Typing of Listeria monocytogenes strains isolated in Italy by inlA gene characterization and evaluation of a new cost-effective approach to antisera selection for serotyping



This article is corrected by:

  1. Errata: Corrigendum Volume 109, Issue 1, 368, Article first published online: 26 May 2010

Giancarlo Ripabelli, Department of Health Sciences, University of Molise, via De Sanctis, 86100 Campobasso, Italy. E-mail: ripab@unimol.it


Aims:  In this study, 105 Listeria monocytogenes strains isolated from humans, foods and environmental samples were characterized using several typing methods. Moreover, serotyping procedure was evaluated, and a cost-effective methodological approach based on preliminary PCRs screening was proposed.

Methods and Results:  The isolates were analysed by conventional serotyping, multiplex-PCRs for serogroup and lineage identification and PCR–RFLP of inlA gene to identify potentially noninvasive L. monocytogenes. Among the strains, only the serotypes 1/2a, 1/2c, 1/2b, 4b and 3a were identified. The isolates were classified into serogroups I (58·10%), II (22·85%), III (12·38%) and IV (6·67%). Among clinical strains, lineage I was more represented (68·75%) than lineage II; whereas, lineage II was more associated with food (90·24%) and environmental (85·72%) isolates. Most of food (89·02%) and environmental (85·71%) isolates were classified into truncated InlA profiles, whereas the 93·75% of clinical strains were associated with a complete form of the protein.

Conclusion:  Molecular techniques were sensitive and specific for classifying strains into serogroup and lineage and in agreement with the serotyping. Moreover, a preliminary PCRs-based screening was proposed to select only the necessary antisera by a flow chart; this methodological approach allows cost saving up to 42%. Our results further suggest the role of InlA protein in human listeriosis, particularly in immunocompetent individuals, and a correlation between truncated protein and serotype.

Significance and Impact of the Study:  This study further validates molecular methods for L. monocytogenes analysis and proposed a new cost-effective approach for serotyping. It could help to improve a national surveillance network for L. monocytogenes infections in Italy.


Listeria monocytogenes is a facultative intracellular human pathogen, widely distributed in the environment, which causes a foodborne illness known as listeriosis, a serious health concern because of the severity of the disease, its high mortality rate and the opportunistic nature of the infection (Wagner and Allerberger 2003; Handa-Miya et al. 2007). Both sporadic cases and common source listeriosis outbreaks are associated with the consumption of contaminated foods (Rocourt et al. 2000; Schlech 2000; Longhi et al. 2003; McLauchlin et al. 2004; Cabedo et al. 2008). The clinical syndromes can occur as a febrile gastroenteritis (Rocourt et al. 2000; Carrique-Mas et al. 2003) or an invasive form, mainly in susceptible individuals (immunocompromized, elderly, pregnant women and infants), characterized by sepsis, meningitis, meningoencephalitis and abortions (Vazquez-Boland et al. 2001). Because of the nature of the pathogen, the identification of the infection source is often difficult (Borucki et al. 2005). The availability of rapid, specific and sensitive diagnostic tests to distinguish L. monocytogenes among Listeria species is an essential tool for disease control, and the development of subtyping procedures is critical for the epidemiological investigation of outbreaks. Listeria somatic (O) antigens have been separated into 15 subtypes (I–XV) and flagellar (H) antigens into four subtypes (A–D). Upon analysis of group-specific Listeria O and H antigens, at least 13 serotypes (1/2a, 1/2b, 1/2c, 3a, 3b, 3c, 4a, 4ab, 4b, 4c, 4d, 4e, 7) are recognized in L. monocytogenes (Kathariou 2002). However, it has been observed that serotypes 1/2a, 1/2b, 1/2c and 4b are responsible for 95–98% of listeriosis cases (Liu 2006). Among these, serotype 4b strains are isolated mostly from epidemic outbreaks (Raybourne 2002; Borucki and Call 2003), whereas 1/2a, 1/2b and 1/2c are linked to sporadic infections and are the most prevalent serotypes in foods and food production environments (Wiedmann et al. 1996). The conventional serological test has become widely used, because it is relatively convenient to perform, but lack of specificity and sensitivity and its low discriminatory ability do not always provide sufficient information in epidemiological investigations (Borucki and Call 2003; Doumith et al. 2004). Recently, molecular methods have been applied for L. monocytogenes typing, as well as two multiplex-PCRs to identify serogroups (Doumith et al. 2004) and phylogenetic lineages (Ward et al. 2004).

Several studies have demonstrated L. monocytogenes virulence in vitro (Roche et al. 2003; Roberts et al. 2005), and a higher mortality rate was shown in patients infected with serogroup 4 strains in comparison with those infected with serogroup 1/2. The molecular basis of this increased virulence is still unknown (Swaminathan and Gerner-Smidt 2007), but it has been reported that internalin, a surface protein associated with the bacterial internalization in the gastrointestinal epithelium and translocation through the foeto–placental barrier (Jonquières et al. 1998; Lecuit et al. 2001), is constantly expressed in full length in serogroup 4 whereas not always in serogroup 1/2 strains (Jacquet et al. 2004). Recently, it has been reported that the inlA gene may contain different nonsense mutations, leading to the expression of a truncated nonfunctional form of internalin (Handa-Miya et al. 2007). The detection of mutations in some L. monocytogenes strains is related to a significant virulence reduction and invasivity in Caco-2 cells (Olier et al. 2002, 2005; Rousseaux et al. 2004). Truncated forms of InlA protein were reported to be mostly distributed in food isolates and less in clinical strains (Jacquet et al. 2004; Nightingale et al. 2005), indicating the critical role of the protein in human listeriosis. Thus, detection of truncated protein could be a new useful tool for risk assessment of contaminated foods.

The purpose of this study was to characterize L. monocytogenes isolates from clinical cases, foods and the environment using several typing methods: serotype detection by traditional agglutination; serogroup identification and lineage designation by two multiplex-PCRs; identification of potentially noninvasive and invasive isolates through PCR–RFLP of a selected fragment of inlA gene. Moreover, serotyping procedure was evaluated, and a cost-effective methodological approach based on preliminary PCR results was proposed.

Materials and methods

Listeria monocytogenes strains and DNA extraction

A total of 105 L. monocytogenes strains, isolated in the north-west Italy from clinical cases (n = 16; median age 63), food products (n = 13 gorgonzola cheese; n = 39 taleggio cheese, n = 30 meat for sausages and salami) and food industry surfaces environmental swabs (n = 7) were analysed using different typing methods. The main risk factors for the patients were iatrogenic (n = 4 isolates) and HIV-related immunocompromission (n = 1 isolate), underlying diseases such as neoplasia, diabetes and renal failure (n = 3) and pregnancy (n = 2). However, for six isolates were not available information regarding the risk factors. Listeria monocytogenes strains were generally isolated from blood cultures, liquor, throat swab and central venous catheter. All isolates were cultured in tryptic soy agar (TSA) (Biolife, Milan, Italy) at 37°C overnight. DNA extraction from growth harvested from TSA was carried out following the Boom method, with some minor modifications (Boom et al. 1990; Ripabelli et al. 2000). Genomic DNA extraction was resuspended in 100 μl and stored at −20°C until use. All the experiments were performed blindly and in duplicate.

Listeria monocytogenes serotypes detection

Serotyping of L. monocytogenes, based on antibodies that specifically react with somatic (O) and flagellar (H) antigens, was performed using commercial Listeria antisera (Denka Seiken, Tokyo, Japan), according to manufacturer’s instructions.

Multiplex-PCR for serogroup identification

A multiplex-PCR assay was carried out to separate the major Lmonocytogenes serovars (1/2a, 1/2b, 1/2c and 4b) into distinct serogroups. The marker genes selected, according to a previous study (Doumith et al. 2004), were lmo0737, lmo1118, ORF2819 and ORF2110. The prs gene, specific for Listeria spp., was used as an internal amplification control. Target genes, primer sequences and PCR products size are listed in Table 1. Briefly, 2 μl of DNA extraction was used for the amplification; each reaction was performed in a 50 μl final volume containing 25 μl of PCR Master Mix 1× (Promega, Milan, Italy) and five primer sets. Primers concentration and amplification conditions used were in agreement with Doumith et al. (2004). Five microlitres of the PCR products was separated on a 2% agarose gel (Eppendorf, Milan, Italy) and visualized on a transilluminator after ethidium bromide staining.

Table 1.   Primers used for serogroup and lineage detection in Listeria monocytogenes
Multiplex-PCR for serogroup detectionMultiplex-PCR for lineage detection
GenesNucleotide sequences (amplicon size-bp)Serovars (serogroup)GenesNucleotide sequences (amplicon size-bp)Serovars (lineage)
prsprs fwd: GCTGAAGAGATTGCGAAAGAAGAll Listeria spp.actAactA1 fwd: AATAACAACAGTGAACAAAGC1/2b, 3b, 4b, 4d, 4e (I)
lmo0737lmo0737 fwd: AGGGCTTCAAGGACTTACCC1/2a, 1/2c, 3a, 3c (I)plcBplcB2 fwd: TTGTGATGAATACTTACAAAC1/2a, 3a, 1/2c, 3c (II)
lmo1118lmo1118 fwd: AGGGGTCTTAAATCCTGGAA1/2c, 3c (II)actA–plcBactA3 fwd: CGGCGAACCATACAACAT4a, 4c (III)
ORF2819ORF2819 fwd: AGCAAAATGCCAAAACTCGT1/2b, 3b, 4b, 4d, 4e (III)   
ORF2110ORF2110 fwd: AGTGGACAATTGATTGGTGAA4b, 4d, 4e (IV)   

Multiplex-PCR for lineage identification

To confirm serotype distribution across lineages, all isolates were classified as one of the three phylogenetic divisions, using a multiplex-PCR as previously described by Ward et al. (2004). Three different sets of primers (actA1, plcB2 and actA3-plcB3) were used (Table 1). PCRs were performed at a final volume of 50 μl consisting of 2 μl of DNA, 25 μl of PCR Master Mix 1× (Promega) and 0·7 μmol l−1 of each primer. Samples were amplified as follows: 94°C for 2 min, followed by 35 cycles of 94°C for 1 min, 54°C for 1 min, 72°C for 1 min with a final extension at 72°C for 5 min. PCR products were resolved on 1·5% agarose gel (Eppendorf), stained with ethidium bromide and visualized on a transilluminator.

Evaluation of serotyping and multiplex-PCR costs

Costs for serotype identification, DNA extraction and amplification were evaluated according to the Italian market reagents prices, to propose a new cost-effective approach for antisera selection based on multiplex-PCRs screening. Costs for microbiological cultivation and staff salary were not estimated.

PCR–RFLP of inlA gene

A PCR–restriction fragment length polymorphism (RFLP) method, based on inlA polymorphisms, was used as previously described by Rousseaux et al. (2004), to identify potentially noninvasive L. monocytogenes strains. A defined inlA fragment (733 bp) was amplified, using seq01 and seq02 primers (Rousseaux et al. 2004). Amplifications were performed with 50 μl volume of reaction mixture containing: 25 μl of PCR Master Mix 1× (Promega), each primer at a concentration of 0·5 μmol l−1, 2 μl of DNA as template. Samples were amplified as follows: 94°C for 4 min, followed by 30 cycles of 94°C for 30 s, 52°C for 1 min, 72°C for 2·5 min, with a final extension at 72°C for 7 min. Two restriction endonucleases AluI (Fermentas, Milan, Italy) and Tsp509I (Fermentas) were used separately for PCR products digestion with an overnight incubation at 37 and 65°C, respectively. PCR–RFLP fragments were separated by standard electrophoresis on a 3·5% agarose gel (Eppendorf). Gels were stained with ethidium bromide and visualized on a transilluminator. Restriction profiles images were recorded for visual analysis by GelDoc System (Bio-Rad, Milan, Italy).


Listeria monocytogenes conventional serotyping

All 105 L. monocytogenes strains belonged to four distinct serotypes, as indicated in Table 2. Of the 16 human isolates, 37·5% (n = 6) were serotype 4b, 31·25% (n = 5) belonged to serotype 1/2b and 31·25% (n = 5) to serotype 1/2a. The majority of food isolates (n = 50) belonged to serotype 1/2a (60·97%), whereas 25·61% (n = 21) were identified as serotype 1/2c, 9·76% (n = 8) as serotype 1/2b and 3·66% (n = 3) as serotype 3a. Among the strains isolated from the environment, serotypes 1/2a and 1/2c were identified in three (42·86%) isolates, respectively, and only one isolate was classified as serotype 4b (14·28%).

Table 2.   Results of serotyping and PCR-based techniques for serogroup and lineage detection in Listeria monocytogenes isolates
Strains originConventional serotypingMultiplex-PCR for serogroup detectionMultiplex-PCR for lineage detection
SerotypesSerogroups*Phylogenetic lineages†
1/2a1/2b4b1/2c3aGroup IGroup IIGroup IIIGroup IVLineage ILineage II
  1. *Group I: 1/2a, 3a; Group II: 1/2c, 3c; Group III: 1/2b, 3b; Group IV: 4b, 4d, 4e.

  2. †Lineage I: 1/2b, 3b, 4b, 4d, 4e; Lineage II: 1/2a, 1/2c, 3a, 3c.

  3. ‡Food isolates: A36 from taleggio cheese, 8 from gorgonzola cheese, 6 from meat products; B3 from taleggio, 3 from meat products, 2 from gorgonzola; Call from meat products; Dall from gorgonzola.

  4. §Human isolates: E3 from patients with underlying diseases, 1 from patient with iatrogenic immunosuppression, 1 from patient with unknown risk factors; F1 from HIV-positive patient, 4 from patients with unknown risk factors; G3 from patients with iatrogenic immunosuppression, 2 from pregnant women and 1 from patient with unknown risk factors.

Foods (= 82)‡ (%)50A (60·97)8B (9·76)0 (0)21C (25·61)3D (3·66)53A, D (64·63)21C (25·61) 8B (9·76)0 (0) 8B (9·76)74A, C, D (90·24)
Humans (= 16)§ (%) 5E (31·25)5F (31·25)6G (37·50) 0 (0)0 (0)5E (31·25)0 (0) 5F (31·25)6G (37·50)11F, G (68·75) 5E (31·25)
Environment (= 7) (%) 3 (42·86)0 (0)1 (14·28) 3 (42·86)0 (0)3 (42·86)3 (42·86) 0 (0)1 (14·28)1 (14·28) 6 (85·72)
Total (n = 105) (%)58 (55·23)13 (12·38)7 (6·67)24 (22·86)3 (2·86)61 (58·10)24 (22·85)13 (12·38)7 (6·67)20 (19·05)85 (80·95)

Multiplex-PCR for serogroup identification

By multiplex-PCR, all L. monocytogenes isolates were classified into four major serogroups, namely I, II, III and IV, using a combination of five primer sets (Table 1). The prs gene fragment, used as a positive PCR control to confirm Listeria spp. amplification, was detected in all strains. Among the isolates, 61 (58·10%) were recognized as belonging to serogroup I (Table 2), which includes serotypes 1/2a and 3a; 24 (22·85%) were grouped into serogroup II, with serotypes 1/2c, 3c; 13 (12·38%) were classified into the group III, which includes serotypes 1/2b, 3b, and seven (6·67%) into the group IV, consisting of serotypes 4b, 4d, 4e. Moreover, isolates classified as serogroup II were further identified by a single-PCR to confirm the amplification of lmo1118 gene, because the multiplex-PCR showed reproducibility problems when performed in duplicate. Most of food isolates (64·63%; 53/82) were serogroup I; while 21 strains (25·61%) were included in the group II; and eight (9·76%) into group III. Among human strains, 6/16 (37·5%) were classified as serogroup IV; five (31·25%) belonged to group I; and five (31·25%) to group III (Table 2). Among environmental strains, 3/7 (42·86%) belonged to serogroup I; three (42·86%) to group II; and one (14·28%) to group IV.

Multiplex-PCR for lineage identification

Additional PCR-based analyses were performed to identify the corresponding serotypes 1/2a, 1/2b, 1/2c and 4b into the major genetic divisions. The multiplex-PCR produced a single amplicon of the correct size for each target gene. The majority of strains (85/105; 80·95%) were grouped into the lineage II (Table 2) consisting of the serotypes 1/2a, 1/2c and 3a; whereas 20/105 (19·05%) were identified as belonging to lineage I including the most important serotypes 1/2b and 4b. Among human strains, 11/16 (68·75%) were identified into the lineage I, while only five (31·25%) were classified as belonging to lineage II. By analysing the 82 food isolates, eight (9·76%) were identified as belonging to lineage I, whereas most (74/82; 90·24%) belonged to lineage II. Among environmental strains, 6/7 (85·72%) were identified into genetic division II, whereas only one isolate (14·28%) was classified into the lineage I (Table 2).

Evaluation of serotyping and multiplex-PCRs costs

The PCR-based screening methods allowed the identification of well-defined clusters with specific serotype distribution. Hence, when PCR-based screening for serogroup identification was applied, on the basis of groups detected was possible to select a subset of specific antisera for serotyping (Fig. 1a), which allowed cost savings per reaction (including PCR amplification and necessary antisera) of between 32% and 39% for serogroups I and II; 15% and 24% for serogroup III; and 20% and 37% for serogroup IV. Furthermore, by using a PCR-based approach to classify the strains into genetic lineage (Fig. 1b), it was possible to make cost savings of between 18% and 22% for isolates belonging to lineage I; 35% and 42% for lineage II; and 30% and 39% for lineage III.

Figure 1.

 Flow charts for Listeria monocytogenes serotyping according to Denka Seiken antisera kit following multiplex-PCR-based screening for serogroup (a) and lineage (b) identification.

PCR–RFLP of inlA gene

A specific fragment of the inlA gene (733 bp) was amplified in all L. monocytogenes strains using the seq01 and seq02 primers. The restriction of this fragment using AluI generated five different profiles (Table 3) according to Rousseaux et al. (2004). Among human strains, 8/16 (50%) were grouped into the profile 2, five (31·25%) into the profile 3, two (12·50%) into the profile 5 and one isolate (6·25%) into the profile 4. Among food isolates, 68/82 (82·93%) were associated with the profile 4, seven (8·53%) were grouped into the profile 2, five (6·10%) into the profile 1 and two (2·44%) were referred to the profile 3. None of the isolates was classified into profile 5. Among environmental strains, 3/7 (42·86%) were identified into the profile 1, three (42·86%) into the profile 4 and one isolate (14·28%) into the profile 2.

Table 3.   PCR–RFLP profiles of Listeria monocytogenes inlA gene
Strains originAluI restriction pattern profilesTsp509I restriction pattern profilesComposite profiles AluI–Tsp509I*
Profile 1Profile 2Profile 3Profile 4Profile 5Profile 1Profile 2Profile 3ABCDENot defined
  1. *Composite profiles: A (1 AluI–3 Tsp509I) and D (4 AluI–2 Tsp509I): truncated InlA protein; B (2 AluI–1 Tsp509I), C (3 AluI–2 Tsp509I) and E (5 AluI–3 Tsp509I): complete InlA protein.

Foods (n = 82) (%)5 (6·10)7 (8·53)2 (2·44)68 (82·93)0 (0)7 (8·53)68 (94·44)7 (8·53)5 (6·10)7 (8·53)0 (0)68 (82·93)0 (0)2 (2·44)
Humans (n = 16) (%)0 (0)8 (50·00)5 (31·25)1 (6·25)2 (12·50)8 (50·00)6 (37·50)2 (12·50)0 (0)8 (50·00)5 (31·25)1 (6·25)2 (12·50)0 (0)
Environment (n = 7) (%)3 (42·86)1 (14·28)0 (0)3 (42·86)0 (0)1 (14·28)3 (42·86)3 (42·86)3 (42·86)1 (14·28)0 (0)3 (42·86)0 (0)0 (0)
Total (n = 105) (%)8 (7·62)16 (15·24)7 (6·67)72 (68·57)2 (1·90)16 (15·24)77 (73·33)12 (11·43)8 (7·62)16 (15·24)5 (4·77)72 (68·57)2 (1·90)2 (1·90)

Restriction by Tsp509I enzyme allowed to differentiate three profiles (Table 3) as previously reported by Rousseaux et al. (2004). Among human strains, the profile 1 was identified in 8/16 (50%) isolates, the profile 2 in six strains (37·5%) and the profile 3 in two strains (12·5%). Most of food isolates (68/82; 94·44%) belonged to the profile 2, seven (8·53%) were identified with the profile 1 and seven (8·53%) showed a profile 3. Among environmental strains, 3/7 (42·86%) were classified into the profile 2, three (42·86%) into the profile 3 and one isolate (14·28%) into the profile 1.

Five distinct composite profiles were detected by the combination of profiles obtained separately using the two enzymes (Table 3). Among human isolates, 8/16 (50%) were characterized as profile B and identified as serotypes 1/2b (4/16) and 4b (4/16); five (31·25%) as profile C, serotype 1/2a; two (12·5%) as profile E, serotype 4b; and only one strain (serotype 1/2b) was identified as profile D. No human strains belonged to profile A. Moreover, 43·75% of human isolates had profile C or E which were not identified among isolates from foods or the environment. The majority of food isolates (68/82; 82·93%) have been grouped into the profile D, including gorgonzola and taleggio cheeses and meat products isolates. Five strains (6·1%) showed the composite profile A, identified only in strains of serotype 1/2a isolated from meat products, and seven (8·53%) belonged to the profile B, identified in one strain isolated from gorgonzola, three from taleggio cheese and three from meat products, all belonging to serotype 1/2b. Interestingly, none of the isolates belonged to profile C or E, while profile B, which was identified in 50·0% of human strains was found only in 8·53% of food isolates. However, for two (2·44%) strains, isolates from gorgonzola cheese, an undefined composite profile was detected, because the combination of 3-AluI and 3-Tsp509I to our knowledge was never reported in literature. Finally, 3/7 (42·86%) environmental strains of serotype 1/2a were classified as profile A, three (42·86%) as profile D, all identified as serotype 1/2c, and only one isolate of serotype 4b (14·28%) was classified as profile B.


Listeria monocytogenes is an opportunistic pathogen which has become a serious cause of human foodborne infections worldwide. Because of its close relationship to other Listeria species, the availability of sensitive tests for its differentiation is helpful for tracking strains involved in listeriosis outbreaks and for the prevention of further disease occurrences. Serological typing, through classical agglutination-based method, is widely used to characterize L. monocytogenes for epidemiological purposes and for surveillance (Wiedmann 2002; Wagner and Allerberger 2003). However, the application of molecular methods is helpful to differentiate micro-organism in separate serogroups, including the principal serotypes (1/2a, 1/2b, 1/2c and 4b), overcoming the need to apply phenotypic and biotyping techniques. Molecular techniques also allowed defining the phylogenetic divisions composition, are useful to understand ecology, genetic diversity and host specificity of L. monocytogenes isolates as well as to determine the distribution of serotypes within a specific lineage (Ward et al. 2004).

Among 105 L. monocytogenes strains tested, high percentages of isolates belonging to the major serotypes (1/2a, 4b, 1/2b and 1/2c) were detected, in agreement with other authors (Kathariou 2002). Serotypes 1/2a and 3a showed the highest and lowest prevalence, respectively. Indeed, serotype 3a is relatively rare in human infections. Serotype 1/2a was the most frequently isolated from foods, followed by serotypes 1/2c, 1/2b and 3a, indicating that contaminated foods may serve as vehicles to transmit potentially virulent L. monocytogenes to humans. In comparison, most human strains were identified as serotype 4b, which cause the majority of human epidemic cases, suggesting that serotype designation is strongly associated with virulence potential (Borucki and Call 2003). However, to confirm the potential pathogenetic capacity of isolates, the presence of specific virulence markers should be always evaluated by bio-molecular methods.

Serological identification of Lmonocytogenes by flagellar and somatic antigenic profiles is often the first step to type isolates for public health and microbiological surveillance of human listeriosis, particularly if molecular expertise is not available. For food industries, where the presence of L. monocytogenes is an important health concern, tracking contaminating strains within the food chain, and the plant environment is of primary importance. However, conventional serotyping may not always be successful because of cross-reactions with other organisms, of a relatively low discriminating power compared to molecular subtyping methods, and then of limited value for epidemiological investigation (Doumith et al. 2004; Ward et al. 2004). Moreover, there are also limitations attributed to the high costs of antisera, the need for technical expertise to perform the assay, as well as the low reproducibility, not always satisfactory for the occasional discrepant results because of the dependence on phenotypic characteristics. Thus, PCR-based techniques constitute a rapid, practical and effective alternative choice to traditional seroagglutination method for differentiating L. monocytogenes isolates (Autio et al. 2003; Martinez et al. 2003), because of greater and intrinsic specificity and sensitivity. However, for definitive serotype identification, the classical serological procedure is still necessary, because molecular markers for recognizing individual serotypes have not yet been identified.

Several studies have also showed that L. monocytogenes comprises at least three primary evolutionary divisions or separate lineages (Ward et al. 2004). Direct correlations between the three lineages and the most common serotypes have been reported (Nadon et al. 2001). A multiplex-PCR previously described by Ward et al. (2004) was used to classify each strain into a specific lineage, to estimate the frequency of the three lineages in clinical, foods and environmental strains, and to evaluate the correlation with serotype identification. Only two distinct phylogenetic lineages, I and II, which represent the most common L. monocytogenes genetic divisions, and generally correspond to clinical and food isolates, were identified. Among the strains isolated from human cases, lineage I was more represented than lineage II; although lineage I includes isolates which caused a large number of human listeriosis cases, it is still unclear whether this reflects an enhanced virulence for humans or ecological adaptations (Ward et al. 2004). Moreover, lineage II was found to be strongly associated with food rather than human isolates, confirming that lineage II strains are found mostly in foods (Hain et al. 2006). Among environmental strains too, the prevalence of lineage II was higher than lineage I for which only one isolate was detected. Although serotypes distribution in environmental strains appeared to be similar to that of food isolates, the limited number of strains analysed in this study does not allow any accurate conclusion with respect to the structure of environmental L. monocytogenes population. None of the strains analysed belonged to lineage III, which is rarely associated with human listeriosis, and most prevalent among animal isolates suggesting a limited virulence potential in humans. The absence of lineage III among human, food and environmental isolates may be explained by better adaptation of these strains to the animal production environment as opposed to the food-processing environment. Thus, indicating less capacity to survive in food production processes than isolates belonging to lineages I and II (Jeffers et al. 2001).

Concordance among lineage identification analysis, serogroup detection and traditional serotyping results was observed. Phylogenetic groups identification test proved to be 100% sensitive and specific in accurately assessing the lineage for all Lmonocytogenes isolates and in agreement with both PCR-based serogroup assay and classical serological test. This method has shown the potential to provide significant insights regarding the population genetics, ecology, epidemiology and biological significance of the lineages. The specific association of genetic lineage with serotype commonly isolated in human listeriosis (4b, 1/2b, 1/2a and 1/2c) confirmed that Lmonocytogenes lineages differ in their pathogenic potential and host specificity (Jeffers et al. 2001).

Our results confirm that the PCR-based approaches are sensitive, reproducible and in perfect concordance with serotyping analysis, thus providing a helpful contribute to correctly detect and confirm the serotype. Moreover, the PCR-based screening allowed the identification of specific serotype distributions which provide invaluable information to overcome the use of all the antisera. In fact, according to the results, any researchers can autonomously follow an appropriate flow chart, cost savings of 42%, with an average of about 30%. This novel typing strategy, which does not increase considerably the laboratory work, is not time consuming and provides an independent control for serotyping analysis, could further improve the availability and the quality of serotypes-based epidemiological data, contributing to a better understanding of L. monocytogenes circulation and control strategies.

Recently, the internalin A has been considered as molecular marker to assess the virulence properties of L. monocytogenes strains, based on the functionality of protein (full length or truncated form) (Jacquet et al. 2004). Virulent and invasive isolates produce InlA of 80 kDa; however, through inlA gene sequencing, several distinct ‘nonsense mutations’ were identified (Olier et al. 2002; Jacquet et al. 2004). Moreover, it has been further demonstrated that these polymorphisms were associated with phenotypes characterized by a reduced invasiveness in human intestinal epithelial cells because of the expression of a truncated InlA (Jonquières et al. 1998; Olier et al. 2003; Nightingale et al. 2005). The attenuated virulence of reference strains resulted in a reduced invasive capacity in Caco-2 cells (Jonquières et al. 1998; Olier et al. 2002, 2003; Rousseaux et al. 2004), confirming the occurrence of nonfunctional forms of InlA, especially in strains belonging to serotypes 1/2a and 1/2c (Jacquet et al. 2004). It has been well documented that a full length of internalin A is overexpressed in clinical isolates than in food isolates (Jacquet et al. 2004), whereas truncated forms are expressed most frequently in food strains. In our study, a high percentage of human strains showed a restriction profile associated with complete InlA, in agreement with Handa-Miya et al. (2007) and Jacquet et al. (2004). These strains were mainly serotypes 4b, followed by serotypes 1/2b and 1/2a. One serotype 1/2b strain, which showed a truncated form of internalin, was isolated from an HIV-positive patient, important risk factor for L. monocytogenes infections. All food strains showed a theoretical nonfunctional form of InlA protein, except seven of these, classified as serotype 1/2b. This finding is consistent with other reports that have also shown that truncation of this protein is not a rare event among food isolates (Jonquières et al. 1998; Nightingale et al. 2005), while it is occasional in human strains. Hence, the prevalence of full length of InlA was highly associated with clinical strains, providing epidemiological evidence in support of the important role of the protein in human listeriosis, and suggesting that the detection of internalin functionality may have a potential use as a biomarker to identify virulent strains for an improved risk assessment. The finding that profile B was identified in 50·0% of human strains and only in 8·53% of food isolates does not represent an incongruence. Probably, the strains have been selected among human isolates by their own capability to cause disease. Conversely, foods may be contaminated with many different potentially nonpathogenic L. monocytogenes strains. Moreover, 43·75% of human isolates had profiles C or E which were not identified among food or environmental isolates, suggesting potential different sources of human infection. Furthermore, among food strains it could be suggested the hypothesis of a relationship between truncated InlA and serotype identity, because the majority of strains with hypothetical truncated form were previously identified as 1/2a, 1/2c or 3a, while strains 1/2b showed a genetic profile referred to a complete protein. This finding was, however, in disagreement with Rousseaux et al. (2004), who did not report correlation between InlA protein structures and the serotype of strains. Moreover, for two isolates from food (gorgonzola cheese), an identical composite profile that to our knowledge has not been defined previously in the literature, was detected; thus, because of food origins of the strains, we hypothesize the occurrence of a new restriction profile associated with the truncated protein. However, any hypothesis will be confirmed by sequencing to assess the presence of premature stop codon mutation in the inlA gene. Among the environmental isolates, two profiles associated with truncated InlA were identified in serotypes 1/2a and 1/2c isolates, and one strain, belonging to serotype 4b, showed a genetic profile related to an entire protein. Hence, the occurrence of complete or truncated InlA could be related to serotype identity also in environmental strains. However, because of the limited number of strains analysed, further experiments are necessary to better evaluate all these relations.

In conclusion, phenotypic methods based on traditional serotyping and DNA-based techniques have been largely utilized for L. monocytogenes characterization. PCR-based methods have been developed for serogroups identification and lineage designation, and these, although did not match yet the final discriminatory power of the serotyping, proved to be a useful and convenient preliminary step to conventional antisera-based slide agglutination typing. With respect to pathogenetic potential of the strains analysed in this study, to further confirm the presence of inlA gene coding for full length or truncated protein, could be useful the evaluation of the in vitro invasion capability in human cells, such as Caco-2 and HepG2 cells.


This work was partially supported by Italian Ministry of University, PRIN 2005, Research Unit no. 2005067184-005. We thank the other Research Units participating to the PRIN project that provided part of the isolates for this study.