Toward Forensic Uses of Microbial Source Tracking

ABSTRACT

The science of microbial source tracking has allowed researchers and watershed managers to go beyond general indicators of fecal pollution in water such as coliforms and enterococci, and to move toward an understanding of specific contributors to water quality issues. The premise of microbial source tracking is that characteristics of microorganisms that are strongly associated with particular host species can be used to trace fecal pollution to particular animal species (including humans) or groups, e.g., ruminants or birds. Microbial source tracking methods are practiced largely in the realm of research, and none are approved for regulatory uses on a federal level. Their application in the conventional sense of forensics, i.e., to investigate a crime, has been limited, but as some of these methods become standardized and recognized in a regulatory context, they will doubtless play a larger role in applications such as total maximum daily load assessment, investigations of sewage spills, and contamination from agricultural practices.

INTRODUCTION

Beyond Indicator Organisms

The first known waterborne pathogen was Vibrio cholerae, the culprit in the London cholera epidemic of 1854 (1). The discovery that disease could be transmitted from feces to humans via drinking water led to the realization that preventing fecal pollution of water could protect public health, leading to the search for a readily culturable bacteria that was always found at high levels in sewage. Klebsiella and Escherichia coli (first termed Bacillus coli) were identified as indicators of sewage in the 1880s (2–4). In cities teeming with open sewers that discharged directly to water bodies, these intestinal “coliform” bacteria served the purpose of detecting fecal contamination in drinking water and alerting authorities of public health risk. Among the first formal regulations for microbial drinking water quality was the U.S. Public Health Service total coliform standard of 1914 (5). Among the first regulations for discharges to ambient (surface) water quality was the Federal Water Pollution Control Act, in 1948, which was amended in 1972 to the Clean Water Act. We now understand that fecal or sewage pollution of water by human or animal sources may spread microbial pathogens such as Cryptosporidium parvum, Giardia lamblia, Salmonella spp., E. coli 0157:H7, Campylobacter spp., Shigella spp., and viruses such as hepatitis A, rotaviruses, adenoviruses, enteroviruses, and noroviruses (6–8).

Societal tolerance for infections contracted from ingestion of or contact with water has diminished greatly over the past century, particularly in developed countries, leading to new areas of regulation, remediation, and potential for litigation. The use of “indicator organisms” such as E. coli and other coliform bacteria to provide a warning of increased risk of pathogen presence has spread from drinking water to wastewater effluent and, finally, to environmental waters, as well as food and a myriad of other products. Indicator organisms are used to assess contamination of water worldwide in a variety of applications, e.g., total maximum daily load (TMDL) programs in U.S. surface waters, in coastal beach waters in compliance with the Beaches Environmental Assessment and Coastal Health Act in the U.S. and the Blue Flag Program in the European Union, and in shell-fishing waters according to the National Shellfish Sanitation Program in the U.S. and the Shellfish Water Directive in the European Union (9, 10).

As the purview of indicator organisms widened from drinking water to environmental waters, the failings of indicator organisms as surrogates for pathogens became ever more apparent. Many recent reviews include extensive information about indicator organisms, including their drawbacks (8, 9, 11, 12). Indicator bacteria belonging to the coliform family (total coliforms, fecal coliforms, and E. coli) and indicators from other taxa such as Clostridium perfringens, enterococci, and coliphages share the characteristic of widespread distribution in the feces of humans and other animals; therefore, they provide no information about contamination sources (7, 13, 14).

The blindness of the indicator organisms to the contamination source is a major stumbling block for many applications in environmental and public health microbiology (1). In the case of TMDL assessment, it is very difficult to determine contamination sources when the indicator is widely distributed among many animal hosts (2, 15). Human health risk from exposure to contaminated water varies according to contamination source, because the pathogens associated with the feces of various animals are often dissimilar, leading to varied estimates of exposure to pathogens depending upon fecal sources (16–19). For example, human fecal contamination is the only source of pathogenic viruses that infect humans, because these viruses are host specific; on the other hand, cattle and poultry feces are likely to contain E. coli O157:H7 and Campylobacter, respectively, and these bacterial pathogens are also sometimes found in sewage (18). Quantitative microbial risk assessment (QMRA) models require information on dose-response (minimum infectious dose) and exposure assessment pathways of contact and magnitude of pathogens (20), making accurate risk assessment highly dependent on knowledge of the contamination source (3). Remediation and prevention of further contamination depends absolutely on knowledge of the source, and (4) legal responsibility for contamination is difficult to ascertain solely on the basis of indicator organism levels, as evidenced in Waterkeeper Alliance, Inc. v. Hudson (see below).

Since the 1990s, research has produced methodologies that link certain fecal indicator organisms to a host group by genetic typing or detecting host-associated genes, leading to the field of microbial source tracking (MST). For recent comprehensive reviews see references 7, 13, 14. Hagedorn et al. defined MST as “a method used to determine the source of fecal bacteria and establish whether fecal bacteria are being introduced into water bodies through human, wildlife, agricultural or pet wastes” (21). This definition is limited, because MST methods can also target viruses (22) or protozoa (23–25), but it captures the required connection between a particular species, strain, or type of microorganism and the host gastrointestinal tract that harbored it prior to its discharge into water.

Theory and Practice of MST

Microbe-host association

The gastrointestinal tract of any one animal species represents a highly diverse group of microhabitats for enteric microorganisms. The gastrointestinal tracts of different animal species provide varying niches for microorganisms, leading to development of microbiota that are characteristic of a given host species but are also influenced by diet and husbandry (26, 27). Long before the development of PCR, high-throughput DNA sequencing, and metagenomics, microbiologists interested in water quality recognized the desirability of discriminating between fecal contamination of water by humans or other animals and recognized the preferential association of some microorganisms with specific hosts.

One of the early attempts at achieving a human-nonhuman dichotomy was the fecal coliform to fecal streptococci ratio (28). A ratio greater than 4.0 was considered indicative of a human sewage source, while lower ratios indicated a nonhuman (other animal) source. The fecal coliform to fecal streptococci ratio was incorporated into Standard Methods for the Examination of Water and Wastewater. This approach was invalidated over time, because research revealed the differential persistence of the groups under various environmental stresses and the variability in the levels of the two indicator groups in various hosts. The 1998 and subsequent editions of “Standard Methods” no longer contain the fecal coliform to fecal streptococci ratio (29).

Some history of MST development has been outlined in other recent publications (13, 30, 31). Over the past two decades, MST methodologies have shifted from library-dependent methods, where fecal indicator bacteria (FIB) such as E. coli and Enterococcus spp. are typed on an isolate-by-isolate basis, to library-independent methods that rely on PCR or quantitative PCR (qPCR) to target specific genes of host-associated bacteria (13, 32). A simplified depiction of the principle behind library-independent MST methods (Fig. 1) begins with the host species of interest in the study area. A microorganism (usually a bacterium or virus) that is commonly found in high numbers in feces of the target host, and that is generally absent or present in very low concentration in the feces of other animals (sometimes termed the “source identifier”), is identified for each host group, e.g., cattle and humans (6, 30). The genetic material (DNA or RNA) is used to derive a genetic marker (nucleic acid sequence) for the source identifier. A separate PCR or qPCR test is then developed for each marker. PCR provides only presence/absence results, while qPCR provides gene copies, which can be used to approximate the number of target microorganisms. Emerging methods for MST include microarray, in which hundreds of pathogen- and host-associated genes can be simultaneously tested (33–35), and metagenomic analysis of microbial populations in feces and water (36–39). The practice and potential for the use of MST methods in environmental forensics is the subject of this article.

FIGURE 1.

A simplified scheme for library-independent MST studies starts with identifying the animals within the study area that are likely to be major contributors to contamination. In this simplified version, we show two sources, cow and human (sewage). In the cow example, a type of fecal bacterium that is strongly associated with cow gastrointestinal tracts is denoted C. The bacterium is detected by extracting DNA and using PCR or qPCR to test for the DNA sequence (marker) that is specific to C. The sewage example follows the same flow, and the human-associated marker is denoted H. Note that in the case of certain viruses with RNA genomes, e.g., Enterovirus, RNA, rather than DNA, is extracted and tested. Water samples, in which the contamination source is unknown (?), can be tested by the MST methods to determine whether (PCR) and how much (qPCR) of the MST marker is present.

MST APPLICATIONS

Environmental Assessment

Microbial forensics, as distinguished from more conventional forensics that typically involve only one species (humans), is inherently complex due to the large number of potential bacterial and viral species implicated and the intricate microbial dynamics involved (40). Thus, the complexity of MST in forensic applications is affected by the number of contributing host species involved (e.g., humans, cattle, poultry), the number of microbial species that must be considered, as well as the environmental survival and resiliency of each group of target organisms over time.

The rapid development of MST approaches and technology has expanded the tool chest for environmental and public health managers and decision-makers, including the legal system. MST techniques can be applied in the area of microbial forensics, a discipline based on microbiology and epidemiology that is used to answer questions within a legal framework (40, 41). Whether the issue is the source of a pathogen, the source of mail items contaminated with a bioterrorism agent, a potential source of environmental contamination, or a standard of proof in a court of law, it is essential to identify as accurately as possible the human and/or animal contributors of fecal pollution. With such information, informed decisions can be made with respect to remediation of contamination, or to define and allocate legal and fiscal responsibility for environmental pollution (42).

MST, like other environmental forensic tools, must be able to accurately identify sources of fecal contamination in the environment. Considerable guidance is needed for when and how to best use the various available assays in specific circumstances (43). Successful application of MST methodologies must include consideration of sample collection, handling and preservation, method selection, case analysis, interpretation of results, validation, and quality assurance to yield useful microbial forensic data (40, 41, 44).

Numerous MST tools and approaches are available, the selection of which is influenced by the complexity of environmental samples and the many variables that affect microbial survival and growth (6, 32, 45, 46), as well as the specific questions at hand. Chemical, culture-based, isolate-by-isolate (library-dependent), sample-specific, and host-specific (library-independent) approaches may be applicable alone or in combination for particular circumstances involving DNA-based methods that use either PCR or qPCR (32, 43, 47–51). Library-dependent MST methods showed early promise (24, 32, 48) and were particularly attractive because (i) FIB such as E. coli or enterococci, which are directly linked to the regulatory rules, could be typed and (ii) the relative contribution of various host species to contamination was estimated from the analysis (source apportionment). Significant limitations of library-dependent methods have been identified; e.g., the construction of the library is costly, constant library maintenance and validation are needed, and libraries that are representative of FIB diversity in host feces must be very large (32, 43).

A wide variety of microorganisms can serve as targets for library-independent MST methods. Table 1 provides examples of MST methods that are based on different host-associated microorganisms. The use of bacteriophages for MST was first proposed in the early 1980s (52–55). Viruses such as human polyomavirus and adenovirus have the advantage of being highly host-specific, but their relatively low concentration in sewage can result in false-negative results (56, 57). Pepper mild mottle virus, while not human specific, is found in high concentrations in sewage and can be useful for indicating sewage contamination (58). Gene fragments representing source-associated genetic markers frequently exist in the bacterial genus Bacteroides and the larger taxonomic group Bacteroidales, which is among the dominant commensal bacterial groups in the human large intestine (43). Members of the Bacteroidales also occur broadly in animal feces (59), but differences in 16S rRNA and other gene sequences can be useful in discriminating among the potential sources. MST using fecal Bacteroidales was used successfully in conjunction with fate and transport modeling techniques in San Pablo Bay to monitor fecal pollution (60). In addition, two indicators of human fecal sources (human Bacteroidales and Lachno2) were used in Lake Michigan water samples to determine the extent of fecal impacts (61). Lee et al. demonstrated successful use of qPCR assays targeting Bacteroidales 16S rRNA markers to identify fecal sources in a quantitative manner when combined with land use data and local weather information (62).

TABLE 1.

Select examples of MST methods currently in use, including some advantages and disadvantages of specific methods^a

Host	Marker	Microorganism	Advantages	Disadvantages	Citation
Bird	GFD	Helicobacter	Detects most bird species		77
Cow	Cow MM3	Bacteroidales	Specific to cattle	Not found in all cattle	124
Human	HF183	Bacteroides dorei	High level in sewage aids sensitivity in environmental samples	Not completely specific, e.g., deer, dogs, chickens	133
Human	HPyVs	Human polyomaviruses	Complete specificity	Low concentration in sewage restricts sensitivity in environmental samples	22
Poultry	LA35	Brevibacterium	High level in poultry litter aids sensitivity in environmental samples	Some cross-reactivity with other bird species	106
Ruminant	CF128	Bacteroidales	Detects a variety of ruminants, e.g., cattle, deer, goats, sheep	May not be specific enough depending on the application	134

^a In the case of HF183, multiple qPCR methods exist and vary largely in primer and probe positions, but the recently developed one cited is the most likely to become a regulatory tool. Note that all of these MST methods should be vetted for specificity and sensitivity toward the target host in the geographic area of interest.

Utilization of MST in Environmental Studies

The health risks that may be associated with fecal contamination of environmental media (e.g., water, soil, sediments) or the food supply, or inhalation of microbial species can range from self-limiting, minor gastrointestinal or respiratory distress to catastrophic illness and fatalities in individuals or populations. Assessment of the potential health risk requires the ability to differentiate among possible contamination sources using robust and reproducible methods, given that fecal contamination from wildlife and other nonhuman sources typically is judged to represent a lower human health risk than human fecal contamination (43, 63–65), although sources such as cattle are also considered high risk due to the amount of waste produced and the prevalence of zoonotic pathogens in their feces (65). The occurrence, frequency, and severity of observed disease typically is related to factors including the intensity of the contamination, the route of exposure, the frequency/duration of exposure, the number and type of microbial species present, and the speed with which the event is recognized, understood, and addressed. The practical ability to address issues of environmental, public health, agricultural, veterinary, or criminal significance often hinge on the availability of techniques for reliable identification and possible quantification of the microbial/pathogen pollution source. This is particularly true when microbial contamination originates from nonpoint sources (e.g., agricultural or urban runoff, compromised pipeline infrastructure, wildlife) as opposed to point sources (e.g., effluent from wastewater treatment plants or known industrial sources).

Recent research has addressed the establishment of a risk-based approach to interpreting the relative health significance of MST biomarker levels in some instances by using QMRA (66, 67). QMRA is intended to quantify relative risks to humans from exposure to recreational water that is impacted by different sources of fecal contamination (43). For example, Boehm et al. found that health risks increased with an increase in the concentration of human qPCR markers (HF183 and HumM2) in waters affected by raw sewage (66). Others have shown a correlation between human-associated MST markers and the detection of adenovirus (68). Oliver et al. pointed out that there is more to consider beyond detection of markers for contamination sources, such as the social and economic ramifications of interpreting such studies for affected beach communities (69). Others have recommended that further epidemiological studies and QMRA be conducted, so that health risks to bathers can be assessed and portrayed accurately (59). Chronic pollution of coastal waters with FIB in some areas has led to increased interest in identifying sources of beach contamination for risk evaluation and remedial purposes; however, the most recent U.S. Recreational Water Quality Criteria are similar to the previous criteria because they rely on quantification of culturable E. coli or enterococci (70) or optional new criteria based on a rapid qPCR method for Enterococcus spp. Although the new criteria are based on epidemiological studies combined with ambient water quality measurements, they do not include assessment of contamination source(s).

Multiple MST methods continue to be recommended as suites of approaches, since each method has strengths and weaknesses that can limit the overall usefulness of PCR (7). As an example, Di Giovanni et al. identified 16 wildlife fecal samples which cross-reacted with the human HF183 Bacteroidales PCR marker (71). Abdelzaher et al. suggested development of a “comprehensive toolbox” to address beach regulation plans in a case study of Virginia Key in Miami-Dade County, Florida, where multiple potential nonpoint microbial sources were identified (72). Those authors stressed source identification and source prevention as keys to establishing a safe beach environment when they found no relationship between the use of FIB and gastrointestinal illness (72). More recent guidance from California recommended a tiered approach to identification of fecal pollution sources, as contained in the Source Identification Protocol Project (73). While common fecal indicators (e.g., E. coli, other coliforms) have been the predominant subject of technical and legal attention as a result of widespread detection and representation in water quality data, the genetic heterogeneity and wide host profile for those common taxonomic groups provide a driving force for development of what may prove to be more focused bacterial probes (74). In addition, Fujioka et al. suggested other alternative fecal indicators (e.g., C. perfringens, coliphages, and Bacteroides), which may work to overcome issues that are often related to the more conventional indicators (59).

Recent studies of watershed fecal contamination highlight the promise of MST for regulatory programs, such as TMDL allocations. A California study used qPCR to test for human-associated (HF183Taqman and HumM2), dog (DogBact), shorebird (Gull2), and horse (HoF597) fecal markers (75). Canine waste was a dominant contributor of fecal contamination among the sources tested. Two caveats from this study are notable: (i) the dog marker was also found in wild canines, such as coyote and fox, and (i) the Gull2 marker is not prevalent in all bird species but occurs most frequently in gulls and pelicans, so other bird species may not have been detected by the Gull2 marker. Use of a more general bird marker such as GFD (76, 77) in this study might have supported a different conclusion concerning the contribution of birds to fecal microorganism occurrence in the watershed.

Several studies illustrate the scope of MST techniques as applied in the global research arena. Boehm et al. analyzed the specificity and sensitivity of 41 MST methods from 27 laboratories, reporting that a number of the assays performed well in identifying ruminant, cow, human, gull, dog, and pig feces and proposed field validation as well (78). Ervin et al. evaluated nine methods involving 12 different sources of fecal loading, reporting that the method of measurement can greatly influence conclusions regarding minor and dominant pollution sources when more than one fecal source is present (79). Gordon et al. further demonstrated the robustness of culture-independent and library-independent MST methods in an analysis of three human markers in a multilaboratory study (80). As pointed out by Badgley and Hagedorn (43), caution is appropriate in MST applications, because research is ongoing on several fronts, including the decay rates of culturable indicators and the utility of various DNA-based markers in environmental scenarios (81–86).

Protocol validation and assessment of reliability are crucial in a given environmental setting (32, 87). In addition, the adaptation of new and scientifically defensible methods for assessing, interpreting, and managing microbial pollution risk is an ongoing process (14, 49). Certain approaches have been replaced for the most part; e.g., library-dependent methods such as antibiotic resistance analysis (88–90) and ribotyping (91) are now infrequently used because of methodological challenges and difficulties in interpretation (7, 13, 14, 32). For example, Moore et al. concluded that the library-dependent MST ribotyping and antibiotic resistance analysis methods were not ideal for determination of fecal pollution source(s) in large urban watersheds and that those techniques were better applied to water bodies affected by a limited number of potential sources, in a limited geographical area, over a short time period (89). Advances in MST continue, with some limitations still apparent. For example, one marker can be used to identify only one specific fecal source, and it is unlikely that a single marker will be identified that is 100% sensitive (found in all individuals of a given host species) and uniquely specific to a contaminant source (43). Environmentally adapted strains of E. coli should be considered as potential sources in MST studies where that indicator organism is the target (92), in recognition of the fact that indicator organisms that may be released from an original source can survive, persist, reproduce, and perhaps change in the extra-enteric environment (7, 9).

TMDL Evaluations

TMDL describes a maximum loading profile for a pollutant that a water body can receive and still meet water quality standards (43, 93). TMDL studies must be performed once a water body has been added to the U.S. EPA 303(d) “impaired list,” after it has been shown not to meet applicable regulatory criteria for FIB in designated use scenarios (e.g., fishing, swimming, supporting aquatic life, supplying drinking water). Comprehensive studies that include MST methods, FIB testing, hydrological surveys, and “boots on the ground” sanitary inspections can provide valuable information concerning fecal contamination sources (24, 33, 93–98).

The development of reliable methods within TMDL programs to distinguish between bacteria from human and animal origins, and to determine load allocations to a specific watershed, continues to improve the value of the program (99, 100). MST methods can be used in the TMDL process to supplement existing tools and methods in the identification of dominant contributors (93). The success of a TMDL program in water quality improvement depends strongly on positive identification of microbiological sources and knowledge regarding relative or quantitative loading attributable to those sources (47). The use of MST is expanding as more researchers become aware of its considerable potential (21). One useful and specific application of MST has been in the assessment and evaluation of agricultural impacts (i.e., livestock) to bodies of water. Beef cattle (24, 88, 101), poultry (83, 102–106), and wildlife (49, 107–109) have been investigated, with methods available to provide reliable predictive data on microbial sources and their relative contributions. State and local officials often use MST data to recommend best management practices for agriculture, with the goals of reducing indicator organism populations and achieving TMDLs (110, 111). Researchers have used Pig-1-Bac and Pig-2-Bac for tracking the off-site spread of swine fecal waste into surface waters downstream of concentrated animal feeding operations, especially during rain events (112).

Fecal contamination in rural watersheds, as with urban watersheds, can originate with domestic animals, wildlife, and concentrated animal feeding operations. As noted, MST can be applied to source water protection, watershed management, beach monitoring/closure, and shellfish bed monitoring/protection (100, 113). Accurate source apportionment, in which a single approach is used for all sources, has not been achieved, but ongoing research suggests that it could be possible in the future (114).

Applications in the Legal System

When improvement of wastewater infrastructure and/or modification of animal waste disposal practices is necessary to improve water quality, controversy may arise due to the cost and difficulty of implementing such changes, and they may not be accepted voluntarily. It is clear that many TMDL source allocation efforts, and accompanying remediation plans, ultimately will enter the administrative and legal sphere. For the reasons listed previously, MST can be a valuable tool in determining the fecal sources involved in environmental microbial contamination. The positive identification of pollution source(s) in a specific case is essential to prevention, interdiction, and remediation of impacts, and the process ultimately will help to assign responsibility in a legal or regulatory context (92, 115–117).

Microbial forensics can be considered a relative newcomer to the legal arena, since such evidence was first produced in a criminal trial in 1998 (118). A judge in that case concluded that the evidence satisfied necessary scientific and evidentiary criteria, as defined by what is known as the “Daubert Decision” (119) (discussed below). In 2001, the techniques of microbial forensics and source tracking were employed in the scientific and legal investigations related to a bioterrorism event in which letters laced with Bacillus anthracis spores caused lethal cases of anthrax in several U.S. individuals (40, 118). Bioterrorism events can create fear and cause great disruption, as well as affect economic welfare (44).

In a manner arguably analogous to terrorism, fecal contamination to water systems can involve significant economic losses and serious disturbances of natural and ecological environments, as well as contentious, complex legal proceedings. Microbial forensics can play a major role in the attribution and deterrence of biological warfare and terrorism, although significant gaps remain in both scientific understanding and operational capability (118). The courtroom presentation and underlying evidentiary processes are exceptionally challenging when dealing with living, self-replicating, and evolving organisms.

State and federal rules may differ among jurisdictions, yet both are designed to ensure the reliability of testimony for use by the trier of fact (i.e., the judge and/or jury) in the courtroom. It is imperative that scientists who may be called upon as expert witnesses understand the legal admissibility requirements and constraints placed on the introduction of any scientific evidence, including MST (115). Blanch et al. pointed out that “determining the source of fecal contamination in aquatic environments is essential for estimating the health risks associated with pollution, facilitating measures to remediate polluted waterways, and resolving legal responsibility for remediation” (23). Experts in all scientific disciplines must use technically defensible and reproducible methods and understand their limitations (40).

MST and microbial forensics represent relatively recent specializations in the use of genetic and biochemical techniques for assessing the relative or specific importance of bacterial and viral sources in samples related to environmental, food safety, or criminal matters. Validated MST techniques are important elements of demonstrable and reproducible identification of the source or sources of environmental microbes and can assist in assessing their health significance. The field has reached a stage at which it is acquiring utility in the regulatory and legal arenas, and it will no doubt play a major role in future water quality litigation (96). Similar to other emerging disciplines, the science of MST is developing more rapidly and completely concerning fundamental knowledge and regulatory applications compared with its use and acceptance in legal forums. Rapidly emerging and powerful methods promise progress in food and product safety, criminology, medicine, environmental evaluation, and pollution management or regulation. Although no formal accreditation options or standardized protocols are in place for MST, researchers are working to define such procedures and data acceptance criteria to improve performance and application of MST (120). Rigorous validation procedures and verification steps remain essential to its continued contribution.

CASE STUDIES

Illinois River Watershed in Oklahoma

Although it is possible that MST evidence has been collected and relied upon by government agencies for administrative enforcement actions or enforcement actions that settle without litigation, available case law indicates that MST-based evidence has played little part in Clean Water Act enforcement to date. In Attorney General of Oklahoma v. Tyson Foods, filed in 2005, the State of Oklahoma attempted to enjoin Tyson Foods and hundreds of its large-scale poultry producers within the Illinois River Watershed (IRW) from discharging fecal contamination to the state’s waterways as a result of their stockpiling and application of poultry litter to agricultural fields (121).

During the trial, expert testimony was presented concerning results of a recently developed qPCR assay for the 16S rRNA gene in Brevibacterium spp., which served as a poultry litter contamination marker in the IRW (106). A correlation was demonstrated between the poultry fecal marker (LA35) and other fecal bacteria, as well as with phosphorus and heavy metals found in used poultry litter (i.e., feces, urine, straw and other absorbents). The LA35 methodology was not published in a peer-reviewed journal when it was considered by the court. Subsequent to the trial, however, several of the studies authored by plaintiff experts were published in peer-reviewed journals (7, 76, 83, 105).

The district court for the Northern District of Oklahoma was unwilling to rely on the state’s MST data under its Daubert analysis, finding the application of the technology to be insufficiently reliable and subsequently denying the state’s motion for preliminary injunction under the Resource Conservation and Recovery Act to enjoin application of the poultry litter (121). On appeal, the Tenth Circuit found no abuse of discretion in the district court’s Daubert analysis and upheld the court’s ruling to afford little weight to the MST evidence and related expert testimony presented by the state (121).

To meet the evidentiary burden for a preliminary injunction against the defendant’s land application and stockpiling of poultry waste under the Resource Conservation and Recovery Act, the state of Oklahoma was required to demonstrate that the poultry litter presented an “imminent and substantial endangerment to health or environment,” but without rising to a showing of “proof of actual harm to health or the environment” (121). The Tenth Circuit held that “Oklahoma failed to link land-applied poultry litter and the bacteria in the IRW, so it could not meet even this low hurdle” (121). The Tenth Circuit upheld the district court’s finding that “Oklahoma could not establish that poultry litter was a contributing cause of the bacteria at all” to the IRW (121). To this end, the court found that the state could not account for other sources of fecal contamination in the IRW and that the state’s expert witnesses could not reliably refute that these other sources could be significant contributing sources of fecal contamination to the waterways of the IRW (121). The court relied on the evidence presented by the defendants demonstrating that the storage and application of the poultry litter ensured that the fecal bacteria had perished and further determined that Oklahoma failed to present a “fate and transport study to establish that any surviving bacteria from poultry litter actually reached the waters of the IRW” (121).

After rejecting the fate and transport theory of fecal contamination presented by the plaintiff, the court considered whether to exclude or afford any significant weight to the MST evidence brought by the state. In doing so, the court applied the Daubert standard for admissibility of scientific evidence, which pursuant to Federal Rules of Evidence, rule 702, considers whether the evidence is both “relevant and reliable” (121). Under Daubert, admission of expert scientific testimony concerns “whether the reasoning or methodology underlying the testimony is scientifically valid and of whether that reasoning or methodology properly can be applied to the facts in issue” (121). To assess the reliability of the methodology, the court considers several factors: “[1] whether the theory or technique has been tested, [2] subjected to peer review, [3] and published, [4] as well as known rate of error” (119).

While the district court admitted the MST evidence pursuant to its Daubert analysis, denying Tyson’s motion to exclude, the court used the same Daubert factors to accord the MST study and the accompanying expert testimony very little, if any, evidentiary weight (121). Oklahoma argued that the Daubert standard should only concern the court’s assessment of the methodologies employed by the expert and the scientific evidence, not the application of those methodologies. However, the district court and Tenth Circuit rejected this analysis of the Daubert standard, finding that the expert’s testimony and conclusions were “insufficiently reliable,” because the expert’s research “had not been peer reviewed or published, and that no one outside this lawsuit… ha[d] either validated or sought to validate [the expert’s] [sic] scientific work” (121). Ultimately, the Tenth Circuit held, persuaded by the defendant’s argument, that Daubert permits a district court to “require further indications of reliability” where even “established methods are employed in new ways” (121). The bench trial (no jury) regarding permanent injunction against poultry litter application commenced in late in 2009. As of March 2017, no decision had been issued in the matter.

Chesapeake Bay in Maryland

The case of Waterkeeper Alliance, Inc. v. Hudson in 2012 illustrates a scenario in which MST would have provided a more rapid and definitive conclusion to the issue of the source of fecal contamination of the Pokomoke River, a tributary of the Chesapeake Bay. The environmental organization Waterkeeper Alliance filed suit against a poultry farm, which it contended was responsible for high levels of FIB (fecal coliform and E. coli) and nutrients downstream of the farm (122). Waterkeeper Alliance based its case on a mound observed by aerial surveillance, which they contended was poultry litter (used poultry house bedding); however, the mound was sewage sludge that was to be used as fertilizer. The plaintiffs offered a second theory that poultry waste was being tracked from the houses and/or blown by fans into the body of water. The defense was based on the theory that the herd of about 85 cows and calves kept on the farm were the source of FIB and nutrients (122) at 4-7. After extensive consideration of competing theories about the fate and transport of fecal contamination on the farm, the district court judge ruled for the defendants, finding that the plaintiff’s FIB evidence, taken together with minimal observations of actual physical discharges from the poultry facilities, could not substantiate their theory of fecal contamination to the receiving waterway (122) at 7.

By the time this litigation took place, the LA 35 MST method for poultry fecal contamination had been published in two peer-reviewed articles (105, 106). Several MST markers for cattle fecal contamination had also appeared in peer-reviewed articles (123–125). This case represents a missed opportunity for the use of MST in matters of environmental pollution by confined animal feeding operations.

As techniques of MST continue to become well-recognized and broadly validated, it is probable that they will be employed in future court proceedings (96). As concluded by Heinzen and Russ, “The science to support successful cases is advancing rapidly, and these cases are becoming increasingly viable as the law catches up with the state of the science” (126). The Oklahoma poultry case study demonstrates that MST methods employed in the legal arena should, at minimum, be published in a peer-reviewed journal and should preferably have been vetted by scientists without an interest in the case. The error rate of the methodology should also be understood, which necessitates testing the specificity (accuracy of a positive result), sensitivity (accuracy of a negative result), and limit of detection of the analytical procedure in the context of the water body or subject in question (56).

CHALLENGES AND FUTURE DIRECTIONS

The MST methodology that will most likely be first to gain traction in the regulatory and legal arena is the use of qPCR to quantify microbial genes that belong to host-associated bacteria or viruses, e.g., HF183, HumM2, human polyomaviruses, and adenoviruses. The bacterial targets tend to be host-associated (not entirely specific), while the viral targets tend to be more specific, but at lower concentrations in sewage, and therefore difficult to detect in environmental waters. A general recommendation is to use these markers in combination, coupling a more sensitive marker (found at higher concentrations in feces/sewage) with a more specific one that has a minimal percentage of false-positive results.

MST methods based on qPCR target a specific gene; therefore, at least one additional analytical method is needed for each fecal source of concern in the water body. Most water bodies are impacted by at least several possible sources of contamination, and costs increase as assays are added to the study plan. Microarrays are an emerging solution to this issue because they simultaneously target many genes and can be engineered to detect a broad array of microbial targets, including 16S rRNA genes for species identification, virulence genes, and antibiotic resistance genes (33, 127, 128). Another microarray strategy, the PhyloChip, targets microbial community 16S rRNA genes and shows promise for MST (35).

Another method for determining microbial community structure, high-throughput (also known as next-generation) DNA sequencing is also being used in MST applications. Compared to microarray methods, this strategy has the disadvantage of generally being unable to detect pathogens in environmental water samples, because they tend to be in low concentrations and are discounted as “noise” in the analysis if they are detected. Recent studies in Australia (36, 129, 130), France (131), and the United States (132) have employed next-generation sequencing and microbial community structure analysis to explore sources of contamination to environmental waters.

All of these nucleic acid-based methods share the challenge of effective concentration of a variety of microorganisms without concentrating inhibitory substances, purification of DNA without loss, and connection of identified fecal sources with human health risk. The microarray and community analysis, in particular, are not quantitative, and accurate source apportionment with all of the methods has not yet been achieved.

CONCLUSIONS

MST methods have proven their usefulness in watershed management. For example, when human-associated microorganisms are consistently detected, and particularly when corroboration with more than one marker is obtained, managers can confidently direct resources in the direction of improvements in sanitation. In contrast, when agricultural animal sources are among the dominant signals, the management of agricultural waste can be targeted to improve water quality. MST is a young science, and existing methodologies continue to be improved, while new ones are under development. As yet, no MST method is approved by a federal regulatory agency. Thus, while these methods will continue to be active areas for research, and useful tools for watershed management, their use in the legal arena is currently limited. Because the science is rapidly developing, one can expect to see these methods in formal legal applications in the near future.