SEPA
                    EPA/600/R-07/123 I December 2007 | vwvw.epa.gov/athens
  United States
  Environmental Protection
  Agency
  Evaluation of Selected DMA-
 based Technology in Impaired
Watersheds Impacted by Fecal
  Contamination from Diverse
             Sources
Ecosystems Research Division, Athens, GA 30605
National Exposure Research Laboratory
Office of Research and Development

-------
                                            EPA/600/R-07/123
                                              December 2007
   Evaluation of Selected DMA-based Technology in
Impaired Watersheds Impacted by Fecal Contamination
                 from Diverse Sources
                            by
                       Marirosa Molina
                 Ecosystems Research Division
              National Exposure Research Laboratory
                      Athens, GA 30605
               U.S. Environmental Protection Agency
               Office of Research and Development
                    Washinton, DC 20460

-------
Notice

The information in this document has been funded by the United States
Environmental Protection Agency.  It has been subjected to the Agency's peer
and administrative review, and it has been approved for publication as an EPA
document.  Mention of trade names of commercial products does not constitute
endorsement or recommendation for use.

Abstract

      Fecal pollution of surface waters is a top reason for impairment, as
reported in the U.S. Environmental Protection Agency's report on the quality of
the Nation's waters. To be able to develop and implement TMDLs for impaired
aquatic resources, it is imperative to determine the sources of the contamination.
One tool used to determine the sources of bacterial fecal contamination is to
apply a microbial source tracking approach to the system of interest.  Microbial
source tracking (MST) approaches are based on the assumption that specific
strains of bacteria, genetic fingerprints, or DNA-based markers are associated
with specific host species. Because accurate source identification of fecal
contamination is essential in MST, more sensitive, selective and reliable
molecular markers are required. The two types of genotypic methods that have
been applied widely in a variety of environments can be classified as library-
independent (LI) and library-dependent (LD). For both types, the temporal and
spatial stability of selected genotypes are aspects that need to be evaluated, and
these aspects are often times missing when applying MST to environmental
samples. LD-MST methods require the development of large  databases
comprised of source-specific isolates.  Once a source-specific fingerprint has
been identified, the temporal and spatial variability of that particular genotype  still
needs to be validated.  LI-MST is based on the application of culture-independent
methods such as amplification of DNAfrom environmental samples using 16S
rDNA markers in combination with polymerase chain reaction (PCR). However,
cross-reactivity of some of the 16S rDNA markers used in field studies has
prompted the development of alternative PCR assays  using metagenomic
markers specific for bovine feces.  In this study, we report on the comparison of
selected LD and LI methodologies, their usability as rapid and reliable methods
for developing and applying markers to various environmental scenarios, and  the
stability of these markers under various spatial and temporal conditions.   From
our results,  we concluded that library production is highly time and resource
consuming.  Its application is probably appropriate in very specific scenarios
where discrimination among a few, selective sources is necessary. In contrast,
application of DNA, PCR-based markers yielded fairly rapid results and has the
capability to screen multiple scenarios in a short period of time. Once stability
and cross-amplification aspects have been addressed, this latter  method can be
a highly efficacious approach to determine sources of contamination in a variety
of scenarios.

-------
 Table of Contents


ENVIRONMENTAL ISSUE	6

RESEARCH GOALS	8

DESCRIPTION OF METHODS USED IN THIS RESEARCH	8

  AMPLIFIED FRAGMENT LENGTH POLYMORPHISM	 8
  Box-PCR ANALYSIS	10
  HOST SPECIFIC 16S-RDNA MARKERS	11
  METAGENOMIC MARKERS	12

GENERAL RESEARCH APPROACH	13

METHODOLOGY	15

  SAMPLING LOCATIONS FOR16S-RDNAAND METAGENOMIC MARKERS	15
  SAMPLE COLLECTION	18
  PHYSICO-CHEMICAL AND MICROBIOLOGICAL METHODS	18
  DMA EXTRACTION AND PCR AMPLIFICATION	19
  SOURCES FOR AFLP AND BOX-PCR ANALYSIS	21
  AFLP AND BOX-PCR PROCEDURES	21

RESULTS AND DISCUSSION	22

  EVALUATION OF LIBRARY-INDEPENDENT METHODS	22
    Comparison of 16S rDNA-based vs. metagenomic marker performance in farms impacted by cattle
    contamination	22
    Relationship between enterococci enumeration and the occurrence of molecular
    markers	24
    Evaluation of human specific 16S-rDNA markers in freshwater streams impacted by rural non-
    point sources in Puerto Rico	28
  EVALUATION OF LIBRARY-DEPENDENT METHODS	34
    Seasonal Distribution of Enterococci Isolates	34
    Some methodological considerations developing the phytogeny of Enterococcus strains using
    AFLP	35
    Phytogeny of E. faecalis, E. hirae ,andE. casseliflavus strains using AFLP.	36
    Comparison of AFLP and BOX-PCR analysis	39

CONCLUSIONS	41

SIGNIFICANCE OF RESEARCH	43

FUTURE DIRECTIONS	44

REFERENCES	45

-------
List of Figures

Figure 1 Main steps of the AFLP Procedure	10

Figure 2  Experimental scheme to perform 16S rDNA and metagenomic marker
analyses	14

Figure 3 General experimental design to isolate and fingerprint enterococcal
species	15

Figure 4 Four sites were sampled  in Farm 1, three sites were located along the
stream while site 4 was located in  a pond used by the cattle for bathing and
drinking.  Aerial photo courtesy of  GlobeXplorer.com	16

Figure 5 Sampling sites related to  WS2. Sites 1-7 are located within the farm
boundary, 8 and 9 are located in a buffer zone between the farm and a
subdivision, and 10-12 are located within a subdivision.  Aerial photo courtesy of
GlobeXplorer. com	17

Figure 6 Relationship between the monthly enterococcal counts and the average
frequency of the DMA markers per month in WS1	27

Figure 7 Relationship between monthly enterococcal counts and the average
frequency of the DMA markers per month in WS2	28

Figure 8 Seasonal distribution of enterococcal species in impacted streams by
cattle contamination	34

Figure 9 Enterococcus hirae phylogenetic tree derived from AFLP fingerprints .38

Figure 10 Typical BOX-PCR gel image produced with E. faecalis isolates	40

-------
List of Tables

Table 1 Frequency (±95% Cl) of 16S rDNA-based Bacteroides and metagenomic
markers in water samples from two watersheds affected by cattle contamination.
Watershed 1 (WS1) receives direct impact from cattle, while watershed 2 (WS2)
only receives contamination through runoff.  Only markers with a frequency
between 0.10 and 0.90 were used for the logistic regression analysis	24

Table 2 Enterococcal abundance {CFU/100ml) in Watershed 1 and Watershed 2.
Sites were divided based on influence by cattle or type of water resource
(streams vs.  ponds)	26

Table 3 Description of samples collected in the Rio Anasco Basin, Anasco,
Puerto Rico from August 3-14, 2006	29

Table 4 Bacteroidetes 16S rRNA gene marker hits in water samples collected in
the Rio Anasco Basin, Anasco, Puerto Rico.  The numbers indicate the times the
individual primer set was found in each water sample after one amplification
round (1x)	32

Table 5  Comparison of the advantages and disadvantages of the BOX-PCR and
AFLP methodologies	41

-------
Environmental Issue



      The U.S.EPA TMDL 303(d) list fact sheet has indicated that fecal pollution



is the #1 cause of impairment in most states, accounting for up to 13% of all



reported impairments. Cost-effective development and implementation of



TMDLs for impaired aquatic resources requires the rapid and accurate



determination of the sources of contamination.  Commonly used microbial water



quality assessment methods measure densities of fecal indicator bacteria, but do



not provide information on the possible sources of contamination producing the



elevated indicator concentration.



      One tool used to determine the sources of bacterial fecal contamination is



to apply a microbial source tracking approach to the system of interest.  Microbial



source tracking (MST) is based on the assumption that specific strains of



bacteria, genetic fingerprints, or DMA-based markers are associated with specific



host species. Because accurate source identification of fecal contamination is



the objective of MST, more sensitive, selective and reliable molecular markers



are required. The two types of genotypic methods that have been applied widely



in a variety of environments can be classified as library-independent (LI) and



library-dependent (LD).  For both types, the temporal and spatial stability of the



selected genotypes are aspects that need to be evaluated, and those aspects



are often times not well characterize when  applying  MST to environmental



samples. LD-MST methods require the development of large databases



comprised of source-specific isolates (Ritter et a/., 2003; Wiggins et al., 2003).



Once a source-specific fingerprint has been identified, the temporal and spatial

-------
variability of that particular genotype still needs to be characterized. LI-MST is



based on the application of culture-independent methods such as amplification of



DMA from environmental samples using polymerase chain reaction (PCR).  One



of the genes that has been widely used for this application is the gene coding for



the 16S rRNA, that has been demonstrated to have host specificity (Bernhard &



Field, 2000; Layton et al., 2006). However, one drawback of this technique is a



degree of cross-reactivity observed with some of the 16S rDNA markers when



used in field studies because they target highly conserved 16S regions (Shanks



et al., 2006; Shanks et al., 2007). This cross-reactivity prompted the development



of alternative PCR assays using metagenomic markers specific for bovine feces



(Shanks et al., 2006b).  These bovine metagenomic markers were successfully



tested in the latter study with little cross reactivity on a large number of bovine



feces collected from a variety of locations  across the U.S.  However, although



some spatial variability was covered in that study, a more detailed  evaluation of



the temporal and spatial variability of the markers was still required to determine



their environmental stability and robustness.



      This  research supports the second  long-term goal (LTG 2) established in



ORD's Water Quality Multiyear Plan for the protection of watersheds and aquatic



communities:  "provide the tools to assess  and diagnose impairment in aquatic



systems and the sources of the associated stressors".

-------
Research Goals



      The overall objective of this research was to evaluate the temporal and



spatial applicability of DMA-based techniques and markers to identify sources of



fecal contamination in a variety of environmental scenarios.








Description of Methods Used in this Research



      The methods evaluated were divided between library-dependent (LD) and



library-independent (LI) approaches.  The library dependent methods were used



with a library of enterococci markers isolated from cattle farms (Molina et a/.,



2007). The two LD methods included amplified fragment length polymorphism



(AFLP) and repetitive fragment polymerase chain reaction (PCR) with Box-PCR



primers (Box-PCR).  Two LI-PCR methods were also compared: 16S rDNA-



based Bacteroidales markers and metagenomic markers, both specific to cattle.



      Amplified fragment length polymorphism.  AFLP consists of selective



amplification of restriction fragments resulting from the digestion of total genomic



DMA using PCR. The technique has the capability to inspect an entire genome



for polymorphisms and is highly reproducible. Molecular genetic polymorphisms



are identified by the presence or absence of fragments after restriction and



amplification of genomic DMA. AFLP involves four basic steps after DMA



extraction from pure cultures: DMA digestion with restriction enzymes; ligation



with oligonucleotide adapters; selective amplification with labeled primers; and



gel-based analysis of amplified fragments. See Figure 1 for a representation of



the procedure. Characteristics of the AFLP procedure include:  PCR and
                                                                      8

-------
fragment analysis are relatively fast to perform if using automated machines; the



entire genome is inspected for polymorphic fragments; uses small amounts of



genomic DMA and the DMA concentration does not affect the reproducibility of



the assay; provides 50 to 200 fragments per genome assayed allowing for easy



identification of polymorphisms; is highly reproducible; and taxon-specific primer



sets are not required (commercially available primers work with a large variety of



genomes).  In addition, the technique can be applied to a large variety of DMA



samples including plants, animals,  human, and microbial genomes. Some of the



most common applications have included generating high resolution genetic



maps in plants and animals, analysis of the genetic diversity in plants and



animals,  characterization of mammalian genotypes, genotypic analysis and



epidemiological typing of bacteria, genotypic classification of fungi, and the



characterization and classification of pathogens (Blears et a/.,  1998).

-------



                                                             of
                                                     adapters
                                                             Ligated

                             Gel
                                                     Amplified
                                                     fragments

                                                 Adapted from Pfaller, 2002
Figure 1 Main steps of the AFLP Procedure


      Box-PCR Analysis.  BOX-PCR is another PCR-based DMA fingerprinting

technique based on amplification of the interspersed repetitive sequences (rep-

PCR) found in the DMA of many bacterial species (Koeuth et a/., 1995). The

BOX element originally described for Streptococcus pneumoniae consists of

three, highly conserved, interspersed, repetitive sequences: boxA, boxB, and

boxC (Martin et a/., 1992) that contain 59, 45 and 50 basepairs in length,

respectively.  BOXA1R and  BOXA2R primers are based on the boxA sequence,

and have been widely applied for rep-PCR amplification of DMA from a wide

variety of bacterial species (Koeuth et a/., 1995), including Enterococcus.  A

comparison of BOX-PCR to pulse field gel electrophoresis (PFGE), identified as
                                                                     10

-------
the gold standard for Enterococcus sp. fingerprinting, indicated that both



techniques yield very similar results at the subspecies level for Enterococcus



faecalis (Malathum etal., 1998).



      Host specific 16S-rDNA markers.  The majority of molecular tools



currently being applied for microbial source tracking rely on the development of



an extensive library of cultured isolates to which DMA fingerprints from



environmental samples can be compared.  The two aforementioned methods fall



into this category. LD methods are labor-intensive and limit the target indicator



bacteria to those that can be readily grown in a laboratory and can also  survive



outside the intestine (Simpson et a/., 2002). Combining technological advances



in molecular biology, such as polymerase chain reaction (PCR) and 16S rDNA



gene sequence analysis, has provided powerful tools for characterizing  microbial



populations without the need for cultivation of the targeted indicators. These



combined techniques have become very useful for of MST application.  For



example, PCR amplification of 16S rRNA gene sequences of the genera



Bacteroides-Prevotella has proven useful for the identification of specific hosts,



such as human, cattle, horses, and pigs  (Allsop & Stickler,  1985, Bernhard &



Field, 2000b,  Dick et a/., 2005, Kreader, 1995). These anaerobic bacteria are



restricted to the intestinal environment of warm-blooded animals. Unlike some



other fecal coliform bacteria, these Bacteroidetes do not survive long in  water,



and make up  30 to 40% of the total fecal bacteria (Harmsen et a/., 1999; Layton



et a/., 2006), which could account for up to 10% of the fecal mass. Therefore,



these anaerobic bacteria could be used as suitable indicators of species-specific
                                                                       11

-------
contamination. Prior to the development of culture-independent molecular



methods, the use of Bacteroides as indicators was limited because of the



difficulty to grow them in culture.



       Metagenomic Markersln addition to PCR amplification of specific 16S



rRNA genes, recent development of another culture-independent technique,



genome fragment enrichment, also seems promising for selecting for host-



specific metagenomic markers (Shanks et a/., 2006a). This technique enriches



for genes that are specific in host organisms by subtracting the genes that are



common in other organisms.  The metagenomic approach not only targets the



16S rRNA gene, but all genes involved in bacterial-host interactions, such as



surface proteins (Shanks et a/., 2006a).  One drawback that the 16S rRNA gene



of Bactero/ctes-like species seems to have is its cross-reactivity with non-target



fecal sources (Lamendella et a/., 2007).  This is especially true for the cattle-



specific markers. The  metagenomic markers developed for bovine sources are a



good alternative that could possibly reduce the identification of false positives



due to that cross-reactivity.  The bovine metagenomic markers developed by



Shanks et al. (2006b) were successfully tested with minimal cross reactivity on



148 different bovine feces collected from a variety of locations across  the U.S.



However, although the latter assays were tested against fecal samples obtained



from different regions, more detailed site tests of the temporal and spatial



variability of  the markers are still required to determine their environmental



stability and  robustness.
                                                                       12

-------
General Research Approach



This research was divided into two general approaches:





   •  Evaluation and comparison of the presence of 16S-rDNA and



      metagenomic markers in both water and sediment samples collected from



      two watersheds associated with cattle farms under different management



      practices (see Figure 2), and from a rural community serviced by



      individual household septic wastewater treatment systems.



   •  Comparison of amplified fragment length polymorphism (AFLP) and



      repetitive polymerase chain  reaction with BOX-primer (BOX-PCR)



      methodologies to genotype an Enterococcus sp. source library, and



      determine the usability of each methodology for host-specific source



      identification (see Figure 3).
                                                                     13

-------
                                      water
                                      sediment
                                      feces
filtration   	enterococci counts
                                                      DMA extraction
                                                      PCR amplification
                                                         Statistical analyses
Figure 2  Experimental scheme to perform 16S rDNA and metagenomic marker
analyses.
                                                                       14

-------
      Two cattle
      farms
      located in
      north east
      Georgia


      First order
      streams
      with
      headwaters,
      7 sampling
      locations
      per stream


      Cattle have
      unrestricted
      access to
      stream
                     5 cow patties/farm
                        t
                      Water addition and
                      membrane filtration
Quantification and isolation
of 35 suspected enterococci
Isolation of 50 suspected
enterococci
                     Identification of individual species using multiplex PCR
                         AFLP and BOX-PCR Analysis
Figure 3 General experimental design to isolate and fingerprint enterococcal
species
Methodology
      Sampling locations for 16S-rDNA and Metagenomic Markers. The
study sites to compare the metagenomic markers and 16S-rDNA primers
consisted of two watersheds associated with cattle farms. Watershed 1 (WS1)
flows across Farm A located in Madison County, GA. Watershed 2 (WS2) starts
in Farm B located in USDA-owned land in Watkinsville, GA. In WS1, samples
were collected from 4 sites along a creek and a pond (Figure  4).
                                                                          15

-------


Figure 4 Four sites were sampled in WS1 (Farm 1), three sites were located along the
stream while site 4 was located in a pond used by the cattle for bathing and drinking.
Aerial photo courtesy of GlobeXplorer.com.

      Cattle had direct access to all sampling sites except for site 1 that was


located upstream from the farm, outside of the property fence approximately 0.13


miles downstream from the origin of the stream. Sites 2 and 3 were located in the


middle and end of the stream crossing the farm, respectively.  Site 4 was located


in a pond used by the cattle for drinking and bathing. On  average, 60 head of


cattle were present on the farm during the course of this  study.  Wildlife, such as


geese and deer, also could contribute to the fecal sources impacting both water


bodies in this farm.


      There were 12 sampling sites in  WS2; seven of the sites were located


along the headwater stream and a pond within Farm B, while five sites were



                                                                         16

-------
located in the same creek downstream, outside the farm (Figure 5). Sites 5 and

11 consisted of agricultural and community ponds, respectively. An average of

140 head of cattle were kept and rotated among 16 fenced pastures in Farm B

during our study. The cattle had no access to the stream or the pond at this farm.

Other possible fecal sources affecting the stream and ponds in this watershed

include wildlife such as deer, geese and raccoons. Neither watershed was

deemed significantly impacted by human fecal pollution.
Figure 5 Sampling sites related to WS2 (Farm B). Sites 1-7 are located within the farm
boundary, 8 and 9 are located in a buffer zone between the farm and a subdivision, and 10-
12 are located within a subdivision. Aerial photo courtesy of GlobeXplorer.com.
                                                                         17

-------
      A separate study site was selected to evaluate the human-specific 16S-



rDNA markers. The site consisted of a rural community located in the town of



Anasco, Puerto Rico.  Water from five sampling locations was collected over a



two week period. The sampling locations included one site along an intermittent



creek that crossed the community; three sites were located in the Casey River



basin (upstream and downstream from the community); and one site consisted of



a shallow well (30 feet).



      Sample collection.  For cattle primers, water and fecal samples were



collected on a monthly basis between September 2005 and February 2007.



Water samples were collected in sterilized 1-liter bottles, kept on ice for transport



to the laboratory and processed for enterococci enumerations and nucleic acid



extractions within 6 hours after collection.  Two fecal samples per sampling event



were collected aseptically from each farm. Fecal samples were stored at -20°C



until processed.



      The samples collected to test the human-specific Bacteroidetes primers



were collected in collaboration with an ongoing study sponsored by the Puerto



Rico Water Resources and Environmental Research Institute in an effort to



provide information for the development of TMDLs for the Rio Anasco.  Water



samples (100, 250, and  500 ml) were filtered through polycarbonate filters (0.2



urn).  The filters were transferred to microcentrifuge tubes and stored at -20° C,



then shipped on ice overnight to our laboratory.



      Physico-chemical and microbiological methods. The temperature and



pH of water samples were measured on-site using a portable pH meter, Orion
                                                                      18

-------
250A plus (Thermo Orion, Beverly, Mass.). Daily precipitation data for WS1 and



WS2 were obtained from station ID 092517 of the National Oceanic and



Atmospheric Administration (httBi/liym^                         and the



Georgia Automated Environmental Monitoring Network



                              respectively. Water sample turbidity was



measured using a 2020 Turbidimeter (LaMotte Co., Chesterfield, MD) according



to the manufacturer's instructions. Enterococcal densities of the water samples



were determined using the membrane filtration technique described in EPA



method 1600. The colonies were counted twice after 24 and 42 hour incubation



at41°C.



      DMA Extraction and PCR amplification. In both WS1 and WS2,



approximately 100 ml water samples, and  0.2-0.25 g of cattle feces were used



for DNA extractions using an UltraClean Soil DNA Kit (MoBio Inc., California)



according to the manufacturer's instructions with some modifications.



Specifically, water samples were filtered onto polycarbonate filter membranes



(0.22 urn; Millipore Inc., Bedford, MA).  Each filter was then transferred to a 6 ml



sterile tube containing bead solution  and solution S1, and vortexed for 10 min.



Inhibitor removal solution (IRS) was added after solution S2, followed by the



steps in the manufacturer's instructions. The nucleic acid fraction was eluted to



65 ul of Tris-EDTA buffer. DNA was quantified photometrically using a NanoDrop



ND-1000 UVA/is spectrophotometer  (NanoDrop Technologies, Wilmington, DE),



and the DNA concentration was adjusted approximately to 10 ng/ul.
                                                                      19

-------
      The Puerto Rico water samples were also extracted using the MoBio kit.



The primers used included a general Bacteroides-Prevotella marker (32F or Gen-



Bac), two human-Bacteroides markers (HF183 and HF654), and two cattle-



Bacteroides primers (CF128 and CF193).  Gels were examined on 1.5%



agarose, mostly for 90 min at 90volts, with one or two exceptions at 100 volts for



60min.



      PCR assays were performed using GoTaq Green master mix (Promega,



Madison, Wl) with either 16S rDNA-based  Bactero/cfa/es-specific primer sets or



six cattle-specific metagenomic primer sets. The annealing temperature for each



PCR assay was determined  using a gradient PCR. The thermal cycling



conditions for the 16S rDNA-based markers were an initial denaturation of 2.5



min at 94°C, followed  by 30 cycles of denaturation at 94°C for 30 sec each,



annealing at an optimized temperature for each primer set for 30 sec plus



extension at 72°C for  1 min,  and a final extension of 5.5 min at 72°C. The thermal



cycling conditions for the metagenomic markers were 3 min of initial denaturation



at 94°C, followed by 35 cycles of 1 min each of denaturation (94°C), annealing



and extension (72°C), and a final extension step of 5 min at 72°C. Amplification



products were visualized on  a 2 % agarose gel stained with 0.2X SYBR Safe



DMA gel stain (Invitrogen). The limit of detection for each molecular marker set



was determined by PCR using serial dilutions of the extracted bovine fecal DMA



as templates, starting  at 10 ng/ul. Negative controls included DMA extracts from



sterilized nanopure water and no DMA template reactions, while DMA extracts
                                                                     20

-------
from feces freshly obtained at each sampling event were used as positive



controls.



      Sources for AFLP and BOX-PCR Analysis.  The library of enterococci



used for the fingerprinting analyses consisted of 1600 isolates collected over a



seasonal cycle at two separate bovine farms where cattle had unrestricted



access to the streams at all times (Molina, 2005). Samples were collected from



pre-farm (non-impacted, upstream-from-the-farm) stream sites, farm stream sites



(impacted), and fecal matter.  The library of enterococcal species was developed



by isolating colonies from mEI plates and identifying them at the species level



using a multiplex PCR procedure (Jackson et a/., 2004).



      AFLP and BOX-PCR Procedures.   Genomic DMA extraction from each



Enterococcus isolate was performed using a Qiagen DNeasy Tissue Kit. The



AFLP procedure was adapted from (Antonishyn et a/., 2000). AFLP restriction



and ligation was performed using /-//ndlll, and Mbo\. Digested genomic DMA was



amplified in parallel reactions using two different selective primer sets, Mbol-AC



and Mibol-CTG. The BOX-PCR procedure was an adaptation from (Malathum et



a/., 1998). BOX-PCR Amplification was  performed using Gitschier buffer (pH



8.0) (Kogan et a/., 1987),  and a BOXA2R primer. To perform fragment analysis,



the PCR products were electrophoresed through 6% polyacrylamide denaturing



gels with a well-to-read distance of 30 cm for 3 hours on a MJ  Research



BaseStation 51 DMA Fragment Analyzer. The parallel reactions were run on



separate gels with a custom size standard in each lane (BioVentures), allowing



accurate sizing of fragments  in the 50-600 bp range.  For the phylogenetic
                                                                     21

-------
analyses, we analyzed gel images using BioNumerics v3.0. Dendrograms were



created from the Mbol-AC and Mbol-CTG fingerprints using a curve-based



similarity coefficient (Pearson correlation) with the unweighted pair group method



(UPGMA).








Results and Discussion



Evaluation of Library-Independent Methods



      Comparison of 16S rDNA-based vs. metagenomic marker



performance in farm waters impacted by cattle fecal contamination. The



general 16S rDNA marker (32F) was detected in all sampling sites at a very high



frequency (81%), except in the sites related to the ponds or their effluents (17%)



(Table 1).  This general marker was followed in frequency order by the cattle-



specific 16S rDNA marker (CF128),  and then by the metagenomic markers Bac



2,1,5, and 3. The metagenomic markers as a whole were found to be 41 -60%



less frequent than the 16S rDNA cattle marker in stream waters under direct



impact (WS1), and between 3-5% less frequent in stream water under indirect



impact (WS2), depending on the sampling site.  These results suggest that the



metagenomic markers are less sensitive than the 16S-rDNA based markers, they



are less stable in the environment, or their presence in cattle is more variable.



The fact that the metagenomic markers were not found in every single cattle



patty sampled at a given time (data not shown) points to the possibility of a



higher variability in cattle manure, but it does not discard the other two



possibilities. Nevertheless, the CF 128 marker was found in relatively low
                                                                    22

-------
frequencies (3-9%) in WS2 versus the directly impacted stream of WS1 (71-94%)



(Table 1).  This indicates a rather low impact of cattle fecal contamination



reaching the stream water in WS2 through run-off compared to the direct inputs



in WS1, even though enterococci numbers were rather high in both farm



streams.
                                                                     23

-------
Table 1. Frequency (±95% Cl) of 16S rDNA-based Bacteroides and metagenomic markers
in water samples from two watersheds affected by cattle contamination. Watershed 1
(WS1) receives direct impact from cattle, while watershed 2 (WS2) only receives
contamination through runoff. Only markers with a frequency between 0.10 and 0.90 were
used for the logistic regression analysis
                           Marker Frequency ±95% Cl

Site     Bac32F   CF128F      Bad        Bac2        Bac3        BacS
WS1-1   0.76 ±0.09
WS1 -2, 3 1.00 ±0.00 0.94 ±0.02   0.62 ±0.08  0.65 ±0.08   0.26 ±0.06  0.59 ±0.08
WS1-4  0.88 ±0.05 0.71 ±0.10   0.12 ±0.05 0.18 ±0.07   0 ±0.00     0.12 ±0.05
WS2-1-4 0.72 ±0.05 0.09 ±0.02   0.03 ±0.01  0.06 ±0.01   0.04 ±0.01  0.04 ±0.01


WS2- 5,
         0.17 ±0.04 0.03 ±0.01        0     0.03 ±0.01        0           0
6,11,12

WS2- 7-
         0.70 ±0.06 0.04 ±0.01        0     0.04 ±0.01        0           0
10
  Relationship between enterococci enumeration and the occurrence of

molecular markers.  Enterococci counts were performed for all locations where

the markers were tested to establish the relationship with the alternative markers

under the two types of farm management.  The geometric mean of the

enterococcal numbers in the areas with the highest probability of cattle impact


                                                                       24

-------
ranged from approximately 24 to 1924 CFU/100ml in WS2 and WS1, respectively



(Table 2).  The counts taken at the ponds or pond outflows were the lowest (site



4 in WS1 and sites 5, 6, 11, and 12 in WS2), being 93 and 4 CFU/100 ml in WS1



and WS2, respectively.  The upstream locations (site 1 in both watersheds),



exhibited counts of 74 and 17 in WS1 and WS2, respectively. In general, WS2



exhibited much lower counts than WS1, which was expected due to best



management  practice implementation in WS2 (fencing cattle out of the stream).



When these results were compared to the observed  DMA marker frequencies, no



significant statistical relationships between the monthly enterococcal counts and



the presence  of the molecular markers in WS1 (Figure 6) were observed.  In



WS2, the enterococcal counts were statistically compared only to the general



marker 32F due to the absence of the other markers from most  sites (Figure 7).



In this case also, no significant relationship was identified between  the marker



and the enterococcal counts.  The only marker that indicated a slightly similar



trend to that observed with the enterococcal  counts was CF 128, and this only



during a brief  time of the sampling period (Dec 05-Feb 06) in WS1.   However,



this relationship didn't persist during the warmer months of the year or the



following winter season.



      In accordance with previous reports, enterococcal counts reported here



could not be related  to the occurrence of microbial source tracking  markers,



suggesting that more information is necessary to understand the dynamics of



DMA source identifiers  in a watershed in relation to the densities of traditional



fecal indicators such as E. coli (Shanks  et a/., 2006b) and enterococci. One
                                                                      25

-------
possible explanation for the discrepancy could be the differences in the

physiological and biochemical features between the two targeted bacterial

groups. Bacteroides are strict anaerobes and have low environmental

persistence, indicating recent contamination (Fiksdal et a/., 1985, Kreader,  1998,

Oshiro & Fujioka, 1995 1995, Ott etal., 2001). Although enumeration of

enterococci provides information on the level of impairment of a system, it does

not identify the specific source of contamination (Scott, 2005). Therefore, it is

recommended to employ a combination of molecular and traditional methods in

field studies to provide more accurate and reliable results in risk assessment and

prevention or reduction of contamination.
Table 2 Enterococcal abundance {CFU/100ml) in Watershed 1 and Watershed 2. Sites were
divided based on influence by cattle or type of water resource (streams vs. ponds).

Site#
WS1 site 1
WS1 sites 2, 3
WS1 site 4
WS2 sites 1-4
WS2 sites 5, 6,
11, 12
WS2 sites 7-10

Geometric mean
74
1924
94
26*
4*
123*
95% Confidence Interval
Lower bound
9
1130
42
10
2
59
Upper bound
640
3275
207
68
9
257
 Zero values in the data were converted to 0.01.
                                                                        26

-------
100 -,
 90 -
 80 -
 70 -
 60 -
 50 -
 40 -
 30 -
 20 -
 10 -
  0 -
                  Enterococci counts vs. molecular markers in WS1
N
                             \
                                                                             J2
                                                                             o
                                                                             o
                 tf
                                      Month
                 ICF178F
                                            Bac3 ^3 Bac5 -*-Ent. count
Figure  6  Relationship  between the  monthly enterococcal  counts and  the  average
frequency of the DMA markers per month in WS1
                                                                            27

-------
                Enterococci counts vs. Molecular markers in WS2
                                                                   1400
                                    Month
           Bac32F
CF128F
lBac2
Bac3
iBacS
•Ent. count
Figure 7 Relationship between monthly enterococcal counts and the average frequency of
the DMA markers per month in WS2
      Evaluation of human-specific 16S-rDNA markers in freshwater

streams impacted by rural non-point sources in Puerto Rico.  In this set of

samples, each primer group was run as follows:  Gen-Bac 32F and HF654 -four

times each; primer HF 183 - six times; and primers CF128 and 193 - twice each

with the objective to determine whether the locations sampled were impacted by

either human (HC) or cattle fecal contamination (CC).
                                                                    28

-------
Table 3 Description of samples collected in the Rio Ahasco Basin, Ahasco, Puerto Rico

from August 3-14, 2006


 0   .   _  .                     Site Description, volume of sample filtered, or sampling
 bampie uoae
 Site A                        Bridge 406



 Site B                        Bridge 430



 Site C                        Shallow well near Bridge 430



 Site D                        Intermittent stream crossing community



 Site E                        Casey River



 wrr*  rl win                Centrifuged sludge from Athens Water Treatment Plant



 Samples 1,4,7,10,13             500 ml of water filtered



 Samples 2, 5, 8, 11, 14             250 ml of water filtered



 Samples 3,6,9,12,15             100 ml of water filtered



 Samples 1 ,2,3                  Sampled on 8/3/06



 Samples 4,5,6                  Sampled on 8/5/07



 Samples 7,8,9                  Sampled on 8/7/06



 Samples 10,11,12               Sampled on 8/9/06



 Samples 13,14,15               Sampled on 8/14/07


*When CF primers were tested, WTC was substituted for Cow Fecal DMA as positive control.






      When indicated, 2x means that 1 ul of PCR product from  a first round (x1 )




was used as template for a second round (x2). Many of the (x2) gels had some




non-specific banding, however, when the correct band size for the primer listed




was present, the gel was scored with a (+);  those gels without the correct band




size, but with non-specific bands, were labeled "M" for multiple bands.  Results




were scored as Clean, Human, Cow, and Human-Cow, based on  the PCR




results.
                                                                           29

-------
      Although 2x amplification assays can increase the signal in those cases



where low initial concentrations of the target DMA are present, our results were



not considered solid enough due to the fact that some gels exhibited a large



number of non-specific bands. Further confirmation, for example through



sequencing of the bands obtained in the 2x amplification, would be necessary to



confirm the presence of the target DMA.



      Results for the 1X runs are presented in Table 4. Although the human



primer was not amplified in all samples collected from sites A and B, the results



indicate that human contamination seems to be present in the system at some



level during most sampling dates. The fact that human contamination was not



indicated in every single sample could indicate a low level of contamination or



presence of inhibitors in some samples. The level of contamination is hard to



assess without a real quantification assay.  Only once during the sampling period



(8/5/06) did the results  indicate that cattle contamination was  present in one of



the sampling locations  (site B). The presence of the Gen-Bac in the absence of



cattle or human contamination may point to another source of contamination



(neither human nor cattle), or cross-amplification with a natural bacterial



population.



      Samples obtained from the well seem to be free from cattle and human



contamination. Only one sample gave a positive human signal, this on the last



day of sampling (out of triplicates), and might not be enough evidence to indicate



an actual human impact. In a situation like this, inadvertent sample



contamination can not be discounted. The stream that crosses the rural
                                                                      30

-------
community (site D) is clearly impacted by human contamination, since that assay



was positive for every triplicate sample collected throughout the whole sampling



period. This community is served mainly by septic systems.  The contamination



observed points to the fact that these septic systems might not be working



properly and are leaking into the intermittent stream traversing the community.



The Casey River also seems to be strongly impacted by human contamination,



but in this case the contamination might be intermittent, since no contamination



was detected in any of the samples collected on the last day of sampling. One



possible explanation for this observation is dilution of the assay signal in the river



or fecal bacterial decay after the initial contamination episode.



      There was no relationship observed between the volume of sample filtered



and the presence of a marker, meaning that sometimes a marker was positive in



the 100 mi-sample while it was absent in the 500 mi-sample and vice versa. This



result could be a function of the amount of inhibitors present  in a sample at a



given time, or it could just reflect sample randomness.  Duplicate and, if possible,



triplicate sample collection is recommended to cover sample variability.
                                                                      31

-------
Table 4 Bacteroidetes 16S rRNA gene marker hits in water samples collected in the Rio
Ahasco Basin, Ahasco, Puerto Rico.  The numbers indicate the times the individual primer
set was found in each water sample after one amplification round (1x)
                               Number of Primer Set Hits

Sample
Neg Control
WT
C/CowCF*
WTD
A-1
A-2
A-3
A-4
A-5
A-6
A-7
A-8
A-9
A-10
A-11
A-1 2
A-1 3
A-1 4
A-1 5
B-1
B-2
B-3

B-5
B-6
B-7
B-8
B-9
B-10
B-11
B-1 2
B-1 3
B-1 4
B-1 5
C-1
C-2
C-3
C-4
C-5
C-6
C-7
General
Bacteroides
Marker
Gen-Bac
0

4
4
1
1
0
1
2
0
0
0
0
0
0
0
0
0
0
1
1
1

2
1
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
Human Human Cattle Cattle
Bacteroides Bacteroides Bacteroides Bacteroides
Marker Marker Marker Marker
HF-183
0

2
3
1
0
0
0
1
0
0
1
0
0
1
2
1
0
1
0
0
1

1
0
0
0
0
0
1
0
0
0
1
0
0
0
0
0
0
0
HF-654
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
CF-128
0

1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
CF-193
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Comments:
Control

Control
Control
Human
General
Clean
General
Human
Clean
Clean
Human
Clean
Clean
Human
Human
Human
Clean
Human
General
General
Human
Human and
cow

Clean
General
Clean
General
Human
Human
Clean
Clean
Human
Clean
Clean
Clean
Clean
Clean
Clean
Clean
                                                                             32

-------
C-8
C-9
C-10
C-11
C-12
C-13
C-14
C-15
D-1
D-2
D-3
D-4
D-5
D-6
D-7
D-8
D-9
D-10
D-11
D-1 2
D-1 3
D-1 4
D-1 5
E-1
E-2
E-3
E-4
E-5
E-6
E-7
E-8
E-9
E-10
E-11
E-1 2
E-1 3
E-1 4
E-1 5
0
0
0
0
0
0
0
0
2
0
0
0
1
2
2
3
0
2
2
0
1
0
0
0
0
0
2
2
0
2
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
2
0
1
2
1
1
2
2
1
2
0
1
2
2
1
1
1
2
2
2
2
1
2
1
0
2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Clean
Clean
Clean
Clean
Clean
Clean
Clean
Human
General
Human
Human
Human
Human
Human
Human
Human
Human
General
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Clean
Human
Human
Clean
Clean
Clean
Clean
33

-------
Evaluation of Library-Dependent Methods
      Seasonal Distribution of Enterococci Isolates. Application of
Pearson's chi-squared statistics to our data indicated that the proportions of the
different bacterial species varied seasonally. However, this variability did not
stand a monthly statistical distribution analysis, which means that the differences
observed were due to sample randomness rather than true seasonal differences.
Nevertheless, some general trends were identified with E. faecalis and E. hirae,
although not with E. casseliflavus.  E. faecalis seems to be the only species
showing a trend of higher occurrence frequency during the warmer months of the
sampling period, April through November 2004 (Figure 8).  E. hirae was present
more commonly during colder months (spring, fall and winter).  E. casseliflavus
indicated no correlation with season, farm sample site or source of sample (water
vs. manure).  In addition to the behavior of the three former species, E. faecium
was found in  higher abundance only during the winter of 2005.  The winter
months also reflected the highest diversity in terms of number of species
identified and the evenness of the different populations.
     100%n
     90%
     80%
     70%
     60%
     50%
     40%
     30%
     20%
     10%
      0%
D other
B flavescens
• mundtii
n hirae
D durans
• faecium
D faecalis
• casseliflavus
           Fall 2003   Spring 2004  Summer 2004   Fall 2004    Winter 2005
Figure 8 Seasonal distribution of enterococcal species in impacted streams by cattle
contamination
                                                                         34

-------
   Some researchers have suggested that species such as E. faecalis can be



used as markers for human contamination (Wheeler et al., 2002).  However,



results from this research suggest that the large seasonal variability exhibited by



the different enterococcal populations identified make the use of individual



Enterococcus species unreliable due to lack of temporal stability. The observed



variability, combined with the observed presence of the same Enterococcus



species  in the cattle farm stream water and the water upstream from the farm,



highlights the fact that enterococci populations are widespread in nature. This



could make their use as markers at the species level undependable.  To evaluate



the suitability of Enterococcus at the subspecies level to serve as markers of



bovine contamination, it was necessary to perform fingerprinting analysis of



some of those subspecies that were observed to be  present in the cattle farm



streams throughout the year.








      Some methodological  considerations developing the phytogeny of



Enterococcus strains using AFLP.  The two primer sets,  Mibo/-CTG and Mbol-



AC, for Hex and Fam, respectively, exhibited congruencies of up to 60%. The



40% incongruence can be explained,  in part, by the dynamic phylogeny



produced due to the high species diversity in the library. Detailed analysis using



band matching and maximum parsimony will need to be performed in order to



obtain more detailed information. The phylogenetic trees produced by each



primer set for E. faecalis yielded the greatest incongruence; however, the two



primers  produced the same basic groupings for both E. hirae and E.
                                                                      35

-------
casseliflavus. These results suggest that E. faecalis exhibits the highest species



diversity in the environment among the three species studied.



      Phylogeny of E. faecalis, E. hirae ,and E. casseliflavus strains using



AFLP.  Using primer set Mbol-CTG, E. faecalis isolates separated into two



distinct clusters depending on the farm from which they were isolated. Other



than the farm differences, isolates were not found to group by source (manure



vs. water), season, or location (stream sites within the farm or stream sites



upstream from the farm). One possible explanation for this division is that cattle



uptake part of their E faecalis fecal population from their drinking water.



Because their drinking water includes the upstream-from-the-farm water, this



possibly explains why the E faecalis isolated from manure could not be



differentiated from that isolated from the upstream water. This observation also



implies that the E. faecalis population present in the wildlife inhabiting each farm



differs from each other, since no similar fingerprints were identified across farms.



      E hirae also showed two distinct clusters, one containing isolates mainly



collected during autumn 2003 in Farm 1 from manure, and the other comprised



of isolates from all seasons and sources (Figure 9).  The autumn 2003 cluster



was not observed at any of the upstream locations, suggesting that it is



composed of species mostly present in the feces of the cattle on Farm 1. This



cluster is in close phylogenetic relationship to a spring cluster found from both



farms, composed of isolates obtained mostly from the water within the farms, but



absent in the water collected upstream of the farms.  Because the fingerprints in



these two clusters are absent in the water upstream from the farms, they could
                                                                        36

-------
be developed and tested as MST markers for cattle fecal contamination.



However, one drawback observed is the fact that these two clusters only



showed-up during autumn and possibly spring, but not during other times of the



year.  This makes them temporally unstable and unreliable. A good indicator



needs to be present throughout all seasons (Simpson et a/., 2002).



      E. casseliflavus isolates also  grouped into two basic clusters, with one



cluster accounting for 73% of the library.  For this species, no seasonal or source



trends were observed, and many fingerprints were found in the water upstream



of the farms.  In addition, no difference was observed  between farms. These



results suggest that E. casseliflavus fingerprints are widespread in the



environment,  making it hard to distinguish contributions of cattle vs. wildlife.
                                                                       37

-------


                                                                                            Oms:

                                                                      bpnaf
                                                                      Spring

                                                                      Spring
                                                                      Spring
                                                                      Spring
                                                                      Spring
                                                                      Spring

                                                                      Sprifli
                                                                      WinJet
                                                                      Wisfif
                                                                      WinJet
                                                                      SprifU;
                                                                      WinJet
                                                                      Wist!

                                                                      Ssia,g
                                                                      Wins:
                                                                      AMtiiBm
                                                                      Spring
                                                                      Wieti
                                                                      »llit!

                                                                      S:iaarn%3
                                                                      AXKITCL
                                                                      Sttrbig
                                                                      Stria Ji
                                                                      Wifcto
                                                                      Sfrriag
                                                                      Sittrysg

                                                                      StTOEg

                                                                      Srsiitf
                                                                      Sfrriag

                                                                      Atftan
                                                                      AKSMK
                                                                      Atitaii
                                                                      Aii&ji)
                                                                      AuJian
                                                                      Astan
                                                                      Axton
                                                                      Aiikmi
                                                                      WiilB
                                                                      Axton
                                                                      Att&un.
F
F
F
I
F
F
F
F

F
1
F
F
F
F
F
F
F
PF
F
F
F
F

F
F
IT
F
I
F
f
F
F

f
F
F
F
I
F
F
F
F
F
F
F
F
F

I
F
F
M
M
M
i
M
I
M
i
M
I
i
i
i
M
1
1
M
M
M
i
i
i
i
i
M
i
i
1
M
M
i
M
M
y
M
y
M
y
M
y
M
M
I

      iwaoof

Figure 9 E. hirae phylogenetic tree derived from AFLP fingerprints of isolates obtained
from water and manure samples collected at two cattle farms with impacted streams.
                                                                                              38

-------
      Comparison of AFLP and BOX-PCR analysis.  The genotyping methods



of BOX-PCR and AFLP each have distinct advantages and disadvantages (Table



5). Our results showed that, in general, AFLP is far superior at discriminating



closely related strains of Enterococcus. AFLP produced a greater number of



bands per PCR reaction, providing greater discriminatory power; had greater



precision in band sizing; and allowed for the use of multiple primer sets.



Additionally, the quality of the AFLP gels was very consistent in  terms of both



band reproducibility and overall gel usability for data analysis. Throughput of



samples was also greater with AFLP due to the high sensitivity of the



fluorescence based-detection, thereby allowing the use of much smaller band



lane widths.



      The advantages of BOX-PCR are: the simplicity of the method (fewer



steps, technically easier); much lower cost of equipment (only a regular thermal



cycler is required) and reagents (BOXA2R primer, enzymes and buffer); and no



production of hazardous waste.  However, the  procedure produced highly



variable results in terms of band detection.  It also produced lower discriminating



power than the AFLP  procedure because for most species we were only able to



obtain between 9 and 18 different bands (Figure 10).  In contrast, the AFLP



analysis consistently produced over 100 bands. The BOX-PCR procedure was



also highly sensitive to the buffer pH, which can affect band detection.  In



addition, sensitivity and band brightness was highly affected by  gel quality.
                                                                      39

-------

Figure 10 Typical BOX-PCR gel image produced with E. faecalis isolates.
                                                                             40

-------
Table 5 Comparison of the advantages and disadvantages of the BOX-PCR and AFLP
methodologies	
                               Methodology
             BOX-PCR
                                  AFLP
  Disadvantages
   Advantages
  Disadvantages
   Advantages
Poor consistency
in gel quality
(affects band
'brightness' or
sensitivity).
Poor
reproducibility (pH
variability affects
band detection;
high PCR assay
variability).
Low band sizing
precision (inability
to discern similarly
sized  bands).
Requires certain
gel  lane width for
accurate detection
(reduces
throughput).
Assay produces
fewer bands (low
discriminatory
power).
Inexpensive (no
expensive
equipment,
primers, or
standards).
Technically
simple.
No hazardous
waste.
Expensive
(machine, primers,
and standard).
Technically more
challenging (more
steps).
Hazardous Waste.
Much greater
consistency in gel
quality (although
not perfect, very
sensitive).
High
reproducibility.
Very high band
sizing precision.
                                      High throughput.
                                      Many bands (high
                                      discriminatory
                                      power).

                                      Option to use
                                      different selective
                                      primers.	
Conclusions

      Application of AFLP methodology vs. DMA markers.  Studies

examining bacterial strain diversity and temporal variability in aquatic and

terrestrial habitats using the level of genetic specificity undertaken in this study

are uncommon. Our work helps fill this void by providing a genotyping study that
                                                                       41

-------
involves hundreds of Enterococcus strains from multiple species, seasons, and



two aquatic systems, as well as a detailed temporal screening of 16S and



metagenomic markers. AFLP genotyping of our Enterococcus strain library



provided a large and robust data set, that supplied many unique fingerprints. We



identified a fingerprint of E. hirae that seems to be fairly specific to cattle manure



samples; however, the fingerprint showed-up only during two out of the five



seasons sampled.  This makes the fingerprint unsuitable for MST applications



due to the lack of temporal stability and reliability.  The fact that E. faecalis



isolates grouped by farm and showed no correlation to source (upstream-of-the-



farms and farm  water,  or manure) suggests that the cattle in our study may



uptake part of their E faecalis population from their drinking water which then



gets transferred to their manure. Such environmental uptake masks identification



of cattle-specific fingerprints of E. faecalis.



      Although the AFLP methodology is very reproducible and has high



discriminating power, its application as a rapid and resource-efficient



methodology is  limited because the library production is highly time and resource



consuming.  Its  application is probably most appropriate in very specific



scenarios where discrimination among few selected sources is necessary. In



contrast, application of DMA, PCR-based markers produced fairly rapid results,



and had the capability to screen multiple scenarios in a short period of time.



Once stability and cross-amplification  aspects have been addressed, it can be a



highly efficacious approach to determine sources of contamination in a variety of



scenarios.
                                                                       42

-------
      From our results, we conclude that a combination of the ruminant-specific



marker, CF128F, with the metagenomic markers, Bad, 2 and 5, may provide a



solid application package for tracking bovine fecal contamination sources to



surface waters.  Because enterococcal counts did not show a strong correlation



with the occurrence of any of the DMA markers, the dynamics of fecal source



tracking markers in a watershed need to be further investigated to be able to



determine their correlation with the densities of traditional indicators of fecal



contamination.








Significance of Research



      This research supports an area of high priority for the Office of Water and



has been listed in the Twenty Needs Report as the highest priority for Regions



and States. This work supports assessment of aquatic systems impairment



under Long Term Goal 2 (LTG 2) of the Office of Research and Development



Water Quality Multiyear Plan. LTG2 provides the tools to assess and diagnose



the causes and pollutant sources of impairment in aquatic systems. Specifically,



the results of this research provide an evaluation of selected LI- and LD-mthods



as to their usability for early and rapid assessment of fecal contamination



sources.  Included is a specific application and comparison of some of the



available  DMA-based methodologies for discriminating among sources of



contamination in impaired surface waters.
                                                                      43

-------
Future Directions



   During the next phase of this research, we will focus on the application of the



Ll-approaches to quantify the loadings of agricultural, human and other non-



human non-point sources of bacterial contaminants into aquatic resources:



   •   Determine the loadings, fate and transport of bacterial contaminants from



      agricultural non-point sources in surface waters using quantitative PCR



      methods that will provide such information in an accurate, fast and



      informed way.



   •   Provide a basis for comparison between traditional fecal indicators, true



      pathogenic bacteria and DMA-based fecal indicators.



   •   Develop and validate a technique by which the recovery of an ecosystem



      from bacterial contamination can be measured, and provide information to



      watershed managers about the effectiveness of alternative BMP



      approaches.
                                                                       44

-------
References
[1] Allsop, K.  and J.D. Stickler (1985) An assessment of Bacteroides fragils
group organisms as indicators of human faecal pollution. J Appl Bacteriology 58:
95-99.
[2] Antonishyn NA, R.R. McDonald, E. L. Chan, G. Horsman, C. E.Woodmansee,
P.S. Falk and A.C.G. Mayhall (2000) Evaluation of Fluorescence-Based
Amplified Fragment Length Polymorphism Analysis for Molecular Typing in
Hospital Epidemiology:  Comparison with Pulsed-Field Gel Electrophoresis for
Typing Strains of Vancomycin-Resistant Enterococcus faecium. Journal of
Clinical Microbiology 38: 4058-4065.


[3] Bernhard,  A.E. and K. G. Field (2000a) A PCR Assay to discriminate human
and ruminant feces on the basis of host differences in Bacteroides-Prevotella
genes encoding 16S rRNA. Appl Environ Microbiol 66: 4571-4574.


 [4] Blears, M., S.A. De Grandis, H. Lee and J. Trevors (1998) Amplified fragment
length polymorphism (AFLP): a review of the procedure and its applications.
Journal of Industrial Microbiology & Biotechnology 21: 99-114.


[5] Dick, L.K., A.M. Bernhard, T.J. Brodeur, J.W. Santo Domingo, J.M. Simpson,
S.P. Walters and K.G. Field (2005) Host Distributions of Uncultivated Fecal
Bacteroidales Bacteria Reveal Genetic Markers for Fecal Source Identification.
Appl Environmental Microbiology 71: 3184-3191.
[6] Fiksdal L, J.S. Maki, S.J. LaCroix and J.T. Staley (1985) Survival and
detection of Bacteroides spp., prospective indicator bacteria. Appl Environmental
Microbiology 49: 148-150.
[7] Harmsen HJM, G.R. Gibson, and P. Elfferich (1999) Comparison of viable cell
counts and fluorescence in situ hybridization using
specific rRNA-based probes for the quantification of human fecal bacteria. FEMS
Microbiol. Lett. 183: 125-129.
[8] Jackson CR, P.J. Fedorka-Cray and J.B. Barrett (2004) Identification of
enterococci using a genus and species multiplex PCR. J Clinical Microbiol 42:
3558-3565.
                                                                      45

-------
[9] Koeuth T., J. Versalovic and J. R. Lupski (1995) Differential Subsequence
Conservation of Interspersed Repetitive Streptococcus pneumoniae BOX
Elements in Diverse Bacteria. Genome Research 5: 408-418.
[10] Kogan S.C., M. Doherrty and J. Gitschier J (1987) An improved method for
prenatal diagnosis of genetic disease by analysis of amplified DMA sequences:
application to hemophilia. A. N. Engl. J. Med. 317: 985-990.
[11] Kreader, C.A. (1995) Design and evaluation of Bacteroides DMA probes for
the specific detection of human fecal pollution. Appl Environmental Microbiology
61: 1171-1179.
[12] Kreader, C.A. (1998) Persistence of PCR-detectable Bacteroides distasonis
from human feces in river water. Appl Environmental Microbiology 64: 4103-
4105.
[13] Lamendella R, J.W. Santo Domingo, D.B. Oerther, J.R. Vogel & D. M.
Stoekel (2007) Assessment of fecal pollution sources in a small northern-plains
watershed using PCR and phylogenetic analyses of Bacteroidetes 16S rRNA
gene. FEMS Microbiol Ecol 59: 651 -660.


[14] Layton A, L. McKay, D. Williams, V. Garrett, R. Gentry and G. Sayler (2006)
Development of Bacteroides 16S rRNA Gene TaqMan-Based Real-Time PCR
Assays for Estimation of Total, Human and Bovine Fecal Pollution in Water. Appl
Environmental Microbiology 72: 4214-4224.
[15] Malathum K, K.V. Singh, G. M. Weinstockand B.E. Murray (1998) Repetitive
sequence-based PCR versus pulsed-field gel electrophoresis for typing of
enterococcus faecalis at the subspecies level. Journal of Clinical Microbiology
36:211-215.
[16] Martin B, 0. Humbert, M. Camara, E. Guenzi, J. Walker, and T. Mitchell
(1992) A highly conserved repeated DMA element located in the chromosome of
Streptococcus pneumoniae. Nucleic Acid Research 20: 3479-3483.


[17] Molina M. (2005) Temporal and Spatial Variability of Fecal Indicator
Bacteria: Implications for the Application of MST Methodologies to Differentiate
Sources of Fecal Contamination. U. S EPA, Athens
                                                                     46

-------
[18] Molina M., M. Cyterski, J. Maimes, J. Fisher and B. Johnson (2007)
Comparison of the temporal variability of enterococcal clusters in impacted
streams using a multiplex polymerase chain reaction procedure. (Hatcher K, ed.),
University of Georgia, Athens Georgia.
[19] Oshiro R. and R. Fujioka (1995) Sand, soil and pigeon droppings-Sources of
indicator bacteria in the waters of Hanauma Bay, Oahu, Hawaii. Water Sci
Technol 31: 251 -254.
[20] Ott E-M, T. Mller, M. Mller, C.M.A.P. Franz, A. Ulrich, M. Gabel and W.
Seyfarth (2001) Population dynamics and antagonistic potential of enterococci
colonizing the phyllosphere of grasses. J Appl Bacteriology 91: 54-66.
[21] Ritter Kj, E. Carruthers,  and C. A. Carson(2003) Assessment of statistical
methods used in library-based approaches to microbial source tracking. Journal
of Water and Health 1: 209-223.

[22] Scott T.M., T.  M. Jenkins, J. Lukasik, and J. B. Rose (2005) Potential use of
a host associated molecular marker in Enterococcus faecium as an index of
human fecal pollution  Environmental Science Technology 39: 283-287.
[23] Shanks OC, J.W. Santo Domingo & J.E. Graham (2006a) Use of competitive
DMA hybridization to identify differences in the genomes of bacteria. J Microbiol
Methods 66: 321 -330.
[24] Shanks O.C., J.W. Santo Domingo, R. Lamdella, C. A. Kelty & J.E. Graham
(2006b) Competitive Metagenomic DMA Hybridization Identifies Host-Specific
Microbial Genetic Markers in Cow Fecal Samples. Appl Environmental
Microbiology 72: 4054-4060.
[25] Simpson J., J.W. Santo Domingo & D. J. Reasoner (2002) Microbial source
tracking:  state of the science. Environmental Science and Technology 36:
5279=5288.
                                                                      47

-------
&EPA
     Environmental Protection
     Agency
     Office of Research
     and Development (8101R)
     Washington, DC 20460
     Official Business
     Penalty for Private Use
     $300
     EPA600/R-07/123
     December 2007
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.

If you do not wish to receive these reports CHECK HERE

D; detach, or copy this cover, and return to the address in
the upper left-hand corner.
PRESORTED STANDARD
 POSTAGE & FEES PAID
          EPA
    PERMIT No. G-35
                                                  Recycled/Recyclable
                                                  Printed with vegetable-based ink on
                                                  paper that contains a minimum of
                                                  50% post-consumer fiber content
                                                  processed chlorine free

-------