A Whole Genome Sequencing-Based Epidemiological Investigation of a Pregnancy-Related Invasive Listeriosis Case in Central Italy

Listeriosis is currently the fifth most common foodborne disease in Europe. Most cases are sporadic; however, outbreaks have also been reported. Compared to other foodborne infections, listeriosis has a modest incidence but can cause life-threatening complications, especially in elderly or immunocompromised people and pregnant women. In the latter case, the pathology can be the cause of premature birth or spontaneous abortion, especially if the fetus is affected during the first months of gestation. The causative agent of listeriosis, Listeria monocytogenes, is characterized by the innate ability to survive in the environment and in food, even in adverse conditions and for long periods. Ready-to-eat food represents the category most at risk for contracting listeriosis. This study presents the result of an investigation carried out on a case of maternal-fetal transmission of listeriosis which occurred in 2020 in central Italy and which was linked, with a retrospective approach, to other cases residing in the same city of the pregnant woman. Thanks to the use of next-generation sequencing methodologies, it was possible to identify an outbreak of infection, linked to the consumption of ready-to-eat sliced products sold in a supermarket in the investigated city.


Introduction
The Gram-positive Listeria monocytogenes is a facultative intracellular bacterium responsible for the food-borne disease listeriosis [1][2][3], an infection predominantly transmitted through the consumption of contaminated food.
Listeriosis has a wide spectrum of infections, from severe invasive listeriosis, non-invasive febrile gastroenteritis [4] to asymptomatic forms [5]. As matter of fact, L. monocytogenes occurs in stool of asymptomatic carriers probably due to a less diverse microbiota compared to

Case Presentation
On 9 January 2020, a 28-year-old pregnant woman, residing in a city near Rome and without a significant medical history, was admitted in the 31st week of gestation to the Obstetrics and Gynecology Department of Casilino Hospital in Rome, due to premature rupture of membranes (PROM) and gastrointestinal symptoms. At the admission, neutrophilic leukocytosis was documented (33,000 white blood cells/mm 3 with 28,080/mm 3 neutrophils), along with an increase in C-reactive protein (CRP) (90.1 mg/L). Blood cultures, serum Listeria IgM, screening for TORCH infections and group B Streptococcus analysis tested negative whereas no cerebrospinal fluid (CSF) examination was performed at that time. Ampicillin 2 gm intravenous (IV) q6h was promptly prescribed; the following day a cesarean section was performed and a 1540 gr female afebrile baby was born. Apgar at one minute was 7, than 8 at five minutes. The baby presented respiratory distress and she was diagnosed with hyaline membrane disease (HMD). Blood tests showed a noticeable increase in inflammatory indices (procalcitonin and CRP measured 0.774 mg/mL and 19.8 mg/L respectively), and the blood samples culture, collected few hours after the birth, documented the growth of L. monocytogenes. Antibiotic therapy with intravenous (IV) weight-adjusted ampicillin and gentamicin was promptly initiated. Lumbar puncture was also performed; cerebrospinal Fluid (CSF) was sterile, opalescent, with 100 white blood cells (WBC), mostly polymorphonuclear leukocytes, with normal glucose and protein levels.
She was transferred to the Neonatal Intensive Care Unit where she underwent to invasive ventilation for 24 h, then to non-invasive positive-pressure ventilation for further 48 h; negativization of blood cultures and normalization of procalcitonin and CRP occurred. The baby was discharged 30 days later in good clinical conditions. After the notification of the listeriosis case, an investigation was conducted by the local health authority, which administrated an epidemiological questionnaire to the mother in order to assess any potential food exposures during the 30 days before illness onset.
The mother declared that she had consumed mainly food purchased in two local supermarkets, one local butcher shop and one fish market. In particular, she reported the purchase of fresh horse meat in the butcher shop, cold cuts at the deli counter as mortadella, baked ham and salami, ready-to-eat bagged salad, other fresh vegetables and fresh fruits (apples and tangerines), dairy products as mozzarella and parmigiano in the two local supermarkets, and smoked salmon and tuna in the fish market. She also claimed to have eaten once at a local restaurant and she denied any travel abroad during the past month.
According to the mother's declarations, the local health authority's attention was focused on the fresh horse meat purchased in the local butchery and cold cuts at the deli counter purchased in one of the two reported supermarkets. During the different inspections, food samples and environmental samples were collected to assess the presence of L. monocytogenes.
The patients' clinical course and the events related to epidemiological investigation are summarized in Figure 1.

Retrospective Epidemiological Investigation
After the notification of the neonatal listeriosis, additional inquiries were carried out to explore the occurrence of further listeriosis cases in the same city of the woman and her daughter (Case A). In the case series of samples collected by our regional center, three strains of L. monocytogenes isolated from clinical cases were found (Case B, C and D) residing in the same city, occurring from May 2019 to May 2020 and all belonging to the same serotype (see below). Where possible, patients were contacted by the competent authority to investigate any epidemiological links.

Retrospective Epidemiological Investigation
After the notification of the neonatal listeriosis, additional inquiries were carried out to explore the occurrence of further listeriosis cases in the same city of the woman and her daughter (Case A). In the case series of samples collected by our regional center, three strains of L. monocytogenes isolated from clinical cases were found (Case B, C and D) residing in the same city, occurring from May 2019 to May 2020 and all belonging to the same serotype (see below). Where possible, patients were contacted by the competent authority to investigate any epidemiological links.
The first case (B) occurred nine months before the above-mentioned episode and involved a patient admitted to the emergency department. A lumbar puncture was performed and a diagnosis of L. monocytogenes meningitis was formulated. CSF cultures reported the growth of L. monocytogenes; antimicrobial susceptibility testing documented ampicillin, erythromycin and meropenem sensibility (MICs determined by E-test: 0.25 µg/mL, 0.38 µ g/mL and 0.094 µ g/mL respectively) and resistance to trimethoprim-sulfamethoxazole (0.094 µ g/mL).
The second case (C) occurred three months later and a patient was hospitalized in April 2020. Blood cultures performed at the time of admission documented the growth of L. monocytogenes. Unfortunately, no data were available on the antimicrobial susceptibility profile.
The third case (D) occurred four months after the above-described neonatal listeriosis case, in May 2020. Seven days after the admission, due to the occurrence of fever, the patient underwent two blood cultures and L. monocytogenes was isolated from both samples. The strain proved to be susceptible to ampicillin (≤0.25 mg/L), erythromycin (≤0.12 mg/L), meropenem (0.19 mg/L), penicillin (0.5 mg/L) and trimethoprim-sulfamethoxazole (0.012 mg/L).
From epidemiological investigation and questionnaire collection it emerged that at least one patient (case B) confirmed their attendance at the same supermarket of the concerned neonatal listeriosis case. No further information could be obtained from the other patients.

Bacterial Identification and Serotyping
Cervicovaginal swabs of the mother performed after identification of Listeria in the newborn were negative. Unfortunately, probably due to antibiotic therapy, it was not possible to isolate the bacterium responsible for the disease, and no more investigation could be carried out on this strain. Clinical information about the first diagnosis of listeriosis The first case (B) occurred nine months before the above-mentioned episode and involved a patient admitted to the emergency department. A lumbar puncture was performed and a diagnosis of L. monocytogenes meningitis was formulated. CSF cultures reported the growth of L. monocytogenes; antimicrobial susceptibility testing documented ampicillin, erythromycin and meropenem sensibility (MICs determined by E-test: 0.25 µg/mL, 0.38 µg/mL and 0.094 µg/mL respectively) and resistance to trimethoprimsulfamethoxazole (0.094 µg/mL).
The second case (C) occurred three months later and a patient was hospitalized in April 2020. Blood cultures performed at the time of admission documented the growth of L. monocytogenes. Unfortunately, no data were available on the antimicrobial susceptibility profile.
The third case (D) occurred four months after the above-described neonatal listeriosis case, in May 2020. Seven days after the admission, due to the occurrence of fever, the patient underwent two blood cultures and L. monocytogenes was isolated from both samples. The strain proved to be susceptible to ampicillin (≤0.25 mg/L), erythromycin (≤0.12 mg/L), meropenem (0.19 mg/L), penicillin (0.5 mg/L) and trimethoprimsulfamethoxazole (0.012 mg/L).
From epidemiological investigation and questionnaire collection it emerged that at least one patient (case B) confirmed their attendance at the same supermarket of the concerned neonatal listeriosis case. No further information could be obtained from the other patients.

Bacterial Identification and Serotyping
Cervicovaginal swabs of the mother performed after identification of Listeria in the newborn were negative. Unfortunately, probably due to antibiotic therapy, it was not possible to isolate the bacterium responsible for the disease, and no more investigation could be carried out on this strain. Clinical information about the first diagnosis of listeriosis and the therapy carried out are untraceable. Blood cultures drawn by the newborn reported the growth of L. monocytogenes, passed from the mother, presumably by vertical uterusplacenta cell-to-cell transmission, before the delivery. In this case it was possible to isolate the strain for further analysis.
For all the clinical strains antimicrobial susceptibility testing (AST) was performed as minimum inhibitory concentration (MIC). All the strains were susceptible to ampicillin (all the strains: The food samples of horse meat from the butcher shop were tested as negative for L. monocytogenes, meanwhile the mortadella samples from the first inspected supermarket were tested as positive, with a colony-forming unit less than 40 UFC/g. Subsequently, environmental samples were collected to identify the extent of contamination. Samples from the two meat-slicers (E1-3, G1-2) present in store were tested as positive for L. monocytogenes, as well as the cold food counter where the cold cuts were exposed to the public (H1-2). The worktable where the staff handled the meat (F), by contrast, was tested as negative for the contamination. All the positive strains were isolated for further analysis.
The isolates collected were assigned to the serotype 1/2b. Sequencing data confirmed serogroup IIb for the isolates and MLST analyses indicated that they all belonged to Sequence Type (ST) 5, Clonal Complex (CC) 5, lineage I.

In Silico Analysis
All the isolates considered in the study, four human cases (A, B, C, D), the environmental samples (E1-3, G1-2, H1-2) and food samples (M1-2), were found to bring the same genes for antibiotic resistance. The identified genes were those of resistance to the lincomycin (lin) with a new allele not described before, quinolone (norB), sulfonamides (sul), fosfomycin (fosX) antibiotics and cationic peptides that disrupt the cell membrane, including defensins (mprF). The virulence features were assessed and all the isolates shared the same virulence genes with the same alleles ( Table 1). The isolates presented the inlA, inlB and actA genes that played a crucial role in the virulence and pathogenesis of L. monocytogenes. The presence of metal, detergent resistance, stress Survival Islands and Listeria Genomic Islands were the same for all the strains and were reported in Supplementary Table S1. Table 1. Complete list of virulence genes found in the analysed strains, in brackets the corresponding lmo code according to BIGSdb-Lm is indicated. For each gene the corresponding allele was indicated (BIGSdb-Lm).

Virulence Gene
Allele Virulence Gene Allele * New allele found.
Pathogens 2022, 11, 667 6 of 15 The cgMLST profile was determined by using two different softwares: cgMLSTFinder (v1.1) and chewBBACA (v2.8.5), described in Material and methods section. In the first case the analysis identified two clusters which differed by 8 alleles (Figure 2A). The first cluster was composed by human isolates from case A and case D and the second one was composed by environmental samples and human isolates from cases B and C. No differences were detected inside each cluster. Using chewBBACA, the same two clusters were inferred ( Figure 2B). The human isolate from case A had 9-11 allelic differences with food and environmental samples, and showed one allelic difference with case D. The second cluster was composed of food and environmental samples, differing from each other by 2-3 alleles, and cases B and C that differed by 1-3 alleles each from environmental strains. The Minimum Spanning Tree showed the schematic differences among the isolates ( Figure 2).
The cgMLST profile was determined by using two different softwares: cgMLSTFinder (v1.1) and chewBBACA (v2.8.5), described in Material and methods section. In the first case the analysis identified two clusters which differed by 8 alleles (Figure 2A). The first cluster was composed by human isolates from case A and case D and the second one was composed by environmental samples and human isolates from cases B and C. No differences were detected inside each cluster. Using chewBBACA, the same two clusters were inferred ( Figure 2B). The human isolate from case A had 9-11 allelic differences with food and environmental samples, and showed one allelic difference with case D. The second cluster was composed of food and environmental samples, differing from each other by 2-3 alleles, and cases B and C that differed by 1-3 alleles each from environmental strains. The Minimum Spanning Tree showed the schematic differences among the isolates ( Figure 2). Further investigation was conducted by searching for ST5 L. monocytogenes isolates from food samples among those received by our laboratory during routine activities. Food samples were found in the records concerning chicken meat (Food Sample 1), pork meat (2), smoked swordfish (3) and salmon (4). Representative strains were used to compare food isolates with the investigated cluster, following the workflow described in materials and methods. Results showed that there was no relation with the main cluster in the core genome, with 58-106 allelic distance calculated with cgMLSTFinder and 66-115 allelic distance with chewBBACA (Supplementary Figure S1).
The BIGSdb-Pasteur, the genomic-based strain taxonomy and nomenclature platform of Institute Pasteur, assigned the same profile for all food, environmental and clinical samples: CT7923. Despite the higher number of allele differences found with the other tools, according to the data available to the platform, all the samples belong to the same CT (cgMLST type) and therefore probably to the same outbreak event.
The SNPs' analysis confirmed the tight relationship of the strains considered in this investigation. The L. monocytogenes strain from mortadella (M) had two SNPs (in whole and core genome) differences with the environmental isolates. The human isolates from cases A and D differed [16][17][18][19] SNPs in whole genome (10-13 in core genome) from food and environmental isolates, and human strains from cases B and C. The SNPs' matrix-based heatmap for whole genome and core genome are reported in Figure 3 and S2.
food isolates with the investigated cluster, following the workflow described in materials and methods. Results showed that there was no relation with the main cluster in the core genome, with 58-106 allelic distance calculated with cgMLSTFinder and 66-115 allelic distance with chewBBACA (Supplementary Figure S1).
The BIGSdb-Pasteur, the genomic-based strain taxonomy and nomenclature platform of Institute Pasteur, assigned the same profile for all food, environmental and clinical samples: CT7923. Despite the higher number of allele differences found with the other tools, according to the data available to the platform, all the samples belong to the same CT (cgMLST type) and therefore probably to the same outbreak event.
The SNPs' analysis confirmed the tight relationship of the strains considered in this investigation. The L. monocytogenes strain from mortadella (M) had two SNPs (in whole and core genome) differences with the environmental isolates. The human isolates from cases A and D differed [16][17][18][19] SNPs in whole genome (10-13 in core genome) from food and environmental isolates, and human strains from cases B and C. The SNPs' matrixbased heatmap for whole genome and core genome are reported in Figures 3 and S2.

Discussion
Herein is described a case of pregnancy-related invasive listeriosis occurring in central Italy, where the infection proved to be linked to contaminated food products and vertically transmitted from mother to baby.
The mother's infection manifested as asymptomatic form [5], with only gastro-intestinal symptoms probably confused with labor contractions.
Since it was not possible to isolate the strain of L. monocytogenes from the mother, the effective differences between the mother's and the child's strains remain unknown, even if a previous study suggested that in such cases no differences occurred, given the vertical transmission of the pathogen [45].
After a thorough completion of the epidemiological questionnaire by the mother, a strong epidemiological link with the mortadella sold in the supermarket was found, despite both the analyses of cgMLST showing allelic differences higher than the generally accepted limit to consider its inclusion in the same cluster. L. monocytogenes strains within seven allelic differences are considered to be linked to the same outbreak and are assigned to a unique cgMLST type (CT) [46], although this cutoff is debatable [45].
The use of different bioinformatics tools of analysis showed slightly different results. This could be due to the Prodigal gene prediction program, integrated in the chewBBACA pipeline, whose function is based on the identification of translation initiation sites [47] and, differently from cgMLSTFinder, is able to recognize new alleles when alternative start codons exist on the analyzed gene (https://github.com/B-UMMI/chewBBACA/issues (accessed on 11 February 2022).
Even if chewBBACA found 9-11 allelic differences between the newborn isolate and the food and environmental samples, thus excluding the possibility of tracing the source of contamination, the cgMLSTFinder results, on the other hand, could put it into question, finding only eight differences, at the limit of the above-mentioned threshold value. Moreover, the Institute Pasteur, whose bioinformatics analysis platform is based on chewBBACA software and which collects and characterizes isolates of L. monocytogenes all over the world, assigned the same CT7923 to all the strains involved in this investigation, suggesting a unique outbreak event. The overall view of their database suggests that the existence of other isolates, which act as a connection between groups, confirmed the existence of a single cluster.
Other considerations may support this hypothesis. First, the sequence type 5 (ST5) of the strains involved in this epidemiological investigation is very rare in human cases in Italy and almost limited to Lazio region (source: ISS IRIDA-ARIES database [48,49] accessed on 22 February 2022). Moreover, based on the sequence data of L. monocytogenes strains isolated from food samples during routine official control activities in the Lazio region by our laboratory, it seems to involve mainly pork meat (data not shown). These peculiar features, together with a strong epidemiological link found between the specific category of food, the geographical area and the habits of the mother of patient A during pregnancy and patient B, could support the hypothesis that all the reported episodes belong to the same outbreak event.
The differences found among samples with SNPs and cgMLST analysis could be due to the different time of isolation. Considering that inspections started one month later than the symptoms' onset in the mother, and given the highly variable incubation period of listeriosis (17-67 days) [50], the estimated exposition time would have been dated up to three and a half months before the positive sample collection.
In a recent work by Luo and colleagues in 2019, [45], it was found one SNP of difference in core genome between mother and child strains isolated 13 days apart, despite the mutation rate of L. monocytogenes core genome is estimated at 2.6 × 10 −7 substitutions per site per year (one substitution every 2.5 years) [46]. Thus, a much higher mutation rate, at least in clinical isolates or in pregnancy-associated listeriosis, may be suspected. As a matter of fact, the clinical isolates could evolve faster than environmental and food strains. Environmental temperature may be extremely variable, but never exceeds 20-22 • C. Food storage temperature is controlled between −20 • C and room temperature. These conditions, although allowing growth of L. monocytogenes, significantly increase the generation time of the organism compared to the optimal growth temperature (typically 30-37 • C) [51]. Furthermore, treatment of surfaces with cleansing and sanitizer could represent a stress factor for L. monocytogenes growth [52,53].
In addition to differences in generation time and variations in population sizes, the direction (i.e., negative or positive) and strength of natural selection could also contribute to differences in the rate of changes per site per year [54]. Given this assumption, the mutation rate could be higher than that estimated, which should explain the longer distance between clinical strains A and D and the suspected food and environmental strains.
Another hypothesis could be the occurrence of the analysis on a different portion or batch of the mortadella originally consumed, with a different strain than the original source of contamination, which could since have been completely sold by the supermarket and not available during the inspection. Alternatively, since it was not possible to isolate the maternal strain, it could have happened that the mother was originally infected with more than one strain of L. monocytogenes, not all transmitted or found to the newborn [55][56][57].
The analysis, through the use of WGS, showed that the three retrospectively identified human cases were closely linked with the newborn case A and with food and environmental samples. In Figure 4a schematic view of the events and relationships was reported. In particular, cases B and C were strictly linked to the supermarket and mortadella. Case B occurred nine months before the inspections and sampling at supermarket and differs 1-3 SNPs from food and environment (2-3 SNPs in core genome). The patient confirmed that he used to attend the concerned supermarket. Case C occurred more than two months after the sampling date, and present two SNPs with food strains and four SNPs with environmental strains two and three SNPs in core genome). Case D occurred more than three months after the supermarket sampling, presented the same amount of divergence with food and environment strains as case A, differing three SNPs in whole genome (one SNP in core genome). Therefore, the molecular investigation suggested that the source of contamination, in any case, had been active for a long period both before and after the official inspection by local heath authority. direction (i.e., negative or positive) and strength of natural selection could also contribute to differences in the rate of changes per site per year [54]. Given this assumption, the mutation rate could be higher than that estimated, which should explain the longer distance between clinical strains A and D and the suspected food and environmental strains.
Another hypothesis could be the occurrence of the analysis on a different portion or batch of the mortadella originally consumed, with a different strain than the original source of contamination, which could since have been completely sold by the supermarket and not available during the inspection. Alternatively, since it was not possible to isolate the maternal strain, it could have happened that the mother was originally infected with more than one strain of L. monocytogenes, not all transmitted or found to the newborn [55][56][57].
The analysis, through the use of WGS, showed that the three retrospectively identified human cases were closely linked with the newborn case A and with food and environmental samples. In Figure 4 a schematic view of the events and relationships was reported. In particular, cases B and C were strictly linked to the supermarket and mortadella. Case B occurred nine months before the inspections and sampling at supermarket and differs 1-3 SNPs from food and environment (2-3 SNPs in core genome). The patient confirmed that he used to attend the concerned supermarket. Case C occurred more than two months after the sampling date, and present two SNPs with food strains and four SNPs with environmental strains two and three SNPs in core genome). Case D occurred more than three months after the supermarket sampling, presented the same amount of divergence with food and environment strains as case A, differing three SNPs in whole genome (one SNP in core genome). Therefore, the molecular investigation suggested that the source of contamination, in any case, had been active for a long period both before and after the official inspection by local heath authority.  After the correct sanitation of premises, sampling performed in self-control and by the local health authority tested negative. Consequently, the deli counter activity was reopened, according to current legislation. The contamination of mortadella was most likely secondary, since the sampled product was already open and no data relating to an intact package of the same batch were available.
Based on that, it could be concluded that the source of contamination in the investigated area was eliminated, and that the actual food source of contamination remains to be investigated. Beyond the strictly sanitary purposes and given the molecular results that are not completely conclusive, further investigations on other probable contamination sources would probably have been useful, in order to assess if the contaminated food found in the inspections was the only one among the products consumed by the mother.
Since the used methodologies have highlighted tools-related results biases, the use of more than one method for analysing data related to an outbreak would be recommended, to avoid errors of underestimation or overestimation of the diversity between the involved strains. Our results also show the possibility that the cut-off of seven allelic differences represents a criterion that is sometimes too stringent for the definition of an epidemic cluster of L. monocytogenes and also suggest other factors, such as the frequency of the investigated sequence type and the presence of other strong epidemiological links.
Herein is described the first outbreak linked to the ST5 in Italy, which also involved a pregnancy-related case. This specific ST, based on national databases, is not common in Italy and seems to be limited in central Italy, in particular the Lazio region (source ISS database). Concerning Europe, sporadic occurrences of ST5 have been reported in literature. Several cases, mostly linked to maternal-fetal listeriosis, were reported during 1992-2003 in France [58].
An investigation into an Austrian cheese processing facility [59] detected the predominance of ST5 during a two-year monitoring of the processing environment (2010-2012). A study carried out in Poland between 2014 and 2017 of different kinds of ready-to-eat food of animal origin and food processing environments showed a high prevalence of ST5, found in 22.9% of samples, all derived from pork meat sausages and their processing plants [60]. Outside Europe, an outbreak was reported in the USA, linked to two ice-cream facilities. Clinically isolated, food and environmental samples all showed the ST5 [61]. Furthermore, this sequence type was found to be very frequent in Asia, especially in the central area of China [62], and was identified in clinical and pregnancy-related cases in Beijing [63]. The importance of reporting outbreaks linked to ST5, in particular when pregnancy-related, is reinforced by the sporadic occurrence of this ST worldwide.
Foods purchased from store delicatessen counters are a risk factor for sporadic listeriosis. Turkish delicatessen meat, pasteurized butter, and hot dogs have all caused outbreaks of L. monocytogenes in the past [18]. This experience shows that continued education about dietary precautions for all pregnant women is required to further reduce the occurrence of listeriosis. In addition, clinicians should consider the possibility of listeriosis whenever a pregnant woman presents with non-specific or flu-like symptoms because early diagnosis and appropriate antibiotic treatment can be lifesaving and significantly improve pregnancy outcome [18].
Finally, this study highlights the importance of a constant surveillance activity of foodborne diseases carried out by health authorities together with clinicians and laboratorians, from notification to investigation. The major aim is to identify and stop the source of contamination as soon as possible, to avoid the extension of a foodborne outbreak.

Food and Environmental Sampling
After the assessment of the epidemiological questionnaire completed by the mother, the Veterinary Service of Hygiene for Foods of Animal Origin (SIAOA) of the local health authority conducted sampling at the butcher shop and the deli counter of the attended supermarket involved. Sampling of horse meat and mortadella was performed according to the Reg. (CE) 2073/2005. Briefly, by means of sterile scalpels four aliquots were aseptically collected, each consisting of five sample units of 120 g, and stored into sterile bags at +4 • C. Environmental sampling was performed by means of sterile swabs and containers. In particular, two meat slicers (E, G), the worktable for handling cold cuts (F), and a refrigerated counter for exposed and store-cold cuts (H) were sampled, each with more than one swab. All the collected samples and times of sampling are summarized in Table 2.

Microbiological Methods for Bacterial Identification and Serotyping
A blood sample of the newborn was collected, inoculated in Bactec Plus aerobic and anaerobic broth, and incubated in the automated culture system, Bactec 9420 (Becton Dickinson, Inc., Sparks, MD, USA). The positivity was assessed upon direct observation under the optical microscope of motile bacilli and stained with Gram dye. An aliquot of the sample was plated using WASP ® automated specimen processing (Copan srl, Brescia, Italy) on Columbia Agar, Columbia CNA Agar, Mac Conkey II Agar, Mannitol Salt Agar, BBL CHROMagar Candida Medium and Chocolate Agar (Becton Dickinson, Inc., Sparks, MD, USA) and incubated at 37 • C for 24 h. Therefore, the multiplex PCR was performed on an aliquot of the positive sample (FilmArray ® bioMérieux SA, Marcy-l'Etoile, France). This technique resulted in the detection of 33 pathogens and 10 antibiotic resistance genes one hour after blood culture positivization. The identification of L. monocytogenes was also confirmed, with a score ≥2, by Maldi-TOF mass spectrometry (Bruker Daltonik GmbH, Bremen, Germany) performed on plate-cultured colonies after 24 h of incubation.
All the isolated clinical strains were sent to the Regional Reference Centre for Pathogenic Enterobacteria (CREP) at the Food Microbiology Unit of IZSLT-central division of Rome.
The food and environmental samples collected by LHA were sent to Food Microbiology Unit laboratories of IZSLT. The detection of L. monocytogenes was performed by Real-time PCR using iQ-Check Listeria monocytogenes II PCR Detection Kit (Bio-Rad Laboratories, Inc., Segrate, Italy). Positive samples were cultured according to UNI EN ISO 11290-1:2017, using ALOA medium and LSM as the second selective medium. L. monocytogenes identification was confirmed by VITEK ® 2 Compact System (bioMérieux SA, Marcy-l'Etoile, France). The biochemical culture confirmation was carried out performing the beta-hemolysis test, the L-rhamnose test, and the D-xylose test.
All the isolated strains were serotyped by using the Mast assure antiserum Listeria 'H' and 'O' (Mast Group) according to the manufacturer's instructions at CREP laboratory of IZLST. Antimicrobial susceptibility testing (AST) for all the clinical strains was performed as minimum inhibitory concentration (MIC) determination by broth microdilution (Trek Diagnostic Systems, Westlake, OH, USA), according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST; http://www.eucast.org (accessed on 28 April 2022), using epidemiological cut-offs. The following drugs were tested: Meropenem, Ampicillin, Erythromycin. Streptococcus pneumoniae ATCC 49619 was used as a quality control strain.

Whole Genome Sequencing and In Silico Analysis
For whole genome sequencing analysis, one colony was selected for each food sample unit tested positive and one colony for each swab used for environmental sampling (Table 2). Genomic DNA was extracted with the automatic extraction system QIAsymphony (Qiagen, Hilden, Germany). Libraries were prepared using Nextera XT DNA Library Prep and pair-end (2 × 300 bp) run with a MiSeq sequencer (Illumina, CA, USA).
The cgMLST analyses were performed using two pipelines. The first one was cgML-STFinder (v1.1 developed by the Center for Genomic Epidemiology), that runs KMA [74] against the Listeria cgMLST database [75] to create core genome allele profiles from raw sequence data (1748 highly conserved core genes). This software assesses the presence of only already known alleles in genomes and considers the possible new one as not found.
The second adopted pipeline was chewBBACA (v2.8.5) [76] based on the same Listeria cgMLST database [75]. This software uses the tool Prodigal [47] to search for already known alleles, but also assigns new alleles if they are not identical. The newly identified cgMLST profiles of the tested samples were added at the Pasteur Institute BIGSdb-Lm. Minimum spanning trees of both results were generated using the MSTreeV2 algorithm in GrapeTree (v1.5.0) software [77].
The SNPs and core genome SNPs' analysis was performed with the pipeline CSI phylogeny (v1.4) accessible from the Center for Genomic Epidemiology [78], starting from trimmed reads and from the core genome found by chewBBACA. The SNPs were filtered according to parameters: a minimum distance of 10 bp between each SNP, a minimum of 10× depth and 10% of the breadth coverage, the mapping quality was above 30, the SNP quality was higher than 25. The SNP matrix was visualized in a heatmap using ClustVis web tool [79].
Identification of acquired antibiotic resistance genes and genes encoding for persistence (metal, detergent resistance, stress Survival Islands and Listeria Genomic Islands) was assessed starting from the assemblies using BLAST against genes database available on the BIGSdb-Lm platform (accessed on 13 January 22). The same method was used to assess the presence for virulence and antimicrobial resistance genes of each strain.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/pathogens11060667/s1, Figure S1: Core genome MLST characterizations of isolates from the main investigation and other ST5 food samples, assessed with the two softwares cgMLSTFinder (A) and chewBBACA (B); Figure S2: SNPs matrix-based heatmap showing the variation in core genome SNPs between the strains considered in this study; Table S1: Schematic summary of the presence (+) and the absence (-) of the principal stress response genes found in the strains.

Informed Consent Statement:
All the involved institutions agreed to the publication of the data presented in the study. For the main case (A) an informed consent was signed. No informed consent for the other cases were needed since it is a descriptive case series.

Data Availability Statement:
The authors confirm that all the supporting data, code and protocols have been provided within the article or through supplementary data files. Raw reads can be found in Sequence Read Archive (SRA) at GenBank database (NCBI) under the BioProject PRJNA814492, BioSamples from SAMN26555605 to SAMN26555617.