Journal of General Microbiology (1983), 129, 557-563.
Printed in Great Britain
557
Differentiation between Gram-negative Anaerobic Bacteria by Pyrolysis
Gas Chromatography of Lipopolysaccharides
By G U N N A R D A H L E N 1 * A N D I N G E R ERICSSON2
Department of Oral Microbiology, Institute of Medical Microbiology, University of Goteborg,
Guldhedsgatan 10, S-413 46 Goteborg, Sweden
Department of Analytical Chemistry, University of Lund, P.O.Box 740, S-22007 Lund, Sweden
(Received 23 March 1982; revised 25 August 1982)
Lipopolysaccharidesextracted by phenol-water from nine strains of Gram-positive anaerobic
bacteria (Bacteroides, Fusobacterium and Veillonella),have been examined by means of pyrolysis
gas chromatography. Lipopolysaccharides were fragmented into a group of low molecular
weight components and four characteristic high molecular fractions probably consisting of
hydrocarbons from the lipid part of the material. The latter fractions were specific for each of
the genera tested. At the species level characteristic differences were also found although a
limited number of strains were tested. Due to the high reproducibility of the technique, the
potential of using the method in differentiating Gram-negative anaerobic bacteria was
indicated.
INTRODUCTION
Gram-negative anaerobic non-sporeforming bacteria represent a major group of the normal
flora on the mucous membranes of man (Rosebury, 1962) and they have been frequently isolated
from clinical material (Finegold, 1977). The identification of anaerobic bacteria is often time
consuming and does not always lead to a definite result due to an incomplete taxonomy and lack
of enough numbers of stable and specific tests for differentiation.
Pyrolysis gas chromatography (Py-GC) has been developed as a method for analysis of
complex organic polymers (Drucker, 1981; Ericsson, 1975, 1977). Py-GC of whole microorganisms has proved useful in the differentiation of several bacterial genera (Haddadin et al.,
1973; Quinn, 1974; Oxborrow et al., 1976; Needleman & Stuchbery, 1977; Emswiler & Kotula,
1978; Stack et al., 1978; Gutteridge & Norris, 1979). Several of these authors have noted that
there are problems in the reproducibility and evaluation of the complex pyrolysates that are
formed from micro-organisms. More reliable information may be obtained by using optimal
conditions for the pyrolyser (Ericsson et al., 1977) but the pyrograms are still very complex.
One way of producing simpler, but diagnostic, pyrograms is to use a macromolecule complex
having genus, species or serological group specificity.A potential macromolecule is the lipopolysaccharide (LPS) present in the outer membranes of Gram-negative bacteria. The polysaccharide part expresses the 0-antigenic specificity that constitutes the basis for differentiation
into chemotypes in the serological classification of species within Enterobacteriaceae
(Kauffmann, 1961, Luderitz et al., 1966). Strains of Gram-negative anaerobic genera such as
Veillonella and Fusobacterium have also been shown to be divided into chemotypes based on the
sugar content of LPS (Fredriksen & Hofstad, 1978, Hofstad, 1978).The lipid part (lipid A) of the
LPS also may contain specific differences between genera or even species of Gram-negative
anaerobic bacteria (Hofstad & Skaug, 1980, Wollenweber et al., 1980). LPS from Gramnegative anaerobic bacteria are genotypically stable macromolecules potentially suitable for
pyrolysis differentiation of micro-organisms.The present study evaluates the possible use of PyGC of LPS for the differentiation of Gram-negative anaerobic bacteria.
Downloaded
from www.microbiologyresearch.org by
0022-1277/83/O001-0488$02.00
01983 SGM
IP: 88.99.165.207
On: Mon, 19 Jun 2017 04:34:28
558
G.
DAHLBN
A N D I . ERICSSON
METHODS
Bacteria. Gram-negative anaerobic bacteria of the genera Bacteroides, Fusobacterium and Veillonella were
obtained from a variety of sources. Bacteroides oralis (Bact-MC3), Fusobacterium necrophorum (Fus-MC4) were
originally isolated together in an infected root canal in a monkey (Fabriciuset al., 1982). Fusobacterium nucleatum
(NCTC 10562), F. mortiferum (VPI 8365) and two strains of Veillonella (Ve 5, Ve 9) were kindly provided by T.
Hofstad (Bergen, Norway). Another two strains of Veillonella(Lim 14 :4 an and Lim 14 :4 cpr) were obtained from
teeth with infected necrotic pulps in humans (DahlCn & Bergenholtz, 1980) and one strain of Veillonella (Veill-H3)
was obtained from human dental plaque. The Veillonella strains were further classified into species according to
the VPI Manual (Holdeman & Moore, 1975) and Bergey's Manual of Determinative Bacteriology (Buchanan &
Gibbons, 1974).
Cultivation and cell preparations. The bacteria were grown at 37 "C in batch cultures, using screw-cap bottles
filled to the top. Brain Heart Infusion Broth (Difco 0037-01) with 0.03%(w/v) cysteine.HCI and 5% (w/v) yeast
autolysate supplemented with 0.01% (w/v) hemin and 0.01% (w/v) menadione according to the VPI Manual
(Holdeman & Moore, 1975), was used for Bacteroides and Fusobacteriumstrains. Cultivation of Veillonella strains
was performed in a medium containing proteose peptone (Difco 0122-01), 40 g; Na2HP04.2H20, 1.0 g;
KH2P0,, 0.4 g; 50% (w/v) sodium lactate, 25 ml and redistilled water, 965 ml. After sterilization, 10 ml of a salt
stock solution was added to the latter medium according to Linder et al. (1974). Cultivation was performed
anaerobicallyand the cells harvested by centrifugationin late-exponentialphase of growth. The cells were finally
washed twice with phosphate-buffered saline (pH 7.2).
Preparation of lipopolysaccharides (LPS). Bacterial cells were treated with equal volumes of phenol-water
(Westphal et al., 1952) (1 g cells to 100 ml liquid) at 20 "C for 15 min. The water phase was removed after
centrifugation and dialysed against tap water. LPS preparations were then lyophilized or further purified by
ultracentrifugation (100OOOg) and treated with DNAase and RNAase (Hofstad & Kristoffersen, 1970).
Chemical analysis. The amount of protein in the LPS preparation was estimated by the Lowry method using
tyrosin as standard.
Preparation of Py-GC samples. Lyophilized LPS preparations were suspended in distilled water (10 pg PI-') and
gently ultrasonicated for 30 s.
Py-GC equipment. The detailed equipment of the pyrolyser was described previously (TydCn-Ericsson, 1973;
Ericsson, 1977). Briefly, a pyrolyser with a platinum foil (2-6 mm wide, 15 mm long and 0.012 mm thick) was
heated by two current pulses. One of the pulses heated the foil to the temperature desired when the other
compensated for the cooling. The amplitude and time of the first pulse also determined the temperature rise time
of the pyrolyser. The platinum foil was placed in a glass cell maintained at a temperature of 175 "C. The pyrolyser
was installed in a flame ionization gas chromatograph (Varian 1860) fitted with a 2 m stainless steel column (id.
1-9mm) packed with 10%Apiezon L on Chromosorb 750. The nitrogen carrier gas and hydrogen flow rates were
both held at 20 ml min-l, and that of the air, approximately at 200 ml min-l. The temperature of the detector was
275 "C and the column temperature programmed from W250 "C at 15 "C min-l.
Py-GC test conditions and analysis. Prior to pyrolysis, 5 p1 of the sample was placed on the foil of the pyrolyser
and evaporated to dryness. Sampleswere pyrolysed to final temperaturesof 600,700or 800 "C during 0-1,2or 10 s
with a temperature rise time of 8 ms. Most of the pyrolyses were performed at a temperature of 700 "C for 2 s if not
stated otherwise.
RESULTS
Chemical analysis of LPS. The protein content of all LPS preparations used was lower than
10%.
Pyrolysis of LPS. The pyrolysis of LPS material showed a general pattern of peaks at very
short retention times, corresponding to volatile low molecular weight fragments and a few
characteristic peaks with longer retention times corresponding to fragments of higher molecular
weight. These latter peaks (designated CI2-Cl5) had the same retention times as normal
saturated hydrocarbons with 12-1 5 carbon atoms. Other characteristic peaks were numbered
1-8. Typical pyrograms of LPS of anaerobic bacterial species are shown in Figs 1 and 2.
Precision. Veillonella(Veill-H3,50p.g) was pyrolysed four times at 700 "Cfor 2 s. The standard
deviations for the heights of the four characteristic peaks are shown in Table 1.
Temperature and time. Pyrograms of LPS extracted from Veillonella strain Veill-H3 at three
different temperatures and three different times showed a similar pattern and only small
quantitative differences could be seen among the characteristic peaks. Maximal peak height
was not seen in the pyrogram of 600°C and 2s. The peaks eluting first (low molecular
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 04:34:28
559
Pyrolysis of lipopolysaccharides
Veillonella
Ve 5
Bacteroides
Bact-MC3
8
I
b
c15
Veillonella
Ve 9
I
2
X
00
c
.
I
Y
1
E
1
I
15
10
4
1
5
15
I
6
5
10
Time (min)
Time (min)
Fig. 1. LPS from one Bacteroides and five VeiZZoneZla strains (50 pg) pyrolysed at 700 "C for 2 s. The
are peaks with the same retention times as
characteristic peaks for each genus are indicated. C12-C16
for hydrocarbons with 12-1 6 carbon atoms, respectively.Other characteristic peaks are numbered 1-8.
A complex of small peaks ( x ) and peak no. 2 are present in the pyrogram of Ve 5 and Veill-H3.
Table 1. The precision obtained for the heights of the characteristicpeaks of LPS from
Veillonella ( Veill-H3)pyrolysed on four diflerent occasions during 2 d
Samples of 50 pg LPS were used, the end temperature was 700 "C and the pyrolysis time was 2 s.
Characteristic peaks (mm)
r
Pyrolysis
Mean value
S.D.
rzJ
1
c12
c
1
3
G.4
CIS
89
73
66
62
154
126
127
138
43
40
41
37
140
115
118
106
72
16
136
10
40
6
120
12
fragments) increased with increasing temperature and time, probably depending on a more
complete degradation to volatile fragments of the polysaccharide part of the LPS. The peaks
from the high molecular fragments remained more constant in height indicating the more stable
character of these components probably derived from lipids. We used 700 "C and 2 s in the
following experiments if not otherwise stated. Dose-responsestudies revealed that 20-50 yg was
suitable for analysis.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 04:34:28
560
G.
DAHLBN
A N D I . ERICSSON
Fusobacterium
15
10
5
0
Time (min)
Fig. 2. LPS (50 pg) from three strains of Fusubacteriurn pyrolysed at 700 "C for 2 s. The characteristic
peaks are indicated. CI2-Cl6are peaks with the same retention times as for hydrocarbons with 12-16
carbon atoms, respectively. Other characteristic peaks are numbered 1-8.
Genus dzferences. Pyrograms of nine strains belonging to three different genera were recorded
and a distinct pattern of characteristic peaks for each genus was found (Figs 1 and 2).
Bacteroides oralis always showed C15,c16, 5, 6 and 7 as characteristic peaks; Fusobacteriurn
strains always showed C13,C14, C I Sand
and VeillonellaCI2,cl3,
C14 and C l 5with no C16.
Strain dzferences. Five strains of Veillonella were tested to examine the homogeneity of the
patterns of characteristic peaks in the pyrograms of LPS (Fig. 1). They all showed the four
characteristic peaks C12,C13, C14 and C I 5that were common to the Veillonellagenus. Two of
the strains (Ve 5 and Veill-H3, both catalase negative indicating a designation as V . parvula)
showed a complex of small peaks in the figure marked with x and a small peak (2) close to the
C I 2peak which were not found in LPS of the three other strains (Ve 9, Lim 14 :4 cpr and Lim
14 : 4 an, all catalase positive indicating a designation of the strains as V . alcalescens). Three
strains of different species of Fusobacteriurn were tested (Fig, 2). They all showed the
characteristic C13,C14, C15 and CI6 peaks. Fusobacterium nucleaturn (NCTC 10562 and F.
necrophorum (Fus-MC4) showed a rather similar pattern, although there were differences that
might indicate strain or even species differences. The strain F . rnortiferurn (VPI 8365) showed a
more complex pattern with a quantitative lower amount of c16 and higher of C13.
DISCUSSION
Pyrolysis procedures have been used in attempts to characterize a variety of products
including micro-organisms. Whole micro-organisms give rise to complex patterns of peaks that
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 04:34:28
Pyrolysis of Iipopolysaccharides
561
can be identified by gas chromatography and/or mass spectrometry but problems are found in
identification and reproducibility (Quinn, 1974; Gutteridge & Norris, 1979). Pyrolysis of cell
fragments such as cell walls, flagella or DNA give less complex pyrograms (Emswiler & Kotula,
1978). The present study shows that the use of LPS gives pyrograms with a limited number of
characteristic peaks which may be more readily identified. The technique with the use of a foil
pyrolyser has been improved for characterization of complex organic materials (Ericsson, 1977 ;
Ericsson et al., 1977). The pyrolyser used in this study is a prototype but it is commercially
available from Dr 1. Ericsson (Lund, Sweden).
The present study showed that patterns and also quantitative results were quite constant
within certain limits with varying times and temperatures and peak heights were proportional to
doses. Specific fragments were therefore obtained over a broad range of conditions. The optimal
conditions seemed to be 700-800 "C in 0-1-2 s. Using other types of column or column material
it might be possible to increase the diagnostic power by resolving the profiles into a greater
number of characteristic peaks. The tendency of LPS to form aggregates in water solution
required efficient ultrasonication before analysis. The presence of small amounts of salts or
other low molecular weight compounds such as phenol residues did not seem to interfere with
the results significantly. Samples pyrolysed after a 1 day interval showed exactly the same
pattern. A slow decrease in sensitivity could be seen with five subsequently pyrolysed samples.
The foil should be regularly cleaned after each sample because of the accumulation of salt
residues from the samples.
The preparation and purification of LPS also seems to have little effect on the pyrolysis
results. It may therefore be possible to use the water phase after the phenol-water extraction
directly as a sample for pyrolysis, thus simplifying the procedures considerably.
LPS seemed to be characteristically fragmented into two distinct groups of fractions. The low
molecular weight fraction group was suggested to contain volatile fragments from sugars
(Gutteridge & Norris, 1979; unpublished observations). The yield of these fragments seems to
be more dependent on the pyrolysis conditions and the content of different sugars and probably
does not give useful characteristic pyrolysis profiles with the chromatographic conditions used.
The high molecular weight group of fractions are probably hydrocarbons formed from the lipid
part of the LPS. Based on the quantitative estimation of the peak heights mentioned, it may be
possible to indicate the sugar/lipid ratio of the LPS. It seems that the pyrolysis products formed
from the lipid part of the LPS are characteristic for each genus.
Hydroxy-fatty acids occur in Gram-negative bacteria where they are constituents of LPS.
Preparations from strains of the genus Bacteroides differ from LPS of other genera in the lipid
part by containing 3-OH-14 : O (3-hydroxymyristic acid) only as a minor component
(Wollenweber et al., 1980). In a gas chromatographic analysis, three major groups of cellular
fatty acids were found on examination of 31 strains of B. oralis and B. ruminicola and others
(Miyagawa & Suto, 1980), indicating the potential for chemotaxonomy in Bacteroides strains.
Hase et al. (1977) and Hofstad & Skaug (1980) showed that 3-OH-14 : O and 3-OH-16 : O were
present as characteristic fatty acids in the LPS of F. nucleaturn. The latter authors found that
LPS preparations from F. necrophorum and F. mortifenrm did not contain 3-OH-16 :0 and that
this was a useful taxonomic difference in identification of F. nucleaturn. Jantzen & Hofstad
(1981) examined fatty acids of 31 Fusobacterium strains which confirmed the species differences.
In the present study differences in characteristic peaks were found between the three strains
(species) of Fusobacterium, however, additional investigations are needed before pyrolysis of
LPS can be considered for subgrouping of Fusobacteriurn strains. It was, however, easy to
distinguish between Fusobacterium and other genera. The Veillonella group contains only two
species, but subgrouping has been suggested (Bergey's Manual of Determinative Bacteriology,
Buchanan & Gibbons, 1974). Hofstad (1978) has shown four different chemotypes based on the
serological patterns of their LPS. No studies of the fatty acids of LPS or the cell walls have been
performed on Veillonella. The present study shows that Veillonella can be easily distinguished
from Bacteroides and Fusobacterium by pyrolysis of phenol-water extracted LPS. Furthermore, it
may be suggested that the strains used fell into two groups, indicated by the pattern of the
pyrograms. These groups corresponded to the two species V. alcalescens and V. paruula.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 04:34:28
562
G . DAHLEN AND I . ERICSSON
At present, gas chromatography of metabolic end-products is widely used as a taxonomical
aid for anaerobic bacteria (Holdeman & Moore, 1975; Sutter et al., 1975), to the genus level.
Analysis of cell envelope components is being included in taxonomical studies to an increasing
extent, indicated by the reports of the GLC of fatty acids (Fritsche, 1974;Miyagawa et al., 1979;
Miyagawa & Suto, 1980; Jantzen & Hofstad, 1981). This method includes methanolysis and
extraction of methyl esters from the micro-organisms. Pyrolysis of phenol-water extracted LPS
is no more laborious although limited in application to Gram-negative bacteria. In conclusion
this study gives evidence that pyrolysis of LPS gives useful information less complex than that
obtained by pyrolysis of whole cells. Pyrolysis of LPS can be used to complement other methods
used in differentiating for these and other Gram-negative bacteria.
REFERENCES
BUCHANAN,
R. E. & GIBBONS,N. E. (1974). Bergey’s HOFSTAD,T. (1978). Chemotypes of Veillonellu
Manual of Determinative Bacteriology, 8th edn.
lipopolysaccharides. Acta pathologica microbiologica
scandinauica BS6, 41-50.
Baltimore: Williams and Wilkins.
HOFSTAD,
T. & KRISTOPFERSEN,
T. (1970). Chemical
DAHLEN,G. & BERGENHOLTZ,
G. (1980). Endotoxin
characteristics of endotoxin from Bacteroides fragilis
activity in teeth with necrotic pulps. Journal of
NCTC 9343. Journal of General Microbiology 61,15Dental Research 59, 1033-1040.
19.
DRUCKER,
D.B. (1981). Microbiological Applications of
HOFSTAD,
T. & SKAUG,N. (1980). Fatty acid and
Gas Chromatography, pp. 364-399. Cambridge :
neutral sugars present in lipopolysaccharides isoCambridge University Press.
lated from Fusobacterium species. Acta pathologica
EMSWILER,
B. S. & KOTULA,
A. W. (1978). Differentimicrobiologica scandinavica Bs8, 115-1 20.
ation of Salmonella serotypes by pyrolysis gas liquid
HOLDEMAN,
L. V. & MOORE,W. E. C. (1975). Anaerobe
chromatography of cell fragments. Applied and
Laboratory Manual, 3rd edn, V.P.I. Anaerobe
Ent.ironmenta1 Microbiology 35, 97-104.
ERICSSON,
I. (1975). Qualitative, and kinetic studies of
Laboratory, Blacksburg, Virginia, U.S.A. : Virginia
Polytechnic Institute and State University.
salts and polymers with a new pyrolyser in combination
with a gas chromatograph. Thesis, University of
JANTZEN,
E. & HOFSTAD,
T. (198 1). Fatty acids of FusoLund, Sweden.
bacterium species:taxonomy implications. Journal of
ERICSSON,
I. (1977). Pyrolysis gas chromatography - a
General Microbiology 123, 163-1 7 1.
potential technique for characterization of complex
KAUFFMANN,
F. (1961). Die Bakterwlogie der Salmoorganic material . A eta pa thologica microbiologica
nella Species. Copenhagen : Munksgaard.
scandinauica B.259,3742.
LINDER,L., HOLME,T. & FROSTELL,G. (1974).
ERICSSON,
I., LARSSON,L. & MARDH,P.-A. (1977).
Hyaluronidase and aminopeptidase activity in culPyrolysis gas chromatography of microorganisms.
tures of Streptococcus mitis ATCC 903. Acta patholoInfluence of various pyrolysis parameters on the
gica microbwlogica scandinavica B82,52 1-526.
LUDERITZ,
O., STAUB,
A. M. & WESTPHAL,
0. (1966).
yield of volatile organic products. Acta pathologica
microbiologica scandinavica B259, 43-47.
Immunochemistry of 0 and R antigens of Salmonella
FABRICIUS,
L., DAHLBN,G., HOLM,S. E. & MOLLER,
and related Enterobacteriaceae. Bacteriological ReA. J. R. (1982). Influence of combinations of oral
views 30, 192-244.
bacteria on periapical tissues of monkeys. ScandinaMIYAGAWA,
E. & Smo, T. (1980). Cellular fatty acid
composition in Bacteroides oralis and Bacteroides
uian Journal of Dental Research 90,200-206.
ruminicola. Journal of General and Applied MicroFINEGOLD,
S. M. (1977). Anaerobic Bacteria in Human
biology 26, 331-343.
Disease. New York: Academic Press.
FREDRIKSEN,
G. & HOFSTAD,
T. (1978). Chemotypes of
MIYAGAWA,
E., AZUMA,
R. & Smo, T. (1979). Cellular
Fusobacterium nucleatum lipopolysaccharides. Acta
fatty acid composition in gram-negative obligately
pathologica microbiologica scandinavica B86, 41-45.
anaerobic rods. Journal of Generaland Applied Microbiology 25, 41-51.
FRITSCHE,
C. D. (1974). Zur Struktur der Lipoide des
Genus Bacteroides. Zentralblatt fur Bakteriologie,
NEEDLEMAN,
M. & STUCHBERY,P. (1977). The
Parasitenkunde, Infektionskrankheiten und Hygiene
identification of micro-organisms by pyrolysis gas(Abteilung I, Originale A ) 228, 100-106.
liquid chromatography. In Analytical Pyrolysis, pp.
GUTTERIDGE,
C. S. & NORRIS, J. R. (1979). The
77-78. Edited by C. E. R. Jones & C. A. Cramers.
application of pyrolysis techniques to the identificaAmsterdam : Elsevier.
tion of micro-organisms. Journal of Applied BacOXBORROW,
G. S., FIELDS,N. D. & PULEO,J. R.
teriology 47, 5-43.
(1976). Preparation of pure microbiological samples
HADDADIN,
J. M., STIRLAND,
R. M., PRESTON,
N. W. &
for pyrolysis gas-liquid chromatography studies.
COLLARD,
P. (1973). Identification of Vibrio cholerae
Applied and Environmental Microbiology 32,302-309.
by pyrolysis gas liquid chromatography. Applied
QUINN,P. A. (1974). Development of high resolution
Microbiology 25, 40-43.
pyrolysis gas chromatography for identification of
HUE, I., HOFSTAD,T. & RIETSCHEL,E. T. (1977).
microorganisms. Journal of Chromatographic Science
Chemical structure of the lipid A component of lipo12,796-806.
polysaccharides from Fusobacter ium nucleatum.
ROSEBURY,
T. (1962). Microorganisms Indigenous to
Journal of Bacteriology 129,9-14.
Man. New York. McGraw-Hill.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 04:34:28
Pyrolysis of lippolysaccharides
STACK,M. V., DONOGHUE,
H. D. & TYLER,J. E.
(1978). Discrimination between oral streptococci by
pyrolysis gas-liquid chromatography. Applied and
Environmental Microbiology 35, 45-50.
SUTTER,
V. L., VARGO,V. L. & FINEGOLD,
S. M. (1975).
Wadsworth Anaerobic Bacteriology Manual, 2nd edn.
Los Angeles: The University of California.
TYDEN-ERICSSON,I. (1973). A new pyrolyser with
improved control of pyrolysis conditions. Chromatographia 6, 353-358.
563
WESTPHAL,O., LUDERITZ,
0.&BISTER,
F. (1952). Uber
die Extraktion von Bakterien mit Phenol/Wasser.
Zeitschrft fur Naturforschung 7b, 148-1 55.
WOLLENWEBER,
H. W., RBTTSCHEL,
E. T., HOFSTAD,
T., WEINTRAUB,A. & LINDBERG,A. A. (1980).
Nature, type of linkage, quantity and absolute
configuration of (3-hydroxy) fatty acids in lipopolysaccharides from Bacteroides fiagilis NCTC 9343
and related strains. Journal of Bacteriology 144,898903.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 04:34:28