Investigation of Campylobacter colonization in three Australian commercial free-range broiler farms.

Campylobacter spp. contaminated poultry products are strongly associated with foodborne illnesses worldwide. Development of effective management strategies to reduce contamination by Campylobacter spp. requires an improved understanding of the numerous factors that drive these contamination processes. Currently, chicken farms are using more free-range chicken meat production systems in response to consumer preferences. However, Campylobacter spp. colonization has rarely been investigated on free-range broiler farms. The present study investigated the temporal and environmental factors influencing Campylobacter spp. colonization of free-range broilers as well as potential sources and genetic diversity of Campylobacter jejuni (C. jejuni) and Campylobacter coli (C. coli) in commercial free-range broiler farms. Genetic linkages among the isolates were analyzed using flaA amplicon analysis. Campylobacter coli was first detected in fecal samples of a commercial free-range broiler flock on day 10 of rearing. Multiple genotypes of C. jejuni and C. coli were identified in this study. The farm environment was identified as a potential source of C. jejuni and C. coli colonization of free-range broilers. The dominant Campylobacter genotype varied between free-range broiler farms over time, with C. jejuni being the most frequently isolated species. These findings enhance the understanding of C. jejuni and C. coli colonization in free-range broiler farms and could inform the development of more effective intervention strategies to help control this important foodborne pathogen.

Introduction

Campylobacter species are important zoonotic pathogens, with Campylobacter jejuni () and Campylobacter coli () being the most 2 common etiological agents of human enteric infections (WHO, 2012). Chickens are commonly considered as natural hosts of Campylobacter spp. as the birds can carry large loads of Campylobacter bacteria in their intestines without showing any clinical signs (Beery et al., 1988; Hermans et al., 2012).

In commercial farms, Campylobacter spp. are often isolated from chickens around 3 wk of rearing (Bull et al., 2006; Yano et al., 2013; Ingresa-Capaccioni et al., 2015; Prachantasena et al., 2016). The production environment of intensive commercial poultry farms is an important source of Campylobacter spp. which colonize chickens (Messens et al., 2009; Ellis-Iversen et al., 2012). Recently, free-range production of broilers has increased in response to consumer demands for products produced by nonintensive systems. However, information on sources and routes of transmission of Campylobacter spp. in broilers from free-range systems is limited, despite the increasing numbers of these types of free-range farms (Miele, 2011; Naald and Cameron, 2011; Singh and Cowieson, 2013; Walley et al., 2015). Templeton (2014) described the diversity of C. jejuni genotypes isolated from cecal contents from Australian intensive and free-range broiler chickens in slaughterhouses but did not investigate colonization and transmission in farms.

The flaA gene has proven to be informative in molecular epidemiological studies (Meinersmann et al., 1997; Petersen and On, 2000; Hiett et al., 2007; Singh and Kwon, 2013; Gomes et al., 2016). In the present study, we used flaA amplicon analysis to investigate the sources, transmission processes, and genetic diversity of C. jejuni and C. coli isolates in commercial free-range broilers in New South Wales (NSW), Australia.

Materials and methods

Farms

Before commencement of this study, appropriate ethics approval was obtained from the Charles Sturt University Animal Care and Ethics Committee (protocol number: 15/057). From May to August 2016, a total of 11 farms were sampled including 8 breeder farms (designated BD–A to BD–H) and 3 free-range broiler farms (designated FB1 to FB3). As illustrated in Figure 1, the 8 breeder farms supplied Ross chicks to the 3 free-range broiler farms. Of the 8 breeder farms, 5 (BD–B, BD–C, BD–D, BD–H, and BD–G) were in NSW and 3 (BD–A, BD–E, and BD–F) were in Queensland (QLD). All farms were part of an integrated poultry production company based in NSW, Australia. The 3 free-range broiler farms were within the same vicinity (approximately 800 m apart), and they were 60 km away from Sydney.

Figure 1

Diagram of free-range broiler and their parent breeder farms in experiments 1 and 2 of this study.1 Indicates all depopulated breeder farms.

After being reared in commercial closed barns in the first 21 d of age, the broiler chickens were free to roam in a fenced outdoor environment through barn flaps during daytime until reaching market weight (Free Range Egg and Poultry Australia, 2012), with the maximum stock density of 28 to 34 kg/m2 (Australian Chicken Meat Federation, 2018). In this study, a flock was defined as the entire population of chickens housed in the same barn.

This study was conducted over 2 free-range broiler farm production cycles (designated experiment 1 and experiment 2; Figure 1). For both experiments, one barn from each broiler farm was selected as the target barn (designated T), focusing on Campylobacter transmission. The adjacent barns (designated A1 and A2) were used to assess potential transmission between flocks.

The codes for free-range broiler barns used in this study were composed of 3 components and presented as “the farm–the barn–the experiment”. Thus, the target barn (T) on free-range broiler farm 1 (FB1) in experiment 1 (Exp.1) was coded as FB1–T–Exp.1; and the adjacent barns (A1 and A2) were separately coded as FB1–A1–Exp.1 and FB1–A2–Exp.1. The farm codes used in the study are listed in Table 1.

Table 1

Summary of the supplied free-range broiler barns and their parent breeder farms for experiments 1 and 2.

Experiment	Broiler farm				Breeder farm
	Farm	Barn	Chickens (n)	Barn code
1	1	Adjacent1	14,670	FB1–A1–Exp.1	BD–C
		Target	14,670	FB1–T–Exp.1	BD–D1
		Adjacent2	15,390	FB1–A2–Exp.1	BD–D1
	2	Adjacent1	15,030	FB2–A1–Exp.1	BD–A
		Target	15,030	FB2–T–Exp.1	BD–A
		Adjacent2	14,850	FB2–A2–Exp.1	BD–A
	3	Adjacent1	11,980	FB3–A1–Exp.1	BD–C
		Target	11,980	FB3–T–Exp.1	BD–B and BD–C
		Adjacent2	15,030	FB3–A2–Exp.1	BD–C
2	1	Adjacent1	15, 480	FB1–A1–Exp.2	BD–F and BD–E1
		Target	14,760	FB1–T–Exp.2	BD–E1
		Adjacent2	14,760	FB1–A2–Exp.2	BD–F
	2	Adjacent1	14,670	FB2–A1–Exp.2	BD–F
		Target	14,670	FB2–T–Exp.2	BD–F
		Adjacent2	15,390	FB2–A2–Exp.2	BD–F and BD–E1
	3	Adjacent1	11,880	FB3–A1–Exp.2	BD–H1
		Target	11,880	FB3–T–Exp.2	BD–H1
		Adjacent2	14,850	FB3–A2–Exp.2	BD–G and BD–H1

Indicates the depopulated breeder farms.

Determination of Sample Size

The sample size was determined using Epitools (AusVet Animal Health Services) with the population size of 12,000, test sensitivity of 0.9, the desired herd sensitivity of 0.95, and the designed prevalence of 0.1 via http://epitools.ausvet.com.au/content.php?page=FreedomFinitePop&Population (accessed on April 2nd, 2016). The designed prevalence of Campylobacter used in this study was justified at 0.1 (10%) to collect 34 fecal samples from chickens of each barn. Owing to time and logistical limitations, a total of 35 fecal samples from the target barns and 10 fecal samples from each of the adjacent barn were collected (Supplementary Table 1). For the breeder farms, 5 fecal samples per barn were collected (Supplementary Table 1). Therefore, a total of 20 or 30 fecal samples per farm were obtained.

Sample Collection

Fresh fecal and cecal excretions were collected and defined as fecal samples. The environment within and surrounding the barn was also selected for sampling and referred to as the environmental samples for Campylobacter.

Fecal samples from the breeder farms were obtained on day 7 after the placement of broiler chicks for logistic reasons. Of further note, farms BD–D, BD–E, and BD–H were completely depopulated and consequently, samples from the 3 farms were not available.

All free-range broiler farms were sampled before chick placement (day 0) and then weekly, starting from the day of chick placement (day 1 or 3) until all fecal samples of the target barns tested were positive for Campylobacter spp. During each visit, fecal and environmental samples were collected from each broiler barn (Supplementary Table 1). All samples were kept in insulated boxes containing ice packs and transported to the laboratory for processing within 24 h.

Fresh fecal samples were randomly collected from each barn using Amies swabs containing charcoal transport medium (Copan Diagnostics Inc., Murrieta, CA) on the day of chick placement (day 1 or 3) and a sterile fecal container with a spoon (Techno Plas, St Marys, SA, Australia) on week 1 (day 8 or 10), 2 (day 15 or 17), and 3 (day 22 or 24). Additional samples using Amies swabs were obtained from the barn wall (swabbing a 100-cm2 area on each side), water and feed pans, and footwear. In addition, drinking water samples (250 mL each) were collected from drinkers in 3 to 6 areas of each barn and kept in separate sterile plastic containers (Techno Plas). Water samples (250 mL each) from the main tank and puddles (outside the barn) were also collected and kept in separate sterile plastic containers. The presence of other potential hosts of Campylobacter (i.e., mammals, insects, and undomesticated birds) on the broiler farms was considered in this study. However, only fresh rodent feces, darkling beetles, and flies were identified on the farms during the sample collection period. Fresh rodent feces (dark in color, soft and moist textures, and spindle-shaped) and insects (darkling beetles and flies) were collected from the anteroom of each barn and placed in separate sterile plastic bags. Floor drag swabs (inside the barn and anteroom) were also collected using sterile tampons (Libra regular; Svenska Cellulosa Aktiebolaget, Springvale, VIC, Australia) moistened with sterile buffered peptone water (Acumedia; Neogen Corporation, Lansing, MI) by swabbing the floor in a zigzag pattern, including the perimeter and center of the room. Soil samples at free-range areas were obtained by drag swabbing a moist sterile tampon along the barn's outside perimeter. The swabs were placed in separate sterile plastic bags. A total of 1,865 samples were collected in this study (Supplementary Table 2).

Campylobacter spp. Isolation

All samples were processed following the standard ISO 10272:2006 method for Campylobacter isolation (ISO, 2006), with slight modifications. Briefly, all fecal samples were directly streaked onto a Campylobacter-selective agar (Campylobacter agar, Skirrow's agar, and Campy Food Agar; BioMérieux, Marcy l’Etoile, France). Water samples were filtered using a membrane that was 47 mm in diameter with a pore size of 0.45 μm (Merck Millipore, Burlington, MA). The membranes were then enriched using 10 mL of Bolton broth (Oxoid, Cambridge, UK). All swab samples were enriched using 10 mL of selective enrichment Bolton broth (Oxoid).

All streaked plates and enriched samples were incubated at 42°C for 48 h under a microaerobic environment generated with a BD GasPak EZ container system (Becton Dickinson Microbiology, North Ryde, NSW, Australia). All enriched samples were screened with VIDAS Campylobacter assay (BioMérieux) for Campylobacter spp. detection before plating onto the selective agar plates and incubating under microaerobic conditions as described earlier.

After incubation, individual bacterial colonies (a maximum of 5 colonies per sample) showing morphological characteristics typical of Campylobacter spp. were selected for species-level identification. Colonies morphologically identified as either C. jejuni or C. coli were plated onto sheep blood agar plates (BioMérieux) and incubated under microaerobic conditions to obtain pure colonies (isolates). Subsequently, all pure C. jejuni and C. coli isolates were stored in the FBP Campylobacter growth medium, as previously described (Gorman and Adley, 2004) at −80°C.

C. jejuni and C. coli Identification

A two-stage approach was used to confirm the identity and differentiate C. jejuni and C. coli colonies in this study. Initially (primary identification), species testing of colonies was performed in an industry laboratory using their established protocols with the VITEK MS system (BioMérieux). Briefly, the VITEK MS system is a matrix-assisted laser desorption ionization time-of-flight mass spectrometer (MALDI-TOF MS) which has an inbuilt capacity to identify C. jejuni or C. coli using the manufacturer's default analysis algorithms. The VITEK MS system was used to screen the individual colonies selected based on morphology to provide a putative identity as either C. jejuni or C. coli. All isolates identified as C. jejuni or C. coli were subsequently transported to an academic laboratory, where PCR assays (tertiary identification) were used to designate each isolate as either C. jejuni or C. coli before genotyping.

Genomic DNA Extraction

DNA was extracted from pure colonies of C. jejuni and C. coli using the PREPMAN Ultra Sample Preparation (Applied Biosystems, Foster City, CA), in accordance with the manufacturer's instructions. DNA samples were stored at −20°C until required.

C. jejuni and C. coli Confirmation

A conventional PCR assay (S1000 Thermal Cycler; Bio-Rad, Australia) was used to confirm C. jejuni and C. coli. A PCR protocol specific for the 16s rRNA (Campylobacter genus), mapA (C. jejuni), and lpxA (C. coli) genes was applied as previously described (Devi, 2019). Each PCR reaction volume was 25 μL, containing 2 U Platinum Taq polymerase (Invitrogen, Carlsbad, CA), 1 × Green PCR Rxn Buffer/MgCl2 (Invitrogen), 1.5 mmol MgCl2 (Invitrogen), 0.2 mmol of dNTPs mixed (Invitrogen), 0.2 μmol of each primer set (Integrated DNA Technologies, Singapore) as described in Table 2, RNase-free water (to a final volume of 24 μL), and 1 μL of DNA template (10–30 ng) as previously described (Devi, 2019).

Table 2

Oligonucleotide primers used for identification of Campylobacter spp., Campylobacter jejuni, and Campylobacter coli (Devi, 2019).

Group or species	Gene	Sequence 5′ to 3′	Amplicon size (bp)
Campylobacter	16S rRNA	Forward: CGTGCTACAATGGCATATACAATGA	113
		Reverse: CGATTCCGGCTTCATGCTC
C. jejuni	mapA	Forward: CACTTTAGACACTGGTATTGCTTTG	191
		Reverse: GATCGTTATTGTCAAGCACAACTATTC
C. coli	lpxA	Forward: GATGATGTTGTTATTGAGGCTTATG	92
		Reverse: GAAAGTATTCTCGCCCCTTG

The PCR cycling conditions consisted of 94°C for 2 min, followed by 40 cycles of 94°C for 10 s, 60°C for 20 s, 72°C for 30 s, and finally 72°C for 5 min. The PCR products were stained with the Midori Green Stain (Nippon Genetics, Duren, Germany) and analyzed using 1.5% gel electrophoresis at 80 V for 40 min. The amplicon sizes were compared with a molecular weight marker (1-kb ladder, New England Biolabs, Ipswich, MA). C. jejuni ATCC 49943 and C. coli ATCC 33559 were used as positive controls for each PCR reaction. RNase water was used as the non-DNA template control.

Genotyping Process

Two processes, partial amplification of the flaA gene and genotyping analysis, were used to genotype study isolates. Based on results of these processes for C. jejuni and C. coli identification, 3 outcomes were used to classify samples: the samples containing C. jejuni, the sample containing C. coli, or the samples containing both C. jejuni and C. coli (different single purified colonies tested). For the purposes of genotyping, if a sample contained either C. jejuni or C. coli, then a single isolate was used for genotyping. If the sample contained C. jejuni and C. coli, then an isolate of each species was used for genotyping.

Amplification of the flaA Gene

C. jejuni and C. coli isolates were assessed for flaA amplification as described by Merchant-Patel et al. (2010) with minor modifications. Each flaA PCR reaction (20 μL) contained 1 × Type-it HRM-PCR kit (Qiagen, Hilden, Germany), 6.6 μL of MilliQ water, 0.7 μmol of flaA primers (Sigma-Aldrich, St. Louis, MO), and 2 μL of DNA template. The PCR assay was performed in a Rotor-Gene Q thermal cycler (Qiagen). The PCR conditions were applied at 95°C for 5 min, 40 cycles of 95°C for 10 s, 60°C for 15 s, followed by 72°C for 30 s.

Genotyping Analysis

The flaA amplicons were commercially sequenced using the Sanger sequencing method (Australian Genomic Research Facility, Sydney, NSW, Australia). The nucleotide sequence alignment was performed using BioEdit Sequence Alignment Editor (version 7.2.5). The flaA allele and peptide numbers were identified by interrogation of each isolate in the Campylobacter flaA database, using the portal: http://pubmlst.org/campylobacter (accessed December 17th, 2019). The C. jejuni and C. coli genotypes were determined by analyzing the flaA allele numbers.

Results

Campylobacter spp. were cultured from 526 (28.3%) of the 1,856 samples collected (Supplementary Table 2). Of these, 465 samples (88.4%) were fecal samples obtained from the breeder (n = 118) and free-range broiler farms (n = 347). The remaining 61 samples (11.6%) were from the environment of free-range broiler farms (Supplementary Table 2).

Campylobacter spp. Isolation From Breeder Farms

Campylobacter spp. were isolated from 118 (98.3%) of 120 fecal samples collected from 5 breeder farms, with the isolation rates ranging from 95 to 100% (Table 3). In this study, 7 isolates from 7 fecal samples initially identified as C. jejuni or C. coli by the VITEK MS system were reassigned by the PCR assays (Supplementary Table 3). The 12 additional isolates from the 7 fecal samples (as reculturable) were tested with the PCR assays. Of the 7 fecal samples retested, 5 and 2 samples were confirmed as C. jejuni and C. coli, respectively. Consequently, of the 118 positive fecal samples, 100 were identified as C. jejuni (n = 67) or C. coli (n = 33), and the remaining 18 contained both C. jejuni and C. coli (C. jejuni and C. coli were separately identified in different single colonies) as shown in Table 3. Hence, the genotypes of 85 C. jejuni (67 samples containing only C. jejuni and 18 samples containing both C. jejuni and C. coli) and 51 C. coli isolates (33 samples containing only C. coli and 18 samples containing both C. jejuni and C. coli) from breeder farms were further assessed with the flaA amplicon analysis. C. jejuni was the most frequently isolated species from the breeder farms in both experiments: Exp.1: BD–A and BD–C and Exp.2: BD–F and BD–G (Table 3).

Table 3

Summary of Campylobacter jejuni and Campylobacter coli isolated from fecal samples from breeder farms based on polymerase chain reaction assays.

Farm	Barn (n)	Samples			Campylobacter species identified
		Tested	Positive	%	C. jejuni	C. coli	C. jejuni and C. coli
BD–A	5	25	25	100.0	14	6	5
BD–B	4	20	19	95.0	8	9	2
BD–C	4	20	20	100.0	10	8	2
BD–F	6	30	30	100.0	13	9	8
BD–G	5	25	24	96.0	22	1	1
Total	24	120	118	98.3	67	33	18

Campylobacter spp. Isolation From Free-Range Broiler Farms

Campylobacter spp. were isolated from 17 of 18 barns, and one barn (FB3–A2–Exp.1) was negative (Table 4). Of the Campylobacter-positive barns, 9 had either C. jejuni or C. coli and 8 were positive for both. Based on the VITEK MS and PCR analyses, the same outcomes for the identification of C. jejuni and C. coli isolates from the broiler farms were identified (Supplementary Table 3). Campylobacter spp. were isolated from 408 (23.5%) samples (Table 4). C. jejuni and C. coli were isolated from 314 (77.0%) and 87 (21.3%), respectively, with 7 (1.7%) containing both species of interest (Table 4). Thus, the genotypes of 321 C. jejuni (314 samples containing only C. jejuni and 7 samples containing both C. jejuni and C. coli) and 94 C. coli isolates (87 samples containing only C. coli and 7 samples containing both C. jejuni and C. coli) from broiler farms were further assessed with the flaA amplicon analysis. C. jejuni was the most frequently isolated species in 14 positive barns and C. coli in the other 3 (Table 4).

Table 4

Summary of Campylobacter jejuni and Campylobacter coli isolated from samples collected from experiments 1 and 2 on 3 free-range broiler farms based on polymerase chain reaction assays.

Barn	Samples			Campylobacter species identified
	Tested	Positive	%	C. jejuni	C. coli	C. jejuni and C. coli
FB1–A1–Exp.1	45	11	24.4	11	0	0
FB1–T–Exp.1	213	42	19.7	42	0	0
FB1–A2–Exp.1	45	20	44.4	20	0	0
FB2–A1–Exp.1	45	11	24.4	10	1	0
FB2–T–Exp.1	211	45	21.3	34	9	2
FB2–A2–Exp.1	45	12	26.7	0	12	0
FB3–A1–Exp.1	34	8	23.5	0	8	0
FB3–T–Exp.1	161	46	28.6	1	45	0
FB3–A2–Exp.1	34	0	0.0	0	0	0
FB1–A1–Exp.2	45	16	35.6	8	6	2
FB1–T–Exp.2	214	42	19.6	42	0	0
FB1–A2–Exp.2	45	21	46.7	20	1	0
FB2–A1–Exp.2	45	11	24.4	11	0	0
FB2–T–Exp.2	210	45	21.4	45	0	0
FB2–A2–Exp.2	45	11	24.4	11	0	0
FB3–A1–Exp.2	45	12	26.7	10	1	1
FB3–T–Exp.2	209	43	20.6	40	3	0
FB3–A2–Exp.2	45	12	26.7	9	1	2
Total	1,736	408	23.5	314	87	7

Genetic Diversity of C. jejuni and C. coli

A total of 551 isolates (C. jejuni, n = 406 and C. coli, n = 145) identified in the breeder and broiler farms were genotyped by flaA amplicon analysis. The flaA nucleotide sequences were assigned into allele numbers. The 406 C. jejuni isolates were grouped into 29 genotypes: 24 recognized flaA alleles and 5 unassigned flaA alleles (Supplementary Table 4). The 145 C. coli isolates were grouped into 20 genotypes: 14 recognized flaA alleles and 6 unassigned flaA alleles (Supplementary Table 5).

Genetic Diversity of C. jejuni and C. coli in Breeder Farms

The C. jejuni isolates (n = 85) from the breeder farms were grouped into 23 genotypes: 18 recognized flaA alleles and 5 unassigned flaA alleles (Supplementary Table 4). Of the 23 genotypes, 5 and 11 genotypes were isolated from Exp.1 and Exp.2, respectively, and the remaining 7 genotypes were isolated from both experiments. By contrast, the C. coli isolates (n = 51) were grouped into 18 genotypes: 12 recognized flaA alleles and 6 unassigned flaA alleles (Supplementary Table 5). Moreover, of these 18 genotypes, 8 and 7 were isolated from Exp.1 and Exp.2, respectively, and 3 were isolated from both experiments.

Genetic Diversity of C. jejuni and C. coli in Free-Range Broiler Farms

The C. jejuni isolates (n = 231) were grouped into 9 genotypes, which belonged to 9 recognized flaA alleles (Supplementary Table 4). Among these, 3 (flaA alleles 14, 18, and 208) and 4 (flaA alleles 2, 18, 105, and 1033) were isolated from Exp.1 and Exp.2, respectively. The remaining 2 (flaA alleles 57 and 239) were isolated from both experiments. In comparison, the C. coli (n = 94) isolates were grouped into 5 genotypes, which were assigned to 5 recognized flaA alleles (Supplementary Table 5). Among these, one was identified exclusively in Exp.1 (flaA allele 769) and one exclusively in Exp.2 (flaA allele 16). The remaining 3 (flaA alleles 30, 36, and 256) were isolated from both experiments.

Dynamics of C. jejuni and C. coli Colonization From Each F-Range Broiler Farm Between Experiments

Free-range Broiler Farm 1

Seventy-three C. jejuni isolates from Exp.1 belonged to flaA alleles 14 (n = 47) and 57 (n = 26), respectively (Figure 2 and Supplementary Table 6). The C. jejuni flaA allele 14 was first isolated from 10 fecal samples of FB1–A2–Exp.1 on day 15. On day 22, this genotype was isolated from fecal samples from all barns (FB1–A1–Exp.1, n = 1; FB1–T–Exp.1, n = 23; and FB1–A2–Exp.1, n = 10), farm boots, and the environment of FB1–T–Exp.1 (drinking water and the free-range area). In addition, the C. jejuni flaA allele 57 was also isolated on the same time (day 22) from fecal samples of 2 barns (FB1–A1–Exp.1, n = 9; and FB1–T–Exp.1, n = 12) as well as the free-range area of FB1–A1–Exp.1 and the internal environment of FB1–T–Exp.1 (floors, walls, and barn boots).

Figure 2

Schematic diagram of dynamics of Campylobacter jejuni and Campylobacter coli flaA types identified on free-range broiler farm 1 (FB1) from experiments 1 and 2.

In comparison, 72 C. jejuni isolates from Exp.2 belonged to flaA alleles 16 (n = 71) and 239 (n = 1) (Figure 2 and Supplementary Table 6). The C. jejuni flaA allele 16 was first isolated from 10 fecal samples of FB1–A2–Exp.2 on day 15. On day 22, this genotype was isolated from fecal samples (FB1–A1–Exp.2, n = 8; FB1–T–Exp.2, n = 35; and FB1–A2–Exp.2, n = 10), free-range areas (FB1–A1–Exp.2 and FB1–T–Exp.2), farm boots, and internal environment of FB1–T–Exp.2 (anteroom, floors, walls, barn boots). However, the C. jejuni flaA allele 239 was isolated only from a fecal sample of FB1–A1–Exp.2 on day 22, whereas the C. coli isolates (n = 9) isolated from FB1–A1–Exp.2 and FB1–A2–Exp.2 belonged to flaA alleles 16 (n = 1), 30 (n = 7), and 36 (n = 1) (Figure 2 and Supplementary Table 6). The C. coli flaA allele 30 was first isolated from 5 fecal samples of FB1–A1–Exp.2 on day 15. On day 22, fecal samples from the same barn were positive for C. coli flaA alleles 30 (n = 2) and 16 (n = 1), whereas the C. coli flaA allele 36 was isolated only from the free-range area of FB1–A2–Exp.2 on day 15.

Free-range Broiler Farm 2

Forty-six C. jejuni isolates from Exp.1 belonged to flaA alleles 14 (n = 15), 18 (n = 1), and 208 (n = 30) (Figure 3 and Supplementary Table 6). The C. jejuni flaA allele 18 was isolated only from rodent feces from FB2–T–Exp.1 on day 8. The C. jejuni flaA allele 14 was first isolated from fecal samples from FB2–A1–Exp.1 (n = 1) and FB2–T–Exp.1 (n = 11) and the environment of FB2–T–Exp.1 (walls and the free-range area) on day 22. At the same time, the C. jejuni flaA allele 208 was first isolated from fecal samples (FB2–A1–Exp.1, n = 9; and FB2–T–Exp.1, n = 20) and rodent feces from FB2–T–Exp.1. Moreover, the C. coli isolates (n = 24) belonged to flaA alleles 30 (n = 15), 256 (n = 8), and 769 (n = 1) (Figure 3 and Supplementary Table 6). The C. coli flaA allele 256 was first isolated from rodent feces from FB2–T–Exp.1 and the free-range area of FB2–A2–Exp.1 on day 1. This genotype was isolated from other samples from the same target barn (FB2-T-Exp.1) at different time points, such as barn boots (day 8) and rodent feces (day 8, 15, and 22) as well as 2 fecal samples of FB2–A2–Exp.1 (day 22). The C. coli flaA allele 769 was isolated only from the free-range area of FB2–A1–Exp.1 on day 8. The C. coli flaA allele 30 was first isolated from a fecal sample of FB2–A2–Exp.1 on day 15. One week later, this genotype coexisted between fecal samples of different barns (FB2–A2–Exp.1, n = 8; and FB2–T–Exp.1, n = 4) and the floors of FB2–T–Exp.1.

Figure 3

Schematic diagram of dynamics of Campylobacter jejuni and Campylobacter coli flaA types identified on free-range broiler farm 2 (FB2) from experiments 1 and 2.

In comparison, the C. jejuni isolates (n = 67) from Exp.2 belonged to flaA alleles 2 (n = 2), 16 (n = 48), 105 (n = 1), 239 (n = 15), and 1033 (n = 1) (Figure 3 and Supplementary Table 6). The C. jejuni flaA alleles 1033, 105, and 239, isolated from rodent feces from FB2–T–Exp.2, were found on day 0, 1, and 8, respectively. One week later, on day 15, the C. jejuni flaA allele 2 was isolated only from rodent feces and the anteroom floor in FB2–T–Exp.2. Then, the C. jejuni flaA allele 16 was first isolated from the barns and the environment on day 22, including from fecal samples (FB2–A1–Exp.2, n = 6; and FB2–T–Exp.2, n = 35), free-range areas of all three barns (FB2–A1–Exp.2, FB2–T–Exp.2, and FB2–A2–Exp.2), the internal environment of FB2–T–Exp.2 (floors and barn boots), and farm boots. At the same time, the C. jejuni flaA allele 239 was isolated from fecal samples of FB2–A1–Exp.2 (n = 4) and FB2–A2–Exp.2 (n = 10).

Free-Range Broiler Farm 3

The C. jejuni flaA allele 239 was isolated only from a sample of rodent feces from FB3–T–Exp.1 on Day 3 (Figure 4 and Supplementary Table 6). Although the C. coli isolates (n = 53) from Exp.1 belonged to flaA alleles 30 (n = 7) and 36 (n = 46) (Figure 4 and Supplementary Table 6), the C. coli flaA allele 36 was first isolated from the free-range area of FB3–A1–Exp.1, farm boots, and samples of FB3–T–Exp.1 (fecal samples; n = 3 and barn boots) on day 10. A week later (day 17), this genotype persisted, being isolated from the farm boots and other samples from FB3–T–Exp.1, such as fecal samples (n = 35), free-range area, and internal environment (floor, walls, water pans, and barn boots). The C. coli flaA allele 30 was isolated from only 7 fecal samples of FB3–A1–Exp.1 (day 17).

Figure 4

Schematic diagram of dynamics of Campylobacter jejuni and Campylobacter coli flaA types identified on free-range broiler farm 3 (FB3) from experiments 1 and 2.

In comparison, the C. jejuni isolates (n = 62) from Exp.2 were isolated only from day 24 and belonged to flaA alleles 16 (n = 1), 57 (n = 36), 239 (n = 22), and 105 (n = 3) (Figure 4 and Supplementary Table 6). The C. jejuni flaA allele 57 was isolated only from FB3–T–Exp.2, including the free-range area, fecal samples (n = 32), and the internal environment (floors and barn boots). The C. jejuni flaA allele 239, previously isolated from samples in Exp.1, was also isolated from farm boots, free-range areas of FB3–A1–Exp.2 and FB3–A2–Exp.2, as well as fecal samples of FB3–A1–Exp.2 (n = 10) and FB3–A2–Exp.2 (n = 9). The C. jejuni flaA allele 105 was found in 2 fecal samples of FB3–T–Exp.2 and a fecal sample of FB3–A2–Exp.2. Furthermore, the C. jejuni flaA allele 16 was found only in a fecal sample of FB3–T–Exp.2. By contrast, the C. coli isolates (n = 8) from Exp.2 were assigned to flaA alleles 36 (n = 7) and 256 (n = 1) (Figure 4 and Supplementary Table 6). The C. coli flaA allele 36, previously isolated in Exp.1, was detected in Exp.2. This genotype was first isolated from the free-range area of FB3–A1–Exp.2 before chick placement (Figure 4). Two weeks later (day 17), this genotype was isolated from rodent feces from FB3–T–Exp.2, a fecal sample of FB3–A2–Exp.2, and farm boots. At a later time point (day 24), this genotype was subsequently isolated from a fecal sample of FB3–A1–Exp.2 and 2 fecal samples of FB3–A2–Exp.2. Moreover, the C. coli flaA allele 256 was only isolated from rodent feces from FB3–T–Exp.2 on day 24 (Figure 4 and Supplementary Table 6).

Similarity of C. jejuni and C. coli Isolates From Breeders and Their Broiler Progeny

Three C. jejuni genotypes (Supplementary Table 4) and 3 C. coli genotypes (Supplementary Table 5) were isolated from samples from both breeder farms and free-range broiler flocks.

Of the 3 C. jejuni genotypes (flaA alleles 18, 105, and 239), only C. jejuni flaA allele 239 was isolated from fecal samples of a breeder farm and in fecal samples from its broiler offspring, despite being located in geographically distant areas. The C. jejuni flaA allele 239 was isolated from 2 fecal samples of BD–F, located in QLD, and one fecal sample of FB1–A1–Exp.2, located in NSW, on day 22 (Figure 5B). However, this genotype was also identified in other linked broilers in FB2 located in NSW within the same experiment (Exp.2) (Figure 5B). However, it was also found in a sample of rodent feces (day 8) in FB2–T–Exp.2, and on day 22, it was identified in fecal samples of FB2–A1–Exp.2 (n = 4) and FB2–A2–Exp.2 (n = 10).

Figure 5

Schematic diagram of similarity of Campylobacter jejuni and Campylobacter coli flaA types between breeder farms and their progeny in experiments 1 (A) and 2 (B).

Notably, 2 of 3 C. coli genotypes (flaA alleles 16 and 30) isolated from fecal samples of 3 breeder farms and their progeny were genetically similar. The C. coli flaA allele 30 was isolated from both breeders and linked broilers, located in NSW and QLD. This genotype was isolated from 2 fecal samples of BD–A based in QLD and the samples from FB2 in Exp.1 based in NSW such as fecal samples of FB2–T–Exp.1 (n = 4, day 22) and FB2–A2–Exp.1 (n = 1, day 17; n = 8, day 22) and 2 floor samples of FB2–T–Exp.1 on day 22 (Figure 5A). This genotype was also isolated from a fecal sample from BD–C located in NSW and 7 fecal samples from FB3–A1–Exp.1 based in NSW (day 17) (Figure 5A). In addition, the C. coli flaA allele 16 was isolated from samples from Exp.2: from breeders and their progeny located in different states. This genotype was isolated from 2 fecal samples of BD–F, located in QLD, and one fecal sample of FB1–A1–Exp.2, located in NSW (day 22) (Figure 5B).

Discussion

In this study, C. jejuni and C. coli were isolated from the breeder and free-range broiler farms, as has been previously reported (Vandeplas et al., 2010; O'Mahony et al., 2011; Prachantasena et al., 2016). Most studies report C. jejuni and C. coli to be the first species isolated after 2 wk of rearing in commercial farms (Bull et al., 2006; Yano et al., 2013; Prachantasena et al., 2016), but these microorganisms have been detected earlier in free-range farms. For example, El-Shibiny et al. (2005) reported that a free-range broiler flock in the UK was colonized by C. jejuni within 1 wk of rearing (day 8). The present study, to the best of our knowledge, is the first to show the early detection of C. coli in fecal samples of chickens in a commercial free-range broiler flock approximately 1 wk (day 10) after chick placement.

The genotypes of C. jejuni and C. coli isolated in the present study were diverse, consistent with previous reports (Colles et al., 2011; Vidal et al., 2016). One reason for this could be that multiple Campylobacter genotypes from various sources can accumulate and persist simultaneously within broiler flocks (Ridley et al., 2008a). In addition, Ridley et al. (2008b) have suggested that C. jejuni could undergo genetic rearrangement by 4 wk after challenge with mixed strains in the birds due to the competitive environment in the chicken gut, thus leading to diverse genotypes. The data also showed more genetic diversity on the breeder farms compared with that on the free-range broiler farms. This has been reported previously and suggests that Campylobacter colonization of breeder chickens is a dynamic process, supported by the notion of repeat exposure in longer-lived breeders compared with broilers (Colles et al., 2011).

Some C. jejuni and C. coli genotypes isolated from the broiler farms were common among chicken feces from the different farms and environments isolated in this study. This suggests that free-range broiler flocks in the same area (although in different farms) are exposed to the same sources of Campylobacter and thus share similar genotypes. Some of these genotypes not only coexisted within a single free-range broiler barn and its environment but were also detected in the adjacent barns and farm environment; this suggests the spread of the microorganisms between the broilers and the surrounding environment. Similar findings have been described previously by Zweifel et al. (2008).

Our data indicated that the dynamics of Campylobacter spp. colonization and the dominant genotypes within a single barn depend on the time of sample collection. For example, the pre-existing dominant C. coli genotype was replaced with a new emerging C. jejuni in some free-range broiler barns. This implies that the newly acquired species could have been more successful in colonizing chickens. A new upcoming C. coli genotype isolated from the environment was unable to replace the pre-existing C. coli genotype, implying that it was less competitive than the pre-existing genotype. Competitive exclusion among Campylobacter species and genotypes in chickens during colonization may lead to one genotype replicating rapidly and becoming dominant (Hook et al., 2005; Colles et al., 2019).

The present study is the first to demonstrate that horizontal transmission (the environment to birds) played an essential role in the colonization of free-range broiler farms in Australia. Importantly, our data showed that the same C. coli genotype from the first production cycle in this study (Exp.1) remained in the environment before the chick placement and was subsequently detected in chicken feces in the associated flock during the following production cycle (Exp.2). This demonstrates the potential for carryover or reintroduction of C. coli between consecutive free-range broiler barns. Improved hygiene practices and appropriate biosecurity measures could reduce Campylobacter transmission in broiler farms (Smith et al., 2016).

As layer breeder chickens supply the eggs for multiple generations of broiler chickens, the possibility of vertical transmission of Campylobacter from layers to broilers is of interest. Carriage of Campylobacter spp. in the eggs, by the previous infection of the eggs within the breeder population, would provide a potential source for vertical transmission. Thus, if vertical transmission was an important source of broiler colonization, Campylobacter control in the layer birds could be an effective intervention point. Cox et al. (2012) suspected that Campylobacter could be transmitted from the breeder flock to the fertile eggs through the hatchery and to the broiler farms. However, a few studies have reported the same C. jejuni or C. coli strains in broiler breeder flocks and their progeny (Cox et al., 2002; Idris et al., 2006), thereby suggesting that the layer hens can be a potential source of Campylobacter spp. in broiler chickens, suggesting the possibility of vertical transmission.

The present study provides some evidence to support the possibility of vertical transmission as some isolates from breeder farms (n = 3) were the same genotypes as the isolates from their progeny in broiler flocks (n = 4) from the same region (approximately 500 km apart) and different regions (approximately 1,000 km apart). However, further studies are required to investigate this, as fecal samples from some breeder farms could only be collected after their corresponding chicks were placed at broiler farms or not at all in this study for commercial reasons. Consequently, it was not possible to determine the specific Campylobacter genotypes, if any, in the breeder farms at the time of egg laying.

Another possible mechanism of Campylobacter spp. transmission in hatching chicks could be the uptake of Campylobacter spp. from hatchery-related samples such as contaminated eggshells and tray liners in hatcheries (Byrd et al., 2007; Messelhausser et al., 2011). In the present study, sampling at the hatchery was not possible for commercial reasons. Because of these factors, directly tracing specific genotypes of Campylobacter spp. through the complete broiler production system was not possible. For future studies, sampling at the hatchery stage should be included to investigate the role of the hatchery in Campylobacter transmission. In addition, other molecular methods such as multilocus sequence typing and whole-genome analyses are required for greater understanding of C. jejuni and C. coli genotypes compared with global epidemiology.

Conclusions

Horizontal transmission was identified as the most frequent mode of colonization of free-range broiler chickens. Although dominant genotypes were identified, all free-range broiler flocks studied were exposed to or colonized by multiple Campylobacter genotypes earlier in the production cycle. Also of interest was the detection of diverse genotypes in the longer-lived layer birds, where it might be expected that the colonizing genotype may stabilize over time. Collectively, these data indicate that the colonization of chickens with Campylobacter is a complex and dynamic process and that effective ongoing control of this critical foodborne pathogen through the broiler production system will require a multifaceted approach.