The Pollution Of The Mississippi Gulf Coast Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

The association of a health risk in coastal waters is determined by the enumeration of fecal coliforms and/or enterococci levels. While these methods have been useful and have protected the public from disease, the fact remains that indicator bacterial levels cannot be associated with a specific animal. There are other problems associated with these standard methods including: 1) the persistence of indicator organisms in waters and sediments; 2) the fact that fecal indicators remain alive in the presence of plant material; 3) the survival of indicator bacteria in beach sand; and 4) the possibility that indicator organisms can exist in areas with no human habitation. Recent investigations have determined that the problems are genuine and inherent in the use of the fecal coliform and enterococci as indicators; however, no suitable alternative to these standard methods has arisen which could serve as a verifiable replacement. Consequently, microbial source tracking has evolved as a way to delineate the possible sources of fecal input in surface waters and to complement viable microbial counts of known indicator species.

There are two forms of microbial source tracking: library-dependent and library-independent methods (Table 1). Library dependent methods isolate indicator organisms (coliforms, enterococci, human enteric viruses, bacteriophage, etc.) from a specific animal source, perform standard biochemical testing to identify the isolate, and carry out genetic fingerprinting on each organism. Several thousand fingerprints of different animal isolates constitute a known source library; unknown isolates are compared to the known library and possibly identified as being from a specific animal. Researchers have demonstrated that there is a significant amount of genetic heterogeneity between environmental and human fecal populations of Eschericia. coli, the use of different statistical procedures can produce conflicting results, and the library based method would require an immense number of sample isolates for it to be viable (Lasalde et al., 2005).

Library-independent methods (LIM) have developed with the goal of identifying animal-specific sources of fecal pollution using a single gene for identification. LIM methods are rapid, specific, simple, economical, and a variety of methods have been developed and tested for use with environmental samples and in a variety of national locations. Our lab developed and tested a wide variety of alternative Archeae molecular markers for the host-specific identification of animal fecal pollution in Mississippi coastal waters. The first use of methanogens as molecular markers was developed in our laboratory and included one sewage specific molecular marker, two domestic ruminant-specific markers, one chicken marker, and one swine marker of fecal pollution (Ufnar et al., 2006). These methods, as well as others listed in Table 2, are currently being examined for application by regulatory agencies as a supplement to the existing standard methods. Future testing may involve testing only for these alternative molecular markers or a combination of these methods and traditional microbial analysis.

Benefits and broad applications of LIM analysis include an improved understanding of the types of fecal pollution that enter the waters of the Northern Gulf of Mexico, enhanced identification of the sources of fecal contamination, and ultimately, better calculation of the risk of increased exposure of the public to human pathogenic microorganisms from contact with Gulf waters. Knowledge of contamination sources is crucial for mitigation and remediation of fecal contamination in coastal waters; thus, the technologies applied and developed by this work have broad application in polluted coastal waters throughout the United States.

This research will compare four, human LI methods and their significance in the determination of fecal pollution along Mississippi coastal beaches. Comparisons will encompass analyses at specific beach sites at which water samples will be collected, transported to the laboratory, filtered to isolate all microbial cells, and extracted to recover total DNA. Primers specific to human Bacteroides spp., Methaonbrevibacter smithii, Fecalibacterium, and Bacteroides thetaiotamicron were employed in formulated Polymerase Chain Reactions reactions which amplify known base pair signals representing the five human markers. Gel electrophoresis and/or MultiNA capillary electrophoretic analysis of PCR products were conducted to determine if the markers are present at particular sites along the coast. Statistical evaluation was conducted to establish possible correlations involving: the individual markers, the specific coastal sites, and the relationship between markers and indicator species, and certain environmental parameters.

Critical to this investigation is the inclusion of samples from freshwater streams that drain to the coastal environment and affect beach water quality. Specific sites within these streams were evaluated for the presence/absence of each marker; fecal coliform and enterococcal counts were also conducted on each stream sample.

Coastal water quality is a fundamental aspect of a vigorous Gulf of Mexico, sustaining the shell fishing industry and providing the basis for an extensive tourism industry. The safety of beaches and quality of life in the Gulf region are highly dependent upon successful stewardship of coastal waters, whose safety with respect to human health is threatened by extensive development and other anthropogenic activities. Water quality is routinely monitored by enumeration of indicator bacteria, which are generally nonpathogenic. These bacteria are associated with a wide variety of fecal inputs from humans and animals, and thus offer no information about the source(s) of pollution that can degrade water quality in coastal areas. This failure impedes the ability of regulatory agencies and managers to protect public health and remediate pollution sources. Microbial (bacterial) source tracking (MST) methods have been developed and tested over the past several decades, showing promise for discriminating between animal and human fecal pollution sources (Field et al. 2003; Meays et al. 2004; Rochelle and De Leon, 2006; Scott et al. 2002; Simpson et al. 2002).

In December 2006, a workshop entitled "Northern Gulf of Mexico Bacterial Source Tracking Workshop" was held in Biloxi, MS (Ellender et al., 2006). Workshop participants, consisting of researchers from around the country and researchers from federal laboratories and Gulf States representatives, concluded that the most promising technologies were: the human Bacteroides (HF8) marker, the M. smithii marker and the analysis of optical brighteners. Since that time, the testing of optical brighteners has diminished and three additional human markers, Human Polyomavirus, Bacteroides thetaiotaomicron and Faecalibacterium spp. have been developed. A summary of the human markers is presented below.

Methanobrevibacter: The genus Methanobrevibacter is a member of the order Methanobacteriales within the domain Archaea of the Kingdom Euryarchaeota (LeFever and Lewis, 2003). Species within this genus occupy very specific environments. They are found in intestinal tracts of warm blooded animals, anaerobic waste water treatment sludge, termite guts, oral cavities, and decaying plant material (Miller and Wolin, 1983; Lai et al, 2004, Miller and Lin, 2002; Gray et al, 2002; Cabiral et al, 2003; Horz and Conrads 2011; Brusa et al, 1993; and Belay et al, 1998). Research supports that they are the dominant methanogens in animal intestines (Lin & Miller, 1998; Lou et al, 2012).

Despite what is known about their distribution in aquatic and terrestrial environments (Miller 1984), animal intestinal tracts (Miller and Wolin., 1983), oral cavities (Belay et al., 1998), and waste water treatment sludge (Gray et al, 2002; Cabiral et al, 2003), little has been gleaned about methanogen presence and persistence in diverse environments (Ferris et al, 1996; van der Maarel et al, 1999; Lai et al, 2004).

Microbiome studies have concluded that the methanogens Methanobrevibacter smithii is only found to inhabit human intestinal and vaginal tract (Miller 1984; Belay et al, 1990). It's unique ability to persist in the complex human microbiome is attributed to the chemical mimicry of its outer surface to carbohydrate formations commonly found in the human (host) digestive track and its ability to regularly express adhesion-like proteins (Buck et al., 2007). M. smithii is very competitive of nutrient sources and is able to mitigate the end fermentation products of other host associated bacterial groups (Buck et al., 2007). Methane emissions from respiration studies have indicated that approximately 33% of the human population in the United States and Great Britain harbor methanogens; M. smithii is the most abundant methanogen in the human gut in order of 107- 1010 per gram (Bond et al, 1971; Lin and Miller, 1998). M. ruminantium is considered the dominant methanogen in the rumen of many animals (bovine, ovine, deer, goat, etc.) and is, therefore, a prime candidate for assessing ruminant-specific fecal pollution (Ufnar et al., 2006).

The ability to amplify the nifH gene of M. smithii from environmental and host samples makes it a good candidate for MST (Ufnar et al., 2006). Researchers have utilized the nifH gene to examine the gut microbial communities of host animals including the termite (Braun et al, 1999; Kirshtein et al, 1991; Ueda et al, 1995; and Widmer et al, 1999). The operon containing the nifH gene is conserved in methanogens and prokaryotes (Ufnar et al., 2006). The fidelity of the nonfunctional nitrogenase nifH gene as being a methanogen specific target has been thoroughly vetted (Okhuma et al, 1999; Raymond et al, 2004; and Ufnar et al, 2007).

In 2006, Ufnar et al. determined that the Methanobrevibacter smithii, assay was rapid, specific, less time consuming and inexpensive when compared to library dependent methods. An assay was developed targeting the nifH gene of M. smithii. This assay was tested against 27 various methanogens, 19 different bacterial species, 548 environmental bacteria, as well as DNA extracts from humans, sewage, cow, sheep, goat, dog, horse, deer, turkey, goose, and chicken feces to determine if the assay was specific for humans and sewage. M. smithii pure culture, human fecal DNA, and sewage were the only samples that tested positive with this assay. In addition, environmental samples collected during a MS coastal sewage spill confirmed the presence of this gene in contaminated waters, and water samples collected prior to the spill were negative for the gene.

Johnston et al. (2010) approached the detection of the nifH gene using a more quantitative real-time qPCR method. In this study the specificity of the primer sets (Mnif 202F 5'- GAA AGC GGA GGT CCT GAA-3' and Mnif 353R 5'- ACT GAA AAA CCT CCG CAA AC 3') were tested against 23 different species of methanogens, 11 of which contained in the genus Methanobrevibacter. The M. smithii target was detected in all environmental water samples that were spiked with sewage. According to Johnston et al. 2010, the M. smithii target sequence was also detected in two water samples spiked with bird guano. More importantly, the detection of the M. smithii target sequence in samples spiked with sewage did not correlate with the detection of culturable E. faecalis and E. coli. Other recent studies have further vindicated the use of this organism in the monitoring of environmental samples for the possible presence of fecal contamination. Rossario et al. (2009) tested the efficacy of using the M. smithii target for monitoring environmental samples in relationship to the detection of a pepper mild mottle virus. The M. smithii target was detected at six marine sites during the course of their study.

Bacteroidales: Bacteroidales are non-spore forming obligate anaerobes, and comprise a large portion of the human intestinal microbial flora. Species of the genus Bacteroides have the most antibiotic resistance mechanisms and the highest resistance rates of all anaerobic pathogens (Wexler, 2007). The use of this organism, as well as other obligate anaerobes, has been impeded by isolation and cultivation problems which are inherent with all conventional fecal anaerobe assays. Standard biochemical assays are being usurped by improved molecular techniques. To circumvent the inability of conventional assays to adequately ascertain the point source of fecal pollution, several labs are utilizing molecular techniques to elucidate the viability of host specific genetic markers in the environment. These molecular based approaches allow the scientific community to approach antiquated laboratory variables with a new found confidence in each experimental design. At the forefront were the molecular techniques for the isolation of Bacteroides spp. as viable fecal indicators, human Bacteroides and Bacteroides thetaiotamicron have emerged as likely candidates. Bacteroides spp. exhibited the characteristics of host specificity that is optimal for identifying the source of fecal contamination (Savichtcheva and Okabe, 2006). Bacteroides spp. existed in higher numbers in human than animal host as compared to the abundance of Enterococci and E. coli spp. (Converse et al. 2009). Kreader (1995) suggested that bacteria from the genus Bacteroides might be used to distinguish human from nonhuman sources of fecal pollution because: a) Bacteroides spp. dominate the human fecal flora, and several species outnumber the coliforms; and b) early experiments designed to better quantify the persistence of Bacteroides spp. in environmental waters supported the efficacy of this organism as a viable indicator. An experiment designed to inoculate Ohio River water samples with whole fecal samples for the PCR detection of B. distasonis indicated that temperature variances and predation were both critical in establishing conventional PCR detection limits (Kreader, 1998). Experiments by Okabe et al. (2007) indicated that there was little correlation between the presence of human specific Bacteroides and the culturable presence of total and fecal coliforms collected from freshwater river samples. This lack of correlation has been confirmed by other research groups working with the Bacteroides 16SrRNA target gene sequence of the human specific Bacteroides group. Human specific Bacteroides presence/absence was not directly correlated to any fecal indicator bacteria (FIB) abundance or interactions when assayed from marine samples (Santoro and Boehm, 2007). This study was of particular interest considering that 1/3 of the positive results for Bacteroides spp. occurred in an area where the confirmed fecal indicators were not of sewage origin; the sampling station farthest from the impacted tidal outlet in question had the highest occurrence human Bacteroides marker (Santoro and Boehm, 2007).

Bacteroides in general are valuable indicators because: 1. the bacterial load of human feces is on the order of 1012 per gram, and the predominant bacteria are of the genus Bacteroides (Zoetendal et al, 1998), outnumbering surrogate indicators such as E. coli and enterococcus species by orders of magnitude. 2. The Bacteroides genome has been sequenced, providing a basis for understanding the symbiotic role and microbial ecology of this microorganism, and enhancing the potential for development of host-specific molecular diagnostics (Xu et al. 2003; Kreader, 1995). 3. There is an established a protocol for detection of Bacteroides spp. that is uniquely associated with human or bovine fecal material (Bernhard and Field, 2000). Their method relies on direct detection of strain-specific 16S rRNA gene sequences. They used a double PCR amplification that employs a primary PCR reaction in which DNA from environmental samples provide the template followed by a second amplification in which a small amount of the primary PCR product serves as the template. This allows for the detection of Bacteroides target sequences in spite of the very low levels of the obligatory anaerobic bacteria present in the surface waters environment. 4. The poor survival of Bacteroides in environmental waters may be a desirable feature since Bacteroides growing or stabilized in the environment is highly unlikely. Thus, the resulting test has a low rate of false positives stemming from material other than recent contamination by/with fecal material. 5. Layton et al. (2006) developed bovine and human-specific primers suitable for qPCR that are highly specific for bovine and human Bacteroides. In the initial development of a library independent method, Bernhard and Field (2000) used 16S ribosomal RNA markers designed to distinguish human and cow fecal pollution, and to also quantify the effective recovery of these markers from natural waters. Further research identified host specific Bacteroides-Prevetella 16S rDNA markers from humans and cows by implementing DNA screening with restriction fragment length polymorphism (T-RFLP). Here, DNA from water samples procured from areas in Tillamook Bay, Oregon, were amplified using Bacteroides-Prevetella primers (Bac32F and Bac708R).

Dick et al. (2004) extended this research using a Taq nuclease assay (TNA) that employed a fluorogenic probe and primer set to determine the capture affinity for Bacteroides 16S rRNA in primary sewage influent. The efficacy of these host specific bacterial markers was further vindicated as it was shown that all possible primers sets should be experimentally exhausted. They employed the use of subtractive hybridization in microplate wells to identify host specific Bacteroides16S rRNA gene fragments and phylogenetic studies were employed to elucidate the endemism of Bacteroides spp. Thus, association of a specific Bacteroides spp. and an individual host would be paramount to its effectiveness as fecal contamination marker. Dick et al. (2004) also tested the 16S rRNA gene sequence analysis of Bacteroides from the feces of eight hosts: human, bovine, elk, pig, dog, cat, gull, and horse. The results revealed both endemic and cosmopolitan distributions of the organisms. Research on the phylogenetic host relationships of the Bacteroides-Prevetella group and their viability in the environment was questioned by Scott et al. (2002), since the persistence of this molecular marker in situ was yet to be fully scrutinized and since little was known concerning the survival and persistence of Bacteroides sp. in the environment.

Recently, the isolation of new Bacteroides sp. and the revisiting of the genomes of previously known species have yielded a plethora of novel possibilities (Robert et al., 2007). These novel species include B. plebeius, B. coprocola, B. helcogenes, B. intestinalis, B. finegoldii and B. doreii. In addition, a toxin produced by enterotoxigenic B. fragilis, which alters the morphology of human intestinal cells in vitro has been sequenced (Chung et al, 1999) and it may be possible to exploit a specific section of this gene when designing genetic probes. Though these species have an extremely high sequence similarity, they may offer unique gene sequences that could better delineate host specificity through stringent primer design and field application.

As noted above, a real time qPCR has been developed for Bacteroides sp. There are several advantages to using qPCR as opposed to conventional PCR including the absence of gel analysis, the ability to simultaneously identify and quantify specific genes, a reduction in the time of assay and the cost effectiveness of the assay (Layton et al., 2006). The detection limit of any marker is of inherent importance. Regardless of the particular assay being implemented, a standard detection limit for that marker must be established. It is fair to assume that from the quantitative data generated by a qPCR assay, there may be an efficient way to set the parameters for a standard detection limit for a marker. Recent experiments performed by Seurinick et al. (2005) attempted to quantify the detection of human specific Bacteroides16S rRNA genetic marker in fresh waters. More recent studies have tested the efficacy of using qPCR to monitor environmental waters for the presence of Bacteroides spp. (Converse et al., 2009; Shanks et al., 2009). The researchers found that these assays were efficient and reliable at targeting human specific Bacteroides spp. in environmental waters. They also discovered that the measured Bacteroides sp. found in sewage spiked samples often surpassed that of culturable Enterococcus sp. Therefore, qPCR would be an indispensable resource for assigning defined detection limits to the molecular detection of specific genetic markers (Converse et al., 2009).

Bacteroides thetaiotaomicron: In contrast to the Bacteroides marker discussed previously (Dick et al. 2005) and to which a variety of procedures has been ascribed to their role in microbial source tracking, B. thetaiotaomicron has not been exhaustively tested as a source marker of human pollution. This organism is present at a much higher percentage in humans than in nonhuman species, making it a strong candidate for MST (Carson et al., 2005). B. thetaiotaomicron became a candidate for a human specific marker when it was realized that it is a dominant species in the human gut and present in a much higher percentage of fecal samples (Carson et al., 2005). This study indicated that there are advantages in using the B. thetaiotaomicron primers instead of the Bernard and Field (2000) human primers since the B. thetaiotaomicron assay was sensitive, exhibited lower species overlap, required fewer PCR cycles, and appeared to be a more precise indicator of human fecal contamination.

This bacterium is known to possess a sizeable enzymatic profile that is of tremendous nutrient value to human metabolism (Xu et al. 2003). This organism is often associated with intra-abdominal sepsis and bacteremia and has been documented to be the second most often encountered disease causing anaerobic gram negative bacillus (Teng et al., 2000). More recently the etiological significance of B. thetaiotaomicron has been established by identifying it as the causative agent of a case of meningitis (Feuillet et al., 2005).

Faecalibacterium: Butyrate-producing bacteria play an important role in the maintenance of intestinal health. The taxonomy, structure and dynamics of these anaerobic bacteria have been extensively studied and documented because they comprise a large percentage of the biomass in the human gut and the fact that they could be exploited as potential MST markers. Recently, there has been research aimed at utilizing a Faecalibacterium spp, formally known as a Fusobacterium sp., as a novel MST marker. Using suppression subtractive hybridization (SSH), a new human specific bacterial gene marker, derived from a 16S rRNA gene region of Faecalibacterium, has been proposed as a potential MST marker (Zheng et al., 2008). Preliminary results indicate that this Faecalibacterium sp. is specific for human sewage, being found in 60.2% of human fecal samples and 100% of sewage samples tested (Zheng et al., 2008).

Human Polyomaviruses: The Polyomaviridae are a family of closed, double stranded DNA viruses that have the propensity to infect a wide range of vertebrates. They have an approximate genome of 5,000 base pairs (bp) and these data have been used to construct oncogenic models. Certain polyomaviruses are unique to humans, namely the JC and BK viruses. They appear to be widespread in the human population and are very host specific. Polyomaviruses of humans are acquired early in life and develop into chronic infections of the kidney persisting indefinitely (Shah, 1996); they are shed in urine and, therefore, are found in sewage (McQuaig et al. 2006).

Bofil-Mas et al. (2000) and Biofil-Mas and Girones (2001) showed that this virus was readily found in sewage, reflecting the potential value of these organisms as a measure of human fecal pollution (Hundesa et al. 2006). A PCR based assay for human polyomavirus was recently described (McQuaig et al. 2006) and has been tested in a validation study in Florida and Mississippi. Recent data suggests that primers specific for the JC and Bk viruses in humans have very little, if any, cross reactivity with bovine and porcine associated viruses (McQuaig et al. 2006). Significant titer volumes have been documented in city sewage waste (Bofil-Mas et al. 2000); the high specificity and sensitivity of these viruses make them appropriate candidates for MST.

A TaqMan based qPCR assay for the detection of polyomaviruses BK and JC in environmental samples was developed (McQuaig et al., 2009). This study concluded that there was a negative statistical correlation between HPyV and bacterial indicators in sewage. This disparity in the rate of decay for HPyV and bacterial indicators in sewage may be more indicative of their relationship in the marine environment.

Use of human and animal markers to detect aquatic pollution: The use of published human and/or animal markers to determine the presence of fecal waste in fresh or salt water bodies, including coastal waters, is a comparatively novel undertaking and has developed using a variety of technological methods (Soule et al. 2006; Shanks et al. 2009; Field and Samadpour 2007; Korajkic et al. 2010). The human Bacteroides marker has been exhaustively vetted in the environment through conventional and real-time qPCR assays (Bower et al. 2005; Layton et al. 2009; Ahmed et al. 2009; Dick et al. 2004; Kildare et al. 2007 Hong et al. 2008; Flood et al. 2011). However, much of this research has yielded conflicting results regarding marker/standard indicator correlations. In addition, much of the research conducted on marker persistence has taken place in the laboratory, greatly inhibiting accurate extrapolations regarding marker sensibilities to pervasive environmental factors.

Experimental designs eventually evolved to include testing nonhuman fecal sources for amplification of the human- specific and general Bacteroides-Prevotella markers (Fogarty et al. 2005; Kildare et al. 2007; Layton et al. 2006). The research conducted by Fogarty et al. (2005), elucidated cross-reactive amplification patterns in chicken and geese samples. Our research has also indicated that there is cross-reactivity between the human specific Bacteroides marker and chicken fecal samples. In addition, we have demonstrated that the marker can be amplified in domestic canine and feline fecal samples.

Persistence and decay of human specific indicators in the natural environment: Relationships affecting the ability of certain organisms to be good predictors of fecal pollution extended beyond their correlation to a host or each other. Once an indicator is exposed to the environment there are numerous biotic and abiotic variables that may affect survivability. Studies to determine which abiotic variables most greatly affect the persistence of detectable Bacteroides spp. target genes determined that salinity and temperature had a significant effect on their survivability (Seurinick et al., 2005; Okabe and Shimazu, 2007; Bell et al., 2008). As additional data has been generated by researchers involved in microbial source tracking, it has become evident that a multi-tiered approach and additional time points and physical variables should be considered when addressing water quality (Santaro and Boehm, 2007). Even when these variables are statistically applied to environmental and microcosm studies, the results are still confounded by geographical and laboratory design variations.

Recently, Balleste and Blanch, 2010, reported that Bacteroides thetaiotaomicron was less oxygen tolerant than Bacteroides fragilis, B. thetaiotaomicron was more thermotolerant in the summer months, and that environmental Bacteroides spp. exhibited a higher survivability rate. To understand the relationships of human specific fecal indicator bacteria it is paramount to design experiments that focus on elucidating specific correlations between these markers and environmental parameters. The data generated utilizing host specific Bacteriodales markers must be scrutinized within the environmental parameters of the sampling area, and they must be compared to previous studies examining their relationship to pathogens and traditional bacterial indicators (Walters et al., 2009). Walters et al. (2006) demonstrated that there is a differential survival rate of bacterial species belonging to the group Bacteriodales. Recently, Flood et al. (2011) showed that the presence of M. smithii and human Bacteroides markers were more prevalent in the coastal creeks that drained directly into the Mississippi Sound, and that these markers did not statistically correlate with the frequency of the markers found in the marine environment. This is indicative of the freshwater environment as a contributor of fecal pollution to the marine environment, but also indicates that the markers experience a differential survival pattern. It has been demonstrated that Bacteroides spp. recovered from sewage had a higher rate of decay than fecal coliforms or enterococci (Balleste and Blanch, 2010).

Questions concerning enterococcal persistence and importance in coastal waters: Conventional indicators of fecal pollution should presumably share common attributes with the pathogens they are proxies for. They should exhibit some correlation with the numbers of pathogens shed by the host, be nonpathogenic, easily assayed and enumerated, and share proportional survivability patterns; they should not persist and grow readily in extra intestinal environments (Scott et al., 2002). The United States Environmental Protection Agency (USEPA) has suggested the use of Escherichia coli and Enterococcus sp. as indicators of water quality for marine waters (USEPA, 2000). As research has progressed in the field of microbial source tracking it has become evident that differential survival rates and the innate ability of these organisms to proliferate and persist in the environment has called into question the efficacy of these organisms as appropriate indicators of fecal pollution. Researchers have attempted to quantify how these organisms react to the many variables encountered when they are introduced to the environment through controlled laboratory microcosm experiments (Lee et al., 2006; Hartz et al., 2008; Anderson et al., 2005). However, it would be more beneficial to strategically sample directly from the environment in a temporally compressed manner to better elucidate enterococcal and human specific marker trends. Marine water sampling strategies should focus primarily on the intertidal wash zone along the beach. These sampling constructs are important for many reasons: 1) the intertidal zone is an area where recreational bathing densities would be the highest; 2) beach sediment should inherently provide enteric bacteria with more nutrients and shelter than the water column; 3) bacterial levels should increase in areas of higher wave energy and higher concentrations of re-suspended particulates. These logical notions are recapitulated throughout the current MST literature. For instance, research conducted by Alm et al. (2003) investigating the efficacy of examining beach sand for higher enteric bacteria concentrations concluded that, compared to water, enterococcal counts were 4-38 times higher and E.coli counts were 3-17 times higher in freshwater sediment samples. The time reported for enterococci survival in sediment varies in the literature. Recent studies by Gast et al. (2011) indicated that enterococci survival may persist in deep sediment (25-70cm). The growth of enterococci in sand seems to be inherently related to the availability of organic matter (Lee et al., 2006). Haller et al. (2009) found that enterococci from sediment could be cultured up to 90 days after the initial sampling event. In addition to protecting the enterococci from predatory grazers, sediment shields the bacteria from prolonged exposure to UV radiation. Solar radiation is thought to be one of the primary factors in inhibiting bacterial populations, especially in shallow seawater (Sinton et al., 2002). The exact mechanism of photo-inactivation can vary for the particular bacteria in question and the environmental waters that the bacterium is recovered from. There is a knowledge gap as to how sunlight actually causes photodamage, either by direct UVB destruction of DNA or the increase in reactive oxygen species, to fecal indicator bacteria in the marine setting (Maraccini et al., 2011). A study by Shibata et al. (2004) spatial concentrations and prevalence of indicator organisms assayed were tied directly to the particular organism, sampling procedure used, and proximity to the beach. These sentiments reiterate the need for further experiments aimed at determining how these organisms react with their environment. More importantly, these notions serve as a warning to investigators when designing experiments and interpreting results.

Direct pathogen detection: A natural and logical progression in the field of microbial source tracking is to adopt methods to assay for the direct presence of pathogens rather than using traditional indicators. Prohibitive costs and intermittent shedding of pathogenic species hinders the implementation of these direct assays in regular environmental monitoring As technology becomes more readily available and costs are lower, researchers are beginning to field test the efficacy of utilizing these direct pathogen measurements. Though these assays still retain the inherent inability to ascertain viability or infectivity, they do represent a direct method for determining the presence of a particular viral, protozoan, or bacterial pathogen (Stewart et al., 2008). It would be more statistically and biologically relevant to elucidate correlation, if any, between current water quality standards (enterococci) and possible pathogens. Carr et al. (2008) demonstrated that the presence of detectable Salmonella spp. did not correlate with enterococci along Mississippi Gulf Coast sampling sites. This discrepancy in correlation between standard fecal indicators and possible pathogenic exposure, whether it is fungal, bacterial, viral, or protozoan in origin, is the current impetus for improved environmental monitoring strategies. Stewart et al. (2008) suggests a tiered approach, ranging from the initial testing of indicators to assays for individual pathogens. A tiered approach should incorporate the known relationships of bacterial indicators, human specific markers (e.g. M. smithii, B. thetaiotaomicron), and pathogens to each other and environmental variables that may influence their persistence. Epidemiological studies for the geographic area in question would ideally mirror the correlation values for the above biological variables and reported beach associated illness. Direct pathogen detection would alleviate much of the uncertainty associated with the current MST methods when assessing water quality.

Field studies indicated that there was a difference in marker recovery from different freshwater sources. This is most readily attributed to components in the water matrix that may inhibit downstream PCR amplification. The slight disparity in the membrane treatment experiment supports the idea that different waters affect recovery rates. Additionally, it elucidates that the mechanical treatment of the membrane at the beginning of the extraction process may affect the recovery of the marker. The most logical conclusion may be that the silica in the sand method may be binding to DNA during the extraction process. The comparison of both the M. smithii and Bacteroides markers in the USM creek samples showed that the markers are not found in similar proportions in lift station samples and was a clear indication that human fecal pollution ocurs in the Payne center creek. This is expected considering that M. smithii is shed in much smaller numbers than Bacteroides.

Examining the use of an internal PCR control: This method of spiking extracted DNA with a control exogenous DNA fragment as a PCR amplification control showed that there were no extraneous PCR inhibitors (i.e. humic acids) from the environment that inhibited the reaction. Samples were chosen at random for this analysis. The 5/2009-11/20089 sample set was amplified under the thermocycler conditions for each one of the four previously described human specific markers to eliminate it as a bias.

For the present, enterococcal measurements are the standard measure of human risk from contact with enteric pathogens in coastal waters; however, recent studies have indicated that there are many factors that mitigate the value of these analyses. For example, enterococci are known to exist in a variety of animals and on plants, and to reproduce in the coastal environment (Signoretto et al., 2004). Furthermore, sediments and beach sand have been shown to harbor enterococci and allow them to persist in the environment (Scott et al., 2002; Hartz et al., 2008). In partial response to the problems experienced by users of the enterococcal standard, researchers developed human and animal markers to identify sources of coastal pollution and allow remediation efforts to occur. The question is: In natural samples, are enterococci a reliable indicator of human fecal pollution, and do human markers correlate well with the levels of enterococci observed in coastal samples? For this geographical area, the answer is no.

This research project represented three distinct data sets collected over a period of three years. The delineations between data represent slightly different experimental designs and questions asked. The augmentations follow logical progressions from subsequently collected data. The sampling locations did not change during the course of these investigations except for the addition of sites CTH and CH in the last two years of the study. Because of the damage caused by hurricane Katrina (2005) there were times early in the study when some of the sites were inaccessible.

Environmental Coastal Samples 8/2007 - 4/2009: During this study, Enterococcus counts at 12 coastal sampling sites were not positively correlated. Unquestionably, creek waters contain substantial enterococcal levels and frequently showed the presence of the human markers; however, these measurements did not statistically translate into associated beach water counts of enterococci or the presence of the human markers. During the same period from August 2008 to April 2009, there were 131 enterococcal exceedances (Mississippi uses a single sample count of ≥104/100ml to designate a polluted beach) associated with the six coastal sites tested. Forty eight exceedances occurred at station 10, followed by 26 at site 10A, 22 at site 9, 17 at site 11, 13 at site 12A and 5 at site 7A. These data imply that a statistical correlation should occur at site 10 which is influenced by sampling sites CC1 and CC2, but it did not exist. Therefore it must indicate that other factors are at play to create this disparity. Differences do exist between the creek and the beach environments including such variables as fresh vs. salt water, the levels of ultraviolet light exposure, the dilution effect as creek water enters the estuary, and tidal transport at beach sites, as well as differences in turbidity and sediment disturbance. All or a portion of these factors could account for the lack of correlations observed (Ufnar et al., 2007).

Similarly, there was a significant difference between EN and FC counts and the presence or absence of the BA or MS marker in either the creek or coastal samples. This is not unexpected since one measurement is a quantifiable bacterial count and MS and BA are measures of presence or absence and represent other microbial genera.

The cross-tabulations indicated that higher percentages were recorded when neither of the markers was present in a sample. In fact, the BA and MS markers agreed more frequently than they disagreed. Differences between capillary electrophoresis and gel electrophoresis were negligible when neither of the markers was present, but agreement between the methods was higher (80%) when the BA marker was analyzed. The MultiNA and the gel procedure were in agreement when the two markers showed different results (61%). In general, these results demonstrate that either marker can be used to evaluate the presence of human coastal water pollution and that either method can be used to generate the results. The advantage of the automated MultiNA method was its sensitivity to small concentrations of DNA in a sample and its ability to evaluate a large number of samples in a short time period. Further, gel staining is not required, avoiding the use of ethidium bromide. The digital gel picture which the instrument presented was a very high resolution image; typically, bands appeared during a MultiNA analysis where none could be seen on an agarose gel. The capillary method has the added advantage of presenting data on the base pair units for each band and the amount (ng/ul) of each DNA fragment in the sample. The instrument requires careful management during the analysis of environmental samples and chip cleaning is often necessary and time consuming. However, if the objective of analysis is to process numerous samples in an abbreviated timeframe, requiring minimal operator attention and inexpensive results, the capillary electrophoresis method would be an appropriate technology.

Despite the fact that these data were part of a local sample population, the conclusion that EN levels did not correlate from sampling site to sampling site nor was their correlation with the levels of two human markers is troubling. Marker analysis has been persistently studied by a variety of international researchers for at least a decade and was considered a complement to enterococcal analysis. However, the random nature of the isolation of both the BA and the MS markers points to the fact that marker presence can be influenced by such factors as dilution, the salt water environment, tidal movements, the presence of sediment in the water column, resiliency to degradation, or other coastal features. This randomness suggests that the analysis of human markers and their relationship to the variable EN count cannot be used to identify and control pollution on coastal beaches.

In the future, a substitute for the measurement of indicator bacterial levels in coastal waters may be a dependable detection of specific microbial pathogens. Several viral pathogens are currently able to be detected by qPCR (McQuaig et al., 2006) and other bacterial and protozoal pathogens can be detected with molecular methods. For the time being, the use of the enterococcal count or the qPRC analysis of the level of this organism in coastal waters will continue, almost certainly in concert with data on one or more of the human markers.

Although enterococcal measurements are the current measure of human risk from contact with enteric pathogens in coastal waters, recent studies have indicated that there are factors that mitigate the value of these analyses. For example, enterococci are known to exist in many animal species, and to reproduce in the coastal environment. Furthermore, sediments and beach sand have been shown to harbor enterococci and allow them to persist in the environment.

Environmental Coastal Samples 5/2009 - 11/2009: During this study there were 14 coastal sampling sites analyzed for the presence of 4 human specific markers as well as enterococci and fecal coliforms. Of these, positive correlations were found between human specific Bacteroides and enterococci, B. thetaiotaomicron and fecal coliforms, and Fecalibacterium and enterococci. B. thetaiotaomicron was found to be positive most often at 32% of the time. Of all the positives for all four organisms, 15% percent of those were found in the creek CC2, followed by 12% at CC1. This creek system is indicative of an area with a large number of anthropogenic inputs. It is also an area that is in close proximity of a sewage lift station. Except for 7A, which receives its effluent from another highly polluted creek system (7ACC & 7ACT), the marine sampling sites had a much lower percentage of positives (<9%) for the 4 organisms that were assayed. From all the statistical analyses performed there does seem to be correlations between 3 of the organisms and standard indicator bacteria. Further analyses were needed to further elucidate this possible relationship.

All sample sites and sub-groupings are the same as designated in the above results section for 5/12/2009-11/19/2009. Because the collected data violated the rules for normality, non-parametric tests were used. The Kruskal-Wallis ranking of the three water types, FWS>NSB>NSM and FWS>NSB≥NSM for EN/100mL and FC/100mL, respectively, was geographically intuitive. If sites were to be significantly different based on indicator organisms measured, this difference, as shown in the results, would be directly tied to presumed bacterial loading based on location. The exact differences from the Mann-Whitney test of indicator bacteria measured from each water type further supported this supposition. This test grouped the water types by two and tested for significant differences between En/100mL and FC/100mL. The recovery of these two bacterial indicators differed significantly for the paired grouping of water types except for NSM and NSB where there was no significant difference for FC/100mL. For En/100mL the relationship again indicated that the major enterococci source of input originated from the freshwater creeks and directly influenced the counts at the terminal effluent sampling points in the NSB. Although it can't be directly proven from this experimental design, the assumption is that the dilution factor of reaching the marine environment was the cause of the NSM variable being ranked lower. It is possible that enterococci are being harbored in sand and eventually re-suspended into the water column by tidal and wave action. Because it was not directly measured, it is unclear if the large data set and robustness of the statistics were powerful enough to dwarf this conflicting variable. The ranking of the three water types based on FC/100mL yielded a slightly different response; unquestionably, the major source input was the freshwater creeks. However, the NSB environment was only ranked slightly higher than the NSM. One explanation for this is that the fecal coliforms do not share the exact same fate between these environments. The results of the Spearman's rho indicated that the highest correlation between En/100mL and FC/100mL was found in the commingling environment of NSB, indicating that as one increased so did the other. Within a geographical context this sampling point represents a confluence of all possible bacterial loads, point and non-point sources, and complicating environmental variables. Proximity to possible fecal pollution inputs and variable temperature, turbidity, uv exposure, bacterial re-suspension, and dilution factors were all normalized at this point. The correlation of these two bacterial indicators could be explained by this global aggregation and mixing of variables. It is interesting to note, however, that the lowest correlation was obtained from the freshwater environment. This could be attributed to the differential input or survivability of these two bacterial indicators in this environmental sampling area. Salinity, turbidity, temperature, and UV exposure were the environmental variables applied to the statistical tests. En/100mL and FC/100mL were both inversely correlated with salinity. The effect of bacterial counts decreasing as salinity values increased could be attributed to their inabilities to mitigate the effects of osmotic pressures. In addition, increases in salinity for these sampling areas were also directly tied to increases in UV exposure and dilutions of nutrient availabilities and bacterial indicator communities in the water column. In an attempt to predict the distribution of the four proposed human specific markers based on salinity, a Spearman's test was run. The marker distribution was ranked by highest negative correlation and significance values in order of B. thetaiotaomicron, human Bacteroides, M. smithii, and Fecalibacterium. This further supported the data that as salinity increased in ppt, bacterial markers were less abundant. B. thetaiotaomicron did have the most significant correlation to salinity at 23%, but this physical variable could not account for the other 77% affecting the organism's presence or absence. The effects of temperature on marker presence showed B. thetaiotaomicron and Fecalibacterium were both positively correlated. M. smithii and human Bacteroides were both negatively correlated. As temperature increased at the study site, Fecalibacterium abundance increased slightly but not significantly. The correlation for all the markers was minimal and yielded essentially uninformative results. The relationship of En/100mL and FC/100mL was weak. The most significant relationship was from the En/100mL data set and it was inversely correlated. This weak correlation could be a result of the precision of the measurements or an artifact of the data and not a true correlation. These results were not shocking as temperature variations remained somewhat consistent during the sampling months. Multiple regression analyses, used to predict the presence of these markers based on current bacterial indicators (En/100mL and FC/100mL), yielded conflicting results. 7% of CFU's for En/100mL can be predicted by the presence of B. thetaiotaomicron. However, there was no significant correlation for FC/100ml, indicating that none of the proposed human specific markers was predicting its presence in any of the coastal environments. It should be noted that if any of the markers were chosen it would have to B. thetaiotaomicron based on its Anova p-Value of 0.032.

A Spearman's correlation was run to test the influence of turbidity on En/100mL, FC/100mL, and all four human specific markers. M. smithii and Fecalibacterium were the only organisms that did not have a significant correlation to turbidity. All other bacterial groups had a significantly low inverse correlation with turbidity, with B. thetaiotaomicron having a low-moderate correlation and En/100mL having a moderate correlation. Even though these correlations are statistically relevant, the calculations were performed in spite of very apparent disrupting outliers. Removing these outliers from the equation yielded essentially a moot turbidity affect. This was surprising considering that a more turbid environment would have provided UV shielding and possibly more abundant nutrient sources. This event could have been related to the unmeasured variable of bacterial attachment and sedimentation. The effects of UV exposure on the bacteria assayed were also variable. Both En/100mL and FC/100mL had a moderate negative correlation to UV exposure, with FC/100mL being slightly higher. All four human specific markers had a negative correlation to UV exposure but the B. thetaiotaomicron marker was the only one to be significant and had the highest correlation. Considering that the UV measurements were collected in the field and were not continuously recorded on a data logger, the variability could be considered a grab sample of the total penetrating radiation for the sample site.

A multiple regression analysis was used to determine if any of the human specific markers were recovered predominately from any one water source and a clear pattern emerged as the markers were recovered in greatest numbers in order of FWS, NSB, and NSM. B. thetaiotaomicron had the highest recovery rate and was directly correlated with the freshwater environment. This pattern was consistent with the rank order of the recovery rates for enterococci and fecal coliforms from different water types. From these data it was concluded that the human specific marker of B. thetaiotaomicron performed best for describing areas that seemed contaminated with fecal pollution. Of the environmental variables tested salinity emerged as the most robust factor influencing the presence or absence of either the bacterial indicators (En/100mL and FC/100mL) or the human specific markers. The primary sources of these bacteria was the freshwater creek sources that spill into the sound (Flood et al. 2011).

Environmental Coastal Samples 4/2010-7/2010: The data followed the same sub-groupings as stated above. Bacterial counts for this portion of the study followed a pattern similar to the above section. Bacterial recovery rates for EN/100mL and FC/100mL were significantly different among the sampling areas and were rank ordered by FWS, NSB, and NSM. For the two bacterial indicator variables there was a difference in recovery rates between freshwater and marine sample sites. The enterococci were recovered in significantly different values between the NSB and FWS sites but fecal coliforms were not. Among the three water sources the highest correlation between these two bacterial indicators was found in the NSB environment. This mirrors the sentiment of this sampling area being a terminal site for all converging variables. Both En/100mL and FC/100mL were inversely impacted by increases in salinity with enterococci having the highest correlation. Between the human specific human markers, B. thetaiotaomicron showed the highest inverse relationship and the most sensitivity to increases in salinity and temperature. En/100mL had a very high inverse correlation to salinity and a low to moderate correlation to temperature. FC/100mL yielded a moderate inverse relationship to both salinity and temperature. Both bacterial indicators had strong inverse relationships to turbidity when data outliers were calculated and graphed. The outliers were calculated due to insufficient reasons to remove them, i.e., there were no transcription errors from written to digital data sets. When these outlier data points were removed from the calculations the relationships became moot. This relationship has proven to be enigmatic when viewed in a purely biologically relevant context and probably needs further testing to draw any real concrete conclusions.

B. thetaiotaomicron was significantly associated with En/100mL and FC/100mL bacterial counts but, based on their magnitude of effect, still failed as a good predictor for these variables. Again, all three human specific markers were recovered more frequently from the freshwater environment. This is further supported by the ranking of recovery rates for the bacterial indicators, the ranking of recovery rates for each marker, and the inverse correlations for each marker with salinity. It is still unclear if the inverse correlations with salinity were due to the organism's ability to mitigate changes in osmotic pressure or if salinity is acting as a proxy for dilution within the sampling area. Salinity was almost never above zero for the FWS sites and remained consistent at 26-33ppt for the NSM and NSB sites. A multiple regression showed that B. thetaiotaomicron, the most prevalent marker, was recovered at a significant rate from the FWS, indicating that the creek systems are responsible for the majority of human fecal input into the study sites. The order of marker recovery (FWS>NSB>NSM) followed the same ranking order of the recovery rates for EN/100mL and FC/100mL. These analyses clearly supported the data supported by the prior two studies. There was a prominent spatial trend for the presence of both the bacterial indicators and the human specific markers, thus, the geographical structure of a study site could be a valuable model parameter when trying to ascertain direct sources of input, probability of host source input, and proper sampling/remediation strategies.

For the three environmental studies described above (8/2007-7/2010) the results are consistent. The probability of recovering either a high bacterial count (CFU/100mL) or a human specific fecal marker can be directly tied to the sampling location and its respective water type. The recovery rate of these biological variables does not appear to be dependent on the presence or absence of one another. Recent MST research supports the opinion that using one bacterial genus to describe the probability of another or the presence of pathogens is flawed. Bacterial communities are in constant flux in the environment; a flux that is directly tied to their host origin, spatial and temporal moments, ability to mitigate detrimental abiotic factors, nutrient requirements, selective predation, and genetic heterogeneity. Our understanding of how these bacteria are able to meliorate environmental stressors (UV damage, osmotic pressures, and temperature) is expanding with studies similar to this one. Carotenoid pigmentation may mitigate the effects of photo-damage by Reactive Oxygen Species (Maraccini et al., 2011).

Regression analyses for human specific markers of fecal pollution and indicator bacteria from environmental samples have demonstrated that relationships can be significant, but have low correlations, for example, human specific Bacteroides marker HF183 were present at low concentrations of indicators (Bonkosky et al., 2009).

Gram (+)(EN) and Gram (-)(FC) bacteria respond differently to predation, osmotic pressure, and photo-inactivation (Solecki et al., 2011). The differential survival characteristics of both indicator bacteria and human specific markers do much to confound a researcher's ability to extrapolate valuable data for hypothesis testing. Under the most utopian settings, with variables scaling at rates which reflect their biological characteristics, choosing the wrong metric for data analysis can further distort experimental results. This is especially true for library-dependent mechanisms of MST. PFGE analysis of E. coli isolates revealed a 27% inconsistency between discriminate analysis and jackknife classification matrices (Lasalde et al., 2005). All of these weaken the efficacy of using a single indicator species collected from a single grab sample to ascertain the probability of detecting a possible pathogen in environmental waters.

Distribution of Markers in Nonhuman Hosts: The efficacy of using any marker of fecal pollution was directly related to host specificity. This study was aimed at determining the probability of amplifying the human specific Bacteroides and M. smithii markers from non-human hosts. The amplification of the Bacteroides marker from the canine samples diminished it validity as a truly specific indicator of human fecal pollution. This is not entirely surprising when considering the close relationship and co-evolution of this animal with humans. The same was true for the feline samples. Conversely, the M. smithii marker was not amplified in any of the canine or feline samples. This further confirms it as a human specific marker for fecal pollution.

The chicken samples analyzed represent an extremely small data set, and as such, contributed a less statistically powerful conclusion. However, the fact remains that the Bacteroides marker was found in 100% of the samples tested. This result appears to hinder this marker's usefulness in host specific MST. As research progresses in this field it is becoming increasingly evident that the human specific Bacteroides marker can be found in hosts other than humans.


This research encompassed 67 sampling trips and the collection and processing of 819 samples. It was our intention to test the efficacy of using published human specific markers to identify areas of fecal pollution along the Mississippi Gulf Coast. And, we endeavored to elucidate relationships of these markers with the current bacterial indicators of water quality and the physical variables that may have affected their presence or absence.

The experimental constructs described indicate, in our view, a logical progression of design and analyses, a progression that went from casting a wide net over many seasons to examining one swimming season with a temporally compressed multi-tiered approach. It is of particular interest that all of these studies resonate the same themes; the current standards for measuring bacterial water quality are failing and the major contributors of fecal bacteria were that of freshwater sources. This research, and that of others, found the same statistical discrepancies between correlations of standard indicator bacteria (enterococci and fecal coliforms) and human specific molecular markers in the environment.

Undeniably, the coastal creek systems tested during this study demonstrated a strong influence of fresh water effluents on the presence or absence of these human specific markers. In addition, there appears to be other variables influencing the ability to assay for the presence of each marker in the marine setting. The most simplistic explanation would simply be a dilution effect on these markers when they reach the marine environment. This research indicated that salinity and temperature were the two main variables influencing both bacterial counts and marker presence. Another explanation could be differential survivorship of these markers/organisms in the natural environment. This concept is currently being vetted within the source tracking community.

Future projects will include performing multivariate analyses incorporating other environmental variables (rainfall, wind direction, tidal/wind action, salinity gradients, and solar exposure) measured at each sampling site in a more temporally and spatially compressed manner. Future investigations aimed at determining the relative differences in transport and fate of each of the indicator organisms, as well as the compliment of human markers, in both the marine and freshwater environments should shed new light on the value of microbial source tracking and its use in the marine environment.