High Level of Heterogeneity Among Listeria monocytogenes Isolates from Clinical and Food Origin Specimens in Greece

Introduction

L isteria monocytogenes infection (listeriosis) is a serious foodborne infection, which mainly affects the elderly, newborns, pregnant women, and the immunocompromised host. Listeriosis typically manifests itself as septicemia, meningitis, and encephalitis, as well as spontaneous abortion during the second or third trimester, or stillbirth (Allerberger and Wagner, 2010).

Approximately 1,600 listeriosis cases are being recorded annually in the United States (CDC, 2012), of which approximately 16% are fatal; among the immunocompromised hosts, total mortality may range up to 30%. In Europe, the annual detection rate is approximately 0.3–7.5 cases per 100,000 inhabitants of the general population, and approximately 12 cases per 100,000 pregnant women (WHO, 2012). Most of the cases are recorded in countries with established epidemiological and surveillance networks, such as Denmark, Germany, The Netherlands, Belgium, Finland, and the United States. In contrast, in Greece, the detection rate reported at the Hellenic Center for Diseases Control (KEELPNO) is negligible, with an average of about 0.05 cases per 100,000 inhabitants annually (KEELPNO, 2012), thus raising concerns of underreporting leading to a lack of epidemiological data.

Epidemiological surveillance of listeriosis is being performed using serotyping and molecular genotyping. A total of 16 serotypes have been recognized, the prevailing ones among clinical isolates being 4b, 1/2a, and 1/2b (Borucki and Call, 2003). Pulsed-field gel electrophoresis (PFGE) is the reference genotyping method at present, which is a demanding and time-consuming procedure (Swaminathan and Gerner-Smidt, 2007). Other molecular techniques, less demanding, comprise Multiple Locus Variable Number Tandem Repeat Analysis (MLVA) and Random Amplification of Polymorphic DNA (RAPD) (Miya et al., 2008; Murphy et al., 2007).

The aim of the present study is the comparative genotyping of a collection of L. monocytogenes isolates from human and food origin in Greece, using molecular serotyping, RAPD and MLVA techniques.

Methods

During the study period (September 2009 to June 2010) a total of 61 non-duplicate L. monocytogenes isolates were collected from different areas of Greece, comprising 19 isolates from human infections and 42 isolates from food. The clinical isolates were collected from an equal number of cases admitted in eight hospitals in Athens and comprised 10, two, and seven isolates from blood, cerebrospinal fluid (CSF), and other specimens, respectively. The food isolates were collected from contaminated food, as part of the respective internal quality control programs implemented in six food processing industries in Athens. Isolates were incubated on 5% sheep blood and chromogenic Listeria Ottavani & Agosti (ALOA) agar plates (Bioprepare, Keratea, Greece). Species identification was performed using the BBL Crystal Gram-positive identification system (Becton Dickinson, Sparks, MD). The National Collection of Type Cultures (NCTC) 10357 strain was used as a reference index.

DNA extraction, from an overnight culture on 5% sheep blood agar plates, was performed using the NucleoSpin Tissue kit (Macherey-Nagel, GmbH & Co. KG, Dueren, Germany) according to the manufacturer's instructions.

Multiplex polymerase chain reaction (PCR) serotyping was performed according to a previously published protocol (Doumith et al., 2004) in 50 μL of reaction volume, using the 2× GoTaq Hot Start PCR Master Mix (Promega Gmbh, Mannheim, Germany) and an annealing temperature of 53°C. PCR products were separated in a 2% agarose gel, stained with ethidium bromide (0.5 μg/mL), and documented visually under UV illumination.

PCR-RAPD was performed as previously described (Aurora et al., 2009), using primers inlA.F and OPM-01 and GoTaq Hot Start PCR Master Mix (Promega). PCR with the primer inlA.F was performed in a 25-μL reaction volume and an annealing temperature of 32°C, whereas PCR with the primer OPM-01 was performed in a 50-μL reaction volume and an annealing temperature of 30°C. PCR products were separated in a 1% agarose gel, stained with ethidium bromide (0.5 μg/mL), and documented visually under ultraviolet illumination.

MLVA typing was performed on all isolates that were assigned to a common RAPD clone. A previously published protocol (Murphy et al., 2007) was followed, in 50-μL reaction volume using GoTaq Hot Start PCR Master Mix, (Promega) and an annealing temperature of 54°C for all primers except LM-TR 6, for which the temperature was adjusted to 52°C. Amplification products were separated in a 2% agarose gel, stained with ethidium bromide (0.5 μg/mL), and documented visually under ultraviolet illumination.

Results

PCR serotyping results of the 61 L. monocytogenes isolates are presented in Table 1. The predominant serotype group among clinical and food isolates was 4b, 4d, 4e. This was followed by serotype group 1/2a, 3°, which was only found in isolates of food origin.

All L. monocytogenes isolates were typed with RAPD and the method was reproducible, taking into account the results of the 20 isolates that were analyzed in duplicate (data not shown). The Simpson index was 0.978 and 0.997 for primers inlA.F and OMP-1, respectively.

Using the primer inlA.F, a total of 31 isolates (50.8%) were grouped in 10 clusters (I–X), whereas the remaining 30 isolates produced unique fingerprints. Prevalent clusters were I, II, IX, III, VIII, and X, comprising six, four, four, three, three, and three isolates, respectively, whereas the remaining four clusters (IV, V, VI, and VII) included two isolates each. Among the 10 different clusters detected, only clusters VIII, IX, and X included isolates from different serotype groups, whereas all seven remaining clusters comprised isolates from the same serotype group, as shown in Tables 2 and 3.

Using primer OPM-1, a total of nine isolates (14.8%) were grouped into four clusters, whereas the remaining 52 strains produced unique fingerprints. Three of the four clusters (ii, iii, and iv) comprised two isolates each, whereas cluster i included three isolates. All four clusters included isolates of a single serotype group, as shown in Tables 2 and 3.

The nine isolates that were grouped into four clusters with the OMP-1 RAPD protocol were subsequently typed using the MLVA method. Five of the nine isolates produced unique MLVA fingerprints (strains 11, 12, 52, 55, and 58), whereas the remaining four were grouped into two clusters, cluster A (strains 44 and 45) and cluster B (strains 37 and 38), corresponding to clusters iii and iv of the OMP-1 RAPD, respectively (Tables 2 and 3).

Discussion

The present study examined the relatedness of a collection of clinical and food L. monocytogenes isolates in Greece, comparing serotyping and three genotyping procedures, RAPD with two different primer-protocol combinations and MLVA. Although all isolates were obtained during the 10-month study period, the moderate scale of the present study (Athens metropolitan area) did not allow extrapolations and conclusions regarding serotype distribution and seasonality in Greece.

Serotype differences between the two groups were detected, as among clinical isolates an overall prevalence of two specific groups was observed (Table 1), whereas the food isolates presented with a more diverse distribution. These findings are in agreement with previous studies (Borucki and Call, 2003; Coillie et al., 2004; Gilbreth et al., 2005; Gray et al., 2004; Lukinmaa et al., 2004; Lunden et al., 2004; Revazishvilli et al., 2004; Vitas and Garcia-Jalon, 2004), which demonstrated that serovars 1/2a, 1/2b, or 4b prevailed among Listeria strains isolated from epidemics or sporadic clinical cases, serogroup 1/2a, 3a was detected among food isolates, and 75–90% of isolates from food and food processing environments tested belonged to serogroup 1/2. Specific serotypes of L. monocytogenes may persist in the food processing environment, subsequently contaminating the finished products (Lukinmaa et al., 2004; Martinez et al., 2003; Revazishvilli et al., 2004). Nevertheless, the ability of all these serotypes to actually cause infection is questionable, thus explaining the less diverse origin of the human isolates, which was mentioned above (Gray et al., 2004).

In several studies, the information obtained from serotyping has been questioned as being of limited value (Gilbreth et al., 2005; Revazishvilli et al., 2004). In that respect, complementary genotyping has been introduced, for further differentiation. Although a simple technique, RAPD is considered a rapid tool for Listeria typing, provided that its standardization and reproducibility problems are solved (Boerlin et al., 1995; Wagner et al., 1999). In the present study, both RAPD protocols gave 100% reproducible results. In addition, improved discrimination was demonstrated when taking into account the combined results of the two different protocols (D=0.999, from 0.978 and 0.997, using primers inlA.F or OMP-1, respectively). In that respect, RAPD with two different protocols, although doubling the actual work which has to be performed, may lead to greater discrimination and is still easier than PFGE.

MLVA typing method is considered to possess higher discriminatory power than RAPD. However, the data presented here only marginally support this observation, as the discriminatory power of the combined two RAPD protocols was only slightly inferior to MLVA (six isolates were grouped into three pairs using the combined RAPD, in comparison to four isolates grouped into two pairs using the MLVA). One of these pairs (no. 44–45, both isolates from meat products) was identified with both RAPD and MLVA protocols, and serotyping, thus being the only cluster confirmed with all typing techniques.

The combined data presented suggest a high genetic heterogeneity of L. monocytogenes isolates originating from food and clinical specimens in Greece. This local observation is described for the first time in such great detail and is in agreement with a previous study conducted in the United States (Gilbreth et al., 2005), which also indicated that clustering of Listeria isolates, either from clinical specimens or from food, may not be common. In contrast, other studies (Coillie et al., 2004; Lawrence et al., 1993), which were based mainly on RAPD, although they showed heterogeneity among food isolates, also indicated more frequent clustering than that detected in the present study, possibly due to the lower discrimination power of the RAPD protocols used. Nevertheless, the majority of these studies showed that clusters usually comprised either clinical or food isolates, and only in a recent work from Ireland (Fox et al., 2012) were mixed clusters detected, which suggested global distribution of specific PFGE patterns.

In conclusion, this study, using molecular typing techniques, showed that Listeria isolates in Greece are heterogeneous in origin and clustering is very uncommon. In addition, the two RAPD protocols, if used in combination, resulted in comparable discriminative power to MLVA, thus offering an alternative technique for Listeria genotyping.