American black bear (Ursus americanus) as a potential host for Campylobacter jejuni

Abstract

The Gram-negative bacterium Campylobacter jejuni is part of the commensal gut microbiota of numerous animal species and a leading cause of bacterial foodborne illness in humans. Most complete genomes of C. jejuni are from strains isolated from human clinical, poultry, and ruminant samples. Here, we characterized and compared the genomes of C. jejuni that were isolated from American black bears in three states in the southeastern United States from 2014 to 2016. Despite the limited sample size (n = 9), the isolates displayed substantial genotypic variability, including eight distinct sequence types (STs) and variable gene content encoding surface glycan structures such as capsular polysaccharides (CPS) and lipooligosaccharides (LOS). Phylogenetic analysis identified several C. jejuni host generalist strains among the isolates from bears that clustered with isolates from domestic poultry, cattle, and environmental sources. Three isolates (SKBC94, SKBC3, SKBC5) clustered with wildlife-associated strains, exhibiting mutations or deletions in loci associated with cytolethal distending toxin production and oxidative stress resistance, potentially influencing host-specific colonization. Additionally, strains SKBC3 and SKBC5 harbored distinct Entner-Doudoroff (E-D) loci, suggesting a potential evolutionary fitness advantage. This study provides the first evidence of C. jejuni colonization in American black bears, highlighting their potential role as reservoirs for diverse C. jejuni lineages from both anthropogenic and environmental sources. Further research is needed to determine the prevalence and host specificity of C. jejuni strains in black bears and their potential implications for public and wildlife health.

Introduction

Campylobacter jejuni is the leading cause of bacterial gastroenteritis worldwide [1–3] and has adapted to the gastrointestinal tract of various avian and mammalian hosts. Human infections with C. jejuni are usually the result of consumption and/or handling of contaminated poultry products, while raw milk and untreated water are also common sources of infection [3–9]. Characterization of strain level variation, using methods such as multilocus sequence typing (MLST), has identified genotypes of C. jejuni strains found in multiple host species (e.g., domesticated poultry, ruminants, etc.) termed host generalists and genotypes of strains isolated primarily from single reservoir species (e.g., domestic cattle or guinea pigs) termed host specialists [10–14]. The genotypes of C. jejuni that are seldom associated with human disease have been identified from non-food production animals [14–19]. In most cases, these genotypes, such as those C. jejuni strains isolated from raccoons and many wild bird species are dissimilar genotypes of strains that cause human disease or colonize livestock animals. These genotypic differences represent signals of host adaptation that are used in public health studies to attribute human infection cases to likely reservoir source(s) [10,19–22]. However, in more urban settings, strains possessing host generalist genotypes can be isolated from multiple hosts species that share the same niche, which complicates source attribution in these cases [12,16,18,19,23–27].

Occasionally, direct transmission of C. jejuni to humans has been linked to wildlife [28–30], and population level analyses of human clinical campylobacteriosis cases attributes a small proportion of infections to strains with wildlife-associated genotypes [12,18,19,31]. Shedding of C. jejuni in animal feces, especially around water, provides a means for the pathogen to enter new ecological niches, exchange between hosts, and cause infections in humans [18,32,33]. In the agricultural environment, the proximity of domestic and wild animals allows certain C. jejuni lineages to be shared among different hosts [12,18]. Indeed, raccoons and guinea pigs harbor C. jejuni strains with both their own host-adapted and also domestic animal-associated genotypes [18,26].

Expanded human activity in, and development of, forested lands increase the interaction of both humans and domestic animals with wildlife. Among wild animals, black bears (Ursus americanus), which range across North America, have been significantly impacted by urban and suburban development [34]. Such development has reduced forest habitat and increased forest fragmentation, leading to increased interactions between humans and bears, as evidenced by observations and property damage [35]. Recently, the foodborne bacterial pathogen Listeria monocytogenes was found to be frequently isolated from black bears [36], and it is possible that bears may act as reservoirs and vehicles for other bacterial pathogens, including C. jejuni.

In the current study, we investigated the complete genome sequences of C. jejuni isolates collected from American black bears in the southeastern United States. Using these sequences, we conducted in silico multilocus sequence typing and performed comparative genomic analyses to determine similarity and phylogeny to worldwide C. jejuni isolates and those associated with the southeastern United States. Upon identifying two isolates with novel sequence types (STs), we explored their genetic content and identified genetic loci that are similar to loci in host-specific C. jejuni strains including C. jejuni subsp. doylei. These results identify American black bears as a new host and will help address the role that wild animals have in the epidemiology of C. jejuni.

Materials and methods

Isolates and culturing

Nine strains of C. jejuni were isolated from fecal and rectal swabs from American black bears that had initially been used to identify Listeria monocytogenes. All sampling locations, sample collection protocols, and permits have previously been described in detail by Parsons et al [36]. Briefly, the bears lived in the southeastern United States, specifically Virginia, North Carolina, and Georgia, and the bear capture and/or handling protocols were approved by the Institutional Animal Care and Use Committee at North Carolina State University (14‐019‐O), the University of Georgia (A2011 10‐004‐A1) and Virginia Tech (12‐112 and 15‐162). Specifically for Campylobacter isolation, the samples (100 mg feces or the entire swab tip) were enriched for general Campylobacter spp. by resuspension in 10 mL Bolton broth (Oxoid Ltd., Hampshire, UK) and incubation at 37°C for 24 h under microaerobic conditions generated with GasPak EZ Campy sachets (Becton, Dickinson and Co., Sparks, MD, USA). The enrichments (100 μL) were plated for thermophilic Campylobacter on modified charcoal cefoperazone deoxycholate agar (mCCDA; Oxoid) with CCDA Selective Supplement (SR0155E, Oxoid), and the plates were incubated microaerobically at 42°C for 48 h. Following incubation, an average of five putative Campylobacter colonies per sample were purified on Mueller-Hinton agar (MHA; Becton, Dickinson, and Co.), as described previously [37]. Campylobacter confirmation was conducted via PCR using primers targeting hipO, as previously described [38].

Bacterial isolate genome sequencing

DNA was extracted from C. jejuni isolates (Table 1) as described previously [39]. Sequencing was carried out using the PacBio RS II (Pacific Biosciences, Menlo Park, CA) and Illumina MiSeq platforms (Illumina Inc., San Diego, CA). For the PacBio platform, SMRTbell libraries were prepared from 10 μg of bacterial genomic DNA with fragmentation in G-TUBEs (Covaris, Woburn, MA), following a BluePippin 10-kb size selection with a 0.75% DF Marker S1 high-pass 6- to 10-kb vs3 cassette (Sage Science, Beverly, MA) with 1 × AMPure cleanup and DNA repair [40]. Single-molecule real-time (SMRT) cells were run with 0.1 nM on-plate concentration, P6/C4 sequencing chemistry, the MagBead OneCellPerWell v1 collection protocol, and 360-min data collection mode. PacBio DNA internal control complex P6 was used as an internal sequencing control, and the read quality control was conducted using FastQC (Pacific Biosciences). The PacBio reads were assembled using the RS Hierarchical Genome Assembly Process (HGAP) v3.0 in SMRT Analysis v2.2.0 (Pacific Biosciences).

Table 1. Strain data, including MLST alleles, sequence types and clonal complexes from black bear C. jejuni isolates.

			MLST alleles
Strain	Year	State	aspA	glnA	gltA	glyA	pgm	tkt	uncA	ST	CC
SKBC1	2014	North Carolina	2	21	5	2	59	1	5	222	206
SKBC3	2014	North Carolina	255	530	279	607	740	585	276	7630
SKBC5	2014	North Carolina	305	756	279	607	479	585	276	10620
SKBC9	2014	Georgia	12	6	61	147	261	32	3	10501
SKBC25	2015	North Carolina	1	6	61	244	797	32	3	10624	179
SKBC41	2015	Virginia	4	7	10	4	1	7	1	45	45
SKBC57	2016	North Carolina	2	1	1	3	2	1	5	21	21
SKBC94	2016	North Carolina	26	2	9	51	8	46	21	682	682
SKBC102	2016	North Carolina	4	7	10	4	1	7	1	45	45

ST in bold are unique as of July 2025.

For the Illumina platform, libraries were prepared with the KAPA LTP library preparation kit (Roche) with Standard PCR Amplification Module (Roche), following the manufacturer’s instructions except for the following changes: 750 ng DNA was sheared at 30 psi for 40 sec and size selected to 700–770 bp following Illumina protocols. Standard desalting TruSeq LT primers were ordered from Integrated DNA Technologies (Coralville, IA) and used at 0.375 µM and 0.5 µM final concentrations, respectively. PCR amplification was reduced to 3–5 cycles. Libraries were quantified using the KAPA Library Quantification Kit, except with 10 µL volume and 90 sec annealing/extension PCR, then pooled and normalized to 4 nM. Pooled libraries were re-quantified by ddPCR on a QX200 system (Bio-Rad), using the Illumina TruSeq ddPCR Library Quantification Kit and following manufacturer’s protocols, except with an extended 2 min annealing/extension time. The libraries were sequenced on a MiSeq instrument (Illumina) using a MiSeq Reagent Kit v2 (500-cycles) at 13.5 pM, following the manufacturer’s protocols.

A final base call validation of the PacBio contigs was performed by mapping Illumina MiSeq reads that were trimmed to quality score threshold of 30 (Q30) or higher using the reference assembler within Geneious Prime 2020.2.1 (Biomatters, Ltd., Auckland, New Zealand). Single nucleotide polymorphisms (SNPs) between the PacBio assembly and the MiSeq reads were addressed using the annotate and predict/find SNPs module, with a minimum coverage parameter of 50 and a minimum variant frequency parameter of 0.8. The final corrected assemblies were annotated for protein-, rRNA-, and tRNA-coding genes using the NCBI Prokaryotic Genomes Annotation Pipeline (PGAP) [41]. The chromosomal and plasmid sequences for each strain were deposited into NCBI under BioProject PRJNA971184.

Determination of Penner capsular types, lipooligosaccharide locus classes, flagellar modification loci and C. jejuni integrated elements

Capsular polysaccharide (CPS)/Penner type, lipooligosaccharide (LOS) locus types, flagellar modification (FM) loci, and C. jejuni integrated elements (CJIEs) were identified by BLASTN analysis using Geneious Prime 2022.2. For CPS/Penner serotype determination, the in silico sequences of 36 PCR products for specific Penner types, as described previously [42,43] were used in a BLASTN screen. For LOS locus classes, the variable sequences for LOS classes A-S [44, 45] were used in BLASTN searches of the nine genomes. For each C. jejuni genome, the LOS locus was compared to the LOS class exhibiting the closest match by BLASTN using Mauve. The genomes were screened for CJIEs, including the four elements identified within strain RM1221 [46] and the element containing the type VI secretion system [47,48] using the BLASTN plugin in Geneious. For CJIEs and FM loci, the complete loci and individual genes were aligned using MAFFT and visualized using Mauve software to determine rearrangements, insertions, and deletions [49].

Molecular typing and comparison

The nine complete genomes were submitted to the PubMLST database (https://pubmlst.org/organisms/campylobacter-jejunicoli) for curation and analysis. The seven-locus MLST sequence types (STs) were assigned as described previously [50,51]. Novel alleles and profiles for two strains, SKBC3 and SKBC5, were assigned new allele numbers and sequence types.

Comparative genomic analyses

The black bear C. jejuni genomes were compared to other thermotolerant Campylobacter spp., including other C. jejuni genomes. To determine average nucleotide identity (ANI) values between Campylobacter spp., we used JSpecies (v. 1.2.1) with default parameters [52]. The relationship between the C. jejuni isolates from black bears and a variety of C. jejuni isolates (Table 2 and S1 Table) was explored using maximum likelihood phylogenies constructed from core genome alignments. Concatenated sequences of 1,107 core genes (present in 99% or more isolates) from 223 total C. jejuni genomes from the following sources: bear isolates (n = 9), cattle (n = 48), poultry (n = 90), environment (n = 6), guinea pig specific (n = 1), human clinical (n = 35), monkey (n = 2), sheep (n = 1), swine (n = 4), mouse (n = 1), turkey (n = 3), vole (n = 1), water (n = 15), wild bird (n = 1), C. jejuni subsp. doylei (n = 4) isolates, and unknown sources (n = 2). Additionally, 143 of these genomes were from sources isolated in the southeastern United States particular Georgia, North Carolina, and Virginia to match where the bear isolates were also collected during the study. Core gene sequences were aligned using the MAFFT module [53] within Roary (v3.12.0), using the following flags: -e (create a multiFASTA alignment of core genes using PRANK); -n (fast core gene alignment with MAFFT); -v (verbose output to STDOUT); and -i 90 (minimum percentage identity for BLASTP; 90%) [54]. Maximum likelihood phylogenies were constructed using Randomized Accelerated Maximum Likelihood (RAxML) with a GTRCAT model with 1,000 bootstraps [55] and visualized in Microreact [56]. Additional phylogenetic comparisons were visualized using iTOL [51,57].

Table 2. Genomic features of black bear C. jejuni isolates.

			Variable genomic loci					Integrated genomic elements
Strain	Genome size (Mbp)	Genome acc. #	Penner	LOS class	cdtABC	E-D	AMRc	CJIE1-Mu-like	CJIE2	CJIE3	CJIE4
SKBC1	1.724	CP125383	unknown	C	Y	N	KTB	Y	Y	N	N
SKBC3	1.765	CP125394	HS:4cx	B2	N	Y	–	Y2d	Y	N	N
SKBC5	1.784	CP125391	HS:16	A2	N	Y	–	Y2	Y	N	N
SKBC9	1.676	CP125390	HS:3,17	O	pseudoa	N	B	N	N	N	Y
SKBC25	1.708	CP125388	HS:1,44	A2	Y	N	B	Y	N	N	N
SKBC41	1.638	CP125386	HS:37	H	Y	N	TB	N	N	N	N
SKBC57	1.681	CP125396	HS:2	C	Y	N	B	N	N	N	Y
SKBC94	1.595	CP125395	HS:1,44	Q	p/2b	N	B	N	N	N	N
SKBC102	1.651	CP125387	HS:55	E	Y	N	TB	N	N	N	N

^a.Pseudo: cdtABC genes present with point mutations.

^b.p/2, cdtABC genes present with point mutations and a second cdtABC operon is present.

^c.K, presence of aph3; T, presence of tet(O); B, blaOXA; -,no AMR markers.

^d.Y2, two Mu-like prophages present

Gene families were identified using Roary software at 90% identity across the entire gene, which generated core (in ≥99% − 100% of all the genomes), soft-core (in ≥95% − 99% of all the genomes), shell (>15% − 95% of all the genomes), and cloud (>0% − 15%) genes, and visualized using the roary_plots.py script.

Phylogenetic dendrogram construction of Entner-Doudoroff loci

The Entner-Doudoroff (E-D) locus from C. jejuni strain SKBC3 (CP125394 from nucleotide 39,023–47,678) was used as bait to identify E-D loci in other Campylobacter genomes using BLASTN specific to the Campylobacter genus within the NCBI non-redundant (nr) database. An E-D locus was previously identified in a C. jejuni strain (GP012) isolated from guinea pigs; however, this locus is not in the NCBI nr database and was added separately. Loci were obtained from the following strains: CP000768 (strain 269.97), CP059375 (2010D-8469), CP125391 (SKBC5), CP125394 (SKBC3), JACRSF000000000 (GP012) and LR134359 (NCTC11951^T) from C. jejuni and C. jejuni subsp. doylei; CP017875 (ZV1224) and KT001110 (CHW475) from C. coli; CP031611 (NCTC 13823^T), CP063536 (USA52) and CP065357 (UF2019SK1) from C. hepaticus; CP059597 (LMG 32306^T) from C. molothri [58]; and CP020867 (LMG 24588^T) from C. cuniculorum were aligned using MAFFT, and a dendrogram was constructed within MEGA v.11 using the Maximum Likelihood method and Tamura-Nei model [59–61].

Results

C. jejuni isolates from American black bears are genotypically diverse

We sequenced the genomes of nine strains of Campylobacter jejuni isolated from 2014 to 2016 from American black bears living in the southeastern United States (Table 1). Eight sequence types (ST) were assigned to these nine C. jejuni genomes by PubMLST (Table 1). Six of the nine strains were members of previously identified C. jejuni clonal complexes. These included CC 21, CC45, CC206, CC179 and CC682, as identified in the strains SKBC57, SKBC41 and SKBC102, SKBC1, SKBC25 and SKBC94, respectively. C. jejuni within clonal complexes CC21, CC45 and CC206 have been described previously as host generalists [13]. Similarly, the non-human isolation sources of C. jejuni within CC-179 includes poultry, cattle and environmental water (Supp. Table 2). It should be noted that strain SKBC9 assigned to ST-10501 and no clonal complex, shares four alleles with ST-179 complex isolate SKBC25. These alleles were shared by 70 strains in PubMLST (S2 Table). The strain SKBC94 (ST-682 complex) would be considered a species-specific C. jejuni with 165/205 ST-682 complex strains from PubMLST isolated from European starlings (Sturnus vulgaris) (S3 Table).

Strains SKBC3 and SKBC5 were assigned novel sequence types, ST-7630 and ST-10620, respectively. These two strains share four alleles (gltA279, glyA607, tkt585 and uncA276), and these alleles are represented at most by only one other strain in the database in Feb 2025 (S4 Table). Moreover, most isolates with shared alleles were from strains that were isolated from environmental waters in North America, except for the aspA allele that were also shared with two human isolates, a dog isolate, and an isolate from an unknown source (S4 Table).

Genomic features of C. jejuni isolated from black bears

The genome sizes of these C. jejuni bear isolates were quite variable with estimated chromosome sizes ranging from 1.595–1.784 Mbp (Table 2) and containing between 1,491 and 1,689 predicted protein-encoding genes. The pangenome for these nine C. jejuni chromosomes was 3,356 genes with 1,256 core genes and 2,100 accessory genes. Three strains (SKBC1, SKBC5 and SKBC25) have plasmids that were not included in the pangenome analysis. Major plasmid features are described below.

The differences in chromosome sizes between the strains was mostly the result of the presence or absence of Campylobacter jejuni integrated elements (CJIEs). CJIEs, which include bacteriophages, plasmid-like elements and transposons have been identified in many C. jejuni strains [46–48,62–64]. The strains with the largest genomes, SKBC1, SKBC3 and SKBC5 (1.724, 1.765 and 1.784 Mb, respectively), possess multiple CJIEs. Bacteriophages similar to the Mu-like phage CJIE1, CJIE2, and CJIE4 were each detected by BLASTN in at least one bear isolate (Table 2).

Mu-like prophage (CJIE1-like) are harbored by four strains (SKBC1, SKBC3, SKBC5 and SKBC25) (Table 2, S1 Fig and S5 Table). These Mu-like prophages are integrated in different genomic locations in these strains (S5 Table) and were classified into two types based on the presence or absence of dns, a gene encoding an extracellular DNase, and major differences in sequence identity of genes encoding DNA transposition protein A, DNA transposition protein B, and Mu-like bacteriophage tail structural and assembly proteins (S1 Fig and S5 Table). Strains SKBC3 and SKBC5 each harbor both types of Mu-like bacteriophages, whereas SKBC1 and SKB25 only harbor one of the two Mu-like bacteriophages. Strains SKBC3, SKBC5 and SKBC25 possess Mu-like bacteriophages containing dns. These bacteriophages have >82% identity to CJIE1 from C. jejuni strain RM1221 (S1A Fig). The other three Mu-like bacteriophages, in strain SKBC1 (near npdA) and the second Mu-like bacteriophages in strains SKBC3 (near ctsT) and SKBC5 (tRNA-Arg gene near aroB and tgt) do not harbor dns, and are similar only for certain viral genes within RM1221 CJIE1, specifically, genes encoding the host-nuclease inhibitor (Gam), the viral baseplate and some viral tail-related proteins (S1B Fig and S5 Table).

The other two bacteriophage families, CJIE2-like and CJIE4-like, are present in more than one of the isolates from the black bears. These bacteriophage families have specific genomic integration sites, with CJIE2-like bacteriophages inserted at the tRNA-Arg gene adjacent to fusA and CJIE4-like bacteriophages inserted at the tRNA-Met gene adjacent to rodA. The CJIE2-like bacteriophages are present in strains SKBC1, SKBC3 and SKBC5. Comparison of these CJIE2-like bacteriophages with CJIE2 in strain RM1221 revealed diversity, identifying large insertions and deletions of genes (S2 Fig). The CJIE4-like bacteriophages are present in strains SKBC9 and SKBC57 and are > 98% identical to CJIE4 from strain RM1221 (S3 Fig).

Other insertion elements were also observed in the C. jejuni from black bears. Although there is an integrated element at the CJIE3 integration site (tRNA-Arg gene near aroB and tgt) in strain SKBC5, this element is quite distinct from the RM1221 CJIE3 with only similarities at the site-specific recombinase (integrase) and TraG-like genes that may have a role in integration site selection and conjugation, respectively (S4 Fig). A CJIE, containing a type VI secretion system (T6SS), possessed by some C. jejuni [47,48] is not present as an integrated element in any of the nine bear isolates; however, strain SKBC25 possesses a plasmid (pSKBC25) with T6SS genes.

Besides differences in genome sizes due to CJIEs, the large hypervariable genomic loci encoding major surface structures, including capsule biosynthesis, lipooligosaccharide biosynthesis and flagellar modification [42,44,65], are mostly distinct between the nine strains. As shown in Table 2, there are eight different capsular biosynthesis loci (Penner serotypes) and seven different LOS biosynthesis loci. Three LOS loci (A2, B2 and C) in five strains are potentially capable of synthesizing sialylated LOSs that have been associated with post infection neuropathies [66]. All nine FM regions are distinct and ranged in size from 27kb to 50kb (S5 Fig).

Other notable variable regions included transposable elements, plasmids and chromosomal loci defined as hypervariable by comparative genomic indexing [64,65]. Strains SKBC41 and SKBC102 each possesses an IS607-like transposable element, containing the tetracycline resistance gene tet(O), inserted near rarA. Only three strains possess plasmids, and none are shared between strains. Strain SKBC1 carries two plasmids: pSKBC1−1 harboring tet(O) and aph(3’)-IIIa and pSKBC1−2. Both pSKBC1−1 and pSKBC1−2 show >98% identity to multiple plasmids. Strain SKBC5 carries two plasmids: pSKBC5−1 harboring putative bacteriocins and pSKBC5−2. pSKBC5−1 is fairly novel, with only a few plasmids identified by BLASTN that aligned with portions of pSKBC5−1. pSKBC5−2 shows >98% identity to pTet-like plasmids but does not harbor the tet(O) gene. Finally, SKBC25 carries pSKBC25, which harbors a T6SS, as mentioned above.

Variation among the bear strains includes regions that are variably present on the chromosome including pantothenate (vitamin B5) biosynthesis locus (panBCD), ggt that encodes gamma-glutamyltranspeptidase, which plays a role in metabolism of glutathione and glutamine, and metAB (metAX) that encode enzymes involved in converting L-homoserine to L-methionine [64,67,68]. The panBCD and metAB loci were detected in the same two strains (SKBC1 and SKBC57), while ggt was detected in four strains (SKBC3, SKBC5, SKBC41 and SKBC102), none of which overlapped with possession of panBCD and metAB. Antibiotic resistance genes were present in some of the bear isolates (Table 2). Strains SKBC41 and SKBC102 possess tet(O) on the chromosome and strain SKBC1 harbors tet(O) and aph(3’)-IIIa on a plasmid, as mentioned above. It should be noted that each of these strains belong to MLST clonal complexes that are associated with domestic animals. Seven of the isolates, SKBC1, SKBC9, SKBC25, SKBC41, SKBC57 and SKBC102, possess a bla_OXA gene near modC. None of the strains have mutations in gyrA or the 23S RNA gene that confer resistances to fluoroquinolones and macrolides, respectively. Strains SKBC3 and SKBC5 have no identifiable antimicrobial resistance markers and no evidence of a remnant of a bla_OXA gene near modC.

ANI and phylogenetic analysis also support carriage of both common and distinct C. jejuni strains

The genomic relationship of the black bear isolates with several other thermotolerant Campylobacter was examined using average nucleotide identity (ANI) analysis. Within the collection here, all C. jejuni strains have ANI values >94% against all other C. jejuni strains, while the ANI values are < 85% against C. coli and C. hepaticus, and <76% with C. upsaliensis, C. vulpis and C. helveticus (Fig 1). Among the bear isolates, seven strains (SKBC1, SKBC9, SKBC25, SKBC41, SKBC57, SKBC94, SKBC102) have ANI values between 96.9% and 99.1% against all other C. jejuni subsp. jejuni strains. However, the bear isolates SKBC3 and SKBC5 have ANI values <97% and <96% against the other C. jejuni subsp. jejuni and C. jejuni subsp. doylei strains, respectively. The ANI values between the two subspecies, C. jejuni subsp. jejuni and C. jejuni subsp. doylei, were between 95 and 96% (Fig 1). Thus, SKBC3 and SKBC5 were slightly distinct from other C. jejuni subsp. jejuni but not as much as C. jejuni subsp. doylei strains, or the guinea pig-associated C. jejuni strain that had ANI values <95% to all other C. jejuni.

Fig 1. Average Nucleotide Identity (ANI) Values.

Twenty-six strains within the genus Campylobacter were numbered (#) in the rows and columns of the table. ANI values of a strain with itself would be 100% and are represented by ---. Values represent averages (in %) for each pair of strains based on BLAST. Values ≥ 90% are shaded in blue; values between 81 and 89% are shaded in white to pink; and values < 81% are shaded in red.

Pan-genome analyses were conducted using Roary analysis on the nine bear and with 71 global C. jejuni genomes, including C. jejuni subsp. jejuni, a strain isolated from a guinea pig, a bank vole, isolates from water, and C. jejuni subsp. doylei and an additional 143 C. jejuni genomes isolated from various sources (e.g., poultry, environment waters, cattle, etc.) in the same southeastern United States region as the bear isolates were obtained (S1 Table). All genomes were downloaded from NCBI and used to gain a better appreciation of the genomic differences of the strains at the individual gene level. Roary analysis was performed at a minimum percentage identity for BLASTP of 90% since the ANI values for several strains were below 95%. At the 90% cutoff, there were 1,107 core genes found in >99%, 236 soft core genes in =>95% to < 99%, 613 shell genes in =>15% <= and < 95% of strains and 3,328 cloud genes >0% to < 15% of strains.

Phylogenetic analysis compared the nine C. jejuni isolates from black bears with 71 global C. jejuni genomes, and 143 C. jejuni genomes from non-clinical strains isolated from the southeastern United States (Georgia, North Carolina and Virginia) (S1 Table). The analysis was performed using 1,107 C. jejuni core genes as determined by Roary analysis above. The resulting phylogeny demonstrated that these nine C. jejuni genomes from black bears were widely distributed among the other C. jejuni genomes (Fig 2). Bear isolates SKBC3, SKBC5, and SKBC94 were located on a major branch that includes C. jejuni subsp. doylei strains and the strains from a guinea pig, bank vole and certain environmental water samples (Fig 2). The other black bear C. jejuni isolates (SKBC1, SKBC9, SKBC25, SKBC41, SKBC57 and SKBC102) were in clusters comprised of other C. jejuni subsp. jejuni isolated from several animal sources (chicken, cattle, pigs, and turkey) and river water.

Fig 2. Phylogenetic and MLST genotypic distribution of C. jejuni isolates from black bears compared to a representative collection of C. jejuni.

A: The relationship between isolates from black bears and a variety of C. jejuni isolates (Table 1 and S1 Table) was explored using a maximum likelihood phylogeny. Concatenated sequences of 1,107 core genes (present in 90% or more isolates) from 214 C. jejuni and C. jejuni subsp. doylei isolates were used to compare the nine bear isolates (purple nodes and indicated with arrows). Isolates from other diverse sources are colored by source, including black bear (n = 9; blue), chicken and poultry farm (n = 96; yellow), cattle and sheep (n = 49; green), turkey (n = 3; orange), water and wild birds (n = 16; light blue), human feces (n = 35; black), human blood (n = 4; red), pigs (n = 4; pink) and other animal sources (n = 7; grey). Core gene sequences were aligned using the MAFFT module [53] within Roary (v3.12.0) [54]. Maximum likelihood phylogenies were constructed using Randomized Accelerated Maximum Likelihood (RAxML), with a GTRCAT model with 1,000 bootstraps [55] and C. jejuni only topology is visualized in microreact [56] (https://microreact.org/project/tYPSuzXxiXyoUxa3R8ytvK-parker-et-alcampylobacter-from-us-black-bears) Major lineages are labeled (ST-clonal complexes), with dotted lines indicating the threshold between CCs B: MLST distribution of black bear isolates (purple shading) in this comparison dataset (dark grey) and those collected from the Southeast USA only (light grey) among sequence types (ST).

Absence of the cdtABC and mfrABE loci within a phylogenetic cluster. Strains SKBC1, SKBC25, SKBC41, SKBC57 and SKBC102 had a cdtABC locus predicted to be functionally complete located adjacent to lctP. This is the same location that has been observed for cdtABC in most C. jejuni subsp. jejuni strains based on BLASTN analysis using lctP-cdtCBA as bait. Strains SKBC9 and SKBC94 had nonfunctional cdtABC loci adjacent to lctP, with both possessing a nonsense mutation in each cdt gene. However, in strain SKBC94, a second cdtABC locus was identified near the lpxB gene. This lpxB-linked cdtABC locus was observed in four of the other 214 C. jejuni genomes used in this study. The lpxB-linked cdtABC is the primary location of the cdtABC locus within C. coli, C. lari and other Campylobacters. BLASTN analysis of the lpxB-linked cdtABC locus from strain SKBC94 shows the closest alignments (~77%) to C. lari strains. Strains SKBC3 and SKBC5 had major deletions within the cdtABC locus adjacent to lctP, with only remnants of cdtC remaining, and neither strain had the lpxB-linked cdtABC locus. It should be noted that all C. jejuni sharing the phylogenetic branch with bear isolates SKBC3, SKBC5, and SKBC94 had nonfunctional lctP-linked cdtABC loci (S6 Fig: orange and blue branches) Among these strains, those within the blue branch (9 strains, including SKBC3 and SKBC5) have major deletions of the lctP-linked cdtABC locus.

The mfrABE operon (previously sdhABC) that encodes methylmenaquinol:fumarate reductase was also mutated in this same cluster of “host specific” strains that includes strains SKBC3 and SKBC5, three C. jejuni subsp. doylei strains, and isolates from a guinea pig, bank vole, and water (S6 Fig: blue branch). Methylmenaquinol:fumarate reductase affects resistance to H₂O₂ [69]. In SKBC3 and SKBC5, mfrA has a nonsense mutation and mfrB and mfrE are deleted. A variety of mfrABE deletions are found in other C. jejuni in this cluster. There are C. jejuni strains outside of this cluster containing mutated mfrABE; however, in these C. jejuni, the mfrABE mutations are nonsense mutations that could more easily revert to a functional methylmenaquinol:fumarate reductase system.

Acquisition of the Entner-Doudoroff pathway

The genomes of strains SKBC3 and SKBC5 possessed genes for the Entner-Doudoroff (E-D) pathway, which catalyzes the conversion of glucose-6-phosphate to pyruvate. These genes are situated within an rRNA locus between the 23S RNA gene and the 16S RNA gene. However, despite the E-D loci from SKBC3 and SKBC5 being 99.8% identical, the specific rRNA loci is distinct between SKBC3 and SKBC5. The E-D locus in SKBC3 is in the rRNA locus linked to pyrG-recJ while the E_D locus in SKBC is in the rRNA locus linked to cfrA-hcrA. Other Campylobacter possess E-D genes, and we performed alignment and maximum likelihood phylogenies to compare the E-D genes from the two bear-associated strains to E-D genes from C. jejuni isolated from a guinea pig, C. jejuni subsp. doylei and other Campylobacter species. Our tree demonstrates that the E-D loci from SKBC3 and SKBC5 are situated phylogenetically between E-D loci from C. jejuni and C. coli. (Fig 3).

Fig 3. Phylogenetic characterization of the Entner-Doudoroff pathway locus.

The phylogenetic tree of the E-D pathway loci obtained from 13 strains within the genus Campylobacter when compared to E-D pathway loci from C. jejuni bear isolates SKBC3 and SKBC5. The dendrogram was created using the Maximum Likelihood method and Tamura-Nei model, and is drawn to scale, with branch lengths measured in the number of substitutions per site.

Discussion

We isolated the foodborne pathogen C. jejuni from American black bears in the southeastern United States. Despite isolating only nine C. jejuni strains, these strains were from three different states and over a three-year period, suggesting C. jejuni colonizes black bears. The nine C. jejuni strains were genotypically variable with eight distinct STs and six CCs. This C. jejuni strain variability may suggest that bears acquire the bacteria from both anthropogenic sources (trash receptacles or agricultural sites) and from non-human environments (water or wild animals).

The multiplatform sequencing, i.e., long reads (PacBio) and short reads (Illumina), provided nine closed and accurate genomes. Genomic comparisons of the isolates revealed substantial variability in chromosome sizes and gene content. The genome sizes ranged from 1.595 to 1.784 Mb and the variation was largely due to the presence of CJIEs as described previously [46–48,62–64]. For instance, strains SKBC1, SKBC3, and SKBC5 had the largest genomes due to multiple CJIEs (Table 2). The presence of these elements not only affects genome size but may also shape the functional potential of the strains by altering gene expression and natural competence [70–72]. Differences in hypervariable loci encoding biosynthesis of surface glycan structures further emphasizes the genomic variability of these strains, with eight different CPS loci, seven different LOS loci and nine different FM loci identified in the nine strains. Campylobacter surface glycans, particularly CPS and LOS, are quite immunogenic and variable [73]. The genetic variation at these glycan loci suggests that the bear isolates exhibit a variety of glycan surface structures. These glycans often play significant roles in virulence, with the best understood of these being the role of molecular mimicry of human gangliosides by particular sialylated LOS structures in the development of post-infection neuropathies [74]. In fact, five of the C. jejuni strains from bears possessed LOS loci that could potentially synthesize sialylated LOS structures. Other variable loci include panBCD, metAB (metAX), and ggt. These genes have been identified as playing a role in colonization of domestic cattle or poultry under certain conditions [67,75–77]; however, their variable presence in cattle and poultry isolates, as described previously [64,65], and here, within bear isolates, suggests that they are not essential for the colonization process in these hosts.

C. jejuni genomes are prone to recombination, and signals of host adaptation can be identified providing information of the strain’s recent ancestry [21,78,79]. Based on their MLST CC, four C. jejuni strains isolated from the bears (SKBC1, SKBC41, SKBC57 and SKBC102) are predicted to be host generalists [21], according to their membership in the clonal complexes CC-21, CC-45, and CC-206. Host generalist C. jejuni are associated with broad host ranges, including domestic poultry and domestic cattle, and they are commonly recovered from human clinical samples [12,80]. Additionally, strains SKBC9 and SKBC25 are part of a group of over 70 strains that share four MLST alleles (S2 Table). The strains in this group are also likely host generalist, with nonclinical strains isolated from both domestic poultry and cattle, and environmental waters. In fact, these generalist strains from bears clustered with poultry and cattle strains within our phylogenetic analysis (Fig 2 and S6 Fig). Furthermore, three of these generalist strains possess antimicrobial resistance (AMR) genes. Surveys of avian C. jejuni strains suggested that C. jejuni from birds more closely associated with anthropogenic settings have a higher prevalence of AMR genes [81]. Given the increased human activity and reduction of wild habitat where black bears live, there has been an increase in interactions between humans and bears [34,35]. Black bears generally stay away from humans; however, in agricultural areas where food is common, bears often feed on crops and can encounter domestic animals and their waste. Considering this, it is likely that black bears were colonized by generalist C. jejuni while foraging around agricultural areas or human habitations.

Three C. jejuni isolates from the black bear exhibited evidence that they were from non-agricultural environments. SKBC94 is in the ST-682 complex that is comprised mostly of strains from European starlings (Sturnus vulgaris). SKBC3 and SKBC5 currently have novel STs but share specific alleles with C. jejuni isolated from environmental waters, and thus unknown animal sources. Infection with C. jejuni from wild birds would not be unusual considering the observations of black bears eating wild bird nestlings [82]. Black bears are also scavengers and could have obtained C. jejuni from a variety of dead animals or from environmental waters. Alternatively, strains SKBC3 and SKBC5 may represent bear-specific C. jejuni; however, there is not enough evidence to determine this currently.

All three of these strains (SKBC94 (ST-682), SKBC3 (ST-7630) and SKBC5 (ST-10620)) cluster phylogenetically with C. jejuni isolated from wildlife and environmental waters, and C. jejuni subsp. doylei strains (Fig 2 and S6 Fig). Beyond the phylogenetic clustering, these three bear isolates and the strains within the cluster (Fig 2 and S6 Fig) also had mutations within the lctP-linked cdtABC locus that would eliminate function. For SKBC94, each lctP-linked cdt gene had a nonsense mutation. On the other hand, SKBC3 and SKBC5 possessed major deletions of the lctP-linked cdtABC locus previously reported for C. jejuni subsp. doylei strains [83] and among C. jejuni subsp. jejuni strains from guinea pigs and wild birds [11,84]. Strains SKBC94 also possessed a second cdtABC operon linked to lpxB that was shared by multiple CC ST-682 strains. The absence of lctP-linked cdtABC and/or the presence of lpxB-linked cdtABC suggests that cytolethal distending toxin may play a major role in host-specific colonization as noted by Guirado [84]. Additionally, the absence of mfrABE in SKBC3, SKBC5 and other strains in the cluster suggests methylmenaquinol:fumarate reductase may play a negative role in host-specific colonization. Methylmenaquinol:fumarate reductase may contribute to the generation of oxidative stress in C. jejuni by creating reactive oxygen species. C. jejuni mutants unable to produce methylmenaquinol:fumarate reductase were demonstrated to be more resistant to H₂O₂ [69]. It is not exactly clear how either cytolethal distending toxin or methylmenaquinol:fumarate reductase affects host specificity. More studies will be required to elucidate their roles.

Finally, we identified the presence of the E-D locus in strains SKBC3 and SKBC5. The E-D locus encodes enzymes involved in the Entner-Doudoroff pathway that catalyze the conversion of glucose to pyruvate and has been identified in C. jejuni, C. coli and other Campylobacter species [11,85]. Vegge et al [85] demonstrated the fitness advantages within C. coli possessing the E-D locus including stationary-phase survival and biofilm production. Despite the physiological advantages, the E-D locus was in less than 2% of C. jejuni/C. coli strains. Moreover, we found that the E-D loci from the bear-associated strains are distinct from E-D from other C. jejuni. The diversity of E-D loci was previously observed by Vegge et al for 113 Campylobacter isolates and they suggested this may be the result of different times of E-D introduction into each strain followed by evolution [85]. Alignment and maximum likelihood trees demonstrate that the E-D loci in SKBC3 and SKBC5 are situated phylogenetically between E-D loci from C. jejuni (e.g., E-D loci in a guinea pig-associated C. jejuni strain and loci from three of C. jejuni subsp. doylei strains) and E-D loci from C. coli.

Conclusions

This study demonstrates for the first time that C. jejuni can be isolated from American black bears. This is not the only human pathogen that has been isolated from bears. Listeria monocytogenes was also identified in feces, rectal and nasal swabs of black bears [36]. Due to the bears attraction to food sources near agricultural settings, we suggest that black bears could serve as a natural vehicle and perhaps a reservoir for a wide variety of C. jejuni lineages from both domestic animals and wildlife. The limited number of C. jejuni strains and the restricted region of collection, i.e., the Southeastern United States, does not permit assessments of the overall prevalence of C. jejuni carriage in black bears or allow rigorous genome-wide association studies. However, we distinguished both domestic animal-associated genotypes and wild bird genotypes among the C. jejuni strains isolated from black bears. Moreover, we identified two strains with novel genotypes that were only associated with black bears. Thus, sequencing and phenotypic analysis of additional C. jejuni isolates from black bears will be necessary to determine if bears are routinely colonized by domestic animal-associated genotypes of C. jejuni and if the novel bear genotypes represent bear specialists or specialists for another host.

Supporting information

S1 Table. Strains used for comparison.

(PDF)

S2 Table. Strains in PubMLST that share four alleles with either SKBC9 or SKBC25.

(PDF)

S3 Table. Strains in PubMLST with CC ST-682.

(PDF)

S4 Table. Strains in PubMLST that share alleles with either SKBC3 or SKBC5.

(PDF)

S5 Table. Similarity to subset of RM1221 CJIE1 (Mu-like) genes.

(PDF)

S1 Fig. Genome alignment of CJIE1 (Mu-like bacteriophages).

(PDF)

S2 Fig. Genome alignment of CJIE2.

(PDF)

S3 Fig. Comparisons of CJIE4.

(PDF)

S4 Fig. Comparison of element at CJIE3 insertion site.

(PDF)

S5 Fig. Comparisons of flagellar modification loci.

(PDF)

S6 Fig. Phylogeny of Campylobacter jejuni strains isolated from various sources in the southeastern United States.

(PDF)

Abbreviations

AMR

antimicrobial resistance

CC

clonal complex

MLST

multi-locus sequence typing

ST

sequence type

WGS

whole genome sequencing

Data Availability

All the data is available in NCBI under BioProject Accession PRJNA971184.

Funding Statement

This research was completely supported through the following sources: USDA-ARS CRIS project 2030-42000-055-00D to Craig Parker, USDA National Institute of Food and Agriculture, award 2018-67017-27927 to Sophia Kathariou, and Technology and Research Initiative Fund (TRIF) provided to Kerry Cooper by the University of Arizona.