Final Report: Detection of Emerging Microbial Contaminants in Source and Finished Drinking Water with DNA Microarrays

EPA Grant Number: R828039
Title: Detection of Emerging Microbial Contaminants in Source and Finished Drinking Water with DNA Microarrays
Investigators: Straub, Timothy M. , Rochelle, Paul A.
Institution: Battelle Memorial Institute, Pacific Northwest Division , Metropolitan Water District of Southern California
EPA Project Officer: Nolt-Helms, Cynthia
Project Period: March 1, 2000 through March 1, 2003
Project Amount: $517,818
RFA: Drinking Water (1999) RFA Text |  Recipients Lists
Research Category: Drinking Water , Water

Objective:

The objectives of this research project are to develop and use DNA arrays for natural, turbid, and processed water supplies. DNA microarrays represent a potentially significant technology and analytical technique for the simultaneous detection of multiple pathogens in a single water sample, with the ability to incorporate live/dead discrimination via mRNA analysis. Cryptosporidium parvum and/or Helicobacter pylori will serve as model organisms.

Summary/Accomplishments (Outputs/Outcomes):

Microarray technology was developed, in part, to address the need to survey the RNA hybridization response profiles of all genes within a group of cells exposed to a stimulus or stressor, simultaneously. The primary application of this technology is the high throughput screening of next generation pharmaceuticals for the treatment of diseases such as cancer. By surveying the hybridization response profiles of all genes in exposed cells, investigators can inexpensively and quickly make predictions regarding the potential efficacy and potential side effects of a new medication before extensive testing on animal models and preclinical trials begin.

The ability to survey the DNA or RNA hybridization response profiles of many genes simultaneously makes microarray technology attractive for the detection of any waterborne pathogen as well. In traditional expression array experiments, all genes come from a single organism (e.g., human genome arrays). Our interest, however, is surveying the DNA hybridization response profile for many genes from many different organisms (see Figure 1). Specifically, we may need to:

  • Determine which strain of C. parvum or Escherichia coli O157:H7 was responsible for a waterborne disease outbreak. This may be manifested by minute sequence differences for a given gene for a given organism.
  • Determine the presence or absence of key virulence factor genes for a given organism. For example, the presence or absence of shiga-like toxin genes in E. coli O157:H7 or pathogenicity islands in H. pylori may make these organisms more or less pathogenic.

Because the response profiles of many genes can be determined, we would like to employ this technology for the detection of any possible threat that may be present in our drinking water.

A Conceptual Microarray for the Detection of Waterborne Pathogens

Figure 1. A Conceptual Microarray for the Detection of Waterborne Pathogens. All probes can be arrayed robotically on a simple glass microscope slide. For waterborne pathogens, we may need to determine the strain (C. parvum), determine the presence or absence of key virulence factors (E. coli O157:H7), and/or determine all threats that may be present (entire array).

Drinking water presents a unique challenge in that, by definition, pathogens should not be present in this source. This challenge becomes manifested in the sample processing steps needed to deliver potential DNA targets in the water sample to the microarray for the detection of these pathogens. It is estimated that the sample DNA must have a minimum of 10,000 copies for each gene that is to be detected on the array. For DNA extracts from these water samples, each gene may only be represented once in the entire genome. This implies that in vitro amplification (PCR or cell culture) techniques are needed to detect a gene of interest, if present, in the sample. Our work on this project ultimately focused on both PCR and combined cell culture hybridization approaches to develop microarray technology for the detection of pathogens in drinking water.

PCR Approach for C. parvum and H. pylori

C. parvum. We found that the PCR approach is most useful for the forensic identification of an organism. Our model organisms were C. parvum and H. pylori. For C. parvum, a 68-probe microarray was designed to detect single nucleotide sequence differences (single nucleotide polymorphisms [SNPs]) between known isolates for the gene coding the 70-kilodalton heat shock protein ([hsp70], Straub, et al., 2002). A single PCR amplicon was generated for hsp70. This amplicon was labeled by the addition of Cy3 to the PCR primers. This amplicon, now carrying the fluorescent dye, then could be detected after hybridization to the microarray probes.

In a variation on this theme, PCR primers were redesigned to carry a promoter sequence recognized by T7 RNA polymerase. After PCR amplification, the PCR is mixed with a cocktail consisting of T7 RNA polymerase, reverse transcriptase, dNTPs, NTPs, and other additives. Upon hybridization to the array, in vitro transcription of the PCR product produces a secondary amplification product consisting of single stranded RNA. This RNA hybridizes to the probes on the array. If the RNA sequence is a perfect complement to the array probe, a primer extension reaction, mediated by reverse transcriptase, occurs. During reverse transcription, the target RNA:microarray probe duplex is labeled by incorporation of fluorescently labeled dNTPs during reverse transcription of the array probe (see Figure 2).

Conceptual Diagram of a New Variation for the Forensic Identification of <em>C. parvum</em> Isolates

Figure 2. Conceptual Diagram of a New Variation for the Forensic Identification of C. parvum Isolates. RNA is generated by in vitro transcription of a PCR product containing a T7 RNA polymerase promoter sequence. The RNA will hybridize to the array probes. Reverse transcriptase will extend and label the microarray probe with fluorescent dNTPs if the RNA hybridized to the probe is a perfect match (middle figure). If the RNA and probe sequence are mismatched, no extension will occur (right figure).

Results for this method provided equivalent results to the results published by Straub, et al. (2002). In most cases, the new method provided improved signal that made sequence discrimination between isolates much easier. For instance, in the published method, statistical analysis was needed to determine if the perfectly matched probe sequence had the greatest hybridization signal intensity of all other probe combinations. For the new method, the results were unequivocal (see Figure 3).

Comparison of Hybridization Response Profiles for the Forensic Identification of <em>C. hominis</em> (formerly <em>C. parvum</em> Genotype I) and <em>C. parvum</em> Genotype II Using the Secondary RNA Amplification Approach

Figure 3. Comparison of Hybridization Response Profiles for the Forensic Identification of C. hominis (formerly C. parvum Genotype I) and C. parvum Genotype II Using the Secondary RNA Amplification Approach

Using the new approach, seven of nine predicted sequences were elucidated for C. hominis. The signal for SNPs 1371 and 1419 was not significantly above background, such that the call regarding a perfect match could be made. For the single nucleotide difference at nucleotide position 1368 between C. hominis and the monkey genotype, the monkey genotype probe was hybridized. This confirmed the result published in Straub, et al. (2002), and is likely a sequence error in GenBank. For C. parvum genotype II, eight of nine single nucleotide differences were correctly determined. The results for 1542 should be reconfirmed to make sure that this failure was not caused by an accidental swap of oligonucleotides during the microarray fabrication process.

H. pylori. A multiplexed PCR product (three different amplicons) could be generated to determine the presence or absence of virulence factor genes in H. pylori (vacA: signal (s) and middle (m) regions, and cagA: pathogenicity island). The microarray then could be used to genotype isolates of H. pylori based on single nucleotide differences within the two virulence genes based on published sequences (see Table 1 and Figure 4; Straub, et al., 2004).

Table 1. SNP Discrimination Within the Signal Region of the vacA Gene of H. pylori. The two genotypes that were tested were s1a/m1/cagA(+) (ATCC 43504) and s1a/m1/cagA(-) (ATCC 700392). The s1a52 probes contain the same degenerate bases; however, those probe sequences start nine bases upstream of the s1a61 probes. Despite this, identical hybridization results were achieved (see Figure 4). This indicates that the first “R” is truly a guanine, but the second “R” remained indeterminate (A or G).

Reported Sequence for s1a61 Probe Suite

G G A G C R T T R G T C A G C A T C A C

Printed Probe Variation s1a61/A (Hybridized) G G A G C G T T G G T C A G C A T C A C
Printed Probe Variation s1a61/B (Hybridized) G G A G C G T T A G T C A G C A T C A C
Printed Probe Variation s1a61/C (Not Hybridized) G G A G C A T T A G T C A G C A T C A C
Printed Probe Variation s1a61/D (Not Hybridized) G G A G C A T T G G T C A G C A T C A C

Hybridization Patterns for Two Strains of H. pylori That Have the s1a/m1 Genotype for the vacA Gene

Figure 4. Hybridization Patterns for Two Strains of H. pylori That Have the s1a/m1 Genotype for the vacA Gene. The presence or absence of the cagA gene is clearly seen. This array shows that it is possible to achieve single nucleotide discrimination and determine the presence or absence of a gene.

Integrated Cell Culture Hybridization Approach

The ability to perform highly multiplexed PCR (e.g., a single PCR that contains more than one primer pair to detect different gene targets) is the single limiting factor that impedes the use of microarrays to detect simultaneously any pathogen in a water sample. Furthermore, PCR does not provide live versus dead discrimination, and this may be an important factor regarding safety
decisions concerning a suspect water source. As stated previously, approximately 10,000 copies of a gene are needed before array-based detection of that gene is possible. For bacteria, addition of a water sample to rich media such as nutrient or tryptic soy broth, followed by overnight incubation, may provide an attractive alternative. Key members of the family Enterobacteriaceae, pseudomonads, aeromonads, and some Vibrio grow relatively quickly in this media, and, depending on their ability to be resuscitated, will have enough “copies” of their key genes to be detected by direct hybridization protocols.

Our research focused on the direct isolation of mRNA from key toxin genes and their direct detection using microarrays. There were several reasons for choosing to isolate RNA versus DNA:

  1. RNA from these genes comes from actively respiring organisms. Thus, culture plus RNA isolation equals viability.
  2. Although there may be just one copy of the gene of interest in genomic DNA, there may be several to thousands of copies of mRNA for that same gene. Empirically, the chances for detection of a gene of interest should increase.
  3. There is the possibility of more efficient binding of RNA to DNA probes based on affinity, and the fact that, as a single stranded species, it would not hybridize with itself like genomic DNA or a double-stranded PCR product can.

We developed methods to isolate and label RNA directly from enriched cultures (see Figure 5). We initially were concerned that the RNA method may not have the specificity of a combined PCR and hybridization technique. However, as shown in Figure 5, microarrays hybridized with labeled RNA showed excellent specificity in that: (1) we were able to genotype different E. coli O157:H7 isolates; (2) we were able to distinguish differences between different organism groups (e.g., Salmonella could be distinguished from Shigella, which also could be distinguished from E. coli); and (3) we confirmed that RNA from nontarget organisms cannot hybridize to this array (e.g., Acinetobacter).

Array-Based Detection of RNA From Members of the Enterobacteriaceae Family and Related Genera. Boxes drawn around the hybridized probes indicate the expected hybridization outcomes

 

Figure 5. Array-Based Detection of RNA From Members of the Enterobacteriaceae Family and Related Genera. Boxes drawn around the hybridized probes indicate the expected hybridization outcomes. The positive results for the fliC gene for Shigella, Salmonella, and E. coli 25922 were confirmed by BLAST, and thus, are not false positive results. A. baumanii belongs to the Moraxellaceae family, and RNA from this organism should not (and did not) hybridize to this array.

The next question we investigated was whether mixed cultures of organisms could be detected using the combined cell culture direct hybridization approach. For our initial tests, roughly equivalent concentrations of E. coli ATCC 25922, S. typhimurium ATCC 14028, and Shigella species that had been grown overnight in tryptic soy broth were mixed. RNA was isolated, labeled, and hybridized to the array. The results shown in Figure 6 illustrate that mixtures can be determined. Future work is needed to characterize environmental samples, where it is likely that the target organisms will not start out at equal concentrations prior to enrichment.

Simultaneous Detection of Three Organisms Using the Integrated Cell Culture, RNA Isolation, and Labeling Approach

Figure 6. Simultaneous Detection of Three Organisms Using the Integrated Cell Culture, RNA Isolation, and Labeling Approach. Shigella can be distinguished from the other two organisms based on the presence of the ipaH gene. Presence of Salmonella can be distinguished based on the hybridization response of invA. The presence of E. coli can be deduced by the hybridization of the uidA3 probe that is absent in Shigella.

Conclusions:

Additional research is needed to develop better methods for universal, in vitro amplification of DNA and/or RNA (if viability information is needed) from environmental water samples before microarray technology can be useful for the detection of waterborne pathogens.

Labeling via multiplexed PCR currently is limited to approximately 10 different primer pairs, and optimization of such reactions often is time consuming given the number of variables that need to be examined (e.g., primer design, primer concentrations for each pair, dNTP concentration, Mg2+, concentration, etc.). Universal PCR approaches based on 16S rRNA (prokaryotic) and 18S rRNA (eukaryotic) have been reported in the literature, but sequence conservation, especially in the Enterobacteriaceae family, may make pathogen discrimination extremely difficult. Additionally, this approach will not work for viral pathogens of interest. True universal methods, such as amplification methods with random primers, need further validation before they can be recommended.

The primary advantage of preculture enrichment techniques is the live versus dead discrimination determination, which is important in making risk-based decisions. Also, with enrichment, the direct labeling and hybridization approach resolve the issue of highly multiplexed detection on the microarrays. Cell culture approaches work well for fast growing organisms, but slow growers or organisms requiring special culturing conditions, such as Campylobacter, H. pylori, and Mycobacterium avium complex could easily be missed. Viral cell culture also is problematic for emerging threats like norovirus.

Finally, research is needed to speed the overall process of detecting waterborne pathogens. Most of the hybridization methods we investigated require only 2 hours, but the amount of time to actually process the sample to get to the microarray hybridization assay often is 8 or more hours (if culture enrichment is used). Few methods exist for the co-concentration of viruses, bacteria, and protozoa threats, and of those methods, there is no method to process the concentrate to remove inhibitors for downstream analysis.


Journal Articles on this Report : 7 Displayed | Download in RIS Format

Other project views: All 27 publications 8 publications in selected types All 7 journal articles
Type Citation Project Document Sources
Journal Article Kingsley MT, Straub TM, Call DR, Daly DS, Wunschel SC, Chandler DP. Fingerprinting closely related Xanthomonas pathovars with random nonamer oligonucleotide microarrays. Applied and Environmental Microbiology 2002;68(12):6361-6370. R828039 (Final)
  • Abstract from PubMed
  • Full-text: AEM
    Exit
  • Other: AEM PDF
    Exit
  • Journal Article Small J, Call DR, Brockman FJ, Straub TM, Chandler DP. Direct detection of 16S rRNA in soil extracts by using oligonucleotide microarrays. Applied and Environmental Microbiology 2001;67(10):4708-4716. R828039 (Final)
  • Abstract from PubMed
  • Full-text: AEM
    Exit
  • Other: AEM PDF
    Exit
  • Journal Article Straub TM, Daly DS, Wunshel S, Rochelle PA, DeLeon R, Chandler DP. Genotyping Cryptosporidium parvum with an hsp70 single-nucleotide polymorphism microarray. Applied and Environmental Microbiology 2002;68(4):1817-1826. R828039 (2002)
    R828039 (Final)
  • Abstract from PubMed
  • Full-text: AEM
    Exit
  • Other: AEM PDF
    Exit
  • Journal Article Straub TM, Chandler DP. Towards a unified system for detecting waterborne pathogens. Journal of Microbiological Methods 2003;53(2):185-197. R828039 (2002)
    R828039 (Final)
  • Abstract from PubMed
  • Full-text: Science Direct Full Text
    Exit
  • Other: Science Direct PDF
    Exit
  • Journal Article Straub TM, Quinonez-Diaz MD, Valdez CO, Call DR, Chandler DP. Using DNA microarrays to detect multiple pathogen threats in water. Water Supply 2004;4(2):107-114. R828039 (Final)
  • Abstract: IWA Online Abstract
    Exit
  • Journal Article Straub TM, Dockendorff BP, Quinonez-Diaz MD, Valdez CO, Shutthanandan JI, Tarasevich BJ, Grate JW, Bruckner-Lea CJ. Automated methods for multiplexed pathogen detection. Journal of Microbiological Methods 2005;62(3):303-316. R828039 (Final)
  • Abstract from PubMed
  • Full-text: Science Direct
    Exit
  • Other: Science Direct pdf
    Exit
  • Journal Article Willse A, Straub TM, Wunschel SC, Small JA, Call DR, Daly DS, Chandler DP. Quantitative oligonucleotide microarray fingerprinting of Salmonella enterica isolates. Nucleic Acids Research 2004;32(5):1848-1856. R828039 (Final)
  • Abstract from PubMed
  • Full-text: NAR
    Exit
  • Other: NAR PDF
    Exit
  • Supplemental Keywords:

    risk assessment, bacteria, signature, effluent, nucleic acids, biology, epidemiology, genetics, pathology, measurement methods, DNA microarrays, Helicobacter pylori, assessment technology, bacteria monitoring, Cryptosporidium, detection, exposure, exposure and effects, microbial contamination, microbial risk assessment, microbiological organisms, microbiology, microorganism, monitoring., RFA, Health, Scientific Discipline, PHYSICAL ASPECTS, Water, Ecosystem Protection/Environmental Exposure & Risk, Health Risk Assessment, Risk Assessments, Monitoring/Modeling, Environmental Monitoring, Physical Processes, Drinking Water, microbial contamination, monitoring, measurement , microbial risk assessment, microbiological organisms, detection, exposure and effects, exposure, bacteria monitoring, other - risk assessment, human exposure, treatment, cryptosporidium , measurement, water quality, microorganism, assessment technology, Helicobacter pylori, human health risk

    Relevant Websites:

    None.

    Progress and Final Reports:

    Original Abstract
  • 2000 Progress Report
  • 2001