Study on Bacteria Isolates and Antimicrobial Resistance in Wildlife in Sicily, Southern Italy

Wild environments and wildlife can be reservoirs of pathogens and antibiotic resistance. Various studies have reported the presence of zoonotic bacteria, resistant strains, and genetic elements that determine antibiotic resistance in wild animals, especially near urban centers or agricultural and zootechnical activities. The purpose of this study was the analysis, by cultural and molecular methods, of bacteria isolated from wild animals in Sicily, Italy, regarding their susceptibility profile to antibiotics and the presence of antibiotic resistance genes. Bacteriological analyses were conducted on 368 wild animals, leading to the isolation of 222 bacterial strains identified by biochemical tests and 16S rRNA sequencing. The most isolated species was Escherichia coli, followed by Clostridium perfringens and Citrobacter freundii. Antibiograms and the determination of resistance genes showed a reduced spread of bacteria carrying antibiotic resistance among wild animals in Sicily. However, since several wild animals are becoming increasingly close to residential areas, it is important to monitor their health status and to perform microbiological analyses following a One Health approach.


Introduction
Sicily is the largest island in the Mediterranean Sea and the biggest Italian region. Located in the middle of the Mediterranean area, Sicily has a strategically important position for bird migration because it represents a natural bridge between Europe and Africa. The Sicilian territory is characterized by a considerable extension of rural areas and natural parks and by the presence of urbanized and industrial areas. These regional parks, such as Parco delle Madonie and Parco dei Nebrodi, as well as several smaller regional reserves, host many wildlife mammal species, including wild boar (Sus scrofa), wild rabbit (Oryctolagus cuniculus), fox (Vulpes vulpes), and porcupine (Hystrix cristata), and many common birds together with endangered species, such as the golden eagle (Aquila chrysaetos), the smaller Bonelli's eagle (Aquila fasciata), and the griffon vulture (Gyps fulvus).
Wildlife diseases represent an aspect of veterinary medicine that is still lacking complete knowledge. Although at the national and international level, research and studies in this field are increasing, it is not always possible to establish how certain diseases occur in wildlife and whether wildlife plays a role in the maintenance and transmission of pathogens [1]. Diseases affecting wild species can occur with different morpho-pathological patterns in a wide range of hosts and populations. Where some occur asymptomatically, with negligible impact on wild populations, domestic animals and humans, they can occasionally cause dramatic epizootic events characterized by high morbidity and mortality [2]. susceptibility profile of some of the isolated bacterial strains was determined by microbiological assays, and the presence of genes coding for antimicrobial resistance was evaluated by molecular analysis.

Sample Collection
Between 2017 and 2019, samples from 193 wild birds, 119 mammals, 35 terrestrial, and 21 swamp tortoises, from occasionally found carcasses, hunting activities, and regional recovery centers, located mainly in the provinces of Messina and Palermo, were analyzed ( Figure 1). The recovery centers collect wounded live animals that need care and carcasses from most of the Sicilian territory, although they are located near two of the largest Sicilian regional parks (Parco delle Madonie and Parco dei Nebrodi). Before sample collection, each animal was identified on a morphological basis to define the species. The 275 carcasses were subjected to necropsy, including gross examinations and tissue sample collections from the organs (intestines, liver, spleen, kidneys, heart, and lungs) for laboratory investigations.
From the live animals, mainly birds and turtles, housed in recovery centers, samples, including 79 swabs (oral, skin, rectal/cloacal, tracheal, of lesions) and 14 feces, were collected. These samples were collected by veterinary practitioners, who visited the rescued animals in the recovery centers, and were sent to the laboratory to facilitate the diagnosis and perform appropriate antibiotic treatments prior to their release into the wild. In addition, feces samples and rectal/cloacal swabs were screened for pathogenic bacteria that were potentially transmissible to other animals housed in the centers.

Bacterial Isolation and Identification
Swabs and feces from live animals and tissue samples collected from carcasses were subjected to bacteriological procedures to isolate bacterial species. Bacteriological cultures were performed using selective and differential medium (blood agar, MacConkey agar, and mannitol salt agar, Thermo Fisher Scientific, Waltham, MA USA) to promote the growth of the main commensal and pathogenic species from the collected samples [19,20]. For the isolation of Salmonella spp. species, a pre-enrichment in buffered peptone water (APT) and two subsequent enrichments in selenite cystine and Rappaport-Vassiliadis broths (Thermo Fisher Scientific, MA USA), incubated, respectively, at 37 °C and 42 °C for The recovery centers collect wounded live animals that need care and carcasses from most of the Sicilian territory, although they are located near two of the largest Sicilian regional parks (Parco delle Madonie and Parco dei Nebrodi). Before sample collection, each animal was identified on a morphological basis to define the species. The 275 carcasses were subjected to necropsy, including gross examinations and tissue sample collections from the organs (intestines, liver, spleen, kidneys, heart, and lungs) for laboratory investigations.
From the live animals, mainly birds and turtles, housed in recovery centers, samples, including 79 swabs (oral, skin, rectal/cloacal, tracheal, of lesions) and 14 feces, were collected. These samples were collected by veterinary practitioners, who visited the rescued animals in the recovery centers, and were sent to the laboratory to facilitate the diagnosis and perform appropriate antibiotic treatments prior to their release into the wild. In addition, feces samples and rectal/cloacal swabs were screened for pathogenic bacteria that were potentially transmissible to other animals housed in the centers.

Bacterial Isolation and Identification
Swabs and feces from live animals and tissue samples collected from carcasses were subjected to bacteriological procedures to isolate bacterial species. Bacteriological cultures were performed using selective and differential medium (blood agar, MacConkey agar, and mannitol salt agar, Thermo Fisher Scientific, Waltham, MA USA) to promote the growth of the main commensal and pathogenic species from the collected samples [19,20]. For the isolation of Salmonella spp. species, a pre-enrichment in buffered peptone water (APT) and two subsequent enrichments in selenite cystine and Rappaport-Vassiliadis broths (Thermo Fisher Scientific, MA USA), incubated, respectively, at 37 • C and 42 • C for 24 h, were performed from the intestine, spleen, liver, feces, and rectal/cloacal swabs. The enrichment broths were then seeded in xylose lysine desoxycholate (XLD) and brilliant green agar (BGA) [6]. The colonies isolated on the different agars were identified using biochemical-enzymatic tests.
For the most significant strains, amplification and nucleotide sequencing of the 16S rRNA genes was performed. Bacterial DNA was extracted using 100 µL of PrepMan™ ultra sample preparation reagent (Thermo Fisher Scientific, MA USA), according to the protocol indicated by the manufacturer. The Q5 ® high-fidelity DNA polymerase kit (New England BioLabs, UK) and a pair of universal primers (5 -CCAGCAGCCGGCGGGTAATACG-3 as the forward primer and 5 -ATCGGYTACCTTGTTACGACTTC-3 as the reverse primer) were used in a polymerase chain reaction (PCR) assay [21].
An aliquot of extracted DNA (5 µL) was added to a PCR reaction mix prepared by mixing a final concentration of 1X of Q5 reaction buffer, 200 µM of dNTPs, 0.5 µM of each universal primer, 0.02 U/µL of Q5 Hot Start high fidelity DNA polymerase, and 1X of Q5 High GC enhancer, in a total volume of 50 µL. After 10 min of denaturation at 98 • C, the reaction mixture was run through 35 cycles of denaturation for 30 s at 98 • C, annealing for 1 min at 55 • C, and extension for 30 s at 72 • C, followed by a final extension of 2 min at 72 • C. Subsequently, 10 µL of the PCR product was used for electrophoresis on a 1% agarose gel to determine the size of the product. The polymerase chain reaction (PCR) products were purified and sequenced at BMR Genomics Srl (Padova, Italy). The nucleotide sequences were identified using the National Center for Biotechnology Information (NCBI, Bethesda, MD, USA) nucleotide basic local alignment search tool (BLASTn) program.
Unfortunately, only 61 strains of the 222 isolates at the first stage were successfully recovered in purity after storage. However, the analyzed strains were representative of the species present in the tested animals.

Detection of Antibiotic Resistance Genes and Class-1 Integron
The presence of bla TEM , bla CTXM , tetA, qnrS, and sull II antibiotic resistance genes and of the mobile element int1 in 61 bacterial strains was determined using PCR. These genes were chosen for their contribution to resistance against the most commonly used antibiotics, especially in veterinary medicine, and to assess any correspondence between the phenotypic resistance found and the presence of these genes in the tested strains. A PCR reaction mix containing 5 µL of the DNA template and a final concentration of 1X DreamTaq buffer, 2 mM of dNTPs, 0.5 µM of the forward primer, 0.5 µM of the reverse primer, and 1.25 U of DreamTaq DNA polymerase (Thermo Fisher Scientific, MA USA) in a total volume of 50 µL was utilized. Subsequently, 10 µL of the PCR product was used for electrophoresis on a 1% agarose gel to determine the size of the product. The primers and the annealing temperatures used are reported in Table 1.

Host Animal Species and Samples Collected
During the three-year period 2017-2019, 275 animal carcasses, 79 fecal samples, and 14 swabs from 368 wild animals were analyzed. Details on the species and number of the animals tested are presented in Table 2. Most of the species analyzed belonged to the class of birds (n = 193), followed by mammals (n = 119), and, with a minor number, reptiles (n = 56).  All carcasses were subjected to autopsy, which revealed several anatomopathological lesions, many of which can be attributable to direct contact with humans. For the birds, indeed, the most frequent causa mortis was from gunshot wounds with fractured forelimbs; moreover, many of the subjects analyzed appeared emaciated and dehydrated. During the autopsy of mammals, subcutaneous hematomas, abdominal hemorrhagic effusions with a ruptured spleen, and fractures were observed, all injuries attributable to impact trauma. Thus, fecal samples and swabs (oral, skin, rectal/cloacal, tracheal, and lesions) from live animals were submitted to the laboratory by veterinary practitioners of the recovery centers who suspected bacterial infections, which were then confirmed by laboratory analyses and subsequently treated during the hospitalization before the release of the animal in the wild.

Antimicrobial Susceptibility Using the Disk Diffusion Method
The susceptibility to the chosen set of antibiotics was determined for 61 strains. The results of the antibiograms performed with the Kirby-Bauer method are reported in Table 4. The highest resistance values were found against ampicillin (41.8%), excluding Enterobacterales species, such as Citrobacter spp. and Klebsiella spp., whose resistance to this antibiotic is indicated as intrinsic by EUCAST. This resistance was present, in particular among strains isolated from mammals (7/12, 58.3%) and birds (22/41, 53.7%). The results obtained for the other antibiotics tested showed high percentages of susceptibility, particularly to chloramphenicol (93%), ceftriaxone (90%), sulfamethoxazole/trimethoprim (85%), and amoxicillin/clavulanic acid (84%).    In addition, the results obtained showed the presence of seven multidrug-resistant strains (MDR), five of which were resistant to three antibiotic classes, one to four, and another to five. Specifically, the MDR strains were: two E. coli from one owl (id. 13) and one tortoise (id. 10) resistant to penicillins, sulfonamides, and tetracyclines; E. coli from one kestrel (id. 28) resistant to penicillins, cephalosporins, and aminoglycosides; Campylobacter jejuni from a crow (id. 30) resistant to penicillins, cephalosporins, and sulfonamides; Pseudomonas aeruginosa from a peregrine falcon (id. 33) resistant to cephalosporins, lincosamides, and sulfonamides; Enterobacter cloacae from a crow (id. 25) resistant to penicillins, cephalosporins, sulfonamides, and tetracyclines; and Staphylococcus simulans from a peregrine falcon (id. 32) resistant to penicillins, phenols, fluoroquinolones, lincosamides, and tetracyclines.

Detention of Resistance Genes and Int1
Molecular analysis conducted for the detection of resistance genes (bla TEM , bla CTXM , tetA, sulII, and qnrS) and mobile element class-1 integron (int1) demonstrated a reduced incidence of these in the bacterial strains isolated. Indeed, of the 61 strains analyzed, only seven (11.47%) were positive for the presence of one or more genes among those investigated. These seven strains, all belonging to the Enterobacterales family, were three strains of E. coli (from one tortoise id. 10 and two peregrine falcons id. 15 and 33) in which only the bla TEM gene was present, an E. coli (from golden eagle id. 5), which harbored two resistance genes, bla TEM and sulII, and E. cloacae (from rabbit id. 25), in which tetA and sulII were present.
Three resistance genes, bla TEM , tetA, and sulII, were found in two strains of E. coli (from owl id. 13 and flamingo id. 26). The presence of bla CTXM and qnrS genes was not found in any of the strains tested. Finally, we found the presence of the mobile element int1 in three strains (5.2%), two E. coli, from golden eagle id. 5 and owl id. 13 and E. cloacae from rabbit id. 25. Table 5 reports the results of the antibiograms and resistance genes from the seven strains that tested positive for one or more of the genes. ; amoxicillin/clavulanic acid (AMC); ceftiofur (FUR); ceftriaxone (CRO); chloramphenicol (C); enrofloxacin (ENR); ciprofloxacin (CIP); gentamycin (CN); amikacin (AK); clindamycin (DA); sulfamethoxazole/trimethoprim (SXT); tetracycline (TE); Resistant (R); susceptible (S); Intermediate (I); a reported as intrinsic resistance [24]; b reported as intrinsic resistance in [24].

Discussion
This study reports the data regarding the presence of antimicrobial resistance in bacteria isolated in 2017-2019 from wildlife in Sicily. Most of the analyzed samples were collected from subjects with death due to anthropic activities (gunshot wounds or impacts with vehicles). Bacteriological analysis allowed the isolation of different bacterial species. Although some of the bacterial species isolated may have pathogenic and/or zoonotic potential, in most cases, they were bacterial species commensal with those animals, without pathogenic activity, and were not related to the anatomopathological findings observed during post-mortem inspection. Only in a few cases, for example, in live animals with wounds or skin lesions infected by Staphylococcus aureus or P. aeruginosa, were the isolated bacteria responsible for the clinical signs of the lesions.
In fact, although wildlife could represent a risk to humans and domestic animals when it acts as a reservoir of disease, intermediate host, or biological amplifier, most interactions between microorganisms and wildlife are harmless and present relatively few risks [1]. The analyses performed indicated that the microorganisms belonging to the family Enterobacterales (76%) represent the species most prevalent in wildlife. This result is in accordance with those reported in previous studies on wild birds in Sicily, in which the most commonly reported species were E. coli, C. freundii, and Klebsiella oxytoca [25]. In addition, C. perfringens, which, after E. coli, was the most isolated species from both birds and mammals, was indicated as a possible part of the microbiota in wild animals by some authors, although more studies on its role in the wild are needed [26,27].
The data we collected showed the percentage of resistance to the antibiotics tested as less than or equal to 16%, except for ampicillin, for which the detected resistance was 41.8%. However, in a previous study conducted on Gram-negative bacteria isolated from wild birds in Sicily in 2013, significant levels of resistance against sulfamethoxa-zole/trimethoprim and amoxicillin/clavulanic acid were reported in a population of live wild birds [25]. High percentages of resistance in Enterobacterales isolated from wild animals were also found in other Italian regions: strains resistant to cephalothin (94.3%), amoxicillin/clavulanic acid (86.9%), and tetracycline (44.6%) were found in wild boars in Tuscany, while, in strains from migratory Passeriformes transiting through the Metaponto territory (Basilicata, Italy), resistances to amoxicillin (64. 8%), ampicillin (63.1%), rifampicin (61.5%), and amoxicillin-clavulanic acid (54.1%) were found [28,29].
In addition, in contrast to our data, a high incidence of resistance to several antibiotics was also reported in studies conducted on wild species from other European areas: Wasyl et al. detected resistance to 11 antibiotic molecules, including streptomycin, tetracycline, sulfamethoxazole/ trimethoprim, and colistin, in E. coli isolated from wild boar and deer in Poland; Smith et al. in a study conducted on herring gulls and deer in Ireland found that all strains isolated were resistant to rifampicin, oxacillin, and penicillin; and a moderately high level of resistance to tetracycline, streptomycin, ampicillin, and sulfamethoxazole/trimethoprim (35%) was found in fecal samples obtained from wild animals in Portugal [30][31][32].
The results in the present study showed that, except for a multidrug-resistant strain from a terrestrial turtle, the other six MDR strains were isolated from wild birds (Table 4), as well as the majority of strains carrying at least one resistant gene (Table 5). These may be explained by the fact that birds come in contact with many different environments, dry and wet areas, and water resources close to human settings. Predatory birds can also feed on carcasses of livestock animals that can carry more AMR strains.
To confirm these data, further sampling, including more subjects from the different classes of animals, will be necessary. However, the data collected regarding the antibiotic susceptibility, determined using the phenotypic method, indicated that resistance at least for the 12 antibiotics tested in this study was not widespread among strains isolated from the wild animals in our study. In contrast to data related to companion animals, which showed a higher prevalence of resistant bacteria [33], the results of this study indicate that, in wildlife, antimicrobial resistance may be quite unusual and that the major responsibility for high resistance is related to the human factor.
A low incidence of genes for antimicrobial resistance was also present in these strains from the wild. Although they were investigated in both Gram-positive and negative strains, these genes have only been found in a few strains (11.4%, 7/61) belonging to the Enterobacterales family, and mainly in E. coli strains (6/7). All these E. coli harbored the bla TEM genes, which, together with the bla CTXM absent in our strains, are the most widespread resistance genes. The bla TEM gene coding for β-lactamase TEM is the most common mechanism of resistance to ampicillin in E. coli.
Although 87% (20/23) of the E. coli strains tested were found to be phenotypically resistant to ampicillin, only for six of these strains, could the resistance be attributable to the presence of resistance genes. In fact, as reported by several authors, there are multiple complex mechanisms that lead a bacterium to be resistant to antimicrobials, and phenotypic expression does not always correspond to genotypic resistance and vice versa [18,34,35]. This is evident in E. coli strains id. 15, 33, and 10, which, although they harbored only the bla TEM gene, had different phenotypic resistance profiles. Indeed, while strain id. 15 was resistant only to ampicillin, the other two strains were also resistant to other antibiotics: strain id. 33 showed resistance also to amikacin and strain id 10, also to sulfamethoxazole/trimethoprim and tetracycline (Table 5).
For the same reason, sulII (responsible for resistance to sulfonamides) and tetA (to tetracyclines) were found in only four (three E. coli and one E. cloacae) and three strains (two E. coli and one E. cloacae), respectively, although not all these strains showed phenotypic resistance to these antibiotics.
Three of the strains that harbored resistance genes were also found to contain class-1 integron, a mobile element carrying resistance cassettes that has been identified in Gramnegative species, mainly E. coli, and some Gram-positive species, such as Staphylococcus spp. and Corynebacterium spp. of clinical origins [36,37]. With regard to the presence of the class-1 integron in wild animals, a study conducted on different animal populations subject to different types of anthropogenic interference demonstrated that its abundance varied depending on the proximity to humans [38]. In addition, class 1 integrons can be spread through urban, agricultural, and livestock wastewater, and, through wastewater treatment plants, they can reach aquatic environments, such as the rivers and estuaries often used by wildlife [39,40]. However, the reduced presence of class-1 integrons (6.5%) detected in our study could be explained by the fact that the genome of bacteria, such as E. coli, is not able to acquire or maintain class-1 integrons without antibiotic pressure [41] and that, also considering the results of phenotypic resistance, in the Sicilian wild environment, the antibiotic pressure does not appear to be high.

Conclusions
The natural microbiota of animals, as well as humans, are subject to numerous antimicrobial pressures due not only to the use and abuse of antibiotic drugs but also to their use in agriculture. In addition, antibiotics and resistant bacteria present in urban, agricultural, and livestock wastewater can also arrive in wild environments where factors, such as the huge number of bacterial species present and the high-density, can promote gene exchange between bacteria and the spread of antibiotic resistance.
Although this study demonstrated a reduced prevalence of pathogenic and/or zoonotic bacteria, as well as a reduced prevalence of phenotypic resistance and genes determining resistance to the main antibiotic classes, in a One Health approach, it is particularly important to monitor wildlife. One Health is an ideal approach to achieve global health by recognizing that human, animal, and ecosystem health are inextricably linked. In fact, antimicrobial resistance is now considered a zoonotic health threat. Worldwide, antimicrobial resistance is on the rise, and, if current trends continue unabated, it has been estimated that this could be the cause of 10 million deaths by 2050 [42].
Wild animals, especially species such as wild boars, foxes, and seagulls, which are increasingly closer to human environments, can be reservoirs and diffusers not only of pathogenic and/or zoonotic bacteria but also of resistant commensal bacteria or bacteria that possess potentially transferable antibiotic resistance genes. Even in aquatic environments, including remote ones, such as glaciers, the presence of antimicrobial-resistant bacteria has been detected [43]. Thus, in addition to the guidelines on antibiotic use in humans and animals, a more comprehensive approach is needed, including constant monitoring of wildlife, aquaculture, and the environment to assess and control the spread of bacteria and the determinants of antimicrobial resistance.  Institutional Review Board Statement: Not applicable. Samples from the live animals were taken by veterinarians at wildlife rescue centers and analyzed at Istituto Zooprofilattico Sperimentale della Sicilia for diagnostic purposes.

Informed Consent Statement: Not applicable.
Data Availability Statement: All data discussed are contained in the article. For 16S rRNA sequences, please contact the corresponding author.