Cryptosporidium Incidence and Surface Water Influence of

2 days ago - Despite its painful-sounding name, the stimulator of interferon genes protein, known as STING, is something... BUSINESS CONCENTRATES ...
0 downloads 0 Views 425KB Size
Article Cite This: Environ. Sci. Technol. XXXX, XXX, XXX−XXX

pubs.acs.org/est

Cryptosporidium Incidence and Surface Water Influence of Groundwater Supplying Public Water Systems in Minnesota, USA Joel P. Stokdyk,† Susan K. Spencer,‡ James F. Walsh,§ Jane R. de Lambert,§ Aaron D. Firnstahl,† Anita C. Anderson,§ Lih-in W. Rezania,§ and Mark A. Borchardt*,‡ †

USGS Upper Midwest Water Science Center, 2615 Yellowstone Drive, Marshfield, Wisconsin 54449, United States USDA-Agricultural Research Service, Environmentally Integrated Dairy Management Research Unit, 2615 Yellowstone Drive, Marshfield, Wisconsin 54449, United States § Minnesota Department of Health, 625 Robert St. N, St. Paul, Minnesota 55164, United States Environ. Sci. Technol. Downloaded from pubs.acs.org by UNIV OF NEW ENGLAND on 03/22/19. For personal use only.



S Supporting Information *

ABSTRACT: Regulations for public water systems (PWS) in the U.S. consider Cryptosporidium a microbial contaminant of surface water supplies. Groundwater is assumed free of Cryptosporidium unless surface water is entering supply wells. We determined the incidence of Cryptosporidium in PWS wells varying in surface water influence. Community and noncommunity PWS wells (n = 145) were sampled (n = 964) and analyzed for Cryptosporidium by qPCR and immunofluorescence assay (IFA). Surface water influence was assessed by stable isotopes and the expert judgment of hydrogeologists using site-specific data. Fifty-eight wells (40%) and 107 samples (11%) were Cryptosporidiumpositive by qPCR, and of these samples 67 were positive by IFA. Cryptosporidium concentrations measured by qPCR and IFA were significantly correlated (p < 0.001). Cryptosporidium incidence was not associated with surface water influence as assessed by stable isotopes or expert judgment. We successfully sequenced 45 of the 107 positive samples to identify species, including C. parvum (41), C. andersoni (2), and C. hominis (2), and the predominant subtype was C. parvum IIa A17G2R1. Assuming USA regulations for surface water-supplied PWS were applicable to the study wells, wells positive for Cryptosporidium by IFA would likely be required to add treatment. Cryptosporidium is not uncommon in groundwater, even when surface water influence is absent.



INTRODUCTION Cryptosporidium is generally considered a surface water contaminant. This view originated when surface water contaminated with the parasite caused the largest waterborne disease outbreak in the United States in Milwaukee, Wisconsin in 1993.1 In response to the outbreak, the U.S. Environmental Protection Agency (USEPA) established a regulatory paradigm that classifies surface water as vulnerable and groundwater as protected unless proven otherwise, which continues to reinforce Cryptosporidium’s status as a surface water contaminant. Public water systems with a surface water source require monitoring and treatment targeting Cryptosporidium, and the definition of groundwater under the direct influence of surface water (GWUDI) was modified to include Cryptosporidium2−5 (see Supporting Information (SI) for the full GWUDI definition.) While the regulatory focus regarding Cryptosporidium has been on surface water, groundwater has been generally overlooked because oocysts are assumed sufficiently large to be removed by the filtering capacity of overlying soil and subsurface sediments.6 Defining a well as GWUDI by the presence of Cryptosporidium, ipso facto, shows the predominant view is that Cryptosporidium is a surface water issue. However, there are few broad, representative assessments of Cryptosporidium in groundwater. Most studies have focused on © XXXX American Chemical Society

wells believed to be vulnerable to contamination, like those in karst aquifers, near surface water, or near contamination sources.7−10 These four studies included 3−26 wells each, and occurrence of Cryptosporidium ranged 0−15%. Studies have also focused on wells in developing regions that may be more vulnerable to contamination due to rudimentary construction or poor hygiene and sanitation conditions.11−13 Cryptosporidium incidence as high as 80% (8 of 10 wells) was reported in these studies, but this may not represent general groundwater conditions in developed countries. Likewise, studies with a limited scope of wells, like those that sampled wells during an outbreak investigation, following a natural disaster, or characterizing a single location (e.g., refs 14−17) may not represent typical groundwater conditions. Hancock et al.18 tested 160 wells in the United States and found that 8% were positive for Cryptosporidium; this is the only broad occurrence study completed in a developed country. Over 140 000 public water systems in the U.S. rely on groundwater,19 and the vulnerability of these systems to Cryptosporidium contamiReceived: September 26, 2018 Revised: January 30, 2019 Accepted: February 19, 2019

A

DOI: 10.1021/acs.est.8b05446 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

for an intended 12 samples. Overall, 964 samples were collected, 529 from community and 435 from noncommunity systems. Only four wells were sampled once; 122 wells were sampled six or more times. Well and Study Area Descriptions. Study wells represented the major aquifer types in Minnesota in proportions similar to public wells statewide (Table 1). Glacial

nation is largely unknown because its presence in groundwater is understudied. In addition, while treatment to remove or inactivate Cryptosporidium is required for all surface water-supplied drinking water in the U.S.2,3 and the level of treatment required depends on the Cryptosporidium concentration in the source,5 treatment for Cryptosporidium is not required for systems using groundwater that is not under the direct influence of surface water. Even when a groundwater-supplied system does disinfect, it usually employs chlorination, which is not effective against the parasite,20 leaving water consumers potentially exposed to infectious oocysts and the risk of illness. While the USEPA sets treatment rules for surface watersupplied systems (including GWUDI), states have discretion in establishing GWUDI criteria. 4 In Minnesota, GWUDI determination is based on an assessment of well construction characteristics and geology, a field evaluation (including fecal source assessment), and water quality measurements that may include total coliform, E. coli, stable isotopes of water, chloride, bromide, tritium, water temperature, and specific conductance. For example, in areas with shallow soil overlying fractured bedrock, systems with a confirmed E. coli incident at the source are considered GWUDI so as to be more protective of public health. In addition to the GWUDI criteria used for regulatory classification, the State has developed a more comprehensive assessment of surface water influence to inform management of its wells. This more inclusive assessment, which is also consistent with USEPA’s guidance on GWUDI, is the definition of surface water influence used in this study. Given states’ latitude in defining GWUDI and that few representative studies exist, the prevalence of Cryptosporidium in groundwater warrants further examination. Our objective was to determine the incidence of Cryptosporidium in public water system wells across the state of Minnesota, U.S. In Minnesota, 96% of community public water systems rely on groundwater, including 27% that do not disinfect, and 99% of noncommunity systems use groundwater sources, including 98% that do not disinfect.21 Like the rest of the U.S., the prevalence of Cryptosporidium in these wells is unknown. We tested wells that varied in geologic setting, construction characteristics, and surface water influence. We measured Cryptosporidium using two methods, quantitative polymerase chain reaction (qPCR) and immunofluorescent assay (IFA). Cryptosporidium-positive samples were further analyzed to identify the species and subtype.

Table 1. Aquifer Types and Geologic Sensitivity for Study Wells Compared to Public Wells Statewide in Minnesota % study wells (n = 145)

% public wells statewide (n = 6640)a

sand and gravel sandstone fractured crystalline rock carbonate rocks mixed sandstone, carbonate rocks, shale

57 23 9

64 13 5

6 5

3 15

very high (hours− months) high (weeks− years) moderate (years− decades) low (decades−a century) very low (>a century)

22

14

13

12

23

15

23

29

19

29

category aquifer type

geologic sensitivity (estimated travel time)

a

Numbers are approximate and only include data for aquifers with two or more wells. Only wells with sufficient information to determine geologic sensitivity (n = 3632) are included in the geologic sensitivity percentages.

sand and gravel aquifers deposited during the Quaternary Period are most common (57% of study wells and 64% of wells statewide). Study wells were completed in aquifers that span a range of geologic sensitivities (Table 1), which represent an assessment of the risk that surface contaminants will reach groundwater. Sensitivities are determined based on the permeability and thickness of the overlying material and are expressed as estimates for the vertical time of travel for water or contaminants to reach the aquifer from the land surface.22 The predominant land use in the one-year time-of-travel groundwater catchment area of study wells was developed land (n = 87), natural areas (n = 39), or agriculture (n = 17), with two wells lacking land use information. Construction characteristics of the study wells are reported in Table 2. Sample Collection and Analysis. Cryptosporidium was concentrated from 140−1783 L (mean, 728 L) of groundwater



MATERIALS AND METHODS Study Design and Well Selection. Samples were collected over two years (May, 2014 through May, 2016) from 145 wells supplying community (n = 88) and noncommunity (n = 57) public water systems across Minnesota. Wells sampled in year one of the study (n = 89) were randomly selected from all year-round nondisinfecting community and nontransient noncommunity systems. Selection for year two (n = 85) was targeted toward wells potentially vulnerable to contamination based on geology, well construction, chemical and stable isotope data, and groundwater flow modeling to ensure such wells were represented. The combination of random and targeted well selection achieved a sample of wells that represented public water systems in the state in terms of aquifers and geology. Wells were sampled every other month for one of the two study years (i.e., sampled six times each), except 29 wells were sampled over both years

Table 2. Summary Statistics for Study Well Characteristics mean depth of well (m) depth cased (m) open/screen interval length (m) depth to bedrock (m)a year drilled

median minimum

maximum missing

64 47 20

46 37 9

6 5 0

192 168 121

0 4 2

48 1985

39 1989

1 1907

158 2015

0 22

For wells with depth-to-bedrock reported as “>” some value (n = 20), the minimum value was used.

a

B

DOI: 10.1021/acs.est.8b05446 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology by dead-end ultrafiltration23 using Hemodialyzer Rexeed-25s filters (Asahi Kasei Medical MT Corp., Oita, Japan). Many reports demonstrate the successful concentration of Cryptosporidium by ultrafiltration (e.g., refs 23−25), which is used with USEPA Method 1623 and outperformed other filters.24,26 Samples were collected at the wellhead prior to treatment or disinfection, shipped on ice, and processed within 48 h. Filter elution, sample concentration, and nucleic acid extraction are described in the Supporting Information. Samples for total organic carbon, nitrate/nitrite, ammonia, bromide, chloride, boron, total coliforms, and the stable isotopes of water, oxygen-18 and deuterium, were collected following Cryptosporidium sample collection and were shipped or transported on ice. Some samples were not collected on the same day as Cryptosporidium sample collection to comply with laboratory hold time requirements. Total coliform samples were analyzed within 30 h of sample collection. qPCR. Samples were tested for Cryptosporidium by qPCR targeting the 18S gene27 using a LightCycler 480 instrument (Roche Diagnostics, Mannheim, Germany) and following procedures and reaction conditions described in Stokdyk et al.28 and Supporting Information. For the standard curve, assay efficiency was 0.92 (Efficiency = 10(−1/slope)-1), mean square error was 0.144, and the highest cycle of quantification was 40. The 95% limit of detection for the qPCR assay as determined by probit analysis was 28 gene copies reaction−1.28−30 Lambda phage DNA was used to evaluate all samples for qPCR inhibition following Gibson et al.;31 inhibition was mitigated for 63 samples by dilution using AE buffer (dilution factors are described in Supporting Information). Negative controls were included for secondary concentration, nucleic acid extraction, and qPCR and must exhibit no fluorescence above the baseline for sample data to be accepted. Modified live bovine herpes virus vaccine (Zoetis Inc., Kalamazoo, MI) was used as an extraction positive control. Species and Subtype Identification. Samples positive for Cryptosporidium by qPCR were sequenced to determine species and subtype. The 18S rRNA gene was targeted to identify species following Xiao et al.32 (SI Table S1). Subtypes were determined by sequencing the GP60 gene33,34 (SI Table S1). Reaction conditions and sequencing are described in Supporting Information. Direct Immunofluorescent Assay (IFA). Oocysts were enumerated by IFA for all qPCR-positive samples and 30 randomly selected qPCR-negative samples. Secondary concentrate (100 μL) was analyzed using a Merifluor Cryptosporidium/Giardia Detection Kit (Meridian Biosciences, Inc., Cincinnati, OH; see Supporting Information). The volume of secondary concentrate analyzed (100 μL) was on average equivalent to 18 L of sampled water (range 5−40 L). Chemical and Physical Parameters. Stable isotope analyses (oxygen-18 and deuterium) were conducted at either the University of Waterloo Environmental Isotope Laboratory or Isotope Tracer Technologies (both in Waterloo, Canada), the former using the LGR Laser method (off-axis integrated cavity output laser spectroscopy) and the latter using cavity ringdown spectroscopy. Tritium was analyzed using electrolytic enrichment followed by liquid scintillation counting. Total organic carbon, nitrate/nitrite, ammonia, bromide, chloride, boron, and total coliforms were analyzed following standard methods (SI Table S2). Oxidation−reduction potential, temperature, conductivity, pH, and dissolved oxygen were

measured during sample collection using YSI or Oakton probes. Surface Water Influence. Wells were placed into one of three categories of surface water influence on groundwater: (1) evidence of an evaporative surface water signature; (2) no evidence of evaporative surface water but there was evidence of rapid infiltration; and (3) no surface water influence as there was no evidence of evaporative water or rapid infiltration (SI Figure S1). Evidence of evaporative surface water was based on stable isotope measurements (oxygen-18 and deuterium). Wells with two or more samples deviating from the local meteoric water line of Landon35 (line-conditioned excess value < −1.0)36 and an oxygen-18 value heavier than the estimated annual precipitation value37 were considered influenced by surface water. Wells with fewer than two isotope samples (n = 8) were not assigned to a category. Evidence of rapid infiltration was determined by expert judgment of State of Minnesota hydrogeologists based on groundwater age and chemical characteristics of the groundwater, including temporal variability of chemical and isotopic parameters, and site-specific hydrogeologic assessments, including vertical hydraulic gradients and geologic sensitivity.22 The Surface Water Treatment Rule broadly defines surface water influence based on the presence of macroscopic organisms (including Cryptosporidium) or water quality characteristics indicative of surface water, directing states to establish specific criteria.4 The assessment of surface water described here is consistent with the criteria outlined by USEPA, and it is currently used in Minnesota for management (but not regulatory) purposes (see Supporting Information). Statistical Analysis. Results from qPCR and IFA analysis were examined to determine if surface water influence was associated with Cryptosporidium detections and concentrations among wells grouped in two ways. First, the proportion of positive wells with an evaporative water signature, evidence of rapid infiltration, or evidence of neither (three groups), was compared using a Chi-squared test, and a Kruskal−Wallis oneway analysis of variance on ranks was used to determine differences in concentrations of positive samples among the three groups (SigmaPlot 13.0). Second, wells with an evaporative water signature or evidence of rapid infiltration were grouped together as both show evidence of surface water influence. Wells in this combined group were compared to wells without evidence of surface water influence (two groups) to evaluate differences in the proportion of positive wells (Chisquared test) and concentrations of positive samples (Mann− Whitney Rank Sum test; SigmaPlot 13.0). Wells (n = 8) and positive samples (n = 3) without a surface water influence assessment were excluded from the analyses, and one nonquantitative positive result (which lacked a record of sample volume) was excluded from the concentration comparison. Linear regression was used to examine the relationship between qPCR and IFA concentrations (Microsoft Excel 2016). The 106 samples with quantitative results were included in the regression. We completed a separate regression restricted to samples that were positive for both methods (n = 66) to examine the effect of IFA-negative samples on the outcome. Likewise, we ran the regression with and without one outlier (concentration of 246 gene copies L−1). C

DOI: 10.1021/acs.est.8b05446 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology Table 3. Number of Times Community and Non-Community Wells Were Positive for Cryptosporidium by qPCR community wells (n = 88)



noncommunity wells (n = 57)

all wells (n = 145)

times well was positive

n

proportion

n

proportion

n

proportion

0 1 2 3 4 5 6 7 8 9 positive at least once positive more than once

54 24 5 4 1 0 0 0 0 0 34 10

0.61 0.27 0.06 0.05 0.01 0.00 0.00 0.00 0.00 0.00 0.39 0.11

33 12 5 3 1 0 1 1 0 1 24 12

0.58 0.21 0.09 0.05 0.02 0.00 0.02 0.02 0.00 0.02 0.42 0.21

87 36 10 7 2 0 1 1 0 1 58 22

0.60 0.25 0.07 0.05 0.01 0 0.01 0.01 0 0.01 0.40 0.15

RESULTS Of 964 samples, 107 (11%) were positive for Cryptosporidium by qPCR. Concentrations for positive samples ranged from 0.05 to 246 gene copies L−1 with a mean of 10.4 gene copies L−1, and most positive samples (77) had concentrations less than one gene copy L−1. Samples collected in year two were qPCR-positive at a higher rate than those collected in year one (14 and 8%, respectively). Of 145 wells, 58 (40%) were positive at least once, and 22 wells were positive more than once, including a single well with nine detections (of 12 samples; Table 3). The percentage of positive wells for community (39%) and noncommunity (42%) systems did not differ, though a higher percentage of noncommunity wells were positive more than once (21% compared to 11% of community wells; Table 3). Wells sampled in year two were qPCR-positive at a higher rate than those sampled in year one (45 and 25%, respectively). Of the 107 qPCR-positive samples, 67 (63%) were positive by IFA, which represents 7% of all 964 samples. Of 30 randomly selected qPCR-negative samples all were negative by IFA. Samples were positive by IFA across the range of qPCR concentrations (0.05−246 gene copes L−1), and only samples with low qPCR concentrations were negative. Specifically, 40 of 75 samples (53%) with qPCR concentrations 0.7 gene copies L−1 (n = 32) were positive. Regression revealed a strong relationship between the two methods (p < 0.001, R2 = 0.906, slope = 13.5 ± 0.4; Figure 1). Exclusion of IFA-negative samples or one outlier produced similar results to analysis of the full data set (R2 = 0.90 and 0.84 and slopes = 13.7 and 11.4, respectively). A total of 96 wells were categorized as potentially influenced by surface water, including 80 with an evaporative surface water signature and 16 with evidence of rapid infiltration (Table 4). Surface water influence was not associated with Cryptosporidium detection as the proportion of wells positive for Cryptosporidium by qPCR did not differ among wells with an evaporative water signature, evidence for rapid infiltration, or neither, χ2 (2, n = 137) = 0.91, p = 0.64. Likewise, concentrations of positive samples did not differ among wells in these groups (H(2) = 0.77, p = 0.68). Of the 96 wells with evidence of surface water influence (evaporative surface water signature or rapid infiltration), 38 (40%) were positive for Cryptosporidium by qPCR while 17 of 41 wells (41%) without surface water influence were positive; these proportions did not differ (χ2 (1, n = 137) = 0.0002, p = 0.99). Similarly,

Figure 1. Relationship between concentrations for samples (n = 106) measured by quantitative polymerase chain reaction (qPCR) and immunofluorescent assay (IFA; p < 0.001, R2 = 0.906). Inset: Regression with one outlier removed.

concentrations of positive samples did not differ among wells in these two groups (U = 1080, p = 0.39). Analysis of IFA data also showed no association with surface water influence for either the proportion of positive wells (χ2 (2, n = 137) = 0.99, p = 0.61) or the concentrations of positive samples (H(2) = 4.17, p = 0.12) when tested using three groups (evaporative surface water, rapid infiltration, and no surface water influence). Likewise, no differences in the proportion of positive wells (χ2 (1, n = 137) = 0.61, p = 0.44) or the concentrations of positive samples (U = 0.389, p = 0.20) were observed when we examined wells in two groups (those with and without evidence of surface water influence). For the statistical analysis of IFA data, we assumed that all qPCR-negative samples were also IFA-negative based on the strong relationship between qPCR and IFA, the greater analytical sensitivity demonstrated by qPCR in this data set, and the fact that all 30 qPCR-negative samples analyzed by IFA were negative. Only two Cryptosporidium-positive wells had been previously categorized by the State of Minnesota as GWUDI for regulatory purposes; the remaining positive wells had not been regulated as GWUDI. Sequencing of the 18S rRNA gene to determine species was successful for 45 of the samples positive by qPCR (62 were unsuccessful). Of these, 41 were identified as C. parvum, two as C. andersoni, and two as C. hominis (Table 5). Results of the BLAST search yielded matches with percent identity ranging from 98 to 100 and 0 for all E-scores. Subtype identification D

DOI: 10.1021/acs.est.8b05446 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Table 4. Number of Cryptosporidium-Positive Wells and Concentration of Positive Samples for Wells with Evidence of Evaporative Surface Water, Rapid Infiltration, Or No Evidence of Surface Water Influencea Cryptosporidium-positive by qPCR category evaporative surface water rapid infiltration no surface water influence

Cryptosporidium-positive by IFA

number of wellsb

number of wells (%)

number of samples

median concentration (gc L−1)

number of Wellsc (%)

number of Samples

median concentration (oocysts L−1)

80

30 (38%)

47

0.15

24 (30%)

34

0.24

16 41

8 (50%) 17 (41%)

20 36

0.18 0.11

5 (31%) 9 (22%)

10 23

1.5 0.19

a gc, gene copies. bEight wells, including 3 qPCR-positive wells, had too few isotope samples for assessment. cPercentages calculated from the number of wells in a category, assuming qPCR-negative samples were negative by IFA.

one or two samples per well were examined in most previous studies). Eleven percent of our samples were positive, which is within the range reported in previous studies. Sampling frequency may help explain the high proportion of positive wells, but the overall incidence of Cryptosporidium in our study challenges the assumption that this parasite is uncommon in groundwater. The unexpected incidence of Cryptosporidium may be the result of sampling wells that were vulnerable to contamination based on geologic setting or well characteristics. However, the range of wells positive for Cryptosporidium suggests that these explanations do not fully account for the high incidence we observed. First, wells across the geographic extent of the state that represented all major aquifer types and a range of geologic sensitivities were positive for Cryptosporidium. Likewise, we detected Cryptosporidium in wells with a range of characteristics like age, depth, and bedrock depth, and wells from community and noncommunity systems were positive at similar rates. The difference in incidence between wells sampled in years one and two may reflect differences in well vulnerability given the inclusion of more potentially vulnerable wells in year two. Alternatively, the difference may be the result of factors that varied between years, like precipitation and source strength, but analyses of those factors were beyond the scope of our study. Regardless of the reason, wells with a broad range of characteristics were contaminated with Cryptosporidium, not only those in sensitive geologic settings or with potentially vulnerable characteristics. While surface water influence is generally associated with groundwater vulnerability, many wells that were positive for Cryptosporidium had no apparent surface water influence. Specifically, only two of the 58 Cryptosporidium-positive wells are defined and regulated as GWUDI, leaving 56 positive wells without a formal surface water classification. Categorization using our more inclusive definition of surface water influence (based on evaporative water signature or rapid infiltration) did not yield a relationship between these measures and Cryptosporidium presence or concentration as measured by qPCR or IFA. Of the wells that lacked any evidence for surface water influence, 41 and 22% were positive for Cryptosporidium by qPCR and IFA, respectively, compared to 40 and 30% for wells with evidence of surface water influence. These percentages reflect our approach to identifying surface water influence and could change if other approaches were applied. For example, microscopic particulate analysis (MPA) can be used for GWUDI determination.4,38 However, given the limitations of MPA,38 we suspect it is less likely to detect surface water influence than the isotope and water quality methods applied here. MPA data were not available for wells in this study. While surface water influence may contribute to

Table 5. Cryptosporidium Species and Subtypes Determined by Sequencinga species

subtype

no. samples

no. wells

C. andersoni C. hominis C. parvum

ND ND IIa A14G1R2 IIa A15G2R1 IIa A16G2R3 IIa A17G2R1 IIa A19G2R1 ND

2 2 1 2 3 16 1 18 62

2 1 1 1 2 6 1 11 33

ND a

ND, not determined (unable to sequence).

based on the GP60 gene was successful for 23 samples. Results based on the GP60 gene were consistent with those from 18S rRNA sequencing (Table 5), and percent identity ranged 99− 100 with all E-scores equal to 0. Like the IFA analysis, sequencing was successful across the range of qPCR concentrations but more often successful for samples with higher concentrations. Specifically, all 23 samples successful for GP60 sequencing had concentrations >2 gene copes L−1, and all 29 samples in that range were successful for 18S rRNA sequencing. When two or more samples from the same well were successfully sequenced, the samples had the same species (n = 10 wells) and subtype (n = 5 wells).



DISCUSSION We analyzed samples from 145 public water system wells across Minnesota for Cryptosporidium using qPCR and IFA and observed a higher incidence of positive wells than previously reported. There are few studies examining Cryptosporidium occurrence in groundwater, presumably because of the assumption that groundwater is protected from contamination by overlying soil and subsurface sediments. Moreover, wells in previous studies were often selected for Cryptosporidium analysis because they were considered vulnerable, such as in close proximity to surface water (e.g., refs 8 and 9) or contamination sources (e.g., refs 7 and 10). Reported Cryptosporidium occurrence for wells ranged from 0 to 15%,8−11,18 whereas 40% of wells in our study were positive by qPCR. However, most positive wells in our study (36 of 58 positive wells) had only a single detection that was accompanied by multiple negative results. Consistent with this result, Gallas-Lindemann et al.9 sampled three wells 22 times each and found only five of the 66 samples were positive. Sample-level incidence may be more comparable to other studies than well-level incidence because we sampled wells more frequently than most other studies (typically six times; E

DOI: 10.1021/acs.est.8b05446 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology groundwater contamination by Cryptosporidium, it does not appear to explain the high incidence we observed. We must also consider potential methodological contributions to the high prevalence of Cryptosporidium in our study as we used qPCR and collected large sample volumes, both of which may increase the sensitivity of our analyses compared to microscopy approaches and small sample volumes. Methodological explanations are insufficient, however. First, results from the conventional IFA analysis support our qPCR results; samples across the qPCR concentration range, including the sample with the lowest concentration, were positive by IFA. In addition, while we collected large sample volumes (mean of 728 L), the mean sample volume equivalent we analyzed by IFA (18 L) was in the lower part of the range of mean volume equivalents analyzed in other studies (0.1−293 L) and described in USEPA method 1623 (10−100 L depending on the filter used13,18,26,39). Moreover, there was no relationship between IFA detections and volume equivalent analyzed (Table 6), which is consistent with observations made by

wells contrasts with the results of existing observational studies. Experimental studies, however, demonstrate that transport of Cryptosporidium through soil and subsurface sediments overlying groundwater is possible. Experiments that used laboratory soil columns have yielded reports of oocyst transport through various types of soil, though the soil retained a significant portion of the oocysts applied.42−46 While soil can attenuate oocysts, it may subsequently serve as a contamination source if conditions allow oocyst release.45−47 Given the high shedding rates by infected hosts and low infective dose of Cryptosporidium, the transport of even a fraction of a source’s oocysts to groundwater represents a potential public health concern, especially as transported oocysts can remain infective.42 Cryptosporidium transport to wells in our study may occur through the soil and subsurface sediments to the aquifer, as demonstrated in experimental studies, or through compromised well structures, such as a casing flaw or compromised grouting. For either transport route, a fecal source is required. The fecal source of Cryptosporidium detected in our study wells can be generally assessed using species and subtype identification, with C. andersoni common in cattle, C. hominis common in humans, and C. parvum, which we detected most, found in cattle, humans, and other mammals, including wildlife.48,49 Species specifically associated with other mammals were not identified, though Cryptosporidium associated with wildlife has been detected in environmental waters (e.g., ref 50). Molecular analyses for Cryptosporidium species and subtype identification are often unsuccessful for environmental samples (unlike fecal specimens). We successfully sequenced 45 and 23 of 107 samples to determine species and subtype, respectively. In comparison, other studies identifying species or subtype by analyzing the 18S rRNA, GP60, or HSP-70 gene using PCR-restriction fragment length polymorphisms or Sanger sequencing was successful for one of 16,8 31 of 92,50 one of nine,11 and three of nine51 water samples; all but three of the reported samples were from surface water as groundwater was infrequently tested. Determining the fecal source associated with contamination in this way is limited because of the low success rate and the possibility that only one of several species present in the water is identified. Livestock are a host of several Cryptosporidium species and are commonly considered a likely source of Cryptosporidium in environmental waters (e.g., refs 7, 8, 11, 50, and 52). The incidence of C. parvum indicates that a cattle fecal source is possible for many of our study wells. However, the zoonotic nature of C. parvum subtype IIa precludes definitive identification of a source because it can infect both cattle and humans. Therefore, human fecal material, the other common source of Cryptosporidium in water, is also possible. While the specific source and transport pathways in our study are unknown, Cryptosporidium’s presence suggests that public health may benefit from a shift away from the regulatory paradigm that considers groundwater an inherently protected resource. Regulatory rules designed for Cryptosporidium do not exist for groundwater in the United States (except GWUDI) as they do for surface water.2,3,5 The USEPA Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR) requires filtered surface water systems with average concentrations above 0.075 oocysts L−1 in their source water to perform additional treatment.5 Among study wells positive by IFA at least once, average concentrations ranged 0.01−3.25 oocysts L−1, including 18 wells with averages >0.075 oocysts L−1; consistent

Table 6. Sample Volume and Volume Equivalent Analyzed for Samples Tested by IFA volume equivalent analyzed (L)b IFA samples

mean sample volume collected (L)

mean

min

max

all (n = 136)a positive (n = 67) negative (n = 70)

712 724 700

17.6 18.0 17.2

4.5 4.5 4.7

40.1 40.1 28.6

a

Includes 106 qPCR-positive and 30 qPCR-negative samples. Represents the equivalent volume of sample analyzed (i.e., viewed under the microscope) based on the sample volume and proportion of the volume concentrate tested, which allows comparison across samples with different volumes collected.

b

Hancock et al.18 in a study that examined a broad range of volume equivalents (0.2−1866 L). Lastly, there was no evidence of laboratory contamination as all no-template controls for qPCR were negative. We demonstrated a strong relationship between qPCR and IFA that supports our qPCR results and that can serve as a reference between the two methods for applications like quantitative microbial risk assessment. The ratio of 13.5 gene copies per oocyst (based on the regression slope) is lower than the ratio of 25 reported by Mary et al.27 The difference may result from analysis of groundwater samples compared to the fecal specimens analyzed by Mary et al.27 Nonetheless, an established relationship between these methods for groundwater samples is informative, useful, and supports the qPCR results. While qPCR was more sensitive, the 26% of study wells positive by IFA is still greater than most previous reports using the same analytical approach. Moreover, the greater sensitivity of qPCR does not imply that our qPCR-based incidence values are overestimates compared to other studies; rather, it indicates a methodological difference40,41 and suggests that the IFA-based results of previous studies may represent underestimates compared to actual incidence. Overall, the IFA results corroborate our qPCR measurements and reveal a robust relationship between the methods. Given the strong relationship between qPCR and IFA results as well as the varied population of wells and hydrogeology represented, the high prevalence of Cryptosporidium in these F

DOI: 10.1021/acs.est.8b05446 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

ment; Final Rule, 40 CFR Parts 9, 141, 142; Office of Water: Washington, DC, 1998. (3) U. S. Environmental Protection Agency. National Primary Drinking Water Regulations: Long Term 1 Enhanced Surface Water Treatment Rule; Final Rule, 40 CFR Parts 9, 141, 142; Office of Water: Washington, DC, 2002. (4) U. S. Environmental Protection Agency. National Primary Drinking Water Regulations, 40 CFR Part 141.2; Office of Water: Washington, DC, 2003. (5) U. S. Environmental Protection Agency. National Primary Drinking Water Regulations: Long Term 2 Enhanced Surface Water Treatment Rule; Final Rule, 40 CFR parts 9, 141, and 142; Office of Water: Washington, DC, 2006. (6) U. S. Environmental Protection Agency. Occurrence and Monitoring Document for the Final Ground Water Rule, EPA 815-R06-012; Office of Water: Washington, DC, 2006. (7) Gaut, S.; Robertson, L.; Gjerde, B.; Dagestad, A.; Brattli, B. Occurrence of Cryptosporidium oocysts and Giardia cysts in Norwegian groundwater wells in bedrock. J. Water Health 2008, 6 (3), 383−388. (8) Budu-Amoako, E.; Greenwood, S. J.; Dixon, B. R.; Barkema, H. W.; McClure, J. T. Occurrence of Cryptosporidium and Giardia on beef farms and water sources within the vicinity of the farms on Prince Edward Island, Canada. Vet. Parasitol. 2012, 184 (1), 1−9. (9) Gallas-Lindemann, C.; Sotiriadou, I.; Plutzer, J.; Karanis, P. Prevalence and distribution of Cryptosporidium and Giardia in wastewater and the surface, drinking and ground waters in the Lower Rhine, Germany. Epidemiol. Infect. 2013, 141 (1), 9−21. (10) Pitkänen, T.; Juselius, T.; Isomäki, E.; Miettinen, I. T.; Valve, M.; Kivimäki, A. L.; Lahti, K.; Hänninen, M. L. Drinking water quality and occurrence of Giardia in Finnish small groundwater supplies. Resources. 2015, 4 (3), 637−654. (11) Daniels, M. E.; Shrivastava, A.; Smith, W. A.; Sahu, P.; Odagiri, M.; Misra, P. R.; Panigrahi, P.; Suar, M.; Clasen, T.; Jenkins, M. W. Cryptosporidium and Giardia in humans, domestic animals, and village water sources in rural India. Am. J. Trop. Med. Hyg. 2015, 93 (3), 596−600. (12) Gamba, R. D.; Ciapina, E. M.; Espíndola, R. S.; Pacheco, A.; Pellizari, H. Detection of Cryptosporidium sp. oocysts in groundwater for human consumption in Itaquaquecetuba city, S. Paulo-Brazil. Braz. J. Microbiol. 2000, 31 (2), 151−153. (13) Haramoto, E.; Yamada, K.; Nishida, K. Prevalence of protozoa, viruses, coliphages and indicator bacteria in groundwater and river water in the Kathmandu Valley, Nepal. Trans. R. Soc. Trop. Med. Hyg. 2011, 105 (12), 711−716. (14) Lee, Y. M.; Johnson, P. W.; Call, J. L.; Arrowood, M. J.; Furness, B. W.; Pichette, S. C.; Grady, K. K.; Reeh, P.; Mitchell, L.; Bergmire-Sweat, D.; Mackenzie, W. R. Development and application of a quantitative, specific assay for Cryptosporidium parvum oocyst detection in high-turbidity environmental water samples. Am. J. Trop. Med. Hyg. 2001, 65 (1), 1−9. (15) Khaldi, S.; Ratajczak, M.; Gargala, G.; Fournier, M.; Berthe, T.; Favennec, L.; Dupont, J. P. Intensive exploitation of a karst aquifer leads to Cryptosporidium water supply contamination. Water Res. 2011, 45 (9), 2906−2914. (16) Betancourt, W. Q.; Rose, J. B. Microbiological assessment of ambient waters and proposed water sources for restoration of a Florida wetland. J. Water Health 2005, 3 (2), 89−100. (17) Damiani, C.; Balthazard-Accou, K.; Clervil, E.; Diallo, A.; Da Costa, C.; Emmanuel, E.; Totet, A.; Agnamey, P. Cryptosporidiosis in Haiti: surprisingly low level of species diversity revealed by molecular characterization of Cryptosporidium oocysts from surface water and groundwater. Parasite 2013, 20 (45), 1−6. (18) Hancock, C. M.; Rose, J. B.; Callahan, M. Crypto and Giardia in US groundwater. J. - Am. Water Works Assoc. 1998, 90 (3), 58−61. (19) U. S. Environmental Protection Agency. 2012. Appendices to the Economic Analysis for the Final Revised Total Coliform Rule, EPA-815R-12-004; Office of Water: Washington, DC, 2012.

with LT2ESWTR, negative samples were included as zeros in the average. The comparison is mismatched, however, because averages for LT2ESWTR are based on at least 48 samples while study wells were sampled 12 or fewer times. If we calculate a theoretical 48-sample average for these wells by conservatively assuming the balance of samples were negative, then seven wells exceed the concentration threshold. The 0.075 oocysts L−1 threshold assumes that the minimum treatment required for all filtered surface water systems in the US is in place. For unfiltered surface water systems, inactivation of Cryptosporidium is required regardless of concentration. All 38 IFA-positive wells would therefore require inactivation if regulated as GWUDI under the LT2ESWTR. The incidence and concentrations of Cryptosporidium in the study wells suggest monitoring and treatment measures should be considered for groundwater-supplied public water systems. Results from this study challenge the current regulatory paradigm regarding the limited occurrence of Cryptosporidium in groundwater. The incidence of Cryptosporidium in wells in this study was higher than previous reports, and occurrence was not associated with surface water influence as defined in this study by evaporative water signature and rapid infiltration. Therefore, these results contradict the use of Cryptosporidium as an indicator of GWUDI and suggest that factors beyond surface water influence should be considered for the potential contamination of groundwater by Cryptosporidium.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.8b05446. Additional methodological details and expanded description of surface water influence (PDF)



AUTHOR INFORMATION

Corresponding Author

*Phone: 715-387-4943; e-mail: [email protected]. ORCID

Mark A. Borchardt: 0000-0002-6471-2627 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Dane Huber, Jared Schmaedeke, Trisha Sisto, Mike Sutliff, and Nathan Gieske for well sampling. Funding was provided by the Minnesota Clean Water, Land, and Legacy Amendment Fund. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. At the time of publication, data have limited availability owing to participating public water systems having the prerogative to communicate raw water quality results. Contact corresponding author for more information.



REFERENCES

(1) Mac Kenzie, W. R.; Hoxie, N. J.; Proctor, M. E.; Gradus, M. S.; Blair, K. A.; Peterson, D. E.; Kazmierczak, J. J.; Addiss, D. G.; Fox, K. R.; Rose, J. B.; Davis, J. P. A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public water supply. N. Engl. J. Med. 1994, 331 (3), 161−167. (2) U. S. Environmental Protection Agency. National Primary Drinking Water Regulations: Interim Enhanced Surface Water TreatG

DOI: 10.1021/acs.est.8b05446 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

hydrologic systems. In Isotopes in Environmental Studies, Aquatic Forum, IAEA-CSP-26; International Atomic Energy Agency: Vienna, 2006; pp 132−135. (37) Bowen, G. J.; Revenaugh, J. Interpolating the isotopic composition of modern meteoric precipitation. Water Resour. Res. 2003, 39 (10), 1299. (38) U. S. Environmental Protection Agency. Consensus Method for Determining Groundwaters under the Direct Influence of Surface Water Using Microscopic Particulate Analysis (MPA), EPA 910/9-92-029; Office of Water: Washington, DC, 1992. (39) U. S. Environmental Protection Agency. Method 1623: Cryptosporidium and Giardia in Water by Filtration/IMS/FA, 815-R05-002; Office of Water: Washington, DC, 2005. (40) Gallas-Lindemann, C.; Sotiriadou, I.; Plutzer, J.; Noack, M. J.; Mahmoudi, M. R.; Karanis, P. Giardia and Cryptosporidium spp. dissemination during wastewater treatment and comparative detection via immunofluorescence assay (IFA), nested polymerase chain reaction (nested PCR) and loop mediated isothermal amplification (LAMP). Acta Trop. 2016, 158, 43−51. (41) Jothikumar, N.; Da Silva, A. J.; Moura, I.; Qvarnstrom, Y.; Hill, V. R. Detection and differentiation of Cryptosporidium hominis and Cryptosporidium parvum by dual TaqMan assays. J. Med. Microbiol. 2008, 57 (9), 1099−1105. (42) Boyer, D. G.; Kuczynska, E.; Fayer, R. Transport, fate, and infectivity of Cryptosporidium parvum oocysts released from manure and leached through macroporous soil. Environ. Geol. 2009, 58 (5), 1011−1019. (43) Darnault, C. J.; Garnier, P.; Kim, Y. J.; Oveson, K. L.; Steenhuis, T. S.; Parlange, J.; Jenkins, M.; Ghiorse, W. C.; Baveye, P. Preferential transport of Cryptosporidium parvum oocysts in variably saturated subsurface environments. Water Environ. Res. 2003, 75 (2), 113−120. (44) Darnault, C. J.; Steenhuis, T. S.; Garnier, P.; Kim, Y. J.; Jenkins, M. B.; Ghiorse, W. C.; Baveye, P. C.; Parlange, J. Preferential flow and transport of Cryptosporidium parvum oocysts through the vadose zone. Vadose Zone J. 2004, 3 (1), 262−270. (45) Hijnen, W. A.; Brouwer-Hanzens, A. J.; Charles, K. J.; Medema, G. J. Transport of MS2 phage, Escherichia coli, Clostridium perfringens, Cryptosporidium parvum, and Giardia intestinalis in a gravel and a sandy soil. Environ. Sci. Technol. 2005, 39 (20), 7860−7868. (46) Mohanram, A.; Ray, C.; Harvey, R. W.; Metge, D. W.; Ryan, J. N.; Chorover, J.; Eberl, D. D. Comparison of transport and attachment behaviors of Cryptosporidium parvum oocysts and oocyst-sized microspheres being advected through three minerologically different granular porous media. Water Res. 2010, 44 (18), 5334−5344. (47) Harter, T.; Wagner, S.; Atwill, E. R. Colloid transport and filtration of Cryptosporidium parvum in sandy soils and aquifer sediments. Environ. Sci. Technol. 2000, 34 (1), 62−70. (48) Xiao, L.; Fayer, R.; Ryan, U.; Upton, S. J. Cryptosporidium taxonomy: recent advances and implications for public health. Clin. Microbiol. Rev. 2004, 17 (1), 72−97. (49) Fayer, R. Cryptosporidium: a water-borne zoonotic parasite. Vet. Parasitol. 2004, 126 (1−2), 37−56. (50) 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 water for improved health risk assessment. Appl. Environ. Microbiol. 2007, 73 (12), 3945−3057. (51) Ehsan, A. M.; Geurden, T.; Casaert, S.; Parvin, S. M.; Islam, T. M.; Ahmed, U. M.; Levecke, B.; Vercruysse, J.; Claerebout, E. Assessment of zoonotic transmission of Giardia and Cryptosporidium between cattle and humans in rural villages in Bangladesh. PLoS One 2015, 10 (2), e0118239. (52) Oates, S. C.; Miller, M. A.; Hardin, D.; Conrad, P. A.; Melli, A.; Jessup, D. A.; Dominik, C.; Roug, A.; Tinker, M. T.; Miller, W. A. Prevalence, environmental loading, and molecular characterization of Cryptosporidium and Giardia isolates from domestic and wild animals along the central California coast. Appl. Environ. Microbiol. 2012, 78 (24), 8762−8772.

(20) 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 (5), 1423−1428. (21) Minnesota drinking water information system database, Minnesota Department of Health; https://data.web.health.state.mn. us/drinking_water_metadata. (22) Geologic Sensitivity Project Workgroup. Criteria and Guidelines for Assessing Geologic Sensitivity of Ground Water Resources in Minnesota; Minnesota Department of Natural Resources, Division of Waters: St. Paul, MN, 1991. (23) Smith, C. M.; Hill, V. R. Dead-end hollow-fiber ultrafiltration for recovery of diverse microbes from water. Appl. Environ. Microbiol. 2009, 75 (16), 5284−5289. (24) Hill, V. R.; Polaczyk, A. L.; Kahler, A. M.; Cromeans, T. L.; Hahn, D.; Amburgey, J. E. Comparison of hollow-fiber ultrafiltration to the USEPA VIRADEL technique and USEPA method 1623. J. Environ. Qual. 2009, 38 (2), 822−825. (25) Polaczyk, A. L.; Narayanan, J.; Cromeans, T. L.; Hahn, D.; Roberts, J. M.; Amburgey, J. E.; Hill, V. R. Ultrafiltration-based techniques for rapid and simultaneous concentration of multiple microbe classes from 100-L tap water samples. J. Microbiol. Methods 2008, 73 (2), 92−99. (26) Rhodes, E. R.; Villegas, L. F.; Shaw, N. J.; Miller, C.; Villegas, E. N. A modified EPA Method 1623 that uses tangential flow hollowfiber ultrafiltration and heat dissociation steps to detect waterborne Cryptosporidium and Giardia spp. J. Visualized Exp. 2012, No. 65, 4177. (27) Mary, C.; Chapey, E.; Dutoit, E.; Guyot, K.; Hasseine, L.; Jeddi, F.; Menotti, J.; Paraud, C.; Pomares, C.; Rabodonirina, M.; Rieux, A.; Derouin, F. Multicentric evaluation of a new real-time PCR assay for quantification of Cryptosporidium spp. and identification of Cryptosporidium parvum and Cryptosporidium hominis. J. Clin. Microbiol. 2013, 51 (8), 2556−2563. (28) Stokdyk, J. P.; Firnstahl, A. D.; Spencer, S. K.; Burch, T. R.; Borchardt, M. A. Determining the 95% limit of detection for waterborne pathogen analyses from primary concentration to qPCR. Water Res. 2016, 96, 105−113. (29) Evaluation of Detection Capability for Clinical Laboratory Measurement Procedures; Approved Guideline, CLSI document EP17eA2, 2nd ed.; Clinical and Laboratory Standards Institute: Wayne, PA, 2012. (30) Wolk, D. M.; et al. Molecular method verification. In Molecular Microbiology: Diagnostic Principles and Practice, 2nd ed.; Persing, D. H., Tenover, F. C., Tan, Y., Nolte, F. S., Hayden, R. T., VanBelkum, A., Eds.; ASM Press: Washington, DC, 2011; pp 861−883. (31) Gibson, K. E.; Schwab, K. J.; Spencer, S. K.; Borchardt, M. A. Measuring and mitigating inhibition during quantitative real time PCR analysis of viral nucleic acid extracts from large-volume environmental water samples. Water Res. 2012, 46 (13), 4281−4291. (32) 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 (12), 5492−5498. (33) Strong, W. B.; Gut, J.; Nelson, R. G. Cloning and sequence analysis of a highly polymorphic Cryptosporidium parvum gene encoding a 60-kilodalton glycoprotein and characterization of its 15- and 45-kilodalton zoite surface antigen products. Infect. Immun. 2000, 68, 4117−4134. (34) Glaberman, S.; Moore, J. E.; Lowery, C. J.; Chalmers, R. M.; Sulaiman, I.; Elwin, K.; Rooney, P. J.; Millar, B. C.; Dooley, J. S.; Lal, A. A.; Xiao, L. Three drinking-water-associated cryptosporidiosis outbreaks, Northern Ireland. Emerging Infect. Dis. 2002, 8, 631−633. (35) Landon, M. K.; Delin, G. N.; Komor, S. C.; Regan, C. R. Relation of pathways and transit times of recharge water to nitrate concentrations using stable isotopes. Groundwater 2000, 38 (3), 381− 395. (36) Landwehr, J. M.; et al. Line-conditioned excess: a new method for characterizing stable hydrogen and oxygen isotope ratios in H

DOI: 10.1021/acs.est.8b05446 Environ. Sci. Technol. XXXX, XXX, XXX−XXX