Introduction

The enterococci are commensal members of the gut microbiota of various animals, and the intestinal tract is considered the major reservoir of these bacteria in nature [1]. The ability to acquire genetic traits, such as virulence and antimicrobial resistance determinants contributes to the adaptability and persistence of these microorganisms in different ecological niches [2]. In this sense, in the last decades, the emergence and spread of multidrug-resistant (MDR) enterococci represent substantial clinical and epidemiological concerns, due to their role as causes of difficult-to-treat healthcare-associated infections (HAIs), and outbreaks worldwide [1, 3,4,5].

Previous studies have indicated that wild birds can be important reservoirs of MDR enterococcal strains [6,7,8,9,10,11,12,13], and suggest they can be good bioindicators of the antimicrobial resistance present in the environmental microbiome [7]. The presence of enterococci harboring resistance determinants and potential virulence factors in the intestinal tract of birds also draws attention to the potential for zoonotic transmission [2], especially to individuals in direct contact with animals, such as veterinarians at wildlife hospitals and rehabilitation centers. Moreover, the enterococci also play a role as pathogens for several animals, including birds [14,15,16,17,18]. Therefore, animal infections caused by antimicrobial-resistant enterococci can led to therapeutic dilemmas during veterinary medical treatment in a similar way to what occurs in human medicine.

Birds represent a major taxon receiving assistance in wildlife veterinary clinics. These animals are kept in captivity for treatment and rehabilitation, so their subsequent environmental reintegration may contribute to the potential dissemination of strains acquired during the captivity management that can be harmful for both the human and animal health, because of the convergence between urban and wild habitats. Additionally, due to their lifestyle and depending on their natural or temporary habitats, many wild birds may become carriers of pathogenic microorganisms, as they eat a large variety of preys, including remains of domestic and wild animals that die of natural or anthropogenic causes, and may have contact with other sources of contamination. Nevertheless, information on the occurrence and dissemination on these potential reservoirs remains largely limited [6,7,8,9,10,11,12,13], and insights into the potential role of different wild birds acting as reservoirs of antimicrobial-resistant microorganisms that can be shed and spread into the environment are of major interest for monitoring public health risks.

In the present study, we examined the distribution of species, antimicrobial resistance and virulence associated characteristics as well as aspects of the genetic diversity of enterococcal strains recovered from the intestinal microbiota of 16 different species of birds receiving attention in two centers for wildlife rehabilitation located in Rio de Janeiro state, Brazil.

Materials and methods

Sampling and specimen processing and bacterial identification

Fecal samples from 113 birds receiving veterinary assistance were collected between January and December 2013 in two wildlife rehabilitation centers (Centro de Triagem de Animais Silvestres do Rio de Janeiro, CETAS-RJ, and Centro de Recuperação de Animais Selvagens da Universidade Estácio de Sá, CRAS/UNESA) located in the state of Rio de Janeiro, Brazil. These centers provide assistance to injured animals found in the environment. Animals under antibiotic treatment were not included in the study.

Birds included 16 species from four families: Accipitridae [Roadside Hawk (Rupornis magnirostris), n = 38; Harris's Hawk (Parabuteo unicinctus), n = 3; Savanna Hawk (Heterospizias meridionalis), n = 2; White-tailed Hawk (Geranoaetus albicaudatus) n = 1; Short-tailed Hawk (Buteo brachyurus) n = 1 and Mantled Hawk (Pseudastur polionotus) n = 1]; Falconidae [Southern Caracara (Caracara plancus), n = 16; American Kestrel (Falco sparverius), n = 6; Aplomado Falcon (Falco femoralis), n = 3; Peregrine Falcon (Falco peregrinus) n = 2; Yellow-headed Caracara (Milvago chimachima) n = 1]; Strigidae [Striped Owl (Asio clamator) n = 15; Tropical Screech-Owl (Megascops choliba) n = 10; Spectacled Owl (Pulsatrix perspicillata) n = 3; Black-banded Owl (Strix huhula) n = 1], and Cathartidae [American black vultur (Coragyps atratus) n = 10]. Data on the age and sex of birds were not available.

Cloacal swab samples were collected by a veterinarian by using cotton swabs dipped into Amies transport medium (Copan, Italy) and transported to the laboratory under refrigeration. All samples were processed on the same day of collection. Each swab containing a fecal sample was homogenized in 1 mL of STGG (Skim milk, Tryptone, Glucose and Glycerin) solution [19] and two aliquots (100 µL each) were inoculated into tubes containing 1 mL of non-supplemented Enterococcosel broth (Becton Dickinson Diagnostic Systems, Sparks, MD, USA) and Enterococcosel broth supplemented with 6 µg/mL of vancomycin (Sigma Chemical Co., St. Louis, MO, USA), respectively. After incubation for 24–48 h at 37 °C, aliquots of broth cultures that showed blackening were streaked on Enterococcosel agar plates (BD Diagnostic Systems). After incubation at 37 °C for 18-24 h, up to six colonies suggestive of Enterococcus were randomly selected from each culture on Enterococcosel agar and streaked on Columbia Blood Agar plates (Plastlabor, Rio de Janeiro, RJ, Brazil) for identification.

The strains were identified based on observation of colony morphology and cellular characteristics after gram staining, and results of biochemical testing as previously described [1]. Additionally, identification was assessed by using PCR-based assays following previous recommendations [20, 21].

Antimicrobial susceptibility testing

Susceptibility to antimicrobial agents was evaluated by the disk diffusion method, according to the Clinical and Laboratory Standard Institute (CLSI, 2021) guidelines [22, 23]. The following 18 antimicrobials (all from Oxoid, Basingstoke, Hampshire, UK) were tested: ampicillin, ciprofloxacin, chloramphenicol, erythromycin, enrofloxacin, fosfomycin, high-level gentamicin (HLR-Ge), levofloxacin, linezolid, norfloxacin, nitrofurantoin, penicillin, quinupristin/dalfopristin, rifampicin, High-level streptomycin (HLR-St), teicoplanin, tetracycline and vancomycin. The minimal inhibitory concentrations (MIC) of vancomycin and linezolid were determined by using E-test strips (bioMérieux, Macy L’Etoile, France). Streptomycin and gentamicin MICs were determined on Mueller–Hinton agar supplemented with streptomycin or gentamicin (concentration ranges: 0.25 to 2048 mg/L for streptomycin and 0.25 to 512 mg/L for gentamicin).The results were interpreted according to the CLSI guidelines [23]. Multidrug resistant (MDR) strains were defined as those shown to be resistant or intermediate to three or more different classes of the antimicrobials tested [24]. Standards strains were used as a quality control: E. faecalis ATCC29212 for high-level aminoglycoside (streptomycin and gentamicin) and Staphylococcus aureus ATCC25923 for other antimicrobials.

DNA extraction

DNAs for all PCR testing were obtained by using the Chelex 100 resin (Bio-Rad Laboratories, Hercules, CA, USA) method, as previously described [25]. Briefly, a loopful of recent bacterial growth was suspended in a solution containing 5% Chelex and 20 mg/mL proteinase K (Invitrogen, Life Technologies, Carlsbad, CA, USA). After incubation at 56 °C for 1 h, followed by incubation at 95 °C for 10 min, the supernatants were used as template DNAs.

Detection of resistance and virulence associated genes

Enterococcal strains that showed resistance or reduced susceptibility were tested by PCR for detection of the following antimicrobial resistance-associated genes: erm(A), erm(B), mef(A/E) and msrC for erythromycin [26, 27]; tet(K), tet(M), tet(L), tet(O) and tet(S) for tetracycline [28]; aac(6′)-aph(2″)-Ia, aph(2″)-Ib, aph(2″)-Ic, aph(2″)-Id for high-level gentamicin [29]; ant(6)-Ia, ant(9)-Ia and ant(9)-Ib and ant(3)-Ia for high-level streptomycin [30,31,32] vat(D) and vat(E) for quinupristin/dalfopristin [33] and vanA, vanB, vanC, vanD, vanE and vanG for vancomycin [20].

Multiplex PCR was used to investigate the occurrence of virulence associated genes, including esp (coding for enterococcal surface protein), cylA (cytolysin/hemolysin), gelE (gelatinase), hyl (glycosyl hydrolase), asa1 and aggA (aggregation substance), efaA (endocarditis-specific antigen A), eeP (pheromone determinant) and ace (collagen-binding protein) [34,35,36].

Strain typing

Analysis of chromosomal DNA restriction profiles by pulsed-field gel electrophoresis (PFGE) was performed for 159 (106 E. faecalis, 26 E. faecium and 27 E. hirae) strains. Their selection was based on differences on antibiotic resistance profiles and in order to test at least one strain per bird. The genomic DNA was obtained, digested with SmaI (New England Biolabs, Ipswich, MA, USA) and subjected to PFGE, following a protocol based on previously described recommendations [11]. Restriction profiles were analyzed by using the BioNumerics software version 7.6 (Applied Maths, Sint-Martens-Latem, Belgium) based on Dice similarity index and Unweighted Pair Group Method with Arithmetic Mean (UPGMA). Strains with indistinguishable band profiles were included in the same pulsotype.

The genetic diversity among 35 selected strains of E. faecalis [37] and 18 E. faecium [38] was also determined by MLST. Sequences Types (ST) were determined and compared with those included in the MLST databases for E. faecalis [39] and for E. faecium [40].

Results

Enterococci were recovered from 104 (92%) out of 113 collected fecal samples collected, and 260 strains were selected for further characterization. Eight different enterococcal species were identified (Table 1). E. faecalis was the predominant species, representing 63.8% (166 strains) of all strains, followed by E. hirae (42 strains; 16.2%), E. faecium (30 strains; 11.5%) and E. gallinarum (14 strains; 5.4%). The other four species, E. avium (4 strains), E. casseliflavus (2 strains), E. cecorum and E. raffinosus (1 strain each) comprised 3.1% of all the strains.

Table 1 Distribution of enterococcal species isolated from fecal samples obtained from wild birds, according to the institution
Full size table

Susceptibility testing revealed that all 260 strains were susceptible to fosfomicin, linezolid, teicoplanin and vancomicin, while variable proportions of the strains were nonsusceptible (either resistant and/or intermediate) to several of the other 14 antimicrobials tested (Table 2). Major percentages of nonsusceptibility (from 11.9% 75.0% of the strains) were observed to quinolones (enrofloxacin, ciprofloxacin, levofloxacin and norfloxacin, in this order), erythromycin, rifampin, nitrofurantoin, tetracycline and streptomycin. Differences on the occurrence or frequency of a few resistance traits were found, according to the species. For instances, resistances to gentamicin and to ampicillin (13.3% each) were only detected among E. faecium strains, while resistance to penicillin was observed among E. faecium and the single E. raffinosus strain.

Table 2 Antimicrobial nonsusceptibility among the different enterococcal species isolated from the fecal microbiota of wild birds in Rio de Janeiro, Brazil
Full size table

Analysis of antimicrobial resistance data allowed to classify 133 (51.2%) strains as MDR, showing a large variety of MDR profiles, composed by simultaneous resistance encompassing 3 to 12 antimicrobials. At least one MDR strain was recovered from 71 (68.2%) of the 104 birds showing to be positive for enterococcal isolation. Considering the distribution of MDR according to the four families of birds investigated, MDR frequency was 57.0% among strains recovered from Cathartidae and Strigidae compared to 46.5% and 48.1% among isolates obtained from Accipitridae and Falconidae, respectively. Simultaneous resistance to enrofloxacin, erythromycin and rifampicin predominated, and was frequently associated to additional resistance traits, as exemplified in Table 3. Complex MDR profiles comprising resistance to at least six antimicrobials were found among the four more frequent enterococci species identified.

Table 3 Antimicrobial nonsusceptibility profiles and MLST sequence types (STs) among selected multidrug-resistant enterococci isolated from the fecal microbiota of wild birds in Rio de Janeiro, Brazil
Full size table

The presence of antibiotic resistance genes was investigated by PCR in all resistant or intermediate enterococci (Table S1). HLR-St was mainly related to the presence of the ant(6)-Ia gene (in fourteen E. faecalis and in two E. faecium and two E. gallinarum strains). The ant(9)-Ia gene was identified in a single strain of HLR-St-E. faecium (MIC = 2048 µg/mL) whereas the ant(9)-Ib gene was identified in two strains of HLR-St-E. faecalis (MIC > 2048 µg/mL). Ten strains presenting HLR-St (≥ 2048 µg/mL) did not harbor any of the genes investigated. In addition, HLR-Ge resistance was detected in four strains of E. faecium; three containing the aph(2’)-Id gene and one carrying the aac(6’)-aph(2″)-Ia gene. Tetracycline-resistant strains (total of 55 strains; 21.1%) harbored tet(M) (23 strains) or a combination of both tet(M) and tet(L) (30 strains), while two strains carried the tet(S) gene.

Resistance to erythromycin (detected in 155 strains; 59.6%) was associated with the presence of the erm(B) gene (19 strains), the msrC (10 strains), the mef(A/E) gene (9 strains), a combination of both erm(B) and mef(A/E) (5 strains), or a combination of both erm(A) and erm(B) (one strain), or both mef(A/E), erm(B) and msrC (3 strains), while 102 strains were negative for all four erythromycin resistance genes tested. The vat(D) gene, associated with streptogramin A resistance, was detected in four (1.5%) strains (three E. hirae and one E. faecium).

The presence of virulence-associated genes was predominant among E. faecalis strains, with 162 (97.6%) harboring at least three virulence genes, simultaneously (Table 4). Most of the 94 strains belonging to species other than E. faecalis did not carry any of the virulence genes investigated, except for thirteen E. hirae (efaA in 10 strains; efaA + eeP in two strains; ace + efaA + eeP + gelE in one strain), nine E. gallinarum (gelE in seven and hyl in two strains), two E. avium (esp) and one E. casseliflavus (hyl) strains.

Table 4 Distribution of the virulence genes among enterococcal species isolated from the fecal microbiota of wild birds in Rio de Janeiro, Brazil
Full size table

The pulsotypes obtained for 106 E. faecalis strains (Fig. 1) isolated from Accipitridae (45 strains), Cathartidae (13 strains), Falconidae (20 strains) and Strigidae (28 strains) were randomly named as Fs1 to Fs84. Twelve groups of pulsotypes were detected, while 72 strains corresponded to unique profiles (Fs13 to Fs84).

Fig. 1
figure 1

Dendrogram showing genotypic relationships based on the analysis of SmaI PFGE profiles, and sequence types (ST) defined by MLST of Enterococcus faecalis strains isolated from the fecal microbiota of wild birds in Rio de Janeiro, Brazil

Full size image

PFGE analysis of 26 E. faecium strains isolated from Accipitridae (seven strains), Cathartidae (six strains), Falconidae (seven strains) and Strigidae (six strains) allowed to observe 15 pulsotypes, randomly named as Fm1 to Fm15. Seven groups of pulsotypes were outlined, while eight profiles (Fm8 to Fm15) were represented by only one strain (Fig. 2). A total of 23 STs were identified among the 35 E. faecalis strains investigated (Figs. 1 and 3) while 14 STs were found among the 18 selected E. faecium strains (Figs. 2 and 3).

Fig. 2
figure 2

Dendrogram showing genotypic relationships based on the analysis of SmaI PFGE profiles, and sequence types (ST) defined by MLST of Enterococcus faecium strains isolated from the fecal microbiota of wild birds in Rio de Janeiro, Brazil

Full size image
Fig. 3
figure 3

Minimum-spanning tree representing the analysis of the allelic profiles generated by MLST for 35 Enterococcus faecalis strains (A) and 17 Enterococcus faecium strains (B) recovered from wild birds in Brazil. Each color represents one of the four families of birds investigated. The size of the nodes indicates the number of strains comprised by each ST. Higher levels of genetic relationships between nodes are indicated by darker lines

Full size image

For PFGE analysis of E. hirae, 27 strains isolated from Accipitridae (six strains), Cathartidae (six strains), Falconidae (six strains) and Strigidae (nine strains) were selected. The pulsotypes were named H1 to H24, and two groups of pulsotypes comprising a total of 5 strains were observed, while 22 strains corresponded to single pulsotype (H3 to H24) (Fig. 4).

Fig. 4
figure 4

Dendrogram showing genotypic relationships based on the analysis of SmaI PFGE profiles of Enterococcus hirae strains isolated from the fecal microbiota of wild birds in Rio de Janeiro, Brazil

Full size image

Discussion

Despite their commensal nature, the enterococci have become one of the most common causes of HAIs, besides being implicated in a number of non-HAIs in humans and animals that frequently are difficult to treat [1, 3,4,5, 35]. In this context, carriage and spread of MDR enterococcal strains by birds may have implications for public health, since handling of these animals and the disposal of their waste represent a hazard for professionals involved in animal care activities, such as veterinarians, biologists, and animal keepers, and for the environment and general population as well.

In the present study, we have found that a considerable proportion of enterococci recovered from fecal samples of wild birds harbored multidrug resistance and virulence traits that can contribute to the role of these microorganisms as opportunistic and difficult to treat pathogens. These findings corroborate the assumption that birds may be good sentinels that help to evaluate the circulation of potential pathogens in certain environments. Carriage of MDR enterococcal strains by free-living birds has been documented in different countries [8, 10, 12, 13]. Moreover, the presence of resistance and virulence determinants among enterococci isolated from animal derived foodstuff, such as chicken and beef, has been reported in various locations [41, 42], including Brazil [43, 44]. Most of the data available are related to E. faecalis and E. faecium, while in the present work we were able to show that other species, particularly E. hirae and E. gallinarum may also contribute to the scenario of potentially hazardous opportunistic pathogens that deserve consideration under the One Health perspective.

PFGE analysis of the three predominant enterococcal species identified in this study (E. faecalis, E. faecium and E. hirae) indicated that some lineages recovered from different birds kept in the same institution (CETAS-RJ or CRAS/UNESA), shared indistinguishable pulsotypes, emphasizing the possibility of transmission of strains between birds during the period of captivity. Nevertheless, E. faecalis pulsotypes Fs1, Fs3, and Fs10 were shared among strains recovered from birds kept at the two institutions investigated. The occurrence of indistinguishable pulsotypes and STs among E. faecalis isolated from free-living birds was previously reported in Poland [12]. Such findings illustrate the difficulties in establishing epidemiologic links among enterococcal strains in some instances [11, 12, 45]. In general, a good correlation between the results obtained by the PFGE and MLST techniques was observed for both E. faecalis and E. faecium strains. However, two E. faecalis strains showed identical pulsotypes and distinct STs (ST40 and ST81), a kind of discrepancy that has previously been reported for E. faecalis [3, 4]. Although both PFGE and MLST are considered highly discriminatory, the occurrence of discrepancies between results is understandable because these typing techniques are based on different principles. PFGE pulsotypes result from the occurrence and changes in restriction sites along the entire chromosomal and involve addition or deletion of DNA, while MLST detects punctual nucleotide changes within seven selected housekeeping genes [46].

Vancomycin-resistant enterococci (VRE) were not identified in this study, in accordance to previous Brazilian studies that had not detected VRE in animals [11, 47,48,49,50,51]. However, dissemination of vanA-containing E. faecium strains belonging to hospital-associated MLST lineages has been reported from urban rivers and retail chicken in Brazil [44, 52], highlighting the importance of performing surveillance of antimicrobial-resistant strains in environmental and animal-derived foodstuff samples.

The presence of virulence-associated genes among enterococcal strains is also of paramount importance to estimate the circulation of variants potentially more harmful to public health. The four virulence genes more frequently detected in the present study (ace, eeP, efaA, and gelE) are commonly reported among enterococcal strains, particularly E. faecalis, isolated from different sources [36, 42, 43]. On the other hand, the haemolysin-cytolysin encoding gene (cylA) was detected in a percentage (33.5%) higher than usually reported among E. faecalis strains recovered from hospitalized patients (6.8%) [56], food (7.2%) [53], and marine animals (8.4%) [47], in Brazil. In our report, 17.5% of the E. faecalis strains and 4.7% of the E. avium strains carried the esp gene. Lower prevalences of the esp gene among E. faecalis and E. faecium have previously been described among strains from non-clinical sources [41, 43, 45, 53, 54].

In the present study, the efaA gene was observed in E. faecalis (157 strains, 94.6%) and E. hirae (13 strains, 30.9%). The antigen A (EfaA) is associated with the adherence of bacteria to both biotic and abiotic surfaces, and with the formation of biofilms [36]. The presence of the antigen A (EfaA) was observed among E. faecalis isolated from cases of endocarditis in dogs in Portugal, suggesting a correlation between presence of efaA and occurrence of endocarditis in these animals [16]. Occurrence of E. hirae is not uncommon among mammals and birds [9, 11, 45, 49, 50, 51]. On the other hand, although human infections caused by this organism appear to be rare, all cases described to date involve bacteremia associated with severe illnesses, including endocarditis [55].

Although the number of strains submitted to MLST in the present work was limited, the results allow to establish comparative considerations. Among the STs observed for E. faecalis strains from birds, nine (ST4, ST47, ST81, ST116, ST300, ST314, ST762, ST763, and ST931) were carried by more than one bird, and five of them (ST47, ST81, ST300, ST314, ST763) were represented by strains carried by birds kept in different institutions. However, E. faecalis strains belonging to the same STs presented distinct virulence and resistance profiles, suggesting that these STs are widely disseminated in the environment. Three of the STs found among birds (ST47, ST116 and ST314) have been described more frequently among strains from hospitalized patients, and have also been documented among isolates recovered from farm animals [39]. Carriage of E. faecalis belonging to ST116 was previously observed among Magellanic penguins in North Coast of Rio Grande do Sul, in the south of Brazil [51]. Also in Brazil, E. faecalis strains belonging to ST116 and ST300 were isolated from vertebral osteomyelitis in broilers [17]. ST762, identified in two birds (Strigidae and Falconidae from CRAS/UNESA), differs from ST21 by the pstS allele. E. faecalis ST21 strains have been detected in hospitalized patients, environmental samples, food of animal origin, and livestock [37, 41, 42, 53]. E. faecalis ST314 appears to be a commensal of the microbiota of animals [39, 41, 57], and it has been found among strains from hospitalized patients in Rio de Janeiro, Brazil [56]. ST314 was also detected after sewage treatment in Tunisia [54] and urban wastewater in Canada [42] reinforcing the concept that water can contribute to the spreading of MDR enterococci in the environment.

E. faecium ST25 was first isolated from a hospitalized patient in France in 1986 and, since then, it has been found among clinical strains in Europe, Asia and Latin America, sometimes carrying the vanA gene [5, 38, 57, 58]. E. faecium ST1274 was isolated from a Tropical Screech-owl [59], a species of bird common in urban areas of Rio de Janeiro, and it is a single locus variant of ST205 and a double loci variant of ST17, belonging to hospital-associated MLST lineages. Other studies in different parts of the world have also reported the presence of STs associated with this E. faecium lineages in the microbiota of wild birds [6,7,8, 10, 12].

Overall, the four families of wild birds investigated in the present study were shown to carry a diversity of potentially harmful enterococcal variants displaying multiple antimicrobial resistance and virulence genes. Due the lifestyles of these animals, including feeding characteristics, they may acquire and become reservoirs and disseminators of opportunistic pathogens for humans and other animals. Thus, considering their potential role as sentinels of the presence of microorganism that may represent hazardous to both animal and human health, birds provide an important target to monitor the characteristics of enterococci circulating in different contexts related to the One Health perspective. This study, however, did not directly investigate the transmission dynamics between wild birds, anthropogenic activities, or environmental context. A deeper understanding of the genetic pathways correlated with virulence and resistance traits, including the mobile genomic elements implicated, may be possible through additional molecular research.