Journal of Microbiological Methods 95 (2013) 245–250

Contents lists available at ScienceDirect

Journal of Microbiological Methods

journal homepage: www.elsevier.com/locate/jmicmeth

Genotyping of outbreak-associated and sporadic Yersinia

pseudotuberculosis strains by novel multilocus variable-number tandem
repeat analysis (MLVA)
Jani Halkilahti a,⁎, Kaisa Haukka a,b, Anja Siitonen a
a
Bacteriology Unit, Department of Infectious Disease Surveillance and Control, National Institute for Health and Welfare (THL), P.O. Box 30, FI-00271 Helsinki, Finland
b
Division of Microbiology, Department of Food and Environmental Sciences, University of Helsinki, P.O. Box 56, FI-00014 Helsinki, Finland

a r t i c l e i n f o a b s t r a c t

Article history: Yersinia pseudotuberculosis human infections caused by serotype O:1 and O:3 isolates have been common in
Received 10 July 2013 Finland and have also caused outbreaks. Epidemiological studies on the outbreaks have been limited by the
Received in revised form 9 September 2013 lack of accurate typing methods. During the recent years, multilocus variable-number tandem repeat analysis
Accepted 9 September 2013
(MLVA) has been successfully applied for molecular typing of several bacterial pathogens. We designed an
Available online 17 September 2013
MLVA scheme based on seven loci for Y. pseudotuberculosis. The method was able to discriminate clinical isolates
Keywords:
of serotypes O:1 and O:3 into several MLVA types. The MLVA profiles were based on the number of 6 to 9 bp long
Genotyping tandem repeats in each locus. The number of repeats varied from 1 to 23 depending on the locus. The loci were all
MLVA located in the bacterial chromosome for stability of the markers. The MLVA method developed was serotype-
VNTR specific and will be a new additional tool for the epidemiological investigations of isolates associated with disease
Yersinia pseudotuberculosis outbreaks and for comparison of sporadic isolates.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction the sources of sporadic infections are largely unknown (Hallanvuo

et al., 2003; Kangas et al., 2008; Nuorti et al., 2004; Rimhanen-Finne
The genus Yersinia contains three species pathogenic for humans: et al., 2009).
Yersinia pestis, Yersinia enterocolitica and Yersinia pseudotuberculosis Epidemiological studies have been limited by the lack of accurate
(Brubaker, 1991). Y. pseudotuberculosis consists of 15 O-serotypes molecular biology typing methods (Vincent et al., 2008; Wobeser
(O:1–O:15) and 10 subtypes (O:1a–O:1c, O:2a–O:2c, O:4a–O:4b, et al., 2009) to compare Y. pseudotuberculosis isolates from humans
O:5a–O:5b) according to their varying lipopolysaccharide O-antigen and from the various environmental sources (Fukushima et al., 2001;
structure (Skurnik et al., 2000). Different serotypes are common in dif- Niskanen et al., 2003). The method that has commonly been applied
ferent geographical areas. Serotypes O:1 and O:3 are most common in to study the genotypes of Y. pseudotuberculosis, is pulsed-field gel elec-
Europe whereas serotypes O:4 and O:5 are more prevailing in the Far trophoresis (PFGE) (Tenover et al., 1995). PFGE has traditionally been
East. Serotypes O:6–O:14 have not been isolated from human infections the golden standard for epidemiological studies of Y. pseudotuberculosis
(Fukushima et al., 2001). and several other bacteria. However, PFGE is time-consuming, labour-
Y. pseudotuberculosis infection in humans is associated with fever intensive and difficult to equalize between laboratories. Furthermore,
and abdominal pain due to mesenteric lymphadenitis, which are easily PFGE does not offer the desired resolution at subtyping level for epide-
interpreted as symptoms of acute appendicitis leading to unnecessary miological study purposes (Jalava et al., 2006; Niskanen et al., 2002;
surgery. Post-infectious complications may include erythema nodosum Nuorti et al., 2004). Therefore, a method that has high level of dis-
and reactive arthritis (Tertti et al., 1984, 1989). In Finland, the annual crimination, low cost and is suited for epidemiological investigation
incidence of yersiniosis caused by Y. pseudotuberculosis has varied is required.
from 28 to 252 according to the Statistical Database of the Infectious Bacterial genomes contain variable number tandem repeats (VNTR),
Diseases Register (www3.thl.fi/stat/). In addition, several outbreaks which typically are highly polymorphic regions and consequently
caused by strains of serotypes O:1 or O:3 have occurred (Jalava et al., suited for subtyping of strains. Multilocus variable-number tandem
2004; Rimhanen-Finne and Sihvonen, 2010; Vasala et al., 2013). In the repeat analysis (MLVA) is based on detecting size variations of several
outbreaks, various vegetables have been vehicles of the pathogen but VNTR loci. These VNTR regions are utilized in MLVA, which is most com-
monly done using capillary electrophoresis and fluorescent dye labelled
⁎ Corresponding author at: National Institute for Health and Welfare, Bacteriology Unit,
primers (Lindstedt, 2005). VNTR-based methods have already been
P.O. Box 30, 00271 Helsinki, Finland. Tel.: +358 29 52 48172. published for Y. pestis, Y. enterocolitica, Escherichia coli, Francisella
E-mail address: jani.halkilahti@thl.fi (J. Halkilahti). tularensis, Mycobacterium tuberculosis, Haemophilus influenzae, Bacillus

0167-7012/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.mimet.2013.09.007
246 J. Halkilahti et al. / Journal of Microbiological Methods 95 (2013) 245–250

anthracis and Salmonella enterica (Gierczyński et al., 2007; Johansson The in silico PCR amplification products were compared with each
et al., 2001; Keim et al., 2000; Klevytska et al., 2001; Lindstedt et al., other (data not shown). Ten VNTR marker candidates that had iden-
2003; Noller et al., 2003; Sihvonen et al., 2011; Skuce et al., 2002; van tical flanking regions per definite locus and differed only in the num-
Belkum et al., 1997) among others. Although some studies have utilized ber of tandem repeats were selected for further analysis. Primer3
the VNTR regions in characterization of Y. pseudotuberculosis strains software (Rozen and Skaletsky, 2000) was used to locate suitable
(Platonov et al., 2013), no detailed MLVA method for epidemiological binding sites for 10 new primer pairs on the complete genome se-
purposes has been described for the species this far. Especially the recent quences of Y. pseudotuberculosis IP32953 and Y. pseudotuberculosis
outbreaks in Finland have shown that with the current methods it is not PB1/+. The number of loci was reduced to seven during the study.
possible to confirm whether certain strains belong to an outbreak or not. It was ensured that the tandem repeats would not be located in a
Therefore, the aim of this study was to develop an MLVA method suit- plasmid. Theoretical annealing temperatures were calculated for
able for more comprehensive subtyping of Y. pseudotuberculosis strains the primers (SantaLucia, 1998). The PCR thermal cycling conditions
of serotypes O:1 and O:3. were optimized by increasing the annealing temperature of each
primer pair higher, than the calculated values to obtain PCR products
2. Materials and methods with specific, single bands detectable in gel electrophoresis. For cap-
illary electrophoresis, the forward primers were labelled with a fluo-
2.1. Bacterial strains rescent 6-FAM dye (Table 1).

A set of 104 human isolates of Y. pseudotuberculosis serotypes O:1 2.3. DNA purification
(n = 61) and O:3 (n = 43) were selected from the 551 Y. pseudotuber-
culosis strains (394 O:1 and 127 O:3 strains) isolated in Finland in The 107 isolates were grown on Drigalski agar plates at 30 °C for
1997–2012. All strains were stored at −70 °C in the culture collection 24 h. A loopful of biomass was inoculated into Penassay broth and incu-
of the Bacteriology Unit of the National Institute for Health and Welfare bated at 30 °C for another 24 h. The genomic DNA purification was car-
(THL), Helsinki. Only one isolate per patient was permitted. The isolates ried out from 1 ml of the bacterial cell suspension using the GeneJET
studied included 17 serotype O:1 isolates associated with an outbreak of Genomic DNA Purification Kit (Fermentas, St. Leon-Rot, Germany)
38 laboratory-confirmed cases in Finland in 2008 (Rimhanen-Finne and according to the manufacturer's instructions. The precipitation phase
Sihvonen, 2010). These 17 isolates were chosen due to a high frequency of the DNA isolation was performed twice to ensure maximum yield
of reactive arthritis associated with this outbreak (Vasala et al., 2013). In of DNA. DNA quality and quantity were assessed with NanoDrop ND-
addition, 32 randomly selected isolates (29 O:1, 3 O:3) from four small 1000 Spectrophotometer (NanoDrop Technologies Inc., Wilmington,
infection clusters were included (within these clusters the isolates had Delaware, USA) using ND-1000 software version 3.1.0. The genomic
temporal and spatial linkage between each other). All other 55 isolates DNA had an A260/280 ratio of ≥1.8 and a concentration of N10 ng/μl.
were from sporadic human infections and represented dispersed
geographical occurrence and the year (between 1997 and 2012) of 2.4. PCR
isolation. Three additional isolates were previously obtained from
Prof. Mikael Skurnik (Department of Bacteriology and Immunology, The PCR amplification was performed with a BioRad DNA Engine
Haartman Institute, University of Helsinki) representing serotypes Dyad Peltier thermal cycler (Hercules, California, USA). The PCR re-
O:1 (Skurnik laboratory strain collection no. 2060 and 2061) and actions were carried out in the total volume of 25 μl containing 1 U
O:3 (Skurnik laboratory strain collection no. 2066) (Skurnik et al., of Taq DNA polymerase (Thermo Scientific, Massachusetts, USA), 1x
2000) and used as controls. The isolates were stored in the THL cul- Taq buffer supplied with the enzyme and 1.5 mM MgCl2, 5 pmol of
ture collection as FE100991, FE100992 and FE100994 bringing the each primer and approximately 15 ng of template DNA. The thermal
total number of isolates in this study to 107. These three isolates cycling conditions were 95 °C initial denaturation for 3 min, 35 cy-
and two randomly selected serotype O:1 isolates (FE73488 and cles of 95 °C denaturation for 30 s, 60 °C annealing for 30 s and
FE82911) and one serotype O:3 isolate (FE81617) were used for 72 °C extension for 60 s ending with a 5 min final extension step at
the VNTR region sequencing. 72 °C. Primer YPb9 had an annealing temperature of 65 °C.

2.2. VNTR markers 2.5. Capillary electrophoresis

The four complete genome sequences of Y. pseudotuberculosis The PCR products were diluted in a ratio of 1:85 in sterile water and
available in the GenBank database in 2011 were used for primer design: run in capillary electrophoresis at SeqLab of the Institute for Molecular
Y. pseudotuberculosis IP32953, Y. pseudotuberculosis IP31758, Y. pseudo- Medicine Finland (FIMM, Helsinki) on an ABI3730xl DNA Analyzer
tuberculosis PB1/+ and Y. pseudotuberculosis YPIII (GenBank accession (Applied Biosystems, Foster City, CA). The GeneScan™ 600 LIZ®
numbers NC_006155.1, NC_009708.1, NC_010634.1 and NC_010465.1, (Applied Biosystems) was used as an internal size standard. The elec-
respectively). The genomes were analysed to locate candidate VNTR tropherograms were analysed using the Local Southern size calling
markers using a tool provided by the Microsatellites/Tandem Repeats method in the Peak Scanner 1.0 software (Applied Biosystems).
database (Bikandi et al., 2004). A total of 42 markers were found. The
potential markers were compared with the Tandem Repeats Finder pro- 2.6. DNA sequencing
gram (Benson, 1999). The tandem repeats that contained a minimum of
four bases, were repeated at least four times in tandem and were lo- Sequencing of the seven VNTR markers was performed for six iso-
cated in the chromosome of Y. pseudotuberculosis were selected. The lates (FE100991, FE100992, FE10094, FE73488, FE82911 and FE81617)
selected repeats were present in at least two of the four complete ge- on both strands using a Big Dye Terminator v1.1 Cycle Sequencing Kit
nome sequences examined. In all, 16 VNTR markers were obtained (Applied Biosystems) with an ABI 3730xl DNA Analyzer (Applied
that did not contain any mismatches or gaps using the Tandem Repeats Biosystems). The sequencing was done with the MLVA primers without
Finder program. The primers for the 16 VNTR markers were designed by the fluorescent label (Table 1). The sequences were aligned by using
using the tool provided by the Microsatellites/Tandem Repeats data- BioEdit version 7.1.3.0 (Hall, 1999). The location of the fragments on
base. An in silico PCR (Bikandi et al., 2004) was performed to narrow the chromosome of Y. pseudotuberculosis IP32953 was determined by
the potential options using Y. pseudotuberculosis IP32953 and Y. pseudo- Basic Local Alignment Search Tool (BLAST) (Zhang et al., 2000) search
tuberculosis PB1/+ complete genome sequences as template. (Table 1). The sizes of the flanking regions of each locus were calculated
 J. Halkilahti et al. / Journal of Microbiological Methods 95 (2013) 245–250 247

from the obtained sequencing data and from the genome sequences in

ORF… 122007-122891 YPTB0107 metF

hypothetical protein (Protein is coded

ORF… 1461648-1462121 YPTB1227 -

ORF… 1703459-1704904 YPTB1422 -

ORF… 2304769-2306190 YPTB1958 -

ORF… 2537434-2538744 YPTB2158 -

ORF… 3321396-3322481 YPTB2812 -

transporter ORF… 2538824-2539822
GntR family transcriptional regulator
zinc uptake transcriptional repressor

major facilitator superfamily tartrate

(Protein is coded in complementary
the GenBank. In addition, locus YPb10 was sequenced from 20 random-

5,10-methylenetetrahydrofolate re-

ORF… 443899-444408 YPTB0371 -

ABC opine/polyamine transporter,

YPTB2159 - hypothetical protein
strand; from 444408 to 443899)
ly selected isolates to investigate the stability of the 3′ flanking region of

in complementary strand; from

Part of gene(s) amplified in Y.

the locus.
pseudotuberculosis IP32953

1704904 to 1703459)

ATP-binding subunit
hypothetical protein
2.7. Data analysis

Simpson's diversity index (DI) (Hunter and Gaston, 1988) was

ductase

calculated for each of the loci by using the PHYLOViZ software

MLVA primers designed for Y. pseudotuberculosis, lengths of the flanking sequences and fragments, repeat sequences and the formula for calculating the repeat number from the size of the amplified fragment.

(Francisco et al., 2012) version 1.0. All the generated MLVA profiles
were used for the calculations. The data on serotypes O:1 and O:3
Fragment size (bp)

were processed separately.

3. Results
179

231

202

197

187
181

231
Ten VNTR loci fulfilled our criteria for minimum of four bases that
were repeated at least four times in a tandem repeat. Three of these
Fragment position in Y. pseudotuberculosis

VNTR markers were omitted due to their poorly repeatable PCR-

results, leaving us with the seven loci suitable for genotyping of the
Y. pseudotuberculosis isolates (Table 1). The different MLVA profiles
were named as a string of seven numbers representing the actual num-
from 3322462 to 3322648
from 1462081 to 1462282

from 1703649 to 1703845

from 2538646 to 2538876

from 2306115 to 2306295

ber of tandem repeat units in each locus. The profiles were based on the
from 122802 to 122980

from 443723 to 443953

number of 6 to 9 bp long tandem repeats in each locus, and the flanking

regions of the loci ranged from 127 to 175 bp (Table 1).
Comparison of the complete Y. pseudotuberculosis genomes available
in the GenBank revealed that locus YPb5 had a flanking region of either
IP32953

152 or 153 bp long (Table 1). Alignment of the 3′-flanking region of

locus YPb5 of all six partially sequenced isolates revealed a one-
base deletion in the same position of the region. This one base differ-
ence was observed in 50% of all available complete genome sequences
(X-[152 or 153])/7

of Y. pseudotuberculosis and also in the six partially sequenced isolates.

Repeat number

The fragment sizes obtained from sequencing (EMBL Nucleotide Se-

(X-143)/9

(X-175)/7

(X-159)/7
(X-147)/7
(X-147)/8

(X-127)/6

quence Database accession numbers HF584708–HF584747) and from

capillary electrophoresis differed from each other in all of the cases
(data not shown). However, the differences between the sequencing
and capillary electrophoresis results were always less than one tandem
ATTTCCTGT
TGGTGGTT

repeat in any of the loci with a mean value of 2,1 bp. Isolate FE100992
TAACGAC

TGAAGTT
GTTAATA
TGCTTTT

ATCAAC
Repeat

had a 7 bp deletion in the 3′-flanking sequence at locus YPb10 when

compared to the 25 other YPb10 loci sequences (data not shown).
The groups of 63 serotype O:1 isolates and 44 serotype O:3 isolates
Length of flanking

were both discriminated into 12 unique, serotype-specific MLVA pro-

files (Table 2). The number of repeats varied from 1 to 23 depending
region (bp)

152 or 153

on the locus. The most frequent, MLVA profiles were obtained for 52 se-
rotype O:1 and 18 serotype O:3 isolates. There was no correlation be-
147

147

143

127

175

159

tween the MLVA profiles and the time of isolation (Table 3).
To study the stability of the VNTR markers, 17 isolates from an
outbreak at North-Eastern part of Finland in 2008 were examined.
5′-6FAM-GCGAAGGGGTGAAGGATT-3′

5′-6FAM-GGAGGGGGTGACACACTG-3′
5′-6FAM-TGGTCAGAACAAAGCGCA-3′
5′-6FAM-CGCTGACTGGCAGGAAAT-3′

5′-6FAM-CCCAGAGTACCACGACCG-3′
5′-6FAM-TGCCTGTTGAGGTACGGC-3′

5′-6FAM-GGGCCTGACACTGCGTAT-3′

These 17 outbreak isolates represented serotype O:1 and all of

them had an identical MLVA profile 4-7-3-5-5-5-2 (Table 3). This
5′-GCGAGCAGAGAAACTCGC-3′

5′-CGAGACGGTGAGCGAGTT-3′
5′-GCGAAAGCCACGATGATT-3′

5′-GGCCATCTGGGTCAACAG-3′
5′-AACAACCGAACCCGATCA-3′

5′-GGCGGATTACCACACCTG-3′

5′-GTCAATCAATCGACGCCC-3′

profile occurred previously in 2000 and since then every year. In

2000, it was found among six of the randomly selected seven O:1 iso-
lates. In 2001, one multilocus variant (4-7-7-3-2-2-3) was found
among 11 O:1 isolates that were thought to belong to the same clus-
Primer sequence

ter. Similarly in 2004, one single locus variant (4-7-2-5-5-5-2) was

found among seven O:1 isolates. Three of the O:3 isolates previously
thought to represent a small outbreak in 2001 belonged to two single
locus variants (5-3-9-2-1-3-2 and 5-7-9-2-1-3-2) (Table 3).
Diversity indices were calculated for individual loci to investigate
YPbR10
YPbF10

their discriminatory power. The DIs for serotype O:1 isolates for
Primer

YPbR1

YPbR3

YPbR5

YPbR7

YPbR8

YPbR9
YPbF1

YPbF3

YPbF5

YPbF7

YPbF8

YPbF9

the loci YPb1, YPb3, YPb5, YPb7, YPb8, YPb9 and YPb10 were
0,7727; 0,8788; 0,8939; 0,7576; 0,8485; 0,8636 and 0,8788, re-
spectively. The DIs for serotype O:3 isolates for the loci YPb1,
YPb10
Locus
Table 1

YPb1

YPb3

YPb5

YPb7

YPb8

YPb9

YPb3 and YPb5 were 0,6212; 0,5758; 0,7576, respectively, and 0,0
for the loci YPb7 to YPb10.
248 J. Halkilahti et al. / Journal of Microbiological Methods 95 (2013) 245–250

Table 2 Table 3
Diversity of the seven VNTR loci in 63 serotype O:1 isolates and 44 serotype O:3 isolates. The year of isolation for the studied 104 clinical Y. pseudotuberculosis strains, their
serotype, MLVA profile and the number of strains with the certain MLVA profile. Isolates
MLVAProfiles O:1 Frequency MLVAProfiles O:3 Frequency FE100991, FE100992 and FE100994 were not included.
4-7-3-5-5-5-2 52 4-6-9-2-1-3-2 18
Isolation year ST MLVA profile No. MLVA
3-16-4-2-5-9-23 1 4-6-10-2-1-3-2 7
profiles
4-10-5-3-3-9-5 1 4-6-12-2-1-3-2 4
4-11-4-6-4-2-8a 1 5-6-9-2-1-3-2 4 1999 O:1 8 - 6 - 5 - 3 - 9 - 4 - 5 1
4-3-6-3-2-2-5a 1 5-7-9-2-1-3-2 3 1999 O:1 4 - 10 - 5 - 3 - 3 - 9 - 5 1
4-7-2-5-5-5-2 1 5-6-10-2-1-3-2 2 1999 O:1 3 - 16 - 4 - 2 - 5 - 9 - 23 1
4-7-7-3-2-2-3 1 12-8-9-2-1-3-2a 1 2000 O:1 4 - 7 - 3 - 5 - 5 - 5 - 2 6
5-11-13-5-3-7-8 1 4-6-11-2-1-3-2 1 2000 O:1 7 - 9 - 3 - 2 - 3 - 7 - 13 1
6-7-3-2-3-7-12 1 4-6-13-2-1-3-2 1 2001 O:1 4 - 7 - 3 - 5 - 5 - 5 - 2 10a
7-9-3-2-3-7-13 1 5-3-9-2-1-3-2 1 2001 O:1 4 - 7 - 7 - 3 - 2 - 2 - 3 1a
8-6-5-3-9-4-5 1 5-5-9-2-1-3-2 1 2002 O:1 4 - 7 - 3 - 5 - 5 - 5 - 2 1
9-6-5-3-8-4-5 1 5-6-8-2-1-3-2 1 2003 O:1 4 - 7 - 3 - 5 - 5 - 5 - 2 7b
2003 O:1 6 - 7 - 3 - 2 - 3 - 7 - 12 1
Unique: 12 Total: 63 Unique: 12 Total: 44 2004 O:1 4 - 7 - 3 - 5 - 5 - 5 - 2 6c
a Isolates received from Prof. Skurnik. 2004 O:1 4 - 7 - 2 - 5 - 5 - 5 - 2 1c
2005 O:1 9 - 6 - 5 - 3 - 8 - 4 - 5 1
2006 O:1 4 - 7 - 3 - 5 - 5 - 5 - 2 4d
2006 O:1 5 - 11 - 13 - 5 - 3 - 7 - 8 1
2008 O:1 4 - 7 - 3 - 5 - 5 - 5 - 2 17e
2009 O:1 4 - 7 - 3 - 5 - 5 - 5 - 2 1
4. Discussion 1997 O:3 5 - 7 - 9 - 2 - 1 - 3 - 2 1
1998 O:3 4 - 6 - 9 - 2 - 1 - 3 - 2 6
1998 O:3 4 - 6 - 10 - 2 - 1 - 3 - 2 1
We designed an MLVA method based on seven loci in the chro- 1999 O:3 4 - 6 - 9 - 2 - 1 - 3 - 2 5
mosome of Y. pseudotuberculosis for comparison of the clinical iso- 2000 O:3 5 - 6 - 9 - 2 - 1 - 3 - 2 4
lates. The method is based on capillary electrophoresis separation 2000 O:3 5 - 6 - 8 - 2 - 1 - 3 - 2 1
2000 O:3 4 - 6 - 10 - 2 - 1 - 3 - 2 1
of PCR products labelled with a fluorescent dye. In assigning the tandem
2000 O:3 4 - 6 - 9 - 2 - 1 - 3 - 2 1
repeat numbers, we followed the style used for Salmonella Typhimurium 2001 O:3 5 - 3 - 9 - 2 - 1 - 3 - 2 1a
(Larsson et al., 2009). A minimum of four bases, that were repeated 2001 O:3 5 - 7 - 9 - 2 - 1 - 3 - 2 2a
at least four times in a tandem repeat were selected. The prokaryotic 2001 O:3 4 - 6 - 11 - 2 - 1 - 3 - 2 1
microsatellites have been proposed to be connected to the slippage 2001 O:3 4 - 6 - 10 - 2 - 1 - 3 - 2 1
2001 O:3 4 - 6 - 9 - 2 - 1 - 3 - 2 4
of the DNA polymerase, which is required for the polymorphism of
2001 O:3 5 - 6 - 10 - 2 - 1 - 3 - 2 2
the VNTR loci and is more frequent with short repeats (Noller et al., 2002 O:3 4 - 6 - 10 - 2 - 1 - 3 - 2 1
2003). Therefore, short tandem repeats were chosen as VNTR markers 2006 O:3 4 - 6 - 10 - 2 - 1 - 3 - 2 1
in this study. In the previous five-loci MLVA method designed for 2007 O:3 4 - 6 - 10 - 2 - 1 - 3 - 2 1
2008 O:3 5 - 5 - 9 - 2 - 1 - 3 - 2 1
Salmonella Typhimurium, one of the VNTR markers is located in a plas-
2008 O:3 4 - 6 - 9 - 2 - 1 - 3 - 2 1
mid (STTR10pl) (Lindstedt et al., 2003). The locus in question had the 2008 O:3 4 - 6 - 10 - 2 - 1 - 3 - 2 1
highest percentage (48%) of missing PCR products in comparison to all 2012 O:3 4 - 6 - 12 - 2 - 1 - 3 - 2 4
the other loci used in the method. The locus STTR10pl also had the 2012 O:3 4 - 6 - 13 - 2 - 1 - 3 - 2 1
highest polymorphism rate. However, plasmids are gained and lost 2012 O:3 4 - 6 - 9 - 2 - 1 - 3 - 2 1

quite easily by Yersinia (Eppinger et al., 2007), especially in laboratory a 11 serotype O:1 isolates and 3 serotype O:3 isolates belonging to the infection clusters in

cultures. Therefore, we did not include any VNTR markers possibly 2001 (Jalava et al., 2004).
b 7 serotype O:1 isolates belonging to the infection cluster in 2003 (Jalava et al., 2006).
located in Y. pseudotuberculosis plasmids, although they might have
c 7 serotype O:1 isolates belonging to the infection cluster in 2004 (Kangas et al., 2008).
added to the polymorphism.
d 4 serotype O:1 isolates belonging to the infection cluster in 2006 Rimhanen-Finne and
The VNTR loci showed variety of polymorphism within our set of Sihvonen, 2010).
Y. pseudotuberculosis isolates: the analysis of the 107 isolates yielded al- e 17 serotype O:1 isolates belonging to the infection cluster in 2008 (Rimhanen-Finne and

together 24 distinct MLVA patterns. The 12 MLVA patterns obtained for Sihvonen, 2010; Vasala et al., 2013).

both serotypes O:1 and O:3 showed that the method presented here is
capable of discriminating within as well as between the isolates of the
two serotypes; namely, all of the MLVA profiles detected from 107 iso- The possibility that one strain evolves into several genotypes during
lates were serotype-specific. an outbreak can pose a problem in the use of the MLVA method
Partial sequencing of six isolates was used to cross-check the theo- (Noller et al., 2003) as in the use of other genotypic subtyping methods.
retical and actual sizes of the flanking regions in each loci. The theoret- In theory, the cellular mechanisms that increase or decrease the number
ical sizes of each locus, which were calculated from the complete of repeats by one in a given loci can also alter the tandem repeat number
genome sequences, matched the actual flanking region sizes obtained by several steps (Bichara et al., 2006). Therefore, an isolate with a
via sequencing. Thus, it is apparent that the flanking region sizes are change in the copy number in one locus might still be part of an out-
conserved throughout the studied isolates. As an exception, the isolate break. Common guidelines, similar to those described by Tenover et al.
FE100992 (serotype O:1) had a 7 bp deletion in the 3′ flanking se- (1995) for PFGE, aiding the interpretation of the different MLVA profiles
quence at locus YPb10 when compared to the other 25 YPb10 loci se- in the epidemiological investigations should be created as more isolates
quenced. Also, the occurrence of flanking regions of the two slightly are analysed.
different sizes (1 bp) at locus YPb5 was observed in our study with Several investigations for epidemiology of Y. pseudotuberculosis
the partially sequenced isolates. These findings should be accounted have been conducted in Finland. They have shown that the non-
for when analysing the results to avoid incorrect size calling and binning outbreak isolates were genetically very closely related to the observed
of the capillary electrophoresis results. outbreak isolates (Hallanvuo et al., 2002). In the 2000s, multiple
In epidemiological investigation, the attention should always be fo- Y. pseudotuberculosis outbreaks, mainly caused by O:1 isolates, occurred
cused on the combination of both pheno- and/or genotypic features of in Finland (Jalava et al., 2004, 2006; Kangas et al., 2008; Rimhanen-Finne
the causative agent as well as on the analytical epidemiological data. and Sihvonen, 2010). Grated carrot was recognized as a vehicle in
 J. Halkilahti et al. / Journal of Microbiological Methods 95 (2013) 245–250 249

some of the outbreaks (Kangas et al., 2008; Rimhanen-Finne et al., References

2009; Rimhanen-Finne and Sihvonen, 2010) but in some the vehicle
of the pathogen remained unknown (Hallanvuo et al., 2002; Jalava Benson, G., 1999. Tandem repeats finder: a program to analyze DNA sequences. Nucleic
Acids Res. 27, 573–580.
et al., 2006; Rimhanen-Finne and Sihvonen, 2010). For example in Bichara, M., Wagner, J., Lambert, I.B., 2006. Mechanisms of tandem repeat instability in
2001, there were altogether 55 cases caused by serotype O:1 isolates bacteria. Mutat. Res. 598, 144–163.
and 34 cases caused by serotype O:3 isolates (Jalava et al., 2004). In Bikandi, J., San Millán, R., Rementeria, A., Garaizar, J., 2004. In silico analysis of complete
bacterial genomes: PCR, AFLP-PCR and endonuclease restriction. Bioinformatics 20,
the data set that was randomly selected (except the 17 outbreak 798–799.
strains from the year 2008) for our study, 11 O:1 isolates and three Brubaker, R.R., 1991. Factors promoting acute and chronic diseases caused by yersiniae.
O:3 isolates represented this outbreak in 2001. There were four dis- Clin. Microbiol. Rev. 4, 309–324.
Eppinger, M., Rosovitz, M.J., Fricke, W.F., Rasko, D. a, Kokorina, G., Fayolle, C., Lindler,
tinct MLVA profiles amongst the outbreak isolates. One serotype O:1 L.E., Carniel, E., Ravel, J., 2007. The complete genome sequence of Yersinia pseu-
MLVA profile (4-7-7-3-2-2-3) differed considerably from the profiles dotuberculosis IP31758, the causative agent of Far East scarlet-like fever. PLoS
of the other 10 serotype O:1 outbreak isolates. The other outbreaks Genet. 3, e142.
Francisco, A.P., Vaz, C., Monteiro, P.T., Melo-Cristino, J., Ramirez, M., Carriço, J.A., 2012.
caused by Y. pseudotuberculosis in Finland from 2003 to 2008 were
PHYLOViZ: phylogenetic inference and data visualization for sequence based typing
all associated with serotype O:1 (Jalava et al., 2006; Kangas et al., methods. BMC Bioinforma. 13, 87.
2008; Rimhanen-Finne et al., 2009; Rimhanen-Finne and Sihvonen, Fukushima, H., Matsuda, Y., Seki, R., Tsubokura, M., Takeda, N., Shubin, F.N., Paik,
2010; Vasala et al., 2013). The isolates from 2003 to 2006 outbreaks I.K., Zheng, X.B., 2001. Geographical heterogeneity between Far Eastern and
Western countries in prevalence of the virulence plasmid, the superantigen
represented two MLVA profiles that were single locus variants. The Yersinia pseudotuberculosis-derived mitogen, and the high-pathogenicity
17 isolates from the outbreak at the North-Eastern part of Finland island among Yersinia pseudotuberculosis strains. J. Clin. Microbiol. 39,
in 2008 had identical MLVA profiles as expected. 3541–3547.
Gierczyński, R., Golubov, A., Neubauer, H., Pham, J.N., Rakin, A., 2007. Development of
Overall, the serotype O:1 MLVA profile 4-7-3-5-5-5-2 was the multiple-locus variable-number tandem-repeat analysis for Yersinia enterocolitica
most common (52 isolates) among all 63 O:1 isolates tested and it subsp. palearctica and its application to bioserogroup 4/O3 subtyping. J. Clin.
was present during several years. This dominant serotype O:1 MLVA Microbiol. 45, 2508–2515.
Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis
profile was associated with all of the O:1 outbreaks. Out of the total program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41, 95–98.
63 serotype O:1 isolates, 14 were clearly sporadic. The most common Hallanvuo, S., Nakari, U., Siitonen, A., 2002. Pulsed-field Electrophoresis in Outbreak In-
serotype O:3 MLVA profile 4-6-9-2-1-3-2 was observed in 18 of the vestigations of Yersinia pseudotuberculosis.
Hallanvuo, S., Nuorti, P., Nakari, U.-M., Siitonen, A., 2003. Molecular epidemiology of the
44 isolates tested. In 2001, three O:3 isolates were associated with sev-
five recent outbreaks of Yersinia pseudotuberculosis in Finland. Adv. Exp. Med. Biol.
eral small infection clusters (Jalava et al., 2004). Two MLVA profiles 529, 309–312.
were obtained from these three outbreak isolates, which differed in Hunter, P.R., Gaston, M.A., 1988. Numerical index of the discriminatory ability of typing sys-
tems: an application of Simpson’s index of diversity. J. Clin. Microbiol. 26, 2465–2466.
two loci from the most common O:3 MLVA profile. The remaining
Jalava, K., Hallanvuo, S., Nakari, U.-M., Ruutu, P., Kela, E., Heinäsmäki, T., Siitonen, A.,
41 O:3 isolates studied were considered sporadic. This finding that Nuorti, J.P., 2004. Multiple outbreaks of Yersinia pseudotuberculosis infections in
the vehicle in five of the outbreaks from 2003 to 2008 was carrots Finland. J. Clin. Microbiol. 42, 2789–2791.
(Kangas et al., 2008; Rimhanen-Finne et al., 2009; Rimhanen-Finne Jalava, K., Hakkinen, M., Valkonen, M., Nakari, U.-M., Palo, T., Hallanvuo, S., Ollgren, J.,
Siitonen, A., Nuorti, J.P., 2006. An outbreak of gastrointestinal illness and erythema
and Sihvonen, 2010) could easily lead to an assumption that the reser- nodosum from grated carrots contaminated with Yersinia pseudotuberculosis.
voir for all mentioned outbreaks are identical or that the strain is widely J. Infect. Dis. 194, 1209–1216.
spread in nature and that the MLVA type has remained stable in that Johansson, A., Göransson, I., Larsson, P., Sjöstedt, A., 2001. Extensive allelic variation
among Francisella tularensis strains in a short-sequence tandem repeat region.
reservoir. Further study is required to elucidate this possibility. J. Clin. Microbiol. 39, 3140–3146.
This study focused on the MLVA typing of the serotype O:1 and O:3 Kangas, S., Takkinen, J., Hakkinen, M., Nakari, U.-M., Johansson, T., Henttonen, H.,
isolates that have been common in human infections in Finland during Virtaluoto, L., Siitonen, A., Ollgren, J., Kuusi, M., 2008. Yersinia pseudotuberculosis
O:1 traced to raw carrots, Finland. Emerg. Infect. Dis. 14, 1959–1961.
the past 15 years. The MLVA method described here is an additional tool Keim, P., Price, L.B., Klevytska, a M., Smith, K.L., Schupp, J.M., Okinaka, R., Jackson,
for subtyping Y. pseudotuberculosis isolates and it is in line with the P.J., Hugh-Jones, M.E., 2000. Multiple-locus variable-number tandem repeat
recent methodological development for other enteric bacteria. analysis reveals genetic relationships within Bacillus anthracis. J. Bacteriol.
182, 2928–2936.
Testing the method with isolates of other serotypes will yield Klevytska, A.M., Price, L.B., Schupp, J.M., Worsham, P.L., Wong, J., Keim, P., 2001. Identifi-
more information on the discriminatory potential and sensitivity cation and characterization of variable-number tandem repeats in the Yersinia pestis
of the Y. pseudotuberculosis MLVA scheme. Multiplexing the PCR genome. J. Clin. Microbiol. 39, 3179–3185.
Larsson, J.T., Torpdahl, M., Petersen, R.F., Sorensen, G., Lindstedt, B.A., Nielsen, E.M., 2009.
primers with the same annealing temperature and labelled with
Development of a new nomenclature for Salmonella typhimurium multilocus vari-
different fluorescent dyes can be done for routine use, which able number of tandem repeats analysis (MLVA). Euro Surveill. 14, 1–5.
would shorten the time necessary to perform the analysis further. Lindstedt, B.-A., 2005. Multiple-locus variable number tandem repeats analysis for genetic
fingerprinting of pathogenic bacteria. Electrophoresis 26, 2567–2582.
Lindstedt, B.-A., Heir, E., Gjernes, E., Kapperud, G., 2003. DNA fingerprinting of Sal-
monella enterica subsp. enterica serovar typhimurium with emphasis on phage
5. Conclusions type DT104 based on variable number of tandem repeat loci. J. Clin. Microbiol.
41, 1469–1479.
Niskanen, T., Fredriksson-Ahomaa, M., Korkeala, H., 2002. Yersinia pseudotuberculosis
During the recent years, MLVA has been successfully applied for mo- with limited genetic diversity is a common finding in tonsils of fattening pigs.
lecular typing of several bacterial pathogens. We designed an MLVA J. Food Prot. 65, 540–545.
scheme based on seven loci, which were able to discriminate clinical Niskanen, T., Waldenström, J., Fredriksson-Ahomaa, M., Olsen, B., Korkeala, H., 2003. virF-
positive Yersinia pseudotuberculosis and Yersinia enterocolitica found in migratory
isolates of Y. pseudotuberculosis serotypes O:1 and O:3 into several birds in Sweden. Appl. Environ. Microbiol. 69, 4670–4675.
MLVA types. In conclusion, the MLVA method developed by us offers a Noller, A.C., McEllistrem, M.C., Pacheco, A.G.F., Boxrud, D.J., Harrison, L.H., 2003.
new molecular tool for epidemiological investigation of Y. pseudotuber- Multilocus variable-number tandem repeat analysis distinguishes outbreak and
sporadic Escherichia coli O157:H7 isolates. J. Clin. Microbiol. 41, 5389–5397.
culosis outbreaks and for comparison of the sporadic isolates.
Nuorti, J.P., Niskanen, T., Hallanvuo, S., Mikkola, J., Kela, E., Hatakka, M., Fredriksson-
Ahomaa, M., Lyytikainen, O., Siitonen, A., Korkeala, H., Ruutu, P., 2004. A widespread
outbreak of Yersinia pseudotuberculosis O:3 infection from iceberg lettuce. J. Infect.
Dis. 189, 766–774.
Acknowledgments
Platonov, M.E., Blouin, Y., Evseeva, V.V., Afanas’ev, M.V., Pourcel, C., Balakhonov, S.V.,
Vergnaud, G., Anisimov, A.P., 2013. Draft genome sequences of five Yersinia
We thank Prof. Mikael Skurnik for providing three Y. pseudotubercu- pseudotuberculosis ST19 isolates and one isolate variant. Genome announce-
losis reference strains and the staff of the Bacteriology Unit, THL, for their ments 1, e0012213.
Rimhanen-Finne, R., Sihvonen, L., 2010. Yersinia. In: Hulkko, T., Lyytikäinen, O., Kuusi, M.,
assistance. The work was supported by a grant (STM/4608/2010) from Seppälä, S., Ruutu, P. (Eds.), Infectious Diseases in Finland 1995–2009. National Insti-
the Finnish Ministry of Social Affairs and Health. tute for Health and Welfare, Helsinki, p. 22.
250 J. Halkilahti et al. / Journal of Microbiological Methods 95 (2013) 245–250

Rimhanen-Finne, R., Niskanen, T., Hallanvuo, S., Makary, P., Haukka, K., Pajunen, S., Siitonen, produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing.
A., Ristolainen, R., Pöyry, H., Ollgren, J., Kuusi, M., 2009. Yersinia pseudotuberculosis J. Clin. Microbiol. 33, 2233–2239.
causing a large outbreak associated with carrots in Finland, 2006. Epidemiol. Infect. Tertti, R., Granfors, K., Lehtonen, O.P., Mertsola, J., Mäkelä, A.L., Välimäki, I., Hänninen, P.,
137, 342–347. Toivanen, A., 1984. An outbreak of Yersinia pseudotuberculosis infection. J. Infect.
Rozen, S., Skaletsky, H., 2000. Primer3 on the WWW for general users and for biologist Dis. 149, 245–250.
programmers. Methods Mol. Biol. 132, 365–386. Tertti, R., Vuento, R., Mikkola, P., Granfors, K., Mäkelä, a L., Toivanen, A., 1989. Clinical
SantaLucia, J., 1998. A unified view of polymer, dumbbell, and oligonucleotide DNA manifestations of Yersinia pseudotuberculosis infection in children. Eur. J. Clin.
nearest-neighbor thermodynamics. Proc. Natl. Acad. Sci. U. S. A. 95, 1460–1465. Microbiol. Infect. Dis. 8, 587–591.
Sihvonen, L.M., Toivonen, S., Haukka, K., Kuusi, M., Skurnik, M., Siitonen, A., 2011. Van Belkum, A., Scherer, S., van Leeuwen, W., Willemse, D., van Alphen, L., Verbrugh, H.,
Multilocus variable-number tandem-repeat analysis, pulsed-field gel electrophoresis, 1997. Variable number of tandem repeats in clinical strains of Haemophilus
and antimicrobial susceptibility patterns in discrimination of sporadic and outbreak- influenzae. Infect. Immun. 65, 5017–5027.
related strains of Yersinia enterocolitica. BMC Microbiol. 11, 42. Vasala, M., Hallanvuo, S., Ruuska, P., Suokas, R., Siitonen, A., Hakala, M., 2013. High fre-
Skuce, R. a, McCorry, T.P., McCarroll, J.F., Roring, S.M.M., Scott, A.N., Brittain, D., Hughes, quency of reactive arthritis in adults after Yersinia pseudotuberculosis O:1 outbreak
S.L., Hewinson, R.G., Neill, S.D., 2002. Discrimination of Mycobacterium tuberculosis caused by contaminated grated carrots. Ann. Rheum. Dis. [Epub ahead of print].
complex bacteria using novel VNTR-PCR targets. Microbiology 148, 519–528. Vincent, P., Leclercq, A., Martin, L., Duez, J.-M., Simonet, M., Carniel, E., 2008. Sudden onset
Skurnik, M., Peippo, A., Ervelä, E., 2000. Characterization of the O-antigen gene clusters of of pseudotuberculosis in humans, France, 2004-05. Emerg. Infect. Dis. 14, 1119–1122.
Yersinia pseudotuberculosis and the cryptic O-antigen gene cluster of Yersinia pestis Wobeser, G., Campbell, G.D., Dallaire, A., McBurney, S., 2009. Tularemia, plague,
shows that the plague bacillus is most closely related to and has evolved from Y. yersiniosis, and Tyzzer’s disease in wild rodents and lagomorphs in Canada: a re-
pseudotuberculosis serotype O:1b. Mol. Microbiol. 37, 316–330. view. Can. Vet. J. 50, 1251–1256.
Tenover, F.C., Arbeit, R.D., Goering, R.V., Mickelsen, P.A., Murray, B.E., Persing, D.H., Zhang, Z., Schwartz, S., Wagner, L., Miller, W., 2000. A greedy algorithm for aligning DNA
Swaminathan, B., 1995. Interpreting chromosomal DNA restriction patterns sequences. J. Comput. Biol. 7, 203–214.