Molecular characterization of Salmonella spp. and Listeria monocytogenes strains from biofilms in cattle and poultry slaughterhouses located in the federal District and State of Goiás, Brazil

Listeria monocytogenes and Salmonella spp. are considered important foodborne pathogens that are commonly associated with foods of animal origin. The aim of this study was to perform molecular characterization of L. monocytogenes and Salmonella spp. isolated from biofilms of cattle and poultry slaughterhouses located in the Federal District and State of Goiás, Brazil. Fourteen L. monocytogenes isolates and one Salmonella sp. were detected in poultry slaughterhouses. No isolates were detected in cattle slaughterhouses. All L. monocytogenes isolates belonged to lineage II, and 11 different pulsotypes were detected. Pulsed-field gel electrophoresis analysis revealed the dissemination of two strains within one plant, in addition to the regional dissemination of one of them. The Salmonella isolate was identified via whole genome sequencing as Salmonella enterica serovar Minnesota ST548. In the sequence analysis, no premature stop codons were detected in the inlA gene of Listeria. All isolates demonstrated the ability to adhere to Caco-2 cells, while 50% were capable of invading them. Antimicrobial resistance was detected in 57.1% of the L. monocytogenes isolates, and resistance to sulfonamide was the most common feature. The tetC, ermB, and tetM genes were detected, and four isolates were classified as multidrug-resistant. Salmonella sp. was resistant to nine antimicrobials and was classified as multidrug-resistant. Resistance genes qnrB19, blaCMY-2, aac(6’)-Iaa, sul2, and tetA, and a mutation in the parC gene were detected. The majority (78.5%) of the L. monocytogenes isolates were capable of forming biofilms after incubation at 37°C for 24 h, and 64.3% were capable of forming biofilms after incubation at 12°C for 168 h. There was no statistical difference in the biofilm-forming capacity under the different evaluated conditions. Salmonella sp. was capable of forming biofilms at both tested temperatures. Biofilm characterization was confirmed by collecting the samples consistently, at the same sampling points, and by assessing biofilm formation in vitro. These results highlight the potential risk of cross-contamination in poultry slaughterhouses and the importance of surveillance and pathogen control maintenance programs within the meat production industry.


Introduction
The hygienic and sanitary conditions of the industrial environment are essential for ensuring food safety and quality [1][2][3]. One of the current challenges encountered in slaughtering and meat processing is the prevention of cross-contamination by microorganisms [4,5]. Crosscontamination can occur via contact with bacterial biofilms due to the performance of improper hygiene and sanitization practices at the slaughter system. This leads to a decrease in food quality, resulting in pathogen transmission and the occurrence of recurring contamination of food [5,6]. Listeria monocytogenes and Salmonella spp. are deemed important pathogens whose dissemination can be controlled to ensure food quality and safety. Their presence is a criterion for evaluation within the various aspects of quality control of the meat production process [7].
L. monocytogenes is recognized as the main etiological agent of listeriosis in humans [8]. It is a psychrophilic bacterium, that can develop on surfaces of meat production industries [9]. The species is subdivided into 13 serotypes, of which 1/2a, 1/2c, and 4b are the most commonly reported serotypes involved in cases of human listeriosis [10]. Owing to the difficulties and limitations encountered with methods of traditional serology, different subtyping methods have been developed, such as the differentiation of the main lineages assessed via polymerase chain reaction (PCR) [11]. Additionally, molecular typing methods with high discriminatory power, such as pulsed-field gel electrophoresis (PFGE), are widely used to assess clonal dissemination within industries [10,12,13]. There is considerable pathogenic potential among the different L. monocytogenes clones. [14]. The role of internalin A in Listeria, encoded by the inlA gene, is exemplary as its functionality is necessary for pathogen internalization [15][16][17]. The presence of premature stop codons (PMSCs) in this gene has been associated with clones with an attenuated invasion phenotype [18].
The genus Salmonella, which is responsible for salmonellosis, includes the Salmonella enterica species. It is a microorganism that is often identified as an etiological agent of food outbreaks, presenting a complex epidemiology regarding its transmission and distribution, and is therefore important for investigation in the assessment of risks to public health. [19]. The species demonstrates adherence to surfaces such as metals, glass and rubber, and exhibits the ability of biofilm formation [20][21][22].
Antimicrobial resistance (AMR) is considered a grave concern and risk to human health [23,24]. To mitigate its risks, the World Health Organization established a global action plan [25]. One of the main strategies outlined in the plan is to monitor the spread of AMR. Considering the importance of monitoring L. monocytogenes and Salmonella spp. dissemination and the necessity of ensuring public health and safety, and owing to the limited investigations related to the presence of biofilms in Brazil, the objectives of this study were to detect biofilminduced contamination points and the presence of these microorganisms in processing plants of cattle and poultry slaughterhouses, to characterize these strains by performing serotyping,

Isolation of L. monocytogenes and Salmonella spp
For the isolation of L. monocytogenes, swab samples were analyzed according to the methodology described by Ryser and Donnely 2015 [31]. Transport tubes containing swabs were homogenized in a tube shaker, with the subsequent addition of 1 mL of transport solution to 9 mL of UVM broth (Acumedia1, USA). After 24 h of incubation at 35˚C, the culture (0.1 mL) was transferred to 9 mL of Fraser broth (Acumedia1, USA). Esculin hydrolysis was performed after incubation at 35˚C for 24-48 h. Each sample was plated on Modified Oxoid (MOX) agar (Difco1, France). Presumptive colonies were incubated in Brain Heart Infusion (BHI) broth (Acumedia1, -USA) at 35˚C for 24 h and were biochemically analyzed. Isolates with characteristics compatible with the selected genus were examined by PCR to confirm the species using the primers LIP1 [32] and LIP2A [33] for amplification of the prfA gene. PCR was performed as per methods described by Kérouanton et al. 2010 [33]. Positive controls were provided by Dr. Ernesto Hofer of the Oswaldo Cruz Foundation located in Rio de Janeiro.
To identify species of Salmonella, swab samples were analyzed as per protocols described by Ryser and Donnely 2015 [31] and ISO 6579/2002 [34]. After homogenization of transport tubes containing the collected swab samples, each homogenate was individually transferred to tubes containing 1% buffered peptone water (Acumedia1, USA). Samples were incubated at 37˚C for 24 h. Aliquots were transferred to tetrathionate broth (Merck1, Germany), selenite cystine broth (Merck1, Germany), and Rappaport-Vassiliadis broth (Merck1, Germany). After incubation in selective media, samples were plated on Brilliant Green Phenol Red Agar (Acumedia1, USA), Xylose Lysine Deoxycholate (Merck1, Germany), and Bismuth Sulfite agars (Acumedia1, USA). Presumptive colonies were cultured on Nutrient agar (Acume-dia1, USA) and were biochemically analyzed using appropriate tests [34]. This was followed by serology analysis using polyvalent anti-Salmonella somatic and flagellar sera (Probac1, Brazil). PCR amplification of the ompC gene was performed using the OMPCF and OMPCR primers [35] for confirmation of the genus [36]. Positive controls were provided by the Oswaldo Cruz Foundation.

Molecular serotyping of L. monocytogenes and Salmonella
L. monocytogenes subtyping was performed using multiplex PCR and primers lmo0737, lmo1118, ORF2819 and ORF2110, as per methods described by Doumith et al. [11] for differentiation into the four main lineages.
Pulsed-field gel electrophoresis of L. monocytogenes PFGE was performed as per guidelines recommended by the Centers for Disease Control and Prevention [39]. The restriction enzyme AscI (Invitrogen1, Lithuania) was used at a concentration of 10 U/μL. Salmonella ser. Braenderup H9812 subjected to digestion with XbaI (Roche1, Germany) at a concentration of 10 U/μL was used as the standard, and was provided by the Enterobacteria Laboratory of the Oswaldo Cruz Foundation. The BioNumerics 7.7. software (Applied Maths) was used to analyze the fragments, with a Dice similarity coefficient of 1.5%. The dendrogram was constructed by analyzing the unweighted pair group method with arithmetic mean (UPGMA) clusters.

Analysis of the inlA gene sequence of L. monocytogenes
Preto, USP. Cells were cultured as per methods described by Gaillard et al. 1987 [40]. Wells in polypropylene plates (Kasvi1) were seeded with cells at a final density of 1.0 × 10 5 cells per well. Adhesion assays were conducted as per protocols described by Moroni et al. 2006 [41]. The wells seeded with the cells were inoculated with a bacterial suspension using a volume adjusted to obtain a multiplicity of infection (MOI) of 100. After adding the Dulbecco's modified Eagle medium (DMEM) (Gibco1, USA) supplemented with 10% fetal bovine serum (Gibco1, USA), the plates were incubated at 37˚C for 2 h in a 5% CO 2 atmosphere. The wells were rinsed three times with 1× phosphate-buffered saline (PBS; Laborclin1, Brazil). The cells were then subjected to treatment with a lysis buffer solution containing 0.1% Triton-X100 (Sigma-Aldrich1, USA), and incubated for 10 min under the aforementioned conditions. Viable bacterial cells were counted by seeding serial dilutions on BHI agar (Difco1, France) incubated at 37˚C for 24 h. The counts were expressed as colony forming units (CFU)/mL. All tests were performed in triplicate. Adhesion of bacteria to Caco-2 cells was calculated (%) as (number of adherent cells × 100) / number of cells adhering to the well [41]. Invasion tests were performed as per methods described by Gaillard et al. 1987 [40] and Moroni et al. 2006 [41]. Cells were added to each well at an MOI of 100. Following incubation, the cells were subjected to washing steps with PBS as described above for the adhesion assay. After subjection to washing steps, 1 mg/mL gentamicin (Sigma-Aldrich1, Israel) was added to the wells and incubation was performed at 37˚C for 1 h in a 5% CO 2 atmosphere. The lysis buffer solution was added to each well, and viable bacterial cells were counted as described above. All tests were performed in triplicate. The percentage of bacterial invasion in Caco-2 cells was determined by using the following formula: % adhesion = (number of internalized recovered cells × 100) / number of cells adhering to the well.

Antibiogram and assessment of antimicrobial resistance genes
The AMR of L. monocytogenes isolates was evaluated by performing the disk diffusion assay using Mueller Hinton agar (Acumedia1, USA), as per methods described by the Clinical and Laboratory Standards Institute [42]. The antibiotics tested were ampicillin (10 μg), ciprofloxacin (5 μg), chloramphenicol (30 μg), doxycycline (30 μg), erythromycin (15 μg), gentamicin (10 μg), sulfonamides (300 μg), and tetracycline (30 μg). The interpretation of the zone of inhibition diameters was based on the standards prescribed for Staphylococcus spp. [43] which were defined by CLSI M100 [42], with the exception of standards for erythromycin and ampicillin, for which the standards defined by the European Committee on Antimicrobial Susceptibility Testing [44] were used for L. monocytogenes. The identification of resistance genes in L. monocytogenes isolates was performed using PCR. Genes related to the resistance to tetracycline (tetA, tetB, tetC, tetM), macrolides (ermA, ermB, ermC, ereA), amphenicols (cat1, cmlA), sulfonamides (sull), beta-lactams (ampC, blaSHV), and aminoglycosides (aac (3)-I) were investigated. Reactions were performed in a final volume of 25 μL, according to the conditions described in the studies reported in Table 1 for each pair of primers.

In vitro biofilm formation capacity
In vitro biofilm formation capacity of L. monocytogenes and Salmonella spp. isolates was evaluated by performing incubation at 37˚C for 24 h and incubation at 12˚C for 168 h (7 days). Tests were performed using polystyrene titration microplates as per methods described by Djordjevic et al. 2002 [49] and modified by Borges et al. 2018 [50]. TSB broth (Kasvi1, Italy) was used to prepare bacterial suspensions at a final density of 3 x 10 8 CFU/mL. All tests were performed in triplicate. After the performance of preparation, incubation, washing steps, staining procedures, and resuspension of each sample, the optical density (OD) of each well was measured using an ELx800 enzyme-linked immunosorbent assay reader (Biotek Instruments) at 490 nm. The OD value for each isolate was determined as the arithmetic mean of the absorbance readings obtained from the three wells. This value was compared with that of the negative control (ODn). The classification proposed by Stepanović et al. 2000 [51] was used to determine the capacity and intensity of biofilm formation. Statistical analysis of data for

PLOS ONE
Salmonella spp. and Listeria monocytogenes from biofilms in cattle and poultry slaughterhouses comparison of the capacity of biofilm formation in the two conditions tested was performed using the SAS software (SAS Inc.).Normality was analyzed using the Shapiro-Wilk test followed by the paired t-test for means comparison between the two groups.

Detection of microorganisms in poultry and cattle slaughterhouses
Detection of L. monocytogenes. A total of 14 L. monocytogenes isolates were detected from 287 swabs collected from poultry and cattle slaughterhouses (4.87%). All isolates were detected in poultry slaughterhouses (A, B, and C), which represented 169 swabs (8.28%). L. monocytogenes was not detected in any of the 118 samples collected from the cattle slaughterhouses (D and E). The detection points for the isolates are listed in Table 2. In the Federal District and State of Goiás region, reports have described the presence of L. monocytogenes isolated from different sources [52,53]. These findings support the results of the present study. The absence of L. monocytogenes in cattle slaughterhouses differs from the findings reported by Palma et al. 2016 [53], who described the presence of L. monocytogenes in samples obtained from cattle slaughterhouses. The difference between the studies may be related to the small number of industries available and the sampling visits that were conducted, which was due to the hesitation of the facilities to participate in the study. The difference may also reflect the adequate hygiene and sanitation conditions prevalent in these establishments.
The detection of L. monocytogenes in poultry slaughterhouses is similar to the findings reported in previous studies described in Brazil [4,[54][55][56]. The detection of L. monocytogenes at the sampling points described in the present study agrees with previously reported detection on stainless steel tables [4,55], mats [4], chutes [56], and drains [12,28]. The usual definition of persistence in the context of foodborne pathogens is the repeated isolation of strains that are identical subtypes identified on different days [2]. In this study, strains were considered persistent when detected at the same sampling location after at least one sanitation process. All collections were performed after the environment was subjected to sanitation procedures at least once, suggesting the persistence of the microorganism, possibly due to the ability to adhere to surfaces and to form biofilms [57][58][59]. At one collection point (a chute in slaughterhouse C), L. monocytogenes was detected in two samples collected on different days, suggesting the presence of biofilms at the site. In two different studies conducted in the United States, Berrang et al. [28,60] performed serial collection from the same points within the slaughter plant at different times to characterize the presence of persistent L. monocytogenes isolates in the environments of poultry slaughterhouses. The authors concluded that the persistence of this microorganism in drains was related to the formation of biofilms [60]. Similar to the methods described in the present study, Sereno et al. 2019 [61] performed serial swab collections of walls, drains, tables, mats, and floors in a swine slaughterhouse in the southern region of Brazil. The persistence of L. monocytogenes strains was detected at the same sampling point over time and was related to the presence of biofilms at said location. Areas of bacterial persistence include drains [28, 60] and a mat [61]. The present study also detected the presence of L. monocytogenes at these sites, although there was no detection at the same site more than once, except for the chute.

Detection of Salmonella spp.
One Salmonella sp. was detected in 287 (0.34%) swabs collected from the cattle and poultry slaughterhouses. The isolate originated from a drain sample that was collected from a dirty area of poultry slaughterhouse B. This is the first investigation of Salmonella spp. in a processing plant of a poultry slaughterhouse located in the Federal District. However, other reports

PLOS ONE
Salmonella spp. and Listeria monocytogenes from biofilms in cattle and poultry slaughterhouses have described the occurrence of Salmonella spp. in poultry carcasses/viscera produced and/or sold in this region [36]. The low detection in poultry slaughterhouses and absence in cattle slaughterhouses may reflect the efficiency of self-control, disinfection, and sanitation programs in these industries, the reduced number of visits resulting from the hesitation of industries to allow internal access for sample collection, and the implementation of control programs by the government with the aim of reducing the incidence of Salmonella spp. to protect consumers from considerable risks [62][63][64].

Serotyping analysis by PCR and evaluation of genetic variability of Listeria monocytogenes by PFGE
All 14 isolates were classified via PCR as 1/2a or 3a serotypes, belonging to lineage II [65][66][67], with amplification of the lmo0737 gene (691pb). These results corroborate reports available in the literature that highlight serotype 1/2a, along with serotypes 1/2b and 1/2c, as one of the most commonly detected serotypes in food samples and food production environments of animal origin [68]. The polymorphic results obtained via PFGE identified 11 distinct pulsotypes that were grouped into four clusters (I, II, III and IV) with a similarity threshold of 80% [69] (Figs 1 and S1). The sample collection points of origin from each pulsotype are shown in Fig 2. To analyze strain dissemination within industries and in the studied regions, clonal variations were considered only for those isolates that presented 100% identity with each other [70]. Pulsotype 9 was the most frequently detected (3/14 isolates), followed by pulsotype 5 (2/14 isolates). The same pulsotype was detected at different sampling points within the same industry (pulsotype 5, detected in a meat chute (72A-2) and an evisceration table (77A-2); and pulsotype 9, detected on a mat in the clean area (59A-2) and a meat chute (78A-2), both present on slaughterhouse C located in Goiás State; Table 2), suggesting the dissemination of these two clones within this establishment. Pulsotype 9 was detected at different sites and at different visits within this slaughterhouse. Though sanitization was performed between sample collections, the fact that the same strain was detected in different collections within the same establishment indicates the presence of biofilms. This finding corroborates those reported by Berrang [61]. Pulsotype 9 was also detected in a drain sample obtained from slaughterhouse B, located in the Federal District, suggesting regional dissemination of the strain. These results highlight the importance of using molecular typing methods with high discriminatory power, such as PFGE, to investigate the dissemination and persistence of strains within food processing plants https://doi.org/10.1371/journal.pone.0259687.g001 [12,13,71,72]. These typing methods are valuable when WGS analysis cannot be performed. Similar to the results obtained in the present study in relation to the use of PFGE, Camargo et al. 2015 [13] described the serial collection of samples during different visits from environmental locations, cuts of meat, and the hands of employees working in a cattle slaughterhouse in the state of Minas Gerais. This aided the identification of cross-contamination between these points and helped determine the persistence of specific pulsotypes. The results of the present study confirm the persistence of two pulsotypes (5 and 9) in slaughterhouse C and possible regional strain dissemination (pulsotype 9). However, due to the low number of isolates identified in slaughterhouses A and B, it was not possible to confirm the persistence of isolates within these establishments. However, the detection of isolates before the commencement of operations in the slaughterhouses (pre-sanitization) and after slaughter (post-sanitization) suggests the continuous presence of this microorganism in existing biofilms in these facilities.

Salmonella sp. serotyping by WGS analysis
WGS analysis of Salmonella sp. isolate identified it as Salmonella enterica serovar Minnesota belonging to ST548 [73]. This finding differs from those reported in other studies in the investigated regions, which indicated a higher occurrence of serovar Enteritidis [36, 74,75]. S. Minnesota ST548 derived from chicken carcasses, chicken feet, and mechanically recovered meat have been reported in Brazil (São Paulo and Minas Gerais, São Paulo, and Federal District, respectively) [76]. On the other hand, this is the first report of its presence in a poultry slaughterhouse located in the Federal District.

Sequencing analysis of the inlA gene in L. monocytogenes isolates
PMSCs were not detected in any of the 14 L. monocytogenes isolates investigated. These results are similar to the results reported by Medeiros et al. 2020 [77]. The authors reported the absence of PMSCs in L. monocytogenes isolates in samples collected from poultry slaughterhouse drains in the Federal District. However, results of the present study differ from those reported previously by Manuel et al. 2015 [78] in the United States and by Camargo et al. 2019 [79] in Brazil. Both studies reported a higher occurrence of PMSCs in strains isolated from food samples and food production environments. The occurrence of PMSCs in L. monocytogenes isolates is associated with the occurrence of attenuated invasion phenotypes [18,80]. Isolates from human listeriosis have less frequent PMSCs in the inlA gene than isolates obtained from food [78,79,81]. The present report indicating the absence of PMSCs is an important

PLOS ONE
Salmonella spp. and Listeria monocytogenes from biofilms in cattle and poultry slaughterhouses finding for public health, since the integrity of the inlA gene is necessary to promote the internalization of this pathogen in host cells [18,82].

Cell invasion and adhesion assays
All 14 L. monocytogenes isolates demonstrated adherence to the Caco-2 cell surface, with an adhesion capacity that varied from 17.38% to 57.14%. Seven of the fourteen isolates (50%) showed ability to invade Caco-2 cells, with an invasion capacity varying from 5.16% to 34.27% (Fig 3). The isolates that exhibited invasion ability belonged to clusters I and III, and pulsotypes 1 (76A-2), 4 (54A-2), 7 (74A-2), 9 (59A-2, 78A-2 and 117A-3), and 10 (88A-2) (Fig 3). The three isolates belonging to pulsotype 9 could invade Caco-2 cells, although presenting with different capacities. The invasion capacity observed in seven of the fourteen isolates, in which the presence of PMSCs was not detected, corroborates with the findings reported by Nightingale et al. 2005 [18]. The latter authors reported that isolates without PMSCs could invade Caco-2 cells with an ability that was significantly more than isolates with PMSCs in the inlA gene. However, the results differed from those reported for the other seven isolates. Despite the absence of PMSCs, these isolates could not invade Caco-2 cells. Several studies have demonstrated the central role of inlA in cell invasion [18,81,83]. However, other elements such as internalin B [84] and listeriolysin O [85] may also be important for the internalization of L. monocytogenes in different cell types. Furthermore, differences in invasion capacity among isolates with the complete gene inlA have been described, indicating that other factors may be related to the efficiency of invasion of this pathogen [86].

Antibiograms and detection of resistance genes
L. monocytogenes. Resistance or intermediate sensitivity was detected for seven of the eight antimicrobials tested ( Table 3). Six of the fourteen isolates (42.9%) were sensitive to all the tested antibiotics. Eight isolates (57.1%) displayed resistance or intermediate sensitivity. The absence of resistance to ampicillin and chloramphenicol corroborates the results observed by other authors in isolates obtained from food sources and food production environments [53,[87][88][89]. Surveillance of the occurrence of ampicillin resistance is important for public health, since this drug is a treatment of choice for human listeriosis [90]. The sensitivity to chloramphenicol can be explained by the ban on the use of this antibiotic for the production of food of animal origin in Brazil since 2003 [91]. The results of the present study on the resistance and intermediate sensitivity towards ciprofloxacin, erythromycin, and gentamicin (42.8%, 35.7%, and 35.7% of L. monocytogenes isolates, respectively) contrasted with those observed by other authors, such as Teixeira et al. 2020 [89], who reported the absence of resistance to the three

PLOS ONE
Salmonella spp. and Listeria monocytogenes from biofilms in cattle and poultry slaughterhouses antibiotics in isolates obtained from beef cuts in the state of Mato Grosso, Brazil. The findings of the present study on resistance to gentamicin and erythromycin indicate a public health concern, since these are the drugs of choice for the treatment of listeriosis in specific cases, such as the use of gentamicin associated with penicillin as an alternative to ampicillin, and the use of erythromycin in the treatment of infections in pregnant women [90,92]. The low occurrence of resistance to doxycycline detected in this study corroborates the findings reported by Vitas et al. 2007 [93], who reported a low occurrence of resistance in isolates of L. monocytogenes obtained from food and human clinical cases in Spain. The low detection of resistance to tetracycline (7.1% of the isolates) is in agreement with the findings reported by Haubert et al. 2016 [88] for isolates obtained from food and food production environments in the southern region of Brazil. However, the findings of the present study differ from those described by Palma et al. 2016 [53] and Camargo et al. 2015 [87], in which resistance to tetracycline was not detected in isolates obtained from samples of beef cuts, cattle slaughterhouse environments, and clinical cases in Brazil. The highest occurrence of AMR detected in this study was in relation to sulfonamides, a characteristic which was found in eight isolates (57.1%). These findings corroborate the results reported by other authors, who also reported a pronounced prevalence of resistance to this class of antimicrobials in isolates obtained from food and food production environments [53,89,94,95]. Sulfonamides associated with trimethoprim are considered as the second-choice of treatment for human listeriosis [90]. Therefore, the detection of resistance to this class of antimicrobials may indicate a risk to human health.
The detection of the ermB gene was in accordance with the finding reported by Haubert et al. [88], who reported the presence of this gene in an isolate of L. monocytogenes obtained from a poultry slaughterhouse environment in the southern region of Brazil. Only two of the four isolates of L. monocytogenes that presented with an erythromycin resistance phenotype in the disk diffusion assay ( Table 3) harbored one of the erythromycin-related resistance genes investigated. It could be possible that the isolates that did not harbor ermA, ermB or ermC genes Table 3 [8]. Alternatively, the resistance of these isolates may be attributable to chromosomal mutations [95]. The presence of the tetM gene in two isolates (14.28%) was similar to the findings reported by Bertrand et al. [96], who reported the presence of this gene in isolates exhibiting phenotypic resistance to tetracycline in strains obtained from human clinical samples and swine and poultry slaughterhouse environments from Belgium and France. Several studies have highlighted the tetM gene as the genotype that is most commonly associated with the development of tetracycline resistance in Listeria isolates [96,97]. These findings differ from findings of the present study that suggested that tetC was the most prevalent gene related to tetracycline resistance. Despite the detection of tetC in eleven of the fourteen isolates (78.6%), only two isolates displayed resistance or intermediate sensitivity to tetracycline in the disk diffusion assay ( Table 3). One of these two isolates harbored two of the four genes related to the development of resistance to this drug (tetC and tetM), and was classified as a resistant isolate based on the

PLOS ONE
Salmonella spp. and Listeria monocytogenes from biofilms in cattle and poultry slaughterhouses antibiogram. The other isolate harbored only the tetC gene and showed intermediate sensitivity to this antimicrobial. The other nine isolates with the tetC gene did not demonstrate resistance to tetracycline in the disk diffusion assay. Therefore, the unique presence of tetC in the isolates did not confer a resistance phenotype to tetracycline, suggesting that the genes did not necessarily trigger the resistance phenotype, which might be related to the presence of mutations that could lead to gene dysfunction [98].
Despite the occurrence of AMR to gentamicin in five of the fourteen isolates in the disk diffusion assay (Table 3), the presence of the aac(3)-1 gene was not detected in any isolate. It could be possible that resistance to gentamicin in these isolates was related to the presence of other genes not investigated in this study, since more than 170 genes related to resistance to aminoglycosides have been described in bacteria [8]. In the present study, resistance to sulfonamides was observed in eight of fourteen isolates in the disk diffusion assay, but no isolates harbored the sulI gene. It could be possible that this resistance was associated with other mechanisms, such as the presence of other genes related to sulfonamide resistance, namely sul2, folP, or thyA genes, with the description of the last two genes reported in L. monocytogenes [99].
Only intermediate sensitivity to chloramphenicol was detected in four of fourteen isolates in the disk diffusion assay. The cmlA and cat1 genes were not detected. The intermediate sensitivity observed in this study could be related to cross-resistance due to the presence of other genes that were not investigated, such as floR [100]. Notably, in Brazil, this antimicrobial has been banned from use in animal production since 2003 [91]. Selection pressure was not a factor. The absence of the ampC and bla SHV genes is consistent with the absence of the ampicillin resistance phenotype in the disk diffusion assay.
Four of the fourteen isolates (28.5%) showed resistance to three or more classes of antimicrobials and were classified as multidrug-resistant isolates [101]. All were isolated from establishment C. Two isolates (54A-2 and 88A-2) showed adherence and invasion in Caco-2 cells. Reports of the presence of L. monocytogenes multidrug-resistant isolates have increased in the literature [88,97,102]. All multidrug-resistant isolates detected in the present study showed resistance to antimicrobials used in the treatment of human listeriosis (gentamicin and erythromycin) [90]. There is a possibility of the transfer of genes from multidrug-resistant isolates to others. Further studies Table 5. Antibiogram and detection of antimicrobial resistance genes from the Salmonella enterica serovar Minnesota isolate from a poultry slaughterhouse located in the Federal District, carried out using the disk diffusion assay (CLSI, 2020) and WGS, respectively.

PLOS ONE
are warranted to elucidate the resistance mechanisms observed in these isolates and their potential as a source of transfer of resistance to other microorganisms. Salmonella enterica Minnesota. Resistance or intermediate sensitivity of Salmonella Minnesota was detected in nine of thirteen antimicrobials tested ( Table 5). The resistance to tetracycline and ampicillin, and the sensitivity to ciprofloxacin and gentamicin were similar to the results obtained by Dantas et al. [103]. The authors reported the same pattern of resistance in isolates of Salmonella spp. obtained from a poultry slaughterhouse environment in the state of São Paulo. The detection of resistance to ampicillin, cephalothin, and sulfonamide bases, in addition to sensitivity to ciprofloxacin, is comparable to the findings reported by Cunha-Neto et al. 2018 [104] who investigated isolates of Salmonella spp. derived from chicken carcasses in a slaughterhouse in the state of Mato Grosso. However, these authors reported the detection of resistance to gentamicin and chloramphenicol, and sensitivity to nalidixic acid, which differed from the results of the present study. The detection of resistance to nalidixic acid in Salmonella spp. isolates has been widely reported [105][106][107]. Meta-analyses of studies published over a period of 20 years in Brazil have revealed the increased occurrence of resistance to quinolone in isolates derived from human samples and chicken meat [108]. Resistance to nalidixic acid may be attributable to the prolonged and widespread use of this antimicrobial in human and veterinary medicine, since this was the first drug in the quinolone class to have clinical use [109]. Similar to nalidixic acid resistance, resistance to tetracycline, sulfonamides, and betalactams may be related to the overuse of these antimicrobials in the different stages of the chicken meat production chain in Brazil [110].
The present study detected resistance to three generations of cephalosporins. Cephalosporins, especially those of the third and fourth generations, are important for human and animal health [111]. Resistance to cephalosporins (including third-and fourth-generation drugs) has been reported in recent years [112,113]. The detection of resistance, especially to ceftazidime, is a public health concern and highlights the need for vigilance and the reduction of the use of antimicrobials in veterinary medicine. In recent years, the Brazilian government has restricted the use of antimicrobials in the production of animal foods, such as via implementation of a ban on the use of chloramphenicol and nitrofurans [91], colistin sulfate [114], and more recently, via the enforcement of a ban on the use of tylosin, lincomycin, and tiamulin [115].
WGS analysis of Salmonella isolate (GenBank accession number JABBEB000000000.1) led to the detection of the tetA, sul2, aac (6')-Iaa, bla CMY-2 , and qnrB19 genes, as well as aided identification of a mutation in the parC gene (T57S, ACC !AGC, causing the mutation T!S). All detected resistance genes have been described in Salmonella spp. isolates in Brazil [76,116]. The concomitant presence of plasmid-mediated quinolone resistance (PMQR) via the qnrB19 gene, and mutations in the region determining quinolone resistance (QRDR) in the parC gene of topoisomerase IV was also observed. These genetic characteristics were confirmed by ascertaining resistance to nalidixic acid, but resistance to ciprofloxacin was not detected in the antibiograms. Despite this, the presence of the qnrB19 gene and other genes related to PMQR may be responsible for the reduced susceptibility to quinolones, which may facilitate the selection of less susceptible isolates and may lead to consequent failure in the treatment with this class of drugs [117]. The detection of PMQR in this study highlights the need for surveillance of the persistence of quinolone resistance genes in the poultry production chain. The bla CMY-2 gene is a plasmid gene related to the production of beta-lactamases (pAmpC), and is a gene that is most commonly detected globally [118]. The presence of this gene corroborates the results of resistance to beta-lactams observed in the disk diffusion assay (ampicillin, cephalothin, cefazolin, and ceftazidime). Its presence is another public health concern considering the importance of beta-lactams in human and animal health, especially thirdgeneration cephalosporins [111]. The presence of the aac-(6')-Iaa gene related to the development of resistance to aminoglycosides did not confer resistance to gentamicin, the only drug from this class of antimicrobials evaluated in the disk diffusion assay. This discrepancy between phenotype and genotype was also observed in Salmonella spp. isolates obtained from different samples in the poultry production chain in the states of São Paulo, Bahia, and Minas Gerais [76]. In contrast, the occurrence of resistance to tetracycline and sulfonamides is related to the presence of the tetA and sul2 genes, respectively. The absence of colistin resistance genes corroborates findings reported by Monte et al. [76], who reported the absence of these genes in Salmonella spp. isolates obtained from samples of swine and poultry production chains from different regions of Brazil, including the Federal District.
The resistance profile observed in this Salmonella sp. isolate enabled its classification as a multidrug-resistant isolate [101]. This characteristic, associated with the presence of mobile genetic elements related to resistance development, such as the presence of PMQR and pAmpC, suggests a potential public health risk. Additionally, the occurrence of resistance to quinolones and cephalosporins, associated with the presence of genetic elements related to this phenotype, is of special concern, considering the importance of these drugs for human health [111].
In vitro biofilm formation capacity L. monocytogenes. During incubation at 37˚C for 24 h, 11 of 14 isolates (78.57%) were capable of biofilm formation and were classified as weak biofilm formers. During incubation at 12˚C for 168 h, nine of 14 isolates (64.3%) were capable of biofilm formation. All nine were deemed weak biofilm formers (Fig 4). The biofilm formation capacity of L. monocytogenes isolates corroborated findings reported in other studies indicating their ability to adhere to abiotic surfaces and to form biofilms on such surfaces [119][120][121]. The poor biofilm formation capacity of most isolates (78.57%) at 37˚C corroborated the findings reported by Harvey et al. 2007 [119]. The authors reported that 90% of the tested L. monocytogenes isolates were weak biofilm-forming bacteria. There was no significant difference (p = 0.2450) in the biofilm formation capacity of the isolates when incubated at 37˚C for 24 h and at 12˚C for 168 h. The finding that nine of the 14 isolates can form biofilms during incubation at 12˚C highlights the risks that these isolates can pose to cattle and poultry slaughter facilities in the event of failures of the hygiene and sanitation practices adopted for the environments, since this is a common temperature documented inside the industry.
There was no association between cluster/pulsotype and in vitro biofilm formation capacity. Pulsotypes 9 and 5, which were the most commonly detected, did not present better in vitro biofilm formation capacity than the other pulsotypes (Table 6). These results are similar to the

PLOS ONE
Salmonella spp. and Listeria monocytogenes from biofilms in cattle and poultry slaughterhouses findings of other studies that reported no difference in the in vitro ability to form biofilms between persistent and non-persistent strains [49,119], although other studies report persistent strains as better in vitro biofilm formers [57,58]. The association between adherence capacity and strain persistence in the industry remains unclear.
Three of the four multidrug-resistant isolates (52A-2, 72A-2 and 88A-2) were capable of in vitro biofilm formation at 37˚C, and none of them formed biofilm at 12˚C. In contrast, 13 out of the 14 isolates that harbored at least one of the resistance genes targeted in this study were capable of biofilm formation in the in vitro assay, at least, at one of the tested temperatures (Table 6). These results highlight the possibility of adherence of these isolates to the surfaces of these industries, posing the risk of acting as resistance gene reservoirs for other strains and species, ultimately posing a threat to public health and causing dissemination of AMR in various strains.
Most isolates investigated in the present study demonstrated the ability to form biofilms under at least one of the conditions tested. These findings, along with the findings of repeated detection of clonal variations within the same industry at different points and different visits, in addition to the identification of the same pulsotype in another slaughterhouse in another region, support the possible existence of L. monocytogenes biofilms in poultry slaughterhouses

PLOS ONE
Salmonella spp. and Listeria monocytogenes from biofilms in cattle and poultry slaughterhouses located in the Federal District and State of Goiás. The detection of the biofilm formation capacity of the isolates, especially the ability documented at 12˚C and observed inside slaughterhouses, indicates the possibility that these isolates can adhere to surfaces, including those of equipment and utensils, which suggests the possibility of biofilm formation in the event of failures in appropriate implementation of hygiene procedures. This creates a potential public health risk because of possible food-associated cross-contamination.

Salmonella spp.
The Salmonella sp. isolate was capable of biofilm formation, albeit weakly, under both conditions tested. These results corroborate those reported by Yin et al. 2018 [122]. The authors reported the ability of biofilm formation of Salmonella spp. isolates obtained from beef production environments in China at different temperatures, including 12˚C and 37˚C. The present results suggest the potential risk to public health by the Salmonella sp. isolate in the slaughterhouse environment, since, in addition to its multidrug-resistance characteristic, this isolate may adhere to abiotic surfaces, and may be considered a potential contaminant of processed foods.

Conclusions
The presence of L. monocytogenes and Salmonella spp. was identified for the first time in poultry slaughterhouses located in the Federal District and in the State of Goiás, Brazil. The bacterial species were detected on equipment and utensils, especially chutes, which were the main routes of contamination by L. monocytogenes within the evaluated plants. Species could not be detected in cattle slaughterhouses. The consistent and repeated collection at the sampling points enabled the characterization of biofilms in the environments of the industries, which was confirmed in 12 of the 14 L. monocytogenes isolates and in Salmonella sp. via the in vitro biofilm formation capacity test that was performed at 12˚C and/or at 37˚C. The PFGE patterns aided the evaluation of the dissemination of specific pulsotypes both within an industry and in the studied regions. Sequencing of the inlA gene derived from L. monocytogenes isolates demonstrated the absence of PMSCs in all isolates. These results, along with the adhesion and cell invasion assay results, suggest their virulence potential. The detection of resistance to antimicrobials of public health importance in L. monocytogenes and Salmonella sp. isolates, in addition to the detection of multidrug-resistant isolates, also underlines the potential risk for the consumer due to the possibility of food-associated cross-contamination via contact with biofilms formed on abiotic surfaces. The results of this study indicate the importance of conducting research that supports surveillance in slaughterhouse environments for possible adjustments in environmental hygiene and sanitation procedures. Additionally, the detection of isolates with pathogenic potential related to the capability of cell invasion, the ability to form biofilms, and AMR development indicates the necessity of assessing the risk for the population as an instrument of public health protection.

Nucleotide sequence accession number
Information on this Whole Genome Shotgun project has been deposited at DDBJ/ENA/Gen-Bank under the accession number JABBEB000000000 (SRA: SRR12606957). The version described in this paper is version JABBEB000000000.1.