Skip to main content
ARS Home » Pacific West Area » Albany, California » Western Regional Research Center » Produce Safety and Microbiology Research » Research » Research Project #430338

Research Project: Ecology and Detection of Human Pathogens in the Produce Production Continuum

Location: Produce Safety and Microbiology Research

2017 Annual Report


1a. Objectives (from AD-416):
1:Elucidate biological factors and molecular mechanisms that enhance or reduce fitness characteristics related to survival and growth of enteric pathogens in the produce production continuum.(A)Phenotypic characterization of shiga toxin-producing E. coli O111 strains of environmental & clinical origins & (B)traits contributing to fitness of Salmonella strains on pre-harvest lettuce.(C)Evaluate formation of STEC persisters & factors contributing to population size in produce production continuum.(D)Compare fitness & physiology of different STEC serotypes & associated indigenous microbiota on produce growing in conventional vs organic soil.(E)EcO157 biofilm formation under stress conditions in MAP lettuce.(F)Identify genes involved in attachment of L. monocytogenes to produce & role of biofilm formation in survival on produce.(G)Evaluate interaction of EcO157 with various lettuce accessions.(H) Investigate evolution of EcO157 in MAP lettuce & (I)interactions between norovirus & native plant-associated bacteria. 2:Identify environmental factors that affect the persistence and transmission of enteric pathogens in the produce production environment for risk assessment.(A)Develop & deploy quantitative assessment of enteric pathogens in surface water in several watersheds (WS)in Salinas region to enhance current incidence data supplied to FDA for risk assessment model.(B)Determine survival & fitness characteristics of enteric pathogens in water & sediment samples from central California coastal (cCc) produce production regions & (C)the impact of indigenous microbial community on environmental persistence of STECs in sediment from WS locations often positive for STEC strains in Salinas region using metagenomics approach.(D)Comparative genomics & transcriptomics applied to characterize environmental Shiga toxin-producing E. coli. (E)Investigate survival & persistence of Salmonella & L. monocytogenes in Central California WS to measure prevalence, identify subtypes & regional traits.(F)Characterize F+ RNA Coliphages in water samples from cCc produce production regions & determine feasibility as indicators for source-tracking enteric pathogens & (G)enteric pathogens transported through aerosols & determine seasonal fluctuations in cCc produce production regions.(H)Determine if wild pigs are hosts for human norovirus (HuNoV). 3:Develop methods for the detection and subtyping of enteric bacterial and viral pathogens from produce production environments; to aid epidemiological investigations and to distinguish pathogenic from non-pathogenic strains. (A)Develop FT-IR identification libraries for serotype analysis.(B)Improve STEC subtyping using sequence of CRISPR & other hypervariable genes to augment existing MLVA method.(C)Examine bacterial & produce-associated factors impacting genome & STEC virulence evolution.(D)Develop better assay to determine inactivation status of Tulane virus (TV) & HuNoV caused by viral genomic damage & (E)accurate, fast, & low-cost multi-amplicon in situ capture qRT-PCR assay determining HuNoV infectivity. 4:Study the ecology of Shiga toxin-producing E. coli (STEC) bacteriophages and its association with bacterial hosts.


1b. Approach (from AD-416):
Plant-microbe model systems in combination with population studies, ecology, molecular methods, genomics, and microbiology will be used to investigate the interaction of human bacterial and viral pathogens with plants and plant-associated bacteria, as well as to develop improved methods for detection and subtyping of human on produce. Pathogenic E. coli is a foodborne pathogen that has been linked to numerous outbreaks of foodborne illnesses, and the illnesses are primarily attributed to the ingestion of Shiga toxin-producing E. coli (STEC). Previous research has indicated the virulence markers such as stx genes, of STEC strains are conferred to stx-encoding bacteriophages and can be transduced into the susceptible bacterial hosts. In order to understand the interplay between STEC-specific phages and their bacterial hosts in the environment to enhance the safety of food products and the prevention of new emerging foodborne pathogens, the initial focuses of the phage research are to isolate, collect and characterize STEC phages and to understand the relationship between phages and their hosts in the environment. Efficient methods for isolation of STEC bacteriophages will be utilized. Characterization of STEC bacteriophages will be established using genomic sequencing and proteomic analyses. The association of fecal contamination with the population of STEC bacteriophages in the environment will be determined. Environmental factors that influence the geographical distribution of STEC bacteriophages will be identified. This will establish a foundation to study biological interactions between phages and their hosts and the association of phages with bacterial evolution as well as to utilize collected phages to develop biosensors and pre-harvest biological controls for STEC to improve the microbiological food safety of the food supplies.


3. Progress Report:
Under Objective 1, fitness and survival of enteric pathogens in produce production environments continue to be assessed. Environmental and outbreak Shiga toxin producing Escherichia coli (STEC) O111 strains were screened for carbon source utilization and compared for growth in lettuce lysates to determine the nutrition patterns of these strains. STEC O111 strains that produce less RpoS, which mediates the stress response, were more efficient users of carbohydrates present in high levels in/on plants, and they multiply more rapidly in lettuce lysates (Sub-obj. 1A). Two projects (mutator phenotype in Salmonella, Sub-obj. 1B, and persister cells of STEC, Sub-obj. 1C) assess the potential of sub-populations of pathogens to tolerate stress in the food production environment and survive to cause illness. We screened over 600 Salmonella strains for mutator phenotypes. Six candidates are in hand for fitness studies on lettuce. In STEC studies we optimized the procedure for quantification of STEC persisters, and tested four STEC outbreak strains in field water, spinach leaf wash water, spinach lysates, and lettuce phylloplane. Increases in persister populations were noted in culture and in growth on field and plant wash waters, indicating that persistence may be involved in STEC presence on produce. A manuscript on the STEC persister work is in preparation. Indigenous microbes present in soil may affect survival of pathogens. In collaboration with scientists at the University of California, Davis, we continued to monitor survival of STEC’s on plants grown in soils managed as organic, mixed, or conventional with different irrigation methods (Sub-obj. 1D). Soils were collected at three points during the growing season (cover crop to cash crop) from farm plots with specific amendments. As in the previous year, indigenous microbial activity was two to four times higher in organic soils, but STEC survival varied depending on soil type and irrigation method. Additional data in the third year will be needed to establish a significant correlation to specific STEC serotypes. Several projects that assess how bacteria interact with plants and tolerate stress are ongoing. Competitive fitness, studies on plants, and chemical analysis of lettuce wounds revealed that E. coli O157:H7 utilizes plant choline to produce an osmoprotectant, glycine betaine, and a mutant that cannot make glycine betaine has lower survival in injured or processed romaine lettuce leaves (Sub-obj. 1E). Our results indicate that this pathogen experiences osmotic stress in plant wounds. A manuscript is in preparation for submission this year. An RpoS mutant of a lettuce outbreak E. coli O157:H7 strain was compared with the parental strain in Modified Atmosphere Packaged Iceberg lettuce and showed a reduced growth rate, suggesting that RpoS may benefit the fitness of the overall E. coli O157:H7 population in processed lettuce (Sub-obj. 1H). To determine if lettuce lines exist that might be resistant to colonization by E. coli (Sub-obj. 1G), four romaine lettuce lines with varying shelf life and resistance to plant diseases such as downy mildew were assessed for expression of several basal plant defense genes and compared with their colonization by E. coli O157:H7. In assessing Listeria monocytogenes attachment to produce (Sub-obj. 1F), biofilm studies continued on strains relevant to produce-related outbreaks, and a 2011 cantaloupe outbreak strain had the best biofilm formation. In collaboration with researchers at North Carolina State University, we are using genetic methods to identify genes needed for attachment and growth of L. monocytogenes on cantaloupe and lettuce. To study human norovirus (HuNoV) attachment to produce (Sub-obj. 1I), we determined that HuNoV could form specific and non-specific complexes with native plant-associated bacteria. A chimeric bacteria-virus complex was constructed with HuNoV capsid proteins presented on the surface of transformed E. coli, and these complexes bound to romaine lettuce leaves. Simple water wash was not sufficient to remove the complexes, but binding was inhibited with salt or mild a detergent. Two manuscripts describing this work were submitted, and one being prepared. A combined Infrared-Hot air technology was highly efficient for the drying of wet whole almond fruit to diminish bacterial proliferation on the fruit during storage. While this approach reduced bacterial contaminant levels on hulls and shells, it did not meet the required 4-log reduction on kernels. As part of Objective 2, we continued projects on pathogen subtypes that persist and are endemic to local regions. This research will identify environmental factors that allow these pathogens to persist in the environment to potentially contaminate produce pre-harvest. With improvements in our method to quantify pathogens from samples collected in the Salinas region, we can now extract DNA from Gram-positive and Gram-negative bacteria equally and efficiently (Sub-obj. 2A). Using a new Polymerase Chain Reaction (PCR) method we are quantifying levels of pathogens (E. coli, STEC, Salmonella, L. monocytogenes) from Moore swab samples, which combine water and sediment samples, by the highly sensitive and quantitative droplet digital PCR (ddPCR) method. The data from this project will help our collaborators at the Food and Drug Administration to do risk assessments of the Salinas region. To determine the indigenous bacteria that live in the region, and how they might allow persistence of STECs (Sub-obj. 2C), we are doing metagenomic analysis of sediment samples from Salinas with samples taken from two sites that demonstrated periodic persistence and lack of persistence of specific STEC subtypes. Statistical analysis indicated a significant difference in the types of microbial populations between the two sample sites. Our collaborator from the Georgia Institute of Technology detected antibiotic resistance genes at levels higher than in similar samples collected elsewhere, indicating that the sediment from the Salinas region may be a significant reservoir of antibiotic resistance. Subtyping work continues on 1,600 Salmonella strains and 1,200 L. monocytogenes strains isolated from our study of the Salinas area watersheds in 2011-2016 (Sub-obj. 2E). Over 50 different Salmonella serovars have been identified. Some of the most common serovars in the region include Typhimurium, Oranienburg, Montevideo, Infantis, Anatum, and Enteritidis. All of these serovars are reported by the Center for Disease Control as being the most common causes of salmonellosis in the U.S. Some Salmonella serotypes (6,8:d:- and Give) show geographical bias in the Salinas region, and others (Typhimurium) do not. The most common L. monocytogenes serotypes found in the region are those commonly responsible for listeriosis outbreaks, with serotype 4b making up 84 percent of the isolates. The Salinas region differs from other regions surveyed in North America where the most common serotype of environmental L. monocytogenes is 1/2a. In a study of a subset of the L. monocytogenes isolates, 90 percent contained intact internalin A (inlA) genes, indicating they have the potential to be virulent. This differs significantly from isolates from foods and food processing plants (40 to 50 percent contained intact inlA). The inlA work was published in PLoS ONE. To source track STECs in the Salinas region (Sub-obj. 2F) we monitored a type of STEC virus (F+ RNA coliphages) as pathogen surrogates in surface waters from leafy green production environments. We developed a highly sensitive, ddPCR method targeting MS2, which is an indicator for animal fecal contamination that detects 5.5 coliphage genomes. We used the method to quantify directly whole phage particles without culturing or RNA extraction. In our project to determine if wild pigs could be hosts for HuNoV (Sub-obj. 2H), we collected 37 fecal samples from wild pigs in California, and are testing them for porcine norovirus, HuNoV, and pig coronavirus. In work to source track and identify Mycobacterium avium subsp. paratuberculosis, the causative agent of Johne’s disease, we developed a matrix-assisted laser desorption/ionization time-of-flight mass spectrometric method to differentiate strains from each other and from other mycobacteria. Obective 3 comprises research to develp methods for detection and subtyping of enteric pathogens. As part of Sub-obj. 3A, our Fourier Transform Infrared Spectroscopy database of Salmonella strains numbers over 600 Salmonella isolates. Analysis of the top 20 disease causing serovars indicated that at least two different databases will be necessary to differentiate them. To detect inactivation of Tulane Virus (TV) and HuNoV (Sub-obj. 3D) we are using Ultraviolet radiation in our assay development to correlate TV levels between a cell-based assay (TCID50 assays) and a digital PCR method. We continue to study conditions that favor acquisition and maintenance of the Shiga toxin (stx) gene in environmental E. coli (Sub-obj. 3C). We replaced the shiga toxin gene in two STEC outbreak strains with an antibiotic resistance gene, and then measured transfer of that antibiotic resistance via stress-induced phages between strains of bacteria. DNA-damaging stresses induced gene transfer, but oxidative and acid stresses did not. Currently we're optimizing methods to quantify gene transfer by stresses common to produce-production environments. Objective 4 comprises three new projects focused on STEC bacteriophage ecology and applications in food safety. From environmental samples from Salinas Valley, we isolated lytic phages targeting the top six STEC serogroups involved in illness. Most phages were found from samples taken in late summer and early fall, but environmental factors (e.g.rainfall, solar radiation) did not affect phage prevalence. Sites positive for phage were negative for the phage's specific host serogroup.


4. Accomplishments
1. Virulence potential of endemic, environmental Listeria monocytogenes strains. Internalin A, encoded by the inlA gene, is a protein essential for L. monocytogenes virulence in humans, and about 40 to 50 percent of the strains isolated from foods and food processing plants contain truncated inlA genes such that they are essentially non-virulent; however, little is known about the inlA genes in environmental isolates. ARS researchers in Albany, California, isolated a collection of L. monocytogenes strains from public access watersheds surrounding a leafy greens growing region on the Central California Coast, and determined the inlA sequence for 112 of these isolates. They found that 90 percent of the isolates had intact inlA genes, which is significantly different from isolates in processing plants. Moreover, the strains carrying these intact genes persisted in the environment for at least three years, possibly cycling through animals. These results may indicate sources for listeriosis outbreaks, and are being used in risk assessment development by the Food and Drug Administration.

2. Attachment of Human Norovirus (HuNoV) to romaine lettuce analysis. Norovirus outbreaks have been caused by consumption of contaminated lettuce, but it is difficult to directly study virus attachment to lettuce leaves. ARS researchers in Albany, California, developed bacterial-virus complexes to characterize the binding of HuNoV to romaine lettuce. Bacterial strains that displayed the most common types of capsid proteins infecting humans were constructed. These were used to characterize the interaction of the viral proteins with compounds on the lettuce leaves, and the bacteria expressing the HuNoV capsid proteins could bind to the surface of the romaine lettuce leaf, including the stoma and the vein. The bacterial-viral complexes could not be washed off the lettuce leaves with water, but could be removed by adding salts or detergents. This research will be important to develop potential intervention methods to remove HuNoV from lettuce surfaces.

3. Curli expression in Escherichia coli O157:H7 may be affected by environmental factors. ARS researchers in Albany, California, evaluated over 450 E. coli O157 strains from outbreaks, sporadic infections, animal fecal matter, and environmental samples for expression of curli, a fitness trait linked with intestinal colonization. More than half of fecal strains and a significant proportion of environmental isolates (82 percent) lacked curli expression. E. coli O157 strains from outbreaks linked with the consumption of contaminated apple juice, produce, hamburgers, steak, and beef also lacked curli expression. Phylogenetic analysis of environmental E. coli strains revealed that curli expression is likely to depend primarily on the type of environmental exposure and isolation source. These results suggest that understanding those environmental factors that effect curli expression would affect the design and choice of land on which to grow produce that would be less exposed to stress resistant E. coli strains. This data is needed for risk assessments to avoid areas prone to contamination of crops with virulent E. coli.

4. Development of method to detect human norovirus (HuNoV). In collaboration with scientists at the Shanghai Jiao Tong University in China, ARS researchers in Albany, California, compared a newly developed In situ Capture reverse transcription-polymerase chain reaction (RT-PCR) molecular method with an older, established RT-PCR method in detecting HuNoV from oysters collected from a market in Shanghai. The investigators established that the new In situ Capture RT-PCR method was more than 10 and 100 times more sensitive for the type GI and GII HuNoV, respectively, than the older method for detection of the human pathogen from oysters. This research was published and is being used to develop diagnostic tests for detection of HuNoV from other foods.

5. Bacteriophages specific to virulent, Shiga toxin-producing Escherichia coli O45, O145 and O157 strains were isolated from pre-harvest, produce environments. Shiga toxin-producing E. coli strains have been isolated from produce-growing regions of the California Central Coast. Bacteriophages are viruses specific to bacteria that can infect and kill certain host bacteria. ARS researchers in Albany, California, assessed the prevalence of bacteriophages specific to virulent, Shiga toxin-producing E. coli strains O45, O145 and O157, which have been isolated from the same environments. These bacteriophages were isolated from several locations along the Central California Coast, and there was a trend that sites that contained bacteriophages specific to O45, O145, or O157 E. coli subtypes also were missing those specific host strains, as if the presence of the bacteriophage affected the presence of the host. This finding may facilitate the use of these bacteriophages in the development of intervention or detection tools to enhance pre-harvest produce safety, and prevent the spread of Shiga toxin-producing E. coli strains in the region.

6. Interaction between bacteria and Human Norovirus (HuNoV), and the role of bacteria in viral heat resistance. One avenue of contamination of produce by HuNoV is binding and protection of the virus with bacteria naturally present on produce surfaces. ARS researchers in Albany, California, used Tulane Virus as a substitute for HuNoV to investigate the interaction between bacteria expressing Human Blood Group Antigens (HBGA), the natural human receptor for HuNoV, and bacteria not expressing HBGA. Tulane Virus could bind the bacteria regardless of whether the bacteria were expressing HBGA, but binding was stronger to bacteria that did express HBGA, indicating that the virus can bind to bacteria. However, binding of virus to the bacteria did not lead to protection from heat-inactivation of the virus. This information may aid in the development of methods for the elimination of HuNoV from foods.


Review Publications
Gorski, L.A., Parker, C., Liang, A.S., Walker, S., Romanola, K. 2016. The majority of genotypes of the virulence gene inlA are intact among natural watershed isolates of Listeria monocytogenes from the Central California Coast. PLoS One. 11(12):e0167566. doi:org/10.1371/journal.pone.0167566.

Wu, V.C., Drummond, F., Tadepalli, S., Camire, M., Davis-Dentici, K., Bushway, A., Yarborough, D.E. 2017. Salmonella spp. dynamics in wild blueberry, Vaccinium angustifolium Aiton. World Journal of Microbiology. 4(1):64-71.

Carter, M.Q. 2016. Decoding the ecological function of accessory genome. Trends in Microbiology. 25(1):6-8. doi:org/10.1016/j.tim.2016.11.012.

Venkitasamy, C., Brandl, M., Wang, B., McHugh, T.H., Pan, Z. 2017. Drying and decontamination of pistachios with sequential infrared drying, tempering and hot air drying. International Journal of Food Microbiology. 246:85-91. doi:10.1016/j.ijfoodmicro.2017.02.005.