Phylogeny and in situ identification of a morphologically conspicuous bacterium, Candidatus...

Phylogeny and in situ identification of a morphologicallyconspicuous bacterium, Candidatus Magnospira bakii,present at very low frequency in activated sludge

Jiri Snaidr, 1,2† Bernhard Fuchs, 1,2 Gunter Wallner, 3

Michael Wagner, 1 Karl-Heinz Schleifer 1 and RudolfAmann 1,2*1Technische Universitat Munchen, Lehrstuhl furMikrobiologie, Arcisstr. 16, D-80290 Munich, Germany.2Max-Planck-Institut fur Marine Mikrobiologie, Celsiusstr.1, D-28359 Bremen, Germany.3GSF-Forschungszentrum fur Umwelt und Gesundheit,Durchflußzytometrie, D-85764 Neuherberg, Germany.

Summary

A morphologically conspicuous bacterium that consti-tuted a very small fraction ( <0.01%) of the total micro-bial community of activated sludge was enriched andanalysed phylogenetically by a combination of culti-vation-independent molecular and physical methods.The large, corkscrew-shaped, filamentous bacteriawere first detected in municipal activated sludge bylight microscopy owing to their unusual rotating glid-ing motility. Various attempts at microbiologicalenrichment and pure culture isolation with traditionaltechniques failed, as did attempts to retrieve the mor-photype of interest by micromanipulation. In situhybridization with the group-specific, rRNA-targetedoligonucleotide probe CF319a indicated a phyloge-netic affiliation to the Cytophaga–Flexibacter groupof the Cytophaga –Flavobacterium–Bacteroides phy-lum. Based on strong morphological resemblance tomembers of the genus Saprospira , additional 16SrRNA-targeted oligonucleotides with more narrow spe-cificity were designed and evaluated for in situ hybridi-zation to the morphotype of interest. Flow cytometriccell sorting based on the fluorescence conferred byprobe SGR1425 and forward scatter enabled a physi-cal enrichment of the helical coiled cells. Subsequentpolymerase chain reaction (PCR) amplification of 16SrDNA fragments from whole fixed sorted cells with aprimer pair based on probes CF319a and SGR1425

resulted in the retrieval of 12 almost identical partial16S rDNA fragments with sequence similarities amongeach other of more than 99.2%. In situ hybridizationsproved that the sequences that showed the highestsimilarity (88.4%) to the 16S rRNA of Saprospiragrandis were indeed retrieved from the corkscrew-shaped filaments. The bacterium is likely to be a mem-ber of a genus of which no species has been culturedhitherto. It was consequently tentatively named ‘Mag-nospira bakii’ and has the taxonomic rank of Candida-tus Magnospira bakii, as the ultimate taxonomicplacement has to await its cultivation. In this study,it was demonstrated that even bacteria occurring atvery low frequencies in highly complex environmen-tal samples can be retrieved selectively without culti-vation for further molecular analysis.

Introduction

The real extent of bacterial diversity remained invisible fordecades, because bacterial identification relied largely oncultivation-dependent methods. Pure cultures isolated fromclinical and environmental samples were classified on thebasis of their biochemical or physiological parameters.They formed the basis for the production of the onlytools for in situ identification, antibodies that were applied,for example, in immunofluorescence assays (Bohlool andSchmidt, 1980). Consequently, bacterial population analy-sis targeted mainly microorganisms that grew readily onartificial nutrient media. However, it has long been recog-nized that only a minority of the microorganisms present innatural and engineered environmental samples can becultivated by standard techniques (Wagner et al., 1993;Amann et al., 1995). Molecular studies indicate an aston-ishing microbial diversity (e.g. Torsvik et al., 1990; Amannet al., 1996). It is the rRNA approach to microbial ecologyand evolution (Olsen et al., 1986; Pace et al., 1986) thatallows the analysis of this high natural microbial diversityin more detail today. Microorganisms can be analysedphylogenetically and identified in situ without cultivationby a combination of polymerase chain reaction (PCR)-assisted rDNA retrieval, comparative sequence analysisand fluorescence in situ hybridization (Amann et al.,1995). Yet, this molecular analysis remained restricted tothe more dominant populations, as the screening and

Environmental Microbiology (1999) 1(2), 125–135

Q 1999 Blackwell Science Ltd

Received 24 August, 1998; revised 11 November, 1998; accepted 16November, 1998. †Present address: Vermicon Engineering andMicrobiology Aktiengesellschaft, Nymphenburger Strasse 81, D-80636 Munich, Germany. *For correspondence at the Max-Planck-Institut, Bremen. E-mail [email protected]; Tel. (þ49) 4212028 930; Fax (þ49) 421 2028 690.

sequencing of numerous clones from a 16S rDNA genelibrary still is quite laborious. Usually fewer than 1000clones are analysed, which leaves the general rRNAapproach unsuitable for the characterization of popula-tions with a low relative abundance. It might be arguedthat the less abundant populations are of low importancefor the overall functioning of an ecosystem. Nevertheless,the very existence of morphologically conspicuous, yetinfrequent bacteria catching the eye of early or modernmicrobe hunters and evading all attempts at pure cultureisolation should be justification enough for developingapproaches for their study. After all, they representunknown bacterial diversity, and their genes and gene pro-ducts are of potential biotechnological interest.

Here, we describe the 16S rRNA-based classificationand in situ identification of such a bacterium, a cork-screw-shaped filament with an unusual rotating glidingmotility that was occurring in low frequencies of less than0.01% in activated sludge. In the course of our study, dif-ferent traditional and modern enrichment strategies weretested: the classical enrichment in liquid media; micro-manipulation of bacteria as pioneered by Skerman (1968)and the use of optical tweezers; and flow cytometric cellsorting. Optical tweezers (Ashkin et al., 1987; Ashkin andDziedzic, 1987) were used successfully by Huber et al.(1995) to isolate staphylococcoid aggregates of Archaeathat made up 0.01% of enrichment dominated by fila-mentous Thermofilum- and Thermoproteus-like cells. Wall-ner et al. (1997) have recently shown the potential of flow

cytometry in sorting cells from activated sludge, lakewater and lake sediment for molecular analysis accordingto differences in light scattering, DNA content and affilia-tion to certain phylogenetic groups, as assessed by fluor-escein-labelled, rRNA-targeted oligonucleotide probes.

Results

Microscopic analysis and cultivation

Live mounts of activated sludge from the two large muni-cipal wastewater treatment plants in Munich contained ahighly conspicuous, large, corkscrew-shaped filament atfrequencies of less than 0.01%. The pitch of the helixwas approximately 2.3 mm. The total length of the cellsranged between 10 mm and 30 mm. The bidirectional glid-ing motility with a speed of up to 5 mm s¹1 was associatedwith a fast rotation of the helix (a video clip can be viewedat http://www.blackwell-science.com/emi). The helixshowed high flexibility. DAPI staining revealed that thehelices actually consisted of chains of curved rods with adiameter of 0.8 mm and a length of approximately 2 mm.In situ hybridization of the original sample with fluores-cently labelled, group-specific oligonucleotide probesshowed binding of the filaments with the bacterial probeEUB338 and CF319a, a probe comprising members of theCytophaga–Flexibacter group of the phylum Cytophaga–Flavobacterium–Bacteroides (Woese, 1985). Cells hybri-dizing to probe CF319a were abundant in the two activated

Q 1999 Blackwell Science Ltd, Environmental Microbiology, 1, 125–135

Table 1. Oligonucleotide probes.

Probe Specificity Sequence (58-38) of probe Targeta site [%] Reference(rRNA positions) formamideb

EUB338 Most Bacteria GCTGCCTCCCGTAGGAGT 16S, 338–355 0 Amann et al. (1990)CF319a Cytophaga–Flexibacter group of CFB TGGTCCGTGTCTCAGTAC 16S, 319–336 35 Manz et al. (1996)

phylumALF1b Alpha-subclass of Proteobacteria, CGTTCGYTCTGAGCCAG 16S, 19–35 20 Manz et al. (1992)

some members of the delta-subclassof Proteobacteria

BET42a Beta-subclass of Proteobacteria GCCTTCCCACTTCGTTT 23S, 1027–1043 35 Manz et al. (1992)GAM42a Gamma-subclass of Proteobacteria GCCTTCCCACATCGTTT 23S, 1027–1043 35 Manz et al. (1992)HGC69a Gram-positive bacteria with a high DNA TATAGTTACCACCGCCGT 23S, 1901–18 35 Roller et al. (1994)

G þ C contentMka220 Bacteria affiliated to clones within the CGCACATCTATCCTCAAGC 16S, 220–237 20 This study

clusters I, II, III and IVMKb221 Bacteria affiliated to clones within the CGCACACCGATCTTTAAC 16S, 221–238 20 This study

clusters I, II, III and IVMKc219 Bacteria affiliated to clones within the CATGCTCATCTTTTGCCG 16S, 219–236 20 This study

clusters I, II, III and IVMKd1111 Bacteria affiliated to clones within the GCAACTAACAACAGGGGT 16S, 1111–1128 20 This study

clusters I, II, III and IVSGR220 Saprospira grandis CGCACACTCATCTTCATCC 16S, 220–217 20 This studySGR1200 Saprospira grandis and relatives ATCATGGCCTTTATGCCC 16S, 1200–1217 0 This studySGR1425 Saprospira grandis and relatives GTCACCATCTTCAGGCAC 16S, 1425–1442 20 This studyBak655 Candidatus Magnospira bakii CGCCAACCCTAACCCATTT 16S, 655–673 0 This studyBak1408 Candidatus Magnospira bakii TCCAAACTCCCATGGCTT 16S, 1408–1425 0 This study

a. Escherichia coli numbering (Brosius et al., 1981).b. Formamide concentration (FA) in the in situ hybridization buffer (v/v).

126 J. Snaidr et al.

sludge samples examined, accounting for approximately12% of the total cells of 1.8 ×109 ml¹1 (Snaidr et al.,1997). The conspicuous bacterium did not hybridize withthe other group-specific oligonucleotides applied (Table 1).Nutrient media developed for the cultivation of membersof the Cytophaga–Flexibacter group (Balows et al., 1992)were applied but did not support significant growth of thecorkscrew-shaped filaments. Cultivation parameters includ-ing light, turbulence, oxygen and temperature were varied,but did not result in reasonable growth of the cells of inter-est. The only slight enrichment achieved, roughly a doubl-ing, followed incubation of the original activated sludge inclosed vials without agitation at 88C.

Based on morphology, the genus Saprospira seemedto resemble the unknown conspicuous bacterium inmany ways (Reichenbach, 1980). This genus of the

Cytophaga–Flexibacter group encompasses severalmarine and freshwater species, of which only Saprospiragrandis can be obtained from public strain collections.Comparative 16S rRNA analysis had shown an affiliationof S. grandis to Haliscomenobacter hydrossis (Woese etal., 1985). A specific 16S rRNA-targeted oligonucleotideprobe, SGR220 (Table 1), was developed that hybridizesto cells of S. grandis in situ. Hybridizations of activatedsludge with this probe revealed no signals, indicatingthat the corkscrew-shaped filaments were not S. grandis.

Cloning and comparative sequence analysis of the16S rDNA of members of the Cytophaga–Flexibactergroup

Several 16S rRNA gene libraries constructed directly from


Fig. 1. Phylogenetic tree showing therelationships of 16S rDNA sequencesretrieved with primer pair 8F and SGR1425with their closest relatives among theCytophaga–Flexibacter group of theCytophaga–Flavobacteria–Bacteroidesphylum. The tree is based on the results ofdistance matrix analyses of all available 16SrRNA primary structures for members of theCytophaga–Flavobacteria–Bacteroidesphylum. The topology of the tree wasevaluated by performing maximum parsimonyand maximum likelihood analyses of the fulldata set and subsets respectively. Only atleast 90% complete sequences were used fortreeing. Alignment positions at which less then50% of sequences of the entire data set sharethe same residues were excluded from thecalculations. The phylogenetic positions of theclones were reconstructed by applying theparsimony criteria without changing theoverall tree topology. The four clustersdetected are numbered from I to IV. Numbersbehind the clone names indicate the length ofthe analysed sequence. The bar indicates10% estimated sequence divergence.

Candidatus Magnospira bakii 127

the activated sludge with a widely used ‘universal’ 16SrRNA PCR primer set, 8F/1546R, did not contain sequencesrelated to the 16S rRNA of members of the Cytophaga–Flexibacter group (Snaidr et al., 1997; J. Snaidr et al.,unpublished), even though cells affiliated to this group con-tributed 12% of the total population. Consequently, a new,more specific PCR primer set was constructed to restrictthe 16S rRNA fragments that were amplified to a subgroupof the Cytophaga–Flexibacter group. Potential 38 primerbinding sites were examined by in situ hybridization withprobes designed to target the 16S rRNA of Saprospiragrandis and additional members of the Saprospira–Halis-comenobacter group. Unlike SGR220, for which a currentprobe check only indicates complementarity to S. grandis,probe SGR1200 is also complementary to the 16S rRNAof Flexibacter elegans, Flexibacter sancti, Cytophagaarvensicola, Haliscomenobacter hydrossis and Flexibacterpolymorphus, and SGR1425 to Haliscomenobacter hydros-sis and Flexibacter polymorphus. Whereas SGR1200 didnot hybridize with the corkscrew-shaped filaments,SGR1425 hybridized to them but also to various othermorphotypes in the activated sludge. Only about 1% ofcells detected with probe SGR1425 had the typical mor-phology. Subsequently, the probe SGR1425 was usedas a PCR primer together with a primer targeting a uni-versal site at the 58 end of bacterial 16S rDNA. Screeningof the resulting clones with digoxigenin-labelled probeCF319a indicated that all contained 16S rDNA affiliatedto the Cytophaga–Flexibacter group. Forty-three randomlychosen clones were partially sequenced (600–800 nucleo-tides) from either the 58 end or the 38 end and analysedphylogenetically. Phylogenetic analysis of the retrievedsequences revealed four clusters (Fig. 1). Cluster I wasaffiliated to the genus Flavobacterium, cluster II containedclones with a high similarity to Microscilla aggregans, clus-ter III was affiliated to Haliscomenobacter hydrossis andcluster IV sequences were more distantly related to thoseof the Saprospira–Haliscomenobacter group. Four newoligonucleotide probes were developed to cover allsequences (Table 1). In situ hybridizations of activatedsludge with these probes resulted in the detection of vari-ous morphotypes that also hybridized with probes CF319aand SGR1425. However, they did not bind to the cork-screw-shaped filaments.

Physical enrichment of the corkscrew-shapedfilaments

As attempts at traditional microbiological enrichment of theconspicuous bacteria had failed, different techniques wereevaluated for physical enrichment. The first two approacheswere microscopic, with cell retrieval based solely on theunique morphology of the corkscrew-shaped filaments.Initially, a motorized micromanipulator attached to an

inverse microscope was used to position a capillary forthe collection of cells of interest. Subsequently, opticaltweezers were tested for their ability to retrieve the cork-screw-shaped filaments. Both techniques failed. One rea-son was that the target cells resided mainly within denseactivated sludge flocs. Furthermore, the concentration ofnon-target cells was so high that they were constantlyco-collected by or passively attaching to the capillary orthe laser beam. Owing to the low abundance of targetcells, it was impossible to dilute the activated sludge further.

Finally, flow cytometry was evaluated. An activatedsludge sample was used that had been enriched slightlyfor the corkscrew-shaped filaments by incubation at lowtemperature for several weeks. After fixation with para-formaldehyde and subsequent hybridization with probeSGR1425 in suspension, the sample was analysed by flowcytometry. The combination of forward scatter and fluores-cent signal conferred by probe SGR1425 resolved a distinctcell population on which a sorting gate was set (Fig. 2).Sorting resulted in a cell suspension in which about one-third of the cells were corkscrew-shaped filaments. Fiveto 10 thousand cells could be sorted in less than 30 min.

Phylogeny and in situ identification of the corkscrew-shaped filaments

Partial 16S rDNA fragments of 1122 nucleotides in lengthwere amplified from the sorted cells using a combination ofthe primers C319S, a shortened derivative of probeCF319a, and SGR1425. Earlier, amplification with the


Fig. 2. Dot plot (x-axis, forward scatter; y-axis, relative probefluorescence) showing the polygonal gate used for the sorting oflarge, probe SGR1425-positive cells from activated sludge. Eachdot represents a single event recorded in list mode. No data areshown below a threshold in forward scatter of 6.


universal 16S rDNA primers 8F/1546R had failed. The com-parative sequence analysis of 12 randomly chosen clonesrevealed sequence similarities between 99.2% and 100%.All clones showed the highest similarity of up to 88.4%to S. grandis (Fig. 3). Two specific oligonucleotide probes,Bak655 and Bak1408 (Table 1), were developed to checkby in situ hybridization whether the sequences originatedfrom the corkscrew-shaped filaments. Both fluorescentlylabelled probes hybridized exclusively to the corkscrew-shaped filaments (Fig. 4).

In situ hybridization was combined with laser scanningmicroscopy to study the probe-positive population at highresolution. The detected cells preferred deeper regionsof the activated sludge flocs (Fig. 4C). All cells that hybri-dized with one of the two specific probes also bound thesecond probe (Fig. 4D), indicating a homogeneous popu-lation. All cells detected had a very similar diameter ofabout 1 mm but, in addition to the typical highly coiled,corkscrew-shaped filaments, more relaxed helices andsingle curved rods also hybridized (Fig. 5). The singlecurved rods probably represent cells released from thefilaments. Reichenbach (1980; 1992) observed similardifferences in the degree of helicity in pure cultures ofSaprospira grandis, in which loss of helical arrangementcoincided with loss of motility.

Discussion

Technical considerations

The principle goal of this study, the demonstration that amorphologically conspicuous, hitherto uncultured bacterium

present at low abundance in a very complex engineeredenvironment such as activated sludge can be retrievedfor molecular analysis, has been achieved. A combinationof physical and molecular enrichment steps were required.Even traditional enrichment contributed to the ultimatesuccess, as an activated sludge sample was used forcell sorting that had been stored for several weeks atlow temperature and contained slightly more corkscrew-shaped filaments. The effect of this storage was, however,small compared with the enrichment factor of approxi-mately 104 achieved by fluorescence-activated cell sort-ing. The frequency of the cells of interest could, in thissingle step, be increased from < 0.01% to approximately30%. After the next step of molecular enrichment, PCRamplification with two group-specific primers, the 12 clonessequenced were almost identical and could be shown tooriginate in the corkscrew-shaped filaments.

The heterogeneities between the 12 sequences are low(< 1%) and could be artifacts. PCR starting from formalde-hyde-fixed material as used in this study has been reportedto have a lower fidelity (DeGiorgi et al., 1994). Small dif-ferences might also originate from slight sequence differ-ences between multiple rRNA genes of one organism(Nubel et al., 1996; Rainey et al., 1996). The complete over-lap of in situ hybridization by probes Bak655 and Bak1408is further evidence that no second closely related popula-tion appears in the examined sludge.

To assess the general applicability of our approach tothe directed retrieval of low-frequency bacterial popula-tions, it is helpful to take a closer look at the individualenrichment steps. The identification of the cells of interest


Fig. 3. Phylogenetic tree showing therelationships of the 16S rDNA sequencesretrieved from the corkscrew-shaped filamentswith the primer combination C319S andSGR1425 from the sorted cells with the 16SrRNA sequences of selected members of theCytophaga–Flexibacter group. For furtherdetails see Fig. 1. The bar indicates 10%estimated sequence divergence.


was based initially solely on the conspicuous morphotype.However, even for such large and morphologically highlyconspicuous cells as our filaments, morphology alone isnot a reliable criterion for sorting. First, as shown inFig. 4A and B, cells of interest might be difficult to detectin the environment and, secondly, a certain degree of

pleomorphy always exists (Fig. 5). This also contributedto the failure of micromanipulation with the capillary andoptical tweezer techniques. These techniques, which haveshown their high potential on other occasions, were notsuited for our case.

All subsequent enrichment procedures did not rely on


Fig. 4. In situ identification of the corkscrew-shaped filaments by fluorescence in situ hybridization.A and B. In situ hybridization of activated sludge with oligonucleotide probe Bak655 (CY3-labelled, red). Phase-contrast (A) andepifluorescence (B) micrographs are shown for identical microscopic fields.C. Simultaneous in situ hybridization with oligonucleotides EUB338 (fluorescein-labelled, green) and Bak655 (CY3-labelled, red). As thecorkscrew-shaped filaments bind both probes, the resulting colour is yellow.D. Simultaneous in situ hybridization of activated sludge with oligonucleotide probes Bak655 (CY3-labelled, red) and Bak1408 (fluorescein-labelled, green). All corkscrew-shaped filaments bind both probes, resulting in the yellow colour. The bar in (C) correlates to (A), (B) and (D).


morphology alone, but were based on the specific bindingof oligonucleotide probes to the cells of interest. Group-specific in situ probing indicated early on that the cork-screw-shaped filament was affiliated to the Cytophaga–Flexibacter group of the Cytophaga–Flavobacterium–Bacteroides phylum. This preliminary assignment washelpful in recognizing the morphological resemblance tothe genus Saprospira. A second round of in situ hybridiza-tion with three probes targeted to sites of the 16S rRNA ofS. grandis with an increasing evolutionary conservationthen indicated that the corkscrew-shaped filament wasnot S. grandis but that it shares at least one variable sitewith S. grandis, the target site of probe SGR1425. In con-trast to probe CF319a, which still detected 12% of all cells,SGR1425 hybridized to less than 1% of all cells and, moreimportantly, about 1% of the SGR1425-positive cells werecorkscrew-shaped filaments. As the corkscrew-shapedfilaments were much larger than most of the other hybri-dizing cells, flow cytometry could clearly resolve eventsthat had large forward scatter, a size-dependent para-meter and high fluorescence conferred by probe SGR1425.This was the basis for the tremendous enrichment achievedby flow cytometric cell sorting. The high sorting efficiencywas only possible because the combination of hybridiza-tion in suspension and flow cytometry had obviouslyresulted in a partial disintegration of activated sludgeflocs (Wallner et al., 1995).

From the highly enriched cell fraction after sorting, itshould have been possible to retrieve almost full-length16S rDNA of the corkscrew-shaped filaments with the

universal PCR primer set but, surprisingly, no amplifica-tion was achieved, although the amount of cells shouldhave been sufficient (Wallner et al., 1997). Clones relatedto the 16S rRNA sequences of members of the Cytophaga–Flexibacter were also absent in a library constructed fromactivated sludge with the same primers (Snaidr et al., 1997).We concluded that at least one of the two primers discrimi-nated strongly against our sequence of interest. Conse-quently, two primers were used for which we knew thetarget sites were present on the corkscrew-shaped fila-ments, CF319S and SGR1425. PCR amplification resultedin an almost homogeneous amplification product, as shownby the high similarity of the 12 sequenced clones.

The specific labelling of low-frequency bacterial popula-tions by means of fluorescence in situ hybridization is notrestricted to morphologically conspicuous bacteria. Anymicrobial population can, in principle, be labelled withfluorescent oligonucleotides targeted to signature siteson the rRNA molecules. The rRNA approach in combina-tion with cell sorting is, however, not only a tool for analys-ing the phylogeny of low-frequency populations in a directedway but, from a biotechnological perspective, it results in astrong enrichment of hitherto, resulting from their low fre-quency, almost inaccessible genes. With the tools of mole-cular biology, these genes can be retrieved and their geneproducts can be analysed for biotechnological potential.

Taxonomic considerations: proposal of CandidatusMagnospira bakii

Bacteria resembling those studied here have beenobserved on several occasions. In 1875, Van Tiegham(1880) observed bacteria on the sediment in an almostdried out mill run that consisted of long helical and tightlycoiled filaments. Locomotion occurred by active rotationof the filaments around the longitudinal axis. In the sameyear, Warming (1875) observed a bacterium on the Dan-ish coast, which was very large and helical, and he called itSpirochaeta gigantea. In 1911, Gross (1911) defined thenew genus Saprospira and described two species, S. grandisand S. nana. The genus Saprospira includes Gram-nega-tive, gliding, strictly aerobic bacteria, which form multi-cellular filaments with a length between 15 mm and 450 mmand a diameter of approximately 1 mm. The filaments consistof cylindrical cells with a length between 1 mm and 5 mm.The altitude of the helix is described as between 3 mmand 10 mm, and the width was 1.5–2.5 mm. Saprospirawas found in marine habitats (Gross, 1911; Lewin andLounsbery, 1969; Lewin and Mandel, 1970) and freshwater (Kolkwitz, 1909; Pringsheim, 1963; Lewin, 1965a)and favours eutrophic conditions (Ashton and Roberts,1987), as they also exist in activated sludge (Sladka andOttova, 1973). Habitat and phenotype, especially thecharacteristic mode of locomotion, suggest placing of


Fig. 5. Black and white confocal laser scanning micrograph of cellsdetected in activated sludge by in situ hybridization witholigonucleotide probe Bak655. Note that, besides the stronglyhelical filament, less coiled filaments and single bent cells alsohybridized with the probes. Bar ¼ 2 mm.


our corkscrew-shaped filaments within the genus Sapro-spira. Although many species of Saprospira have beendescribed, only one species, Saprospira grandis, can beobtained from public strain collections. This is probablybecause of the difficulty of cultivation, even though a spe-cific medium for S. grandis has been described (Reichen-bach, 1992).

Should the corkscrew-shaped filament now be classifiedas a new species of the genus Saprospira? We believenot. Despite some common phenotypical characteristics,the relatively low 16S rRNA similarity of below 90%(88.4% to Saprospira grandis) supports placement as anew species in a new genus. We would like to proposeMagnospira bakii, Magnospira to indicate the large sizeand helical shape of the bacterium and bakii after the Ger-man microbiologist, Friedhelm Bak, who had a great inter-est in the enrichment and cultivation of morphologicallyconspicuous bacteria. As no pure culture is availableand, according to Murray and Schleifer (1994), the officialtaxonomic status is that of Candidatus Magnospira bakii.The short description of Candidatus Magnospira bakii isas follows: phylogenetic position, member of the Sapro-spira–Haliscomenobacter group within the Cytophaga–Flavobacterium–Bacteroides phylum; not yet cultivated;Gram-staining, negative; morphology, corkscrew-shapedfilament, altitude of the helix approximately 2.2 mm; singlecells are bent rods with a diameter of about 0.8 mm and alength of approximately 2 mm; identification with the oligo-nucleotide probes Bak655 58-CGCCAACCCTAACCCA-TTT-38 and Bak1408 58-TCCAAACTCCCATGGCTT-38;authors, Snaidr et al., this study).

In situ monitoring and ecological aspects

The origin of motility in Candidatus Magnospira bakii andSaprospira is not yet fully understood. Both are surroundedby a slime belt, which, in the case of Saprospira, wasdescribed as consisting of different carbohydrates, includ-ing glucose, rhamnose, xylose, mannose and galactose(Lewin, 1965b). The breakage of a filament into two differentlarge pieces, first described in 1911 (Gross, 1911), couldalso be observed with cells of Candidatus Magnospirabakii (a video clip can be viewed at http://www.blackwell-science.com/emi.htm). Lewin (1997) observed Saprospiragrandis not only to live heterotrophically on dissolvedorganic substrate but also to subsist on other microbes,catching motile bacteria by their flagella before digestingthem externally. Small cells get trapped in the slime beltof the filament and are transported with the slime. Thesame was observed for Candidatus Magnospira bakii.The movement of the attached cells also provided evi-dence for bidirectionality of the gliding, which has alsobeen reported for other gliding bacteria (Pate and Chang,1979; Burchard, 1980; Bloodgood, 1990).

More detailed insights into the physiology and ecology ofCandidatus Magnospira bakii must await its cultivation. Asthe morphologies of bacteria grown on nutrient mediaoften differ significantly from the in situ morphologies,the two oligonucleotide probes specific for CandidatusMagnospira bakii might assist and direct the screeningprocess.

Experimental procedures

Sampling

Grab samples of mixed liquor were collected from the high-load aeration basins of two municipal sewage plants (MunichI, Grosslappen and Munich II, Dietersheim). For in situ hybri-dization, activated sludge was fixed for 3 h with paraformalde-hyde as described previously (Amann et al., 1990). For PCRamplification of 16S rDNA, unfixed aliquots of activatedsludge were frozen at ¹208C immediately after sampling.

Enrichment and cultivation

Liquid enrichments and plates with solidified media wereinoculated with various amounts of activated sludge. HD,Stokes, CY, LB, LEW3, LEW4, LEW5, WAT and SP6 mediumwere applied as described previously (Balows et al., 1992).Plates and liquid media were incubated at 88C, 148C, 238C,308C and 378C aerobically and microaerobically.

Oligonucleotide probes and DAPI staining

Fluorochrome- and digoxigenin-labelled oligonucleotide probeswere purchased from Interactiva. Probe sequences, hybridi-zation conditions and references are given in Table 1. Stainingwith 48,6-diamidino-2-phenylindole (DAPI) was performed asdescribed by Wagner et al. (1993).

In situ hybridization

Fixed sludge samples were immobilized on glass slides by airdrying and dehydrated in 50%, 80% and 96% (v/v) ethanol for3 min each. Fluorescence in situ hybridizations were per-formed as described recently by Snaidr et al. (1997).

For flow cytometric analysis, cells were hybridized in sus-pension in a buffer containing 0.9 M sodium chloride, 0.1%SDS, 20 mM Tris-HCl (pH 7.2) and 2 ng of oligonucleotide at468C (Wallner et al., 1993; 1995). After 3 h of incubation,cells were centrifuged, washed in hybridization buffer for0.5 h at 468C and resuspended in PBS containing 1 mM DAPI.

Confocal laser scanning microscopy

A Zeiss LSM 510 scanning confocal microscope (Carl Zeiss)equipped with a UV laser (351 nm, 364 nm), an Ar-ion laser(450–514 nm) and two HeNe lasers (543 nm, 633 nm) wasused to record optical sections. Image processing was per-formed with the standard software package delivered withthe instrument (version 1.5). Reconstructed and processed



images were printed using the software package MicrosoftPower Point (version 7.0) on a Kodak printer 8650 PS.

Micromanipulation

A Zeiss micromanipulator MR-mot. linked to an inverse epi-fluorescence microscope (Axiovert 135; Zeiss) was used toposition glass capillaries with tip diameters of 5–15 mm nextto the cells of interest. Subsequently, a slight vacuum wasapplied manually to transfer the cell into the capillary. Alterna-tively, a one-beam optical tweezer (S þ l) with a power of 450mW was used to singularize cells of interest.

Flow cytometry and sorting

Flow cytometric analysis and cell sorting were performed witha FACStar Plus (Becton Dickinson) as described by Wallneret al. (1997). The 488 nm emission line of an Ar-ion laserwith an output power of 500 mW was used for measuringforward-angle light scatter (FSC; 488 nm bandpass filter fordetection), as well as for right-angle light scatter (SSC;488 nm bandpass) and fluorescence of the fluorescein-labelled probes (530 nm bandpass). The collected data werestored in list mode as pulse height signals (four decades inlogarithmic scale each). In order to achieve both high purityand recovery, the Normal-R-modus was chosen for cell sort-ing. The drop drive frequency was set at approximately27 kHz, three drops were deflected simultaneously and drop-let delay was set between 12 and 15. For sorting, a lowsodium chloride concentration (0.1%) was used as sheathfluid to avoid the interference of salts with the PCR reaction.Sort criteria were defined by drawing polygonal gates inbivariate histograms (forward scatter versus probe fluores-cence) with the LYSIS II software package (Becton Dickinson).Fixed hybridized cells were first sorted directly onto micro-scopic slides to check whether the selected populationincluded the cells of interest and were subsequently collectedin sterile 1.5 ml reaction tubes. The salt concentration wasdecreased by a 30 min dialysis of the sorted cell suspensionon a cellulose acetate filter (Millipore) before PCR.

Amplification and cloning of 16S rRNA gene

Aliquots of frozen activated sludge and whole fixed sortedcells were used for PCR amplification with a thermal capillarycycler without prior treatment (Idaho Technology). The primersequences used to obtain almost full-length bacterial 16SrRNA gene fragments were 8F 58-AGAGTTTGATYMTGG-CTCAG-38 (Escherichia coli 16S rDNA positions 8–27; Brosiuset al., 1981) and 1546R 58-CAKAAAGGAGGTGATCC-38(E. coli 16S rDNA positions 1529–1546; Brosius et al.,1981). Two additional primers were used for the amplificationof partial 16S rDNA fragments of a length of 1100 nucleotides:C319S 58-GTACTGAGACACGG-38 (E. coli 16S rDNA positions319–332; Brosius et al., 1981) and SGR1425 58-GTCAC-CATCTTCAGGCAC-38 (E. coli 16S rDNA positions 1425–1442; Brosius et al., 1981). Reaction mixtures were preparedaccording to the manufacturer’s recommendations in a totalvolume of 50 ml using the 20 mM MgCl2 reaction buffer. Ther-mal cycling was carried out as follows: an initial denaturation

step at 948C for 30 s followed by 30 cycles of denaturation at948C for 15 s, annealing at 488C for 20 s and elongation at728C for 30 s. Cycling was completed with a final elongationstep at 728C for 1 min, and PCR products were analysed byelectrophoresis in 1% (w/v) agarose gel and staining with ethi-dium bromide, excised from the gel, purified with an agarosegel extraction kit (Boehringer Mannheim) and subsequentlyligated according to the manufacturer’s recommendationsinto the cloning vector pCRII supplied with the Original TAcloning kit (Invitrogen). Detection occurred via blue/whitescreening using Xgal. Plasmids were purified using the QIA-prep-spin kit (Diagen), and the presence of right-sized insertswas checked by restriction analysis with Not I.

16S rDNA sequencing

Nucleotide sequences were determined by the dideoxynu-cleotide method (Chen and Seeburg, 1985). Plasmids weresequenced with a ThermoSequenase cycle sequencing kit(Amersham) and analysed with an automated DNA sequen-cer (Li-Cor).

Data analysis

Sequences were added to the 16S rRNA sequence databaseof the Technical University Munich using the program packageARB (Strunk and Ludwig, 1997). The tool ARB_EDIT was used forsequence alignment. The alignment was checked by eye andcorrected manually. Phylogenetic trees were reconstructedbased on the results of distance matrix analyses (Jukes andCantor, 1969) of all available 16S rRNA primary structuresfor members of the Cytophaga–Flavobacterium–Bacteroidesphylum. The topologies of the different trees were evaluatedby performing maximum parsimony and maximum likelihoodanalyses of the full data set and subsets respectively. Align-ment positions at which less then 50% of sequences of theentire data set shared the same residues were excludedfrom the calculations. The phylogenetic positions of organ-isms presented by partial sequences were roughly recon-structed by applying the parsimony criteria without changingthe overall tree topology.

Nucleotide sequence accession numbers

The 16S rRNA sequences obtained in this study are availablein GenBank under accession numbers AF087043–AF087097.

Acknowledgements

This work was supported by a grant from the Deutsche For-schungsgemeinschaft to R.A. (Am 73/2-4) and by theResearch Center for Fundamental Studies of Aerobic Biologi-cal Wastewater Treatment (SFB 411). We thank SibylleSchadhauser for excellent technical assistance, and HansReichenbach for the supply of a pure culture of Saprospiragrandis. Claudia Beimfohr is acknowledged for critical readingof the manuscript.

References

Amann, R.I., Binder, B.J., Olson, R.J., Chisholm, S.W.,Devereux, R., and Stahl, D.A. (1990) Combination of 16S



rRNA-targeted oligonucleotide probes with flow cytometryfor analyzing mixed microbial populations. Appl EnvironMicrobiol 56: 1919–1925.

Amann, R., Ludwig, W., and Schleifer, K.H. (1995) Phylo-genetic identification and in situ detection of individualmicrobial cells without cultivation. Microbiol Rev 59: 143–169.

Amann, R.I., Snaidr, J., Wagner, M., Ludwig, W., and Schlei-fer, K.H. (1996) In situ visualization of high genetic diversityin a natural bacterial community. J Bacteriol 178: 3496–3500.

Ashkin, A., and Dziedzic, J.M. (1987) Optical trapping andmanipulation of viruses and bacteria. Science 235:1517–1520.

Ashkin, A., Dziedzic, J.M., and Yamane, T. (1987) Opticaltrapping and manipulation of single cells using infraredlaser beams. Nature 330: 769–771.

Ashton, P.J., and Roberts, R.D. (1987) Apparent predation ofMicrocystis aeruginosa Kutz. Emend. Elenkin by a Sapro-spira-like bacterium in a hypertrophic lake (HartbeespoortDam, South Africa). J Limnol Soc South Afr 13: 44–47.

Balows, A., Truper, H.G., Dworkin, M., Harder, W., andSchleifer, K.H. (1992) The Prokaryotes, 2nd edn, Vol. 4,Div. F, Chapters 199 and 200. New York: Springer-Verlag.

Bloodgood, R.A. (1990) Gliding motility and flagellar glyco-protein dynamics in Chlamydomonas. In Ciliary and Flagel-lar Membranes. Bloodgood, R.A. (ed.). New York: PlenumPress, pp. 91–128.

Bohlool, B.B., and Schmidt, E.L. (1980) The immunofluores-cence approach in microbial ecology. Adv Microb Ecol 4:203–241.

Brosius, J., Dull, T.J., Sleeter, D.D., and Noller, H.F. (1981)Gene organization and primary structure of a ribosomalRNA operon from Escherichia coli. J Mol Biol 148: 107–127.

Burchard, R.P. (1980) Gliding motility of bacteria. Bioscience30: 157–162.

Chen, E.Y., and Seeburg, P.H. (1985) Supercoil sequencing:a fast and simple method for sequencing plasmid DNA.DNA 4: 165–170.

DeGiorgi, C., Sialer, M.F., and Lamberti, F. (1994) Formalin-induced infidelity in PCR-amplified DNA fragments. MolCell Probes 8: 459–462.

Gross, J. (1911) Uber freilebende Spironemaceen. Mitt ZoolStat Neapel 20: 188–203.

Huber, R., Burggraf, S., Mayer, T., Barns, S.M., Rossnagel,P., and Stetter, K.O. (1995) Isolation of a hyperthermo-philic archaeum predicted by in situ RNA analysis. Nature376: 57–58.

Jukes, T.H., and Cantor, C.R. (1969) Evolution of proteinmolecules. Mammalian Protein Metabolism. Murano, N.H.(ed.). New York: Academic Press, p. 21.

Kolkwitz, R. (1909) Kryptogamenflora der Mark Branden-burg. Leipzig: Gebruder Borntraeger, Band V, p. 137.

Lewin, R.A. (1965a) Freshwater species of Saprospira. Can JMicrobiol 11: 135–139.

Lewin, R.A. (1965b) Isolation and some physical features ofSaprospira thermalis. Can J Microbiol 11: 77–86.

Lewin, R.A. (1997) Saprospira grandis: a flexibacterium thatcan catch bacterial prey by ‘ixotrophy’. Microb Ecol 34:232–236.

Lewin, R.A., and Lounsbery, D.M. (1969) Isolation, cultiva-tion and characterization of flexibacteria. J Gen Microbiol58: 145–170.

Lewin, R.A., and Mandel, M. (1970) Saprospira toviformisnov. spec. (Flexibacteriales) from a New Zealand sea-shore. Can J Microbiol 16: 507–510.

Manz, W., Amann, R., Ludwig, W., Wagner, M., and Schlei-fer, K.H. (1992) Phylogenetic oligodeoxynucleotide probesfor the major subclasses of proteobacteria: problems andsolutions. Syst Appl Microbiol 15: 593–600.

Manz, W., Amann, R., Ludwig, W., Vancanneyt, M., andSchleifer, K.H. (1996) Application of a suite of 16S rRNA-specific oligonucleotide probes designed to investigate bac-teria of the phylum cytophaga–flavobacter–bacteroides inthe natural environment. Microbiology 142: 1097–1106.

Murray, R.G.E., and Schleifer, K.H. (1994) Taxonomic notes:a proposal for recording the properties of putative taxa ofProcaryotes. Int J Syst Bacteriol 44: 174–176.

Nubel, U., Engelen, B., Felske, A., Snaidr, J., Wieshuber, A.,Amann, R.I., et al. (1996) Sequence heterogeneities ofgenes encoding 16S rRNA in Paenibacillus polymyxadetected by temperature gradient gel electrophoresis. JBacteriol 178: 5636–5643.

Olsen, G.J., Lane, D.J., Giovannoni, S.J., Pace, N.R., andStahl, D.A. (1986) Microbial ecology and evolution: a ribo-somal rRNA approach. Annu Rev Microbiol 40: 337–365.

Pace, N.R., Stahl, D.A., Lane, D.J., and Olsen, G.J. (1986)The analysis of natural microbial populations by ribosomalRNA sequences. Adv Microb Ecol 9: 1–55.

Pate, J.L., and Chang, L.Y.E. (1979) Evidence that glidingmotility in procaryotic cells is driven by rotary assembliesin the cell envelopes. Curr Microbiol 2: 59–64.

Pringsheim, E.G. (1963) Farblose Algen. Stuttgart: GustavFischer-Verlag, p. 111.

Rainey, F.A., Ward-Rainey, N., Janssen, P.H., Hippe, H., andStackebrandt, E. (1996) Clostridium paradoxum DSM7308T contains multiple 16S rRNA genes with hetero-geneous intervening sequences. Microbiology 142: 2087–2095.

Reichenbach, H. (1980) Saprospira grandis (Leucotrichales).Growth and movement. Publik Wiss Film, Sekt Biol, Ser 1326: 1–21.

Reichenbach, H. (1992) The genus Saprospira. In The Pro-karyotes, 2nd edn. Balows, A., Truper, H.G., Dworkin,M., Harder, W. and Schleifer, K.H. (eds). New York:Springer-Verlag, pp. 3676–3687.

Roller, C., Wagner, M., Amann, R., Ludwig, W., and Schlei-fer, K.H. (1994) In situ probing of Gram-positive bacteriawith high DNA GþC content using 23S rRNA-targeted oli-gonucleotides. Microbiology 140: 2849–2858.

Skerman, V.B.D. (1968) A new type of micromanipulator andmicroforge. J Gen Microbiol 54: 287–297.

Sladka, A., and Ottova, V. (1973) Filamentous organisms inactivated sludge. Hydrobiologica 43: 285–299.

Snaidr, J., Amann, R., Huber, I., Ludwig, W., and Schleifer,K.H. (1997) Phylogenetic analysis and in situ identificationof bacteria in activated sludge. Appl Environ Microbiol 63:2884–2896.

Strunk, O., and Ludwig, W. ARB: a software environment forsequence data (WWW document). URL http.//www.mikro.-biologie.tu-muenchen.de.



Torsvik, V., Goksoyr, J., and Daae, F.L. (1990) High diversityof DNA of soil bacteria. Appl Environ Microbiol 56: 782–787.

Van Tiegham, P. (1880) Observations sur les Bacteriaceesvertes, sur des Phycochromacees blanches, et sur les affi-nites de ces deux familles. Bull Soc Bot France 27: 174–179.

Wagner, M., Amann, R., Lemmer, H., and Schleifer, K.H.(1993) Probing activated sludge with proteobacteria-specificoligonucleotides: inadequacy of culture-dependent methodsfor describing microbial community structure. Appl EnvironMicrobiol 59: 1520–1525.

Wallner, G., Amann, R., and Beisker, W. (1993) Optimizingfluorescent in situ hybridization with rRNA-targeted

oligonucleotide probes for flow cytometric identification ofmicroorganisms. Cytometry 14: 136–143.

Wallner, G., Erhart, R., and Amann, R. (1995) Flow cyto-metric analysis of activated sludge with rRNA-targetedprobes. Appl Environ Microbiol 61: 1859–1866.

Wallner, G., Fuchs, B., Spring, S., Beisker, W., and Amann,R. (1997) Flow sorting of microorganisms for molecularanalysis. Appl Environ Microbiol 63: 4223–4231.

Warming, E. (1875) Om nogle ved Danmarks Kyster levendeBakterier. Videnskabelige Meddelelser fra den Naturhistor-iske Forening i Kjobenhaven for Aaret 1875: 307–420.

Woese, C.R., Stackebrandt, E., Macke, T.J., and Fox, G.E.(1985) A phylogenetic definition of the major eubacterialtaxa. Syst Appl Microbiol 6: 143–151.



Phylogeny and in situ identification of a morphologically conspicuous bacterium, Candidatus...

Documents

Transcript of Phylogeny and in situ identification of a morphologically conspicuous bacterium, Candidatus...