Real-time quantitative PCR assay development and application for assessment of agricultural surface water and various fecal matter for prevalence of Aliarcobacter faecis and Aliarcobacter lanthieri

Background Aliarcobacter faecis and Aliarcobacter lanthieri are recently identified as emerging human and animal pathogens. In this paper, we demonstrate the development and optimization of two direct DNA-based quantitative real-time PCR assays using species-specific oligonucleotide primer pairs derived from rpoB and gyrA genes for A. faecis and A. lanthieri, respectively. Initially, the specificity of primers and amplicon size of each target reference strain was verified and confirmed by melt curve analysis. Standard curves were developed with a minimum quantification limit of 100 cells mL− 1 or g− 1 obtained using known quantities of spiked A. faecis and A. lanthieri reference strains in autoclaved agricultural surface water and dairy cow manure samples. Results Each species-specific qPCR assay was validated and applied to determine the rate of prevalence and quantify the total number of cells of each target species in natural surface waters of an agriculturally-dominant and non-agricultural reference watershed. In addition, the prevalence and densities were determined for human and various animal (e.g., dogs, cats, dairy cow, and poultry) fecal samples. Overall, the prevalence of A. faecis for surface water and feces was 21 and 28%, respectively. The maximum A. faecis concentration for water and feces was 2.3 × 107 cells 100 mL- 1 and 1.2 × 107 cells g− 1, respectively. A. lanthieri was detected at a lower frequency (2%) with a maximum concentration in surface water of 4.2 × 105 cells 100 mL− 1; fecal samples had a prevalence and maximum density of 10% and 2.0 × 106 cells g− 1, respectively. Conclusions The results indicate that the occurrence of these species in agricultural surface water is potentially due to fecal contamination of water from livestock, human, or wildlife as both species were detected in fecal samples. The new real-time qPCR assays can facilitate rapid and accurate detection in < 3 h to quantify total numbers of A. faecis and A. lanthieri cells present in various complex environmental samples.

various complex environmental niches. Non-viable or nonculturable cells of Gram-negative bacteria can potentially contaminate water by producing virulence-associated factors and toxins that can pose health risks to humans [17,18].
Real-time quantitative PCR (qPCR) assays have provided more rapid and robust tools to detect and quantify Aliarcobacter spp. in pure culture, fecal, hide, food, and complex environmental samples [19][20][21][22]. None of these developed real-time qPCR assays were capable of differentiating and quantifying A. faecis and A. lanthieri directly from environmental matrices, partly because of their unknown status and/or low abundance in these niches.
Therefore, it is necessary to develop fast and accurate methods for detecting these species in complex environmental matrices, since conventional methods are not always accurate measures for recovery and may fail to detect these species when prevalent at low concentrations and when competing with other Aliarcobacter spp. The main objectives of this study were to: i) develop and optimize species-specific direct real-time qPCR assays to quantitatively detect A. faecis and A. lanthieri in environmental niches; and ii) validate and apply these qPCR assays to detect, quantify, and assess the prevalence of A. faecis and A. lanthieri in agricultural surface water and fecal samples from human and animal sources.

Results
Optimization of species-specific real-time qPCR assays and development of standard curves Two novel real-time PCR assays were developed and optimized using A. faecis LMG 28519 and A. lanthieri LMG 28516 reference strains. The assays were further validated by applying to the field isolates of each target species (A. faecis: n = 29; and A. lanthieri: n = 10). The primers specifically amplified to their target sequences with expected melting peaks of 74°C for A. lanthieri and 79°C for A. faecis ( Fig. 1A and B) and typical amplicon sizes 152 bp and 72 bp, respectively ( Figure S1A and B). Moreover, no amplification signals were observed for any other Aliarcobacter spp. or other bacterial reference species and strains that could potentially occur in water and fecal matter ( Table 1).
The limit of detection for quantitative analysis of each optimized real-time PCR assay was determined by developing standard curves of reference strains of A. faecis and A. lanthieri DNA templates extracted from spiked water and dairy cow manure samples, in units of number of cells mL − 1 ( Fig. 2A and B) and cells g − 1 (Fig. 3A and B). Although a minimum of 10 cells mL − 1 or 10 cells g − 1 were also used for the quantitative assay, amplification was observed at ≥40 Cq value; therefore, Cq value ≥42 thresholds were considered as negative or indeterminate.
qPCR assay validation and application for detection and quantitation of A. faecis and A. lanthieri in agricultural surface water and fecal sources The qPCR assays were further validated and applied by analysing a total of 804 environmental (fecal and surface water) samples. Of the total 588 agricultural surface water samples, A. faecis was detected at a frequency of 21% (n = 124), while A. lanthieri (n = 13) was only detected in 2% of samples. Similarly, of the total 216 (human, n = 19; animals, n = 197) fecal samples, A. faecis (n = 61) was detected more commonly (28%) than A. lanthieri (10%; n = 22).
Further comparative analyses showed that the rate of A. faecis prevalence was significantly higher (p < 0.05) than A. lanthieri in agricultural sites (Table 2). Interestingly, only A. faecis (19% frequency), not A. lanthieri, was detected at the reference sampling site 24 (Table 2). Overall, the frequency of co-occurrence of these two target species was low and observed in only one single surface water sample, which was collected from an agricultural drainage ditch with upstream proximity to dairy livestock operations. Moreover, among the 11 agriculturally impacted sites, A. faecis was detected at a significantly (p < 0.05) higher frequency (> 20%) at sampling sites 5, 6, 10, 18 and 20 as compared to sampling sites 1,8,9,19,21 and 253 (< 20%). There was no significance (p > 0.05) difference in the occurrence of A. lanthieri among the sites.
Of the total 216 fecal samples collected from human and various animal fecal sources, 28% (n = 61) and 10% (n = 22) samples were positive for A. faecis and A. lanthieri, respectively. Among these different fecal samples, A. faecis was detected at higher frequencies in human, cat, cow, dog, and pig, compared to A. lanthieri which was detected at lower frequency (Table 3). Interestingly, only one fecal sample from chicken was positive for A. lanthieri whereas one fecal sample from sheep was positive for A. faecis . On the other hand, duck, goat, and pony fecal samples were negative for both target species. Similar to the water samples, a low frequency of co-occurrence of both species in only four (cow: n = 2; human: n = 1; pig: n = 1) fecal samples was observed. Additional comparative analysis showed, overall, no significant (p > 0.05) difference in the rate of prevalence of A. faecis and A. lanthieri between human and animal fecal samples was observed. Similarly, no significant difference between the rate of prevalence of A. faecis and A. lanthieri was found among human, cat, and dog fecal samples. However, a significantly higher frequency of occurrence (p < 0.05) of A. faecis than A. lanthieri was observed between cow and pig fecal samples.

Discussion
Conventional culture-based multiplex PCR assays for the detection of A. faecis and A. lanthieri, along with four other closely related Aliarcobacter spp., were developed by Khan et al. [15]. In the present study, we further established species-specific direct DNA-based realtime quantitative PCR assays to improve the detection method for rapid identification and quantification of total number of (viable and non-viable) cells of A. faecis and A. lanthieri in surface water and fecal samples. Each species-specific qPCR assay is rapid, sensitive, and reliable for quantitative analysis of A. faecis and A. lanthieri DNA. The assay has a reproducible detection limit per reaction with linear amplification over a wide range of seven to eight orders of magnitude. qPCR assays are less time-and labor-intensive than culture-based methods, and have minimum potential for cross-contamination; therefore, the assays developed here are more robust and useful in diagnostic and analytical settings, especially when the cells of the target species are present at low concentrations [23,24]. The other advantage is that these assays do not require post-PCR confirmation, and possess the ability to provide quick results which are more desirable for high-throughput studies [25,26]. In addition, the fluorescent dye SYBR Green was used in the developed assays, which is more cost-effective than fluorogenic probes. qPCR assays can also detect and quantify total (viable and non-viable) number of cells, which is important as the non-viable cells can generate human immunological responses despite these cells being incapable of causing infection. Therefore, the present qPCR assays we have developed allow quantitative detection of these species from complex environmental samples even when they are present at low levels.
To validate the newly developed assays, this study analyzed 588 water samples from an agriculturally dominated watershed and 216 samples from various fecal sources, and A. lanthieri and A. faecis were detected and quantified. Overall, we found that A. faecis was more    [27] where potential fecal inputs from adjacent farm lands and wildlife can occur readily due to tile drainage and surface runoff [28]. Levican et al. [29] found that cell counts for adhesion and invasion of different Aliarcobacter spp. were possible above the limits of 1.7 × 10 4 CFU mL − 1 and 1.7 × 10 2 CFU mL -1 , respectively. The cell concentrations of A. lanthieri and A. faecis that we detected here ranged as high as 10 7 cells 100 mL − 1 . Our findings are in congruence with a previous study [30] where a comparable range of concentration (2.0 × 10 5 to 1.2 × 10 9 cells 100 mL − 1 ) of Arcobacter spp. in various water sources was reported.
In order to compare the rate of prevalence of these species in agriculture and non-agricultural surface waters, site 24 was chosen as a reference site, as it is not impacted by any known direct anthropogenic activity [31]. However, A. faecis was detected at this site which suggests that there may be alternate sources of water contamination, possibly from wildlife. However, in previous studies human-specific bacterial markers were detected at site 24 [32,33]. Throughout the sampling period, among agriculturally dominated SNR sites, A. faecis was most frequently detected at sites 5, 6, 10, 18, and 20 that have dairy operations in the upstream vicinity (Table 3). Additionally, A. lanthieri   (Table 3), where dairy-based farming operations occur along the drainage ditch.
The prevalence of other microbial species in the SNR watershed has previously been examined, which add value in our capacity to detect A. lanthieri and A. faecis in the same study area. For example, Lyautey et al. [34] investigated the prevalence of Listeria monocytogenes, and also found that occurrence was associated with proximity to dairy farming operations. The authors found that sites 9 and 18 had the highest prevalence of L. monocytogenes. However, our results showed high frequency of A. faecis in site 18 compared to site 9. Frey et al. [35] detected Campylobacter spp. and Salmonella spp. at the same SNR watershed sites where cattle fecal markers were detected. A. lanthieri and A. faecis were originally isolated from human and fecal sources [4,5], and in this study both species were detected in human and livestock feces, as well as in agricultural surface water. This strongly indicates that contamination of water by fecal matter from livestock, particularly cattle, could be linked to the prevalence of A. lanthieri and A. faecis.

Conclusions
The qPCR assays designed here can accurately detect the prevalence and quantify the total number of cells of A. faecis and A. lanthieri in complex environmental niches. It is critical to develop alternative methods other than the widely-used culture-based techniques for the detection of gram-negative bacteria in environmental or clinical samples, as the presence of virulence, antibiotic resistance and toxin (VAT) genes can still pose a health risk even when cells are in a non-viable state. The study results suggest that routine quantitative testing of water sources for microbial contamination is important, especially in areas such as agricultural and urban communities where fecal contamination risks are higher. The developed assays could, therefore, provide rapid DNAbased tools for early and reliable detection of target species in field samples, which would help in improving water quality and intervention for reducing and eliminating the risk of contamination of A. faecis and A. lanthieri in aquatic sources.

qPCR assay development and optimization Bacterial species and culture conditions
For testing the specificity and sensitivity of primers and real-time qPCR assays for the detection and identification of A. faecis and A. lanthieri, two reference strains of A. faecis LMG 28519 and A. lanthieri LMG 28516, were used as positive controls (Table 1). Six other Aliarcobacter spp., nine species from genus Arcobacter, Haloarcobacter, Malacobacter and Pseudoarcobacter, and 50 other bacterial reference species and strains were used as negative controls (Table 1). In addition to the two LMG strains above, 29 A. faecis and 10 A. lanthieri cultures of our lab collection, isolated from various human and animal fecal and water samples, were used as positive controls. All control reference strains were grown on selective media according to appropriate aerobic and microaerophilic culture conditions. A. faecis and A. lanthieri strains were grown in Arcobacter media broth and incubated at 30°C under microaerophilic (85% N 2 , 10% CO 2 and 5% O 2 ) conditions with continuous shaking at 125 rpm. The DNA from pure cultures of reference strains and field isolates was extracted using a boiling method [36] where a single colony was suspended in 75 μL TE (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) buffer, boiled for 10 min and centrifuged. The supernatant containing DNA was quantified using a Qubit 3.0 fluorometer (Thermo Fisher Scientific, Waltham, MA, USA), transferred to a sterile tube, and stored at − 20°C for further PCR analysis.

Spiked assay for standard curve development and quantitation
A spiking experiment was carried out to develop standard curves using A. faecis LMG 28519 and A. lanthieri LMG 28516 reference strains to assess the purity of nucleic acid in terms of yield, concentration, reproducibility and removal of potential PCR-inhibitory compounds. The experiment also helped to quantify and measure the sensitivity (least number of cells mL − 1 ) of the qPCR assays. A. faecis and A. lanthieri cells were grown under microaerophilic conditions as described above. The cells were collected by centrifugation at room temperature and re-suspended in 1 mL TE buffer. The cell concentration mL − 1 of each target species was measured on modified Arcobacter Agar Medium (m-AAM; Oxoid) containing selective antimicrobial agents (cefoperazone, amphotericin B, and teicoplanin) and incubated under conditions as described above. The known quantity of A. faecis or A. lanthieri reference strain cells (10 8 cells mL − 1 ) was then simultaneously spiked and serially (10-fold) diluted from 10 8 to 10 1 cells mL − 1 in autoclaved agricultural watershed water and cow manure samples. Each spiked water sample with known cell concentration was filtered through a 0.22 μm sterile nitrocellulose filter.
Total genomic DNA was extracted from each spiked filter and 0.5 g manure sample with known cell concentration using DNeasy PowerSoil Kit (Qiagen; formerly MoBio PowerSoil DNA Isolation Kit) following the manufacturer's instructions. The purity and concentration of DNA was measured by Qubit 3.0 fluorometer and 1% agarose gel electrophoresis using 1X TAE (0.04 M Trisacetate, 0.001 M EDTA, pH 7.8) buffer.

Primer design and qPCR assay conditions
Real-time qPCR assays were developed and optimized for A. lanthieri by designing primer pairs from the variable region of the gyrase (gyrA) gene. The primers were designed based on alignment analysis of gyrA gene sequences of Aliarcobacter and other reference species and strains belong to other genera available in the Gen-Bank database. On the other hand, the real-time PCR assay for A. faecis was optimized by using primers from the rpoB gene encoding β-subunit of RNA polymerase previously designed by Khan et al. [15].
The reactions were run on a Lightcycler® 480 Instrument II (Roche, Indianapolis, IN, USA) with an initial denaturation at 98°C for 3 min followed by 50 cycles of denaturation at 98°C for 15 s, annealing temperatures of 58°C for A. faecis and 55°C for A. lanthieri for 30 s, and extension at 72°C for 30 s. The amplified product obtained from each cell number was confirmed by melt curve analysis where the melting peak was 79 and 74°C for A. faecis and A. lanthieri, respectively (Table 4). Due to expected small amplicon sizes, the amplified products were further confirmed on a 2% agarose gel matrix, stained (ethidium bromide 0.5 μg mL − 1 ) and visualized on a UV transilluminator using an Alpha Imager (Fisher Scientific) gel documentation system.
Validation and application of qPCR assays Study site description, and surface water and fecal sample collection The developed qPCR assays were further validated and applied to investigate the rate of prevalence and concentration of A. faecis or A. lanthieri cells in surface water and fecal samples. In order to assess the purity of total DNA in terms of removal of PCR inhibitors (such as humic acid, organic and inorganic compounds) and sensitivity of assays for quantitation of total number of cells, surface water samples were collected from the South Nation River (SNR) watershed, located near Ottawa, Ontario in eastern Canada [37]. The watershed covers an area of~3900 km 2 , of which approximately 60% is used for agricultural purposes, primarily related to dairy farming. A detailed description of the watershed and sampling sites have been previously reported by Wilkes et al. [37,38] and Lapen et al. [39] ( Table 2). For this study, a total of 12 sites of varying stream orders were selected for sampling, based on their proximity to agriculturallyimpacted areas. In addition, one site with no known upstream anthropogenic activity was selected as a reference site (Site 24; Edge et al.) [31] (Fig. 6). A total of 588 (from 2013 to 2018) surface water samples were collected on a bi-weekly basis between April and November. In addition, a total of 216 fecal samples from human (n = 19) and various animal (n = 197) sources including cat (n = 20); chicken (n = 8); cow (n = 68); dog (n = 18); duck (n = 1); goat (n = 4); pig (n = 75); pony (n = 2) and sheep (n = 1) were collected in the same region. The surface water and fecal samples were collected in sterile polypropylene bottles and bags, placed in coolers and delivered to Agriculture and Agri-Food Canada-Ottawa, Ontario Laboratory where the samples were processed within 24 h of their collection for microbiological analysis. Water samples were filtered through 0.22 μm sterile nitrocellulose filters. The DNA from filters and fecal samples were extracted using DNeasy PowerSoil Kit and quantified by Qubit 3.0 fluorometer.

Quantitation of A. faecis and A. lanthieri cell concentration in environmental sources
The two optimized real-time qPCR assays described above were validated, using the developed standard curves, by detecting and quantifying the total number (viable and non-viable) of A. faecis and A. lanthieri cells 100 mL − 1 from agricultural surface water and fecal samples. The specificity and quality of amplified products were confirmed by analyzing and comparing the melting curves to the standard melting peaks obtained for A. faecis and A. lanthieri amplicons. In addition, the amplification quality was also validated by agarose gel electrophoresis using 100 bp DNA size marker (Thermo Fisher Scientific) (Fig. S2A&B). The gel was stained, visualized, and photographed as described in the preceding section.

Data analysis
McNemar Chi-square Contingency and Fisher's Exact tests were applied to compare the rate of prevalence and identify significant differences (p < 0.05) of A. faecis and A. lanthieri among different agricultural and non-agricultural sites, surface water and fecal samples using STATISTICA (StatSoft, Inc., 2013) [40].