Abstract
Rabbit meat has outstanding dietetic and nutritional properties. However, few data on microbiological hazards associated with rabbit productions are available. In this study, the presence of Listeria monocytogenes was determined in 430 rabbit carcasses, 256 rabbit meat cuts and products, and 599 environmental sponges collected from four Italian rabbit slaughterhouses over a period of 1 year. Prevalence of L. monocytogenes among the 1285 rabbit meat and environmental samples was 11%, with statistically significant differences between slaughterhouses. The highest prevalence (33.6%) was observed in rabbit meat cuts and products; the majority of positive environmental samples were collected from conveyor belts. Overall, 27.9% and 14.3% of rabbit cuts and carcasses, respectively, had L. monocytogenes counts higher than 1 colony-forming unit (CFU)/10 g. A selection of 123 isolates from positive samples was genotyped and serotyped to determine genetic profiles and diversity among L. monocytogenes isolates contaminating different slaughterhouses and classes of products investigated. Discriminatory power and concordance among the results obtained using multilocus variable-number tandem-repeat analysis (MLVA), multilocus sequence typing (MLST), pulsed-field gel electrophoresis (PFGE), automated EcoRI ribotyping, and serotyping were assessed. The isolates selected for typing were classified into serotypes 1/2a (52.8%), 1/2c (32.5%), and 1/2b (14.6%). The majority of the isolates were classified as ST14 (34.1%), ST9 (35.5%), ST121 (17.9%), and ST224 (14.6%). The greatest discriminatory power was observed with the MLVA typing, followed by MLST, PFGE, and ribotyping. The best bidirectional concordance was achieved between PFGE and MLST. There was 100% correlation between both MLST and MLVA with serotype. Moreover, a high unidirectional correspondence was observed between MLVA and both MLST and PFGE, as well as between PFGE and both MLST and serotyping. The results of this study show for the first time in Italy prevalence and genetic profiles of L. monocytogenes isolated in rabbit products and slaughterhouses.
Introduction
R
L. monocytogenes is widespread in nature. It can survive and grow over a wide range of environmental conditions, including refrigeration temperatures. Even if meat products are cooked before consumption, L. monocytogenes can survive and grow during storage, posing cross-contamination issues in the kitchens of consumers. Differences in virulence between L. monocytogenes strains may influence infection and clinical outcome. Therefore, it is important to investigate which L. monocytogenes strains contaminate meat products. The aims of this study were to investigate L. monocytogenes prevalence in rabbit carcasses, rabbit meat products, and environmental sites collected within four rabbit slaughterhouses in the Veneto region of Italy over a 1-year period. Moreover, genetic profiles and diversity of the L. monocytogenes isolates were assessed by multilocus variable-number tandem-repeat analysis (MLVA), multilocus sequence typing (MLST), pulsed-field gel electrophoresis (PFGE), automated EcoRI ribotyping, and serotyping. The discriminatory power of serotyping and genotyping methods to differentiate L. monocytogenes isolates was quantified using the Simpson's discriminatory index (D), as described by Hunter and Gaston (1988). Moreover, the bidirectional and unidirectional concordance between serotyping and typing results was calculated by the adjusted Rand (AR) coefficient (Carriço et al., 2006) and the adjusted Wallace (AW) coefficient (Severiano et al., 2011) with corresponding confidence intervals (CIs).
Materials and Methods
Sampling procedure and bacteriological analysis
The presence of L. monocytogenes was determined in 1285 samples collected, as part of autocontrol programs, from four rabbit slaughterhouses, labeled as A to D, located in the Veneto region, in Italy. Overall, 92 samplings were performed between November 2005 and December 2006. Samples were represented by 430 rabbit carcasses, 256 rabbit meat cuts and products, as well as 599 environmental sponges. Rabbit carcasses were randomly sampled at the slaughterhouses, at the end of processing, after the air cooling operations. Rabbit meat cuts and products of 250 g were randomly collected at the slaughterhouses, within the cutting areas. The environmental samples were sterile sponges premoistened with 1% peptone water, containing 0.85% NaCl, before sampling 100 cm2 of different sites. All samples and sponges were kept refrigerated from sampling until the arrival in the laboratory, where they were immediately analyzed for the presence of L. monocytogenes according to the standard ISO procedure 11290-1 (Anonymous, 2004) using Agar Listeria Ottaviani Agosti (ALOA), incubated at 37°C for 24–48 h. In rabbit carcasses, as well as meat cuts and products, L. monocytogenes was also enumerated, according to the ISO procedure 11290-2 (Anonymous, 2004).
Isolate identification, serotyping, and genotyping
Up to five colonies isolated on ALOA plates, showing the expected L. monocytogenes morphology, were purified through three serial cultivation steps on Brain Heart Infusion agar (Oxoid) plates, incubated at 37°C for 24 h. One purified isolate from a selection of positive samples was subjected to automated EcoRI ribotyping in the RiboPrinter® (Bruce, 1996), as previously described (De Cesare et al., 2001; Manfreda et al., 2005). The isolates confirmed as L. monocytogenes, by ribotyping, were subjected to DNA extraction, serotyping, and PFGE. The genomic DNA was extracted using the DNeasy® blood and tissue kit (Qiagen GmbH, Hilden, Germany). Serotyping was performed using the DENKA-SEIKEN kit (Tokyo, Japan) (Ueda et al., 2002). PFGE was executed according to the PulseNet protocol, with ApaI (Graves and Swaminathan, 2001). The DNA extracted from each isolate was subjected to MLVA and MLST. A panel of five variable-number tandem repeat loci was subjected to analysis (Lindstedt et al., 2008) and allele numbers were assigned according to Pasteur Institute MLVA database. Sequence types (STs) were identified by polymerase chain reaction amplification using previously described primers, with the exception of the ldh gene, for which primers from a modified MLST scheme (
Calculation of discriminatory power and congruence among typing methods
The discriminatory index (D) was calculated using the Simpson's diversity index described by Hunter and Gaston (1988), taking into account the whole set of typing results obtained by MLVA, MLST, PFGE, automated EcoRI ribotyping, and serotyping for each tested isolate. The AR and AW coefficients, along with the respective CIs, were calculated using the Comparing Partitions website (
Statistical analysis
The data collected were analyzed with the Statgraphics® package (version 5.1) (StatSoft, Inc.). Prevalence of L. monocytogenes in different slaughterhouses and samples was compared by using the Pearson's chi-square test. A p value ≤0.05 was considered statistically significant.
Results
Prevalence of L. monocytogenes in the rabbit meat and environmental samples
The overall prevalence of L. monocytogenes, assessed by using the ISO method up to confirmation, was 11% (Table 1) and it was significantly different in meat cuts (33.6%), rabbit carcasses (9.8%), and environmental sponges (2.2%). Taking into account all classes of samples tested, prevalence of positive samples was significantly higher in slaughterhouse A than in slaughterhouses D, B, and C. Different classes of samples as well as slaughterhouses investigated also showed statistically significant differences. The environmental sponges collected from slaughterhouses B, C, and D and the carcasses of slaughterhouse B had lower number of positive samples. On the contrary, the highest prevalence was found in meat cuts collected from slaughterhouses A and B (Table 1). Prevalence of L. monocytogenes in the same class of samples collected from different slaughterhouses did not show any statistically significant difference. The only exception was represented by rabbit carcasses collected from slaughterhouse B (Table 1).
Numbers without common superscripts (a–d) differ significantly (p ≤ 0.05).
Overall, 92.7% of the positive carcasses were obtained in slaughterhouses A and D (Table 2). L. monocytogenes counts higher than 1 colony-forming unit (CFU)/10 g were obtained in 14.3% of positive carcasses (Table 2). The L. monocytogenes contamination level among rabbit meat cuts ranged between 13% for loins and 66.6% for deboned products. On the contrary, the level of contamination of rabbit meat-based products was always higher than 40% (Table 2). The concentration of L. monocytogenes was determined to be greater than 1 CFU/10 g in 29.9% of meat cuts, whereas only one rabbit meat-based product had counts exceeding that limit (Table 2).
Samples positive for L. monocytogens and showing L. monocytogens enumeration results >1 CFU/10 g are in bold.
Overall, 13 positive environmental samples were collected: 10 were from the conveyor belt No. 2 located in the cutting area of slaughterhouse A during 3 months; 1 positive sample was obtained in the same location, but on conveyor belt No. 1; 2 additional positive samples were collected the same day on a Teflon cutting board, located in the cutting area, and a calibration tank, located in the evisceration area, both within slaughterhouse D (Table 3).
Genotypes and serotypes
A total of 123 isolates (i.e., 1 from each selected positive sample) obtained from rabbit carcasses (N = 41), rabbit meat cuts and products (N = 71), and environment sponges (N = 11) were typed (Table 4). The 41 isolates isolated from carcasses were obtained in slaughterhouses A (N = 26), B (N = 1), C (N = 2), and D (N = 12) (Table 4). Within slaughterhouse A, 21 isolates were classified as ST14, 4 as ST224, and 1 as ST121. All isolates with ST14 shared serotype 1/2a and the same MLVA and PFGE profile. Moreover, all but two isolates shared the same ribotype. The genetic profile identified among the majority of carcass isolates collected from slaughterhouse A was that associated with the majority of isolates from rabbit products tested in the same slaughterhouse. All carcass isolates collected in September displayed the same genetic profile and serotype 1/2b, whereas isolates collected in May were serotyped as 1/2a and MLVA and pulsotype different from those of the other isolates obtained in the same slaughterhouse. Within slaughterhouse D, the genetic profiles associated with the 12 carcass isolates that were characterized as STs 121 (N = 5) and 224 (N = 7) corresponded to those identified among the 13 product isolates during the same sampling period. The same result was observed for one carcass isolate obtained in slaughterhouse C with ST9, sharing the same genetic profile with the product isolate obtained in the same slaughterhouse, except for the ribotyping profile. Finally, the only carcass isolate obtained in slaughterhouse B, and classified as ST9, shared the same genetic profiles, with 9.7% of the product isolates collected from the same slaughterhouse.
MLVA, multilocus variable-number tandem-repeat analysis; PFGE, pulsed-field gel electrophoresis; ST, sequence type.
The 71 isolates collected from rabbit products were obtained in the slaughterhouses A (N = 25), B (N = 32), C (N = 1), and D (N = 13) (Table 4). Within slaughterhouse A, 18 isolates were classified as ST14, 3 as ST121, 3 as ST9, and 1 as ST11. The isolates with ST14 were the same serotype (i.e., 1/2a), MLVA, PFGE, and ribotype. They were collected from both rabbit cuts and rabbit products. In particular, they were obtained from rabbit hamburgers (made with ground rabbit meat) and rabbit cuts during the same sampling in February; hamburgers, cuts, and legs in the same sampling in March; in three different rabbit products in the second half of March, and again from sausages and legs in the same sampling day in April. The remaining isolates were collected from different rabbit cuts during different sampling days. The isolates classified as ST121 had the same serotype (i.e., 1/2a), MLVA, MLST, and PFGE, but different ribotype. They were obtained on different rabbit cuts. The isolates with ST9 had the same serotype (i.e., 1/2c), but different MLVA and two different PFGE and ribotyping profiles. Finally, the isolate with ST11 shared typing profiles identical to those of isolates with ST14. Within slaughterhouse B, all isolates, but one, had ST9, serotype 1/2c, and identical pulsotype. However, they were characterized by five different MLVAs and three different ribotypes. The only exception was represented by one isolate classified as ST224, serotype 1/2b, and MLVA, PFGE, and ribotyping profiles were never observed among isolates with ST9 in the same slaughterhouse. The only product isolate obtained in slaughterhouse C was classified as ST121, serotype 1/2a, and MLVA, as well as PFGE profiles, not associated with the two carcass isolates from the same slaughterhouse. The 13 isolates collected from slaughterhouse D were classified into two clusters: one containing 7 isolates from rabbit cuts, classified as ST121, serotype 1/2a, identical MLVA, PFGE profiles, and ribotype, except for 1 isolate with different ribotype; the second containing 6 isolates from rabbit cuts, classified as ST224, serotype 1/2b, and the same MLVA, PFGE, and ribotype, except for 1 isolate with different ribotype.
The 11 typed environmental isolates were collected from slaughterhouses A (N = 9) and D (N = 2) (Table 4). In slaughterhouse A, the isolates were obtained in the cutting area on two different conveyor belts, between December 2005 and March 2006. Five of those isolates shared the same profile (i.e., ST121, serotype 1/2a, MLVA 1, PFGE 4, and ribotype 347S1), whereas two isolates shared a different profile (i.e., ST14, serotype 1/2a, MLVA 2, PFGE 3, and ribotype 347S2). Finally, two isolates had the same ST9, MLVA, and serotype, but different PFGE and ribotyping profiles. The isolates collected from slaughterhouse D were obtained the same day in the evisceration area on a calibration tank and in the cutting area on a Teflon cutting board. However, they had different ST profiles (i.e., 9 vs. 14) as well as different genotypic and serotyping results.
Discriminatory power and concordance among typing methods
The discriminatory indexes (Ds) of the applied typing methods ranged between 0.810 for MLVA and 0.598 for serotyping (Table 5). The D calculated for MLVA was significantly higher than for all others, whereas that calculated for serotyping was significantly lower. Concerning the bidirectional concordance between methods, the AR coefficient value for PFGE versus MLST (AR 0.891) was significantly higher than all others (Table 6). The AW coefficient, concerning the unidirectional concordance between methods, was significantly higher for MLST versus serotyping (AW 1.0), MLVA versus MLST (AW 0.960), MLVA versus serotyping (AW 1.0), MLVA versus PFGE (0.927), PFGE versus MLST (AW 0.929), and PFGE versus serotyping (AW 0.949) (Table 7).
Numbers without common superscripts (a–c) differ significantly (p < 0.001).
CI, confidence interval; MLST, multilocus sequence typing.
Numbers without common superscripts (a–c) differ significantly (p < 0.001).
Numbers without common superscripts (a–d) differ significantly (p ≤ 0.05).
Discussion
Microbiological data, including typing results, concerning L. monocytogenes contaminating rabbit meat products and rabbit slaughterhouse environments are limited. Prevalence of L. monocytogenes among the 1285 rabbit meat and environmental samples investigated in this research was 11%. The highest prevalence was observed in rabbit meat cuts and products, suggesting that L. monocytogenes contamination increases during processing, as a result of cross-contaminations because of cutting and handling. In slaughterhouse A, the 35.9% prevalence in the meat cuts corresponded to a prevalence of 15.7% and 14.3% on rabbit carcasses and environmental sponges, respectively. However, in slaughterhouse B, the 33.9% prevalence among meat cuts did not reflect the low or zero prevalence on carcasses and in the environment. In slaughterhouse C, a few carcasses and meat samples were tested, and in slaughterhouse D, the 33.3% prevalence on meat cuts reflected the 21.7% prevalence on rabbit carcasses.
Overall, 27.9% of the rabbit meat cuts and 14.3% of carcasses had L. monocytogens counts higher than 1 CFU/10 g. Rabbit meat must be cooked before consumption. However, the high concentration of the pathogen on raw meat, or after storage at refrigeration temperature, can pose cross-contamination problems during meat preparation and handling by consumers. Concerning the role of rabbit as primary source of L. monocytogens contamination, intensive farmed rabbits and their meat tested negative in a German study (Van Treel, 2006). Similarly, Kohler et al. (2008) did not find L. monocytogens during testing of 500 rabbit fecal samples and 500 rabbit carcasses within a slaughterhouse in Switzerland. Khalafalla (1993) reported 10% prevalence of L. monocytogenes among rabbit meat products at retail level in Egypt. Busani et al. (2005) described 4.2% of positive samples among rabbit and pigeon meat samples tested in Italy. Moreover, Rodríguez-Calleja et al. (2006) found L. monocytogenes in 5.9% of 51 samples collected in Spain. According to Chaitiemwong et al. (2010), the highest prevalence of positives was obtained on conveyor belts, which are probably difficult to clean and represent a possible site for L. monocytogenes forming biofilms (Wong, 1998; Bremer et al., 2001; Wilks et al., 2006).
There are 13 serotypes of L. monocytogenes, but most outbreaks of human disease are caused by serotypes 1/2a, 1/2b, and 4b (Orsi et al., 2011). In this study, the identified serotypes were 1/2a (52.8%), 1/2c (32.5%), and 1/2b (14.6%). Similar results were reported in previous articles on meat products and environments (Giovannacci et al., 1999; Chasseignaux et al., 2001; Thevenot et al., 2006; Martín et al., 2014). As reported in previous studies (Thevenot et al., 2006; Ferreira et al., 2011; Meloni et al., 2012; Kramarenko et al., 2013), our results revealed a higher prevalence of serotype 1/2a in meat products. MLST is well established as a population biology tool and provides a standard operational definition of clones as CCs associated with specific sources, such as foods and humans (Chenal-Francisque et al., 2011; Ragon et al., 2008). The majority of our isolates were classified as ST14 (i.e., 34.15%), followed by ST9 (35.52%), ST121 (17.89%), and ST224 (14.63%). Finally, one isolate was classified as ST11. Overall, ST9, ST14, and ST121 were associated with isolates collected from environment, carcasses, and rabbit products. On the contrary, ST11 was associated with a rabbit product and ST224 was never associated with environmental isolates. Among the identified STs, ST14 has been previously identified in meat products and cured fish (Parisi et al., 2010); ST11 belongs to the epidemic clone III (Ragon et al., 2008), whereas ST224 has been associated with both food and human isolates (Jensen, 2016). This result shows that rabbit products, as other meat matrixes, can harbor L. monocytogenes isolates potentially pathogenic to humans. More information is available for ST9 and ST121, classified in the corresponding CCs: CC9 and CC121. The strains belonging to these CCs may have genetic determinants supporting contamination of foods and food processing environments as preferential sites, but seem to be less pathogenic to humans (Henri et al., 2016). Similar conclusions, regarding ST121, were recently drawn from MLST analysis performed on a large collection of food and clinical isolates (Maury et al., 2016). ST121 strains have previously been isolated from food and food processing facilities over several years, in processing plants in Denmark (Holch et al., 2013), Austria, and Belgium (Hein et al., 2011). ST9 and ST121 strains have also been reported in several European countries from different environmental, clinical, and food sources (Ragon et al., 2008; Parisi et al., 2010; Holch et al., 2013). MLVA was the typing method showing the best discriminatory power, whereas the best bidirectional concordance was achieved for PFGE versus MLST. In relation to the unidirectional concordance, both MLST and MLVA can predict the serotyping results. Our MLVA and MLST results can be easily compared, in terms of similarities and differences between strains, with sequencing data, representing the most promising technique in the analysis of L. monocytogenes (Kwong et al., 2016).
Overall, this study shows for the first time in Italy prevalence and genetic profiles of L. monocytogenes isolated in rabbit products and slaughterhouses, indicating that rabbit meat contamination increases during slaughtering, and contaminating strains belong to serotypes and CCs able to cause human diseases.
Footnotes
Disclosure Statement
No competing financial interests exist.