Progress 09/01/23 to 08/31/24

Outputs
Target Audience:Our primary target audience continues to be agricultural centers and extensions. In collaboration with Prof. Channah Rock's group at the University of Arizona, side-by-side study of agricultural water samples in Yuma AZ was conducted between the ALERT Lab and the EPA approved method (Colilert Quantitray/2000). We also investigated the presence and effects of bacterial aggregates in Ag waters by running measurements with 5um filters. In addition, we have presented selected results from our ALERT Lab Ag water quality monitoring studies at conferences, such as WEFTEC 2023 and NEMC 2024, with audience that includes the USEPA, utilities, environmental and other groups. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Following the training for the ALERT LAB instrument, we continue to provide support to Prof. Channah Rock's research group at the University of Arizona as well as the extension specialists at the Maricopa and Yuma Ag Center, on the data analysis and interpretation as well as the design of additional experiments. We also have regular discussions to review and analyze the measurement data together. One of our scientists, David Wanless, also spent one week in the lab of Prof. Channah Rock at the Maricopa Ag center to work with the team there to conduct experiments related to agricultural applications. The University of Arizona team has been trained on the use of our ALERT Ag++ instrument prototypes and has used them autonomously in the field. The handheld autonomous E.coli field quantification instrument that is the direct result of the project will also be tested both by extension specialists and by partnering organizations and produce associations. How have the results been disseminated to communities of interest?We continue to have periodic discussions on the ALERT study results with our collaborators at the University of Arizona who are actively involved with the agricultural community and growers. Having one of our scientists spend a week working with Prof. Channah Rock's team at the Maricopa Ag Center allows for sharing and exchange of knowledge with communities of interest. The Ag water measurement results have also been presented at water quality conferences, such as the National Environmental Monitoring Conference (NEMC) 2024. What do you plan to do during the next reporting period to accomplish the goals?In the next reporting period, we will start testing the new qPCR methods with pure culture pathogenic E. coli. Once the sensitivity and the effectiveness of the method have been confirmed with pure culture targets, the next step would be to incorporate low number of target (pathogenic) cells in a high background (e.g. 10^6) of non-pathogenic E. coli and media. The final step will utilize actual agriculture water samples spiked with low levels of pathogenic E. coli. These samples will be grown in the ALERT instrument, diluted, and cell-popped. Along with the STX-duplex qPCR assays, confirmation using other molecular detection methods (dd-PCR) will be run on these samples. We plan to travel to the University of Arizona to perform these experiments using locally sampled agricultural water together with Prof. Channah Rock's group as well as the extension specialists at the Maricopa and Yuma Agricultural Center.

Impacts
What was accomplished under these goals? Objective 1: Operational field-deployable protocol In the last reporting period, our experiments have shown that target DNA (E. coli) could be detected in a background of non-target cells. During this reporting period, we tested using cell-pop on cellular targets in a background of non-target cells. The target would be E. coli cells which would more closely mimic the final product of cell-pop LAMP for the detection of STX-1&2 genes. For this experiment, two types of cells were grown-up in ALERT instruments. One sample had only E. coli bacteria and the other had only Enterococci bacteria. After testing the two cultures for cross-contamination, sample dilutions were made with decreasing amounts of target cell (E. coli) in a background of Enterococci at a specific concentration (1x10^6 cells/µl). The results showed that with the LAMP assay for universal E. coli, we had successful detection from 10,000 cells/µl down to 100 cells/µl. From the experiments conducted in our collaborator's lab at the University of Arizona, we confirmed that the STX-1&2 duplex LAMP assay was able to detect O157:H7, the "big six" STEC cultures, and the Yuma 2018 outbreak strains using a simple cell-pop lysis. These experiments show that the LAMP assay works not only with gDNA but also with cells and crude cell-pop lysates. Objective 3: Functional instrumentation prototypes For concept validation, we demonstrated in a set of experiments that liquid culture from an ALERT instrument could be moved from one vial into a 1:10 dilution chamber, followed by incubation and LAMP reaction. In addition, we showed that mixing was necessary for timely and reproducible results, based on fluorescence vs time-to-threshold curves of three sets of samples that have undergone different degrees of mixing. Another experiment showed that the STX-1&2 duplex assay works effectively in this 1:10 cell pop crude lysate. LAMP has always been viewed as a method that works well in the presence of inhibitors, and that its single amplification temperature was advantageous over the cycling parameters of qPCR. After the purchase of a qPCR machine in our lab, LAMP and qPCR experiments were conducted side-by-side to compare their performance. Dilutions of target DNA were tested using both qPCR and LAMP chemistries. HF183 assay was used for these comparisons, with the results for standards and LA River samples measured using qPCR and LAMP. The results showed that qPCR was more sensitive (down to 2 copies/µl) than LAMP (down to 200 copies/µl). Following the experiments comparing LAMP and qPCR with the latter shown to be more sensitive, primers and probes were subsequently designed for a duplex qPCR assay. STX-1 and STX-2 results from diluted pathogenic E.coli gDNA in a background of a water sample grown in the ALERT Lab show that these qPCR assays were effective in detecting various concentrations of the pathogenic targets in our experiments. Objective 4: Design and implementation of regulatory approval strategy Agricultural waters in the Maricopa area are sampled and measured by the University of Arizona on a regular basis to gauge the number of bacteria that are present, using EPA-approved methodology (IDEXX Colilert Quantitray-2000). Recent outbreaks of pathogenic bacteria have intensified this demand. The Total Coliforms level in the waters are also monitored pre- and post- chlorine treatment, for measuring treatment abatement factors for all bacteria. Extension staff from Prof. Channah Rock's team who are based at the Yuma Ag center performed a series of side-by-side tests using the fluidion Alert Lab along with standard IDEXX methods to detect E. coli and Total Coliforms bacteria. 132 samples were collected, most having an unfiltered sample and 3µm-filtered sample run in parallel using both measurement methods. The 3µm filtered samples were representatives of free-floating (or planktonic) bacteria, which can pass through the filter, while larger suspended sediment particles and fecal aggregates, if present, are retained. The unfiltered samples contained all the suspended matter, which were comprehensively measured by ALERT technology (comprehensive count), whereas IDEXX only identified the total number of aggregates, without the ability to provide the total count of bacteria attached to these aggregates. The data show that there was good correlation between the ALERT Lab and IDEXX Colilert results. There was some over-estimation on the Alert Lab for E. coli and Total Coliforms and we attribute the difference being a result of the ability of the ALERT Lab to perform a comprehensive count of all bacteria, including aggregates, in the water sample. When the measured values for both E. coli and Total Coliforms in these Ag water samples were entered into the EPA's Alternative Methods Calculator Tool (AltCalc Tool available here: https://www.epa.gov/system/files/documents/2021-11/site-specific-alt-calculator-tool-factsheet.pdf), index of agreement (IA) for E.coli and Total Coliforms is 0.80 and 0.71, respectively, which demonstrate that the side-by-side comparison was successful, and the ALERT can be substituted for the standard IDEXX method. The same results were also analyzed by plotting separately the planktonic and comprehensive counts of E. coli and Total Coliforms, against the results provided by the standard method (IDEXX Colilert Quantitray-2000 performed on raw unfiltered samples). All samples below the detection limit were considered to be at the detection limit for all methods. In both cases we observe a major contamination event on August 14, 2023 (sample # 38-40), which was recorded and confirmed by all available methods. This demonstrates the capability of ALERT-Ag++ technology to identify high risk situations when present. As a general tendency, we observe that most of the planktonic counts were at or below the detection limit of the ALERT technology (4 E.coli/100mL, due to the 25mL sample analysis volume used). Comprehensive counts, however, show consistently higher numbers than the planktonic counts and also significantly higher than IDEXX counts, suggesting that, when E.coli are present they do have a tendency to be aggregated on larger particles which are not visible using standard lab methodology. This tendency is confirmed when ALERT (planktonic and comprehensive counts) and IDEXX data are plotted against each other or as Bland-Altman plots. While the planktonic counts show a small mean difference compared to IDEXX (0.20 log10 units or 158%), the comprehensive count shows a much higher mean difference (0.76 log10 units or 575%). This is consistent with data we previously published, from sampling performed in the CA Central Valley. We note however that the numerous results that are at or below the detection limit, especially for planktonic counts, do not allow for highly accurate statistics. Using the STV threshold of 410 E.coli/100mL that is imposed by the FSMA regulations, we can calculate the confusion matrices for ALERT Comprehensive and Planktonic E.coli counts, compared to IDEXX. These show excellent sensitivity and negative predictive value, along with high specificity and accuracy. The lower precision values are a consequence of the very low number of true positives (3 samples total). The Total Coliform data show similar tendencies, with the exception that total coliforms are present in much higher numbers, providing better statistics. Also, very few measurements were at the detection limit. This results in significantly better side-by-side correspondence, and balanced Bland-Altman plots. We also observe that a large number of coliforms are present on particles that are larger than 3µm, which results in a lower mean measured for ALERT planktonic counts than for IDEXX (-0.32 log units or 47%). The comprehensive counts are, again, significantly higher than IDEXX counts (0.55 log10 units, or 354%).

Publications

Progress 09/01/22 to 08/31/23

Outputs
Target Audience:Agricultural centers and Extensions. We have been in collaboration with Prof. Channah Rock's group at the University of Arizona to conduct side-by-side study of agricultural samples in Yuma AZ between the ALERT LAB and EPA approved method (Colilert Quantitray/2000) and to investigate the presence of bacterial aggregates in Ag waters from the Yuma growing region. To this goal we provided an ALERT LAB instrument and online training to the Yuma Ag center staff, as well as on-going technical support. We are also in contact with Dr. Trevor Suslow, who was one of collaborations during the USDA Phase I project, and is interested in conducting further studies using the ALERT LAB with local growers in the Central CA region. We recently initiated contact with the World Health Organization and presented some of our recent finding related to bacterial aggregates and how our method allows improving on current quantitative risk assessment methods for surface waters. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?We have provided training for the ALERT LAB instrument to Prof. Channah Rock's research group at the University of Arizona as well as the extension specialists at the Maricopa and Yuma Ag Center. We also have regular discussions to review and analyze the measurement data together. How have the results been disseminated to communities of interest?We have periodic discussions on the ALERT study results with our collaborators at the University of Arizona who are actively involved with the agricultural community and growers. The Ag water measurement results have also been included in our presentations at water quality conferences, such as the National Environmental Monitoring Conference (NEMC) 2022 and Center for Watershed Protection (CWP) Conference 2023. What do you plan to do during the next reporting period to accomplish the goals?In the next reporting period, we will continue our collaboration with the University of Arizona in the side-by-side study of ALERT vs Colilert Quantitray/2000 using agricultural samples, and in particular, gain a better understanding of the effect of particles in the E.coli quantification between the two methods. We plan to publish the results and present at relevant conferences for both environmental monitoring and agriculture. We will reach out to regulating agencies but also to produce safety officers within relevant Ag organizations with our findings. We will also reach out to the wider agricultural community through the International Fresh Produce Association (IFPA) and learn more about current food safety practices and challenges within the produce industry. We will also apply the feedback from the agricultural community in our development of functional instrumentation prototypes and continue our validation through extensive testing. Finally, with results from side-by-side studies, we will design and implement a test plan towards EPA approval.

Impacts
What was accomplished under these goals? Objective 1 Designed LAMP assays for targeting the two Shiga toxin genes of STX-1, STX-2, and a virulence gene for shiga toxin E. coli (STEC) eae gene, and a universal E. coli LAMP assay for the gene uidA. Specific primers for all the targets, including generic E.coli and STEC O157:H7 were designed and optimized. All assays amplified their positive control gBlocks, with the limit of detectionfound to be: 100 copies/ul for STX-1 and universal E. coli gene uidA, and 10 copies/ul for STX-2 andeae virulence gene. Our LAMP assays were also tested using commercially available gDNA for the E.coli strain, O157:H7 EDL 933, and shown to be positive with higher sensitivity compared togBlock controls. The limit of detection for the assays were found to be close to 10 copies/ul for STX-1, STX-2,eae gene, and the universal E. coli gene uidA. Giventhat the three genes, STX-1, STX-2, and eae, are markers for shiga toxin producing E. coli (STEC), their LAMP assays were combined into one tri-plex assay and wastested using the gDNA from O157:H7 EDL 933. The results showed higher sensitivity for the tri-plex assay, down to 1 copy/ul. Through our collaborator at the University of Arizona (UoA), we were provided with the "Big Six" non-O157 toxin producing E. coli strains: O103:H11, O121:H19, O145, O45:H2, O26:H11, and O111.These strains contained at least one STX toxin gene along with the eae gene. Our intial experiments showed that four of the strains which had the corresponding STX gene in its genome (O103:H11, O121:H19, O26:H11, and O111) were sensitive to 100 copies/ul while two of the strains (O145 and O45:H2) were sensitive to 1,000 copies/ul. All the strains tested contain the eae gene, however, only two of the two strains were sensitive to 100 copies/ul. Since the eae gene is also present in non-toxigenic strains of E. coli, the eae LAMP assay was not further optimized at this time. Our experiments showed that the STX-duplex assay worked well in detecting all six strains from UoA, with only two strains (O145: and O45:H2) having slightly decreased sensitivity. Next, we explored different cell lysis methods: 1) heating the culture to 95°C for 5-10 min (cell-pop) and filtering the crude lysate through a 0.45um filter, 2) pelleting the cells, replacing the media with water, and cell-popping the resulting mixture, and 3) starting with method 2, followed by method 1. Based on our LAMP experiments, the best results were obtained uinsg cell lysis method 2. The cell lysis method 2 was applied to generic E.coli obtained from enriched ALERT instrument culture. The sample was serially diluted, cell-popped, and then added to the E. coli uidA LAMP assay for analysis. Our results showed that the culture could be detected down to 1:1000 dilution, and that the removal of the media in the 1:1 dilution was necessary to remove its inhibitory effects. We anticipate that the pathogenic E. coli strains will perform similarly to the non-pathogenic E. coli. To mimic a live sample, 100 copies/ul of gDNA from each of the target strains of pathogenic E. coli were added to a background (from the ALERT instrument culture) of non-target bacteria that was cell-popped. Then the crude lysate with gDNA spike was added to the STX-duplex LAMP assay. Of the "Big 6" strains of E.coli, only O45:H2 was not detected. The rest of the strains were detected with a concentrated cell-pop background. We also verified that the non-pathogenic E. coli gDNA was not detected. Objective 2 A new disposable cartridge concept has been designed, and is currently being discussed with industrial plastic molding partners to optimize for large-volume manufacturability. The new disposable cartridge will allow integrating an in-line filter and checkvalve at the end of the current ALERT V2 cartridges, and will also change the position of some of the fluidic elements within the cartridge, such as the hydrophobic filter and reagent diffuser, to optimize operation in the new proposed configuration. Upon sampling, the vacuum-drawn sample will automatically be filtered inline, thus providing the measurement of the placktonic bacteria. Filtered and non-filtered cartridges can either be alternated within a single instrument, or installed in two separate instruments deployed side by side. To accommodate the additional hardware (filter and checkvalve) which extend further out, an extension of instrument's feet will also be installed. The designed interface allows for samples to be clearly labeled (filtered / unfiltered) for later identification and interpretation of the results. The new Ag interface (in development) will have additional features allowing the users to easily navigate the data and collect both particle-bound and planktonic bacterial concentration data. In collaboration with Prof. Channah Rock's group at UoA, side-by-side testing of agriculturalwater samples obtained in Ag-relevant field locations in Yuma by extension specialists has been ongoing since May 2023. The samples were tested using both the ALERT LAB instrument (provided by Fluidion) along with Colilert Quantitray/2000. Both filtered and unfiltered samples were tested (filter size: 3µm) to understand the partition of E.coli bacteria between the planktonic and particulate (or aggregated) forms and their signature in both ALERT LAB and Colilert Quantitray/2000. This study will continue until Dec 2023. Objective 3 An initial prototype was designed allowing touch-screen operation. In parallel, we also developed the first prototypes of a fully-autonomous instrument for analyzing single samples, which can operate without requiring network connection, by performing all measurements and analysis internally. The single-sample analyzer can find important applications for smaller farms in remote locations, which rely on a small number of water sources and may not have the resources to afford the ALERT LAB instrument. We are currently working on the industrialization of the single-sample prototype, and on optimizing the electronics and user interface for the touch screen version. Approved bacterial enumeration methods in use today (MPN or MF) are not designed to measure the particle-bound fraction, and as a result, may drastically underestimate global bacterial charge. This is especially important in agriculture, where attached bacteria could be shielded from UV light and disinfectants, and will release their bacterial load later, as the particle degrade. So they tend to have longer persistence on produce, and is a major risk for leafy greens. As we have learned during our collaboration with Dr. Trevor Suslow's group at UC Davis, there are more particles with attached bacterial charge in canal sources, and as a result, filtering the sample would give a lower bacteria count compared to unfiltered sample, suggesting a high particle content, and potential higher attached bacterial charge, in those samples. With the ALERT technology, we can measure both free and attached E. Coli, which can have important impact in agricultural water monitoring. With the ALERT LAB instrument being used by the UoAteam for measurements of agricultural water samples in the field, we have received valuable feedback from the perspectives of agricultural applications. Factors to take into consideration for our development of E.coli analyzers for Ag use include: field conditions such as high temperatures, limited cell service in many field locations, materials for measurement sample vials for field use, ideal measurement sample capability per day for growers, data management and reporting preferences for Ag industry. As we continue our collaboration with the UoA and the grower community on the use of the ALERT instrument, it would be greatly beneficial to receive feedback relevant to agricultural use.

Publications

• Type: Other Status: Published Year Published: 2022 Citation: WHO/Unicef Joint Monitoring Programme, Laboratory evaluation of portable water quality testing kits: Fluidion ALERT LAB, Independent laboratory evaluation, Phase 1 and Phase 2. Oct.2022, https://washdata.org/reports/laboratory-assessment-fluidion-alert-lab