Prevalence of ESBL, AmpC and Carbapenemase-Producing Enterobacterales Isolated from Raw Vegetables Retailed in Romania

(1) Background: As β-lactamase-producing Enterobacterales are no longer exclusively associated with the health care system, investigating the potential risk they pose to the integrity of the environment and food safety has become of utmost importance. This study aimed to determine the prevalence of extended-spectrum β-lactamase (ESBL), AmpC, and carbapenemase-producing Enterobacterales isolates from retailed raw vegetables and to determine if household washing is an effective method of lowering bacterial load; (2) Methods: Seasonal vegetables (n = 165) were acquired from supermarkets (n = 2) and farmer markets (n = 2) in Romania. Following sample processing and isolation, identification of Enterobacterales was performed by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF). Polymerase chain reaction (PCR) multiplex was used to ascertain the presence of the main ESBL, AmpC, and Carbapenemase genes. Phenotypic antibiotic resistance profiles of isolates were determined by extended antibiograms. Enterobacteriaceae colony-forming units (CFU) counts were compared between vegetable types; (3) Results: Beta-lactamase producing bacteria were observed on 7.9% of vegetables, with 5.5% displaying ESBL/AmpC phenotype and 2.4% identified as Carbapenemase producers. The most frequently detected β-lactamase genes were blaSHV (n = 4), followed by blaCTX-M and blaTEM (each with n = 3). Phenotypic antibiotic resistance analysis showed that 46% of isolates were multiple drug resistant, with aminoglycosides (38.5%) the most prevalent non-β-lactam resistance, followed by first-generation quinolones (38.5%). (4) Conclusions: The present study has described for the first time the presence of β-lactamase producing Enterobacterales in fresh produce retailed in Romania.


Introduction
The "antibiotic era", ushered in by Paul Ehrlich and Alexander Fleming, marked the beginning of a golden age for medicine. Since then, antibiotics, their development, and use, have become vegetables and vegetable products [19], raises the need to investigate the potential of fresh produce to harbor and disseminate antibiotic resistance genes.
The purpose of the present study was to determine the prevalence of ESBL, AmpC, and carbapenemaseproducing Enterobacterales from fresh vegetables in Romania.

Sample Collection
In order to determine the prevalence of ESBL, AmpC, and carbapenemase-producing Enterobacterales on fresh produce, during April-May of 2019, a total of 165 vegetables were acquired from supermarkets (n = 2) and farmer markets (n, =; 2) in the city of Cluj-Napoca, Romania. The 11 types of vegetables that were sampled from each market included: radish (n = 15), spring onions (n = 15), cabbage (n = 15), cucumbers (n = 15), cornichons (n = 15), lettuce (n = 15), spinach (n = 15), carrots (n = 15), parsley (n = 15), peppers (n = 15) and, tomatoes (n = 15). Only vegetables that were labeled as "grown in Romania" were chosen. Selection of produce was based on the dietary preferences of the Romanian population. Samples that had visible damage were excluded from the study. Vegetable samples were processed within 3 h of their acquisition.

Enterobacteriaceae Colony-Forming Unit Count
Two grams of each vegetable sample (pulp and peel) was mashed with a sterile mortar and pestle. Each mashed sample was added to 18 mL of peptone water. Following this, one µL of the vegetable mixture was diluted in 1 mL sterile saline. One µL of this dilution was then inoculated onto an EBM agar plate and incubated for 24 h at 37 • C. A colony-forming unit (CFU) count was performed by two individual researchers, using the ImageJ software version 1.52 r (Laboratory for Optical and Computational Instrumentation, University of Wisconsin, Madison, USA). The dilution was performed to allow accurate CFU counts on the agar plate.

Isolation and Identification of Enterobacterales
After the sample for the CFU count was taken, peptone-enriched samples were incubated for 24 h at 37 • C. Five µL of the enriched samples were streaked onto Brilliance UTI Agar plates (Oxoid, Johannesburg, SA) in order to allow for morphological and color-coded distinction of individual bacteria according to the specifications of the producer. A maximum of 6 morphological distinct colonies were recovered from each plate, isolated and then purified. Identification of isolates was performed by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF) (Bruker, Bremen, Germany). Only identifications that displayed a single result with a confidence score ≥2 were considered acceptable. Non-Enterobacterales isolates were not included in further analysis.
Each of the isolates phenotypically identified as ESBL, AmpC, and carbapanemase producers were placed in a BHI (brain heart infusion) broth and allowed to incubate overnight at 37 • C. A 7% solution of dimethyl sulfoxide (DMSO) was added to each liquid culture (1:3 ratio). The isolates were stored at −80 • C until molecular processing (maximum six months) [22].

Genomic DNA Extraction and Polymerase Chain Reaction (PCR) Amplification
PCR multiplex was used to ascertain the presence of the main ESBL, AmpC, and carbapenemase genes. Single colonies of each isolate were cultured in BHI broth for 18 h at 37 • C. The bacterial cells were pelleted by centrifugation (14,500 g for 5 min). DNA extraction was performed using the Wizard ® Genomic DNA Purification kit (Promega, Madison, USA), and DNA concentration was determined using the Eppendorf BioPhotometer plus (Eppendorf AG, Hamburg, Germany). Total DNA (1 µL) was subjected to each multiplex PCR in a 25 µL reaction mixture containing 12.5 µL PCR Master Mix (ThermoFisher Scientific, Johannesburg, SA), 1 µL bovine serum albumin (ThermoFisher Scientific, Johannesburg, SA), 8.3 µL specific primer and 2.2 µL nuclease-free water. Amplification was carried out as follows: initial denaturation at 94 • C for 10 min; 30 cycles of 94 • C for 40 s, 60 • C for 40 s and 72 • C for 1 min; and a final elongation step at 72 • C for 7 min [23]. Amplicons were visualized after running at 100 V for 1.5 h on a 2% agarose gel (MetaPhorTM Agarose, Lonza Bioscience, Basel, Switzerland). A 100 bp DNA ladder (ThermoFisher Scientific, Johannesburg, SA) was used as a size marker. Specific primers for each of the genes were ordered from ThermoFisher Scientific and are described in Table 1 [24]. Positive and negative controls were used to test the multiplex PCR protocol Figure S1.

Statistical Analysis
Statistical analysis was performed using GraphPad Prism software version 8.4.3 (GraphPad, CA, USA). Descriptive results were presented as mean and standard deviation (SD). In order to assess the hypothesis that some vegetable types could routinely harbor a more diverse selection of bacteria than others, the average number of distinct bacterial colonies isolated from each vegetable type was determined and compared. To achieve this, one-way ANOVA was used, followed by Tukey's post-hoc tests for multiple comparisons. Finally, an evaluation of the total Enterobacteriaceae CFU count of vegetables in relation to the origin of the sample (farmer's market vs supermarket) was conducted. This was achieved by applying either a parametric test (unpaired t-test) or a non-parametric test (Kolmogorov-Smirnov test), depending on the normal or skewed distribution of the compared series. The threshold for statistical significance was set at α = 0.05. Table 1. Group-specific primers used for polymerase chain reaction (PCR) multiplex assay [24].

Bacterial Diversity
A total number of 6 bacterial colony types were identified on Brilliance UTI agar, based on the instructions provided by the manufacturer. For each vegetable sample, we recorded the number of distinct colony types observed and isolated. Tomato samples showed the lowest amount of contamination, yielding an average of 2.2 distinct colony types, followed closely by bell pepper and spring onion samples with 2.4 as can be noted in Figure 1. On the other hand, gherkin, spinach, and parsley exhibited a more diverse flora, generating an average of 3.9, 4.1, and 4.9 distinct colonies per sample, respectively. High statistical significance was observed when comparing tomatoes to the above mentioned (p < 0.0005, p < 0.0001, p < 0.0001 respectively). Bell peppers and spring onions showed highly similar results (p < 0.005, p = 0.0001, and p < 0.0001 respectively). Overall, leafy vegetables harbored more morphologically distinct bacteria than the other types of fresh produce included in the study.

Enterobacteriaceae Colony-Forming Unit (CFU) Count
All vegetable samples exhibited similar Enterobacteriaceae CFU counts Table 2. All in all, vegetable samples presented similar Enterobacteriaceae CFU counts whether they were acquired from supermarkets or farmer's markets Figure 2. No relevant statistical significance was observed, with the exception of tomato samples, which showed higher CFU counts in farmer markets (4.8 log) compared to supermarkets (2.6 log) (p < 0.005).

Bacterial Species
Vegetable Sample Store Type Antibiotic Resistance Phenotype Genetic Determinants

Antibiotic Resistance Genotype Profile
Genes encoding β-lactamases were detected in 69.2% (9/13) of isolates obtained from fresh produce. The most frequently detected b-lactamase genes were bla SHV (n = 4), followed by bla CTX-M (n = 3), bla TEM (n = 3), bla KPC (n = 2), bla DHA (n = 1) and OXA-48 (n = 1).Five isolates harbored more than one gene; E. coli displayed the bla CTX-M gene in association with bla TEM , two isolates (S. marcescens and P. vulgaris) carried the bla TEM and bla SHV genes, one E. cloacae presented bla CTX-M with bla SHV , and lastly, K. oxytoca with bla KPC and bla SHV Table 3. Despite repetitive PCR assays, in 4 of the samples phenotypically identified as β-lactamase producers, the resistance genes could not be determined.

Antibiotic Susceptibility Patterns
As shown in Table 3, 46% (6/13) of isolates presented MDR profiles. The most prevalent non-βlactam resistance was against aminoglycosides (9/13 isolates), followed by first-generation quinolones (5/13). The E. coli isolate was the only one to present resistance to tetracyclines. Similarly, resistance to sulfonamides was observed in one E. cloacae isolate.

Discussion
The presence of bacteria with increased antimicrobial resistance on fresh vegetables has been receiving increased attention from the medical community, the food industry, and the general public. To our knowledge, there is little to no information regarding the presence of AR-E from raw vegetables retailed in Romania. In order to fill this gap in knowledge, the principal aim of the present study was to detect and record the prevalence of ESBL, AmpC, and carbapenem-producing Enterobacterales in raw vegetables grown and distributed in Romania. To achieve this, 165 vegetable samples (11 different types) purchased from farmer markets and supermarkets in Cluj-Napoca, Romania were analyzed. Beta-lactamase producing bacteria were observed on 7.9% of vegetable samples, with 5.5% displaying ESBL/AmpC phenotype and 2.4% identified as carbapenemase producers. The vegetables that displayed the highest contamination rates with β-lactamase producing bacteria were carrots and spinach (20% of samples contaminated), followed by cucumber, lettuce and parsley (13.3%) and lastly cabbage (6.6%). The percentage of ESBL/AmpC is similar to that observed by Reuland et al. (2014) [23] on retail vegetables from the Netherlands (6%) but considerably lower than the 25.4% reported by Zurfluh et al. (2015) [25] for vegetables imported by Switzerland. Richter et al. (2019) also reported a high prevalence (17.4%) of ESBL/AmpC bacteria isolated from raw vegetables in Gauteng Province, South Africa [26]. Regarding carbapenem-resistant Enterobacterales, Liu et al. (2018) reported a similar prevalence in ready-to-eat vegetables from China, as 10 out of 411 samples analyzed were contaminated [27].
While in the past, studies on fresh produce have identified mainly environmental bacteria carrying chromosomal β-lactamase resistance genes [5,28], recent studies paint a very different picture, with a high prevalence of pathogenic bacteria (Salmonella spp.) and opportunistic pathogens such as E. coli, Klebsiella pneumoniae, Citrobacter spp. and Enterobacter spp. [8,29,30]; bacteria that are not only able to cause community-acquired human infections, but often carry resistance genes located on transmissible plasmids. Similar to recent publications, in the present study the overwhelming majority of β-lactamase producing bacteria were Enterobacter spp. (n = 6), followed by Citrobacter spp. (n = 2) and E. coli, S. marcescens, M. morganii, K. oxytoca, P. vulgaris (each with n = 1). It is important to note that a recent study by Farkas et al. (2019) concerning the antibiotic resistance profile of Enterobacteriaceae showed that the most prevalent pathogenic bacteria isolated from humans in the city of Cluj-Napoca, Romania, were E. coli and Enterobacter spp. [31].
ESBLs include CTX-M-, TEM-, SHV-, and OXA-type enzymes [33]. Their prevalence and distribution have shifted since the 1980s when 3rd-generation cephalosporin resistance among Enterobacterales was mainly due to TEM-and SHV-type ESBLs [34]. For the last 20 years, CTX-M has dominated SHV and TEM, and is currently the most frequent ESBL worldwide [35]. However, due to the variance in antibiotic consumption, therapies, and strategies employed, the prevalence of ESBLs tends to differ from country to country. In Romania, CTX-M is the most common type of ESBL isolated from humans in clinical settings. Maciuca et al. (2015) also highlighted that E. coli isolated from poultry in Romania had the highest international prevalence of CTX-M-15, with 53% of isolates carrying this gene [36]. In the present study, 5 isolates were identified as ESBL producers. Genotypic characterization by PCR assays revealed bla SHV to be the most frequently isolated ESBL, followed by bla CTX-M and bla TEM . This result is quite different from previous studies from other countries that detected either bla CTX-M [37][38][39][40] or bla TEM [41,42] as the predominant ESBL gene in their isolates. Although according to phenotype, 4 isolates were identified as AmpC producers, genotyping was successful for only one E. cloacae isolate which presented the bla DHA gene. Enterobacter spp. carrying bla DHA AmpC β-lactamases have been previously described in isolates originating from fresh produce [37] and humans [43]. As previously mentioned, Romania is currently facing a rapid emergence of carbapenem-resistant bacteria, all of the major carbapenemases being previously detected in K. pneumoniae isolates collected from Romanian patients [44]. Thus, the present study also analyzed the prevalence of carbapenemase-producing Enterobacterales from fresh produce. A total of 4 isolates were identified as carbapenemase producers; PCR assay genotyping revealed a K. oxytoca isolate to harbor bla KPC , E. cloacae with OXA-48 and M. morganii displaying bla KPC . Previous studies have also isolated carbapenemase-producing Enterobacteriaceae from seafood [45], animals [27], retail meat [46] and fresh vegetables [27,47,48].
The presence of bacteria pertaining to the Enterobacteriaceae family is used to assess the general hygiene level of various foods. The Health Protection Agency (United Kingdom of Great Britain) has set a guideline for assessing the microbiological safety of ready to eat foods, classifying Enterobacteriaceae count in 3 levels: unsatisfactory (>4 log CFU/g), borderline (2-4 CFU/g) and satisfactory (<2 CFU/g) [49]. Although their presence on fresh produce is expected and generally not considered hazardous, high numbers can be an indication of poor practices [49]. The Enterobacteriaceae population levels in the present work ranged from 4.51 log (tomatoes) to 5.51 log (carrot), results which are in accordance with previous publications that have analyzed the microbial load of fresh vegetables [50,51]. While this does not necessarily reflect poorly on the microbiological safety of the vegetables analyzed, it has been previously observed that a high Enterobacteriaceae count is associated with the application of manure and/or wastewater prior to harvesting, and excessive handling during harvest, shipment and distribution [52].
Another aim of this study was to determine if high microbial load in vegetables is correlated with a broader microbial diversity. For this, as previously described, a maximum of 6 morphologically distinct colonies were isolated and purified from each sample. It is important to note that a single parsley sample yielded 6 types of colony. The vegetables that displayed the lowest number of distinct bacteria were tomato (an average of 2.2 distinct colony types), bell pepper (2.4), and spring onion (2.4). Incidentally, these also registered the lowest CFU counts of all vegetable types. This can be attributed to the fact that tomatoes and bell peppers grow above ground and come in contact less often with manure and wastewater that is used as fertilizer, thus limiting their contamination with fecal Enterobacteriaceae. In the case of spring onions, it has been previously demonstrated that their juices and vapors are able to inhibit the growth of numerous bacteria including E. coli, Bacillus cereus, Staphylococcus aureus, and Salmonella spp. [53]. Leafy and bulb vegetables displayed the highest microbial diversity which was correlated with high CFU counts. In addition to possible contamination with various bacteria found on the soil, due to their structure, leafy vegetables are more vulnerable to microdamage caused by environmental factors (e.g., wind, heavy rainfall). These lesions in turn can act as an entry portal for bacteria that attach preferentially to cut edges [54].
Lastly, we compared vegetables acquired from supermarkets and those from farmers' markets in order to discern if there are any meaningful differences from a microbiological point of view. Vegetable samples displayed similar microbial loads whether they were acquired from supermarkets or farmers' markets. The only exception were tomatoes, which displayed an average of 2.2 log decrease in CFU count in samples from supermarkets. This could be attributed to practices that are often employed by supermarkets in order to prolong shelf-life and increase the visual appeal of produce, such as thoroughly washing and applying edible coatings on produce. Tomatoes in retail markets are often waxed in order to prevent external microbial contamination, moisture desorption/absorption, all the while maintaining their organoleptic properties for extended periods of time [55]. Furthermore, various substances that act as natural antimicrobials (such as essential oils) are being incorporated increasingly in edible coatings [56]. Regarding AR-E, 61.5% of isolates originated from farmers' markets and 38.5% from supermarkets. Given the present results, we could not find any statistically significant difference in the microbiological quality of vegetables available in farmers' markets and supermarkets.

Conclusions
The present study has described for the first time the presence of β-lactamase producing Enterobacterales in fresh produce retailed in Romania. The results obtained indicate that further investigation of different vegetable types from other regions of Romania is necessary. These findings suggest that under certain circumstances, produce that is often consumed raw and considered beneficial for the human body, can pose serious health risks. In order to properly ascertain the true impact of the increasing prevalence of antibiotic resistance genes along the food chain, future studies should focus on identifying the effects and occurrence rate of human colonization with ARBs originating from fresh produce.