Occurrence and Molecular Characterization of Multidrug-Resistant Vegetable-Borne Listeria monocytogenes Isolates

Fresh vegetables play a significant role in the human diet. However, ready-to-eat (RTE) vegetables have been associated with increasing foodborne outbreaks including L. monocytogenes, which is a common human pathogen associated with foodborne infections resulting in listeriosis. This study aims to assess the resistance of vegetable-borne L. monocytogenes to antibiotics. L. monocytogenes was isolated and molecularly characterized using polymerase chain reaction (PCR) from 17 RTE vegetable samples. The confirmed L. monocytogenes was further assessed for phenotypic and genotypic antibiotic resistance using the disc diffusion test and PCR primers targeting six antibiotic classes and thirty-one related antibiotic resistance genes (ARGs), respectively. The results revealed that Listeria counts ranged from 1.60 to 3.44 log10 CFU/g in the samples. The isolates exhibited high resistance against penicillin G, erythromycin, vancomycin, tetracycline, trimethoprim-sulfamethoxazole, and nitrofurantoin among the 108 isolates tested. A total of 71 multiple antibiotic resistance (MAR) phenotypes were observed in the isolates, which ranged from resistance to 3 to 13 antibiotics. The MAR index was ˃0.2 in 97% of the isolates. Some of the highly detected ARG subtypes included SulI (100%), TEM (76.9%), tetA (59%), and tetM (54.7%). The findings show a high occurrence of multidrug-resistant L. monocytogenes and clinical ARGs in fresh vegetables, which constitutes an immediate danger for the health security of the public.


Introduction
Listeria monocytogenes is a common Gram-positive facultative anaerobic bacillus that is recognised as one of the most important foodborne pathogens by the World Health Organisation [1]. Based on latest classification, the genus Listeria is comprised of 20 major species [2] and is organised into two groups centred on their relationship with L. monocytogenes [3]-a known pathogen of the human foodborne infection, listeriosis [4]. L. monocytogenes is an invasive pathogen that can infect human and animals. It can cause severe listeriosis, leading to meningoencephalitis, sepsis, and foetal infection or miscarriage in pregnant women, with a mortality rate of 20-40% [1]. Antibiotic therapies usually employed in the treatment of listeriosis include penicillin, ampicillin, gentamicin, rifampicin, chloramphenicol, erythromycin, tetracycline, or trimethoprim with sulphamethoxazole in combination or alone [5,6]. However, most L. monocytogenes strains possess natural

Incidence of Vegetable-Borne Listeria monocytogenes
A total of 189 presumptive Listeria isolates were screened for the presence of the prs and prfA genes for the identification of Listeria genus and L. monocytogenes, respectively, using a duplex PCR assay. While 59% (112/189) of the presumptive isolates were con-

Incidence of Vegetable-Borne Listeria monocytogenes
A total of 189 presumptive Listeria isolates were screened for the presence of the prs and prfA genes for the identification of Listeria genus and L. monocytogenes, respectively, using a duplex PCR assay. While 59% (112/189) of the presumptive isolates were confirmed as Listeria spp., 57% (108/189) were confirmed as L. monocytogenes (Figure 2).

Incidence of Vegetable-Borne Listeria monocytogenes
A total of 189 presumptive Listeria isolates were screened for the presence of the prs and prfA genes for the identification of Listeria genus and L. monocytogenes, respectively, using a duplex PCR assay. While 59% (112/189) of the presumptive isolates were confirmed as Listeria spp., 57% (108/189) were confirmed as L. monocytogenes ( Figure 2).

Antibiotic Susceptibility Patterns of L. monocytogenes
One hundred and eight positive L. monocytogenes isolates were assessed for their phenotypic antibiotic resistance patterns against 16 different antibiotics ( Figure 3). All isolates showed resistance against penicillin G. High levels of resistance were equally observed against erythromycin, vancomycin, tetracycline, and trimethoprim-sulfamethoxazole.

Antibiotic Susceptibility Patterns of L. monocytogenes
One hundred and eight positive L. monocytogenes isolates were assessed for their phenotypic antibiotic resistance patterns against 16 different antibiotics ( Figure 3). All isolates showed resistance against penicillin G. High levels of resistance were equally observed against erythromycin, vancomycin, tetracycline, and trimethoprim-sulfamethoxazole. However, the isolates were found to be highly sensitive to amikacin, gentamicin, meropenem, kanamycin, chloramphenicol, levofloxacin, and amoxicillin-clavulanic acid. However, the isolates were found to be highly sensitive to amikacin, gentamicin, meropenem, kanamycin, chloramphenicol, levofloxacin, and amoxicillin-clavulanic acid.

Multiple Antibiotic Resistance Phenotypes (MARP) of L. monocytogenes
The MAR phenotype patterns and the multiple antibiotic resistance index (MARI) of all L. monocytogenes isolates are given in Table 1. The isolates exhibited a total of 71 pat-

Multiple Antibiotic Resistance Phenotypes (MARP) of L. monocytogenes
The MAR phenotype patterns and the multiple antibiotic resistance index (MARI) of all L. monocytogenes isolates are given in Table 1. The isolates exhibited a total of 71 patterns of MARPs ranging from resistance to 3 to 13 antibiotics, with a high number of single appearances. Ninety-seven percent of the isolates had MARI values that were greater than the 0.2 threshold value, which is the recommended benchmark for MARI as set by Krumperman [15].

Discussion
Although only 17 vegetable samples were investigated for Listeria spp. and L. monocytogenes in the study, multiple isolates were picked from single samples to explore the variability among isolates that could be cultured from the same vegetable type-which also allowed for the investigation of diversity in susceptibility to antibiotics that may be present among the isolates.
The Listeria species detected and identified from the vegetable samples ranged from 3.04-4.38 log 10 CFU/g. This suggests that the vegetables could have been contaminated on farms via irrigation water and other farm practices. Additionally, the count in the samples exceeded the 2 log 10 CFU/g (100 CFU/g) maximum limit of L. monocytogenes for low-risk foods, as well as the "zero-tolerance" limit for ready-to-eat (RTE) foods. Thus, the vegetables are unsafe for raw consumption without thorough preparation and must be properly cooked before consumption. Public Health England (2014) reported that food samples containing ≥100 CFU/g (2 log 10 CFU/g) of Listeria species are considered of unsatisfactory microbial quality and should be further investigated. Therefore, the vegetable samples studied pose a threat to consumers' health and make them prone to the risk of contracting listeriosis from consumption of contaminated vegetables. However, the range of Listeria counts from the vegetables studied was lower compared to the 2.98-5.32 log 10 CFU/g) previously reported by Bilung and colleagues [5] in vegetables.
The Listeria genus and L. monocytogenes contamination in the vegetables studied was 59% and 57%, respectively. The incidence of Listeria species from vegetables in this study was higher than that reported by Bilung and colleagues (6.7-8%) [5], Goni and colleagues (21%) [16], and Sangeetha and Shubha (1.81%) [17]. However, the Listeria incidence was lower than the Listeria incidence (69.2%) reported by Onyilokwu and colleagues in vegetables [18]. The occurrence of L. monocytogenes known to cause listeriosis in both humans and animals [19], at 57% among the vegetable-borne Listeria isolates, could be linked to pollution of irrigation source water with effluents from poultry and other husbandries-as seen upstream of the farm where some of the vegetables were collected. Similarly, irrigation water pollution or organic fertilizer usage might be prevalent in the farms that supply supermarkets and street vendors with vegetables in the catchment. L. monocytogenes contamination of vegetables has been found to occur through irrigation water used in fresh produce cultivation via surface contamination and internalization through the roots and subsequent survival in crops [20]. Raw vegetables are generally prone to L. monocytogenes contamination [21,22]. The 57% incidence of L. monocytogenes among the vegetable-borne isolates in this study is higher compared with the L. monocytogenes incidence of 4.18% and 13.6% reported by Moreno et al. [23] and Cetinkaya et al. [24], respectively, in vegetables.
The 108 L. monocytogenes profiled for antibiotic susceptibility showed resistance against penicillin G (100%), followed by erythromycin (98.2%), vancomycin (94.5%), tetracycline (80.7%), trimethoprim-sulfamethoxazole (78.9%), and nitrofurantoin (54.1%). This is an indication that the isolates originated from sources where high level of antibiotics have been used. Additionally, L. monocytogenes is known and capable of acquiring resistance against erythromycin and tetracycline from lactic acid bacteria [6], which are common in decaying vegetables and plant materials on farms as well as in food materials. The low resistance of the L. monocytogenes against gentamicin, one of the antibiotics of choice for the treatment of listeriosis in South Africa, ensures hope for the treatment of its infections in the locality. The gentamicin susceptibility profiles of the isolates are comparable to those reported by Bilung et al. [5] and Li et al. [25] in vegetables. The present results demonstrate a decrease in the efficacy of some of the antibiotics against L. monocytogenes. Thus, the vegetables represent a potential source of multidrug-resistant L. monocytogenes infections in the locality. Kuan et al. [26] observed 100% resistance to penicillin G-which is line with the findings of this work-and also found gentamicin (91.4%) to be effective in restraining the growth of L. monocytogenes. The reduced susceptibility of penicillin G as a first line drug may be caused by the indiscriminate use or misuse of this antibiotic, as reported by Kuan et al. [26]. The high level of resistance to penicillin G and vancomycin is very concerning as these antibiotics are used in the treatment of listeriosis during pregnancy and in treating listerial meningitis, respectively [27]. It is also worrisome that high resistance was recorded against trimethoprim-sulfamethoxazole as this antibiotic, together with vancomycin, can be used as an alternative therapy for patients allergic to penicillin [28].
The MAR phenotypes compiled for L. monocytogenes indicated a high degree of multiple antibiotic resistance, with a range of resistance to 3 to 13 antibiotics. The most prevalent MARPs were observed in MARP 6. E-PG-T-TS-VA (n = 6, 5.5%), E-NI-PG-T-TS-VA (n = 6, 5.5%), AUG-CXM-KF-C-E-K-NI-PG-T-TS-VA (n = 6, 5.5%), and AUG-E-NI-PG-T-TS-VA (n = 5, 4.6%) were the most predominant MARPs. This is indicatory of the possible failure of combinational therapy with these antibiotics should human or animal infections occur through them. The health risk associated with the spread of antibiotic resistance in an environment is estimated using MARI [29,30], and the MARI values obtained in this study signify an overuse or misuse of antibiotics in the environment from which the samples were collected. A MARI value of 0.2 (arbitrary) was used to differentiate between a low-and high-health risk. A value that is greater than 0.2 suggests that the resistant L. monocytogenes isolates originated from an environment where there is high contamination or overuse of antibiotics [29,30]. The MARI values of most of the vegetable-borne L. monocytogenes in this study was ≥0.2. This suggests that the L. monocytogenes originated from farms using irrigation water polluted with high levels of antimicrobial substances, antibiotic residue-laden effluents from livestock farming, or the application of organic manures with high loads of antimicrobial residues and antibiotic-resistant bacteria. This could further increase the spread of multidrug-resistant Listeria isolates to humans via the consumption of contaminated vegetables. This is similar to the L. monocytogenes MARPs reported from fruits and vegetables by Kayode and Okoh (2022) [31].
This study also investigated the prevalence of antibiotic resistance genes in L. monocytogenes isolates. A total of 14/25 antibiotic resistance genes were detected across six different antibiotic classes. Some of the isolates exhibited one or more antibiotic resistance genes that may act as a pool of resistance genes for other commensal and pathogenic bacteria in vegetable farm environments [32]. SulI (100%) was the only detected gene conferring resistance to sulfanomides. Sulfanomide-resistance genes signify that the isolates originated from animal sources. The most prevalent were genes conferring resistance to tetracyclines, including tetA (59%), tetM (54.7%), tetC (43%), and tetD (43%; Table 2). In beta-lactams, bla TEM (76.9%) was among the other genes that were detected, including aminoglycosides aphA2 (41.7%), aadA (33.3%), and several other genes conferring resistance against phenicols, cephems, and aminoglycosides. However, high resistance against sulfamethoxazole (80.58%), amoxicillin (58.25%), and erythromycin (49.52%) was observed. About 85.44% of Lm isolates showed multidrug-resistant phenotypes against the tested antibiotics [31].
Conversely, some of the L. monocytogenes isolates displayed phenotypic resistance to multiple antibiotics but did not contain antibiotic resistance genes. L. monocytogenes isolates may have acquired genes for antibiotic resistance through antibiotic selection pressure or through various gene transfer mechanisms from other bacteria in the farm area [32]. Studies have shown the conjugative transfer of antibiotic resistance, i.e., the acquisition of enterococcal and streptococcal plasmids into the genus Listeria and subsequent transfer of these plasmids within the genus, including transmission to Listeria monocytogenes [33]. According to Srinivasan et al. [32], the expression of different genes of resistance to antibiotics does not always correlate with the phenotypic antibiotic resistance of foodborne pathogens.

Materials and Methods
This study was conducted in the Sarah Baartman District Municipality (SBDM), the largest district municipality in the Eastern Cape, with the geographical coordinates: 33 • 57 00" S; 25 • 36 00" E. Samples were collected from one farm, two supermarkets, and one street vendor. The selected farm, which is located near a river, supplies fresh vegetables to a number of Supermarkets across the SBDM, while the supermarkets are the most commonly used by the majority of residents from Grahamstown and its surrounds.
The samples were aseptically collected in September 2018 using sterile stomacher bags and were transported to the laboratory, where they were analysed within 6 h of collection. Ethical clearance for the research was obtained from the University of Fort Hare Research Ethics board under the reference number: OKO011SNTS01.
For enrichment, 0.1 mL of each homogenized sample was cultured in 10 mL tryptic soy broth (TSB) and incubated at 37 • C for 18-24 h (Merck, Johannesburg, South Africa). For secondary enrichment, a 0.1 mL aliquot was transferred from the TSB suspension into 10 mL Fraser broth (Oxoid Ltd., Bsingstoke, UK) and incubated at 37 • C for 48 h. A loopful of the enriched culture was streaked onto chromogenic Listeria agar (LCA; Oxoid Ltd., Basingstoke, UK), and supplemented using an OCLA (ISO) selective supplement with an OCLA (ISO) differential supplement, following the standards defined by Ottaviani and Agosti (ALOA) in ISO 11290-1:1997, and incubated at 37 • C for 24 h. After incubation, the petri dishes were observed for typical Listeria species colonies (blue/green colonies with a sunken centre), which were picked and subsequently purified onto nutrient agar (NA; Oxoid Ltd., Basingstoke, UK) and incubated at 37 • C for 24 h. Pure distinct colonies were further inoculated in sterile nutrient broth (NB) and incubated at 37 • C for 18-24 h. The overnight culture was stored at -80 • C in a 25% glycerol stock until further analysis.
DNA was extracted from reactivated bacterial cells using the boiling method [34]. The bacterial cells were reactivated by inoculation into nutrient broth (NB), followed by incubation at 37 • C for 18-24 h. A loopful of the NB culture was further streaked onto NA and incubated at 37 • C for 24 h. Single distinct bacterial colonies were homogenised in 200 µL sterile nuclease-free water contained in an Eppendorf tube and boiled using a heating block (MS2a Dri-Block DB.2A Techne, Bibby Scientific LTD, Staffordshire, UK) at 100 • C for 10 min. Homogenates were further centrifuged at 13,500 rpm for 10 min to separate the liquid from the cell debris. The supernatant was kept at 4 • C and later used as a DNA template in PCR. Primers targeting the prs (F: 5 -GCT GAA GAG ATT GCG AAA GAA G-3 ; R: 5 -CAA AGA AAC CTT GGA TTT GCG G-3 ) and prf A (F: 5 -GAT ACA GAA ACA TCG GTT GGC-3 ; R: 5 -GTG TAA TCT TGA TGC CAT CAG-3 ) genes for Listeria genus and L. monocytogenes, respectively, were used for the amplification of DNA in PCR. PCR was performed in a 25 µL reaction volume consisting of 12.5 µL master mix (Quick-Load BioLabs), 4.5 µL nuclease-free water, 1 µL forward primer, 1 µL reverse primers, 5 µL DNA, 0.5 µL MgCl 2 , and buffer. The cycle condition included initial denaturation (94 • C, 5 min), followed by 33 cycles of denaturation (95 • C, 45 s), annealing (60 • C, 30 s), extension (72 • C, 1 min), and a single final extension (72 • C, 5 min). PCR products (5 µL) were electrophoresed in 1.5% agarose gel, stained with ethidium bromide solution, and visualised under a UV illuminator (Alliance 4.7 XD-79.WL/26MX, Paris, France).
The antibiotic resistance profiles of the confirmed isolates were assessed using the sixteen antibiotic panels recommended for the clinical treatment of L. monocytogenes infections by the Clinical and Laboratory Standards Institute [35]  . From a 24 h culture of L. monocytogenes isolates on nutrient agar, a single colony was picked and inoculated into 5 mL normal saline and vortexed. The turbidity of the suspension was adjusted to 0.5 McFarland standard (corresponding to 1.5 × 10 8 CFU/100 mL) and spread-plated onto Mueller-Hinton (MH) agar (Neogen, Lansing, MI, USA) using sterile swabs dipped into the suspension. The antibiotic disks were dispensed onto the surface of the inoculated MH agar using an antibiotic disk dispenser and incubated at 37 • C for 24 h. Following incubation, the Petri dishes were examined for clear inhibition zones, which were measured (mm) and interpreted using the CLSI guidelines for Staphylococcus spp. and Enterococcus spp. as a surrogate (since resistance criteria is unavailable in the CLSI guidelines for Listeria spp.) [36].
Multiple antibiotic-resistant phenotypes (MARPs) were generated for bacterial isolates that were resistant to more than two antibiotics, using the method from Krumperman [15]. Resistance patterns, the number of antibiotics, and percentages were also recorded.
The mathematical expression from Krumperman [15] was used to calculate the multiple antibiotic resistance index (MARI) for each listerial isolate: where n is the number of antibiotics against which resistance was shown by a listeria isolate and m is the total number of antibiotics against which each listeria isolate was tested.
Three multiplex PCRs and one simplex PCR were used for the screening of 10 antibiotic resistance genes (ARGs) encoding Extended Spectrum Beta-Lactamases (ESBL), as previously described by Dallenne et al. [37]. Table 3 gives a summary of the group-specific primers and cycling conditions used for the detection of ESBL ARGs. Supplementary Material Table S1 gives a summary of all primers used for the screening of all the 31 target ARGs. The ARGs included tetA, tetB, tetC, tetD, tetK, tetM, aacC2, aphA1, aphA2, aadA, strA, bla TEM , blaZ, ampC, TEM, SHV, OXA1-like, GES, PER, VEB, ACC, FOX, MOX, DHA, CIT, EBC, cmlA1, catI, catII, SulI, and sulII. The PCR assays were either simplex, duplex, or multiplex. Briefly, 1 µL DNA was amplified in a 25 µL reaction mix made up of 12.5 µL master mix (Quick-Load BioLabs), 4.5 µL nuclease-free water, 5 µL DNA, 0.5 µL MgCl 2 , and buffer), a variable concentration of specific-group primers (Table 3 and Supplementary  Material Table S1), and 1 U Taq polymerase (Sigma Aldrich, Johannesburg, South Africa). Amplification cycles were as presented in Table 3. Amplicons were electrophoresed at 100 V for 60 min using a 2% agarose gel spiked with 2 µL ethidium bromide. A 100 bp DNA ladder (New England Biolabs, Ipswich, MA, USA) was used as a size marker.  The data obtained were subjected to descriptive analysis using frequencies. The distribution of presumptive Listeria species in vegetable samples was represented via composite violin-box plots created using the ggpubr package (https://cloud.r-project.org/ package=ggpubr, accessed on 7 June 2022) in R version 3.4.4 and compared by Kruskal-Wallis tests.

Conclusions
The presence of L. monocytogenes in fresh vegetables indicates a potential risk for consumers-especially the elderly, the immunocompromised, and pregnant women. The findings in this study indicate a high percentage of L. monocytogenes in the fresh vegetables studied and proves that fresh vegetables could be a reservoir of multidrug-resistant L. monocytogenes strains. Multidrug-resistant L. monocytogenes may serve as carriers of antibiotic resistance determinants that may be easily transferred to other bacteria in different environments, and possibly acquired by human pathogens through the ingestion of contaminated vegetables. To diminish the contamination of vegetables with multidrug-resistant microorganisms, it is imperative that vegetables are continuously monitored. It is also important to find ways to diminish the antibiotic selection pressure in order to reduce the dissemination of antibiotic-resistant foodborne pathogens.