Bacteria richness and antibiotic-resistance in bats from a protected area in the Atlantic Forest of Southeastern Brazil

Bats play key ecological roles, also hosting many zoonotic pathogens. Neotropical bat microbiota is still poorly known. We speculate that their dietary habits strongly influence their microbiota richness and antibiotic-resistance patterns, which represent growing and serious public health and environmental issue. Here we describe the aerobic microbiota richness of bats from an Atlantic Forest remnant in Southeastern Brazil, and the antibiotic-resistance patterns of bacteria of clinical importance. Oral and rectal cavities of 113 bats from Carlos Botelho State Park were swabbed. Samples were plated on 5% sheep blood and MacConkey agar and identified by the MALDI-TOF technique. Antibiotic susceptibility tests were performed using Kirby-Bauer’s antibiotic disc diffusion technique.We identified 596 isolates at the genus level and tentatively to the species level. Proteobacteria was the most abundant phylum in all the dietary guilds, representing 87% of the total identified samples. The most common bacteria within bat individuals were Escherichia coli, Klebsiella oxytoca and Serratia marcescens, and within bat species were Serratia marcescens, Pseudomonas sp. and Staphylococcus sp. Frugivores presented the most diverse microbiota. In general, the antibiogram results indicated a low occurrence of resistance on eigth potentially pathogenic bacteria species. The resistance to antibiotics found on our samples was related mostly to the intrinsic resistance of the tested species.The low occurrence of resistant bacteria in our samples could be related to the well preserved environment where bats were caught. Once the major causes of resistance-acquiring are related to anthropic activites, the controlled access of tourists on certain regions of the Park seems to be effectively protecting the environment.


Introduction
Bats as a group are distributed worldwide, with more than 1300 species, representing ca. 20% of the world mammals [1]. They are highly diversified ecologically, bringing together the most diversified feeding strategies among terrestrial vertebrates. Dietary strategies include frugivory, hematophagy, insectivory, nectarivory, carnivory, piscivory and omnivory [1,2]. Some species allocated in one of these categories include different food sources in their diet [1,2]. Due to this diversified diet they provide important ecosystem services such as seed dispersal, pollination and pest control, but also carry many pathogens, some of them of zoonotic potential [3,4]. Little is known about Neotropical bat microbiota, which is in great part studied for Old World species and mostly related to the gastrointestinal diversity [5][6][7][8][9][10][11];. Also, studies focused on the interaction, influence and ecologic role of bats oral and rectal microbiota are scarce, despite their importance on the digestion, vitamin synthesis, protection against harmful microorganisms and also public health [12][13][14][15][16][17].
Previous studies of bat gut microbiota showed that the bacteria diversity is in part related to the host diet, with a partial overlap between species in different dietary guilds, once these species can compensate the lack of some requirements with different food sources during resources shortages [9,18]. Besides the microbiota diversity, the bacteria antibiotic-resistance patterns could be also modulated by dietary habits [19][20][21]. Among the major causes of resistance acquiring is the contact with anthropic environments [21,22]. Antimicrobial resistant bacteria are a growing and serious problem to the public health and environment, and are reported to be present even on remote habitats [17,23]. The presence of antimicrobial resistance in wildlife brings implications, as it can drive animals to become potential reservoirs of resistant bacteria, and also impose limits to the efficiency of antibiotics used on the control of human and wildlife diseases [17,21].
Against that background, we aimed (1) to describe the oral and rectal aerobic microbiota richness of bats in five dietary guilds from the Carlos Botelho State Park (CBSP), a protected area on the Atlantic Forest of Southeastern Brazil, focusing bacteria of clinical importance; (2) to identify the antibiotic-resistance profile of eight potentially pathogenic bacteria for those bats; and (3) to evaluate whether the protected area is preserving the wildlife from antibiotic resistant bacteria.

Sampling
Fieldwork was conducted monthly from October 2016 to September 2017 on the Carlos Botelho State Park (CBSP; 24˚12'-24˚4'S, 47˚47'-48˚7'W), which is a protected area in the Brazilian Southeastern Atlantic Forest, created in 1982. The phytophysiognomy is mostly represented by the ombrophilous forest, with ca. 23,300 ha composed by pristine forests [24]. Bats were captured using with mist-nets and during searches for roosts, under the permits SISBIO/ICMBIO 54.381-1/2016 and COTEC/SMA-IF 260108006.479/2016. Monthly, oral and rectal cavities of one bat of each species captured were swabbed with sterile cotton swabs, which were then separately transported in Stuart's transport medium and refrigerated. Samples used in this study were collected from 113 bats of 33 [25]. The database of this technique is mostly composed by pathogenic species, therefore a great part of the identifications tend to result on pathogenic bacteria species. Isolates were analyzed using a formic acid-based direct, on-plate preparation method. Small amounts of a single colony were smeared directly onto a spot of the MALDI-TOF MS steel anchor plate. Each spot was then overlaid with one microliter of 70% formic acid and allowed to dry. The dried mixture was overlain with 1 μl of matrix solution (α-cyano-4-hydroxycinnamic acid [HCCA]) dissolved in 50% acetonitrile, 47.5% water, and 2.5% trifluoroacetic acid and allowed to dry prior to analysis using the MALDI Biotyper. An Escherichia coli (ATCC 25922) isolate was used for instrument calibration. Two positive controls (Escherichia coli ATCC 25922 and Staphylococcus aureus ATCC 25923) were included with each run [26].

Antibiotic sensitivity
Antibiotic susceptibility tests were performed on Mueller Hinton agar using Kirby-Bauer's antibiotic disc diffusion technique [27]. The tests were performed for the most potentially pathogenic bacteria species Acinetobacter baumannii, Escherichia coli, Klebsiella oxytoca, Pseudomonas aeruginosa, Salmonella sp., Serratia marcescens, Stenotrophomonas maltophilia, and Stenotrophomonas sp. The antibiotics used on the tests were selected according to the bacteria characteristics [28], and the discs were firmly placed on the seeded plates, which were incubated at 36˚C for 24h. The susceptibility of each isolate for different antibiotics was evaluated by the zones of inhibition, which were measured and compared with the susceptibility pattern of each antibiotic defined by the Clinical and Laboratory Standards Institute [28].

Oral and rectal microbiota
We isolated 830 morphotypes of bacteria from bats in five different dietary guilds (carnivores, frugivores, insectivores, nectarivores and sanguivores). A total of 596 morphotypes were identified at the genus level and tentatively to the species level by the MALDI-TOF methodology, including 243 from the oral cavity and 353 from the rectal cavity. Successfully identified isolates from the oral cavity are represented by: 14 isolates from two species of carnivores; 15 isolates from two species of sanguivores; 25 isolates from three species of nectarivores; 75 isolates from 14 species of insectivores; and 113 isolates from 10 species of frugivores (Table 1). Insectivore -

Subfamily Desmodontinae
Desmodus rotundus ( Successfully identified isolates from the rectal cavity are represented by: 11 isolates from two species of carnivores; 27 isolates from two species of sanguivores; 60 isolates from two species of nectarivores; 90 isolates from 15 species of insectivores; and 165 isolates from 11 species of frugivores ( Table 2). Isolates belong to four bacteria phyla, divided into 15 families. Proteobacteria was the most abundant phylum in all the dietary guilds, representing 87% of the total samples, followed by Firmicutes with 12%, and Actinobacteria and Bacteriodetes counting together 1% of the total identified samples. The family Enterobacteriaceae represented 73% of the samples, followed by Pseudomonadaceae, with 7%, and the other 20% are composed by small sums of the families Aeromonadaceae, Bacillaceae, Brucellaceae, Burkholderiaceae, Caulobacteraceae, Enterococcaceae, Lysobacteraceae, Microbacteriaceae, Micrococcaceae, Moraxellaceae, Neisseriaceae, Sphingobacteriaceae, Staphylococcaceae and Streptococcaceae. The phylum Actinobacteria, represented by Arthrobacter sp. and Microbacterium sp. was found only in the oral cavity, while the phylum Bacteroidetes is represented only by Sphingobacterium sp. in the rectal cavity.
Sixty-two taxa of bacteria were identified in the oral cavity and 72 in rectal cavity of the bats. The Venn diagram analysis (Figs 1 and 2) indicates that the major proportion of the bacteria within the dietary guilds is shared between two or more guilds. The oral richness shared between guilds varies from 49% to 75%, whereas the rectal richness varies from 59% to 87% of bacteria taxa shared with at least one other guild. However, only the species S. marcescens is shared between all five guilds when the oral richness is analyzed alone, and only the species K. oxytoca and S. marcescens are shared between all the guilds when considered the rectal richness. Comparing the dietary guilds, higher richness was found on frugivores (58 taxa), followed by insectivores (50 taxa), nectarivores (37 taxa), sanguivores (21 taxa) and carnivores (11 taxa).

Bacteria richness
Gram-negative bacteria in the phylum Proteobacteria seems to be common in bat microbiota on studies based both on culture protocols and DNA sequencing, being isolated from oral and rectal cavities [8], intestine [9,18] and saliva [10]. The phyla Actinobacteria, Bacteriodetes and Firmicutes were also previously reported as common on bats [10,11,18]. The mammalian gut microbiota diversity is related to the host diet, and should increase from animal-based diets to omnivorous to herbivore diets [29]. In our results, the frugivores microbiota was the most diverse among the five analyzed dietary guilds, and agrees to the mammalian gut microbiota theory. The less diverse microbiota in our survey was found in carnivores, which is also in agreement to the mammal microbiota theory. However, insectivores also showed high microbiota richness, and diverge from the expected, which could be explained by the inclusion of different alimentary items, rather than insects, on the diet of many species classified as insectivores. Species such as Glyphonycteris sylvestris, Lampronycteris brachyotis, Micronycteris microtis and Myotis nigricans analyzed in this study are reported to complement their diet with fruits and/or pollen [30][31][32], which could increase the general microbiota richness of the insectivore bats guild analyzed here.
Another possible explanation for the richness observed in the different bat guilds lies within the number of bats sampled for each guild, whereas the most diverse guilds are also the ones with more bat captures.Though most of the results are in agreement to other studies based on DNA sequencing, the general bacteria richness of bats from CBSP may be biased by the identification technique and the culture step. On the other hand, some bacteria genera, including pathogenic ones, are hard to speciate using DNA sequencing techniques [11], making comparisons even harder.
Some bacteria genera, such as Arthrobacter, Burkholderia, Microbacterium, Neisseria and Rahnella were found only in the oral cavity. Arthrobacter is composed by soil bacteria, and was also found on bats' wing sacs, chin and axillae by other authors [33][34][35]; strains of Arthrobacter and Rahnella were identified as effective inhibitory antagonists of the growth of Pseudogymnoascus destructans, the fungus that causes white-nose syndrome, a letal bat disease [36]. Burkholderia and Microbacterium were previously found on bats' saliva, urine, faeces, and intestine [9,10]. Neisseria was previously found on bat saliva samples [10], and is closely related to mucosal and dental surfaces, being a consistent component of human oral microbiota and also found in different mammals [37]. The rectal cavity exclusive genus Enterococcus was also isolated from bats' wings [35]. Brevundimonas, found only on the rectal cavity, was originally isolated from water and hospital-related material. This bacterium has been previously reported for marine mammals and is not common in bats [38,39].
Bacteria genera observed within different dietary guilds were also divergent, with some exclusive occurrences. Edwardsiella was found only in sanguivores. This bacterium was previously isolated from bovine faeces and latter from cattle meat, wild mammals and birds [40,41]. Thus, the occurrence of this bacterium only in this guild appears to be related to the feeding habit, which is based on blood from domestic and wild mammals and birds [42]. Plesiomonas, Proteus and Yokenella were identified only in insectivores. The genus Proteus, however, was also found in sanguivores and frugivores in other studies [8,43]. The genus Plesiomonas is reported to be isolated from freshwater and surface water samples [44], and many species of insectivores are associated to these environments [45][46][47], where they forage and could be exposed to bacteria. Yokenella was previously isolated from the intestinal tracts of insects and faeces of insect-feeding animals, including bats [48,49]; therefore it is probably related to this kind of diet. Vagococcus, here observed only in carnivores, has been recovered from animals, water, soil and human sources [50]. S.marcescens and K. oxytoca were found on all five dietary guilds. S.marcescens was reported in other studies and various dietary guilds, including frugivores [7], sanguivores [8,43], and insectivores [11,35]. K. oxytoca was previously reported in frugivores [7,51] and insectivores [14,52]; however, K. oxytoca was highly related to vespertilionid (insectivores) bats rather than to any other Australian mammal on previous studies [53].

Antibiotic sensitivity
Generally, the resistance to antibiotics found on our samples was related to the intrinsic resistance of the tested species [54] and independent of dietary guilds of the bats. The species P. aeruginosa and S. maltophilia did not show any resistance besides their expected intrinsic resistance patterns. The species A. baumannii and Stenotrophomonas sp. showed resistance to the antibiotics ciprofloxacin and ceftazidime, respectively; those resistances are not intrinsic and could be acquired from both clinical or environmental antibiotic resistance genes sources, disseminated on the environment. A. baumannii is one of the most important pathogens in hospitals, and the development of multidrug-resistant strains has become of great concern for antibiotic therapies. Ciprofloxacin is a very potent antibiotic used as first line agaist A. baumannii infections [55][56][57] and previous studies have isolated high rates of ciprofloxacin resistant strains of A. baumannii [58][59][60]. The development of resistance on A. baumannii strains has been previously related to mutations in the quinolone resistance determining regions and efflux pump mechanisms [58,59]. The only strain of A. baumannii was isolated from an insectivore bat that was found during the day on the floor of a Visitors Center on CBSP, and the

Molossus rufus INS Acinetobacter baumannii
https://doi.org/10.1371/journal.pone.0203411.t003 Bat microbiota in SE Atlantic Forest possible contact of the bat with human leavings could have influenced on the acquiring of resistant strains. The Stenotrophomonas sp. resistant strains were found on frugivorous and nectarivorous and could outcome from the contact with water or fruits and even casual ingestion of insects [19][20][21]. Additionally, once the contact with anthropic and agricultural environments is one of the major sources of acquired resistance, the activity pattern and diet of carnivores, insectivores and sanguivores bats would make them more susceptible to exposure to antimicrobials [21, 61,  The tested K. oxytoca isolates presented only one resistant strain (5%) and none of the S. marcescens isolates presented any resistance besides the intrinsic ones. The small rates of resistant bacteria observed on CBSP in consistently different from those observed in other studies Bat microbiota in SE Atlantic Forest conducted on areas influenced by anthropic activities [17,20,63,64]. A study conducted on Krakatau Islands found a great number of resistant bacteria on local bats and rats, which they correlated, in part, to anthropic influence on the local islands [20]. The antibiotic-resistance pattern found for E. coli isolates from Nigerian bats also showed a great number of resistant isolates; the resistance was attributed to the use of antibiotics on poultry feed or on poultry itself [17]. Analyzing all of our tested isolates, 71 out of 81 (87%) did not present any resistance besides the expected from the intrinsic pattern, which could be related to the effectiveness of CBSP on the conservation of the wildlife and environment present on the preserved area of the Park. Once some of the sampling sites were close to the Park limits and some Brazilian bats are know to forage on distances of 0.5 to 15 km [65][66][67], it seens that bats from CBSP prefer to forage on the pristine environments rather than anthropized surroundings. Moreover, the restriction of the contact to antibiotics would not lead to the decline of acquired resistances; therefore, it is reasonable to expect that resistance patterns on CBSP were always similar to the results presented here and no previous chronic exposures existed [68]. This result is in agreement to previous studies [22,69], which reported a lack of human-acquired antibiotic resistance on environments with minimal anthropic influence and no chronic exposure to antibiotics.
Besides direct exposure to antibiotics, bacterial resistance can be originated through horizontally mobile elements such as conjugative plasmids, integrons and transposons [21]. Therefore, the low rate of resistance found on the Enterobacteriaceae from CBSP also suggests a small probability of the diffusion of acquired resistance on the Park. Many authors reported that bacteria from remote areas could work as sentinels and help to evaluate the impact of anthropic pressure on wildlife and the role of wild-species and natural environments on the process of resistance acquiring, which includes not only the exposure to antibiotics but also horizontal transference [21][22][23]. Our findings reinforce the need of monitoring antimicrobial resistance in wildlife from remote areas, appearing to be an effective tool to evaluate the environment responses to anthropic pressures. On this way, more efforts should be carried out on the Park to better evaluate local resistance patterns, the impact that the human activites of the surroundings on the Park environment and the role of wildlife as reservoirs of resistant bacteria.