https://orcid.org/0000-0002-5953-8089
Figures
Abstract
Many rural communities in Alabama’s Black Belt region lack adequate sanitation, resulting in wastewater discharges that may pose risks to residents. To understand the scope of the problem in one community, we conducted three cross-sectional surveys in a small town with limited sanitation infrastructure in 2023. We measured a range of enteric pathogens in environmental samples by multi-parallel qPCR as well as fecal indicator bacteria E. coli and Enterococcus by culture and molecular methods. We examined soil samples (n = 58) from sites near failing septic systems or suspected direct surface discharges and comparison soil (n = 10) far from potential discharges to estimate sanitation-related pathogen hazards. We examined surface water samples from community (n = 8) and localized (n = 20) sites that may have been impacted by wastewater discharges. Comparing impacted and unimpacted soil samples revealed greater fecal contamination near known or suspected discharges, compared with control samples. The mean culturable E. coli count in impacted soils was 224 MPN/g (95% CI 0-510.5 MPN/g) and in unimpacted soils was 0.5 MPN/g (95% CI 0-1.5 MPN/g). We detected several pathogens via qPCR in impacted soil and surface water, including Acanthamoeba spp., Balantidium coli, Blastocystis spp., Cryptosporidium spp., and rotavirus. In community-level surface waters, 88% of samples were positive for E. coli by culture (n = 8, mean 3.04 x 105, 95% CI 0-8.96 x 105 MPN/100 mL); 100% were positive for Enterococcus by culture (n = 4, mean 1.10 x 104, 95% CI 0-2.55 x 104 MPN/100 mL); and we detected Acanthamoeba spp., Blastocystis spp., Cryptosporidium spp., Plesiomonas shigelloides., rotavirus, and Yersinia enterocolitica, suggesting community-level wastewater discharges may degrade local surface water quality. Evidence suggests sanitation failures contribute to enteric pathogen hazards in this community.
Citation: Harmon OA, Lott MEJ, Elliott M, McGlohn E, Brown J (2026) Environmental pathogen hazards reveal need for improved sanitation infrastructure in Alabama's Black Belt. PLOS Water 5(4): e0000393. https://doi.org/10.1371/journal.pwat.0000393
Editor: Nuhu Amin, ICDDR B: International Centre for Diarrhoeal Disease Research Bangladesh, BANGLADESH
Received: May 22, 2025; Accepted: February 8, 2026; Published: April 22, 2026
Copyright: © 2026 Harmon et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The datasets generated and analyzed during this study are available at https://osf.io/9kvfa/.
Funding: This study was supported by Columbia World Projects, which supported all authors. Olivia Harmon was additionally supported by a training grant from the National Institute of Environmental Health Sciences (T32ES007018). The funders had no role in the study design, data collection, data analysis, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Sanitary sewer connections are unavailable for many rural or disadvantaged communities within high-income countries [1,2]. Residents without access to sewer connections typically rely instead on onsite sewage disposal systems (OSDS), though these systems are vulnerable to failure due to poor silting, poor draining, and climate-related disasters [3]. The Transforming Wastewater Infrastructure in the United States project, supported by Columbia World Projects, aims to address these issues by presenting alternative wastewater technology solutions that are best suited for use in under-resourced settings.
As part of this project, a team will be piloting the installation of a decentralized wastewater system in a small town in the Black Belt region of Alabama. While most households in this town are connected to the county water supply, there is no existing municipal wastewater treatment system. Communities in the Black Belt region have received national recognition for failing or nonexistent wastewater infrastructure [4,5]. Since shrink-swell clay, common to this area, lead to hydraulic failure of septic systems, a high proportion of residents rely instead on straight pipes and cesspools [3]. This led to the Department of Justice’s first Environmental Justice investigation under Title VI of the Civil Rights Act of 1964 [6]. This investigation concluded with the citing of two areas of concern: the use of fines and law enforcement to punish people with inadequate wastewater treatment systems and inadequate action to assess and address health risks from raw sewage [4]. The decentralized cluster system that will be piloted in this town is intended to model a novel system designed specifically to meet the needs of rural and resource-limited communities.
The primary objective of this study was to compare fecal indicator bacteria (E. coli and Enterococcus) between sites near failing septic systems or suspected direct surface discharges (“impacted”) and sites far away from buildings, standing water, ditches, and other potential discharges (“unimpacted”). We aimed to establish baseline levels of fecal indicator bacteria (E. coli and Enterococcus) and enteric pathogens in soil, surface water, and standing water prior to the installation of a decentralized wastewater system. Environmental samples were analyzed using culture-based methods and multiplex molecular techniques to detect and quantify microbial targets.
Materials and methods
We conducted three cross-sectional surveys in a small town in Alabama’s Black Belt region during January, February, and May of 2023 to assess enteric pathogens and fecal indicator hazards in soil, surface water, and standing water samples. We collected samples across three different time points to capture potential microbial variations related to rainfall. Based on visual identification and local knowledge, we collected samples near suspected failing septic systems or suspected direct surface discharges. We categorized these samples as “impacted.” Additionally, we collected samples as far away as possible from known or identified wastewater discharge sites. We considered these “unimpacted.” We collected approximately 2 g of soil 2 inches below the soil surface and placed them into sterile WhirlPak bags. We collected 100–500 mL of surface and standing water and placed them into sterile WhirlPak bags. We stored all soil samples at –20C until downstream processing. Since all of the samples were collected in public spaces, no permits were obtained or required.
Culture of fecal indicator bacteria from soil
We cultured E. coli and Enterococcus from soil samples using the Quanti-Tray 2000 System. Briefly, we diluted 3 g of soil into 30 mL of phosphate-buffered saline (PBS). After homogenization by shaking, we diluted the solution 1:10 and 1:100 in PBS. We prepared the 1:10 and 1:100 dilutions for culture using the IDEXX Colilert-18 and IDEXX Enterolert media and IDEXX QuantiTrays. We incubated the QuantiTrays at either 35°C (Colilert-18) or 41°C (Enterolert) for 18–24 hours. We scored the wells and determined the most probable number (MPN) concentration using the IDEXX guidelines. Raw culture data for environmental samples analyzed via IDEXX can be found in S1 Data.
Culture of fecal indicator bacteria from water
We cultured E. coli and Enterococcus from surface water and standing water samples using the Quanti-Tray 2000 System. We prepared 1:10 and 1:100 dilutions of each water sample in sterile PBS. We prepared the samples with either the IDEXX Colilert-18 media or the IDEXX Enterolert (Enterococcus) media in the Quanti-Tray 2000 trays. Once we sealed and labeled the trays, we incubated them at 41°C for 24 hours for Enterolert or 35°C for 18 hours for Colilert-18. Raw culture data for environmental samples analyzed via IDEXX can be found in S1 Data.
Soil and surface water processing
For surface water samples transported to UNC, we thawed them at 4°C for 72 hours before processing. From each sample, we filtered 25 mL to 200 mL onto a 47-mm uM Millipore HA membrane, as shown in S1 Fig. For soil samples, we weighed out 0.189 grams to 0.485 grams and placed them into a 2 mL tube. Exact processing volumes and weights are included in S2 Data. To compare across sample types despite different processing volumes and weights, we converted template concentrations to normalized units, copies per 100 mL of water or copies per gram of soil, prior to analysis seen in equations 1 and 2, respectively.
Isolation of total nucleic acids and analysis
From the soil and water samples, we isolated total nucleic acids using the ZymoBIOMICS DNA/RNA Miniprep Kit (Zymo Research). Immediately prior to extraction, we spiked all samples with 10 μL of Bovine herpesvirus (BHV) and Bovine respiratory syncytial virus (BRSV), carried in the Zoetis Inforce-3 vaccine, which served as the extraction control. For each batch of extractions, we included a negative extraction control (PCR-grade water), and a positive extraction control (Inforce-3). From the extraction eluates, we measured the concentration of double-stranded DNA (dsDNA) and RNA using the Qubit High Sensitivity dsDNA and Qubit RNA High Sensitivity, Broad Range Assay Kits (Invitrogen, Carlsbad, CA, USA), respectively.
We analyzed the isolated total nucleic acids from soil and surface water samples, using a custom TaqMan Array Card (TAC) (ThermoFisher Scientific, Waltham, MA) targeting 39 enteric pathogens and fecal markers including: 20 types of bacterial gene targets (Acanthamoeba spp., Campylobacter jejuni/coli, Clostridium difficile, E. coli O157:H7, E. coli ybbW, Enteroaggregative E. coli aaiC, Enteroaggregative E. coli aatA, Enterococcus spp., Enteropathogenic E. coli bfpA + , Enteropathogenic E. coli eae + , Enterotoxigenic E. coli heat-labile, Enterotoxigenic E. coli STh, Enterotoxigenic E. coli STp, Helicobacter pylori, Plesiomonas shigelloides, Salmonella enterica, Shiga toxin producing E. coli Stx1, Shiga toxin producing E. coli Stx2, Shigella spp. and enteroinvasive E. coli, and Yersinia enterocolitica), 9 virus gene targets (Adenovirus 40/41, Astroviridae, Hepatitis A, Norovirus GI, Norovirus GII, Rotavirus, Sapovirus I/II/IV/V, SARS-CoV-2, Zika Virus), 6 helminth gene targets (Ancylostoma duodenale, Ascaris lumbricoides, Enterobius vermicularis, Necator americanus, Strongyloides stercoralis, and Trichuris trichiura), 6 protozoan gene targets (Balantidium coli, Blastocystis spp., Cryptosporidium spp., Cyclospora cayetanensis, Entamoeba histolytica, and Giardia spp.), and 2 MST Makers (Human Mitochondrial DNA and Human-specific HF183 Bacteroides 16S rRNA genetic marker). More details on the targets included are in S1 Table. We added 33 μL of nucleic acid extraction and 67 μL of master mix (Hepatitis G, Exo IPC Mix (10X), Exo IPC DNA (50X), AgPath Buffer (2X), and AgPath Enzyme) into one port in each TAC. We spiked in Hepatitis G to the master mix as the qPCR control.
Once we filled each TAC port, we performed one-step reverse transcription qPCR using a QuantStudio 7 (Thermo Fisher Scientific, Waltham, MA) at the following thermocycling conditions: 45°C for 20 minutes and 94°C for 10 minutes, followed by 45 cycles of 95°C for 15 seconds and 60°C for 1 minute, with a ramp rate of 1°C/second between each step. Once complete, we compared the curves and multicomponent plots with the control plots to ensure only positive amplifications were included by either flagging them or omitting unreliable results [7].
Standard curves
We generated standard curves by running a 7-fold dilution series of positive control material to determine assay efficiency. Next, we compiled the raw TAC data into a single data frame. We then implemented quality control steps and flagged failed assays where the IPC did not amplify (435/6315, 6.9%). We set all Ct values above the y-intercept of the standard curve for that particular target to “NA” in order to remove suspicious data points (136/6315, 2.2%) [8]. We performed quality assurance checks, including evaluating amplification of the different targets across different TAC cards in the positive PCR control and identifying failed targets based on Ct thresholds (10 < Ct > 40) (178/6315, 2.8%). We also assessed the amplification of the Hepatitis G target, which was included in the PCR master mix, across all samples and excluded any sample that did not meet the Ct thresholds (10 < Ct > 40) (661/6315, 10.5%). We calculated recovery efficiency for BHV and BRSV by comparing sample Ct values against the average value of each target in the positive extraction control samples.
Quality control
The targets Enterococcus spp. lsrRNA and/or E. coli ybbW appeared in 7/8 (88%) of the NEC (negative extraction control) and 6/8 (75%) PEC (process extraction control) (S3 Data). We identified the matching environmental samples with the same extraction group for each target detected in NEC or PEC controls, then subtracted the corresponding copies per μL. If both controls showed detection, the higher of the two was subtracted to account for background or contamination, improving the accuracy of quantification. This final data frame is included in S4 Data.
Results
Across the three sampling campaigns, we collected and processed 123 samples (soils weighed and membrane filtration of water samples). Of the samples collected, 35% (43 of 123) were cultured for E. coli via IDEXX, 25% (31 of 123) were cultured for Enterococcus via IDEXX, 88% (108 of 123) underwent nucleic acid extraction, and 74% (91 of 123) were analyzed using multi-parallel qPCR. These data are summarized in Table 1 below. Not all samples were cultured, extracted, and/or run on TAC due to budget restraints.
Detection and quantification of enteric pathogens and fecal indicator bacteria of soil samples
To assess patterns of microbial contamination across different environmental contexts, we evaluated soil from sites identified as impacted or unimpacted by wastewater. Fig 1 summarizes the comparison of cultured fecal indicator bacteria (E. coli, Enterococcus, and total coliform) and molecular screening via multi-parallel qPCR of enteric pathogens and human-associated markers in impacted and unimpacted soil samples.
Enteric pathogens are found in detectable concentrations in impacted and unimpacted soil. Wilcoxon rank sum p-values are shown for pairwise comparisons between impacted and unimpacted soil samples analyzed via culture by IDEXX and molecular methods by multi-parallel qPCR (p < 0.05).
Of the 19 soil samples examined by culture, total coliforms were detected in 19 samples (100%) at a range of 1,120 to > 2,419.6 MPN/g, E. coli was detected in 12 samples (63%) at a range of 2–2,247 MPN/g, Enterococcus was detected in 19 samples (100%), at a range of 41 to > 2,419.6 MPN/g (Table 2).
Neither the prevalence nor the concentration of total coliform or Enterococcus was significantly different between the sites assumed to be impacted and the sites assumed to be unimpacted by failing or non-existent wastewater treatment systems (Fig 1). However, the prevalence and concentration of E. coli were significantly higher for samples assumed to be impacted by failing or non-existent wastewater treatment systems than for samples assumed to be unimpacted, but not significant after a false discovery rate (FDR) correction as seen in S5 Data.
From the 49 soil samples analyzed by multi-parallel qPCR that met all the quality control requirements, 11 samples (22%) were found to contain detectable levels of the ybbW gene target for E. coli. Of the 51 soil samples that met all the quality control requirements, 31 samples (61%) were found to contain detectable levels of the gene targets specific to Enterococcus spp. lsrRNA (Table 3). The concentration of E. coli ranged from 1.23 x 103 copies per gram to 8.27 x 105 copies gram. The concentration of Enterococcus ranged from 1.60 x 102 copies per gram to 1.13 x 106 copies per gram.
There was no significant difference in prevalence or concentration of these indicators, E. coli. ybbW and Enterococcus spp. lsrRNA, between the sites assumed to be impacted and the sites assumed to be unimpacted by failing or non-existent wastewater treatment systems, as seen in Fig 1.
The only pathogen detected in both impacted and unimpacted soil samples besides E. coli ybbW and Enterococcus spp. lsrRNA was Acanthamoeba spp. Of the 38 impacted soil samples analyzed that met all the quality control requirements, 37 samples (97%) were found to contain detectable levels of the gene target selected for the Acanthamoeba spp. assay with concentrations ranging from 1.22 x 103 copies per g to 6.06 x 105 copies per g. Of the 7 unimpacted soil samples analyzed that met all the quality control requirements, 7 samples (100%) were found to contain detectable levels of Acanthamoeba spp. with concentrations ranging from 2.86 x 104 copies per g to 2.61 x 105 copies per g. There was no significant difference in prevalence or concentration of Acanthamoeba spp. between the sites assumed to be impacted and the sites assumed to be unimpacted by failing or non-existent wastewater treatment systems, as seen in Fig 1.
Blastocystis spp., Cryptosporidium spp., and Balantidium coli were only detected in impacted soil samples. Blastocystis spp. were detected in 3 of 33 impacted soil samples that met all the quality control requirements (9%), with concentrations ranging from 2.12 x 103 copies per g to 4.08 x 103 copies per g. Cryptosporidium spp. were detected in 1 of 35 impacted soil samples that met all the quality control requirements (3%) at 7.33 x 102 copies per g. Balantidium coli was detected in 1 of 43 impacted soil samples that met all the quality control requirements (2%) at 2.87 x 104 copies per g. There was no significant difference in prevalence or concentration of Blastocystis spp., Cryptosporidium spp., or Balantidium coli between the sites assumed to be impacted and the sites assumed to be unimpacted by failing or non-existent wastewater treatment systems, as seen in Fig 1.
Rotavirus and Astroviridae were only detected in unimpacted soil samples. Rotavirus was detected in 1 of 7 unimpacted soil samples that met all the quality control requirements (14%) at 2.45 x 103 copies per g. Astroviridae were detected in 1 of 7 unimpacted soil samples that met all the quality control requirements (14%) at 5.57 x 103 copies per g. There was a significantly greater difference in prevalence and concentration of Rotavirus and Astroviridae at the sites assumed to be unimpacted and the sites assumed to be impacted by failing or non-existent wastewater treatment systems, but not significant after a false discovery rate (FDR) correction as seen in S5 Data.
All soil samples were also screened for host-associated genetic markers (MST markers). Two soil samples were found to have a detectable amount of human mitochondrial DNA. Human mitochondrial DNA is a genetic marker that is found in all human cells and can be used to indicate the presence of human biological material and fecal matter [9]. Of the 40 impacted soil samples analyzed that met all the quality control requirements, 1 sample (3%) was found to contain detectable levels of human mitochondrial DNA at 7.35 x 102 copies per g. Of the 6 unimpacted soil samples analyzed that met all the quality control requirements, 1 sample (17%) was found to contain detectable levels of human mitochondrial DNA at 8.76 x 102 copies per g. There was no significant difference in prevalence or concentration of human mitochondrial DNA between the sites assumed to be impacted and the sites assumed to be unimpacted by failing or non-existent wastewater treatment systems, as seen in Fig 1.
Detection and Quantification of Enteric Pathogens and Fecal Indicator Bacteria of Surface Water Samples:
While all the surface water samples are from “impacted” sources, the sampling locations represent two distinct spatial contexts: “localized” (collected near known or suspected discharge points) or “community” (collected from bodies of water such as ponds and streams in the community representative of broader environmental conditions). Fig 2 compares fecal markers, analyzed by culture and molecular methods, and pathogen concentrations in surface water by proximity to suspected discharges, comparing samples classified as either community or localized.
Wilcoxon rank sum p-values are shown for pairwise comparisons between surface water samples collected in the community and localized sites via culture by IDEXX and molecular methods by multi-parallel qPCR (p < 0.05).
Of the 24 surface water samples examined by culture, total coliforms were detected in 24 samples (100%) at a range of 109 to > 2,419.6 MPN/100 mL. Of the 24 surface water samples examined by culture for E. coli, E. coli was detected in 23 samples (96%) at a range of 10 to > 2,419.6 MPN/100 mL. Of the 12 samples examined by culture for Enterococcus, Enterococcus was detected in 12 samples (100%), at a range of 100 to > 2,419.6 MPN/100 mL (Table 2). All surface water samples collected and analyzed by culture were assumed to be in areas impacted by failing or non-existent wastewater treatment systems. The EPA criteria for recreational water quality are 33 MPN/100 mL of Enterococcus and 126 MPN/100 mL of E. coli [10]. All of the surface water samples are above the guideline for Enterococcus (100%), and all but six (75%) are above the guidelines for E. coli Table 4.
Neither the prevalence nor the concentration of E. coli or Enterococcus was significantly different between the community and localized sites (Fig 2). However, the prevalence and concentration of total coliform were significantly greater between the community sites and localized sites, but not significant after a false discovery rate (FDR) correction as seen in S5 Data.
From the 25 surface water samples analyzed by multi-parallel qPCR that met all the quality control requirements, 24 samples (96%) were found to contain detectable levels of the ybbW gene target for E. coli. Of the 24 surface water samples that met all the quality control requirements, 24 samples (96%) were found to contain detectable levels of the gene targets specific to Enterococcus spp. lsrRNA (Table 5). The concentration of E. coli ranged from 1.34 x 103 copies per 100 mL to 7.45 x 107 copies per 100 mL. The concentration of Enterococcus spp. lsrRNA ranged from 2.80 x 101 copies per 100 mL to 9.02 x 106 copies per 100 mL.
There was no significant difference in prevalence or concentration of these indicators, E. coli. ybbW or Enterococcus spp. lsrRNA, between the community and localized sites as seen in Fig 2.
In addition to E. coli ybbW and Enterococcus spp. lsrRNA, the enteric pathogens Acanthamoeba spp., Blastocystis spp., Cryptosporidium spp., and Plesiomonas shigelloides were detected in surface water classified as community and localized samples. Of the 20 surface water samples classified as localized that met all the quality control requirements, 18 samples (90%) were found to contain detectable levels of Acanthamoeba spp. with concentrations ranging from 1.95 x 102 to 3.30 x 104 copies per 100 mL. Of the 6 community surface water samples analyzed that met all the quality control requirements, 4 samples (67%) were found to contain detectable levels of Acanthamoeba spp. with concentrations ranging from 1.24 x 103 to 2.78 x 103 copies per 100 mL. Of the 16 surface water samples classified as localized that met all the quality control requirements, 11 samples (69%) were found to contain detectable levels of Blastocystis spp. with concentrations ranging from 2.04 x 102 to 3.10 x 105 copies per 100 mL. Of the 4 community surface water samples analyzed that met all the quality control requirements, 3 samples (75%) were found to contain detectable levels of Blastocystis spp. with concentrations ranging from 9.96 x 102 to 8.93 x 104 copies per 100 mL. Of the 20 surface water samples classified as localized that met all the quality control requirements, 15 samples (75%) were found to contain detectable levels of Cryptosporidium spp. with concentrations ranging from 4.41 x 102 to 8.30 x 104 copies per 100 mL. Of the 6 community surface water samples analyzed that met all the quality control requirements, 6 samples (100%) were found to contain detectable levels of Cryptosporidium spp. with concentrations ranging from 1.51 x 102 to 1.17 x 104 copies per 100 mL. Of the 20 surface water samples classified as localized that met all the quality control requirements, 1 sample (5%) was found to contain detectable levels of Plesiomonas shigelloides. with concentrations at 3.76 x 102 copies per 100 mL. Of the 6 community surface water samples analyzed that met all the quality control requirements, 2 samples (33%) were found to contain detectable levels of Plesiomonas shigelloides. with concentrations ranging from 1.09 x 102 to 3.09 x 103 copies per 100 mL. There was no significant difference in prevalence or concentration of Acanthamoeba spp., Blastocystis spp., Cryptosporidium spp., or Plesiomonas shigelloides. between the community and localized sites, as seen in Fig 2.
Rotavirus and Yersinia enterocolitica were only detected in community surface water samples. Rotavirus and Yersinia enterocolitica were found in 1 of 6 surface water samples at community sites that met all the quality control requirements (17%) at 8.00 x 102 and 2.94 x 102 copies per 100 mL, respectively. There was no significant difference in prevalence or concentration in Rotavirus or Yersinia enterocolitica between the community and localized sites, as seen in Fig 2.
Adenovirus 40/41 and Balantidium coli were only detected in surface water samples at localized sites. Adenovirus 40/41 was found in 1 of 19 localized surface water samples that met all the quality control requirements (5%) at 1.95 x 102 copies per 100 mL. Balantidium coli was found in 2 of 14 localized surface water samples that met all the quality control requirements (14%), ranging from 1.91 x 106 to 2.33 x 106 copies per 100 mL. There was no significant difference in prevalence or concentration in Adenovirus 40/41 or Balantidium coli between the community and localized sites as seen in Fig 2.
All surface water samples were also screened for host-associated genetic markers (MST markers). Seven surface water samples were found to have a detectable amount of human mitochondrial DNA. Of the 20 surface water samples classified as localized that met all the quality control requirements, 6 samples (30%) were found to contain detectable levels of human mitochondrial DNA with concentrations ranging from 2.72 x 102 to 5.45 x 102 copies per 100 mL. Of the 5 community surface water samples analyzed that met all the quality control requirements, 1 sample (20%) was found to contain detectable levels of human mitochondrial DNA with concentrations at 1.04 x 102 copies per 100 mL. There was no significant difference in prevalence or concentration of human mitochondrial DNA between the community and localized sites, as seen in Fig 2. Six surface water samples were found to have a detectable amount of Human-specific HF183 Bacteroides. Human-specific HF183 Bacteroides is a molecular marker from the human gut microbiome that is used to detect human fecal contamination [11]. Of the 19 surface water samples classified as localized that met all the quality control requirements, 2 samples (11%) were found to contain detectable levels of Human-specific HF183 Bacteroides with concentrations ranging from 2.02 x 103 to 5.34 x 103 copies per 100 mL. Of the 6 community surface water samples analyzed that met all the quality control requirements, 4 samples (68%) were found to contain detectable levels of Human-specific HF183 Bacteroides with concentrations ranging from 1.64 x 102 to 1.18 x 103 copies per 100 mL. The prevalence and concentration of Human-specific HF183 Bacteroides were significantly greater at the community sites when compared to localized sites, but not after a false discovery rate (FDR) correction as seen in Appendix 7.
Effect of rainfall on the detection and quantification of fecal indicator bacteria analyzed multi-parallel qPCR in impacted soil and surface water
We collected samples across three different time points (January, February, and May of 2023) to capture potential microbial variations related to rainfall. We collected samples on January 23rd, after heavy rainfall (38.6 mm in the past 48 hours), February 25th, after minimal rainfall (1.01 mm in the past 48 hours), and May 13th, after moderate rainfall (12.7 mm in the past 48 hours) [12]. The American Meteorological Society (AMS) classifies rain in three categories: light (0–6 mm/day), moderate (6–18 mm/day), and heavy (more than 18 mm/day) [13]. Therefore, according to the AMS, we sampled during all three rainfall categories: heavy (January), light (February), and moderate (May). As seen in S2 Fig, rainfall categories did not significantly impact Enterococcus spp. lsrRNA in either sample type or E. coli ybbW concentrations in impacted soil samples. However, E. coli ybbW concentrations in impacted surface water were significantly greater during moderate when compared to light rainfall conditions. Due to the fact that there was only one sample type where there was a significant difference between rainfall categories, the rest of this paper will be combining the time points for analysis.
Discussion and limitations
The presence and extent of E. coli and Enterococcus in both the culture and molecular data suggest widespread fecal contamination in the study setting. To put the results of this study in context, we can compare the soil culture results to a study conducted in Maputo, Mozambique, where researchers collected and cultured soils from latrine entrances, solid waste areas, and dishwashing areas from 30 household clusters [14]. This is an imperfect comparison for several reasons: the two settings are very different, our soil samples were collected from the broader environment whereas the soil samples in the Capone et al. study were collected at the household level, and the soil samples from each study were cultured using different methods. Despite these differences, the magnitude of culturable E. coli found in soil samples from these two studies is comparable. In the Capone et al. study, log-transformed CFU counts per gram ranged from 0 to 5.3 with a mean of 3.2. In our study, the soils showed log-transformed MPN counts ranging from 0 to 3.35, with a mean of 2.25. These findings suggest that the fecal contamination observed in this study is substantial and comparable in magnitude to levels observed in settings with known sanitation challenges.
While the prevalence of the enteric pathogens examined was low, several targets important to public health were identified, such as Cryptosporidium spp. and Blastocystis spp. Cryptosporidium spp. is a protozoan parasite that can infect a wide range of animal hosts, including birds, reptiles, and mammals. The assay for Cryptosporidium spp. targeted the 18s rRNA region of all human-pathogenic crypto species, including C. hominis, C. parvum, C. meleagridis, C. canis, C. felis, C. muris, and C. suis. It is important to note the possibility of non-specific amplification since the assay is not species-specific. Future analyses will include species-specific assays to interrogate the likely source (human or animal) of Cryptosporidium spp. within these environmental samples. Blastocystis spp. are enteric parasites, known to infect a range of human and animal hosts [15]. Infections of Blastocystis spp. can cause gastrointestinal distress; carriers may also be asymptomatic. Blastocystis spp. are considered to be the cause of emerging infectious disease [15], and their prevalence is thought to be higher than Giardia spp. and Cryptosporidium spp. in the United States and globally [15]. In a study that examined enteric pathogens in children in the Black Belt, the prevalence of Blastocystis spp. in stool was found to be 3.7% (18/488) in a cohort of children from this study region [16]. This gives more evidence that this target is likely to be valuable to measure the effect of the sanitation intervention.
Other pathogens of interest were detected besides Cryptosporidium spp. and Blastocystis spp. Acanthamoeba spp. are amoebae found in soil and water. These amoebae are pathogens that have been implicated in cases of amebic encephalitis, an infection of the central nervous system, and amebic keratitis, an infection of the eye. However, since these amoebae are free-living and commonly found in the environment, we cannot conclusively relate their detection to sanitation infrastructure in the region. Instead, we may continue to monitor this pathogen in future work as a control. Plesiomonas shigelloides has been implicated in cases of childhood diarrhea, and it is commonly found in surface water samples [17]. Balantidium coli, while mostly presenting as asymptomatic, can lead to persistent diarrhea and occasionally to dysentery. It presents most commonly in reservoirs where animals are kept and in areas where sanitation is poor [18]. Astrovirus is also very common globally; it has been shown that 90% of children will be infected with this virus, through the fecal-oral route, by the time they are 9 years old [19]. Rotavirus is another virus that is transmitted through the fecal-oral route and is commonly linked to limited sanitation service [20]. All of these pathogens were also found in the stool of children living in the Black Belt in a previous study [16].
Our study had several limitations. First, we determined assay lower limits of detection (LLOD), using criteria outlined in Sahoo et. al, which defines the y-intercept of the standard curve as “the theoretical limit of detection of the assay” [8] and may differ from other methods of empirically deriving detection limits. We therefore excluded some samples with Ct values near the LLOD as a conservative approach to pathogen detection via molecular methods. We necessarily omitted some presumptive detections at the LLOD as potentially spurious (e.g., one detection of Giardia spp. in surface water) as a conservative approach. Details on samples can be found in S6 Data. Second, the typology of impacted and unimpacted sampling sites was limited by the potential subjectivity of researcher observations in the field. Since all environmental samples were collected in public spaces around the community, we were not certain to what degree each site was impacted by the sanitation status in the community, beyond directly observable criteria such as proximity to a failing septic system or direct surface discharge. Sites may have been categorized incorrectly, for example, if unobserved discharges were present. Third, limiting our sampling to publicly accessible sites may have constrained our analysis by excluding potentially important (but inaccessible) sites on private property. For example, households with failing septic systems may have sewage pooling in the yard around the tank or drain fields. In this case, the “impacted” site would be on private property and therefore unavailable, given the restraints of this study. In future sampling, researchers will be sampling from sites at households that have enrolled in a prospective impact study. Finally, the sample size for culture data was relatively small, limiting statistical power. Out of the 123 samples that were collected across three time points, only 43 samples (35%) were cultured at all, and out of the samples that were cultured, not all of them were cultured for the same organism. This more than likely affected the analysis of the culture data and limited the conclusions we could draw about what these samples tell us about the potential exposure relevance of sanitation infrastructure in this area.
Despite limitations, initial results show that widespread sanitation deficits in the study’s setting present pathogen hazards in the community. These hazards are most commonly associated with uncontained fecal waste discharges. Planned sanitation infrastructure expansion may be effective in limiting the potential for exposures to residents, which in turn could improve overall health and well-being.
Supporting information
S1 Table. qPCR primer and probe sequences for TaqMan array card.
https://doi.org/10.1371/journal.pwat.0000393.s001
(DOCX)
S1 Fig. Example of the membrane filtration set up.
https://doi.org/10.1371/journal.pwat.0000393.s002
(DOCX)
S2 Fig. Quantification of Fecal Indicator Bacteria (E. coli ybbW and Enterococcus spp. lsrRNA) from impacted Environmental Samples in January, February, and May 2023.
https://doi.org/10.1371/journal.pwat.0000393.s003
(DOCX)
S1 Data. Raw culture data for environmental samples analyzed via IDEXX.
https://doi.org/10.1371/journal.pwat.0000393.s004
(XLSX)
S2 Data. General descriptive information for environmental samples.
https://doi.org/10.1371/journal.pwat.0000393.s005
(XLSX)
S3 Data. Quality control findings: Detection of targets in NEC and PEC samples.
https://doi.org/10.1371/journal.pwat.0000393.s006
(CSV)
S4 Data. Raw qPCR data for environmental samples analyzed via TaqMan array card.
https://doi.org/10.1371/journal.pwat.0000393.s007
(CSV)
S6 Data. Excluded samples with Ct values near the limit of detection.
https://doi.org/10.1371/journal.pwat.0000393.s009
(CSV)
Acknowledgments
We are deeply grateful to the community members who generously welcomed us to their town and made this project possible through their time and trust. In addition to those that the authors (Harmon and Lott) collected and analyzed, some samples were collected with the help of Todd Hester, MS and analyzed by Corinne Baroni, EIT. We thank Kevin Zhu, PhD who improved this work by offering his expertise.
References
- 1. Maxcy-Brown J, Capone D, Elliott MA. Characterizing the nature and extent of access to unsafely managed sanitation in the United States. Nat Water. 2023;1(11):915–28.
- 2. Flowers CC, Ward JK, Winkler IT. Flushed and forgotten: sanitation and wastewater in rural communities in the US. Institute for the study of human rights. Accessed 2022 November 17. https://www.humanrightscolumbia.org/news/flushed-and-forgotten-sanitation-and-wastewater-rural-communities-us
- 3. Maxcy-Brown J, Elliott MA, Krometis LA, Brown J, White KD, Lall U. Making waves: Right in our backyard- surface discharge of untreated wastewater from homes in the United States. Water Res. 2021;190:116647. pmid:33310443
- 4.
Office of Public Affairs | Departments of Justice and Health and Human Services Announce Interim Resolution Agreement in Environmental Justice Investigation of Alabama Department of Public Health. 2023. Accessed 2023 December 12. https://www.justice.gov/opa/pr/departments-justice-and-health-and-human-services-announce-interim-resolution-agreement
- 5. Okeowo A. The heavy toll of the Black Belt’s wastewater crisis. The New Yorker. 2020.
- 6.
Office of Public Affairs. Justice department announces environmental justice investigation into alabama department of public health and lowndes county health department. 2021. Accessed 2024 August 18. https://www.justice.gov/opa/pr/justice-department-announces-environmental-justice-investigation-alabama-department-public
- 7. Baker KK, Senesac R, Sewell D, Sen Gupta A, Cumming O, Mumma J. Fecal fingerprints of enteric pathogen contamination in public environments of kisumu, kenya, associated with human sanitation conditions and domestic animals. Environ Sci Technol. 2018;52(18):10263–74. pmid:30106283
- 8. Sahoo S, Mandal S, Das P, Bhattacharya S, Chandy M. An analysis of the standard curve parameters of cytomegalovirus, BK virus and hepatitis B virus quantitative polymerase chain reaction from a clinical virology laboratory in eastern India. Indian J Med Microbiol. 2022;40(1):81–5. pmid:34602291
- 9. Zhu K, Suttner B, Pickering A, Konstantinidis KT, Brown J. A novel droplet digital PCR human mtDNA assay for fecal source tracking. Water Res. 2020;183:116085. pmid:32750535
- 10.
Environmental Protection Agency. Recreational water quality criteria.
- 11. Seurinck S, Defoirdt T, Verstraete W, Siciliano SD. Detection and quantification of the human-specific HF183 Bacteroides 16S rRNA genetic marker with real-time PCR for assessment of human faecal pollution in freshwater. Environ Microbiol. 2005;7(2):249–59. pmid:15658992
- 12.
National Centers for Environmental Information. National Centers for Environmental Information (NCEI). Accessed 2025 March 20. https://www.ncei.noaa.gov/
- 13. Sala CM, Zorzetto E, Li L. The contrasting roles of land and oceanic moisture contributions to summer rainfall over the southeastern United States. J Geophys Res Atmospheres. 2025;130(3).
- 14. Capone D, Bivins A, Knee J, Cumming O, Nalá R, Brown J. Quantitative Microbial Risk Assessment of Pediatric Infections Attributable to Ingestion of Fecally Contaminated Domestic Soils in Low-Income Urban Maputo, Mozambique. Environ Sci Technol. 2021;55(3):1941–52. pmid:33472364
- 15. Wawrzyniak I, Poirier P, Viscogliosi E, Dionigia M, Texier C, Delbac F, et al. Blastocystis, an unrecognized parasite: an overview of pathogenesis and diagnosis. Ther Adv Infect Dis. 2013;1(5):167–78. pmid:25165551
- 16. Capone D, Bakare T, Barker T, Chatham AH, Clark R, Copperthwaite L, et al. Risk factors for enteric pathogen exposure among children in black belt region of Alabama, USA. Emerg Infect Dis. 2023;29(12).
- 17.
Janda JM. Plesiomonas shigelloides. In: Reference module in food science. Elsevier; 2023.
- 18.
Percival SL, Yates MV, Williams DW, Chalmers RM, Gray NF. Chapter Fifteen - Balantidium coli. In: Microbiology of waterborne diseases. London: Academic Press; 2014. 277–86.
- 19. Moser L, Schultz-Cherry S. Encycl Virol. 2008. 204–10.
- 20.
Centers for Disease Control and Prevention. Learn more about rotavirus. 2021. Accessed 2023 December 14. https://www.cdc.gov/rotavirus/index.html