Introduction
Many sporadic and outbreak cases of human listeriosis are closely related to the consumption of
Listeria monocytogenes (
L. monocytogenes) contaminated ready-to-eat (RTE) products, such as milk, cheese, sausages, fruits, and salads [
1–
3].
L. monocytogenes can survive in a wide range of pH (4.0−9.5) [
4], therefore it can overcome the gastric fluids and macrophage phagosomes of the host’s stomach when it enters the digestive system along with the contaminated food [
5]. For immunocompetent groups,
L. monocytogenes usually results in gastrointestinal infections. For immunocompromised individuals, it can invade the epithelial barrier and enter into the blood system, resulting in bacteremia, sepsis, or focal infection like arthritis, hepatitis, and peritonitis [
6–
8].
L. monocytogenes can also traverse the blood-cerebrospinal barrier, especially of fetuses and newborns, leading to central nervous system (CNS) infections [
9].
L. monocytogenes may transfer across the placental barrier of pregnant women and cause maternal intrauterine infections that cause abortion, stillbirth, and premature delivery [
10].
L. monocytogenes is a robust pathogen that is widely distributed in soil, irrigation water, seawater, crops, animals, as well as food processing and retailing environments, thereby facilitating transmission to humans [
5]. RTE products being consumed directly without treatment, are well-recognized as one type of high risk food products for listeriosis. In Europe,
L. monocytogenes in RTE food should not exceed 100 CFU/g throughout its shelf-life; meanwhile, in RTE products for infants or for special medical purposes, absence of this pathogen in 25 g of samples is required [
11].
L. monocytogenes in raw food products may cross-contaminate the final RTE products through the handling of the operation personnel or through the food-processing environment, such as machines, tools, and package materials. Moreover, owing to the capability of forming biofilms and stress tolerance,
L. monocytogenes was found to be persistently colonized in the food processing plants leading to continuous contamination of the final products [
12].
Among the foodborne illnesses in the USA,
L. monocytogenes infection ranked the third leading cause of death (19%) and the fifth cause of disability-adjusted life years (8800) [
13,
14]. The thirteen serotypes of the
L. monocytogenes differ in terms of the listeriosis-causing frequency; serotype 4b, 1/2b, and 1/2a, are responsible for the majority of clinical cases, and serotype 4b is responsible for most of the outbreaks [
3]. However, the key virulence-related genes on the
Listeria pathogenicity island 1 (LIPI-1) (i.e.,
actA,
hly,
mpl,
plcA,
plcB, and
prfA), the internalin genes (
inlA,
inlB,
inlC, and
inlJ), and the
iap (invasion associated protein) genes are widely distributed in all of the serotypes, leading to the misconception that
L. monocytogenes are equally virulent in general [
15]. In the recent years, the findings of the novel pathogenicity factors that were unevenly distributed in different subtypes of
L. monocytogenes demonstrated their heterogeneity in terms of virulence [
15,
16]. Listeriolysin O, which is coded by the
hly gene on LIPI-1, is harbored by all serotypes of
L. monocytogenes [
16]. However, the peptide heamolysin listeriolysin S (LLS), which can enhance the hemolytic and cytotoxic activity of
L. monocytogenes, is only possessed by some lineage I strains [
16,
17]. The gene cluster of LLS is named pathogenicity island 3 (LIPI-3) [
16].
Listeria pathogenicity island 4 (LIPI-4) is a newly found gene cluster and expresses the cellobiose-family phosphotransferase system (PTS). Through the PTS gene cloning and deletion experiments, PTS contributes to the neural and placental infections of complex 4 (CC4)
L. monocytogenes [
15]. In this study, polymerase chain reaction (PCR) amplification of the LIPI-3 gene fragment
llsX and the LIPI-4 gene fragment (we named it
ptsA which is responsible for encoding the maltose-6′-phosphate-glucosidase) was used to identify the potential hypervirulent
L. monocytogenes that was isolated from cooked foods and listeriosis cases.
This study investigated the prevalence of L. monocytogenes in cooked products from RTE food shops and mutton restaurants in Zigong City, China. Serotyping and multilocus sequence typing (MLST) were utilized to analyze the diversity of L. monocytogenes. LIPI-3 and LIPI-4 genes (llsX and ptsA, respectively) were screened for the potential hypervirulent L. monocytogenes. The molecular characteristics of the food-source L. monocytogenes in this city were further compared with those of the 33 clinic-sourced strains from different districts of China, including four local listeriosis strains in Zigong City.
Material and methods
Sampling, isolation, and identification of L. monocytogenes
A total of 46 RTE food shops and 17 mutton restaurants, which are distributed in 6 districts of Zigong City, were sampled monthly with an average of 70 samples (ranging from 30 to 90) collected each month in 2015. RTE meat pieces and RTE vegetables were sampled from the food trays in the RTE food shops. A total of 560 RTE samples were collected, including 334 RTE pork, 153 RTE vegetables, 11 RTE beef, 25 RTE chicken wings, and 37 cooked mutton or haggis pieces (Table 1). Moreover, 313 samples were collected from the 17 mutton restaurants, including 125 cooked haggis (cooked sheep organ meat pieces, e.g., head, lung, heart, liver, kidney, stomach, and intestine), 102 cooked mutton and 37 environmental swabs on meat trays, refrigerators, tables, and knives (Table 1). In mutton restaurants, cooked mutton and haggis pieces are stored at 4 °C or −20 °C in refrigerators and would mostly likely be reboiled before consumption. The same isolation and identification methods of
L. monocytogenes were used as in our previous study in Zigong City [
18].
To compare the molecular characteristics of cooked food sources of
L. monocytogenes in Zigong City and clinical listeriosis strains, we included 33 clinic-source strains from different districts in China during the past 10 years in this study. Among the 33 strains, 23 strains were previously reported by Wang
et al. [
19].
Serotyping and MLST of isolates
DNA extraction of all
L. monocytogenes isolates were performed using boiling methods as indicated in our previous study [
18]. All strains were serotyped with multiplex PCR assays, followed by antigen serum agglutination, which was done according to the method reported by Doumith
et al. [
20], and according to the manufactures’ instructions (Denka Seiken, Tokyo, Japan). MLST based on PCR amplification and sequencing of the seven housekeeping genes (
abcZ,
bglA,
cat,
dapE,
dat,
ldh, and
lhkA) was applied to analyze the sequence types of
L. monocytogenes that was collected from different sources, according to the reports of Ragon
et al. [
21]. A minimum spanning tree was constructed using the BioNumerics 4.0 (Applied Maths, Kortrijk, Belgium) to compare the sequence types of cooked food isolates in Zigong City and the 33 clinical strains of listeriosis collected from different districts of China during the past 10 years (Fig. 1). A column chart was constructed using Microsoft Excel 2010 to describe the infection types of the 33 clinical strains in different clonal complexes (CCs) (Fig. 2).
Identification of hypervirulent genes and phylogenetic tree construction
PCR amplification of the LIPI-3 and LIPI-4 genes (
llsX and
ptsA, respectively) was used to identify the potential hypervirulent
L. monocytogenes among all the 33 clinic-source isolates and the 141 cooked food source isolates. Primers for
ptsA were designed using the Primer-Blast of NCBI. The primers for PCR amplification of
ptsA and the annealing temperature were indicated in Table S1. The primers for PCR amplications of
llsX has been described in Clayton
et al.’s report [
22]. Each PCR reagent was a 25 mL system, including 2× Taq PCR Master Mix 12.5 mL (Takara, Japan), 1 mL forward primer (10 mmol/L), 1 mL reverse primer (10 mmol/L), and 1.5 mL DNA template.
A maximum likelihood-based phylogenetic tree was constructed based on the seven concatenated housekeeping gene sequences of 40 STs with Mega 5.04 software to analyze the phylogenetic relationship of different STs and the distribution of hypervirulence islands (Fig. 3). Among the 40 STs, 22 STs occurred in this study, and the dominant STs of the top 30 clones among the 6633 food and clinical isolates in Maury
et al.’s study and ST382 in Chen
et al’s report were included [
15,
23]. The seven housekeeping genes of each ST were aligned using the EditSeq software (DNASTAR/Lasergene). All of the bootstrap values (1000 replicates) were shown on the branch nodes of the tree.
Data analysis
χ2 test and Fisher’s exact test were applied to analyze the seasonal variations for the prevalence of L. monocytogenes from different sources and to compare the dominant L. monocytogenes subtypes of raw pork meat and RTE pork meat isolates. Based on the STs, R software was used to calculate the Simpson’s Diversity Index of L. monocytogenes from the RTE shops and mutton restaurants.
Results and discussion
Prevalence of L. monocytogenes in RTE products
In this study, the overall prevalence of
L. monocytogenes in the RTE shops and mutton restaurants was 16.2% (141/873). The occurrence of
L. monocytogenes in RTE food shops samples was 11.8% (66/560) (Table 1), which was relatively higher than the results of another Chinese study on RTE foods (6.3%) [
24]. Moreover, 63.7% of the RTE foods were RTE pork meat (334/523). Prevalence of
L. monocytogenes in the RTE pork meat was 13.5% (Table 1). No statistic difference was observed between the raw and RTE pork meat (
χ2=0.057,
P>0.05) compared with the occurrence of
L. monocytogenes in raw pork meat (13.0%) collected from Zigong City [
18]. Occurrence of
L. monocytogenes in either cooked haggis or cooked mutton was higher than 24.0% (Table 1). These data indicated that although raw meat products underwent the process of heating, the occurrence of
L. monocytogenes in cooked meat did not decrease substantially. Because
L. monocytogenes is susceptible to heat treatment (as high as 100°C in China) [
25], the high occurrence of
L. monocytogenes in cooked products can only be explained by the contamination that took place during the processing stage after the heat treatment. Studies found that
L. monocytogenes contamination in the manufacturing environments of RTE foods is the source of recurrent contamination in the final products [
12,
26]. Thus, sanitary conditions in the RTE food shops and mutton restaurants should be improved. The prevalence of
L. monocytogenes in RTE vegetables was 6.5% (10/153) (Table 1). Many listeriosis outbreaks are related to the consumption of
L. monocytogenes-contaminated RTE meat or RTE vegetables [
27–
29]. Surveillance and preventive measures should be enhanced on the handling of cooked products, especially of RTE foods.
In RTE food shop samples that were sold at room temperature, the prevalence of
L. monocytogenes in spring and winter (15.8%, 41/260) was significantly higher than that in summer and autumn (8.3%, 25/300) (
χ2=7.41,
P<0.01) (Table 2), which was similar to the results of raw pork meat in Zigong City as reported in the previous study [
18]. The psychrophilic feature of
L. monocytogenes was considered to be related to this phenomenon [
18]. No seasonal variation was observed in cooked mutton or haggis pieces that were stored in refrigerators (at 4 °C or -20 °C) (
χ2=0.01,
P>0.05) (Table 2).
Subtyping characteristics of L. monocytogenes
In this study, the 141
L. monocytogenes isolates from cooked products were subtyped into 5 serotypes; serotype 1/2b (64, 45.4%), 1/2a (47, 33.3%), and 1/2c (20, 14.2%) comprised 92.9% of the isolates (Table 3). Serotypes 4b and 3a were relatively rare with only 4.3% (6/141) and 2.8% (4/141), respectively. These results were similar to those of other studies in China [
18,
24,
30]. However, in Wu
et al.’s report on the RTE foods from 24 cities of China, the serotype 4b (CC1) strains were prevalent but were only collected from two cities [
31].
All the food-source
L. monocytogenes were subtyped into 15 CCs and one singleton ST619, with CC3 and CC87 being the predominant subtypes (Fig. 1). However, since food samples in this study were monthly collected in 2015, it can introduce an overestimation for the prevalence of certain CCs which were repeatedly isolated from the same sampling sites. For example, more than half of the ST3 strains were isolated from the one mutton restaurant in which the samples were collected in different months of 2015 (Table 3). By analyzing the occurrence of
L. monocytogenes in the 46 RTE food shops and 17 mutton restaurants that were used in the study (63 in total), we found CC87 (23.8%, 15/63) and CC9 (20.6%, 13/63) were the two types that occurred frequently in different restaurants or shops, which were followed by CC8 (12.7%, 8/63) and CC3 (11.1%, 7/63) (Table 3). Similarly, these four CCs were prevalent in raw pork meat samples in Zigong City [
18]. CC9 and CC121 are the most prevalent CCs in food products in Switzerland, according to Ebner
et al.’s report [
32], whereas CC121 was relatively less frequent in this study and other reports in China [
30,
31]. ST3 was found only in 3.8% (8/212) of the
L. monocytogenes isolates in food products from 12 provinces of China in Wang
et al.’s report [
30].
The strains of L. monocytogenes isolated from the samples in the mutton restaurants were subtyped into 10 CCs and 1 singleton ST619, with the Simspon’s diversity index of 0.8393 (based on STs). It showed similarity to the strains from the RTE food shops that were subtyped into 12 CCs and 1 singleton ST619, with the Simspon’s diversity index of 0.8376 (Table 3). During the multiple sampling done in different months, the recurrence of identical STs in the same sampling sites (like CC3) may indicate the possibility of persistent contamination. Subtyping methods of higher discriminative power is required for the persistence identification of L. monocytogenes.
The 33 clinic-source strains from different districts of China were subtyped into 13 sequence types (STs) and 11 CCs (Fig. 2). CC87 and CC3, which were prevalent in the cooked products, were the most common types in the clinical isolates (Table 3, Fig. 1). Moreover, four known listeriosis cases in Zigong City (during 2014 and 2015) were caused by CC8 in two cases, and CC3 and CC87 in one case [
33]. CC8 was a prevalent CC in cooked products in this study (Table 3) and was predominant in several surveys done in China [
18,
30]. In Althaus
et al.’s report on 98 clinical listeriosis cases in Switzerland, CC8 (serotype 1/2a) and CC1 (serotype 4b) were the first two subtypes [
34]. The prevalent CCs of isolates in cooked products, especially in the RTE foods in this city, were similar to that of clinical isolates. But it should be noted that the national surveillance system of listeriosis has not been established in China, and the number of clinic-source
L. monocytogenes strains collected was relatively few despite its frequent presence in food products [
18,
35].
The prevalent clinical subtypes of
L. monocytogenes in Western countries are different from those prevalent in the Chinese mainland and Taiwan. In Western countries, CC1, CC6, CC2, and CC4 (all of which belonged to serotype 4b) are the frequent types in human listeriosis cases, especially in outbreaks [
15]. In contrast, CC87 (serotype 1/2b) is the most prevalent type in clinical cases in both Chinese mainland and Taiwan, as reported in literature [
19,
36]. CC87 has caused two outbreaks in Spain, but it is relatively uncommon in Western countries [
15,
37]. The low occurrence of serotype 4b CCs in listeriosis cases in China may be partially attributed to the relatively low prevalence of serotype 4b in Chinese food products compared with that of Western countries [
18,
30,
38–
40]. More attention should be paid to the potential hypervirulent serotype 4b CCs strains, especially in the imported products reported in the literature to be predominately contaminated by serotype 4b
L. monocytogenes [
41].
Presence of LIPI-3 and LIPI-4 genes in L. monocytogenes isolates
A recent study has revealed that
L. monocytogenes is heterogeneous in virulence, which is closely associated with the CCs by MLST [
15]. LIPI-4 is a newly identified cluster of hypervirulent genes, which encodes the putative cellobiose-family PTS presented in CC4
L. monocytogenes [
15]. All of the CC87 strains of
L. monocytogenes in this study, including ST87 and the novel ST1166, were positive for the LIPI-4 gene fragment
ptsA (Table 3, Fig. 3). Moreover, the whole LIPI-4 cluster existed in the genome of a clinical ST87 strain ICDC-LM188 of China (accession number PRJNA319288). Except for CC87 and ST619, all the other CCs of
L. monocytogenes were negative for the putative PTS. Among the 8 clinical strains of CC87, 6 strains were responsible for maternal-neonatal infections (MN infections) and 1 strain for CNS infection (Fig. 2). Although the number of CC87 cases was limited, it supported the findings of Maury
et al. that LIPI-4 is closely related with the neural and placenta infection in humanized mouse [
15].
CC3 (ST3 and ST778), CC1 (ST1 and ST308), and ST619
L. monocytogenes were positive for the listeriolysin S (LLS) or LIPI-3 gene fragment
llsX (Table 3, Fig. 3). Previous studies had reported that CC3 and CC1 harbored listeriolysin S [
15,
31], which has enhanced hemolytic and cytotoxic reactions in in vitro and in vivo experiments [
16].
ST619 (serotype 1/2b) that was isolated from the RTE meat and cooked haggis was positive for both LIPI-3 and LIPI-4 genes, which was similar with the hypervirulent
L. monocytogenes CC4 that belonged to serotype 4b (Fig. 3) [
15]. According to the systematic phylogenetic tree based on the seven housekeeping genes, both CC87 and ST619 (belonging to serotype 1/2b) were genetically distant to the hypervirulent CC4 (Fig. 3). ST619 is not a dominant sequence type in the district, but it is present in the raw pork meat of this city, as reported in the previous study [
18]. However, neither ST619 nor hypervirulent serotype 4b clones were found in listeriosis cases in Zigong City. This result may be attributed to the lack of established systematic listeriosis surveillance system in China. The establishment of the clinical listeriosis surveillance system in China is necessary to better know and control the prevalence of clinic listeriosis, especially of the invasive infection of
L. monocytogenes.
The presence of LIPI-4 and LIPI-3 genes in CC87 and CC3, respectively, may increase the prevalence of these two types of
L. monocytogenes in the listerisis cases in China. The prevalence of these two likely hypervirulent types in cooked foods, especially in the RTE products, is a great threat to the local consumers. CC7 is also a common type in listeriosis patients, whereas it is relatively rare in food source isolates (Table 3) [
30,
31]. Other unknown virulence mechanisms may exist in this CC.
Conclusions
The relatively high prevalence of L. monocytogenes in cooked foods, especially in RTE meat and vegetables, is a threat to the local consumers in Zigong City. CC87, CC9, CC8, and CC3 are prevalent in cooked products from this city. CC87 and CC3 are also frequent in the 33 clinic-sourced strains from different districts of China. The presence of LIPI-4 genes ptsA in CC87 strains and LIPI-3 genes llsX in CC3 strains may increase the prevalence of these two types in L. monocytogenes isolates from clinical listeriosis patients and further underline the risk of infection owing to the consumption of cooked foods, especially the RTE products in this city. This study firstly compared the molecular characteristics (based on the MLST subtyping and the screening of the hypervirulent genes) of the L. monocytogenes from food products and listeriosis cases in China. The results of this study would be helpful to control and prevent foodborne L. monocytogenes contamination and clinical infections.
Higher Education Press and Springer-Verlag GmbH Germany, part of Springer Nature