Differences in microbial composition of litter and water line biofilm of broiler farms as influenced by water quality history

Abstract

Microbiome profiling of litter and drinking water lines of poultry environments is essential to understanding their role in the persistence and transmission of opportunistic and pathogenic microorganisms. While the litter environment is well characterized, the drinking water system has received less attention, despite its potential as a control point for Salmonella. In this study, we identified two groups of five farms based on historic water quality issues: normal sulfur-iron (NSIF) and high sulfur-iron (HSIF) content with the aim of assessing the incidence/quantity of Salmonella and characterize the microbiota of the litter and water line biofilms within the two farm groups. Our findings showed no differences in Salmonella incidence and quantity between the litter of both groups, while the biofilm of one of the HSIF farms had Salmonella. Common serotypes, such as Enteritidis, Infantis, Kentucky, and Typhimurium, were identified in the litter of both groups, while Agona, Alachua, and Schwarzengrund were unique to the NSIF group and Ouakam and Worthington to the HSIF group. Generally, the litter exhibited high microbial diversity compared to the biofilm. The litter of the NSIF group had lower alpha diversity than the HSIF group. Conversely, NSIF biofilms had higher alpha diversity than the HSIF biofilms. Distinct differences in microbial community composition were observed in the litter and biofilm of the two farm groups. Aerococcus, Lactobacillus, and Staphylococcus dominated the litter of the two groups, whereas probiotic (NSIF) and pathogenic (HSIF) Bacillus was the most prevalent in the biofilms. These results suggest that water quality issues could influence the microbiota of poultry house environments, particularly in water line biofilms. Effective removal of these biofilms is crucial for controlling Salmonella at pre-harvest poultry production.

Introduction

Poultry litter and water lines are potential reservoirs for microbial contamination, including foodborne pathogens such as Salmonella at pre-harvest poultry production (Chinivasagam et al., 2010; Marin et al., 2009; Wingender and Flemming, 2011; Zimmer et al., 2003). For instance, in the United States where several flocks are raised on a single bed of layered litter topped by layers of new bedding material between flocks for at least a year, the litter is likely to maintain microbiological record of every past flock (Dumas et al., 2011). Poultry litter consists mainly of bird excreta, feed, bedding material, manure, and feathers (Crippen et al., 2021; Dunn et al., 2022). Several studies have reported on the presence and persistence of Salmonella in poultry litter environment (Chinivasagam et al., 2010; Dunn et al., 2022; Obe et al., 2023). Moreover, opportunistic and disease-causing microorganisms originating from organic matter in water lines can persist in the drinking water systems, facilitating their transmission to flocks (Van Eenige et al., 2013; Sparks, 2009).

Water quality is of critical importance to animal health and production, especially food-producing animals like poultry (Amaral, 2004). Drinking water supplied to birds should be clean, hygienic, and of acceptable mineral composition level to prevent microbial growth in the drinking water system because microbial water quality can directly influence poultry health and water intake (Maharjan et al., 2016; Münster and Kemper, 2024; Mustedanagic et al., 2023). Maximum recommended levels for poultry drinking water quality standards are less than 0.3 ppm for iron, 200 ppm for sulfate, and 5-8 for pH, among other water quality parameters (Watkins, 2008). However, sulfate concentrations of 50 ppm or more in combination with high levels of magnesium and chloride in water could negatively affect bird performance (Carter and Sneed, 1996). Moreover, the water supplied in commercial poultry houses containing high levels of iron and manganese can serve as nutrients for the survival of some pathogens (Maharjan et al., 2016). This is problematic, especially for farms in rural areas that use well water and have the history of such water quality issues. While direct health concerns resulting from high iron content in poultry drinking water have not been reported (Fairchild et al., 2006; Hairston and Stribling, 1995), solid particulates like iron oxide from iron can accumulate in the water lines which may increase microbial biofilm formation (Prakash et al., 2003).

Poultry water lines are a major component of the poultry drinking water system, typically made of polyvinylchloride (PVC) material (Maharjan et al., 2017). The formation of microbial biofilm in the drinking water lines of poultry houses occurs gradually over time from the build-up of free-floating microbes (mostly bacteria), minerals, dirt, rust, and algae (Maharjan et al., 2016) providing a protective cover (slimy layer of extracellular polymeric matrix) for the pathogens against adverse environmental conditions inside the water lines (Szewzyk et al., 2000). Several studies have shown that bacteria can form biofilm in poultry drinking water systems (Buswell et al., 1998; Marin et al., 2009; Trachoo et al., 2002). The bacterial cells, instead of occurring freely in water, attach to the inner surface of pipes to form biofilms (Aboelseoud et al., 2021). More so, biofilm can harbor several microorganisms including Pseudomonas, Acinetobacter, Sphingomonas, and Klebsiella (Maes et al., 2019) and may favor multiplication of opportunistic microbes such as coliforms, Mycobacterium, Legionella and Aeromonas (Huang et al., 2018; van der Kooij et al., 2003). Consequently, the detachment of these pathogens into the drinking water system can result in contamination, which enhances the vulnerability of birds, particularly chicks to microbial challenges leading to poor flock productivity (Zimmer et al., 2003). Although the microbial load of drinking water is monitored regularly at the source on broiler farms and occasionally at the end of the water lines, not much has been done to assess the inside of drinking water pipelines for the presence and composition of biofilm, therefore information is still lacking in this area, proposing the need for microbial biofilm characterization (Van Eenige et al., 2013; Maes et al., 2019; Vermeulen et al., 2002). Such characterization could reveal information pertinent to establishing sanitation strategies for biofilm removal and its potential impact on flock health, particularly to enhance food safety.

Furthermore, the microbiome of different units of the poultry house environments has a key role to play in production management issues, yet little is known about the total microbial community within the broiler houses (Crippen et al., 2021). Previous studies have reported that the genera associated with broiler guts are usually included in the bacterial component of the litter microbiome. For example, Lactobacillus, Escherichia, Bacteroides, and Brevibacterium have been found in broiler litter and are common members of the broiler gut microbial community (Bucher et al., 2020; Cressman et al., 2010; Dumas et al., 2011; Enticknap et al., 2006; Lovanh et al., 2007; Lu et al., 2003). It is still unclear what constitutes the microbial ecology of the biofilm of poultry drinking water systems due to the difficulty of accessing the internal surface of the water lines, therefore most microbial diversity studies in poultry drinking water systems are limited to bulk water samples (Pinto et al., 2012). Although these studies provide a very good resource for the background transmission of microorganisms, they do not convey information regarding biofilm (Douterelo et al., 2014). To that end in our previous study (Ogundipe et al., 2024), we reported a novel method to visualize the biofilm inside the poultry drinking water lines and developed a modified protocol to collect the biofilm for further analysis.

Currently, research studies on the microbial ecology of biofilm of drinking water systems are insufficient and limited. Consequently, it is critical to develop research under representative conditions to determine the microbial ecology of poultry house environments as this could provide useful information to improve flock health. Therefore, two objectives were designed for this study: 1) to characterize the microbial ecology of the litter environment and biofilm in poultry water lines and 2) to investigate the impact of historic water quality issues on the microbial composition of the biofilm. To achieve this, we sampled two farm groups with well water source, one composed of five farms with a history of normal sulfur-iron water quality (NSIF) and the other five with a history of high sulfur-iron water quality (HSIF), over three months. Litter and water line (biofilm) samples were collected, and their microbiome profiles, including Salmonella incidence and quantity were analyzed to characterize their compositional differences.

Materials and methods

Farm demographics and operational practices

Ten commercial broiler farms were visited for this study (Table 1). These farms were divided into two groups of five farms, each based on the history of water quality issues obtained from the integrator. One group had a history of normal sulfur-iron (NSIF) content and the other had high sulfur-iron (HSIF) content water quality, although this was not observable in the mineral analysis of these farms water at the time of sampling (Table 2). The commercial farms were located within a radius of 60 miles of each other and operated an intensive poultry production system. Each farm had about four to six houses with at least four water lines per house. Within each farm, three houses were randomly selected, and, in each house, boot swab samples were collected from the litter and biofilm swab samples from the water lines. The water lines were constructed of PVC pipe with an internal diameter of 1.90 cm (about 0.75 in). Other farm demographic information has been published in Ogundipe et al. (2024).

Table 1.

Farm demographic.

Water quality	Farm	Age of flock	Water source	Sanitation	Sanitizer
NSIF	1	34	Well	Unknown	Peroxide
	2	33	Well	Unknown	Chlorine Dioxide
	3	31	Well	Unknown	None
	4	32	Well	Unknown	Chlorine Dioxide
	5	32	Well	Unknown	Chlorine Dioxide
HSIF	1	33	Well	Unknown	Chlorine Dioxide
	2	34	Well	Unknown	Peroxide
	3	32	Well	Unknown	Chlorine Dioxide
	4	33	Well	Unknown	Chlorine
	5	34	Well	Unknown	Chlorine Dioxide

Age of flock means age at sampling and sanitation means the sanitation practice at time of sampling. Sanitizer is the antimicrobial sanitizer the growers use at each farm historically.

Table 2.

Mineral analysis of water from farms.

Water quality	Farm	Minerals (ppm or mg/l)
		Cal	Mag	Fe	Man	Cl	Na	SO2-4	S	Pb	Cu	Zn	pH
NSIF	1	28.1	7.3	0.0	ND	51.0	ND	7.0	2.3	< 0.01	< 0.01	0.0	6.8
	2	30.5	9.8	0.0	ND	124.5	38.8	39.3	13.1	ND	0.0	0.0	6.9
	3	24.1	9.8	0.0	ND	4.6	6.6	12.1	4.0	ND	ND	0.0	6.6
	4	31.6	12.2	0.0	ND	269.2	70.7	32.4	10.8	ND	0.2	0.5	6.8
	5	29.1	9.3	ND	ND	132.9	35.0	14.8	4.9	< 0.01	0.0	0.1	6.8
HSIF	1	37.5	10.0	0.0	ND	100.6	30.6	13.2	4.4	< 0.01	0.0	0.1	7.0
	2	23.2	10.1	ND	ND	5.3	10.8	19.0	6.4	ND	ND	0.0	6.9
	3	31.0	12.2	0.0	ND	123.3	32.3	47.2	15.8	ND	0.0	0.0	6.7
	4	25.2	8.9	0.0	ND	0.6	6.9	47.7	15.9	ND	ND	0.1	6.6
	5	21.1	7.5	0.0	ND	40.7	ND	4.1	1.4	< 0.01	ND	0.0	6.7
Unacceptable Level		121.0	137.5	0.3	0.1	165.0	165.0	220.0		0.1	0.7	1.7	8.0

Abbreviations: ND, none detected; Cal, Calcium; Mag, Magnesium; Fe, Iron; Man, Manganese; Cl, Chloride; Na, Sodium; SO2-4, Sulfate; S, Sulfur; Pb, Lead; Cu, Copper; Zn, Zinc.

Sample collection

Litter samples were collected using two pairs of boot swabs to walk between the feeder and water lines on the right and left sides of the house. The biofilm determination and sample collection in the water lines were previously described in Ogundipe et al. (2024). Briefly, a borescope (Teslong TD500, Teslong, Irvine, CA) was used to view the inside of the water lines to confirm the presence and magnitude of biofilm, after which a sterile gauze was attached to a sterile 24-inch hemostat, long enough to swab the biofilm on the surface of the PVC pipe covering 12 to 18 inches inside the water lines. Four water lines (two at the fan end and the cool cell end) were swabbed in each house using one sterile hemostat and gauze per water line. Collectively, six litter and 12 biofilm swabs were collected for each farm visit, and a total of 60 litter and 120 biofilm samples were collected for analysis in this study. All samples at each farm visit were placed in sterile Whirl-Pak bags and transported to the lab in a cooler with icepacks for immediate microbial analysis.

Salmonella isolation, quantification, and identification

A 200 mL buffered peptone water (BPW) pre-enrichment broth was added to each litter sample in Whirl-Pak bags and mixed in a stomacher at 230 rpm for 30 s. The biofilm swabs were removed from the Whirl-Pak bags into 50 mL tubes containing 30 mL BPW broth to allow the full immersion of the swabs in the broth. The swabs were vortexed vigorously for two minutes to release the microbial content into solution and aliquots were taken from all samples into separate containers for microbiome analysis.

Pre-warmed MP media was added to the samples at 1:1 mixture and then placed in the incubator at 42°C for 10 h for initial Salmonella recovery. After incubation, the AOAC 081201 protocol for quantification of Salmonella using the BAX® System SalQuant™ (Hygiena, Camarillo, CA, USA) was followed and Salmonella lysates were prepared for all the samples accordingly for BAX real-time Salmonella assay. Following quantification, the samples were further incubated at 42°C to complete the 24 hours of pre-enrichment for prevalence testing. After pre-enrichment, 1 mL of each pre-enriched culture were inoculated into 10 mL each of Rappaport Vassiliadis (RV) and Tetrathionate (TT) broth at 37°C for 20-24 h. An aliquot of all enriched broths was streaked onto Xylose-Lysine-Tergitol 4 (XLT4) for Salmonella isolation. All plates were incubated at 37°C for 24 h. Colonies on XLT4 plates presumed to be Salmonella were selected and purified, and at least two colonies were picked for confirmation using serum agglutination with BD Difco™ Salmonella Antiserum Poly A – I & Vi (BD Difco, Fisher Scientific, Hampton, NH). The confirmed isolates were serogrouped with O antigen serum (BD Difco, Fisher Scientific, Hampton, NH) as follows: O:4 (B), O:7 (C₁), O:8 (C_2—C₃), O:9,46 (D₁-D₂), O:13 (G), and O:35 (O), all isolates from different serogroups were sub-cultured on Tryptic Soy Agar (TSA) plates and further serotyped.

DNA extraction and sequencing

The Zymo Quick-DNA Miniprep Plus Kit (Zymo Research, Irvine, CA, USA) was used to extract genomic DNA (gDNA) from sample pellets according to the manufacturer’s instructions. The quality of the extracted DNA was examined using 0.8 % w/v agarose gel electrophoresis, and the gDNA purity and concentrations were determined using a NanoDrop One spectrophotometer (Thermo Scientific, Wilmington, DE, USA). The full-length 16S rRNA was amplified using 20 ng gDNA template and 16S full-length primers 16S Forward (Sequence: 5′- CGGTTACCTTGTTACGACTT-3′) and 16S Reverse (Sequence: 5′- AGAGTTTGATCCTGGCTCAG-3′). The PCR reaction included 12.5 µL of Phusion High-Fidelity PCR Master Mix (Thermo Fisher Scientific, Waltham, MA, USA), 1.0 µL of each 10 µM forward and reverse primer, 1.0 µL of normalized gDNA (20 ng/ μL), and nuclease-free water in a total volume of 25 µL. The reaction started with 2 min at 98°C followed by thirty-five cycles of 98°C for 10 s, 60°C for 15 s, and 72°C for 1 min. The final extension lasted 5 min at 72°C. The yield and quality of the PCR product were then assessed using 1 % w/v agarose gel electrophoresis and subsequently purified with AMPure XP beads (Beckman Coulter, Brea, CA, USA). The cleaned amplicons were quantified using a NanoDrop One spectrophotometer (Thermo Scientific, Wilmington, DE, USA), end-prepped using the NEBNext Ultra II End Repair/dA-Tailing Module (New England Biolabs, Ipswich, MA, USA), and then natively barcoded using the Native Barcoding Kit (EXP-NBD196; Oxford Nanopore Technologies, Oxford, UK) according to the manufacturer’s instructions. The barcoded samples were pooled together and ligated with a Nanopore sequencing adaptor (Adaptor Mix II) using the Ligation Sequencing Kit (SQK-LSK109; Oxford Nanopore Technologies, Oxford, UK), followed by sequencing on a Flongle flow cell using the Nanopore MinION sequencer (Oxford Nanopore Technologies, Oxford, UK) for 24 h.

Bioinformatic and statistical analyses

Guppy (v4.3.4) with the Super-accurate model was used for basecalling, generating FASTQ files (Wick et al., 2019). Full-length 16S rRNA gene amplicons were classified using EMU software (v3.2.0) with default settings (Curry et al., 2022). Alpha diversity was analyzed using Chao1 and Shannon indices (vegan R package) (Oksanen et al., 2020). The Wilcoxon test followed by Dunn tests with Benjamini-Hochberg correction (FSA and companion packages) assessed differences in alpha diversity. Relative abundance differences in bacterial genera and species were examined similarly. Beta diversity was calculated using Bray-Curtis dissimilarity (vegan package) and visualized with Principal Coordinate Analysis (PCoA) at the species level (Anderson, 2001). Permutational multivariate analysis of variance (PERMANOVA, adonis function, vegan package) analyzed microbial community differences between sample types, with Benjamini-Hochberg correction for multiple comparisons (Anderson, 2005). Visualization was done using ggplot2.

Results

Salmonella incidence, quantity, and identification

Salmonella incidence and quantity were determined for the 10 commercial broiler farms visited in this study. A total of 58/60 bootswab (litter) and 5/120 biofilm samples were positive for Salmonella with an incidence rate of 96.67 % and 4.17 % and an average quantity of 4.07 and 0.50 Log₁₀ CFU/sample, respectively (Table 3). Collectively, there were no differences in Salmonella incidence and quantity of the litter between the NSIF and HSIF groups (P > 0.05). However, individual differences were observed in the Salmonella quantity of the litter of the farms within the HSIF group (Fig. 1) (P = 0.0001). For the biofilm samples, Salmonella was detected exclusively in Farm 1 of the HSIF group (HSIF1) with an average quantity of 0.50 Log₁₀ CFU/sample (Fig. 1).

Table 3.

Salmonella incidence and average quantity (± standard error of mean) of litter and biofilm samples from 10 commercial broiler farms.

Group	Sample type	No. of positives/total	Salmonella incidence a (%)	Salmonella quantity b (Log CFU/sample)
NSIF	Litter	30/30	100	3.64 ± 0.40
	Biofilm	0/60	0	-
HSIF	Litter	28/30	93.33	4.58 ± 0.44
	Biofilm	5/60	8.33	0.50 ± 0.78
Both	Litter	58/60	96.67	4.07 ± 0.30
	P value		0.15	0.12
Both	Biofilm	5/120	4.17	0.50 ± 0.78
	P value		0.02	-

NSIF, normal sulfur-iron farms. HSIF, high sulfur-iron farms.

^aData was analyzed using Chi-square test.

^bData was analyzed using Least squares test. Significant difference (P < 0.05).

Fig. 1.

Salmonella quantity of the litter and biofilm of each farm in the NSIF (normal sulfur-iron) and HSIF (high sulfur-iron) groups. Overall, no significant difference was detected in the Salmonella quantity of the NSIF and HSIF litter (P = 0.12) even though the Salmonella quantity of the farms within the HSIF group differed (P = 0.0001). Salmonella was only detected in the biofilm of farm 1 of the HSIF group (HSIF1). The quantity was determined by BAX PCR analysis in Log₁₀ CFU/sample. Distinct letters above boxes indicate significant differences.

A total of 40 Salmonella isolates were recovered with 22 from the NSIF group and 18 from the HSIF group. The isolates represented diverse serogroups, including B (31.8 %, 7/22), C₁ (4.5 %, 1/22), C_2—C₃ (31.8 %, 7/22), D₁ (27.3 %, 6/22), and O (4.5 %, 1/22) in the NSIF group; and B (5.6 %, 1/18), C₁ (27.8 %, 5/18), C₂ (11.1 %, 2/18), D₁ (44.4 %, 8/18), D₂ (5.6 %. 1/18), and G (5.6 %, 1/18) in the HSIF group. These were further serotyped, and the serotypes were mostly comparable between the NSIF and HSIF groups, with both groups containing serotypes Typhimurium (B), Infantis (C₁), Kentucky (C₂), and Enteritidis (D₁). However, Alachua (O), Schwarzengrund (B), and Agona (B) were unique to the NSIF group, while Ouakam (D₂) and Worthington (G) were exclusive to the HSIF group (Fig. 2).

Fig. 2.

Distribution of Salmonella serogroups in the litter of the NSIF (A) and HSIF (B) groups. Isolates were grouped according to O antigen serum testing. Each group had varying percentages of similar serogroups except for serogroup O in NSIF group, and serogroups D₂ and G in HSIF group. NSIF serogroups were represented by seven serotypes, including Agona, Alachua, Enteritidis, Infantis, Kentucky, Typhimurium, and Schwarzengrund, whereas the HSIF serogroups had six serotypes, including Enteritidis, Infantis, Kentucky, Ouakam, Typhimurium, and Worthington.

Alpha diversity

Alpha diversity denotes the relative abundance and number of species found within a specific community or ecosystem. The alpha diversity of the litter (A) and biofilm (B) of the farms within the NSIF and HSIF groups is represented in Fig. 3. The alpha diversity indices observed were richness, Chao1, evenness, and Shannon. Richness represents the total number of different species of microorganisms present in the sample while evenness measures the balance of the abundance of each species present. The Shannon index considers both species richness and evenness, and Chao1 emphasizes the presence of rare species by accounting for the ratio of singletons to doubletons in the sample.

Fig. 3.

Alpha diversity metrics of the litter (A) and biofilm (B) of individual farms (F1-F5) within the NSIF (normal sulfur-iron) and HSIF (high sulfur-iron farms) groups. The NSIF litter has lower alpha diversity than the HSIF litter (P = 0.04) while the NSIF biofilms had higher alpha diversity than the HSIF biofilms (P = 0.007).

The litter of the NSIF group had a significantly low microbial diversity compared to the HSIF group (P < 0.05) in terms of species richness and abundance (P = 0.004) and distribution of the species population (P = 0.007). On the other hand, the biofilm of the NSIF group had significantly higher microbial population than the HSIF group (P < 0.05), with differences observed in the abundance and rarity of species in both groups (P < 0.05) including the evenness of the species distribution (P = 0.02). Data was analyzed using the Kruskal-Wallis test.

Beta diversity and relative abundance

The variations in the microbial community composition between the farms of the NSIF and HSIF groups based on the litter (A) and biofilm (B) were evaluated by Bray-Curtis distances and visualized using Principal Coordinates Analysis (PCoA) as shown in Fig. 4. The PCoA plot showed clear separation of the litter of the NSIF and HSIF group, indicating prominent differences in their microbial composition (P = 0.0001). Similarly, the microbial community of the biofilm of the NSIF group differed significantly from that of the HSIF group (P = 0.0001), suggesting the influence of historic water quality dynamic on biofilm microbiome. The axes 1 and 2 of the plot annotated the percentage of variances seen in the samples with axis 1 explaining the most variance observed. Data was analyzed using PERMANOVA.

Fig. 4.

Principal coordinates analysis plot of the litter (A) and biofilm (B) of the NSIF (normal sulfur-iron farms) and HSIF (high sulfur-iron farms) groups. Distinct clusters of samples show significant differences in the microbial community structure between the two farm groups (P < 0.05).

Microbial community dynamics

Firmicutes dominated the litter of both farm groups, however, the NSIF group had detectable presence of Proteobacteria (Fig. 5). The relative abundance of the top microorganisms with higher proportions is displayed at the genus (Fig. 6) and species level (Fig. 7). At the genus (species) level, Lactobacillus (L. crispatus, L. johnsonii, L. reuteri, and L. salivarius), Aerococcus (A. urinaeequi and A. viridans), and Staphylococcus (S. cohnii, S. lentus, S. nepalensis, and S. simulans) were the most abundant genera accounting for about 50 % of the total population of the litter microbial community in both NSIF and HSIF groups. In addition, Anaerococcus, Atopostipes, Enterococcus, Facklamia, Jeotgalicoccus (J. halotolerans), Salinicoccus, Streptococcus, and Weissella (W. jogaejeotgali) were common to both farm groups. Notably, Bacillus (B. licheniformis), Brochothrix (B. thermosphacta), and the distinctive genus Escherichia (E. coli) under the phylum Proteobacteria, were also found in the NSIF group, while Peptostreptococcus (P. russellii) was the only addition to the HSIF group.

Fig. 5.

Top phyla in the litter and biofilm of the NSIF (normal sulfur-iron farms) and HSIF (high sulfur-iron farms) groups.

Fig. 6.

Top genera in the litter (A) and biofilm (B) of individual farms (F1-F5) within the NSIF (normal sulfur-iron farms) and HSIF (high sulfur-iron farms) groups.

Fig. 7.

Top species of the litter (A) and biofilm (B) of individual farms (F1-F5) within the NSIF (normal sulfur-iron farms) and HSIF (high sulfur-iron farms) groups.

Furthermore, the biofilm samples from the two farm groups had both phylum Firmicutes and Proteobacteria, with the majority being Firmicutes (over 75 % in NSIF and about 98 % in HSIF) and the remaining consisting of Proteobacteria (around 25 % and 2 % in NSIF and HSIF group, respectively). In the biofilm of the NSIF group, the phylum Proteobacteria was primarily represented by Escherichia (E. coli), Enterobacter and Cupriavidus (C. metallidurans). In contrast, Escherichia (E. coli) appeared to be the only genus from Proteobacteria in the biofilm of the HSIF group. Bacillus was the most prominent Firmicutes accounting for about 62.5 % and over 70 % of the total microbial population in the biofilm of the NSIF and HSIF groups, respectively. Additionally, Enterococcus and Lysinibacillus (Lysinibacillus spp. YS11) of the phylum Firmicutes were common to both groups. Furthermore, Paraclostridium and Peptostreptococcus (P. russellii) were also detected in the NSIF group. However, in the biofilm of the HSIF group, Staphylococcus (S. cohnii, S. lentus, S. nepalensis, and S. simulans) was the second most prominent genus followed by Aerococcus (A. urinaeequi and A. viridans), under the phylum Firmicutes, accounting for almost 25 % and about 4 % of the total microbial community, respectively. Interestingly, the species level revealed Bacillus licheniformis and Bacillus cereus as the predominant species in the biofilm of the NSIF and HSIF groups, respectively.

Discussion

This study observed a 96.7 % incidence rate of Salmonella in the litter of 10 commercial broiler farms housing birds aged 30 to 40 days, with an average concentration of 4.1 Log₁₀ CFU/sample. This finding emphasizes the significant incidence of Salmonella in poultry litter. Previous studies conducted over the years have reported high variability of Salmonella prevalence in poultry litter at different phases of production (Dunn et al., 2022; Gutierrez et al., 2020; Payne et al., 2006; Roll et al., 2011). For instance, a 2006 survey of three commercial broiler farms (two houses per farm) in North Carolina detected Salmonella with a concentration range of ≤ 1.0 to 3.6 Log MPN/g in 50 % of the 24 litter samples collected during four to six weeks of production (Payne et al., 2006). Berghaus et al. (2013) recorded a 68.2 % incidence rate and a concentration of 2.3 Log₁₀ MPN in boot swabs from 50 commercial broiler flocks aged 36 to 43 days. Conversely, Lu et al. (2003) failed to detect Salmonella in the litter of four poultry farms with birds ranging from a few days old to six weeks old, where litter had not been removed between flocks. These variabilities may be due to numerous factors, including geographic location, litter type, bird species, bird age, temperature, and environmental/management practices (Dunn et al., 2022). The high incidence of Salmonella in the litter observed in this study could be attributed to factors such as inadequate sanitary measures and the role of contaminated litter as a significant reservoir for pathogens such as Salmonella from fecal shedding of colonized birds. (Brooks et al., 2016; Hubbard et al., 2020; Kubasova et al., 2022).

The incidence of Salmonella in water line biofilm in this study was 4.2 %, with an average quantity of 0.5 Log₁₀ CFU/sample. This is a novel finding, as previous research has primarily focused on the prevalence of Salmonella in drinking water rather than biofilms within water lines. While the overall incidence of Salmonella in poultry water lines may not always be significant, ranging from 0 % to 11 % as reported by various studies (Alali et al., 2010; Bailey et al., 2001; Liljebjelke et al., 2005; Wang et al., 2023), the ability of Salmonella to produce biofilm and reside within a biofilm matrix in the water supply systems on poultry farms represents a potentially problematic source of contamination. Biofilms can provide a protective environment for Salmonella, allowing survival and persistence despite rigorous sanitation efforts (Liu et al., 2023; Obe et al., 2024). Our findings highlight the need for further research to explore the mechanisms of biofilm formation by Salmonella and other microorganisms within the drinking water lines to develop targeted interventions to reduce the risk of microbial contamination in the poultry drinking water system, thereby enhancing poultry health and food safety.

The top Salmonella serotypes identified in this study were Kentucky, Enteritidis, Typhimurium, and Infantis, while the less frequent serotypes included Schwarzengrund, Agona, Alachua, Ouakam, and Worthington. According to the recent U.S. Department of Agriculture - Food Safety Inspection Service (USDA-FSIS) food safety key performance indicator (KPI), serotypes Enteritidis, Typhimurium, and Infantis are the most prominent serotypes associated with human illness (USDA, 2022). These serotypes were present in both farm groups in this study. Serotype Enteritidis was the most abundant serotype, accounting for 27.3 % and 44.4 % of the isolates in the NSIF and HSIF groups, respectively. Serotype Typhimurium and Infantis exhibited varying prevalence: 31.8 % and 4.5 % in the NSIF group, and 5.6 % and 27.8 % in the HSIF group, respectively. Previous studies have reported Enteritidis and Typhimurium as the most studied serotypes due to their association with foodborne outbreaks (Finstad et al., 2012; Lamas et al., 2018; Park et al., 2014; Punchihewage-Don et al., 2022). These two serovars have shown the broadest distribution in poultry samples across the Americas (Diaz et al., 2022). In 2018, the European Food Safety Authority (EFSA) identified Salmonella Infantis as the most frequent serotype in broiler flocks, attributing this to the rapid spread of an emergent clone carrying a large mega plasmid (pESI) with the extended-spectrum beta-lactamase gene (Tyson et al., 2021). Serotype Kentucky was found in 31.8 % and 11.1 % of isolates in the NSIF and HSIF groups, respectively. A systematic review of poultry samples from various regions in America revealed Salmonella Kentucky and Heidelberg as the most common serotypes in environmental samples, including litter, feed, and water (Diaz et al., 2022). However, serotype Heidelberg was not detected in our study, this may be due to the methodology used, where a few colonies were randomly picked for serotyping, which might have led to some serotypes being overlooked in mixed Salmonella populations (Siceloff et al., 2022). The review also mentioned that serotypes such as Agona, Alachua, Ouakam, Schwarzengrund, and Worthington were present but at lower prevalence, and all these serotypes were associated with at least one antimicrobial resistance (AMR). These findings stress the importance of monitoring and controlling Salmonella at the pre-harvest stage of poultry production to mitigate the risk of food-borne outbreaks and address antimicrobial resistance.

Some studies have indicated that poultry litter contains a dynamic and complex microbiome, which is majorly composed of the bacteria from the gut and poultry house environment depending on the litter management protocols (Crippen et al., 2021; Lovanh et al., 2007; Lu et al., 2003). In this study, Firmicutes dominated the NSIF and HSIF litter with similar bacteria composition of Lactobacillus, Aerococcus, Staphylococcus, Anaerococcus, Atopostipes, Enterococcus, Facklamia, Jeotgalicoccus, Salinicoccus, Streptococcus, Weissella, and Bacillus. However, the NSIF litter also contained Proteobacteria primarily composed of Escherichia. Beta diversity further revealed differences in the bacteria community structure. These results align with the findings of Kubasova et al. (2022) that reported similar bacterial composition of the litter during the first 2 months of broiler production and highlighted that the abundance of E. coli and Lactobacilli during the first month probably originated from the chicken intestinal tract. Several studies also stated that as the litter environment develops, Firmicutes from the orders Lactobacillales (including Aerococcus, Facklamia, Enterococcus, and Lactobacillus), Staphylococcales (Staphylococcus, Jeotgalicoccus, and Salinicoccus), Actinobacteria from the orders Corynebacteriales (Corynebacterium or Dietzia), Micrococcales (Brevibacterium or Brachybacterium), and Bacteroidetes from the order Sphingobacteriales (Sphingobacterium) becomes the most frequent litter microbiota members (Bucher et al., 2020; Cressman et al., 2010; De Cesare et al., 2020; Johnson et al., 2018; Lu et al., 2003; Oakley et al., 2013; Torok et al., 2009; Valeris-Chacin et al., 2021; Wang et al., 2016). Moreover, Zwirzitz et al. (2023) observed that the microbial community in the litter constantly changed throughout the grow-out period and the bacterial beta diversity in the litter was significantly influenced by the age of the broilers. The lack of Salmonella 16S rRNA gene reads in the litter samples was unexpected but this could mean that Salmonella was in low abundance in the microbial population. Also, we observed a relative abundance of beneficial probiotic bacteria, such as Bacillus licheniformis (about 12.5 % in NSIF) and Lactobacillus (around 25-27 % in both NSIF and HSIF) in the litter communities which have been demonstrated to positively impact microbial decontamination in spent litter (De Cesare et al., 2019).

Microbial build-up in drinking water systems often leads to biofilm formation, and this has been detected in the water lines of poultry houses on commercial broiler farms (Ogundipe et al., 2024). Although bacteria are the most studied microorganisms in biofilms, our previous study revealed the presence of aerobic bacteria, Enterobacteriaceae, and yeasts and molds within the biofilm of poultry drinking water systems (Ogundipe et al., 2024). Additionally, this follow-up study demonstrated that the microbial composition of the biofilm of poultry drinking water system differed when characterized based on historic water quality issues of the farms. According to the integrator, some of the farms are located in areas with high iron and/or sulfur levels in the well water and this sometimes contributes to algae bloom, affecting the drinking water quality. Although this was not evident in the water quality analysis we performed because growers could resolve this water quality issue by various means including the addition of an iron filtration system. The NSIF biofilms exhibited greater bacterial diversity which was mostly composed of ≥ 60 % Bacillus and 6 % Escherichia. In contrast, the HSIF biofilms showed reduced bacterial diversity dominated by almost 70 % Bacillus and around 23 % Staphylococcus. Beta diversity also indicated the microbiome of the biofilm of the two farm groups were significantly different. Interestingly, the species level revealed the dominance of Bacillus licheniformis and Bacillus cereus in the NSIF and HSIF biofilms, respectively. This finding suggests that the water quality history of the farms may have influenced the microbial population of the biofilm of the poultry drinking water lines. Most microbiome studies on drinking water systems have focused on drinking water alone rather than the biofilms in the water system as the source of microbial contamination, making this study one of the few attempts to characterize the microbiome of the biofilms in poultry water lines, thus paving the way for further research. Few studies corroborate these findings, for instance, Rychlik et al. (2023) identified a variety of bacteria in drinking water microbiomes, including Gammaproteobacteria and genera such as Escherichia, Acinetobacter, Pseudomonas, Flavobacterium, Chryseobacterium, Sphingobacterium (phylum Bacteroidetes), and Sulfurospirillum (phylum Campylobacterota), along with phylum Firmicutes (Lactobacillus aviarius and Veillonella magna) and Alphaproteobacteria. However, Goraj et al. (2021), in a seven-year study of a constructed drinking water system model found the genera Acinetobacter, Legionella, Enterobacter, Mycobacterium, Pseudomonas, and Staphylococcus, with a predominance of Proteobacteria in the biofilm on the inner surface of the constructed model (made of PVC and cast iron). Furthermore, Liu et al. (2020) explained that sampling strategy can significantly influence the resulting biofilm bacterial community in the drinking water distribution pipe and proposed a consistent multi-sectional swabbing strategy for future drinking water biofilm sampling. This approach may be more effective with the recently published modified swabbing technique described in our previous study (Ogundipe et al., 2024). Additionally, the materials used in drinking water piping systems, such as asbestos cement, cast iron, polyvinylchloride, polyethylene, stainless steel, and concrete, can influence biofilm formation (Liu et al., 2017). While drinking water might not be the preferred niche for most microorganisms, their presence suggests contamination from other sources where they are most abundant like the litter, feces, dust, and aerosol (Bindari et al., 2021; Chen et al., 2021), posing potential infection risks to the chicken intestinal tract and raising food safety concerns (Rychlik et al., 2023).

Conclusion

Although the water quality history of the farm groups did not significantly influence the overall Salmonella incidence in the litter and biofilm, the differences in the microbial ecology, particularly in the biofilm of the two farm groups highlight the importance of improved management of the poultry drinking water system. The presence of Bacillus species, probiotic (B. licheniformis) and pathogenic (B. cereus) in the NSIF and HSIF groups, respectively, is noteworthy and suggests the need to develop a model to strategically control biofilms in the poultry drinking water lines. To that end, further studies exploring the potential of biological agents like B. licheniformis and others are critical to developing these controls. Using these biological agents would also be of significant benefit to bird health as many probiotics have been shown to improve bird performance and intestinal microbial population.

CRediT authorship contribution statement

Tolulope T. Ogundipe: Data curation, Formal analysis, Investigation, Writing – original draft. Samantha Beitia: Methodology, Resources. Li Zhang: Data curation, Formal analysis, Methodology, Writing – review & editing. Xue Zhang: Data curation, Formal analysis, Methodology, Writing – review & editing. Tomi Obe: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Validation, Writing – original draft, Writing – review & editing.

Disclosures

The authors declare no conflict of interest.