Antibiotic Resistance and Virulence Profiles of Gram-Negative Bacteria Isolated from Loggerhead Sea Turtles (Caretta caretta) of the Island of Maio, Cape Verde

Previous studies revealed high levels of antimicrobial resistance (AMR) in loggerhead sea turtles (Caretta caretta), describing this species as prime reservoir of antimicrobial-resistant bacteria. This study aimed to characterise, for the first time, the AMR and virulence profiles of Gram-negative bacteria isolated from 33 nesting loggerhead turtles of the island of Maio, Cape Verde. Cloacal, oral, and egg content swab samples (n = 99) were collected and analysed using conventional bacteriological techniques. Shewanella putrefaciens, Morganella morganii, and Vibrio alginolyticus were isolated from the samples under study. The isolates obtained from this loggerhead subpopulation (North-East Atlantic) revealed lower levels of AMR, compared with the results of studies performed in other subpopulations (e.g., Mediterranean). However, the detection of resistance to carbapenems and multiple antimicrobial resistance indices higher than 0.20, raises concern about the potential association of these animals to points of high antimicrobial exposure. Furthermore, virulence phenotypic characterisation revealed that the isolates presented complex virulence profiles, including the ability to produce biofilms. Finally, due to their pathogenic potential, and considering the evidence of illegal consumption of turtle-related products on the island of Maio, the identified bacteria may represent a significant threat to public health.

and tetracyclines [2,3]. The evidence of resistant strains in nesting sea turtles is of great concern in the context of One Health, due to the potential transmission of these strains to humans, mainly local populations [4,5]. Furthermore, ARB, including Vibrio spp., Morganella spp., Aeromonas hydrophila, Pseudomonas aeruginosa, and Escherichia coli, isolated from loggerheads, were previously associated with several sea turtles' diseases [6][7][8].
Caretta caretta is globally categorised by the International Union for the Conservation of Nature (IUCN) Red List of Threatened Species as vulnerable [9]. However, the Caretta caretta subpopulation of Cape Verde (North-East Atlantic subpopulation) is classified as endangered [10]. Moreover, recent data show that this loggerhead subpopulation might be the largest worldwide, which supports the importance of the archipelago of Cape Verde for the conservation of the species [11]. Furthermore, after the Island of Boa Vista, the island of Maio is, together with the Island of Sal, the most important nesting site in the archipelago [12,13].
Due to the endangered status of the loggerhead colony of the island of Maio, there is an increasing need to study this subpopulation to reinforce conservation and surveillance strategies. To the best of our knowledge, there are no studies available describing the antimicrobial-resistance profile of loggerhead's bacteria of the North-East Atlantic subpopulation. Therefore, this study aims to evaluate the antimicrobial resistance and virulence profiles of Gram-negative bacteria isolated from cloacal, oral, and egg content samples of loggerhead turtles of the island of Maio. The present study also aims to evaluate the impact of these bacterial species on sea turtles' health and conservation, as well as the underlying public health risks resulting from interactions with these animals and the consumption of turtle-derived products. Our findings revealed a lower level of AMR for this loggerhead subpopulation compared with other subpopulations addressed in previous studies [2,3,14]. However, the detection of multiple antimicrobial resistance indices higher than 0.20, suggests that Maio's nesting loggerheads may contact with points of high antimicrobial exposure. Finally, the identified isolates revealed the ability to produce several virulence factors, which raises concern about their pathogenic potential.

Results
Cloacal (oviductal fluid), oral, and egg content swab samples were collected from 33 animals, comprising a total of 99 samples. From the 99 samples under study, it was possible to obtain a total of 49 isolates (49.49%) from 24 samples.
Considering the culture media from which were isolated and the type of sample, 19 Gram-negative bacilli were selected for further characterization, including Non Enterobacteriaceae (n = 12) and Enterobacterales isolates (n = 7). The Gram-negative bacilli were selected based on the animal (respective nesting turtle) and on the type of sample from which they were collected (cloacal, oral, or egg content) and considering the media from which they were obtained, their macro and microscopic characterization, oxidase, and catalase reactions, being representative of all the isolates obtained.

Characterisation of Isolates' Antimicrobial Resistance Profile
A considerable number of isolates under study (n = 13) were resistant or intermediately resistant to at least one of the twelve antimicrobial compounds tested. Higher levels of resistance were detected for tetracyclines (nº. of isolates =2), and none of the isolates presented resistance or intermediate resistance to aminoglycosides (amikacin, gentamicin, and tobramycin). Some isolates showed intermediate resistance or resistance to imipenem and enrofloxacin (Table 1).   MAR indices higher than the cut-off value (MAR Index = 0.20) were detected for two isolates (Isolates number 1 and 14) ( Table 2). According to Magiorakos et al. [15] classification, no multidrug-resistant (MDR) isolates were detected, as none was nonsusceptible to at least three antimicrobial agents of different categories.

Discussion
To the best of our knowledge, this study represents the first description of ARB isolated from nesting loggerhead turtles of the North-East Atlantic subpopulation. The present study also represents the first characterisation of the virulence phenotypic profile of sea turtles' bacteria, underlining the role of loggerhead turtles as carriers of potentially pathogenic bacteria.
Despite the low number of isolates obtained, the results of this study contribute to the highly required body of evidence on antimicrobial resistance, especially concerning marine wildlife and the marine environments.
For the collection of oral and cloacal samples, Amies swabs were used, being described as reliable, effective, non-traumatic techniques for the isolation and posterior identification of aerobic and facultative anaerobic Gram-negative bacteria of loggerhead sea turtles [3,16]. Oliveira et al. [17] showed that similar collection and transport methods permit the isolation and characterisation of Gram-negative bacteria, even when requiring large distances and processing periods. Moreover, the processing of samples from three different anatomic sites (cloaca, oral cavity, and egg content) allowed the isolation of distinct bacterial species, e.g., the isolation of A. hydrophila/caviae and B. cepacia from a swab sample of the cloacal and oral cavity, respectively.
In the present study, S. putrefaciens was the most prevalent species found in Caretta caretta, as previously reported by Blasi et al. [18]. Unexpectedly, no Pseudomonas spp. isolates were detected, even though P. aeruginosa was one of the most prevalent species isolated from sea turtles in previous studies [4,19]. This finding may emphasise the importance of monitoring AMR in distinct bacterial species regarding the loggerhead subpopulation under study.
All identified bacterial species have been previously isolated from both injured and stranded sea turtles, as well as healthy wild animals [8,18,20], except for B. vesicularis, which was isolated for the first time from the oviductal fluid of sea turtles in the present study.
B. cepacia, S. putrefaciens and V. alginolyticus, isolated from egg samples in this study, were previously associated with unhatched eggs [21][22][23]. Craven et al. [22] suggested that opportunist pathogens found in adult females could infect sea turtle eggs and cause embryonic mortality. Therefore, the presence of these bacterial species in the egg content of Maio's loggerheads may represent a potential threat to successful embryonic development and the overall reproductive success of this loggerhead colony.
Despite the small number of isolates, the level of AMR detected for the isolates in this study is considerably lower compared with previous studies [2][3][4]24]. No MDR bacteria were detected, which is in line with a previous study performed in juvenile hawksbill sea turtles (Eretmochelys imbricata) and green turtles (Chelonia mydas) from potential coincident feeding grounds [4], but discordant with previous studies in other loggerhead and green turtles' populations [2,3,14,24].
Following previous studies, higher resistance levels were observed for tetracyclines and lower ones for the aminoglycoside class [2,3,14,17,24]. To the best of our knowledge, no resistance to imipenem was previously described for loggerhead sea turtles' bacteria. Here, A. hydrophila/caviae, B. vesicularis, S. putrefaciens, and M. morganii presented intermediate resistance to imipenem, with M. morganii also showing intermediate resistance to meropenem. Regardless of the low incidence of resistance to carbapenems, this finding should be further assessed due to the categorisation of these antimicrobial compounds as last resort options for the treatment of serious Gram-negative infections, being of major importance for human medicine [25].
Sea turtles living in ecosystems affected by humans' activities are at higher risk of being exposed to antimicrobial environmental pressure [14,26]. The island of Maio is mostly characterised by a pristine environment, being less affected by anthropogenic impacts, such as the discharge of wastewater carrying high levels of antimicrobials, associated with aquaculture, intensive farms, and medical facilities [27,28].
The lower anthropogenic influence observed in the island of Maio may explain the lower levels of AMR in this loggerhead subpopulation compared with other subpopulations addressed in previous studies [2,3,14]. However, the detection of isolates with MAR indices equal to or higher than 0.20 from animals not directly exposed to antimicrobial compounds suggests previous contact to points of high antimicrobial exposure [29]. Therefore, although their nesting sites are less exposed to anthropogenic impacts, Maio's loggerheads may not be fully protected from antimicrobial environmental pressure. Moreover, the highly migratory nature of sea turtles may expose them to a broad range of marine environments, promoting contact with sources of contamination.
Virulence characterisation showed that the isolates from this study could express virulence traits that may contribute to the evasion of the host immune system and host tissue colonisation and damage [30]. Moreover, most isolates showed the ability to produce biofilms. Biofilm synthesis is one of the most important virulence factors in bacteria, playing a prominent role in AMR. In fact, the antimicrobials concentration required to eliminate bacterial biofilms can be up to 1000-fold higher in comparison with their free-swimming, planktonic counterparts, which makes these microbial communities extremely difficult to control [31,32]. The expression of a high number of virulence factors may play an essential role in the pathogenesis of infections [30].
Considering Singh et al. [33] classification, a considerable number of isolates (n = 11) in this study were considered a high threat (MAR Index ≥ 0.20; V. Index ≥ 0.50) or moderate threat (MAR Index < 0.20; V. Index ≥ 0.50) to human/animal host or both. These findings suggest that the isolates obtained from the loggerhead turtles under study can pose a threat as potential pathogens, especially those revealing high virulence indices, such as A. hydrophila/caviae, B. vesicularis, and S. putrefaciens.
Despite the existent conservation efforts, the slaughter of nesting loggerheads for consumption is frequently practised in the island of Maio [11]. The consumption of turtlerelated products represents a risk behaviour for public health, which is supported by the detection of pathogenic and antimicrobial-resistant bacteria isolated from loggerhead turtles in this study.
In conclusion, the lower levels of AMR detected for the Cape Verdean loggerhead subpopulation, compared with the levels identified for other subpopulations, represent positive and encouraging results regarding the current context of AMR. However, the presence of potentially pathogenic Gram-negative bacteria expressing several virulence factors may represent a risk to sea turtles' health and consequently affect the conservation of this endangered species. Finally, due to their pathogenic potential, the identified bacteria may represent a significant threat to public health. This threat arises mainly through the unsafe and illegal consumption of turtle-derived products.
Further studies are encouraged to characterise the identified bacterial species at the molecular level, as well as to assess the genetic determinants imparting AMR and virulence. It would be equally relevant to further study the resistance mechanisms involved in the isolates' antimicrobial resistance profile, especially concerning resistance to carbapenems.

Sample Collection
A total of 33 nesting loggerhead sea turtles (Caretta caretta) were sampled during August 2019. Oral, cloacal, and egg content samples were obtained from each female turtle, using Amies swabs 1814-002 (VWR, Leuven, Belgium).
The three samples were collected sequentially, in the following order: cloacal, oral, and egg. For cloacal samples, the swab was gently inserted approximately 5 cm into the cloaca, and with a rotational movement, the oviductal fluid from the internal surface was collected.
Oral sampling was performed by opening the rhamphotheca with a previously disinfected (ethylic alcohol 70%) wooden pry bar. The soft tissue of the mouth (tongue and palate) was gently swabbed for approximately 5 s.
For the egg sample, an egg was collected directly from the cloaca without contacting the surrounding environment. A small surface of the shell was sterilised with a fire-heated bistoury, and a circle shape window was cut. A sterilised Pasteur pipette was used to collect approximately 1.5 mL of the egg content (yolk and albumen), which was introduced directly in the transport medium of the Amies swabs. The samples were identified with date, time, type of sample, and flipper tag number and then safely placed in a thermic bag at 4 °C.
After the sampling period, the collected samples were transported to the Microbiology and Immunology Laboratory of the Veterinary Faculty, University of Lisbon, Portugal, for further processing. The following information was collected regarding each sample from each animal: date and time of sampling, local of sample collection, flipper tag identification number, Passive Integrated Transponder (PIT) number (when available), and type of sample (cloaca, oral cavity, and egg content) (See Supplementary file 1). The handling time did not exceed 5 min, before and after which the animals were observed from a safe distance to ensure that oviposition proceeded normally.
The collection of samples was conducted under Maio Biodiversity Foundation guidelines and by the permits of the Environmental National Authority DNA (Direção Nacional do Ambiente). Research protocols were performed per the IUCN Policy Statement on Research Involving Species at Risk of Extinction [41] approved by the 27th Meeting of IUCN Council, Gland Switzerland, 14 June, 1989, and the Sea Turtle Research Techniques Manual [42].

Isolation and Identification of Gram-Negative Bacteria
After pre-enrichment in Buffered Peptone Water (Oxoid, Basingstoke, UK) at 37 °C for 24 h, aerobic and facultative anaerobic Gram-negative bacteria were isolated from collected samples using Glutamate Starch Red Phenol (GSP) Agar plates supplemented with 100,000 UI penicillin g/L (Merck, Darmstadt, Germany) and MacConkey Agar (Oxoid, Basingstoke, UK), incubated at 37 °C for 24-48 h [2]. Positive bacterial colonies were isolated in Columbia agar supplemented with 5% sheep blood (bioMérieux, Marcy-l'Étoile, France). Isolates were characterised regarding their macro and microscopic morphology, Gram staining, catalase test and oxidase reaction. For further characterisation, 19 Gramnegative bacilli were selected and identified through the biochemical identification galleries API 20E and 20NE (bioMérieux, Marcy-l'Étoile, France), according to the manufacturer's instructions.
Multiple antimicrobial resistance (MAR) indices were calculated for the selected isolates as follows: nº. antimicrobials to which isolates were resistant/nº. antimicrobials tested. A MAR index equal or greater than 0.20 was used as a cut-off to differentiate between high and low-risk contamination [29].

Evaluation of Isolates' Virulence Profile
Isolates were characterised regarding their phenotypic virulence profile by assessing the production of enzymes associated with bacterial pathogenic potential.
The virulence indices were calculated for the tested isolates, as follows: nº. positive virulence factors/nº. virulence factors tested [33]. A virulence index equal to or higher than 0.50 was used as a cut-off to evaluate the threat levels for the selected isolates, regarding their pathogenic potential [33]. According to Singh et al. [33], isolates were categorised as a high threat-isolates with virulence and MAR indices greater than or equal to the cut-off values (MAR index ≥0.20, V. Index ≥0.50); moderate threat-isolates having virulence index ≥0.50 but MAR index <0.20; and no threat-isolates having virulence and MAR indices below the cut-off values.