Fecal Source Tracking, Fecal Typing

Bacterial Source Tracking (BST)

BST Methodologies

INTRODUCTION

The BST methods described below can be subdivided into three basic groups: Molecular, Biochemical, and Chemical. Molecular (genotype) are all referred to as "DNA fingerprinting" and are based on the unique genetic makeup of different strains, or subspecies, of fecal bacteria. Biochemical (phenotype) methods are based on an effect of an organism's genes that actively produce a biochemical substance. The type and quantity of these substances produced is what is actually measured. Chemical methods are based on finding chemical compounds that are associated with human wastewaters, and would be restricted to determining is sources of pollution were human or not. Of the methods below, 5 or 6 appear to have the greatest potential at present (see BST PUBLICATIONS). There is little doubt that variations of those presented below and new methods (to be added below) will appear over time.

The following parameters should be kept in mind when considering any one BST method.

At this time there is no evidence that any one BST method will emerge as the single best method for all situations. A "toolbox" approach seems warranted where the method chosen depends on the types of fecal sources that one has to deal with (or the questions one needs answered). Obtaining similar results with different BST methods also improves the chances that the source identifications are correct

As BST methodology development is so new, comparative tests between methods with the same collections of isolates has yet to be done on a sufficiently large scale. Only when such research is performed will the relative strengths and weaknesses of each BST method be understood.

To be most usefule to the scientific community, details of BST methods should be published in peer-reviewed journals and validated by successful use in other laboratories.

BST methods have been used primarily with E. coli and the fecal streps, and to a lesser degree with bifidobacteria, Bacteroides-Prevotella, and coliphages. While regulatory agencies often prefer E. coli (because that is the most widely used indicator standard), there are good reasons to consider other fecal organisms, depending on the situation. For example, E. coli is difficult to detect in composted poultry litters or in high quality biosolids, while the fecal streps are very numerous in both materials. Also, there is increasing evidence of regrowth of E. coli and fecal streps in tropical and/or subtropical waters and sediments. Under these conditions, other organisms such as the bifidobacteria, Bacteroides-Prevotella, or coliphages may be better organisms for source traking (and perhaps also as better indicators of fecal pollution).

With all BST methodologies (except chemical) it is necessary to first build a library or database of isolates taken from known sources (e.g. human, cow, deer, etc.). The size of the library will be partly determined by the number of potential major sources of fecal pollution in the target area. Nobody knows at this point how large a library has to be or how similar libraries from different geographical areas may be. Publications to date indicate that at least a few hundred isolates from each major known source will be needed. Once the known source library has been developed, and correct source identifications are sufficiently high for the desired purpose, then the second step can be taken of comparing fecal isolates of unknown origin against the library to obtain source identification. It is also necessary to regularly recheck the database with known source isolates to be sure that correct source identification remains high.

Lastly, to achieve the best source separation results from any BST method it will be necessary to become very familiar with some type of statistical analysis package. Database files tend to be quite large, so computer skills in database management and statistical analysis are essential. At this point, cluster analysis and discriminant analysis (both subroutines of ANOVA) are the most widely used, but are not the only options.

MOLECULAR BST METHODS (Genotype)

The fecal bacteria in any two animals (including humans) are very much genetically the same. There are unique differences, but the differences are only in a small percentage of an organism's total DNA. The key to molecular BST methods is finding these differences against a large background of similarity. It is thought that the distinctions between fecal bacteria from different animals (including humans) occur because the intestinal environments (selective pressures) are not the same, and fecal bacteria develop with detectable differences that can be related to sources.

All molecular techniques require highly trained personnel and there are many places in the procedures where mistakes can happen that could produce different band patterns. Automation and miniaturization should increase the sample volume and reduce per isolate costs over the next few years.

Molecular techniques can be divided between those that use Restriction Endonucleases (REs) and those that use Polymerase Chain Reaction (PCR) to amplify DNA. For both approaches, the DNA must first be carefully extracted, purified, and quantified.

Ribotyping

Ribotyping involves the bacterial genes that code for ribosomal RNA. Because such genes are highly conserved in microorganisms, ribotyping has been widely accepted for microbial identification. Ribotyping involves cutting the total genomic bacterial DNA with different DNAases, or restriction enzymes, followed by gel electrophoresis. Following electrophoresis, Southern blotting is performed to blot the DNA bands onto nylon membranes from the gels. DNA probes must be prepared for bacterial 16S and 23S rRNA and labeled with some type of detection sytem. Membrane hybridization is then performed to hybridize the probes with the appropriate DNA bands on the nylon filter. Difference in the size and location of the ribosomal RNA bands on the filters can then be used to differentiate between the sources that the fecal bacteria were obtained from.

Pulse Field Gel Electrophoresis (PFGE)

PFGE methodology (commercially developed by BioRad) involves a unique gel apparatus where electric current is passed through a gel in different directions at low voltage for 10-12 hrs to achieve the best level of band separation possible (Genepath apparatus). Another modification of PFGE involves embedding the bacterial DNA in a agarose plug. This eliminates many of the contamination concerns that can be very troublesome with molecular methods. The DNA is digested while in the plug, and the plugs are placed in hollow gel combs and become part of the gel as the gel is cast around the combs. Gels are stained and photographed after electrophoresis.

Randomly Amplified Polymorphic DNA (RAPD)

RAPD methodology involves identifying unique polymorphisms within the DNA of fecal bacteria. Arbitrary primers are used to identify randomly selected polymorphisms, and amplification occurs via polymerase chain reaction (PCR). Performing RAPDs first involves very careful DNA isolation, purification, and quantification; followed by addition or primers to the DNA, then amplification with PCR followed by gel electrophoresis. This method requires screening primers (there are over 1,200 commercially available) to find sets of polymorphisms that are either unique to fecal bacteria from a given source or occur in a given source to a large and predictable degree. Once such sets of polymorphisms have been found, fecal bacteria can be "sourced" by comparison.

BIOCHEMICAL BST METHODS (Phenotype)

"All higher order mechanisms are suspect until confirmed at the molecular level" (F. Crick. 1988, What Mad Pursuit, Harper Collins, NY). Most of the phenotypic methods described below offer certain advantages over molecular methods (above). The advantages include less required training for lab personnel, lower per isolate cost (in time and materials), and the potential to perform the methods on hundreds of isolates per week (a few dozen isolates per week is more typical for molecular methods). For sampling sites that are highly contaminated, it may be necessary to perform BST on several hundered isolates so that the results will be representative of the fecal population in the sample.

There is some controversy over the use of molecular vs. non-molecular methods. These should be viewed as complimentary methods rather than opposing methods. While it would appear that molecular methods are more precise, the published results to date indicate that at least one non-molecular method (ARA see below) is just as (if not more) precise in identifying sources. Referring to Francis Crick's admonishment, perhaps the correct approach is to validate a given non-molecular method with a molecular method (actually, this cross-validates both methods). As long as the two are in agreement there is no reason not to enjoy the advantages that some non-molecular techniques offer. This is why the "toolbox" approach mentioned above is so important. One can use the speed and large isolate capabilities of a non-molecular method and then regularly validate the results with a molecular method (on perhaps 5 to 10% of the total isolates). Obtaining the same answers with two different methods also increases the confidence in those answers.

Antibiotic Resistance Analysis (ARA)

This method uses fecal streptococcus (including the enterococci) and/or E. coli and patterns of antibiotic resistance for separation of sources. The premise is that human fecal bacteria will have the greatest resistance to antibiotics and that domestic and wildlife animal fecal bacteria will have significantly less resistance (but still different) to the battery of antibiotics and concentrations used. Most investigators are testing each isolate on 30 to 70+ antibiotic concentrations. Fecal bacteria are grown in wells in microtiter trays and then replica-plated onto a series of agar plates, each containing one specific antibiotic concentration. Forty-eight cultures can be transferred to each agar plate simultaneously with a stainless steel replicator. After incubation, each isolate is scored for growth or no growth on each plate and a resistance pattern emerges that can be used in source differentiation.

F-Specific (F+ or FRNA) Coliphage

Considered perhaps more indicative of viral contamination, the FRNA coliphages are pathogens of E. coli and infect the pilus of male E. coli strains. These coliphages can be differentiated using serology. There are four antigenically distinct serogroups of FRNA coliphages, and those predominating in humans (groups II and III) differ from those predominating in animals (groups I and IV). Hence, it may be possible to distinguish between human and animal wastes by serotyping FRNA coliphage isolates. However, there is a problem with separation between human serotypes and serotypes associated with pigs which also contain group II. Additionally not all animals have FRNA coliphage associated with their respective E. coli. The coliphage is persistent in the environment for less than a week and survival is a function of sunlight and water temperature. Ultraviolet light denatures the virus and below 25 degrees C F-pilus synthesis ceases. The coliphage does not replicate in the environment, only in the presence of F-pilus E. coli, and is not found in sediments, just in the water column. DNA fingerprinting of FRNA coliphages may be able to resolve some of the problems with serological typing.

Sterols or Fatty Acid Analysis

Sterols are constituents of the fatty acids in cell walls and membranes. This method is to differentiate between the types and quantities of sterols in human E. coli cell walls and membranes verses those in other animals. This method is currently under development and there are no published reports of its use in fecal sourcing. Access to Gas and/or HPLC chromatography equipment is required to perform fatty acid analysis. Fatty acids are first converted to fatty acid methyl esters (FAMEs) by chemical methods prior to perfoming gas chromatography.

Nutritional Patterns

This technique is based on differences among bacteria in their use of a wide range of carbon and nitrogen sources for energy and growth. This method works well in the laboratory. However, there are many environmental factors in a watershed that can affect bacterial nutrient requirements that may make this method impractical for field determination. The BIOLOG system involves allows the user to rapidly perform, score, and tabulate 96 carbon source utilization tests per isolate and is widely used in the medical field for microbial identification. Another modification of the nutritional pattern concept is the use of human-specific (sorbitol fementing) bifidiobacteria as indicators of non-point source human fecal pollution.

Fecal Bacteria Ratios

This procedure is based on the ratios (presence and numbers) of many different types of stomach and intestinal bacteria, not just fecal coliform bacteria, to develop a ratio coefficient that could be usefule in source identification. While the traditional fecal coliform-fecal streptococcus ratio is no longer considered reliable for accurate source identification, ratios may still be useful as a general indicator of human verses non-human fecal bacterial contamination, and the ratio concept could perhaps be found more reliable if other microbes were used in developing the ratios (e.g. Bacteroides, Prevotella, and Clostridium).

CHEMICAL BST METHODS

Chemical methods do not detect fecal bacteria. These methods are designed to detect chemical compounds that are associated with humans. These chemicals are often found in wastewaters such as septic tank effluent. Therefore, if the compound(s) are found in a water body then there is likely a human source. The follwoing two examples would only be useful indetermining if a source is human or not.

Optical Brighteners

This method detects the optical brighteners that are in all laundry detergents. They are persistent in the environment and are detected using low-tech black lights or mass spectroscopy. Sample collection is accomplished by placing optical brightener-free cotton in a wire mesh trap and placing the trap in the stream for a few days. After the trap is recovered the cotton is examined with a black light to see if it glows. The fluorescent cotton can then be examined with mass spectroscopy to verify the presence of the compounds. It is postulated that if these chemicals can be detected then there must be a human source. The problem with persistent chemicals is that they may not reflect recent pollution.

Caffeine Detection

This method is currently being developed to identify areas where a human source is suspected. Hopefully this will result in a low-cost test that can be used to detect caffeine in a water sample. Caffeine passes through the human digestive system and could be used as an indicator chemical. The major problems with this method now is that it is expensive running about $100 per sample. Also, there are some other plants that have significant levels of caffeine in (e.g. watermelon), and could confuse results. Lastly, caffeine is easily degraded by soil microbes, so it is not known what proportion of human sources actually contain detectable levels of caffeine.

CONCLUSIONS

BST technologies are quickly becoming proven and should be utilized by federal and state regulatory agencies to address sources of fecal bacterial pollution in water. Even though still experimental, BST methods represent the best tools available for determining TMDL load allocations, TMDL implementation plan development, and watershed restoration. These technologies can also be utilized to direct implementation of effective BMP's to remove impaired segments from state's 303d lists. The use of BST technologies can also help agencies display their commitment to water quality and conservation to the legislature and the citizens of their respective states and the EPA. Since the EPA will require reasonable assurances of TMDL implementation, BST should be an important methodology that states can utilize to meet this requirement.

There will probably not be any one BST method that will be the answer to every situation. The toolbox approach of using a biochemical and/or chemical method(s) to run large numbers of samples, and then confirm the results by assaying some subset of samples by a molecular procedure appears to be the most cost and time efficient use of these methodologies at present. Such an approach will have the benefit of more than one method producing the same or similar results. These results, when coupled with the output of computer models, will give considerable scientific justification for TMDL allocation scenarios and implementation efforts for any particular watershed contaminated by fecal bacteria.

| BST HOMEPAGE | BST OVERVIEW | BST METHODS

BST website maintained by Charles Hagedorn, Professor of Crop and Soil Environmental Sciences, Va Tech.