Virulence genes, antibiotic resistance and integrons in Escherichia coli strains isolated from synanthropic birds from Spain

Abstract

The aim of this study was to determine the presence of virulence genes and antibiotic resistance profiles in 164 Escherichia coli strains isolated from birds (feral pigeons, hybrid ducks, house sparrows and spotless starlings) inhabiting urban and rural environments. A total of eight atypical enteropathogenic E. coli strains were identified: one in a house sparrow, four in feral pigeons and three in spotless starlings. Antibiotic resistance was present in 32.9% (54) of E. coli strains. The dominant type of resistance was to tetracycline (21.3%), ampicillin (19.5%) and sulfamethoxazole (18.9%). Five isolates had class 1 integrons containing gene cassettes encoding for dihydrofolate reductase A (dfrA) and aminoglycoside adenyltransferase A (aadA), one in a feral pigeon and four in spotless starlings. To our knowledge, the present study constitutes the first detection of virulence genes from E. coli in spotless starlings and house sparrows, and is also the first identification worldwide of integrons containing antibiotic resistance gene cassettes in E. coli strains from spotless starlings and pigeons.

Introduction

Water and food-borne infections carried by pathogenic Escherichia coli strains, such as the Shiga toxin producing E. coli (STEC) or enteropathogenic E. coli (EPEC), are an important health issue. STEC infections can cause severe life-threatening diseases in humans and are the fourth most frequently reported zoonoses in the European Union (European Food Safety Authority [EFSA], 2012). EPEC can cause diarrhoea in humans (Zhang et al., 2002). In EPEC strains the intimin, an outer membrane protein encoded by the eaeA gene, binds the host's intestinal cells (Kobayashi et al., 2009). EPEC strains are considered “typical” (tEPEC) when they possess an EPEC adherence factor (EAF) plasmid and a bundle-forming pilus (bfpA) gene (Gunzburg et al., 1995) or are considered “atypical” (aEPEC) when strains are negative.

Another health concern is the presence of antibiotic-resistant bacteria, an emerging global problem in human and veterinary medicine (Levy & Marshall, 2004). Bacteria can become resistant to antibiotics through mobile genetic elements such as plasmids, transposons and integrons (Prescott, 2000), the latter of which are able to integrate or excise gene cassettes in their structures. These genes could encode for antibiotic resistance (Stokes & Hall, 1989). The class 1 integrons remain the most common integrons found in members of the family Enterobacteriaceae (Ruiz-Martínez et al., 2011).

Wild animals are regarded as a source of zoonotic and antibiotic-resistant bacteria such as certain E. coli strains (Kruse et al., 2004; Dolejska et al., 2007; Wallensten et al., 2011). Wild species could become sentinels for the spread of resistant bacteria from farms into surrounding ecosystems (Sayah et al., 2005). Wild bird species play an important role in the faecal contamination of drinking water and of agricultural crops (Lillehaug et al., 2005). They may also come into close contact with domestic birds, thereby facilitating the direct transfer of infectious agents and antibiotic resistance genes, especially to backyard poultry (Lillehaug et al., 2005).

Although there is considerable concern about antibiotic resistance in human and farm-animal bacteria, the spread of resistance into wider ecosystems has received much less attention (Livermore et al., 2001). Given the relevance of this issue, it is essential to improve our knowledge of the behaviour of antibiotic resistance in wildlife, particularly in humanized environments. Thus, the objective of this study was to investigate the role of synanthropic birds as reservoirs of virulence genes and antibiotic-resistant E. coli.

Materials and Methods

Samples and strains

The E. coli strains investigated in this study were part of the bacterial culture collection of the Animal Health Research Centre (CISA-INIA, Madrid, Spain). The strains were collected from previous technical studies of birds in which zoonotic pathogens such as Campylobacter jejuni, Chlamydophila psittaci (for example, Vázquez et al., 2010) and enterobacteria were screened.

A total of 228 individuals—106 feral pigeons (Columba livia) (42 from urban environments and 64 from rural environments), 51 house sparrows (Passer domesticus), 50 hybrid ducks (Anas spp.) and 21 spotless starlings (Sturnus unicolor)—were sampled. Samples of cloacal enemas were cultured in peptone broth at 37°C for 16 h, and subsequently in VRGB media (Oxoid, Basingstoke, UK). Pink colonies were selected and further identified by RapId^® One strips (Remel Inc., Lenexa, KS, USA). An overall prevalence of 71.9% E. coli strains was found: 78.6% for urban pigeons, 78.1% for rural pigeons, 52.9% for house sparrows, 84.0% for hybrid ducks and 57.1% for spotless starlings. All of the E. coli strain growths were analysed (n = 164) in the present study (Table 1).

Table 1. Detection of virulence genes and antibiotic resistance according to the species and the origin of the strain.

	Urban species		Rural species
	Hybrid duck	Feral pigeon	Feral pigeon	House sparrow	Spotless starlings	All species
n	42	33	50	27	12	164
eaeA	0	2 (6.0%)	2 (4.0%)	1 (3.7%)	3 (25%)	8 (4.9%)
stx-1^a	0	0	0	0	0	0
stx-2^a	0	0	0	0	0	0
bfp^a	0	0	0	0	0	0
Resistant strains (%)	47.6	33.3	26.0	14.8	50.0	32.9
AMC (%)	0.0	0.0	4.0	0.0	0.0	0.6
AMP (%)	16.7	21.2	20.0	3.7	33.3	19.5
TIC (%)	16.7	21.2	8.0	3.7	16.7	15.2
TIM (%)	0.0	3.0	8.0	0.0	8.3	2.4
KF (%)	0.0	0.0	8.0	3.7	33.3	5.5
FOX (%)	0.0	0.0	0.0	0.0	0.0	0.6
EFT (%)	0.0	0.0	0.0	3.7	0.0	0.6
CN (%)	0.0	3.0	0.0	0.0	0.0	0.6
AK (%)	0.0	0.0	0.0	0.0	0.0	0.0
K (%)	7.1	3.0	0.0	0.0	0.0	3.7
NA (%)	0.0	6.1	4.0	0.0	0.0	1.8
CIP (%)	0.0	3.0	0.0	0.0	0.0	0.6
TET (%)	21.4	24.2	20.0	3.7	33.3	21.3
C (%)	2.4	15.2	0.0	0.0	8.3	4.9
RL (%)	16.7	21.2	28.0	7.4	25.0	18.9
STX (%)	16.7	12.1	12.0	3.7	25.0	14.0

AMC, amoxicillin–clavulanic acid; AMP, ampicillin; TIC, ticarcillin; TIM, ticarcillin–clavulanic acid; KF, cephalothin; FOX, cefoxitin; EFT, ceftiofur; CN, gentamicin; AK, amikacin; K, kanamycin; NA, nalidixic acid; CIP, ciprofloxacin; TET, tetracycline; C, chloramphenicol; RL, sulfamethoxazole; STX, trimethoprim–sulphamethoxazole.

^a Only analysed in eaeA-positive strains.

Virulence genes

To detect zoonotic strains of E. coli, the genes from STEC (stx-1, stx-2 and eaeA) and EPEC (eaeA and bfp) were detected by polymerase chain reaction (PCR). Briefly, DNA was extracted by heat (De Medici et al., 2003). First, a conventional PCR was applied to detect the eaeA gene (Vidal et al., 2004). All eaeA-positive samples were then tested for the presence of stx1 (Chui et al., 2010) and stx2 (Blanco et al., 2003) as per the EFSA protocols, which consider a sample STEC to be positive when at least one E. coli strain containing stx and eaeA genes is present (EFSA, 2009). Finally, eaeA-positive strains and stx-1-negative and stx-2-negative strains were tested for the bfp gene by conventional PCR (Vidal et al., 2004) in order to separate between aEPEC and tEPEC. The amplified products were separated in 2% agarose gel stained with Sybr^® Green (Molecular Probes, Inc., Eugene, OR, USA).

Antibiotic resistance

The susceptibility of E. coli strains to 16 antibiotics used in human and veterinary medicine was determined using the disk-diffusion method following the recommendations of the Clinical and Laboratory Standards Institute, formerly the National Committee for Clinical Laboratory Standards (NCCLS, 2001). Intermediate susceptible isolates were categorized as susceptible (Dyar et al., 2012).

All bacterial isolates were examined for resistance to amoxicillin–clavulanic acid, ampicillin, ticarcillin, ticarcillin–clavulanic acid, cephalothin, cefoxitin, ceftiofur, gentamicin, amikacin, kanamycin, nalidixic acid, ciprofloxacin, tetracycline, chloramphenicol, sulfamethoxazole and trimethoprim-sulphamethoxazole (Oxoid, Basingstoke, UK). E. coli strains resistant to tetracycline–ampicillin–sulfamethoxazole (see Results and Discussion) were examined for their susceptibility to streptomycin (10 µg), as a routine procedure for selection of strains for integron analysis (Herrera-León, personal communication, 2010). The streptomycin-resistant strains were screened for the presence of integrase gene int1 and gene cassettes within class 1 integrons by PCR and sequencing (Saenz et al., 2004).

Statistical analysis

Pearson's chi-square test was used to measure the bivariate probability of association between species and the percentage of antibiotic resistance. The same test was also used to measure the bivariate probability of differences according to the environment (urban or rural) and to compare the resistances between species to the antibiotics ampicillin, tetracycline, ticarcillin, sulfamethoxazole and trimethoprim–sulfamethoxazole. Resistance was categorized as follows: no resistance, resistance to one or two antibiotic classes, or multi-resistance—that is, resistance to ≥3 antibiotic classes (Karczmarczyk et al., 2011). P < 0.05 was considered significant. All statistical studies were conducted using SPSS v.15.0 software (SPSS Inc, Chicago, IL, USA). To avoid a species-based bias, only feral pigeons were compared when testing whether there were statistically significant differences between urban and rural strains.

Results and Discussion

Four virulence genes (eaeA, stx-1, stx-2 and bfp) associated with diarrhoeagenic E. coli, belonging to pathotypes EPEC or STEC, were searched for using PCR (Table 1). The eaeA gene was detected in eight E. coli strains. All of these eaeA-positive strains were analysed to determine the presence or not of stx-1, stx-2 and bfp. All were negative and were classified as aEPEC. Other studies have isolated tEPEC and aEPEC strains from humans and animals (Trabulsi et al., 2002; Blanco et al., 2005), while EPEC has been isolated from poultry (57%) and pigeons (7%) in Finland (Kobayashi et al., 2002). According to our results, feral pigeons, spotless starlings and house sparrows could be considered carriers of aEPEC strains. This study is the first to report eaeA in spotless starlings and house sparrows. The results of the present study indicate the importance of synanthropic birds as reservoirs and possible vectors of virulence genes and potential zoonotic agents.

The results for detection of virulence genes and antibiotic resistance in terms of location and species are summarized in Table 1. The most frequent resistance found was to tetracycline (21.3%), followed by ampicillin (19.5%), sulfamethoxazole (18.9%), ticarcillin (15.2%) and trimethoprim–sulphamethoxazole (14.0%). For amoxicillin–clavulanic acid, cefoxitin, ceftiofur, ciprofloxacin and gentamicin, only one isolate was resistant in each case. All isolates were susceptible to amikacin. The most frequent profile of antibiotic resistance was tetracycline–ampicillin (14.0%). Feral pigeons were selected to compare the differences between urban and rural environments given that they are present in both. Apart from the statistically significant (P = 0.014) exception of ticarcillin in urban and rural feral pigeons (21.2% vs. 4.0%), the differences between urban and rural feral pigeons were non-significant. The reason for this could be that ticarcillin is commonly used in the treatment of Pseudomonas aeruginosa infections in human patients (Paul et al., 2010). This bacterium is frequently associated with nosocomial infections and immunocompromised hosts (Driscoll et al., 2007).

There were statistically significant differences in the percentage of resistant strains between hybrid ducks and rural feral pigeons (P = 0.031), between hybrid ducks and house sparrows (P = 0.005), and between spotless starlings and house sparrows (P = 0.020). Spotless starlings tended to have the highest percentage of resistant E. coli isolates. In hybrid ducks, the percentage of resistant strains (47.6%) was comparatively high since findings from a similar study in Poland indicated that just 27% of E. coli isolated from mallards (Anas platyrhynchos) living in city parks were resistant to at least one antibiotic (Literak et al., 2010). One of the reasons for this discrepancy between studies could be that the hybrid ducks included in the present study lived in urban parks that use water from sludge-sewage treatment plants. It is known that wastewater frequently contain antibiotics and bacteria with resistance genes (Valverde et al., 2008). In the studied house sparrows, only 14.8% of E. coli strains were resistant, a result that is slightly higher than the 9% resistance found in house sparrows in Poland (Dolejska et al., 2008).

In the present study, the most frequent form of resistance was to tetracycline, followed by resistance to ampicillin and to sulfamethoxazole. These resistance phenotypes are frequently found in livestock (Guerra et al., 2003) and humans (Phongpaichit et al., 2008) worldwide. All of the strains simultaneously resistant to tetracycline, ampicillin and sulfamethoxazole (11.0%) were examined for their susceptibility to streptomycin; in all, 16 of the isolates (9.8%) were resistant. These resistant strains were analysed by PCR for integrase and gene cassettes. Five out of 16 isolates (31.2%) were positive to integrons with gene cassettes. All of the integrons were identified as class 1 integrons containing the integrase 1 gene (int1) (Table 2). Gene cassettes were detected in all of these isolates and encoded dihydrofolate reductase A (dfrA) and aminoglycosides adenyltransferase A (aadA) with resistance to trimethoprim and aminoglycosides, respectively. The subtypes dfrA1, dfrA7, dfrA12, dfrA17, aadA1, aadA2 and aadA5 were found. The resistance profile of integron-positive isolates was relatively broad (resistant to six or more antibiotics; see Table 2). To our knowledge, the present study is the first worldwide to have detected integrons containing antibiotic resistance gene cassettes in E. coli strains from spotless starlings and pigeons. It is not clear whether these birds can truly act as reservoirs, in which E. coli colonizes the gut, or whether they act as mechanical vectors, by the result of recent infections and shedding them. Experimental studies are required to understand their possible role as reservoirs.

Table 2. Integrons found in relation to species and the observed resistance phenotype.

Species	Resistance phenotypes	Integron	Gene cassette
Spotless starling	AMP, TET, TIC, STR, RL, STX	Class 1	dfrA1, aadA1
Urban feral pigeon	AMP, TET, TIC, STR, RL, STX, TIM	Class 1	dfrA1, aadA1
Urban feral pigeon	AMP, TET, TIC,STR, RL, STX	Class 1	dfrA7
Urban feral pigeon	AMP, TET, TIC, STR, RL, STX, C, CN, NA, CIP	Class 1	dfrA17, aadA5
Rural feral pigeon	AMP, TET, TIC, STR, RL, STX	Class 1	dfrA12, aadA2

AMP, ampicillin; TET, tetracycline; TIC, ticarcillin; STR, streptomycin; RL, sulfamethoxazole; STX, trimethoprim–sulfamethoxazole; TIM, trimethoprim; C, chloramphenicol; CN, gentamicin; NA, nalidixic acid; CIP, ciprofloxacin.

Our results show that synanthropic birds can carry E. coli containing pathogenic and antibiotic-resistant genes, and therefore may spread these genes to other birds, farm animals and humans. The birds may also acquire them from a variety of sources. Monitoring studies on synanthropic birds could be useful for early detection of these genes in the environment.

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgements

This work was supported by the MARM-INIA under Grant CC08-C20, by the CAM under Grant S2009/AGR-1489 and by the INIA under Grant RTA2010-00066-C02-01. The authors want to thank Dr Michael Lockwood for the English edition of the manuscript.