Occurrence of Putative Virulence Genes in Arcobacter Species Isolated from Humans and Animals

Abstract

Interest in arcobacters in veterinary and human public health has increased since the first report of the isolation of arcobacters from food of animal origin. Since then, studies worldwide have reported the occurrence of arcobacters on food and in food production animals and have highlighted possible transmission, especially of Arcobacter butzleri, to the human population. In humans, arcobacters are associated with enteritis and septicemia. To assess their clinical relevance for humans and animals, evaluation of potential virulence factors is required. However, up to now, little has been known about the mechanisms of pathogenicity. Because of their close phylogenetic affiliation to the food-borne pathogen Campylobacter and their similar clinical manifestations, the presence of nine putative Campylobacter virulence genes (cadF, ciaB, cj1349, hecA, hecB, irgA, mviN, pldA, and tlyA) previously identified in the recent Arcobacter butzleri ATCC 49616 genome sequence was determined in a large set of human and animal Arcobacter butzleri, Arcobacter cryaerophilus, and Arcobacter skirrowii strains after the development of rapid and accurate PCR assays and confirmed by sequencing and dot blot hybridization.

INTRODUCTION

Arcobacters are increasingly being isolated from a wide range of food products all over the world. These Gram-negative bacteria have been classified into the family Campylobacteraceae (35), although a recent annotation of the Arcobacter butzleri genome suggests a closer phylogenetic relation to Sulfurimonas denitrificans and Wolinella succinogenes, both members of the Helicobacteraceae, as well as to the deep-sea vent Epsilonproteobacteria members Sulfurovum and Nitratiruptor (26). At present, 13 Arcobacter species have been characterized, of which 6 were isolated from mammals. In humans, A. butzleri is predominantly associated with enteritis and septicemia (24, 30, 46), though Arcobacter cryaerophilus and Arcobacter skirrowii have also been isolated from diarrheal stool specimens (21, 33, 44). The other three species, Arcobacter cibarius (15), Arcobacter thereius (14), and Arcobacter trophiarum (4) are present in farm animals and on food of animal origin but have not yet been isolated from human specimens.

Contaminated drinking water is identified as a major source of human Arcobacter infection in developing regions (1), whereas in industrialized countries, infections are assumed to be food-borne. Close contact with pets and person-to-person transmission are the other potential risk factors (9, 36). Arcobacters seem to be commonly present on food of animal origin, with the highest prevalence reported for poultry followed by pork and beef (38, 42). The origin of the contamination on poultry products is still debated (39), but for pork and beef, feces transmitted during the slaughter process is regarded as the initial source of contamination (40, 44).

The A. butzleri ATCC 49616 genome revealed that this strain has putative virulence determinants such as genes cadF and cj1349, coding for fibronectin binding proteins; the invasin gene ciaB; the virulence factor mviN; the phospholipase gene pldA; the hemolysin gene tlyA; the irgA gene coding for an iron-regulated outer membrane protein; the gene hecA, a member of the filamentous hemagglutinin (FHA) family; and the gene hecB, coding for a hemolysin activation protein (26).

At present, there are no tools for the rapid and accurate detection of those genes, and the distribution of those genes in A. butzleri, A. cryaerophilus, and A. skirrowii strains associated with human infection is not known. Therefore, the aims of the present study were to develop PCR-based assays to detect those putative virulence genes and to assess their presence in a set of 319 Arcobacter strains.

MATERIALS AND METHODS

Arcobacter reference strains.

For the development and validation of the primer sets used in the PCR assays, 30 strains of the three human- and animal-associated Arcobacter species were obtained from the Belgian Coordinated Collections of Microorganisms (BCCM), Laboratory for Microbiology (LMG), Ghent University (Ghent, Belgium) (Table 1). In this set, the A. butzleri genome strain ATCC 49616 was included as a positive control (25). The strains were grown on blood agar plates (Mueller-Hinton [CM 337; Oxoid, Basingstoke, United Kingdom] and 50 ml/liter defibrinated horse blood [E&O Laboratories Ltd., Bonnybridge, Scotland]) and incubated for 48 h at 28°C under microaerobic conditions by evacuating 80% of the normal atmosphere and introducing a gas mixture of 8% CO₂, 8% H₂, and 84% N₂ into the jar. This incubation atmosphere was used for all further Arcobacter cultivations.

Table 1.

Presence of virulence determinants in Arcobacter reference strains used in the study

Arcobacter species	Accession no.a	Biological origin	Geographic origin	Virulence gene
				cadF	ciaB	cj1349	irgA	hecA	hecB	mviN	pldA	tlyA
A. butzleri	LMG 6620	Blood, 65-yr-old women	UK	+b	+b	+b	+	−	+b	+b	+b	+b
	LMG 9869	Deadborn piglet	UK	+	+	+	−	−	−c	+	+	+
	LMG 9939	Bovine fetus	Canada	+	+	+	+	−	−c	+	+	+
	LMG 10220	Feces, pig with diarrhea	Canada	+	+	+	−c	−	−	+	+	+
	LMG 10223	Liver, cow	Canada	+	+	+	−	−	−	+	+	+
	LMG 10240	Feces, horse with diarrhea	Canada	+	+	+	−	−	+	+	+	+
	LMG 10828T = ATCC 49616	Feces, human with diarrhea	USA	+b	+b	+b	+b	+b	+b	+b	+b	+b
	LMG 10900	Feces, child with diarrhea	Italy	+	+	+	+	−	+	+	+	+
	LMG 11119	Feces, child with diarrhea	Italy	+	+	+	+	−	+	+	+	+
	LMG 11632	Infected wound of hand, child	Canada	+	+	+	+b	+	−c	+	+	+
A. cryaerophilus 1	LMG 9065	Placenta, ovine fetus	UK	−c	+	+b	−	−	−c	+b	+	+
	LMG 9863	Placenta, ovine fetus	UK	−c	+	+	−	−	−	+	−	−
	LMG 9865	Eye, porcine fetus	UK	+b	+	+	−	−	−	+	+b	+b
	LMG 9904T	Brain, bovine fetus	UK	+	+	+	−	+b	−	−	+	+
	LMG 10210	Bovine fetus	Canada	+	+b	+	−	−	−c	+	+	+
A. cryaerophilus 2	LMG 9867	Spleen, equine fetus	UK	+	+	−c	−	−	+	+	+	+
	LMG 9947	Bovine fetus	Canada	+b	+	+	−	−	+b	+	+	+
	LMG 10209	Bovine fetus	Canada	+	+	+	−	−	−	+	+	+
	LMG 10212	Bovine fetus	Canada	+	+	+	−	−	−	+	+	−
	LMG 10216	Placenta, porcine fetus	Canada	+	+	+	−	−	−c	+	+	+
	LMG 10225	Colon, pig	Canada	+	+	+	−	−	−c	+	+	+
	LMG 10228	Organs, porcine fetus	Canada	+	+	+b	−	−	−	−	+	+
	LMG 10242	Organs, porcine fetus	Canada	+	+b	+	−	−	+	+b	+b	+b
	LMG 10244	Organs, porcine fetus	Canada	+	+	+	−	−	−c	+	+	+
A. skirrowii	LMG 6621T	Feces, lamb	UK	+b	+b	+	−	+b	−c	+	+	+b
	LMG 9801	Preputial fluid, bull	Belgium	+	+b	+b	−	−	+b	+	+	+
	LMG 9911	Thoracic fluid, porcine fetus	UK	+	+	−c	−	+b	−	+	+	+
	LMG 9912	Placenta, bovine fetus	UK	+	+	+	−	−	−	+	+	+
	LMG 10234	Organs, porcine fetus	Canada	+	+	+	−	−	+	+	+	+
	LMG 14985	Preputial washing of bull	USA	+	+	−c	−	−	−	+b	+b	+

^aLMG, Belgian Coordinated Collections of Microorganisms, Laboratory for Microbiology, (BCCM/LMG), Ghent University, Ghent, Belgium.

^bBidirectionally sequenced samples.

^cNegative PCR result/positive dot blot result.

Primer design and optimization of the PCR assays.

Nine primer pairs were designed using the free software program FastPCR (20). Primers were designed based on the A. butzleri ATCC 49616 genome sequence (deposited in GenBank [National Institutes of Health] as RM4018 [accession number NC 009850]) with gene accession numbers AB0483 (cadF), AB1555 (ciaB), AB0070 (cj1349), AB0729 (irgA), AB0941 (hecA), AB0940 (hecB), AB0876 (mviN), AB0859 (pldA0), and AB1846 (tlyA). To avoid amplification of nonspecific PCR products, in silico analyses by blasting the selected primers on the complete Arcobacter genome were performed. The primer sequences, expected amplicon lengths, and their positions compared to the genome numbering of A. butzleri ATCC 49616 are listed in Table 2. All PCRs were performed in a reaction mixture (50 μl final volume) containing 2 μl of DNA extract (50 ng/μl), 1.5 U of Taq DNA polymerase (catalog number [cat. no.] 10342-020; Invitrogen, Carlsbad, CA), 0.2 mM each deoxyribonucleotide triphosphate, 50 pmol of each primer set, 5 μl of 10× PCR buffer (cat. no. 10297-018; Invitrogen), and 1.3 mM MgCl₂ for the primer sets for ciaB, cj1349, hecA, irgA, and mviN. For the primer sets for cadF, hecB, pldA, and tlyA, 5 μl 10× PCR buffer and 1.5 mM MgCl₂ were used. PCR involved 32 cycles of denaturation (94°C, 45 s); primer annealing at 56°C for 45 s for the primer sets for ciaB, cj1349, hecA, irgA, and mviN and at 55°C for 45 s for the primer sets for cadF, hecB, pldA, and tlyA; and a chain extension (72°C, 45 s) followed by final elongation for 3 min at 72°C. Prior to cycling, samples were heated at 94°C for 3 min. For all experiments, a Veriti 96-well thermal cycler (Applied Biosystems, Foster City, CA) was used. Amplified products were detected by electrophoresis in 1.5% agarose in 0.5× Tris-borate-EDTA (TBE) buffer at 100 V for 60 min. Gels were stained with ethidium bromide (1 μg/ml). A UV transilluminator (Digi-Doc system; Bio-Rad, California) with an analyst computer program (Quantity One 4.1 software; Bio-Rad) was used for visualization.

Table 2.

Sequences and positions of the primers designed for the detection of the Arcobacter putative virulence genes

Target gene	Primer pair	Sequence of primers (5′ to 3′)	Amplicon size (bp)	Nucleotide positiona
cadF	cadF-F	TTACTCCTACACCGTAGT	283	489366–489383
	cadF-R	AAACTATGCTAACGCTGGTT		489649–489630
ciaB	ciaB-F	TGGGCAGATGTGGATAGAGCTTGGA	284	1560931–1560955
	ciaB-R	TAGTGCTGGTCGTCCCACATAAAG		1561215–1561192
cj1349	cj1349-F	CCAGAAATCACTGGCTTTTGAG	659	79205–79226
	cj1349-R	GGGCATAAGTTAGATGAGGTTCC		79864–79842
irgA	irgA-F	TGCAGAGGATACTTGGAGCGTAACT	437	737079–737103
	irgA-R	GTATAACCCCATTGATGAGGAGCA		737516–737493
hecA	hecA-F	GTGGAAGTACAACGATAGCAGGCTC	537	944614–944638
	hecA-R	GTCTGTTTTAGTTGCTCTGCACTC		945151–945128
hecB	hecB-F	CTAAACTCTACAAATCGTGC	528	937588–937607
	hecB-R	CTTTTGAGTGTTGACCTC		938116–938099
mviN	mviN-F	TGCACTTGTTGCAAAACGGTG	294	878897–878917
	mviN-R	TGCTGATGGAGCTTTTACGCAAGC		879191–879168
pldA	pldA-F	TTGACGAGACAATAAGTGCAGC	293	865140–865161
	pldA-R	CGTCTTTATCTTTGCTTTCAGGGA		865433–865410
tlyA	tlyA-F	CAAAGTCGAAACAAAGCGACTG	230	1842810–1842831
	tlyA-R	TCCACCAGTGCTACTTCCTATA		1843040–1843019

^aThe nucleotide positions of primers were compared to the genome numbering of A. butzleri ATCC 49616^T (26).

DNA sequencing.

To verify the specificity of the primers in Arcobacter strains, at least one amplicon of each gene per Arcobacter species was selected (Table 1), and the sequences were determined by DNA sequencing. The amplicons were purified by QIAquick spin column (cat. no. 28104; Qiagen, Hilden, Germany) and were cloned in pGEM-T vector according to the instructions of the manufacturer (Promega, Madison, WI). Bidirectional sequencing was performed with an ABI 3730xl sequencer (Applied Biosystems, Foster City, CA) with the Prism BigDye Terminator cycle sequencing kit (Applied Biosystems) and the primers T7 and SP6. After alignment of the resulting sequences using BioEdit 7.0.9 software (7), all amplicons were blasted using the National Center for Biotechnology Information network.

Dot blot hybridization.

From all 30 reference strains, 250 ng of DNA was dot blotted on Hybond-N+ nylon membrane filters (Amersham, Amersham, United Kingdom). Nylon filters were air dried at room temperature and baked at 80°C for 2 h to immobilize the DNA. AB0483 (cadF), AB0729 (irgA), AB0941 (hecA), AB0942 (hecB), AB0876 (mviN), AB0859 (pldA), AB1555 (ciaB), AB0070 (cj1349), and AB1846 (tlyA) amplicons were used as a DNA probe after conjugation with horseradish peroxidase (HRP), according to the instructions of the manufacturer (ECL direct nucleic acid detection system; Amersham). The heat-denatured DNA in 10 μl of distilled water (100 ng of PCR product) was mixed with a DNA-labeling reagent mixture consisting of 10 μl of a horseradish peroxidase-p-benzoquinone-polyethyleneimine complex and 10 μl of glutaraldehyde solution. This mixture was incubated at 37°C for 10 min. The filter with the target DNA was incubated with 5 ml of hybridization buffer. After incubation for 15 min at 42°C, it was further incubated at 42°C overnight with 100 ng of the conjugated probe. The filter was washed three times for 10 min at 42°C with 42 ml of washing solution consisting of 5× saline-sodium citrate buffer (SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate]) and was rinsed twice in 48 ml of 2× SSC at room temperature for 5 min. The following procedures were performed in a dark room. A mixture containing equal volumes of detection reagents 1 and 2 was added to the filter and incubated for 1 min at room temperature. To detect chemiluminescence, the filter was wrapped with Saran Wrap and exposed to X-ray film (Hyperfilm-ECL; Amersham), for 10 to 30 min at room temperature. A. butzleri ATCC 49616 was used as a positive and Escherichia coli (LMG 8223) as a negative control.

Human and animal Arcobacter strains.

To assess the presence of the nine genes in a large collection of Arcobacter isolates, 75 A. butzleri isolates and 1 A. cryaerophilus isolate from human patients were obtained from the National Reference Center for Enteric Campylobacter, Department of Microbiology, Saint-Peter University Hospital, Brussels, Belgium, where since 1995 all stool samples from patients have been tested for the presence of Campylobacteraceae using the nonselective membrane filtration technique (37). In addition, 3 A. butzleri and 14 A. cryaerophilus isolates from stool samples of adult and infant patients of the Saint-Lucas Hospital, Ghent, Belgium, were obtained, using the Arcobacter-selective isolation method of Houf and Stephan (16). The single European A. skirrowii strain available, isolated from a patient in Brussels, Belgium, was also included (45), as were seven A. cryaerophilus strains isolated from healthy persons during a survey in 2005 (16).

Since 2001, arcobacters have routinely been isolated from food and feces of food-producing animals in the Department of Veterinary Public Health, Ghent University, Belgium, using the Arcobacter-selective isolation methods of Houf et al. (13) and van Driessche et al. (41). A selection of this isolates was included in this study.

Identification and typing.

All isolates were subcultured onto blood agar plates and incubated for 48 h at 28°C under microaerobic conditions. Cell suspensions were prepared in 10 ml of sterile water with an optical density of about 0.074 ± 0.002 (measured at 660 nm) which corresponded to a concentration of approximately 10⁷ CFU/ml. Template DNA was extracted from a 0.5-ml cell suspension of each isolate in phosphate-buffered saline (PBS) buffer (Sigma-Aldrich, Irvine, Ayrshire, United Kingdom). Before extraction, all cell suspensions were centrifuged for 5 min at 17,900 × g (model 5417-R centrifuge; Eppendorf, Hamburg, Germany) to pellet the cells, and the supernatants were discarded. The pellets were resuspended in 100 μl Tris-EDTA buffer, and genomic DNA was extracted by the guanidiniumthiocyanate method described by Pitcher et al. (31). Five microliters of each DNA preparation was size separated by electrophoresis in 1% gels to evaluate the integrity of the DNA extracted. The concentration of each DNA template was determined spectrophotometrically at A₂₆₀ and adjusted to 50 ng μl⁻¹. The DNA templates were stored at −20°C.

For identification at the species level, an Arcobacter species-specific multiplex PCR assay developed by Douidah et al. (5) was performed in a reaction mixture at a 50-μl final volume composed of water (W4502; Sigma-Aldrich), 5 μl 10× PCR buffer (Invitrogen), 1.5 U Taq polymerase (Invitrogen), and a deoxynucleotide triphosphate mixture at a final concentration of 0.2 mM each (Invitrogen), 1.5 mmol of MgCl₂, and 50 pmol of each primer for ButR, SkiR, TherR, CibR, ArcoF, GyrasF, and GyrasR. The PCR assay involved 30 cycles of denaturation (94°C, 45 s), primer annealing (58°C, 45 s), and chain extension (72°C, 2 min). Arcobacter isolates that did not react in the multiplex PCR were subjected to partial 16S rRNA gene sequencing.

To avoid the inclusion of identical strains, all isolates were further characterized below species level by a modified enterobacterial repetitive intergenic consensus (ERIC)-PCR (10). Therefore, 1 μl of DNA extract was added to a 49-μl PCR volume. The ERIC motifs 1R (5′-ATGTAAGCTCCTGGGGATTCAC-3′) and 2 (5′-AAGTAAGTGACTGGGGTGAGCG-3′) were used at concentrations of 25 pmol each. The PCR products were size separated by electrophoresis in 2% agarose gels in TBE buffer at 100 V for 2 h. The banding patterns used to determine the genotypes comprised DNA fragments between 100 and 2,072 bp. Computer-based normalization and interpolation of the DNA profiles and numerical analysis using the Pearson product moment correlation coefficient, with 1% position tolerance, were performed using the GelCompar 4.2 software package (Applied Maths, Sint-Martens-Latem, Belgium). Dendrograms were constructed using the unweighted-pair group average linkage analysis (UPGMA) method. For convenience, the correlation level was expressed as a percentage of similarity. As shown in previous studies, DNA patterns that differed in one or more DNA fragments were regarded as different genotypes (10, 43).

A selection of 102 A. butzleri, 73 A. cryaerophilus, and 37 A. skirrowii strains from broiler carcasses and meat, cattle, pig, horse, and sheep feces, as well as 2 A. butzleri and 4 A. cryaerophilus strains isolated from dogs, were included in the present study (9, 11, 12, 41–43) (Table 3). All strains were stored on beads (Microbank; Pro-Lab Diagnostics, Richmond Hill, ON, Canada) at −80°C until further examination.

Table 3.

Presence of virulence-associated genes in Arcobacter isolates

Species	Biological origin	No. of strains examined	No. of strains generating specific gene amplicona
			cadF	ciaB	cjl349	irgA	hecA	hecB	mviN	pldA	tlyA
A. butzleri	Human	78	78	78	78	27	16	53	78	78	78
	Chicken	36	36	36	36	9	8	16	36	36	36
	Pig	33	33	33	33	4	6	14	33	33	33
	Cattle	29	29	29	29	14	17	25	29	29	29
	Sheep	2	2	2	2	0	0	2	2	2	2
	Horse	2	2	2	2	0	0	1	2	2	2
	Dog	2	2	2	2	0	0	0	2	2	2
A. cryaerophilus	Human	22	14	20	16	1	2	8	22	13	12
	Chicken	34	5	33	11	1	1	0	32	1	5
	Pig	23	6	21	15	1	0	1	21	6	14
	Cattle	9	4	8	3	0	1	0	7	5	4
	Sheep	6	0	5	3	0	0	0	5	0	0
	Horse	1	1	1	0	0	0	0	1	1	0
	Dog	4	4	4	3	0	0	1	4	0	0
A. skirrowii	Human	1	1	1	1	0	0	0	1	1	1
	Chicken	−	−	−	−	−	−	−	−	−	−
	Pig	13	2	12	2	0	1	3	1	2	5
	Cattle	23	4	23	2	0	0	5	8	3	1
	Sheep	1	1	1	1	0	0	0	0	1	1
	Horse	−	−	−	−	−	−	−	−	−	−
	Dog	−	−	−	−	−	−	−	−	−	−

^a−, no strains available.

Statistical analysis.

For statistical analysis, a chi-square test was performed to analyze the association of the nine putative genes in human and animal Arcobacter strains with their biological origin. A P value of <0.05 was considered as statistically significant.

RESULTS

Validation of the PCR assays.

Specific primer sets were designed based on the genomic information available for A. butzleri ATCC 49616 (Table 2). Subsequently, 30 reference strains, including 10 A. butzleri, 14 A. cryaerophilus and 6 A. skirrowii strains, were included to validate the PCR conditions (Table 1). The specificity of the PCR assays was assessed by sequencing PCR amplicons (Table 1). For all species, only the expected gene fragments were obtained. Within A. butzleri, a sequence similarity from 93 to 100% with the targeted gene fragment in A. butzleri ATCC 49616 was obtained, while the similarity of the amplicons of A. skirrowii and A. cryaerophilus ranged from 75 and 88 to 100%, respectively (data not shown).

The occurrence of the putative virulence genes in the 30 reference strains is shown in Table 1. In general, the genes cadF, ciaB, cj1349, mviN, pldA, and tlyA were detected in more than 85% of the Arcobacter strains. In contrast, the genes hecA and hecB were less frequently detected, and irgA was detected only in A. butzleri (Table 1).

The specificity and sensitivity of the PCR assays were further evaluated by cell suspension dot blot hybridization (Fig. 1). The PCR results were 100% in agreement with the dot blot results for the genes ciaB, hecA, mviN, pldA, and tlyA (Table 1; Fig. 1). No false-positive PCR results were detected for any of the nine genes in comparison to the dot blot results. False-negative PCR results were observed for hecB (n = 3) and irgA (n = 1) in A. butzleri and for cadF (n = 2), cj1349 (n = 3), and hecB (n = 6) in A. cryaerophilus and A. skirrowii strains (Table 1). Although hecA has the most variable gene sequence, the dot blot results were 100% in agreement with the PCR results. In contrast, nine strains were negative by PCR compared to blotting for hecB, although a sequence fragment homology of more than 98% was observed.

Fig 1.

Cell suspension dot blot hybridization of Arcobacter reference strains for the genes pldA, ciaB, hecA, and hecB. (A1) Positive control, A. butzleri (LMG 10828^T); (A2) negative control, E. coli (LMG 2092^T); (A3) A. skirrowii (LMG 6621^T); (A4) A. butzleri (LMG 9869); (A5) A. butzleri (LMG 9939); (A6) A. butzleri (LMG 10220); (A7) A. butzleri (LMG 10223); (A8) A. butzleri (LMG 10240); (B1) A. butzleri (LMG 10828); (B2) A. butzleri (LMG 10900); (B3) A. butzleri (LMG 11119); (B4) A. butzleri (LMG 11632); (B5) A. cryaerophilus (LMG 9065); (B6) A. cryaerophilus (LMG 9863); (B7) A. cryaerophilus (LMG 9865); (B8) A. cryaerophilus (LMG 9904^T); (C1) A. cryaerophilus (LMG 10210); (C2) A. cryaerophilus (LMG 9867); (C3) A. cryaerophilus (LMG 9947); (C4) A. cryaerophilus (LMG 10209); (C5) A. cryaerophilus (LMG 10212); (C6) A. cryaerophilus (LMG 10216); (C7) A. cryaerophilus (LMG 10225); (C8) A. cryaerophilus (LMG 10228); (D1) A. cryaerophilus (LMG 10242); (D2) A. cryaerophilus (LMG 10244); (D3) A. skirrowii (LMG 6621^T); (D4) A. skirrowii (LMG 9801), D5) A. skirrowii (LMG 9911), D6) A. skirrowii (LMG 9912); (D7) A. skirrowii (LMG 10243); (D8) A. skirrowii (LMG 14985).

Arcobacters isolated from humans and animals.

The occurrence of the nine putative virulence genes in this set of 182 A. butzleri, 99 A. cryaerophilus, and 38 A. skirrowii strains from various origins is shown in Table 3. Twenty-six (14.8%) A. butzleri strains carried all nine genes, in contrast to A. cryaerophilus and A. skirrowii, where none of the strains possessed all nine genes. The genes cadF, ciaB, cj1349, mviN, pldA, and tlyA were detected in all A. butzleri strains (Table 3). For A. cryaerophilus and A. skirrowii, the cadF gene was detected in 34 and 21%, ciaB in 93 and 97%, cj1349 in 52 and 16%, mviN in 93 and 26%, pldA in 26 and 18%, and tlyA in 35 and 21%, respectively. The genes irgA, hecA, and hecB were more frequently detected in A. butzleri strains, although the inclusion of an especially lower number of A. skirrowii representatives may bias this observation. The irgA gene, which was in the reference strains detected only in A. butzleri, was also detected in three A. cryaerophilus strains.

Statistical analyses of the gene distribution related to the Arcobacter strain origin showed for A. butzleri no significant differences for the genes cadF, ciaB, cj1349, mviN, pldA, and tlyA. The gene hecA was significantly highly detected in strains isolated from cattle compared to those from humans, pigs, and chickens (P < 0.05). In contrast, hecB was also significantly highly detected in arcobacters from cattle compared to those from pigs and chickens (P < 0.05) but showed no significant difference in occurrence between strains isolated from cattle and humans. The gene irgA was significantly less represented in pig strains than in human and cattle strains (P < 0.05). For A. cryaerophilus, the genes cadF, cj1349, pldA, and tlyA were significantly more related to arcobacters isolated from humans than to those from chickens (P < 0.01). The gene hecB was more detected in strains isolated from humans than in those from chickens, cattle, and pigs (P < 0.05). The gene pldA was also significantly more detected in strains isolated from cattle than in those from chickens, while the tlyA gene was more detected in arcobacters associated with pigs than in those associated with chickens (P < 0.05). Due to the low numbers of A. skirrowii strains, statistical analyses of the distribution patterns of the putative virulence genes in A. skirrowii related to their biological origin were not performed.

DISCUSSION

Arcobacters are currently not considered microorganisms of major public health concern, but increasing data suggest that their significance in human infections may be underestimated. Little is known about their pathogenic mechanisms or potential virulence factors. However, several studies have already described an adhesive, invasive, and toxigenic capacity in vitro (8, 19). The mechanisms by which A. butzleri induces enteritis have been reported and comprise epithelial barrier dysfunction by a reduced expression of claudin-1, -5, and -8 as well as an induction of epithelial apoptosis (3). The induction of the expression of the cytokine interleukin-8, which is also considered a major virulence factor, has been reported for A. butzleri, A. cryaerophilus, A. skirrowii, and A. cibarius (8). The presence of the nine putative virulence genes has been reported in the A. butzleri ATCC 49616 genome (26). The Arcobacter putative virulence genes share similarities with virulence genes of other bacteria (see Table S1 in the supplemental material). The cadF gene detected in Campylobacter spp. (57% similarity with the corresponding Arcobacter gene) and Wolinella succinogenes (59%), and the cj1349 gene found in Sulfurimonas denitrificans (66%), Campylobacter fetus subsp. fetus (64%), and Campylobacter curvus (62%), encode outer membrane proteins which facilitate cell-to-cell contact with intestinal epithelial cells by adherence to fibronectin (22, 27). cadF and cj1349 sequenced fragments were more than 97.6% conserved in all Arcobacter species tested. The ciaB gene (Campylobacter invasive antigen B) is involved in host cell invasion (23) and is present in various organisms, like Campylobacter spp. (59%), Sulfurovum spp. (65%), Nitratiruptor spp. (65%), Wolinella succinogenes (62%), and Caminibacter mediatlanticus (62%). This invasive gene was the most frequently detected gene in all Arcobacter strains. (>92%). The pldA gene encodes outer membrane phospholipase A, a protein which hydrolyzes acyl ester bonds and can be found in various bacteria and fungi (18, 34). The highest protein similarity of this gene is found in Campylobacter curvus (60%), Campylobacter concisus, and Campylobacter fetus subsp. fetus (58%). The tlyA gene detected in Arcobacter shows similarity to the tlyA gene found in Caminibacter mediatlanticus (73%), Campylobacter lari (70%), Campylobacter upsaliensis (69%), Campylobacter jejuni subsp. jejuni (69%), and Sulfurovum spp. (68%) and is considered a hemolysin gene. The mviN gene was first identified in Salmonella enterica serovar Typhimurium (52%) as a chromosomal gene which causes a typhoid-like disease in mice. A relatively high homology of the Arcobacter mviN gene with its representatives in Campylobacterales bacterium (76%), Sulfurimonas denitrificans (74%), and Wolinella succinogenes (73%) was observed. The gene encodes a sigma factor necessary for the expression of the flg operon, resulting in flagellum synthesis (29). An mviN homolog gene was also found in E. coli (53%) and is involved in peptidoglycan and thus cellular membrane synthesis (17). The irgA gene encodes an outer membrane receptor for enterobactin (6) that is an abundant iron-regulated protein in many bacteria, such as Marinobacter aquaeolei (57%). irgA is identified as a TonB-dependent outer membrane receptor (28) and plays a role in the pathogenesis of urinary tract infections by uropathogenic E. coli (51%) (2). The hecA gene belongs to the filamentous hemagglutinin (FHA) family, while hecB genes encode a hemolysin activation protein. They are found in a wide range of Enterobacteriaceae and are widely spread in both plant (Pseudomonas syringae, Ralstonia solanacearum) and animal (Burkholderia cepacia, Acinetobacter species, and uropathogenic E. coli) pathogens. In Erwinia chrysanthemi, a plant pathogen, the hecA gene participates in attachment and aggregation and is also implicated in epidermal cell killing (32).

In the present study, PCR assays were developed to detect those reported putative virulence genes in a large set of Arcobacter strains of the three species associated with human disease. Whether or not the above-described genes actually become expressed in Arcobacter strains under various environmental conditions and what their role is in Arcobacter pathogenicity need to be further investigated.

In contrast to A. butzleri, no A. cryaerophilus and A. skirrowii strains carried all nine genes, which may indicate a different pathogenic behavior or might be due to a higher heterogeneity in their genomes, making screening of the presence of the putative virulence genes by PCR tests less accurate. Also, the false-negative results of the PCR assay targeting the hecB gene are probably due to a natural variation in the genes involved, as also illustrated by sequencing results. All strains carrying irgA, hecA, and hecB genes also tested positive for the presence of the other six genes.

The distribution pattern of the putative virulence genes in mammalian arcobacters showed no clear correlation with the origin of tested Arcobacter strains, hampering further identification of human infection sources.

This study validated a new, rapid, and accurate PCR tool for the detection of reported putative virulence genes in Arcobacter. In addition, it has been shown that virulence factor homologs to well-known virulence factors of other bacterial pathogens are present in a wide range of Arcobacter strains. Further studies are needed to elucidate the role of those genes in Arcobacter virulence.