Over the last decade, the Southeastern Cooperative Wildlife Disease Study has performed 264 bacterial cultures on over 50 species of free-ranging birds that were submitted for necropsy through the Southeastern Cooperative Wildlife Disease Study diagnostic service. Eighteen Salmonella isolates were obtained from those 264 bacterial cultures, giving a prevalence of 6.8%. Fourteen of the 18 cases were diagnosed as clinical salmonellosis, suggesting that these birds were not unapparent carriers. Nine of those 18 isolates were preserved, allowing this retrospective comparison of the Salmonella isolates found in free-ranging birds. Necropsy of these wild birds revealed that lesions associated with salmonellosis in passerine birds were distinctly different from the disease observed in other free-ranging and captive nondomestic birds. Is the unique pathology associated with salmonellosis in passerine birds, like goldfinches, due to a single Salmonella clone, or is the disease syndrome a function of the host species? To address the former question, we typed Salmonella isolates by pulsed-field gel electrophoresis (PFGE) and random amplified polymorphic DNA (RAPD) analysis. We further characterized these isolates as to the distribution of virulence and antibiotic resistance-associated genes. In this report, we identify a single S. enterica serovar Typhimurium genetic type commonly associated with disease in passerine species. Also, we identify the first reported case of multiple-drug-resistant S. enterica serovar Typhimurium DT104 from two psittacine companion birds, indicating a potential disease risk for both bird owners and fanciers in the United States.

Bacterial isolates. All of the Salmonella isolates used in this study are listed in Table 1. S. enterica serovar Typhimurium SR11 and Escherichia coli HB101 were included in PCR and invasion assays as positive and negative controls.

TABLE 1 Gross and histologic lesions associated with salmonellosis in nondomesticbirds

Antimicrobial susceptibility determination. Antimicrobial MICs of E. coli isolates were determined via the Sensititre automated antimicrobial susceptibility system (Trek Diagnostic Systems, Westlake, Ohio) and interpreted according to the National Committee for Clinical Laboratory Standards guidelines for broth microdilution methods (39, 40). Sensititre susceptibility testing was performed in accordance with the manufacturer's instructions. The following antimicrobials were assayed: amikacin, amoxicillin-clavulanic acid, ampicillin, ceftiofur, cephalothin, chloramphenicol, ciprofloxacin, gentamicin, kanamycin, nalidixic acid, streptomycin, sulfamethoxazole, tetracycline, and trimethoprim-sulfamethoxazole.

RAPD analysis. Salmonella isolates were typed by RAPD analysis using primer 1254 (23). RAPD PCR was performed using the Rapidcycler hot-air thermocycler (Idaho Technologies, Idaho Falls, Idaho) (60). The RAPD conditions used were those of Maurer et al. (36). DNA was separated on 1.5% agarose and 1× TAE gel with ethidium bromide (0.2 μg/ml) at 100 V for 1 h (50). The 100-bp ladder (GIBCO/BRL, Gaithersburg, Md.) was used as a molecular weight standard for determining the molecular weights of the PCR products.

PFGE. Agarose-embedded bacterial genomic DNA was digested with 10 U of restriction enzyme XbaI overnight at 37°C, and DNA fragments were separated by PFGE (3, 49) in a 1% agarose gel (Bio-Rad, Hercules, Calif.) using the CHEF DR-II electrophoretic apparatus (Bio-Rad). Electrophoresis was done for 25 h with a voltage of 200 V and a linearly ramped pulse time of 2 to 40 s (3). A lambda ladder (582 to 48.5 kb; Bio-Rad) served as the MW markers for PFGE. The restriction enzyme XbaI has proven useful in the typing of Salmonella (2, 42) and other gram-negative bacteria (3) by PFGE. For isolates with indistinguishable XbaI PFGE DNA patterns, RAPD analysis (23) and a second PFGE using BlnI as the restriction enzyme (57) were done to confirm their genetic relatedness.

Identification of S. enterica serovar Typhimurium DT104 multidrug resistance locus by PCR mapping. Since florfenicol resistance gene flo has been mapped to a site between two type 1 integrons in S. enterica serovar Typhimurium DT104 (11), PCR primers were designed to amplify a portion of flo and antibiotic resistance genes mapping upstream (qacΔE) and downstream (tetR) of this gene (GenBank accession no. AF071555). Primers flomap1F (GATGTCTAACAATTCGTTCAG) and flomap1R (CAACGTGAGTTGGATCATAG) anneal to positions 2751 of qacDE and 4609 of flo, respectively, to amplify a 1,858-bp DNA fragment. The second set of primers flomap2F (GGGATCGGCGAACTTTAC) and flomap2R (TGTGGTCGGTTCCGTTCTC), anneal at positions 5330 of flo and 6073 of tetR to produce a 744-bp amplicand by PCR. The PCRs were done using the Idaho Technologies Rapidcycler. The 10 PCR mixture consisted of 3 mM MgCl₂ (flomap1 primers) or 2 mM MgCl₂ (flomap2 primers), 50 mM Tris (pH 7.4), bovine serum albumin at 0.25 mg/ml, 0.1 mM primer, 0.2 mM deoxynucleoside triphosphates (Boehringer Mannheim; Indianapolis, Ind.), 0.5 U of Taq polymerase (Boehringer Mannheim), and 100 ng of the DNA template. The program parameters for the PCR using the flomap1 primers consisted of a hold at 94°C for 15 s; 30 cycles of 94°C for 15 s (denaturation), 40°C for 1 min (annealing), and 72°C for 1 min (extension) with a slope of 2; and a final extension at 72°C for 4 min. The PCR parameters for the flomap2 primers was 30 cycles of 94°C for 1 s (denaturation), 46°C for 1 s (annealing), and 72°C for 15 s at a slope of 2.

Detection of Salmonella virulence genes and antibiotic resistance-associated genes. Four different sets of primers for Salmonella virulence genes invA (54), spvC (54), sefC (7), and pef (7) were used to generate DNA probes by PCR as described by Dodson et al. (15). The conditions for Southern hybridization were followed as previously described (15, 50), with hybridization and washes carried out at 68°C. DNA probes for antibiotic resistance-associated genes intI1 (5, 29), aadA1 (5, 29), and merA (5, 32) were also generated by PCR. Conditions for PCR and Southern analysis were performed as described by Bass et al. (5).

Salmonella cell invasion assay. Budgie abdominal tumor cells (BATCs) were used in the invasion assays with the salmonellae. This is a permanent avian epithelial cell line derived from an abdominal tumor in a parakeet (22). Salmonella invasion was done by the procedure of Henderson et al. (22). The invasion results were reported as a ratio of the bacterial count that invaded the BATCs divided by the inoculum concentration. These values were then normalized such that the positive control, S. enterica serovar Typhimurium SR11, represented 100% invasion. All invasion levels greater than 1% were considered to be positive for invasion of BATCs (15).

Clinical syndrome associated with Salmonella infection of nondomestic birds. Salmonella isolates from free-ranging birds were submitted from four southeastern states (Georgia, Florida, West Virginia, and Kentucky) and one western state (Wyoming). Five of the cases in wild birds were epizootics that resulted in morbidity and mortality of numerous birds. The four additional cases involving wild birds were individual animal submissions.

The lesions classically associated with salmonellosis in songbirds consist of necrotic plaques in the esophagus and crop (18, 25). In this study, this multifocal, necrotizing esophagitis and ingluvitis were only seen in passerine species (cowbirds, goldfinches, and English sparrows) (Fig. 1) and S. enterica serovar Typhimurium was isolated from the necrotic foci in all cases. Histologically, the epithelial surface was focally ulcerated, leaving a thick layer of necrotic cellular debris admixed with degenerate and intact leukocytes and myriads of gram-negative bacterial rods. The underlying submucosa had a focally intense infiltrate of heterophils and fewer lymphocytes and plasma cells. No evidence was found of other known pathogens, such as avian poxvirus, that can cause esophageal lesions in passerine species. Trichomonas gallinae, a flagellate protozoan that causes esophageal disease in columbiform species, such as mourning doves and, less commonly, predator birds, has not been reported in songbird species (9).

FIG. 1 Gross pathology of an American cowbird that died of salmonellosis. The arrows delineate multiple necrotic plaques on the esophageal mucosa from which S. enterica serovar Typhimurium was isolated.

Among all of the examined bird species, the most consistent lesion reported at necropsy was emaciation but autolysis may have obscured subtle lesions in many of the cases (Table 1). Liver lesions, which were found grossly or histologically in 11 of the 22 cases, consisted of randomly scattered, small, pale foci that, histologically, were small foci of hepatocellular necrosis and inflammation. When present, lower gastrointestinal (GI) tract lesions consisted of acute hemorrhagic or necrotizing enteritis or enterocolitis. The caseous cecal plugs occasionally seen in domestic poultry were not found in these nondomestic birds. Other gross lesions less commonly encountered included small 1- to 2-mm caseous granulomas (found in the caudal airsacs of the quail and one cowbird). Other histologic lesions included orchitis and conjunctivitis (both diagnosed in cowbirds), meningitis (diagnosed in the pigeon), and septicemia (diagnosed in a cowbird, the African grey parrot, and the Senegal parrot).

GI lesions are common in the lower GI tract in poultry (20). However, the necrotic esophageal plaques seen in passerine birds have not been reported in domesticated gallinaceous birds. Disease was generally reported for adult birds, which contrasts with salmonellosis in domestic poultry where illness is generally observed in younger (1- to 7-day-old) birds (20). The Salmonella isolates from these diseased birds were genetically typed to determine if there was a pathogenic clone associated with this illness in passerine birds.

Genetic diversity among Salmonella strains associated with nondomestic birds. Five distinct XbaI DNA patterns were observed for the 23 Salmonella isolates typed by PFGE (Table 2). Isolates with seven or more different DNA fragments were assigned to a specific PFGE XbaI genetic type, A to E, as recommended by Tenover et al. (55). The majority (70%) of these Salmonella isolates belonged to PFGE XbaI genetic type A. Within this PFGE group, there were additional differences of four to six DNA fragments within PFGE pattern A to further differentiate these related Salmonella isolates into three subgroupings, A1 to A3. A group of closely related isolates were identified that differed by only two or three DNA fragments to justify a final subcategory, A1_A-B. Using a second restriction enzyme, BlnI, we were able to distinguish among Salmonella isolates by PFGE that produced similar or identical PFGE patterns, earlier, with XbaI (Table 2). PFGE using both restriction enzymes XbaI and BlnI identified nine distinct genetic types among Salmonella isolates from nondomestic birds. However, half of these isolates appeared to be possibly related, according to their PFGE pattern. Five distinct genetic types were identified among the Salmonella isolates from Georgia. Four of these five Salmonella genetic types were found only in that state. Likewise, other Salmonella genetic types, as defined by PFGE, appeared to be unique to other locales from which these specimens were obtained. Certain Salmonella genetic types therefore appear to be endemic.

TABLE 2 Salmonella isolates used in thisstudy

There are other genetic types, however, that do not appear to be geographically restricted. For example, closely related Salmonella genetic types were identified by PFGE among isolates obtained from Georgia and Wyoming. We were unable to definitively separate Salmonella isolates A7582, 98A-24966-2, and 98-28238 further using another molecular typing tool, RAPD PCR (Table 2). These genetically related salmonellae were all isolated from passerines.

We also noted that Salmonella isolates from the Lorikeet and Senegal parrot were indistinguishable by their PFGE and RAPD patterns as well. Both birds were from the same pet owner. Since Salmonella isolates from the lorikeet and the Senegal parrot had a pentadrug resistance pattern similar to that of S. enterica serovar Typhimurium DT104 (21), we compared their PFGE patterns. S. enterica serovar Typhimurium DT104 isolates are genetically related, producing PFGE patterns that are indistinguishable (1). We noted that our Salmonella isolates had PFGE patterns indistinguishable from the DNA fingerprint of S. enterica serovar Typhimurium DT104 (data not shown). In addition to similarities in PFGE patterns, these Salmonella isolates were similar to S. enterica serovar Typhimurium DT104 with regard to the genetic organization of the multidrug resistance locus (11). The Salmonella isolates from the lorikeet and the Senegal parrot had the florfenicol resistance gene flo, a marker for multidrug-resistant S. enterica serovar Typhimurium DT104 (10). In these isolates, the florfenicol resistance gene flo is flanked by quaternary ammonium resistance gene qacΔE and tetracycline resistance gene tetR. PCR analysis revealed that the psittacine Salmonella isolates had an arrangement of antibiotic resistance genes similar to that of the S. enterica serovar Typhimurium DT104 multidrug resistance locus. Phage typing confirmed that these isolates were S. enterica serovar Typhimurium DT104.

Antimicrobial susceptibility and drug resistance genes present in Salmonella isolates from nondomestic birds. Most of the Salmonella isolates from nondomestic birds (15 of 22) in this study were sensitive to 16 antibiotics. All of the isolates, including multidrug-resistant Salmonella isolates 98A-3397, 98-3398, 97-9746, and 98A-9551, were sensitive to amikacin, apramycin, ceftiofur, ceftriaxone, cephalothin, ciprofloxacin, gentamicin, nalidixic acid, and trimethoprim. Seven of the 22 Salmonella isolates were resistant to sulfamethoxazole (MIC, >512 μg/ml), and of the seven sulfamethoxazole-resistant Salmonella isolates, four were also resistant to streptomycin (MICs, 64 to >256 μg/ml). We also observed resistance to other drugs, including ampicillin (MIC, >32 μg/ml), tetracycline (MIC, ≥32 μg/ml), chloramphenicol (MIC, >32 μg/ml), and kanamycin (MIC, >64 μg/ml). With one exception, Salmonella isolates from passerine birds were sensitive to antibiotics while those from quail and psittacines were resistant to one or more antibiotics. The drug resistance patterns and percentages were clearly different in Salmonella isolates cultured from nondomestic birds compared to those from poultry (38). In addition to being resistant to sulfamethoxazole and streptomycin, the poultry Salmonella isolates are also resistant to tetracyclines and gentamicin (38).

Rapid dissemination of drug resistance is often attributed to integrons, genetic elements involved in swapping or combining of antibiotic resistance into the integration site (52). Class 1 integrons, and resistance genes associated with them, are widely distributed in nature. In our analysis of Salmonella isolates from nondomestic birds, 9 of the 22 isolates contained the integrase, intI1, gene, a marker for class 1 integrons (Table 3). In addition to intI1, Salmonella isolates also contained other markers associated with class 1 integrons, including the streptomycin-spectinomycin resistance gene aadA1 and the mercuric reductase gene merA (31), but at a lower frequency than intI1. Twelve of the 22 Salmonella isolates did not have intI1 or the other antibiotic resistance-associated genes, aadA1 and merA.

TABLE 3 Salmonellagenotypes

Virulence potential of Salmonella isolates from nondomestic birds. Salmonella pathogenesis is dictated by a series of genes responsible for colonization (56), invasion (48), and spread (4, 30) within its avian host. Host adaptation is influenced by the distribution of fimbrial and nonfimbrial adhesins among the salmonellae (7). To assess potential virulence of Salmonella isolates by the presence or absence of genes, PCR and Southern hybridization were used to detect Salmonella virulence genes. All of the Salmonella isolates contained the invasion gene invA. Most of the Salmonella isolates (17 of 22) from diseased nondomestic birds also contained the virulence plasmid, as is evident from the incidence of spvC among these isolates (Table 3). Of the 17 Salmonella isolates positive for spvC, 15 also contained the Salmonella plasmid-encoded fimbrial gene pef (7). pef appears to be restricted in its distribution to a few Salmonella serotypes, Typhimurium, Paratyphi, Choleraesuis, and Typhi (7). All of the Salmonella isolates were negative for sefC (data not shown), a fimbrial gene found primarily in the avian-adapted salmonellae, S. enterica subsp. enteritidis, S. enterica subsp. pullorum, and S. enterica subsp. gallinarum (7, 14, 15). The S. enterica serovar Typhimurium isolates from free-ranging birds do not appear to be any different from the broad-host-range S. enterica serovar Typhimurium strains associated with other avian and mammalian hosts with regard to the distribution of adhesion (7) and invasion (54) genes.

We also examined these Salmonella isolates for phenotypic differences in the ability to invade epithelial cells. All of the Salmonella isolates were considered to be invasive for avian epithelial cells, with levels ranging from 8 to 175% invasion compared to the control, S. enterica serovar Typhimurium SR11 (Table 2). These Salmonella isolates were more efficient at invading avian epithelial cells than avian host-specific S. enterica subsp. pullorum (15). Variability in the invasion phenotype did not correlate with a specific genetic type, as defined by PFGE pattern, or virulence genotype.

According to genotype and antibiotic resistance pattern, the Salmonella strains associated with captive and free-ranging nondomestic birds represent a potential threat to humans. Most of the Salmonella isolates possessed the 90-kb virulence plasmid, enabling the organisms to colonize and spread in an avian or nonavian host. The other troubling news is the identification of multidrug-resistant S. enterica serovar Typhimurium DT104 from pet birds. As far as we are aware, this is the first reported case of Salmonella DT104 from exotic birds in the United States. Nondomestic birds can obviously serve as a reservoir for transmission of salmonellae to pet owners and bird watchers. Bird feeders are popular in the United States, and determining the incidence of salmonellae at these feeders is important in order to assess their health risk to the general population.

This research was funded by a grant from the U.S. Poultry and Egg Association, the State of Georgia's Veterinary Medicine Agricultural Research Grant, and USDA Formula Funds.

We thank Stanley Vezey and Susan Little for their comments and critique of the manuscript. We also thank Sherry Ayers for her technical assistance. Thanks to Elizabeth Williams of the Wyoming State Veterinary Diagnostic Laboratory for providing the four western Salmonella isolates.