Genotyping of Francisella tularensis strains by pulsed-field gel electrophoresis, amplified fragment length polymorphism fingerprinting, and 16S rRNA gene sequencing.

Abstract

We evaluated three molecular methods for identification of Francisella strains: pulsed-field gel electrophoresis (PFGE), amplified fragment length polymorphism (AFLP) analysis, and 16S rRNA gene sequencing. The analysis was performed with 54 Francisella tularensis subsp. holarctica, 5 F. tularensis subsp. tularensis, 2 F. tularensis subsp. novicida, and 1 F. philomiragia strains. On the basis of the combination of results obtained by PFGE with the restriction enzymes XhoI and BamHI, PFGE revealed seven pulsotypes, which allowed us to discriminate the strains to the subspecies level and which even allowed us to discriminate among some isolates of F. tularensis subsp. holarctica. The AFLP analysis technique produced some degree of discrimination among F. tularensis subsp. holarctica strains (one primary cluster with three major subclusters and minor variations within subclusters) when EcoRI-C and MseI-A, EcoRI-T and MseI-T, EcoRI-A and MseI-C, and EcoRI-0 and MseI-CA were used as primers. The degree of similarity among the strains was about 94%. The percent similarities of the AFLP profiles of this subspecies compared to those of F. tularensis subsp. tularensis, F. tularensis subsp. novicida, and F. philomiragia were less than 90%, about 72%, and less than 24%, respectively, thus permitting easy differentiation of this subspecies. 16S rRNA gene sequencing revealed 100% similarity for all F. tularensis subsp. holarctica isolates compared in this study. These results suggest that although limited genetic heterogeneity among F. tularensis subsp. holarctica isolates was observed, PFGE and AFLP analysis appear to be promising tools for the diagnosis of infections caused by different subspecies of F. tularensis and suitable techniques for the differentiation of individual strains.

Full text

Page 1

JOURNAL OF CLINICAL MICROBIOLOGY, Aug. 2002, p. 2964–2972
0095-1137/02/$04.00?0DOI: 10.1128/JCM.40.8.2964–2972.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Vol. 40, No. 8
Genotyping of Francisella tularensis Strains by Pulsed-Field Gel
Electrophoresis, Amplified Fragment Length Polymorphism
Fingerprinting, and 16S rRNA Gene Sequencing
N. García Del Blanco,1M. E. Dobson,2A. I. Vela,3V. A. De La Puente,1C. B. Gutie ´rrez,1
T. L. Hadfield,2P. Kuhnert,4J. Frey,4L. Domínguez,3
and E. F. Rodríguez Ferri1*
Section of Microbiology and Immunology, Department of Animal Health, Faculty of Veterinary Medicine, Leo ´n,1and
Section of Microbiology and Immunology, Department of Animal Health, Faculty of Veterinary Medicine,
Madrid,3Spain; Infectious and Parasitic Diseases Department, Armed Forces Institute of Pathology,
Washington, D.C. 203062; and Institute for Veterinary Bacteriology, Bern, Switzerland4
Received 12 December 2001/Returned for modification 11 February 2002/Accepted 18 May 2002
We evaluated three molecular methods for identification of Francisella strains: pulsed-field gel electrophore-
sis (PFGE), amplified fragment length polymorphism (AFLP) analysis, and 16S rRNA gene sequencing. The
analysis was performed with 54 Francisella tularensis subsp. holarctica, 5 F. tularensis subsp. tularensis, 2 F. tu-
larensis subsp. novicida, and 1 F. philomiragia strains. On the basis of the combination of results obtained by
PFGE with the restriction enzymes XhoI and BamHI, PFGE revealed seven pulsotypes, which allowed us to
discriminate the strains to the subspecies level and which even allowed us to discriminate among some isolates
of F. tularensis subsp. holarctica. The AFLP analysis technique produced some degree of discrimination among
F. tularensis subsp. holarctica strains (one primary cluster with three major subclusters and minor variations
within subclusters) when EcoRI-C and MseI-A, EcoRI-T and MseI-T, EcoRI-A and MseI-C, and EcoRI-0 and
MseI-CA were used as primers. The degree of similarity among the strains was about 94%. The percent sim-
ilarities of the AFLP profiles of this subspecies compared to those of F. tularensis subsp. tularensis, F. tularensis
subsp. novicida, and F. philomiragia were less than 90%, about 72%, and less than 24%, respectively, thus
permitting easy differentiation of this subspecies. 16S rRNA gene sequencing revealed 100% similarity for all
F. tularensis subsp. holarctica isolates compared in this study. These results suggest that although limited
genetic heterogeneity among F. tularensis subsp. holarctica isolates was observed, PFGE and AFLP analysis
appear to be promising tools for the diagnosis of infections caused by different subspecies of F. tularensis and
suitable techniques for the differentiation of individual strains.
The family Francisellaceae, which belongs to the ? subclass
of the class Proteobacteria, includes closely related organisms
within the single genus Francisella. There are two recognized
species, Francisella tularensis and F. philomiragia (25). Of
these, F. philomiragia is relatively rare, is often associated with
water, and is less virulent than F. tularensis (14). The more
common species, F. tularensis, has four subspecies, all of which
have been related to the acute infectious disease known as
tularemia, although they differ in their virulence for humans
and rabbits (16).
F. tularensis subsp. tularensis (also called F. tularensis subsp.
nearctica or biovar type A), the causal agent of tularemia in
North America, has recently been observed for the first time in
Europe (Slovakia) (13). This subspecies is a highly virulent
pathogen for many mammalian species, including humans, and
is mainly associated with tick-borne tularemia in lagomorphs
(16). F. tularensis subsp. palaearctica (also called F. tularensis
subsp. holarctica or biovar type B) is more widely distributed in
nature and is found in Europe, Asia, and to a lesser degree in
North America. This subspecies is linked to waterborne dis-
ease of rodents and hares, it is considered less pathogenic for
mammals than F. tularensis subsp. tularensis, and it usually
causes the ulceroglandular form of tularemia in humans (28).
F. tularensis subsp. mediaasiatica has been found only in the
former Soviet republics of Central Asia (23). The fourth sub-
species, F. tularensis subsp. novicida, has been recovered from
water samples and has rarely been associated with human
illness (14).
The subspecies of F. tularensis have been differentiated by
biochemical analysis, and the two main subspecies have been
differentiated by 16S rRNA hybridization (8, 23). However,
such tests have obvious drawbacks for unambiguous strain
classification. In addition, Francisella strains have similar anti-
genic compositions, and it is not possible to distinguish the
different subspecies from each other by serological methods
(21). Therefore, the development of new molecular typing
methods is of special interest.
F. tularensis has been described as a genetically homoge-
neous species, despite differences in virulence and geographic
origins (18), with a limited diversity among isolates. In this
respect, some molecular techniques have been able to discrim-
inate the species within the genus but not the subspecies within
the species (9, 15, 17); however, other recent taxonomic work
based on either repetitive element sequence-based PCR (rep-
PCR), random amplified polymorphic DNA (RAPD) analysis
* Corresponding author. Mailing address: Facultad de Veterinaria,
Departamento de Sanidad Animal, Universidad de Leo ´n, 24007-Leo ´n,
Spain. Phone: 34 987 291 297. Fax: 34 987 291 304. E-mail: dsaerf
@unileon.es.
2964

Page 2

(4, 16), or allele size determinations (6, 18) has successfully
demonstrated variability in subspecies and individual strains of
F. tularensis. In addition, a capture enzyme-linked immunosor-
bent assay based on lipopolysaccharide-specific monoclonal
antibodies and a handheld immunochromatographic assay
have been applied to the detection of F. tularensis subsp. tula-
rensis and F. tularensis subsp. holarctica (12).
Some recently developed molecular methods with strong
discriminatory power, such as macrorestriction pulsed-field gel
electrophoresis (PFGE) and amplified fragment length poly-
morphism (AFLP) analysis, have been used to differentiate
species and clinical isolates of a variety of gram-positive and
gram-negative organisms other than Francisella (7, 11, 30).
PFGE provides a highly reproducible restriction profile that
typically shows distinct, well-resolved fragments representing
the entire bacterial chromosome in a single gel (20). AFLP
analysis is a genome fingerprinting technique based on the
selective amplification of a subset of DNA fragments gener-
ated by restriction enzyme digestion. It was originally applied
to the characterization of plant genomes and more recently has
been applied to the typing of bacteria (29).
This report describes the genetic characterization by PFGE
and AFLP analysis of Spanish F. tularensis isolates recovered
from hares, humans, voles, and ticks and compares the genetic
profiles with those of other F. tularensis strains recovered from
patients and animals with tularemia in the United States and
Europe, as well as with those of F. philomiragia isolates. These
two molecular techniques are also compared with a more tra-
ditional typing method, 16S rRNA gene sequencing. Although
the latter method usually shows limited variability between
strains of a bacterial species (22), any mutation in the genes of
this highly conserved region could be of interest for epidemi-
ological purposes.
MATERIALS AND METHODS
Bacterial strains and culture conditions. Sixty-two Francisella tularensis strains
were studied (Table 1). Forty-two F. tularensis subsp. holarctica isolates were
recovered from a tularemia epidemic (which involved ulceroglandular, typhoidal,
glandular, pneumonic, oculoglandular, and atypical clinical forms diagnosed in
humans) that occurred in Castilla y Leo ´n, which is in northwestern Spain, be-
tween 1997 and 1999 (4, 10). Of these isolates, 31 were from hares, 8 were from
humans, 1 was from a vole, and 2 were from ticks. The geographical origins of
these isolates are shown in Table 1. In addition, three Czech, two French, two
Russian, and five U.S. isolates were included in the study. Five F. tularensis subsp.
tularensis, two F. tularensis subsp. novicida, and one F. philomiragia strains were
also included. All isolates, which had been stored frozen at ?80°C, were grown
aerobically at 37°C for 2 days on modified Thayer-Martin agar plates containing
GC medium base (36 g/liter), hemoglobin (10 g/liter), and Vitox supplement
(2 vials/liter; Oxoid Ltd., Basingstoke, England).
16S rRNA gene sequencing. Sequencing of the 16S rRNA genes was per-
formed with universal primers directed at a 1.4-kb conserved region of the gene
(19). The PCR amplification products were purified in CentriSep (Princeton
Separations, Inc., Adelphia, N.J.) columns and were sequenced with the Taq Dye
Deoxy Terminator Cycle Sequencing kit on a model 310 genetic analyzer (Ap-
plied Biosystems, Foster City, Calif.) according to the instructions of the man-
ufacturer. Sequence comparisons were done with the BLAST (version 2.0) sys-
tem (1).
Analysis of chromosomal DNA restriction pattern by PFGE. The Francisella
strains were grown as described above. A loop filled with cells from each strain
was suspended in SE buffer (25 mM EDTA, 75 mM NaCl [pH 7.5]). The cells
were harvested by centrifugation (3,500 ? g, 10 min, 4°C) and washed twice with
the same buffer. Agarose plugs were made from a 1:1 mixture of 2% low-melting-
point agarose and the cell suspension. The plugs were lysed in buffer (50 mM
Tris-HCl, 50 mM EDTA, 1% [vol/vol] lauroyl sarcosine, 500 ?g of lysozyme per
ml [pH 9.5]) for 24 h at 37°C. The cells were then treated for 24 h at 56°C with
the same volume of a solution (50 mM Tris-HCl, 50 mM EDTA, 1% [vol/vol]
lauroyl sarcosine [pH 9.5]) containing proteinase K at 500 ?g/ml and were
washed three times with Tris-EDTA buffer for 1 h at 4°C. XhoI, BamHI, SpeI,
XbaI, SmaI, BglII, and SacII (all from MBI Fermentas) and ApaI (Promega
Corp.) were used for restriction endonuclease digestion in accordance with the
instructions of the manufacturers. The fragments were resolved by PFGE in elec-
trophoresis-grade agarose (1%; Boehringer Mannheim) by using a CHEF-DR
III system (Bio-Rad). The following parameters were used: running time, 24 h;
temperature, 14°C; voltage gradient, 200 V; included angle, 120°; an initial pulse
time of 0.1 s and a final pulse time of 10 s for XhoI, SpeI, XbaI, BglII, and SacII;
an initial pulse time of 0.1 s and a final pulse time of 15 s for BamHI; and an
initial pulse time of 0.1 s and a final pulse time of 25 s for SmaI and ApaI. The
gels were stained with ethidium bromide (0.5 ?g/ml) for 15 min, destained in
distilled water, and photographed under UV light. A bacteriophage lambda
ladder PFGE marker (concatemers of bacteriophage lambda cI857 Sam7;
Boehringer Mannheim) was used for molecular size determination. Any non-
identity in terms of the presence, absence, or apparent mobility of a band was
considered one difference. Variations in band intensity were not counted as a
difference.
AFLP analysis. AFLP analysis of DNA was performed in duplicate with the
AFLP Microbial Fingerprinting kit (Applied Biosystems) according to the rec-
ommendations of the manufacturer. Briefly, 10 ng of purified DNA was simul-
taneously restricted with EcoRI and MseI endonucleases (New England Biolabs,
Beverly, Mass.) and ligated to the adapter-linker sequences with T4 DNA ligase
(New England Biolabs) for 2 h at 37°C in a final volume of 11 ?l. After dilution,
4 ?l was amplified with the EcoRI and MseI core primer sequences in a final
volume of 20 ?l. Selective amplification was performed with primers EcoRI-C
and MseI-A, EcoRI-T and MseI-T, EcoRI-A and MseI-C, and EcoRI-0 and
MseI-CA by using 1.5 ?l of the diluted preselective amplification reaction mix-
ture in a final volume of 10 ?l. All reactions were performed in an Applied
Biosystems model 9600 thermal cycler with 200-?l reaction tubes.
Separation and detection of the AFLP fragments were performed with an
Applied Biosystems model 310 genetic analyzer equipped with a 47-cm capillary
and by using polymer POP-4. Samples were injected for 8 s at 15 kV and
separated for 28 min at 15 kV and 60°C by using denaturing conditions. Each
sample contained GeneScan 500[ROX] (Applied Biosystems) as an internal size
standard. Determination of the sizes of the fragments was performed automat-
ically with GeneScan software (version 3.1; Applied Biosystems). Additional
analysis and construction of dendrograms and trees were done with GelCompar
II software (version 2.0; Applied Maths, Sint-Martens-Latem, Belgium). The
dendrograms were constructed by binary band matching by using the Dice cor-
relation and the unweighted pairwise group method with mathematical averaging
(UPGMA). The unrooted maximum parsimony tree was constructed by using
500 bootstrap simulations.
RESULTS
16S rRNA gene sequencing. Only the 42 Spanish F. tularensis
subsp. holarctica isolates were tested by the 16S rRNA gene
sequencing method. All strains tested were found to have the
same 16S rRNA gene sequence, and that sequence was com-
pared with the 16S rRNA sequences previously reported to
GenBank. The sequence detected in this study shared 100%
similarity to that of F. tularensis subsp. holarctica strain LVS
(Table 2) and 99% similarity to those of strains from another
Spanish outbreak of tularemia (2). In the previous study (2),
the amplicons were directly amplified from clinical or environ-
mental samples but not from F. tularensis subsp. holarctica
isolates. The similarities observed between our sequences and
those from the other F. tularensis subspecies listed in Table 2
were also 99%; the sequences of the strains shared lower
percent similarities with those of F. philomiragia strains.
PFGE analysis. Only 49 Francisella strains were studied by
PFGE (Table 1). Of the eight restriction enzymes tested, SmaI
and ApaI provided profiles with only four bands, and these
profiles could be used because of the small number of bands.
On the other hand, SpeI, XbaI, BglII, and SacII gave patterns
which were extremely difficult to interpret because of the large
VOL. 40, 2002TYPING OF F. TULARENSIS2965

Page 3

TABLE 1. Characteristics of the strains of the genus Francisella used in this study
Strain
no.
Sourcea
Species or subspecies
Geographic origin,
yr of isolation
Host
PFGE results
AFLP
profile
XhoI BamHIPulsotype
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
AFIP1
CAPM1
AFIP2
ALG1
ALG2
ALG3
ALG4
ALG5
ALG6
ALG7
ALG8
ALG9
ALG10
ALG11
ALG12
ALG13
ALG14
ALG15
ALG16
ALG17
ALG18
ALG19
ALG20
ALG21
ALG22
CAPM2
F. tularensis subsp. tularensis
F. tularensis subsp. tularensis
F. tularensis subsp. tularensis
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
Utah, 1920
Strain Schu (CAPM 5600), Ohio, 1941
Strain A88R160, United States
Valladolid, Spain, 1997
Valladolid, Spain, 1998
Valladolid, Spain, 1998
Leo ´n, Spain, 1998
Palencia, Spain, 1998
Leo ´n, Spain, 1998
Leo ´n, Spain, 1998
Palencia, Spain, 1998
Palencia, Spain, 1998
Valladolid, Spain, 1998
Valladolid, Spain, 1998
Zamora, Spain, 1998
Zamora, Spain, 1998
Palencia, Spain, 1998
Valladolid, Spain, 1998
Valladolid, Spain, 1998
Segovia, Spain, 1998
Palencia, Spain, 1998
Palencia, Spain, 1998
Palencia, Spain, 1998
Palencia, Spain, 1998
Valladolid, Spain, 1998
Strain 130 (CAPM 5536), Czech
Republic
Strain 2713 (CAPM 5537), Czech
Republic
Strain T-1/59 (CAPM 5151), Czech
Republic
Palencia, Spain, 1998
Valladolid, Spain, 1998
Soria, Spain, 1998
Zamora, Spain, 1999
Zamora, Spain, 1998
A ´vila, Spain, 1998
Valladolid, Spain, 1998
Leo ´n, Spain, 1998
Palencia, Spain, 1998
Valladolid, Spain, 1998
Valladolid, Spain, 1998
Valladolid, Spain, 1998
Palencia, Spain, 1998
Leo ´n, Spain, 1998
Zamora, Spain, 1998
Zamora, Spain, 1998
Zamora, Spain, 1998
Chateneaux, France
St. Germaine, France
Zamora, Spain, 1998
Zamora, Spain, 1998
Valladolid, Spain, 1998
Strain 503, Russia
Strain Grousse, unknown
Reference strain ATCC 29684, Russia
Reference strain ATCC 15482, Utah,
1950
Texas, 1991
Reference strain ATCC 25015, Utah,
1959
Strain 99A-2628, California
Strain 99A-9419, California
Strain 89A-7092, California
Human
Human
Unknown
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
NDb
A
ND
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
ND
ND
ND
C
A
A
A
B
A
A
C
C
A
A
A
A
A
A
A
A
A
A
B
A
A
A
ND
I
ND
IV
III
III
III
II
III
III
IV
IV
III
III
III
III
III
III
III
III
III
III
II
III
III
III
A1
A1
A1
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B2
27CAPM3 F. tularensis subsp. holarcticaHareBA IIIB2
28 CAPM4F. tularensis subsp. holarcticaHareBAIIIB2
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
LEO1
LEO2
LEO3
LEO4
LEO5
LEO6
LEO7
LEO8
LEO9
ALG23
ALG24
ALG25
ALG26
HLE1
HZA1
HZA2
HZA3
AFIP3
AFIP4
ALG27
LEO10
LEO11
CAPM5
AFIP5
AFIP6
CAPM6
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. novicida
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Hare
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Vole
Tick
Tick
Tick
Unknown
Vole
Water
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
ND
ND
B
B
B
C
ND
ND
D
B
A
A
A
A
A
A
A
B
A
A
A
A
A
A
A
A
ND
ND
A
A
A
D
ND
ND
ND
II
III
III
III
III
III
III
III
II
III
III
III
III
III
III
III
III
ND
ND
III
III
III
V
ND
ND
VI
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B2
B3
B2
C
55
56
AFIP7
FOA1
F. tularensis subsp. novicida
F. philomiragia
Human
Muskrat
ND
E
ND
ND
ND
VII
C
D
57
58
59
CDHS1
CDHS2
CDHS3
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
Human
Monkey (capuchin)
Monkey (squirrel
monkey)
Human
Human
Human
ND
ND
ND
ND
ND
ND
ND
ND
ND
B1
B1
B1
60
61
62
CDHS4
CDHS5
AFIP8
F. tularensis subsp. holarctica
F. tularensis subsp. tularensis
F. tularensis subsp. tularensis
Strain 96A-10640, California
Strain 00A-416, Arkansas
Reference strain ATCC 6223,
United States
ND
ND
ND
ND
ND
ND
ND
ND
ND
B1
A1
A2
aAFIP, Armed Forces Institute of Pathology, Washington, D.C; CAPM, Collection of Animal Pathogenic Microorganisms, Brno, Czech Republic; ALG, Central
Laboratory of Animal Health, Algete, Madrid, Spain; LEO, Central Laboratory of Animal Health, Leo ´n, Spain; HLE, Department of Medical Microbiology, Hospital
Princesa Sofı ´a, Insalud, Leo ´n, Spain; HZA, Laboratory of Microbiology, Hospital Virgen de la Concha, Insalud, Zamora, Spain; FOA, National Defence Research
Establishment, Umeå, Sweden; CDHS, California Department of Health Services, Sacramento.
bND, not done.
2966GARCI´A DEL BLANCO ET AL.J. CLIN. MICROBIOL.

Page 4

number of bands (more than 30 bands, with all bands having
molecular sizes below 40 kb). XhoI was used as the primary
enzyme, and five different DNA fragment profiles were gener-
ated, with a mean of 20 bands per isolate, all of which were less
than 100 kb (Table 1 and Fig. 1A). The 42 Spanish F. tularensis
subsp. holarctica clinical isolates exhibited the same XhoI pat-
tern (designated pattern B), irrespective of the host from
which they were recovered; this profile was also shared by the
three Czech F. tularensis subsp. holarctica isolates. The human
F. tularensis subsp. tularensis isolate, the F. tularensis subsp.
holarctica strain isolated from a tick in Russia, F. tularensis
subsp. novicida ATCC 15482, and F. philomiragia were as-
signed profiles A, C, D, and E, respectively.
BamHI also provided good discrimination and was used with
F. tularensis subsp. holarctica isolates to help interpret minor
differences in the PFGE patterns obtained with XhoI. There-
fore, it was applied only to the Spanish and Czech isolates and
Russian strain 51, yielding four distinguishable patterns, with
12 to 16 bands with sizes ranging from about 27 to 291 kb.
These profiles differed from each other by three or fewer re-
striction fragments. Thirty-five of the Spanish isolates (83.3%)
as well as the Czech hare isolates were designated type A, four
Spanish hare isolates recovered in 1998 were designated type B
(9.5%), three other Spanish hare isolates recovered between
1997 and 1998 (7.1%) were designated type C, and finally, the
Russian tick isolate was designated type D (Table 1 and Fig.
1B).
When the results obtained with XhoI and BamHI were an-
alyzed together, the 49 isolates studied could be divided into
seven pulsotype groups (Table 1). Most strains (77.6%) be-
longed to pulsotype III, which included the majority of the
Spanish F. tularensis subsp. holarctica isolates and Czech iso-
lates. The F. tularensis subsp. tularensis isolate tested by PFGE
was assigned to pulsotype I, a limited number of Spanish iso-
lates were referred to as pulsotypes II and IV, the Russian
F. tularensis subsp. holarctica strain isolated from a tick was
assigned to pulsotype V, and the F. tularensis subsp. novicida
and F. philomiragia strains were grouped into pulsotypes VI
and VII, respectively.
AFLP analysis. All 62 strains were analyzed by AFLP anal-
ysis with four primer pairs: EcoRI-T and MseI-T, EcoRI-0 and
MseI-CA, EcoRI-C and MseI-A, and EcoRI-A and MseI-C.
These primers produced a total of 192 DNA fragments, of
which 92%, or 177, were polymorphic for one or more strains.
The patterns seen for F. philomiragia were quite different from
those seen for the F. tularensis strains and contributed greatly
to the percentage of polymorphic bands (data not shown). If
the contributions from F. philomiragia are removed, there re-
FIG. 1. PFGE fingerprint patterns of Francisella strains after diges-
tion with XhoI (A) and BamHI (B). (A) Lanes: 1, F. tularensis subsp.
holarctica isolate 26 (pattern B); 2, F. tularensis subsp. tularensis strain
Schu (isolate 1 in this study; pattern A); 3, F. tularensis subsp. holarc-
tica strain 503 (isolate 51 in this study; pattern C); 4, F. tularensis subsp.
novicida reference strain ATCC 15482 (isolate 54 in this study; pattern
D); 5, F. philomiragia reference strain ATCC 25015 (isolate 56 in this
study; pattern E); 6, bacteriophage lambda ladder PFGE marker.
(B) Lanes: 1 and 6, bacteriophage lambda ladder PFGE marker; 2,
F. tularensis subsp. holarctica isolate 8 (type A); 3, F. tularensis subsp.
holarctica isolate 11 (type B); 4, F. tularensis subsp. holarctica strain
CAPM 5536 (isolate 26 in this study; type C); 5, F. tularensis subsp.
holarctica strain 503 (isolate 51 in this study; type D).
TABLE 2. 16S rRNA gene sequence similarity among the 42 Spanish F. tularensis subsp. holarctica isolates tested in
this study and other Francisella strains
Species or subspeciesOrigin
GenBank
accession no.
% Similarity
with the isolates
tested in this
study
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. holarctica
F. tularensis subsp. tularensis
F. tularensis subsp. tularensis
F. tularensis subsp. novicida
F. philomiragia
F. philomiragia
Strain LVS
Spanish isolate from freshwater stream
Spanish isolate from river crayfish
Spanish isolate from human lymph node
ATCC 6223
Strain SCHU
ATCC 15482
ATCC 25015
ATCC 25017
L26086
AF27312
AF27313
AF27314
Z21931
Z21932
L26084
L26085
Z21933
100
99
99
99
99
99
99
98
97
VOL. 40, 2002 TYPING OF F. TULARENSIS 2967

Page 5

mained 68 polymorphic bands among the total of 125 bands, or
54%. Of these 68 polymorphic bands, a little more than one-
half (n ? 35) were due to F. tularensis subsp. novicida.
Figure 2 shows the dendrogram obtained after clustering by
UPGMA when the data obtained with all four primer pairs
were analyzed as a composite. A total of four primary clusters
were observed among the 62 Francisella strains tested. The five
F. tularensis subsp. tularensis strains were assigned to primary
cluster A. Two subclusters were seen, one for isolate 62 (sub-
cluster A2) and the other for the remaining F. tularensis subsp.
tularensis isolates (subcluster A1). A genetic similarity of 91%
was observed between these two subclusters. In addition, a
minor variation between isolate 1 and the other three isolates
in subcluster A1 was seen. Primary cluster B included all the
F. tularensis subsp. holarctica isolates. Three major subclusters
were detected within cluster B (with similarities among them of
about 94%). Subcluster B1 contained four isolates recovered
from monkeys and humans in the United States (greater than
96% similarity), subcluster B2 contained the Czech and Rus-
sian isolates (greater than 98% similarity), and subcluster B3
included the Spanish and French isolates, as well as an isolate
(isolate 52) of unknown origin (greater than 99% similarity).
The isolates in primary cluster B shared less than 90% simi-
larity with the isolates in primary cluster A. The two F. tula-
rensis subsp. novicida isolates, which were slightly different
from each other, were assigned to primary cluster C and had
about 72% similarity to the other F. tularensis subspecies. Fi-
nally, a genetically distant cluster (cluster D) was generated for
F. philomiragia (less than 24% similarity to the F. tularensis
strains). This species had just 15 AFLP bands in common with
the other strains (data not shown).
An unrooted maximum parsimony tree is presented in Fig. 3.
Three primary branches from a common origin can be seen.
One branch contains the F. tularensis subsp. tularensis (subclus-
ters A1 and A2) and F. tularensis subsp. holarctica (subclusters
B1, B2, and B3) clusters, the second branch contains the F. tu-
larensis subsp. novicida (cluster C) cluster, and the third branch
represents the unique species, F. philomiragia (cluster D). The
lengths of the lines between the nodes represent the relative
degrees of dissimilarity.
Comparative analysis of the results obtained by PFGE and
AFLP analysis. A comparative analysis of the results obtained
by PFGE and AFLP analysis for the 49 Francisella strains
analyzed by both methods is presented in Table 1. The F.
tularensis subsp. tularensis isolate belonging to PFGE cluster I
had AFLP profile A1. The four strains yielding PFGE cluster
II had AFLP profile B3. The 35 strains belonging to PFGE
cluster III could also be grouped into AFLP profile B3, and 3
strains belonging to PFGE cluster III were assigned to AFLP
profile B2. The three strains belonging to PFGE cluster IV had
AFLP profile B3. Finally, the strains that were grouped into
PFGE clusters V, VI, and VII had AFLP profiles B2, C, and D,
respectively.
DISCUSSION
Although the genus Francisella comprises two species, al-
most all knowledge originates from studies with F. tularensis
and, more concretely, from studies with two of the subspecies
of F. tularensis (F. tularensis subsp. tularensis and F. tularensis
subsp. holarctica), which are extremely infectious bacteria that
cause the zoonotic disease tularemia. In this study, PFGE,
AFLP analysis, and 16S rRNA gene sequencing were used as
tools to characterize F. tularensis subsp. holarctica isolates of
different origins and to compare them with other subspecies
and species of the genus Francisella.
No nucleotide differences were detected among the ampli-
cons of the 42 Spanish isolates by 16S rRNA sequencing. This
single sequence matched base by base the 16S rRNA sequence
of F. tularensis subsp. holarctica strain LVS (GenBank acces-
sion no. L26086), and therefore, this method confirmed the
identities (at the species and subspecies levels) of the Fran-
cisella isolates recovered from patients involved in the tulare-
mia outbreak that occurred in northwestern Spain between
1997 and 1999, but it did not allow us to discriminate individual
isolates.
It is interesting that the sequence reported in this study was
not identical to those recently obtained in Spain from a second
outbreak of tularemia; that outbreak was associated with cray-
fish (Procambarus clarkii) fishing (2) and was unrelated to the
first hare-associated outbreak, the isolates from which were
tested in our investigation. The published sequences of the
isolates involved in the second Spanish outbreak were reported
to be 100% identical to that of F. tularensis subsp. holarctica
strain LVS (2). However, our analysis showed that all these
other sequences differed by one base from the sequence of the
LVS reference strain, and, consequently, they also differed at
the same base from the sequences of the 42 Spanish isolates
listed in Table 1. This slight difference was found at a position
equivalent to base 1066 of the published sequence of the LVS
reference strain, for which the sequences reported by Anda
et al. (2) lacked an A residue. This finding suggests that the
isolates recovered from humans, hares, ticks, and voles in-
volved in the first Spanish tularemia outbreak were different
from those recovered from humans, crayfish, and water in-
volved in the second outbreak. This means that two variants of
F. tularensis subsp. holarctica are present in Spain, and so far
they have been involved in completely unrelated outbreaks
from an epidemiological point of view. In addition, the lack of
variability in the 16S rRNA genes of Francisella species ob-
served in previous studies (9, 15, 23) makes this finding of
special interest.
The initial PFGE experiments were done with restriction
endonucleases routinely used for gram-negative organisms
(XhoI, SpeI, XbaI, SmaI, ApaI, BglII) (27). However, the re-
sults obtained were difficult to analyze because of either two
few or too many bands, with the exception of the patterns
produced by XhoI. Paradoxically, a high discriminatory power
(four types among 46 F. tularensis subsp. holarctica isolates)
was obtained after digestion with BamHI, an endonuclease
normally used with gram-positive organisms. This fact could be
explained by the low G?C contents of the DNAs of members
of the genus Francisella (33 to 36 mol%), which are more
similar to those of some gram-positive organisms (for instance,
gram-positive cocci) than to those of other closely related
gram-negative organisms.
The banding pattern generated for F. philomiragia isolates
by PFGE was completely different from those generated for
F. tularensis isolates, and similarly, the pulsotypes observed for
the three subspecies within the latter species were easily dis-
2968GARCI´A DEL BLANCO ET AL. J. CLIN. MICROBIOL.