Environ. Sci. Technol. 2009, 43, 4267–4272
Source Tracking Identifies Deer and Geese as Vectors of Human-Infectious Cryptosporidium Genotypes in an Urban/Suburban Watershed KRISTEN L. JELLISON,* AMY E. LYNCH, AND JOSEPH M. ZIEMANN Department of Civil and Environmental Engineering, Lehigh University, Bethlehem, Pennsylvania 18015
Received January 10, 2009. Revised manuscript received April 8, 2009. Accepted April 17, 2009.
This study identified Cryptosporidium genotypes in the Wissahickon watershed from May 2005 to April 2008. We analyzed 129 samples from Wissahickon Creek, 83 effluent samples from wastewater treatment plants (WWTPs), and 240 fecal droppings. Genotyping was based on the hypervariable region of the 18S rRNA gene. Oocysts were detected year-round, independent of wet weather events, in 22% of Wissahickon Creek samples, 5% of WWTP effluents, and 7% of fecal samples. Of the genotypes detected, 67% were human-infectious: 30% C. hominis or C. hominis-like, 12% C. parvum, 14% cervine genotype, 9% skunk genotype, and 1% chipmunk I genotype. Similar genotype profiles were detected in Wissahickon Creek each year, and human-infectious genotypes were detected yearround. Unusual genotypes detected in a deer (a C. hominislike genotype) and geese (C. hominis-like genotypes, C. parvum, and muskrat genotype I) show that these animals are vectors of human-infectious genotypes in this watershed. Results suggest that deer, geese, and WWTPs are appropriate targets for source water protection in the Wissahickon watershed.
water source for Philadelphia, Pennsylvania, from May 2005 to April 2008. Samples from two locations on Wissahickon Creek, three treated wastewater effluents, and animal fecal droppings within the watershed were analyzed regularly. Objectives of the study were to (i) determine the frequency of Cryptosporidium presence in Wissahickon Creek, (ii) determine the genotypes and likely sources of Cryptosporidium in the watershed, and (iii) identify the times of year when oocysts, particularly those genotypes that have been associated with human disease, are prevalent in Wissahickon Creek.
Materials and Methods Water/Wastewater Sampling. Sampling in Wissahickon watershed (Figure 1) occurred from May 2005 to April 2008. Wissahickon Creek is 24.1 miles long and drains about 64 square miles of lower Montgomery and northwest Philadelphia counties before it empties into the Schuylkill River just 0.5 miles upstream of the Philadelphia Water Department’s (PWD) Queen Lane WTP intake. The watershed is urban (lower reaches) and suburban (upper reaches), with major land use characterized as residential (52%), wooded (17%), recreational (7%), and agricultural (6%); the remaining land use is a mix of light industry, commercial development, and parking and transportation (8). Most of the land immediately adjacent to the creek within the city limits is wooded or associated with recreational uses. Approximately 54 million liters per day of treated wastewater are discharged to the upper Wissahickon Creek by five municipalities in the watershed; in fact, most of Wissahickon Creek baseflow consists of treated wastewater (9).
Introduction Cryptosporidium is a protozoan parasite, transmitted via ingestion of fecally contaminated food and water, responsible for a potentially fatal gastrointestinal disease in immunocompromised people. Numerous species/genotypes have been associated with human disease, including C. parvum, C. hominis, C. felis, C. meleagridis, C. suis, C. canis, C. muris, and the chipmunk I, cervine, skunk, horse, and rabbit genotypes (1-3). Currently available medications for the treatment of cryptosporidiosis have not been proven effective for immunocompromised patients (4). Furthermore, removal/ inactivation of Cryptosporidium oocysts in water treatment plants (WTPs) is not completely effective because the oocysts are small enough (4-8 µm) to pass through sand filters under suboptimal coagulation/flocculation regimes (5) and are resistant to chlorine disinfection (6, 7). Understanding the sources and species/genotypes of Cryptosporidium in raw water supplies is necessary to identify public health risk and develop appropriate source water protection plans. This study provides information on Cryptosporidium genotypes detected in the Wissahickon watershed, a drinking * Corresponding author phone: 610-758-3555; fax: 610-758-6405; e-mail:
[email protected]. 10.1021/es900081m CCC: $40.75
Published on Web 05/08/2009
2009 American Chemical Society
FIGURE 1. The Wissahickon watershed. Yellow circles indicate the two surface water (Wiss 140 and Wiss 410) and three WWTP sampling sites. (Figure courtesy of PWD Office of Watersheds). VOL. 43, NO. 12, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4267
FIGURE 2. Frequency of oocyst detection (A) and number of genotypes detected in Wissahickon Creek per season (B). Hatched bars indicate number of total samples analyzed; black bars indicate number of positive samples. Numbers above bars in A denote percent positive samples; numbers inside black bars in B denote number of genotypes detected. Two surface water sites, Wiss 140 and Wiss 410, were sampled biweekly beginning in May 2005. Wiss 410, located just outside of Philadelphia, is heavily influenced by five wastewater treatment plant (WWTP) discharges. Wiss 140 is further downstream, within the city limits, after the creek has flowed through Fairmount Park. WWTP effluents were added to the sampling plan as the project progressed: WWTP1 was sampled monthly from January 2006 to December 2006, and biweekly from January 2007 to April 2008; WWTP2 and WWTP3 were sampled biweekly from February 2007 to April 2008. From May 2005 to January 2008, a third monthly sample at each location was attempted if a wet weather event (defined as a minimum of 0.05 in of rainfall over the preceding 48 h) occurred. A total of 129 surface water samples (64 and 65 at Wiss 410 and Wiss 140, respectively) and 83 WWTP effluent samples (37, 25, and 21 at WWTP1, WWTP2, and WWTP3, respectively) were analyzed (Figure 2, panel A). For each water and WWTP sample collected, sample volume and turbidity (10) were measured. All samples were filtered with Pall Envirochek HV filter capsules (Pall Life Sciences, Ann Arbor, MI) by staff from the PWD Office of Watersheds. Sample volumes (35 L avg. ( 31 L s.d. for surface water samples; 19 L avg. ( 16 L SD for WWTP effluent samples) were variable, depending upon how quickly the filters clogged, and were measured with a FTB4000 turbine meter for water totalization (Omega Engineering, Inc., Stamford, CT). Filters were stored at 1-5 °C for less than 24 h prior to processing. A pre-elution step with 5% (w/v) sodium hexametaphosphate, to remove the excess ferric chloride used in the wastewater treatment process, was performed for the WWTP filters as previously described (11). All filters were subsequently eluted according to standard manufacturer recommendations. Eluted filter pellets were shipped overnight in insulated cartons containing freezer packs to Lehigh University for further processing. Fecal Sampling. Fecal collection within the watershed was performed by a staff biologist from the PWD Office of Watersheds. A total of 240 fecal samples from deer (Odocoileus virginianus, n ) 91), geese (Branta Canadensis, n ) 75), sheep 4268
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 12, 2009
(Ovis aries, n ) 15), horses (Equus caballus, n ) 11), cows (Bos taurus, n ) 11), calves (Bos taurus, n ) 8), ducks (Anas platyrhyncha, n ) 8), donkeys (Equus asinus, n ) 6), dogs (Canis lupus familiariz, n ) 6), raccoons (Procyon lotor, n ) 2), opossums (Didelphis virginiana, n ) 2), groundhogs (Marmota monax, n ) 2), a skunk (Mephitis mephitis, n ) 1), and a cat (Felis catus, n ) 1) were analyzed (one additional fecal sample from an unknown source was also analyzed). When possible, fecal samples were collected immediately following visual identification of host excretion; when visual identification of host excretion was not possible, the freshest available fecal samples were collected from the ground and verified using a field guide (12). All fecal samples were handled with disposable latex gloves and sterile polystyrene spatulas; to avoid cross-contamination, gloves were changed between samples, and a new sterile spatula was used for each sample. Samples were placed in individual, labeled Whirl-pack bags in a cooler with ice (approximately 4 °C) and transported to the PWD laboratory for shipping. Fecal samples in the original Whirl-pack bags were shipped overnight in insulated cartons containing freezer packs to Lehigh University for further processing. Immunomagnetic Separation. All water and fecal samples were processed the same day that they were received by Lehigh University. Eluted water pellets were resuspended in 5 mL of Millipore water (Milli-Q Biocel System; Millipore Corporation, Bedford, MA) for each 1 mL of solids. A 2 g sample of each fecal pellet was suspended in 40 mL of Millipore water and vortexed for 30 s to create a fecal slurry. The fecal slurry was settled for 1 min to remove large particles, and the top 20 mL of slurry were transferred to a new centrifuge tube. The slurry was centrifuged (15 min at 1100g) and the supernatant reduced to 5 mL of water per mL of solids. Oocysts were purified from water pellets and fecal slurries by immunomagnetic separation (IMS) with the Aureon IMS kit (ImmTech, Inc., New Windsor, MD) according to the manufacturer’s recommendations. If the total volume of water or fecal solids exceeded 1 mL, the sample was split into multiple IMS tests to ensure that the entire sample was analyzed. Oocysts were dissociated from the magnetic beads using 0.05 M HCl; if multiple IMS tests were performed for a given sample, the samples were recombined after the oocyst dissociation step. Oocyst suspensions were neutralized with 0.5 M NaOH, pelleted for 3 min at 13 000g, and resuspended in 50 µL of Millipore water. Positive IMS controls consisted of 4.5 mL of Millipore water spiked with 500 mL of a 104 oocyst per mL suspension; negative IMS controls consisted of 5 mL of Millipore water. IMS controls were processed with each set of samples. DNA Extraction. Oocyst DNA was extracted from IMS products with phenol chloroform and precipitated with ethanol (13). Positive and negative DNA extraction controls were included with each set of field samples (13). Polymerase Chain Reaction. Nested PCR amplification of a hypervariable region of the 18S rRNA gene (approximately 434-bp in length) was performed (14). Positive and negative PCR controls were included with each set of water or fecal samples (14). Cloning and Sequencing. Secondary PCR products positive for Cryptosporidium were cloned into the pGEM-T Easy Vector System (Promega Corporation, Madison, WI) and used to transform z-competent DH5R E. coli cells (Zymo Research, Orange, CA) (14). Ten to 12 clones per PCR reaction were digested with NdeI to identify any heterogeneity among the clones (13, 14). Representative clones of the secondary PCR products were shipped overnight to the University of Pennsylvania DNA sequencing facility (Philadelphia, PA). If multiple NdeI digestion patterns existed among clones from a given sample, three clones of each digestion pattern were
sequenced (in some cases this was not possible because only one or two clones had a differing digestion pattern). At least three clones for each positive sample were sequenced, and data were confirmed by sequencing both strands of each clone. When multiple clones from a single sample were sequenced with less than 1% difference, the consensus sequence was used in the phylogenetic analysis (to avoid the inclusion of PCR errors in the analysis). All sequences derived from this work were submitted to GenBank (15) under accession numbers FJ607871-FJ607946. Phylogenetic Analysis. Sequences were aligned manually using MacClade 4.06 (Sinauer Associates, Inc., Sunderland, MA). Secondary structures were the basis for aligning the two variable-length helices (spanning nucleotides 620-670 and 674-703 of GenBank no. AF093490) in the hypervariable region and were determined using the RNAfold web server hosted by the Institute for Theoretical Chemistry at the University of Vienna, Austria. MEGA4 (16) was used to create a neighbor-joining tree; evolutionary distances were calculated by the Kimura two-parameter analysis. All positions containing alignment gaps were eliminated only in pairwise sequence comparisons. Statistical support for the resulting tree was tested using 1000 pseudoreplicates of the bootstrap test; only values above 50% were reported. Statistics. The Shapiro-Wilk test was performed to determine whether the turbidity and volume data sets (for positive, negative, wet, and dry samples) were normally distributed. With the exception of positive sample volumes (p ) 0.14), all data sets were non-normally distributed (p e 0.005); therefore, nonparametric statistical analyses were performed to investigate differences and correlations among the data. The Wilcoxon-Mann-Whitney test was used to evaluate differences in median sample volumes and turbidities for (i) positive versus negative samples and (ii) wet versus dry samples, as well as differences in median rainfall on positive versus negative sampling days. Spearman correlation was used to test the degree of correlation between sample volume and turbidity. The Pearson chi-square test was used to evaluate differences in the proportion of (i) positive versus negative samples and (ii) potentially infectious versus noninfectious genotypes collected in spring (March 20 to June 20), summer (June 21 to September 21), fall (September 22 to December 20), and winter (December 21 to March 19) months. The Fisher exact test was used to evaluate a difference in the proportion of (i) positive versus negative samples on wet versus dry days, (ii) C. hominis/C. parvum versus Cryptosporidium sp. wildlife genotypes on wet versus dry days, and (iii) potentially infectious versus noninfectious genotypes detected at Wiss 410 versus Wiss 140. All statistical analyses were performed with the Analyzeit add-in (Analyze-it Software, Ltd., Leeds, England) for Microsoft Excel.
Results and Discussion A 10-month study in 1995 confirmed the presence of Cryptosporidium in 1 of 10 source water samples from the Queen Lane WTP; no protozoa were detected in any finished water samples (17). Additional data from the PWD Bureau of Laboratory Services confirmed the presence of Cryptosporidium in Wissahickon Creek at concentrations ranging from 10 to 20 oocysts per 100 L in 7 of 14 samples analyzed from May 2002 to October 2004 (18). Results of initial Cryptosporidium monitoring, as required by the Long-Term 2 Enhanced Surface Water Treatment Rule, placed the Queen Lane WTP in Bin 2 (19). The PWD has taken a proactive approach to improve source water quality by supporting the current Cryptosporidium genotyping and source tracking study in Wissahickon Creek. In the current study, oocysts were detected in 22% (28 of 129) of Wissahickon Creek samples. Annual frequency of
oocyst detection was consistent at Wiss 140 and Wiss 410 (oocysts were detected in 18.2-23.8% and 20.0-27.3% of samples at Wiss 140 and Wiss 410, respectively) but more variable in WWTP effluents, ranging from 0 to 9.5% (Figure 2, panel A). No apparent seasonal trend of oocyst detection was observed, and no difference (p ) 0.50) was observed between the proportion of positive creek samples collected in the spring, summer, fall, or winter months (Figure 2, panel B). The data do not show an impact of sample volume or turbidity on oocyst detection. Moderately negative Spearman correlation coefficients (rs ) -0.57 and -0.63) were observed between turbidity and sample volume of Wissahickon Creek and WWTP samples, respectively. There was no difference between the median turbidity of positive (3.30 ntu, n ) 27) versus negative (2.97 ntu, n ) 100) Wissahickon Creek samples (p ) 0.60) or positive (5.76 ntu, n ) 4) versus negative (3.14 ntu, n ) 79) WWTP samples (p ) 0.28). The median analyzed volume of negative Wissahickon Creek samples (34.2 L, n ) 90) was higher (p ) 0.02) than that of positive samples (22.2 L, n ) 28), but there was no difference between the median analyzed volume of positive (15.9 L, n ) 4) versus negative (15.1 L, n ) 79) WWTP samples (p ) 0.88). Wiss 140 and Wiss 410 were sampled on 53 dry days and 14 wet days. The median turbidity of Wiss 140 and Wiss 410 samples was higher (p < 0.0001) on wet days (20.2 ntu, n ) 25) than dry days (2.5 ntu, n ) 102). It follows that the median sample volume collected and analyzed was higher (p < 0.0001) on dry days (35.5 L, n ) 100) than wet days (8.0 L, n ) 18). Samples collected on wet days (n ) 26) were not any more or less likely (p ) 0.63) to be positive for Cryptosporidium than samples collected on dry days (n ) 103), although perhaps a difference in oocyst detection on wet and dry days would be observed if the sample volumes on wet and dry days were comparable. There was also no difference in the genotype profiles detected on wet versus dry days: the proportion of C. hominis and C. parvum versus Cryptosporidium wildlife genotypes detected on wet and dry days was not statistically different (p ) 0.27) and, therefore, wet and dry samples are not differentiated in the following data analysis. Although the average cumulative rainfall over the 48 h prior to sampling was higher for positive (0.68 in, n ) 28) than negative (0.38 in, n ) 101) creek samples, the median rainfalls of positive versus negative samples were not statistically different (p ) 0.14). The lack of relationship between wet weather events and Cryptosporidium detection is in contrast to previous studies which have shown a correlation between rainfall and waterborne disease outbreaks in general (20) and rainfall and Cryptosporidium detection in particular (21-26). The lack of relationship observed in this study may be attributable to the fact that there were not an equivalent number of wet and dry sample dates (the majority of sample dates were dry days) as well as the difference in sample volumes analyzed on wet versus dry days. Of the 76 Cryptosporidium sequences detected from creek, WWTP, and fecal samples (Figure 3), 51 (67%) were identified as human-infectious genotypes (1, 3) and thus indicate a potential public health risk: 23 (30%) were C. hominis or C. hominis-like, 9 (12%) were C. parvum, 11 (14%) were the cervine genotype, 7 (9%) were the skunk genotype, and 1 (1%) was the chipmunk genotype I. Human-infectious genotypes comprised 65% (30 of 46), 88% (7 of 8), and 64% (14 of 22) of the Cryptosporidium sequences detected in Wissahickon Creek, WWTP, and fecal samples, respectively. In addition, multiple genotypes were detected in 46% (13 of 28), 75% (3 of 4), and 24% (4 of 17) of positive Wissahickon Creek, WWTP, and fecal samples, respectively. Detection of multiple genotypes has also been reported in other studies (e.g., 44 and 70% of positive stormwater samples (27, 28) and 10% of raw wastewater samples (29) had mixed genotypes) and shows that (i) multiple oocyst VOL. 43, NO. 12, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4269
FIGURE 3. Neighbor-joining phylogenetic tree based on the hypervariable region of the 18S rRNA gene for Wissahickon Creek, WWTP, and fecal samples collected May 2005 to April 2008 (GenBank accession numbers FJ607871-FJ607946) and reference sequences from GenBank (identified by Cryptosporidium species/genotype and accession number). Creek and WWTP samples are identified by the location and date of collection; if multiple genotypes were detected, they are distinguished by a number following the date (e.g., Wiss 410 (1/3/06) #3 ) the 3rd distinct genotype detected at Wiss 410 on 1/3/06). Fecal samples are identified by the animal host and date of collection. Bootstrap values greater than 50% are indicated at each respective node. sources can impact a single sample location and (ii) a single oocyst source (WWTP effluent or animal host) may release multiple oocyst genotypes into the environment. 4270
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 12, 2009
Similar genotype profiles were detected in Wissahickon Creek each year: approximately 1/3 (28, 34, and 25% in years 1, 2, and 3, respectively) of genotypes detected were the
human pathogens C. hominis (or C. hominis-like) or C. parvum, 1/3 (33, 33, and 44% in years 1, 2, and 3, respectively) were other genotypes associated with human disease (i.e., cervine, skunk, or chipmunk I genotypes), and 1/3 (39, 33, and 31% in years 1, 2, and 3, respectively) were other Cryptosporidium genotypes not associated with human disease (i.e., duck, snake, rat, muskrat, deer mouse III, squirrel, and W19 genotypes; and a novel sequence most closely related to the skunk genotype). The genotypes detected in Wissahickon Creek indicate human and wildlife sources of Cryptosporidium, which makes sense for this urban/suburban watershed with residential land use, wooded/recreational areas, and five WWTP discharges. Slightly different genotype profiles were reported in several recent watershed studies, with C. andersoni being detected commonly (30) and, in some cases, most frequently (22, 31); however, those studies were performed in watersheds where agriculture was a predominant land use, and agriculture is a very small fraction of land use in the Wissahickon watershed. Similar to our findings, Yang et al. (31) reported no C. andersoni at a site influenced by urban wastewater. Genotypes associated with human illness made up a higher percentage of the genotypes detected at Wiss 410 (74%) than Wiss 140 (57%), and although this difference was not statistically significant (p ) 0.35), the data suggest that public health risk may be attenuated as water flows through the lower watershed (which does not have wastewater discharges and is contained within parkland). For the three combined years of sampling at Wiss 410, 35% of genotypes detected were the human pathogens C. hominis (or C. hominis-like) or C. parvum, 39% were other genotypes associated with human disease (i.e., cervine or skunk genotypes), and 26% were other Cryptosporidium genotypes not associated with human disease (i.e., duck, deer mouse III, and a novel sequence most closely related to the skunk genotype). For the three combined years of sampling at Wiss 140, 22% of genotypes detected were the human pathogens C. hominis (or C. hominis-like) or C. parvum, 35% were other genotypes associated with human disease (i.e., cervine, skunk, or chipmunk I genotypes), and 43% were other Cryptosporidium genotypes not associated with human disease (i.e., duck, snake, rat, muskrat, deer mouse III, squirrel, and W19 genotypes). Previous studies (27, 28, 31) have also found that the distribution of Cryptosporidium genotypes can vary between different sample locations within the same watershed and even the same brook; these differences have been attributed to different land uses in different parts of the watershed. The differences between Wiss 410 and Wiss 140 make sense in the context of land use within the watershed; the five WWTPs discharge upstream of Wiss 410, while most of the land immediately adjacent to the creek upstream of Wiss 140 is wooded and associated with wildlife and recreational uses. Seasonal detection of oocyst genotypes in Wissahickon Creek was observed (Figure 4). C. hominis or C. hominis-like genotypes were detected year-round, but C. parvum was only detected in the spring. The cervine and skunk genotypes were detected in the fall, winter, and spring, and chipmunk genotype I was detected in the fall only. Genotypes associated with human disease were detected in Wissahickon Creek year-round, and although not statistically significant, the proportion of potentially infectious Cryptosporidium genotypes exceeded (p ) 0.97) the number of noninfectious genotypes detected for every season. Cryptosporidium was detected in 17 (7%) of the fecal samples, including nine geese, two deer, three sheep, one calf, one raccoon, and the unknown fecal sample. Multiple Cryptosporidium genotypes were detected in a single deer or goose fecal sample (Figure 3), and several unusual genotypes were detected: C. hominis-like genotypes were found in one deer (on 10/16/05) and six geese (four on 8/20/ 07; two on 9/25/07), C. parvum was detected in three geese
FIGURE 4. Seasonal Cryptosporidium genotype detection in Wissahickon Creek. Hatched columns represent the number of potentially infectious genotypes detected in each season over 3 years; black columns represent the number of noninfectious genotypes detected in each season over 3 years. (all on 8/20/07), and the Cryptosporidium muskrat genotype I was detected in one goose (on 3/17/08). In contrast, Feng et al. (32) detected Cryptosporidium in 20.5% (111 of 541) of wildlife fecal samples in New York and did not detect any C. hominis or C. parvum in wildlife. The higher oocyst prevalence reported by Feng et al. likely indicates true host-adapted Cryptosporidium infections in New York wildlife compared to the wildlife in Wissahickon Creek. It is likely that deer and geese are simply mechanical vectors of the unusual genotypes found in this study, rather than truly infected hosts. Previous studies have reported a low (10.2%) prevalence of C. parvum and C. hominis in geese (33) and have shown that geese can disseminate infectious C. parvum oocysts in the environment (34, 35). Crosstransmission of oocysts between multiple hosts in a watershed is very likely, given the proximity of deer and geese to areas inhabited by humans, domesticated animals, and other wildlife. The potential for wildlife to serve as vectors of human-infectious oocysts means that identification and control of these vectors may be as important as the identification and control of the original oocyst source for watershed protection. Most of the genotypes detected in the deer and geese fecal samples were closely related to genotypes identified in Wissahickon Creek (Figure 3), highlighting the importance of matching genotypes in water with genotypes from animals or other point sources in the watershed to more accurately identify both the sources and vectors of waterborne oocysts which should be targeted for watershed control. Only 4 (5%) of 83 WWTP effluent samples were positive for Cryptosporidium, which is lower than the 25-28% oocyst prevalence reported by others in raw wastewater samples (29, 30). While the wastewater treatment process may reduce the prevalence of oocysts in treated effluent, it is likely that oocyst prevalence in WWTP effluents exceeded our observations and that oocyst detection limits in WWTP effluents were not as sensitive as those in creek water (all WWTP effluent samples were processed according to McCuin and Clancy (11), who reported a 9.3% ( 0.9% mean oocyst recovery from secondary wastewater effluents using the Aureon IMS kit). C. parvum, C. hominis, and the squirrel genotype were detected in the WWTP1 effluent. C. hominis, a C. hominis-like genotype, and the Cryptosporidium cervine genotype were detected in the WWTP2 effluent. All of the genotypes detected in WWTP effluents were closely related to genotypes identified in Wissahickon Creek (Figure 3), suggesting that WWTPs are also a source of Cryptosporidium oocysts in Wissahickon Creek. Although the percent of fecal and WWTP samples which tested positive for Cryptosporidium were similar (7% and 5%, respectively), the fact that VOL. 43, NO. 12, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4271
WWTP effluents are a continuous discharge to the creek that can make up 60% or more of the flow suggests that WWTP effluents pose a more significant threat to oocyst contamination in Wissahickon Creek. Results of this study show that Cryptosporidium oocysts are present in the Wissahickon watershed year-round, independent of wet weather events, and that WWTP effluents, geese, and deer are all reasonable targets for watershed management efforts to reduce the risk of waterborne exposure to potentially infectious oocysts. The consistency of oocyst detection and genotyping results among all three years indicates that the genotyping methods used in this study are robust enough to repeatedly detect the most prevalent genotypes in complex, multiuse watersheds. Results from this study also have important implications for other watersheds, showing that oocyst source tracking which depends solely on identification of host-adapted genotypes from surface water may overlook the importance of vector organisms which transport humaninfectious oocysts from host to water. A more robust approach to oocyst source tracking should include comparison and matching of Cryptosporidium genotypes in water, animals, and other potential point sources in a watershed.
Acknowledgments We acknowledge the following Philadelphia Water Department personnel for their assistance: Geoffrey L. Brock and Gary A. Burlingame for project management; Ken Sarkis, Ivanna Szpilczak, Cindy Rettig, and Anne Harvey for elution of water filters; and Kelly Anderson and Phil Duzinski for field sample collection and watershed information. This work was financed in part by the Philadelphia Water Department and several grants from the Commonwealth of Pennsylvania, Department of Community and Economic Development, through the Pennsylvania Infrastructure Technology Alliance.
Literature Cited (1) Nichols, G. In Cryptosporidium and Cryptosporidiosis, 2nd ed.; Xiao, L., Ed.; CRC Press: Boca Raton, FL, 2008; p 560. (2) Robinson, G.; Elwin, K.; Chalmers, R. M. Unusual Cryptosporidium genotypes in human cases of diarrhea. Emerging Infect. Dis. 2008, 14, 1800–1802. (3) Feltus, D. C.; Giddings, C. W.; Schneck, B. L.; Monson, T.; Warshauer, D.; McEvoy, J. M. Evidence supporting zoonotic transmission of Cryptosporidium spp. in Wisconsin. J. Clin. Microbiol. 2006, 44, 4303–4308. (4) Abubakar, I.; Aliyu, S. H.; Arumugam, C.; Usman, N. K.; Hunter, P. R. Treatment of cryptosporidiosis in immunocompromised individuals: Systematic review and meta-analysis. Br. J. Clin. Pharmacol. 2007, 63, 387–393. (5) Edzwald, J.; Kelley, M. Control of C. parvum: from reservoirs to clarifiers to filters. Water Sci. Technol. 1998, 37, 1–8. (6) Korich, D. G.; Mead, J. R.; Madore, M. S.; Sinclair, N. A.; Sterling, C. R. Effects of ozone, chlorine dioxide, chlorine, and monochloramine on Cryptosporidium parvum oocyst viability. Appl. Environ. Microbiol. 1990, 56, 1423–1428. (7) Finch, G. R.; Black, E. K.; Gyurek, L.; Belosevic, M. Ozone inactivation of C. parvum in demand-free phosphate buffer determined by in-vitro excystation and animal infectivity. Appl. Environ. Microbiol. 1993, 59, 4203–4210. (8) Personal Communication to K. Jellison from the Office of Watersheds, Philadelphia Water Department, 2008. (9) Crockett, C. S.; Haas, C. N. Understanding protozoa in your watershed. J. Am. Water Works Assoc. 1997, 89, 62–73. (10) Standard Methods for the Examination of Water and Wastewater, 20th ed.; Eaton, A. D., Ed.; American Public Health Association: Washington, DC, 1998. (11) McCuin, R. M.; Clancy, J. L. Methods for the recovery, isolation, and detection of Cryptosporidium oocysts in wastewaters. J. Microbiol. Methods 2005, 63, 73–88. (12) Halfpenny, J.; Bruchac, J. Scats and Tracks of the Mid-Atlantic: A Field Guide to the Signs of Seventy Wildlife Species; The Globe Pequot Press: Guilford, CT, 2006. (13) Jellison, K. L.; Hemond, H. F.; Schauer, D. B. Sources and species of Cryptosporidium oocysts in the Wachusett Reservoir watershed. Appl. Environ. Microbiol. 2002, 68, 569–575.
4272
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 12, 2009
(14) Jellison, K. L.; Distel, D. L.; Hemond, H. F.; Schauer, D. B. Phylogenetic analysis of the hypervariable region of the 18S rRNA gene of Cryptosporidium oocysts in feces of Canada geese (Branta canadensis): Evidence for five novel genotypes. Appl. Environ. Microbiol. 2004, 70, 452–458. (15) Benson, D. A.; Karsch-Mizrachi, I.; Lipman, D. J.; Ostell, J.; Wheeler, D. L. GenBank. Nucleic Acids Res. 2008, 36, D25-30. (16) Tamura, K.; Dudley, J.; Nei, M.; Kumar, S. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol. Biol. Evol. 2007, 24, 1596–1599. (17) Crockett, C. S.; Haas, C. N. Understanding the occurrence and behavior and Giardia and Cryptosporidium in the Philadelphia watershed; Drexel University: Philadelphia, 1995. (18) Brock, G. L., Personal Communication to K. Jellison from the Bureau of Laboratory Services, Philadelphia Water Department, 2004. (19) Burlingame, G. A. Addressing emerging pathogens: Philadelphia’s reflections on Giardia and Cryptosporidium. J. New Engl. Water Works Assoc. 2008, 122, 117–129. (20) Curriero, F. C.; Patz, J. A.; Rose, J. B.; Lele, S. The association between extreme precipitation and waterborne disease outbreaks in the United States, 1948-1994. Am. J. Public Health 2001, 91, 1194–1199. (21) Keeley, A.; Faulkner, B. R. Influence of land use and watershed characteristics on protozoa contamination in a potential drinking water resources reservoir. Water Res. 2008, 42, 2803–2813. (22) Ruecker, N. J.; Braithwaite, S. L.; Topp, E.; Edge, T.; Lapen, D. R.; Wilkes, G.; Robertson, W.; Medeiros, D.; Sensen, C. W.; Neumann, N. F. Tracking host sources of Cryptosporidium spp. in raw waste for improved health risk assessment. Appl. Environ. Microbiol. 2007, 73, 3945–3957. (23) Montemayor, M.; Valero, F.; Jofre, J.; Lucena, F. Occurrence of Cryptosporidium spp. oocysts in raw and treated sewage and river water in north-eastern Spain. J. Appl. Microbiol. 2005, 99, 1455–1462. (24) Fayer, R.; Trout, J. M.; Lewis, E. J.; Xiao, L.; Lal, A.; Jenkins, M. C.; Graczyk, T. K. Temporal variability of Cryptosporidium in the Chesapeake Bay. Parasitol Res 2002, 88, 998–1003. (25) Kistemann, T.; Classen, T.; Koch, C.; Dangendorf, F.; Fischeder, R.; Gebel, J.; Vacata, V.; Exner, M. Microbial load of drinking water reservoir tributaries during extreme rainfall and runoff. Appl. Environ. Microbiol. 2002, 68, 2188–2197. (26) Hansen, J. S.; Ongerth, J. E. Effects of time and watershed characteristics on the concentration of Cryptosporidium oocysts in river water. Appl. Environ. Microbiol. 1991, 57, 2790–2795. (27) Xiao, L.; Alderisio, K.; Limor, J.; Royer, M.; Lal, A. A. Identification of species and sources of Cryptosporidium oocysts in storm waters with a small-subunit rRNA-based diagnostic and genotyping tool. Appl. Environ. Microbiol. 2000, 66, 5492–5498. (28) Jiang, J.; Alderisio, K. A.; Xiao, L. Distribution of Cryptosporidium genotypes in storm event water samples from three watersheds in New York. Appl. Environ. Microbiol. 2005, 71, 4446–4454. (29) Zhou, L.; Singh, A.; Jiang, J.; Xiao, L. Molecular surveillance of Cryptosporidium spp. in raw wastewater in Milwaukee: implications for understanding outbreak occurrence and transmission dynamics. J. Clin. Microbiol. 2003, 41, 5254–5257. (30) Xiao, L.; Singh, A.; Limor, J.; Graczyk, T. K.; Gradus, S.; Lal, A. Molecular characterization of Cryptosporidium oocysts in samples of raw surface water and wastewater. Appl. Environ. Microbiol. 2001, 67, 1097–1101. (31) Yang, W.; Chen, P.; Villegas, E. N.; Landy, R. B.; Kanetsky, C.; Cama, V.; Dearen, T.; Schultz, C. L.; Orndorff, K. G.; Prelewicz, G. J.; Brown, M. H.; Young, K. R.; Xiao, L. Cryptosporidium source tracking in the Potomac River watershed. Appl. Environ. Microbiol. 2008, 74. (32) Feng, Y.; Alderisio, K. A.; Yang, W.; Blancero, L. A.; Kuhne, W. G.; Nadareski, C. A.; Reid, M.; Xiao, L. Cryptosporidium genotypes in wildlife from a New York watershed. Appl. Environ. Microbiol. 2007, 73, 6475–6483. (33) Zhou, L.; Kassa, H.; Tischler, M. L.; Xiao, L. Host-adapted Cryptosporidium spp. in Canada geese (Branta canadensis). Appl. Environ. Microbiol. 2004, 70, 4211–4215. (34) Graczyk, T. K.; Cranfield, M. R.; Fayer, R.; Trout, J.; Goodale, H. J. Infectivity of Cryptosporidium parvum oocysts is retained upon intestinal passage through a migratory water-fowl species (Canada goose, Branta canadensis). Trop. Med. Int. Health 1997, 2, 341–347. (35) Graczyk, T. K.; Fayer, R.; Trout, J. M.; Lewis, E. J.; Farley, C. A.; Sulaiman, I.; Lal, A. A. Giardia sp. cysts and Cryptosporidium parvum oocysts in the feces of migratory Canada geese (Branta canadensis). Appl. Environ. Microbiol. 1998, 64, 2736–2738.
ES900081M
