric assignments, strain histories and properties o f pure cultures o f cyanobacteria. J G en M ic ro b io l 111: 1-61.

3. M uro-Pastor A M , Hess W R (2012) Heterocyst differentiation: from single mutants to global approaches. Trends M icro b io l 20: 548-557 .

4. Stal LJ (2009) Is the distribution o f n itrogen-fixing cyanobacteria in the oceans related to tem perature? E nviron M ic ro b io l 11: 1632-1645.

5. V in tila S, E l-Shehaw y R (2007) A m m o n iu m ions inh ib it nitrogen fixation bu t do not affect heterocyst frequency in the b loom -form ing cyanobacterium Nodularia spumigena strain A V I . M icrob io logy 153: 3704-3712.

6. Suikkanen S, K aartoka llio H , Hällfors S, H u ttu n e n M , Laam anen M (2010) Life cycle strategies o f b loom -form ing , filam entous cyanobacteria in the Baltic Sea. Deep-Sea Res I I 57: 1 99 -2 09 .

branch length = 7.8464659 is shown. The percentage o f replicate trees in w h ich the associated taxa clustered in the bootstrap test (10,000 replicates) are shown next to the branches i f > 6 0 . The tree is draw n to scale, w ith branch lengths in the same units as those o f the evolutionary distances used to in fer the phylogenetic tree and are in the units o f the num ber o f am ino acid substitutions per site. A ll positions w ith less than 50% site coverage were elim inated. There were a to ta l o f 716 positions in the fina l dataset.(PPTX)

Figure S2 F usion p ro te in s b e tw een an Is iA /C P 43 ho­m o lo g and PsaL in A n abaen a 7120 and N . sp u m ig en aCCY9414. A . Sequence alignm ent o f the CP43-PsaL fusion proteins from N spumigena C C Y 9414 (jVij&37500) and Anabaena PCC7120 (A114002) and the respective PsaL proteins (jVij&40050 and A110107). B. Prediction o f transmembrane helices fo r the jVij&37500 fusion pro te in (numbered I to IX ). The topology and possible transmembrane helices were predicted using T M H M M 2.0 at h ttp ://w w w .c b s .d tu .d k /s e rv ic e s /T M H M M /. PsbC-PsaL h yb rid proteins sim ilar to Nsp'iTNX) exist in on ly nine other cyanobacteria: in Anabaena PCC 7120, Moorea producens (.Lyngbya majuscula 3L), Leptolyngbya sp. PCC 7375, Fischerella sp. JSC-11, Trichodesmium erythraeum IM S101, Synechococcus spp. JA -2 -3B ’a(2—3) and JA -3-3A b, Oscillatoria sp. PCC 6506 and Crocosphaera watsonii W H 0003.(PPTX)

Table SI F a m ilies o f IS e lem en ts in N. sp u m ig en a CCY9414.(X LS X )

Table S2 D eta ils o f 608 gen e c lu sters th at are com m on to th ree w ell-stu d ied N o sto ca les (Fig. 2A) b u t n o t foundin N . sp u m ig en a CCY9414. The acronyms are as follows:N_punct, Nostoc punctiforme sp. PCC 73102; A_var, Anabaena variabilis sp. A T C C 29413; N_7120, Anabaena PCC 7120, based on M C L clustering o f B L A S T p results (m in im um e-value: IO-8 ). (X LS X )

Table S3 L ist o f p red ic ted N. sp u m ig en a CCY9414 p ro te in s n o t p re sen t in A n abaen a PCC 7120, N o sto c p u n c tifo rm e sp . PCC 73102, or A n abaen a v a r ia b ilis sp . ATCC 29413.(X LS X )

Author ContributionsConceived and designed the experiments: MH IJS WRH. Performed the experiments: HB FM MH LJS. Analyzed the data: BV HB DPF MK FM FH RES PH BB KS ED DJS MH LJS WRH. Contributed reagents/ materials/analysis tools: BV MK. Wrote the paper: BV HB DPF PH BB KS ED DJS MH LJS WRH.

7. W a n g H , Sivonen K , R ouhia inen L , Fewer D P , Lyra C et al. (2012) G enom e- derived insights into the biology o f the hepatotoxic b loom -form ing cyanobac­terium Anabaena sp. strain 90. B M C Genomics 13: 613.

8. Larsson U , H a jd u S, W a ive J , E lm gren R (2001) Baltic Sea nitrogen fixation estimated from the sum m er increase in upper m ixed layer total nitrogen. Lim n o l O ceanogr 46: 811 -820 .

9. Degerho lm J , Gundersen K , H a jd u S, Bergm an B, Söderbäck E (2008) Seasonal significance o f N 2 fixation in coastal and offshore waters o f the north­western Baltic Sea. M a rin e Ecol Prog Series 360: 73-84.

10. Sivonen K , K onon en K , C arm ichael W W , D a h lem A M , R in e h art K L et al. (1989) O ccurrence o f the hepatotoxic cyanobacterium Nodularia spumigena in the Baltic Sea and structure o f the toxin. A p p l E nviron M ic ro b io l 55: 1990-1995.

11. Stal LJ, A lbertano P, Bergm an B, von Bröckeld K , G allone J R et al. (2003) B A S IC : Baltic Sea cyanobacteria. A n investigation o f the structure and dynamics o f w ater blooms o f cyanobacteria in the Baltic Sea - responses to a changing environm ent. C o n t Shelf Res 23: 1695-1714.

PLOS ONE I w w w .p loso n e .o rg 19 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

12. E d ler L , Ferno S, L in d M G , Lundberg R , Nilsson P O (1985) M o rta lity o f dogs associated w ith a b loom o f the cyanobacterium Nodularia spumigena in the Baltic Sea. O phe lia 24: 103-109.

13. Sim óla O , W ib e rg M , Jokela J , W ahlsten M , Sivonen K et al. (2012) Pathological findings and identification o f the toxin in acute cyanobacterial (,Nodularia spumigena) in toxication in a dog. V eterinary Pathol 49: 755-759.

14. Sellner K G (1997) Physiology, ecology, and toxic properties o f m arine cyanobacteria blooms. L im n o l O ceanogr 42: 1089-1104.

15. H u b e r AT; (1984) Nodularia (cyanobacteriaceae) akinetes in the sediments o f the Peel-H arvey estuary, W estern Australia: potential inoculum source fo r Nodularia blooms. A p p l E nviron M ic ro b io l 47: 234-238.

16. H u b e r AT; (1985) Factors affecting the germ ination o f akinetes o f Nodularia spumigena (Cyanobacteriaceae). A p p l E nviron M ic ro b io l 49: 73-78.

17. Paerl H W , Prufert L E , Am brose W W (1991) Contem poraneous N 2 fixation and oxygenic photosynthesis in the nonheterocystous m at-fo rm ing cyanobac­terium Lyngbya aestuarii. A pp l E nviron M ic ro b io l 57: 3086-3092.

18. M itschke J , G eorg J , Scholz I , Sharm a G M , D ienst D et al. (2011a) A n experim entally anchored m ap o f transcriptional start sites in the m odel cyanobacterium Synechocystis sp. P C C 6 80 3 . Proc N a d A cad S c iU S A 108: 2124- 2129.

19. M itschke J , V io que A , Haas F , Hess W R , M uro-Pastor A M (201 lb ) Dynam ics o f transcriptional start site selection during n itrogen stress-induced cell differentiation in Anabaena sp. P G G 7120 . Proc N a d A cad Sei U S A 108: 2 0130-20135.

20. D iih r in g U , A x m an n I M , Hess W R , W ild e A (2006) A n in ternal antisense R N A regulates expression o f the photosynthesis gene isiA. Proc N a d A cad Sei U S A 103: 7054-7058.

21. Eisenhut M , G eorg J , K la h n S, Sakurai I , M ustila H et al. (2012) T h e antisense R N A A s l_ flv 4 in the cyanobacterium Synechocystis sp. P C C 6803 prevents prem ature expression o f the flv4-2 operon upon shift in inorganic carbon supply. J B io l G hem 287: 33153 -33162.

22. Finsinger K , Scholz I , Serrano A , M ora les S, U rib e -L o rio L et al. (2008) Characterizadon o f true-branching cyanobacteria from geotherm al sites and hot springs o f Costa R ica. E nviron M ic ro b io l 10: 460-473 .

23. Stucken K , John U , Cem bella A , M u rillo A A , Soto-Liebe K et al. (2010) Th e smallest know n genomes o f m ulticellu lar and toxic cyanobacteria: comparison, m in im a l gene sets fo r linked traits and the evolutionary implications. PLoS O ne 5: e9235.

24. X u M Q , K ä th e S D , G oodrich-B lair H , N ierzw icki-B auer SA, Shub D A (1990) Bacterial origin o f a chloroplast intron: conserved self-splicing group I introns in cyanobacteria. Science 250: 1566-1570.

25. Biniszkiewicz D , Cesnaviciene E , Shub D A (1994) Selfhyphen;splicing group I in tron in cyanobacterial in itiator m ethionine tR N A : evidence for lateral transfer o f introns in bacteria. E M B O J 13: 4629-4635.

26. Siguier P, V a ra n i A , Perochon J , C handler M (2012) Exp loring bacterial insertion sequences w ith ISfm der: objectives, uses, and future developments. M ethods M o l B io l 859: 91-103.

27. Bench S R , Ilikchyan IN , T r ip p HJ, Z e h rJ P (2011) T w o strains o f Crocosphaera watsonii w ith highly conserved genomes are disdnguished by strain-specific features. Fron t M ic ro b io l 2.

28. Guo H , Tse L V , Barbalat R , S ivaam nuaiphorn S, X u M et al. (2008) D iversity- generating retroelem ent hom ing regenerates target sequences for repeated rounds o f codon rew riting and protein diversification. M o l C e ll 31: 813-823 .

29. Doulatov S, Hodes A , D a i L , M a n d h a n a N , L iu M et al. (2004) T rop ism switching in Bordetella bacteriophage defines a fam ily o f diversity-generating retroelements. N a tu re 431: 476 -481 .

30. G ladyshev E A , A rk h ip o v a I R (2 0 11 ) A w idespread class o f reverse transcriptase-related cellular genes. Proc N a d A cad Sei U S A 108: 20311- 20316.

31. H ig a K C , C allahan S M (2010) Ectopic expression o f hetP can partia lly bypass the need for hetR in heterocyst differentiation by Anabaena sp. strain P C C 7120. M o l M ic ro b io l 77: 562 -574 .

32. Leng X , Z h u W , J in J , M a o X (2011) Evidence tha t a chaperone-usher-like pathw ay o f Myxococcus xanthus functions in spore coat form ation. M icrobiology 157: 1886-1896.

33. Hayes P K , Barker G L A (1997) Genetic diversity w ith in Baltic Sea populations o f Nodularia (Cyanobacteria). J Phycol 33: 919-923 .

34. Lyra C , Laam anen M , L eh tim aki J M , Surakka A , Sivonen K (2005) Benthic cyanobacteria o f the genus Nodularia are non-toxic, w ithou t gas vacuoles, able to glide and genetically m ore diverse than planktonic Nodularia. In t J Syst Evol M icro b io l 55: 555 -568 .

35. Stanley D N , Raines G A , K e rfe ld C A (2013) Com parative analysis o f 126 cyanobacterial genomes reveals evidence o f functional diversity am ong homologs o f the redox-regulated C P 12 protein. P lant Physiol 161: 824 -835 .

36. Lechno-Yossef S, Fan Q , W o jc iuch E , W o lk C P (2011) Iden tification o f ten Anabaena sp. genes that, under aerobic conditions, are required fo r growth on dinitrogen bu t not for grow th on fixed nitrogen. J Bacteriol 193: 3482-3489.

37. H e rre ro A , M u ro -P a s to r A M , F lores E (2 0 01 ) N itro g e n co n tro l in cyanobacteria. J Bacteriol 183: 411 -425 .

38. M u ro -P a s to r A M , V a llad ares A , Flores E , H e rre ro A (2002 ) M u tu a l dependence o f the expression o f the cell differentiation regulatory protein H etR and the global nitrogen regulator N tc A during heterocyst developm ent. M o l M ic ro b io l 44: 1377-1385.

39. R am asubram anian T S , W e i T F , O ld h am A K , G o ld e n J W (1996) Transcrip ­tion o f the Anabaena sp. strain P C C 7120 ntcA gene: m ultip le transcripts and N tc A binding. J Bacteriol 178: 922-926 .

40. Buikem a W J, Haselkorn R (2001) Expression o f the Anabaena hetR gene from a copper-regulated prom oter leads to heterocyst differentiation under repressing conditions. Proc N a tl A cad Sei U S A 98: 2729-2734.

41. Rajagopalan R , C a llahan S M (2010) T e m p o ra l and spatial regulation o f the four transcription start sites o f hetR from Anabaena sp. strain P C C 7120. J Bacteriol 192: 1088-1096.

42. Valladares A , M uro-Pastor A M , H e rre ro A , Flores E (2004) T h e N tcA - dependent P i prom oter is utilized for glnA expression in N 2-frxing heterocysts o f Anabaena sp. strain P C C 7120. J Bacteriol 186: 7337-7343.

43. Frias J E , Flores E , H e rre ro A (1994) R equirem ent o f the regulatory protein N tc A fo r the expression o f nitrogen assimilation and heterocyst developm ent genes in the cyanobacterium Anabaena sp. P C C 7 12 0 . M o l M ic ro b io l 14: 823- 832.

44. Turner N E , Robinson SJ, Haselkorn R (1983) D ifferen t prom oters fo r the Anabaena glutam ine synthetase gene during grow th using m olecular or fixed nitrogen. N a tu re 306: 337-342.

45. G alm ozzi C V , Saelices L , Florencio IJ , M uro-Pastor M I (2010) Posttranscrip- tional regulation o f glutam ine synthetase in the filam entous cyanobacterium Anabaena sp. P C C 7120: differential expression betw een vegetative cells and heterocysts. J Bacteriol 192: 4701-4711.

46. R am asubram anian T S , W e i T F , G o ld e n J W (1994) T w o Anabaena sp. strain P C C 7120 D N A -b in d in g factors interact w ith vegetative cell- and heterocyst- specific genes. J Bacteriol 176: 1214-1223.

47. Schneider GJ, L a n g J D , Haselkorn R (1991) Prom oter recognition by the R N A polymerase from vegetative cells o f the cyanobacterium Anabaena 7120. G ene 105: 51-60 .

48. N ierzw icki-B auer SA, Curtis SE, Haselkorn R (1984) Cotranscription o f genes encoding the small and large subunits o f ribulose-l,5-bisphosphate carboxylase in the cyanobacterium Anabaena 7120. Proc N a tl A cad Sei U S A 81: 5961-5965.

49. H agem ann M (2011) M o lecu lar biology o f cyanobacterial salt acclim ation. F E M S M ic ro b io l R ev 35: 87-123.

50. H agem ann M , M a r in K (1999) Salt-induced sucrose accum ulation is m ediated by sucrose-phosphate-synthase in cyanobacteria. J P lan t Physiol 155: 424-430 .

51. C um ino A C , M a rco zz i C , Barreiro R , Salerno G L (2007) C arbon cycling in Anabaena sp. P C C 7120. Sucrose synthesis in the heterocysts and possible role in nitrogen fixation . P lant Physiol 143: 1385-1397.

52. Porchia A C , Salerno G L (1996) Sucrose biosynthesis in a prokaryotic organism: Presence o f two sucrose-phosphate synthases in Anabaena w ith rem arkable differences com pared w ith the p lan t enzymes. Proc N a tl A cad Sei U S A 93: 13600-13604.

53. C u ra tti L , Flores E , Salerno G (2002) Sucrose is involved in the diazotrophic m etabolism o f the heterocyst-form ing cyanobacterium Anabaena sp. FEBS Le tt 513: 175-178.

54. W o lf A , K ra m e r R , M o rb a c h S (2003) T h re e pathways fo r trehalose m etabolism in Corynebacterium glutamicum A T C C 13032 and their significance in response to osmotic stress. M o l M ic ro b io l 49: 1119-1134.

55. H igo A , K a to h H , O h m o ri K , Ikeuchi M , O h m o ri M (2006) T h e role o f a gene cluster for trehalose m etabolism in dehydration tolerance o f the filamentous cyanobacterium Anabaena sp. P C C 7120. M icrob io logy 152: 979-987 .

56. H igo A , Suzuki T , Ikeuchi M , O h m o ri M (2007) D ynam ic transcriptional changes in response to rehydration in Anabaena sp. P C C 7120. M icrobiology 153: 3685-3694.

57. K a to h H , A sthana R K , O h m o ri M (2004 ) G ene expression in the cyanobacterium Anabaena sp. P C C 7 1 2 0 under desiccation. M ic ro b E co l 47: 164-174.

58. Kappes R M , K e m p f B, K n e ip S, Boch J , G ade J et al. (1999) T w o evolutionarily closely related A B C transporters m ediate the uptake o f choline for synthesis o f the osm oprotectant glycine betaine in Bacillus subtilis. M o l M icro b io l 32: 203-216.

59. Lucht J M , Brem er E (1994) A daptation o f Escherichia coli to high osmolarity environments: osm oregulation o f the h igh-affin ity glycine betaine transport system pro U. F E M S M ic ro b io l R ev 14: 3-20.

60. Stal LJ, Staal M , V illb ra n d t M (1999) N u tr ie n t control o f cyanobacterial blooms in the Baltic Sea. A q u at M ic ro b Ecol 18: 165-173.

61. K a to h H , H ag ino N , Grossman A R , O gaw a T (2001) Genes essential to iron transport in the cyanobacterium Synechocystis sp. strain P C C 6803. J Bacteriol 183: 2779-2784.

62. Stevanovic M , H a h n A , Nicolaisen K , M irus O , Schleiff E (2012) T he components o f the putative iron transport system in the cyanobacterium Anabaena sp. P C C 7120. E nviron M ic ro b io l 14: 1655-1670.

63. Nicolaisen K , M ariscal V , Bredem eier R , P em il R , Moslavac S et al. (2009) T h e outer m em brane o f a heterocyst-form ing cyanobacterium is a perm eability barrie r for uptake o f metabolites that are exchanged betw een cells. M o l M icro b io l 74: 58-70.

64. Burnap R L , T ro y a n T , Sherm an L A (1993) T h e highly abundant chlorophyll- protein com plex o f iron-defic ient Synechococcus sp. P C C 7 94 2 (C P 4 3 ’) is encoded by the isiA gene. P lant Physiol 103: 893-902 .

65. Laudenbach D E , Straus N A (1988) C haracterization o f a cyanobacterial iron stress-induced gene sim ilar to psbC. J Bacteriol 170: 5018-5026.

PLOS ONE I w w w .p loso n e .o rg 20 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

66. C hauhan D , Folea I M , Jolley C C , K o u ril R , Lubner C E et al. {2011) A novel photosynthetic strategy fo r adaptation to low -iron aquatic environments. Biochemistry 50: 686-692 .

67. H a va u x M , G uedeney G , H agem ann M , Y erem enko N , M atth ijs H C et al. (2005) T h e chlorophyll-b inding protein Is iA is inducible by high light and protects the cyanobacterium Synechocystis P C C 6 8 0 3 from photooxidative stress. FEBS L e tt 579: 2289-2293.

68. G eiß U , V in n em e ierJ , Schoor A , H agem ann M (2001) T h e iron-regulated isiA gene o f Fischerella muscicola strain P C C 73103 is linked to a likewise regulated gene encoding a Pcb-like chlorophyll-b inding protein . F E M S M ic ro b io l Le tt 197: 123-129.

69. Shih P M , W u D , La tifi A , A xen S D , Fewer D P et al. (2013) Im proving the coverage o f the cyanobacteria l phy lu m using divers ity-driven genom e sequencing. Proc N a d A cad Sei U S A 110: 1053-1058.

70. K o u ril R , A rte n i A A , L a x J , Y erem enko N , D ’H aen e S et al. (2005) Structure and functional ro le o f supercomplexes o f Is iA and Photosystem I in cyanobacterial photosynthesis. FEBS L e tt 579: 3 2 5 3 -3 2 5 7 .

71. V in tila S, Selao T , N o re n A , Bergm an B, E l-Shehaw y R (2011) C haracter­ization o f n ifH gene expression, m odification and rearrangem ent in Nodularia spumigena strain A V I . F E M S M ic ro b io l Ecol 77: 449 -459 .

72. L indberg P, H ansel A , L indb lad P (2000) hupS and hupL constitute a transcription un it in the cyanobacterium Nostoc sp. P C C 73102. A rc h M ic rob io l

174: 129-133.73. Zhao M - X , J iang Y -L , H e Y -X , C hen Y -F , T e n g Y -B et al. (2010) Structural

basis for the allosteric control o f the global transcription factor N tc A by the nitrogen starvation signal 2-oxoglutarate. Proc N a d A cad Sei U S A 107: 12487- 12492.

74. K im Y , Joach im iak G , Y e Z , Binkowski T A , Zhang R et al. {2011) Structure o f transcription factor H etR required for heterocyst d ifferentiation in cyanobac­teria. Proc N a d A cad Sei U S A 108: 10109-10114.

75. Black T A , C a i Y , W o lk C P {1993) Spatial expression and autoregulation o f hetR, a gene involved in the control o f heterocyst developm ent in Anabaena. M o l M icro b io l 9: 77-84.

76. Buikem a W J, Haselkorn R {1991) Characterization o f a gene controlling heterocyst d ifferentiation in the cyanobacterium Anabaena 7120. Genes D ev 5: 321-330.

77. Buikem a W J, Haselkorn R {2001) Expression o f the Anabaena hetR gene from a copper-regulated prom oter leads to heterocyst differentiation under repressing conditions. Proc N a t A cad Sei U S A 98: 2729-2734.

78. C a i Y , W o lk C P {1997) Anabaena sp. S train P C C 7120 responds to nitrogen deprivation w ith a cascade-like sequence o f transcriptional activations.

J Bacteriol 179: 267-271.79. W o n g F C Y , Meeks J C {2001) T h e hetF gene product is essential to heterocyst

differentiation and affects H etR function in the cyanobacterium Nostoc punctiforme. J Bacteriol 183: 2654-2661.

80. Risser D D , C a llahan S M {2008) H e tF and P a tA control levels o f H etR in Anabaena sp. strain P C C 7120. J Bacteriol 190: 7645-7654.

81. Ionescu D , Voss B, O re n A , Hess W R , M uro-Pastor A M {2010) Heterocyst- specific transcription o f N s iR l, a non-coding R N A encoded in a tandem array

o f direct repeats in cyanobacteria. J M o l B io l 398: 177-188.82. Z hang W , D u Y , Khudyakov I , Fan Q , G ao H et al. {2007) A gene cluster that

regulates both heterocyst differentiation and pattern fo rm ation in Anabaena sp. strain P C C 7120. M o l M ic ro b io l 66: 1429-1443.

83. D u Y , C a i Y , H o u S, X u X {2012) Idendficadon o f the H etR recognition sequence upstream o f hetf, in Anabaena sp. strain P C C 7120. J Bacteriol 194: 2297-2306.

84. Jones K M , Buikem a W J, Haselkorn R {2003) Heterocyst-specific expression o f patB , a gene required fo r nitrogen fixation in Anabaena sp. S train P C C 7120. J Bacteriol 185: 2306-2314.

85. L ian g J , Scappino L , Haselkorn R {1992) T h e patA gene product, w hich contains a region sim ilar to C h e Y o f Escherichia coli, controls heterocyst pattern form adon in the cyanobacterium Anabaena 7120. Proc N a d A cad Sei U S A 89: 5655-5659.

86. L ian g J , Scappino L , Haselkorn R {1993) T h e patB product, required for grow th o f the cyanobacterium Anabaena sp. strain P C C 7120 under nitrogen- l im it in g conditions, contains fe rro d o x in and h e lix -tu rn -h e lix dom ains. J Bacteriol 175: 1697-1704.

87. N agara ja R , Haselkorn R {1994) Protein H U from the cyanobacterium Anabaena. B iochem ie 76: 1082-1089.

88. Khudyakov I , W o lk C P {1996) Evidence that the hanA gene coding for H U protein is essential for heterocyst differentiation in, and cyanophage A -4 {L) sensitivity of, Anabaena sp. S train P C C 7120. J Bacteriol 178: 3572-3577.

89. W a n g Y , X u X {2005) R eguladon by hetC o f genes required for heterocyst differentiation and cell division in Anabaena sp. strain P C C 7 1 2 0 .J Bacteriol 87: 8489-8493.

90. X u X , W o lk C P {2001) Role fo r hetC in the transidon to a nondivid ing state during heterocyst differentiation in Anabaena sp. J Bacteriol 183: 393-396.

91. L iu D , G o ld en J W {2002) hetL overexpression sdmulates heterocyst form adon in Anabaena sp. strain P C C 7120. J Bacteriol 184: 6873-6881.

92. G olden J W , Robinson SJ, Haselkorn R {1985) Rearrangem ent o f nitrogen fixation genes during heterocyst differentiation in the cyanobacterium Anabaena. N a tu re 314: 419 -423 .

93. Henson BJ, H a rtm a n L , W atson L E , B arnum S R {2011) E voludon and variation o f the nifD and hupL elements in the heterocystous cyanobacteria. In t J Syst E vo l M ic ro b io l 61: 2938-2949.

94. Henson BJ, W atson L E , B arnum S R {2005) Characterizadon o f a 4 kb variant o f the nifD elem ent in Anabaena sp. strain A T C C 33047. C u rr M ic ro b io l 50: 129-132.

95. Carrasco C D , H o llid ay S D , Hansel A , L indb lad P, G olden J W {2005) Heterocyst-specific excision o f the Anabaena sp. strain P C C 7120 hupL elem ent requires xisC. J Bacteriol 187: 6031-6038.

96. V in t i la S, E l-S h eh a w y R {2010) V a r ia b il i ty in the response o f the cyanobacterium Nodularia spumigena to nitrogen supplem entadon. J E nviron M o n ito r 12: 1 88 5-18 9 0 .

97. H ecky R E , K ilh a m P {1988) N u trie n t lim ita tion o f phytoplankton in freshwater and m arine environments: A review o f recent evidence on the effects o f enrichm ent. L im n o l O ceanogr 33: 796-822 .

98. A m m e rm a n J W , H o o d R R , Case D A , C o tner J B {2003) Phosphorus deficiency in the A tlantic: an em erging paradigm in oceanography. E O S 84: 165-170.

99. D y h rm an S T , H a ley S T {2006) Phosphorus scavenging in the unicellular m arine diazotroph Crocosphaera watsonii. A p p l E nviron M ic ro b io l 72: 1452-1458.

100. Su Z , O lm a n V , X u Y {2007) C om putational prediction o f Pho régulons in cyanobacteria. B M C Genomics 8: 156.

101. O rch a rd E D , W ebb E A , D y h rm a n S T {2009) M o lecu lar analysis o f the phosphorus starvation response in Trichodesmium spp. E nviron M ic ro b io l 11: 2400-2411.

102. Scanlan DJ, Ostrowski M , M a z a rd S, Dufresne A , G arczarek L et al. {2009) Ecological genomics o f m arine picocyanobacteria. M ic ro b io l M o l B io l R ev 73: 249-299.

103. Nausch M , Nausch G , W asm und N , N age l K {2008) Phosphorus pool variations and their relation to cyanobacteria developm ent in the Baltic Sea: a three year study. J M a r Syst 71: 99—111.

104. V ah te ra E , A udo R , K aartoka llio H , Laam anen M {2010) Phosphate addition to phosphorus-deficient Baltic Sea p lankton com m unides benefits nitrogen- fixing Cyanobacteria. A q u at M ic ro b Ecol 60: 43-57.

105. Jonasson S, V in tila S, Sivonen K , E l-Shehaw y R {2008) Expression o f the nodularin synthetase genes in the Baltic Sea b loom -form er cyanobacterium Nodularia spumigena strain A V I . F E M S M ic ro b io l Ecol 65: 31-39.

106. P itt F D , M a z a rd S, H um phreys L , Scanlan DJ {2010) Funcdonal character­izadon o f Synechocystis sp. strain P C C 6803 jbstl and p st2 gene clusters reveals a novel strategy fo r phosphate uptake in a freshwater cyanobacterium . J Bacteriol 192: 3512-3523.

107. A ib a H , M izu n o T {1994) A novel gene whose expression is regulated by the response-regulator, SphR, in response to phosphate lim itadon in Synechococcus species P C C 7 94 2 . M o l M ic ro b io l 13: 2 5 -3 4 .

108. M a n n N H , Scanlan D J {1994) T h e S p h X protein o f Synechococcus sp. P C C 7 94 2 belongs to a fam ily o f phosphate-binding proteins. M o l M ic ro b io l 14: 595-596 .

109. C utter G A , C utter LS , Featherstone A M , Lohrenz SE {2001) A n d m o n y and arsenic biogeochemistry in the western A dantic O cean. D eep Sea Res P art I I 48: 2895-2915.

110. O rdóñez E , Letek M , V a lbuena N , G il J A , M ateos L M {2005) Analysis o f genes involved in arsenic resistance in Coynebacterium glutamicum A T C C 13032. A pp l E nviron M ic ro b io l 71: 6 2 0 6 -6 2 1 5 .

111. Silver S, Phung L {2005) Genes and enzymes involved in bacterial oxidation and reduction o f inorganic arsenic. A p p l E nviron M ic ro b io l 71: 599—608.

112. M a rd n e z A , Tyson G W , D elong E F {2010) W idespread know n and novel phosphonate utilization pathways in m arine bacteria revealed by funcdonal screening and m etagenom ic analyses. E nviron M ic ro b io l 12: 222-238.

113. W a n n e r B L {1994) M o lecu lar genetics o f carbon—phosphorus bond cleavage in bacteria. B iodegradadon 5: 175— 184.

114. W h ite A K , M e tc a lf W W {2004) T w o C -P lyase opérons in Pseudomonas stutzeri and their roles in the oxidadon o f phosphonates, phosphite, and hypophos- phite. J Bacteriol 186: 4730-4739.

115. Adam s M M , G om ez-G arc ia M R , Grossman A R , Bhaya D {2008) Phosphorus deprivation responses and phosphonate u tilization in a therm ophilic Synecho­coccus sp. from m icrobial mats. J Bacteriol 190: 8171-8184.

116. G om ez-G arc ia M R , Davison M , B la in -H artnung M , Grossman A R , Bhaya D {2011) A lternadve pathways for phosphonate m etabolism in therm ophilic cyanobacteria from m icrobial mats. IS M E J 5: 141-149.

117. Q u in n J P , Ku lakova A N , Cooley N A , M c G ra th J W {2007) N e w ways to break an old bond: the bacteria l carbon-phosphorus hydrolases and their role in biogeochem ical phosphorus cycling. E nviron M ic ro b io l 9: 2392-2400.

118. M e tc a lf W W , W o lfe R S {1998) M o lecu la r genetic analysis o f phosphite and hypophosphite oxidadon by Pseudomonas stutzeri W M 8 8 . J Bacteriol: 5547-5558.

119. M a rd n e z A , O sburne M S , Sharm a A K , D e L o n g E F , Chisholm S W {2012) Phosphite u tiliza tio n by the m a rin e p ico cyano bacterium Prochlorococcus M IT 9 3 0 1 . E nviron M ic ro b io l 14: 1363-1377.

120. R ay J M , Bhaya D , Block M A , Grossman A R {1991) Isolation, transcription, and inactivadon o f the gene fo r an atypical alkaline phosphatase o f Synechococcus sp. strain P C C 7942. J Bacteriol 173: 4297-4309.

121. Sebastian M , A m m e rm a n J W {2009) T h e alkaline phosphatase P h o X is m ore widely distributed in m arine bacteria than classical PhoA. IS M E J 3: 563-572 .

122. Tom m assen J , E ig lm eier K , Cole S T , O verdu in P, Larson T J et al. {1991) C haracterizadon o f two genes, glpQ_ and u gp(f encoding glycerophosphoryl diester phosphodiesterases o f Escherichia coli. M o l G en G enet 226: 321-327.

123. Francis G {1878) Poisonous Australian lake. N a tu re 18: 11 -12 .

PLOS ONE I w w w .p loso n e .o rg 21 M arch 2013 | V o lu m e 8 | Issue 3 | e60224

G enom e and T ranscrip tom e o f a Toxic C yanobacte rium

124. M o ffitt M C , N e ila n B A (2004) Characterization o f the nodularin synthetase gene cluster and proposed theory o f the evo lu tion o f cyanobacteria l hepatotoxins. A p p l E nviron M ic ro b io l 70: 6353-6362.

125. T ille tt D , D ittm an n E , E rhard M , von D ö h re n H , Börner T et al. (2000) Structural organization o f microcystin biosynthesis in Microcystis aeruginosa P C C 7 80 6 : an integrated peptide-polyketide synthetase system. C h e m B iol 7: 753-764.

126. Fu jii K , Sivonen K , Adachi K , Noguchi K , Sano H et al. (1997a) C om parative study o f toxic and non-toxic cyanobacterial products: novel peptides from toxic Nodularia spumigena A V I . Te trahedron L e tt 31: 5525-5528 .

127. L e h tim ak iJ , Ly ra C , Suom alainen S, Sundm an P, Rouhia inen L et al. (2000) Characterization o f Nodularia strains, cyanobacteria from brackish waters, by genotypic and phenotypic methods. In t J Syst E vol M ic ro b io l 50 P t 3: 1043- 1053.

128. Laam anen M J , Gugger M F , L eh tim äki J M , H au kka K , Sivonen K (2001) Diversity o f toxic and nontoxic Nodularia isolates (cyanobacteria) and filaments from the Baltic Sea. A p p l E nviron M ic ro b io l 67: 4638-4647 .

129. Fewer D P , K öykka M , H a lin e n K , Jokela J , Lyra C et al. (2009) C u lture- independent evidence for the persistent presence and genetic diversity o f m icrocystin-producing Anabaena (C yanobacteria) in the G u lf o f F inland . E nviron M ic ro b io l 11: 855-866 .

130. Fewer D P , Jokela J , R ouhia inen L , W ahlsten M , Koskenniem i K et al. (2009) T h e non-ribosom al assembly and frequent occurrence o f the protease inhibitors spumigins in the b loom -form ing cyanobacterium Nodularia spumigena. M o l M icro b io l 73: 924 -937 .

131. R ouhia inen L , Jokela J , Fewer D P , U rm a n n M , Sivonen K (2010) T w o alternative starter modules for the non-ribosom al biosynthesis o f specific anabaenopeptin variants in Anabaena (Cyanobacteria). C h e m B iol 17: 265-273.

132. Ishida K , Christiansen G , Yoshida W Y , K u rm a ye r R , W e lker M et al. (2007) Biosynthesis and structure o f aeruginoside 126A and 126B, cyanobacterial peptide glycosides bearing a 2-carboxy-6-hydroxyoctahydroindole m oiety. C h e m B iol 14: 565 -576 .

133. Ishida K , W e lker M , Christiansen G , Cadel-S ix S, Bouchier C et al. (2009) Plasticity and evolution o f aeruginosin biosynthesis in cyanobacteria. A pp l E nviron M ic ro b io l 75: 2017-2026.

134. K a u lm an n U , Flertweck C (2002) Biosynthesis o f polyunsaturated fatty acids by polyketide synthases. A ngew C h e m In t E d E ngl 41: 1866-1869.

135. Jenke-K odam a H , Sandm ann A , M u lle r R , D ittm an n E (2005) E volutionary im plications o f bacterial polyketide synthases. M o l Biol E vo l 22: 2027-2039.

136. W o lk C P (2000) Heterocyst fo rm ation in Anabaena. In : Shimkets LJ, editor. Prokaryotic Developm ent. W ashington D C : A m erican Society for M ic ro b io l- ogy. pp. 83-104.

137. D o n ia M S , R ave l J , Schm idt E W (2008) A global assembly line for cyanobactins. N a t C h e m B iol 4: 341-343.

138. Sivonen K , Leikoski N , Fewer D P , Jokela J (2010) Cyanobactins-ribosom al cyclic peptides produced by cyanobacteria. A p p l M ic ro b io l B iotechnol 86: 1213-1225.

139. Philmus B, Christiansen G , Yoshida W Y , Hem scheid t T K (2008) Post- translational m odification in m icrov irid in biosynthesis. Chem biochem 9: 3066-3073.

140. Z iem e rt N , Ishida K , L ia im e r A , H ertw eck C , D ittm an n E (2008) Ribosom al synthesis o f tricyclic depsipeptides in b loom -form ing cyanobacteria. Angew C h e m In t E d Engl 47: 7756-7759.

141. L i B, Sher D , K e lly L , Shi Y , H u a n g K et al. (2010) Catalytic prom iscuity in the biosynthesis o f cyclic peptide secondary metabolites in planktonic m arine cyanobacteria. Proc N a d A cad Sei U S A 107: 10430-10435.

142. W a n g H , Fewer D P , Sivonen K (2011) G enom e m in ing demonstrates the widespread occurrence o f gene clusters encoding bacteriocins in cyanobacteria. PLoS O n e 6: e22384.

143. Balskus E P , W alsh C T (2008) Investigating the in itia l steps in the biosynthesis o f cyanobacterial sunscreen scytonemin. J A m C h e m Soc 130: 15260-15261.

144. Balskus EP, W alsh C T (2010) T h e genetic and m olecular basis for sunscreen biosynthesis in cyanobacteria. Science 329: 1653-1656.

145. G ao Q , G arcia-P ichel F (2010) A n A TP -grasp ligase involved in the last biosynthetic step o f the im inom ycosporine shinorine in Nostoc punctiforme A T C C 29133. J Bacteriol 193: 5923-5928.

146. Balskus E P , Case RJ, W alsh C T (2011) T h e biosynthesis o f cyanobacterial sunscreen scytonem in in in te rtid a l m icrob ia l m a t com m unities. F E M S M icro b io l Ecol 77: 322-332.

147. Jones A C , M o n ro e E A , Podell S, Hess W R , Klages S et al. (2011) G enom ic insights in to the physio logy a nd ecology o f the m a rin e filam entous cyanobacterium Lyngbya majuscula. Proc N a d Acad Sei U S A 108: 8815-8820.

148. Fu jii K , Sivonen K , Adachi K , N oguchi K , Shim izu Y et al. (1997b) C om parative study o f toxic and non-toxic cyanobacterial products: a novel glycoside, suomilide, from non-toxic Nodularia spumigena H K W . Tetrahedron Lett 38: 5529-5532.

149. M enze lla H G , R e id R , C arney J R , C handran SS, Reisinger SJ et al. (2005) C om binato ria l polyketide biosynthesis by de novo design and rearrangem ent o f m odular polyketide synthase genes. N a t B iotechnol 23: 1171-1176.

150. Arnison P G , Bibb M J , B ierbaum G , Bowers A A , Bugni T S et al. (2013) Ribosom ally synthesized and post-translationally m odified peptide natural products: overview and recom m endations for a universal nom enclature. N a t Prod R ep 30: 108-160.

151. G oldberg S M , Johnso nJ, Busam D , Feldblyum T , Ferriera S et al. (2006) A Sanger/pyrosequencing hybrid approach for the generation o f high-quality draft assemblies o f m arine m icrob ia l genomes. Proc N a d A cad Sei U S A 103: 11240-11245.

152. C hevreux B, W e tte r T , Suhai S. G enom e sequence assembly using trace signals and additional sequence inform ation. C om put. Sei. Biol.; 1999. pp. 45-56.

153. Pop M , Kosack D S , Salzberg SL (2004) H ie rarch ica l scaffolding w ith Bambus. G enom e Res 14: 149-159.

154. A z iz R K , Bartels D , Best A A , D e jo n g h M , D isz T et al. (2008) T h e R A S T Server: rap id annotations using subsystems technology. B M C Genomics 9: 75.

155. H o ffm an n S, O tto C , K u rtz S, Sharm a C M , K ha itov ich P et al. (2009) Fast m apping o f short sequences w ith mismatches, insertions and deletions using index structures. PLoS C o m p u t B io l 5: e l0 0 0 5 0 2 .

156. Bachm ann B O , R avel J (2009) M ethods fo r in silica prediction o f m icrobial polyketide and nonribosom al peptide biosynthetic pathways from D N A sequence data. M ethods E nzym ol 458: 181-217.

157. R o ttig M , M e d e m a M H , B lin K , W e b e r T , Rausch C et al. (2011) N R P S p red ic to r2—a web server for predicting N R P S adénylation dom ain specificity. N ucleic Acids Res 39: W 3 6 2 -3 6 7 .

158. T a m u ra K , Peterson D , Peterson N , Stecher G , N e i M et al. (2011) M E G A 5 : m olecular evolutionary genetics analysis using m axim um likelihood, evolution­ary distance, and m axim um parsim ony methods. M o l B io l E vo l 28: 2731-2739.

159. G aud in C , Zhou X , W illiam s K P , Felden B (2002) Tw o-p iece tm R N A in cyanobacteria and its structural analysis. N ucleic Acids Res 30: 2018-2024.

160. M a o C , Bhardw aj K , Sharkady S M , Fish R I , D riscoll T et al. (2009) Variadons on the tm R N A gene. R N A B iol 6: 355-361.

161. A x m an n I M , H o ltze n d o rffJ , Voss B, Kensche P, Hess W R (2007) T w o disdnct types o f 6S R N A in Prochlorococcus. G ene 406: 69-78.

162. G ierga G , Voss B, Hess W R (2009) T h e Y fr2 n c R N A fam ily , a group o f abundant R N A molecules w idely conserved in cyanobacteria. R N A B iol 6: 222-227.

163. Kaneko T , N a ka m u ra Y , W o lk G P, K u ritz T , Sasamoto S et al. (2001) C om plete genomic sequence o f the filam entous nitrogen-fixing cyanobacterium Anabaena sp. strain P G G 7120. D N A Res 8: 205-213.

164. Nicolaisen K , M oslavac S, Samborski A , V aldeben ito M , H an tke K et al. (2008) A lr0 3 9 7 is an outer m em brane transporter for the siderophore schizokinen in Anabaena sp. strain P C C 7120. J Bacteriol 190: 7500-7507.

PLOS ONE I w w w .p loso n e .o rg 22 M arch 2013 | V o lu m e 8 | Issue 3 | e60224