Polynucleobacter cosmopolitanus sp. nov., free-living planktonic bacteria inhabiting freshwater lakes and rivers

Five heterotrophic, aerobic, catalase- and oxidase-positive, non-motile strains from freshwater habitats located in Austria, France, Uganda, P. R. China, and New Zealand were characterized. The strains share 16S rRNA gene similarities ≥ 99.3%. The strains grew on NSY medium over a temperature range of 10–35 °C (two strains also grew at 5 °C, and one strain also grew at 38 °C) and a NaCl tolerance range of 0.0–0.3 % (four strains grew up to 0.5% NaCl). The predominant fatty acids were C_16:0, C_18:1ω7c, C_12:0-3OH, and feature 3 (including C_16:1ω7c). The DNA G+C content of the type strain MWH-MoIso2^T was 44.9 mol%. Phylogenetic analysis of 16S rRNA gene sequences demonstrated that the five strains formed a monophyletic cluster closely related to Polynucleobacter necessarius (96-97% sequence similarity). This cluster also harbours other isolates as well as environmental sequences, which have been obtained from several habitats. Investigations with taxon-specific FISH probes demonstrated that these bacteria dwell as free-living, planktonic cells in freshwater systems. Based on the revealed phylogeny and pronounced chemotaxonomic differences to P. necessarius (presence of>7 % C_12:0-3OH, absence of C_12:0 and C_12:0-2OH), we propose to establish for the investigated strains the new species Polynucleobacter cosmopolitanus sp. nov. with the type strain MWH-MoIso2^T (= DSM 21490^T, = CIP 109840^T). The proposed species belongs to the minority of described species of free-living bacteria for which both in situ data from their natural environments and culture-based knowledge is available.

K. Heckmann and H.-J. Schmidt described the genus Polynucleobacter for bacterial endosymbionts of freshwater ciliates within the genus Euplotes, and the species Polynucleobacter necessarius for obligate endosymbionts living in the cytoplasm of E. aediculatus (Heckmann and Schmidt, 1987). Phylogenetic analysis of the 16S rRNA gene sequence of an endosymbiotic P. necessarius strain in an E. aediculatus culture revealed the affiliation of the strain to the class Betaproteobacteria (Springer et al., 1996). Recently, it was demonstrated that the obligate endosymbionts representing the species P. necessarius are closely related with obligately free-living strains (Hahn, 2003; Vannini et al., 2007). Therefore, the description of the genus Polynucleobacter and its sole species P. necessarius was emended by addition of descriptions of free-living strains (Hahn et al., in press). Due to the very pronounced differences in lifestyle of the obligately endosymbiotic and obligately free-living strains, the placement of these organisms in the two subspecies P. necessarius subsp. necessarius (for endosymbionts of E. aediculatus and E. harpa) and P. necessarius subsp. asymbioticus (for obligately free-living strains) was proposed (Hahn et al., in press). The obligately free-living P. necessarius strains represent aerobic chemoorganotrophic, non-motile bacteria. Cultivation-independent investigations by fluorescent in situ hybridization (FISH) with P. necessarius-specific, probes demonstrated that the free-living strains possess a planktonic lifestyle and inhabit a large variety of freshwater habitats (Hahn et al., 2005; Wu & Hahn, 2006; Salcher et al., 2008; Alonso et al., in press).

In previous studies, strains were isolated or sequences were obtained by cultivation independent approaches, which were phylogenetically closely related to P. necessarius but did not cluster in trees together with P. necessarius strains (Hahn, 2003; Watanabe et al., 2009; Crump et al., 1999; Zwart et al., 2002). Therefore, these isolates were grouped in separate species-like operational taxonomic units (designated subclusters A, B, C, and D), which were defined by phylogenetic reconstructions (Hahn, 2003). In the present study, we characterize five strains affiliated with the so-called subcluster D (PnecD) of the Polynucleobacter cluster (Hahn, 2003; Wu and Hahn, 2006). These strains share many phenotypic, chemotaxonomic, and ecological traits with strains of P. necessarius subsp. asymbioticus but can be distinguished from those by phylogenetic and chemotaxonomic traits. Based on these differences, we propose to establish for these strains the species P. cosmopolitanus sp. nov. as a second species within the genus Polynucleobacter.

Isolation and characterization

All five investigated strains were isolated by using the filtration-acclimatization method (Hahn, 2003, Hahn et al., 2004) from freshwater habitats located in Austria, France, New Zealand, China, and Uganda, (Table 1). These habitats are located in temperate, subtropical and tropical climatic zones. Four strains were isolated from the surface waters of lakes, but one strain was isolated from a navigable channel fed with water from the Loire River, France. A recent investigation demonstrated that strains closely related to the five strains characterized here, could be directly isolated from agar plates with modified R2A medium (Watanabe et al., 2009).

Table 1.

Traits characterizing the five investigated P. cosmopolitanus sp. nov. strains. All strains have the following characteristics in common. DAPI stained cells show only rarely nucleoid-like structures; non-motile; catalase (strain MWH-TaW3 only weakly positive) and oxidase positive. −, Negative; +, positive; w, weakly positive; n.d., not determined.

	MWH-MoIso2T (DSM 21490T)	MWH-CaK1 (DSM 21494)	MWH-VicM1 (DSM 21486)	MWH-TaW3 (DSM 21487)	MWH-NZ8W13 (DSM 21488)	P. necessarius subsp. asymbioticus
Origin
Country	Austria	France	Uganda	P.R. China	New Zealand	-
Habitat	Lake Mondsee temperate	Canal Roanne a Digoin temperate	Lake Victoria tropical	Lake Taihu subtropical	Lake near Aviemore temperate	-
Climatic zone						-
Morphological traits
Cell morphology	short curved rods	curved rods	short curved rods	short curved rods	short curved rods	straight or curved rods
Cell length (μm)	0.5 - 1.2	0.5 - 1.4	0.4 - 1.1	0.5 - 1.2	0.5 - 1.2	0.5 - 2.9
Cell width (μm)	0.3 - 0.5	0.3 - 0.5	0.3 - 0.5	0.3 - 0.5	0.3 - 0.5	0.3 - 0.5
Physiological traits
Min. temp. of growth (°C)#	5(w)	5	10	5(w)	5	5
Max. temp. of growth (°C)#	35(w)	35(w)	38	38?	35	30 - 35
Max. NaCl concentration (% NaCl, w/v)	0.5(w)	0.5(w)	0.5	0.5	0.3	0.3 - 0.5
Anaerobic growth	+	+	+	−	w	+/−
Growth in mineral medium   with acetic acid and B12	−	w	w	−	n.d.	w/−
Other traits
Genome size (Mb)	n.d.§	n.d.§	n.d.§	n.d.§	n.d.§	2.1 -2.5
DNA G+C content (mol%)	44.9	n.d.	n.d.	n.d.	n.d.	44 - 46

^#investigated temperature range 5 - 38°C,

^§genome sizes of 1.9-2.1 Mbp were determined for other strains presumably belonging to P. cosmopolitanus spec. nov. (Vannini et al., 2007)

All investigated strains were routinely grown on NSY medium (Hahn et al., 2004) with strength of 3 g L⁻¹. All strains also grew on R2A medium, and Bacto peptone at least when concentrations were lowered to 3 g L⁻¹ (Hahn, 2003). Growth at different temperatures, and growth under anoxic conditions in an anaerobic chamber were examined on NSY agar. NaCl tolerance was determined using NSY agar supplemented with different NaCl concentrations (0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 1.0, 1.25, 1.5, 1.75, and 2.0%, w/v). The temperature range supporting growth was tested on NSY agar plates exposed to different temperatures (5, 10, 15, 20, 30, 33, 34, 35, 36, and 38 °C). Two of four tested strains grew weakly on an inorganic mineral medium (Hahn et al., 2004) supplemented with acetate as sole carbon and energy source and vitamin B12 (Table 1). However, all three tested strains utilized acetate (Table 2) when this substrate was offered in combination with the complex NSY medium (see below). The investigated strains share the weak growth performance on medium with a sole carbon source with previously investigated P. necessarius subsp. asymbioticus strains (Hahn et al., in press). Due to this weak performance and due to the sake of comparability with the previously investigated Polynucleobacter strains, tests on utilization of various organic compounds were neither performed with sole substrate media, nor by using commercially available test kits. Instead, experiments were performed by applying the method previously used for P. necessarius strains (Hahn et al., in press). Briefly, growth enabled by utilization of a specific substrate was determined by comparison of OD established in liquid one-tenth-strength NSY medium (0.3 gL⁻¹) with and without 0.5 g L⁻¹ test substance. OD differences of <10%, of 10-50%, and of >50% of the OD established on the medium without test substance were scored after 10 days of growth as no utilization (−), weak utilization (w) and good utilization (+), respectively. Sequencing and phylogenetic analyses of 16S rRNA genes was performed as described previously (Hahn, 2003; Hahn et al., 2005). Neighbour-Joining trees were calculated by using the software MEGA4 (Tamura et al., 2007), and Maximum-Likelihood trees were generated by using the RaxML web server (Stamatakis et al., 2008). The morphological characterization of the proposed type strain by electron microscopy was performed as described previously (Yakimov et al., 1998). The G+C content of DNA was determined as described by Tóth et al. (2008). Fatty acid methyl esters (FAMEs) were obtained as described previously (Kämpfer & Kroppenstedt, 1996) and separated by a gas chromatograph (model 6890, Hewlett Packard). Peaks were automatically computed using the Microbial Identification standard software package (Sasser, 1990).

Table 2.

Results of utilization experiments with three investigated strains as compared to data reported for free-living P. necessarius strains (Hahn et al., in press). All tested strains have with P. necessarius subsp. asymbioticus in common that they did not utilize glycolate, oxalate, and L-serine, and that they did utilize acetate, pyruvate, and succinate. For. P. necessarius subsp. asymbioticus results from four investigated strains are presented. Only one symbol is presented if all four strains showed a uniform response, otherwise the different responses are presented separated by a slash. −, Negative; +, positive; w, weakly positive;

Assimilation of	MWH-MoIso2T (DSM 21490T)	MWH-CaK1 (DSM 21494)	MWH-VicM1 (DSM 21486)	P. necessarius subsp. asymbioticus
Urea	+	−	+	+/−
Thiosulfate	+	−	+	+/−
Formic acid	−	−	−	w/−
Glyoxylic acid	−	w	−	w/−
Glycolic acid	−	−	−	−
Acetic acid	+	+	+	+
Oxalic acid	−	−	−	−
Propionic acid	w	+	+	+/−
Pyruvic acid	+	+	+	+
Malonic acid	−	+	+	+/−
Oxaloacetic acid	+	+	+	+/−
Malic acid	+	+	+	+/w
Succinic acid	+	+	+	+
Fumaric acid	+	+	+	+/w
Levulinic acid	w	−	−	w/−
Citric acid	−	+	−	−
D-Mannose	−	w	−	w/−
D-Glucose	−	w	−	w/−
D-Galacturonic acid	+	+	w	w
D-Galactose	w	w	w	w/−
D-Lyxose	−	−	−	w/−
D-Fructose	−	−	−	w/−
D-Fucose	−	−	−	w/−
D-Sorbitole	−	−	−	w/−
L-Glutamate	−	w	−	+/−
L-Aspartate	−	−	−	+/−
L-Cysteine	+	+	+	+/w
L-Alanine	+	w	+	w/−
L-Serine	−	−	−	−
L-Asparagine	−	−	−	w/−
Betaine	−	−	−	w/−

The five investigated strains have with P. necessarius subsp. necessarius in common that they produce only low biomass yields in relatively rich media (Hahn et al., in press). Biomass yields are usually one order of magnitude lower than in other Burkholderiaceae. The results of the chemotaxonomic and phenotypic characterization of the investigated strains are presented in Tables 1, 2, and 3. Experiments on utilization capabilities of three strains resulted only in 68% of the 37 tested substrates in identical results for all three strains (three categories; i.e. no, weak and good utilization) (Table 2). Testing of four P. necessarius subsp. asymbioticus strains with the same substrates even yielded only in 25% of tested substrates identical results (Hahn et al., in press). Despite of the relatively large number of tested substrates no discriminative traits between the three investigated strains and the three P. necessarius subsp. asymbioticus strains were found (Table 2). Therefore, utilization patterns of the remaining two strains, i.e. MWH-TaW3 and MWH-NZ8W13, were not examined. All strains grew in NSY medium as curved rods with small cell sizes (Fig. 1). Doubling times for growth in liquid NSY medium at 20°C were for strains MWH-VicM1, MWH-MoIso2^T, and MWH-CaK1 4.5, 5.7, and 5.9 hours, respectively. All three strains grew under these conditions with average cell volumes of < 0.1 μm³ (Hahn, 2003).

Table 3.

Whole cell fatty acid composition of P. cosmopolitanus sp. nov. strains as compared to the type strain of P. necessarius subsp. asymbioticus, QLW-P1DMWA-1^T, (Hahn et al., in press). Values are percentages of the summed fatty acids named in the peak library of the MIDI system (contents > 0.2%). Strains were grown on R2A agar plates for 3-5 days at 28° C.

Fatty acid	MWH- MoIso2T	MWH- CaK1	MWH- VicM1	MWH- TaW3	MWH- NZ8W13	QLW- P1DMWA- 1T
Saturated
Cl2:0	-	-	-	-	-	3.4
Cl4:0	0.7	0.6	0.9	2.3	1.3	0.9
Cl5:0	0.2	-	0.2	-	-	0.3
Cl6:0	15.4	11.0	11.1	14.9	13.5	22.2
Cl8:0	0.8	0.7	0.5	1.1	1.1	1.2
Unsaturated
C14:1 ω5C	0.6	0.4	0.6	-	1.1	-
C16:1 ω5c	0.3	0.5	0.5	1.1	-	0.9
C17:1 ω6c	0.5	0.3	0.7	-	-	-
C18:1 ω9C	0.3	-	0.2	2.0	-	-
Cl8:1 ω7C	28.7	37.5	35.6	33.7	38.1	12.9
C18:1ω5c	0.2	0.2	0.2	-	-	-
1 1Methyl-C18:1 ω7c	3.7	2.2	0.4	1.0	2.3	3.1
Hydroxylated
C12:0-2OH	-	-	-	-	-	2.5
C12:0-3OH	11.1	9.9	7.1	11.2	10.5	-
Summed features
Feat. 1 (C12:0 ALDE?)	0.1	-	0.2	-	-	0.4
Feat. 2 including C14:0-3OH	0.6	0.7	3.9	0.8	0.6	9.6
Feat. 3 including C16:1 ω7c	34.7	35.4	36.5	32.0	31.5	41.3
Feat. 7 including C19:1 ω6c	1.5	0.7	0.6	-	-	0.4

Fig. 1.

Transmission-electron survey view of Pt-C shadow-cast cells of strain MWH-MoIso2^T. The arrowheads indicate the direction of shadow casting.

Major components of whole cell fatty acids were C_16:1ω7c, C_18:1ω 7c, C_16:0 and C_12:0-3OH (Table 3). The pattern of P. necessarius was dominated by the same non-hydroxylated acids (Table 3 and Hahn et al., in press). The occurrence of several other fatty acids in minor amounts was shared by P. necessarius and all studied strains of P. cosmopolitanus. These are C_14:0, C_18:0, and 11-methyl-C_18:1ω7c.

On the other hand, the two species could be clearly distinguished by the absence of C_12:0 and C_12:0-2OH and the presence of high amounts (higher than 7 %) of C_12:0-3OH in P. cosmopolitanus. In all P. necessarius strains studied yet, no C_12:0-3OH was detected, and the contents of C_12:0 and C_12:0-3OH were higher than 3 % and 1 %, respectively (Hahn et al., in press). Furthermore, MIDI-feature 2 including C_14:0-3OH occurred in P. cosmopolitanus strains in much lower amounts (lower than 4 %) than in P. necessarius strains (higher than 9 %), and the fatty acid C18:ω7c amounted higher than 28 % in P. cosmopolitanus while it amounted less than 21 % in P. necessarius.

The genome size of the five investigated strains was not determined, however, the genome sizes of strains tentatively assigned to the proposed taxon are ranging from 1.9 to 2.1 Mbp (Vannini et al., 2007), which may indicate that P. cosmopolitanus strains possess smaller genome sizes than P. necessarius subsp. asymbioticus strains.

Phylogeny

The phylogenetic analyses of 16S rRNA genes of the five strains revealed clustering of the strains in a monophyletic clade representing a sibling taxon of P. necessarius (Fig. 2). The minimal 16S rRNA sequence similarity of the five strains is 99.3 % (sequence stretches ranging from E. coli position 28 to 1542 were analysed). The sequence similarity between 16S rRNA genes of the proposed type strain MWH-MoIso2^T and the type strain of P. necessarius subsp. asymbioticus and a sequence representing P. necessarius subsP. necessarius ‘E24’ are 97.3% and 96.8%, respectively. BLAST searches with the 16S rRNA sequences of the investigated strains and subsequent phylogenetic analyses revealed that the five investigated strains form together with 99 cultured and uncultured strains a tight monophyletic cluster (data not shown). This cluster is phylogenetically identical with the previously characterized ‘Polynucleobacter subcluster D’ (Hahn, 2003) for which a minimal within-cluster sequence similarity of 99.1% was reported previously (Wu and Hahn, 2006). Phylogenetic analysis of 16S-23S ITS sequences confirmed the separate clustering of the five investigated strains and previously investigated P. necessarius strains, and demonstrated tight clustering of sequences within both groups, respectively (Fig. 3).

Fig. 2.

Neighbour-joining (NJ) tree based on almost complete 16S rRNA gene sequences, reconstructing the phylogenetic position of the five investigated strains within the family Burkholderiaceae. Bar indicates 0.02 substitutions per nucleotide position. A maximum likelihood (ML) tree (not shown) calculated with the same sequence set differed from the NJ tree in the positions of B. pertussis and L. thiooxidans. Bootstrap values obtained by the NJ (first value) and the ML algorithm (second value) are shown if at least one value was > 60 %. PnecA, ‘Polynucleobacter subcluster A’; PnecB, ‘Polynucleobacter subcluster B’.

Fig. 3.

Neighbour-joining (NJ) tree based on complete 16S-23S ITS sequences (approx. length 500 bp). Bar indicates 0.05 substitutions per nucleotide position. The five P. cosmopolitanus sp. nov. strains are shown in bold face. Sequences of strains affiliated with the ‘Polynucleobacter subclusters A’ (PnecA) and B (PnecB) were included in this analysis. A maximum likelihood tree (not shown) calculated with the same sequence set confirmed the tight clustering of the P. cosmopolitanus sp. nov. strains, as well as the tight clustering of the P. necessarius strains (bootstrap value 100%). Bootstrap values obtained by the NJ (first value) and the ML algorithm (second value) are shown if at least one value was > 60 %.

Genotypic traits

All members of the ‘Polynucleobacter subcluster D’ share the presence of the oligonucleotide sequence 5′-AA(T/G) CCC T(A/T)A GGG GGA AA-3′ within the 16S rRNA gene (E. coli. position 181-197). In almost all cases, the presence of this sequence can be determined by fluorescent in situ hybridization of whole cells by using a mixture of the fluorescently labelled oligonucleotide probes PnecD1-181 and PnecD2-181 (Hahn et al., 2005). The sequences of these two probes differ in the second variable position (E. coli. position 188) of the diagnostic sequence, but not in the first variable position (E. coli. position 183). Analysis of 104 sequences forming the phylogenetic cluster representing the proposed taxon P. cosmopolitanus demonstrated that only one strain differed in the first variable position of the diagnostic sequence from the probe sequences. Besides the FISH probe, a reverse line blot hybridization probe (PolynucD123-144) for detection of the proposed taxon in environmental samples has been developed and successfully applied (Wu et al., 2007).

Biogeography of P. cosmopolitanus sp. nov

Strains affiliated to the phylogenetic cluster represented by the proposed species were isolated from freshwater habitats located in Eurasia, South America, Africa, and Oceania (Hahn, 2003; Hahn et al., unpublished data). Even from a freshwater habitat located on the Hawaiian Archipelago a strain could be isolated (strain MWH-Haw2W10a, accession number AM110098). Recently, many strains were isolated from various lakes located in Japan (Watanabe et al., 2009). Furthermore, cultivation-independent investigations indicated the presence of this taxon at various places including Northern America (Crump et al., 1999; Zwart et al., 2002; Simpson et al., 2004; Crump & Hobbie, 2005), a river estuary of northern Taiwan (Liao et al., 2007), and a running water in Germany (Beier et al., 2008). Even in a pond located at the Tibetan Plateau at an altitude of 3210 m.a.s.l the phylogenetic group representing the proposed species was detected by cultivation-independent methods (Wu et al., 2006). The geographic distribution of habitats colonized by strains phylogenetically belonging to the proposed new species clearly indicates a cosmopolitan distribution of the new taxon over all continents and all climatic zones.

Ecology of P. cosmopolitanus sp. nov

Members of the phylogenetic cluster represented by the proposed species were isolated and detected in the water columns of running and stagnant freshwater habitats like ponds, lakes and streams. A recent investigation compared the salinity tolerance determined by ecophysiological experiments with six pure culture strains of the proposed taxon with the occurrence of the taxon along a salinity gradient of sixteen lakes (Wu et al., 2006). This gradient ranged from 0.2 g L⁻¹ (freshwater) to 222 g L⁻¹ salinity (hypersaline). While the tested strains grew in the laboratory up to maximal salinities of 3.6 to 5.6 g L⁻¹ (note that the salinity tolerance considers the total amount of dissolved salt, while the NaCl tolerance mentioned above considers exclusively the NaCl added to the medium but not the other salt contained in the medium), the taxon was in situ detected by cultivation-independent methods exclusively in habitats with salinities < 1 g L⁻¹. On the other hand, the new taxon was detected in a metagenomic study in two estuaries, and for one of these estuaries (Chesapeake Bay, MD) a salinity of 3.5 g L⁻¹ (oligosaline) was reported (Shaw et al., 2008). The detection at terrestrial or offshore marine sites was never reported. All these observations indicate that the new taxon mainly inhabits the water column of freshwater habitats but may also occur in oligosaline habitats, and is absent in eusaline (oceans) and hypersaline environments.

The new taxon could be detected by the FISH probes PnecD1-181/PnecD2-181 in several habitats (Wu and Hahn, 2006, Wu and Hahn, unpublished data). In a survey of 60 freshwater lakes, the proposed taxon was detected by FISH in 15 habitats. Usually, detections ranged from 0 % (below detection limit by FISH) to 0.9 % of total prokaryotic numbers. The highest recorded relative abundance of 8.1 % of total prokaryotic cells was observed in a freshwater pond located at the Tibetan Plateau (see above). All cells detected by the probes in the various investigated habitats possessed morphologies of curved rods. In a seasonal study on shallow, eutrophic Taihu Lake located in subtropical China, two population peaks both with 0.3 % relative abundance (equalling 1 × 10⁴ cells mL⁻¹) in winter and late spring could be observed (Wu and Hahn, 2006).

In a study focusing on the vulnerability of strains affiliated with the new taxon to predation by bacterivorous flagellates a relatively low vulnerability was observed (Boenigk et al., 2004). Predation by such flagellates represent in general a major mortality factor of planktonic freshwater bacteria (Hahn & Höfle, 2001). The revealed low vulnerability was explained by the small cell sizes of the strains (< 0.1 μm³, ultramicrobacteria), however, a complete resistance to predation was not observed. Thus, factors determining the occurrence and relative abundance of the new taxon in various freshwater habitats remain to be revealed.

Results from the phylogenetic analysis and chemotaxonomic investigations demonstrated pronounced differences between the five investigated strains and strains of the species P. necessarius (Figs 2 and 3, Table 4). Furthermore, the determined 16S rRNA gene sequence similarities between the type strain of the proposed new species and strains representing the two P. necessarius subspecies of ≤ 97.3 % leads us to the conclusion that strain MWH-MoIso2^T represents a new species (Stackebrandt & Ebers, 2006). Coherence of the taxon represented by the five investigated strains could not be checked by DNA-DNA hybridization experiments, because of limitations in production of DNA amounts sufficient for such investigations. However, the phylogenetic analysis of 16S rRNA gene sequences and of the much less conserved 16S-23S ITS sequences of Polynucleobacter spp. strains (Fig. 3) strongly indicates that the five investigated strains represent a coherent taxon. The latter phylogenetic marker is much less conserved than 16S rRNA gene sequences, thus significant intra-taxon heterogeneity would be expected to be indicated by this marker. We propose to establish for strain MWH-MoIso2^T the new species Polynucleobacter cosmopolitanus sp. nov., and to assign tentatively the other four investigated strains to this new species.

Table 4.

Discriminative characteristics of P. cosmopolitanus sp. nov. and P. necessarius subspecies asymbioticus. Note that only an incomplete phenotypic and chemotaxonomic description of P. necessarius subsp. necessarius is available because of its obligately endosymbiotic lifestyle.

	P. cosmopolitanus sp. nov.	P. necessarius subsp. asymbioticus
Fatty acid C12:0	not detected	> 3 %§
Fatty acid C18:1ω7c	> 28 %§	< 21 %§
Fatty acid C12:0-2OH	not detected	> 1 %§
Fatty acid C12:0-3OH	> 7 %§	not detected
Fatty acid feat. 2 (including C14:0-3OH)	< 4 %§	> 9 %§
Probe PnecC-16S-445#	negative	positive
Probe PnecD-181*	positive	negative

^§Percentages of the summed fatty acids named in the peak library of the MIDI system

^#Hahn et al., 2005

^*equal mixture of probes PnecDl-181 and PnecD2-181 (Hahn et al., 2005)

Description of Polynucleobacter cosmopolitanus sp. nov

Polynucleobacter cosmopolitanus (cos.mo.po.li’ta.nus. Gr. adj. kosmopolites, an inhabitant of the world, a cosmopolitan; N.L. masc. adj. cosmopolitanus cosmopolitan). Curved, non-motile rods, 0.5–1.4 μm in length and 0.3–0.5 μm in width. Chemoorganotrophic, aerobic strains, at least some of the strains are able to grow anaerobically. Can be cultivated on NSY, R2A, and Bactopeptone medium, and some other complex media with concentrations of ≤ 3g L⁻¹. Colonies grown on NSY agar are unpigmented, circular and convex with smooth surface and reach a diameter of 1 mm after 5-10 days at 20-24° C. Growth occurs at 10–35 °C, four strains also grow at 5 °C, and two strain also grow at 38 °C. Grows without NaCl. Maximum NaCl concentration tolerated is strain dependant 0.3 - 0.5 % (w/v). Utilizes acetate, pyruvate, propionate, oxaloacetate, malate, succinate, fumarate, D-galacturonic acid, D-galactose, L-cysteine, and L-alanine when these substrates are provided in a medium containing low amounts of NSY. None of the strains utilizes formiate, glycolate, oxalate, D-lyxose, D-fructose, D-fucose, D-sorbitole, L-aspartate, L-serine, L-asparagine, and betaine. Major cellular fatty acids are C_18:1ω7c, C_16:0 and C_12:0-3OH. No 2-hydroxylated compounds detected. The DNA G+C content of the type strain is 44.9 %. Fractions of natural populations pass through 0.2μm filters. Can be distinguished from P. necessarius by using the P. necessarius-specific FISH probe PnecC-16S-445 (Hahn et al., in press), as well as by differences in their fatty acids profiles (Table 3). In contrast to P. necessarius, exclusively free-living strains and no endosymbiotic strains were reported, however, the existence of endosymbiotic strains affiliated to P. cosmopolitanus sp. nov. cannot be completely excluded at present. Tentatively, all strains possessing the oligonucleotide sequence 5′-AA(T/G) CCC T(A/T)A GGG GGA AA-3′ (E. coli. position 181-197) within the 16S rRNA gene shall be assigned to the proposed species. In almost all cases (one known exception), the presence of the diagnostic sequence can be determined by FISH with a mixture of the two probes PnecD1-181 and PnecD2-181 (Hahn et al., 2005). The type strain is MWH-MoIso2^T (= DSM 21490^T = CIP 109840^T), isolated from the water column (pelagic zone) of Lake Mondsee located in the Salzkammergut Lake District in Austria (Hahn, 2003).

Abbreviations

FISH

Fluorescent in situ hybridization

PFGE

pulsed field gel electrophoresis

Mbp

mega base pairs

OD

optical density