ABSTRACT

There is increasing evidence that diversity changes in bacterial communities of beef cattle correlate to the presence of Shiga toxin–producing Escherichia coli (STEC). However, studies that found an association between STEC and bacterial diversity have been focused on preslaughter stages in the beef supply chain. This study was designed to test a hypothesis that there are no differences in bacterial diversity between samples with and those without the presence of the top 7 STEC (O26, O45, O103, O111, O121, O145, and O157) throughout processing in an integrated (abattoir A) and a fragmented (abattoir B) Australian beef abattoir. Slaughter and boning room surface samples from each abattoir were analyzed using 16S rRNA amplicon sequencing and tested for the top 7 STEC following the Food Safety and Inspection Service protocol. Potential positives through slaughter were similar between the abattoirs (64 to 81%). However, abattoir B had substantially reduced potential positives in the boning room compared with abattoir A (abattoir A: 23 and 48%; abattoir B: 2 and 7%). Alpha diversity between the sample groups was not significantly different (P > 0.05) regardless of different STEC markers. Nonmetric multidimensional scaling of slaughter samples showed that the bacterial composition in fecal and hide samples shared the least similarity with the communities in carcass and environmental samples. Surface samples from slaughter (carcass and environmental) and boning (carcass, beef trim, and environmental) all appeared randomly plotted on the scale. This indicated that the STEC presence also did not have a significant effect (P > 0.05) on beta diversity. Although presence of STEC appeared to correlate with changes in diversity of fecal and hide bacterial communities in previous studies, it did not appear to have the same effect on other samples throughout processing.

HIGHLIGHTS
  • Bacterial diversity and STEC were analyzed through two Australian beef abattoirs.

  • Similar alpha diversity was found in STEC potential positives and nonpotential positives.

  • The top 7 STEC markers did not correlate with beta diversity differences.

  • There was no correlation between the top 7 STEC testing status and bacterial diversity.

Beef cattle are identified as a natural reservoir of Shiga toxin–producing Escherichia coli (STEC), and beef products have been implicated in STEC infections and multiple foodborne outbreaks across the globe over the last few decades (3, 8, 14). Since then, stringent STEC screening procedures have been implemented in the Australian red meat industry to meet regulatory standards for beef export. The U.S. Department of Agriculture's Food Safety and Inspection Service (FSIS) lists seven serotypes (O26, O45, O103, O111, O121, O145, and O157, known as the top 7 STEC) as adulterants in ground beef (17). The regulatory procedure demands that raw nonintact beef products for the United States be deemed free of the top 7 STEC to gain and maintain access to the market.

Pathogenesis of STEC relies on attachment to host epithelial cells in the intestinal lining (eae genes) and release of the cytotoxic Shiga toxins (stx genes) (9, 16, 26, 28). Researchers have used these virulence genes to understand the STEC prevalence in, and subsequently the contamination of, beef in the supply chain (4, 5, 12, 13, 15, 25, 32, 33, 38). These studies typically relied on molecular analysis and cultivation of STEC using selective media. The findings contributed to the hypothesis that microbial contaminants transfer from fecally contaminated hides onto the carcasses and ultimately onto the beef product (2, 4, 12, 15, 30). More recently, high-throughput sequencing technologies and bioinformatic software have enabled researchers to understand STEC contamination of beef at resolutions that were not achievable with conventional microbiological methods (7, 18, 21, 23). Advances in high-throughput sequencing and population analysis techniques have given researchers the capability to investigate the ecological patterns in beef microflora that may indicate the presence of STEC within the community.

Numerous studies have explored bacterial diversity and prevalence of STEC in the beef supply chain between production and slaughter of the animals. Chopyk et al. (11) examined the association between STEC and corresponding differences in bacterial diversity on preharvest hides. The study reported that a correlation exists between prevalence of STEC serotypes and low bacterial community diversity in hide-specific cohorts of bacterial populations. It also showed that there were variations in bacterial community composition depending on combinations of the top 7 serotypes. Zhao et al. (42) found that high levels of STEC shedding were associated with lower fecal bacterial diversity, especially in younger calves with lower butyrate-producing bacteria in the lower gastrointestinal tract. In another study, local bacterial diversity differences were present along the digestive tract in supershedders and nonsupershedders, but no correlation was found between fecal bacterial diversity and O157 shedding status (41). A study by Xu et al. (39) demonstrated that fecal bacterial richness and diversity was higher in supershedders compared with nonsupershedders. These studies were able to describe the diversity differences observed in response to the presence of STEC by analyzing 16S rRNA amplicon sequences in combination with enumeration of STEC.

We recently reported bacterial community profiles from two Australian abattoirs in relation to indictor bacterial counts (18, 19). In the present study, we examine the metagenomic data from the two abattoirs in light of STEC and the influence of diversity on the top 7 STEC. Knowledge of the relationship between STEC and the structure of beef microflora during slaughter and boning of the animals is lacking. The present study used 16S rRNA amplicon sequencing, in combination with cultivation methods and PCR, to examine the fundamental changes in bacterial diversity through processing of beef cattle. The hypothesis is that there is no correlation between bacterial diversity and presence of the top 7 STEC; therefore, the emphasis of the analysis was placed on diversity differences between samples that tested positive and those that tested negative for STEC virulence markers. The findings from this study could contribute to understanding of the ecological changes that surrounds the top 7 STEC in the beef supply chain.

MATERIALS AND METHODS

Sample collection

Samples were collected from two abattoirs with differing levels of integration in the supply chain. An integrated abattoir (abattoir A) receives beef cattle that have been through the same production system, typically from a small number of sources. A fragmented abattoir (abattoir B) processes cattle that are provided from multiple producers. Abattoirs A and B were each visited four times over a period of 14 months. Each abattoir was visited twice for collection of slaughter samples between January and July 2018 and twice again for boning room samples between November 2018 and March 2019. Per visit, 90 samples were collected from the slaughter floor. The samples collected consisted of 10 fecal samples, 15 hide samples before hide-pull, 15 carcass samples immediately after hide-pull, 15 carcass samples postevisceration, 15 carcass samples immediately before chilling (prechill), and 20 environmental samples throughout the slaughter floor. For the boning room, 90 samples were collected over two sampling days. The first sampling day was spent collecting 15 prechill carcass samples. Samples collected on the second day consisted of 15 chilled carcasses, 5 forequarters, 5 hindquarters, 30 beef trim that were boned from the carcasses (manufacturing beef), and 20 environmental surfaces. In abattoir A, trim samples from forequarters (n = 12) and hindquarters (n = 12) and boxes of manufacturing beef (n = 6) were collected. In abattoir B, trim samples were collected from forequarters (n = 16) and hindquarters (n = 14). Carcass and beef trim samples were not matched, but the samples were collected from the same lot of animals.

Sampling method and preparation

Fecal samples were collected from internal content of freshly excreted fecal pats in holding pens before slaughter using sterile yellow-capped plastic jars (Sarstedt, Numbrecht, Germany) and a sterile spoon. A large-area sampling technique was used for sampling of the carcasses as previously described (10). The carcass samples were collected aseptically using the large-area sampling technique (3,000 cm2). Environmental samples were obtained from swabbing an area covering approximately 900 cm2. For beef trim, approximately 100 cm2 of the most outer fat surface was collected in a large stomacher bag. Swabbing of the surface samples was performed using a Whirl-Pak Speci-Sponge (Nasco, Fort Atkinson, WI) premoistened in 25 mL of sterile buffered peptone water (BPW; Oxoid, Basingstoke, Hampshire, UK). All samples were immediately placed on ice for transport to the laboratory. At the laboratory, samples collected with a sponge had an additional 75 mL of BPW added before being stomached for 30 s at eight strokes per second (Interscience, Saint Nom, France). Fecal samples were diluted 10-fold by adding 90 mL of BPW to 10 g of feces. Approximately 100 cm2 of the most outer layer of trim samples was excised and added to 100 mL of BPW. The samples were then stomached for 1 min. Following stomaching, aliquots from the samples were kept for 16S rRNA sequencing analysis at −80°C and the remaining portion of the bacterial-BPW suspension was used for detection of STEC.

Confirmation of the top 7 STEC

All samples were processed for the presence of the top 7 STEC following the export regulatory guideline for meat products by the FSIS (35). The guideline evolved around detection of stx, eae, and gene fragments that were specific for the top 7 serotypes throughout the methods. In brief, enriched bacterial-BPW suspensions were tested for detection of the top 7 STEC virulence markers. The samples that contained the markers were indicative of STEC presence, and these samples were processed for isolation of colonies. The samples were confirmed as positive if the isolated colonies were pure and tested positive for all STEC virulence markers.

Detection of STEC serotypes

Bacterial-BPW suspensions were enriched overnight at 37°C using BPW as the enrichment media. A stepwise PCR-based approach was used to screen the enriched samples using the ViiA 7 Real-Time PCR System (Applied Biosystems, Scoresby, Victoria, Australia) for all PCR tests. The initial phase of PCR detection of STEC involved screening for stx and eae markers. The samples that were stx and eae positive were subsequently tested for the presence of genetic loci specific for individual O serotypes of the top 7 STEC (O26, O45, O103, O111, O121, O145, and O157). DNA was extracted from the enriched samples using the Wizard Genomic DNA Purification Kit (Promega, Alexandria, New South Wales, Australia). Extracted DNA was used as a template with the PCR procedures and sequences of primers and probes from previously published methods (29, 36). Protocols for O157 and non-O157 STEC were adopted from Perelle et al. (29) and the FSIS Microbiology Laboratory Guidebook (36). Samples that were positive for stx, eae, and one or more STEC serotypes were regarded as potential positives (PPs) and continued to the isolation and characterization phase.

Isolation and characterization of STEC from slaughter

Isolation of the top 7 STEC was carried out for PPs from slaughter room visits using immunomagnetic separation (IMS) with commercially available Assurance GDS MPX Top 7 STEC (AMSL, Chatswood, New South Wales, Australia) beads following the manufacturer's instructions. The bead-bacteria complexes at the completion of IMS for O157 were plated onto three media: sorbitol MacConkey agar with cefixime-tellurite supplement (0.05 mg/L cefixime and 2.5 mg/L potassium tellurite; Oxoid); Rainbow agar (Cell Biosciences, Preston, Victoria, Australia) supplemented with 0.05 mg/L cefixime trihydrate, 0.15 mg/L potassium tellurite, and 5.0 mg/L sodium novobiocin; and CHROMAgar STEC (Dutec Diagnostics, Croydon, New South Wales, Australia). Non-O157 bead-bacteria complexes were plated on modified Rainbow agar (as earlier) and CHROMAgar STEC. All media were incubated for 18 to 24 h at 37°C.

Following IMS, up to five phenotypically distinct colonies from each plate were streaked onto tryptic soy agar (Oxoid) and incubated for 18 to 24 h at 37°C. All isolates were initially tested for stx1, stx2, eae, and ehxA using conventional multiplex PCR as described by Paton and Paton (27). The isolates with stx and eae were then characterized targeting the serotypes that were captured during detection using real-time PCR. PCR master mix without DNA material was used as negative control and known pure top 7 STEC isolates were used as positive control for all PCR tests.

DNA extraction for 16S rRNA amplicon sequencing

A 40-mL aliquot of preenriched bacterial-BPW suspension from each sample was centrifuged at 10,000 × g for 5 min to pellet the cells. The pellets were washed twice with BPW before DNA extraction using the QIAamp PowerFecal DNA kit (Qiagen, Valencia, CA) following the manufacturer's protocol with modifications in the bead-beating step as previously described by Kang et al. (18).

Preparation of 16S rRNA library

Extracted DNA was used to construct a library of 16S rRNA amplicons targeting the V4 region using the protocol from a previous study by Kozich et al. (22). Briefly, 5 μL from each DNA template was amplified with dual-index primers by PCR, and the amplicons were manually normalized by comparing the intensity of the DNA band against the GeneRuler 100 bp Plus DNA Ladder (ThermoFisher, Scoresby, Victoria, Australia). The DNA materials were stained with ethidium bromide and visualized under UV light in 2% agarose gel. Each template contained barcode sequences at the 5′ and 3′ ends of the PCR amplicon to enable demultiplexing of sequence reads. Each template acted as a positive and negative control for each PCR reaction. Approximately 50 ng of rRNA DNA amplicon from each sample was combined to create a pooled library for sequencing. A microbial community DNA standard, ZymoBIOMICS (Zymo Research, Irvine, CA), was normalized using the same approach and added to the library. The library was purified using Agencourt AMPure XP magnetic beads (Beckman Coulter, Brea, CA) following the manufacturer's instructions. The concentration of amplicons was measured before and after purification using a Qubit 2.0 fluorometer (Life Technologies, Scoresby, Victoria, Australia). The NanoPhotometer (Implen, Southend-on-Sea, Essex, UK) was used to measure the purity of the library, and the final purified library was sent for sequencing.

Sequencing using Illumina MiSeq

Sequencing of the pooled library of amplicons was performed at the Ramaciotti Centre for Genomics (University of New South Wales) using the Illumina MiSeq platform (Illumina, San Diego, CA) with a paired-end 300 base pair sequencing protocol as previously described (18). All sequencing procedures were monitored through the Illumina BaseSpace Web site.

Bioinformatic analysis of 16S rRNA profile

Both demultiplexed R1 and R2 sequencing read files (approximately 250 bp in length) were acquired from the Illumina BaseSpace Web site. Bioinformatic processing of raw sequence data was performed using the v1.40.5 MOTHUR pipeline (31) following the MiSeq standard operating procedures (22). The paired end reads were generated and clustered before assembly into operational taxonomic unit (OTU) tables, with 97% as the cutoff for identification. The taxonomic classification of OTUs was performed using the SILVA 16S rRNA database (v132). The 16S rRNA sequences were converted into relative abundance for each sample in Microsoft Excel (v16.0) and used for the subsequent analysis. For investigation of bacterial diversity within groups (alpha diversity), Simpson's index (SI) and the number of OTUs at the genus level were calculated using the ANALYSE→DIVERSE tool on PRIMER-7 v7.0.13 (Primer-E, Ivybridge, UK) and used to create bar plots in Microsoft Excel. SI is scaled from 0 to 1, where 1 indicates high diversity within the community. For analysis of bacterial diversity between sample groups (beta diversity), nonmetric multidimensional scaling (nMDS) and analysis of similarity (ANOSIM) were used. Relative abundance of genus OTUs was used to create similarity ranks (resemblance matrix) using the Bray-Curtis similarity. The matrix was then used for nMDS ordination and ANOSIM (with >999 permutations) on ANALYSE tools within PRIMER-7. The samples were assigned a sample type and STEC testing status as a function of diversity during estimation of bacterial population composition.

RESULTS AND DISCUSSION

STEC PPs and confirmed isolates

The presence of PPs during slaughter at the abattoirs was similar and ranged between 64 and 69% (58 to 62 of 90) with the exception of the second visit to abattoir B that had 81% (73 of 90) of the samples as PPs (Table 1). The number of PPs generally decreased in the boning room, and abattoir B had a substantially lower percentage of PPs, at 7% (6 of 90) and 2% (2 of 90) compared with 23% (21 of 90) and 48% (43 of 90) in abattoir A samples, despite having equivalent or higher levels of STEC in the slaughter phase. From 360 samples from slaughter, 10 STEC isolates belonging to O157 and O111 serotypes were obtained (Table 2). Two O157 isolates were retrieved from the first abattoir A visit in a hide and a fecal sample with the same profile of virulence markers (stx2, eae, and ehxA). The remaining eight isolates were from the visits to abattoir B. O157 and O111 STEC were isolated from fecal and hide samples from the first visit, but only the hide samples from the second visit had O157 and O111. Three O157s were also isolated from carcass and environmental samples from the second visit. Prevalence of the top 7 STEC in the Australian beef industry is low and O157 is the most common STEC, whereas non-O157 STECs such as O26 and O111 have also been identified but are less common (4, 25, 37). Thus, it was not surprising to find as few as 10 isolates that were O157 and O111 STEC from the slaughter samples.

TABLE 1

Potential positives of the top 7 STEC from samples collected from two visits to slaughter and boning in two abattoirs with an integrated (abattoir A) and a fragment (abattoir B) supply chaina

Potential positives of the top 7 STEC from samples collected from two visits to slaughter and boning in two abattoirs with an integrated (abattoir A) and a fragment (abattoir B) supply chaina
Potential positives of the top 7 STEC from samples collected from two visits to slaughter and boning in two abattoirs with an integrated (abattoir A) and a fragment (abattoir B) supply chaina
TABLE 2

Sources and virulence markers of the top 7 STEC isolated from samples collected on the slaughter floor at an integrated abattoir (A) and a fragmented abattoir (B)a

Sources and virulence markers of the top 7 STEC isolated from samples collected on the slaughter floor at an integrated abattoir (A) and a fragmented abattoir (B)a
Sources and virulence markers of the top 7 STEC isolated from samples collected on the slaughter floor at an integrated abattoir (A) and a fragmented abattoir (B)a

Table 2 suggests that there were possibly higher proportions of STEC contamination within abattoir B on the second visit and that the contamination control strategies were not able to manage the subsequent increase in STEC concentration through slaughter. It also suggests that the transfer of STEC from feces to hides was minimized during the other three visits. Hereafter, PPs and non-PPs are used to compare and analyze the changes in bacterial community diversity. The hypothesis is that lower bacterial diversity should be observed in samples that are indicative of STEC presence if higher bacterial diversity is truly correlated with lower prevalence of STEC.

Alpha diversity by STEC PCR in slaughter samples

The alpha diversity (SI) was not significantly different (P > 0.05) between STEC-positive and STEC-negative samples regardless of the processing stage or sample type (Fig. 1). All fecal samples from abattoir A were PPs as opposed to the fecal samples from abattoir B, which had PPs and non-PPs (Fig. 1A and 1B). Regardless of STEC status, the fecal group in both abattoirs had similar OTUs and SI. Fecal PPs in abattoir A had an average of 148 OTUs and SI of 0.95. In abattoir B, fecal PPs had an average of 152 OTUs and SI of 0.93, whereas the negative group had an average of 140 OTUs and SI of 0.92, showing no significant differences (P > 0.05) in alpha diversity (Fig. 1B). Previous studies have analyzed bacterial diversity in fecal samples by comparing community structure in feces from animals with different levels of STEC shedding (39, 41, 42). Conclusions on the correlation of diversity and STEC prevalence varied among these studies. In this study, the comparison of diversity in fecal samples would have benefited from having a larger pool of both STEC-positive and STEC-negative samples to consolidate and compare the findings to the previous studies.

FIGURE 1

Number of OTUs at the genus level and SI (alpha diversity) in different sample groups in slaughter and boning of an integrated (abattoir A) and a fragmented (abattoir B) Australian beef abattoir. Abattoir A visits to slaughter and boning are shown in panels A and C, respectively. Abattoir B visits to slaughter and boning are shown in panels B and D, respectively. Sample groups are categorized by STEC quantitative PCR signals. STEC status of the groups is labeled + and −; some groups do not show both because of absence of the signal in the samples. Alpha diversity within each group is represented in SI.

FIGURE 1

Number of OTUs at the genus level and SI (alpha diversity) in different sample groups in slaughter and boning of an integrated (abattoir A) and a fragmented (abattoir B) Australian beef abattoir. Abattoir A visits to slaughter and boning are shown in panels A and C, respectively. Abattoir B visits to slaughter and boning are shown in panels B and D, respectively. Sample groups are categorized by STEC quantitative PCR signals. STEC status of the groups is labeled + and −; some groups do not show both because of absence of the signal in the samples. Alpha diversity within each group is represented in SI.

Hide samples from the slaughter floor had the highest average number of OTUs and SI values in comparison to the other samples in both abattoirs (Fig. 1A and 1B). Hide samples from abattoirs A and B had an average of 257 and 312 OTUs with an average SI of 0.96 and 0.94, respectively. Chopyk et al. (11) investigated bacterial diversity in preharvest hide samples and found that STEC presence correlated with lower bacterial diversity. The study eliminated fecal OTUs from hides and created a hide-specific 16S profile. The authors then used the hide-specific communities to measure diversity within and between sample groups. Chopyk et al. (11) added that the bacterial composition in hides without STEC varied from the hide communities with the presence of one, two, three, or more STEC serotypes. Comparison of diversity in hide microflora between the previous study and this study may provide insight into the differences in ecological niches that ultimately can affect diversity differences, with STEC as a function of diversity. However, there were insufficient positive and negative hide samples in this study to draw statistically significant conclusions or comparisons with the previous study (11). The aim of this study was not to compare changes in bacterial diversity at specific regions of processing (feces or hides) to the previous findings but to observe bacterial diversity in light of the top 7 STEC through the processing of beef cattle.

In abattoir A, the average number of OTUs in PP environmental samples (212) was marginally higher than that of the non-PP samples (186), but both groups had an average SI of 0.90 (Fig. 1A). Environmental PPs in abattoir B had a higher number of OTUs by more than 50 and a higher SI value by 0.06 than the negative group, but the difference in alpha diversity was not statistically significant (P > 0.05). In abattoir A, the average number of OTUs and SI were slightly higher in the non-PP group. However, the alpha diversity between the two groups was not significantly different (P > 0.05).

Presence of PPs in carcass samples from the boning phase did not have a significant effect (P > 0.05) on the alpha diversity in both abattoirs. In abattoir A, carcass PPs had an average SI of 0.87, whereas the non-PPs had an average of 0.86 (Fig. 1C). In abattoir B, the SI was noticeably lower in comparison to abattoir A at 0.73 in the positive group and 0.71 in the negative group (Fig. 1D). PPs from beef trim samples had an average SI value of 0.86, whereas the STEC-negative group had an average of 0.77 (Fig. 1C). However, the difference in alpha diversity between the two groups was not statistically significant (P > 0.05). The apparent numerical difference in alpha diversity of beef trim PPs and non-PPs from abattoir B may have occurred because a range of SI values was observed across the beef trim samples. Only one sample from beef trim (3%) and environmental surfaces (5%) tested positive for STEC; therefore, comparison of alpha diversity in different STEC testing groups of beef trim in abattoir B was not valid.

This study showed that alpha diversity in the boning room was generally lower in comparison to slaughter in both abattoirs (Fig. 1). In abattoir A, the average SI decreased from 0.91 on the slaughter floor to 0.83 in the boning room. In abattoir B, the reduction in alpha diversity across the samples was more substantial, from 0.89 in slaughter to 0.66 in boning. This correlated with a substantial decrease in the number of OTUs from slaughter to boning rooms in both abattoirs: 189 to 118 in abattoir A and 193 to 113 in abattoir B. Changes in bacterial community composition in the boning room were investigated in a different study (data not shown). In that study, the community structure throughout the boning room in abattoir A had several predominant bacterial populations (Clostridiales, Enterobacteriales, Lactobacillales, and Pseudomonadales) in contrast to the ecological makeup in abattoir B, where Enterobacteriales was the most dominant (19). Relative abundance of Enterobacteriales averaged at 39.5 and 51.2% in carcasses and 56.1 and 53.9% in beef trim from two visits to the boning room in abattoir B. In abattoir A, the four groups (Clostridiales, Enterobacteriales, Lactobacillales, and Pseudomonadales) dominated 65.5 to 72.8% of carcass composition and 63.6 to 73.3% in beef trim. Yang et al. (40) demonstrated that the proportion of E. coli among other bacterial groups increased in the beef trim community as diversity decreased. These findings shed light on the observation of alpha diversity in trim samples from abattoir B in this study and indicate that communities dominated by a small number of species can limit the richness of bacterial populations in beef microflora. Despite a large proportion of the community being Enterobacteriales in the boning room of abattoir B, the number of STEC PPs was comparably lower than in abattoir A, which suggests that the relative abundance of STEC in that population was minimal.

Previous studies have hypothesized that the degree of competitive exclusion increased as the microflora became more diverse and reduced the chance of STEC becoming prevalent within the community (11, 17). Another hypothesis was there may be a specific composition of bacterial populations that either outcompete or produce inhibitory substances that prevent STEC from becoming prevalent in beef microflora (2, 11, 24, 39, 41). Observations from this study do not provide evidence to validate these hypotheses. On the contrary, results from this study and previous studies appear to show that higher alpha diversity does not affect the STEC prevalence throughout beef processing (18, 19).

Relationship between STEC and beta diversity

Figure 2 shows the differences in bacterial diversity between the sample groups (beta diversity) in relation to the presence of STEC. The statistical significance of STEC markers as a function of diversity in bacterial communities is shown in Figure 3. Before the analysis of diversity in PPs and non-PPs, the samples with confirmed O157 and O111 STEC were compared with non-STEC samples, and no differences were observed. The confirmed positives were included as PPs during comparison of bacterial diversity in PP and non-PP groups, and the two groups did not show significant differences in bacterial diversity through slaughter (P > 0.05; Fig. 3). In abattoir A, all fecal and hide samples were PPs, and most of these samples overlapped on nMDS plots, indicating that the bacterial diversity in the two sample groups was similar (Fig. 2A). In abattoir B, the fecal and hide samples were separated from the rest on the nMDS plot, with hide samples closer to the carcass samples (Fig. 2B). The fecal group from both abattoirs shared commonality in that they were ranked away from the other sample types on nMDS plots. The carcass and environmental samples from slaughter rooms in both abattoirs were predominantly grouped without distinguishable trends between the groups, regardless of the STEC status (Fig. 2A and 2B).

FIGURE 2

Comparison of beta diversity of bacterial communities in all samples by STEC status using nMDS. Slaughter samples (fecal, hide, carcass, and environmental) from abattoirs A and B are shown in panels A and B, respectively. (C and D) Samples from the boning room (carcass, trim, and environmental) in abattoirs A and B, respectively. Blue labels indicate samples that tested positive for STEC, and red labels indicate negative STEC samples. Two-dimensional stress of nMDS ranged from 0.1 to 0.24.

FIGURE 2

Comparison of beta diversity of bacterial communities in all samples by STEC status using nMDS. Slaughter samples (fecal, hide, carcass, and environmental) from abattoirs A and B are shown in panels A and B, respectively. (C and D) Samples from the boning room (carcass, trim, and environmental) in abattoirs A and B, respectively. Blue labels indicate samples that tested positive for STEC, and red labels indicate negative STEC samples. Two-dimensional stress of nMDS ranged from 0.1 to 0.24.

FIGURE 3

ANOSIM using a combined factor (STEC markers and sample type) for comparing bacterial diversity within and between sample groups. Slaughter visits for abattoirs A and B are shown in panels A and B, respectively; boning room visits for abattoirs A and B are shown in panels C and D, respectively. The R value for each ordination is indicated by the dotted vertical line. All plots had a P value of 0.001, showing significance for the difference between frequency of permutated samples and R value from actual samples.

FIGURE 3

ANOSIM using a combined factor (STEC markers and sample type) for comparing bacterial diversity within and between sample groups. Slaughter visits for abattoirs A and B are shown in panels A and B, respectively; boning room visits for abattoirs A and B are shown in panels C and D, respectively. The R value for each ordination is indicated by the dotted vertical line. All plots had a P value of 0.001, showing significance for the difference between frequency of permutated samples and R value from actual samples.

The boning room samples provided a similar outcome, with no ecological differences observed between STEC PP and non-PP groups. In abattoir A, the carcass and environmental samples clustered relatively close to each other, and the distribution of PPs and non-PPs on nMDS appeared to be randomly placed (Fig. 2C). Some non-PPs of beef trim were separated from the rest of the trim samples, suggesting that the bacterial structure was slightly different in comparison to the other samples but was not statistically significant (P > 0.05; Fig. 3C). Carcass, trim, and environmental PPs from the boning room of abattoir B were grouped into one tight cluster; non-PPs separated into two defined areas with some outliers (Fig. 2D). The distribution of the samples in the two areas may result from differences observed in bacterial community composition between the boning room samples from visit 1 and those from visit 2 (data not shown). The samples that were grouped in the left population on nMDS were from the first visit, and the other population represented samples from the second visit (Fig. 2D). Therefore, the clustering of the samples on the nMDS plot may more be a function of variability between sampling days than of differences in beta diversity.

ANOSIM demonstrated that differences in beta diversity between sample groups—regardless of STEC testing status, processing phase, or supply chain—was not significant (P > 0.05; Fig. 3). R values in the slaughter phase from both abattoirs (abattoir A: 0.314; abattoir B: 0.395) showed greater differences in beta diversity (higher R value) in comparison to the boning phase (abattoir A: 0.134; abattoir B: 0.064). As discussed earlier, the fecal group in both abattoirs shared the least similarity in bacterial composition compared with the other groups (Fig. 2A and 2B). The higher R values indicate that the bacterial composition of feces is substantially different from that of the other sample groups (Fig. 2A and 2B), which increased the overall diversity differences during the ANOSIM calculation. Therefore, this result is indicative of the ecological influence of the fecal groups being responsible for the stretch of the R value (Fig. 3A and 3B) but not necessarily the hypothesized impact of STEC presence on bacterial diversity (11, 42). Previous STEC and diversity correlation studies have explored the diversity in a single sample type (i.e., feces or hide), and this correlation does not necessarily apply across sample types (11, 39, 41, 42). Kang et al. (18) demonstrated that the fecal bacterial populations shared the least similarity in bacterial composition (<60%) with carcass and environmental samples from slaughter rooms. This supports the impact of fecal bacteria on the analysis of overall diversity differences between slaughter groups in this study.

ANOSIM of bacterial diversity in the boning room added statistical confidence to the hypothesis of this study: that there is no correlation between STEC prevalence and bacterial diversity throughout beef processing. The histograms showing the frequency of permutations were similar between abattoirs regardless of the processing phase, but the R values from the actual samples in the boning room reflected the higher level of similarity in diversity in comparison to slaughter rooms (Figs. 2 and 3). The R values for abattoirs A and B were 0.134 and 0.064, respectively, and both analyses had a P value less than 0.05 (P = 0.001; Fig. 3C and 3D). A previous study by Kang et al. (19) demonstrated that carcass and trim communities in the boning room consistently shared more than 70% similarity in composition and more than 50 to 70% with environmental microflora. This study took STEC testing status into consideration and produced an outcome in which the distribution of high and low ranks of dissimilarity within and between carcass, trim, and environmental groups was closer to being even. This indicated that the bacterial diversity of samples collected throughout the slaughter and boning process could not be used to predict the presence or absence of STEC.

In conclusion, correlation between bacterial diversity and presence of the top 7 STEC did not appear to exist through the slaughter and boning phases in two Australian beef abattoirs with different levels of fragmentation in the supply chain. The major difference of this study compared with previous studies is that this study investigated the processing stages in the beef supply chain. It is possible that the previous studies may have observed correlations of diversity or dominant bacterial species in response to the top 7 STEC presence, because there were opportunities for bacterial populations to establish a community in the samples that were collected (feces and hides in feedlots). A stable community is less likely to grow and become fully established on meat in the processing environment, because microbial groups within the microflora are consistently changing and exposed to intervention strategies used in the processing facilities (1, 6, 18, 20). This study used 16S rRNA amplicon sequencing to show that the processing environment may not be the optimal part of the beef chain to better understand the possible relationship between STEC and diversity. Future studies using techniques with increased resolution, such as shotgun metagenomics, would be useful to provide a deeper understanding of the relationship between STEC presence and microbial diversity.

ACKNOWLEDGMENTS

This study was funded by Australian Meat Processors Corporation, and we thank the participating Australian beef abattoirs.

REFERENCES

REFERENCES
1.
Arthur,
T. M.,
Bosilevac
J. M.,
Nou
X. W.,
Shackelford
S. D.,
Wheeler
T. L.,
Kent
M. P.,
Jaroni
D.,
Pauling
B.,
Allen
D. M.,
and
Koohmaraie
M.
2004
.
Escherichia coli O157 prevalence and enumeration of aerobic bacteria, Enterobacteriaceae, and Escherichia coli O157 at various steps in commercial beef processing plants
.
J. Food Prot
.
67
:
658
665
.
2.
Arthur,
T. M.,
Brichta-Harhay
D. M.,
Bosilevac
J. M.,
Kalchayanand
N.,
Shackelford
S. D.,
Wheeler
T. L.,
and
Koohmaraie
M.
2010
.
Super shedding of Escherichia coli O157:H7 by cattle and the impact on beef carcass contamination
.
Meat Sci
.
86
:
32
37
.
3.
Barlow,
R. S.,
Gobius
K. S.,
and
Desmarchelier
P. M.
2006
.
Shiga toxin–producing Escherichia coli in ground beef and lamb cuts: results of a one-year study
.
Int. J. Food Microbiol
.
111
:
1
5
.
4.
Barlow,
R. S.,
and
Mellor
G. E.
2010
.
Prevalence of enterohemorrhagic Escherichia coli serotypes in Australian beef cattle
.
Foodborne Pathog. Dis
.
7
:
1239
1245
.
5.
Bosilevac,
J. M.,
Guerini
M. N.,
Brichta-Harhay
D. M.,
Arthur
T. M.,
and
Koohmaraie
M.
2007
.
Microbiological characterization of imported and domestic boneless beef trim used for ground beef
.
J. Food Prot
.
70
:
440
449
.
6.
Brashears,
M. M.,
and
Chaves
B. D.
2017
.
The diversity of beef safety: a global reason to strengthen our current systems
.
Meat Sci
.
132
:
59
71
.
7.
Brightwell,
G.,
Boerema
J.,
Mills
J.,
Mowat
E.,
and
Pulford
D.
2006
.
Identifying the bacterial community on the surface of Intralox belting in a meat boning room by culture-dependent and culture-independent 16S rDNA sequence analysis
.
Int. J. Food Microbiol
.
109
:
47
53
.
8.
Browne,
A. S.,
Midwinter
A. C.,
Withers
H.,
Cookson
A. L.,
Biggs
P. J.,
Marshall
J. C.,
Benschop
J.,
Hathaway
S.,
Haack
N. A.,
Akhter
R. N.,
and
French
N. P.
2018
.
Molecular epidemiology of Shiga toxin–producing Escherichia coli (STEC) on New Zealand dairy farms: application of a culture-independent assay and whole-genome sequencing
.
Appl. Environ. Microbiol
.
84
:
e00481-18.
9.
Castro,
V. S.,
Carvalho
R. C. T.,
Conte
C. A.,
and
Figuiredo
E. E. S.
2017
.
Shiga-toxin producing Escherichia coli: pathogenicity, supershedding, diagnostic methods, occurrence, and foodborne outbreaks
.
Compr. Rev. Food Sci. Food Saf
.
16
:
1269
1280
.
10.
Chandry,
S.
2016
.
Metagenomic analysis to explore the mechanisms of carcass contamination
.
11.
Chopyk,
J.,
Moore
R. M.,
DiSpirito
Z.,
Stromberg
Z. R.,
Lewis
G. L.,
Renter
D. G.,
Cernicchiaro
N.,
Moxley
R. A.,
and
Wommack
K. E.
2016
.
Presence of pathogenic Escherichia coli is correlated with bacterial community diversity and composition on pre-harvest cattle hides
.
Microbiome
4
:
9
.
12.
Elder,
R. O.,
Keen
J. E.,
Siragusa
G. R.,
Barkocy-Gallagher
G. A.,
Koohmaraie
M.,
and
Laegreid
W. W.
2000
.
Correlation of enterohemorrhagic Escherichia coli O157 prevalence in feces, hides, and carcasses of beef cattle during processing
.
Proc. Natl. Acad. Sci. USA
97
:
2999
3003
.
13.
Fegan,
N.,
Higgs
G.,
Duffy
L. L.,
and
Barlow
R. S.
2009
.
The effects of transport and lairage on counts of Escherichia coli O157 in the feces and on the hides of individual cattle
.
Foodborne Pathog. Dis
.
6
:
1113
1120
.
14.
Fegan,
N.,
Higgs
G.,
Vanderlinde
P.,
and
Desmarchelier
P.
2004
.
Enumeration of Escherichia coli O157 in cattle faeces using most probable number technique and automated immunomagnetic separation
.
Lett. Appl. Microbiol
.
38
:
56
59
.
15.
Fegan,
N.,
Higgs
G.,
Vanderlinde
P.,
and
Desmarchelier
P.
2005
.
An investigation of Escherichia coli O157 contamination of cattle during slaughter at an abattoir
.
J. Food Prot
.
68
:
451
457
.
16.
Ferdous,
M.,
Zhou
K.,
Mellmann
A.,
Morabito
S.,
Croughs
P. D.,
de Boer
R. F.,
Kooistra-Smid
A. M. D.,
Rossen
J. W. A.,
and
Friedrich
A. W.
2015
.
Is Shiga toxin–negative Escherichia coli O157:H7 enteropathogenic or enterohemorrhagic Escherichia coli? Comprehensive molecular analysis using whole-genome sequencing
.
J. Clin. Microbiol
.
53
:
3530
3538
.
17.
Fujikawa,
H.,
and
Sakha
M. Z.
2014
.
Prediction of competitive microbial growth in mixed culture at dynamic temperature patterns
.
Biocontrol Sci
.
19
:
121
127
.
18.
Kang,
S.,
Ravensdale
J.,
Coorey
R.,
Dykes
G. A.,
and
Barlow
R.
2019
.
A comparison of 16S rRNA profiles through slaughter in Australian export beef abattoirs
.
Front. Microbiol
.
10
:
2747
.
19.
Kang,
S.,
Ravensdale
J.,
Coorey
R.,
Dykes
G. A.,
and
Barlow
R.
.
Changes in bacterial composition determined using 16S rRNA amplicon sequencing in the boning room of Australian beef export abattoirs
.
Unpublished data.
20.
Kennedy,
T. G.,
Giotis
E. S.,
and
McKevitt
A. I.
2014
.
Microbial assessment of an upward and downward dehiding technique in a commercial beef processing plant
.
Meat Sci
.
97
:
486
489
.
21.
Kergourlay,
G.,
Taminiau
B.,
Daube
G.,
and
Champomier Verges
M. C.
2015
.
Metagenomic insights into the dynamics of microbial communities in food
.
Int. J. Food Microbiol
.
213
:
31
39
.
22.
Kozich,
J. J.,
Westcott
S. L.,
Baxter
N. T.,
Highlander
S. K.,
and
Schloss
P. D.
2013
.
Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform
.
Appl. Environ. Microbiol
.
79
:
5112
5120
.
23.
Lagkouvardos,
I.,
Joseph
D.,
Kapfhammer
M.,
Giritli
S.,
Horn
M.,
Haller
D.,
and
Clavel
T.
2016
.
IMNGS: A comprehensive open resource of processed 16S rRNA microbial profiles for ecology and diversity studies
.
Sci. Rep
.
6
:
33721
.
24.
Matthews,
L.,
McKendrick
I. J.,
Ternent
H.,
Gunn
G. J.,
Synge
B.,
and
Woolhouse
M. E.
2006
.
Super-shedding cattle and the transmission dynamics of Escherichia coli O157
.
Epidemiol. Infect
.
134
:
131
142
.
25.
Mellor,
G. E.,
Fegan
N.,
Duffy
L. L.,
McMillan
K. E.,
Jordan
D.,
and
Barlow
R. S.
2016
.
National survey of Shiga toxin–producing Escherichia coli serotypes O26, O45, O103, O111, O121, O145, and O157 in Australian beef cattle feces
.
J. Food Prot
.
79
:
1868
1874
.
26.
Nguyen,
Y.,
and
Sperandio
V.
2012
.
Enterohemorrhagic E. coli (EHEC) pathogenesis
.
Front. Cell Infect. Microbiol
.
2
:
90
.
27.
Paton,
A. W.,
and
Paton
J. C.
1998
.
Detection and characterization of Shiga toxigenic Escherichia coli by using multiplex PCR assays for stx1, stx2, eaeA, enterohemorrhagic E. coli hlyA, rfbO111, and rfbO157
.
J. Clin. Microbiol
.
36
:
598
602
.
28.
Paton,
J. C.,
and
Paton
A. W.
1998
.
Pathogenesis and diagnosis of Shiga toxin–producing Escherichia coli infections
.
Clin. Microbiol. Rev
.
11
:
450
479
.
29.
Perelle,
S.,
Dilasser
F.,
Grout
J.,
and
Fach
P.
2004
.
Detection by 5′-nuclease PCR of Shiga-toxin producing Escherichia coli O26, O55, O91, O103, O111, O113, O145 and O157:H7, associated with the world's most frequent clinical cases
.
Mol. Cell. Probes
18
:
185
192
.
30.
Pointon,
A.,
Kiermeier
A.,
and
Fegan
N.
2012
.
Review of the impact of pre-slaughter feed curfews of cattle, sheep and goats on food safety and carcase hygiene in Australia
.
Food Control
26
:
313
321
.
31.
Schloss,
P. D.,
Westcott
S. L.,
Ryabin
T.,
Hall
J. R.,
Hartmann
M.,
Hollister
E. B.,
Lesniewski
R. A.,
Oakley
B. B.,
Parks
D. H.,
Robinson
C. J.,
Sahl
J. W.,
Stres
B.,
Thallinger
G. G.,
Van Horn
D. J.,
and
Weber
C. F.
2009
.
Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities
.
Appl. Environ. Microbiol
.
75
:
7537
7541
.
32.
Stromberg,
Z. R.,
Baumann
N. W.,
Lewis
G. L.,
Sevart
N. J.,
Cernicchiaro
N.,
Renter
D. G.,
Marx
D. B.,
Phebus
R. K.,
and
Moxley
R. A.
2015
.
Prevalence of enterohemorrhagic Escherichia coli O26, O45, O103, O111, O121, O145, and O157 on hides and preintervention carcass surfaces of feedlot cattle at harvest
.
Foodborne Pathog. Dis
.
12
:
631
638
.
33.
Svoboda,
A. L.,
Dudley
E. G.,
Debroy
C.,
Mills
E. W.,
and
Cutter
C. N.
2013
.
Presence of Shiga toxin–producing Escherichia coli O-groups in small and very-small beef-processing plants and resulting ground beef detected by a multiplex polymerase chain reaction assay
.
Foodborne Pathog. Dis
.
10
:
789
795
.
34.
U.S. Department of Agriculture, Food Safety and Inspection Service
.
2011
.
Shiga toxin–producing Escherichia coli in certain raw beef products
.
Fed. Regist
.
76
(182)
:
58157
58165
.
35.
U.S. Department of Agriculture, Food Safety and Inspection Service
.
2019
.
Microbiology laboratory guidebook 5C.00: Detection, isolation and identification of top seven Shiga toxin–producing Escherichia coli (STECs) from meat products and carcass and environmental sponges
.
36.
U.S. Department of Agriculture, Food Safety and Inspection Service
.
2019
.
Microbiology laboratory guidebook 5C.00 appendix 4.00: Primer and probe sequences and reagent concentrations for non-O157 Shiga toxin–producing Escherichia coli (STEC) real-time PCR assay
.
37.
Vally,
H.,
Hall
G.,
Dyda
A.,
Raupach
J.,
Knope
K.,
Combs
B.,
and
Desmarchelier
P.
2012
.
Epidemiology of Shiga toxin producing Escherichia coli in Australia, 2000–2010
.
BMC Public Health
12
:
63
.
38.
Vanderlinde,
P. B.,
Shay
B.,
and
Murray
J.
1998
.
Microbiological quality of Australian beef carcass meat and frozen bulk packed beef
.
J. Food Prot
.
61
:
437
443
.
39.
Xu,
Y.,
Dugat-Bony
E.,
Zaheer
R.,
Selinger
L.,
Barbieri
R.,
Munns
K.,
McAllister
T. A.,
and
Selinger
L. B.
2014
.
Escherichia coli O157:H7 super-shedder and non-shedder feedlot steers harbour distinct fecal bacterial communities
.
PLoS One
9
:
e98115
.
40.
Yang,
X.,
Noyes
N. R.,
Doster
E.,
Martin
J. N.,
Linke
L. M.,
Magnuson
R. J.,
Yang
H.,
Geornaras
I.,
Woerner
D. R.,
Jones
K. L.,
Ruiz
J.,
Boucher
C.,
Morley
P. S.,
and
Belk
K. E.
2016
.
Use of metagenomic shotgun sequencing technology to detect foodborne pathogens within the microbiome of the beef production chain
.
Appl. Environ. Microbiol
.
82
:
2433
2443
.
41.
Zaheer,
R.,
Dugat-Bony
E.,
Holman
D. B.,
Cousteix
E.,
Xu
Y.,
Munns
K.,
Selinger
L. J.,
Barbieri
R.,
Alexander
T.,
McAllister
T. A.,
and
Selinger
L. B.
2017
.
Changes in bacterial community composition of Escherichia coli O157:H7 super-shedder cattle occur in the lower intestine
.
PLoS One
12
:
e0170050
.
42.
Zhao,
L.,
Tyler
P. J.,
Starnes
J.,
Bratcher
C. L.,
Rankins
D.,
McCaskey
T. A.,
and
Wang
L.
2013
.
Correlation analysis of Shiga toxin–producing Escherichia coli shedding and fecal bacterial composition in beef cattle
.
J. Appl. Microbiol
.
115
:
591
603
.