Environmental Occurrence of the Enterococcal ... - ACS Publications

Oct 3, 2008 - in nearshore sand; 0% in backshore sand; 24.4 and 0% in. Cladophora sp. ..... 48, cumulative rainfall of 6.45 cm in 72 h) (Table 1). Sim...
0 downloads 0 Views 717KB Size
Download PDF
Environ. Sci. Technol. 2008, 42, 8014–8020

Environmental Occurrence of the Enterococcal Surface Protein (esp) Gene is an Unreliable Indicator of Human Fecal Contamination MURULEEDHARA N. BYAPPANAHALLI,* KATARZYNA PRZYBYLA-KELLY, DAWN A. SHIVELY, AND RICHARD L. WHITMAN U.S. Geological Survey, Great Lakes Science Center, Lake Michigan Ecological Research Station, 1100 Mineral Springs Road, Porter, Indiana 46304

Received February 16, 2008. Revised manuscript received July 15, 2008. Accepted August 13, 2008.

The enterococcal surface protein (esp) gene found in Enterococcus faecalis and E. faecium has recently been explored as a marker of sewage pollution in recreational waters but its occurrence and distribution in environmental enterococci has not been well-documented. If the esp gene is found in environmental samples, there are potential implications for microbial source tracking applications. In the current study, a total of 452 samples (lake water, 100; stream water, 129; nearshore sand, 96; and backshore sand, 71; Cladophora sp.(Chlorophyta),41;andperiphyton(mostly Bacillariophyceae), 15) collected from the coastal watersheds of southern Lake Michigan were selectively cultured for enterococci and then analyzed for the esp gene by PCR, targeting E. faecalis/ E. faecium (espfs/fm) and E. faecium (espfm). Overall relative frequencies for espfs/fm and espfm were 27.4 and 5.1%. Respective percent frequency for the espfs/fm and espfm was 36 and 14% in lake water; 38.8 and 2.3% in stream water; 24 and 6.3% in nearshore sand; 0% in backshore sand; 24.4 and 0% in Cladophora sp.; and 33.3 and 0% in periphyton. The overall occurrence of both espfs/fm and espfm was significantly related (χ2 ) 49, P < 0.0001). Post-rain incidence of espfs/fm increased in lake and stream water and nearshore sand. Further, E. coli and enterococci cell densities were significant predictors for espfs/fm occurrence in post-rain lake water, but espfm was not. F+ coliphage densities were not significant predictors for espfm or espfs/fm gene incidence. In summary, the differential occurrence of the esp gene in the environment suggests that it is not limited to human fecal sources and thus may weaken its use as a reliable tool in discriminating contaminant sources (i.e., human vs nonhuman).

Introduction One of the key steps to the remediation of polluted swimming water is the identification of contaminant sources. Reference library-based source-tracking techniques have been criticized as being too expensive and tedious, requiring highly trained personnel and sophisticated laboratories (1, 2). Consequently, research emphasis has been directed toward nonreference * Corresponding author phone: (219) 926-8336; fax: (219) 9295792; e-mail: [email protected]. 8014

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 21, 2008

library-based source-tracking techniques for differentiating human from nonhuman sources or specific animal species (e.g., Bacteroides ruminant-specific molecular marker (3)). Both culture- and PCR-based methods for identifying hostspecific molecular markers associated with certain enteric bacteria and viruses (bacteriophages) have been pursued. Promising examples include Bacteroides (3-5), the enterococcal surface protein (esp) gene (6), and bacterial viruses such as F+ coliphages (7, 8). There have been several attempts to validate these methods using known human and animal fecal sources (see Field and Samadpour (9)); however, few, have examined the non-host occurrence of these molecular markers in the environment. Enterococci have been widely used as indicators of sewage contamination even though they are commonly found in secondary habitats, such as soil, sand, water, and plants (10-12). Scott et al. (6) have suggested the detection of Enterococcus faecium esp gene (espfm) as a rapid, nonlibrarydependent molecular marker for sewage. Later, Whitman et al. (13) surveyed the relative frequency of the esp gene in enterococci derived from humans, domesticated animals, and wildlife of the southern Lake Michigan area. They observed that the esp gene was consistently found in raw sewage influent (93-100%); however, the frequency was rather inconsistent in septic and pit toilet wastes. The detection of the esp gene in the fecal materials of animals, as diverse as dogs and gulls, implied that the gene is not necessarily limited to enterococci from human fecal wastes. The current study investigates the frequency and the pattern of the esp gene occurrence in enterococci from the environment under various hydrological conditions to explore its usefulness as a marker of human pollution. A total of 904 assays were run, targeting the esp genessespfm and espfs/fms in ambient sources such as lake and stream water, nearshore and backshore sand, an aquatic macroalgae (Cladophora), and periphyton from several coastal watersheds along southern Lake Michigan. The relationship of the esp gene with other fecal indicators (E. coli, F+ coliphage), as well as reliability and distribution of the esp gene as a marker of human pollution, were tested statistically using predictive modeling and hypothesis testing.

Materials and Methods Sample Collection. Most of the sampling in this research occurred in the watersheds of southern Lake Michigan between May and October 2006. The sites were selected because of the availability of historical water quality data and many of these sites are popular recreational beaches (e.g., Indiana Dunes State Park, West Beach, Washington Park, and the 63rd Street Beach). All samples were analyzed for enterococci and the esp gene (espfm, espfs/fm). Lake Water. A total of 100 water grab samples (approximately 500 mL each) were collected from 11 Lake Michigan beaches of Illinois and Indiana, including 63rd Street Beach (63rd, Chicago, Illinois), Hammond (HM), Lake Street (LS), Marquette Park (MQ), and Wells Street (WS) (Lake County, Indiana), West Beach (WB), Ogden Dunes (OD), Indiana Dunes State Park (SP), Dunbar (DU), and Central Avenue (CA) (Porter County, Indiana), and Washington Park (WP, La Porte County, Indiana) (Figure 1). Samples were collected by dipping a sterile polyethylene bag approximately 18 cm below the surface, taking care not disturb the underlying sediments. Some of the lake water samples were collected post-rain event (n1) and under dry weather conditions (n2) from LS, MQ, WS, WB, and OD and were analyzed for enterococci cell 10.1021/es800481p CCC: $40.75

 2008 American Chemical Society

Published on Web 10/03/2008

FIGURE 1. Most of the sampling took place in Illinois and Indiana: 63rd Street Beach (63rd), Hammond (HM), Lake Street (LS), Marquette Park (MQ), Wells Street (WS), West Beach (WB), Ogden Dunes (OD), Indiana Dunes State Park (SP), Dunbar (DU), Central Avenue (CA), and Washington Park (WP). In addition, samples were collected from the beaches of Door County (DC) and Racine (R), and the Sleeping Bear Dunes National Lakeshore (SB) (inset). densities and the esp gene (espfm, espfs/fm) (n1 ) 48, n2 ) 28); E. coli (n1 ) 41, n2 ) 28); and the F+ coliphage (n1 ) 43, n2 ) 13). post-rain and dry weather sampling conditions in this study are described as follows: a cumulative rainfall of 0.25 cm or more in 24 h and less than 0.25 cm in the preceding 24 h, respectively. Stream water. A total of 129 grab water samples (∼250 to 500 mL) were collected from several creeks: Dunes Creek, Derby Ditch, Kintzele Ditch, Munson Ditch, Salt Creek, Trail Creek, Little Calumet River, and Burns Ditch (Figure 1). Water samples from Dunes Creek watershed were analyzed for enterococci and the esp gene post-rain (n1 ) 42) and under dry weather conditions (n2 ) 17); E. coli (n1 ) 12, n2 ) 10); and F+ coliphage (n1 ) 42). Sand. Moist, subsurface sand samples (2-10 cm depth; n ) 96) were collected 1-2 m from the shoreline from the following beaches: WP, CA, DU, SP, OD, WB, WS, MQ, LS, and 63rd (Figure 1). Nearshore sand samples collected postrain and dry weather conditions from five northern Indiana beaches were analyzed for culturable enterococci, the esp gene (n1 ) 41, n2 ) 38), E. coli (n1 ) 35, n2 ) 35), and the F+ coliphage (n1 ) 6). In addition, 450 purified enterococci isolates previously collected from 71 backshore sand samples from DU and WB between January 2002 and May 2003 (10) were included in the analysis. In this manuscript, we refer to the backshore beach (sand) as the relatively flat area behind the berm and isolated from all but the highest lake surges. Macrophytic Algae (Cladophora). Forty-one Cladophora samples were analyzed in this study, including 24 samples collected from the following beaches: WP, OD, SP, and 63rd (Figure 1). Additional samples were collected from Sleeping Bear Dunes National Lakeshore (SB, Michigan) and Door County (DC), and Racine (R, Wisconsin). Also, 55 purified

enterococci isolates from 17 Cladophora samples collected in 2002 from DC, R, 63rd, OD, WP, and SB (14) were included in the assay. Periphyton. Fifteen periphyton samples, a complex mixture of algae, cyanobacteria, and heterotrophic microorganisms, were analyzed for the esp gene. Microscope slides were buried in sediment in July 2006 near the mouth of Trail Creek, which discharges into Lake Michigan at WP, and were retrieved for analysis in October 2006 (Figure 1). Periphyton growth on each slide was thoroughly scraped, the contents were transferred to a 50 mL test tube and then diluted with phosphate buffered water, PBW (KH2PO4 and MgCl2 · 6H2O, final pH 7.0 ( 0.2) (15) to a known volume. The resulting mixture was analyzed for enterococci and the esp gene.

Microbiological Analysis Enterococci, E. coli, and F+ Coliphage. All samples were first analyzed for enterococci by membrane filtration (MF) using mEI media (16). Filters with typical enterococci colonies were then transferred into a 15 mL centrifuge that contained 6 mL azide dextrose broth, ADB (Becton Dickinson, Sparks, MD) and were further enriched for 4 h at 35 °C (13). Aliquots (1 mL) were used as a template for polymerase chain reaction (PCR; see below). Selected lake and stream water and nearshore sand samples collected post-rain and dry weather conditions were also tested for E. coli by Colilert-18 method (IDEXX Laboratories Inc., Westbrook, ME) and for F+ coliphages using EPA Method 1602 (17). Coliphages (especially, the F+ coliphage), have been suggested as reliable indicators of sewage contamination in environmental waters (18). Bacteria from algae, periphyton, and sand were first elutriated by shaking vigorously by hand for two minutes or by using a wrist-arm shaker for 10 min. Elutriate was then VOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8015

TABLE 1. Incidence Rates of espfs/fm and espfm Relative to Sewage Indicators in Lake Water, Stream Water, And Nearshore Beach Sand during Dry and Post-Rain Weather Conditions substrate lake water stream water nearshore sand

weather conditions

Log enterococci cfu/ 100 mL or 100 g (N)

Log E. coli cfu/100 mL or 100 g (N)

F+coliphage pfu/ 100 mL (N)

espfs/fm+/N (%)

espfm+/N (%)

dry post-rain dry post-rain dry post-rain

1.4 ( 0.1 (28) 2.9 ( 0.1 (48)a 2.3 ( 0.1 (17)a 4.0 ( 0.1 (42)a 2.4 ( 0.1 (38)a 3.1 ( 0.1 (41)a

2.1 ( 0.1 (28) 2.5 ( 0.1 (41)a 2.3 ( 0.2 (10)a 3.3 ( 0.3 (12)a 2.8 ( 0.1 (35) 3.0 ( 0.1 (35)

59 ( 29 (13) 74 ( 13 (43) ND 238 ( 195 (42) ND 0 (6)

1/28 (3.6) 24/48 (50) 1/17 (5.9) 25/42 (59.5) 0/38 (0) 11/41 (26.8)

0/28 (0) 8/48 (16.7) 0/17 (0) 0/42 (0) 0/38 (0) 0/41 (0)

a

a

a Significant differences (ANOVA, P < 0.05) in bacteria concentrations between samples collected in dry and post-rain weather conditions. ND, not determined.

diluted as necessary for further analysis to achieve enterococci counts, usually in the range of 150-500/filter. All E. coli and enterococci analyses included control strains (E. coli ATCC 25922 and E. faecalis 29212) and blanks. Unless otherwise stated, E. coli and enterococci counts were expressed as colony-forming units (cfu) or most probable number (mpn) per 100 mL (lake and stream water, periphyton) or per 100 g fresh/dry weight (Cladophora, sand). Coliphage numbers were expressed as plague-forming units (pfu) per 100 mL. Esp Gene Detection by PCR. Template DNA for PCR Reactions. Cultures from ADB broth were transferred to 1.5mL centrifuge tubes and centrifuged at 12 000 rpm (4830g) for 5 min. After discarding the supernatant, the resultant pellet was washed (2×) with PCR-grade water (Integrated DNA Technologies, Inc., Coralville, IA) and later suspended in 300 µL AE elutriation buffer (Qiagen Gmbh, Hilden, Germany). Five microliters of the homogenized suspension was used as the DNA template. PCR Primers and Reaction Conditions. Two sets of primers were used for PCR amplification of the esp gene; the primer design and PCR protocols are explained in detail elsewhere (6, 13, 19). Positive- and negative-control strains were included with each set of amplifications. For espfs/fm, the positive and negative strains were 594 and OG1RF (kindly provided by Nathan Shankar, University of Oklahoma). For espfm, the positive-control was either sewage, where this marker has consistently been found (6), or E. faecium C68 (kindly provided by Joan Rose, Michigan State University); no template (5 µL PCR grade water) served as the negative control. PCR products espfs/fm, 950 bp; espfm, 680 bp) were separated on a 1.5% agarose gel stained with ethidium bromide and viewed under ultraviolet light. Gel images were captured using the KODAK Gel Logic 100 Imaging System. Band positions were determined with respect to a standard, 1-kb DNA ladder. Sequencing. Selected PCR products positive for the espfs/ fm and/or espfm gene from the current and a previous study (13) were partially sequenced to confirm that the PCR assays were amplifying the right target sequences and to identify the likely source of the esp gene (E. faecium or E. faecalis). Sixteen samples were chosen for sequencing the espfm gene (five dogs feces, three septic trucks, three sewage influent, five lake water) and 12 samples for sequencing the espfs/fm gene (fecal material from three songbirds, one mouse, four gulls, and two pit toilets and two sewage influent samples). Statistical Analysis. All statistical analyses were performed using SPSS v.12 statistical software. Chi-square test of independence was performed to compare the esp frequency among substrates. One-way analysis of variance and Spearman’s R correlations were conducted to compare enterococci concentrations based on positive or negative esp signal as well as post-rain event or dry weather conditions. Two logistic regression models were developed to characterize the relationship between the esp gene incidence and 8016

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 21, 2008

densities of other fecal indicators. The first logistic regression model included all six substrates tested (n ) 425) and was fit for the incidence of both markers (i.e., espfs/fm, 1a; and espfm, 1b) in relation to enterococci densities, where the logit of the incidence of the marker was the response, and the predictor variables and their two-way interaction included logit(marker) ∼ intercept + substrate + log Ent + substrate × log Ent The AIC (Akaike’s Information Criterion) was used to select the best fitting subset model for each marker (the lower the AIC value, the better the model). The relation from the best model was plotted and the goodness-of-fit statistic was reported (20). Best fitting subset models for both markers consisted of logit(espfm) ∼ intercept + substrate + log Ent logit(espfs⁄fm) ∼ intercept + substrate + log Ent + substrate × log Ent A second logistic regression model was also fit for incidence of both markers (i.e., espfs/fm, 2a; and espfm, 2b) in three substrates (i.e., lake water, stream water, and nearshore sand) in relation to enterococci, E. coli, and F+ coliphage densities under dry and post-rain weather conditions (n ) 214), with the following variables: logit(marker) ∼ intercept + condition + substrate + log Ent + log Ecoli + log F+ + condition × substrate + condition × log Ent + condition × log Ecoli + condition × log F+ + substrate × log Ent + substrate × log Ecoli + substrate × log F+ For the espfs/fm marker, the best fitted subset model (lowest AIC) consisted of logit(espfs⁄fm) ∼ intercept + condition + substrate + log Ent + log Ecoli and for espfm marker logit(espfm) ∼ intercept + condition + substrate

Results The espfs/fm Gene Incidence. The overall espfs/fm incidence in all samples was 27.4% (124/452). The espfs/fm gene was not detected in enterococci derived from backshore sand but was found in enterococci from all of the remaining samples analyzed, with the highest frequency in stream water (50/ 129, 38.8%), followed by lake water (36/100, 36%), nearshore sand (23/96, 24%), Cladophora (10/41, 24.4%), and periphyton (5/15, 33.3%). While only one positive for espfs/fm sample was detected in lake water during dry weather conditions (n ) 28), its incidence reached as high as 50% in post-rain samples (n )

48, cumulative rainfall of 6.45 cm in 72 h) (Table 1). Similarly, the incidence of the espfs/fm gene in stream water increased from 5.9% (n ) 17) in dry weather conditions to 59.5% (n ) 42) in post-rain samples (cumulative rainfall of 6.68 cm in 96 h). A similar pattern was observed for the nearshore sand: 0% (n ) 38) incidence rate during dry weather plus calm lake conditions to 26.8% (n ) 41) following storm and high wave surges (cumulative rainfall of 8.03 cm in 96 h). The taxonomic identity of the espfs/fm gene detected in 8 animal feces (3 songbirds, 4 gulls, 1 mouse) and 4 human fecal samples (2 sewage influent, 2 pit toilets) was verified by DNA sequencing. The PCR primers (esp11 and esp12) amplified approximately 960 bp region of the espfs/fm gene, encompassing nucleotides 1232 - 2154 (Enterococcus faecalis surface protein precursor gene). Basic Local Alignment Search Tool (BLAST) analysis indicated that the sequenced region from all 12 isolates had >98% nucleotide identity to the sequence of Enterococcus faecalis surface protein precursor gene (Accession number AF034779). The espfm gene Incidence. Overall, the espfm frequency was rather sporadic (23/452, 5.1%) in the substrates tested. The espfm gene was found in lake water (14/100, 14%), nearshore sand (6/96, 6.3%), and stream water (3/129, 2.3%) but not in backshore sand, Cladophora, or periphyton samples. The espfm incidence in lake water increased postrain, reaching 16.7% (n ) 48) (Table 1). No espfm was detected in any of the other substrates tested during dry weather conditions. The taxonomic identity of the espfm gene detected in samples from post-rain Lake Michigan water (five), dog (five), and human fecal material (three septic trucks, three sewage influent) was verified by sequence analysis of DNA. The PCR primers (espF and espR) amplified an approximately 680 bp region of the espfm gene, encompassing nucleotides 43515008 (Enterococcus faecium isolate E734 putative enterococcal surface protein (esp) gene). BLAST analysis indicated that the sequenced regions of all five lake water samples, six human fecal samples, and fecal material of five dogs had 97-99% nucleotide identity to the sequence of Enterococcus faecium isolate E734 putative enterococcal surface protein (esp) gene (Accession number AY322499S1). Taken together, these results indicate that the espfm gene detected in lake water samples is very closely related with that derived from human and dog fecal material. Fecal Indicators. Mean log ((SE) enterococci densities were as follows: lake water 2.3 ((0.1) (n ) 96); stream water 3.2 ((0.1) (n ) 120); nearshore sand 2.8 ((0.1) (n ) 86); backshore sand 1.2 ((0.1) (n ) 71); Cladophora 4.6 ((0.2) (n ) 37); and periphyton 2.6 ((0.1) (n ) 15). Enterococci densities in stream water were significantly higher than in lake water (P < 0.0001); the counts in nearshore sand were significantly higher than in backshore sand (P < 0.0001). Mean log enterococci densities in lake water, stream water, and nearshore sand post-rain were significantly higher relative to those during dry weather conditions (Table 1). Mean log E. coli counts in lake and stream water, but not in nearshore sand, were significantly higher post-rain than those during dry weather conditions (Table 1). F+ Coliphage. In lake water, mean F+ coliphage densities (pfu/100 mL) post-rain were higher but not significantly different from those of dry weather conditions (Table 1). While no F+ coliphage was measured in streams during dry weather conditions, the overall mean densities post-rain reached 238 ( 195 pfu/100 mL, with the highest counts detected in Dunes Creek (8200 pfu/100 mL). No F+ coliphage was detected in nearshore sand post-rain (n ) 6). Relations among Fecal Indicators and Markers. Overall, the espfs/fm frequency was approximately 6 times that of espfm. The overall occurrence of these two markers was significantly related (χ2 ) 49.0, P < 0.0001), with 77% agreement. The

frequencies of the two markers in lake water, stream water, and nearshore sand were significantly related (P < 0.05). Substrates in which both the espfm and espfs/fm genes were not detected were excluded from this analysis (Cladophora, periphyton, and backshore sand). The occurrence of both espfs/fm and espfm in post-rain tested sources was also significantly related (χ2 ) 10.0, P ) 0.001), with 60% agreement. There was no significant relationship between espfs/fm or espfm and F+ coliphage occurrence in samples collected post-rain (χ2 ) 0.03, P ) 0.875 and χ2 ) 2.90, P ) 0.09, respectively). In general, enterococci counts were positively correlated with espfs/fm incidence (Spearman’s R ) 0.37, P < 0.0001, n ) 425); however, the same relationship was not evident for the espfm gene. The espfs/fm incidence was positively correlated with enterococci densities for the following sample types: lake water (R ) 0.53, P < 0.0001, n ) 96), stream water (R ) 0.26, P ) 0.004, n ) 120), and nearshore sand (R ) 0.33, P ) 0.002, n ) 86); in contrast, espfm incidence and enterococci counts were correlated only for lake water (R ) 0.27, P ) 0.007, n ) 96). In the current research, post-rain lake water samples were collected either after a combined sewage overflow (CSO) or during days when there was visible stream plume following a heavy rainfall. Under these conditions, the espfs/fm and espfm were detected in 50 and 17% of the samples, respectively; 79% of these samples were also positive for the F+ coliphage. In contrast, the esp gene frequency during dry weather conditions was sporadic (1/28, espfs/fm). In samples collected in dry and post-rain conditions, the espfs/fm gene incidence was correlated with higher enterococci densities in lake water (R ) 0.45, P < 0.0001, n ) 76), stream water (R ) 0.27, P ) 0.042, n ) 59), and nearshore sand (R ) 0.39, P < 0.0001, n ) 79). The espfm gene incidence was not correlated with enterococci concentrations in any of the sources tested. E. coli densities were correlated with espfs/fm incidence (R ) 0.19, P ) 0.014, n ) 161) but not with espfm. Incidence of espfs/fm in the lake water and nearshore sand was significantly correlated with E. coli densities (R ) 0.30, P ) 0.012, n ) 69 and R ) 0.25, P ) 0.038, n ) 70, respectively) but not incidence in stream water. Generally, the incidence of the espfs/fm gene in lake water samples had the highest correlations with both enterococci and E. coli concentrations; as the logistic regression showed, densities of both of these bacteria were significant predictors for the espfs/fm gene occurrence (Figure 2A) but not the espfm gene incidence (Figure 2B). To characterize the relationship between esp incidence and indicator bacteria concentrations, we developed two logistic regression models (for incidence of espfs/fm and espfm separately). The first model (1a) demonstrates that the espfs/fm gene occurrence in different sources is influenced by three important factors: enterococci source, enterococci density, and the interaction between the two (Figure 3A). For instance, with enterococci densities of 10 000 cfu/100 mL or 100 g, the probabilities of espfs/fm incidence were as follows: lake water, 90%; stream water, 50%; periphyton, 40%; nearshore sand, 30%; algae, 20%; and backshore sand, 0%. The goodness-of-fit statistic was not significant (P ) 0.105, n ) 425), indicating that the model was adequate. The espfm model (1b) showed that sample source and bacterial counts affected the gene incidence but not the relationship between the two (P ) 0.648, n ) 425) (Figure 3B). The second logistic regression model (2a) shows the espfs/fm gene incidence in various substrates during dry weather and post-rain relative to sample type, weather conditions, and densities of enterococci and E. coli. The fit was not significant; thus, the model was adequate (P ) 0.963). In post-rain samples, probabilities of the espfs/fm gene occurrence increased with both enterococci and E. coli densities, with higher incidence rates in lake water relative VOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8017

FIGURE 2. E. coli and enterococci as significant predictors for (A) espfs/fm and (B) espfm frequency in Lake Michigan water samples during post-rain and dry weather samplings. The size of the symbols is connected to the weather conditions: small, dry weather; large, post-rain.

FIGURE 3. Logistic regression model showing the expected incidence rates of the esp genes, (A) espfs/fm and (B) espfm, relative to enterococci densities in the ambient samples: a, lake water; b, nearshore sand; c, stream water; d, periphyton; e, Cladophora; f, backshore sand. to stream water. For lake water samples collected post-rain, the probabilities of the espfs/fm occurrence were between 60 and 70%, when E. coli and enterococci counts ranged from 8018

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 21, 2008

1000 to 5000 cfu/100 mL. In stream water collected postrain, with E. coli and enterococci densities of 5000 cfu/100 mL, the probabilities of the espfs/fm gene occurrence were

only 20-30%. The espfm gene occurrence during post-rain conditions was very sporadic, and the model 2b was not adequate (P < 0.0001). The concentrations of F+ coliphage were not significantly correlated with the incidence of both genes and therefore not included in the second regression model (2a and 2b).

Discussion E. coli and enterococci are commonly found in a variety of substrates across northern Indiana watersheds: Cladophora mats (14, 21), beach sand (10, 22), riparian soils and sediments (23, 24), wetland plants (25), and groundwater seeps (23, 26). The relative frequency of host-specific genetic markers within the Enterococcus spp. could provide a greater ability to discriminate sources and characterize contaminated sediments, soils, and other potential substrates. We inspected the occurrence and distribution of one such proposed sewage markersthe esp genesin enterococci from ambient samples and related our findings to its use as a marker of sewage pollution (6, 13, 27, 28). In our study, samples collected post-rain event in lake and stream water and nearshore sand had higher esp incidence rates relative to those collected during dry weather conditions, an observation consistent with other related studies (27, 29, 30). Among lake water post-rain samples, 50% tested positive for espfs/fm, 17% for espfm, and 79% for F+ coliphage. Six of those samples contained esp markers and F+ coliphage, collectively suggesting sewage contamination. Determination of sewage contamination may prove to be more problematic when fewer sewage indicators are detected. The majority of stream water samples during dry weather and post-rain conditions were acquired from Dunes Creek, which is not known to have any significant point-source inputs. Espfm was not detected in samples collected during either dry or post-rain (6.68 cm rainfall in 96 h, with enterococci counts reaching 140 000 cfu/100 mL) conditions. Nonetheless, human contamination seems likely. The espfs/fm gene was far more prevalent (60% incidence) in the creek post-rain; 38% of these samples were also positive for F+ coliphage. Previous studies have shown that high E. coli and enterococci levels in Dunes Creek are mostly from nonpoint sources, such as runoff, soil and sediments, and wild animals (23, 24, 31). There are a number of pit toilets along Dunes Creek at Indiana Dunes State Park, and many of the residences along the lower reaches of the creek use septic systems. We previously reported 80 and 30% incidence of the espfs/fm gene in wastes from pit toilets and septic tanks, respectively (13). Their proximity to the stream and postrain detection of the esp gene implies that these sources are potential candidates of contamination. Wildlife cannot be excluded as a source since the espfs/fm gene was recovered in enterococci from songbirds (9.1%), mice (13.6%), and gulls (29.4%) (13). Non-detection of the espfm gene in Dunes Creek may be explained by its infrequent occurrence in some human wastes, such as pit toilets and septic tank waste (0 and 30% respectively), and its absence in wildlife (13). A larger more complex watershed with known human input is the Little Calumet River system of northwest Indiana and part of northeastern Illinois. The espfm gene was not found in samples collected after a sewage bypass event, while espfs/fm was consistently found for three days following the event. Concrete conclusions are difficult because we lack a control area above any contamination site (the river is urbanized through the branch in question) and we did not take corresponding dry weather samples. Sewage bypassing might suggest a known contamination event, especially since we had positive results from the raw influent itself, but the

absence of espfm could be a false-negative resulting from experimental/sampling error or from the large dilution caused by the flood waters. The absence of the espfm gene during this pollution event was unexpected because of its common occurrence in sewage (93-100%) (6, 13). Regardless, only a small percentage of enterococci seem to harbor this gene (Joan Rose, Michigan State University, East Lansing; personal communication). Considering these factors, we remain unconvinced that we can reliably detect a bypass event using the esp gene as a sewage marker without more intense replicate sampling over time. Residual presence of the espfs/fm gene in lake and stream water during dry weather conditions (1/28 and 1/17, respectively) may reflect incidental occurrence from wildlife or human sources. Interestingly, among samples collected at 63rd on two dry weather occasions, 4/6 lake water, 7/8 nearshore sand, and 2/4 Cladophora samples tested positive for espfs/fm; corresponding espfm frequencies were 1/6, 5/8, 0/4, respectively. These findings suggest that substrates such as nearshore sand and Cladophora may become sources of enterococci and associated pathogens to the lake water once they enter the water column. Possible sources in dry weather conditions may include dogs; strays are common and dogs are allowed along a nearby beach. Gulls are also common at 63rd and Cladophora and associated periphyton builds up in nearshore waters and along the shoreline annually. In the current study, we detected 33.3% of espfs/fm in periphyton retrieved from artificial substrates suspended in Trail Creek; half of the stream water samples also tested positive for this gene under dry weather conditions. Previously, we have shown that the espfs/fm gene occurs at a greater frequency than espfm in human and animal fecal sources (13). The occurrence of this gene in animal fecal material weakens its use as a microbial source tracking tool. However, the combination of the espfs/fm gene occurrence plus indicator bacterial levels, F+ coliphage, as well as other proposed molecular markers (e.g., Bacteroides) could increase confidence in such investigations. Application of predictive models for espfs/fm (see Figures 2A and 3A), might be useful in water quality investigations when managers may only have fecal indicator bacteria data available. Unfortunately, this relationship is not strong for the espfm gene occurrence and indicator concentrations, lessening its use in source-tracking studies. In conclusion, the differential occurrence of the esp gene in the environment and its association with enterococci from nonhuman sources may weaken its use as a reliable marker of sewage contamination.

Acknowledgments M.N.B. and R.L.W. contributed equally to this work. We appreciate the assistance of Rebecca Prosser, Emily Garritson, and Katelyn VanderPol in field and laboratory work. We thank Meredith Nevers for reviewing this manuscript and providing valuable comments. We are grateful to Michael Casteel, Joy Marburger, and Doug Wilcox for their critical review of this manuscript. Special thanks to Jean Adams, who helped with statistical analyses. This article is Contribution 1494 of the USGS Great Lakes Science Center.

Literature Cited (1) Scott, T. M.; Rose, J. B.; Jenkins, T. M.; Farrah, S. R.; Lukasik, J. Microbial source tracking: Current methodology and future directions. Appl. Environ. Microbiol. 2002, 68, 57965803. (2) Stoeckel, D. M.; Mathes, M. V.; Hyer, K. E.; Hagedorn, C.; Kator, H.; Lukasik, J.; O’Brien, T. L.; Fenger, T. W.; Samadpour, M.; Strickler, K. M.; Wiggins, B. A. Comparison of seven protocols to identify fecal contamination sources using Escherichia coli. Environ. Sci. Technol. 2004, 38, 6109–6117. (3) Bernhard, A. E.; Field, K. G. A PCR assay to discriminate human and ruminant feces on the basis of host differences in BacteroiVOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8019

(4) (5)

(6)

(7) (8)

(9) (10)

(11) (12) (13)

(14)

(15) (16)

(17)

(18)

8020

des-Prevotella genes encoding 16S rRNA. Appl. Environ. Microbiol. 2000, 66, 4571–4574. Field, K. G.; Bernhard, A. E.; Brodeur, T. J. Molecular approaches to microbiological monitoring: Fecal source detection. Environ. Monit. Assess. 2003, 81, 313–326. Okabe, S.; Okayama, N.; Savichtcheva, O.; Ito, T. Quantification of host-specific Bacteroides-Prevotella 16S rRNA genetic markers for assessment of fecal pollution in freshwater. Appl. Microbiol. Biotechnol. 2007, 74, 890–901. Scott, T. M.; Jenkins, T. M.; Lukasik, J.; Rose, J. B. Potential use of a host associated molecular marker in Enterococcus faecium as an index of human fecal pollution. Environ. Sci. Technol. 2005, 39, 283–287. Kirs, M.; Smith, D. C. Multiplex quantitative real-time RT-PCR for the F+ specific RNA coliphages: A method for use in microbial source tracking. Appl. Environ. Microbiol. 2007. Stewart-Pullaro, J.; Daugomah, J. W.; Chestnut, D. E.; Graves, D. A.; Sobsey, M. D.; Scott, G. I. F+RNA coliphage typing for microbial source tracking in surface waters. J. Appl. Microbiol. 2006, 101, 1015–1026. Field, K. G.; Samadpour, M. Fecal source tracking, the indicator paradigm, and managing water quality. Water Res. 2007, 41, 3517–3538. Byappanahalli, M. N.; Whitman, R. L.; Shively, D. A.; Evert Ting, W. T.; Tseng, C. C.; Nevers, M. B. Seasonal persistence and population characteristics of Escherichia coli and enterococci in deep backshore sand of two freshwater beaches. J. Water Health 2006, 4, 313–320. Hardina, C. M.; Fujioka, R. S. Soil: The environmental source of Escherichia coli and enterococci in Hawaii’s streams. Environ. Toxicol. Water Qual. 1991, 6, 185–195. Muller, T.; Ulrich, A.; Ott, E. M.; Muller, M. Identification of plant-associated enterococci. J. Appl. Microbiol. 2001, 91, 268– 278. Whitman, R. L.; Przybyla-Kelly, K.; Shively, D. A.; Byappanahalli, M. N. Incidence of the enterococcal surface protein (esp) gene in human and animal fecal sources. Environ. Sci. Technol. 2007, 41, 6090–6095. Whitman, R. L.; Shively, D. A.; Pawlik, H.; Nevers, M. B.; Byappanahalli, M. N. Occurrence of Escherichia coli and enterococci in Cladophora (Chlorophyta) in nearshore water and beach sand of Lake Michigan. Appl. Environ. Microbiol. 2003, 69, 4714–4719. APHA Standard Methods for the Examination of Water and Wastewater, 20th ed.; American Public Health Association: Washington, DC 1998. USEPA. Improved Enumeration Methods for the Recreational Water Quality Indicators: Enterococci and Escherichia coli, EPA/ 821/R-97/004; U.S. Environmental protection Agency: Washington, DC, 2000; p 49, (http://www.epa.gov/waterscience/ beaches/files/rvsdman.pdf). USEPA. Method 1602: Male-Specific (F+) and Somatic Coliphage in Water by Single Agar (SAL) Procedure, EPA 821-R-01-029; U.S. Environmental protection Agency: Washington, DC, 2001, (http://www.epa.gov/nerlcwww/1602ap01.pdf). Havelaar, A. H.; Furuse, K.; Hogeboom, W. H. F-specific RNAbacteriophages as model viruses in water hygiene: Ecological aspects. Water Sci. Technol. 1988, 20, 399–407.

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 21, 2008

(19) Shankar, V.; Baghdayan, A. S.; Huycke, M. M.; Lindahl, G.; Gilmore, M. S. Infection-derived Enterococcus faecalis strains are enriched in esp, a gene encoding a novel surface protein. Infect. Immun. 1999, 67, 193–200. (20) Hosmer, D. W.; Hosmer, T.; Le Cessie, S.; Llemeshow, S. A comparison of goodness-of-fit tests for the logistic regression model. Stat. Med. 1997, 16, 965–980. (21) Byappanahalli, M. N.; Shively, D. A.; Nevers, M. B.; Sadowsky, M. J.; Whitman, R. L. Growth and survival of Escherichia coli and enterococci populations in the macro-alga Cladophora (Chlorophyta). FEMS Microbiol. Ecol. 2003, 46, 203–211. (22) Whitman, R. L.; Nevers, M. B. Foreshore sand as a source of Escherichia coli in nearshore water of a Lake Michigan beach. Appl. Environ. Microbiol. 2003, 69, 5555–5562. (23) Byappanahalli, M.; Fowler, M.; Shively, D.; Whitman, R. Ubiquity and persistence of Escherichia coli in a midwestern coastal stream. Appl. Environ. Microbiol. 2003, 69, 4549–4555. (24) Byappanahalli, M. N.; Whitman, R. L.; Shively, D. A.; Sadowsky, M. J.; Ishii, S. Population structure, persistence, and seasonality of autochthonous Escherichia coli in temperate, coastal forest soil from a Great Lakes watershed. Environ. Microbiol. 2006, 8, 504–513. (25) Whitman, R. L.; Byers, S. E.; Shively, D. A.; Ferguson, D. M.; Byappanahalli, M. Occurrence and growth characteristics of Escherichia coli and enterococci within the accumulated fluid of the northern pitcher plant (Sarracenia purpurea L.). Can. J. Microbiol. 2005, 51, 1027–1037. (26) Whitman, R. L.; Nevers, M. B.; Byappanahalli, M. N. Examination of the watershed-wide distribution of Escherichia coli along southern Lake Michigan: An integrated approach. Appl. Environ. Microbiol. 2006, 72, 7301–7310. (27) McDonald, J. L.; Hartel, P. G.; Gentit, L. C.; Belcher, C. N.; Gates, K. W.; Rodgers, K.; Fisher, J. A.; Smith, K. A.; Payne, K. A. Identifying sources of fecal contamination inexpensively with targeted sampling and bacterial source tracking. J. Environ. Qual. 2006, 35, 889–897. (28) McQuaig, S. M.; Scott, T. M.; Harwood, V. J.; Farrah, S. R.; Lukasik, J. O. Detection of human-derived fecal pollution in environmental waters by use of a PCR-based human polyomavirus assay. Appl. Environ. Microbiol. 2006, 72, 7567–7574. (29) Betancourt, W.; Fujioka, R. Evaluation of the Enterococcal Surface Protein (Esp) an Alternative Indicator of Sewage Contamination of Hawaii’s Streams and Coastal Waters, May 21-25, 106th ed.; General Meeting of the American Society for Microbiology: Orlando, FL, 2006; (abstract). (30) Tang, R.; Waldron, M.; Breault, R.; Weiskel, P.; Stoner, R.; DiBara, M.; Dunn, W.; Duerring, C.; Beskenis, J.; Chase, R.; DiPietro, P.; Callaghan, T.; Celona, M.; Gray, D.; Pancorbo, O. Assessment of Sewage Pollution in Massachusetts Rivers and Beaches Using a Sewage-Specific Marker PCR Assay Targeting a Putative Virulence Factor (esp Gene) in Enterococcus faecium, May 21-25, 106th ed.; General Meeting of the American Society for Microbiology: Orlando, FL, 2006; (abstract). (31) Whitman, R. L.; Gochee, A. V.; Dustman, W. A.; Kennedy, K. J. Use of coliform bacteria in assessing human sewage contamination. Nat. Areas J. 1995, 15, 227–233.

ES800481P