Final Report: Assessment of Microbial Pathogens in Drinking Water using Molecular Methods Coupled with Solid Phase Cytometry

EPA Grant Number: R833830
Title: Assessment of Microbial Pathogens in Drinking Water using Molecular Methods Coupled with Solid Phase Cytometry
Investigators: Pyle, Barry H , Camper, Anne
Institution: Montana State University - Bozeman , Little Big Horn College
EPA Project Officer: Klieforth, Barbara I
Project Period: March 1, 2008 through February 28, 2011 (Extended to February 28, 2013)
Project Amount: $599,996
RFA: Development and Evaluation of Innovative Approaches for the Quantitative Assessment of Pathogens and Cyanobacteria and Their Toxins in Drinking Water (2007) RFA Text |  Recipients Lists
Research Category: Drinking Water , Water


The determination of microbial safety of drinking water is still done by assessing the presence of fecal contamination through culturing indicator bacteria, such as E. coli, rather than by direct detection of pathogens from the water. Ideally, the absence of the indicators should always signal the absence of pathogens; however, this is not always the case. Because pathogens and indicator bacteria react differently to environmental stresses, pathogens can be present in drinking water even when no indicator bacteria have been detected (1, 2). Additionally, because the methods for detection of indicator organisms rely on culture techniques, it can take days before indicator bacteria can be identified. The goal of this study was to find a rapid method to detect multiple pathogens directly from water. Although this is a global issue, the project was structured toward identifying microbial contamination and its significance within Montana, particularly on the Crow Indian Reservation east of Billings. The objectives of this study were to: (1) develop a sensitive method for rapid/real-time detection and enumeration of target microbial pathogens, including assessment of viability and infectivity using fluorescent in situ hybridization (FISH) coupled with Solid Phase Laser Cytometry, (SPLC), which facilitates very rapid, sensitive and reliable enumeration and immediate confirmation by epifluorescent microscopy (EFM); (2) evaluate this method for identification and detection of Escherichia coli O157:H7, Helicobacter pylori, Legionella pneumophila, Mycobacterium avium, Aeromonas hydrophila, Giardia lamblia, and Cryptosporidium parvum; (3) examine the possibility of detection of multiple pathogens at the same time; and (4) determine if the method could be used in environmental conditions in collaboration with the Crow community.

Summary/Accomplishments (Outputs/Outcomes):

Objective 1 Detection of FISH labeled organisms with the ScanRDI
The ScanRDI is a solid phase laser cytometer that detects appropriately labeled cells on a 25 mm diameter black polycarbonate membrane filter. The instrument scans an entire filter in 3-4 minutes, detecting and mapping individual fluorescent particles. The software system discriminates between cells and debris, and then locates particles to be viewed with an epifluorescent microscope, allowing the operator to directly validate the results. The ScanRDI can detect a single fluorescent particle on the membrane, giving it a very low detection limit, essentially restricted by the amount of filterable sample.
Fluorescent in situ hybridization (FISH) is a technique that uses oligonucleotide probes with the sequences specific for a target organism that hybridize to portions of the ribosomal RNA (3). These probes can be labeled with fluorophores with the emission spectrum that can be detected by the ScanRDI. Methods from previous studies that used the ScanRDI to detect cells labeled by FISH (4-6), were adapted during the initial optimization of the FISH technique, to produce microorganisms that had sufficient fluorescent intensity to be detected with the SPLC. Prior to processing, the cells were filtered through the specific membranes necessitated by the ScanRDI (AES Chemunex CB04) and all steps were done by placing the membrane on the reagents in a stepwise fashion.
Because the ultimate objective of this project was to screen for multiple pathogens on a single membrane, optimization of methods focused on protocols which could be used for a range of organisms. E. coli was chosen to optimize the FISH technique on the membrane filter for use with the ScanRDI due to the availability of information on in situ hybridization of E. coli (5, 7-12) and the relationship of this organism to drinking water safety (13).
Three elements were key to optimizing the FISH procedure to allow successful detection of FISH labeled E. coli on membrane filters by the ScanRDIA prefixation incubation of 2 to 4 hours on nalidixic acid was essential to providing sufficient numbers of ribosomes for the signal to have sufficient fluorescent intensity to be detectable with the SPLC. Two hours was sufficient for laboratory controls, but 4 hours was chosen as the standard because cells from environmental samples would likely have fewer ribosomes than lab strains and would need the longer incubation to provide an adequate signal.
The second important step was the use of tyramide signal amplification (TSA). The TSA technique method uses oligonucleotide probes that can be coupled either directly or indirectly to horse radish peroxidase (HRP). HRP converts fluorescein labeled tyramide into an activated intermediate that binds to adjacent proteins. Multiple deposition of fluorescein occurs in a short time producing a large signal amplification, which enabled detection of the individual cells by the ScanRDI. To reduce extraneous signal, the oligonucleotides used in this study were directly coupled with HRP rather than using a biotin-streptavidin-HRP combination or a probe-FITC - α FITC antibody-HRP combination, although both combinations were explored because of the lower cost.
Thirdly, a lysozyme permeabilization allowed entry of the probe with the large horseradish peroxidase label into the cell. An hydrochloric acid step was introduced to stop lysozyme activity after 10 minutes and prevent cells from becoming leaky. This step also served the purpose of  eliminating any endogenous peroxidase.
Direct epifluorescent microscopy was used during this optimization to determine which methods increased the fluorescent signal. In addition to enumeration, images were taken with the Zeiss Axioskop and analyzed with the Metamorph Software (Molecular Devices) for fluorescent intensity data. Because previous work has demonstrated that SYBR Green fluoresces at sufficient intensity to be detected by the ScanRDI (12), fluorescent intensity of cells labeled by an experimental method were compared to fluorescent intensity of SYBR Green.
Optimization of the protocol was begun using the probe ECO541  since ECO541 was reported to provide a bright signal when compared to other potential 16S probes for E. coli. It has been postulated that the brightness of the signal was likely due to accessibility of the target sequence in the ribosome that allowed the FISH probe to bind easily (10). The methods developed by Baudart, et al (4), describing the use of a 16S rRNA probe for Enterobacteriaceae detected by theScanRDI, were adapted for use with the ECO probe and E. coli. Experiments showed that this probe coupled with horseradish peroxidase followed by the tyramide signal amplification protocol could be detected by the ScanRDI, but it was found that this probe was not specific for E. coli, in fact it also probedAeromonas hydrophila, which was used a as control organism. Another probe designated Colinsitu (COL) was shown by the Regnault group to specifically label E. coliEscherichia fergusonii and Shigella spp. (9), organisms that would be significant water contaminants, but the target was in a fairly inaccessible region of the ribosome (10) and thus was unlikely to provide an adequately strong fluorescent signal for the ScanRDI detection. Fuchs, et al., (14) found that by adding unlabeled ‘helper’ probes designed to bind to positions flanking the labeled probe target region in the ribosome, it was possible to increase the fluorescent signal by binding to an otherwise inaccessible region. We adapted the methods described by Baudart, et al., (5), using a 2- to 4-hour prefixation incubation with the antibiotic nalidixic acid in R2A broth, a 90-minute hybridization with the combination of Colinsitu probe labeled with horseradish peroxidase and unlabeled helper probes in the hybridization buffer followed by amplification of the signal with tyramide. The basic hybridization buffer contained 0.9 M NaCl, 20 mM Tris HCl pH7.2, 0.01% sodium dodecyl sulfate with 0.5 pmol/µl probe-HRP added. When helper probes were used, each unlabeled helper probe concentration was 2.5 pmol/µl. The addition of formamide (final concentration 20%) gave a brighter signal.
The parameters of the ScanRDI software can be adjusted so that the characteristics the scanner uses to differentiate between target and debris more closely matches the characteristics of the particular fluorescent label of the cells. Table 1 gives the settings for the application developed in this lab for FISH labeled E. coli. After adjusting the application software of the ScanRDI to detect the FISH probed cells, we found good correlation between plate counts and enumeration of E. coli with the ScanRDI, consistently identifying between 70 and 112% of the E. coli compared to cells enumerated by plate counts. A summary of the labeling method is shown in Table 2.

Table 1. ScanRDI Frequency Thresholding and Discriminant Settings used for Detection of FISH Labeled Cells and Oocysts


Frequency Table Thresholding

Trig Delay


Peak Value
2D Gaussian
Half Width
Table 2. Summary of protocol used for FISH hybridization for detection of E. coli
FISH Hybridization
R2A broth with nalidixic acid 2 – 4 hours
4% paraformaldehyde 1 hour
50 % 80% 94% ETOH baths 4 min each
60 µl TE rinse 5 min
Lysozyme 10 min
60 µl 0.02 N HCL 10 min
60 µl TE rinse 5 min
Hybridize 90 min at 48 °C
Wash buffer 100 µl 30 min 48°C
Rinse TNT buffer 5 min
FITC tyramide signal amplification 75 µl 10 min
Rinse TNT buffer 5 min
Store dry on cellulose pad at 4˚C until analyzed
Objective 2 Evaluation of this method for detecting other target organisms
Once the method was developed for use with E. coli, other organisms were probed, with some adjustments made to the hybridization protocol and the ScanRDI parameters when necessary.
Optimization of FISH with Cryptosporidium parvum was chosen from the list of candidate organisms because of the interest of the Crow Environmental Health Steering Committee, (CEHSC), a group that meets to discuss public health issues and solutions on the Crow Reservation in south central Montana. They had particular interest in whether Cryptosporidium parvum might be present in the source water for the water treatment plant on the reservation. During optimization, it was determined that the method used for E. coli shown above using the probe, CRY-HRP, without the use of helper probes and without formamide in the hybridization buffer produced brightly labeled Cryptosporidium parvum oocysts detectable with the ScanRDI with the application developed for E. coli labeled with FISH (Table 1). For quantification of oocysts, the FISH/SPLC  method was compared to the EPA fluorescent antibody/DAPI/ differential interference contrast method.
The results on control oocysts are shown in Table 3. The advantage of the FISH method is that the both the identification and viability of oocysts can be ascertained from the results of the FISH probes (15); DIC or DAPI staining is not required. There is also less cross reactivity so the preparations have fewer false positives. When the ScanRDI is used for enumeration, preparations can be read in a much shorter time than is possible with the fluorescent antibody method.
Table 3. Number of Control Oocysts Detected by the EPA FAb method or ScanRDI



Legionella pneumophila  and Aeromonas hydrophila were also examined using the FISH method.  Detection of Legionella pneumophila with the ScanRDIrequired adjustment of the basic protocol including; prehybridization incubation on chloramphenicol rather than nalidixic acid, incubation on 0.2% Tween 20 following the lysozyme permeabilization step to increase permeability, the addition of 1 mg/µL of dextran sulfate to the hybridization buffer and the TSA solution. Addition of helper probes to the hybridization buffer gave a slight increase in fluorescence. Although these changes improved the overall fluorescent intensity of the Legionella cells, a new application with a lower fluorescent intensity threshold was required for detection by the ScanRDI.  With these changes, 89.1% of the L. pneumophilaenumerated by plate counts were detected with the ScanRDI method. This is an improvement in detection primarily because culture procedures take at least 3 days and when using FISH and the ScanRDI results can be obtained within 10 hours.
Aeromonas hydrophila was FISH probed using the procedures for E. coli and scanned with the SPLC.  The number of cells enumerated by this method were 39% of those enumerated by plate counts. Adaptation of the method were not attempted. A list of all probes used in this study are found in Table 5.
The remaining target organisms Helicobacter pylori, Mycobacterium avium, and Giardia lamblia were not able to be examined during this project. The pathogenic E. coli O157:H7 was not specifically probed as literature reviews suggested that FISH could not differentiate between nonpathogenic and pathogenic strains.
Objective 3 Detection of multiple pathogens at the same time
Because of the diverse requirements for FISH probing of the different organisms (e.g., different antibiotics, different composition of the hybridization buffer, different ScanRDI applications), the goal of multiple organisms on one membrane filter was not met with one exception.
Cryptosporidium parvum and E. coli could be probed using the same protocol (Table 2) and detected by the SPLC using the same application (Table 1). Cryptosporidium fluorescence and detection improved after incubation at 37˚C for 4 hours on R2A with nalidixic acid, a condition which was required for optimal detection of E. coli using FISH with the ScanRDI. Probes for both organisms including helper probes for the Colinsitu probe were added to the hybridization buffer concurrently. Because of the differences in morphology, it was easy to differentiate between the 2 organisms (see Figure 2 below).
Figure 2. A) E. coli cells and B) Cryptosporidium oocysts after 4 hour incubation on R2A broth with nalidixic acid

Objective 4  Use of method in environmental conditions in collaboration with the Crow community

As noted above, the Crow Environmental Health Steering Committee specifically requested that we look for Cryptosporidium parvum in the Little Big Horn River. This river is the source water for the water treatment facility at Crow Agency, Montana and may be contaminated by both cattle operations and rural houses upstream from the treatment plant. Nine water samples were taken from the source water and the treated drinking water over the course of 1 year, from June 2011 through June 2012. The samples were collected using filtration, concentrated by centrifugation and purified with immunomagnetic separation as described in EPA Method 1623. The resulting sample was split evenly and labeled with either the FAb/DAPI/DIC and examined microscopically or labeled with the newly developed FISH method and examined with the ScanRDI. At each sampling point, both river water and drinking water were also tested for coliforms and E. coli using Hach mColiBlue.

Cryptosporidium parvum oocysts were found in all river water samples tested by both EPA and ScanRDI methods, with highest numbers found in June and July, 2012 (Table 4). Figure 3 shows that the number of oocysts labeled with FISH detected by the SPLC was higher than the number of oocysts labeled with FAb/DAPI enumerated by the manual microscopy for every sample taken from the river.

There was no correlation between the number of E. coli or coliforms and the number of oocysts found in a sample.

Table 4. Results of Sampling the Source Water from the Little Big Horn River The column labeled EPA shows oocysts/liter as detected by EPA method FAb/DAPI and DIC and the column labeled ScanRDI shows oocysts/liter as detected by the ScanRDI after FISH probing with CRY-HRP. *ScanRDI result lost due to lab error.
Coliform (CFU/100 ml)
E. coli (CFU/100ml)
EPA #/liter
Oocysts ScanRDI#/liter
Total oocysts in Sample (Calculated)
Total oocysts/liter (Calculated)
Figure 3. Comparison of Oocysts Detected by FISH/ScanRDI and Oocysts Detected by FAb/DAPI/DIC. Line represents the line of equality
In May, 2012, there was a major flood event which carried river water directly into the water treatment plant, preventing sampling during that month. Disturbingly, following this flood, although no coliforms or E. coli were found in the treated water, Cryptosporidium parvum oocysts were found in the treated water using the ScanRDI method in June and by both methods in July. By the August sampling and throughout the rest of the study, no oocysts were found in the treated water by either method.


We have demonstrated that the FISH method as described above can be used with the ScanRDI for detection of E. coli from water in numbers that are not significantly different than enumeration with plate counts. With no alterations, this method can also be used for detection of Cryptosporidium parvum oocysts. Both identification and viability can be ascertained using FISH. It can also be used to identify and enumerate 89% of the Legionella pneumophila found using culture methods within 10 hours compared to 3 or more days. Detection of  Aeromonas hydrophila using FISH and the ScanRDI were only 39 % of plate counts. Other target organisms were not tested within the time frame of this project. 
E. coli and Cryptosporidium were enumerated concurrently using the FISH/ScanRDI method. The method allowed determination of both identification and viability for each organism. Because of differences in requirements for FISH, concurrent detection of other pathogens remained problematic.
This method was successfully used for environmental testing of water for Cryptosporidium after filtration, concentration and purification. During environmental testing, sampling for Cryptosporidium parvum using the FISH/ScanRDI method detected more oocysts in every sample than the currently used FAb/DAPI/DIC method. Source water from the Little Bighorn River contained Cryptosporidium in each of the nine water samples tested and the treated water from the Crow Agency water treatment plant contained viable oocysts following flooding. 
Future Study
It was found that purification and concentration of the organisms was necessary prior to detection of pathogens with the SPLC. Direct filtration of 100 ml of clear pathogen-free water sometimes produced over 10,000 fluorescent events that were detectable with the ScanRDI FISH application making enumeration too difficult to be practicable. A method which allows concentration all the target pathogens from large volumes of water needs to be developed. One possibility is using the less specific immunomagnetic separation technique using antibodies for concentration, along with FISH probing for identification. The addition of large volumes of paramagnetic beads to liters of water would be cost prohibitive. Instrumentation that allows flow of water past immobile antibody-labeled paramagnetic beads similar to the Pathatrix® instrument (Life Technologies), used for detection of pathogens from food, may provide an answer to this problem. This would eliminate the difficulty of concentration of particles along with pathogens which is a problem with filtration methods.
One important outcome of this work is that the Crow community is seeking funding to address the risk documented by the presence of Cryptosporidium in their drinking water due to breakthrough during the flood event in May 2012. If regular sampling during this study had not been done, the breakthrough event would have gone undetected. Improvement in treatment is necessary due to the high numbers of Cryptosporidium in the source water as determined during this project.
Table 5.  List of FISH Probes used in this study
Legionella pneumophila
Helper for L. pneumophila
LEG H343
This lab
Helper for L. pneumophila
LEG H307
This lab
Aeromonas hydrophila
Helper for A. hydrophila
This lab
Helper for A. hydrophila
This lab
Cryptosporidium parvum
E. coli non specific
E. coli
Helper for E. coli
Helper for E. coli
Negative control
Positive control (Bacteria)
EUB 338



1.         Wu J, Long SC, Das D, Dorner SM. 2011. Are microbial indicators and pathogens correlated? A statistical analysis of 40 years of research. Journal of Water & Health 9:265-278.
2.         Nasser AM, Zaruk N, Tenenbaum L, Netzan Y. 2003. Comparative survival of Cryptosporidium, coxsackievirus A9 and Escherichia coli in stream, brackish-and sea waters. Water Science and Technology 47:91-96.
3.         Amann RI, Krumholz L, Stahl D. 1990. Fluorescent-Oligonucleotide Probing of Whole Cells for Determinative, Phylogenetic, and Environmental Studies in Microbiology. Journal of Bacteriology 172:762-770.
4.         Baudart J, Coallier J, Laurent P, Prévost M. 2002. Rapid and Sensitive Enumeration of Viable Diluted Cells of Members of the Family Enterobacteriaceae in Freshwater and Drinking Water. Applied & Environmental Microbiology 68:5057.
5.         Baudart J, Lebaron P. 2010. Rapid detection of Escherichia coli in waters using fluorescent in situ hybridization, direct viable counting and solid phase cytometry. Journal of Applied Microbiology 109:1253-1264.
6.         Rompre A, Servais P, Baudart J, de-Roubin MR, Laurent P. 2002. Detection and enumeration of coliforms in drinking water: current methods and emerging approaches. Journal of Microbiological Methods 49:31-54.
7.         Prescott AM, Fricker CR. 1999. Use of PNA oligonucleotides for the in situ detection of Escherichia coli in water. Molecular and Cellular Probes 13:261-268.
8.         Stender H, Broomer AJ, Oliveira K, Perry-O'Keefe H, Hyldig-Nielsen JJ, Sage A, Coull J. 2001. Rapid detection, identification, and enumeration of Escherichia coli cells in municipal water by chemiluminescent in situ hybridization. Applied and Environmental Microbiology 67:142-147.
9.         Regnault B, Martin-Delautre S, Lejay-Collin M, Lefevre M, Grimont PAD. 2000. Oligonucleotide probe for the visualization of Escherichia coli/Escherichia fergusonii cells by in situ hybridization: specificity and potential applications. Research in Microbiology 151:521-533.
10.       Fuchs BM, Wallner G. 1998. Flow Cytometric Analysis of the In Situ Accessibility of Escherichia coli 16S rRNA for. Applied & Environmental Microbiology64:4973.
11.       Tang YZ, Gin KYH, Lim TH. 2005. High-temperature fluorescent in situ hybridization for detecting Escherichia coli in seawater samples, using rRNA-targeted oligonucleotide probes and flow cytometry. Applied and Environmental Microbiology 71:8157-8164.
12.       Lepeuple A-S, Delabre K, Gilouppe S, Intertaglia L, de Roubin M-R. 2003. Laser scanning detection of FISH-labelled Escherichia coli from water samples. Water Science and Technology 47:123-129.
13.       Standard Methods for the Examination of Water and Wastewater, 2005, 21 ed. American Public Health Association; American Water Works Association; Water Environment Federation Baltimore, Maryland.
14.       Fuchs BM, Glockner FO, Wulf J, Amann R. 2000. Unlabeled helper oligonucleotides increase the in situ accessibility to 16S rRNA of fluorescently labeled oligonucleotide probes. Applied and Environmental Microbiology 66:3603-3607.
15.       Graczyk TK, Grimes BH, Knight R, Da Silva A, J., Pieniazek NJ, Veal DA. 2003. Detection of Cryptosporidium parvum and Giardia lamblia carried by synanthropic flies by combined Fluorescent in situ Hybridization and a Monoclonal Antibody, p. 228-232, The American Society of Tropical Medicine and Hygiene, vol. 68.
16.       Grimm D, H. Merkert, W. Ludwig, K. H. Schleifer, J. Hacker, and B. C. Brand. 1998. Specific detection of Legionella pneumophila: construction of a new 16S rRNA-targeted oligonucleotide probe. Appl. Environ. Microbiol.:2686-2690.
17.       Kikuchi YaJG. 2007. Spatial and temporal population dynamics of a naturally occurring two-species microbial community inside the digestive tract of the medicinal leech. Appl. Environ. Microbiol:1984-1991.
18.       Lemos V, T. K. Graczyk, M. Alves, M. L. Lobo, M. C. Sousa, F. Antunes, and O. Matos. 2005. Identification and determination of the viability of Giardia lamblia cysts and Cryptosporidium parvum and Cryptosporidium hominis oocysts in human fecal and water supply samples. Parasitology Research:48-53.
19.       Fuchs BM, G. Wallner, W. Beisker, I. Schwippl, W. Ludwig, and R. Amann. 1998. Flow cytometric analysis of the in situ accessibility of Escherichia coli 16S rRNA for fluorescently labeled oligonucleotide probes. Appl. Environ. Microbiol. :4973-4498.
20.       Lin X, S. G. Wakeham, I. F. Putnam, Y. M. Astor, M. I. Scranton, A. Y. Chistoserdov, and G. T. Taylor. 2006. 2006. Comparison of vertical distributions of prokaryotic assemblages in the anoxic Cariaco Basin and Black Sea by use of fluorescence in situ hybridization. Appl. Environ. Microbiol.:2679-2690.
21.       Amann RI, Binder, B. J., Olson, R. J., Chisholm, S. W., Devereux, R. & Stahl, D. A. 1990. Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl Environ Microbiol: 1919–1925.

Journal Articles:

No journal articles submitted with this report: View all 6 publications for this project

Supplemental Keywords:

Water, drinking water, groundwater, wastewater, effluent, feedlot, health effects, organisms, pathogens, bacteria, protozoa, biology, microbiology, monitoring, measurement methods, Northwest, Montana, MT, EPA Region 8, Native American, Crow Reservation, CryptosporidiumE. coliLegionella,Aeromonas hydrophila, fluorescent in situ hybridization, solid phase laser cytometer.

Progress and Final Reports:

Original Abstract
  • 2008
  • 2009 Progress Report
  • 2010 Progress Report
  • 2011