Development and Application of Optimized Isolation Methods and Diagnostic PCR Protocols for the Detection of Pathogenic Enterococcus cecorum Isolated from Broiler Chickens Open Access

Pathogenic strains of Enterococcus cecorum (EC) cause outbreaks of fatal septicemia and paralytic disease in broiler chicken flocks worldwide, resulting in substantial economic losses (1). Commensal EC strains colonize the chicken gastrointestinal tract starting at 3–4 wk of age and are prevalent in the normal gut flora (2). In contrast, pathogenic EC strains can be isolated from the gastrointestinal tract and escape this niche, causing sepsis, as early as 1 wk of age (3,4). The paralytic form of the disease, enterococcal spondylitis, occurs later in production as a sequela of EC bacteremia, predisposing the chicken to osteochondritis dissecans lesions of the free thoracic vertebra (4). However, variant strains have recently emerged and are associated with earlier mortality due to fatal sepsis before the paralytic disease is observed (5). To date, no practical treatment or preventative interventions for EC have been developed, and the source of infection and transmission dynamics are largely unknown despite regional endemicity of the disease worldwide (1).

Molecular typing and phylogenetic analysis of EC strains isolated from poultry have revealed distinct lineages of pathogenic and commensal EC variants. Genomic comparisons of strains isolated from the normal flora of healthy chickens to strains isolated from chickens with sepsis or enterococcal spondylitis lesions have revealed unique features of pathogenic EC (5,6). These include host tissue-binding proteins, an epa-like locus, and a core capsular polysaccharide biosynthesis gene cluster with a 12-kbp downstream region that is highly conserved in pathogenic EC strains but variable or entirely absent in commensal strains (6). Pathogenic EC strains isolated from early outbreaks are genetically clonal and exhibit an increased range of antimicrobial resistance, typically to tetracycline and erythromycin among others (7,8), and are intrinsically resistant to sulfonamides (1,9). As such, these distinguishing features of pathogenic EC are attractive targets to aid in EC epidemiologic, molecular typing, and clinical diagnostic investigations.

Multiple sampling and molecular detection methods have been developed to identify EC in poultry. EC is readily cultured from spleens of bacteremic birds and spinal lesions (4). The EC-specific sodA gene has been employed to speciate EC through conventional and multiplex PCR methods (Fig. 1) (10,11). More recently, a species-specific quantitative real-time PCR targeting the 16S ribosomal RNA gene of EC has been employed to better elucidate colonization patterns of EC during disease outbreaks (12). However, molecular typing methods have not been developed to differentiate commensal from pathogenic EC strains. Further, isolation of pathogenic EC from heterogeneous environmental samples with an elevated background of environmental and fecal flora has proven to be difficult (4). Therefore, the first goal of the present study was to differentiate pathogenic from commensal EC strains in clinical or experimental samples. The second goal was to couple a sensitive culture enrichment procedure with a high-throughput PCR assay that is specific to pathogenic EC. Finally, this method was implemented in an epidemiologic field investigation to identify potential sources of pathogenic EC transmission in poultry production.

Field and reference EC strains used in this study are listed in Table 1. Reference strains CE1, CE2, CE3, and CE4 were initially isolated from the gastrointestinal tract of healthy birds from unaffected flocks (6). Reference strains SA1, SA2, and SA3 were isolated from spinal lesions of birds in three epidemiologically distinct outbreaks among three separate integrated commercial broiler flocks across the southeastern United States by plating samples on Columbia agar with colistin and nalidixic acid with 5% sheep blood (CNA; BD no. 212123, Franklin Lakes, NJ) prior to incubation at 37 C with 5% CO₂ for 16–24 hr (6,13). Cream- to gray-colored, mucoid colonies with slight alpha hemolysis were presumptively identified as EC and utilized for subsequent PCR and sequencing applications (5,14). All strains were stored in tryptic soy broth with 25% glycerol at −80 C and revived on trypticase soy agar (TSA) with 5% sheep blood incubated at 37 C with 5% CO₂ for 16–24 hr. Broth cultures were grown in Todd Hewitt Broth with 1% yeast extract (THBY) and incubated standing at 37 C with 5% CO₂ for 12–16 hr. For enumeration of EC, aliquots of sample suspensions were serially diluted in phosphate-buffered saline (1X PBS), plated on TSA with 5% sheep blood, and incubated at 37 C with 5% CO₂ for 16–24 hr.

A sample enrichment protocol was developed to optimize detection of pathogenic EC from environmental and hatchery samples and tested in a simulated egg slurry model. First, overnight cultures of CE1, SA1, SA3, or LB5872 were serially diluted 10-fold in 1X PBS. These dilutions were used to spike a medium consisting of 1 ml of egg slurry made from unincubated hatching eggs from a commercial poultry hatchery and 9 ml of THBY with antibiotics at the following concentrations: nalidixic acid (15 μg/ml), neomycin (5 μg/ml), trimethoprim (1 μg/ml), and sulfamethoxazole (20 μg/ml). The suspension was incubated standing at 37 C with 5% CO₂ for 16–24 hr. A secondary enrichment was completed by diluting the primary enrichment 1:10 in THBY with the above antibiotics followed by standing incubation at 37 C with 5% CO₂ for 16–24 hr. The secondary enrichment broth was cultured on CNA for recovery of isolates and used in downstream PCR applications. Sample enrichment procedures and the diagnostic workflow to prepare PCR templates are depicted in Fig. 2. For the outbreak investigation described below, slight modifications were made to the enrichment procedure. For these samples, a 60-µl aliquot of transfer residue was added to 540 µl of THBY plus the above antibiotics in the first enrichment step followed by a similar dilution of 60 µl of culture to 540 µl of broth for the second step. In this setup, the template for real-time PCR was prepared by adding 50 µl of the secondary enrichment culture to 450 µl of molecular-grade water prior to boiling. The downstream techniques for the PCR remained the same.

To prepare PCR templates for conventional PCR from EC cultures on solid media, a colony was suspended in 100 μl of molecular-grade water. These preparations were stored at 4 C and remained stable for 30 days. The secondary enrichment broths from the sample enrichment procedure above were used directly as templates for conventional PCR. To prepare real-time PCR templates, the above colony suspensions or enrichment broths were first diluted 1:10 in molecular-grade water. The diluted samples were then incubated in a Bio-Rad thermocycler (Hercules, CA) at 95–100 C for 10 min, held at 4 C for 10 min, and then centrifuged for 5 min in a benchtop centrifuge at approximately 2000 rpm to pellet cell debris. Aliquots of the supernatant were used immediately in downstream PCR reactions or stored at 4 C for up to 30 days.

Primer sets were optimized for multiplex reactions and cross-platform application in conventional and real-time PCR assays. Species-specific EC primers targeted the superoxide dismutase (sodA) gene and were modified from Jackson et al. (10). The sodA sequence of 23 Enterococcus (E.) species (E. asini, E. raffinosus, E. sulfureus, E. faecium, E. solitarius, E. pseudoavium, E. gallinarum, E. gilvus, E. flavescens, E. casseliflavus, E. mundtii, E. saccharolyticus, E. faecalis, E. durans, E. dispar, E. avium, E. hirae, E. malodoratus, E. pallens, E. seriolicida, E. villorum, E. columbae, and E. cecorum) was downloaded from GenBank (14) and aligned using SnapGene version 6.2.1 (http://www.snapgene.com/; Fig. 1). Primers were generated from the EC sodA region with the least homology to other species and optimized for downstream multiplex PCR. The EC species-specific sodA primers selected for validation were ECsodA Fwd: 5′-GAA AAA CAT CCA GAG TTG C-3′ and ECsodA Rev: 5′-GCG AAT ATC AGC AGG CAC TT-3′, which generated a 78-bp amplicon. To differentiate commensal EC from pathogenic EC, primers were developed to target the cpsO gene in the downstream variable region of the EC capsular locus that is conserved in pathogenic strains but absent in commensal strains (6). The pathogenic EC primers selected for validation were cpsO Fwd: 5′-GCG ATT GTT CCA AAG GTG TTA G-3′ and cpsO Rev: 5′-AGT TTG AAT GGC AAA GCT AAT TC-3′, which generated a 195-bp amplicon.

Conventional multiplex PCR reaction mixtures were prepared by combining 12.5 μl of master mix (no. 201445; Qiagen), 1 μl of each forward and reverse ECsodA and cpsO primer at 10-μM final concentration, 1.5 μl of template, and 7 μl of molecular-grade water to achieve a final volume of 25 μl. A Bio-Rad thermocycler was used for the PCR reaction with the following parameters: a single initial denaturation at 95 C for 4 min followed by 30 cycles of denaturation at 95 C for 30 sec, annealing at 50 C for 30 sec, and extension at 72 C for 30 sec. A final extension step at 72 C for 7 min completed the PCR reaction. PCR products were electrophoresed at 74 V for 1.5 hr in a 2% agarose gel with incorporated ethidium bromide in tris/borate/ethylenediaminetetraacetic acid buffer. A low-molecular-weight DNA (25–766 bp) ladder (no. N3233; New England Biolabs, Ipswich, MA) was used as a reference to estimate the size of PCR products. Gels were visualized on a Bio-Rad Gel Doc™ XR+ (no. 170-8,195). Real-time PCR reaction mixtures for each primer set were prepared by combining 5 μl of SYBR Green master mix (no. A25742; Applied Biosystems Waltham, MA), 0.5 μl of the forward and reverse primers each at 5-μM final concentration, and 4 μl of template DNA to achieve a final reaction volume of 10 μl. A Step-One Plus thermocycler (Applied Biosystems) was used for the real-time PCR reaction with the following parameters: denaturation at 95 C for 3 min followed by 50 cycles of denaturation at 95 C for 30 sec, annealing at 60 C for 10 sec, and extension at 72 C for 15 sec. Step and hold melt curve analysis ranged from 70 C to 83 C with 0.5 C incremental increases for 15 sec per step.

Ten Enterococcus spp. ATCC reference strains with overlapping sodA regions (Fig. 1) and 12 field isolates from our laboratory reference collection (Table 1), which included four commensal EC strains CE1–CE4 and eight pathogenic EC strains SA1–SA4, SA7, LB5759, LB5859, and LB5872, were used to demonstrate sodA assay specificity. The lower limit of detection (LLD) for EC was established with the CE1, SA1, SA3, and LB5872 strains. Overnight cultures grown in THBY were serially diluted 10-fold in 1X PBS and plated in triplicate on TSA with 5% sheep blood for enumeration. These dilutions were concurrently diluted 1:10 in molecular-grade water to serve as templates in the real-time PCR protocol described above.

Hatchery transfer residue, boot sock swabs of hatchery trays, and postmortem samples from euthanized birds exhibiting clinical signs of disease were collected from 14 poultry complexes experiencing outbreaks of EC-related disease in the southeastern United States from 2020 to 2022. The selective culture enrichment protocol combined with the real-time PCR assay were employed to isolate and detect pathogenic EC from the samples from five commercial broiler hatcheries within the 14 sampled broiler complexes. Transfer residue consisted of pooled samples of whole eggs identified as infertile or nonviable after 17 days of incubation. The transfer residue samples (n = 44) were placed in sterile, leak-proof bottles and immediately shipped on ice to an off-site laboratory for processing. For sample hatchery trays, a boot sock swab (no. 10001906 [BTSW001SM]; Romer Labs) was placed over a clean gloved hand and used to wipe down the tray including hatch debris and meconium. The boot sock swabs were returned to their original sterile bags and shipped on ice. The egg residues and fluid from the boot sock swabs were processed through the enrichment protocol. For the postmortem samples, direct culture methods were used. Tissues including spleen and liver were aseptically collected in sterile Whirl-pak® bags (no. B01062; Nasco, Chicago, IL). Stuart swabs (no. 220099; BD, Sparks, MD) were used to aseptically sample heart, air sac, yolk sac, bone, and spinal lesions (n = 262). To culture possible spinal lesions, the free thoracic vertebra was exposed and sprayed with 70% ethanol. A transverse incision was then made through the vertebra using aseptic technique. The incised vertebra, including the vertebral medulla, was sampled with a swab. Tissues and swabs were shipped on ice to an off-site laboratory for processing. Macerated tissues and swabs were directly plated on CNA and incubated at 37 C with 5% CO₂ for 16–24 hr for isolation, template preparation, and real-time PCR as described above.

Isolates (n = 56) collected and typed as pathogenic EC during the outbreak investigation using the above methods were genotyped using pulsed-field gel electrophoresis (PFGE) according to methods previously described for Streptococcus suis (15) with modifications described by Borst et al. (13). Two internal control reference strains, SA3 and CE1, were included. Dendrograms were generated with a 1% tolerance for band matching using band-based cluster analysis and the unweighted pair group method using Bionumerics software version 7.5 (Applied Maths, Austin, TX). An 80% similarity coefficient was used to assign subgroups, and strains with 100% similarity were designated as genetically clonal.

Species-specific sodA primers produced a 78-bp amplicon, while pathogenic EC-strain-specific cpsO primers generated a 195-bp amplicon (Fig. 3A). A single distinct band was visualized in lanes with commensal EC, and two distinct bands were visualized for pathogenic EC strains (Fig. 3A). An E. faecalis negative control template did not amplify.

The specificity of the EC sodA primers in the real-time PCR was validated against E. avium, E. columbae, E. dispar, E. durans, E. faecalis, E. hirae, E. malodoratus, E. pullens, E. seriolicida, and E. villorum. The specificity of the cpsO primers was validated against four commensal EC strains CE1–CE4, all PCR negative, and eight pathogenic EC strains, SA1–SA4, SA7, LB5759, LB5859, and LB5872, which were all positive for cpsO. CE1 and SA3 were used to measure the sensitivity of the sodA and cpsO PCR assays. Prior to the 1:10 dilution made for template preparation, the LLD for CE1 was 40 colony-forming units (CFU)/ml for sodA, and the LLD for SA3 was 80 CFU/ml for sodA and cpsO. The sodA assay was determined to be positive with confidence at a cycle threshold (CT) value <43 with a melt temperature of 75 C, and the cpsO assay was determined to be positive at a CT value <40 with a melt temperature of 75 C (Fig. 3B).

Four strains, CE1, SA1, SA3, and LB5872, were tested in the enrichment sensitivity. Overnight broths of each strain were serially diluted 10-fold in 1X PBS. The primary enrichment medium consisting of 1 ml of egg slurry and 9 ml of THBY with nalidixic acid (15 μg/ml), neomycin (5 μg/ml), trimethoprim (1 μg/ml), and sulfamethoxazole (20 μg/ml) was inoculated with 100 μl of the diluted strains. Inoculum concentrations spanning 10⁴ to 10⁻² CFU were tested. All inocula containing 10⁰ CFU (1 to <10 bacteria) or more produced visible growth in the primary enrichment broth, while no growth was visible in the primary enrichment seeded with less than 10⁰ CFU. This pattern was replicated when 1 ml of the primary enrichment was transferred to 9 ml of THBY plus the above antibiotics for the secondary enrichment. The titers for the inoculum, the primary and secondary enrichments, and the sodA and cpsO CT values are presented in Table 2 for the 10¹, 10⁰, and 10⁻¹ CFU inocula.

Postmortem tissue and swab samples including spleen, liver, heart, air sac, yolk sac, bone, and spinal lesions were collected from broilers showing clinical signs of EC-related disease from 14 complexes over a 2-yr period. Pathogenic EC samples were identified from all 14 complexes and all five hatcheries. Two-hundred of the 262 cultures identified as EC sodA positive were also cpsO positive in the real-time PCR, identifying them as pathogenic strains of EC. Forty-four transfer residue and two hatchery tray samples from hatcheries associated with the 14 poultry complexes were cultured in the two-step enrichment protocol and then subjected to real-time PCR. All 46 samples were sodA positive, and 37 samples, including the hatchery trays, were also cpsO positive. A subset of the cpsO-positive samples, representing 37 tissue or swab samples and 18 transfer residue samples, was selected for PFGE. Using an 80% similarity coefficient to assign subgroups, four clades were revealed. Clade 1 was comprised of isolates from spinal lesions, while Clade 3 contained isolates from early sepsis mortality. Most of the transfer residue samples were assigned to Clade 2, which also included SA3, a standard spinal isolate, and several early mortality isolates. Within this group, two transfer residue isolates shared 100% similarity with two early mortality isolates.

Pathogenic EC has plagued the broiler industry worldwide since it was first described in 2002 (16). Clinical manifestations of pathogenic EC disease range from sepsis to paralysis due to infection of the free thoracic vertebra. There has been no effective treatment to date, and detriments to animal welfare and economic losses persist. Further, the inability to differentiate between commensal and pathogenic EC has hindered clinical diagnosis and research models and made tracing the source of pathogenic EC impossible. The goal of this work was to develop easily accessible diagnostic tools to identify pathogenic EC and enrichment techniques to aid in the tracing of the source of EC infection.

When EC first emerged as a poultry pathogen, it was difficult to identify either commensal or pathogenic strains of EC by commercially available platforms, which required training to make accurate EC identifications (1). The differentiation of enterococcal species by PCR amplification of sodA was developed by Jackson et al. (2004) and encompassed enterococcal species found in poultry, including E. faecalis, E. faecium, E. gallinarum, and EC (10). The species-specific sodA PCR proved to be an invaluable first step in identifying EC but did not separate commensal from pathogenic strains. To distinguish pathogenic strains, we relied on previous comparative whole genome sequencing that revealed unique sequences conserved in pathogenic strains (6). A 12-kbp region located downstream from the four core capsule biosynthesis genes, cpsACDB, was found to be highly conserved in pathogenic strains but variable or absent in commensal strains. This region contains genes encoding polysaccharide biosynthesis, phosphotransferases, and others associated with capsule production found in other enterococcal species (17). After screening several candidate genes, the cpsO gene was selected as the target gene in this capsular biosynthesis region to identify pathogenic EC strains. The protein encoded by cpsO belongs to the oligosaccharide flippase family that exports capsular polysaccharide to the cell surface.

The multiplex PCR has also been valuable in interpreting culture results in experimental trials involving challenges with pathogenic EC strains. In these trials, aseptic collection of samples is of utmost importance, but contamination from the environment or fecal matter can be unavoidable. These contaminated cultures can still be evaluated if the multiplex PCR is implemented to differentiate a pathogenic challenge strain from contaminating commensals.

A need to increase the throughput and turnaround time for pathogenic EC identification for accurate diagnostic testing and specific epidemiologic investigations was identified (12). To address this need, we developed a real-time PCR assay specific to pathogenic EC. Primers targeting EC sodA and cpsO were modified to decrease the amplicon sizes for real-time PCR and allow for PCR product differentiation in standard gel electrophoresis PCR applications. The specificity of the EC sodA primers was validated against enterococcal species with similar sodA sequences: E. avium, E. columbae, E. dispar, E. durans, E. faecalis, E. hirae, E. malodoratus, E. pullens, E. seriolicida, and E. villorum (Fig. 1). The cpsO PCR identified both pathogenic spinal isolates and strains isolated from poultry affected by sepsis before spinal lesions developed.

These standard multiplex PCR and real-time PCR protocols, neither requiring separate DNA isolation steps, allow for accurate identification of pathogenic strains. Tracing the source and spread of pathogenic EC outbreaks is paramount to limiting or eradicating EC-related disease, as other control measures have been unsuccessful (1). Standard vaccine approaches are unlikely to be effective as chicks are colonized and become septic in the first 2 wk after hatching (4). A previous study showed that vaccinating breeder hens with an EC bacterin resulted in the transfer of antibodies to the chicks, but these antibodies were not protective against EC challenge (18). In fact, the failure of primary chicken macrophages to phagocytose even opsonized pathogenic EC suggests that vaccines will not aid in the control of EC disease (18). Other approaches to lowering EC morbidity and mortality include direct fed prebiotics and probiotics (19). Some of these products have shown promise in decreasing EC morbidity and are now commercially available (20), but this intervention does not eliminate pathogenic EC from production.

A process in which pathogenic EC is cultured from production systems and combined with genomic analysis is required to follow transmission of specific strains. During early outbreak investigations, attempts to isolate EC on direct culture from poultry house environmental samples or from late-dead embryos from the associated hatchery were unsuccessful (21). In another study, environmental sampling followed by broth culture enrichment and plate cultures also failed to isolate EC (4). Exhaustive environmental sampling by Tessin et al. (2024) from two broiler farms through six production cycles was able to identify EC through qPCR using 16s RNA primers, but not by culture (22). The two-step enrichment protocol presented in the current study is based on ideal growth conditions and medium for EC and includes an antibiotic cocktail to prevent the overgrowth of Gram-negative contaminants (8). The sensitivity of the enrichment protocol was tested experimentally with one commensal EC strain (CE1) and three pathogenic EC strains (SA1, SA3, and LB5872). For all four strains, the sensitivity was found to be 10⁰ (Table 2). This was equivalent to less than 10 bacteria in the initial 1 ml of egg slurry. The culture broth obtained from environmental samples or colonies of cultured EC can be directly used to prepare templates for the real-time PCR assay without requiring DNA isolation steps. This greatly expedites the screening of samples prior to bacterial isolation in culture.

A proof-of-concept study was employed to explore the prevalence of pathogenic EC in an integrated broiler production system. The enrichment protocol described in this study was used to culture transfer residue or hatchery trays from five hatcheries. Pathogenic EC was isolated from each hatchery from egg transfer residue. Based on PFGE analysis, two of the hatchery isolates (5876 and 5878) shared 100% similarity with two isolates from septic broilers (5800 and 5880 isolated from the heart and air sac, respectively) at farms within the production system (Fig. 4). Although vertical transmission has been suspected, the source and reservoir of pathogenic EC have remained elusive (21). The clonality of these strains is the first evidence to support vertical transmission of pathogenic EC. The PFGE analysis also revealed that pathogenic EC isolated from spinal lesions cluster together in Clade 1, while heart isolates from cases of pericarditis in suspected sepsis mortality primarily cluster in Clade 3. This cluster pattern suggests that EC strains isolated from spinal lesions are genetically distinct from EC strains isolated from pericarditis lesions. Further genotyping is warranted to identify pathogenic EC strain differences isolated from different clinical presentations of disease.

The identification of cpsO as a marker for pathogenic EC makes it possible for the first time to differentiate between commensal and pathogenic strains of EC. Standard multiplex PCR and real-time PCR protocols based on EC sodA and cpsO are presented here to easily identify EC and to distinguish pathogenic strains from the common commensal EC strains. Culture and enrichment methods described in this work also facilitate the detection and isolation of EC from diseased birds and hatchery samples. Further, isolation of clonal strains of pathogenic EC from hatchery trays, transfer residue, and diseased birds is an initial step towards proving vertical transmission via contaminated eggs laid by broiler breeders harboring pathogenic EC. This work will allow broiler production systems, researchers, and veterinarians to trace the transmission of pathogenic EC to hopefully aid in its eradication.

We would like to thank Dr. Charlene Jackson for her support and for generously sharing her reference strains with our laboratory.

CFU =

colony-forming units;

CNA =

Columbia agar with colistin and nalidixic acid with 5% sheep blood;

EC =

Enterococcus cecorum;

LLD =

lower limit of detection;

PBS =

phosphate-buffered saline;

PFGE =

pulsed-field gel electrophoresis

THBY =

Todd Hewitt Broth with 1% yeast extract

TSA =