Evaluation of surface sampling techniques for collection of Bacillus spores on common drinking water pipe materials

Abstract

Drinking water utilities may face biological contamination of the distribution system from a natural incident or deliberate contamination. Determining the extent of contamination or the efficacy of decontamination is a challenge, because it may require sampling of the wetted surfaces of distribution infrastructure. This study evaluated two sampling techniques that utilities might use to sample exhumed pipe sections. Polyvinyl chloride (PVC), cement-lined ductile iron, and ductile iron pipe coupons (3 cm × 14 cm) cut from new water main piping were conditioned for three months in dechlorinated Cincinnati, Ohio tap water. Coupons were spiked with Bacillus atrophaeus subsp. globigii, a surrogate for Bacillus anthracis. Brushing and scraping were used to recover the inoculated spores from the coupons. Mean recoveries for all materials ranged from 37 ± 30% to 43 ± 20% for brushing vs. 24 ± 10% to 51 ± 29% for scraping. On cement-lined pipe, brushing yielded a significantly different recovery than scraping. No differences were seen between brushing and scraping the PVC and iron pipe coupons. Mean brushing and scraping recoveries from PVC coupons were more variable than mean recoveries from cement-lined and iron coupons. Spore retention differed between pipe materials and the presence of established biofilms also had an impact. Conditioned PVC coupons (with established biofilms) had significantly lower spore retention (31 ± 11%) than conditioned cement-lined coupons (61 ± 14%) and conditioned iron coupons (71 ± 8%).

---

Environmental impact

Treated water available to communities through a water supply network is an important resource that not only impacts the human health in an area, but also the natural environment. Ultimately, generated wastewater must be collected, treated, and discharged into the watershed. Thus, if the system is intentionally or unintentionally contaminated with a persistent pathogen, understanding the fate and transport of the pathogen within the system, as well as outside the system, becomes crucial. It is vital to assess the extent of deposition on wetted internal surfaces like pipes, as well as the efficacy of any attempts to decontaminate those surfaces. In order to take into account pathogen and particle interaction with corrosion and/or biofilm micro-environments, it is essential to evaluate surface sampling methods on various materials, each with unique surface properties.

Introduction

In the past decade, especially following the terrorist events of September 11, 2001 and subsequent anthrax mailings, the possibility that persistent biological agents might be used to disable critical infrastructures has received increasing attention. This paper examines the biological aspect of this scenario with a specific focus on bacterial endospores. Even in the presence of chlorine, bacterial endospores are persistent on internal pipe surfaces. It has been shown that they are able to persist for long periods even in the presence of chlorine following a contamination event.^1,2 It has also been shown that the presence of biofilms and corroding surfaces in drinking water distribution systems may influence adhesion of planktonic bacteria.³

There are currently protocols for collecting bulk water samples from a distribution system to detect microbial contamination. However, there are no standardized methods that include sampling of the internal surfaces of drinking water infrastructure for either intentional or unintentional contamination. The purpose of this study was to evaluate the precision and recovery efficiency of two different surface sampling techniques that could potentially be used to sample the internal surfaces of water distribution pipes contaminated with bacterial endospores. Research on drinking water biofilms has generally used standard microbiological surface sampling techniques such as swabbing to recover the biofilm from wetted pipe surfaces.⁴ However, to evaluate other surface sampling methods, brushing and scraping of three different types of pipe material were compared to determine which was the most effective in removing spores. Spore retention on different pipe materials was also evaluated.

Experimental

Pipe material

PVC, unlined ductile iron, and cement-lined ductile iron were used in this study. These materials were chosen based on the results of a database search of the 2002 American Water Works Association (AWWA) survey of 337 small, medium-sized, and large utilities in the United States and Canada. The search showed that PVC, iron (including cast iron), and cement-lined pipes were found in the United States' and Canada's distribution systems 8%, 15%, and 50% of the time, respectively.⁵ Each of the pipe materials came from new 20.32 cm water mains obtained from IPEX (Mississauga, ON) (PVC) and American Ductile Iron Pipe Company (Birmingham, AL) (ductile iron and cement-lined ductile iron). Coupons (3 cm × 13 cm) were cut from 20.32 cm pipes using a high-pressure water jet cutter, which kept the coupon surfaces cool as they were cut. Each coupon provided a 39 cm² area to sample. Prior to use, coupons were washed and disinfected using the National Sanitation Foundation (NSF) International procedure for conditioning drinking water contact materials prior to determining the toxicity of the surface.⁶ This involved scrubbing the coupons in tap water with a test tube brush to remove any debris, spraying the coupons with 200 mg L⁻¹ sodium hypochlorite solution coating all surfaces of the coupon, and then after 30 min, rinsing the coupons in deionized water. The coupons were then placed into the recirculation tank.

Recirculation tank

A 1136-liter recirculation tank, with a residence time of approximately 14 h, was used to expose the coupons to dechlorinated Cincinnati tap water. The average chlorine concentration of the incoming tap water was 0.99 mg L⁻¹ with a standard deviation of 0.11 during the time period. A chemical feed pump delivered approximately five grams of sodium thiosulfate into the recirculation tank every 24 h to achieve 0.01 to 0.05 mg L⁻¹ of detectable free chlorine in the tank. The tank chlorine levels were measured weekly and the sodium thiosulfate feed rate was adjusted to keep the chlorine residual in the tank between 0.01 and 0.05 mg L⁻¹. The tank was gently mixed with a small electric mixer, which exposed the coupons to some shear forces.

Stainless steel racks were built to suspend each type of coupon material vertically in the tank. Each rack supported approximately 30 coupons stacked on top of each other and held in place with a stainless steel rod running through the middle of the rack. The curvature of the coupons made it possible for the rod to hold the coupons in place with contact only to the edges of the coupon. The rack and coupons are shown in Fig. 1.

Fig. 1 Racks holding coupons in the recirculation tank. Pipe material (from left to right) is PVC, cement-lined, and iron.

Biofilm enumeration

Three coupons of each type of pipe material were sampled to collect biofilms for enumeration. The biofilm was brushed from the surfaces using adult soft toothbrushes (CVS Pharmacy, Inc., Woonsocket, RI, catalog no. 100982). Each coupon was brushed and subsequently washed a total of four times using a total of 50 mL volume of the phosphate buffer (Standard Methods Section 9216 B).⁷ Samples were diluted in phosphate buffer and cultured on R2A agar using the spread plate method and incubated at room temperature (21–24) for 7 days (Section 9215 C).⁷ Mean HPC counts for the biofilm developed on each material were: PVC, 1 × 10⁵ CFU/cm², cement-lined, 4 × 10⁵ CFU/cm², and iron, 3 × 10⁸ CFU/cm². Simoes et al.⁸ found that biofilms grown in chlorine-free tap water on PVC under laminar conditions generated 10⁴ CFU/cm². Also, according to a literature review conducted by Ollos et al.,⁹ a bacterial cell count of 10⁴ CFU/cm² or higher represents a viable drinking water biofilm.

Bacillus spore selection

A literature search was conducted to find a spore-forming Bacillus species that could be used as a surrogate for B. anthracis. One study showed that Bacillus atrophaeus subsp. globigii (BG) spores have a mean CT (C is the concentration of chlorine in mg/liter, and T is the exposure time in minutes) value higher than the mean CT values for B. anthracis Sterne, B. cereus, and B. thuringiensis along with the virulent Ames strain of B. anthracis.¹⁰ Because of its increased resistance to chlorine, BG was selected since it represents a conservative surrogate for future drinking water pipe disinfection research.

Spore preparation, growth, and enumeration

BG spores were obtained from the U.S. Army's Dugway Proving Ground (Dugway, Utah) and were grown according to the methods described in Coroller et al.¹¹ and Nicholson and Setlow.¹² Generic spore media was inoculated with vegetative BG cells and incubated for five days at 35 °C with gentle shaking in a rotary shaker. Purified BG endospores were produced using gradient separation as described by Nicholson and Setlow¹² and the presence of spores was confirmed using phase contrast microscopy (<0.1% vegetative cells). Spores were stored in 40% ethanol at 4 °C until use. Spores used in this study came from the same batch and container.

Spores were enumerated using serial dilution and the spread plate method (Section 9215 C)⁷ with 0.1 mL of the sample on tryptic soy agar (TSA). The diluent used was phosphate buffer (Standard Methods Section 9216 B)⁷ with 0.01% Tween® 80 (Fisher Scientific, Pittsburgh, PA, catalog no. AC278630000). Tween® 80 is a surfactant that has been shown to improve spore recovery.¹³ All samples, including quality control (QC) samples, were spread plated in triplicate with the exception of the tap water spore suspension, which was spread plated using 5 replicate plates to generate a more precise value. Spore samples from coupons were heat shocked at 80 °C for 10 min to inactivate vegetative organisms. Samples were cooled and vortexed prior to plating. Plates were incubated for 24 ± 2 h at 35 °C. Because it forms orange colonies, BG was easily distinguished from other biofilm organisms that survived the heat shock.

Tap water spore suspensions

Fresh spore suspensions were made on the day of the experiment. Suspensions were made by adding 40 μL of the concentrated spore stock suspension, described above, to 50 mL of dechlorinated Cincinnati tap water resulting in a 6 × 10⁵ CFU/mL suspension. Dechlorinated water from the tank was used as the diluent so that the recirculation tank water quality was represented.

Water parameter monitoring

Conductivity, temperature, and pH of the recirculation tank water were measured at each coupon sampling event. Average values of 340 μS cm⁻¹ (SD = 190) for conductivity and 13.7 °C (SD = 0.015) for temperature were recorded using an Extech conductivity meter (Extech instruments, Waltham, MA, catalog no. 407303). pH ranged from 8.22–9.00 with a median value of 8.27 and was measured using an Orion 555A pH/ORP/Conductivity Meter (Thermo Fisher Scientific Inc., Waltham, MA). Total chlorine was measured twice a week using the Hach total chlorine method 8167(4500-Cl Chlorine DPD Colorimetric Method)⁷ using Hach DPD free chlorine reagent AccuVac® ampules (Hach, Loveland, CO, Catalog no. 25020-25), with a Nalco 2800 DR spectrophotometer (Nalco Company, Naperville, IL). Chlorine was found to be less than 0.05 mg L⁻¹ in the tank due to the introduction of sodium thiosulfate. On four occasions, additional recirculation tank water samples were collected, heat shocked, and spread plated on TSA to determine whether spores were present in the bulk water in the tank.

Coupon collection and spore loading

Individual coupons were removed from the coupon rack and placed into a sterile 1000 mL beaker while still submerged in the tank. Each coupon was then placed in a sterile 14 cm diameter Petri dish and was leveled using a sterile level. A two step procedure was used to load the coupons with spores: the suspension on the coupons was spiked; then, the coupons were rinsed to remove spores that had not been retained on the surface.

Spore spiking. After the coupon was leveled, it was spiked with 0.5 mL of dechlorinated tap water containing approximately 3 × 10⁵ spores. The spore suspension on the coupon was then spread with a sterile 250 μm pipette tip to ensure that all parts of the coupon were exposed to the suspension. The suspension was left on the coupon for 20 min.

Coupon rinsing. After 20 min, the contaminated surface of the coupon was manually rinsed with 100 mL of dechlorinated tap water. Dechlorinated tap water was pumped from a sterile 1 litre beaker using a peristaltic pump calibrated to pump approximately 120 mL per minute. The rinsing procedure was timed to ensure that 100 mL of water was used in the rinsing process. Spores rinsed off the coupon were collected in the same Petri dish that was used to hold the coupon while it was spiked with the spore suspension. This rinsing procedure removed any spores not retained (or weakly retained) on the coupon surface.

Coupon sampling

Sampling techniques were compared on the three different pipe surfaces to determine differences in recovery and precision. These methods were chosen due to their low cost and availability, as well as their potential to remove biofilm and loose corrosion from porous and non-porous surfaces.

Toothbrush. A search was conducted to find low-cost, disposable brushes with bristles on one side that could be used to remove loose biofilm and scale from porous wetted surfaces. It was found that brushes designed for scientific uses, such as test tube brushes, were either too expensive or were designed with bristles all around the shaft. Soft adult toothbrushes were the right design and relatively inexpensive (CVS Pharmacy, Inc., Woonsocket, RI, catalog no. 100982). These toothbrushes, available in bulk for less than a dollar per unit, were used for all of the brushing experiments.

Cell scraper. The cell scraper was chosen because, like the toothbrush, it was disposable, easy to use, and relatively inexpensive when purchased in bulk at less than $2.50 a unit. Designed for use in harvesting cells from culture vessels, it was logical choice for removing corrosion, biofilms, and adhered contaminants from drinking water pipes. Sterile Fisherbrand cell scrapers (Fisher Scientific, Pittsburgh, PA, catalog no. 08-773-2) with plastic 1.8 cm blades were used for the scraping experiments.

Spore sampling from coupon surface. The brushing and scraping procedure was the same for each coupon. Coupons were brushed with 10 downward strokes, ensuring that all of the material brushed from the coupon was collected in the Petri dish. After the 10 brush strokes, the coupon and brush were washed with phosphate buffer and 0.01% Tween® 80 contained in a sterile 150 mL squirt bottle. The scraping procedure consisted of scraping the coupon from both directions horizontally so that the biofilm/spores were scraped to the center of the coupon. After the entire surface of the coupon was scraped, it was washed with phosphate buffer and 0.01% Tween® 80 in the same manner that was done with the brushing samples.

Each coupon was brushed or scraped, then washed, four times. A combined 50 mL of the phosphate buffer and 0.01% Tween® 80 was used in the process, resulting in a 50 mL sample that was collected in the Petri dish. This suspension was transferred to a 100 mL sample container. The quantity of spores found in this sample was compared to the number of spores that were calculated to be on the surface prior to sampling, resulting in a recovery efficiency value.

Other sampling and quality control checks

Additional spore sampling was conducted to determine where losses may have occurred throughout the sampling and analysis process. Quality control checks were also routinely done to check for potential contamination issues.

Sides and back check. To determine the extent of contamination on the backs and sides of the coupons throughout sampling, the sides and back of each coupon were brushed using 50 mL of phosphate buffer with 0.01% Tween® 80 as a diluent to rinse the brush and coupon during and after brushing. This procedure for sampling the sides and back of each coupon was the same as the brushing procedure described in the previous section. The resulting spore suspension was collected in a sterile Petri dish and transferred to a 100 mL sample container.

Petri dish check. In order to determine the extent of spore retention on the walls and bottom of the Petri dishes used to collect samples, the dishes were swabbed using sterile macrofoam swabs (ITW Texwipe® CleanTip® Swabs, Fisher Scientific, Pittsburgh, PA, catalog no. 18-359) after transferring the samples to 100 mL sample containers. Each swab was placed in a sterile 50 mL test tube containing 10 mL of phosphate buffer and 0.01% Tween® 80. The tube was then vortexed for 2 min using 10-second intervals to remove the spores from the swabs. Swabs were then rolled on the inside of the tubes to remove as much liquid from the swab as possible and the swabs were discarded. Tubes containing spore suspensions were then heat shocked and 0.1 mL of the suspension was then spread plated on TSA to enumerate spores.

Brush and scraper check. Spore losses due to retention on brushes and scrapers were checked after the brushes and scrapers were used to sample the coupons. Each brush or scraper used in the experiments was placed in a sterile 50 mL test tube containing 10 mL of phosphate buffer and 0.01% Tween® 80 and was vortexed as described above for the swabs. Spores removed from the brushes and scrapers were enumerated using the same procedure as was used with the swabs.

Matrix effect check. Corrosion and/or biofilm collected from one conditioned coupon of each pipe material were used to determine the impact of the matrix on spore recovery. 6 × 10⁴ CFU were added directly to 100 mL of phosphate buffer containing the scraped matrix and 0.01% Tween® 80. Undiluted samples from these treatments and a phosphate buffer/Tween® 80 control were enumerated for BG spores and recoveries were calculated.

Conditioning control coupons. To determine the effect the 3 month conditioning process had on spore retention, 12 additional coupons (4 each of PVC, unlined ductile iron, and cement-lined ductile iron) were washed and disinfected using the NSF International procedure⁶ described in the pipe material section. The coupons were then placed in 3 L buckets of autoclave sterilized Cincinnati tap water for 24 h to hydrate the surfaces. These coupons were then spiked with spores and rinsed using the same procedure described in the proceeding section.

Contamination checks. To ensure that TSA, R2A plates were sterile, blank plates were incubated along with samples during every sampling event. Phosphate buffer used in the experiments was spread plated on TSA to determine sterility. On four occasions, coupons of each pipe material were collected from the recirculation tank to take HPC counts as described in the biofilm enumeration section. In addition to biofilm enumeration, the biofilm/corrosion collected from the coupons was heat shocked and the resulting sample was spread plated on TSA to verify that the coupons were not contaminated with BG spores prior to experimentation. Fresh spore suspensions were made from the same stock suspension each time coupons were spiked. Each suspension made was spread plated to check spore viability and to enumerate the number of spores per unit volume. The mean spore concentration for all of the spore suspensions made was 3.1 × 10⁵ with a standard deviation of 4.8 × 10⁴. Additionally, prior to conducting sampling experiments, biofilm and corrosion from one coupon of each pipe material was collected, spiked with BG spores, and enumerated to determine the matrix effect from the corrosion and biofilm.

Mass balance calculation

A mass balance approach was used to calculate the recovery efficiency of the sampling methods. The specific components (Spike C_o, rinse C_r, brush or scrape C_s, and sides and back C_b) are summarized in Table 1.

Table 1 Samples from spore suspensions, coupons, and sampling equipment used in mass balance calculation^a

	Samples	Description	Reason for sample
a All samples are spread plated in triplicate.
Coupon inoculation spore suspension (Co)	Sample	Tap water spore suspensions used to inoculate coupons were quantified to determine the initial number of spores inoculated onto coupon	To quantify the total number of spores that were spiked onto the coupon
Rinse (Cr)	Sample	Following inoculation, coupons were rinsed with 100 mL of dechlorinated tap water to remove spores that have not been retained on the surface	To quantify the number of spores that were rinsed off of the coupon with dechlorinated tap water
	Petri dish check	Determination of the number of spores that stay attached to the Petri dish	To quantify spores retained to Petri dish from rinse sample
Brush or scrape (Cs)	Sample	Following rinsing, the side of the coupon corresponding to the internal pipe surface was brushed or scraped	To quantify spores that were brushed or scraped from the coupon
	Toothbrush or scraper check	Determination of the number of spores that stay attached to the toothbrushes and scrapers	To quantify spores on toothbrush or scraper
	Petri dish check	Determination of the number of spores that stay attached to the Petri dish	To quantify spores retained to Petri dish from toothbrush or scraper sample
Sides and back (Cb)	Sample	Following brushing or scraping, the sides and back of each coupon were brushed	To quantify the number of spores found on the sides and back of each coupon
	Toothbrush check	Determination of the number of spores that stay attached to the toothbrushes and scrapers	To quantify spores on toothbrush
	Petri dish check	Determination of the number of spores that stay attached to the Petri dish	To quantify spores retained on Petri dishes from sides and back sample

---

Spore recovery efficiency (RE) was computed as shown in eqn (1) using the terms defined in Table 1. Percent retention (PR) of spores was computed as shown in eqn (2).

Experimental design

Seventy two coupons including quality control coupons were used in the experiments. Twelve of the 72 coupons were conditioning control coupons described above. An additional twelve of the recirculation tank conditioned coupons were sacrificed in order to enumerate biofilm and conduct quality control checks. The remaining 48 coupons were used for the sampling experiments as shown in Table 2.

Table 2 Experimental design for pipe coupons^a

	Conditioning control coupons (no biofilm)	Recirculation tank conditioned coupons (established biofilm)			TOTAL
		Biofilm and procedural blank samples	Brushing	Scraping
a Numbers in parentheses are actual numbers of coupons used in recovery efficiency and percent retention experiments.
PVC	4	4	8 (7)	8 (7)	24 (18)
Cement - lined	4	4	8 (8)	8 (7)	24 (19)
Iron	4	4	8 (8)	8 (7)	24 (19)
					72 (56)

---

Statistical analysis and calculations

The experiment was designed for analysis using a 2-way analysis of variance (ANOVA). However, spore recovery efficiency (RE) data were not equally variable among pipe materials (see data in Fig. 2), violating one of the underlying assumptions for an ANOVA analysis. Therefore, recovery efficiency data were analyzed with t-tests using SigmaPlot® software (SYSTAT Software, Inc., San Jose, CA) with a 0.05 significance level. The experimental design enabled testing of the null hypothesis that assumed there were no differences in recovery efficiencies between the sampling techniques for each pipe material. Percent retention (eqn (2)) was tested using a one-way analysis of variance. The null hypothesis for this design was the assumption that there would be no significant differences in retention among pipe materials.

Fig. 2 Recovery of B. globigii spores from PVC (P), cement-lined (C), and iron (I) pipe coupons using brushing (Br) and scraping (S). The line in the middle of the box indicates the median and the ends of the boxes define the 25th and 75th percentiles. Whisker bars are not shown due to sample size ≤ n = 8.

Results

Recovery efficiency

Recovery efficiency values were calculated using eqn (1) and are shown in Table 3 and Fig. 2. Data were then arcsin squareroot transformed prior to conducting statistical tests. Mean recovery efficiency values for brushing and scraping for combined pipe materials were 0.41 (sd = 0.20) for brushing, and 0.40 (sd = 0.21) for scraping. Combined recovery efficiency for brushing and scraping on all pipe materials was 0.40 (sd = 0.20). t-test results showed significant differences between cement-lined brushing and scraping (P = 0.003). There were no significant differences between brushing and scraping for PVC (P = 0.092) and iron (P = 0.691).

Table 3 Mean, median, and standard deviation values for recovery efficiency of B. globigii spores from conditioned PVC, cement-lined, and iron pipe coupons

	PVC		Cement		Iron
	Brush	Scrape	Brush	Scrape	Brush	Scrape
Sample size	7	7	8	7	8	6
Mean	0.37	0.51	0.42	0.24	0.43	0.46
Median	0.29	0.47	0.42	0.21	0.44	0.46
SD	0.30	0.29	0.09	0.10	0.20	0.04

---

Percent retention

The percent retention of B. globigii was calculated using eqn (2). Fig. 3 illustrates the differences in retention between the various pipe materials.

Fig. 3 Comparison of B. globigii retention on PVC (P), cement-lined (C), and iron (I) conditioned (Con) and unconditioned pipe coupons. The line in the middle of the box indicates the median and the ends of the boxes define the 25th and 75th percentiles. Whisker bars define the 10th and 90th percentiles.

There were significant differences between PVC and cement-lined and PVC and iron coupons that had been conditioned. In contrast, no significant differences between conditioned cement-lined and iron pipe coupons were found. Additionally, significant differences were seen between conditioned and unconditioned PVC, and conditioned and unconditioned iron pipe material.

Sides and back. As part of the mass balance calculation, the sides and backs of all coupons were brushed and rinsed to determine the number of spores that may have migrated there during the spiking procedure or during the rinsing, or the brushing or scraping process. It was observed that spore suspensions on the PVC coupons pooled up on the surface, whereas the cement-lined pipe coupons either absorbed or wicked the suspension off the surface during the 20 min contact time. Data generated from analyses of the cement-lined pipe coupon sides and back show that 11.8% of the inoculated spores were recovered there. In contrast, 0.14% and 0.10% of the spores were recovered from the sides and back of the iron and PVC, respectively.

Relative proportions of spores in each sampling step. Mean spore recovery for each of the steps is summarized in Fig. 4. The height of the bar represents the mean number of spores spiked onto the coupon. Sections within the bar represent the relative proportions of spores that were recovered in each of the sampling steps described in Table 1.

Fig. 4 Mean spore recovery for all techniques. Each bar represents the mean number of spores spiked on the coupon (8 brushed (Br) and 7 scraped (S) coupons). Relative numbers of spores recovered as mean values for C_r, C_s, and C_b are shown in relation to the number of spores spiked on PVC (P), cement-lined (C), and iron (I) coupons.

Spore losses to petri dishes, toothbrushes, and scrapers

Spore losses to retention on Petri dishes, toothbrushes, and scrapers, are shown in Table 4. Percentages indicate the number of spores recovered with respect to the total number spiked on the coupons.

Table 4 Mean counts of spores sampled from Petri dishes, toothbrushes, and scrapers

	Rinse Petri dishes (no Tween®)	Scrape Petri dishes	Sides and back Petri dishes	Toothbrushes	Scrapers
Avg. No. spores recovered (SD)	13557 (10172)	4078 (4851)	3423 (5651)	515 (482)	568 (912)
Percentage of spores recovered	4.30%	1.30%	1.10%	0.20%	0.20%

---

Impact of Tween® 80 and corrosion on spore recovery

In preliminary studies, it was determined that there were significant losses due to spores clumping and adhesion to sample containers. Thus, 0.01% Tween® 80, was included in all phosphate buffer used to collect and enumerate spores. The purpose of the spore spiking and coupon rinsing steps was to simulate contaminated tap water and subsequent water movement in a pipe. Therefore, out of necessity, the spore spiking and rinsing steps did not include Tween® 80. To confirm the effect that adding a surfactant has on BG spore recovery, a phosphate buffer spore suspension was split into two identical samples and Tween® 80 was added to one of the samples. The suspensions were then spread plated on TSA. Approximately two times the number of spores were recovered in the presence of Tween® 80 (3153 per mL with vs. 1603 spores per mL without).

In contrast to the effects of Tween® 80 on spore recovery, iron corrosion was shown to have an adverse effect. Corrosion and/or biofilm was collected from uncontaminated but conditioned pipe coupons using the same procedure used for sampling coupons via brushing without Tween® 80. Then an aliquot from the same spore stock suspension was spiked into each sample to determine spore recovery in the presence of the corrosion and/or biofilm. Fig. 5 shows that there is a considerable difference between spore recovery in the presence of iron corrosion and recovery in the presence of biofilm alone when brushed from the PVC and cement-lined pipe coupons. Furthermore, if samples containing iron corrosion were diluted by half phosphate buffer solution (and, later Tween® 80), spores were able to grow more readily on TSA. The dilution and plating protocol for the iron coupons was adjusted accordingly for the reported experiments.

Fig. 5 Spore enumeration in the presence of biofilm and corrosion (3 plates per sample). Error bars represent standard deviations.

Discussion

This study examined two ways to sample attached bacterial spores from wetted porous and non-porous surfaces. When the mean recovery for all the brushing data is compared with the mean recovery for all the scraping data (for all three pipe materials), it is clear that one sampling technique did not stand out as being more effective than the other. Brushing vs. scraping recovery efficiency data for all three pipe materials combined showed that one sampling technique was not more effective than the other. However, brushing cement-lined pipe coupons was shown to be more effective than scraping cement-lined pipe coupons. Results for recovery efficiency also differed in variability among the three pipe materials, 0.29 for PVC vs. 0.10 and 0.12 for cement-lined and iron, respectively. This suggests that pipe material influences the variability of the recovery efficiency. Recovery efficiency standard deviations for iron coupons varied for brushing and scraping (0.22 vs. 0.04) while PVC (0.30 vs. 0.29) and cement-lined (0.09 vs. 0.10) coupons had approximately the same standard deviation for brushing and scraping, respectively. This difference between brushing and scraping observed for the iron coupons could be attributed to the layer of rust that absorbed and stabilized the spore suspension. It was noted that scraping removed the majority of the biofilm and rust layer with one stroke. In contrast, brushing required 3–4 strokes to remove the same amount of corrosion. Although there were no statistically significant differences between the mean recovery efficiencies for brushing and scraping on iron, variability in the data indicate that scraping may be the more desirable technique due to the higher level of precision.

One of the limitations in this study is that the corroded iron coupons may not be representative of tuberculated iron pipes found in real world distribution systems. Based on research involving characterization of iron pipe surface deposits found in pilot scale and real distribution systems,¹⁴ the layer of tuberculation would likely be harder to remove than the iron corrosion formed on the iron coupons used in this study. Thus, for exhumed tuberculated iron pipe, a better surface sampling strategy might be to brush the outer layer of pipe scale to collect the soft corrosion and biofilm present on the surface. If spores are able to penetrate past the outer layer of corrosion, removing deeper layers of corrosion using a heavier chisel, scalpel, or spatula may be required to remove the entire deposit.¹⁴ Due to the large amounts of iron corrosion that would be present in these samples spore growth would most likely be inhibited. Therefore, understanding the impact of corrosion on spore recovery from exhumed pipe should be verified through the use of positive controls, matrix spikes, and standard additions.

Recovery results for the cement-lined pipe coupons showed similar brushing and scraping variability. The significant differences between mean recovery efficiency of brushing and scraping were expected (42% vs. 24%, respectively) due to the porous cement-lined surface. The toothbrush was likely more effective because it removed the spores that may have been stuck in the interstitial spaces and could not be accessed by the scraper. Therefore, for hard porous surfaces, brushing appears to be a more effective sampling technique that will produce the highest recoveries.

Recovery efficiency variability was highest for PVC pipe coupons for both brushing and scraping. Additionally, conditioned PVC had a lower level of spore retention than cement-lined or iron. This was expected for two reasons. First, PVC had lower HPC (less biofilm) than the cement-lined coupons. Second, the PVC surface was notably smoother, providing less opportunity for spore retention. Iron pipe coupons had a higher HPC (3 × 10⁸ CFU/cm²). The large amount of corrosion and biofilm on the iron coupons may have provided additional surface area for immobilizing the spores within the matrix.

The 24 h conditioned iron coupons did not have the opportunity to develop the same amount of corrosion and surface roughness as did the coupons conditioned for three months, and, not surprisingly, they did not retain spores as well as the iron coupons with more rust. PVC conditioning had the opposite effect. Spores were retained more readily on coupons with minimal conditioning than they were on the conditioned coupons with 3 months of biofilm growth (1 × 10⁵ CFU/cm² for PVC). This may have been due to the physiochemical properties of the material such as charge or hydrophobicity of the PVC. Spores are generally hydrophobic and negatively charged which may account for a stronger attraction to the unconditioned PVC.^15,16 Also, while not directly observed, biofilm sloughing may have occurred more readily from the smooth, hydrophobic surface of the PVC than from the cement-lined or iron coupons, taking the spores with it during the rinse.

The unclosed mass balances (Fig. 4) could be attributed to a number of different causes in addition to sampling inefficiency. First, spores collected during coupon rinsing (C_r) were exposed to tap water without Tween® 80. As seen by increased Petri dish swab counts for the rinse samples, a proportion of the spores that were not recovered probably had been retained on the sample containers and Petri dishes. This illustrates the importance of using a surfactant such as Tween® 80 in every step to increase recovery. The second loss may be attributed to incomplete recovery of spores from the sides and backs of the coupons. Recovery efficiency of spores from the sides and backs was not quantifiable. Thus, the proportion of the spores that were actually removed and enumerated may not have been representative of the true number of spores present on the sides and backs.

The reported data illustrate the potential for spore retention on different wetted pipe materials and associated sampling recovery from those materials. Since spore losses can be encountered throughout the sampling and analysis process, it is important to consider the recovery efficiency of the method used. Furthermore, in real-world situations, very low concentrations of surface contamination may be present making high recovery efficiency and precision of recovery all the more important. In selecting sampling techniques, this study shows that the pipe material characteristics and the presence of a biofilm will influence recovery efficiency and recovery precision. Accordingly, the sampling technique should be selected based on its ability to remove spores from the particular surface with an acceptable level of precision. Brushing vs. scraping cement-lined coupons illustrates the differences in recovery efficiency that may be obtained for the same surface. Bacterial spore retention and sample collection from wetted surfaces is complex. Future work in this area should evaluate additional surface sampling methods for wetted pipe surfaces and methods for quantifying pathogens in the presence of pipe corrosion.

Disclaimer

The U.S. Environmental Protection Agency (EPA) through its Office of Research and Development funded and conducted the research described. It has been reviewed by the Agency but does not necessarily reflect the Agency's views. No official endorsement should be inferred. EPA does not endorse the purchase or sale of any commercial products or services.

Acknowledgements

The authors would like to thank the following people and organizations for their support, without which this work would not have been possible: Dr. Hiba Ernst, U.S. EPA/NHSRC, Dr. Alan Lindquist, U.S. EPA/NHSRC, Dr. Nick Ashbolt, U.S.EPA/NERL, Dr. Jeff Szabo, U.S.EPA/NHSRC, John Hall, U.S.EPA/NHSRC, Noreen Adcock, U.S.EPA/NRMRL

References

1. J. G. Szabo, E. W. Rice and P. L. Bishop, Appl. Environ. Microbiol., 2007, 73, 2451–2457.
2. J. B. Morrow, J. L. Almeida, L. A. Fitzgerald and K. D. Cole, Water Res., 2008, 42, 5011–5021.
3. L. C. Simões, M. Simões, R. Oliveira and M. J. Vieira, J. Basic Microbiol., 2007, 47, 174–183.
4. N. B. Hallam, J. R. West, C. F. Forster and J. Simms, Water Res., 2001, 35, 4063–4071.
5. American Water Works Association, Water:\Stats, 2003 edition on CD-ROM, ISBN 1-58321-306-6-Catalog No. 53007.
6. NSF/American National Standards Institute (ANSI), 61-2000a Drinking Water System Components - Health Effects, NSF International, Ann Arbor, Michigan.
7. Standard Methods for the Examination of Water and Wastewater, ed. A. D. Eaton, L. S. Clesceri, E. W. Rice and A. E. Greenberg, American Public Health Association, Water Environment Federation, and American Water Works Association, 21st edition, 2005.
8. L. C. Simões, N. Azevedo, A. Pacheco, C. W. Keevil and M. J. Vieira, Biofouling, 2006, 22, 91–99.
9. P. J. Ollos, P. M. Huck and R. M. Slawson, J. Water Works Assoc., 2003, 95, 87–97.
10. M. Sivaganesan, N. J. Adcock and E. W. Rice, J. Water Supply: Res. Technol.—AQUA, 2006, 55, 33–43.
11. L. Coroller, I. Leguerinel and P. Mafart, Appl. Environ. Microbiol., 2001, 67, 317–322.
12. W. L. Nicholson and P. Setlow, in Molecular Biological Methods for Bacillus, ed. C. Harwood and S. Cutting, John Wiley & Sons Ltd., Chichester, England, 1990, 391–450.
13. G. S. Brown, R. G. Betty, J. E. Brockmann, D. A. Lucero, C. A. Souza, K. S. Walsh, R. M. Boucher, M. Tezak, M. C. Wilson and T. Rudolph, Appl. Environ. Microbiol., 2007, 73, 706–710.
14. S. E. Smith, J. S. Colbourne, D. Holt, B. J. Lloyd, A. Bisset, Water Quality Technology Conference proceedings AWWA 1996.
15. D. A. Lytle, C. H. Johnson and E. W. Rice, Colloids Surf., B, 2002, 24, 91–101.
16. G. Tauveron, C. Slomianny, C. Henry and C. Faille, Int. J. Food Microbiol., 2006, 110, 254–262.

---

Footnotes

† Part of a themed issue dealing with water and water related issues.

‡ Electronic supplementary information (ESI) available: Quality control details. See DOI: 10.1039/b917570a

---