bs_bs_banner

Environmental Microbiology (2015)

doi:10.1111/1462-2920.12907

Single-cell genomics of uncultivated deep-branching magnetotactic bacteria reveals a conserved set of magnetosome genes

Sebastian Kolinko,1*† Michael Richter,2 Frank-Oliver Glöckner,2,3 Andreas Brachmann1 and Dirk Schüler1,4** 1 Department of Biology I, LMU Biozentrum, Ludwig-Maximilians University Munich, Großhaderner Str. 2-4, Planegg-Martinsried 82152, Germany. 2 Microbial Genomics and Bioinformatics Research Group, Max Planck Institute for Marine Microbiology, Celsiusstr. 1, Bremen 28359, Germany. 3 Department of Life Sciences & Chemistry, Jacobs University Bremen, Campus Ring 1, Bremen 28759, Germany. 4 Department of Microbiology, University Bayreuth, Bayreuth, Germany. Summary While magnetosome biosynthesis within the magnetotactic Proteobacteria is increasingly well understood, much less is known about the genetic control within deep-branching phyla, which have a unique ultrastructure and biosynthesize up to several hundreds of bullet-shaped magnetite magnetosomes arranged in multiple bundles of chains, but have no cultured representatives. Recent metagenomic analysis identified magnetosome genes in the genus ‘Candidatus Magnetobacterium’ homologous to those in Proteobacteria. However, metagenomic analysis has been limited to highly abundant members of the community, and therefore only little is known about the magnetosome biosynthesis, ecophysiology and metabolic capacity in deep-branching MTB. Here we report the analysis of single-cell derived draft genomes of three deep-branching uncultivated MTB. Single-cell sorting followed by whole genome amplification generated draft genomes of Candidatus Magnetobacterium bavaricum and Candidatus Received 1 February, 2015; revised 10 May, 2015; accepted 14 May, 2015. For correspondence. *E-mail [email protected]; Tel. +4989218074514; Fax +4989218074515. **E-mail dirk.schueler@ uni-bayreuth.de; Tel. +49921552729; Fax +49921552727. †Present address: Physical Biosciences Division, Lawrence Berkeley National Laboratory, Joint BioEnergy Institute (JBEI), 1 Cyclotron Road, Berkeley, CA, USA.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd

Magnetoovum chiemensis CS-04 of the Nitrospirae phylum. Furthermore, we present the first, nearly complete draft genome of a magnetotactic representative from the candidate phylum Omnitrophica, tentatively named Candidatus Omnitrophus magneticus SKK-01. Besides key metabolic features consistent with a common chemolithoautotrophic lifestyle, we identified numerous, partly novel genes most likely involved in magnetosome biosynthesis of bullet-shaped magnetosomes and their arrangement in multiple bundles of chains. Introduction Magnetotactic bacteria (MTB) are a phylogenetically and morphologically diverse group of prokaryotes that biosynthesize unique membrane-enveloped nano-sized ferromagnetic crystals (Bazylinski and Frankel, 2004; Bazylinski et al., 2013). By arranging these so-called magnetosomes into a chain-like structure, each cell is able to align to the earth’s geomagnetic field. This facilitates their navigation within chemically stratified freshwater and marine sediments (Frankel et al., 1997; Bazylinski and Frankel, 2004). Magnetosomes exhibit intricate structures that result from a complex, stepwise biosynthesis, including vesicle formation of special intracellular membrane vesicles, iron transport and crystallization, and magnetosome alignment along a cytoskeletal filament (Bazylinski and Frankel, 2004; Jogler and Schüler, 2009). The genetic determinants of magnetite magnetosome biosynthesis have been first discovered in the Alphaproteobacterium Magnetospirillum gryphiswaldense (Grünberg et al., 2001; Jogler and Schüler, 2009), in which most of the > 30 magnetosome genes are clustered in several operons within a large genomic magnetosome island (MAI) (Schübbe et al., 2003; Ullrich et al., 2005; Richter et al., 2007). Many magnetosome genes are involved in iron uptake, redox control and chain alignment, whereas the functions of many others are still unknown. The MAI-like clusters were also identified in cultured MTB of the alpha, gamma and delta lineages of the Proteobacteria (Bazylinski et al., 2013; Lefèvre and Bazylinski, 2013). Despite gross conservation, they display variability between different species with respect

2

S. Kolinko et al.

to molecular organization and gene content (Jogler et al., 2009a; Abreu et al., 2011; Lefèvre and Bazylinski, 2013; Lefèvre et al., 2013; Kolinko et al., 2014b). However, most MTB, which are abundant in natural environments, cannot be cultivated due their unknown lifestyle, nutritional requirements and adaptation to chemically stratified aquatic sediments. Nevertheless, previous approaches utilized the active magnetotactic swimming motility of uncultivated MTB by which they can be collected from environmental samples and separated from contaminating organisms (Jogler et al., 2009b; 2011). This revealed a high morphological variability of magnetosomes including hexa- and cubo-octahedral, elongated-prismatic and bullet shapes of ferromagnetic crystals consisting of either the magnetic minerals magnetite (Fe3O4) and/or greigite (Fe3S4) (Bazylinski et al., 2013). This great diversity is matched also by an enormous genetic diversity, exceeding that found in cultivated species and indicating the existence of further and novel potential biosynthesis pathways of magnetosome formation (Bazylinski and Frankel, 2004; Jogler et al., 2009a,b; 2010; 2011). Candidatus Magnetobacterium bavaricum, one of the most intriguing uncultivated bacteria, is a giant rod from the Nitrospirae phylum that was originally discovered due to its conspicuous morphology by electron microscopy and rDNA sequencing (Vali et al., 1987; Spring et al., 1993). The ubiquitous distribution and high contribution to the entire microbial biomass in the habitat (up to 30%) of Ca. M. bavaricum and closely related MTB points to a great potential for iron cycling in natural environments (Spring et al., 1993; Lin et al., 2014). In addition, these bacteria exhibit a highly unusual cell morphology and ultrastructure, for example displaying a multilayered cell wall and up to 1000 bullet-shaped magnetite magnetosomes arranged into multiple bundles of chains, indicating a distinct and potentially more complex biosynthesis than in cultivated MTB from the Proteobacteria (Jogler et al., 2011). Despite considerable efforts for several decades, it proved recalcitrant to any attempts to isolate and grow it axenically in the laboratory. However, analysis of metagenomic libraries from magnetic enrichments revealed first insights into the general metabolism of Ca. M. bavaricum (Jogler et al., 2010). In addition, first magnetosome genes with homology to those in cultivated MTB were identified, which were found colocalized with novel and unknown genes within the incomplete sequence of a MAI-like gene cluster (Jogler et al., 2011). Shortly before completion of this study, the almost complete genome sequence of the close relative Candidatus Magnetobacterium casensis was determined by a metagenomic approach, which revealed additional information, such as an apparently extended MAI (Lin et al., 2014). However, metagenomic approaches of uncultivated MTB are limited only to the most abundant members of

natural communities. This limitation has been overcome by targeted analysis of single MTB cells (Kolinko et al., 2012; 2013). For example, by combining microsorting of individual cell aggregates with whole genome amplification (WGA), a draft genome sequence of an uncultivated magnetotactic multicellular prokaryote (MMP) from a North Sea mudflat (Germany) was de novo assembled, revealing the presence of magnetite and greigite magnetosome gene clusters (Kolinko et al., 2014b). A similar approach recently also identified new and unexpected microbial diversity that has been inaccessible by metagenomic techniques (Kolinko et al., 2012; 2013). Besides new magnetotactic representatives of the Nitrospirae phylum, a novel ovoid MTB of the candidate phylum Omnitrophica (former candidate division OP3), has been described (Kolinko et al., 2012). However, these studies were limited to phylogenetic and ultrastructural analysis, and no information concerning magnetosome biosynthesis or metabolism was revealed. Here we used a targeted genomic analysis of different conspicuous magnetotactic morphotypes by magnetic collection followed by micromanipulation and WGA to assemble three draft genomes of three MTB, including the giant rod Ca. M. bavaricum, the ovoid Ca. Magnetoovum chiemensis of the Nitrospirae phylum and the ovoid Candidatus Omnitrophus magneticus, the first draft genome of a magnetotactic representative of the candidate phylum Omnitrophica that has been recently discovered in sediments of Lake Chiemsee (Kolinko et al., 2012; 2013). Analysis of retrieved draft genomes revealed a similar chemolithoautotrophic lifestyle shared by all MTB of the deep-branching phyla, illuminating the metabolism and physiology of these so far poorly understood MTB. In addition, we discovered novel genetic determinants of magnetosome biosynthesis, including MAI-like gene clusters of both Nitrospirae and the first magnetosome genes from a representative of the candidate division Omnitrophica that are highly similar to those in the Nitrospirae. Phylogenetic analysis of magnetosome genes of deep-branching MTB showed a divergent evolution compared with 16S rDNA sequences, indicating an ancient event of horizontal gene transfer. Results Microcosms set up with mud from Lake Chiemsee contained rich populations of MTB with a variety of morphotypes, including spirilla, rods and cocci, similar as observed before (Kolinko et al., 2012; 2013). Among MTB that could be magnetically collected, we occasionally observed (< 1% abundance) a conspicuous ovoid ‘melon’shaped magnetic responsive morphotype, virtually identical to the previously described uncultivated magnetotactic bacterium SKK-01 (Kolinko et al., 2012). In addition, large

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Genomic analysis of an uncultivated multicellular magnetotactic prokaryote

3

Table 1. General genomic features of the draft genomes of Ca. Omnitrophus magneticus SKK1, Ca. Magnetoovum chiemensis CS-04, Ca. Magnetobacterium bavaricum and for comparison Ca. Magnetobacterium casensis (Lin et al., 2014).

Parameter

Candidatus Magnetovum chiemensis CS-04

Candidatus Magnetobacterium bavaricum

Candidatus Omnitrophus magneticus SKK01

Candidatus Magnetomorum HK-1

Candidatus Magnetobacterium casensis

Total genome size (Mb) GC content (%) Number of contigs N50 (kb) Maximum contig length (kb) Genome completeness (%) Estimated genome size (Mb) Number of coding sequences (CDS) Number of tRNAs Number of copies of rRNA operon Total MAI size (kb) GC content MAI (%) Number of putative magnetosome genes

3.9 40.4 1402 9.651 43.149 74.1 5.3 4527 45 1 27.9 39.5 21

6.3 47.40 4211 4.571 67.959 74.8 8.4 8277 69 1 46 50.20 30

3.4 35.80 1120 15.846 133.939 87.1 3.9 3439 42 1 4.8 38.10 6

14.7 34.7 3197 17.5 122.255 95.1 15.5 14 451 49 1 50.8 38.2 51

3.4 48.90 70 90.253 239.626 N.A. 3.4 3140 40 1 39.2 49.80 31

magnetotactic rods virtually identical to the previously identified Candidatus Magnetobacterium bavaricum (Spring et al., 1993; Kolinko et al., 2013) were abundant in the same microcosm. A further microcosm contained a large magnetotactic coccus morphologically resembling the previously described uncultivated MTB CS-04, which was also first identified by single-cell techniques (Kolinko et al., 2013), a similar organism described recently (Lin et al., 2012), and Candidatus Magnetoovum mohaviensis strain LO-1 (Lefèvre et al., 2011). Cells with similarity to SKK-01, Ca. M. bavaricum and CS-04 were targeted by micromanipulation for single-cell analysis and the extracted DNA amplified into individual single amplified genomes (SAGs) as previously described (Woyke et al., 2011; Kolinko et al., 2012; 2013; 2014b). Only SAGs yielding a single partial 16S rDNA sequence, indicating clonality and the absence of contaminations, qualified for individual sequencing and de novo genome assembly. De novo assemblies were validated by BLAST (Altschul et al., 1997), tetra-nucleotide frequency and single copy marker gene analysis (Rinke et al., 2013). The size of obtained SAGs ranged between 0.6 Mb and 3.7 Mb (2.3 ± 0.9 Mb) composed of an average of 2100 ± 1221 contigs, with an average size of 1327 ± 657 bp. The estimated genome completeness of SAGs varied between 7.3% and 60% (44.5 ± 0.2%) (Fig. S1, Table S1). The SAGs contained primarily small contigs, indicated by an N50 value < 3 kb (2.9 ± 2.1 kb) and an average contig size < 1.5 kb, which prevented the identification of large contiguous magnetosome gene clusters (Table 1). Each SAG contained a full rDNA operon with the standard 16S, 23S-5S gene arrangement, and the sequences were identical to those of the morphologically similar strains SKK-01, Ca. M. bavaricum and CS-04 respectively. SKK-01 assigned to the candidate division

OP3 (Hugenholtz et al., 1998; Kolinko et al., 2012), which belongs to the Planctomycetes-VerrucomicrobiaChlamydiae (PVC) superphylum and was recently renamed into candidate phylum Omnitrophica (Wagner and Horn, 2006; Rinke et al., 2013). A described conserved PVC signature gene of unknown function encoding a hypothetical protein (Gupta et al., 2012) (NP_219933) was detected in individual SAGs. In the following, we refer to this MTB as Candidatus Omnitrophus magneticus strain SKK-01. The large magnetotactic coccus CS-04 belongs, like Ca. M. bavaricum, to the deep-branching phylum Nitrospirae and was first identified by single-cell techniques from sediments of Lake Chiemsee (Kolinko et al., 2013). Due to its similarity to Ca. M. mohaviensis and its habitat, we refer to this MTB as Candidatus Magnetoovum chiemensis strain CS-04 in the following. Combining the individual SAGs of Ca. O. magneticus SKK-01, Ca. M. bavaricum and Ca. M. chiemensis CS-04 (6, 6 and 4 SAGs, respectively) significantly increased the draft genome sizes (3.9, 6.3 and 3.4 Mb, respectively) and estimated genome completeness (74%, 75% and 87%, respectively), facilitating automatic gene annotation and magnetosome gene cluster detection (Fig. S1, Table 1, Table S1). Genomic reconstruction of general key metabolic features Besides their ability to form magnetosomes (for detailed analysis, see below), the three analysed MTB share several common characteristics, such as a respiratory metabolism and the ability of autotrophy. In addition, functional analysis of their draft genomes revealed a range of metabolic differences and peculiarities between organisms, some examples of which will be described below.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

4

S. Kolinko et al.

Candidatus Magnetobacterium bavaricum A versatile metabolism was predicted by the draft genome of Ca. M. bavaricum, which putatively can utilize both oxidized nitrogen and sulfur compounds, but not oxygen as electron acceptors. Complete respiratory pathways for either denitrification (DN) or dissimilatory nitrate reduction to ammonia (DNRA) (Dong et al., 2011) were reconstructed. Under nitrogen-limiting conditions, the DNRA pathway is used to reduce cytoplasmic nitrate to nitrite by a membrane-bound nitrate reductase (Nar) (SAG4_002195–SAG4_002187) and further reduced by a cytoplasmic nitrite reductase (NirB) (SAG4_006891) to ammonia. For DN, periplasmic nitrate is successively reduced to its gaseous products nitrous oxide (N2O) and N2 by the nitrate reductase (Nap) (SAG4_006956), nitrite reductase (NirS) (SAG4_007402), nitric oxide reductase (FprA) (SAG_000686) and nitrous oxide reductase (Nos) (SAG4_007532), which are encoded in the draft genome. The utilization of nitrate or nitrite as nitrogen source by assimilatory nitrate reduction is indicated by the presence of an assimilatory nitrate reductase (SAG4_001651). In contrast to Ca. M. casensis, no genes involved in nitrogen fixation were identified in Ca. M. bavaricum (Lin et al., 2014). Sulfur globule accumulation and consumption under oxic conditions in the cytoplasm was previously hypothesized to indicate a putative sulfur-oxidizing metabolism (Jogler et al., 2010). Whereas no genes encoding sulfur oxidation proteins (Sox) were found, a gene coding for a putative sulfide : quinone oxidoreductase (SAG4_006205) was identified, which is postulated to oxidize sulfide to elemental sulfur (Dahl et al., 2005). Also genes are present coding for a dissimilatory sulfite reductase (dsr) (SAG4_005384, SAG4_000906), which further oxidizes sulfur to sulfite. The oxidation of sulfite to sulfate is proposed to be mediated by an adenosine 5′-phosphosulfate (APS) reductase (Apr) and adenosine 5′-triphosphate (ATP) sulfurylase (Sat) (SAG4_003142SAG4_003143). Additionally, an assimilatory sulfate reducing pathway with cysteine is present in Ca. M. bavaricum. The previously identified Ribulose-1,5-bisphosphatcarboxylase (RuBisCo), like protein of Ca. M. bavaricum, was found to be encoded in the draft genome (SAG4_004452) (Jogler et al., 2010). Identical to the closely related Ca. M. casensis (Lin et al., 2014), we identified genes coding for CO2 fixation via the reductive acetyl-CoA pathway (Wood–Ljungdahl pathway) and the reductive citric acid cycle in Ca. M. bavaricum, indicative for autotrophic growth. Since intracellular biomineralization of up to 1000 magnetite (Fe3O4) crystals requires the accumulation of large amounts of iron, metabolic pathways for its uptake and

handling, as well as their regulation, are of particular interest. We identified an iron transport operon with two adjacent genes coding for a ferrous iron transporter protein A (FeoA), followed downstream by a gene encoding FeoB (SAG4_005787-SAG4-005789). These proteins had highest similarity to the magnetosome proteins Mad17 and Mad30 previously identified in Ca. M. casensis (Lin et al., 2014). Besides a gene (SAG4_006900) coding for a general transporter for divalent metal ions of the NRAMP transporter family, we found genes with considerable similarities to TonBdependent transporters. Additionally, we identified the coding potential for bacterial ferritins (KJU83420), which are responsible for intracellular iron homeostasis. A phosphate-rich ferric hydroxide phase, similar to bacterial ferritins (SAG4_005573), is hypothesized to be a precursor for magnetite mineralization (Baumgartner et al., 2013). Previous ultrastructural analyses revealed a highly unusual cell wall structure in Ca. M. bavaricum, including a multilayered cell boundary composing a cytoplasmic membrane, an outer membrane, one layer of peptidoglycan, and a bipartite outer layer resembling a capsular structure, which was predicted to consist of polysaccharides (Jogler et al., 2011). Besides genes involved in outer membrane and peptidoglycan biogenesis as in the two draft genomes, Ca. M. bavaricum indeed contains two sets of genes encoding the biosynthesis of putative polysaccharides (spsBC2E2F2G2) (SAG4_007242 – SAG4_007262), similar to those in the Gram-negative soil bacterium Rhizobium leguminosarum, which was shown to produce large amounts of exopolysaccharide important for successful nitrogen-fixing symbiosis with legume plants (Bonomi et al., 2012; Marczak et al., 2013). The presence of a polysaccharide layer could also be the reason for the observed poor chemical lysis of Ca. M. bavaricum (Jogler et al., 2011), which caused its strong underrepresentation in previous metagenomic cloning approaches despite its high numeric dominance (Jogler et al., 2011). Candidatus Magnetoovum chiemensis CS-04 In the Ca. M. chinensis CS-04 draft genome, we identified genes coding for the DNRA pathway, including a (membrane-bound) nitrate reductase (Nar) (SAG2_004472-SAG2_002275) and (cytoplasic) nitrite reductase. In contrast to Ca. M. bavaricum, neither genes for a DN nor for an assimilatory nitrate reducing pathway were found, but most genes were involved in nitrogen fixation, like a nif-operon (nifHDKNXHD1HD2BF,nifSU) and a gene encoding a Nif-specific ferredoxin III (SAG2_003423 – SAG2_003431, SAG2_000687, and SAG2_002127).

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Genomic analysis of an uncultivated multicellular magnetotactic prokaryote Consistent with the presence of sulfur-rich inclusions (Kolinko et al., 2013), genes encoding a sulfur oxidation pathway were identified in part with a putative sulfide : quinone oxidoreductase (SAG2_003639), a dissimilatory sulfite reductase Sat (SAG2_001606, SAG2_004249) and Apr (SAG2_00042 – SAG2_000044). Genes coding for all proteins necessary for CO2 fixation via the Wood–Ljungdahl pathway imply the potential for autotrophy for Ca. M. chiemensis. As in Ca. M. bavaricum, genes coding for Mad17, Mad30, a general metal2+ transporter (NRAMP), TonB-dependent transporters and bacterial ferritin were also present in Ca. M. chiemensis. Additionally, we found genes for a ferric uptake regulator (Fur) (SAG2_003095) and the ferrous iron transporter FeoB (SAG2_004218) on discontinuous contigs. Candidatus Omnitrophus magneticus SKK-01 In contrast to the two Nitrospirae genomes described in this work, no indications of a respiratory pathway with nitrogen compounds were found in SKK-01. However, genes governing nitrogen fixation were identified coding for NifBHD1HD2SU (SAG01_000292, SAG01_000293, SAG01_000295, SAG01_000231, SAG01_001446) and a Nif-specific ferredoxin III (SAG01_000294) encoded within an operon together with nif genes. The identification of sulfur-rich inclusions in Ca. O. magneticus previously led to the hypothesis of a sulfur-oxidizing metabolism (Kolinko et al., 2012). In fact, a reversible dissulfoviridintype dissimilatory sulfite reductase (dSir) (SAG1_002184) was found in the draft genome. Under reducing conditions, dSiR transfers six electrons from the hydrolysation of H2 by a periplasmic and a membrane-bound hydrogenase (SAG1_001262 – SAG1_001265) to sulfite, reducing it to sulfide. Sulfur-oxidizing bacteria, such as Thiobacillus, oxidize sulfur components by a reverse reaction of dSiR (Steuber and Kroneck, 1998), which could be the case for Ca. O. magneticus, too. Instead of a Wood–Ljungdahl pathway for CO2 fixation as in Ca. M. bavaricum and Ca. M. chiemensis, we identified two key genes coding for a fumarate reductase (SAG1_001169) and a 2-oxoglutarate ferredoxin oxidoreductase (SAG1_002145), implying CO2 fixation through the reductive citric acid cycle. Unexpectedly, we failed to identify any iron uptake or regulation systems, apart from an iron-dependent repressor (SAG1_003364) and a TonB-dependent receptor plug (SAG1_000510). Magnetosome biosynthesis Contigs encoding magnetosome proteins in all draft genomes differed only slightly with respect to G + C content to their respective core genomes (Table 1).

5

Whereas the MAI of Ca. M. bavaricum had a G + C content of 50%, those of Ca. M. chiemensis CS-04 and Ca. O. magneticus SKK-01 had ∼ 40%. One contig (23 385 bp) in the draft genome of Ca. M. bavaricum contained a mamAB-like operon, which is essentially identical (> 99%) to that found on a previous metagenomic clone of Ca. M. bavaricum (Jogler et al., 2011), and had a similar gene order and high nucleotide identity (91.2%) to the mamAB-like operon in Ca. M. casensis (Lin et al., 2014). Magnetosome genes mamK, mad11 and mad28, which are part of the mamAB-like operon in Ca. M. casensis, were found on a short accessory contig (Fig. 1). In addition to mamPMQBAIEQO-Cter, man5, man6, mad2, mad20 and mad23-26, which were previously identified in Ca. M. bavaricum, we found a mamK and homologues to mad11 and mad28 with high identities to Ca. M. bavaricum (> 98%) and to Ca. M. casensis (> 80%) (Table 2). A partial mamAB-like operon is also present in the draft genome of Ca. M. chiemensis CS-04 encoding magnetosome proteins MamEMQBAPM and Mad23-26 with an identity of approximately 50% to other magnetotactic Nitrospirae and Deltaproteobacteria (Table 3). Our failure to assemble magnetosome gene contigs > 1.5 kb in Ca. O. magneticus SKK-01 prevented the analysis of gene order within the putatively larger clusters (Fig. 1). Encoded magnetosome proteins MamEKBM and Mad11 have a high identity of approximately 60% to their orthologues in magnetotactic Nitrospirae (Table 4). Furthermore, Mam and Mad proteins of Ca. O. magneticus SKK-01 seem to be closely related to those of Ca. M. chiemensis CS-04 within the deep-branching Nitrospirae (Figs 1 and 2). Additionally, we identified several discontinuous contigs with magnetosome genes putatively involved in magnetosome biosynthesis outside the mamAB-like operons of Ca. M. bavaricum and Ca. M. casensis. The genomes of both Ca. M. bavaricum and Ca. M. chiemensis CS-04 contain a gene encoding a further Mad28 homologue (Mad28-2) (KJR44079, KJU83011), which is accompanied downstream by a highly conserved gene (KJR44078). This gene encodes a putative penicillin-binding protein (PBP) with a multimodular transpeptidase-transglycosylase activity. A homologous magnetosome gene tandem was also found outside of the described mamAB-like operon of Ca. M. casensis, and is therefore likely to be conserved in all magnetotactic Nitrospirae (Fig. 1, Tables 2 and 3). Another contig (4 kb) of Ca. M. bavaricum contained five genes organized in an operon. Two of them code for actin-like proteins Mad28-3 (KJU86277) and Mad28-4 (KJU86282), and one encodes a putative ATPase Mad29 (KJU86280). This operon was, therefore, termed

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

mad26

mad26

mad26

mad25

mad25

mad25

mad24

mad24

mad24

mamOCter

mad genes

man genes

mamK mamP

mad11

Conserved hypothetical genes

Hypothetic genes

mad30

5 kb

mad17-2

mad28-5 mad17-1

mad29-2

Already published magnetosome genes

Candidatus Magnetoovum chiemensis CS-04

mad17-1

Candidatus Magnetobacterium casensis

mad17-2 mad30

mad28-4

mamB-2

mad28-3

mad28-2

mamM

mad31

mad23-2

mamM mamQ-2

mamE

Fig. 1. Molecular organization of identified magnetosome genes in Ca. Magnetobacterium bavaricum, Ca. Magnetoovum chiemensis and Ca. Omnitrophus magneticus compared with the Nitrospirae MTB Ca. Magnetobacterium casensis. Bars between clusters indicate homologous genes. Magnetosome genes published in previous publications are indicated by yellow.

mamOCter

mad23

mad23-1

mad23

mam genes

mad11 mad11

Candidatus Magnetobacterium bavaricum

mamA

man6

man6

mad2

man5

man5

mamB-1

mamQ-1 mamM

mamI mamI

mamA

mamA

mad2

mad2

mad20 man4 mamQ-2

mamQ-1 mad31

mad28-1 mad28-1

mamB mamB

mamE mamE mamE

mamM mad11

mamP man3 mad10

mamB mamP

mamQ-2 mamM

mad31 mamK

man2

mad28-1

mad29-1 mad28-3

mamK mamK mad29

mad28-2 mad17

mad29 mad28-2

Candidatus Omnitrophus magneticus SKK-01

mad30

6 S. Kolinko et al.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Accession number

KJU82680 KJU82741 KJU82967 KJU82968 KJU83011 KJU83257 KJU83258 KJU83259 KJU83601 KJU83602 KJU83603 KJU84496 KJU84833 KJU84834 KJU84835 KJU84836 KJU84838 KJU84839 KJU84840 KJU84843 KJU84845 KJU84846 KJU84847 KJU84848 KJU84849 KJU84850 KJU84851 KJU86277 KJU86280 KJU86282 KJU86423 KJU86492

Putative protein

Mad29-2 MamI Mad11 Mad28-1 Mad28-2 Mad30 Mad17-1 Mad17-2 MamO Cter MamQ-1 Man4 Mad28-6 Mad26 Mad25 Mad24 Mad23 Man6 Man5 Cter Man5 Nter MamE MamA Mad2 MamB MamQ-2 Mad31 MamM MamP Mad28-3 Mad29-1 Mad28-4 MamK Mad28-5

Mcas Mbav Mcas Mcas MLMS-1 Mcas Movum Mcas Mbav Mbav Mbav Mcas Mbav Mbav Mbav Mbav Mbav Mbav Mbav Mbav Mbav Mbav Mbav Mbav Mbav Mbav Mbav Mcas Movum Mcas Mbav Mcas

JMFO00000000.1 FQ377626 AIM41329 AIM41330 WP_040867321 KJR44073 KJR44074 JMFO00000000.1 FQ377626 FQ377626 FQ377626 AIM41330 FQ377626 FQ377626 FQ377626 FQ377626 FQ377626 FQ377626 FQ377626 FQ377626 FQ377626 KJU86753 FQ377626 FQ377626 FQ377626 FQ377626 FQ377626 AIM41330 KJR44077 AIM41330 AFX60119 AIM41330

Bacterium with protein with highest sequence identity Accession 89.67 100.00 61.29 80.52 58.56 53.44 53.81 88.26 100.00 100.00 100.00 70.59 93.01 100.00 100.00 100.00 100.00 100.00 100.00 99.45 98.77 100.00 99.63 99.45 100.00 100.00 98.54 91.40 47.65 91.72 97.48 92.68

0.00E+00 7.27E-42 4.69E-66 0.00E+00 1.81E-128 2.22E-87 0.00E+00 0.00E+00 3.21E-170 7.88E-126 1.07E-80 4.23E-52 8.45E-112 1.60E-148 1.49E-122 7.18E-32 0.00E+00 2.70E-62 4.50E-169 4.23E-118 3.93E-55 3.48E-27 3.33E-180 9.35E-136 2.65E-127 0.00E+00 0.00E+00 0.00E+00 1.76E-85 0.00E+00 1.10E-78 0.00E+00

Identity (%) E-value Movum Mcas Movum MLMS-1 RS-1 BW-1 BW-1 Mcas Mcas Mcas Mcas MLMS-1 Mcas Mcas Mcas Mcas Mbav Mbav Mbav Mcas Mcas Mcas Mcas Mcas Mcas Mcas Mcas MLMS-1 BW-1 Mcas Mcas MLMS-1

2.76E-84 3.76E-45 9.05E-15 1.32E-148 1.69E-126 5.09E-71 0.00E+00 0.00E+00 4.54E-145 2.32E-151 6.47E-77 8.78E-20 1.53E-106 3.13E-126 3.71E-118 5.83E-173 0.00E+00 1.07E-63 2.59E-19 0.00E+00 7.00E-119 3.23E-52 3.07E-170 1.26E-116 5.38E-124 6.01E-168 0.00E+00 3.98E-141 1.76E-85 0.00E+00 3.24E-109 3.98E-141

Identity (%) E-value

KJR44077 46.98 93.67 AIM41316 KJR44081 27.23 WP_040867321 63.80 WP_013162962 58.56 45.02 AGG16188 52.11 AET24920 JMFO00000000.1 79.69 88.72 AIM41315 92.92 AIM41314 91.10 AIM41313 WP_040867321 61.76 88.83 AIM41305 80.83 AIM41306 87.75 AIM41307 84.42 AIM41308 KJU85628 100.00 KJU85629 99.04 KJU85629 97.13 88.47 AIM41315 92.97 AIM41317 78.38 AIM41318 92.88 AIM41319 82.56 AIM41320 87.22 AIM41321 95.19 AIM41322 81.00 AIM41323 WP_007290995 66.45 47.65 CCO06735 JMFO00000000.1 91.40 96.82 AIM41328 WP_007290995 66.45

Bacterium with protein with second highest sequence identity Accession CCO06735 AHG23891 AFZ77031 KJR44079 AGG16228 ADV17384 ADV17385 KJR44074 KJR42514

WP_013162962 KJR42127 KJR42126 ADV17380 BAH77580 AIM41311 AIM41312 AIM41312 KJJ83152 KJR43885 KJR43883 AIM41320 KJR43881 WP_024080586 KJR WP_013162962 BAH77573 WP_040867321 ADV17375 AGG16228

D. alkaliphilus Movum Movum MMP RS-1 Mcas Mcas Mcas Omag Movum Movum Movum Movum MSR-1 Movum D. alkaliphilus RS-1 MLMS-1 MMP FH-1

Accession

RS-1 IT-1 ML-1 Movum FH-1 MMP MMP Movum Movum

Bacterium with protein with third highest sequence identity

66.78 37.30 38.10 35.77 57.22 65.50 38.72 66.45 63.92 56.73

2.11E-134 3.77E-37 2.91E-34 1.06E-40 7.61E-66 5.90E-130 8.90E-67 2.01E-132 1.48E-61 9.47E-121

4.60E-19 4.85E-19 7.92E-35 5.55E-10 1.53E-78 0.00E+00 3.25E-40 6.81E-142 1.89E-90 7.87E-63

3.49E-69 0.00E+00 0.00E+00 9.00E-36

50.88 52.59 52.80 38.98

63.64 31.55 41.21 27.39 38.25 82.98 73.08 85.02 57.52 55.25

3.93E-76 4.64E-11 7.19E-09 2.92E-135

41.89 43.59 35.57 50.12

Identity (%) E-value

Table 2. Comparison of Mam, Mad, Man and putative magnetosome proteins involved in magnetosome formation identified in the draft genome of Ca. M. bavaricum against draft genomes of Ca. Magnetoovum chiemensis CS-04, Ca. M. Casensis, Ca. Omnitrophus magneticus SKK-01 and the GenBank database.

Genomic analysis of an uncultivated multicellular magnetotactic prokaryote

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

7

MamP, partial MamM-2 Mad23 Mad24 Mad25 Mad26 Mad28-1 MamE MamM-1 Mad31 MamQ-II MamB-1 Mad2 MamA MamB-2 Mad30 Mad17 Mad29 Mad28-2 Mad11 MamK

KJU84851 AIM41322 KJU84836 ADV17380 AIM41306 AIM41305 KJU83011 KJJ83152 KJU84850 KJU84849 AIM41320 KJU84847 AIM41318 KJU84845 KIM00465 JMFO00000000.1 JMFO00000000.1 CCO06735 AIM41330 JYNY00000000 AIM41328

KJR41428 KJR41429 KJR42124 KJR42125 KJR42126 KJR42127 KJR43667 KJR43878 KJR43880 KJR43881 KJR43882 KJR43883 KJR43884 KJR43885 KJR44072 KJR44073 KJR44074 KJR44077 KJR44079 KJR44081 KJR44083

Putative protein

Mbav Mcas Mbav MMP Mcas Mcas Mbav Omag Mbav Mbav Mcas Mbav Mcas Mbav MS-1 Mcas Mcas BW-1 Mcas Omag Mcas

Bacterium with protein with highest Accession sequence number identity Accession 57.22 55.21 39.93 31.69 41.24 34.09 56.08 73.27 41.41 38.10 40.13 66.78 29.91 52.85 44.60 59.07 55.06 48.00 58.43 41.76 63.28

2.38E-66 1.73E-20 7.41E-51 7.02E-13 2.52E-33 1.10E-16 9.48E-132 1.78E-77 9.84E-24 8.56E-33 6.43E-30 2.23E-124 1.20E-11 2.13E-63 2.11E-77 1.68E-97 0.00E+00 1.85E-88 2.11E-136 4.34E-45 1.77E-72

Identity (%) E-value Mcas Mbav Mcas Mcas Mbav Mbav MLMS-1 Mcas Mcas Mcas Mbav Mcas ML-1 Mcas SO-1 Mbav Mbav Mbav Mbav Mcas Mbav AIM41323 KJU84850 JMFO00000000.1 AIM41307 KJU84834 KJU84833 WP_040867321 AIM41315 AIM41322 AIM41321 FQ377626 AIM41319 AFZ77013 AIM41317 WP_040477222 KJU83257 KJU83258 KJU86280 KJU82968 AIM41329 KJU82966

Bacterium with protein with second highest sequence identity Accession 57.22 54.17 45.61 28.17 40.68 32.58 56.85 53.73 41.41 37.98 38.75 66.07 33.61 51.30 44.25 55.13 53.81 47.65 58.43 28.79 62.71

4.95E-60 2.28E-20 1.03E-54 2.00E-08 2.52E-33 5.04E-17 4.90E-129 3.78E-90 1.34E-23 2.20E-32 6.18E-35 2.41E-110 2.19E-07 2.16E-62 1.21E-76 1.86E-91 0.00E+00 2.85E-91 6.24E-132 2.18E-11 2.10E-74

Identity (%) E-value

WP_041642553 AFX_88990 CCO06723 KJU84835 CCO06725 ADV17378 WP_015862704 KJU84843 WP_043601159 KJU84848 WP_011713872 CCO06674 CAV30807 AGG16216 AET24920 JMFO00000000.1 EAT05033 KJU82967 AFX88979

MC-1 SS-5 BW-1 Mbav BW-1 MMP RS-1 Mbav MV-1 Mbav MC-1 BW-1 MV-1 FH-1 BW-1 Mcas MLMS-1 Mbav ML-1

Bacterium with protein with third highest sequence identity Accession

42.16 45.02 53.69 46.28 55.56 25.79 56.50

37.50 32.53 32.52

44.00 47.19 39.78 27.46 36.99 33.33 49.42 51.29 33.77

8.33E-68 1.31E-61 0.00E+00 2.35E-89 5.28E-131 4.84E-14 1.14E-63

1.92E-29 3.55E-36 1.09E-06

1.00E-40 1.32E-20 1.42E-41 1.38E-07 8.72E-21 6.33E-08 1.44E-123 6.07E-84 1.16E-07

Identity (%) E-value

Table 3. Comparison of Mam, Mad and putative magnetosome proteins involved in magnetosome formation identified in the draft genome of Ca. Magnetoovum chiemensis CS-04 against draft genomes of Ca. M. bavaricum, Ca. M. casensis, Ca. Omnitrophus magneticus SKK-01 and the GenBank database.

8 S. Kolinko et al.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

2.19E-90 2.65E-23 2.45E-58 2.04E-13 9.03E-76 58.33 66.67 64.75 35.76 65.19 KJU84843 KJU84847 KJU84850 KJU82967 AIM41328 Mbav Mbav Mbav Mbav Mcas 1.17E-90 1.57E-24 8.45E-59 6.26E-16 2.29E-78 72.41 67.86 65.47 36.25 65.19 KJR43878 AIM41319 AIM41322 AIM41329 KJU82966 Movum Mcas Mcas Mcas Mbav 8.26E-95 3.63E-33 2.04E-24 4.95E-35 1.90E-83 60.78 71.43 70.15 43.53 71.02 AIM41315 KJR43883 KJR40361 KJR44081 KJR44083

E-value Identity (%) Accession Accession number

KJJ83152 KJJ83153 KJJ83154 KR349326 KR349327

Putative protein

MamE MamB, partial MamM Mad11 MamK, partial

Mcas Movum Movum Movum Movum

E-value Identity (%) Accession Accession

Identity (%)

E-value

Bacterium with protein with third highest sequence identity Bacterium with protein with second highest sequence identity Bacterium with protein with highest sequence identity

Table 4. Comparison of Mam and Mad magnetosome proteins involved in magnetosome formation identified in the draft genome of Ca. Omnitrophus magneticus SKK-01 against draft genomes of Ca. M. bavaricum, Ca. M. casensis, Ca. Magnetoovum CS-04 and the GenBank database.

Genomic analysis of an uncultivated multicellular magnetotactic prokaryote

9

mad28-29 operon. The genes mad28-3 and mad29 are separated by a hypothetical gene (KJU86279), which was found to be conserved in magnetotactic Nitrospirae and encodes a protein putatively involved lipid A biosynthesis (Table 2). In Ca. M. casensis, a homologous mad28-29 operon is present outside the MAI, and comprises as well a third mad28 and a mad29 gene, which are separated by a homologous gene to KJU86279. In Ca. M. chiemensis CS-04, the mad28-29 operon is also present but much larger (12 560 bp) and contains besides a second mad28 (KJR44079) a gene encoding a hypothetical protein (KJR44078), a mad29 (KJR44077), and several previously described magnetosome genes, like mamKB, mad11, mad17 and mad30 (Fig. 1). Noteworthy, magnetosome genes involved in iron uptake (mad17, mad30 and mamB) are in close genomic proximity. Two genes (KJR44080, KJR44071) have high identities to genes in other magnetotactic bacteria. KJR44071 encodes a lipoprotein that is highly conserved in magnetotactic Nitrospirae and Alphaproteobacteria, but missing in the class of Deltaproteobacteria. The second gene encodes a CheY-like protein (KJR44080) and is similar to an orthologue in Ca. M. bavaricum (KJU83260), which is in close proximity to magnetosome genes mad17-2 (KJU83259), mad17-1 (KJU83258) and mad30 (KJU83257) (Table 2) (Fig. 1). A third conserved hypothetical protein is located upstream of the two copies of mad17 in Ca. M. bavaricum (KJU83256) and Ca. M. casensis, but was not identified in Ca. M. chiemensis CS-04 (Fig. 1). A further highly conserved gene was identified in the mamAB-like operons of Ca. M. bavaricum, Ca. M. casensis and Ca. M. chiemensis CS-04 (KJR43881), always localized between mamM and mamQ-2 coding for a putative thioredoxin. This gene was also present in a gene cluster together with mamK and several mad genes of the recently described multicellular magnetotactic prokaryote Ca. Magnetomorum HK-1 (Kolinko et al., 2014a,b). Due to its high conservation among MTB of the Nitrospirae phylum, its presence in the Deltaproteobacteria class, but absence in the Alphaproteobacteria, it was termed mad31. Thioredoxins are small ubiquitous proteins with a highly conserved active site sequence (Holmgren, 1985; 1995; Martin, 1995) that have functions in redox regulation and oxidative stress response (Zeller and Klug, 2006), and therefore may be involved in the maintenance of physicochemical conditions necessary for magnetite biomineralization. Individual Mam and Mad proteins of Ca. M. bavaricum, Ca. M. casensis, Ca. M. chiemensis CS-04, Ca. O. magneticus SKK-01 and Ca. Magnetomorum HK-1, together with Ca. Magnetoglobus multicellularis, Ca. Desulfamplus magnetomortis, Desulfovibrio magneticus, Magnetospirillum gryphiswaldense and Magnetococcus

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

0.2

51

69

86

88

66

93

75

100

100

94

100 Ca. M. casensis

Ca. M. bavaricum

Ca. M.chiemensis CS-04

Ca. O.magneticus SKK-01

Ca. M. multicellularis*

Ca. D. magnetomortis BW-1*

Ca. Magnetomorum HK-1*

ML-1

D.magneticus RS-1.

Ca. M. casensis

M. marinus MC-1 M. gryphiswaldense MSR-1

Ca. D. magnetomortis BW-1

D. magneticus RS-1

Ca. Magnetomorum HK-1

88

Ca. Magnetomorum HK-1

100

Ca. M. bavaricum

Ca. M.chiemensis CS-04

Ca. D. magnetomortis BW-1.

Ca. M. multicellularis

#

Ca. Magnetomorum HK-1

B

A

92 87

Nitrospirae

0.05

99

100

100

51

68

100

93

Ca. M. multicellularis (EF014726)

Gamma proteobacterium SS-5 (HQ595729)

Gamma proteobacterium BW-2 (HQ595728)

M. gryphiswaldense MSR-1 (Y10109)

M. marinus MC-1 (LO6456)

D. magneticus RS-1 (AP01904)

Ca. D. magnetomortis BW-1 (JN252194)

Ca. Magnetananas tsingtaoensis (HQ857738)

Ca. Magnetomorum HK-1 (PRJNA252699)

Ca. M. mohaviensis LO-1 (GU797422)

Deltaproteobacteria

Alpha- und Gammaproteobacteria

Omnitrophica

uncultured bacterial clone OP3/2 (FP245540)

uncultured bacterial clone OP3/1 (FP245538)

Ca. O. magneticus SKK-01 (JN412733)

uncultured bacterial clone OP3/3 (FP245539)

98

MWB-1 (JN630580)

Ca. M. chiemensis CS-04

Ca. M. bavaricum (X71838)

100 Ca. M. casensis (JMFO00000000)

MHB-1 (AJ863136)

Thermodesulfovibrio yellowstonii (NC 011296)

41

45

100

32

100

73

C

Fig. 2. Phylogenetic maximum-Likelihood tree based on concatenated amino acid sequences of Mad (Mad23, Mad24, Mad25 and Mad26) (A) and Mam (mamaqbe) (B) genes. *Magnetosome genes likely involved in greigite biomineralization. #Mad genes closely associated to putatively mam genes likely involved in greigite biomineralization in Ca. Magnetomorum strain HK-1. (C) Phylogenetic tree of 16S rDNA gene sequences. Alpha- and Gammaproteobacteria are indicated in blue, Deltaproteobacteria are indicated in green, Nitrospirae are indicated in yellow, and Omnitrophica are indicated in pink.

39

26

10 S. Kolinko et al.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Genomic analysis of an uncultivated multicellular magnetotactic prokaryote marinus, involved in magnetite and greigite biomineralization, were analysed separately and in concatenation (Fig. 2). This indicated that Mam proteins of the Nitrospirae and the Deltaproteobacteria could be more closely related to each other than to those of the Alphaproteobacteria. Noteworthy, individual and concatenated Mam and Mad magnetosome proteins of all Nitrospirae clustered within the Deltaproteobacteria and seem to be more closely related to those putatively involved in greigite than to those in magnetite biomineralization, which was previously described for Mam proteins only (Abreu et al., 2014). Mam and Mad proteins of Ca. O. magneticus SKK-01, the only known magnetotactic representative of the deepbranching candidate phylum Omnitrophica, seemed to be closely related to those of Ca. M. chiemensis CS-04 within the deep-branching Nitrospirae. Discussion Recently a growing number of uncultivated MTB of the Nitrospirae and Omnitrophica phyla from diverse habitats in Germany, USA and China (Spring et al., 1993; Flies et al., 2005; Lefèvre et al., 2011; Lin et al., 2011; 2014; Kolinko et al., 2013) have been discovered. Because of their postulated importance in environmental sulfur and iron cycling, their conspicuous cell architecture, lifestyle, and the biosynthesis of hundreds of bullet-shaped magnetosomes, the genetic analyses of these MTB are of significant interest. In contrast to non-magnetic uncultivated microorganisms, MTB can be directly collected from environmental samples and separated from sediment particles and other microorganisms by using their magnetically directed motility (Flies et al., 2005). Metagenomic analysis of magnetically collected MTB revealed highly conserved magnetotactic signature genes in MTB of the deep-branching phylum Nitrospirae, with homology to those in Proteobacteria (Jogler et al., 2011). Recently, the genome of the uncultivated MTB Ca. M. casensis has been partly sequenced from highly enriched magnetic enrichments (Lin et al., 2014). This genome sequencing approach used highly enriched magnetic collections of uncultivated MTB and relied on high cell numbers, and thus is limited to abundant representatives. Since most uncultivated MTB with conspicuous and unique morphologies have low abundance, different abundanceindependent techniques must be applied (Kolinko et al., 2014b). Similar to recently published cultivationindependent targeted approaches applying single-cell techniques (Rinke et al., 2013; Kolinko et al., 2014b), we recovered draft genomes of Ca. O. magneticus strain SKK-01 of the deep-branching candidate phylum Omnitrophica (formerly candidate division OP3) (Rinke et al., 2013) and the magnetic Nitrospirae Ca. M. bavaricum and Ca. M. chiemensis CS-04 respec-

11

tively. Although the obtained draft genomes contained several small contigs and were not complete, they were sufficient for functional genome analysis. General metabolism All analysed MTB shared a putative autotrophic lifestyle fixing CO2 through the reverse citric acid cycle or the wood-ljungdahl (WL) pathway, which was only present in magnetic Nitrospirae Ca. M. bavaricum and Ca. M. chiemensis CS-04, and is consistent with the identification of the WL pathway and the reverse citric acid cycle in the draft genome of Ca. M. casensis (Lin et al., 2014). The WL pathway can be found in acetogens, methanogens and many sulfate reducers, and may represent a growth advantage due to its energetic convenience (Peréto et al., 1999; Sousa et al., 2013). Despite ultrastructural identification of sulfur inclusions in magnetic Nitrospirae and their consumption during prolonged microaerobic storage in Ca. M. bavaricum, indicating a sulfur-oxidizing metabolism (Jogler et al., 2010; Kolinko et al., 2012; 2013), we identified genes coding for enzymes Sat, Apr and Dsr, which are known to be involved in dissimilatory sulfate reduction, but were suggested to operate also in reverse direction (Zhou et al., 2011; Lin et al., 2014). Additionally, the oxidation of sulfide to sulfur is putatively catalysed by the sulfide : quinone oxidoreductase, which was encoded in their draft genomes. The reversibility of sulfur oxidation and the vertical distribution of Ca. M. bavaricum in sediments, forming two peaks at the oxic-anoxic transition zone, with one at the microarophilic and the other at the anoxic zone (Jogler et al., 2010). These findings suggest an electron shuttling between sulfate and sulfide depending on redox conditions, as reported for some bacteria, like Desulfovibrio desulfuricans und Desulfovibrio vulgaris (Dilling and Cypionka, 1990; Dannenberg et al., 1992; Cypionka, 1994). The identification of a DNRA pathway in both Nitrospirae and an additional DN pathway in Ca. M. bavaricum, as well as a partial respiratory chain, indicates the oxidation of reduced sulfur compounds at the OATZ with nitrate and oxygen respectively. In deeper and therefore more reduced sediment layers, sulfate may be used as terminal electron acceptor. These postulations are consistent with the findings for Ca. M. casensis (Lin et al., 2014). In contrast to Ca. M. bavaricum and Ca. M. chiemensis CS-04, Ca. O. magneticus SKK-01 seems to be less flexible. The presence of sulfur-rich inclusions indicated a sulfur-oxidizing metabolism in the magnetotactic Omnitrophica SKK-01 and is in agreement with the identification of a dissimilatory sulfite reductase (dSiR). The dSiR is postulated to be involved in the oxidation of sulfide under certain, putatively more oxidized conditions

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

12 S. Kolinko et al. (Steuber and Kroneck, 1998), as found at the OATZ where most Ca. O. magneticus SKK-01 cells were present in 10 mm sediment depth, coinciding with a transition in colour of the stratified sediment and indicating a change from oxic to reduced conditions (Kolinko et al., 2012). Therefore, a sulfur cycling between sulfate and sulfide, depending on redox conditions, may be performed by Ca. O. magneticus SKK-01, similar to Ca. M. bavaricum and Ca. M. chiemensis CS-04, but further studies need to prove this hypothesis of energy conservation for deep-branching MTB. Ca. M. bavaricum and Ca. M. chiemensis CS-04 display complex metabolic strategies similar to Ca. M. casensis, allowing them to quickly adapt to changing environmental conditions, which may lead to a growth advantage of these microorganisms. The metabolic accumulation of large amounts of iron is required for magnetosome biosynthesis; therefore, the intracellular iron content may account for up to 4% of the dry weight (Uebe et al., 2011). Ca. M. bavaricum and Ca. M. bavaricum and Ca. M. chiemensis CS-04 seem to accumulate ferrous iron applying the NRAMP and Feo system, which is supported by the identification of genes of the NRAMP transporter family and genes feoA (mad17) and feoB (mad30) in both genomes. The genes mad17, mad30 and mamB were contiguously present in Ca. M. chiemensis CS-04, whereas in Ca. M. bavaricum one mad30 was accompinied upstream by two mad17 genes, which were found together on one contig. The presence of TonB-dependent transporters indicates as well the uptake of ferric iron, despite no genes involved in siderophore biosynthesis were encoded. Magnetosome biosynthesis The mamAB operon of magnetite-producing MTB contains a conserved set of core genes (mamABEIKLMOPQ), which include an overlapping set of universal functions for magnetosome biosynthesis (Murat et al., 2010; Lefèvre and Bazylinski, 2013; Lefèvre et al., 2013; Lohsse et al., 2014) and crystal formation in a heterologous host (Kolinko et al., 2014a). We identified these genes (mamABEIKMOPQ) in the draft genome of Ca. M. bavaricum organized in a mamAB-like operon. This mamAB-like operon shows only minor divergences to the full mamAB-like operon of Ca. M. casensis with respect to gene order, G + C content (∼ 50%) and average nucleotide sequence identity (Lin et al., 2014) (Fig. 1, Table 2). These remarkable similarities between the geographically distant Ca. M. bavaricum and Ca. M. casensis indicate a strong conservation of mamAB-like operons in the Ca. Magnetobacterium genus. The mam genes (mamABQME) found in the partial mamAB operon of Ca. M. chiemensis CS-04 were arranged similar to

Ca. M. bavaricum, except for mamE being located upstream of mamM and not downstream of mamA. Further, partial magnetosome genes mamP and a second mamM were found together on a small contig. Despite a similar gene arrangement, a significant sequence divergences at the 16S rDNA level and a differing G + C content (40%) between Ca. M. chiemensis CS-04 and Ca. M. bavaricum, and Ca. M. casensis was discovered, suggesting a higher diversity within the magnetotactic Nitrospirae as previously speculated (Lefèvre et al., 2011; Kolinko et al., 2013). Although the draft genome of Ca. O. magneticus SKK-01 had the highest estimated genome completeness of all analysed MTB, only few partial magnetosome genes (mamEBMK) could be identified on small discontinuous contigs, and therefore no gene order could be conferred. Apart from the possibility that magnetosome genes escaped our detection, this may suggest that further magnetosome genes of the deep-branching candidate phylum Omnitrophica share low or even no homology to those found in the Proteobacteria and Nitrospirae. Despite the recent identification of a conserved cluster of magnetosome genes potentially involved in greigite formation in an uncultivated member of the phylum Latescibacteria (Lin and Pan, 2015), greigite magnetosome formation is only described for MTB of the Deltaproteobacteria and was never observed in MTB of the Nitrospirae phylum. Consistently, the absence of respective genes contraindicates a genetic potential for greigite biomineralization in Nitrospirae MTB. A number of previously identified alphaproteobacterial magnetosome genes, such as mamHRSTNOLJ, were absent from all analysed species, including Ca. M. casensis. Interestingly, we identified in all draft genomes several mad genes, either interspacing or adjacent to mam genes. Lefèvre and colleagues found a set of 18 mad genes specific to magnetotactic Deltaproteobacteria that biomineralize bullet-shaped magnetosomes (Lefèvre and Bazylinski, 2013; Lefèvre et al., 2013). By combined analysis of deep-branching MTB, including Ca. M. casensis, we found a set of 11 proteins (Mad2, Mad10, Mad11, Mad17, Mad23, Mad24, Mad25, Mad26, Mad28, Mad29 and Mad30) shared with the magnetotactic Deltaproteobacteria. The gene mad11 was also present in Ca. M. casensis in which it was annotated as man1. Interestingly, a short cluster comprising magnetosome genes mad23, mad24, mad25 and mad26 was found in all Nitrospirae and Deltaproteobacteria genomes of MTB synthesizing bullet-shaped magnetosomes. Except for Mad23, which is putatively involved in protein interaction, no functions are predicted for Mad24, Mad25 and Mad26. Nonetheless, it is possible that these genes specific to the deep-branching MTB might govern accessory functions, such as control of size,

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Genomic analysis of an uncultivated multicellular magnetotactic prokaryote shape and organization of magnetosomes, similar to the mamGFDC operon in magnetic Alphaproteobacteria. The magnetosome gene mad11 is also conserved in all deepbranching MTB and the Deltaproteobacteria, and is always adjacent to mamK (Lefèvre and Bazylinski, 2013; Lefèvre et al., 2013), including Desulfovibrio magneticus, Ca. Magnetomortis desulfamplus, bacterium FH-1 and in the recently published draft genome of the MMP Ca. Magnetomorum HK-1 (Kolinko et al., 2014b), indicating a putative function in magnetosome chain formation. However, as no conserved domains or motifs could be identified and no function for Mad11 was predicted, an accessory function in magnetosome biosynthesis can only be hypothesized. The Nitrospirae Ca. M. bavaricum, Ca. M. casensis and Ca. M. chiemensis CS-04 biosynthesize multiple chains of magnetosomes traversing the cell along their cell length axis (Spring et al., 1993; Hanzlik et al., 1996; Lin et al., 2014). Apparently, all MAIs encode the actin-like protein MamK, which forms a cytoskeletal element and is known to be involved in magnetosome chain assembly (Komeili et al., 2006; Katzmann et al., 2010). Noteworthy, we found up to five copies of mad28 coding for a further actin-like protein (Lefèvre and Bazylinski, 2013; Lefèvre et al., 2013) randomly scattered in our draft genomes and in Ca. M. casensis. This high number of genes for potential cytoskeletal proteins might be associated with the ability of magnetic Nitrospirae to arrange hundreds of magnetosomes into complex structures and to withstand strong magnetic interactions. These observations are consistent with the demonstration of filamentous structures in Ca. M. bavaricum by cryo-scanning electron microscopy (SEM) and cryo-transmission electron microscopy (TEM), which unlike alphaproteobacterial filaments did not form a network of filament bundles but appeared to form ordered tubular structures with a hollow interior, around which the individual magnetosome strands are arranged (Jogler et al., 2011). Interestingly, mad28-2 is accompanied by a conserved gene coding for a membranous putative penicillin-binding protein (PBP1A), which is involved in peptidoglycan synthesis as a multimodular transpeptidase-transglycosylase (Sauvage et al., 2008; Banzhaf et al., 2012). The interaction of cytoskeletal elements with peptidoglycan synthesis might facilitate in overcoming the magnetostatic interactions between asymmetrically distributed separating daughter chains during cell division, as described for M. gryphiswaldense (Katzmann et al., 2011). The gene mad29 encodes a putative ATPase and is located downstream of a mad28 homologue in all magnetic Nitrospirae and the Deltaproteobacteria Ca. D. magnetomortis and D. magneticus (Lefèvre and Bazylinski, 2013; Lefèvre et al., 2013). As an ATPase, Mad29 might interact with Mad28 filaments and might therefore be involved

13

in magnetosome repositioning as described for M. gryphiswaldense (Draper et al., 2011). Phylogenetic trees indicated that magnetosome proteins, such as MamE, MamB, MamM, MamK and Mad11 of Ca. O. magneticus SKK-01, branch together with their homologues of MTB relatives from the Nitrospirae, forming a branch within the class of magnetic Deltaproteobacteria. Due to the fact that these three magnetotactic groups are only distantly related and separated by non-magnetotactic species, a common magnetotactic ancestor for Proteobacteria, Nitrospirae and Omnitrophica seems very unlikely, as this scenario would require multiple events of loss of magnetosome genes in all three phyla. Another possibility is represented by horizontal gene transfer of magnetosome genes between MTB of the Proteobacteria, Nitrospirae and Omntitrophica. In contrast to phylogenetic trees based on magnetosome proteins, phylogenetic 16S and 23S rDNA gene trees showed a clear separate branching of Ca. O. magneticus SKK-01 from Nitrospirae and Proteobacteria within the PVC superphylum (Kolinko et al., 2013). Horizontal gene transfer of conserved magnetosome genes between Proteobacteria, Nitrospirae and Omnitrophica provides a more likely explanation of this contradiction. Metagenomic clones, which were assigned to the candidate phylum Omnitrophica based on rDNA analysis, contained a high proportion of open reading frames (ORFs) with best matches to homologues from Deltaproteobacteria and may have evolved similar capabilities (Glöckner et al., 2010). In contrast to metabolic genes, genes involved in magnetosome biosynthesis of Ca. O. magneticus SKK-01 yielded best hits to those of the Nitrospirae phylum, again indicating horizontal gene transfer. The observed minor difference in G + C content between MAI (35.8%) and draft genome (38.1%) in Ca. O. magneticus SKK-01 argues against a recent horizontal gene transfer event and rather for a more ancient event. In contrast to the first identification of mad genes, this set of genes is not restricted to magnetotactic Deltaproteobacteria but is ubiquitously present in all deep-branching MTB and should be redefined as ‘magnetosome associated deep-branching’.

Experimental procedures Sampling Freshwater sediments were collected from Lake Chiemsee (Germany, Bavaria) (47°51′08″N, 12°24′00″E) and stored at room temperature as previously described (Kolinko et al., 2012). Magnetic enrichments from microcosms were performed as previously described (Jogler et al., 2010) and analysed for the presence of morphologically conspicuous MTB like Ca. M. bavaricum, SKK-01 and CS-04.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

14

S. Kolinko et al.

WGA and SAG screening Micromanipulation and WGA of single cells were performed as previously described (Kolinko et al., 2012). Briefly, single conspicuous cells were selected and washed under strict microscopic control via micromanipulation to eliminate contaminations and extracellular DNA for subsequent WGA. UV irradiation of WGA sample and reaction buffers (illustra GenomiPhi V2 DNA amplification kit, GE Healthcare, Little Chalfont, UK) were performed as previously described (Woyke et al., 2011) for decontamination. Each WGA reaction with five individual cells was carried out at 30°C for 10 h, which yielded ∼ 600 ng DNA. Partial 16S rDNA was amplified from SAGs applying bacteria-specific primers 27F and 1492R (Lane, 1991), cloned into pJET1.2/blunt Cloning Vector (Fermentas, Waltham, MA, USA) and sequenced with an ABI3730 (Life Technologies) system according to manufacturer’s instructions to identify SAGs for WGA.

Sequencing and draft genome assembly For SAG sequencing, DNA libraries were generated with the Nextera XT Kit (Illumina) according to manufacturer’s instructions. Sequencing (2× 150 bp, 2× 250 bp, and 2× 300 bp) was performed with a MiSeq sequencer (Illumina) (Fig. S1). Quality trimming and filtering have been done with trimmomatic version 0.30 and fasta toolkit version 0.13.2. Passed single and paired reads were assembled by using SPAdes. SPAdes has been performed with standard parameters in the multiple displacement amplification (MDA) (single cell) modus with activated error correction (Bankevich et al., 2012; Nurk et al., 2013).

Genome completeness estimation Genome size and completeness was estimated using a set of 139 conserved single copy genes (CSCG), which have been determined from finished bacterial genome sequences (n = 1516) in the IMG database (Markowitz et al., 2012; Rinke et al., 2013). These genes occurred only once in at least 90% of all genomes, based on hits to the protein family (Pfam) database (Punta et al., 2012). The estimated genome completeness of individual and combined SAGs was estimated as the ratio of identified CSCGs to total CSCGs, and was used to calculate the estimated complete genome size by dividing the estimated genome coverage by the total assembly size (Rinke et al., 2013).

Phylogenetic analysis Alignments of sequences were performed using CLUSTALW multiple alignment accessory application in the MEGA version 6 sequence alignment editor (Tamura et al., 2013). Phylogeny was analysed as previously described (Abreu et al., 2014). In brief, phylogenetic trees were constructed using MEGA version 6 applying the maximum likelihood method based on the Whelan and Goldman model (Whelan and Goldman, 2001). The tree with the highest log likelihood is shown. Bootstrap values were calculated with 1000 replicates. The tree topology space was searched using

the best tree by Subtree Pruning and Regrafting algorithms, starting from five random trees generated by the BIONJ algorithm (Guindon and Gascuel, 2003; Guindon et al., 2010).

Acknowledgements This work was supported by Deutsche Forschungsgemeinschaft (Grant DFG Schu1080/11-1 to D.S.). This Whole Genome Shotgun project has been deposited at the DDBJ/EMBL/GenBank under the accessions JYNY00000000, LACI00000000 and JZJI00000000. The version described in this paper is version JYNY01000000, LACI00000000 and JZJI00000000.

References Abreu, F., Cantao, M.E., Nicolas, M.F., Barcellos, F.G., Morillo, V., Almeida, L.G., et al. (2011) Common ancestry of iron oxide- and iron-sulfide-based biomineralization in magnetotactic bacteria. ISME J 5: 1634–1640. Abreu, F., Morillo, V., Nascimento, F.F., Werneck, C., Cantao, M.E., Ciapina, L.P., et al. (2014) Deciphering unusual uncultured magnetotactic multicellular prokaryotes through genomics. ISME J 8: 1055–1068. Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402. Bankevich, A., Nurk, S., Antipov, D., Gurevich, A.A., Dvorkin, M., Kulikov, A.S., et al. (2012) SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J Comput Biol 19: 455–477. Banzhaf, M., van den Berg van Saparoea, B., Terrak, M., Fraipont, C., Egan, A., Philippe, J., et al. (2012) Cooperativity of peptidoglycan synthases active in bacterial cell elongation. Mol Microbiol 85: 179–194. Baumgartner, J., Morin, G., Menguy, N., Perez Gonzalez, T., Widdrat, M., Cosmidis, J., and Faivre, D. (2013) Magnetotactic bacteria form magnetite from a phosphaterich ferric hydroxide via nanometric ferric (oxyhydr)oxide intermediates. Proc Natl Acad Sci USA 110: 14883–14888. Bazylinski, D.A., and Frankel, R.B. (2004) Magnetosome formation in prokaryotes. Nat Rev Microbiol 2: 217–230. Bazylinski, D.A., Lefèvre, C.T., and Schüler, D. (2013) Magnetotactic bacteria. In The Prokaryotes. Rosenberg, E., DeLong, E.F., Lory, S., Stackebrandt, E., and Thomson, F. (eds). Berlin, Germany: Springer, pp. 453–494. Bonomi, H., Posadas, D., Paris, G., Carrica, M.d.C., Frederickson, L., Pietrasanta, L. et al. (2012) Light regulates attachment, exopolysaccharide production in Rhizobium leguminosarum through a LOV-histidine kinase photoreceptor. Proc Natl Acad Sci USA 109: 12135– 12140. Cypionka, H. (1994) Novel metabolic capacities of sulfatereducing bacteria, and their activities in microbial mats. In Microbial Mats. Stal, L., and Caumette, P. (eds). Berlin, Germany: Springer Heidelberg, pp. 367–376. Dahl, C., Engels, S., Pott-Sperling, A.S., Schulte, A., Sander, J., Lubbe, Y., et al. (2005) Novel genes of the dsr gene cluster and evidence for close interaction of Dsr proteins

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Genomic analysis of an uncultivated multicellular magnetotactic prokaryote during sulfur oxidation in the phototrophic sulfur bacterium Allochromatium vinosum. J Bacteriol 187: 1392–1404. Dannenberg, S., Kroder, M., Dilling, W., and Cypionka, H. (1992) Oxidation of H2, organic compounds and inorganic sulfur compounds coupled to reduction of O2 or nitrate by sulfate-reducing bacteria. Arch Microbiol 158: 93–99. Dilling, W., and Cypionka, H. (1990) Aerobic respiration in sulfate-reducing bacteria. FEMS Microbiol Lett 71: 123– 127. Dong, L.F., Sobey, M.N., Smith, C.J., Rusmana, I., Phillips, W., Stoff, A., et al. (2011) Dissimilatory reduction of nitrate to ammonium, not denitrification or anammox, dominates benthic nitrate reduction in tropical estuaries. Limnol Oceanogr 56: 279–291. Draper, O., Byrne, M.E., Li, Z., Keyhani, S., Barrozo, J.C., Jensen, G., and Komeili, A. (2011) MamK, a bacterial actin, forms dynamic filaments in vivo that are regulated by the acidic proteins MamJ and LimJ. Mol Microbiol 82: 342– 354. Flies, C.B., Peplies, J., and Schüler, D. (2005) Combined approach for characterization of uncultivated magnetotactic bacteria from various aquatic environments. Appl Environ Microbiol 71: 2723–2731. Frankel, R.B., Bazylinski, D.A., Johnson, M.S., and Taylor, B.L. (1997) Magneto-aerotaxis in marine coccoid bacteria. Biophys J 73: 994–1000. Glöckner, J., Kube, M., Shrestha, P.M., Weber, M., Glöckner, F.O., Reinhardt, R., and Liesack, W. (2010) Phylogenetic diversity and metagenomics of candidate division OP3. Environ Microbiol 12: 1218–1229. Grünberg, K., Wawer, C., Tebo, B.M., and Schüler, D. (2001) A large gene cluster encoding several magnetosome proteins is conserved in different species of magnetotactic bacteria. Appl Environ Microbiol 67: 4573–4582. Guindon, S., and Gascuel, O. (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol 52: 696–704. Guindon, S., Dufayard, J.F., Lefort, V., Anisimova, M., Hordijk, W., and Gascuel, O. (2010) New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol 59: 307–321. Gupta, R.S., Bhandari, V., and Naushad, H.S. (2012) Molecular signatures for the PVC clade (Planctomycetes, Verrucomicrobia, Chlamydiae, and Lentisphaerae) of bacteria provide insights into their evolutionary relationships. Front Microbiol 3: 327. Hanzlik, M., Winklhofer, M., and Petersen, N. (1996) Spatial arrangement of chains of magnetosomes in magnetotactic bacteria. Earth Planet Sci Lett 145: 125–134. Holmgren, A. (1985) Thioredoxin. Annu Rev Biochem 54: 237–271. Holmgren, A. (1995) Thioredoxin structure and mechanism: conformational changes on oxidation of the active-site sulfhydryls to a disulfide. Structure 3: 239–243. Hugenholtz, P., Pitulle, C., Hershberger, K.L., and Pace, N.R. (1998) Novel division level bacterial diversity in a Yellowstone hot spring. J Bacteriol 180: 366–376. Jogler, C., and Schüler, D. (2009) Genomics, genetics, and cell biology of magnetosome formation. Annu Rev Microbiol 63: 501–521.

15

Jogler, C., Kube, M., Schübbe, S., Ullrich, S., Teeling, H., Bazylinski, D.A., et al. (2009a) Comparative analysis of magnetosome gene clusters in magnetotactic bacteria provides further evidence for horizontal gene transfer. Environ Microbiol 11: 1267–1277. Jogler, C., Lin, W., Meyerdierks, A., Kube, M., Katzmann, E., Flies, C., et al. (2009b) Toward cloning of the magnetotactic metagenome: identification of magnetosome island gene clusters in uncultivated magnetotactic bacteria from different aquatic sediments. Appl Environ Microbiol 75: 3972– 3979. Jogler, C., Niebler, M., Lin, W., Kube, M., Wanner, G., Kolinko, S., et al. (2010) Cultivation-independent characterization of ‘Candidatus Magnetobacterium bavaricum’ via ultrastructural, geochemical, ecological and metagenomic methods. Environ Microbiol 12: 2466–2478. Jogler, C., Wanner, G., Kolinko, S., Niebler, M., Amann, R., Petersen, N., et al. (2011) Conservation of proteobacterial magnetosome genes and structures in an uncultivated member of the deep-branching Nitrospirae phylum. Proc Natl Acad Sci USA 108: 1134–1139. Katzmann, E., Scheffel, A., Gruska, M., Plitzko, J.M., and Schüler, D. (2010) Loss of the actin-like protein MamK has pleiotropic effects on magnetosome formation and chain assembly in Magnetospirillum gryphiswaldense. Mol Microbiol 77: 208–224. Katzmann, E., Müller, F.D., Lang, C., Messerer, M., Winklhofer, M., Plitzko, J.M., and Schüler, D. (2011) Magnetosome chains are recruited to cellular division sites and split by asymmetric septation. Mol Microbiol 82: 1316– 1329. Kolinko, S., Jogler, C., Katzmann, E., Wanner, G., Peplies, J., and Schüler, D. (2012) Single-cell analysis reveals a novel uncultivated magnetotactic bacterium within the candidate division OP3. Environ Microbiol 14: 1709–1721. Kolinko, S., Wanner, G., Katzmann, E., Kiemer, F., Fuchs, B.M., and Schüler, D. (2013) Clone libraries and single cell genome amplification reveal extended diversity of uncultivated magnetotactic bacteria from marine and freshwater environments. Environ Microbiol 15: 1290–1301. Kolinko, I., Lohsse, A., Borg, S., Raschdorf, O., Jogler, C., Tu, Q., et al. (2014a) Biosynthesis of magnetic nanostructures in a foreign organism by transfer of bacterial magnetosome gene clusters. Nat Nanotechnol 9: 193–197. Kolinko, S., Richter, M., Glöckner, F.O., Brachmann, A., and Schüler, D. (2014b) Single-cell genomics reveals potential for magnetite and greigite biomineralization in an uncultivated multicellular magnetotactic prokaryote. Environ Microbiol Rep 6: 524–531. Komeili, A., Li, Z., Newman, D.K., and Jensen, G.J. (2006) Magnetosomes are cell membrane invaginations organized by the actin-like protein MamK. Science 311: 242–245. Lane, D.J. (1991) 16S/23S sequencing. In Nucleic Acid Techniques in Bacterial Systematics. Stackebrandt, E., and Chichster, M.G. (eds). New York, NY, USA: Wiley & Sons, pp. 115–175. Lefèvre, C.T., and Bazylinski, D.A. (2013) Ecology, diversity, and evolution of magnetotactic bacteria. Microbiol Mol Biol Rev 77: 497–526. Lefèvre, C.T., Frankel, R.B., Abreu, F., Lins, U., and Bazylinski, D.A. (2011) Culture-independent

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

16

S. Kolinko et al.

characterization of a novel, uncultivated magnetotactic member of the Nitrospirae phylum. Environ Microbiol 13: 538–549. Lefèvre, C.T., Trubitsyn, D., Abreu, F., Kolinko, S., Jogler, C., de Almeida, L.G., et al. (2013) Comparative genomic analysis of magnetotactic bacteria from the Deltaproteobacteria provides new insights into magnetite and greigite magnetosome genes required for magnetotaxis. Environ Microbiol 15: 2712–2735. Lin, W., and Pan, Y. (2015) A putative greigite-type magnetosome gene cluster from the candidate phylum Latescibacteria. Environ Microbiol Rep 7: 237–242. Lin, W., Jogler, C., Schüler, D., and Pan, Y. (2011) Metagenomic analysis reveals unexpected subgenomic diversity of magnetotactic bacteria within the phylum Nitrospirae. Appl Environ Microbiol 77: 323–326. Lin, W., Li, J., and Pan, Y. (2012) Newly isolated but uncultivated magnetotactic bacterium of the phylum Nitrospirae from Beijing, China. Appl Environ Microbiol 78: 668– 675. Lin, W., Deng, A., Wang, Z., Li, Y., Wen, T., Wu, L.F., et al. (2014) Genomic insights into the uncultured genus ‘Candidatus Magnetobacterium’ in the phylum Nitrospirae. ISME J 12: 2463–2477. Lohsse, A., Borg, S., Raschdorf, O., Kolinko, I., Tompa, E., Pósfai, M., et al. (2014) Genetic dissection of the mamAB and mms6 operons reveals a gene set essential for magnetosome biogenesis in Magnetospirillum gryphiswaldense. J Bacteriol 196: 2658–2669. Marczak, M., Dz´wierzyn´ska, M., and Skorupska, A. (2013) Homo- and heterotypic interactions between Pss proteins involved in the exopolysaccharide transport system in Rhizobium leguminosarum bv. trifolii. Biol Chem 4: 541– 559. Markowitz, V.M., Chen, I.M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y., et al. (2012) IMG: the Integrated Microbial Genomes database and comparative analysis system. Nucleic Acids Res 40: D115–D122. Martin, J.L. (1995) Thioredoxin – a fold for all reasons. Structure 3: 245–250. Murat, D., Quinlan, A., Vali, H., and Komeili, A. (2010) Comprehensive genetic dissection of the magnetosome gene island reveals the step-wise assembly of a prokaryotic organelle. Proc Natl Acad Sci USA 107: 5593–5598. Nurk, S., Bankevich, A., Antipov, D., Gurevich, A.A., Korobeynikov, A., Lapidus, A., et al. (2013) Assembling single-cell genomes and mini-metagenomes from chimeric MDA products. J Comput Biol 20: 714–737. Peréto, J.G., Velasco, A.M., Becerra, A., and Lazcano, A. (1999) Comparative biochemistry of CO2 fixation and the evolution of autotrophy. Int Microbiol 2: 3–10. Punta, M., Coggill, P.C., Eberhardt, R.Y., Mistry, J., Tate, J., Boursnell, C., et al. (2012) The PFAM protein families database. Nucleic Acids Res 40: D290–D301. Richter, M., Kube, M., Bazylinski, D.A., Lombardot, T., Glöckner, F.O., Reinhardt, R., and Schüler, D. (2007) Comparative genome analysis of four magnetotactic bacteria reveals a complex set of group-specific genes implicated in magnetosome biomineralization and function. J Bacteriol 189: 4899–4910.

Rinke, C., Schwientek, P., Sczyrba, A., Ivanova, N.N., Anderson, I.J., Cheng, J.F., et al. (2013) Insights into the phylogeny and coding potential of microbial dark matter. Nature 499: 431–437. Sauvage, E., Kerff, F., Terrak, M., Ayala, J.A., and Charlier, P. (2008) The penicillin-binding proteins: structure and role in peptidoglycan biosynthesis. FEMS Microbiol Rev 32: 234– 258. Schübbe, S., Kube, M., Scheffel, A., Wawer, C., Heyen, U., Meyerdierks, A., et al. (2003) Characterization of a spontaneous nonmagnetic mutant of Magnetospirillum gryphiswaldense reveals a large deletion comprising a putative magnetosome island. J Bacteriol 185: 5779– 5790. Sousa, F.L., Thiergart, T., Landan, G., Nelson-Sathi, S., Pereira, I.A., Allen, J.F., et al. (2013) Early bioenergetic evolution. Philos Trans R Soc Lond B Biol Sci 368: 20130088. Spring, S., Amann, R., Ludwig, W., Schleifer, K.H., van Gemerden, H., and Petersen, N. (1993) Dominating role of an unusual magnetotactic bacterium in the microaerobic zone of a freshwater sediment. Appl Environ Microbiol 59: 2397–2403. Steuber, J., and Kroneck, P.M.H. (1998) Desulfoviridin, the dissimilatory sulfite reductase from Desulfovibrio desulfuricans (Essex): new structural and functional aspects of the membranous enzyme. Inorganica Chim Acta 275: 52–57. Tamura, K., Stecher, G., Peterson, D., Filipski, A., and Kumar, S. (2013) MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Mol Biol Evol 30: 2725–2729. Uebe, R., Junge, K., Henn, V., Poxleitner, G., Katzmann, E., Plitzko, J.M., et al. (2011) The cation diffusion facilitator proteins MamB and MamM of Magnetospirillum gryphiswaldense have distinct and complex functions, and are involved in magnetite biomineralization and magnetosome membrane assembly. Mol Microbiol 82: 818–835. Ullrich, S., Kube, M., Schübbe, S., Reinhardt, R., and Schüler, D. (2005) A hypervariable 130-kilobase genomic region of Magnetospirillum gryphiswaldense comprises a magnetosome island which undergoes frequent rearrangements during stationary growth. J Bacteriol 187: 7176– 7184. Vali, H., Förster, O., Amarantidis, G., and Petersen, N. (1987) Magnetotactic bacteria and their magnetofossils in sediments. Earth Planet Sci Lett 87: 389–400. Wagner, M., and Horn, M. (2006) The Planctomycetes, Verrucomicrobia, Chlamydiae and sister phyla comprise a superphylum with biotechnological and medical relevance. Curr Opin Biotechnol 17: 241–249. Whelan, S., and Goldman, N. (2001) A general empirical model of protein evolution derived from multiple protein families using a maximum-likelihood approach. Mol Biol Evol 18: 691–699. Woyke, T., Sczyrba, A., Lee, J., Rinke, C., Tighe, D., Clingenpeel, S., et al. (2011) Decontamination of MDA reagents for single cell whole genome amplification. PLoS ONE 6: e26161. Zeller, T., and Klug, G. (2006) Thioredoxins in bacteria: functions in oxidative stress response and regulation of thioredoxin genes. Naturwissenschaften 93: 259–266.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Genomic analysis of an uncultivated multicellular magnetotactic prokaryote Zhou, J., He, Q., Hemme, C.L., Mukhopadhyay, A., Hillesland, K., Zhou, A., et al. (2011) How sulphatereducing microorganisms cope with stress: lessons from systems biology. Nat Rev Microbiol 9: 452–466.

Supporting information Additional Supporting Information may be found in the online version of this article at the publisher’s web-site:

17

Fig. S1. Estimated genome completeness for single amplified genomes (grey) and combined draft genomes (black) of Ca. Omnitrophus magneticus SKK-01 (SAG01), Ca. Magnetoovum hydrogenophilus CS-04 (SAG-02) and Ca. Magnetobacterium bavaricum (SAG-04). Table S1. Sequencing and assembly statistics for Ca. Magnetobacterium bavaricum, Ca. Magnetoovum chiemensis and Ca. Omnitrophus magneticus.

© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology