{"paper_id":"c91e461d-c9c0-4d92-a46a-ed043e6605c4","body_text":"The World Health Organization (WHO) defines zoonoses as “any disease or infection that is naturally transmissible from vertebrate animals to humans” [ 1 ]. Zoonotic viruses, bacteria, fungi, and parasites can be transmitted through direct contact or through food, water, vectors, or the environment [ 2 ], contributing to 61% of human infections, or even more in the case of emerging and reemerging diseases, with approximately 75% of them being related to animals [ 3 , 4 ]. Zoonotic diseases are not limited to rural areas and can also appear in urban settings, even if there are not noticeable animal populations in the surroundings, because of food and water contamination [ 5 ]. While zoonotic pandemics have been affecting humans since neolithic times, since humanity started to domesticate animals and plants, their relevance to human health has been particularly highlighted by recent highly virulent infections with pandemic potential, such as the 2005 H5/N1 avian influenza outbreak, the 2009 “swine flu” H1/N1 influenza pandemic, the 2013–2016 West African Ebola outbreak, and the 2019 coronavirus disease (COVID-19) pandemic, as well as local outbreaks of “neglected zoonoses” that can also have significant consequences [ 6 , 7 ].\nRelated to this, the introduction of foreign species into new habitats, mainly caused by the live animal trade stimulated by the growing tendency of owning exotic pets, increases the risk of zoonotic infections that can be life-threatening, particularly in children and immunocompromised people [ 8 , 9 ]. Some of these infections could be incorrectly prevented, diagnosed, or threatened because of the lack of knowledge of local health systems, which could have insufficient experience or training with these kinds of diseases [ 10 , 11 ]. Native fauna can also be infected with pathogens carried by introduced species, as well as being affected by predation, competition, and habitat alteration [ 12 ], especially species that belong to insular territories, due to the lack of competitors and predators that facilitate the settlement of invasive species and the spread of zoonotic diseases [ 13 , 14 ].\nThe Canary Islands (Spain) are located in north-west Africa, near the Morocco coastline (13°23′–18°80′ W and 27°37′–29°24′ N), and present the ideal life conditions for exotic species: warm temperatures, high availability of resources, and absence of predators. For these reasons, the archipelago harbors more than 340 invasive (or potentially invasive) animal and plant species, such as the barbary ground squirrel ( Atlantoxerus getulus ), the rose-ringed parakeet ( Psittacula krameri ), and the crimson fountaingrass ( Cenchrus setaceus ) [ 15 , 16 ]. Exotic reptiles are also present in the Canaries, the main example being the California kingsnake ( Lampropeltis californiae ), which has decreased the population of endemic lizards due to its dietary habits, and has been recognized as a carrier of zoonotic bacteria and parasites [ 17 , 18 , 19 ]. Moreover, exotic freshwater turtles have also been found on the islands, mostly pond sliders ( Trachemys scripta ) [ 20 ]. This chelonian, included in the Spanish catalog of invasive exotic species [ 21 ], is native to the southeastern United States and is the most widely distributed species in comparison with its endemic area, being found all around the world except for Antarctica [ 22 ]. In general, pet turtles can carry different bacteria without showing any clinical signs and shed them through their feces, posing an emerging public health concern [ 23 ]. However, despite all the investigations conducted on  Salmonella  in reptiles, which is the most contracted pathogen from this animal group, the presence and prevalence of other bacteria remain unclear [ 24 ]. Even though the zoonotic risk of other exotic reptiles found in the Canary Islands has been studied [ 19 , 25 , 26 ], there is no research regarding the microflora of wild freshwater turtles in this territory. For this reason, the aim of this study was to determine the presence of pathogenic bacteria in these animals and to evaluate their health risk to humans and local fauna.\n\nA total of 42 adult freshwater turtles (2  Graptemys pseudogeographica , 7  Mauremys  spp., 3  Pseudemys peninsularis , 30  T. scripta ) from the islands of Tenerife and Gran Canaria ( Figure 1 ) were euthanized, frozen, and donated by the staff of the “Red de Alerta Temprana de Canarias para la Detección e Intervención de Especies Exóticas Invasoras” (REDEXOS) during 2021–2023, affiliated with a company Gestión y Planeamiento Territorial y Medioambiental, S.A. (GESPLAN) (Santa Cruz de Tenerife, Spain). This action was carried out following authorization from the “Dirección General de Lucha Contra el Cambio Climático y Medio Ambiente, Gobierno de Canarias” (Las Palmas de Gran Canaria, Spain).  Table 1  shows the species studied and the location of their capture.\nThe processing of the samples was performed after thawing the specimens at 4 °C and under aseptic conditions, using a class II biological safety cabinet (Telstar, Barcelona, Spain). Two cloacal samples were taken from each sampled animal using sterile swabs (Deltalab, Barcelona, Spain), and were incubated with 5 mL of Buffered Peptone Water (BPW) (Labkem, Barcelona, Spain). One sample was incubated at 37 °C for 24 h, and the other one at 42 °C for 18 h under microaerophilic conditions (using a 5 mL Eppendorf tube to reduce the oxygen percentage) for  Campylobacter  spp. isolation. An additional sample was obtained under the same conditions and stored for 8 h at 37 °C in 5 mL of Alkaline Peptone Water (APW) (1% NaCl, pH = 8.4) for  Vibrio  spp. Subsequently, 100 µL of liquid culture were inoculated onto different selective culture media: Baird–Parker agar (Labkem) for  Staphylococcus  spp., Cetrimide agar (VWR International, Leuven, Belgium) for  Pseudomonas  spp., Cefsulodin Irgasan Novobiocin agar (CIN) (Merck, Darmstadt, Germany) for  Yersinia enterocolitica , Oxford agar for  Listeria monocytogenes  (Labkem), Thiosulfate citrate bile salts sucrose agar (TCBS) (VWR International) for  Vibrio  spp., sorbitol supplemented MacConkey agar (Scharlab, Barcelona, Spain), and Tryptone Bile X-glucuronide chromogenic agar (TBX) (Labkem) for  Escherichia coli . Every plate was incubated for 24 h at 37 °C, except for the CIN, which was stored at 30 °C. In the case of  Salmonella  spp., 500 µL of BPW culture was transferred to 4.5 mL of Rappaport–Vassiliadis broth (VWR International) and incubated for 20 h at 42 °C. Then, 100 µL of the liquid culture was later incubated in  Salmonella – Shigella  agar (Merck) for 24 h at 37 °C.\nThe colonies obtained were suspended in 1 mL of PBS under sterile conditions, followed by centrifugation at 12,000×  g ; then, the supernatant was discarded, and the process was repeated again. The resulting pellet was subjected to DNA extraction following López et al.’s [ 27 ] protocol. The same methodology was applied for the DNA isolation of  Mycobacterium  spp. and  Campylobacter  spp. using 1 mL of each BPW culture.\nDifferent polymerase chain reaction (PCR) techniques were performed for the identification of relevant zoonotic bacteria, along with resistance and virulence genes:\nSix pairs of primers were employed for  Campylobacter  spp. (23s rRNA fragment) confirmation and  Campylobacter coli  ( glyA  gene),  Campylobacter fetus  ( sapB2  gene),  Campylobacter jejuni  ( hipO  gene),  Campylobacter lari  ( glyA  gene), and  Campylobacter upsaliensis  ( glyA  gene) identification, according to Wang et al. [ 28 ].\nFollowing the protocol described by Blanco et al. [ 29 ], some  E. coli  pathotypes were identified through the amplification of  stx 1 ,  stx 2 , and  eae  virulence genes, responsible for Shiga-like toxins and intimin protein synthesis, respectively.\nListeria monocytogenes  that grew in Oxford agar was identified by a simple PCR of a region of the  iap  gene, which codifies the p60 invasion-associated protein, as described by Jaton et al. [ 30 ].\nMycobacteria identification was carried out using a multiplex PCR described by Kim et al. [ 31 ]. This protocol also allows the differentiation between the  Mycobacterium tuberculosis  complex and the atypical mycobacteria group by amplifying the  rpoB  gene, and the regions of difference (RD) RD1 and RD8.\nThe colonies from cetrimide agar were irradiated with UV light, and the fluorescent ones were tested for  Pseudomonas aeruginosa  through simultaneous amplification of lipoprotein-coding genes:  oprI  and  oprL , as described by De Vos et al. [ 32 ].\nThe identification of  Salmonella enterica  serotypes important to human health was carried out following De Freitas et al.’s [ 33 ] protocol. Two different PCRs were performed: one to identify  Salmonella  Enteritidis ( sdfI  gene) and  Salmonella  Typhi ( ViaB  gene), and a second one for  Salmonella  Typhimurium detection ( Spy  gene). In both cases,  Salmonella  spp. ( OMPC  gene) was tested.\nA single m-PCR described by Campos-Peña et al. [ 34 ] was used for the identification of six  Staphylococcus  species:  Staphylococcus aureus  ( nucA  gene),  Staphylococcus epidermidis  ( sep  gene),  Staphylococcus haemolyticus  ( mvaA  gene),  Staphylococcus hominis  ( hom  gene),  Staphylococcus lugdunensis  ( fbl  gene), and  Staphylococcus saprophyticus  ( sap  gene), as well as for the detection of methicillin ( mecA  gene) and mupirocin ( ileS2  gene) resistance genes.\nAccording to Liu et al. [ 35 ], a PCR assay was performed to detect all bacteria belonging to the  Vibrio  genus by amplifying a fragment of 16s rDNA. A more specific PCR described by Neogi et al. [ 36 ] was performed with the positive samples, based on  toxR  gene amplification to identify  Vibrio cholerae  and  Vibrio parahaemolyticus , and the  vvhA  gene for  Vibrio vulnificus .\nThe colonies grown in CIN agar were tested for the  ail  (attachment and invasion locus) gene to identify pathogenic and non-pathogenic  Y. enterocolitica  strains, according to Wannet et al. [ 37 ].\nAll PCR assays were evaluated with 1.5% agarose gel electrophoresis (Fisher Bioreagents, Madrid, Spain) at 90 V for 1 h. SiZer-100 DNA Marker (iNtRON Biotechnology, Seongnam-Si, Republic of Korea) was used as molecular size marker and Real-Safe (Durviz SL, Valencia, Spain) as DNA stain. The gels were revealed with a ChemiDocTM XRS+ (Bio-Rad, Hercules, CA, USA) system.\nPositive controls were employed in all PCR assays, using bacterial strains from the American Type Culture Collection (ATCC). These strains were stored at −70 °C and incubated for growth for 18 to 24 h in Tryptic Soy Broth (TSB) (Labkem) at 37 °C under aerobic conditions, or microaerophilic conditions in the case of  Campylobacter  spp. Subsequently, they were submitted to DNA extraction using the same method used for the samples. For the negative controls, nuclease-free molecular biology grade water (VWR International) was used instead of DNA.\nThe chi-square test and Fisher’s exact test were applied, establishing a  p -value of 0.05, to compare the prevalence between the turtle species and the islands where the studied animals were found. This was performed using the statistical Windows software “Statistical Package for the Social Sciences” (SPSS) 29.0.1.0 (IBM Corporation, Armonk, NY, USA). The 95% Clopper Pearson confidence intervals (95% CI) were calculated using the approximate or exact method, as appropriate.\n\nMycobacterium  spp. was the most isolated pathogen in the forty-two turtles studied, being identified in eleven out of nineteen animals (57.9%; 33.5–79.7), followed by  Y. enterocolitica  in eight out of nineteen (42.1%; 20.3–66.5), and virulent  E. coli  in fourteen out of forty-two (33.3%; 19.6–49.5). In contrast, none of the thirty-six turtles tested for  L. monocytogenes  showed positive results.  Table 2  summarizes all positive results isolates.\nNone of the most clinically relevant  Campylobacter  species ( C. coli ,  C. fetus ,  C. jejuni ,  C. lari , and  C. upsaliensis ) were detected in this study; however, four isolates were identified at the genus level. Three of them were from Gran Canaria (one  P. peninsularis  and two  T. scripta ) and one from  T. scripta  from Tenerife, with no statistical differences observed between the islands.\nVirulent  E. coli  genes were detected in fourteen out of forty-two turtles (33.3%; 19.6–49.5). The most prevalent gene was  stx 2 , found in seven animals (16.7%; 7.0–31.4), followed by  eae  and  sxt 1  being found in six (14.3%; 5.4–28.5) and four (9.5%; 2.7–22.6) turtles, respectively. Three  T.scripta  showed the coexistence of two different genes:  eae  +  sxt 1  (Tenerife),  eae  +  stx 2  (Gran Canaria), and  sxt 1  +  stx 2  (Tenerife). Detailed data are shown in  Table 3 . There were no statistical differences between the prevalences of turtle species or island.\nThirty-six turtle samples were tested for  L. monocytogenes , and all tested negative.\nThe exclusive amplification of the  rpo B gene in eleven out of nineteen (57.9%; 33.5–79.7) turtles evidences the presence of atypical (non-tuberculous) mycobacteria in these specimens. No significative differences were found between the prevalences by species or island. Detailed results are shown in  Table 4 .\nPseudomonas  spp. was detected in two out of nineteen (10.1%; 1.3–33.1) animals tested ( T. scripta  from Tenerife) with no statistical differences between species or islands. The amplification of both  oprI  and  oprL  genes indicated the presence of  P. aeruginosa  in one (5.3%; 0.1–26.0) of them.\nSalmonella  spp. was detected in thirteen out of forty-two (31.0%; 17.6–47.1) turtles. More specifically,  S . Typhi and  S . Typhimurium serotypes were found coinfecting one (2.4%, 0.06–12.6)  T. scripta  from Tenerife, but  S . Enteritidis were not identified. Results are shown in  Table 5 . No statistical differences were found between the prevalences by turtle species or island.\nNine positive isolates were obtained for  Staphylococcus  spp. from the forty-two chelonians studied (21.4%; 10.3–36.8), with  S. aureus  being identified in eight cases (19.0%; 8.6–34.1). The remaining isolate (2.4%; 0.06–12.6), from one  T. scripta  from Tenerife, was characterized as mupirocin-resistant  S .  hominis . All results are shown in  Table 6 . No statistical differences were observed between the prevalences by species or island.\nThirty-six animals were tested for  Vibrio  sp., yielding positive results in one  P. peninsularis  and one  T. scripta  from Gran Canaria (5.6%; 0.7–18.7); however, a posterior PCR resulted negative for  V. cholerae ,  V. parahaemolyticus , and  V. vulnificus .\nEight out of nineteen turtles tested for  Y. enterocolitica  were positive for this bacterial species, with no statistical differences between islands. The presence of  ail  gene was not observed in any sample.  Table 7  shows detailed results.\n\nSome  Campylobacter  species are well-known zoonotic agents with importance for human and veterinary health. The gastrointestinal disease they cause (campylobacteriosis) is one the most common bacterial illnesses and its incidence has been increasing over the last decade, causing symptoms like fever, abdominal pain, vomiting, diarrhea, and, in fewer cases, extraintestinal infections and/or autoimmune disorders [ 38 , 39 ].  Campylobacter jejuni  and  C. coli  are the most frequent species that cause human infection, which are part of the microbiome that infect warm-blooded animals such as pigs and poultry, and are normally asymptomatic. For this reason, the main entry way for  Campylobacter  spp. into hosts is through the host’s consumption of contaminated animal products [ 40 ]. Other species, such as  C. upsaliensis  or  C. lari , have been found to cause human disease; however, their actual clinical importance remains unknown because the specific  Campylobacter  species involved are not usually identified [ 41 ]. Four species have been isolated from reptiles to date:  C. fetus ,  Campylobacter geochelonis ,  Campylobacter hyointestinalis , and  Campylobacter iguaniorum , of which only  C. fetus  has been associated with human disease [ 42 ].\nIn our study, four (9.5%)  Campylobacter  spp. isolates were obtained, but none could be identified at the species level using our PCR protocol, which was designed for bacteria commonly involved in human campylobacteriosis. They may correspond to the previously mentioned reptile-associated species:  C. geochelonis ,  C. hyointestinalis , and  C. iguaniorum . Recent investigations have identified other species as responsible for disease [ 43 , 44 ], suggesting that the risk for humans and warm-blooded animals in contact with this turtle population could be higher. In general, the presence of  Campylobacter  in turtles is low, with reported prevalences ranging from 10.4% [ 45 ] to 1.1% or there even being a complete absence [ 46 , 47 ], which aligns with the results obtained in the Canary Islands. An exception is the study by Gilbert et al. [ 48 ] who reported 60.4% positive samples using PCR as the detection method. These authors noted that such differences could be due to the varying isolation and identification methods used in each study (specific media or PCR), as well as the intermittent shedding of microorganisms.\nEscherichia coli  is widely distributed among vertebrates, especially warm-blooded animals and reptiles, showing different prevalences between species and being remarkably high in humans compared to reptiles [ 49 ]. Although  E. coli  is an important component of the intestinal microflora, its genetic plasticity has allowed the acquisition of multiple virulence factors, leading to different pathotypes that can cause both intestinal and extraintestinal infections, the latter mainly occurring in the urinary tract, but occasionally resulting in meningitis or endocarditis [ 50 , 51 , 52 ]. The  stx  and  eae  genes are responsible for a large part of the virulence of these pathogenic  E. coli  strains [ 53 ]; the first ones encode for verotoxins or Shiga-like toxins, divided into  stx 1  and  stx 2  with various subtypes each, and are related to bloody diarrhea and the life-threating hemolytic–uremic syndrome, especially  stx 2  [ 54 , 55 ]. Furthermore, the  eae  gene encodes for intimin, a protein that facilitates the attachment of  E. coli  to the intestinal epithelium, which is necessary for colonization [ 56 ].\nNot many studies have tested reptiles for  E. coli  due to their association with warm-blooded animals, and even fewer have addressed its virulence [ 57 ]. The scarce data available show a low presence of Shiga-like  E. coli  (STEC) and/or intimin; for instance, Dec et al. [ 58 ] identified 32 out of 67 (47.8%) positive samples for  E. coli  among turtles, lizards, and snakes from Poland (with significantly similar prevalences between groups), half of which showed virulence factors, but none contained  stx 1 ,  stx 2 , or  eae  genes. Martinez et al. [ 59 ] analyzed 20 ocellated lizards ( Timon lepidus ) from Spain without describing any positive STEC sample, and Bautista-Trujillo et al. [ 60 ] found low prevalences of virulent  E. coli  in 240 green iguanas ( Iguana iguana ) sampled in Mexico: 10%  stx 1 , 0.4%  stx 2 , and 0.8%  eae . The results of our study are considerable higher than these and could suggest that, although reptiles are not major carriers of zoonotic  E. coli , freshwater turtles from the Canary Islands could suppose a risk. However, further research is needed to confirm this hypothesis, along with the identification of other virulent factors and the characterization of the strains involved and their adaptability to the human host.\nListeria monocytogenes  is an opportunistic pathogen that mainly affects immunocompromised individuals, pregnant women and newborns, and can be found in soil, water, various food products, humans, and animals. This microorganism colonizes the intestinal tract through the ingestion of contaminated food and then disseminates to other organs, causing gastroenteritis, meningitis, encephalitis, mother-to-fetus infections, and septicemia, with a death rate of 25–30%. Even though listeriosis is rare compared to other foodborne infections, its high mortality makes this bacterium an important public health concern [ 61 , 62 ]. It is important to note that a cutaneous form of listeriosis can be contracted by veterinarians and farm workers from the animals they handle, and potentially spread the disease to their pets [ 63 ].\nThe adaptability of  L. monocytogenes  to different ranges of temperature, salinity, and pH allows its development in cold-blooded animals like reptiles [ 64 ]; even so, not many studies have been conducted on its prevalence in these animals, showing relatively low infection percentages. In wildlife animals from New York, Chen et al. [ 65 ] obtained a prevalence of 5.6% (18/324) overall, with the prevalence in reptiles (12%, 2/17) being slightly higher to mammals (8%, 5/64) and birds (4.5%, 11/242), while Nowakiewicz et al. [ 66 ] tested 130 European pond turtles ( Emys orbicularis ) from Poland, finding just two cases of  L. monocytogenes  in adults (1.5%, 2/130). The authors of these works comment that the observed differences could be due to the isolation methods applied. Of the 36 turtles tested in this study, none of them showed positive results for  L. monocytogenes , suggesting that listeriosis infection may not be a major preoccupation to consider in freshwater turtles from the Canary Islands; nevertheless, further investigation is required to confirm this statement, taking into account the small sample size.\nMycobacteriosis are a group of diseases with different symptomatology caused by various  Mycobacterium  species. These species are classified into the  M. tuberculosis  complex (primarily affecting the lungs) and non-tuberculous mycobacteria (NTM); these are also referred to as atypical or environmental mycobacteria because they can be isolated from water, soil, dust, and plants and frequently affect lymphatic, skin, and soft tissues [ 67 , 68 , 69 ]. Bacteria belonging to the atypical group (e.g.,  Mycobacterium chelonae ,  Mycobacterium fortuitum ,  Mycobacterium kansasii ) can infect reptiles through cutaneous lesions or the ingestion of contaminated food and/or water; even so, reptiles appear to be naturally resistant, and in most cases are asymptomatic [ 24 ]. Many NTMs have been characterized as antibiotic-resistant, and despite not frequently affecting humans, case reports have been published; for this reason, children and individuals with compromised immune systems should take special precautions and avoid close contact with reptiles to minimize the risk of exposure [ 70 ].\nWhile mycobacteriosis is more frequently reported in chelonians than in other reptile groups, mainly due their association with aquatic environments, most studies consist of case reports of sea turtles or, less frequently, freshwater turtles that show granulomatous lesions on viscera, bone or joint tissues, usually found postmortem [ 70 ]. In this study, non-tuberculous mycobacteria were detected in 11 out of 19 (57.9%) specimens. In contrast, the only study found searching for mycobacteria in wild freshwater turtles, conducted in Poland [ 71 ], reported a prevalence of 24.8% (31/125). This difference could be attributed to the warmer temperatures in the Canary Islands compared to Central Europe, which facilitates bacterial development, as well as the smaller sample size and the methodology employed [ 72 , 73 ]. Although our PCR protocol could not differentiate between species within this group, the most probable species present in the tested specimens is  M. chelonae , as the most frequently identified mycobacterium in these animals, along with  Mycobacterium marinum  and  Mycobacterium haemophilum  [ 74 , 75 ]. Further studies should aim to identify the  Mycobacterium  species by amplifying and sequencing of other DNA fragments.\nBacteria belonging to the  Pseudomonas  genus are known for their capacity to colonize a wide variety of environments, both aquatic and terrestrial, due to their metabolic and physiological adaptability [ 76 ]. Among all species,  P. aeruginosa  is the most extensively studied because of its pathogenic characteristics in plants and animals (including humans). It is often described as an opportunistic pathogen and is one of the most common causes of nosocomial infection, especially in individuals with compromised immune system, burns or wounds, or those using implants or indwelling medical devices [ 77 , 78 ]. In humans,  P. aeruginosa  can affect multiple organs including the skin, brain, eyes, ears, urinary tract, and lungs; however, urinary tract and pulmonary infections are the most common due to their ability to form biofilms on catheters and intubation equipment. Additionally, this bacterium possesses other virulence factors such as toxins, proteases, hemolysins, and antibiotic resistance mechanisms [ 79 , 80 ].\nRegarding animals,  P. aeruginosa  can cause different symptomatology such as otitis in dogs, respiratory infections in cats, mastitis in cows or endometriosis in horses [ 81 , 82 ]. In reptiles, it is part of their oral and intestinal microflora but acts as an opportunistic pathogen too, although few reports have been published and these are focused mainly on lizards and snakes in which it causes ulcerative stomatitis, necrotizing enteritis, cloacitis, dermatitis, abscesses, and septicemia, among other symptoms [ 83 ]. The prevalence of  P. aeruginosa  found in this study (5.3%; 1/19) matches other works conducted in continental Spain such as Mengistu et al. [ 84 ], who found 2 out of 91 (2.2%) positive isolates from wild  T. scripta  samples, or Muñoz-Ibarra et al. [ 85 ], reporting 62 out of 345 (18.0%) positive reptiles (Testudines and Squamata) samples from Spain and Portugal; in this latter study, the authors analyzed other anatomical locations besides feces, like the skin or nose, which, along with the origin of the samples (from diseased animals belonging to clinics), might explain the higher percentage. In Italy, a study showed 9 positive  P. aeruginosa  isolates out of 218 (4.1%) healthy pet reptiles [ 86 ], suggesting that  P. aeruginosa  does not constitute a great threat to consider in freshwater turtles from the Canary Islands; however, more studies need to be conducted to affirm this hypothesis, considering the small sample size analyzed.\nSalmonella  spp. is one of the most frequent causes of foodborne disease in humans, mainly through the consumption of poultry and eggs which are the primary sources of salmonellosis outbreaks [ 87 ]. Infections caused by this bacterium can be classified according to their pathogenicity: human-restricted serotypes ( S . Typhi,  Salmonella  Paratyphi, and  Salmonella  Sendai) cause an invasive, life-threatening systemic disease known as typhoid or enteric fever, whereas nontyphoidal serotypes such as  S . Enteritidis or  S . Typhimurium normally cause self-limited gastroenteritis associated with intestinal inflammation and diarrhea lasting 5–7 days in immunocompetent individuals [ 88 , 89 ]. In animals, the most common clinical manifestation is a gastrointestinal disease, although acute septicemia, abortion, arthritis or respiratory disease can also be observed. However, infection often remain asymptotic, making control in farms and herds challenging [ 90 ].\nReptiles carry  Salmonella  spp. as part of their normal microbiota, with prevalences reaching up to 90% according to some studies, along with a wide variety of serotypes, some of them zoonotic [ 91 ]. This makes salmonellosis the most frequent zoonotic disease transmitted by pet reptiles [ 92 ]. In studies conducted on freshwater turtles from continental Spain, Hidalgo-Vila et al. [ 93 , 94 ] reported prevalences of 13,2% (10/76) and 6.6% (5/78) in free-living endemic turtles, and 6.38% (6/94) and 5.1% (2/39) in free-living and pet exotic turtles, respectively. They detected only one isolate of the zoonotic  S . Typhimurium serotype in a single  T. scripta  turtle, with no presence of  S . Typhi or  S . Enteritidis. Marin et al. [ 95 ] reported that none of the 37 freshwater turtles ( E. orbicularis  and  Mauremys leprosa ) nor the 34 sea turtles ( Caretta caretta ) analyzed were positive for  Salmonella  spp. In contrast, tortoises tested in that study showed a prevalence of 36% (29/81), similar to the findings of Hidalgo-Vila et al. [ 93 ], where all samples (100%; 16/16) from  Testudo graeca  tortoises tested positive for this bacterial genus. The authors attributed these differences to the longer persistence of  Salmonella  spp. in terrestrial environments compared to aquatic ones, as well as to the geophagic and coprophagic habits of tortoises.\nThe results of this study (31.0%; 13/42) are considerably higher compared to continental Spain, which could mean that the turtles analyzed were pets liberated by their owners or escaped from households where they cohabited with more turtles in the same terrarium, as can occur in zoos and especially in pet shops, which has been noted in different studies [ 96 , 97 , 98 , 99 , 100 ]. The finding of  S . Typhi in one  T. scripta  specimen from Tenerife, a serotype considered human-restricted [ 101 , 102 ], indicates a contamination of the animal environment with human feces; also, the isolation of the zoonotic  S . Typhimurium in the same specimen indicates the risk of infection to the surrounding human populations and, especially, animal handlers. More studies need to be carried out to identify other  Salmonella  serotypes in freshwater turtles from the Canary Islands.\nMany  Staphylococcus  species are part of the normal microflora of humans and animals (mostly skin and mucosa), traditionally divided according to their capacity to produce coagulase enzymes [ 103 ]. Staphylococci can be transmitted through direct contact with infected humans, animals or unclean sanitary equipment, as well as through ingestion of contaminated food or water [ 104 ], with  S. aureus  being the most studied species due to its virulence and antimicrobial resistance. In healthy humans, it predominantly colonizes the nose, throat, axillae, and groin, causing minor skin infections that do not usually require antibiotic treatment but, in hospitalized or immunocompromised patients, especially those with skin lesions, the severity of the infection is higher, producing abscesses, lung infections, bacteremia, endocarditis, or osteomyelitis, requiring antibiotic therapy; this makes  S. aureus  one of the most common causes of hospital-associated infection along with  P. aeruginosa  [ 105 , 106 ].\nIn warm-blooded animals, the infection is similar to that of humans; however, staphylococci in reptiles often produce cutaneous diseases as well as gastrointestinal ones, such as stomatitis, dental and liver disease, or cloacitis [ 107 ]. Most of the related bibliography uses skin reptile samples for  Staphylococcus  spp. identification, which cannot be directly compared to our study; even so, interesting results have been obtained. For example, Strompfová et al. [ 108 ] used the MALDI-TOF technique on 40 skin samples from 17 different reptile species and isolated 51 coagulase-negative staphylococci, mainly  Staphylococcus xylosus  (22/40; 55%) and  Staphylococcus sciuri  (16/40; 40%), both commensal of skin and mucosa with the second one identified as an occasional opportunistic pathogen; however, no  S. aureus  was identified. Using the same method, Brockmann et al. [ 109 ] analyzed skin samples from 235 different reptiles and found 25 (10.6%) positive  Staphylococcus  spp. samples. Of the few studies found that analyze fecal samples, one, conducted by Espinosa-Gongora et al. [ 110 ] in the Copenhagen Zoo, did not isolate any  S. aureus  in the 21 reptiles analyzed (chelonians, lizards and snakes), which was the same result as Almeida et al.’s study [ 111 ], where all 66  Chelonoidis carbonaria  tortoises were negative for this bacterium; however, 48 (72.7%) showed positive results for coagulase negative staphylococci, mostly  S. sciuri  and  S. xylosus .\nThe notable difference in the prevalences of  S. aureus  in relation to our study (19.0%) indicates that the freshwater turtles from Tenerife and Gran Canaria could present a health risk to handlers, and especially, to people with a deficient immune system; however, more studies are needed to clarify the origin of the bacterium (e.g., contaminated water environment, contact with infected animals, etc.). One turtle was found carrying mupirocin-resistant  S. hominis . This bacterium is typically found on human skin and rarely causes dermatological diseases, with some studies suggesting that it protects against the development of opportunistic pathogens [ 112 , 113 ]; therefore, its presence can indicate human contamination of the environment where the turtles inhabit. It is also important to mention that the PCR protocol used could lead to false negatives for less-common  Staphylococcus  species due to limited identification.\nThe  Vibrio  genus comprises almost 200 described species that inhabit a wide range of aquatic environments [ 114 , 115 ]. Several species are pathogenic, with  V. cholerae ,  V. parahaemolyticus , and  V. vulnificus  being the most important for human health; the first two cause gastroenteric disease with severe diarrhea (cholera), while the third causes wound infections and septicemia [ 116 , 117 ]. Vibriosis primarily occurs through the ingestion of raw or undercooked seafood contaminated with the bacteria or through wound contact with contaminated water, especially in warm seas [ 118 , 119 ]. Different  Vibrio  species have been detected in turtles; however, limited information is available regarding their symptomatology in these animals. They appear to be associated with skin and gastrointestinal lesions and may also cause bloodstream infection. Among the most common species identified in these reptiles,  Vibrio alginolyticus  can cause serious disease in humans, whereas  Vibrio harveyi  is an emerging opportunistic pathogen that affects many aquatic animals worldwide [ 120 , 121 ].\nStudies focused on  Vibrio  spp. in turtles are mainly centered on sea turtles, particularly in Asia, where sea turtles are considered part of the human diet [ 122 ]. These studies show variable infection rates in cloacal samples, with  V. alginolyticus  and  V. parahaemolyticus  being frequently isolated, highlighting sea turtle consumption risk [ 121 , 123 , 124 , 125 ]. In this study,  Vibrio  prevalence was low, with only 2 positive turtles out of 36 (5.6%) being different from  V. cholerae ,  V. parahaemolyticus , and  V. vulnificus . This is probably due to  V. alginolyticus ’s reported abundance in similar animals; for this reason, more studies need to be conducted to precisely identify the species involved and determine the human infection risk.\nYersiniosis is a foodborne disease mainly caused by  Y. enterocolitica  that manifests as gastrointestinal inflammation, fever, vomiting, and diarrhea; it is especially intense in children under 5 years of age [ 126 ]. This bacterium can be found in a wide variety of animals and environments, with pigs (and pork products) being the main reservoir of human pathogenic strains; however, virulent serogroups have also been isolated in dogs, sheep, wild rodents, and water [ 127 , 128 ].  Yersinia enterocolitica  is resistant to cold temperatures, allowing it to develop in refrigerated food, which is its main infection pathway [ 129 ]; despite this, not many studies have investigated its presence in cool-blooded animals.\nThe prevalence of  Y. enterocolitica  found in reptiles is considerably low: Silveira et al. [ 130 ] obtained one positive sample in 1  Pantherophis guttatus  snake out of 23 (4.3%) wild reptiles from a rehabilitation center in Brazil; Nowakiewicz et al. [ 66 ] did not obtain any positive isolate from the 96 juvenile  E. orbicularis  turtles in a breeding center and only obtain positive isolate from 2 out of the 34 (5.9%) wild adult ones, both from Poland; and Kumar & Sharma [ 131 ] obtained negative results when analyzing 101 urban  Hemidactylus flaviviridis  geckos from India. These low results contrast with those obtained in this work (42.1%; 8/19); however, another study performed with exotic reptiles in the Canary Islands [ 26 ] obtained similar results, where 11 out of 29 (52.4%) veiled chameleons ( Chamaeleo calyptratus ) from Gran Canaria were positive for  Y. enterocolitica , with only one isolate expressing the  ail  gene, which is typically found in pathogenic strains. In the case of the aquatic turtles, this gene was not detected; this absence may indicate an environmental origin and decrease the risk of human and animal infection.\nThe most frequently isolated bacteria were, in order of prevalence,  Mycobacterium  spp.,  Y. enterocolitica , and virulent  E. coli . The detection of species and serotypes such as  S . Typhi or  S. hominis , which are considered specific or are mostly found in humans, may indicate close contact between the turtles and people, possibly due to contamination of the chelonian environment with wastewater from nearby households that, in some cases, are not connected to a sewerage system [ 132 ]. This finding could also suggest that some turtles captured in the wild had been released by their owners or had escaped from households where they were kept under poor livings conditions (i.e., overcrowded terrariums or unclean water supply). This would help explain the higher prevalence of some pathogens found, such as  Y. enterocolitica  or non-tuberculous mycobacteria, compared to previously published data. In addition to the risk to humans, the bacteria identified could affect other animals, even those far from the turtles’ habitat, through the contamination of watercourses.\nFree-living turtles have been less studied than pet or zoological turtles, making it difficult to compare the results of this study with those of others. As far as is known, this is the first study conducted in the Canary Islands investigating the presence of zoonotic bacteria in wild freshwater turtles and providing relevant data on the public health risks posed by these species. However, the small sample size of some of the species examined, which limits statistical analysis, and the PCR methods employed, which did not allow to identify all the isolates obtained, constitute its main limitations. For these reasons, further studies should include larger numbers of specimens from different species, target additional pathogens or serotypes, and incorporate antibiotic resistance and virulence factor assays. It would also be useful to investigate the role of turtles in bacteria dissemination and to distinguish between reservoirs and mechanical carriers. The analysis of skin samples in addition to cloacal swabs could be valuable, as some bacteria species are more abundant in that anatomical part.\nThe results of this study highlight the risk that aquatic turtle populations in the Canary Islands pose to human health, especially to animal handlers in close contact with these animals and their environment, as well as to immunocompromised individuals. Therefore, the use of protective equipment such as gloves and face masks when handling these reptiles is recommended, and direct contact with the animals or their surroundings with bare hands should be avoided.\n\nExotic freshwater turtle populations from the islands of Tenerife and Gran Canaria carry diverse bacteria relevant to human and veterinary health. Among the pathogens detected, Shiga-like toxin producer  E. coli , non-tuberculous mycobacteria, and  S. aureus  pose the greatest threat to people, especially animal handlers, children, and the elderly; additionally, other animals can also become infected. Considering the preliminary results obtained in the study, further research is required to analyze additional bacterial species and serotypes in larger sample sizes in order to gain a better understanding of the health risks posed by exotic turtles in the Canary Islands.","source_license":"CC-BY-4.0","license_restricted":false}