EPA 600/R-11/103 | October 2011 | www.epa.gov/research
United States
Environmental Protection
Agency
Comparison of Ultrafiltration
Techniques for Recovering
Biothreat Agents in Water
i**
IIX
Office of Research and Development
National Homeland Security Research Center
Centers for Disease Control and Prevention
Atlanta, Georgia
-------
EPA600/R-11/103
Comparison of Ultrafiltration Techniques for Recovering
Biothreat Agents in Water
October 2011
Centers for Disease Control and Prevention
Atlanta, GA 30341
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
Table of Contents
List of Figures iii
List of Tables iv
Disclaimer v
Foreword vii
List of Acronyms viii
Acknowledgements x
Executive Summary xi
1.0 Introduction 1
1.1 Background 1
1.2 Study Objectives 3
2.0 Methods and Materials 5
2.1 Water Sample Preparation 5
2.2 Microorganisms and Assays 6
2.2.1 Microbes and Seed Levels for Experiment Suites 6
2.2.2 Suite 1 and 2 Microbe Sources and Seeding Procedures 7
2.2.3 Suite 3, 4 and 5 Microbe Sources and Seeding Procedures 9
2.2.4 Post-Ultrafiltration Processing and Assays 9
2.3 CDC/LRN Ultrafiltration Set-Up 11
2.3. lUltrafilter Blocking 12
2.3.2 Ultrafilter Blocking 13
2.4 EPA Ultrafiltration Set-Up 13
2.4.1 Ultrafilter Blocking Solution 15
2.4.2 Sample Processing 15
2.5 Data Analysis 17
2.6 Blanks and Controls 18
3.0 Results 19
3.1 Water Quality 19
3.2 Operations and Safety 19
3.3 Microbial Recoveries 20
3.3.1 Bacterial Recovery Efficiencies 20
-------
3.3.2 Viral Recovery Efficiencies 24
3.3.3 C. parvum and G. intestinalis Recovery Efficiencies 24
3.4 Project Data Quality Objectives and Overall Microbial Recovery Efficiencies 25
4.0 Discussion 27
5.0 Conclusions 31
6.0 Presentations and Other Activities 33
7.0 References 34
Appendix A: Quality Assurance/Quality Control 36
11
-------
List of Figures
Figure 1. Schematic of CDC UF set-up 12
Figure 2. EPA Water Sample Concentrator set-up at CDC laboratory facility 14
Figure 3. Schematic of UF set-up for EPA method 16
Figure 4. View of the Water Sample Concentrator monitoring screen as see during an ultrafiltration
run 17
in
-------
List of Tables
Table 1. Differential characteristics between the EPA and CDC/LRN UF methods 3
Table 2. Framework for study experiments 5
TableS. Water Quality data for 100-L tap water samples 19
Table 4. Operational data for EPA and CDC/LRN UF methods for 100-L water samples 20
Table 5. Average microbial recovery efficiencies for the CDC and EPA ultrafiltration procedures 22
IV
-------
Disclaimer
This report was prepared by the Centers for Disease Control and Prevention, Atlanta, Georgia, in
support of the U.S. Environmental Protection Agency's National Homeland Security Research
Center in Cincinnati, Ohio. U.S. Environmental Protection Agency (EPA) partially funded and
collaborated in this research under Interagency Agreement DW-75-92259701 with Centers for
Disease Control (CDC), U.S. Department of Health and Human Services. It has been reviewed by
EPA but does not necessarily reflect EPA's views. EPA does not endorse commercial products or
services. The findings and conclusions in this presentation have not been formally disseminated by
CDC and should not be construed to represent any agency determination or policy.
Use of trade names and commercial sources is for identification only and does not imply
endorsement by the Centers for Disease Control and Prevention or U.S. Department of Health and
Human Services.
Questions concerning this document or its application should be addressed to:
Erin Silvestri, MPH (EPA Project Officer)
U.S. Environmental Protection Agency
National Homeland Security Research Center
26 W. Martin Luther King Drive, MS NG16
Cincinnati, OH 45268
513-569-7619
Silvestri.Erin@epa.gov
H.D. Alan Lindquist, PhD (EPA Co-Lead)
U.S. Environmental Protection Agency
National Homeland Security Research Center
26 W. Martin Luther King Drive, MS NG16
Cincinnati, OH 45268
513-569-7192
Lindquist.Alan@epa.gov
Vicente Gallardo, MS (EPA Co-Lead)
U.S. Environmental Protection Agency
National Homeland Security Research Center
26 W. Martin Luther King Drive, MS NG16
Cincinnati, OH 45268
513-569-7176
Gallardo.Vincente@epa.gov
Vincent R. Hill, PhD, PE (CDC Technical Lead and Project Manager)
Centers for Disease Control and Prevention
Division of Foodborne, Waterborne, and Environmental Diseases
National Center for Emerging and Zoonotic Infectious Diseases
-------
1600 Clifton Road, ME, Mailstop D-66
Atlanta, GA 30329
404-718-4151
veh2@cdc.gov
VI
-------
Foreword
The National Homeland Security Research Center (NHSRC), part of the U.S. Environmental
Protection Agency's (EPA's) Office of Research and Development, is focused on developing and
delivering scientifically sound, reliable, and responsive products. These products are designed to
address homeland security information gaps and research needs that support the Agency's mission
of protecting public health and the environment. A portion of NHSRC's research is directed at
decontamination of indoor surfaces, outdoor areas, and water infrastructure. This research is
conducted as part of EPA's response to chemical, biological, and radiological contamination
incidents. NHSRC has been charged with delivering tools and methodologies (e.g. sampling and
analytical methods, sample collection protocols) that enable the rapid characterization of indoor and
outdoor areas, and water systems following terrorist attacks, and more broadly, natural and
manmade disasters.
NHSRC recently developed a field-portable ultrafiltration (UF) method and automated UF system.
NHSRC funded, and collaborated with, the Centers for Disease Control and Prevention (CDC) to
compare the performance of the EPA developed method and device with the established CDC
Laboratory Response Network UF method for five suites of biothreat agents and/or their surrogates.
This project determined if either method was associated with significantly higher recovery
efficiencies for biothreat agents and microbial surrogates that had been seeded into 100-L samples
of tap water. Having an understanding of the relative microbial recovery performance for the two
methods may allow for potential interchangeability of the methods for use during a bioterrorism
event.
This report represents a summary of methods and materials and results of the CDC and EPA UF
method comparison.
Jonathan Herrmann,
Director, National Homeland Security Research Center
vn
-------
List of Acronyms
ANO VA Analy si s of Vari ance
ATCC™ American Type Culture Collection
BGMK Buffalo Green Monkey Kidney
BSL Biosafety Level
CDC Centers for Disease Control and Prevention
CPU Colony-forming unit
CHAB-A Cysteine Heart Agar with Chocolatized 9% Sheep Blood and Antibiotics
CIN Cefsulodin-Irgasan Novobiocin
CT Crossing threshold
CV Coefficient of variation
DMEM Dulbecco's Modified Eagle Medium
DNA Deoxyribonucleic acid
DPD Division of Parasitic Diseases
DQO Data Quality Objectives
EMEM Eagle's Minimum Essential Medium
EPA Environmental Protection Agency
FA Immunofluorescence assay
FBS Fetal bovine serum
FIPC Heterotrophic plate count
IMS Immunomagnetic Separation
kDa Kilodaltons
LRN Laboratory Response Network
LVS Live vaccine strain
MWCO Molecular weight cut-off
NCTC National Collection of Type Cultures
NHSRC National Homeland Security Research Center
NTU Nephelometric turbidity units
ORD Office of Research and Development
PBS Phosphate buffered saline
PFU Plaque-forming unit
PLET Polymyxin, Lysozyme, EDTA, Thallous Acetate
PPE Personal Protective Equipment
psig Pound-force per square inch gauge
R2A Reasoner's2A
RPM Revolutions per minute
SAM Standardized Analytical Methods for Environmental Restoration Following
Homeland Security Events, Revision 5.0
NaPP Sodium polyphosphate
qPCR Quantitative (real-time) Polymerase Chain Reaction
TOC Total organic carbon
TSA Trypticase soy agar
TSB Trypticase soy broth
UF Ultrafiltration
EPA Environmental Protection Agency
viii
-------
WSC Water sample concentrator
IX
-------
Acknowledgements
The following researchers assisted with experiments and were critical to the success of this project:
Centers for Disease Control and Prevention, National Center for Emerging and Zoonotic
Infectious Diseases
Vince Hill
Suresh Pai
Tina Lusk
Bonnie Mull
Amy Kahler
The following individuals and organizations served as members of the Project Team and are
acknowledged:
U.S. Environmental Protection Agency (EPA), Office of Research and Development (ORD),
National Homeland Security Research Center (NHSRC)
Vicente Gallardo
H. D. Alan Lindquist
Sarah Perkins
Sanjiv Shah
Erin Silvestri (EPA Project Officer)
-------
Executive Summary
This is the final report for the U.S. Environmental Protection Agency (EPA) and Centers for
Disease Control and Prevention (CDC) Biological Sample Preparation Collaboration Project to
compare EPA and CDC ultrafiltration techniques for recovering biothreat agents in water. Hollow-
fiber ultrafiltration (UF) is increasingly accepted as an effective sampling technique for
simultaneous recovery of diverse microbes from water, including drinking water samples collected
during water-related emergency response events. In this study, a laboratory-based UF method
established by CDC for the Laboratory Response Network (LRN), a network of labs that can
respond to biological and chemical terrorism, and other public health emergencies, was compared to
a field-portable UF method developed by EPA for use with an automated UF system [the Water
Sample Concentrator (WSC)]. Five suites of experiments were performed. For Suite 1 to 3
experiments, sodium polyphosphate (NaPP) was added as a sample amendment to water samples
that were used for both the CDC and the EPA methods. For Suite 4 and 5 experiments, NaPP was
added only to water samples processed with the CDC UF method. Suite 4 and 5 experiments were
conducted to see if there was a measurable effect in adding NaPP to the water samples on the EPA
method as had been done in Suite 1-3 experiments. Microbial recovery efficiencies were
determined for the following microbes seeded into 100-L water samples which were then processed
by each method:
Suite 1: Bacillus anthracis (Sterne) spores, Yersiniapestis (Al 122), Francisella tularensis
LVS (i.e., live vaccine strain), Enterococcusfaecalis, and Clostridiumperfringens spores
Suite 2: MS2 bactedophage, phi XI74 bacteriophage, echovirus type 1, high seed
Cryptosporidium parvum oocysts, high seed Giardia intestinalis (aka G. lamblid) cysts,
ColorSeed™ [containing 100 C. parvum and 100 G. lamblia fluorescent (oo)cysts]
Suite 3: B. anthracis (Sterne) spores, Bacillus atrophaeus subsp. globigii., F. tularensis LVS
and Brevundimonas diminuta
Suite 4: B. anthracis (Sterne) spores, B. atrophaeus subsp. globigii, Y. pestis (Al 122), F.
tularensis LVS and B. diminuta
Suite 5: E.faecalis, MS2 bacteriophage, phi XI74 bacteriophage, echovirus type 1, C.
parvum oocysts, and G. intestinalis cysts
After performing the respective UF methods, samples were further concentrated and assayed using
microbe-specific techniques, including membrane filtration and agar culture (for bacteria),
microconcentrators and cell culture plaque assays (for viruses), and centrifugation and fluorescence
microscopy (for parasites). In general, both the CDC and the EPA UF methods achieved greater
than 50% recovery efficiencies during the Suite 1, 2 and 3 experiments:
Suite 1: B. anthracis spores (85 and 100%, respectively), Y. pestis (70 and 70%), E.faecalis
(97 and 100%) and C. perfringens (100 and 110%)
Suite 2: MS2 (110 and 120%, respectively), phi X174 (100 and 95%), echovirus 1 (68 and
47%), C. parvum (82 and 73%) and G intestinalis (99 and 85%)
Suite 3: B. anthracis spores (65 and 92%, respectively), B. atrophaeus subsp. globigii (57
and 99%) and B. diminuta (83 and 84%)
xi
-------
F. tularemis was the most challenging microbe to recover during the Suite 1, 2 and 3 experiments,
with average recovery efficiencies of 13-17% for the CDC/LRN method and 25-29% for the EPA
method. When UF concentrates were exposed to 1% ammonium chloride for 2 h before culture, F.
tularensis culturability was significantly improved (and measured recovery efficiencies increased
by 35-120%). While both methods were found to be similarly effective overall, statistical analysis
indicated that the bacterial recoveries obtained using the EPA automated UF method were
significantly higher (a < 0.05) when Suite 1, 2, and 3 data were combined and analyzed.
ColorSeed™ (BRF Precise Microbiology, Pittsburgh, PA) recoveries were similar for the EPA and
CDC/LRN methods, with C. parvum oocyst recoveries of 30 and 38%, respectively, and G. lamblia
recoveries of 44 and 42%, respectively.
In the Suite 4 and Suite 5 experiments, which included NaPP only in water samples processed using
the CDC UF method, recovery efficiencies were also generally greater than 50% for both the CDC
and EPA UF methods:
Suite 4: B. anthracis spores (74 and 96%, respectively), B. atrophaeus subsp. globigii (47
and 89%), Y. pestis (100 and 76%), B. diminuta (82 and 78%)
Suite 5: E.faecalis (100 and 63%, respectively), MS2 (99 and 69%), phi XI74 (110 and
86%), echovirus 1 (79 and 37%), C. parvum (72 and 110%), and G. intestinalis (78 and
110%)
When Suite 4 and 5 microbial recovery data were combined and analyzed, no statistically
significant difference between the EPA and CDC/LRN UF methods was observed. However,
statistically significant different (a < 0.05) recovery efficiencies were measured for a number of
individual microbial analytes as follows. Higher recovery efficiencies were measured for the EPA
UF method for B. anthracis spores, B. atrophaeus subsp. globigii spores, C. parvum oocysts, and G.
intestinalis cysts while higher recovery efficiencies were measured for the CDC/LRN UF method
for E.faecalis, MS2 b acted ophage, phi XI74 bacteriophage, and echovirus 1 (See Table 5).
Operationally, filtrate rates for the WSC were slightly higher than for the CDC method. The higher
filtrate rates and automation of the procedure resulted in the EPA procedure requiring
approximately 20 fewer minutes to complete than the CDC/LRN UF procedure. Overall, despite
physical, operational, and procedural differences between the two methods, the data from this study
demonstrate that the EPA and CDC/LRN UF methods are highly efficient for recovering diverse
microbes, including biothreat agent surrogates, and provide similar recovery performance.
xn
-------
1.0 Introduction
1.1 Background
Intentional contamination of drinking water supplies is a concern for water utilities, federal, state,
and local agencies tasked with protecting human health and the environment. Because relatively
low levels of biothreat agents can cause human health effects (1), sensitive detection of these agents
in drinking water is needed. However, most rapid response analytical techniques [e.g.,
immunological "dipstick" methods, real-time polymerase chain reaction (qPCR)] assay small
sample volumes or require high concentrations of analytes. Therefore, to enable sensitive detection
of biothreat agents large volumes of water (on the order of 10-100 L) should be collected and
concentrated. Alternative large-volume water sampling techniques have been published for viruses
(e.g., various adsorption-elution techniques), bacteria (membrane filtration), and parasites
(microfiltration cartridges), but the effectiveness of these methods are generally optimized for
particular microbes types (i.e., viruses, bacteria or parasites). However, in the event of a biological
attack on a drinking water system, the biothreat agent may not be known with certainty and
deployment of multiple sampling techniques would be a logistical challenge and resource intensive.
For this reason, the U.S. Environmental Protection Agency (EPA) and U.S. Centers for Disease
Control and Prevention (CDC) have worked together to investigate methods to enable rapid and
sensitive analysis of water samples for diverse, unidentified biothreat agents. This is the final report
for the EPA and CDC Biological Sample Preparation Collaboration Project to compare EPA and
CDC ultrafiltration techniques for recovering biothreat agents in water.
Homeland Security Presidential Directive 9 requires the development of a nationwide,
interconnected network of federal and state laboratories that integrate resources and use
standardized analytical procedures when supporting responses to homeland security incidents. The
Laboratory Response Network was launched by CDC in 1999. Another key component of this
directive is the Standardized Analytical Methods for Environmental Restoration Following
Homeland Security Events (SAM) (2), which contains suggested assays for use by the LRN, the
laboratories tasked with performing confirmatory analysis of environmental samples following a
homeland security event (SAM is published by EPA's National Homeland Security Research
Center (NHRSC) along with other EPA divisions and sister agencies). Though the manual details a
variety of sample assays, it does not describe a method for sampling large volumes of water for an
unidentified biothreat agent [e.g., viruses, bacteria, spores, parasite (oo)cysts, toxins]. Further,
development of a field-deployable sampling method would make it unnecessary to manually collect
large-volume water samples (e.g., in 20-L carboys) that would need to be shipped to an analytical
laboratory at great expense and effort.
Ultrafiltration (UF) has become an established technique for co-concentrating diverse microbes
(including viruses, bacteria and parasites) in large-volume water samples. Ultrafiltration has been of
particular interest for bioterrorism preparedness because the technique can be used to capture
unidentified biothreat agents. Since 2003, numerous studies have reported the effectiveness of UF
for co-concentrating diverse microbes in water, including potential biothreat agents (3-9). In
general, the ultrafiltration techniques reported within the last 10 years have used cross-flow
-------
recirculation of water samples through hollow-fiber ultrafilters to concentrate 10- to 100-L water
samples down to volumes on the order of 200-500 mL. While the recirculating flow UF technique
can be performed in the field (10), it requires training and experience to perform effectively and
consistently, and can be a challenge to set up under field conditions.
Since about 2003, EPA and CDC have been investigating UF methods for water-related
bioterrorism preparedness. In 2006, CDC researchers developed an ultrafiltration method, and
associated secondary sample processing protocols for the Laboratory Response Network (LRN).
The LRN ultrafiltration and water processing procedure is a laboratory-based protocol. During the
same time frame, EPA was also investigating recirculating flow UF methods, but with a focus on
developing a field-portable and automated UF device. EPA and Idaho National Laboratory
succeeded in developing an automated UF instrument, referred to as the water sample concentrator
(WSC) in this report.
The CDC ultrafiltration method used in these experiments was established by the Laboratory
Response Network in a document entitled "Filter Concentration for the Detection of Bioterrorism
Threat Agents in Potable Water Samples" (11). The method involves the use of five, 20-L carboys,
an ultrafilter, and a pump to filter a 100-L drinking water sample. The method is completely
manual, with all steps performed by the technician. After 100 L is filtered, an elution step is
performed to recover microbes that are adsorbed or otherwise retained in the filtration system; the
final UF concentrate sample is then further processed and/or analyzed for the target microbe(s)
using standard microbiological methods.
The WSC (approximate dimensions: 31 x 20 x 16 inches [795 x 518 x 393 mm])was developed as a
field portable instrument to improve ease of use, safety, and consistency of the ultrafiltration
concentration process. The device was controlled by software that was installed on a personal
computer. As with the LRN method, the WSC used a hollow fiber filtration cartridge which was
pre-treated prior to use. In addition the WSC, similar to the LRN method, used an elution
procedure after filtration and prior to the final UF concentrate sample recovery and analysis for the
target microbe(s) using standard microbiological methods.
Beyond the laboratory-based versus field-portable nature of the CDC/LRN and EPA UF methods,
the two UF procedures differ in a few other potentially important ways (Table 1).
The LRN method was developed using a Masterflex® L/S® peristaltic pump (< 2.9 L/min
pumping rate) (Cole-Parmer Instrument Company, Vernon Hills, IL) versus the larger
Masterflex® I/P® -sized pump (< 8 L/min pumping rate) used in the WSC. Thus, cross-flow
rates and filtrate rates are higher for the WSC.
The LRN method uses pre-treatment of water samples with NaPP, a dispersing agent. No
sample pre-treatment is performed with the WSC procedure.
Ultrafilters are blocked (i.e., pre-treated) with fetal bovine calf serum (FBS) in the LRN
method, whereas in the EPA UF system the blocking is achieved by exposure of the filter
to a solution containing Tween® 80 (ICS Americas, Foster, KY), Antifoam A, and NaPP.
Both methods use elution to desorb microbes that may have become attached during the
ultrafiltration procedure, but the LRN method uses an elution solution containing Tween®
-------
80, Antifoam A (or Antifoam Y-30 Emulsion) and NaPP, whereas the WSC method uses a
solution containing only Tween 80.
Table 1. Differential Characteristics Between the EPA and CDC/LRN Ultrafiltration Methods
Differentiating
Characteristic
Sample Amendment
Filter Blocking
Elution
Setup
Mode of control
Pump size
Filter type
EPA Method
None
With solution containing 0.055%
Tween 80, 0.001% Antifoam A,
0.1% NaPP
With solution containing 0.001%
Tween® 80
Field-portable
Computer-controlled
Masterflex I/P®
(industrial/process scale)
REXEED™25-S filter (for this
study)
CDC/LRN Method
With 0.01% NaPP
With 5% FBS
With solution containing 0.01%
Tween® 80, 0.01% NaPP, 0.001%
Y-30 Antifoam Emulsion
Laboratory method
Manual
Masterflex L/S®
(lab scale)
Fresenius F200NR filter (for this
study)
Because the CDC/LRN and EPA UF methods were developed to achieve the same basic goal (rapid
recovery of diverse biothreat agents in large-volume drinking water samples) it is important to
understand the relative microbial recovery performance for the two methods. Such method
comparison data will be useful to both the EPA and CDC for understanding the relative strengths of
each method and the potential interchangeability of the methods if either—or both—are used during
a bioterrorism response. In this study we compared the use of the laboratory-based LRN UF
method to the EPA's field-portable WSC UF device to concentrate 100-L tap water samples for five
suites of biothreat agents and/or their surrogates. Pathogens of concern in environmental matrices
were selected from SAM, 4.0 and are those that result in adverse human health effects upon
infection or exposure.
1.2 Study Objectives
The primary objective of this project was to compare the CDC/LRN UF protocol and EPA UF
device protocol to determine if either is associated with significantly higher recovery efficiencies
for microbes seeded into 100-L samples of tap water. Pathogens and biothreat agent surrogates
used in this study were Bacillus anthracis (Sterne) spores, Yersiniapestis (Al 122), Francisella
tularensis LVS, echovirus type 1, Cryptosporidium parvum oocysts, and Giardia intestinalis cysts.
In addition, the following microbes were studied because they have been suggested as potential
biothreat agent surrogates or UF method proficiency parameters: Enter ococcusfaecalis, Bacillus
-------
atrophaeus subsp. globigii, Brevundimonas diminuta, Clostridium perfringens spores, MS2
bacteriophage, and phi XI74 bacteriophage. Secondary goals of this study included:
Comparing the recovery efficiencies of the two UF methods when water samples contained
C. parvum and G. intestinalis at high [~105 (oo)cysts] and low [100 (oo)cysts] seed levels
Evaluating use of 1% ammonium chloride for improving the culturability of F. tularensis in
UF concentrates (measured as a change in recovery efficiency)
Comparing average processing times associated with the two UF methods
For each experiment, physical and chemical water quality parameters were measured to enable
evaluation of potential water quality influences on the performance of the UF procedures and
analytical assays. The data quality objectives for this project included coefficient of variation (CV)
goals for percent recovery efficiency data sets for each high seed microbe (CV < 25% for each UF
method) and for each low seed microbe (CV <50%, reflecting higher data variability associated
with the multiple procedures [UF and secondary processing] employed for low-seed microbe).
-------
2.0 Methods and Materials
2.1 Water Sample Preparation
Experiments were performed using five independent microbial "Suites". Experiments using
microbial Suites 1, 2, and 3 were performed from May 2009 to April 2010 with tap water
samples obtained from the CDC Waterborne Disease Prevention Branch Environmental
Microbiology Laboratory on the CDC's "Chamblee Campus" (Table 2). In July, 2010 the
WDPB Environmental Microbiology Laboratory moved to a new laboratory facility located on
the CDC's "Roybal Campus." Experiments using microbial Suites 4 and 5 were performed
using tap water from the laboratory on the Roybal Campus. Tap water samples were collected in
sterile, 35-gallon high-density polyethylene tanks that were calibrated to 100-L using 10-L
gradations. Prior to collecting each water sample, the tap was fully opened for 5 minutes to draw
fresh water through the building distribution system. Two 100-L tap water samples were
collected at the same time from two taps in the same laboratory room. To ensure that the same
quality water was used to perform both the CDC/LRN and EPA methods, a third 35-gal tank was
used to mix 50-L from each of the other tanks. Free chlorine was measured in each tank to assess
initial chlorine residual using Hach® DPD (Division of Parasitic Diseases) Methods 8021 (Hach
Companyl, Loveland, CO) and 8167, respectively (equivalent to Standard Method 4500-C1 G),
and a Hach® DR/2400 spectrophotometer (12). A volume of 50 mL of 10% w/v stock of sodium
thiosulfate solution was then added to each tank to quench the chlorine. Free chlorine was read
again for each tank to confirm quenching. Additional sodium thiosulfate was added to each water
sample, if necessary, until no free chlorine could be detected. A 500-mL sample of water was
collected by obtaining 250 mL of water from each tank. For each experiment, this water sample
was seeded with the same numbers of study microbes added to each 100-L sample and the
sample was assayed in conjunction with the CDC/LRN and EPA UF concentrate samples at the
end of the experiment. The data from this 500-mL "control sample" was used to quantify the
microbe seeding levels for each experiment.
Table 2. Framework for Study Experiments
Site Microbial Suites CDC/LRN Method EPA Method
Chamblee 1,2, and 3 With NaPP sample With NaPP sample
amendment (per amendment (not
established protocol) established protocol)
Roybal 4 and 5 With NaPP sample No NaPP sample
amendment (per amendment (per
established protocol) established protocol)
NaPP, sodium phosphate
When the chlorine had been quenched in each tank, 75 mL of water was collected from each tank
and combined for water quality analysis. All water samples were characterized using the
following water quality parameters: specific conductance, temperature, pH, turbidity, total
hardness, total organic carbon (TOC), and heterotrophic plate count (FtPC) of bacteria. A 50-mL
-------
portion of this sample was tested for specific conductance and temperature using an Oakton®
CON 100 Conductivity/°C meter (Oakton Instruments, Vernon Hills, IL). This conductivity
meter was calibrated weekly using vendor instructions for conductivity and temperature
calibration. The pH of water samples was measured using a Fisher Scientific™ Accumet®
Research AR25 Meter (Fisher Cat. No. 13-636-AR25A, Fisher Scientific, Pittsburgh, PA) and
Accumet Standard Size Combination Electrode (Fisher Cat. No. 13-620-285). Turbidity was
measured using a Hach Model 2100N Laboratory Turbidimeter (Cat. No. 4700000, Hach
Company). All measurements were collected using the "Signal Averaging" function on the
turbidimeter. Total hardness was measured using Hach Method 8213 with a Hach Hardness
(Ca/Mg) Reagent Set (Cat. No. 24480-00, Hach Company) and Hach Model 16900 digital
titrator (12). TOC was measured using Hach Method 10129 with a Hach Low Range TOC
Reagent Set (Cat. No. 2760345) and the Hach DR/2400 Portable Spectrophotometer (12). HPC
bacteria were measured in duplicate assays using a Standard Method (13). For the HPC tests, one
30-CFU (colony-forming unit) K coli [NCTC 9001 (Pall Supor Acrodisc 11775)] Bioball®
purchased from BTF Pty. Ltd. (Australia) was used as a positive control and 10 mL sterile wash
phosphate buffer saline (PBS) was used as the negative control for the HPC count.
For Suite 1, 2 and 3 experiments, sodium polyphosphate (NaPP), a chemical dispersant, was
added at a 0.01% w/v ratio to the 100-L water samples that were processed by the CDC and EPA
UF methods. For Suite 4 and 5 experiments, NAPP was added only to the 100-L water samples
processed with the CDC method (and not to the water sample processed using the EPA method).
Suite 4 and 5 experiments were performed in a laboratory facility at the CDC's Roybal Campus,
while Suite 1, 2 and 3 experiments were performed in a laboratory at CDC's Chamblee Campus,
but both laboratory facilities were served by the same water treatment plant (Dekalb County
Water and Sewer's Scott Candler Water Treatment Plant). The Scott Candler Water Treatment
Plant produces drinking water that is conventionally treated before chlorine addition (to achieve
disinfectant residual) and caustic soda (sodium hydroxide) addition for corrosion control in the
distribution system. Tap water samples processed using the CDC/LRN UF method were pumped
into 5, 20-L Cubitainers® to perform the method.
2.2 Microorganisms and Assays
2.2.1 Microbes and Seed Levels for Experiment Suites. In order to limit the number of
microbes assayed for each experiment, five suites of microbes were used in separate
experiments. After appropriate dilutions were made in diluent PBS (0.01M) containing 0.01%
Tween 80, each bacterial stock used to create the seed spike for an experiment was filtered
through a 5-|im Pall Supor Acrodisc syringe filter (Model No. 4650; Pall Corporation, Port
Washington, NY).
Suite 1 consisted of 10 replicate experiments with water seeded with the following microbes:
B. anthracis (Sterne) spores - 3,600 ± 1,700 CFU
Y. pestis Al 122 - 70,000 ± 16,000 CFU
F. tularensisLVS - 90,000 ± 100,000 CFU
E.faecalis (ATCC™ 29212 from BTF Multishot-550 BioBall®) - 1,100 CFU
-------
C. perfringens spores (NCTC 8798 from HighDose-lOK BioBall®, BTF Pty. Ltd.,
Australia)-110 ±56 CPU
Suite 2 included 11 replicate experiments with water seeded with the following microbes:
MS2 bacteriophage (ATCC 15597-B1) - 45,000 ± 29,000 PFU (plaque-forming unit)
phi X174 bacteriophage (ATCC 13706-B1) - 11,000 ± 2,500 PFU
echovirus 1 (Farouk strain, ATCC VR-1038) - 3,600 ± 1,600 PFU
C. parvum oocysts (Waterborne, Inc., New Orleans, LA) - 180,000 ± 100, 000 oocysts
G. intestinalis cysts (aka Giardia lamblia; Waterborne, Inc.) - 200,000 ± 110,000 cysts
ColorSeed™ [containing 100 (oo)cysts each of fluorescent C. parvum and G.
intestinalis] (BTF Pty) - 1 vial containing 100 (oo)cysts each of fluorescent C. parvum
and G. intestinalis.
Suite 3 consisted of 10 replicate experiments with water seeded with the following microbes:
B. anthracis (Sterne) spores - 6,600 ± 1,500 CFU
B. atrophaeus subsp. globigii spores - 9,300 ± 2,200 CFU
F. tularensis LVS - 81,000 ± 91,000 CFU
Brevundimonas diminuta - 42,000 ± 22,000 CFU
Suite 4 consisted of 9 replicate experiments with water seeded with the following microbes:
B. anthracis (Sterne) spores - 5,200 ± 690 CFU
B. atrophaeus subsp. globigii spores - 9,800 ± 3,700 CFU
Y. pestis - 5,100 ± 5,700 CFU
F. tularensis - 46,000 ± 44,000 CFU
B. diminuta - 5,100 ± 3,300 CFU
Suite 5 consisted of 8 replicate experiments with water seeded with the following microbes:
E. faecalis - 780 ± 72 CFU
MS2 bacteriophage - 110,000 ± 23,000 PFU
Phi X174 bacteriophage - 12,000 ± 2,000 PFU
Echovirus type 1 - 45,000 ± 14,000 PFU
C. parvum oocysts - 150,000 ± 24,000 oocysts
G. intestinalis cysts - 180,000 ± 46,000 cysts
2.2.2 Suite 1 and 2 Microbe Sources and Seeding Procedures
For each experiment, a seed stock was made that consisted of the study microbes for the
experiment. One third of the stock was added to a 500-mL control sample that was drawn from
the two 100-L water samples, one third was added to the 100-L "EPA Method" tap water sample,
and one third was added to the "CDC/LRN Method" water sample (for this method an equal-
volume aliquot of the microbial stock was added to each of the 20-L Cubitainers® [i.e., ~l/5th of
7
-------
the stock volume was added to each of the 5, 20-L Cubitainers®]) (Hedwin Corporation,
Baltimor, MD). Microbial seed stocks were prepared and added to water samples as described
below.
Bacteria. B. anthracis spores were acquired from CDC Division of Healthcare Quality
Promotion (Matt Arduino and Laura Rose) and were produced by culture and sporulation on AK
Agar #2 (Sporulating Agar) (BD Diagnostics; Franklin Lakes, NJ) and purified by centrifugation
through 58% Hypaque®-76 (Nycomed, Inc., Switzerland). B. anthracis spore stocks were stored
at 4 °C in 40% (v/v) ethanol. Y. pestis stocks were acquired from CDC Division of Healthcare
Quality Promotion (Matt Arduino and Laura Rose) and were produced on tryptic soy agar (TSA)
containing 5% sheep blood (CDC Scientific Resources Program). F. tularemis stocks were
acquired from CDC Division of Healthcare Quality Promotion (Matt Arduino and Laura Rose)
and were produced on Chocolate II agar (CDC Scientific Resources Program). Six, 550-CFUE1.
faecalis BioBalls® (ATCC 29212) were used for each experiment (two for each 100-L water
sample and two for the control sample). Although a BioBall® containing 10,000 CPU of C.
perfringens spores was used for each experiment, when cultured on mCP agar these BioBalls®
yielded an average of-330 CPU. C. perfringens BioBalls® were processed following
manufacturer's guidelines to disaggregate spores; they were vigorously shaken at 600
oscillations/min in diluent PBS for 30 min using a Pall Gelman laboratory shaker (Model No.
4821).
The total seeding amount for each bacterial analyte was determined by membrane filtration of
appropriate volumes of the seeded 500-mL control sample and selective agar culture. These seed
levels were selected to enable quantification of each microbe in control and UF concentrate
samples at per-plate counts of 20-80 CPU when sample volumes of approximately 0.1-10 mL
were assayed. Each bacterial stock used to create the seed spike for an experiment was filtered
through a 5-|im Pall Supor Acrodisc (Model No. 4650) to remove bacterial aggregates before
appropriate dilutions were made in diluent PBS (0.01M; CDC Scientific Resources Program)
containing 0.01% Tween 80.
Viruses. Microbial seed dilutions of the stocks of MS2 and phi XI74 bacteriophage were made
in diluent 0.01M phosphate-buffered saline (PBS; Dulbecco's modification, pH 7.40), 0.01%
(w/v) Tween 80 (Fisher), and 0.001% (w/v) Antifoam Y-30 emulsion (Sigma) to disperse viral
particles. The stocks were vortexed vigorously for 30 seconds before making the dilutions and
vortexed 10-15 seconds between dilutions. The bacteriophages and echovirus 1 were filtered
through a 0.1-jim Acrodisc filter before seeding. A clone of echovirus 1 (Farouk strain) was
prepared from a strain obtained from the American Type Culture Collection (ATCC, Manassas,
VA) and propagated in BGMK (Buffalo Green Monkey Kidney) cells (Scientific Resources
Program, CDC). Cell lines were maintained in either Eagle's Minimum Essential Medium
(EMEM) or Dulbecco's Modified Eagle Medium (DMEM) as described previously (14).
Parasites. Before use in an experiment, C. parvum and G. intestinalis stocks from Waterborne,
Inc. were diluted to achieve a diluted stock concentration of 100,000 (oo)cysts/mL. Three mL of
each stock dilution were heat-treated for 10 min at 60 °C to inactivate the (oo)cysts. The stocks
were then shaken on a Pall Gelman laboratory shaker for 30 min to disaggregate the (oo)cysts
-------
before adding 1 mL of each stock to each 100-L water sample and the control sample for an
experiment.
2.2.3 Suite 3, 4 and 5 Microbe Sources and Seeding Procedures
In Suite 3, 4, and 5 experiments, the microbes studied and seeding procedures used were
the same as used in Suites 1 and 2 for B. anthracis spores, 7 pestis, F. tularensis, E. faecalis,
MS2, phi XI74, echovirus 1, C. parvum, and G. intestinalis. In Suite 3 and 4 experiments, water
samples were seeded with B. atrophaeus subsp. globigii spores and B. diminuta. B. atrophaeus
subsp. globigii spores were obtained from EPA (Cincinnati) and were propagated using Generic
Spore Media as previously described (6). B. atrophaeus subsp. globigii spore stocks were stored
at 4 °C in 40% (v/v) ethanol. A kanamycin-resistant isolate of B. diminuta was obtained from
ATCC (#19146). C. perfringens spores and ColorSeed™ (oo)cysts were not studied in Suite 3, 4,
and 5 experiments (ColorSeed™, BTF Precise Microbiology, Inc., Pittsburgh, PA) .
2.2.4 Post-Ultrafiltration Processing and Assays
Bacteria. UF concentrate samples and the input control sample for each experiment
were assayed for each bacterial analyte using two or more sample volumes and duplicate assays
for each sample volume. When each UF procedure was completed, UF concentrates were
assayed immediately for F. tularensis by membrane filtration using 0.2-|im Supor® membranes
(Pall Life Sciences, #66234) and culture of the filters on plates of Cysteine Heart Agar with
Chocolatized 9% Sheep Blood and Antibiotics (CHAB-A) (15), which were prepared by CDC's
Division of Scientific Resources. In addition, aliquots of the experiment control sample, CDC
UF concentrate, and the EPA UF concentrate samples were also exposed to 1% ammonium
chloride (final concentration with water sample added) for 2 h before membrane filtration and
incubating on CHAB-A plates. CHAB-A plates were incubated for 4-7 days at 37°C before
inspecting for characteristic^, tularensis colonies (yellow, mucoid). Assays for B. anthracis, Y.
pestis, E. faecalis, and C. perfringens were performed after culture assays for F. tularensis were
completed. Membrane filtration was performed for each of these bacteria using 0.45-|im mixed-
cellulose ester membrane filters. B. anthracis spores and B. atrophaeus subsp. globigii spores
were cultured on plates of Polymyxin B-Lysozyme-EDTA-Thallous Acetate (PLET) agar
(prepared by CDC's Division of Scientific Resources) incubated at 37°C for 24 hours (16) and
inspected for characteristic B. anthracis colonies (pink/cream) and B. atrophaeus subsp. globigii
colonies (orange). 7 pestis was cultured on plates of (Cefsulodin-Irgasan -Novobiocin) CIN
agar (prepared by CDC's Division of Scientific Resources) incubated at 27°C in an
environmental chamber for 2-3 d (17). C. perfringens spores were cultured on plates of mCP
agar (Acumedia #7477A) incubated in an anaerobic jar at 41°C for 18-24 h (18). Bacterial
colonies on plates of mCP agar were exposed to ammonium hydroxide in fume hood and
characteristic pink colonies were counted as C. perfringens. E.faecalis was cultured on plates of
mEI agar (mE agar [Becton Dickinson #233320] with 0.075% [w/v] indoxyl P-D glucoside)
incubated at 41 °C for 24 h (19). B. diminuta was enumerated using R2A agar (Reasoner's 2A)
(Remel #R454372) containing 0.4 |ig/mL of kanamycin and 0.08 |ig/mL of tetracycline (to
minimize growth of background microbes) and incubated at 30°C for 48 h.
-------
Viruses. UF concentrate samples and the input control sample for each experiment were
assayed for each virus analyte using two or more sample volumes and duplicate assays for each
sample volume. When each UF procedure was completed, MS2 and phi XI74 were assayed in
the experiment control sample and UF concentrates using the single agar plaque assay method
using the E. coli CN-13 (ATCC 700609) and Famp (ATCC 700891) host cells, respectively,
according to EPA method 1602 (20). According to Method 1602, the appropriate bacterial host
was inoculated into separate water sample aliquots and incubated briefly. The appropriate molten
agar for each bacterial host was then added to each water sample, swirled to mix and then poured
onto 150-mm Petri dishes. After cooling on a bench top for -15 min, plates were then incubated
at 37 °C for approximately 17 h.
Because echovirus 1 was seeded into water samples at a relatively low seed level,
quantification of echovirus 1 recovery efficiencies required concentration of viruses in UF
concentrates. UF concentrates produced by both the CDC/LRN and EPA methods were further
processed for echovirus 1 analysis using Centricon Plus-70 microconcentrators. The
manufacturer's procedure was followed with the exception that two 70-mL volumes of sample
were processed for each UF method (140 mL total). Echovirus 1 was quantified in Centricon®
(Millipore Corp., Billerica, MA) concentrates by plaque assay by inoculating 10-fold dilutions
onto BGMK cell monolayers in 60 mm2 dishes (9). After 1-h adsorption at 37 °C and 5% CC>2,
the infected cells were overlaid with 5 mL maintenance medium containing 0.5% agarose.
Following a 2-day incubation, a second overlay containing 2% neutral red was added to visualize
plaques within 4 h. For echovirus 1, 0.25 mL of a 10-fold dilution was assayed per plate.
Parasites. UF concentrate samples and the input control sample for each experiment
were assayed for high seed C. parvum and G. intestinalis (oo)cysts in duplicate assays.
Recovery efficiencies for C. parvum and G. intestinalis were based on direct fluorescence
microscopy analysis of UF concentrates and the experiment control without immunomagnetic
separation (IMS) processing. Microscopy slides were prepared with 300 jiL of each sample using
SuperStick™ slides (Waterborne, Inc, New Orleans, LA). Oocysts and cysts were stained using
Easy Stain™ (BTF, Australia) according to the manufacturer's instructions and observed using a
fluorescence microscope at 400X magnification. In addition to adding C. parvum and G.
intestinalis (oo)cysts at high seeding levels, low level seeding was also performed using
ColorSeed™ (BTF) to enable comparative evaluation of the EPA and CDC/LRN water
processing methods for a water-related biothreat agent present at a low concentration. To assay
ColorSeed™ (oo)cysts, 250 mL from each UF concentrate sample was further concentrated by
centrifugation according to the procedure of Lindquist et al. (4). ColorSeed™ (oo)cysts were
recovered from the pellet using immunomagnetic separation (IMS) (Dynabeads® GC-Combo;
Life Technologies/Invitrogen, Carlsbad, CA) according to the procedures in EPA Method 1623
(21) and counted on SuperStick™ slides by immunofluorescence assay microscopy. One
immunofluorescence assay was performed for each UF concentrate sample (i.e., duplicate assays
were not performed). For the ColorSeed™ sample, an initial control was not performed because
ColorSeed™ is warranted by the manufacturer to contain 100 C. parvum oocysts and 100 G.
intestinalis cysts. To calculate recovery efficiencies, microscopy counts were compared to this
value and a percent recovery was determined.
10
-------
2.3 CDC/LRN Ultrafiltration Set-Up
The CDC/LRN method was performed in accordance with the LRN protocol, Filter
Concentration and Detection of Bioterrorism Threat Agents in Potable Water Samples (Rev
09/21/2007). The procedure was performed on a bench top in a BSL-2 (biosafety level 2)
laboratory (no microbes were used that required a BSL-3 facility). A Cole-Parmer model 7550-
30 Masterflex® L/S peristaltic pump and high performance, platinum-cured L/S 36 silicone
tubing (Masterflex; Cole-Parmer Instrument Co., Vernon Hills, IL) were used to pump water
from a 20-L Cubitainer® through the ultrafilter (Fig. 1; 1 of 5 Cubitainers® shown). The CDC UF
method was performed using Fresenius F200NR polysulfone single-use dialysis filter (Fresenius
Medical Care, Lexington, MA) because this is the filter that was used during LRN validation
testing for the method. F200NR dialyzers have an approximate molecular weight cut-off
(MWCO) of 30 kDa and surface area of 2.0 m2. The CDC/LRN UF procedure included ultrafilter
blocking (pre-treatment), sample amendment with NaPP, sample filtering, and a filter elution
step.
11
-------
Pump
Dialysis filter
Sample
Injection
20L Carboy
Sample Retentate Reservoir
Sample Filtrate
Reservoir
Figure 1. Schematic of CDC ultrafiltration set-up.
2.3.1 Ultrafilter Blocking. The ultrafilter was positioned vertically with a ring stand
and clamp. A 50-mL syringe was then connected to the filter's inlet port using a piece of tubing
approximately 6 inches in length and a DIN adapter. Using the syringe, approximately 150 mL
of a 5% calf serum (Invitrogen catalog no. 16170-078) solution was injected into the ultrafilter.
Both ends of the filter were capped and covered with Parafilm® (Pechiney, Stamford, CT) and
both side ports were tightened to prevent leaks. The filter was placed on the rotisserie at room
temperature for a period of at least 30 minutes. Immediately prior to performing an experiment,
the blocking solution was flushed from the ultrafilter by pumping 1 L of a 0.01% NaPP (Sigma,
catalogue #305553) (Sigma, St. Louis, MO) solution through the ultrafilter filtrate and retentate
ports.
12
-------
2.3.2 Sample Processing. As described in Section 2.1, the CDC/LRN UF method was
performed using 100 L of tap water that had been dechlorinated, amended with NaPP to a
concentration of 0.01% w/v, and distributed into 5, 20-L Cubitainers®. Per the LRN UF
procedure, the target filtrate rate was 60% of the peristaltic pump flow with the balance of the
pump flow exiting the filter through the retentate line and recycled back into the system. The
average filtrate flow rate achieved for the CDC/LRN method during the study was 1,700 ±180
mL/min (58% of the nominal pump flow rate of 2,900 mL/min). Sample water in each
Cubitainer® was concentrated by ultrafiltration until a volume of <500 mL remained in the
Cubitainer, at which point the pump was stopped. The outlet tubing was moved to a 1-L glass
beaker and the retentate in the Cubitainer® was pumped into the beaker. The retentate in the
beaker was then concentrated further until there was no retentate left in the beaker, at which
point the pump was stopped and the tubing was moved to the next Cubitainer®. The beaker was
then set aside and the next Cubitainer was processed. When processing the sample from the last
Cubitainer (i.e., the 5th of 5 Cubitainers®), the retentate in the beaker was reduced to as low a
volume as possible. Then the inlet tubing was removed from the sample with the flow regulator
open to let the peristaltic pump run until all the sample from the filter was pushed out. At the end
of the sample concentration procedure, retentate sample volumes were 260 ±36 mL.
After the entire 100-L sample was processed and the retentate sample collected, the filter was
then eluted using an elution solution containing 0.01% NaPP, 0.01% Tween 80, and 0.001 %
Antifoam Y-30 Emulsion (Sigma). The inlet and the outlet tubing from the filter were placed in
the 500 mL elution solution. The screw clamp was loosened and the pump flow rate was set to
2000 mL/min. The elution solution was recirculated until the system began to draw up air. The
inlet tubing was then removed and eluent remaining in the ultrafilter and tubing was recovered in
a glass beaker. The elution process was repeated until the volume was as close as possible to 250
mL. For the study, the final UF concentrate volumes (retentate + eluent) were 490 ±38 mL. The
time required to perform the filtration and elution procedures was measured using a watch. The
time was noted when the filtration procedure was started and when the elution procedure was
completed (using the same watch).
2.4 EPA Ultrafiltration Set-Up
The EPA UF procedure was performed on a bench top in a BSL-2 laboratory (no microbes were
used that required a BSL-3 facility) (Figure 2). For each experiment, the EPA UF method was
performed using the EPA-developed WSC and its associated UF operational protocols (Figure
3). At the time of this study, the WSC was not commercially available.
13
-------
Figure 2. EPA water sample concentrator set-up at CDC laboratory facility.
The WSC primarily consisted of a modified peristaltic pump [Masterflex® I/P pump drive, (Cole
Farmer model 77401-00) and I/P Easy-Load® pump head (Cole Farmer model 77601-00)],
tubing pinch valves, sensors, DC power supplies for the valves and sensors, and data acquisition
modules that facilitated communication between the computer and the various electrical
components. Pre-made filtration assemblies were installed into the device prior to a
concentration run. A filtration assembly consisted of a Rexeed™-25S single-use dialysis filter
(Asahi Kasei Kuraray Medical Co. Ltd., Tokyo, Japan), sample bottle, tubing, fittings, clamps,
and pressure sensor. REXEED-25S dialyzers have an approximate MWCO of 30,000 daltons
and a surface area of 2.5 m2. The pump tubing was Tygon® Lab tubing R-3603 [9.5 mm (3/8 in)
ID x 16 mm (5/8 in) OD]. The pump tubing and the filter were connected by a coil of Tygon®
Lab tubing [6.3 mm (1/4 in) x 13 mm (1/2 in)], in order to dampen pulsations from the pump.
The remainder of the tubing was Tygon® silicone tubing 3350, of the following sizes, 6.3 mm
(1/4 in) x 9.5 mm (3/8 in), 6.3 mm (1/4 in) x 11 mm (7/16 in) and 9.5 mm (3/8 in) x 16 mm (5/8
in).
A key design feature of the filtration assembly was that the parts that came into contact with the
sample water were single use items (although for this study, some of these parts were disinfected
and recycled into new assemblies to save supply costs). Thus the valves used in the device were
solenoid pinch valves which resulted in only the tubing, and not the valve body, coming into
contact with sample water. Similarly, the water level in the sample bottle was measured via a
load cell, which the bottle rested on; this weight-based method allowed monitoring without a
sensor contacting the sample water.
The computer software controlled the multi-step concentration process by directing the operation
of the pump, valves, and by monitoring pressure, filtrate flow rate, and the amount of water in
the sample bottle. The inlet pressure was set at 30 psig and the filtrate pressure was
approximately at atmospheric pressure. If the inlet pressure exceeded 30 psig, the pump speed
14
-------
would decrease which resulted in a decrease in pressure. Similarly if the pressure was below 30
psig, the pump speed would increase to increase pressure. The filtrate flow rate typically started
off at -2,800 mL/min but decreased as the run progressed as the peristaltic pump tubing was
broken in. Likewise, an inlet pressure of 30 psig was maintained initially, but as the run
progressed, (and pump flow decreased), the pressure eventually decreased to below 30 psig
despite the pump running at maximum (650 RPM) speed. The average flow rate for a
concentration run was -1,700 mL/min; the average pressure was 25 psig. The retentate flow rate
was not measured but in previous work under similar conditions had been measured to be -4,000
mL/min on average over the course of a run.
2.4.1 Ultrafilter Blocking Solution. The REXEED-25S filters used in the EPA method
were blocked according to the EPA protocol NHSRC 004 [Reagent Preparation - Filter Blocking
Solution (0.055% Tween 80, 0.001% Antifoam A, 0.1 % NaPP)] prior to each experiment. This
blocking solution is recirculated through the ultrafilter for 3 min, after which most the solution is
removed from the system, but 250 mL is retained in the retentate bottle. Then the influent tubing
is placed in the sample container to begin sample processing.
2.4.2 Sample Processing. The EPA method was performed by processing the 100-L tap
water sample in the tank at an average system pressure of 25 psig and a flow rate of 1738
mL/minute. During each experiment, the water sample in the tank was manually stirred every 10
minutes. The process began with filter blocking and ended with the elution of the filter with an
elution solution. One day prior to the experiment, fresh 1-L volumes of both the blocking and
elution solutions were made. At the start of the run, a pre-made filtration assembly (provided by
EPA) was installed into the WSC per instructions in the operator's manual (22) (Figure 3). After
installation of the assembly, a volume of filter blocking solution was drawn up through the
sample inlet port. After a 3 minute recirculation period, the blocking solution exited the system
through the filtrate port. This was followed by the device drawing up and concentrating the
water sample, and then by a drawing up multiple volumes of elution solution through the sample
inlet port. During the UF process the software would prompt the user to perform simple steps
such as placing the sample inlet into the filter block solution, water sample, and elution solution
(0.001% Tween 80). The software also monitored operational parameters, including sample
volume processed, system pressure, and filtrate flow rate (Figure 4). The final target volume for
UF concentrates using the WSC was 450 ± 25 mL after elution. The time was noted when the
filtration procedure was started and when the elution procedure was completed using the same
watch.
15
-------
nitrate
va
V4
V3
retentate
filter, Rexeed, 2.5 m2
surface area 29 kDa
silicone, 3/8x5/8
Tygon Lab, 3/8x5/8
Tygon Lab, 1/4x1/2
silicone, 1/4x3/8
silicone, 1/2x7/1 G
quick disconnects — i
1 liter
polypropylene
bottle
Forward dir
Masterflex pump, size I/P,
Easy Load pump head
sample
100 L water sample
,10 ml pipette with tip
removed
Figure 3. Schematic of ultrafiltration set-up for EPA method ("V" indicates valve
location).
16
-------
Manual Pump Speed
AUTOMATED ULTRAFILTRATION CONCENTRATOR PROGRAM
Pressure
PAUSE
1,00-
0.900-.
0.800 -_
0.700-
0.600 -i
0300 -.
0.400 -I
O.ZDQ-H
o.ioo ; 1
Figure 4. View of the water sample concentrator monitoring screen as seen during an
ultrafiltration run.
2.5 Data Analysis
Calculation of microbial counts in each sample was performed using calculated concentration
data and total sample volume data for input samples (i.e., non-concentrated, seeded 500-mL
Control Samples) and output samples (i.e., UF concentrates, centrifuge concentrates for low-seed
Cryptosporidium and Giardid). Concentration data were calculated on a per mL basis, using
total microbial counts for a plate/slide [e.g., 20-60 CPU, 20-100 PFU, 20-100 (oo)cysts] per the
sample volume assayed. Total counts of each microbe for each sample were calculated by
multiplying the calculated concentration by the total sample volume.
Percent recovery efficiency was computed for each microbe using the following equation:
R= 100x(N/T)
Where: R = percent recovery, N = number of the microbe calculated to be in concentrated
sample, T = number of the microbe calculated to be in the control sample (i.e., non-concentrated
sample).
17
-------
Comparative recovery efficiency data were statistically analyzed for Suite 1 to Suite 5
experiments using paired t-tests at an alpha level of 0.05. CDC and EPA method data were
paired based on date of experiment. The difference between pairs was checked for normality
using the Shapiro-Wilk W test (IMP 9.0.2, used for all statistical analyses). When the data was
not normally distributed the Wilcoxon signed rank test was used instead. When comparing
different suites for a single method, as well as water quality data from experiments performed at
the CDC Chamblee and Roybal facilities, analysis of variance (ANOVA) was used. A two-sided
F-test for variance was performed to determine the appropriate statistical procedure (t-test,
ANOVA). The Bonferroni correction for multiple comparisons (n = 30) was used when
performing the ANOVA test for difference in microbial recovery efficiencies between the EPA
and CDC/LRN methods. The Wilcoxon rank sum test was used when comparing different suites
for which data was not normally distributed.
2.6 Blanks and Controls
For every five UF experiments performed, one 100-L tap water procedural blank was processed
for both the EPA and CDC/LRN UF methods. This quality control measure enabled evaluation
for potential background contamination (e.g., from laboratory environment or from drinking
water system).
Sample analyses were performed with an analytical positive control and negative control for
each analytical parameter. Positive control data (e.g., number of CFU on a B. anthracis positive
control plate) were compared against expected results to determine whether analytical conditions
were appropriate.
18
-------
3.0 Results
3.1 Water Quality
After collection of the 100-L water samples for an experiment, a suite of water quality tests were
performed to characterize the water samples. These water quality data are summarized for the
two laboratory facilities (Chamblee and Roybal campuses) in Table 3. Free chlorine residuals at
both facilities were within normal ranges for drinking water. Post-dechlorination testing
demonstrated that water samples contained no free chlorine when microbes were seeded into the
water samples. Post-dechlorination free chlorine results were equal to or below the method
detection limit for the analytical method (0.03 mg/L). The average pH of the water samples (8.8 -
9.0) reflected the higher pH employed by the water utility to control corrosion in the drinking
water distribution system. Turbidity of the tap water varied from 0.078 nephelometric turbidity
units (NTU) to 1.06 NTU, but average turbidity levels were similar at the two facilities. Total
hardness and specific conductance data indicate that the water used in this study would be
classified as soft with a low ionic strength. TOC concentrations in tap water at both facilities
were also similar (average 3.1 and 2.7 mg/L). Heterotrophic plate count (HPC) bacteria levels
were low in tap water from both facilities, but concentrations were more variable in tap water at
the Chamblee campus. While it appears that tap water quality at the Chamblee and Roybal
facilities was similar, data for the following parameters was found to be significantly different
statistically: turbidity (p = 0.03), specific conductance (p <0.0001), pH (p = 0.0001), and free
chlorine (p<0.0001).
Table 3. Water Quality Data for 100-L Tap Water Samples
Source
Chamblee
Roybal
Avg
SD
n
Avg
SD
n
Free
Chlorine
(mg/L)
0.94
±0.36
40
1.2
±0.14
19
pH
9.0
±0.33
41
8.8
±0.37
19
Temp
22
±3.3
41
24
±0.81
9
Turb
(NTU)
0.20
±0.13
43
0.32
±0.26
19
SC
117
±38.8
42
141
±54.1
19
T. Hardness
(mg/L Ca as
CaCO3)
16
±2.5
20
16
±1.6
19
TOC
(mg/L
asC)
3.1
±1.9
34
2.7
± 1.1
18
HPC
(CFU/mL)
12
± 14
40
3.3
±2.9
17
CPU, colony forming units; HPC, heterotrophic plate count; SC, specific conductance; SD; standard deviation; T.,
total; TOC, total organic carbon; Turb, turbidity
3.2 Operations and Safety
During operation of the WSC instrument to perform the EPA UF method, the system software
automatically monitored pressure and flow rate. For the CDC/LRN UF procedure, the filtrate
rate was set by manually adjusting a tubing clamp on the return tubing at the start of an
19
-------
experiment; system pressure monitoring was not performed. No evidence of filter clogging was
observed (i.e., no increase in pressure or decrease in filtrate rate with cumulative increase in
sample volume filtered) for either method/filter type (Table 4). Filtrate rates for the EPA method
were slightly higher than for the CDC method; the higher filtrate rates and automation of the
procedure resulted in the EPA procedure requiring approximately 20 fewer minutes to complete
than the CDC/LRN UF procedure.
While some software and maintenance issues were encountered when operating the WSC
instrument early in the project, these issues were readily resolved with improvements to the
software and changes to the system components. During Suite 4 and 5 experiments, few
operational problems were encountered when operating the WSC instrument. No explicit safety
issues were encountered when performing either UF procedure, but it should be noted that the
WSC instrument performs filtration in a contained system. Unless tubing connections are loose
(risking sample leakage or spraying), there is no risk for sample exposure. The CDC/LRN
method does not employ a completely contained system; tubing must be manually handled to
switch-out carboys, retentate is collected in open beakers, and sample concentration is performed
in open-mouth containers instead of a bottle enclosed with a vented cap. The potential risk for
aerosol exposure with the CDC/LRN method must be controlled by performing the procedure in
a laboratory having an appropriate biosafety level (BSL) and through the use of appropriate
personal protective equipment (PPE) such as a powered air purifying respirator.
Table 4. Operational Data for EPA and CDC/LRN Ultrafiltration (UF) Methods for 100-L
Water Samples
Method
EPA
CDC/LRN
Filter
REXEED
25 SX
Fresenius
F200NR
System
Pressure
(psig)
25 ±2.0
NA
Filtrate Flow
Rate (mL/min)
1800 ±250
1700 ± 180
UF
Concentrate
Volume (mL)
420 ± 72
490 ±38
UF
Processing
Time (min)
60 ± 10
80 ±8
NA: Not analyzed
3.3 Microbial Recoveries
3.3.1 Bacterial Recovery Efficiencies
B. anthracis and B. atrophaeus subsp. globigii. Recovery efficiencies for B. anthracis
spores in Suite 1 and 3 experiments were significantly higher for the EPA method than for the
CDC method (p = 0.01 and 0.001, respectively) (Table 5). The average recovery of B. anthracis
spores by the EPA method was similar in the Suite 1 and Suite 3 experiments (p = 0.07). The
average recovery of B. anthracis spores by the CDC method were significantly lower during
Suite 3 experiments (65 ± 14%) than during Suite 1 experiments (85 ± 17%) (p = 0.02).
20
-------
In Suite 3 experiments, B. atrophaeus subsp. globigii spore recovery was also significantly
higher for the EPA method (99 ± 11%) than the CDC/LRN method (57 ± 15%) (p<0.0001). Suite
3 recovery efficiency data for both CDC/LRN and EPA methods indicated that B. atrophaeus
subsp. globigii spores were effective surrogates for recovery of B. anthracis spores from tap
water samples, based on no significant difference in recovery efficiencies between the two (p =
0.07 and p = 0.18 for CDC and EPA method data, respectively).
In Suite 4 experiments, for which NaPP was only added for filter blocking but was not added to
the water sample processed using the EPA UF method, the EPA method resulted in significantly
higher recoveries of B. anthracis spores and B. atrophaeus subsp. globigii spores than the
CDC/LRN method (p = 0.001 and 0.0002, respectively). As observed for the Suite 3
experiments, Suite 4 recovery efficiency data for the EPA method indicated that B. atrophaeus
subsp. globigii spores and B. anthracis spores were recovered from 100-L water samples at
similar efficiencies (p = 0.27), but recovery efficiencies were significantly different when the
CDC/LRN method was used (p = 0.01).
C perfringens spores. The percent recovery of the C. perfringens spores was similar for
both the CDC (100 ± 22%) and the EPA (110 ± 27%) methods (p = 0.33). The high percent
recoveries measured were likely due to disaggregation of cell aggregates during ultrafiltration,
despite attempts to produce monodispersed C. perfringens spore seed stocks by vigorous mixing
and filtration through 5-|im filters.
E.faecalis. Percent recoveries ofE.faecalis were similar for both the CDC/LRN method
(97 ± 12%) and the EPA method (100 ± 12%) in Suite 1 experiments (p = 0.4829). In Suite 5
experiments, E. faecalis recovery efficiencies associated with the CDC/LRN procedure were
similar to Suite 1 data (p = 0.43). E.faecalis recovery efficiencies for the EPA method were
significantly lower in Suite 5 experiments than in Suite 1 experiments (p = 0.0049). Suite 5 E.
faecalis recovery efficiencies were significantly higher for the CDC/LRN method than the EPA
method (p = 0.0001). In addition to measuring lower E. faecalis recovery efficiencies for the
EPA method in Suite 5 experiments, it was also observed that E. faecalis colony development
was substantially slower for the EPA UF concentrate samples. When EPA sample plates were
held an additional 24 h (i.e., 48 h incubation), 19% more colonies were counted on EPA agar
plates but no additional colonies were counted on CDC/LRN plates.
Y. pestis. Percentage recoveries of Y. pestis were similar for both the CDC/LRN method
(70 ± 16%) and the EPA method (70 ± 18%) in Suite 1 experiments (p = 0.93). In Suite 4
experiments, Y. pestis recovery efficiencies for the CDC/LRN method were higher (100 ± 38%),
but statistical analysis versus Suite 1 data did not find a significant difference (p = 0.15), likely
due to the higher variability in experimental results from the Suite 4 experiments. While Suite 4
Y. pestis recovery data for the EPA method (76 ± 22%) was also found to not be significantly
different than in Suite 1 (p = 0.53), the difference in Suite 4 recovery efficiencies between the
two UF methods was significantly different (p = 0.03).
21
-------
Table 5. Average Microbial Recovery Efficiencies for the CDC and EPA Ultrafiltration Procedures
Microbe*
B. anthracis spores
B. atrophaeus subsp.
globigii spores
E.faecalis
Y. pestis
C. perfringens spores
F. tularensis
F. tularensis (1%
NH4CI)
B. diminuta
MS2
PhiX174
Echovirus 1
C. pan/urn
Color Seed C. pan/urn
G. intestinalis
Color Seed G.
intestinalis
Methods
CDC
Average Percent Recovery; CV (coefficient of variance)
Suite 1 Suite 2
85**; 20
97; 13
70; 23
100; 22
17*; 41
23*; 38
110; 34
100*; 13
68*; 38
82; 36
38; 33
99*; 18
42; 25
Suite 3 Suite 4 Suite 5
65*; 21 74*; 35
57*; 26 47*; 48
100*; 12
100*; 37
13*; 67 29; 72
29*; 41 62; 15
83; 18 82; 19
98*; 5.7
110*; 11
79*; 35
72*; 14
78*; 34
EPA
Average Percent Recovery; CV
Suite 1 Suite 2 Suite 3 Suite 4
100*; 12 92*; 11 96*; 35
99*. n 89*. 34
100; 10
70; 26 76*; 28
110; 24
29*; 52 25*; 75 27; 61
39*; 38 40*; 44 46; 48
84; 11 78; 16
120; 28
95*; 12
47*; 33
73; 39
30; 72
85*; 17
44; 53
Suite 5
63*; 33
69*; 17
86*; 16
37*; 62
110*; 17
^^H
110*; 16
^^H
*Bacillus anthracis (Sterne) spores, Bacillus atrophaeus subsp. globigii, Enter ococcus faecalis, Yersinia pestis (A1122), Clostridium
perfringens spores, Francisella tularensis LVS, Brevundimonas diminuta, MS2 bacteriophage, phi X174 bacteriophage, echovirus type
Cryptosporidium parvum oocysts, and Giardia intestinalis cysts
* * Significant differences between CDC and EPA methods for an organism in a particular suit
1,
22
-------
F. tularensis andB. diminuta. Recovery of culturable F. tularensis was challenging for
both UF methods. Initially (for Suite 1), frozen stocks were used to seed water samples, but this
procedure was associated with CDC/LRN method recoveries on the order of 1% and EPA method
recoveries on the order of 10% (data not shown). After instituting use of an overnight culture ofF.
tularensis to seed water samples in Suite 1, recovery efficiencies were substantially higher for both
UF methods. However, while use of an overnight culture improved the culturability ofF. tularensis
in UF concentrates this procedure also resulted in highly variable input seeding levels (which
ranged from 3,000 to 290,000 CFU). Suite 1 F. tularensis recovery efficiencies using the EPA
method (29 ± 15%) were significantly higher (p = 0.0096) than the CDC/LRN method (17 ± 7.0%).
In Suite 3 experiments, F. tularensis recovery was investigated again, but in conjunction with a
potential surrogate microbe, B. diminuta. For the Suite 3 experiments, F. tularensis recovery
efficiencies using the EPA method (25 ± 19%) were significantly higher than the CDC method (13
± 9%) (p = 0.01). B. diminuta recovery efficiencies were high for both the EPA method (84 ± 9%)
and CDC/LRN method (83 ± 15%), and no significant difference was found in recovery efficiencies
between the two methods (p = 0.85).
In Suite 4 experiments, F. tularensis recovery efficiencies for the CDC method were higher (29 ±
21%) than in Suite 1 and Suite 3 experiments, but the differences were not significant (p = 0.052).
The use of NaPP only as a blocking agent in the Suite 4 experiments for the EPA method was not
associated with a significant effect on F. tularensis recovery (27 ± 16% versus 29 ± 15% and 25 ±
19% in Suites 1 and 3, respectively) (p = 0.98). In Suite 4, B. diminuta recovery efficiencies were
again very similar between the EPA and CDC/LRN UF methods (p = 0.30). B. diminuta recovery
efficiencies were found to be significantly higher than F. tularensis recovery efficiencies for both
the EPA and CDC/LRN methods when Suite 3 and 4 experiment data were combined for statistical
analysis (p = 0.001 for CDC/LRN method and p = 0.0012 for EPA method).
In an attempt to improve the culturability ofF. tularensis in UF concentrates, aliquots of UF
concentrates produced using each UF method were exposed to 1% ammonium chloride for 2 h prior
to membrane filtration and CHAB-A agar culture, as suggested by Valentine et al. (23). For the
CDC/LRN method, in which water samples were always amended with NaPP, exposure to 1%
ammonium chloride was associated with significantly higher recovery efficiencies (p <0.0001).
For EPA method experiments in which water samples were amended with NaPP (i.e., Suites 1 and
3), exposure to 1% ammonium chloride was also associated with a significant increase in recovery
efficiencies (p = 0.002). In Suite 4 experiments, in which water samples processed using the EPA
method were not amended with NaPP, exposure to 1% ammonium chloride appeared to be
associated with higher F. tularensis recovery efficiencies, but the differences were not significant (p
= 0.07). Incorporation of 1% ammonium chloride into the culture protocol did not impact relative F.
tularensis recovery efficiency differences between the two UF methods in Suite 1 and 3
experiments, for which the EPA method recoveries were still significantly higher (p = 0.0002). In
Suite 4 experiments, F. tularensis recovery efficiencies were higher for the CDC/LRN method (62
± 9.4%) than the EPA method (46 ± 22%) when UF concentrates were exposed to 1% ammonium
chloride prior to culture, but the differences were not significant (p = 0.23).
23
-------
3.3.2 Viral Recovery Efficiencies
MS2. In Suite 2 experiments, recovery efficiencies for the CDC/LRN and EPA UF methods
were high and variable (110 ± 38% and 120 ± 33%, respectively). No significant difference was
found between the two methods (p = 0.11). In Suite 5 experiments, MS2 recovery efficiencies for
the CDC/LRN method were similar to recovery efficiencies from Suite 2 experiments, but EPA
method MS2 recovery efficiencies were significantly lower in Suite 5 than in Suite 1 (p = 0.0004).
Suite 5 MS2 recovery efficiencies for the CDC/LRN method were significantly higher than the EPA
method (p = 0.003).
Phi X174. As found for MS2, phi XI74 recovery efficiencies in Suite 2 experiments were
also high for both the CDC/LRN method (100 ± 13%) and EPA method (95 ± 11%), but variability
in the data was much lower than for MS2. Consequently, while recovery efficiencies were similar
between the two methods, the CDC/LRN recovery efficiencies were found to be significantly higher
than for the EPA method (p = 0.02). In Suite 5 experiments, phi X174 recovery efficiencies
associated with the CDC/LRN method remained high (110 ± 12%). Recovery efficiencies for the
EPA method were lower (86 ± 14%), but were not significantly different than in Suite 2
experiments (p = 0.13). As determined for Suite 2, Suite 5 phi X174 recovery efficiencies were
significantly higher for the CDC/LRN method than the EPA method (p = 0.008).
Echovirus 1. Echovirus 1 recovery data for this study reflect recovery of the virus after
performing UF and secondary concentration using Centricon® Plus-70 microconcentrators.
Recoveries of echovirus 1 were significantly higher in Suite 2 for the CDC UF method (68 ± 26%)
than the EPA method (47 ± 15%) (p=0.03). In Suite 5, echovirus 1 recovery efficiencies remained
high for the CDC/LRN method (79 ± 27%), but recovery efficiencies for the EPA method were
slightly lower (37 ± 23%) than in Suite 2 (p = 0.25). As determined for the two bacteriophages,
echovirus 1 recovery efficiencies for the CDC/LRN method were significantly higher than for the
EPA method in Suite 5 experiments (p = 0.0008).
3.3.3 C. parvum and G. intestinalis Recovery Efficiencies
Average recoveries of high seed C. parvum oocysts were similar for the CDC/LRN (82 ± 29%) and
EPA (73 ± 28%) UF methods in Suite 2 experiments (p = 0.21). Recoveries of high seed G.
intestinalis cysts were also high (99 ± 18% and 85 ± 14% for the CDC/LRN and EPA methods,
respectively), and the difference between the methods was significantly different (p = 0.03). In
Suite 5 experiments, C. parvum and G. intestinalis (oo)cyst recovery efficiencies were significantly
higher for the EPA method (0.0047 and 0.0043, respectively). The C. parvum oocyst recovery
efficiencies for the EPA UF method were found to be significantly higher than the CDC/LRN UF
method (p = 0.0003). G. intestinalis cyst recovery efficiencies were also significantly higher for the
EPA UF method than the CDC/LRN UF method for Suite 5 (p = 0.02).
In Suite 2, ColorSeed™ C. parvum and G. intestinalis (G. lamblia) (oo)cysts were concentrated by
UF, followed by centrifugation, and finally IMS before fluorescence microscopy analysis.
ColorSeed™ C. parvum oocyst recoveries associated with the EPA method (30 ± 22%) were not
significantly different than oocyst recoveries associated with the CDC/LRN method (38 ± 12%) (p
= 0.23). ColorSeed™ G. intestinalis cyst recoveries associated with the EPA method (44 ± 24%)
24
-------
were also not significantly different than cyst recoveries associated with the CDC/LRN method (42
± ll%)(p = 0.64).
3.4 Project Data Quality Objectives and Overall Microbial Recovery Efficiencies for Each
UF Method
In general, both the EPA and CDC/LRN UF methods recovered >50% of seeded microbes. To
enable effective statistical comparisons of recovery efficiencies for the two methods, the goal for
this project was to produce recovery efficiency data having coefficients of variation (CV) values
less than 25%. This data quality objective was achieved for 23 of the 46 (50%) of the recovery
efficiency percentages reported in Table 5 for high-seed microbial parameters. The most
challenging microbe to recover for both methods was Francisella tularemis. CV values for F.
tularensis recovery efficiencies were also generally above the target ceiling for the study. Other
microbes for which recovery efficiency CV data was higher than data quality objectives were B.
anthracis and B. atrophaeus subsp. globigii in Suite 4 experiments, Y. pestis in Suite 4, E.faecalis
in Suite 5 (EPA method only), MS2 in Suite 2, C. parvum in Suite 2, and G. intestinalis in Suite 5
(CDC/LRN method only). Recovery efficiency CV data were also relatively high for echovirus 1
and ColorSeed™ (oo)cysts, but these relatively higher CV values were expected because echovirus
1 and ColorSeed™ required additional sample processing steps (having additional processing
inefficiencies and variability).
ANOVA analysis (with Bonferroni correction) of microbial recovery data from Suite 1, 2 and 3
experiments indicated that the EPA method was associated with a significantly higher (5.4% higher)
overall microbial recovery efficiency than the LRN method. This performance difference between
the two methods was largely driven by differences in method performance for recovering B.
anthracis spores and B. atrophaeus subsp. globigii spores. With data for these two microbes
removed from the analysis, no significant difference in overall microbial recovery was observed
between the two UF methods. For Suite 4 and 5 experiment data, including B. anthracis and B.
atrophaeus subsp. globigii spore data, ANOVA analysis (with Bonferroni correction) found no
significant difference between the EPA and CDC/LRN UF methods. When testing for the potential
effect of performing Suite 4 and 5 experiments at a different laboratory facility, no significant
difference in overall microbial recovery efficiency was found for the CDC/LRN method (p = 0.45).
The same analysis for EPA method data found that combined Suite 1, 2 and 3 microbial recovery
efficiencies were not significantly different than Suite 4 and 5 microbial recovery efficiencies (p =
0.39).
The change in laboratory facilities between Suite 1-3 and Suite 4-5 experiments was also associated
with a change in laboratory protocol when NaPP was not used as a sample amendment for water
samples processed using the EPA UF method. This change in lab location and NaPP protocol was
not associated with significantly different recovery efficiencies for the EPA method when B.
anthracis and B. atrophaeus subsp. globigii data was grouped together (98% combined recovery
efficiency for Suites 1 and 3 versus 93% combined recovery efficiency for Suite 4) (p = 0.20). A
similar analytical approach determined that EPA method recovery efficiencies for the virus
parameters (MS2, phi XI 74 and echovirus 1 grouped together) were significantly higher in Suite 2
(grouped average = 87%) than in Suite 5 (grouped average = 64%) (p = 0.01). However, the
opposite association was found for the parasite parameters. High seed (oo)cyst recoveries for the
25
-------
EPA method were significantly lower in Suite 2 (grouped average = 79%) than in Suite 5 (grouped
average = 109%) (p < 0.0001).
26
-------
4.0 Discussion
The data from this study demonstrate that the EPA and CDC/LRN UF methods were effective at
recovering diverse microbes in 100-L drinking water samples. Despite significant differences
between the EPA and CDC/LRN UF methods (e.g., blocking solution composition and method,
pump size and recirculation flow rate, ultrafilter type, elution solution and method, operating
pressure) both UF methods were able to recover >50% of seeded bacteria (with the possible
exception of F. tularensis\ viruses, and parasites. For echovirus 1 and ColorSeed™ (oo)cysts, total
method recoveries (including secondary processing steps after UF) were generally above 30%.
Although NaPP was added to water samples processed by both methods in Suite 1, 2 and 3
experiments (resulting in a concentration of 0.01% NaPP in the water samples), the EPA UF
method for the WSC does not typically include sample amendment with NaPP. Suite 4 and 5
experiments were more exemplary of EPA's likely methodology for the WSC, as NaPP was not
added to water samples processed using the EPA method (but NaPP was added to water samples
processed by the CDC/LRN UF method in Suite 4 and 5 experiments). The filter blocking method
used in the EPA protocol did result in some residual NaPP remaining in the system after the
blocking procedure was completed, but the corresponding amount of NaPP (< 0.00025%) was at
least 40-fold lower than the concentration of NaPP resulting from the sample amendment
procedure. The presence of NaPP at a level of 0.001% has been suggested in previous research as
being relatively ineffective for recovering microbes (e.g., E. coif) in water samples using tangential-
flow UF (9).
While both methods were found to be similarly effective overall, there were microbial recovery
performance differences between the methods for certain analytes. In particular, the data from this
study demonstrate that the EPA WSC UF method was more effective at recovering B. anthracis
spores than the CDC/LRN UF method. B. atrophaeus subsp. globigii spores were found to be good
surrogates for B. anthracis spores for both UF methods and were also determined to be more
effectively recovered using the EPA UF method versus the CDC/LRN UF method. In a 2007 study
report in which the EPA method was performed manually (before the WSC device was developed),
B. atrophaeus subsp. globigii spore recoveries were also found to be similar to B. anthracis spore
recoveries, although recoveries of both microbes (average = 26-32%) were lower than achieved
during the present study (4). The data from the present study support EPA's ongoing initiative to
develop quality control (QC) criteria for B. atrophaeus subsp. globigii for determining UF method
performance proficiency (24). Another study performed using a manual version of the EPA UF
method reported higher B. anthracis spore recoveries (80 ± 44%) when a Fresenius F200NR
ultrafilter was used (3). For 7 pestis, no performance difference between the methods was found,
except in the Suite 4 experiments when recovery efficiencies for the CDC/LRN UF method were
substantially higher than in Suite 1 and were found to be significantly higher than paired recovery
efficiencies for the EPA method. Yersiniapestis recovery efficiencies for the EPA UF method in the
present study (average = 70-76%) were similar to 7 pestis recovery efficiencies reported previously
for a similar EPA UF method (84 ± 38%) when Fresenius F200NR filters were used (3). For F.
tularensis, it was found that recovery efficiencies were significantly higher for the EPA UF method
than the CDC/LRN method when NaPP was used as a sample amendment for both methods. When
NaPP was not used as a sample amendment for water samples processed using the EPA UF method,
there was no significant difference between the two UF methods, but this was due to higher
recovery efficiencies for the CDC/LRN method in Suite 4 experiments than were obtained in Suite
27
-------
1 and 3 experiments. F. tularensis recovery efficiencies for the EPA UF method in the present study
were within the range of average recovery efficiencies (18-103%) reported for a previous EPA UF
method when using Fresenius F200NR ultrafilters to process seeded water samples from three US
cities (10). In all F. tularensis experiments from the present study, it was found that average
recovery efficiencies for both the EPA and CDC/LRN methods were higher when UF concentrates
were exposed to 1% ammonium chloride for 2 h prior to culturing. These data indicate that
protocols for culturing F. tularensis from water samples should include this ammonium chloride
exposure technique. A recent report from Pacific Northwest National Laboratory researchers
indicated that exposure to 1% ammonium chloride was of significant benefit in preserving forensic
analysis specimens for F. tularensis testing (23).
For the non-biothreat agent bacterial parameters investigated in this study, significant differences
were found between the EPA and CDC/LRN UF methods for some of the analytes, but there were
no consistent overall trends. No significant differences were found between the EPA and CDC/LRN
UF methods for recovering C. perfringens spores or B. diminuta. B. diminuta was of interest in this
study because it is small and similar in size to F. tularensis (0.2-0.3 jim). However, recovery/culture
efficiencies were significantly higher for B. diminuta than F. tularensis, which raises the issue of
whether it is a useful indicator of the effectiveness of a UF procedure for recovering/culturing F.
tularensis. The other bacterial parameter in this study, E. faecalis, was included because it has been
proposed as a QC parameter for establishing proficiency for the CDC/LRN UF method (24). In the
present study, no significant difference was found between the EPA and CDC/LRN UF methods for
recovering E. faecalis when NaPP was used as a sample amendment for both methods. But when
NaPP was not used as a sample amendment for water samples processed using the EPA method, E.
faecalis recovery efficiencies were significantly lower for the EPA method than the CDC/LRN
method. E. faecalis recovery efficiencies for the CDC/LRN UF method in the present study were
similar to E. faecalis recovery efficiencies reported by Hill et al. for a similar UF method applied to
tap water (8) and by EPA for a multi-lab oratory study of E. faecalis recovery as a QC parameter for
the CDC/LRN UF method (24).
Bacteriophages MS2 and phi XI74 were used as model enteric viruses in the present study based on
prior research recommending them as useful models for water sampling methods based on
morphological and surface charge characteristics (25). MS2 was recovered at a high level by both
UF methods. In Suite 5 experiments, MS2 recovery efficiencies were significantly lower for the
EPA UF method than in Suite 1 and were significantly lower than MS2 recovery efficiencies for the
CDC/LRN method. This apparent association of higher MS2 recovery efficiencies with the use of
NaPP as a sample amendment has been reported previously (6, 9). The CDC/LRN MS2 recovery
efficiencies were similar to MS2 recovery efficiencies previously reported by Hill et al for a similar
UF method (120 ± 22%) and were greater than MS2 recovery efficiencies reported by 13
laboratories (67 ± 8.3%) for a QC study for the CDC/LRN UF method (8, 24). Average MS2
recovery efficiencies for the EPA UF method (using the REXEED-25S filter, without NaPP sample
amendment) in the present study (69 ± 12%) were slightly greater than average MS2 recovery
efficiencies reported previously for a manual version of this method (52 ± 34%) (3). Phi XI74 was
also recovered at a high level by both UF methods, but recoveries were found to be significantly
higher for the CDC/LRN method. The CDC/LRN phi XI74 recovery efficiencies were slightly
higher than phi XI74 recovery efficiencies previously reported by Hill et al for a similar UF method
(86 ± 13%) (8). Average phi XI74 recovery efficiencies for the EPA UF method (using a
28
-------
REXEED-25S filter, without NaPP sample amendment) in the present study (86 ± 14%) were
greater than average phi XI74 recovery efficiencies reported previously for a manual version of this
method (57 ± 34%) (3).
The third viral parameter included in this study, echovirus 1, was seeded at a relatively low level
(approximately 1,000 PFU) that only enabled determination of recovery efficiencies after secondary
concentration. Reported echovirus 1 recovery efficiencies for each UF method in Table 3 were
lower than for MS2 and phi XI74, but these relatively lower recovery efficiencies for echovirus 1
reflect additional sample processing losses associated with the Centricon procedure. Total method
recovery efficiencies for echovirus 1 from the present study (average = 37-47% for EPA method,
68-79% for CDC/LRN method) were similar to or higher than echovirus 1 recovery efficiencies
reported in other UF studies (6, 26). The factors associated with the significantly higher echovirus 1
recoveries for the CDC/LRN method versus the EPA method are not clear. Higher recoveries for
MS2 and phi XI74 were also observed for the CDC/LRN method.
Recovery of parasite (oo)cysts by tangential flow UF have been studied extensively. In the present
study, C. parvum and G. intestinalis (oo)cyst recoveries were similar between the two UF methods,
and were generally above 70%. Similar UF recovery efficiencies for large-volume water samples
have been previously reported (4, 8, 6, 26). ColorSeed™ (oo)cyst recovery efficiencies in the
present study were also similar to low-seed C. parvum oocyst recoveries reported by Holowecky et
al. (3), and were slightly lower than C. parvum and G. intestinalis recovery efficiencies reported by
Hill et al. (26). ColorSeed™ recovery efficiencies were lower than for high seed C. parvum and G.
intestinalis recoveries, but this was expected because of (oo)cyst losses inherent in additional
sample processing (centrifugation, IMS) that was required for enumeration of ColorSeed™
(oo)cysts in water sample concentrates. In the present study, when NaPP was used as a sample
amendment the recovery efficiencies for the EPA and CDC methods were similar for the high seed
and ColorSeed™ (oo)cysts. In Suite 5 experiments, when NaPP was not used as a sample
amendment for water samples processed by the EPA UF method, (oo)cyst recoveries were
significantly higher than in Suite 1 and were significantly higher than for the CDC/LRN UF
method. The apparent association of NaPP sample amendment with lower C. parvum oocyst
recoveries for tangential flow UF was not expected based on previous published research results
indicating that NaPP and associated polyphosphates are effective in dispersing Cryptosporidium
oocysts, which should enable more efficient recovery during UF. Previous studies have reported
that higher C. parvum oocyst recovery efficiencies were obtained during tangential flow UF when
NaPP was used as a water sample amendment (6, 9). Other researchers have reported using NaPP to
disperse Cryptosporidium oocysts prior to flow cytometry (27).
The results of this study demonstrate that the EPA and CDC/LRN UF procedures are effective at
recovering diverse microbes from 100-L drinking water samples. When recovery data for all the
microbial analytes were combined, a statistically significant difference between the two methods
was observed for Suite 1, 2 and 3 experiments, indicating that the EPA UF method obtained higher
recovery efficiencies for these experiments than the CDC/LRN UF method. This apparent
difference in method performance was driven by consistently higher recoveries of B. anthracis
spores and B. atrophaeus subsp. globigii spores. No significant difference between the EPA and
CDC/LRN methods was observed when Suite 4 and 5 data were statistically analyzed. Recovery
and culturability of F. tularensis was challenging for both UF methods, but exposure of UF
29
-------
concentrates to 1% ammonium chloride for 2 h was found to consistently increase the culturability
of F. tularemis by approximately 35-120%. The move to a new laboratory facility for Suite 4 and 5
experiments did not appear to affect experimental results based on potential water quality effects.
Although some significant differences in water quality were measured, the significant differences
were considered to be more reflective of low water quality variability rather than reflecting
biologically or chemically plausible differences that could cause differential recovery efficiencies
for UF. The move to a new laboratory facility was also associated with a change in protocol; NaPP
was not used as a sample amendment for water samples processed by the EPA UF method in Suite
4 and 5 experiments. The NaPP experimental variable was not associated with consistent trends in
microbial recovery efficiencies for the EPA UF method. When NaPP was used as a sample
amendment, higher recovery efficiencies were measured for E. faecalis and MS2, but lower
recovery efficiencies were measured for C. parvum and G. intestinalis.
30
-------
5.0 Conclusions
1. Data from this study indicate that the CDC and the EPA UF methods can be similarly
effective for the recovery of diverse biothreat agents in large-volume drinking water samples
and thus validation of the EPA method for use during a response event is recommended.
2. UF recovery efficiencies were >50% for both methods for B. anthracis spores, 7 pestis, E.
faecalis, C. perfringem spores, B. diminuta, MS2 bactedophage, phi XI74 bacteriophage, C.
parvum., and G. intestinalis.
3. The lowest UF recovery efficiencies obtained in this study were for F. tularensis, but the use
of 1% ammonium chloride was found to significantly increase the culturability of F.
tularensis in UF concentrates.
4. ColorSeed™ recoveries were similar for the EPA and CDC/LRN methods, with C. parvum
oocyst recoveries of 30 and 38%, respectively, and G. lamblia recoveries of 44 and 42%,
respectively.
5. In general, data quality objectives (DQOs) for this project were met, including generating
recovery efficiency data for high seed microbes with standard deviations < 20% and CV
values < 25%; standard deviation and CV value DQOs for low seed microbes were < 25%
and < 50%, respectively. Microbial data of note that did not meet DQOs were: B. anthracis
and B. atrophaeus subsp. globigii (Suite 4, both methods), F. tularensis (throughout the
study), E.faecalis (Suite 5, EPA method), 7 pestis (Suite 4, EPA method), MS2 (Suite 2,
both methods), C. parvum (Suite 2, both methods), and G. intestinalis (Suite 5, CDC/LRN
method).
6. No significant difference in overall microbial recovery efficiency was observed between the
EPA and CDC/LRN UF recovery methods in Suite 4 and 5 experiments. However,
differences for individual microbial parameters were observed.
7. Significantly higher recovery efficiencies for the EPA UF method were found for B.
anthracis spores, B. atrophaeus subsp. globigii spores, C. parvum., and G. intestinalis.
8. Significantly higher recovery efficiencies for the CDC/LRN UF method were found for E.
faecalis, MS2, phi XI74, and echovirus 1.
9. The use of NaPP did not appear to be associated with a consistent trend in microbial
recovery efficiency, but effects on recovery of individual analytes was indicated. When
NaPP was used as a sample amendment, higher recovery efficiencies were measured for E.
faecalis and MS2, but lower recovery efficiencies were measured for C. parvum and G.
intestinalis. Use of NaPP as a sample amendment when operating the WSC under field
conditions does not appear to be warranted considering the additional sample process
complexity associated with adding NaPP in-line as a water sample is being processed.
-------
10. The time required to concentrate 100 L of tap water using the CDC method (average = 80
min for filtration and elution) was approximately 20 min longer than the time required by
the EPA method (average = 60 min for filtration and elution).
32
-------
6.0 Presentations and Other Activities
Results from this project were presented in part as a poster at the 2010 American Society for
Microbiology Biodefense and Emerging Diseases Research Meeting, Baltimore, MD. The poster
presentation was entitled "Comparative Performance of Hollow-Fiber Ultrafiltration Procedures for
Recovery of Biothreat Agents from 100-L Tap Water Samples" and was co-authored by S. Pai, T.
Lusk, V. Gallardo, S. Shah, H.D.A. Lindquist, and V.R. Hill.
Funds for this project were also used for travel by the Principal Investigator, Vincent Hill, to
Virginia to participate in a workshop on persistence of biothreat agents in the environment. Dr. Hill
led the discussion on persistence of viruses.
33
-------
7.0 References
1. Burrows, W. D., and S. E. Renner. 1999. Biological warfare agents as threats to potable
water. Environmental Health Perspectives 107(12):975-984.
2. U.S. Environmental Protection Agency (EPA). 2009. Standardized Analytical
Methods for Environmental Restoration Following Homeland Security Events (SAM).
EPA/600/R-04/126E September 2009. Cincinnati, Ohio: United States Environmental
Protection Agency, National Homeland Security Research Center.
3 Holowecky, P. M., R. R. James, D. P. Lorch, S. E. Straka, and H. D. A. Lindquist. 2009
Evaluation of ultrafiltration cartridges for a water sampling apparatus. J Appl Microbiol
106(3):738-747.
4. Lindquist, H. D., S. Harris, S. Lucas, M. Hartzel, D. Riner, P. Rochele, and R. DeLeon.
2007. Using ultrafiltration to concentrate and detect Bacillus anthracis, Bacillus atrophaeus
subspecies globigii, and Cryptosporidium parvum in 100-liter water samples. J Microbiol
Methods 70:484-492.
5. Morales-Morales, H. A., G. Vidal, J. Olszewski, C. M. Rock, D. Dasgupta, K. H.
Oshima, and G. B. Smith. 2003. Optimization of a reusable hollow-fiber ultrafilter for
simultaneous concentration of enteric bacteria, protozoa, and viruses from water. Appl
Environ Microbiol 69(7):4098-4102.
6 Polaczyk, A. L., J. Narayanan, T. L. Cromeans, D. Hahn, J. M. Roberts, J. E.
Amburgey, and V. R. Hill. 2008. Ultrafiltration-based techniques for rapid and
simultaneous concentration of multiple microbe classes from 100-L tap water samples. J
Microbiol Methods 73(2):92-99.
7. Smith, C. M., and V. R. Hill. 2009. Dead-end hollow-fiber ultrafiltration for recovery of
diverse microbes from water. Appl Environ Microbiol 75:5284-5289.
8. Hill, V. R., A. M. Kahler, N. Jothikumar, T. B. Johnson, D. Hahn, and T. L. Cromeans.
2007. Multistate evaluation of an ultrafiltration-based procedure for simultaneous recovery
of enteric microbes in 100-liter tap water samples. Appl and Environ Microbiol
73(13):4218-4225.
9. Hill, V. R., A. L. Polaczyk, D. Hahn, J. Narayanan, T. L. Cromeans, J. M. Roberts, and
J. E. Amburgey. 2005. Development of a rapid method for simultaneous recovery of
diverse microbes in drinking water by ultrafiltration with sodium polyphosphate and
surfactants. Appl and Environ Microbiol 71(ll):6878-6884.
10 O'Reilly, C. E., A. B. Bowen, N. E. Perez, J. P. Sarisky, C. A. Shepherd, M. D. Miller,
B. C. Hubbard, M. Herring, S. D. Buchanan, C. C. Fitzgerald, V. Hill, M. J. Arrowood,
L. X. Xiao, R. M. Hoekstra, E. D. Mintz, and M. F. Lynch. 2007 A waterborne outbreak
of gastroenteritis with multiple etiologies among resort island visitors and residents: Ohio,
2004. Clin Infect Dis 44(4):506-512.
11. LRN. 2007. Filter Concentration for the Detection of Bioterrorism Threat Agents in Potable
Water Samples. Centers for Disease Control and Prevention.
12. Hach. 2002. DR/2400 Spectrophotometer Procedure Manual. Hach Company.
13. American Public Health Association, American Water Works Association, and Water
Environment Federation. 2005. Standard Methods for the Examination of Water and
Wastewater, 21 ed. American Public Health Association, New York.
34
-------
14. Cromeans, T. L., A. M. Kahler, and V. R. Hill. 2010. Inactivation of adenoviruses,
enteroviruses, and murine norovirus in water by free chlorine and monochloramine. Appl
Environ Microbiol 76(4): 1028-1033.
15 Petersen, J. M., M. E. Schriefer, K. L. Gage, J. A. Montenieri, L. G. Carter, M.
Stanley, and M. C. Chu. 2004. Methods for enhanced culture recovery ofFrancisella
tularensis. Appl Environ Microbiol 70(6):3733-373 5.
16 Marston, C. K., C. Beesley, L. Helsel, and A. R. Hoffmaster. 2008 Evaluation of two
selective media for the isolation of Bacillus anthracis. Lett Appl Microbiol 47(1):25-30.
17 Russell, P., M. Nelson, D. Whittington, M. Green, S. M. Eley, and R. W. Titball. 1997
Laboratory diagnosis of plague. Br J Biomed Sci 54(4):231-236.
18. Bisson, J. W., and V. J. Cabelli. 1979. Membrane filter enumeration method for
Clostridiumperfringens. Appl Environ Microbiol 37(l):55-66.
19. EPA. 2005a. Method 1600: Enterococci in Water by Membrane Filtration Using
membrane-Enterococcus Indoxyl-Beta-D-Glucoside Agar (mEI). EPA-821-R-04-023.
Washington, D.C.: U.S. Environmental Protection Agency,Office of Water.
20. EPA. 2001. Method 1602: Male-specific (F+) and Somatic Coliphage in Water by Single
Agar Layer (SAL) Procedure. EPA 821-R-01-029. Washington, D.C.: U.S. Environmental
Protection Agency, Office of Water.
21. EPA. 2005b. Method 1623: Cryptosporidium and Giardia in water by filtration/EVIS/FA.
EPA 815-R-05-002. Washington, D.C.: U.S. Environmental Protection Agency, Office of
Water.
22. EPA. 2010b. Operator's Manual for the Water Sample Concentrator. EPA/600/X-10/017.
For limited distribution. Contact EPA author V. Gallardo, to request a copy of the manual."
23 Valentine, N. B., S. C. Wunschel, C. O. Valdez, H. Kreuzer, R. A. Bartholomew, T. M.
Straub, and K. L. Wahl. 2011. Preservation of viable Francisella tularensis for forensic
analysis. J Microbiol Methods 84(2):346-348.
24. EPA. 2010c. Draft. Quality Control (QC) Criteria Development for the Laboratory
Response Network (LRN) Filter Concentration (Ultrafiltration) Protocol Report. U.S.
Environmental Protection Agency, Office of Water.
25. Polaczyk, A.L., J.M. Roberts, and V.R. Hill. 2007. Evaluation of 1MDS electropositive
microfilters for simultaneous recovery of multiple microbe classes from tap water. J
Microbiol Methods 68(2):260-266.
26 Hill, V. R., A. L. Polaczyk, A. M. Kahler, T. L. Cromeans, D. Hahn, and J. E.
Amburgey. 2009. Comparison of hollow-fiber ultrafiltration to the USEPA VIRADEL
technique and USEPA method 1623. J Environ Qual 38(2):822-825.
27. Lepesteur, M., S. Blasdall, and N. J. Ashbolt. 2003. Particle dispersion for further
Cryptosporidium and Giardia detection by flow cytometry. Lett Appl Microbiol 37(3):218-
229.
35
-------
Appendix A: Quality Assurance/Quality Control
Data Quality Objectives
QA/QC Implemented
Data for comparison of the Environmental
Protection Agency (EPA) and Centers for
Disease Control and Prevention
(CDC)/Laboratory Response Network
(LRN) Ultrafiltration (UF) method
performance should be obtained in paired
experiments for statistical analysis.
Comparative recovery efficiency data was statistically analyzed
for Suite 1 to Suite 5 experiments using paired t-tests at an alpha
level of 0.05. CDC and EPA method data were paired based on
date of experiment. The difference between pairs was checked
for normality using the Shapiro-Wilk W test. When the data was
not normally distributed the Wilcoxon signed rank test was used
instead.
Percent recovery efficiency data for each
microbe studied should have a standard
deviation < 20% for each UF method.
Coefficients of variation (CV) values for
percent recovery efficiency data sets for
each microbe should be < 25%, with the
exception of F. tularensis (for which highly
variable recovery efficiency data is
anticipated).
hi general, data quality objectives (DQOs) for this project were
met. For high seed microbes recovery efficiency data had
standard deviations < 20% and CVvalues < 25%. For low seed
microbes, standard deviation and CV value DQOs were < 25%
and < 50%, respectively. Microbial data of note that did not meet
DQOs were: B. anthracis and B. atrophaeus subsp. globigii
(Suite 4, both methods), F. tularensis (throughout the study), E.
faecalis (Suite 5, EPA method), Y. pestis (Suite 4, EPA method),
MS2 (Suite 2, both methods), C. parvum (Suite 2, both methods),
and G. intestinalis (Suite 5, CDC/LRN method). See Section 3.4
for mo re information.
Real-time polymerase chain reaction (RT-
PCR) triplicate data for an assay should
have interassay CVvalues (standard
deviation divided by the mean) of < 4%.
Mean cycle threshold (CT) values should
be below a value of 40 facilitate
reproducibility and statistical analysis
(positive RT-PCRresults are generally
limited to CT values of < 42).
RT-PCR was performed early in the project to evaluate whether
this technique could be effective as an additional measure for
characterizing performance differences between the CDC/LRN
and EPA methods. This DQO was met for F. tularensis, but not
Y. pestis orB. anthracis (for which seed levels were too low for
consistent detections below CT = 40. It was determined that this
technique was not useful or needed for this project. Description
of this real-time PCR work was not incorporated into the main
body of the report in order to maintain clarity for readers by
focusing on culture- and microscopy-based data.
Procedural blanks and negative controls
should ensure that background or
introduced contamination does not affect
experimental data used for method
comparison analysis.
For every five UF experiments performed, one 100-Ltap water
procedural blank was processed for both the EPA and CDC/LRN
UF methods. This quality control measure enabled evaluation
for potential background contamination (e.g., from laboratory
environment or from drinking water system). Sample analyses
were performed with an analytical positive control and negative
control for each analytical parameter.
Positive control data should indicate that
analytical as say conditions met
performance expectations.
Sample analyses were performed with an analytical positive
control and negative control for each analytical parameter.
Positive control data (e.g., number of colony forming units on a
B. anthracis positive control plate) were compared against
expected results to determine whether analytical conditions were
appropriate.
36
-------
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |