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