Pet Reptiles in Poland as a Potential Source of Transmission of Salmonella

Reptiles are considered a potential source of Salmonella transmission to humans. The aim of this research was to determine the incidence of Salmonella in pet reptiles in Poland and to examine Salmonella isolates with regard to their biochemical characteristics, serotype, antimicrobial susceptibility, and pathogenic and zoonotic potential. The research material consisted of 67 reptile faeces samples. The taxonomic affiliation of the Salmonella isolates was determined by MALDI-TOF mass spectrometry, biochemical analyses, and serotyping; whole genome sequencing (WGS) analysis was performed on three isolates whose serotype could not be determined by agglutination. The antimicrobial susceptibility of the Salmonella isolates was determined by the broth dilution method, and in the case of some antimicrobials by the disk diffusion method. The pathogenic and zoonotic potential of the identified serotypes was estimated based on available reports and case studies. The presence of Salmonella was confirmed in 71.6% of faecal samples, with the highest incidence (87.1%) recorded for snakes, followed by lizards (77.8%) and turtles (38.9%). All isolates (n = 51) belonged to the species S. enterica, predominantly to subspecies I (66.7%) and IIIb (25.5%). Among these, 25 serotypes were identified, including 10 that had previously been confirmed to cause reptile-associated salmonellosis (RAS). Salmonella isolates were susceptible to all antimicrobial substances used except streptomycin, to which 9.8% of the strains showed resistance. None of the strains contained corresponding resistance genes. The study demonstrates that pet reptiles kept in Poland are a significant reservoir of Salmonella and contribute to knowledge of the characteristics of reptilian Salmonella strains. Due to the risk of salmonellosis, contact with these animals requires special hygiene rules.


Identification of Salmonella by MALDI-TOF Mass Spectrometry
The collected isolates grown on TSB agar were identified by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (MS) using a standard ethanol/formic acid extraction method [12]. The mass spectra obtained from each isolate were processed with the MALDI Biotyper 3.1 software package (Bruker, Germany), and the results were shown as the top 10 identification matches along with confidence scores ranging from 0.000 to 3.000, according to the manufacturer's criteria (www.bruker.com accessed on 1 March 2022).

Biochemical Profiles of Salmonella Strains
In order to assign Salmonella isolates to species and subspecies, the API 20E test (Biomerieux, PL) and the series of biochemical tests listed in the White-Kauffmann-Le Minor scheme were performed [13]. The ability to produce β-glucuronidase was determined by culturing the isolates on Tryptone Bile X-Glucuronide (TBX, BTL, PL) agar. Utilization of o-nitrophenyl-6-D-galactopyranoside (ONPG) and sorbitol was assessed using the API 20E test (bioMerieux, PL). Ewing's malonate modified broth (Biomaxima, PL) was used to evaluate the ability of the strains to utilize sodium malonate. The utilization of galacturonate and mucate was determined using homemade broths [14]. To assess the ability of Salmonella strains to utilize dulcitol, lactose and salicin, the isolates were inoculated in peptone water containing 1% sugar and a pH indicator (a 2% solution of bromocresol purple was added to the broth in a volume of 100 µL per 5 mL). The results were read after 24 and 48 h of incubation at 36 • C. The change in the colour of the medium from red to yellow or yellow-brown indicated the utilization of carbohydrates.
To determine the ability of Salmonella isolates to utilize L(+)-tartrate (dT+ strains), an indirect method was used, consisting of detection of the sequence located in the region between the STM 3357 and STM 3356 genes in the genome of the strains. The STM 3356 gene encodes a protein responsible for the transport of cations in the L(+)-tartrate fermentation pathway. To discriminate between dT-and dT+ strains, multiplex PCR was performed using primers 166, 167, ST11 and ST15 and DreamTaq polymerase (Thermo Fisher Scientific, USA) [15]. ST11 and ST15 enable the detection of DNA regions characteristic of the genus Salmonella, and primers 166 and 167 were used to amplify regions associated with L(+)tartrate utilization.

Serotyping
To determine Salmonella, antigenic structure strains were streaked on Nutrient Agar (NA) (Biomaxima, PL) and AKG plates [16] (home-made, PIWET, PL) and incubated at 37 • C for 24 h. Then, isolates were serotyped with specific O-and H-antisera (Si-fin, DE; SSI Diagnostica, DK) by using agglutination slide tests. The presence of granular agglutination after the application of isolate with specific antisera was classified as a positive result. Based on these results, the composition of O-and H-antigens was established, and serovar names were identified according to the White-Kauffmann-Le Minor scheme [13], including the latest supplement (no. 48) [5].

Whole Genome Sequencing (WGS) and Bioinformatics Tools
Whole genome sequencing was performed for the three strains (nos. 33B, 50, and 51) whose serovar could not be determined in the agglutination test, as well as for one strain of S. Muenchen (no. 33A). DNA was extracted using the Maxwell ® RSC Cultured Cells  [17] was used to remove the low-quality sequences at the ends of reads. Corrected reads were assembled de novo by SPAdes v.3.15.3 [18]. The SeqSero 1.2 tool was used to identify the Salmonella serovars based on curated databases of Salmonella serotype determinants (rfb gene cluster, fliC and fljB alleles) [19].

Evaluation of Pathogenic and Zoonotic Potential of Identified Salmonella Serovars Based on Review of Cases
To assess the pathogenic and zoonotic potential of the Salmonella serotypes identified in this study, a review of literature from PubMed and other online resources was performed. Reports from the Center for Disease Control and Prevention (CDC), European Food Safety Authority (EFSA) and others were searched for information on the frequency of human infections with specific serotypes. A database search to determine zoonotic potential was performed using the search terms "Salmonella", "infection", "reptile", and "zoonosis", in addition to the names of specific serovars.
Salmonella serotypes that have been reported to cause illness with a frequency of at least 0.05 cases per 100,000 population annually (which corresponds to >150 cases in the USA and >220 cases in the EU annually) were considered to have high pathogenic potential. All serotypes that have been reported to induce RAS were considered to have high zoonotic potential.

Evaluation of Antimicrobial Susceptibility of Salmonella Isolates
The drug susceptibility of Salmonella isolates to 16 antimicrobial substances was determined using the broth dilution method [21] (for ampicillin (AMP), cefotaxime (CTX), ceftazidime (CFZ), meropenem (MEM), nalidixic acid (NA), ciprofloxacin (CIP), chloramphenicol (CHL), gentamicin (CN), trimethoprim (W), sulfamethoxazole (SMX), tetracycline (T), colistin (CT), and azithromycin (AZM)) or the disk diffusion method [22] (for streptomycin (S, 10 µg), amikacin (AK, 30 µg) and nitrofurantoin (F, 100 µg)). Ready-to-use Sensititre EU Surveillance Salmonella/E. coli EUVSEC microplates (ThermoFisher Scientific, Waltham, MA, USA) were used for the microdilution method. A bacterial inoculum was prepared by suspending the pure culture grown on NA agar (24 h, 37 • C) in 0.9% NaCl to obtain a density of 0.5 McFarland. A 10 µL volume of the inoculum was added to 11 mL of Muller-Hinton broth (ThermoFisher Scientific, USA). Microplates were inoculated with 50 µL of the bacterial suspension and incubated at 35 • C for 18 h. The E. coli ATCC 25922 reference strain was used as a quality control. Strains were considered resistant (non-wild-type, NWT) if the minimum inhibitory concentration (MIC) for a given antimicrobial substance was above the epidemiological cut-off value (ECOFF) established by European Committee on Antimicrobial Susceptibility Testing (EUCAST) [21]. Salmonella with MIC values equal or below the ECOFF were recognized as susceptible (wild type, WT; WT ≤ z mg/L, NWT >z mg/L). WHONET software (v.22.7.21) was used for MIC data management (https://whonet.org/, accessed on 25 July 2022). As EUCAST does not indicate epidemiological cut-offs for SMX and CT, the results for these antimicrobials were interpreted based on the clinical cut-offs according to the Clinical and Laboratory Standards Institute (CLSI) [22] (for SMX) and EUCAST [23] (for CT). For S and AK, the strains were classified as susceptible, intermediate or resistant according to CLSI guidelines [22], while EUCAST guidelines [23] were used for nitrofurantoin.

Detection of Aminoglycoside Resistance Genes
All of the Salmonella spp. strains that showed phenotypic resistance or intermediate susceptibility to streptomycin were tested for the presence of corresponding resistance genes, i.e., strA/strB, aadA, aac(3)-IV and aphA1, by uniplex PCR [24]. Resistance genes were detected using the primers and annealing temperatures shown in Table S1.

Statistical Analysis
The results of the biochemical tests on Salmonella isolates were subjected to statistical analysis, which in total included the results of 28 assays, i.e., lactose, dulcitol, salicin, malonate, mucate, galacturonate, β-glucuronidase, L(+)-tartrate and 20 reactions from the API 20E test, including ONPG (Table S2). Positive and negative results were coded as 1 and 0, respectively, in a data matrix (Excel, Microsoft Office 2019), and the unweighted pair group method with arithmetic averages (UPGMA) was used to generate a dendrogram of dissimilarity.
The relationships between the reptile type (snakes, lizards, or turtles) and the presence of Salmonella were determined using the chi-square test. Normal distribution of data was tested using the Shapiro-Wilk test. The level of significance was set at p < 0.05. All statistical analyses were performed using Statistica 14.0.0 (TIBCO Software Inc., USA).

Identification and Prevalence of Salmonella Isolates
The identification of putative Salmonella isolates (n = 51) grown on XLD, SS or BGA plates was confirmed by MALDI-TOF mass spectrometry. All isolates were identified only up to the Salmonella genus. The discrimination power of MALDI-TOF MS was insufficient to determine the species or subspecies of Salmonella. For 14 strains (27.45%), the log (score) identification value was in the range of 2.300-3.000, for 32 isolates (62.74%) it was between 2.000 and 2.299, and for 5 strains (9.80%) it was in the range of 1.700-1.999.

Biochemical Profiles of Salmonella Strains
Detailed results of the biochemical assays of all Salmonella isolates as well as subspecies and serotypes corresponding to specific biochemical profiles are presented in the Supplementary Materials (Table S2). Summarized data on the biochemical analysis are presented in Table 2.
The results of the tests with dulcitol, salicine, sorbitol, galacturonate, β-glucuronidase, and malonate were largely consistent with the information on the biochemical characteristics of individual Salmonella subspecies provided in the WHO Collaborating Center for Reference and Research on Salmonella guidelines [13]. Surprisingly, none of the Salmonella isolates tested produced gelatinase. According to the above-mentioned guidelines [13], gelatinase production is characteristic of the subspecies salamae, arizoneae and diarizonae. Activity of β-galactosidase (ONPG test) was demonstrated in all (n = 13) strains of S. enterica subsp. diarizonae and S. enterica subsp. arizonae (n = 2), 1 strain of S. enterica subsp. salamae (50%) and 10 strains of S. enterica subsp. enterica, including 8 strains representing the Lattenkamp serotype. The positive results of the ONPG test for enterica subspecies strains are surprising, as the vast majority of strains of subsp. I do not produce β-galactosidase and are lactose-negative [13]. Interestingly, 5 of 26 ONPG+ strains did not utilize lactose (Table S2, Table 2). L(+)-tartrate utilization sequences located in the region between the STM 3357 and STM 3356 genes were detected in all S. enterica subsp. enterica strains, as well as in one (50%) salamae strain (No. 36a) (Tables 2 and S2, Figure 1). the Lattenkamp serotype. The positive results of the ONPG test for enterica subspecies strains are surprising, as the vast majority of strains of subsp. I do not produce β-galactosidase and are lactose-negative [13]. Interestingly, 5 of 26 ONPG+ strains did not utilize lactose (Table S2, Table 2). L(+)-tartrate utilization sequences located in the region between the STM 3357 and STM 3356 genes were detected in all S. enterica subsp. enterica strains, as well as in one (50%) salamae strain (No. 36a) (Tables 2 and S2, Figure 1). Based on the results of the biochemical tests, and in the case of several strains additionally by serotyping, all isolates (n = 51) were assigned to the species Salmonella enterica and to the following subspecies: enterica (66.7%, n = 34), diarizonae (25.5%, n = 13), salamae (3.9%, n = 2), and arizonae (3.9%, n = 2) ( Table S1). The greatest variation in terms of Salmonella subspecies was noted in snakes; among 29 isolates derived from this type of reptile, 15 (52%) belonged to subspecies I, 11 (38%) to IIIb, 2 (7%) to IIIa, and one (3%) to subspecies II. Among 15 isolates isolated from lizards, as many as 13 (86.7%) belonged to subspecies I, and two (13.3%) represented subspecies IIIb. In strains from turtles (total n = 7), subspecies I was most often identified (n = 6; 86%); only one isolate (14%) belonged to subspecies II. The vast majority of isolates of the diarizonae (IIIb) subspecies were derived from snakes (n = 11; 85%), and only two isolates (15%) from lizards (Table S2).
The similarity between Salmonella isolates based on their biochemical profiles, obtained from the results of 28 biochemical assays, is shown in the dendrogram derived from UPGMA cluster analysis (Figure 2).
At the level of 23% dissimilarity, the strains formed two main clusters-one comprised all S. enterica subsp. I and S. enterica subsp. II strains and one strain of S. enterica subsp. IIIa, while the other comprised S. enterica subsp. IIIb strains and one strain of S. enterica subsp. IIIa. This analysis showed that the biochemical tests used in the study are helpful in establishing the taxonomic identity of Salmonella strains, but they do not always allow for the unambiguous assignment of a strain to a subspecies.
At the level of 23% dissimilarity, the strains formed two main clusters-one comprised all S. enterica subsp. I and S. enterica subsp. II strains and one strain of S. enterica subsp. IIIa, while the other comprised S. enterica subsp. IIIb strains and one strain of S. enterica subsp. IIIa. This analysis showed that the biochemical tests used in the study are helpful in establishing the taxonomic identity of Salmonella strains, but they do not always allow for the unambiguous assignment of a strain to a subspecies.

WGS
The total length of the final assemblies of four S. enterica strains ranged from 5,376,408 bp to 5,984,163 bp. The three strains, whose serotype could not be determined by agglutination (nos. 33B, 50, and 51), were identified as S. enterica subsp. diarizonae 16:z 10 :e,n,x,z 15 (Table 4).  The presence/absence of the lacZ gene encoding β-galactosidase was correlated with the results of the ONPG test for the four analysed strains. Interestingly, despite the presence of the gene encoding the MFS oligosaccharide transporter, strain 33B did not ferment lactose within 48 h of incubation. Comparative analysis of the lactose operon sequence of strains 33A, 50 and 51 revealed the presence of a non-synonymous mutation (substitution CCG→CTG at nt position 290) in the gene encoding MFS transporter in strain 33B, leading to a Pro97→Leu change.
The aac(6 )-Iaa gene, coding for chromosomal-encoded aminoglycoside acetyltransferase, and the kdeA, mdtM, emrD and acrf genes, encoding multidrug efflux transporters, were detected in all strains (Table 4). However, the presence of these genes did not correlate with the phenotypic antimicrobial susceptibility profiles.

Pathogenic and Zoonotic Potential of Salmonella Serovars Identified in Reptiles
Reports of human Salmonella infections in the USA [25], the EU [1] and Queensland, Australia [26], confirm the prevalence of salmonellosis caused by 19 (76%) of the 25 serotypes identified in this work in reptiles ( Table 5). The most accurate data come from the USA, where in 2006-2016 salmonellosis was most often caused by S. Newport (~1.4/100,000), S. Oranienburg (~0.23/100,000), S. Paratyphi B var. Java (~0.13/100,000) and S. Poona (~0.11/100,000). S. Tennessee (~0.04/100,000) and S. Muenchen (~0.03/100,000) were recorded slightly less frequently. Infections caused by serotypes of the subspecies arizonae, diarizonae and salamae identified in this study were reported sporadically (<14 cases per year, <0.004/100,000). In the EU (2018-2020), only 2 of the 25 serotypes, i.e., S. Newport (~0.17/100,000) and Muenchen (~0.05/100,000), were confirmed to have caused infections in humans. In Queensland, over 10 years (2007-2016), infections caused by S. Virchow were recorded with high frequency (~5.6/100,000), and significantly fewer cases were caused by S. Paratyphi B var. Java, S. Newport and S. Oranienburg (>0.17/100,000) ( Table 5). In summary, on the basis of available reports and the adopted criteria, the serotypes with high pathogenic potential (>0.05 cases/100,000 annually) include Newport, Oranienburg, Paratyphi B var. Java, Poona, Virchow, and Muenchen. a-the serotype was not identified or listed in the report; cases of infections caused by rarely identified or unknown Salmonella serotypes were included in the report as one group. b-probable source of infection, not scientifically proven Ten (40%) of the 25 serotypes under consideration have documented zoonotic potential. These are the eight serotypes representing subspecies I, i.e., Abony, Cotham, Muenchen, Newport, Oranienburg, Paratyphi B var. Java and Poona, and Tennessee, as well as one serotype of S. enterica subsp. arizonae, i.e., 41:z 4 ,z 23 :-, and one serotype of S. enterica subsp. diarizonae, i.e., 47:i:z 53 . The pathogen was most often transmitted as a result of contact with turtles, less often with snakes or lizards. The age of patients with RAS ranged from one month to 87 years, but the median estimated in many reports indicates that children aged 3-10 years were most often affected (Table 5).

Prevalence of Salmonella spp. in Captive Reptiles in Poland
Over the past 20 years, there have been a number of reports showing that reptiles, both in the wild and in captivity, are carriers of Salmonella [6]. However, the frequency of occurrence of these zoonotic bacteria is highly varied, ranging from 2.1% [38] to 87.5% [39], depending on the geographic region and type of reptiles analysed [6]. The results of the present study showed that pet reptiles bred in Poland are asymptomatic carriers of Salmonella enterica strains (71.64% positive samples), with a prevalence of 87%, 78% and 39% in snakes, lizards and turtles, respectively. These findings are largely consistent with the observations of Zając et al. [9], who showed a very high frequency of Salmonella in captive reptiles bred in Poland; it was highest in snakes (92.2%), followed by lizards (83.7%) and turtles (60.0%). Our results are also partially in line with the findings of Geue and Löschner [40] and Nakadai et al. [41], who showed the presence of Salmonella in 54.1% and 74.1% of faecal samples, respectively, from caged reptiles reared in Germany and Austria and in Japan. The trend of snakes as the most common carrier of Salmonella spp. (69.7-76.0% ± 9.3%), lizards as slightly less common (61-69.0% ± 6.7%) and turtles as the least common carriers (2.6-24.3%) has also been confirmed by several other teams of researchers [40,42,43]. A significantly lower frequency of S. enterica than that recorded in this study was reported in captive reptiles in Italy (13.61%) [44], Taiwan (30.9%) [42], Indonesia (10%) [45] and Croatia (13%) [46]. Moreover, the Croatian study showed a much higher Salmonella prevalence in lizards (48.4%) than in snakes (8.9%) and chelonians (3.8%). The large variation in the frequency of Salmonella prevalence in reptiles in different countries may depend on the type of reptile included in the study and the type of diet associated with it. In carnivorous reptiles, Salmonella is more likely to occur, as these bacteria may be found in the meat (mice, rats) that these animals are fed. In addition, antibiotics are commonly used on reptilian farms in some countries; antimicrobial substances can destroy the gut microflora of reptiles, including Salmonella.
The lactose fermentation capacity of Lattenkamp strains noted in the present study is a feature rarely found in S. enterica subsp. I, in contrast to strains of subspecies IIIa and IIIb [47]. The lack of correlation in several strains between the presence of β-galactosidase and the capacity to utilize lactose may be the result of mutations in the genes encoding lactose transporter or in the regulatory genes of the lactose operon [48]. More detailed studies are needed to confirm the effect of the Pro97→Leu mutation in the MFS transporter gene on the protein's ability to transport lactose into Salmonella cells.

Zoonotic Potential of Identified Salmonella Serovars
All serotypes detected in this study were non-typhoid and belonged to either S. enterica subsp. enterica, which is responsible for over 99% of human infections, and to subspecies II, IIIa and IIIb, which are much less likely to cause human salmonellosis [6]. The Virchow and Newport serovars identified in the study are among the seven serotypes most frequently causing infections in Poland (in 2016 and 2017) [3], while S. Newport and S. Muenchen rank among the top 11 serotypes most commonly recorded in humans in the EU [1].
Statistical data on the occurrence of RAS in the EU, including Poland, are not available, as there is no obligation to interview patients regarding possible contact with reptiles. The scale of the problem is therefore unknown, although cases of Salmonella infections attributed to direct or indirect contact with captive reptiles have been reported in many European countries (in 2008-2017) [28,30,[33][34][35][36][54][55][56][57]. Meletiadis et al. [58] estimated the frequency of RAS in Italy at 3% to 7%, and in the 0-21 age group even up to 11.7%. In the USA, regular reports of reptile-associated salmonellosis were conducted from 1994 to 2002 [37,59,60]. Based on the available reports and case studies, some serotypes found in reptiles in Poland, i.e., S. Abony, S. Cotham, S. Muenchen, S. Newport, S. Oranienburg, S. Paratyphi B var. Java, S. Poona, S. enterica subsp. arizonae 41:z 4 ,z 23 :-and S. enterica subsp. diarizone 47:i:z 53 , can be considered to have high zoonotic potential, causing infections not only in infants and babies (under 2 years old) but also in older children and adults with normal immunity [10,11,27,36,61]. Occasionally, RAS in infants and young children (<1-5 years) caused by some of these serotypes (S. Abony, S. Cotham, S. Muenchen, S. Newport, and S. Oranienburg) has occurred as invasive infections such as septicaemia, meningitis and colitis [28][29][30].
Many RAS reports point to turtles as the most common source of the pathogen [10,11,27,28,34,35]. However, our research shows that only 38.9% of turtles kept in Poland are Salmonella carriers, and in Germany, snakes or lizards are most often identified as transmitters of Salmonella causing RAS [36].
The diversity of Salmonella strains in terms of pathogenic potential is confirmed not only by the analyses of salmonellosis in humans but also by the results of in vitro studies. McWhorter et al. [62] showed that all considered reptile Salmonella strains representing 30 different serotypes (including strains of the subspecies I, II, IIa and IIIb) were invasive into both human intestinal epithelial (Caco2) and mouse macrophage cell lines (J774A.1). One of the most invasive was the Paratyphi B ver. Java serotype, and a strain of another serotype identified in our study, S. Adelaide, was moderately invasive. Similar results were obtained by Pasmans et al. [63], who demonstrated that all tested reptile Salmonella strains representing 44 serotypes belonging to subspecies I, II, IIIb, and IV were able to invade human intestinal epithelial Caco-2 cells, although isolates belonging to subspecies I invaded the Caco-2 cells to a higher extent than those from the other subspecies. What is more, the human isolates invaded the Caco-2 cells to a lesser degree than their saurian counterparts. These in vitro findings correlate with reports of more hospitalizations and cases of invasive infections among RAS patients than in non-RAS ones [30,64].

Antimicrobial Susceptibility
In the present study, we demonstrated that the majority of Salmonella strains were sensitive to all tested antimicrobials, except for single strains resistant to streptomycin. Our findings are largely in line with the results of Chen et al. [42], who reported that the prevalence of streptomycin-intermediate-resistant and resistant Salmonella strains in captive reptiles in Taiwan was 23.1% and 14%, respectively. The frequency of resistance to other antimicrobial substances was low, ranging from 0% to 9.2%. A relatively low frequency (32.8%) of drugresistant Salmonella strains in captive reptiles in Poland was also shown by Zając et al. [9]. The prevalence of streptomycin-resistant strains reported by these authors was twice as high (25%) as that recorded in this study, while the percentage of strains showing resistance to other antimicrobial substances ranged from 0.2% to 8.1%. A low frequency (12.8%, n = 34/39) of resistant Salmonella strains was also reported in captive reptiles in the Czech Republic [65]. Completely different results regarding the drug susceptibility of reptilian Salmonella strains were obtained in Italy and Spain, where the frequency of drug-resistant strains (to at least one antimicrobial substance) ranged from 93.1% to 100% [43,44], and the percentage of multidrug resistant strains was about 70%. Moreover, the Italian study [44] showed a very high percentage of Salmonella isolates with resistance (89.7%) and intermediate susceptibility (79%) to streptomycin; however, in contrast to our results, this was not the only type of resistance in the strains tested. The presence of streptomycin-resistant Salmonella strains in reptiles (75%) was also reported by Arnafia et al. [45].
The resistance of Salmonella strains to streptomycin noted in this study cannot be explained by the use of antibiotic therapy in reptiles, as streptomycin is not used in treatment of these animals in Poland. It also appears that the reduced susceptibility to streptomycin recorded in the study is not a result of acquired resistance, as none of the strains contained the genes that determine streptomycin resistance in Salmonella (strA, strB, aadA, aac(3)-IVa, and aphA1) [66,67]. For all strains, a small zone of inhibition of growth was recorded around the streptomycin discs, i.e., up to 13.5 mm, while a diameter ≤11 mm indicates resistance. Hence, it appears that this reduced susceptibility of reptilian Salmonella strains to streptomycin may have an evolutionary basis. The same trait has previously been observed in reptilian strains of E. coli [24]. The lack of correlation between the presence of the aac(6 )-Iaa gene encoding chromosomal aminoglycoside N(6 )-acetyltransferase (conferring resistance to tobramycin, kanamycin and amikacin) and the phenotypic susceptibility of the Salmonella strains (none of the strains were resistant to amikacin) confirms previous reports about the non-functionality of this gene in Salmonella [68]. Common occurrence of the aac(6 )-Iaa cryptic gene in Salmonella enterica strains, including subsp. diarizonae, has previously been confirmed by several authors [69,70].

Conclusions
The research provided valuable data complementing existing knowledge of the prevalence of Salmonella strains in reptiles, as well as knowledge of their serotypes, drug susceptibility, and biochemical characteristics. We showed that the pet reptiles kept in households in Poland are common carriers of Salmonella strains, including serotypes that had previously been confirmed to cause reptile-associated salmonellosis in humans. Identification of the Salmonella reservoir is important in assessing the risk associated with the spread of this zoonotic agent and potential infection.
Reptiles should always be considered a potential source of Salmonella transmission, and special hygiene rules should be followed when handling these animals and cleaning terrariums. It is advisable to wash your hands after each contact with reptiles or objects in their living areas. Reptiles should not move freely around the house, especially in the food preparation area, due to possible environmental contamination. When washing terrariums, use disinfectants and disposable gloves. Due to the increased risk of RAS and hospitalization, children under the age of five, immunocompromised people, and the elderly should not touch reptiles or their environment. There is also a need to educate the public, especially reptile owners, about RAS and related preventive measures. Despite the zoonotic risk, the low level of drug resistance of reptilian Salmonella strains gives hope for effective antibiotic therapy in the event of infection.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/pathogens11101125/s1, Table S1: Sequences of primers used to detect resistance genes; Table S2: Biochemical profiles of Salmonella strains; Table S3. Salmonella susceptibility test results.  Data Availability Statement: All the data presented in the study are included in the article. Further enquiries can be directed to the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.