The Use of Pulsed Field Gel Electrophoresis in Listeria monocytogenes Sub-Typing – Comparison with MLVA Method Coupled with Gel Electrophoresis

Out of the several molecular methods currently available, pulsed field gel electrophoresis (PFGE) is one of the most discriminatory and reproducible methods for the sub-typing of Listeria monocytogenes (L. monocytogenes) (Kerouanton et al., 1998; Brosch et al., 1996). The combination of restriction endonucleases AscI and ApaI has shown excellent discrimination for L. monocytogenes (Brosch et al., 1996). Thus, the PFGE method, using these two enzymes, is considered to be the international standard for sub-typing (Graves and Swamminathan, 2001). However, although the protocol has been shortened to 30 hours from the time a pure culture of the bacteria has been obtained (Graves and Swamminathan, 2001), PFGE remains a manual, time-consuming and labor intensive subtyping method. It also requires highly skilled operators and does not offer standardized reagents.

In the MLVA schemes developed so far, fragment detection is performed by CE. Nevertheless, Murphy et al (2007) demonstrated that it is possible to detect VNTR loci on agarose gels. However, most of the 45 isolates tested in this study had the same origin (food origin) and the same serotype (1/2a). Moreover, out of the six VNTR loci described by Murphy et al. (2007), four were excluded from the PulseNet USA MLVA protocol because two loci have low diversity and two others display sequence variability in flanking regions. The purpose of the present study was to evaluate the feasibility of a MLVA protocol coupled with conventional gel electrophoresis. The results were compared with those obtained by PFGE.

Serotyping
Species identification was performed using agar Listeria according to Ottaviani & Agosti (ALOA) plates (AES, Combourg, France) and the CAMP Test (McKellar 1994). Each strain was serotyped by agglutination using commercially available antisera (Denka, Eurobio, Les Ulis, France), after adapting the manufacturer's instructions and using the procedures outlined by Seeliger & Hohne (1979). Our laboratory has been certified by the French Accreditation Committee (COFRAC) for this serotyping method as an internal method (accreditation no. 1-22465, Section Laboratories, www.cofrac.fr). Determination of the Oantigen was performed from a pure culture [instead of a bacterial suspension]. Determination of the H-antigen was performed using semi-liquid brain heart infusion (BHI) media with 0.5% agar [instead of 0.2%].

Molecular serotyping
Molecular serotyping was performed using the protocol developed by Kerouanton et al. (2010).

PFGE
PFGE was performed using the standard CDC PulseNet protocol (Graves & Swamminathan, 2001) with minor modifications. Each strain was grown overnight on tryptone soya agar with yeast extract (TSAYE) plates instead of BHI. For the DNA digestion step in agarose plugs using ApaI and AscI enzymes, 10 units of enzyme were used per plug [instead of 25 units of enzyme per plug for AscI] and 160-200 units of enzyme per plug for ApaI in the PulseNet protocol. Plugs were incubated with restriction enzymes for 4 h [instead of 5 h]. Gels were then stained with ethidium bromide and banding patterns were visualized under UV light, using the Gel Doc EQ system and Quantity One software (Bio-Rad). DNA patterns were analyzed with BioNumerics software (ver. 6.5, Applied Maths, Kortrijk, Belgium). The recommendations of Barrett et al. (2006) were followed for gel analysis: gels including partial digestions, or unclear bands were not analyzed. All bands with sizes lower than 33 kb were systematically removed. A similarity value of 97.0% was established as a cut-off to consider two profiles as indistinguishable in UPGMA dendrograms using the Dice coefficient, with a 1% tolerance limit and 1% optimization. If the similarity value was strictly less than 97%, the two profiles were considered as different. The dendrogram settings used were chosen according to PulseNet Europe recommendations (Martin et al., 2006). The similarity value taken as the cut-off was established according to the EURL database settings. Each PFGE profile was arbitrarily assigned a number. Our laboratory has been certified by COFRAC for PFGE analyses (Accreditation no. 1-22465, Section Laboratories, www.cofrac.fr).

Locus selection
VNTR loci found in the literature with a repeat size greater than or equal to 9 bp were selected. New VNTR loci were selected from the complete genome of the three reference strains. The genomes of strains EGDe (1/2a), F2365 (4b) and CLIP 80459 (4b) were individually screened using the Tandem Repeat Finder (TRF) program (http://tandem.bu.edu/). The tandem repeat databases http://mlva.u-psud.fr and http://www.hpa-bionum.org.uk/VNTRUK/ were then used to compare the genomes.

Primer design
The primer sets were either similar to those described in the literature (Table 2), or designed in regions flanking the VNTR locus, (Table 3), using AlleleID® software (Premier Biosoft International, USA). All the primers were synthesized by Eurogentec (France).

Amplification of VNTR loci
The VNTR loci were amplified on DNA from strains EGDe and F2365. The amplification products were electrophoresed on two gels run independently.
For each primer set, the final mix contained 1 U HotStart Taq Polymerase (Roche), 2 or 3 mM MgCl 2 , 0.2 mM desoxynucleotide triphosphate, 1X PCR buffer, PCR grade water, 0.3 µM each primer, and 1 µl of DNA in a 25 µl reaction mixture. PCR was performed on a thermal cycler (GeneAmp PCR System, 9700, PE, Applied Biosystems). For Lm-8, the parameters used were those described by Sperry et al. (2008): initial denaturation at 95°C for 5 min, followed by 35 cycles of denaturation at 94°C for 20 s, annealing at 50°C for 20 s, extension at 72°C for 20 s and a final extension at 72°C for 5 min. For LMCEB 02, 06, 12, 14 and Lm-26: initial denaturation at 95°C for 5 min, followed by 35 cycles of denaturation at 94°C for 20 s, annealing at 57°C for 30 s, extension at 72°C for 30 s and a final extension at 72°C for 7 min. For LMCEB 05 the annealing was performed at 54°C. For JLR-4, the parameters used were those described by Larsson et al. (2010). At SSI, amplification for JLR-4 and Lm-8 was performed according to Larson et al. (2010).

Detection of VNTR loci
Aliquots (5 µl) of amplified products were electrophoresed on 2% agarose gels (Resophor, Eurobio, France) in 1X TBE buffer (0.45 mM Tris-HCl, 0.45 mM boric acid, 1 mM EDTA, pH 8). Electrophoresis was performed in 12 cm long gels and run at 80 V for 30 min followed by 90 V for 4 h. In each run, the 20 bp DNA Ladder (Bio-Rad, France) and the PCR products from the two strains EGDe and F2365 were systematically included at least twice to facilitate the sizing of amplified DNA fragments. Each run included a negative/water control to ensure the absence of contamination.

www.intechopen.com
The Use of Pulsed Field Gel Electrophoresis in Listeria monocytogenes Sub-Typing -Comparison with MLVA Method Coupled with Gel Electrophoresis

303
The gels were stained in 2 µg/ml ethidium bromide for 90 min and photographed under UV illumination (Gel Doc EQ R Bio-Rad, France). The length of each amplified VNTR locus was measured using Quantity One software (Bio-Rad, France). An allele number string based on the estimated number of tandem repeats at each locus was assigned to the amplified DNA fragments from each isolate. Detection of PCR products by capillary electrophoresis was performed according to Larsson et al. (2010).

Data analysis
The allele strings were imported into BioNumerics software. Dendrograms were constructed using a categorical coefficient and UPGMA clustering. Allele nomenclature was that recommended by PulseNet USA. No amplification was coded as negative (-1). Efficient amplification with no VNTR detected was coded as "zero" (0). Partial repeats were rounded down to the closest whole number.

Sequence verification
The loci and flanking regions were amplified in both directions with high-fidelity HotStart Taq Polymerase (Roche). Amplification products were sequenced by Eurofins (MWG Operon, France). The sequence analysis was performed with the CodonCode Aligner software (CodonCode Corporation, USA).

Stability determination
The stability test was performed according to Sperry et al. (2008): the strains EGDe, F2365 and CLIP 80459 were tested 45 times. All DNA were tested for MLVA.

Reproducibility
The reproducibility of the MLVA method was determined from the results obtained from the two reference strains included in each run, and with the four TS strains represented in duplicate and the epidemiologically related strains included in this study. Moreover, amplification products were systematically run on two independent gels. At least two independent PCRs were performed from a given DNA extract from the reference strains.

Selection of VNTR loci from the literature
A total of 16 VNTRs have been described in the literature (Table 1). Although some VNTRs are common to different MLVA schemes, their nomenclature is different. Moreover, the primer pairs used for the amplification of a given locus can differ among studies.

Locus name Primer names References
Lm
The locus Lm-26 had already been published but has not been used previously due to its low diversity (Sperry et al., 2008). For each of the nine VNTR loci, primers were designed in the regions flanking the locus (Table 4).

Amplification and detection of the selected VNTR loci from the two strains EGDe and F2365
For Lm-3, Lm-10, Lm-11, Lm-15, LMV6-JLR and LMV9-JLR, the size of the amplification products obtained with all the primer pairs tested, observed in the same run and in two different runs differed from the true length by up to 18 bp (data not shown). For this reason, other primer pairs were designed and tested. Sizing discrepancies were nevertheless observed (data not shown).  Table 4. Primers and characteristics of PCR amplification products in the reference strains for each of the nine VNTR loci selected through a bioinformatics-based search.
observed on the gel of strains EGDe and F2365 were solely related to the differences in repeat number, and not nucleotide variation in the flanking regions. For each locus, the repeat number was very similar to that indicated in the databases.

Screening of VNTR loci on the total strain panel
The 11 VNTR loci JLR4,LMCEB01,02,03,04,05,06,12,14) were tested on the total test strain panel to evaluate the polymorphism of each VNTR locus. The loci LMCEB01, 03 and 04 exhibited no diversity (Table 5) and were therefore removed from the study. The eight remaining VNTR loci displayed between two and six alleles. Locus JLR4 had the highest diversity.

Comparison of data obtained with conventional electrophoresis and those obtained with CE
Two loci Lm-8 and JLR-4 were tested at Serun Statens Institute on the common panel of 40 strains using CE. Except for two strains, all showed the same repeat number. For Lm-8, one strain from SSI, 20092474, had a real repeat number of 2.7 in CE and 2.4 in agarose gel electrophoresis. For JLR4, one strain from SSI, 20082357, had a real repeat number of 3 in CE and 3.56 in agarose gel electrophoresis.

MLVA stability
The stability of each locus was evaluated to determine the effect of laboratory passage. The copy number was determined to be 100% reproducible (data not shown). Each of the eight loci tested on three reference strains were stable.

MLVA assay applied on the test panel of strains
Based on MLVA results, the 72 isolates were divided into 21 types (Figure 1). MLVA types were clustered into two groups. All the isolates of serotypes 1/2a, 3a, 1/2c, 3c were classified in one group, while all the isolates of serotypes 4b, 1/2b, 3b, 4d were in another group (Figure 1). Nineteen of the 21 types contained isolates of the same serotype ( Figure 1). Type "10" contained isolates of two serotypes 1/2a and 1/2c and the autoagglutinable strain. Type "7" contained isolates of two serotypes 4b and 1/2b and one isolate of the 1/2c serotype (Figure 1).

PFGE data
For PFGE, the two-enzyme combination divided the isolates into 48 distinct profiles ( Figure 2). All the isolates of serotype 1/2a, 3a, 1/2c, 3c were classified in one group, while all the isolates of serotypes 4b, 1/2b, 3b, 4d were in another group. Combined PFGE types contained isolates of the same serotype, except the type "70/25", which contained isolates of serotypes 1/2c and 3c.

MLVA data compared with PFGE data
Six different MLVA types were encountered for nine distinct epidemiological groups. A single ApaI/AscI PFGE type was observed for each epidemiological group (Figure 1).

Discussion
The objective of this work was to evaluate the feasibility of an MLVA scheme coupled with conventional agarose gel electrophoresis for subtyping L. monocytogenes. This type of scheme would be very useful for L. monocytogenes surveillance, because it can be implemented by any molecular laboratory and does not require an expensive capillary electrophoresis system.
Out of the 16 VNTRs published, only eight Lm-3, Lm-8, Lm-10, Lm-11, Lm-15, JLR4, LMV6-JLR and LMV9-JLR were selected here because (1) their repeat length was greater than or equal to 9 bp as demonstrated on a large panel of human and food strains (Sperry et al., 2008;Larson et al. 2010;Lindstedt et al., 2008;Murphy et al., 2007). For six out of eight loci (Lm-3, Lm-10, Lm-11, Lm-15, LMV6-JLR and LMV9-JLR), the size of the amplification products observed on the agarose gels differed between the runs. This result was observed for different primer sets, both previously published and newly designed. This result was surprising, particularly regarding loci Lm-3 and Lm-10, for which Murphy et al. (2007) observed accurate detection on agarose gels. In this study, agarose gel electrophoresis does not appear to be sufficiently accurate for determining repeat number for these six loci. In contrast, agarose gel electrophoresis was suitable for loci Lm-8 and JLR-4. The sizing discrepancies need to be normalized to develop a standardized agarose gel protocol uding all the VNTRs selected here.
For Lm-8, the amplification protocol used here was as similar to that described by Sperry et al. (2008). The repeat number obtained here for the 34 "TS" strains on agarose gel was exactly the same as that obtained on the same panel on a CE Beckman Coulter CEQ 8000 genetic analyzer (Sperry et al., 2008). For Lm-8 and JLR-4, of 39 strains from a panel of 40, the repeat number on agarose gels was exactly the same as that obtained on the ABI 3130 genetic analyzer (Applied Biosystems) at SSI (Larsson et al., 2010). For only one strain, a low difference (maximum 0.56) was observed in the number of base pairs. We demonstrated here that the change in equipment used for the detection of JLR4 and Lm-8 did not affect the determination of repeat number. These data confirm the reliability of these two loci.
However, locus Lm-8 revealed low levels of diversity (2 alleles (2008). This locus overlaps with locus LM-TR2, included in the scheme of Murphy et al. (2007). It had the lowest diversity index in comparison to the five other VNTR loci.
The five VNTR loci found here, LM 02, 05, 06, 12, 14, were identified from the sequenced genomes of three reference strains. They have never been described before. Our results demonstrate that these loci show reliable amplification.
With 71 of 72 strains, our MLVA scheme of eight loci (Lm 02, 05, 06, 12, 14, Lm-8, Lm-26 and JLR-4) confirmed the division of L. monocytogenes strains into two distinct genetic lineages. One strain of the 1/2c serotype showed an MLVA type common to strains of serotype 1/2b and 4b. This strain belonged to molecular serogroup IIc and has a combined PFGE profile specific to 1/2c and IIc strains. Other molecular methods are needed to further investigate the genetic profile of this strain.
These data indicate that the MLVA scheme developed here was less discriminating than ApaI/AscI PFGE. However, the eight VNTR loci selected in this study have proved useful and can be included in a larger MLVA scheme coupled with CE, including VNTR loci with shorter repeat motifs and with higher polymorphism. The more polymorphic loci were excluded from this study, either because they are too short to be visible on agarose gels or because sizing discrepancies were observed on agarose gels. It is absolutely necessary to normalize these sizing discrepancies for accurate and standardized detection on agarose gels. Moreover, in the future, it is necessary to compare all the data obtained in different laboratories and to harmonize VNTR loci and allele naming for a standardized L. monocytogenes MLVA scheme. www.intechopen.com