Application of quasimetagenomics methods to define microbial diversity and subtype Listeria monocytogenes in dairy and seafood production facilities

IMPORTANCE

In developed countries, the human diet is predominated by food commodities, which have been manufactured, processed, and stored in a food production facility. Little is known about the application of metagenomic sequencing approaches for detecting foodborne pathogens, such as L. monocytogenes, and characterizing microbial diversity in food production ecosystems. In this work, we investigated the utility of 16S rRNA amplicon and quasimetagenomic sequencing for the taxonomic and phylogenetic classification of Listeria culture enrichments of environmental swabs collected from dairy and seafood production facilities. We demonstrated that single-nucleotide polymorphism (SNP) analyses of L. monocytogenes metagenome-assembled genomes (MAGs) from quasimetagenomic data sets can achieve similar resolution as culture isolate whole-genome sequencing. To further understand the impact of genome coverage on MAG SNP cluster resolution, an in silico downsampling approach was employed to reduce the percentage of target pathogen sequence reads, providing an initial estimate of required MAG coverage for subtyping resolution of L. monocytogenes.

Use of Whole Genome Sequencing by the Federal Interagency Collaboration for Genomics for Food and Feed Safety in the United States

ABSTRACT

This multiagency report developed by the Interagency Collaboration for Genomics for Food and Feed Safety provides an overview of the use of and transition to whole genome sequencing (WGS) technology for detection and characterization of pathogens transmitted commonly by food and for identification of their sources. We describe foodborne pathogen analysis, investigation, and harmonization efforts among the following federal agencies: National Institutes of Health; Department of Health and Human Services, Centers for Disease Control and Prevention (CDC) and U.S. Food and Drug Administration (FDA); and the U.S. Department of Agriculture, Food Safety and Inspection Service, Agricultural Research Service, and Animal and Plant Health Inspection Service. We describe single nucleotide polymorphism, core-genome, and whole genome multilocus sequence typing data analysis methods as used in the PulseNet (CDC) and GenomeTrakr (FDA) networks, underscoring the complementary nature of the results for linking genetically related foodborne pathogens during outbreak investigations while allowing flexibility to meet the specific needs of Interagency Collaboration partners. We highlight how we apply WGS to pathogen characterization (virulence and antimicrobial resistance profiles) and source attribution efforts and increase transparency by making the sequences and other data publicly available through the National Center for Biotechnology Information. We also highlight the impact of current trends in the use of culture-independent diagnostic tests for human diagnostic testing on analytical approaches related to food safety and what is next for the use of WGS in the area of food safety.

HIGHLIGHTS

Microbiome Population Dynamics of Cold-Smoked Sockeye Salmon during Refrigerated Storage and after Culture Enrichment

ABSTRACT

Cold-smoked salmon is a ready-to-eat seafood product of high commercial importance. The processing and storage steps facilitate the introduction, growth, and persistence of foodborne pathogens and spoilage bacteria. The growth of commensal bacteria during storage and once the product is opened also influence the quality and safety of cold-smoked salmon. Here we investigated the microbial community through targeted 16S rRNA gene and shotgun metagenomic sequencing as means to better understand the interactions among bacteria in cold-smoked salmon. Cold-smoked salmon samples were tested over 30 days of aerobic storage at 4°C and cultured at each time point in a buffered Listeria enrichment broth (BLEB) commonly used to detect Listeria in foods. The microbiomes were composed of Firmicutes and Proteobacteria, namely, Carnobacterium, Brochothrix, Pseudomonas, Serratia, and Psychrobacter. Pseudomonas species were the most diverse species, with 181 taxa identified. In addition, we identified potential homologs to 10 classes of bacteriocins in microbiomes of cold-smoked salmon stored at 4°C and corresponding BLEB culture enrichments. The findings presented here contribute to our understanding of microbiome population dynamics in cold-smoked salmon, including changes in bacterial taxa during aerobic cold storage and after culture enrichment. This may facilitate improvements to pathogen detection and quality preservation of this food.

HIGHLIGHTS

Research Note: Effect of butyric acid glycerol esters on ileal and cecal mucosal and luminal microbiota in chickens challenged with Eimeria maxima

Coccidiosis is one of the most prevalent diseases seen in the poultry industry leading to excessive economic losses. The aim of this study was to investigate the effect of butyric acid glycerol esters (BE) on the ileal and cecal microbiota in birds challenged with Eimeria maxima (EM). Ross 708 male broilers were fed a diet supplemented with 0 (control) or 0.25% BE from day 1. On day 21, half of the birds were infected with 10³ EM oocysts. For determing microbiota, ileal and cecal contents and epithelial scrapings were collected at 7 and 10 D postinfection (PI). Alpha diversity of bacterial communities was mostly affected (P < 0.05) by time PI and EM infection. The richness of luminal bacterial populations in the ileum and ceca was affected (P < 0.05) by addition of BE and by time PI × EM × BE interaction, respectively. In the ileal and cecal luminal and mucosal bacterial communities, permutational multivariate analysis of variance (PERMANOVA, unweighted UniFrac) showed significant (P < 0.05) differences because of time PI and interaction between time PI, EM, and BE. Significant (P < 0.05) differences in taxonomic composition at the family level were observed in microbiota of luminal and mucosal populations of the ileum and ceca owing to time PI, EM, BE, and their interactions. The bacterial community present in the cecal lumen was characterized by the lowest number of differential bacteria, whereas the cecal mucosal community was characterized by the highest number of differentially abundant bacteria. In conclusion, our results show that EM infection and time PI has the biggest impact on microbial diversity in the chicken gut. The presence of BE in the diet had a limited effect on gut microbiota.

Microbiomes Associated With Foods From Plant and Animal Sources

Food microbiome composition impacts food safety and quality. The resident microbiota of many food products is influenced throughout the farm to fork continuum by farming practices, environmental factors, and food manufacturing and processing procedures. Currently, most food microbiology studies rely on culture-dependent methods to identify bacteria. However, advances in high-throughput DNA sequencing technologies have enabled the use of targeted 16S rRNA gene sequencing to profile complex microbial communities including non-culturable members. In this study we used 16S rRNA gene sequencing to assess the microbiome profiles of plant and animal derived foods collected at two points in the manufacturing process; post-harvest/pre-retail (cilantro) and retail (cilantro, masala spice mixes, cucumbers, mung bean sprouts, and smoked salmon). Our findings revealed microbiome profiles, unique to each food, that were influenced by the moisture content (dry spices, fresh produce), packaging methods, such as modified atmospheric packaging (mung bean sprouts and smoked salmon), and manufacturing stage (cilantro prior to retail and at retail). The masala spice mixes and cucumbers were comprised mainly of Proteobacteria, Firmicutes, and Actinobacteria. Cilantro microbiome profiles consisted mainly of Proteobacteria, followed by Bacteroidetes, and low levels of Firmicutes and Actinobacteria. The two brands of mung bean sprouts and the three smoked salmon samples differed from one another in their microbiome composition, each predominated by either by Firmicutes or Proteobacteria. These data demonstrate diverse and highly variable resident microbial communities across food products, which is informative in the context of food safety, and spoilage where indigenous bacteria could hamper pathogen detection, and limit shelf life.

High-Resolution Microbiome Profiling for Detection and Tracking of Salmonella enterica

16S rRNA community profiling continues to be a useful tool to study microbiome composition and dynamics, in part due to advances in next generation sequencing technology that translate into reductions in cost. Reliable taxonomic identification to the species-level, however, remains difficult, especially for short-read sequencing platforms, due to incomplete coverage of the 16S rRNA gene. This is especially true for Salmonella enterica, which is often found as a low abundant member of the microbial community, and is often found in combination with several other closely related enteric species. Here, we report on the evaluation and application of Resphera Insight, an ultra-high resolution taxonomic assignment algorithm for 16S rRNA sequences to the species level. The analytical pipeline achieved 99.7% sensitivity to correctly identify S. enterica from WGS datasets extracted from the FDA GenomeTrakr Bioproject, while demonstrating 99.9% specificity over other Enterobacteriaceae members. From low-diversity and low-complexity samples, namely ice cream, the algorithm achieved 100% specificity and sensitivity for Salmonella detection. As demonstrated using cilantro and chili powder, for highly complex and diverse samples, especially those that contain closely related species, the detection threshold will likely have to be adjusted higher to account for misidentifications. We also demonstrate the utility of this approach to detect Salmonella in the clinical setting, in this case, bloodborne infections.

Early Recovery of Salmonella from Food Using a 6-Hour Non-selective Pre-enrichment and Reformulation of Tetrathionate Broth

Culture based methods are commonly employed to detect pathogens in food and environmental samples. These methods are time consuming and complex, requiring multiple non-selective and selective enrichment broths, and usually take at least 1 week to recover and identify pathogens. Improving pathogen detection in foods is a primary goal for regulatory agencies and industry. Salmonella detection in food relies on a series of culture steps in broth formulations optimized to resuscitate Salmonella and reduce the abundance of competitive bacteria. Examples of non-selective pre-enrichment broths used to isolate Salmonella from food include Lactose, Universal Pre-enrichment, BPW, and Trypticase Soy broths. Tetrathionate (TT) and Rappaport-Vassiliadis (RV) broths are employed after a 24-h non-selective enrichment to select for Salmonella and hamper the growth of competitive bacteria. In this study, we tested a new formulation of TT broth that lacks brilliant green dye and has lower levels of TT . We employed this TT broth formulation in conjunction with a 6-h non-selective pre-enrichment period and determined that Salmonella recovery was possible one day earlier than standard food culture methods. We tested the shortened culture method in different non-selective enrichment broths, enumerated Salmonella in the non-selective enrichments, and used 16S rRNA gene sequencing to determine the proportional abundances of Salmonella in the TT and RV selective enrichments. Together these data revealed that a 6-h non-selective pre-enrichment reduces the levels of competitive bacteria inoculated into the selective TT and RV broths, enabling the recovery of Salmonella 1 day earlier than standard culture enrichment methods.

Recent and emerging innovations in Salmonella detection: a food and environmental perspective

Salmonella is a diverse genus of Gram‐negative bacilli and a major foodborne pathogen responsible for more than a million illnesses annually in the United States alone. Rapid, reliable detection and identification of this pathogen in food and environmental sources is key to safeguarding the food supply. Traditional microbiological culture techniques have been the ‘gold standard’ for State and Federal regulators. Unfortunately, the time to result is too long to effectively monitor foodstuffs, especially those with very short shelf lives. Advances in traditional microbiology and molecular biology over the past 25 years have greatly improved the speed at which this pathogen is detected. Nonetheless, food and environmental samples possess a distinctive set of challenges for these newer, more rapid methodologies. Furthermore, more detailed identification and subtyping strategies still rely heavily on the availability of a pure isolate. However, major shifts in DNA sequencing technologies are meeting this challenge by advancing the detection, identification and subtyping of Salmonella towards a culture‐independent diagnostic framework. This review will focus on current approaches and state‐of‐the‐art next‐generation advances in the detection, identification and subtyping of Salmonella from food and environmental sources.

Cilantro microbiome before and after nonselective pre-enrichment for Salmonella using 16S rRNA and metagenomic sequencing

BACKGROUND

Salmonella enterica is a common cause of foodborne gastroenteritis in the United States and is associated with outbreaks in fresh produce such as cilantro. Salmonella culture-based detection methods are complex and time consuming, and improvments to increase detection sensitivity will benefit consumers. In this study, we used 16S rRNA sequencing to determine the microbiome of cilantro. We also investigated changes to the microbial community prior to and after a 24-hour nonselective pre-enrichment culture step commonly used by laboratory analysts to resuscitate microorganisms in foods suspected of contamination with pathogens. Cilantro samples were processed for Salmonella detection according to the method in the United States Food and Drug Administration Bacteriological Analytical Manual. Genomic DNA was extracted from culture supernatants prior to and after a 24-hour nonselective pre-enrichment step and 454 pyrosequencing was performed on 16S rRNA amplicon libraries. A database of Enterobacteriaceae 16S rRNA sequences was created, and used to screen the libraries for Salmonella, as some samples were known to be culture positive. Additionally, culture positive cilantro samples were examined for the presence of Salmonella using shotgun metagenomics on the Illumina MiSeq.

RESULTS

Time zero uncultured samples had an abundance of Proteobacteria while the 24-hour enriched samples were composed mostly of Gram-positive Firmicutes. Shotgun metagenomic sequencing of Salmonella culture positive cilantro samples revealed variable degrees of Salmonella contamination among the sequenced samples.

CONCLUSIONS

Our cilantro study demonstrates the use of high-throughput sequencing to reveal the microbiome of cilantro, and how the microbiome changes during the culture-based protocols employed by food safety laboratories to detect foodborne pathogens. Finding that culturing the cilantro shifts the microbiome to a predominance of Firmicutes suggests that changing our culture-based methods will improve detection sensitivity for foodborne enteric pathogens.

Cronobacter: an emergent pathogen causing meningitis to neonates through their feeds

The recognition of Cronobacter as a public health concern was raised when powdered infant formula (PIF) was linked to several neonatal meningitis outbreaks. It is an opportunistic pathogen that causes necrotising enterocolitis, infantile septicaemia, and meningitis which carries a high mortality rate among neonates. It has been also linked with cases of infection in adults and elderly. Over the past decade, much focus has been made on developing sensitive and specific characterisation, detection, and isolation methods to ascertain the quality of foods, notably contamination of PIF with Cronobacter and to understand its ability to cause disease. Whole genome sequencing has unveiled several putative virulence factors, yet the full capacity of the pathogenesis of Cronobacter has not yet been elucidated.

Pan-genome analysis of the emerging foodborne pathogen Cronobacter spp. suggests a species-level bidirectional divergence driven by niche adaptation

Background

Members of the genus Cronobacter are causes of rare but severe illness in neonates and preterm infants following the ingestion of contaminated infant formula. Seven species have been described and two of the species genomes were subsequently published. In this study, we performed comparative genomics on eight strains of Cronobacter, including six that we sequenced (representing six of the seven species) and two previously published, closed genomes.

Results

We identified and characterized the features associated with the core and pan genome of the genus Cronobacter in an attempt to understand the evolution of these bacteria and the genetic content of each species. We identified 84 genomic regions that are present in two or more Cronobacter genomes, along with 45 unique genomic regions. Many potentially horizontally transferred genes, such as lysogenic prophages, were also identified. Most notable among these were several type six secretion system gene clusters, transposons that carried tellurium, copper and/or silver resistance genes, and a novel integrative conjugative element.

Conclusions

Cronobacter have diverged into two clusters, one consisting of C. dublinensis and C. muytjensii (Cdub-Cmuy) and the other comprised of C. sakazakii, C. malonaticus, C. universalis, and C. turicensis, (Csak-Cmal-Cuni-Ctur) from the most recent common ancestral species. While several genetic determinants for plant-association and human virulence could be found in the core genome of Cronobacter, the four Cdub-Cmuy clade genomes contained several accessory genomic regions important for survival in a plant-associated environmental niche, while the Csak-Cmal-Cuni-Ctur clade genomes harbored numerous virulence-related genetic traits.

Antibodies to intimin and Escherichia coli secreted proteins A and B in patients with enterohemorrhagic Escherichia coli infections

Enterohemorrhagic Escherichia coli produce an attaching and effacing lesion upon adhering to the intestinal epithelium. Bacterial factors involved in this histopathology include the intimin adhesin and E. coli secreted proteins (Esps) A and B. In this study we investigated the serum antibody responses to recombinant E. coli O157:H7 intimin, EspA, and EspB by immunoblotting. Canadian patients with O157:H7 infection (n=10), Swedish patients with O157:H7 (n=21), non-O157 (n=18), or infection from which the serotype was not available (n=3), and asymptomatic household members (n=25) were studied and compared with Canadian (n=20) and Swedish controls (n=52). In Canadian patients, IgG antibodies to intimin, EspA, and EspB were analyzed, in Swedish patients and their household members IgA, IgG, and IgM antibodies to EspA and EspB were studied. Patients and household members mounted an antibody response to the antigens. Significantly more patients developed an acute response to EspB compared with controls (P<0.01 Canadian patients, P<0.0001 Swedish patients). EspB IgA, IgG, and IgM had a specificity of 100%, 86%, and 86%, positive predictive value of 100%, 83%, and 81%, and sensitivity of 57%, 69%, and 63%, respectively, and appear to be an appropriate assay for the detection of EHEC infection. In cases of hemolytic uremic syndrome or hemorrhagic colitis this assay may be useful when a fecal strain has not been isolated, or in epidemics of non-O157 infection.

Identification of lngA, the structural gene of longus type IV pilus of enterotoxigenic Escherichia coli

Human enterotoxigenic Escherichia coli (ETEC) produces a type IV pilus termed longus which is encoded on large plasmids in association with colonization factor antigens (CFAs) and enterotoxins. A plasmid-derived 7 kbp BamHI DNA fragment hybridizing with an oligonucleotide probe designed from the amino-terminal amino acid sequence of the denatured 22 kDa structural longus pilin subunit was subcloned and sequenced. DNA sequencing analysis revealed an open reading frame, designated lngA, whose predicted amino acid sequence matched perfectly the N-terminal sequence of LngA obtained by Edman degradation. lngA is the first gene described of the longus gene cluster. Cloned lngA encoded and expressed a prepilin protein of 236 residues with a calculated mass of 25.17 kDa. The prepilin is apparently processed into a mature pilin of 206 residues with a calculated mass of 21.5 kDa. The predicted peptide sequence of lngA showed 78.8 and 37% identity to CFA/III pilin (CofA) of ETEC and the toxin-coregulated pilus (TcpA) of Vibrio cholerae. Peptide sequence homology between lngA and cofA was more prominent towards the amino terminus than within the carboxy region. Like other type IV pilins, LngA contains two cysteine residues towards the carboxy-terminal region. Transmission electron microscopy and immunoblot analysis of ETEC strains expressing either longus or CFA/III detected antigenic differences between native and denatured epitopes of these pili. In addition, differential regulation of pilus expression was identified when ETEC strains were grown in different media. Our data indicate that longus and CFA/III are two distinct but yet highly related type IV pili of ETEC.

Longus pilus of enterotoxigenic Escherichia coli and its relatedness to other type-4 pili--a minireview

Longus is a long pilus produced by human enterotoxigenic Escherichia coli (ETEC) which shares significant structural and biochemical features with class-B type-4 pili. These pili include the toxin-coregulated pilus (TCP) of Vibrio cholerae, the bundle-forming pilus (BFP) of enteropathogenic E. coli and both longus and the colonization factor antigen III (CFA/III) of ETEC. These pili are produced under defined growth conditions indicating that they are under the control of different regulatory elements. While TCP is chromosomally encoded, the remaining pili are encoded on large virulence plasmids. Longus and CFA/III are closely related pili although certain DNA and protein differences also exist between them. This may account for the differences in the regulation, surface presentation, antigenicity, and prevalence of these two pilins among ETEC. Neighboring lngA, a second open reading frame termed lngB was found which encodes a protein with significant homology to proteins which are part of a type-II secretory system such as XcpV, OutC, and PulO of Pseudomonas aeruginosa, Erwinia chrysanthemi, and Klebsiella pneumoniae, respectively. This suggests that lngB may be an accessory gene involved in biogenesis of longus.

Genetics of virulence of enteropathogenic E. coli

Enteropathogenic E. coli (EPEC) are a major cause of diarrhea in infants throughout the world. Although this pathogen was described 50 years ago, it is only recently that the pathogenic mechanisms employed by this organism have been elucidated. The characteristic histopathology induced by this organism, called "attaching and effacing", consists of intimate adherence of the bacterium to the epithelial cell with marked cytoskeletal changes including effacement of microvilli. A 35 kb region of chromosomal DNA, called the LEE for locus of enterocyte effacement has recently been described which contains all known genes necessary for production of this characteristic histopathology. Within this region is the eae gene encoding intimin, a 94 kDa OMP involved in intimate adherence. Also within this region are genes encoding proteins secreted extracellularly by EPEC (esp) and a type III secretion apparatus (sep) which shares homology with similar systems in Yersinia, Shigella, and Salmonella. Additional genes on a 60 MDa plasmid encode a type IV pilus (BFP) and a positive transcriptional activator (per) of multiple chromosomal and plasmid virulence genes.

Secretion of extracellular proteins by enterohemorrhagic Escherichia coli via a putative type III secretion system

Enterohemorrhagic and enteropathogenic Escherichia coli (EHEC and EPEC) infections result in attaching and effacing lesions on intestinal epithelial cells. Secretion of extracellular proteins via a type III secretion apparatus is necessary for the formation of attaching and effacing lesions by EPEC. We now show that EHEC also secretes polypeptides via a putative type III secretion system. The secreted EHEC proteins are recognized by rabbit antiserum raised against the proteins secreted from EPEC and by human serum from a patient infected with an EHEC O157:H7 strain.

Enteropathogenic Escherichia coli contains a putative type III secretion system necessary for the export of proteins involved in attaching and effacing lesion formation

Enteropathogenic Escherichia coli (EPEC) causes a characteristic histopathology in intestinal epithelial cells called the attaching and effacing lesion. Although the histopathological lesion is well described the bacterial factors responsible for it are poorly characterized. We have identified four EPEC chromosomal genes whose predicted protein sequences are similar to components of a recently described secretory pathway (type III) responsible for exporting proteins lacking a typical signal sequence. We have designated the genes sepA, sepB, sepC, and sepD (sep, for secretion of E. coli proteins). The predicted Sep polypeptides are similar to the Lcr (low calcium response) and Ysc (yersinia secretion) proteins of Yersinia species and the Mxi (membrane expression of invasion plasmid antigens) and Spa (surface presentation of antigens) regions of Shigella flexneri. Culture supernatants of EPEC strain E2348/69 contain several polypeptides ranging in size from 110 kDa to 19 kDa. Proteins of comparable size were recognized by human convalescent serum from a volunteer experimentally infected with strain E2348/69. A sepB mutant of EPEC secreted only the 110-kDa polypeptide and was defective in the formation of attaching and effacing lesions and protein-tyrosine phosphorylation in tissue culture cells. These phenotypes were restored upon complementation with a plasmid carrying an intact sepB gene. These data suggest that the EPEC Sep proteins are components of a type III secretory apparatus necessary for the export of virulence determinants.

A genetic locus of enterocyte effacement conserved among diverse enterobacterial pathogens

Enteropathogenic Escherichia coli (EPEC) and enterohemorrhagic E. coli O157:H7 are intestinal pathogens that profoundly damage the microvilli and subapical cytoskeleton of epithelial cells. Here we report finding in EPEC a 35-kbp locus containing several regions implicated in formation of these lesions. DNA probes throughout this locus hybridize to E. coli O157:H7 and other pathogens of three genera that cause similar lesions but do not hybridize to avirulent members of the same species. The EPEC locus and a different virulence locus of uropathogenic E. coli insert into the E. coli chromosome at the identical site and share highly similar sequences near the point of insertion.

Isogenic P-fimbrial deletion mutants of pyelonephritogenic Escherichia coli: the role of alpha Gal(1-4) beta Gal binding in virulence of a wild-type strain

Escherichia coli strains causing acute pyelonephritis often express multiple fimbrial types and haemolysin, which may contribute to their ability to adhere to, and interact with, kidney epithelial cells. Strain CFT073, a pap+, sfa+, pil+, hly+ pyelonephritis strain, previously established as virulent in the CBA mouse model of ascending urinary tract infection and cytotoxic for cultured human renal epithelial cells, was selected for construction of isogenic strains. From a gene bank of this strain, two distinct copies of the pap operon were isolated. The two P-fimbrial determinants were subcloned into pCVD442, a positive selection suicide vector containing the sacB gene of Bacillus subtilis. Deletion mutations were introduced into each of the two constructs, within papEFG of one operon and papDEFG of the other. Suicide vectors carrying pap deletions were mobilized from E. coli SM10 lambda pir into CFT073 (NalR) and cointegrates were passaged on non-selective medium. The first pap mutation was identified by screening a Southern blot of DNA from sucrose-resistant colonies using a papEFG probe. This mutant retained the MRHA+ phenotype since a second functional copy of pap was still present. A double pap-deletion mutant, UPEC76, confirmed by Southern blotting, was unable to agglutinate human type O erythrocytes or alpha Gal(1-4)beta Gal-coated latex beads. CBA mice (N = 100) were challenged transurethrally with 10(5), 10(6), 10(7), or 10(9) cfu of strains CFT073 or UPEC76. After one week, quantitative cultures of urine, bladder, and kidney were done and histologic changes were examined. No substantive differences in organism concentration or histological findings between parent and mutant were detected in urine, bladder, or kidney at any challenge concentration. We conclude that adherence by P fimbriae of uropathogenic E. coli strain CFT073 plays only a subtle role in the development of acute pyelonephritis in the CBA mouse model.

Distribution of the bundle-forming pilus structural gene (bfpA) among enteropathogenic Escherichia coli

Enteropathogenic Escherichia coli (EPEC) express an inducible bundle-forming pilus (BFP) associated with the presence of the EPEC adherence factor (EAF) plasmid and localized adherence (LA) on HEp-2 cells. The cloned structural gene (bfpA) encoding BFP was found to be specific for EPEC, as homologous sequences were found only in EPEC and not in other enteropathogens. The bfpA probe was slightly more sensitive than the EAF probe; among EPEC strains with LA, the bfpA and EAF probes hybridized with 99% and 96% of the strains, respectively. Immunoblotting of whole cell lysates of BFP-positive, EAF-positive or -negative, and LA-positive or -negative EPEC strains revealed variations in the size (18,500-21,000) of the expressed structural subunit of BFP, suggesting differences in processing that may account for discrepancies between the bfpA, EAF, and LA. Because the bfpA probe consists primarily of sequences encoding an important EPEC virulence factor, in contrast to the unknown function of sequences contained in the EAF probe, the bfpA probe may be useful in epidemiologic studies to detect EPEC.