Specific Oligonucleotide Probes for in situ Detection of a Major Group of Gram-positive Bacteria with low DNA G+C Content

System. Appl. Microbiol. 22,186-196 (1999) SYSTEM4llC AND © Urban & Fischer Verlag _ht...:..tp_://_w_w_w_.u_rb_a_nf_isc_h_e_r.d_e..:../jo_u_rn_a_ls/_s_am _______________ APPLIED MICROBIOlOGY

Specific Oligonucleotide Probes for in situ Detection of a Major Group of Gram-positive Bacteria with low DNA G+C Content

HARALD MEIER 1,2, RUDOLF AMANN1,\ WOLFGANG LUDWIG!, and KARL HEINZ SCHLEIFER1

lLehrstuhl fur Mikrobiologie, Technische Universitat Munchen, Germany 2GSF-Forschungszentrum fur Umwelt und Gesundheit, Neuherberg, Germany 3MPI fur marine Mikrobiologie, Bremen, Germany

Received February 15, 1999

Summary

Almost one thousand 16S rRNA sequences of Gram-positive bacteria with a low DNA G+C content from public databases were analyzed using the ARB software package. A signature region was identified between positions 354 and 371 (E. coli numbering) for the Bacillus sub-branch of the Gram-positive bacteria with a low DNA G+C content, the former orders Bacillales and Lactobacillales. Three oligonu­cleotide probes, namely LGC354A, LGC354B, and LGC354C, were designed to target this diagnostic site. Their fluorescent derivatives were suitable for whole cell detection by fluorescence in situ hybridiza­tion (FISH). Hybridization conditions were adjusted for differentiation of target and related non-target reference species. When applying FISH to whole bacterial cells in a sample of activated sludge from a communal wastewater treatment plant, members of the Bacillus sub-branch were detected at levels from 0.01 % of cells in samples fixed with paraformaldehyde to over 8 percent in the same samples fixed with ethanol and treated with lysozyme. The problems of quantitative in situ analysis of Gram-positive bacte­ria with a low DNA G+C content in biofilm floes are discussed and recommendations made. Members of the Bacillus sub-branch were detected in different abundances in activated sludge samples from differ­ent wastewater plants.

Key words: fluorescence-in-situ-hybridization, FISH, rRNA, DNA probes, activated sludge, Gram-posi­tive bacteria, identification, rRNA database

Introduction

Ribosomal RNA molecules have been shown to be very suitable for investigating the phylogeny of bacteria (Fox et ai., 1980). Comparative analyses of 16S and 23S rRNA primary structures enable the identification and classification of bacteria into phyla, subphyla, genera and even species according to their phylogeny (Fox et ai., 1980; WOESE, 1987; SCHLEIFER and LUDWIG, 1989). Ribosomal RNA molecules contain signature sequences unique to phylogenetic groups of bacteria (WOESE, 1987), making them an ideal target for hybridization of complementary oligonucleotide probes. Applying fluo­rescently labeled oligonucleotides, the so-called phylo­genetic stains, individual whole fixed cells of bacteria can be identified by fluorescence in situ hybridizations (FISH) (DELONG et aI., 1989; AMAN et ai., 1990a, 1995). Group-specific probes had been developed for different sub-classes of Proteobacteria (MANZ et aI., 1992), for the CytophagalFlavobacteriumlBacteroides group

0723-2020/99/22/02-186 $ 12.00/0

(MANZ et ai., 1996) and for the Gram-positive bacteria with high DNA G+C content (GPB-HGC) (ROLLER et ai., 1994). They were successfully applied for whole cell identification and enumeration in activated sludge (WAGNER et aI., 1993, 1994; MANZ et ai., 1996), olig­otrophic lakes (GLOCKNER et ai., 1996; ALFREIDER et ai., 1996), drinking water biofilms (MANZ et ai., 1993), and soil (ZARDA et aI., 1997). In each of these environments, total numbers of bacteria and relative abundances of the main bacterial phyla could be determined quantitatively by epifluorescence microscopy without prior cultivation or nucleic acid extraction. So far, in situ abundances of another important bacterial phylum, the Gram-positive bacteria with a low DNA G+C content (GPB-LGC), could not be determined due to the lack of a group spe­cific probe.

Comparative analyses of 16S and 235 rRNA se­quences strongly indicate that GPB-LGC and GPB-HGC

In situ Detection of Gram-positive Bacteria with low mol% G+C 187

(C) Streptococcu8 (C)

(A) LeuconOiltoc LactococcU8 Cllmobacterlum (8)

(A) We/.-1I11

LGC354A-C

(8) SfllphylOcoccutl and relatlvea

(8) B.elllutl tlubtilltl and relatives

(8) Baclllu8 tlphaerleutl and relatives

(8) Baclllutl Mearothennophllull and relatives

(8) S.elllua cereull and relatives

(A) Sporolactobaefllua and relatives

(A) Exlguobaeterlum

Aneurinibllelllua (8) and relatives

and relatives Enterococcull

LactobaelllUII/ PedIOCOCCUII (A)

Paenlbllcl/lutl

(8) Alicyeiobaclllull and relatives

(8)

Fig. 1. Phylogenetic tree of Gram-positive bacteria with a low DNA G+C-content and outgroup bacteria based on 16S rRNA pri­mary structure. GPB-LGC showing target sequence perfectly complementary to probes LGC354A-C are indicated by black branch­es. GPB-LGC detected by probe LGC354A-C despite single mismatch are indicated by stippled branches. The bar indicates 10% of phylogenetic distance.

are biphyletic groups (LUDWIG et al., 1992; VAN DE PEER et al., 1994; DE RUK et al., 1995). The GPB-LGC phylum encompasses at least four distinct major phylogenetic groups of bacteria (Fig. 1): (i), the Bacillus sub-branch, including Bacillus and related genera, Listeria, Staphylo­coccus, and the lactic acid bacteria (SCHLEIFER and LUD­WIG, 1995; STILES et al., 1997), including the genus Strep­tococcus, (ii), Mycoplasma and relatives, (iii), Clostridi­um and related bacteria, and (iv), phylogenetic ally deep branching bacteria with a cell wall type, similar to Gram­negative bacteria, e.g. the genera Selenomonas, Pectina­tus, and Heliobacterium.

The aim of this study was to develop group-specific probes for detection of GPB-LGC by FISH and their ini­tial application to highly complex environmental sam­ples, namely activated sludge.

Materials and Methods

Organisms and media. The control strains used in this study are listed in Table 1. All strains were grown in liquid culture ac­cording to the protocols given by the respective strain collec­tions (ATCC, American Type Culture Collection, Rockville, MD, USA; WS, Stammsammlung des Instituts flir Mikrobiolo­gie, Forschungszentrum fiir Milch und Lebensmittel, Technische Universitat Miinchen, Freising-Weihenstephan, Germany; DSM, Deutsche Sammlung von Mikroorganismen und Zellkul­turen, Braunschweig, Germany; NCTC, National Collection of Type Cultures, London, UK; LMG, Belgium Coordinated Col­lections of Microorganisms, Laboratorium voor Microbiologie, Gent, Belgium) or in (1) Yeast-Dextrose-Medium (in gil: tryp­tone 10; NaCI 7; yeast extract 1.5; dextrose 1; pH 7.2), (2) BHI-Medium (Difco, Detroit, USA), (E) LB nutrient broth (in gil tryptone iO; NaCI 10; yeast extract 5; pH 7.0), (4) MRS-Me-

188 H. MEIER et a!.

Table 1. Bacterial strains used in this study.

Species

Bacillus alvei (4) Bacillus cereus (4) Bacillus polymyxa (4) Bacillus subtilis (1) Brevundimonas diminuta (1) Burkholderia cepacia (1) Campylobacter jejuni (2) Campylobacter coli (2) Campylobacter lari (2) Carnobacterium earn is (5) Carnobacterium divergens (5) Carnobacterium piscicola (5) Caulobacter crescentus Comamonas testosteroni (1) Enterococcus durans (2) Enterococcus faecalis (2) Enterococcus faecium (2) Enterococcus gallinarum (2) Enterococcus hirae (2) Escherichia coli (1)

Strain

WS 1535 WS3P WS 1538 ATCC 6633 DSM 1635 LMG 1222! NCTC 11322 NCTC 11366 NCTC 11352 WS2296 WS 2297 WS2299 ATCC 15252T

LMG 1800T

DSM 20633T

DSM 20478T

DSM 20477T

DSM 20628T

DSM 20160 l

DSM 30083T

Species

Hyphomonas jannaschiana Lactobacillus casei (4) Lactobacillus casei (4) Lactobacillus lindneri (4) Lactococcus garviae (2) Lactococcus lactis ssp. cremoris (4) Lactococcus lactis ssp. lactis (4) Leuconostoc mesenteroides ssp. cremoris (4) Leuconostoc mesenteroidess sp.lactis (4) Listeria monocytogenes Rhodococcus rhodochrous Pectinatus frisingensis Pediococcus dammosus (4) Pediococcus inopinatus (4) Pseudomonas aeruginosa (1) Staphylococcus au reus (3) Staphylococcus carnosus Streptococcus salivarius (2) Streptococcus bovis (2) Streptococcus equinus (2)

Strain

ATCC 33882 LMG9091 WS 1020 WS2533 WS 1029 DSM20069T DSM 20481T

WS 1037 WS 1036 DSM2060QT DSM43752 DSM20465 WS2530 WS2529 DSM 50071T

DSM 20231 T

DSM20501T DSM2056QT DSM 20480T

DSM 20558T

(1) Yeast-Dextrose-Medium; (2) BHI-Medium; (3) LB-Medium; (4) MRS-Medium; (5) Carnobacterium-Medium

dium (Difco 0881), (5) Carnobacterium-Medium (in gil: casein peptone 10; meat extract 8; yeast extract 4; saccharose 20; K2HP04 2; sodium acetate (triple hydrate) 5; di-ammoniumhy­drogen citrate 2; Mg2S04 • 7H20 2; MnS04 0.05; 1 ml of Tween 80; pH 8.5). When necessary, media were solidified by addition of 15 gil agar (Gibco BRL, Eggenstein, FRG).

Sequence analysis and phylogenetic trees. Aligned primary structures of about 3600 almost complete 16S rRNA sequences were analyzed using the software package ARB (STRUNK et a!., 1998). Phylogenetic trees were calculated from 16S rRNA se­quences of GPBLGC using the ARB parsimony tool including distance calculation. The Deinococci were chosen as outgroup. The topology of trees was compared with those obtained by ap­plying maximum likelihood methods (FELSENSTEIN, 1981) on se­lected sequences and, if necessary, corrected by drawing unsta­ble branchings as multifurcations. Phylogenetic distances were compared with those obtained by the use of the neighbor-join­ing method (JUKES and CANTOR, 1969; SAITOU and NEI, 1987) and corrected where necessary. All tools used for analysis are availabel in the ARB software package.

Design, specificity check, and labeling of oligonucleotide probes. Oligonucleotide probes comprising 18 nucleotides were designed using computer aided comparative 16S rRNA se­quence analysis with the ARB tool PROBE-DESIGN (STRUNK et a!., 1998). Probe specificities were evaluated by comparison of the probe specific target sequences with all sequences within the database using the ARB tool PROBE-MATCH.

Table 2. Probes specific for "Bacillus "-subbranch of GPB-LGC.

Probe

LGC354A LGC354B LGC354C

+ SUGGS et a!. (1981)

5'-sequence-3'

TGGAAGATTCCCTACTGC CGGAAGATTCCCTACTGC CCGAAGATTCCCTACTGC

Synthesis and labeling of oligonucleotide probes with fluo­rescein or carboxytetramethyl-rhodamine (CT) were performed as described earlier (AMANN et a!., 1990a; ZARDA et a!., 1991). Probe sequences, target organisms, and target sites are shown in Table 2 and Fig. 2.

Membrane filtration and DNA-staining with DAPI for deter­mination of total bacterial counts. Membrane filtration and staining with 4',6 diamidino-2-phenylindole (DAPI) (PORTER and FEIG, 1980) was performed as described earlier (WAGNER et a!., 1993). Total cell counts were determined by counting 10 in­dependent fields with at least 200 cells using a lOx eyepiece with a counting grid and a 100x Plan-Neofluar objective (Zeiss, Jena, Germany).

Fixation of bacterial cells/samples from activated sludge and whole cell hybridizations. Cell were harvested from liquid cul­tures during log phase and fixed with 4% para formaldehyde (PFA) (AMANN et a!., 1990a) or with 50% ethanol (ROLLER et aI., 1994). Samples from activated sludge were fixed with PFA as described by WAGNER et al. (1993) or with ethanol (WAGNER et aI., 1994). The fixed samples were spotted onto glass slides, dried at 46°C for 30 min, and immersed for 3 min each in 50, 80 and 100% ethanol. Samples were pretreated on the slide with 50 ).II lysozyme solution (final concentration: 133-1330 U/).II of lysozyme (Sigma, Deisenhofen, Germany) in 100 mM TrisIHCI, 50 mM EDTA, pH 7.2) before hybridization for 1 to 10 min at 0 °C as described by BEIMFOHR et a!. (1993). Whole cell hybridization was done as described by AMANN et al.

GC-content

50% 55.6% 55.6%

Calculated +T 0 (0C)

54 56 56

In situ Detection of Gram-positive Bacteria with low mol% G+C 189

LGC354A: 5'-GCAGUAGGGAAUCUUCCA-3'

Lactobacillus casei: CGGGAGGCA-==================-CAAUGGACG

Lactobacillus lindneri: CGGGAGGCA-==================-CAAUGGACG

Leuc. mesenteroides ssp. lactis: CGGGAGGCA-==================-CAAUGGGCG

PediocQCCUs inopinatus

Hvohomonas jannaschiana:

Brevundimonas diminuta:

Campylobacter jejuni:

Rhodococcus rhodochrous:

LGC354B:

Bacillus subtilis:

Bacillus polymyxa:

Bacillus alvei:

Carnobacterium carnis: Carnobactqium piscicola'

Listeria monocytogenes:

Staphylococcus aureus:

PectinatUS frisingensis: Caulobactq crescentus:

Brevundimonas pnuta:

Campylobacter jejuni:

Campylobactq lari: Campylobacter coli:

Hvohomonas jannaschiana:

Rhodococcus rhodochrous:

LGC354C:

Lactococcus lactis sub. lactis:

CGGGAGGCA-==================-CAAUGGGCG

CGGGAGGCA-=====g=========G==-CAAUGGGCG

CGGGAGGCA-=====g=========G=g-CAAUGGGCG

CGGGAGGCA-============a==G=g-CAAUGGGCG

CGGGAGGCA-=====g======a==G==-CAAUGGGCG

5'-GCAGUAGGGAAUCUUCCG-3'

CGGGAGGCA-==================-CAAUGGGCG

CGGGAGGCA-==================-CAAUGGGCG

CGGGAGGCA-==================-CAAUGGGCG

CGGGAGGCA-==================-CAAUGGACG

CGGGAGGCA-==================-CAAUGGACG

CGGGAGGCA-==================-CAAUGGGCG

CGGGAGGCA-==================-CAAUGGGCG

CGGGAGGCA-=====g============-CAAUGGGCG

CGGGAGGCA-=====g=========G==-CAAUGGGCG

CGGGAGGCA-=====g=========G==-CAAUGGGCG

CGGGAGGCA-============a==G==-CAAUGGGCG

CGGGAGGCA-============a==G==-CAAUGGGCG

CGGGAGGCA-============a==G==-CAAUGGGCG

CGGGAGGCA-=====g=========G=a-CAAUGGGCG

CGGGAGGCA-=====g======a==G=a-CAAUGGGCG

5'-GCAGUAGGGAAUCUUCGG-3'

CGGGAGGCA-==================-CAAUGGACG

20%

20%

20%

0%

n.d.

20%

20%

20%

20%

20%

20%

0%

Lactococcus lactis sub. cremoris: CGGGAGGCA-==================-CAAUGGACG

Lactococcus garviae:

Streptococcus salivarius: Enterococcus durans:

Entqococcus faecalis: Enterococcus faecium:

Enterococcus galIinarum:

Enterococcus hirae: Streptococcus bovis:

Streptococcus eguinus:

Brevundimonas dirninuta:

Campylobacter jejuni:

Hvohomonas jannaschiana:

CGGGAGGCA-==================-CAAUGGGGG

CGGGAGGCA-==================-CAAUGGGGG

CGGGAGGCA-==================-CAAUGGACG

CGGGAGGCA-==================-CAAUGGACG

CGGGAGGCA-==N===============-CAAUGGACG

CGGGAGGCA-==g===============-CAAUGGACG

CGGGAGGCA-==N===============-CAAUGGACG

CGGGAGGCA-==N===============-CAAUGGGGG

CGGGAGGCA-==N===============-CAAUGGGGG

CGGGAGGCA-=====g=========GC=-CAAUGGGCG

CGGGAGGCA-============a==GC=-CAAUGGGCG

CGGGAGGCA-=====g=========GCa-CAAUGGGCG

n.d.

n.d.

n.d.

n.d.

n.d.

20%

20%

20%

Fig. 2. Alignment of target sites of probes LGC354A-C in target- and nontarget bacteria. The stringency at which nontarget species showed no hybridization signal, is given after the sequences (in % formamide in hybridization buffer). n.d. = cannot be differentiat­ed from target bacteria even at 50% formamide .

(1990b) with fluorescently labeled probes EUB338 (AMANN et aI., 1990b) of LGC354 (Table 2) or both simultaneously. The stringency of whole cell hybridization was adjusted by addition of formamide to the hybridization buffer and an equivalent de­crease of the NaCI concentration in the washing buffer (MANZ et aI., 1992). Staining with 4,6 diamino-2-phenylindole (DAPI, 0.5 ~g1ml in ddHzO) was performed afterwards directly on the

slides (WAGNER et ai., 1993). Slides were embedded in Citifluor (Citifluor Ltd., Canterbury, UK) before microscopic examina­tion with a Zeiss Axioplan microscope (Zeiss). Filter sets HQ41007, HQ41001, and HQ82000 (Chroma Tech. Corp., Brattleboro, VT, USA) and 01 (Zeiss) were used. Color and black/white photomicrographs were made on Kodak Ek­tachrome 1600 and Kodak TMAX 400 (Rochester, NY, USA).

190 H. MEIER et al.

Probe specific cell counts in an activated sludge sample simulta­neously hybridized with EUB338-FLUOS and stained with DAPI or simultaneously hybridized with probes EUB338-FLUOS and LGC354A-C-CT were determined by counting 10 independent fields as described above. In additional samples of activated sludge, the proportion of Gram-positive bacteria was estimated by comparing the number of cells detected by LGC354!-C-CT with the number of DAPI-stained bacteria.

Results and Discussion

Design of oligonucleotide probes. Probe design and specificity analyses were performed by using an ARB se­quence database containing a total of 3600 almost full­length bacterial 16S rRNA sequences (more than 1400 determined positions) of which 896 were GPB-LGC. Comparative 16S rRNA sequence analysis using the ARB tool PROBE-DESIGN resulted in several suggestions for target sites common in GPB-LGC. None of them was perfect in the sense that one probe could detect all mem­bers of the GPB-LGC and distinguish them from all other bacteria. A site at 16S rRNA positions 354 to 371 (E. coli - numbering, BROSIUS et aI., 1981) was most promis­ing. A set of three probes, LGC354A-C, with small dif­ferences at the 5' end (Table 2) were therefore targeted to

that site to enable identification of a large group of GPB­LGC (at least 43% of all GPB-LGC sequences in the database). Probes LGC354A and LGC354B differ at po­sition 371 (T and C), probes LGC354B and LGC354C at position 370 (G and C), and the probes LGC354A and LGC354C differ at positions 370-371 (TG and CC). A total of 112 sequences were perfectly matched to probe LGC354A, 184 to LGC354B and 50 to LGC354C.

The target region of LGC354A-C is located in a highly conserved region of the 16S rRNA. It overlaps by two nucleotides (positions 354-355) with the binding site of probe EUB338 specific for members of the kingdom Bac­teria. With the GPB-LGC the target encompasses the fol­lowing signatures: A at position 359 in 63% of 896 GPB-LGC 16-S rRNA sequences analyzed and in only 6.5% of the 2700 sequences of non-GPBLGC, C at posi­tion 366 in 51 % of GPB-LGC and only in 3% of non­GPB-LGC sequences, C at position 369 in 61 % of GPB­LGC and in only 6% of the rest (Table 3).

Results obtained with the ARB tool Probe Match are shown in Fig. 2 and Table 4. Bacterial species and genera listed are those present in the ARB database with 16S rRNA sequences which match the probe target sequences exactly or have only minor differences and are still likely to bind to LGC354A-C.

Table 3. Signature positions of GPB-LGC in target of probes LGC354A-C. 3556 full 16S rRNA sequences were analyzed (GPB­LGC: 896 sequences; non GPB-LGC: 2660 sequences ).

Signature (position in E. coli)

A (359) C (366) C (369) A + C (359 + 366) A+C+C (359 + 366 + 369)

Sequences of GPB-LGC with the signature bases (relative to a1l16S rRNA sequences of GPB-LGC in the analyzed ARB database)

566 (=63% ) 460 (=51%) 551 (=61 %) 400 (=45%) 383 (=43%)

Non-GPB-LGC sequences with the signature bases (relative to a1l16S rRNA sequences of rRNA sequences of non GPB-LGC in the analyzed ARB database)

174 (=6.5%) 78 (=3%)

160 (=6%) 6 (=0.2%) o

Table 4. Bacterial genera and species showing the perfect or almost perfect target site for probes LGC354A-C in the analyzed database (ARB-project, STRUNK et al., 1998).

Probe

LGC354A Alloiococcus spp., Bacillus sphaericus and relatives, Caryophanon spp., Dolosigranulum spp., Exiguobacterium spp., Filibacter limicola Kurthia spp., Lactobacillus spp., Leuconostoc spp., Mycoplasma mobile, Oxalophagus spp., Pedio­coccus spp., Sporolactobacillus spp., Sporosarcina spp., Sulfobacillus spp., Weissella spp., single G-T mismatch: Haloanaerobium spp., Haloincola saccharolyticum

LGC354B Abiotrophia spp., Aerococcus spp., Alicyclobacillus spp., Aneurinibacillus sp., most Bacillus spp., Carnobacterium spp., Carnococcus allantoicus, Gemella spp., Globicatella sanguis, Halobacillus spp., Planococcus spp., Lactobacillus thermophilus, Lactobacter thermoamylovorans, Lactosphaera pasteurii, Listeria spp., Paenibacillus spp., Saccharo­coccus thermophilus, Salmococcus spp., Staphylococcus spp., single G-T mismatch: Desulfotomaculum spp., Pectina­tus spp., Heliobacterium spp., Selenomonas ruminantium, Acidaminococcus fermentans, Peptococcus niger, Sporo­musa spp., Veillonella spp., Halobacteroides spp., Phascolarctobacterium faecium, Megasphaera, Desulfotomaculum spp., Orenia marismortui, Sporohalobacter lortetii, Zymopgilus paucivorans

LGC354C Brochothrix spp., Enterococcus spp., Lactococcus spp., Melissococcus spp., Streptococcus spp., Tetragenococcus halophilus, Vagococcus spp.

In situ Detection of Gram-positive Bacteria with low mol% G+C 191

The group of organisms showing a perfect match for probes LGC354-C is the well defined Bacillus sub­branch within the phylum GPB-LGC (Fig. 1). Deeper branching GPB-LGC, e.g. the mycoplasmas and the clostridia, have different sequences at that position. Some of the sequences of Mycoplasma spp. have one base difference from the perfect target for LGC354A. Fif­teen species, including M. pulmonis, have U instead of C at position 366, 26 sequences of Mycoplasma spp. have an A at the same position (e.g. M. hominis). Two mis­matches to probe LGC354A are found in the 16S rRNA of all Ureoplasma spp., Spiroplasma spp., and some My­coplasma spp. (positions 366 and 369, U instead of C) .

A large number of Clostridium spp. 16S rRNA se­quences (170) show three mismatches to probes LGC354A or LGC354B (359, G instead of A; 366, A in­stead of C; 369, G instead of C). Other species show even more mismatches.

Permeabilization of GPB-LGC cells for fluorescently labeled oligonucleotide probes. Four different pretreat­mentlfixation-protocols were performed to optimize pen­etration of fluorescent probes into GPB-LGC cells. Large differences in the strength of the FISH-signal were no­ticed for most strains of GPB-LGC when different fixa­tion and pretreatment methods were applied. Generally, the probe-conferred fluorescence was higher after ethanol fixation compared to PFA fixation. Ethanol.­treated cells of carnobacteria, Lactobacillus lindneri and Enterococcus hirae were permeable to the probes with­out further enzymatic treatment. All other GPB-LGC strains used as references had to be pretreated with 400 U/pl of lysozyme for 4 min at 0 °C to achieve high and uniform fluorescence in the cells detected, except for Listeria monocytogenes and Staphylococcus aureus. L. monocytogenes was permeabilized by incubation with 665 U/Jll for 10 min at O°C, whereas S. aureus could not be permeabilized with lysozyme, even after longer incu­bations at higher concentrations.

Optimization of hybridization conditions. Optimal hybridization conditions were determined by monitoring probe hybridization to target and non-target cells at dif­ferent stringencies, set by the concentration of for­mamide in the hybridization buffer Carnobacterium car­nis and Bacillus subtilis were selected as target cells for probe LGC354B. Brevundimonas diminuta and Campy­lobacter jejuni from the (X- and £-subclasses respectively

of Proteobacteria (two mismatches each) as well as Rhodococcus rhodochrous from GPB-HGC (four mis­matches) were used as non-target references with rela­tively high similarity in the probe target sequence (Fig. 2). Cells of the genera Anabaena, Micro cys tis, Nostoc, Oscillatoria which have only two mismatches in the binding site of the probes have not been tested since cyanobacteria can be easily discriminated from GPB­LGC by epifluorescence microscopy due to their red aut­ofluorescence after excitation with blue or green light.

The fluorescence conferred by probe LGC354B-CT to target cells at 0, 20 and 35% formamide was stable and similar in intensity to that conferred by the bacterial probe EUB338-CT. It was significantly reduced at 50% formamide and almost undetectable at 60% formamide in hybridization buffer. R. rhodochrous did not bind de­tectable levels of LGC354B-CT at 0% and the other non­target references, e.g. Brevundimonas diminuta were dis­criminated at 20% formamide (Table 5/Fig. 2/3A). Con­sequently, 35% formamide was chosen as the optimal concentration.

Hybridization conditions for probes LGC354A-CT and LGC354C-CT were evaluated in a similar way and 35% formamide gave optimal specificity and sensitivity in these cases too. When probes LGC354A-CT and EUB338-FLUOS were applied simultaneously at 35% formamide, the hybridization signal of EUB338-FLUOS was significantly reduced (data not shown). This effect was reproducible and probably caused by probe compe­tition. We cannot, however, explain why this did not occur with probes LGC354B and LGC354C. A group of deep branching GPB-LGC with cell wall types similar to Gram-negative bacteria, e.g. Selenomonas spp., Veil­lonella spp., Pectinatus sp., Helicobacterium spp., Desul­fotomaculum spp. and others, have only one mismatch (G-T at E. coli position 359) with LGC354A (six se­quences) or LGC354B (51 sequences). Fixed cells of Pectinatus frisingensis a representative of this group of bacteria, were still detected at 35% formamide. It is, therefore, likely that all of these strains will be detected with LGC354A-C at 35 % formamide (Table 4, Fig. 2). Some Mycoplasma spp. (e.g. M. pulmonis) or Mycoplas­ma-like organisms (e.g. Clostridium spiroforme, Achole­plasma spp., Anaeroplasma spp.) may also react positive­ly with LGC354A or LGC354C for the same reason. In contrast, species belonging to the MycoplasmalUreaplas-

Table 5. Fluorescence intensity in target and nontarget cells after FISH with probe LGC354B-CT estimated by microscopical obser­vation.

Bacteria % Formamide 0 20 25 35 50 60

Carnobacterium earnis +++ +++ +++ +++ + (+) Bacillus subtilis +++ +++ +++ +++ + (+) Pectinatus (risingensis ++ ++ ++ ++ nt nt Brevundimonas diminuta + nt nt nt Campylobaeter jejuni + nt nt nt Rhodoeoeeus rhodoehrous nt nt nt

+++ bright fluorescence; ++ average fluorescence; + weak fluorescence; (+) detection limit; (-) no fluorescence; nt = not tested

192 H. MEIER et al.

Table 6. Abundances of cells detected by LGC3S4-CY3 in activated sludges from different wastewater plants.

Sample/Sampling date

Augsburg AS 1117.06. 1997 Augsburg AS 1I117. 06. 1997 Kraftisriedl03. 03. 1998 Lindau AS/OS. 06. 1997 Kraftisriedl06. 05. 1998 Grosslappen AS 1102.06. 1998 Lenggries ASI17. 06. 1997 Dietersheim AS 1/03.03. 1998

AS = activated sludge

Type of treatment plant

communal communal animal carcasses processing communal animal carcasses processing communal communal communal

ma/Spiroplasma group as well as the true clostridia are discriminated by probes LGC354A-C as they have more mismatches than, for example Hyphomonas jannaschi­ana (Fig. 2), which was not detected at 20% or 35% for­mamide.

A reliable differentiation between subgroups of the target bacteria by probes LGC354A, LGC354B and LGC354C was impossible even at very stringent condi­tions (50-60% formamide). The only exception was dis­cnmmation of LGC354C-target cells by probe LGC354A in the presence of probe EUB338 at 35% for­mamide. Lactococcus lactis, for example, has two termi­nal mismatches to probe LGC354A, additional destabi­lization is probably caused by the presence of probe EUB338 which overlaps with probes LGC354A-C by two nucleotides (data not shown).

In conclusion, it is recommended that an equimolar mixture of LGC354A, LGC354B, LGC354C (= LGC354) is used for detection of GPB-LGC, excluding the clostridia and mycoplasmalureaplasmalspiroplasma, and that whole cell hybridization is conducted in the standard hybridization buffer containing 35 % for­mamide. Simultaneous hybridization with probe EUB338 should be. avoided in an unknown habitat under these stringent conditions, but can be performed in the presence of 25% formamide without significant probe competition effects.

Detection of members of the Bacillus sub-branch in activated sludge. In activated sludge from the first aera­tion stage of a municipal wastewater treatment plant (15. 11. 1995, Munich 1, Grosslappen) 74.4% ± 6% of the total cells (1.33 x 109 ± 1.18 X 109 cells(ml) could be identified as members of Bacteria by DAPI staining and FISH with probe EUB338-FLUOS. Subsequently, the

Estimated abundance of members of Bacillus­subbranch of GPB-LGC

2-3% 1-2%

<0.01 % 2-3%

<0.01 % 3-4%

<-1% 2-3%

fraction of GPB-LGC was determined by scoring cells which had been simultaneously hybridized with EUB338-FLUOS and LGC354-CT in fixed samples. In PFA-fixed samples far less than 0.1 % all bacteria could be detected with probe LGC354-CT (Fig. 3B). However, about 4% (4% ± 1.8%) of all bacteria were identified as GPB-LGC when PFA-fixed sample were pretreated with lysozyme (400 U/1l1 for 4 min at 0 0c). When ethanol­fixed samples were used for FISH, 3.1 % ± 0.7% of all bacteria were identified as GPB-LGC (Fig. 3C). The pro­portion of bacterial cells binding probe LGC354-CT specifically could be increased up to 8.8% ± 5.4% when ethanol-fixed samples were treated with lysozyme (400 U/IlI for 4 min at 0 0c) before FISH (Fig. 3D). Obviously, the relative abundance of GPB-LGC determined was de­pendent on the fixation and the permeabilization proto­col applied. The highest proportion determined cannot be interpreted as the most realistic abundance of these organisms in activated sludge, since the EUB338-FLUOS specific counts per microscope field were significantly lower in ethanol-fixed, lysozyme-treated samples (182-638; N = 10, 0 = 305) than in ethanol-fixed un­treated samples (312-745; N = 10,0 = 534). This indi­cates that lysozyme treatment results in a loss of bacte­ria, especially when the samples are ethanol-fixed. There­fore the proportion of GPB-LGC, which are more resis­tant to lysozyme treatment than Gram-negative bacteria, was overestimated. This problem could be overcome by direct counting on membrane filters, though homoge­nization methods have to be developed to disperse acti­vated sludge flocs and improve the statistical distribution of bacteria on the filter surface. Only by comparison of the total cell counts (e.g. DAPI counts without permeabi­lization) with absolute numbers of GPB-LGC-positive

Fig. 3. (A) In situ differentiation between Carnobacterium carnis and Brevundimonas diminuta. Left pannel, mixture of fixed cells in phase contrast. Right pannel, same field in epifluorescence double exposure. Cells of Carnobacterium carnis are stained red by specific binding of LGC3S4B-CT and Brevundimonas diminuta fluorescing green by ALF968-FLUOS, specific for Alpha-Proteobac­teria (NEEF, 1997). Yellow colour is caused by overlapping of red- with green-stained cells. (B-D) Whole cell detection and identifi­cation of members of the Bacillus-subbranch in activated sludge (Grosslappen I, Munich) by probe LGC3S4-CT depends on sample fixation and lysozyme treatment. Left, phase contrast; right, epifluorescence exposure. (B) Sample fixed with PFA, (C) sample fixed with ethanol, (D) sample fixed with ethanol and treated with lysozyme before hybridization. Bars indicate 10 pm (A, B, C) and 20 pm (D).

In situ Detection of Gram-positive Bacteria with low mol% G+C 193

194 H. MEIER et al.

cells using an optimized permeabilization (e.g. ethanol­fixation plus lysozyme treatment) can the relative abun­dance of GPB-LGC be determined reliably in environ­mental samples.

By examination of ethanol-fixed, lysozyme-treated ac­tivated sludges from different wastewater treatment plants combining whole cell hybridization with LGC354-CT and DAPI staining on slides, it became clear that GPB-LGC are a small, but common part of the microbial consortia in communal wastewater. In all activated sludge samples from communal wastewater treatment plants the proportions of LGC354-CT-positive bacteria were calculated at between one and three percent of total bacteria (Table 6, Fig. 4A). This is not surprising since GPB-LGC such as enterococci, lactobacilli, streoptococci and peptostreptococci are common inhabitants of the human gastrointestinal tract (KNOKE and BERNHARDT, 1985) and are generally present in human feces in high

numbers (MOORE and HOLDEMAN, 1974; MCCARTNEY et al., 1996). Previous studies have detected streptococci and relatives by whole cell habridization in concentra­tions of about 107/g in human feces (FRANKS et al., 1998) and cells of Streptococcus thermophilus were detected in activated sludge by a fluorescently labelled rRNA-target­ed specific probe (BEIMFOHR et al., 1993).

In contrast, less than 0.01 % of all cells were detected by probes LGC354-CT in activated sludge samples from a facility processing animal carcasses on two different sampling dates (Table 6, Fig. 4B). As generally only dead animals are processed in such facilities, small amounts of feces will be present and this may explain the very low levels detected compared with communal wastewater sludges.

The development of the probes LGC354 closes an im­portant gap in the set of rRNA-targed, group specific probes. These probes allow in situ detection of a phylo-

Fig. 4. Fluorescence exposures of activated sludge after simultaneous DAPI-staining (left) and whole cell hybridization with probes LGC354-CT (right). (A) Grosslappen I, Munich, municipal wastewater treatment plant and (B) Kraftisried, wastewater treatment plant of an animal carcassis processing facility. Bars indicate 5 flm.

In situ Detection of Gram-positive Bacteria with low mol% G+C 195

genetically defined, large group of GPB-LGC in environ­mental samples and, thereby, a more comprehensive analysis of the microbial community composition by FISH.

Acknowledgement This work was granted by the Deutsche Forschungsgemein­

schaft Am 73/2-3 and the EC-project FAIR-CT96-1037. The excellent technical assistance of SIBYLLE SCHAD HAUSER, BERN­HARD FUCHS and MARKUS SCHMID is gratefully acknowledged. We thank MARTIN ADAMS (School of Biological Sciences, Uni­versity of Surrey, Guildford, UK) for improving the linguistic quality of the manuscript.

References

ALFREIDER, A., PERNTHALER, ]., AMANN, R., SATTLER, B., GLOCKNER, EO., WILLE, A., PSENNER, R.: Community analy­sis of the bacterial assemblages in the winter cover and pelag­ic layers of a high mountain lake by in situ hybridization. App!. Environ. Microbiol. 62, 2138-2144 (1996).

AMANN, R. I., KRUMHOLZ, L., STAHL, D. A.: Fluorescent­oligonucleotide probing of whole cells for determinative, phylogenetic, and environmental studies in microbiology. ]. Bacteriol. 172,762-770 (1990a).

AMANN, R. I., BINDER, B.]., OLSON, R. J:, CHISHOLM, S. W., DE­VEREUX, R., STAHL, D. A.: Combination of 16S rRNA-targed oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl. Environ. Microbiol. 56, 1919-1925 (1990b).

AMANN, R. I., LUDWIG, W. SCHLEIFER, K.-H.: Phylogenetic iden­tification and in situ detection of individual microbial cells without cultivation. Microbial Review 59,143-169 (1995).

BEIMFOHR, c., KRAUSE, A., AMANN, R., LUDWIG, W., SCHLEIFER, K. H.: In situ Identification of Lactococci, Enterococci and streptococci. System. App!. Microbio!. 16,450-456 (1993).

BROSIUS,]., DULL, T. L., SLEETER, D. D., NOLLER, H. E: Gene or­ganization and primary structure of a ribosomal RNA oper­on from Escherichia coli. J. Molec. BioI. 148, 107-127 (1981).

DELONG, E. E, WICKHAM, G. S., PACE, N. R.: Phylogenetic stains: ribosomal RNA-based probes for the identification of single microbial cells . Science 43,1360-1363 (1989).

DE RI]K, P., VAN DE PEER, Y., VAN DEN BROECK, I., DE WACHTER, R.: Evolution according to large ribosomal subunit RNA ]. Mol. Evol. 41 (3),366-375 (1995).

FELSENSTEIN, ].: Evolutionary trees from DNA sequences: A maximum likelihood approach. ]. Mol. Evol. 17, 368-376 (1981).

Fox, G. E, STACKEBRANDT, E., HESI'ELL, R. B., GIBSON, ]., MANILOFF,]., DYER, T. A., WOLFE, R. S., BALCH, W. E., TAN­NER, R. S., MAGRUM, L. ]., ZABLEN, L. B., BLAKEMORE, R., GUPTA, R., BONEN, L., LEWIS, B. J., STAHL, D. A., LUEHRSEN, K. R., CHEN, K. N., WOESE, C. R.: The phylogeny of prokaryotes. Science 209, 457-463 (1980).

FRANKS, A. H., HARMSEN, H. J. M., RAA NGS, C. c., JANSEN, G. J., SCHUTS, E, WELLING, C. W.: Variations of bacterial popu­lation sin human feces measured by fluorescent in situ hy­bridization with group-specific 16S rRNA-targed oligonu­cleotide probes. Appl. Environ. Microbiol. 64, 3336-3345 (1998).

GLOCKNER, F. 0., AMANN, R., ALFREIDER, A., PERNTHALER, J., PSENNER, R., TREBESIUS, K.-H., SCHLEIFER, K.-H.: An In situ Hybridization Protocol for Detection and Identification of

Planktonic Bacteria. System. App!. Microbiol 19, 403-406 (1996).

JUKES, T. H., CANTOR, C. R.: Evolution of protein molecules, pp. 21-132. In: N. H. MURANO (ed.), Mammalian protein metabolism, Academic Press, New York 1969.

KNOKE, M., BERNHARDT, H.: Spezielle Mikrookologie des Men­schen, Standorte, pp. 78-112. In: Mikrookologie des Men­schen (U. TAUBENECK, ed.) Akademie-Verlag Berlin 1985.

LUDWIG, W., KIRCHHOF, G., KLUGBAUER, N., WEIZENEGGER, M., BETZL, D., EHRMANN, M., HERTL, c., JILG, S., TATZL, R., ZITZEKSBERGER, H., LIEBL, S., HOCHBERGER, M., SHAH, J., LANE, D., WALLNOFER, P. W., SCHLEIFER, K. H.: Complete 23S ribosomal RNA sequences of Gram-positive bacteria with a low DNA G+C content. System. Appl. Microbiol. 15, 487-501 (1992).

MANZ, W., AMANN, R., LUDWIG, W., WAGNER, M., SCHLEIFER, K.-H.: Phylogenetic oligodeoxynucleotide probes for the major subclasses of proteobacteria: problems and solutions. System. Appl. Microbiol. 15,593-600 (1992).

MANZ, W., SZEWZYK, u., ERICSON, E., AMANN, R. I., SCHLEIFER, K.-H., STENSTROM, T.-A.: In situ identification of bacteria in drinking water and adjoining biofilms by hybridization with 16S- and 23S-rRNA-directed fluorescent oligonculeotide probes. Appl. Environ. Microbio!. 59,2293-2298 (1993).

MANZ, W., AMANN, R., LUDWIG, W., VANCANNEYT, M., SCHLEI­FER, K.-H.: Application of a suite of 16S rRNA-specific­oligonucleotide probes designed to investigate bacteria of the phylum Cytophaga-Flavobacterium-Bacteroides in the natu­ral environment. Microbiol. 142, 1097-1106 (1996).

MCCARTNEY, A. L., WENZHI, W., TANNOCK, G. W.: Molecular analysis of the composition of the Bifidobacterial and lacto­bacillus microflora of Humans. Appl. Environ. Microbiol. 62,4608-4613 (1996).

MOORE, W. E. c., HOLDEMAN, L. Y.: Human fecal flora: the normal flora of 20 Japanese-Hawaiians. App!. Microbiol. 27,961-979 (1974).

NEEt~ A.: Anwendung der in situ-Einzelzellidentifizierung von Bakterien zur Populationsanalyse in komplexen mikrobiellen Biozonosen. PhD thesis at the Technical University Munich, Germany (1996).

PORTER, K. G., FEIG, Y. S.: The use of DAPI for identifying and counting aquatic microflora. Limnol. Oceanogr. 25, 943-948 (1980).

ROLLER, c., WAGNER, M .. , AMANN, R., LUDWIG, w., SCHLEIFER, K. H.: In situ probing of Gram-positive bacteria with high DNA G+C content by using 235 rRNA-targeted oligonu­cleotides. Microbiology 140,2849-2858 (1994).

SAITOU, N., NEI, M.: The neighbor-joining method: A method for reconstructing phylogenetic trees. Mol. BioI. Evol. 4, 1906-425 (1987).

SCHLEIFER, K. H., LUDWIG, W.: Phylogenetic relationships among bacteria, pp. 103-117. In: The Hirachy of Life (B. FERNHOLM, K. BRENNER, and H. JORNVALL, eds.), Elsevier Science Publishers, Amsterdam (1989).

SCHLEIFER, K.-H., LUDWIG, W.: Phylogeny of the genus Lacto­bacillus and related genera. System. Appl. Microbiol. 18, 461-467 (1995).

STILES, M. E., HOLZAPFEL, W. H.: Lactic acid bacteria of foods and their current taxonomy. Int. J. Food Microbio!' 36: 1-29 (1997).

STRUNK, 0., GROSS, 0., REICHEL, B., MAY, M., HERRMAN, S., STUCKMANN, N., NONHOFF, B., GINHART, T., VILBIG, A., LENKE, M., LUDWIG, T., BODE, A., SCHLEIFER, K.-H., LUDWIG, W.: http:\\www.mikro.biologie.tu-muenchen.de. Department of Microbiology, Technische Universitat Miinchen, Munich, Germany (1998).

SUGGS, S. V., HIROSE, T., MIYAKE, T., KAWASHIMA, E. H.,

196 H. MEIER et a!.

JOHNSEN, M. J., ITAKURA, K., WALLACE, R. B.: Use of synthet­ic oligodeoxynucleotides for the isolation of specific cloned DNA sequences, pp. 683-693. In: Developmental Biology Using Purified Genes. (D. BROWN, and C. F. Fox, eds.), Aca­demic Press, New York (1981).

VAN DER PEER, Y., NEEFS, J.-M., DE RI]K, P., DE Vos, P., DE WACHTER, R.: About the order of divergence of the major bacterial taxa during evolution. System. App!. Microbio!. 17, 32-38 (1994).

WAGNER, M., AMANN, R., H . LEMMER, SCHLEIFER, K. H.: Prob­ing activated sludge with proteobacteria-specific oligonu­cleotides: inadequacy of culture-dependent methods for de­scribing microbial community structure. App!. Environ. Mi­crobio!. 59.1520-1525 (1993).

WAGNER, M ., ERHART, R., MANZ, W., AMANN, R., LEMMER, H., WEDI, D., SCHLEIFER, K. H.: Development of an rRNA-targed oligonucleotide probe for the genus Actinobacter and its ap­plication for in situ monitoring in activated sludge. App!. En­viron. Microbio!. 60, 792-800 (1994).

WOESE, C. R.: Bacterial evolution. Microbio!. Rev. 51: 221-271 (1987).

ZARDA, B., AMANN, R., WALLNER, G., SCHLEIFER, K. H. : Identifi­cation of single bacterial cells using digoxigenin-Iabelled, rRNA-targeted oligonucleotides. J. Gen. Microbio!. 137, 2823-2830 (1991).

ZARDA, B., HAHN, D., CHATZINOTAS, A., SCHONHUBER, W., NEEF, A., AMANN, R. I., ZEYER, J.: Analysis of bacterial com­munity structure in bulk soil by in situ hybridization. Arch. Microbiol. 168, 185-192 (1997).

Corresponding author's present address: Dr. HARALD MEIER, GSF-National Research Center for Environ­ment and Health, Flow Cytometry Group, Ingolstadter Land­stralSe 1, 85764 Neuherberg, Germany