Ten farms within the southeastern United States were sampled over a period of 3 years (from 2014 to 2016), representing 37 pasture-raised broiler flocks. Farm descriptions are available in Table 1. Soil and feces samples were collected from the pasture where the flock was currently residing at the time of sampling. Samplings occurred three times during grow-out: (i) within a few days of being placed in the pasture, (ii) halfway through their time on pasture, and (iii) on the day the flock was processed. At each sampling time, the pasture area was divided into five separate sections, and five subsamples in each section were pooled into a single sample for each section (a total of five soil samples and five feces samples were collected on each sampling day). Soil samples were collected from the surface (0–7 cm) with sterile scoops, and feces samples were collected from fresh droppings on the soil surface. Gloves and scoops were changed between sample types and between sampling areas. Samples were transported back to the lab on ice and processed within 2 h of collection. A total of 1,110 samples (555 feces samples and 555 soil samples) were collected over the 3-year study period.

Enrichment and isolation of Listeria from these environmental samples were performed using a modified version of the USDA-FSIS MLG 8.10 method (36). Three grams of fresh soil or feces were added to 9 ml of buffered peptone water (Acumedia, Lansing, MI, USA) in a filtered stomacher bag and were vigorously shaken for 30 s. As a pre-enrichment step, the stomached homogenates remained in the filtered stomacher bag and were incubated overnight at 35°C. This pre-enrichment step was followed by two enrichments in University of Vermont Modified Listeria Enrichment Broth (UVM; Remel, Lenexa, KS, USA) and Fraser Broth (Oxoid CM0895, Basingstoke, UK), both requiring overnight incubation for 24 h at 30°C. One loopful of the Fraser’s enrichment culture was streaked on Listeria selective agar (LSA, Oxoid CM0856, Basingstoke, UK) for the isolation of Listeria colonies. These plates were incubated overnight at 30°C, and on each plate three Listeria-like colonies per positive samples were picked and kept for further identification tests. Stock cultures were prepared by growing Listeria strains in tripticase soy broth (TSB; Acumedia, Lansing, MI, USA) at 37°C. After washing in sterile water, the cell pellet was suspended in a brain heart infusion (BHI) broth (Acumedia, Lansing, MI, USA) with 25% of glycerol, aliquoted (300 µl in microtubes) and frozen at −80°C until further utilization.

The species of presumptive Listeria colonies recovered on LSA were determined by multiplex PCR (37). In short, speciation occurred using two multiplex PCR reactions, based on the size of PCR amplicons. Pool 1 contained the primers for the identification of L. ivanovii, L. grayi, and L. innocua, and pool two contained primers for the identification of L. welshimeri, L. monocytogenes, and L. seeligeri. A 25 µl PCR reaction was composed of 1× EconoTaq PLUS 2× Master Mix (Lucigen Corporation, Middleton, WI, USA), 1 µM of each livN, Igr, and lin2 reverse and forward primers (for Pool 1) or 1 µM of each lwe, Lmo, and lse reverse and forward primers (for Pool 2) and quantity sufficient (qs) of water. For negative controls, sterile water was added instead of template DNA. The cycling program consisted of 1 cycle at 95°C for 9 min; 30 cycles at 94°C for 30 s, at 60°C for 30 s, and at 72°C for 1 min; and 1 cycle at 72°C for 7 min. The serovar group of isolates classified as L. monocytogenes was determined by multiplex PCR using five sets of primers (38). Briefly, one colony of L. monocytogenes isolates was thoroughly mixed in a 25 µl PCR reaction containing: 1× EconoTaq PLUS 2× Master Mix (Lucigen Corporation, Middleton, WI, USA), 1 µM of each Lmo0737, ORF2819, and ORF2110 reverse and forward primers, 1.5 µM of Lmo1118 reverse and forward primers, 0.2 µM of prs reverse and forward primers and qs water. For negative controls, sterile water was added instead of template DNA. PCR was performed with an initial denaturation step at 94°C for 3 min; 35 cycles of 94°C for 0.40 min, 53°C for 1.15 min and 72°C for 1.15 min; and 1 final cycle for 72°C for 7 min. PCR reactions were performed in an Eppendorf Mastercycler EP Gradient S (Eppendorf). After the completion of all cycles, 18 µl of PCR product was mixed with 3 µl of BlueJuice™ loading buffer (Invitrogen, Carlsbad, CA, USA) and separated on a 2% E-gel^® with SYBR-safe™ (Invitrogen) along with 12 µl of E-Gel™ 1 kb Plus DNA Ladder (Invitrogen).

Growth curves were plotted based on OD values over time. Each bacterial growth curve was fitted to a modified Gompertz model using Matlab 2007b. The model equation is as follows: y=Ae{−e[μmax⋅eA(λ−t)+1]}, where y is the OD value measured, t is the time (h), μ_max is the maximum specific growth rate (h⁻¹), A is the maximum OD value attained, and λ is the lag time (h). Within the m-file written in Matlab, the lsqcurvefit function (a nonlinear least-squares solver for data fitting) was utilized to fit the growth curves by first using the following Gompertz equation: y=Ae−e(B−Cx).

Then, the A, B, and C terms were used to determine the growth parameters of interest in the model equation as follows: μm=A ⋅ Ce, λ=B−1C.

The raw data for each growth curve were graphed along with the resulting fit, and the R² value (coefficient of determination) for each resulting fit was calculated. A four-way analysis of variance (ANOVA) was performed separately on each growth parameter (λ, μ_max, and A), followed by Tukey’s post hoc test in R software v3.2.1. Factors included in the model were the Listeria strains, the culture medium, the inoculum concentration, and the incubation temperature. For the coculture experiments, μ_max and stationary phase cell densities (equivalent to OD_max) were log10-transformed before ANOVA, and a Tukey’s post hoc test was used to group treatments. For all analyses, differences among groups were considered significant if p ≤ 0.05.

A total of 1,110 samples (555 feces samples and 555 soil samples) were collected from 37 flocks on 10 pastured poultry farms over a 3-year period, and the distribution of Listeria species varied according to the sampling year (Figure 1A), broiler farm (Figure 1B), and sample type (Figure 1C). Overall, Listeria species were detected on all the farms and isolated in 15% of samples (83 from feces and 85 from soils), which is in the range of Listeria species prevalences reported in poultry-related environmental samples (from 1.4 to 53%) such as broiler litter, farm feed, farm drinking water, soil, and grass (25, 26, 39, 40) as well as in poultry feces (4.7–17%) (23, 25). In our study, three species were isolated including L. innocua (65.7%), L. monocytogenes (17.4%), and L. welshimeri (15.1%), and each of these species were recovered from at least half of the broiler farms (80, 50, and 90%, respectively; Figure 1B). Although different Listeria species distributions were observed between farms, at least two Listeria species were recovered from all but one farm (Farm M), and all three species were recovered from soil samples in all 3 years of the study (Figure 1C). Listeria innocua has been previously shown to be the predominant species isolated from the broiler farm environment (23–26, 40), while the detection of other non-pathogenic Listeria species, such as L. welshimeri, remains infrequent (23, 40), mostly because studies only focus on L. monocytogenes (41–43).

Listeria monocytogenes was isolated from 5.8, 0.3, and 1.0% of all samples collected in 2014, 2015, and 2016, respectively, with 87% (26/30) recovered during 2014 (Figure 1A) and 57% (17/30) of all L. monocytogenes isolates coming from the only flock sampled on Farm D in 2014 (Figure 1B). Overall, three L. monocytogenes serovar groups were identified: 1/2a-3a (70%), 1/2b-3b-7 (20%), and 4b-4d-4e (10%). Interestingly, over the 3-year sampling period, only one L. monocytogenes-positive flock (Farm I, 2014) harbored more than one serotype, demonstrating the potential clonal nature of L. monocytogenes within a flock or on a farm (44). The overall prevalence of L. monocytogenes on these 10 farms was low compared with other grow-out farm environments where 0–46.2% of the environmental and feces samples were L. monocytogenes positive (45), but the distribution of the serotypes was consistent with other studies that have characterized L. monocytogenes serotypes in broiler flocks (26, 41, 43). The prevalence of L. monocytogenes contamination may be dependent on the type of production system. A significant difference between caged- and floor-reared hens was observed with a greater detection of L. monocytogenes in dust samples from floor-reared hens in L. monocytogenes-positive flocks (41). In alternative systems, broilers are raised in less controlled environments than conventional systems and are more likely to be in contact with L. monocytogenes known to be widely spread in soil and vegetation (35).

Poultry farms frequently have other animals (beef cattle, sheep, goats, or swine) and pets present on the production site (35). These animals can be reservoirs for and play a role in the proliferation and deposition of L. monocytogenes into the environment. In our study, all but two farms had other animals raised in close proximity to the broiler flocks during the sampling period, but we did not investigate the possible genotype matching between animal species. Generally, the presence of other animals on the farm increase the risk factor associated with pathogenic bacteria contamination of poultry flocks (42, 46). This has been shown with Campylobacter spp. where adjacent broiler flocks and cattle appear to be the most frequently identified animals with broiler-flock matching Campylobacter spp. isolates (47). Another study has reported an increased risk of L. monocytogenes contamination in laying hen flocks when pets were present on the production site (42).

Using the cultural conditions for the initial enrichment step for the USDA-FSIS MLG 8.10 L. monocytogenes enrichment method (UVM, 30°C) we found that L. innocua exhibited a significantly shorter lag time than the L. monocytogenes strains in monocultures, especially for low initial inoculum concentrations (10² CFU/ml). To determine if L. innocua has any direct competitive growth advantages over L. monocytogenes in UVM, coculture experiments were performed using the L. innocua strain and the L. monocytogenes 1/2a-3a strain (the most prevalent serovar group from the farm surveys). When both strains were inoculated into the coculture at 10⁵ CFU/ml (Figure 5A), there were no significant differences in λ, μ_max, or stationary phase cell density (similar to OD_max), although L. monocytogenes densities did begin to exceed L. innocua cell densities after 24 h. When both strains started at the lower inoculum level (10² CFU/ml), while λ and μ_max were similar, L. innocua reached a significantly higher stationary phase cell density compared with L. monocytogenes (F = 31, p < 0.01; Figure 5B). Conversely, when the cocultures inoculated at the lower initial concentrations were grown in TSB, L. monocytogenes demonstrated significantly higher stationary phase cell densities compared with L. innocua (F = 19, p < 0.05), with L. monocytogenes cell densities being ~3× greater than L. innocua (Figure 5C). When comparing the growth curve parameters among the three coculture experiments, only the stationary phase cell density was significantly effected at the species-level (Table 2).

Unlike the monoculture results using the UVM protocol from the USDA-FSIS MLG 8.10 method, no significant lag time difference was observed between L. innocua and L. monocytogenes at low initial inoculum concentrations; however, L. innocua still maintained a competitive growth advantage under these cultural conditions represented by significantly higher stationary phase cell densities (Table 2; Figure 5B). When grown under the same conditions in TSB (Figure 5C), L. monocytogenes grew at significantly higher levels than L. innocua, indicating that there is a enrichment media-specific effect on Listeria growth within these cocultures. When looking at the differences in stationary phase cell densities across all cocultures, there was no significant difference for L. innocua between UVM and TSB in the Low:Low cocultures, but there was over a 1 log reduction in stationary phase cell density for L. monocytogenes grown in TSB (7.17 ± 0.06 log₁₀ CFU/ml) compared with UVM (6.38 ± 0.05 log₁₀ CFU/ml) under those growth conditions (Table 2). While it appears that L. innocua has a competitive growth advantage in UVM with low initial inoculum concentrations, it is possible that this advantage comes more from a disadvantage of L. monocytogenes growing under these conditions, rather than a specific advantage that L. innocua possesses, and previous studies have shown that enrichment/culture media can differentially effect L. innocua and L. monocytogenes (28, 31, 33).

However, considering the UVM enrichment is the first of two enrichments in the USDA-FSIS MLG 8.10 protocol, having significantly higher densities of L. innocua compared with L. monocytogenes would artificially increase the likelihood of isolating L. innocua from samples with equivalent levels of L. innocua and L. monocytogenes. Considering the UVM enrichment is used within the USDA-FSIS MLG 8.10 method, and USDA-FSIS is responsible for the testing of foodborne pathogens from broiler production farms and processing plants, this enrichment bias could potentially or partially explain the prevalence of L. innocua as the dominant Listeria spp. found on poultry farms (15, 23–26).