An official website of the United States government.
E. coli and several other closely related species of coliform bacteria have been used historically as indicator organisms to assess the microbiological risk and safety of food and water. Use of these species as indicators was predicated on their primary habitat being within the mammalian gastrointestinal tract and the fact that these species are easy to detect with simple microbiological methods. Recent ecological surveys of the GI microbiota, however, consistently show that the Enterobacteraceae, the taxonomic family to which these species belong, comprise only a minor fraction of the GI microbiota (less than 0.1%). Moreover, coliform genera within the Enterobacteraceae can also be found in certain terrestrial environments and are often epiphytes of plants. With the recent outbreaks of foodborne illness associated with leafy greens and other fresh vegetables and fruits, it is becoming increasingly clear that alternative or adjunct methods are needed to accurately assess contamination risks. The experiments outlined in this proposal will examine the validity of a new microbiological testing concept, called community profiling, which can be used as an alternative or adjunct to indicator organism testing. Community profiling is a non-culture-based approach which analyzes the composition of the entire microbial community from a sample and determines if the community is expected or predicted for that food matrix or if it bears species from fecal matter or other sources. The analysis is performed by PCR amplification of the 16S rDNA from total DNA extracted from a food sample, followed by massively parallel pyrosequencing and bioinformatics analysis of hundreds of thousands of the resulting PCR products. Using spinach as an experimental model, this experiments in this project will test two general criterion that must be met to validate the community profiling concept: i) the normal composition of the microbiota of a food matrix should be explained by simple temporal/spatial factors, and ii) fecal contamination or temperature abuse should result in community profiles that bear obvious signatures of fecal species or of abusive storage conditions.
NON-TECHNICAL SUMMARY: The indicator organism concept is based on use of a given microbial species or group of species as an indicator of the relative risk that a sample is contaminated with pathogenic species. Indicator organisms such as generic E. coli have traditionally been used to assess the relative microbiological risk that food or water is contaminated with fecal matter since the mammalian gastrointestinal tract is the primary habitat of this species. One of the problems with this approach, however, is the fact that the species E. coli is only a very minor fraction of the total mass of microorganisms present in the mammalian GI tract, making up only less than 0.1% of the total population. While other more abundant species would certainly provide more sensitivity, most currently accepted microbiological methods rely on culturing of the organisms from the sample and the vast majority of species in the GI tract are difficult, if not impossible to culture routinely. We therefore propose to test the validity of a new non-culture-based approach which combines the powerful PCR-based amplification of target genes from each microbe in a sample, coupled with massively parallel DNA sequencing of hundreds of thousands of these target genes to enumerate the different species that are present. This approach therefore provides a profile of the entire community of microorganisms in a sample, allowing the analyst to determine if the community of microorganisms present is comprised of those that are expected from that type of food or if the community contains species that are characteristic of other environment such as feces or soil. The experiments in this proposal are designed to test two criterion that must be met in order to validate this community profiling concept. First, the microbial community of a given commodity must be predictable and distinguishable from fecal and soil communities. Second, the approach must be able to detect the presence of species from feces when present. These criteria will be evaluated using spinach as an experimental model with plants derived from different cultivars, climates, and growing seasons. <P> APPROACH: The composition of the microbial communities will be assessed by washing bacteria from the surface of spinach leaves, followed by extraction of total DNA and 454-based pyrosequencing of PCR products amplified from the V1 region of the 16S rRNA genes. The V1 region will be amplified from the total DNA extract using modified versions of the A8 and B357 PCR primers, which amplify the first 350 bases of the 16S rRNA gene. The modified primers will contain the Roche A or B sequencing adapter sequence upstream of the A8 and B357 sequence. The A primer will also contain an additional 8-base bar-code such that multiple combinations of A primers can be used to amplify individual samples and the resulting PCR products will all carry a bar-code unique to that sample. For pyrosequencing, we will pool the PCR products from 10 different samples followed by attachment of the individual amplicons from the pool to magnetic capture beads for emulsion PCR (emPCR). Purified beads from the emPCR are then deposited into the wells of a pico-titre plate along with pyrosequencing enzymes. Successive flows of nucleotides are then deposited over the surface of the plate on the Roche-454 GS-FLX Instrument with a CCD camera capturing the light flashes produced during each flow cycle. The order of flashes from each bead that occur during the nucleotide flows is then determined by the on-instrument software and converted to DNA sequence for each of the beads. Each run of the instrument produces sequence form hundreds of thousands of the original 16S rRNA amplicons from the pooled PCR reactions, with about 40,000 sequences being obtained from each of the pooled samples. Bioinformatics analysis of the resulting sequences then allows taxonomic assignment of each sequence and simple binning of the sequence reads by taxonomic rank (e.g. Phylum, Class, Order, Family, Genus) then quantitatively reveals the composition of the community. For objective 1, a factorial treatment design will be used to account for covariation among multiple dependent variables (species of microbiota) and multiple independent variables (spinach cultivar/variety, geography, climactic conditions). A total of 96 samples will be used for a single replication of the factorial with built-in technical replicates to evaluate heterogeneity within a single spinach sample. To test the second criterion (objective 2), subsamples of the spinach will be spiked with cattle manure (fecal contamination) or stored at room temperature versus 4 degrees C (temperature abuse). One replication of the experiment will be conducted with a factorial design with a total factorial of 96. Because the unspiked and non-temperature abused control samples in this part of the study are subsets of the same samples from the first objective, pyrosequencing analysis need only be performed on the spiked or abused samples.