System. App!. Microbio!. 21,251-259 (1998) g ________________ _©_G_us_ta_v_Fis_ch_er_V_er_la_ SYSTEIVIAllC AND APPLIED MICROBIOLOGY Classification and Identification of Propionibacteria Based on Ribosomal RNA Genes and PCR GOTTFRIED DASEN, JITKA SMUTNY, MICHAEL TEUBER and LEO MEILEl Institut fur Lebensmittelwissenschaften, Labor fur Lebensmittelmikrobiologie, Eidgenossische Technische Hochschule, ETH-Zentrum, Zurich, Switzerland Received February 15, 1998 Summary A rapid method was developed to differentiate the genus Propionibacterium from other genera by using a modified multiplex-PCR (MPCR) approach. Three 16S rRNA-targeted oligonucleotide primers were designed to amplify simultaneously two DNA-fragments in the MPCR assay. The universal primer pair bakll wand bak4 (corresponding to the E. coli 16S rRNA positions 8-25 and 1522-1540, respectively) was used in combination with the primer pair bak4 and gdl (5'- TGCTTTCGATACGGGTTGAC -3'). The later sequence corresponding to a 16S rRNA motif that is unique for the genus Propionibacterium. Propionibacteria were identified by the amplification of a Propionibacterium-genus specific 900-bp fragment whereas MPCR with DNA from other bacteria generated only a DNA fragment of 1500 bp in amplifications with the two universal primers. The whole procedure including cell lysis, MPCR amplification and analysis can be performed within 1 day, detection limits are at approximately 10 3 cfu propionibacteria (or 35 pg DNA). In addition, the taxonomic situation of the genus Propionibacterium was reexamined using a cycle sequencing strategy. Based on the 16S rDNA, a phylogenetic tree of all the Propionibacterium type strains was reconstructed. Key words: Propionibacterium, 16S rRNA, multiplex PCR, phylogenetic tree, cycle sequencing Introduction Propionibacteria form a relatively homogenous taxonomic cluster (BRITZ & RIEDEL, 1991) within the class Actinobacteria that was recently defined by STACKEBRANDT et al. (1997). They are Gram-positive, rodshaped and catalase-variable bacteria that prefer anaerobic growth conditions. Their most characteristic property is the production of propionic and acetic acid as main metabolites from lactic acid as substrate. Propionibacteria are found in dairy products like Swiss-type cheese and on the skin or within the human body. They are divided in two main groups: the dairy and the medically important (or cutaneous) propionibacteria. Dairy propionibacteria are important in the cheese ripening whereas cutaneous strains are involved in certain human diseases (for example acne, endophthalmitis). Propionibacteria are also discussed to be utilized in the industrial production of vitamin Bll and propionic acid (CUMMINS & JOHNSON, 1992; GLATZ, 1992). Various approaches to classify propionibacteria have been carried out. Recently the new species Propionibacterium cyclohexanicum, isolated from orange juice, has been proposed (KU5ANO et al., 1997), strains have been reclassified within the genus (DE CARVALHO et al., 1995) or from other genera (CHARFREITAG et al., 1988). Propionibacteria were grouped into 5 clusters using chemotaxonomy (BRITZ & RIEDEL, 1994). Time consuming methods based on proteinprofiles, serology and immunology have been described (BAER & RYBA, 1991, 1992) to differentiate between species. Another approach for that was ribotyping (DE CARVALHO et al., 1994; RIEDEL & BRITZ, 1996). An analysis of propionibacteria based on 16S rRNA sequences has been performed (CHARFREITAG & STACKEBRANDT, 1989) and restriction analysis of 16S rRNA genes as a tool to distinguish between species was proposed (RIEDEL et al., 1994; GAUTIER et al., 1996). A recent study (ROSSI et al., 1997) reported the use of restriction analysis of PCR-amplified 16S-23S rDNA spacer regions to separate between the different species of dairy propionibacteria. Reliable and fast detection of production strains as well as wild type strains of propionibacteria is not easy to perform and can be shortened and simplified by using 252 G. DASEN et al. DNA-probe techniques. In raw milk, these techniques could proof the occurrence of wild type strains that could replace the starter cultures and cause undesired side effects (for example red spots) during the ripening of cheese. Primers for the detection of cutaneous propionibacteria using peR or hybridization techniques have been described (GREISEN et a!., 1994; HYKIN et a!., 1994). However, primers for the detection of both medical and dairy propionibacteria by means of peR have so far not been investigated. The aim of this work is to provide a method for rapid and specific confirmation of propionibacteria by means of peR and the development of a fast and reliable sequencing strategy for 165 rRNA genes amplified with peR using an automatic DNA-sequencer. Materials and Methods Bacterial strains and maintenance: The strains that were used in this study are listed in Table 2. Propionibacteria were inoculated aerobically and grown under anaerobic conditions either in YEL-medium (MALIK et al. 1968) or in MRS-medium supplied with 0.5 gil L-cysteine-hydrochloride (Fluka). Anaerobic growth conditions were achieved by using gas-tight 20-ml McCartney tubes (broth media) or anaerobic jars (Oxoid) containing the Anaerocult A system (Merck). Dairy propionibacteria were grown at 30°C, cutaneous strains at 37 °C for 2 to 7 days. Strains were stored either at 4 °C or at -72 °C in 30% (v/v) glycerol. Strain purity was checked regularly by microscopy, colony morphology, Gram-test and catalase activity. All other strains were cultivated as recommended in the DSM catalogue of strains. DNA-isolation and quantification: Total DNA was extracted as described previously (LEENHOUTS et aI., 1989). DNA was quantified by gel-electrophoresis of the sample together with /...DNA (Gibco) as a standard. DNA isolation for PCR: We used the simple DNA extraction method of GOLDENBERGER et al. (1995) mediated by SDS and Proteinase K to obtain template DNA for PCR. Bacteria as templates for PCR: With a sterile toothpick, cells were selected from agar-plates and added instead of template DNA to the PCR-mixture. Alternately, 1-10 pi suspension of a liquid culture was added to the PCR-mixture and amplified. Multiplex-PCR-conditions: Target-DNA was amplified in 0.2 ml thin-walled tubes using thermocyclers (Personal Cycler, Biometra; Genius, Techne) equipped with heated lids. The 50 pi reaction mixture contained 0.2 mM each of dATP, dCTP, dGTP, dTTP (Pharmacia), 1 Unit Taq DNA-Polymerase (Pharmacia), 5 pi lOx standard PCR-buffer (Pharmacia), 2% Tween 20 (Merck), sterile bidestilled water, 1 pM forward primers bakll w, gd1 and 1 pM reverse primer bak4 (Table 1) and template DNA (simple method DNA or 1-2 ng total DNA). The amplification was performed after an initial denaturation step at 95°C for 3 min. by 40 cycles (95°C for 15 sec., 56°C for 30 sec., 72 °C for 2 min.). After a final polymerization step at 72 °C for 7 min., the samples were cooled to 4°C. PCR-products were analyzed by gel-electrophoresis (Pharmacia) on a 2 % agarose-gel (Gibco BRL) in Ix TBE-buffer and stored at -20°C. Table 1. Synthetic oligonucleotides, position and origin, targeted on the 16S rRNA gene Name Sequence E. coli - Position specificity gd1 gd6r bak4 TGCTTTCGATACGGGTTGAC GTGCCWAGGCATCCACCG AGGAGGTGATCCARCCGCA 632-651, fw 2003-2020, rev 1522-1540, rev bak11w uni1088 uni1392 uni1392r uni515 uni785 eubac338 ms350r AGTTTGATCMTGGCTCAG GGTTAAGTCCCGCAACGAGC GTACACACCGCCCGTCA TGACGGGCGGTGTGTAC ACCGCGGCTGCTGGCAC GGMTTAGATACCCTGGTAGTCC ACTCCTACGGGAGGCAGC CTGCTGCCTCCCGTA 8-25, fw 1088-1107, fw 1392-1408, fw 1392-1408, rev 515-531, rev 785-806, fw 338-355, fw 343-357, rev fw ongm application this study MPCR Genus Propionibacterium Propionibacteria, 23S rDNA this study Sequencing GREISEN et al. 1994 MPCR universal16S rDNA (modified) universal16S rDNA GOLDENBERGER 1997 MPCR AMANN et al. 1995 Sequencing universal16S rDNA universal16S rDNA LANE 1991 Sequencing LANE 1991 Sequencing universal16S rDNA universal16S rDNA LANE 1991 Sequencing universal16S rDNA AMANN et al. 1995 Sequencing universal16S rDNA AMANN et al. 1995 Sequencing universal16S rDNA LANE 1991 Sequencing =forward; rev =reverse; E. coli positions according to BROSIUS et al. 1978 (referred to start of 16S rDNA). Table 2. Bacterial strains and multiplex-PCR signals using primers gd1, bak11w and bak4 respectively. (1) (2) (3) (4) (5) (6) Strain Origin' Multiplex b Propionibacterium acidipropionici (arabinosum) Propionibacterium acidipropionici (pentosaceum) Propionibacterium acidipropioniclcr Propionibacterium acnesT Propionibacterium avidumT Propionibacterium cyclohexanicumT DSM20273 DSM20272 DSM4900 DSM 1897 DSM4901 lAM 14535 + + + + + + Classification and Identification of Propionibacteria Based 253 Table 2. (Continued). (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) (28) (29) (30) (31) (32) (33) (34) (35) (36) (37) (38) (39) (40) (41) (42) (43) (44) (45) (46) (47) (48) (49) (50) (51) (52) (53) (54) (55) (56) (57) (58) (59) (60) (61) (62) Strain Origin a Multiplex b Propionibacterium freudenreichii Propionibacterium freudenreichii Propionibacterium freudenreichii Propionibacterium freudenreichii Propionibacterium freudenreichii Propionibacterium freudenreichii Propionibacterium freudenreichii subsp. freudenreichi? Propionibacterium freudenreichii subsp. shermanii Propionibacterium freudenreichii subsp. shermaniiT Propionibacterium granulosum T Propionibacterium jensenii Propionibacterium jensenii (petersonii) Propionibacterium jensenii (rubrum) Propionibacterium jensenii (zeae) Propionibacterium jenseniiT Propionibacterium lymphophilum T Propionibacterium propionicum T', Propionibacterium thoeniif Propionibacterium sp. JS53 Propionibacterium sp. MI20 Propionibacterium sp. MI6 Propionibacterium sp. Z/S Acetobacter diazotrophicus T Actinomyces georgiaeT Arthrobacter citreusT Bacillus subtilis Bacillus thuringiensisT Bifidobacterium lactis T Bifidobacterium minimumT Brevibacterium linens T Cellulomonas udaT Corynebacterium variabilisT Enterobacter sp. Enterococcus faecalis F01 Enterococcus faecalis JH2-2c Enterococcus faecalis RE25 Enterococcus sp. RE39 Escherichia coli S17-1 Escherichia coli XU-blue, with pUC18 Gardnerella vaginalis T Lactobacillus acidophilus Lactobacillus casei Lactobacillus rhamnosus Lactococcus lactis subsp. cremoris AC1 Lactococcus lactis subsp. lactis BU2-60 Lactococcus lactis subsp. lactis K214 Leuconostoc mesenteroides Microbacterium lacticum T Dermacoccus nishinomiyaensisT (Micrococcus nishinomiyaensis) Micromonospora sp. Microbacterium arborescensT Mycobacterium chlorophenolicum (Rhodococcus chlorophenolicusT ) Nocardia sp. Streptococcus thermophilus 55n RlS Streptococcus thermophilus S203 Streptomyces azureusT FAM 1409 FAM 1410 FAM 1411 FAM 1412 FAM 1413 FAM 1414 DSM20271 DSM 20270 DSM4902 DSM20700 DSM 20278 DSM20279 DSM20275 DSM20274 DSM 20535 DSM4903 DSM 43307 DSM20276 LME LME LME LME DSM5601 DSM 6843 DSM 20133 Wiesby 168 DSM2046 DSM 10140 DSM 20102 DSM 20425 DSM 20107 DSM20132 LME LME LME LME LME LME LME DSM4944 Wiesby 145 Wiesby 160 Wiesby 744 LME LME LME GD14 DSM 20427 DSM20448 DSM 1043 DSM 20754 DSM 43826 DSM 43280 LME LME DSM 40106 + + + + + + + + + + + + + + + + + + + + + H H (-) DSM - Deutsche Sammlumg fur Mikroorganismen und Zellkulturen, Braunschweig, Germany; FAM - Forschungsanstalt fur Milchwirtschaft, Liebefeld-Bern, Switzerland; LME - Laboratory of Food Microbiology, ETH, Zurich, Switzerland. lAM - Institute of Molecular and Cellular Biosciences, University of Tokyo, Japan. Wiesby - Wiesby GmbH & Co. KG, Niebull, Germany. b Multiplex PCR results: + Genus Propionibacterium (900-bp Fragment.), - other Bacteria, nd not determined. (-) DNA as Template 'f "Correct" latin form of the name. The strain is usually cited as P. propionicus. T Type strain. Names in brackets: former strain designation. a 254 G. DASEN et al. Table 3. 16S rRNA sequences used in this study and amplicon sizes of MPCR reactions with propionibacteria. Strain Bifidobacterium lactis T Escherichia coli Luteococcus japonicus IF012422 P. acidipropioniciTDSM 20272 P. acnesTDSM 1897 P. avidumTDSM 4901 P. avidumTDSM 49012 P. cyclohexanicumTIAM 14535 P. shermaniiTDSM 4902 P. freudenreichii TDSM 20271 P. granulosumTDSM 20700 P. granulosumTDSM 20700 2 P. jenseniiTDSM 20535 P.lymphophilumTDSM 4903 P. lymphophilumTDSM 490}2 P. propionicumTDSM 43307 P. propionicumTDSM 433072 P. thoenizTDSM 20276 Ace. no. (GenEmbl) DSM 10140 J01695 DS5487 X53221 M61903 U02905 AJ003055 D82046 YI0819 X53217 U02906 AJ003057 X53219 U02907 AJ003056 X53216 AJ003058 X53220 length! Added 2 (bp) (bp) XS9513 750S (0) 1506 (1) 1351 (29) 1530 (18) 385 (0) 1504 (0) 1471 (0) 5116 (28) 1462 (26) 385 (1) 1514 (0) 1364 (33) 388 (0) 1502 (0) 1179 (148) 1506 (0) 1467(33) 2104 (0) bak11w bak4" (bp) gd1 bak4" (bp) 1522 1513 915 895 26 1510 1508 1507 1506 895 891 889 889 126 1517 1514 900 892 1508 893 1512 1515 896 893 134 54 Type strain. ambiguous bases 2 The whole 16 rDNA or parts of it were resequenced in this study. " Length of amplified PCR-fragment (including primers) using the respective primer combination. T 1 Purification of PCR-products: PCR products were purified with the Nucleospin-Extract kit (Macherey & Nagel) and eluated with 10 mM Tris-HCI, pH 8.5 (Merck). Cycle Sequencing: PCR-products covering almost the complete 16S rDNA were generated using the multiplex-PCR protocol with primers bakll wand gd6r. 200-300 ng of the purified PCR products (8 pi) were sequenced using the "Thermo Sequenase fluorescent labelled primer cycle sequencing kit with 7deaza dGTP" (Amersham). Sequencing primers for conserved regions of the 16S rRNA gene were synthesized and labeled with cy5 (Microsynth, Balgach, Switzerland). The amplification program consisted of 25 cycles with 30 sec. at 95°C, 30 sec. 50 °C and 1 min. at 72 0c. Samples were cooled to 4 °C and 5 pi stop solution (Amersham) was added. The products were analyzed on an ALF-Express automatic DNA-sequencer (Pharmacia). Data searches, analysis and phylogenetic calculations. DNA sequences were analyzed with the programs provided in the GCG-package version S.O (Genetics Computer Group, Madison, Wisconsin, USA). 16S rDNA sequences and accession numbers are listed in Table 3. Results and discussion Multiplex peR Using a standard PCR protocol with primers gdl and bak4 alone, for propionibacteria a single fragment of 900 bp approximately (889-915 bp, Table 3) was amplified whereas with template-DNA from other organisms no amplification product was detected. Since it was not possible to determine whether the absence of a signal was due to no amplification or limited DNA extraction it was necessary to run a second PCR using primers bak 11 w and bak4. In order to simplify the method we developed a multiplex-PCR like approach. Multiplex PCR (MPCR) is commonly used to amplify 2 or more loci simultaneously. The method has been already been applied to detect pathogens in food samples (DENG & FRATAMICO, 1996; WITHAM et ai., 1996). The difference between "classical" MPCR and our method is that the same gene (165 rDNA) is the target for all primers which results in a more complex and less predictable amplification. Propionibacteria could be distinguished from other organisms by amplification of a specific 900 bp fragment amplified with primers gdl and bak4. All other organisms showed only a "universal" fragment of 1500 bp resulting of amplification with primers bakll wand bak4. The detection limits for propionibacteria were 2x 10 3 colony forming units (cfu) (Figure 5). In our MPCR approach using Propionibacterium DNA oligonucleotides gdl, bakllw and bak4 should theoretically form 2 fragments (Figure 1, approx. 900 bp and 1500 bp), but under the selected conditions and DNA-concentrations only the specific fragment of 900 bp was amplified in sufficient amounts. Depending on DNA-concentrations and annealing temperature both fragments could be amplified (Figure 3 and Figure 4) for propionibacteria. In all tested strains of propionibacteria the specific fragment was amplified and no false positive signals were found to occur for other organisms (Figure 2 and Table 2). Since the aim of our study is the differentiation of the genus Propionibacterium from other gen- Classification and Identification of Propionibacteria Based bakllw 255 gd1 2000 1 -1---1---1---1---+--+--+--+--+--+--+--+--+--+--+-1--+---+---+---+---tl .... • -1 gd6r bak4 889 bp 1508 bp } Multiplex peR Sequencing 1891 bp Fig. 1. Graphical representation of oligonucleotide-positions and amplification products applied on P. freudenreichii subsp. shermanii (Accession no. YI0819). Multiplex PCR reaction targeted on 16S rDNA resulting in a 889 (primers bak11 w, gdl) and a 1508 bp-sized (primers bakllw, bak4) fragment. Primers bakllw, gd6r are indicated for the amplification of the 16S rDNA including spacer DNA (up to 5'-end of 23S rDNA) resulting in a fragment of 1891 bp. Fig. 2. Detection of Genus Propionibacterium using Multiplex PCR (primers bak11 w, gdl, bak4). DNA from propionibacteria and other organisms (isolated with a simple method) was submitted to the MPCR protocol and analyzed on a 2 % TBE agarose gel. Propionibacteria (lanes 1-28) are characterized by the amplification of a specific 900 bp fragment. For other organisms (lanes 29-62) only a universal 1500 bp fragment is amplified. The numbering of the lanes corresponds to the strains listed in Table 2. The lane "kb" contains a kb-ladder (Gibco). The lane (-) means MPCR-reaction without DNA as a negative control. 256 G. DASEN et al. Fig. 3. "Single" PCR-reactions in comparison to MPCR-reactions and effect of varying annealing temperatures (TA ) on PCR-product formation. 2 ng DNA from P. shermanii were submitted to the MPCR reactions (Lanes 1-3: T A = 56°C, Lanes 4-6: T A = 60°C) to demonstrate a temperature dependent formation of an universal 1500 bp fragment for propionibacteria (fragment in lane 5, marked with an arrow). MPCRreactions (lanes 2 and 4: primers bak11 w, gd1, bak4) resulting in the same amplification products as the corresponding standard ("single") PCR reactions (lanes 1 and 4: primers bak11 wand bak4; lanes 3 and 6: primers gd1, bak4). Fig. 4. Effect of template-DNA quantity on multiplex PCR reaction. To amplify both the universal and the specific fragment in propionibacteria MPCR reactions with various template DNA concentrations of P. freudenreichii subsp. shermanii were performed (lanes 1 to 5: Multiplex PCR product with template DNA concentrations of 3.5 pg, 35 pg, 350 pg, 1.75 ng and 3.5 ng respectively). At a template concentration of 1. 75 ng DNA the universal fragment was amplified (lane 4, marked with an arrow).The lane "kb" contains a,kb-Iadder (Gibco). The lane (-) means MPCR-reaction without DNA as a negative control. Fig. 5. Detection limit of multiplex PCR. To determine the detection limits of our MPCR method, different cell-concentrations (lanes 1-6: 2x106, 2x105, 2x10\ 2x103, 2x102, 2x10 1 du) of P. freudenreichii subsp. shermanii were prepared and analyzed with our method. The lane "kb" contains a kb-Iadder (Gibco). The lane (-) means MPCR without DNA as a negative control. B.lactis E. coli P. Iymphophilum P. subsp.shermanii 1% P. P. cyclohexanicum P. acidipropionici P. granulosum Fig. 6. Phylogenetic tree of the genus Propionibacterium based on 16S rDNA sequences. The sequences were aligned from base 28 to 1521 (E. coli numbering) with the program eclustalw (EGCG-package).The tree was constructed based on all positions of the alignment using a maximum likelihood algorithm with molecular clock (dnamlk, PHYLIP-package). Classification and Identification of Propionibacteria Based 257 cld prOp10nlC1 .cn $ P. P. P. P. P. P. P. P. P. P. P. Fig. 7. Alignments of 165 rDNA sequences from propionibacteria (Table 3) with the genus-specific primers gd1 (this study) and RDR514 (GREISEN et ai., 1994). Ambiguous or nonmatching positions are marked in bold type. (*) marks sequences we got in our study and which differ from those in the databases. era which is achieved by the presence of the specific 900bp fragment, it was decided that a further optimization (HENEGARIU et aI., 1997) of the protocol was not necessary. Primer specificity Different researchers have already described specific primers for the detection of propionibacteria. The primer RDR514 (E. coli position 1376-1400) was tested and proposed as hybridization probe to detect P. acnes, P. avidum, P. lymphophilum and P. granulosum (GREISEN et aI., 1994). However this primer was not tested with the classical strains, but an alignment with these 16S rDNA sequences shows certain mismatches (Figure 7). Since the primer is relatively long (25 bp) its suitability for PCR may not be optimal. PCR-primers for the detection of P. acnes have been described (HYKIN et aI., 1994) but they are limited to the detection of this species. Primers for the amplification of Propionibacterium 16S rDNA fragments have been used (RIEDEL et aI., 1994) but these primers are not specific for the genus Propionibacterium. The alignment of our primer, gdl (E. coli position 632-651), is shown in Figure 7. In a FASTA-comparison of this primer with all sequences of the GenEmbl database (2,073,516 entries, March 1998) only 15 sequences showed at least 95% homology (in all 20 nucleotides) with the target sequence. From these sequences 8 were propionibacteria, 4 belonged to organisms that were not further identified to genus level, 1 sequence of Actinomyces israelii serotype 1 (Accession no. X53228), Mycobacterium chitae (Accession no. M29560) and Eubacterium combesii (Accession no. L34614) showed 100% identity. These sequences show lower similarity to other sequences from the same genera than to propionibacteria: A. israelii has a similarity of 96.3 % to P. acnes with only 84% similarity to other species from the genus Actinomyces. M. chitae is closer related to P. acnes (98.3%) than to other Mycobacterium species (87%). E. combesii clusters with P. thoenii (92.3%) and shows only 78% similarity with other Eubacterium species. That in- dicates either the need of reclassification and re-evaluation of these strains or sequencing problems. The 16S rDNA of the new classified strain P. cyclohexanicum shows 1 mismatch with our primer sequence. Since this strain is relatively new (KUSANO et aI., 1997) its occurrence in other habitats than orange juice has not been investigated. The strain could be detected using the MPCR technique by lowering the annealing-temperature from 60°C to 56 °C resulting in a slightly lower specificity for the detection and, in certain cases, the formation of a third fragment of approximately 2 kb (Figure 2, P. avidum and P. cyclohexanicum), due to unspecific binding of the oligonucleotides at the lower annealing temperature. This problem could possibly be solved by using degenerated primers with a mixture of A and C at position 18 of the oligonucleotide gdl but was not further investigated. Phylogenetic analysis of the genus Propionibacterium The 16S rDNA of all type strains was completed or resequenced as far as possible. This was necessary to obtain as much nucleotides as possible to compute an optimal alignment (LUDWIG & SCHLEIFER, 1994). A phylogenetic tree of the genus including the medical strains was calculated (maximum likelihood method) based on the alignment from E. coli positions 28 to 1521. The strains could be grouped in different clusters (Figure 6): The classical species P. acidipropionici, P. jensenii and P. thoenii form 3 closely related and P. freudenreichii (both subspecies) and P. cyclohexanicum 2 clusters. These data were in accordance with the phylogenetic trees of KuSANO et al. (1997) and CHARFREITAG & STACKEBRANDT (1989). The cutaneous species separate in 3 closer related clusters with P. acnes, P. avidum and P. propionicum. P. granulosum forms a separate cluster between the classical and cutaneous groups whereas P. lymphophilum is the most distant relative of the whole genus. In fact, the 16S rRNA sequence of P. lymphophilum shows more similarity (91.1 %) with Luteococcus japonicus than with other propionibacteria (90.7% with P. freudenreichii). 258 G. DASEN et a!. Sequencing strategy The sequencing strategy using PCR-products instead of cloned fragments allowed us to amplify parts of the 165 rDNA of unknown organisms in a short time. Within 2 days an organism could be identified according to the similarity of its 165 rRNA with known organisms of the database. Almost the complete 165 rRNA gene could be sequenced, except the bases corresponding to the E. coli positions 1-8 could not be sequenced since the amplification-product of primers bak11 wand gd6r did not cover this region. To avoid mistakes due to the Taq polymerase, different PCR products were used as templates for the sequencing reactions. Control-sequencing of already known sequences showed a good agreement of the data we obtained. The same strategy was applied to obtain a 235 rDNA sequence of P. freudenreichii subsp. shermanii. A PCR product using primers gd1 and gd5sr ranging from the start of the 165 rDNA to the middle of the 55 rDNA was used as template. 5equencing primers were selected from conserved regions of the 235 rRNA. The obtained sequence showed a high similarity with the 235 rRNA sequence of P. freudenreichii subsp. freudenreichii (W. LUDWIG, unpublished). Hybridization studies Further studies to quantify propionibacteria directly in mixed cultures are under investigation by using an approach similar to methods developed in our laboratory for other Actinobacteria (KAUFMANN et a!., 1997; KOLLOFFEL et a!., 1997). Acknowledgement This work is supported by a Grant from the Swiss National Fonds SPP Biotechnology 5002-041663/1. References AMANN, R. 1., LUDWIG, W., SCHLEIFER, K. H.: Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbio!. Rev. 59, 143-169 (1995). BAER, A., RYBA, 1.: Identification of Propionibacteria and Streptococci by immunoblotting. Milchwissenschaft 46, 292-294 (1991). BAER, A., RYBA, 1.: Serological identification of Propionibacteria in milk and cheese samples. Int. Dairy Journal 2, 299-310 (1992). BRITZ, T. ]., RIEDEL, K.-H. ].: A numerical taxonomic study of Propionibacterium strains from dairy sources. ]. App!. Bacteriol. 71,407-416 (1991). BRITZ, T. J., RIEDEL, K.-H. ].: Propionibacterium species diversity in Leerdammer Cheese. Int. J. Food Microbiology 22, 257-267 (1994). BROSIUS, ]., PALMER, M. L., KENNEDY, P. J., NOLLER, H. E: Complete nucleotide sequence of a 16S ribosomal RNA gene from Escherichia coli. Proc. Natl. Acad. Sci. USA 75, 4801-4805 (1978). DE CARVALHO, A. E, GAUTIER, M., GRIMONT, E: Identification of dairy Propionibacterium species by rRNA gene restriction patterns. Res. Microbiol. 145,667-676 (1994). DE CARVALHO, A. E, GUEZENEC, S., GAUTIER, M., GRIMONT, P. A. D.: Reclassification of "Propionibacterium rubrum" as P. jensenii. Res. Microbiol. 146,51-58 (1995). CHARFREITAG, 0., COLLINS, M. D., STACKEBRANDT, E.: Reclassification of Arachnia propionica as Propionibacterium propionicus comb. nov. Int.]' Syst. Bacteriol. 38, 354-357 (1988). CHARFREITAG, 0., STACKEBRANDT, E.: Inter- and intrageneric relationships of the genus Propionibacterium as determined by 16S rRNA sequences. J. Gen. Microbiol. 135, 2065-2070 (1989). CUMMINS, C. S., JOHNSON, J. L.: The genus Propionibacterium pp. 834-849. In: The Prokaryotes (A. BALOWS, H. G. TRDPER, M. DWORKIN, W. HARDER, K. H. SCHLEIFER, eds.) 2nd ed., Vol. II, New York, Springer-Verlag (1992). DENG, M. Y., FRATAMICO, P. M.: A multiplex PCR for rapid identification of shiga-like toxin-producing Escherichia coli 0157:H7 isolated from foods. ]. Food Prot. 59, 570-576 (1996). DSM: Catalogue of strains. (1993). GAUTIER, M., DE CARVALHO, A., ROUAULT, A.: DNA fingerprinting of dairy propionibacteria strains by pulsed-field gel electrophoresis. Curr. Microbio!. 32,17-24 (1996). GLATZ, B. A.: The classical propionibacteria: Their past, present and future as industrial organisms. ASM news 58, 197-201 (1992). GOLDENBERGER, D., PERSCHIL, I., RITZLER, M., ALTWEGG, M.: A simple "universal" DNA extraction procedure using SDS and Proteinase K is compatible with direct PCR amplification. PCR Methods and Applications 4, 368-370 (1995). GOLDENBERGER, D.: Detection of bacterial pathogens in clinical specimens by broad-range PCR amplification and direct sequencing of part of the 16S rRNA gene. Thesis. Swiss Federal Institute of Technology, Zurich, Switzerland (1997). GREISEN, K., LOEFFELHOLZ, M., PUROHIT, A., LEONG, D.: PCR primers and probes for the 16S rRNA gene of most species of pathogenic bacteria, including bacteria found in cerebrospinal fluid.]. Clin. Microbiol. 32, 335-351 (1994). HENEGARIU, 0., HEEREMA, N .A., DLOHY, S. R., VANCE, G. H., VOGT, P. H.: Multiplex PCR: Critical parameters and stepby-step protocol. BioTechniques 23, 504-511 (1997). HYKIN, P. G., TOBAL, K., McINTYRE, G., MATHESON, M. M., TOWLER, H. M. A., LIGHTMAN, S. L.: The diagnosis of delayed post-operative endophthalmitis by polymerase chain reaction of bacterial DNA in vitreous samples. J. Med. Microbiol. 40, 408-415 (1994). KAUFMANN, P., PFEFFERKORN, A., TEUBER, M., MElLE, L.: Identification and quantification of Bifidobacterium species isolated form food with genus-specific 16S rRNA-targeted probes by colony hybridization and PCR. Appl. Environ. Microbiol. 63, 1268-1273 (1997). KOLLOFFEL, B., BURRI, S. MElLE, L., TEUBER, M.: Development of 16S rRNA oligonucleotide probes for Brevibacterium, Micrococcus/Arthrobacter and Microbacterium/Aureobacterium used in dairy starter cultures. System. Appl. Microbio!. 20,409-417 (1997). KUSANO, K., YAMADA, H., NIWA, M., YAMASATO, K.: Propionibacterium cyclohexanicum sp. nov., a new acid-tolerant 0)cyclohexyl fatty acid-containing Propionibacterium isolated from spoiled orange juice. Int. J. Syst. Bacterio!' 47, 825-831. (1997). LANE, D.].: 16S/23S rRNA sequencing, p.115-175. In: Nucleic acid techniques in bacterial systematics. (E. STACKEBRANDT, M. GOODFELLOW, eds.). Chichester, Wiley & Sons Ltd (1991). LEENHOUTS, K .J., KOK, J., VENEMA, G.: Campbell-like integration of heterologous plasmid DNA into the chromosome of Lactococcus lactis subsp. lactis. Appl. Environ. Microbiol. 55,394-400 (1989). Classification and Identification of Propionibacteria Based LUDWIG, W., SCHLEIFER, K. H.: Bacterial phylogeny based on 16S and 23S rRNA sequence analysis. FEMS Microbiol. Rev. 15,155-173 (1994). MALIK, A. C., REINBOLD, G. W., VEDAMUTHU, E. R.: An evaluation of the taxonomy of Propionibacterium. Can. J. Microbi01. 14, 1185-1191 (1968). RIEDEL, K.-H. J., BRITZ, T. J.: Justification of the classical Propionibacterium species concept by ribotyping. System. Appl. Microbiol. 19,370-380 (1996). RIEDEL, K.-H. J., WINGFIELD, B. D., BRITZ, T. J.: Justification of the "classical" Propionibacterium species concept by restriction analysis of the 16S ribosomal RNA genes. System. Appl. Microbiol.17, 536-542 (1994). ROSSI, F., SAMMARTINO, M., TORRIANI, S.: 16S-23S ribosomal spacer polymorphism in dairy propionibacteria. Biotechnology Techniques 11, 159-161 (1997). 259 SAMBROOK, J., FRITSCH, E. F., MANIATIS, T.: Molecular cloning, a laboratory manual. Cold Spring Harbour Laboratory Press, Cold Spring Harbour, New York (1989). STACKEBRANDT, E., RAINEY, F. A., WARD-RAINEY, N. L.: Proposal for a new hierarchic classification system Actinobacteria classis nov.. Int. J. Syst. Bacteriol. 47, 479-491 (1997). WITHAM, P. K., YAMASHIRO, C. T., LrVAK, K. J., BATT, C. A.: A PCR-based assay for the detection of Escherichia coli shigalike toxin genes in ground beef. Appl. Environ. Microbiol. 62,1347-1353 (1996). Corresponding author: Dr. LEO MElLE, Institut fur Lebensmittelwissenschaften, Labor fur Lebensmittelmikrobiologie, ETH, LFO G20, CH - 8092 Zurich Fax: +41 1 632 12 66; e-mail: meile@ilw.agr.ethz.ch