Stable flies are bona fide carriers of mastitis-associated bacteria

Abstract

ABSTRACT

Hematophagous Stomoxys (stable) fly populations in dairy barns are sustained by a constant availability of cattle hosts and manure, which serve as major reservoirs of both zoonotic and opportunistic bacterial pathogens. However, the composition of the Stomoxys fly microbiota, the mechanisms by which flies acquire their microbiome, and the ability of potentially pathogenic bacteria to colonize and persist in fly hosts remain to be investigated. Here, we longitudinally collected fly and manure samples from two connected dairy facilities. High throughput 16S rRNA gene amplicon sequencing was then used to characterize and compare bacterial communities present on or within flies and in manure collected from the same facility, while culture-dependent methods were used to verify the viability of clinically relevant bacteria. Bacterial alpha diversity was overall higher in manure samples as compared to fly samples, with manure-associated bacterial communities being dominated by members of the Bacteroidales, Eubacteriales, and Oscillospirales. In contrast, flies harbored relatively low-complexity communities dominated by members of the Enterobacterales, Staphylococcales, and Lactobacillales. Clinically relevant bacterial strains, including Escherichia spp. and other taxa associated with mastitic cows housed in the same facilities, were detected in paired fly and manure samples but exhibited dramatically elevated abundances in fly samples as compared to manure samples. Viable colonies of Escherichia, Klebsiella, and Staphylococcus spp. were also readily isolated from fly samples, confirming that flies harbor culturable mastitis-associated bacteria. This study identifies biting flies as bona fide carriers of opportunistically pathogenic bacterial taxa on dairy farms.

IMPORTANCE

Disease prevention on dairy farms has significant implications for cattle health, food security, and zoonosis. Of particular importance is the control of bovine mastitis, which can be caused by diverse bacteria, including Klebsiella, Escherichia coli, Streptococcus, and Staphylococcus spp. Despite being one of the most significant and costly cattle diseases worldwide, the epidemiology of bovine mastitis is not well understood. This study provides parallel culture-independent and culture-dependent evidence to support the carriage of opportunistically pathogenic bacteria by Stomoxys flies on dairy farms. We further show that the fly microbiota is enriched in clinically relevant taxa—the vast majority of which can be traced to the manure habitats in which flies breed. Altogether, our results identify biting flies as underrecognized carriers of bacterial taxa associated with environmental bovine mastitis and other opportunistic infections in vertebrates and offer important insights into mechanisms of microbial acquisition by these and other medically important insects.

INTRODUCTION

Animal-borne (zoonotic) diseases pose a serious risk to both human health and global food security. Bio-surveillance across dairy farms is especially important as cow manure is a well-established reservoir of major human bacterial pathogens, including enterohemorrhagic Escherichia coli, Salmonella, Listeria, Campylobacter, as well as economically important bovine bacterial pathogens, including the causative agents of mastitis, enteritis, and neonatal diarrhea (1–5). The growing threat of zoonosis, which accounts for most new emerging infectious diseases (6), further highlights the need for a better understanding of pathogen ecology and potential mechanisms of bacterial dispersal and persistence on dairy farms and in other agricultural settings.

Bovine mastitis, an inflammatory response to intramammary infection in lactating cows, is the most prevalent disease on dairy farms, with the USDA National Animal Health Monitoring System (NAHMS) reporting cases in 99.7% of all U.S. dairy operations (7). In brief, bacterial invasion of the mammary gland triggers the recruitment of neutrophils and expansion of proinflammatory cytokines to control the spread of infection (8, 9). This host immune response leads to an increase in the abundance of somatic cells (i.e., neutrophils) in the udder tissue, which can then shed into produced milk (10, 11). Not only does low-quality milk need to be discarded, but dairy farmers must also pay for costs associated with lost labor, as well as the treatment and potential culling of sick cows (12). The economic losses associated with mastitis cases are especially burdensome, costing the global dairy industry an estimated 35 billion dollars annually (13).

Intramammary infections can result from either direct contact with contaminated milking equipment or through environmental exposure post-milking. Contagious, mastitis-causing bacterial pathogens, including host-adapted strains of Staphylococcus aureus and Streptococcus agalactiae, are primarily spread from cow to cow during the milking process (2, 3). Environmental pathogens, in contrast, cause opportunistic infections through contact with soiled bedding, manure, or other contaminated areas. A diverse array of environmental bacteria can induce clinical and subclinical forms of mastitis; however, strains of Enterobacteriaceae, Staphylococcaceae, and Streptococcaceae are among the most prevalent causative agents of intramammary infections and are of particular concern to the dairy industry (2, 3). As treatment of many environmental mastitis pathogens is often viewed as ineffective and infection dynamics are primarily host-driven (14, 15), there is need for a deeper comprehension of pathogen ecology, management strategies, and epidemiology to mitigate the disease burden.

Members of the insect family Muscidae (Insecta: Diptera), including flies of the genera Musca (house and face flies) and Stomoxys (stable flies), have long been implicated in the carriage of manure-associated bacteria, including opportunistic environmental mastitis pathogens (16–20). Adult muscid flies are covered in hair-like projections, which may lead to external dissemination of environmental microbes collected from manure or other debris (21). Dissemination of internally associated bacteria can also occur, either through defecation or regurgitation of saliva when feeding (22, 23). Muscid flies are most prevalent on dairy farms during the warmest months of the year, and incidences of mastitis in dairy cattle peak in this same timeframe (24, 25). Stomoxys flies are also obligate blood feeders, and the dense livestock populations present in most barns provide ample access to nutritional blood meals for biting flies (19). Large fly populations are additionally sustained by the constant availability of raw manure and soiled bedding, which serve as their preferred breeding sites and are also likely to harbor environmental pathogens. In a typical commercial dairy farm setting, adult female flies deposit eggs on both managed manure piles and fresh pats or slurries of manure, where larvae develop through a series of three progressively larger immature stages known as instars before undergoing metamorphosis to the adult stage (26). In addition to providing a source of undigested carbohydrates, proteins, and other nutrients to developing larvae, both adult male and female flies are strongly attracted to and may opportunistically ingest cattle manure to support their breeding activities (27, 28).

Although historical studies have provided circumstantial evidence linking muscid fly abundance to the prevalence of mastitis pathogens on dairy farms, very little is known about the native microbiota of muscid flies. To date, studies attempting to characterize the muscid fly microbiota have focused near-exclusively on M. domestica and examinations of Stomoxys-associated bacteria have relied primarily on selective bacterial culturing (29–31). The overall composition and diversity of Stomoxys-associated bacterial communities, the impact of manure habitats in shaping fly bacterial community assembly, and the relative contribution of potential opportunistic pathogens to the fly microbiota, remain to be investigated. In this study, we provide the first parallel culture-independent and culture-dependent analysis of any biting muscid fly microbiota through longitudinal sampling of adult Stomoxys flies and environmental manure across two connected dairy facilities. High throughput 16S rRNA gene amplicon sequencing was used to characterize bacterial communities in fly and manure samples collected on a weekly basis during peak fly season in both facilities. Culture-dependent methods were then employed to verify the viability of highly prevalent taxa, including mastitis-associated Escherichia coli, Staphylococcus, and Klebsiella spp., in fly homogenates plated on selective enrichment media. Our results demonstrate that the bacteria associated with Stomoxys flies nearly completely overlap with the bacteria present in their associated manure habitats, although community diversity is lower in flies and the abundance of specific community members differs. Notably, mastitis-associated taxa present at low relative abundances in manure were highly enriched in paired fly samples, suggesting that biting flies represent important, yet understudied carriers of opportunistic bacterial pathogens within and off of animal husbandry facilities.

RESULTS

Stomoxys flies harbor a distinct, low-complexity microbiota

In order to characterize the Stomoxys fly microbiota in relation to environmental manure, Stomoxys flies and manure samples were collected on a weekly basis across two dairy facilities in South Central Wisconsin: the UW-Madison Dairy Cattle Center (DCC) and the Arlington Agricultural Research Station Blaine Dairy Cattle Center (Arlington) (Fig. 1). Flies recovered from the same trap were pooled into groups of up to 25 individuals to collect externally associated bacteria before homogenizing subgroups of 3–5 individuals to collect internal bacteria. A total of 183 Stomoxys pools (697 flies total) and 106 manure samples were processed for high throughput 16S rRNA gene amplicon sequencing, of which 126 internal fly samples, 32 external fly samples, and 106 manure samples, with a respective median read count of 7,108, 5,807, and 26,050, respectively, were selected for further analysis (Table S1). Across both facilities, bacterial richness was significantly greater in manure samples compared to both internal and external fly samples (Fig. 2), irrespective of sampling date (Fig. S1). Bacterial richness in Arlington-derived external fly samples was also higher than internal fly samples (Fig. 2); however, no significant differences were observed between external and internal fly samples derived from the DCC (Fig. 2; Fig. S1). No significant differences in bacterial richness were observed between manure and fly samples collected from different sampling locations within either facility (Fig. S2), although richness estimates for DCC-derived manure samples were consistently higher than those for Arlington-derived manure samples (Fig. 2; Fig. S1 and S2).

Fig 1

Collection sites for the 16S rRNA libraries and glycerol stocks prepared from manure and fly samples. The upper portion of the figure shows the locations of the collection sites at the Arlington Agricultural Research Station Blaine Dairy Cattle Center (Arlington), while the lower portion of the figure shows the locations of the collection sites at the UW-Madison Dairy Cattle Center (DCC). White squares and black circles delineate fly trap and manure sampling locations in both facilities, respectively. The locations of the manure scraper systems and pen housing sick cattle (Sick pen) at Arlington are further identified using black diamonds/lines and a white rectangle, respectively (top panel only). Both facilities are located in South Central Wisconsin on or in close proximity to the UW-Madison campus.

Fig 2

Community-level ASV richness in Arlington-derived (left) and DCC-derived (right) samples, by sample source (x-axis; yellow bars: internal fly samples, blue bars: external fly samples; red bars: manure samples). Bars represent 95% CI (±1.96 * SE).

Altogether, we identified a total of 7,666 bacterial ASVs in manure samples and 1,539 and 1,233 ASVs across the internal and external fly samples we sequenced, respectively (Table S2). These ASVs corresponded to a total of 191 bacterial orders being represented across all samples, although the relative abundance of each order varied greatly between fly and manure samples (Fig. 3; Table S2). Internal fly samples from both Arlington and the DCC on average harbored high relative abundances of taxa within the Enterobacterales (29.6% Arlington, 37.1% DCC) and Lactobacillales (18.3% Arlington, 20.3% DCC), while taxa within the Staphylococcales were overall more abundant in Arlington internal fly samples (27.7%) than in DCC internal fly samples (11.2%) (ANCOM-BC, log fold change > 1.5 and FDR-adjusted P value < 0.05) (Fig. 3). Taxa within the Enterobacterales (20.3% Arlington, 27.1% DCC) and Pseudomonadales (11.4% Arlington, 12.1% DCC) were similarly highly abundant in external fly samples from both facilities (Fig. 3). In contrast, manure samples from both Arlington and the DCC on average contained high relative abundances of taxa within the Bacteroidales (15.3% Arlington, 21.9% DCC) and Oscillospirales (21.0% Arlington, 29.4% DCC), while taxa within the Lactobacillales were highly abundant in Arlington manure samples (16.7%) but not DCC manure samples (2.7%) (ANCOM-BC, log fold change > 1.5 and FDR-adjusted P value < 0.05) (Fig. 3).

Fig 3

Relative abundance of bacterial orders in Arlington-derived (top) and DCC-derived (bottom) fly and manure samples, by sampling date (x-axis). Libraries derived from the same sample source on a given sampling date were pooled across sampling locations for the bar graphs presented. Colored bars present the proportion of sequencing reads assigned to a given bacterial order. Low abundance orders (<2%) are represented by the “Other” category.”

Principal coordinates analyses using the Bray-Curtis dissimilarity index further revealed significant clustering of bacterial communities by sample source, with both Arlington and DCC manure-associated communities forming distinct clusters separate from bacterial communities in flies collected from the same facilities (Fig. 4), again irrespective of sampling date (Fig. S3 and S4). Arlington-derived internal and external fly samples also formed distinct clusters (Fig. 4), with bacterial communities in Arlington-derived external fly samples generally being more similar to those present in manure collected from the same facility (Fig. 4). In keeping with our bacterial richness results, bacterial community composition was largely consistent across manure and fly samples collected from different sampling locations within a given facility (Fig. S5). The only exceptions to this were the manure samples collected from the sick pen at Arlington, which harbored notably higher relative abundance of bacterial taxa within the order Burkholderiales and lower relative abundances of taxa within the Lactobacillales and Corynebacteriales than manure collected from different sampling locations in the same facility (ANCOM-BC, log fold change > 1.5 and FDR-adjusted P value < 0.05) (Fig. S5 and S6).

Fig 4

PCoA of Bray-Curtis dissimilarities of community relative abundances, colored by sample source (internal fly samples, yellow; external fly samples, blue; manure samples, red). Each point represents one manure or fly sample community. PERMANOVA identified significant clustering by sample source for both Arlington-derived samples (left; P = 0.001, R² = 0.21) and DCC-derived samples (right; P = 0.001, R² = 0.28). Significant post hoc pairwise comparisons between internal/external fly and manure samples were also identified for both facilities (P < 0.01). In contrast, a significant pairwise comparison between internal and external fly samples was identified for Arlington (P = 0.001) but not the DCC (P = 0.2450).

Bacterial communities in Stomoxys flies are highly enriched in mastitis-associated taxa

To develop a more detailed understanding of fly-associated bacterial communities and their relationship to those detected in manure, we next identified bacterial taxa shared between communities in manure and flies collected from the same facility and asked what is the relative abundance of these bacterial taxa in manure and fly samples, respectively. On average, both Arlington and DCC-derived manure and fly samples shared ~20 bacterial ASVs, which accounted for a large percentage (up to 100%) of the total reads in individual fly samples (Fig. 5), irrespective of sampling date or location (Fig. S7 to S10). In contrast, these taxa on average represented only ~39% of the total reads in individual manure samples, although the average relative abundance of shared taxa was significantly lower in manure samples collected from the DCC as compared to those collected from Arlington (Fig. 5; Fig. S11 and S12).

Fig 5

Bacterial ASV analysis. (A) Average number of ASVs shared between Arlington-derived (left) or DCC-derived (right) fly (internal or external) and manure samples. (B) Average relative abundance of fly associated ASVs (internal or external) shared with manure collected from the same facility. (C) Average relative abundance of manure associated ASVs shared with flies collected from the same facility. Bars in all panels represent 95% CI (±1.96 * SE).

We also examined whether there were ASVs that were frequently shared by manure and fly samples collected from the same facility. A total of 17 and 12 ASVs were shared in at least 25% of Arlington- and DCC-derived manure-fly sample pairs, respectively (Fig. 6). This included representatives of genera within the bacterial families Enterobacteriaceae (Order Enterobacterales), Enterococcaceae and Streptococcaceae (Order Lactobacillales), Pseudomonadaceae (Order Pseudomonadales), and Staphylococcaceae (Order Bacillales), all of which are commonly associated with environmentally acquired mastitis (Fig. 6) (15). Perhaps most interesting, however, was the observation that, while the bacterial taxa present in both manure and fly samples represented only a small fraction of manure bacterial communities, they exhibited very high relative abundances in fly bacterial communities (Fig. 7; Fig. S13 to S16). Highly enriched taxa were also cultured from mastitic cows housed in the same facilities during the sampling period (Table S3).

Fig 6

Frequency plot of the 17 and 12 bacterial ASVs shared in at least 25% of Arlington-derived (left)- and DCC-derived (right) manure-fly sample pairs, respectively, along with their taxonomy at the genus level. Taxa highlighted in red were identified as commonly shared ASVs in both facilities.

Fig 7

Relative abundance of commonly shared ASVs in Arlington-derived (left) and DCC-derived (right) samples, by sample source (x-axis). Libraries derived from a given sample source were pooled across sampling locations and dates for the bar graphs presented. Colored bars present the proportion of sequencing reads assigned to a given ASV. ASV taxonomy is shown at the genus level. Taxa highlighted in red were identified as commonly shared ASVs in both facilities.

Stomoxys flies serve as bona fide carriers of viable mastitis-associated bacteria

To confirm that flies carry mastitis-associated bacterial taxa, we screened glycerol stocks generated from homogenates of pooled internal fly samples for the growth of Staphylococcus, Escherichia, and Klebsiella bacteria. Viable colonies were recovered from 88.5% and 100% of fly homogenates streaked onto MacConkey and mannitol salt agar plates, which are generally permissive for the growth of Gram-negative and Gram-positive bacteria, respectively (Table S4). Restreaking of individual colonies on selective agar (followed by morphological or molecular identification) confirmed the presence of E. coli in 22.95% (28 out of 122) of all plated homogenates, which represents viable growth in 55.81% (24 out of 43) of samples with corresponding sequence reads assigned to the genus Escherichia-Shigella (Tables S2 and S4). The same methods also confirmed the presence of Klebsiella spp., which were likely defined as unidentified members of the Enterobacteriaceae in our sequencing data set, in 21.31% (26 out of 122) of all plated homogenates, which represents 27.47% (25 out of 91) of samples with potential corresponding sequence reads (Tables S2 and S4). Similarly, 75.41% (92 out of 122) of all plated homogenates, representing 79.28% (88 out of 111) of samples with corresponding sequence reads, contained viable colonies of Staphylococcus spp. (Tables S2 and S4).

DISCUSSION

Muscid flies have long been implicated in the mechanical transmission of environmental bacterial pathogens responsible for bovine mastitis and other vertebrate diseases (32, 33). Large fly populations are sustained by the high availability of raw manure, which serves as both the preferred oviposition site for adult flies and as a major reservoir of microbial pathogens (19). However, previous experimental studies on the vectorial capacity and microbial communities of muscid flies have focused primarily on the house fly, Musca domestica. This work collectively indicates that bacterial isolates can proliferate within the house fly digestive tract leading to the continual excretion of bacteria via regurgitation and defecation (23, 33–36). Across sampling locations, Staphylococcus and Weissella appear to be highly dominant members of the M. domestica microbiome; zoonotic pathogens, including E. coli O157, Salmonella enterica, and Campylobacter are also routinely isolated from house fly samples (37–43). In contrast, far less is known about the microbiota of biting Stomoxys flies. To date, there has been no culture-independent analysis of the microbiota of any biting muscid fly and only a handful of studies have used culture-dependent methods to survey Stomoxys fly-associated microbial communities, identifying strains of Bacillus, Staphylococcus, and Enterobacterales as among the most common bacterial taxa isolated from adult flies (29–31). Prior research on Stomoxys fly microbiota has also focused exclusively on fly-associated microbes, without also examining potential acquisition from the surrounding barn environment or environmental breeding sites.

The first goal of this study was to characterize the Stomoxys fly microbiota in relation to environmental manure collected longitudinally over a 3-month sampling period across two dairy barns in South Central Wisconsin. Our analyses revealed that adult Stomoxys flies harbor a low-complexity microbiota dominated by relatively few taxa, which included members of the Enterobacterales, Lactobacillales, and Staphylococcales. In contrast, manure samples harbored much more diverse bacterial communities dominated by taxa within the Bacteroidales, Eubacteriales, and Oscillospirales. Beta diversity analyses likewise revealed that Stomoxys bacterial communities clustered separately from manure samples, suggesting some degree of environmental filtering of microorganisms in the fly. Similar patterns of microbial diversity, including both the divergence of host-associated bacterial communities from environmental food sources as well as the relative dominance of specific taxa across fly samples, have also been observed in other terrestrial dipteran species (38, 44–47).

We found that sequenced environmental and Stomoxys fly bacterial communities were largely similar across trap locations and collection dates for both facilities; however, there were observable differences in the relative abundances of specific bacterial taxa across sites. This variability could potentially be explained by either geographical location or differences in sample sizes due to the DCC being a smaller facility in an urbanized area. Cows housed at the DCC are also kept in individual tie-stalls requiring manual manure collection, while Arlington utilizes a free-stall barn system where manure is collected on motorized scraper systems. Cattle and waste management practices impact manure microbiota, which could explain observed differences between bacterial communities in Arlington and DCC manure samples (48–51). Within Arlington manure samples, we also observed a significant decrease in the abundance of Lactobacillales and Corynebacteriales from samples collected within the sick pen. Manure is not collected by a bulk scraper system in the sick pen and likely represents more recently deposited fecal samples. Differences in diversity could alternatively be attributed to host infection status, which can lead to dysbiosis and increased shedding of microbes in the gastrointestinal tract (52–55).

Despite differences in alpha and beta diversity, the vast majority of bacteria in fly samples could be traced to their corresponding manure samples, suggesting that, like other dipteran insects including mosquitoes, Stomoxys flies acquire their microbiota from the environments in which they breed (summarized in references 56–58). However, while our results support that flies visit manure and externally acquire bacteria, whether the internal adult microbiota is primarily shaped by adult feeding activities or transstadial transmission of bacteria from larvae, which develop in manure, is currently unknown. Laboratory studies of M. domestica and other dipteran insects indicate that, despite expulsion of midgut contents and remodeling of the larval gut to produce the adult gut during metamorphosis, a subset of the larval gut microbiota persists to the adult stage (59, 60–62). Members of the larval microbiota may also be re-acquired by adults when they ingest material from the larval habitat immediately following emergence, as has been demonstrated in adult mosquitoes (57). Additional work is therefore necessary to tease apart the relative contribution of transstadial transmission and adult feeding activities in shaping the Stomoxys microbiota.

Of particular significance in this study was the observation that shared ASVs contributed to a significant portion of reads in the fly microbiome but were found in relatively low abundance across manure samples, suggesting enrichment of specific manure-associated bacterial strains within the Stomoxys fly internal environment. Notably, shared ASVs enriched in fly microbial communities included multiple taxa (Staphylococcus, Pseudomonas, Enterococcus, and Enterobacterales) associated with environmental bovine mastitis. The high abundance of mastitis-associated lineages could be driven by physiological conditions within the dipteran digestive tract (pH, nutrient availability, etc.), which act as a strong selective pressure on microbial community assembly (58, 63). Spatial organization of the digestive tract, including colonization of the crop organ, may also allow for the persistence and proliferation of ingested microbes, as has been previously reported in M. domestica and Drosophila (23, 64). In mosquitoes (Diptera: Culicidae), hematophagous activities result in both a decrease in microbial diversity and the selective enrichment of Enterobacteriaceae (65, 66). Similar selective processes likely impact the Stomoxys fly microbiota and could explain the high dominance of Enterobacteriaceae in collected fly samples. We fully recognize that our microbiota data do not distinguish community members in adult Stomoxys flies that reside in the gut versus other tissues since our sequencing libraries were prepared from whole-body samples. This decision was driven in large part by the methodology used to collect flies in this study, which made explanting the gut with no possibility of contamination difficult. However, prior studies in related muscid flies and other dipteran insects have experimentally demonstrated that multiple species contain no bacteria as larvae before hatching from eggs, but rapidly acquire a gut microbiota from the environment in which they hatch by feeding (summarized in references 56, 57). This, combined with the fact that (i) almost all of the reads present in our surface-sterilized internal fly samples represented ASVs that were present in manure samples collected from the same facility on the same sampling date, and (ii) viable colonies of commonly shared taxa were readily isolated from internal fly samples, strongly suggests that most of the bacteria we identified in adult Stomoxys flies in the current study are gut community members.

In the present study, concurrent farm-specific mastitis incidence records showed that Staphylococcus and Enterobacteriaceae spp. were routinely isolated from mastitic cows housed in the same facilities from which fly and manure samples were collected. Although our study could not address whether fly-associated strains were genetically identical to isolates from these mastitis cases, the enrichment of bacterial lineages associated with bovine mastitis in the Stomoxys microbiota strongly suggests that flies act as bona fide carriers of taxa with the potential to serve as opportunistic bovine pathogens. The only exception to this was Streptococcus spp., which constitute important contagious and environmental bovine mastitis pathogens but were rare in Stomoxys fly-associated bacterial communities sequenced from both Arlington and the DCC. Additional work is necessary to establish whether this pattern is simply due to the low prevalence and relative abundance of Streptococcus spp. in these facilities or reflects an inability of at least some Streptococcus spp. to colonize and persist on or within Stomoxys flies. Additional work is also necessary to assess whether the enrichment of mastitis-associated taxa is consistent across flies collected from broader geographic regions, and to validate the dissemination of individual bacterial strains by flies within dairy barn environments. Such studies are particularly challenging for Stomoxys spp., due to the inability to maintain colonies in continuous culture in the laboratory.

The second goal of this study was to verify that fly-associated taxa represent viable and culturable bacterial isolates. In flies, enzymatic activity and other digestive processes can facilitate the lysis of bacteria within the gut (67, 68). The use of 16S rRNA sequencing data has been critical to our understanding of animal microbiomes; however, amplification of DNA can be from either viable or dead cells within samples (69, 70). Fly pools were therefore screened for culturable Klebsiella, E. coli, and Staphylococcus spp., which represent key mastitis-associated bacterial pathogens from highly dominant lineages across fly samples. Our results show that taxa associated with mastitis were not only observed in the sequence data but could be readily cultured from a subset of the corresponding fly homogenates. Taken together, these results confirm that Staphylococcus and Enterobacteriaceae are prominent members of the Stomoxys microbiota. The culture-dependent methodology is likely an underrepresentation of the true incidence rates as culturing was performed on an enriched aliquot of each fly homogenate pool. We also observed culturable isolates in a subsection of samples without corresponding sequence reads, which could indicate those taxa were rare in the unenriched homogenates used for DNA extraction and amplicon sequencing.

Collectively, our results provide evidence for a potential role for Stomoxys flies in the carriage of opportunistic pathogens in dairy barns. We found that mastitis-associated bacterial taxa were highly dominant across fly samples, but relatively rare across corresponding manure samples. Viable colonies of mastitis-associated taxa were also readily isolated from fly samples, indicating that the Stomoxys fly may represent an important, yet understudied, element of disease transmission in agricultural settings. Future work will elucidate the role of biting flies in the transmission of not only bovine mastitis pathogens, but other farm-associated zoonotic pathogens including E. coli O157, Brucella, and Salmonella (4, 5). Future work will also leverage these results to provide novel insights into the ecology and evolution of these and other dipteran insects more broadly, including species that serve important roles as bioindicators, biocontrol agents, sources of nutrition for other organisms, and/or nuisance pests, as well as other vectors of disease (56).

MATERIALS AND METHODS

Field sampling

Samples of Stomoxys flies and manure were collected on a weekly basis from July through September 2021 across two focal dairy farms in south central Wisconsin: the UW-Madison Dairy Cattle Center (DCC) and the Arlington Agricultural Research Station Blaine Dairy Cattle Center (Arlington) (Fig. 1). Stomoxys flies were caught on adhesive alsynite fiberglass traps (Olson Products, Medina, OH, USA), which selectively attract Stomoxys flies through reflection of ultraviolet light (71, 72). A single trap was set adjacent to an outdoor cow pen at the DCC, while four fly traps were set at Arlington outside the free-stall barn structures (Fig. 1). Traps at both farms were retrieved and replaced weekly. For manure samples, at least 25 mL of material was collected in a sterile conical tube for downstream processing (Thermo Fisher Scientific, Waltham, MA, USA). A total of five manure samples were collected each week from the DCC—one pooled from material collected along each quadrant of the associated tie-stall barn and one containing material collected from an outdoor pen (Fig. 1). Six manure samples were collected each week from Arlington—one pooled from material collected from each of five manure scraper systems and one containing material from an isolated pen housing sick cattle in the same facility (Fig. 1). Upon retrieval, all samples were transported directly to the laboratory on ice. Fly samples were immediately stored at −20°C until processing; manure samples were diluted 1:1 with sterile 1× phosphate-buffered saline (PBS) and vortexed to ensure homogenization before storing at −80°C (73).

Biting flies belonging to the genus Stomoxys were identified with the assistance of taxonomic keys available in the Manual of Nearctic Diptera (74). Whole bodies of flies from each trap were carefully retrieved from adhesive linings with surface-sterilized featherweight tweezers and pooled together into groups of up to 25 individuals. Pools were then vortexed gently for 40 s in 10 mL PBS-T (1× PBS + 0.01% Tween 80), and flies were removed before centrifugation of the PBS-T solutions to isolate the external fly microbiota. To isolate the internal fly microbiota, the same flies were subsequently surface sterilized with successive washes in 70% ethanol, 0.05% bleach, and water, and homogenized in groups of three to five flies by bead-beating with 3 × 5 mm² stainless steel beads (Qiagen, Hilden, Germany) in 1 mL 1× PBS. For each homogenate, a 40 µL aliquot was plated onto both tryptic soy agar (TSA) and brain heart infusion (BHI) agar plates before incubating at 30°C for 3 days. The resulting microbial growth was collected in 1 mL sterile 1× PBS and stored as a 40% glycerol stock at −80°C for use in bacterial isolations. The remainder of each homogenate was then centrifuged at 1,000 rcf (× g) to concentrate any remaining insect integument, which could interfere with DNA extractions and PCR amplification (75, 76), and supernatants were transferred to new tubes prior to centrifugation at 21,300 rcf (× g) for 20 min to pellet any cells. All cell pellets were stored at −20°C prior to DNA isolation and sequencing.

Bacterial 16S rRNA library construction and sequencing

Total genomic DNA was isolated from fly-derived cell pellets, 200 µL of homogenized manure samples, and associated extraction controls using a Qiagen DNeasy Blood and Tissue Kit prior to one-step PCR amplification of the V4 region of the 16S rRNA gene using barcoded primers as described previously (77). No-template reactions as well as reactions using templates from blank DNA extractions served as negative controls. With the exception of 16 external and 6 internal fly samples amplified at 35 cycles, PCR amplification was performed in 25 µL reactions under the following conditions: initial denaturation cycle of 95°C for 3 min, followed by 30 cycles at 95°C for 30 s, 55°C for 30 s and 72°C for 30 s, and a final extension step at 72°C for 5 min. PCR products were visualized on 1% agarose gels, purified using a MagJET NGS Cleanup and Size Selection Kit (Thermo Fisher Scientific, Waltham, MA, USA), and quantified with a Quantus fluorometer (Promega, Madison, WI, USA). Purified libraries were then combined in equimolar amounts prior to paired-end sequencing (2  ×  250 bp) on an Illumina MiSeq at the University of Wisconsin-Madison.

Sequence processing and data analysis

Paired-end demultiplexed sequences were imported into QIIME2 2022.2.0 for processing (78). Sequence quality scores were assessed and denoising was performed via DADA2 (79), followed by multiple sequence alignment and phylogenetic tree construction using Mafft and FastTree2, respectively (80, 81). Taxonomy was assigned using a naive-Bayes classifier natively implemented in Qiime and pre-trained against the SILVA reference database (82). The resulting Qiime2-generated taxonomy table, ASV table, and phylogenetic tree were then imported into R (version 4.1.3) and merged with sample metadata as phyloseq objects (83). Suspected contaminants were identified and removed from the ASV table using the “decontam” R package, which compares both the frequency and prevalence of ASVs between samples, extraction controls, and PCR controls (84). ASV reads classified as “Archaea,” “Chloroplast,” or “Mitochondria,” as well as samples containing fewer than 100 reads, were also removed prior to further analysis.

Data analysis was performed in R version 4.3.1 (http://www.r-project.org/), using “ggplot2” (85) for data visualization. Bacterial community richness was estimated using the weighted linear regression model of ASV richness estimates to calculate 95% confidence intervals (CIs) for group means using the betta() function in “breakaway” (86), interpreting only groups with non-overlapping 95% CIs. Beta diversity was visualized for Bray-Curtis dissimilarities (87) of relative abundance data using principal coordinates analysis (PCoA) via the ordinate() function in “phyloseq” (83). To test for a significant effect of sampling source or sampling date on community composition, we used permutational multivariate analysis of variance (PERMANOVA) to partition Bray-Curtis dissimilarity matrices among sources of variation using the adonis() function in “vegan” (88) and permutational multivariate analysis of dispersion (PERMDISP) to test for homogeneity of dispersion using the betadisper() function (89). Significant PERMANOVA and PERMDISP results were then subjected to post hoc pairwise comparisons, adjusting P values using the Benjamini-Hochberg False Discovery Rate (FDR) method (90) to identify significant differences between groups. Finally, the package “ANCOM-BC” (91) was used to identify significant changes in taxa abundance between sample groups, interpreting only taxa with an effect size (log fold change) > 1.5 and FDR-adjusted P value < 0.05 for a given group comparison.

Isolation of mastitis-associated bacterial taxa

Selective microbial growth media was used to isolate select clinically relevant bacterial taxa from fly-derived glycerol stocks. In brief, stocks were streaked onto MacConkey agar (lactose base) plates to enrich the growth of Gram-negative enteric bacteria. Presumptive E. coli colonies (i.e., showing both lactose fermentation and bile salt precipitation) were re-streaked onto eosin methylene blue (EMB) agar plates where they were verified as E. coli if a distinctive green, metallic sheen was observed (92). Klebsiella strains were isolated on MacConkey-inositol-carbenicillin agar, which allows for the enrichment and subsequent differentiation of Klebsiella spp. from other Enterobacteriaceae (93). Presumptive Klebsiella isolates were confirmed by comparison against the NCBI BLAST database after performing Sanger sequencing of the 16S rRNA gene using the universal 1492R primer (94). Mannitol salt agar was used as a primary selective media to enrich the growth of Staphylococcus spp. Presumptive Staphylococcus species were verified via colony PCR using the Staphylococcus-specific primers TStaG422 and TStag765 (95).