filtration methods for recovery of bacillus anthracis spores spiked into source and finished water

13
Water Research 39 (2005) 5199–5211 Filtration methods for recovery of Bacillus anthracis spores spiked into source and finished water Abril Perez a , Christina Hohn b , James Higgins b, a City of Phoenix Water Services Laboratory, 2474 South 22nd Avenue, Phoenix, AZ 85009, USA b USDA-ARS, Bldg 173, 10300 Baltimore Blvd, Beltsville, MD 20705, USA Received 1 November 2004; received in revised form 3 October 2005; accepted 4 October 2005 Abstract Spores of Bacillus anthracis Sterne strain were recovered from 100 ml and 1 L volumes of tap and source waters using filtration through a 0.45 um filter, followed by overnight culture on agar plates. In a set of experiments comparing sheep red blood cell (SRBC) plates with a chromogenic agar formulation designed by R & F Laboratories, with a spiking dose of 47 plate-enumerated spores in 100 ml tap water, the mean spore recoveries were 34.0 and 30.8 spores, respectively. When a spiking dose of 100 fluorescence activated cell sorter(FACS)-enumerated spores was used in 100 ml potable water, the average recovery with SRBC plates was 48 spores. Detection efforts with spiking doses of 35 and 10 spores in 1 L tap water were successful, but recovery efforts from spiked 1 L volumes of source water were problematic due to the concomitant growth of normal spore-forming flora. Recoveries were also attempted on 10 L volumes of tap water. For a spiking dose of 100 spores, mean recovery from six replicates was 11 spores (76.8, range 2–20), and for a spiking dose of 10 spores, mean recovery from six replicates was 2.3 spores (73.5, range 0–9). Efforts were also made to ‘‘direct detect’’ spores via polymerase chain reaction (PCR) on washes from filters. When spiking 534 spores in 100 ml, 9/9 replicates of spiked tap water, 6/6 source water replicates, and 0/3 unspiked controls were positive by lef PCR. When 534 spores were spiked into 1 L tap water, the lef PCR was unsuccessful; however, using the nested vrrA PCR resulted in 4/9 spiked samples, and 0/3 unspiked controls, testing positive. Our results indicate that an inexpensive and user- friendly method, utilizing filtration apparatus commonly present in many water quality testing labs, can readily be adapted for use in detecting this potential threat agent. Published by Elsevier Ltd. Keywords: Bacillus anthracis; Recovery; Filtration 1. Introduction In the United States in the Fall of 2001, a terror attack using Bacillus anthracis spores distributed through mailing envelopes resulted in the deaths of five people from inhalation anthrax, and infection with either pulmonary or cutaneous anthrax in another 17 indivi- duals (Jernigan et al., 2002). To date, the perpetrator(s) of the attack have not been identified. Conventional speculation on the use of a weaponized B. anthracis envisaged scenarios in which an aerosolized formulation was distributed via airplane spraying or some other form of vehicle-mounted, specialized dispersal method. How- ever, this bioterror event provided graphic evidence of the ability of a relatively crude delivery method to cause widespread apprehension among the public, including ARTICLE IN PRESS www.elsevier.com/locate/watres 0043-1354/$ - see front matter Published by Elsevier Ltd. doi:10.1016/j.watres.2005.10.009 Corresponding author. Tel.: +301 504 6443; fax: +301 504 6608. E-mail address: [email protected] (J. Higgins).

Upload: independent

Post on 22-Jan-2023

1 views

Category:

Documents


0 download

TRANSCRIPT

ARTICLE IN PRESS

0043-1354/$ - se

doi:10.1016/j.w

�Correspondfax: +301 504 6

E-mail addr

Water Research 39 (2005) 5199–5211

www.elsevier.com/locate/watres

Filtration methods for recovery of Bacillus anthracis sporesspiked into source and finished water

Abril Pereza, Christina Hohnb, James Higginsb,�

aCity of Phoenix Water Services Laboratory, 2474 South 22nd Avenue, Phoenix, AZ 85009, USAbUSDA-ARS, Bldg 173, 10300 Baltimore Blvd, Beltsville, MD 20705, USA

Received 1 November 2004; received in revised form 3 October 2005; accepted 4 October 2005

Abstract

Spores of Bacillus anthracis Sterne strain were recovered from 100ml and 1L volumes of tap and source waters using

filtration through a 0.45 um filter, followed by overnight culture on agar plates. In a set of experiments comparing sheep

red blood cell (SRBC) plates with a chromogenic agar formulation designed by R & F Laboratories, with a spiking dose

of 47 plate-enumerated spores in 100ml tap water, the mean spore recoveries were 34.0 and 30.8 spores, respectively.

When a spiking dose of 100 fluorescence activated cell sorter(FACS)-enumerated spores was used in 100ml potable

water, the average recovery with SRBC plates was 48 spores. Detection efforts with spiking doses of 35 and 10 spores in

1L tap water were successful, but recovery efforts from spiked 1L volumes of source water were problematic due to the

concomitant growth of normal spore-forming flora. Recoveries were also attempted on 10L volumes of tap water. For

a spiking dose of 100 spores, mean recovery from six replicates was 11 spores (76.8, range 2–20), and for a spiking dose

of 10 spores, mean recovery from six replicates was 2.3 spores (73.5, range 0–9). Efforts were also made to ‘‘direct

detect’’ spores via polymerase chain reaction (PCR) on washes from filters. When spiking 534 spores in 100ml, 9/9

replicates of spiked tap water, 6/6 source water replicates, and 0/3 unspiked controls were positive by lef PCR. When

534 spores were spiked into 1L tap water, the lef PCR was unsuccessful; however, using the nested vrrA PCR resulted in

4/9 spiked samples, and 0/3 unspiked controls, testing positive. Our results indicate that an inexpensive and user-

friendly method, utilizing filtration apparatus commonly present in many water quality testing labs, can readily be

adapted for use in detecting this potential threat agent.

Published by Elsevier Ltd.

Keywords: Bacillus anthracis; Recovery; Filtration

1. Introduction

In the United States in the Fall of 2001, a terror attack

using Bacillus anthracis spores distributed through

mailing envelopes resulted in the deaths of five people

from inhalation anthrax, and infection with either

e front matter Published by Elsevier Ltd.

atres.2005.10.009

ing author. Tel.: +301 504 6443;

608.

ess: [email protected] (J. Higgins).

pulmonary or cutaneous anthrax in another 17 indivi-

duals (Jernigan et al., 2002). To date, the perpetrator(s)

of the attack have not been identified. Conventional

speculation on the use of a weaponized B. anthracis

envisaged scenarios in which an aerosolized formulation

was distributed via airplane spraying or some other form

of vehicle-mounted, specialized dispersal method. How-

ever, this bioterror event provided graphic evidence of

the ability of a relatively crude delivery method to cause

widespread apprehension among the public, including

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–52115200

disruption of mail delivery, and has resulted in the

implementation of irradiation protocols for mail ad-

dressed to upper-level politicians and government

officials.

One aspect of a bioterror event involving the use of

B. anthracis that has not received a great deal of

attention is water-borne transmission. We are aware

of no official account of such an event, but we believe it

prudent to ask if such an event can be dismissed as

unlikely. This is particularly relevant in light of the

arrest in the US in July, 2002, of two Islamic militants,

James and Mustafa Ujaama, on suspicion of plotting to

contaminate potable water sources (‘‘Feds Arrest Al

Qaida Suspects With Plans to Poison Water Supplies’’,

by Carl Cameron, Fox News Service, Tuesday July 30,

2002). (The Ujaamas were subsequently released after

agreeing to cooperate with authorities in their investiga-

tions of other suspected US-based terror cells). Another

relevant report, based on an article in an Arabic-

language newsmagazine, indicates that Al-Qaida opera-

tives are contemplating poisoning American water

supplies (‘‘magazine reports Al-Qaida Threat to US

Water Supply’’, Associated Press, Baltimore Sun, May

30, 2003). In addition to a bioterrorism scenario, covert

usage of bioweapons during warfare also cannot be

ruled out; during the campaign against Iraq in March/

April, 2003, Jordanian police arrested six Iraqis on

suspicion of plotting to poison a water tank located in

the desert near the border with Iraq. The tank supplied

water to US troops (‘‘Jordan arrests Iraqis in Plot to

Poison Water, and Probes Others’’, by Emily Wax,

Washington Post, Tuesday April 2, 2003).

There are several aspects of the biology of B. anthracis

that would make it more suitable than other threat

agents for deliberate introduction into source or

drinking water supplies: the environmental hardiness

of the spores, their resistance to levels of chlorine used to

treat drinking water (Burrows and Renne, 1999; Rose et

al., 2005), the protean symptomatology of gastrointest-

inal (GI) tract anthrax (Sirisanthana and Brown, 2002),

and the observation that mortality of this form of the

infection can approach 50% (Friedlander, 1997). Exist-

ing literature on the detection of B. anthracis spores in

water is scarce; a Russian study published in 1966

indicated that 100ml river water spiked with the

equivalent of 20 spores per 1ml could be detected by a

combination of filtration and overnight culture on

peptone/beef serum agar plates (Tomov, 1966). Studies

done on other species of the genus Bacillus indicate that

spores can be recovered from 1–100ml volumes of

source and finished waters using membrane filtration

and agar plate culture (Francis et al., 2001; Ostensvik

et al., 2004).

We investigated several methods for the detection of

B. anthracis spores in source and tap water. Our

approach was colored by the following considerations:

first, the methods should ideally be relatively inexpensive

and require modest equipment in order to be feasible for

routine use by testing laboratories associated with

municipal water facilities. Secondly, they should require

skills that are either already present among water quality

testing laboratory personnel, or alternatively, require a

modest amount of instruction in order to effectively

implement. Finally, we looked to design detection

protocols that were amenable to ramping-up in terms

of scale and number, as might be required during

periods when there is reason to perform frequent testing

of water samples for the presence of B. anthracis (i.e., in

response to threats, however believable, that this agent

has been introduced into water supplies).

This report describes the use of an overnight, culture-

based method, and a more immediate, polymerase chain

reaction (PCR)-based ‘‘direct detection’’ method, for the

detection of B. anthracis spores in source and tap water.

2. Materials and methods

2.1. B. anthracis spores

B. anthracis Sterne strain was a kind gift of Dr.

Catherine Fenselau, University of Maryland at College

Park, MD. This BSL-2 strain lacks the pXO2 plasmid

that codes for the capsule genes, and historically was

used as a veterinary vaccine. Aliquots (100–200 ml) of

spores in sterile, reagent-grade water were quantitated

by heat shock at 60–65 1C for 20min and plating on

sheep red blood cell (SRBC) agar plates (Remel, Lenexa,

Kansas). The plates were incubated overnight at 37 1C

and examined the next morning for the appearance of

colony forming units (cfu) with an appearance sugges-

tive of B. anthracis Sterne (i.e., gray or whitish colonies

with a ‘‘ground glass’’ appearance, with weak or absent

evidence of hemolytic activity). Another agar formula-

tion, supplied by R&F Laboratories (West Chicago, IL),

also was used to culture some of the spores recovered

from spiked water samples; with this formulation, the B.

anthracis colonies initially appear as white colonies with

irregular contours and hairlike projections; after �24 h

incubation at 37 1C, the centers assume a dark blue

coloration. In contrast, nonB. anthracis spp. of Bacillus

(e.g., B. cereus) grow as colonies with a well-defined

outline and assume a uniform blue coloration through-

out the colony.

When the number of colonies exceeded �50–70/plate,

an automated colony-counting platform was used

(Q-Count, Spiral Biotech, Norwood, MA).Throughout

the manuscript, when quantities of spores are men-

tioned, this refers to the cfu yielded by plating of the

spore quantity in question. However, for spiking

experiments conducted at the City of Phoenix Water

Services Laboratory, we used spores enumerated via

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–5211 5201

fluorescence activated cell sorter (FACS) by the

Wisconsin State Laboratory of Hygiene facility in

Madison, WI, aliquoted as 100 spores in 100 ml volumes

of water.

2.2. Spore filtration protocol

Tap water was obtained from the faucets in building

173 of the Beltsville Agricultural Research Center,

Beltsville, MD, and from the distribution system at the

City of Phoenix Water Services Laboratory in Phoenix,

AZ. Source water came from the Wye River and the

Middle Patuxent River, Maryland, and from the 24th

Street Water Treatment facility in Phoenix, AZ. Source

water turbidity was measured using at least three, 20ml

aliquots on an Orbeco-Hellige turbidometer (Orbeco

Analytical Systems, Inc., Farmingdale, NY). Filters

(nitrocellulose, gridded, 47mm diameter, 0.45mm, cat-

alog no. E04WG047S4) were purchased from Osmonics,

Westborough, MA. A Millipore ( catalog no. 1504700;

Bedford, MA) ground glass filtration apparatus capable

of holding a 47mm diameter filter, with a 250ml volume

funnel and 1L volume collection flask, was used for

filtrations. The Millipore apparatus was autoclaved

for 20min at 121 1C between uses. Experiments con-

ducted in Phoenix used a Millipore three-position

vacuum manifold with membrane supports and sterile

funnels (Millipore catalog nos. MIA-CO3P-01 and

MZHAWG1-01) and Advantec 0.45mm filters (Hardy

Diagnostics, Santa Maria, CA).

The filtration and plate culture protocol for 100ml

and 1L volumes of water (as well as for 10L volumes) is

illustrated in Fig. 1. Briefly, the 47mm 0.45 mm filter was

pre-wetted with sterile water, overlaid on the Millipore

apparatus with the grid side facing up (i.e., receiving the

water), and the water sample passed through with the

aid of a vacuum line or vacuum pump (Fig. 1, step 3). In

general, 100ml of tap or source water passed through

the filter ino10 s. One liter volumes of tap water

required several minutes to passage. After filtering, the

filter was removed from the Millipore apparatus and

placed on a test tube rack or sterile Petri dish and

Fig. 1. Diagram of protocol used to recover spores from spiked

water samples. Ten liters of tap water were passed through a

0.2mm filter capsule using a peristaltic pump at a rate of 0.7 L

per min. (step 1). Spores were eluted from the capsule via two

consecutive elution runs involving the addition of �90ml

elution buffer and agitation in a benchtop shaker at �1200 rpm

for 5min (step 2). The aliquots of elution buffer were combined

and subjected to filtration using a Millipore apparatus and a

0.45ml nitrocellulose filter (step 3). The filter was heated at

65 1C for 15min and placed facedown on a sheep red blood cell

agar plate (step 4) and incubated overnight at 37 1C. Note that

for recovery of spores from 100ml and 1L volumes of water,

the protocol commenced with step 3 followed by step 4.

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–52115202

incubated at 60–65 1C for 20min in a small oven

(Quincy Laboratory Oven, Model 20 AF, Fisher

Scientific, Pittsburgh, PA) or incubator. The filter was

removed from the oven and placed grid-side down on a

SRBC plate and incubated overnight at 37 1C (Fig. 1

step 4).

For direct detection of DNA extracted from spores

recovered via filtration, the filter was transferred to a

15ml polypropylene conical centrifuge tube containing

2ml of phosphate buffered saline (pH 7.4) with 1%

Tween-20 (polyoxyethylenesorbitan monolaurate, Sig-

ma, St. Louis, MO). The tube was vortexed for three,

10 s periods using a Vortex GenieTM 2 apparatus (Fisher

Scientific, Pittsburgh, PA ) and the PBS-T20 decanted

into a 2.0ml microfuge tube. The tube was centrifuged

at 16,168� g (equivalent to 13,000 rpm on a microcen-

trifuge) for 10min to pellet spores. All but �100ml ofsupernatant was removed , the spores resuspended by

vortexing, and the contents transferred to a 1,2,3

RAPID DNA extraction kit bead tube, containing

250ml of a liquid suspension of 0.1mm-diameter

zirconia–silica beads (Idaho Technology, Inc., Salt Lake

City, UT). The bead tube was agitated on a Vortex

Genie 2 apparatus, using an adaptor for 2.0ml tubes, for

5minutes. The bead tubes were then centrifuged at

16,168g for 5min to pellet the beads and the supernatant

transferred to a sterile 1.5ml microfuge tube, where it

was subjected to DNA extraction using phenol–chlor-

oform–isoamyl alcohol, isopropanol precipitation, and

70% ethanol wash. The resultant DNA pellet suspended

in 50 ml of sterile water. This method was used for the

initial three replicates in which 534 spores were spiked

into 100ml water (Results section, below), but all

subsequent replicates used the QiagenTM DNeasys

Tissue kit protocol, which does not require phenol and

is thus more acceptable from the point of view of

laboratory safety and hazardous waste disposal. Extrac-

tion of lysed spore DNA followed the manufacturer’s

protocols, with the exception that only 50 ml of DNA

elution buffer AE, rather than the recommended

100–200 ml, was used. Whether the phenol–chlorofor-

m–isoamyl alcohol or DNeasy Tissue kit methods were

used, 10–15 ml out of the total 50ml volume of DNA

were used as template for PCR.

For extraction of DNA from colonies on SRBC

plates, a sample of cells was taken with a sterile,

disposable plastic loop and transferred to a 1.5ml

microfuge tube containing 200ml of InstageneTM matrix,

a 6% solution of Chelexs 100 resin (Bio-Rad, Hercules,

CA). The loop was gently twirled to dislodge the cells,

and the microfuge tube then subjected to heating steps

of 56 1C for 15min and 100 1C for 8min, after which it

was centrifuged at 16,168g for 5min to pellet the matrix.

A 5ml aliquot of supernatant was used as template for

PCR. Note that it was often impractical to sample every

suspect colony on a given plate, particularly when

recoveries were high; the investigator usually attempted

to get a loopfull of cells from one or two colonies or

several portions of the bacterial lawn, with a morphol-

ogy highly suggestive of that of B. anthracis, as well as

colonies of nontarget organisms that were also present

on the plate, to serve as DNA extraction controls.

2.3. Recovery of spiked spores from 10 L volumes

Carboys containing 10L volumes of tap water

(obtained from faucets in the USDA-ARS laboratory

in building 173, Beltsville, MD, or houses located in

Howard and Charles Counties, MD, and from faucets

from the distribution system at the City of Phoenix

Water Services Laboratory in Phoenix, AZ) were spiked

with 1 ml aliquots containing discrete quantities of

spores of B. anthracis Sterne strain (Table 4). The

carboys were capped and agitated to disperse the spores

throughout their contents. A Masterflexs 77601-00

model peristaltic pump (Cole Parmer Instrument Co.,

Vernon Hills, IL) was used to pass the spiked water

through a 0.2 mm capsule filter (part no. 12117, Pall

Gelman Laboratories, Anne Arbor, MI) at a rate of

0.7 L per min (attempts to use a faster flow rate resulted

in backup of the flow) (Fig. 1 step 1). After filtration the

capsule received ‘‘USEPA Method 1623’’ elution buffer

(1 L contains 1 g Laureth-12 [PallGelman Laboratories],

10ml 1M Tris, pH 7.4, 150 ml Antifoam A reagent

[Sigma Chemical Co., St Louis, MO], and 2ml 0.5M

EDTA); an initial quantity of�125ml was sufficient to

completely cover the filter inside the plastic capsule

housing. The capsule was sealed and shaken on a

MistralTM multi-mix instrument (Lab-Line Instruments,

Inc., Melrose Park, IL) for 5min at a setting of

�1000 rpm (Fig. 1 step 2). The buffer was decanted

and saved, and another �90–100ml of elution buffer

added; the capsule was placed in the Mistral shaker, but

this time the capsule was rotated 451, and again shaken

for 5min. The capsule was rotated another 451 in the

Mistral shaker, and shaken for 5min. The buffer was

combined with the previous aliquot and this�200–

225ml volume was filtered through a 0.45mm filter

using the Millipore apparatus described in the section

above (Fig. 1 step 3). The filter was heated at 65 1C for

15mins and placed facedown on an SRBC agar plate

(Fig. 1 step 4) and incubated overnight at 37 1C; the filter

was removed and the plate allowed to incubate for

another 2–3 h (this allowed B. anthracis Sterne colonies

to grow larger and provide more biomass for DNA

extraction). After enumeration colonies were subjected

to DNA extraction.

If a capsule first was used to filter an unspiked control,

and no B. anthracis colonies appeared on the plate, the

capsule was used a second time on spiked samples (the

per-capsule price of�$98 US dictated this economizing

measure).

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–5211 5203

To reduce carryover of spores from one experiment to

the next, several measures were employed. After use in a

spiking experiment carboys were autoclaved. The

efficacy of the autoclaving process was confirmed by

placing within one of the carboys a 0.5ml microfuge

tube containing 50ml (�1� 106) spores of B. anthracis

Sterne. After autoclaving the spore solution was spread

onto an SRBC plate and incubated at 37 1C for 36 h. As

another measure to control for possible carryover of

spores, unspiked 10L tap water samples were periodi-

cally interspersed with spiked samples, and subjected to

the same procedure used to examine filtrate for the

presence of B. anthracis spores. Any colonies visible on

the plate following overnight incubation were subjected

to DNA extraction and PCR amplification as for

B. anthracis colonies.

2.4. Polymerase chain reaction

Several published primer sets were evaluated. The

GPR 1 and GPR 2 (outer) and GPR 4 and GPR 5

(inner) primer pairs for the variable repeat region A

(vrrA) gene, the LEF-2F and LEF-2R primers for the

lethal factor (lef) gene, and the CAPA-F and CAPA-R

primers for the capsule (capA) gene, are from Jackson et

al. (1997, 1998). The primer set targeting the vrrA gene

was chosen for evaluation because, while historically

used for genotyping purposes, it is a nested PCR and

thus may provide greater sensitivity; the outer GPR 1

and GPR 2 primers amplify a �1.1 kB fragment of the

gene, and the inner GPR 4 and GPR 5 primer pair

amplify a 377–245 bp fragment. Cycling conditions for

the GPR 1 and GPR 2 set were: 2min at 95 1C, then 40

cycles of 15 s at 95 1C, 1min 30 s at 55 1C, and 2min at

72 1C. Cycling conditions for the inner GPR 4 and GPR

5 , LEF-2F and LEF-2R, and CAPA-F and CAPA-R

primer sets were: 2:00min at 95 1C, then 40 cycles of 15 s

at 95 1C, 30 s at 55 1C, and 45 s at 72 1C. PCR

amplification reactions were examined via agarose gel

electrophoresis and ethidium bromide staining.

Real time PCR analysis of samples was conducted

using the lef TaqMan probe and primer protocol of

Dang et al. (2001) on the Stratagene Mx4000 instrument

(La Jolla, CA) or Applied Biosystems 7900HT Sequence

Detection System (Foster City, CA). (To distinguish

between the two lef PCR assays, note that throughout

the manuscript the Dang et al. assay is referred to

as the ‘‘real time’’ lef PCR, while the LEF-2F/LEF-2R

assay of Jackson et al. is simply referred to as the ‘‘lef’’

PCR).

Sequencing of select PCR amplicons was done using

two forward and two reverse sequencing reactions using

Big Dyes v3.0 dye-terminator master mix on an ABI

3100 automated fluorescence sequencing instrument

(Applied Biosystems, Foster City, CA); sequence assem-

bly was done using LaserGeneTM software (DNAS-

TAR, Inc., Madison, WI). Identity of sequences was

done using the NCBI nBLAST program.

For spiking experiments involving 10L tap water

samples (all but one of which were conducted at

Beltsville), we assayed suspect colonies (as well as non-

anthracis-appearing controls) with three different PCR

assays: the GPR 1 and 2 primers, targeting the vrrA

gene; the real-time lef primers; and the conventional lef

primers. Our rationale was that, as the volume of spiked

water increased, the likelihood of nontarget organisms

appearing on plates and overgrowing and/or intermin-

gling with legitimate Sterne strain-derived colonies

would be greatly increased, particularly if more pro-

longed incubation times were needed to allow for

maximal growth of recovered B. anthracis spores. In

such circumstances, the ability to selectively pick out

independent colonies of B. anthracis may be difficult. As

a consequence, primers and probes demonstrated to

show a satisfactory degree of specificity with prepara-

tions of DNA from discrete isolates of anthracis , would

be presented with heterogeneous mixtures of Bacillus

(and other bacterial flora) DNAs. We were thus

interested in the comparative performance of these

primer sets with this (arguably more challenging)

DNA template. All PCR assays conducted at the City

of Phoenix Water Services Laboratory used the real-

time lef PCR assay, exclusively.

Sequence analyses depicted in Fig. 2C were done

using vrrA Genbank depositions for B. anthracis Sterne

(L48553), B. mycoides (L48556) and B. cereus (L48555)

(Anderson et al., 1996).

3. Results

3.1. Culture-based detection protocol: 100 ml and 1 L

In general, with a larger spiking dose the B. anthracis

Sterne colonies appearing on plates after an overnight

incubation were readily distinguishable from those of

other, nonanthracis species, some of which generated a

‘‘halo’’ of lysed red cells. However, with smaller spiking

doses it was often necessary to allow extra incubation

time for colonies of B. anthracis to grow to sufficient size

and clarity to be accurately identified; not surprisingly

this was equally advantageous to nontarget organisms

present in the sample. Consequently, colonies of B.

anthracis were sometimes intermingled with those of

other species, or alternatively, colonies derived from

spores in the spiking dose generated what appeared to be

accompanying satellite colonies, creating the potential

for inflated counts. Both of these factors made colony

counts (particularly in the spiked 1L and 10L samples)

somewhat objective.

Selection of PCR primers also influenced interpreta-

tion of recovery data. For example, while the vrrA outer

ARTICLE IN PRESS

Fig. 2. Comparative performance of PCR assays for confirmation of B. anthracis recovered from 10L tap water spiked with 10 spores

using capsule filtration. Panel A: plot of results of lef real time PCR (Dang et al., 2001) conducted on DNAs, extracted from suspected

B. anthracis and other Bacillus spp. colonies, associated with spiked (sp) and unspiked (unsp) replicates (labeled as A, B, C, etc.); NTC

is the no template control and POS the B. anthracis Sterne strain positive control. Cycle number is plotted on the x-axis and normalized

fluorescence on the y-axis. Samples whose amplification plots are above the threshold (horizontal line at�250 dR on the y-axis) at cycle

40 are deemed positive, thus the spA and spB samples (and positive control) are the only positives. Panel B: agarose gel image of results

of Bacillus spp. vrrA gene PCR (Jackson et al., 1997) conducted on the same panel of DNAs. Lane L: DNA ladder with rung sizes

indicated. Lane ntc: no template control. Lane Pos: B. anthracis Sterne strain positive control. (note that spiked sample C has an

amplicon of size similar to that of the positive control). Panel C: clustal W nucleotide sequence alignment -based phylogenetic tree

depicting relationship of vrrA gene sequence (1006 nt) from spiked sample C, to relevant vrrA sequences from B. anthracis, B. cereus,

and B. mycoides.

A. Perez et al. / Water Research 39 (2005) 5199–52115204

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–5211 5205

and inner sets frequently generated amplicons of a size

noticeably different from those of B. anthracis from

unspiked controls and hemolytic colonies, we observed

one instance, using the outer ‘‘GPR 1/GPR 2’’ set, in

which a nonlytic colony with an appearance suggestive

of B. anthracis generated an amplicon of visibly

equivalent size to B. anthracis (Fig. 2B). However,

subsequent lef real time PCR on this colony was

negative (Fig. 2A). We obtained readable sequence for

1006 nucleotides from the vrrA PCR product and

observed 98% similarity to B. cereus and 89% similarity

to B. anthracis Sterne and B. mycoides (Fig. 2C). We

therefore relied on the conventional and real-time lef

PCR assays to provide a more accurate confirmation of

B. anthracis colonies on our plates.

Table 1 provides results of a series of assays in which

two different agar formulations were tested, using B.

anthracis Sterne spores (x ¼ 47) enumerated via plating,

spiked into 100ml volumes of tap water and recovered

using filtration and overnight incubation. There was no

statistically significant difference between agar formulation

in terms of mean recovery of spores (t ¼ 2:579, 4 df;

P ¼ 0:0614). We noted that it was slightly easier to identify

discrete colonies on the SRBC plates, since those on the R

& F plates showed a greater degree of coalescence.

After436h incubation the colonies on the R & F plates

assumed the color profile indicative of B. anthracis. While

one of the SRBC unspiked control replicates displayed

contaminant colonies, none of the R & F plates did so.

DNA extracted from these SRBC plate contaminants was

negative when assayed via lef PCR, while DNA from

putative B. anthracis colonies from the SRBC and R & F

plates was lef PCR positive (data not shown).

Table 1

Recoveries of B. anthracis Sterne strain spores from spiked

100ml volumes of tap water: comparison of sheep red blood cell

(SRBC) agar plates and R&F agar plates

Spore inoculum SRBCa (%) R&F (%)

47 21 44 22 46

34 72 29 61

38 80 36 76

45 95 41 87

32 68 26 55

Unspiked

control

0 0

0 0

0 0

6 NDb

0 ND

SRBC x ¼ 3478.8, R & F x ¼ 30.877.6, t ¼ 2.579,

P ¼ 0.0614 ns.aAll colonies appearing on the plate were counted.bND ¼ Not done.

When spores were used to spike 100ml volumes

source water taken from the Patuxent River and

Chesapeake Bay, the observable cfu on the SRBC plates

exceeded those calculated to be in the spiking dose

proper, reflecting the adventitious growth of other

waterborne flora. The co-presence of these bacteria

made accurate enumeration of recovered B. anthracis

spores difficult. However, in all three replicates with a

spike dose of 267 spores, and three replicates with a

spike dose of 35 spores, the majority of colonies

observed on the plates were morphological similar to

those of B. anthracis (data not shown). When DNA was

extracted from putative B. anthracis colonies, as well as

non-anthracis colonies, positive LEF-2F/LEF-2R PCR

signals were observed only for the B. anthracis colonies

(data not shown). DNA sequencing of two amplicons

derived from these experiments indicated 100% homol-

ogy (over 387 bp) with Genbank depositions of

B. anthracis (data not shown).

Results of a series of assays conducted at the Phoenix

Water Services Laboratory in which 100 FACS-enum-

erated B. anthracis Sterne spores were spiked into 100ml

volumes of finished water samples, and recovered using

filtration and overnight incubation on SRBC plates, are

shown in Table 2. The mean recovery was 48 spores

(78.4, range 31–61, n ¼ 18 replicates); nonanthracis

colonies were not observed on the plates. DNA

extracted from 20 of the B. anthracis colonies

present on several of the plates derived from spore-

spiked water, were all positive by real time lef PCR (data

not shown).

Table 2

Recoveries of FACS-enumerated B. anthracis Sterne spores

(n ¼ 100) spiked into 100ml volumes of finished water

Replicate No. spores recovered

1A 44

2A 49

3A 61

4A 54

5A 53

6A 57

7A 40

8A 43

9A 46

10A 61

1B 31

2B 46

3B 53

4B 35

5B 46

6B 51

7B 48

8B 59

Mean 4878.X2¼ 485, df ¼ 17.

ARTICLE IN PRESS

Table 4

Recoveries of B. anthracis Sterne spores spiked into 10 L

volumes of tap water

Spike dose Water volume No. PCRb

A. Perez et al. / Water Research 39 (2005) 5199–52115206

When we attempted to spike 10 replicates of 100ml

source water obtained from the Phoenix area with 100

FACS-enumerated spores, intense growth of nontarget

flora was noted on the plates; colony DNA yielded a

positive real time lef PCR result for only five of the 10

replicates (data not shown).

Experiments done using spores enumerated by plat-

ing, spiked into larger volumes of source and tap waters,

were conducted at the USDA laboratory. When 1L of

tap water was spiked with 35 spores, the recovery

efficiency was 43%, and for a spiking dose of just 10

spores, 72% (Table 3). However, attempts to recover

spores spiked into 1L volumes of source water were

sometimes hampered by the intense overgrowth of the

plates by other spore-forming flora (Table 3). Also, the

quantity of suspended solids in this greater volume

resulted in clogging of the filter to such an extent that it

took as long as 30–45min to filter the entire liter (mean

turbidity of three replicates 3.870.2 NTU). Attempts to

perform lef PCR on suspect colonies from a spiked

source water sample resulted in just one, weakly positive

PCR out of three replicates (Table 3). We did not

observe B. anthracis colonies in the unspiked controls

(Table 3).

3.2. Culture-based detection protocol: 10 L

Six replicates were done with a spiking dose of �100

spores in 10L of tap water. All six replicates were

Table 3

Recoveries of spores spiked into 1L volumes of water

Spike dose

(spores)

Water

volume

Type No.

recoveredaPCRb

35 1L Tap 17 +

19 +

10 +

35 1L Source TNTCc —

TNTC —

TNTC —

10 1L Tap 11 +

7 +

9 NDd

9 ND

0 —

Unspiked 1L Tap 3 ND

0 ND

0 ND

21 —

4 ND

aRefers to all putative B. anthracis colonies observed.bVia lef gene PCR assay; refer to text for details.cToo numerous to count.dND ¼ not done (appearance of colony was definitive

enough to be assigned as B. anthracis).

successful in generating colonies of B. anthracis Sterne

after overnight incubation; while the co-presence of

colonies derived from nontarget organisms and the

coalescence of Sterne colonies made counting colonies

more difficult than with the 100ml or 1L volumes, a

mean of 11 colonies (76.8, range 2–20) was observed

(Table 4).

Two unspiked control runs resulted in the generation

of non- B. anthracis colonies; DNA from these failed to

amplify with the lef PCR primers (Table 4).

The spiking dose was subsequently reduced to 70

spores, and both replicates were successful in generating

B. anthracis colonies (mean of 8), which were in turn

positive by lef PCR (Table 4). A decision was made to

reduce the spiking dose further, to just 10 spores; we

observed that this small a spike dose led to problems

identifying colonies on most of the plates, because the

ratio of other Bacillus colonies to those of B. anthracis

was increased in favor of the former. Extended

incubation times (i.e.,418 h), although equally advanta-

geous to the growth of other Bacillus spp., were often

needed to allow for emergence of suspected B. anthracis

colonies from the background flora. Also, we could not

(spores) Recovereda

100 10L 11 +

5 +

17 +

20 +

2 +

13 NDc

70 10L 9 +

7 ND

10 10L 2 +

9 +

0 —

0 —

0 —

3 +

Unspiked 10L 0 —

0 —

3 +

TNTCd —

TNTC —

TNTC —

TNTC —

21 —

aRefers to all putative B. anthracis colonies observed.bVia lef gene PCR assay; refer to text for details.cND ¼ not done (appearance of colony was definitive

enough to be assigned as B. anthracis).dToo numerous to count.

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–5211 5207

be certain that the sampled colonies were entirely free of

bacteria from nontarget organisms. More so than in

experiments using a higher spiking dose and smaller

volumes of water, PCR was required to confirm the

identity of these suspect colonies.

When 10 spores were spiked into 10L volumes of tap

water, only three of six replicates were positive, with a

mean recovery of 2.3 spores (73.5, range 0–9) (Table 4).

After the initial three replicates of the 10-spore spiking

assays, an unspiked control was performed; we observed

several B. anthracis colonies on this plate; these tested

positive by vrrA and lef PCRs, but negative for capA

PCR , indicating they were ‘‘carryover’’ contaminants of

our Sterne strain spiking spores, and not B. anthracis

naturally present in the water (Table 4). Spiking

experiments were suspended and four more unspiked

control runs performed; all had florid growth of

nontarget organisms (the tap water was collected during

periods of heavy rainfall in the Beltsville area in late

March/early April 2004), which made identification and

sampling of B. anthracis-appearing colonies difficult.

vrrA and lef PCRs were negative on these four unspiked

replicates; accordingly we resumed our 10 spore spiking

experiments and completed three more replicates, one of

which was positive (Table 4). A final unspiked control

was conducted after these spiked replicates; while over

20 colonies were observed on the plate, none had an

appearance suggestive of B. anthracis, and none tested

positive by lef PCR (Table 4).

Fig. 3. Results of PCR assays performed on B. anthracis Sterne strain

from spiked water samples. Panel A: lef gene PCR results on filter w

spores (replicates Sp1, Sp2, and Sp3) and an unspiked control. Panel B

1L tap water samples spiked with 534 spores (replicates Sp1, Sp2, and

sizes indicated. Lane ntc: no template control. Lane Sterne: B. anthra

3.3. Direct detection of recovered spores

In contrast to a culture-based approach to detection

of spores, so-called ‘‘direct detection’’ relies on removal

of spores from the filter and their subsequent concen-

tration, lysis, and DNA extraction. Initially, we used a

protocol in which the filter was incubated for 1 h in 2ml

PBS-T20 on a rocking platform, followed by extraction

of spores liberated from the filter into the PBS-T20

solution. When large doses of spores (26,000+ and

2600+) were spiked into 100ml tap or source water, all

spiked replicates (n ¼ 9 for tap and n ¼ 9 for source)

were positive by lef PCR (data not shown). Use of

smaller spiking doses was unsuccessful, and plating

experiments indicated that insufficient quantities of

spores were being dislodged from the filter. The protocol

was subsequently altered, with vortexing substituted for

incubation on a rocking platform. When the spiking

dose was reduced to 534 spores in 100ml, 9/9 replicates

of spiked tap water, 6/6 source water replicates, and 0/3

unspiked controls were positive by lef PCR; an example

of the lef PCR results obtained from one such replicate

are presented in Fig. 3A. When 534 spores were spiked

into 1L tap water, the lef PCR was unsuccessful;

however, using the nested vrrA PCR resulted in 4/9

spiked samples, and 0/3 unspiked controls, positive

(results from one of these replicates is depicted in

Fig. 3B). When the spiking dose was reduced to 100

spores in 100ml tap water, even a nested PCR for the

DNA extracted from filter washings containing spores recovered

ashings derived from 100ml tap water samples spiked with 534

: nested vrrA gene PCR results on filter washings derived from

Sp3) and an unspiked control. Lane L: DNA ladder with rung

cis Sterne strain positive control.

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–52115208

vrrA gene failed to generate a detectable signal for three

replicates (data not shown).

4. Discussion

We have described a two-pronged approach to the

detection of B. anthracis spores in tap and source water

samples ranging from 100ml to 10L in volume, via the

use of either overnight culture, or a ‘‘direct detection’’

method based on PCR. The culture method should be

suitable for routine use on relatively large numbers of

samples (i.e., 10 or more a day) by even modestly

equipped water testing laboratories, while the direct

detection method could be employed in those situations

in which a more timely (but admittedly less sensitive)

detection method is needed. The protocol for recovering

spores from 100ml and 1L volumes of water is designed

to be relatively inexpensive (i.e., less than $3.00/sample)

and can be readily adapted from existing protocols, such

as the Standard Method 9222D (fecal coliform mem-

brane filter procedure) and USEPA Method 1605

(detection of Aeromonas spp. in finished water).

Attempting direct detection of spores washed from the

filter was also designed to be as inexpensive and user-

friendly as possible. The protocol for recovering spores

from 10L volumes of tap water is obviously more labor-

intensive and expensive: the capsules cost �$98.00 and

their use requires a pump and tabletop shaking plat-

form. However, since larger water quality testing

laboratories already have these instruments on hand in

order to perform the USEPA Method 1623 protocol for

detecting parasite cysts in water samples, it may not be

particularly burdensome for such labs to adapt these

instruments for use in spore detection.

Our approach was successful in culturing B. anthracis

Sterne from 10L volumes of tap water spiked with as

few as 100, 70, or even 10 spores. One drawback to the

protocol was the presence of B. anthracis in an unspiked

control associated with the 10 spore spiking cohort,

which indicated that ‘‘leakage’’ of spores from previous

replicates was occurring. While followup control assays

appear to indicate that carryover was eliminated by

measures such as repeated autoclaving of tubing and

carboys, with such a small spiking dose we cannot

definitively rule out confounding of our results due to

this phenomenon; thus our 10 spore spike recoveries

should be interpreted with some degree of caution. We

are somewhat heartened by the fact that none of our

recoveries were in excess of the spiking dose, which

would seem to indicate that if and when it occurred, the

extent of spore carryover was limited in scope.

We did not attempt to detect spores in 10L volumes

of source water using capsule filtration; such an

approach would confront formidable obstacles, parti-

cularly the co-presence of other spore-forming flora,

which, on some occasions, were so numerous in our tap

water samples that they covered the SRBC plate in a

lawn of colonies. Interestingly, the presence of signifi-

cant numbers of Bacillus spp. spores in Norwegian water

has been reported by Ostensvik et al. (2004), who used

membrane filtration to recover the spores from water

samples. These investigators observed 38 and 15 cfu/

100ml in two drinking water samples from urban and

rural sources, respectively, and as many as 1400 cfu/

100ml in source water samples. In a study conducted in

the United Kingdom, Francis et al. (2001) examined

aerobic spore-forming bacilli recovered from Severn

River water samples using membrane filtration, and

reported 11–109 cfu per 100 ml river water cultured on

nutrient agar plates (when diluted in 20ml of distilled

water). When potable water was sampled, cfu ranged

from 11 to 30/100ml.

Could the use of a more selective media enhance

production of B. anthracis colonies and reduce that of

competing organisms? Previous experiences in the

course of screening46000 environmental samples, taken

in association with the anthrax bioterror attack in the

Washington, DC area in 2001–2002, indicated that

PLET agar could support profuse growth of nonan-

thracis organisms (Higgins et al., 2003). We evaluated a

new formulation of chromogenic agar from R & F

Laboratories designed to allow for discrimination

between B. anthracis and other species of Bacillus. This

formulation was comparable to SRBC plates in cultur-

ing spores recovered from spiked tap water samples;

however, further testing will be needed to see how well it

performs with regard to spiked source or raw water

samples. One strategy that may improve sensitivity

would be to selectively remove B. anthracis spores from

those of related Bacillus species, perhaps via use of

immunomagnetic separation, in a manner similar to that

available for selective recovery of Cryptosporidium

oocysts and Giardia cysts from water concentrates and

feces. However, such a protocol would require the

antibody reagent to reproducibly capture small quan-

tities of B. anthracis spores against a high background of

nontarget spores, and we are not aware of any such

reagent currently available for commercial use. Another

obstacle presented by source water is its potentially high

turbidity and/or suspended solid content (particularly

after rainfall events), which can lead to clogging of the

capsule filter and severely reduce its filtering capacity

(DiGiorgio et al., 2002). Pre-filtration or low-speed

centrifugation may help to ameliorate this phenomenon

but will obviously increase the time and labor involved

with the recovery process. Finally, regardless of the

water source, the presence of other spore-forming

bacteria in the sample can obviously influence the

success of growth of B. anthracis on the SRBC plates.

Production of bacteriocins by Bacillus spp. has been well

documented (Zheng et al., 1999) and may hinder or

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–5211 5209

prevent establishment of colonies of target organisms,

thus leading to potential false-negative results. For this

reason, extensive re-use of the capsule filters may not be

feasible, as they may accumulate an ever-increasing

quantity of competing microorganisms with successive

runs.

Obviously, in the event of a genuine waterborne

outbreak of GI anthrax, it is difficult to know if our

detection limits have relevance to what would constitute

a dose of spores sufficient to cause morbidity and

mortality. Using overnight culture, our protocols are

capable of detecting as few as 10 spores in 10L of tap

water; with PCR-based direct detection, the sensitivity of

the assay is reduced to �500 spores in 1L tap water. We

are assuming that vegetative cells would not be employed

in a scenario involving deliberate contamination of

potable water supplies, since chlorination would inacti-

vate them. However, there is much we do not know

about GI anthrax; in recent years there has been only one

small outbreak of this form of the disease in the US

(Anonymous, 2000) and investigations reported from

outbreaks overseas posit that ingestion of vegetative cells

in undercooked meat is responsible for establishing

infection (Sirisanthana and Brown, 2002; Friedlander,

1997). Experiments underway in several laboratories,

using animal models, may provide useful information on

the ability of ingested spores to cause infection.

The ‘‘direct detection’’ protocol we devised lacked the

sensitivity of the culture-based methodology; this was

not surprising in light of the fact that the ‘‘biological

amplification’’ inherent in culture usually can provide

the investigator with sufficient material for molecular

analyses. Nonetheless we feel that a detection limit of

534 spores in 1L will provide the end-user with some

degree of versatility when confronted with the need to

rapidly detect spores in given water samples. Obviously,

there is room for improvement in this area, but the

successful methodology will have to be able to

concentrate the spores from the water sample—be it

100ml, 1 L, or even 10L—into a volume small enough

to be manipulable using the liquid volumes common to

most molecular biology protocols, i.e., several ml (or

less). DNA resulting from these extraction procedures

will need to be of sufficient quality and quantity to

participate in 20–100 ml PCR reactions. This is a

significant challenge with existing protocols, and we

note that a number of promising technologies for

concentrating/capturing threat agents from water or

other liquids are in development, originating from firms

such as CreatVMicrotek, IQuum, Nanosphere, Micro-

Fluidic Systems, Inc., Idaho Technology, Inc., and

Nanolytics, among others.

The choice of PCR primers and target gene for

confirmation of the identity of suspect colonies is

complicated by the fact that other species of Bacillus

can be present on the agar plate, possibly commingled

with B. anthracis cells, and thus may contribute nucleic

acids to the tested samples. Our results indicate that, at

least with B. anthracis Sterne strain, primers targeting

the lef gene can provide a satisfactory degree of

specificity. An extended discussion of the best choices

for assays targeting virulent strains of B. anthracis (e.g.,

Ames) is beyond the scope of this report, particularly in

light of the increased number of published PCR

detection assays over the past several years; however,

some of the papers of relevance to this issue are those of

Bell et al. (2002), Blackwood et al. (2004), Hoffmaster et

al. (2002) and Luna et al. (2003). We note that while

many published probe and primer sets show satisfactory

true-negative reactions against nontarget species, speci-

ficity assays often are done using DNA templates

extracted from purified, homogeneous isolates of Bacil-

lus spp. Cultures from water concentrates may present

the investigator with a spectrum of Bacillus species and/

or genotypes that, in the course of the incubation step,

may or may not have exchanged genetic material. The

recent finding that B. cereus can host a plasmid

(pBCXO1) with 99.6% homology to that of B. anthracis

(pXO1), and cause pneumonia in an otherwise healthy

patient, would seem to indicate that this is not an

unrealistic possibility, with interesting implications for

diagnostic protocols (Hoffmaster, et al., 2004).

Once the presence of anthracis spores in a water

supply has been documented, the troubling issue of their

removal and/or inactivation will naturally arise (Rose

et al., 2005). Possible remediation methods include the

use of ozone, filtration, or UV light, all of which would

be very expensive to implement on a large scale (Morato

et al., 2003). However, we note that recent work suggests

that Bacillus spores, far from being inert structures, can

increase in size due to hydration in response to increased

relative humidity (Westphal et al., 2003). This suggests

that organic substances present in the water may gain

access to the interior of the spore, and perhaps affect the

viability of the dormant bacterium. This may allow for

the administration of select inactivating agents to the

contaminated water.

5. Summary

1.

One hundred FACS-enumerated spores of B. anthra-

cis were spiked into 100ml volumes of finished water

and recovered using simple membrane filtration,

followed by overnight culture on red blood cell agar

plates and visual inspection and counting of colonies.

The mean number of recovered spores was 48 spores.

2.

With large spiking doses of4250 spores, recoveries

from 100ml of source water were successful; how-

ever, reduction of spiking dose to 100 spores or less

gave comparatively poor results due to the over-

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–52115210

growth of the plates with other species of Bacillus

native to the water.

3.

When 35 or 10 spores were spiked into 1L volumes

of tap water, mean recoveries were 15.3 and 7.2

spores, respectively. Attempts to recover spores

spiked into 1L volumes of source water were

unsuccessful.

4.

A combination of capsule (0.2 mm) and membrane

filtrations were used to recover spores from spiked 10

L volumes of tap water. For spikes of 100, 70, and 10

spores, mean recoveries were 11.3, 8, and 2.3 spores,

respectively.

5.

When ‘‘direct detection’’ of recovered spores was

made via DNA extraction and PCR in lieu of

overnight culture, a spiking dose of 534 spores in

100ml of tap water generated positive results.

Detection of 534 spores spiked into 1L tap water

could be achieved using nested PCR. Assays were

negative when the spiking dose was reduced to 100

spores in 100ml tap water.

Acknowledgements

The authors would like to thank Jeff Karns for

preparing the spore formulations, Yakov Pachepsky and

Boris Devin for translation of the paper by Tomov,

Darcy Hanes for her review of the manuscript, and

Martin Klass, Michelle Klass, Dorrine Felix, and the

Molecular Diagnostics Laboratories of the Southern

Arizona Veterans Affairs Health Care System, for

providing technical assistance. Drs G. Smith and L.

Restaino, R&F Laboratories, kindly supplied chromo-

genic agar plates for evaluation of spore culture methods.

References

Andersen, G.L., Simchock, J.M., Wilson, K.H., 1996. Identi-

fication of a region of genetic variability among Bacillus

anthracis strains and related species. J. Bacteriol. 178,

377–384.

Anonymous. Human ingestion of Bacillus anthracis—contami-

nated meat—Minnesota, August 2000. 2000. MMWR 49:

813-816.

Bell, C.A., Uhl, J.R., Hadfield, T.L., David, J.C., Meyer, R.F.,

Smith, T.F., Cockerill, IIIR.F., 2002. Detection of Bacillus

anthracis DNA by lightcycler PCR. J. Clin. Microbiol. 40,

2897–2902.

Blackwood, K.S., Turenne, C.Y., Harmsen, D., Kabani, A.M.,

2004. Reassessment of sequence-based targets for

identification of Bacillus species. J. Clin. Microbiol. 42,

1626–1630.

Burrows, W.D., Renne, S.E., 1999. Biological warfare agents as

threats to potable water. Environ. Health Perspect. 107,

975–984.

Dang, J.L., Heroux, K., Kearney, J., Arasteh, A., Gostomski,

M., Emanuel, P.A., 2001. Bacillus spore inactivation

methods affect detection assays. Appl. Environ. Microbiol.

67, 3665–3670.

DiGiorgio, C.L., Gonzalez, D.A., Huitt, C.C., 2002. Cryptos-

poridium and Giardia recoveries in natural waters by using

environmental protection agency method 1623. Appl.

Environ. Microbiol. 68, 5952–5955.

Francis, C.A., Lockley, A.C., Sartory, D.P., Watkins, J., 2001.

A simple modified membrane filtration medium for the

enumeration of aerobic spore-bearing bacilli in water.

Water Res. 35, 3758–3761.

Friedlander, A., 1997. Anthrax. In: Sidell, F.R., Takafuji, E.T.,

Franz, D.R. (Eds.), Medical Aspects of Chemical and

Biological Warfare. Borden Institute, Washington, DC,

pp. 467–478.

Higgins, J.A., Cooper, M., Tucker, L., Black, S., Miller, D.,

Karns, J.S., Manthey, E., Breeze, R., Perdue, M.L., 2003. A

field investigation of Bacillus anthracis contamination of

USDA and other Washington, DC buildings during the

anthrax attack of October 2001. Appl. Environ. Microbiol.

69, 593–599.

Hoffmaster, A.R., Nine others, 2002. Evaluation and validation

of a real-time polymerase chain reaction assay for rapid

detection of Bacillus anthracis. Emerg. Infec. Dis. 8,

1178–1181.

Hoffmaster, A.R., 21 others. 2004. Identification of anthrax

toxin genes in a Bacillus cereus associated with an illness

resembling inhalation anthrax. Proc. Natl. Acad. Sci. USA

101, 8449–8454.

Jackson, P.J., Walthers, E.A., Kalif, A.S., Richmond, K.L.,

Adair, D.M., Hill, K.K., Kuske, C.R., Wilson, K.H., Hugh-

Jones, M.E., Keim, P., 1997. Characterization of the

variable-number tandem repeats in vrrA from different

Bacillus anthracis isolates. Appl. Environ. Microbiol. 63,

1400–1405.

Jackson, P.J., Hugh-Jones, M.E., Adair, D.M., Green, G., Hill,

K.K., Kuske, C.R., Grinberg, L.M., Abramova, F.A.,

Keim, P., 1998. PCR analysis of tissue samples from the

1979 Sverdlovsk anthrax victims: the presence of multiple

Bacillus anthracis strains in different victims. Proc. Natl.

Acad. Sci. USA 95, 1224–1229.

Jernigan, D.B., and 41 others, 2002.Investigation of bioterror-

ism-related anthrax, United States, 2001: epidemiologic

findings. Emerg. Infec. Dis. 8, 1019-1028.

Luna, V.A., King, D., Davis, C., Rycerz, T., Ewert, M.,

Cannons, A., Amuso, P., Cattani, J., 2003. Novel sample

preparation method for safe and rapid detection of Bacillus

anthracis spores in environmental powders and nasal swabs.

Appl. Environ. Microbiol. 41, 1252–1255.

Morato, J., Mir, J., Codony, F., Mas, J., Ribas, F., 2003.

Microbial response to disinfectants. In: Mara D., Horan,N.

(Ed.), The Handbook of Water and Wastewater Micro-

biology. Academic Press, Boston, MA, pp. 658–693.

Ostensvik, O., From, C., Heidenreich, B., O’ Sullivan, K.,

Granum, P.E., 2004. Cytotoxic Bacillus spp.

belonging to the B. cereus and B. subtilis groups

in Norwegian surface waters. J. Appl. Microbiol. 96,

987–993.

Rose, L.J., Rice, E.W., Jensen, B., Murga, R., Peterson, A.,

Donlan, R.M., Arduino, M.J., 2005. Chlorine inactivation

of bacterial bioterrorism agents. Appl. Environ. Microbiol.

71, 566–568.

ARTICLE IN PRESSA. Perez et al. / Water Research 39 (2005) 5199–5211 5211

Sirisanthana, T., Brown, A.E., 2002. Anthrax of the gastro-

intestinal tract. Emerg. Infec. Dis. 8, 649–651.

Tomov, A.T., 1966. An accelerated combined method of

Anthrax sibirica spores determination in fluvial water. Voen,

Med. Zh. 7, 49–51 [in Russian].

Westphal, A.J., Price, P.B., Leighton, T.J., Wheeler, K.E.,

2003. Kinetics of size changes of individual Bacillus

thuringiensis spores in response to changes in rela-

tive humidity. Proc. Natl. Acad. Sci. USA 100,

3461–3466.

Zheng, G., Yan, L.Z., Vederas, J.C., Zube, P., 1999. Genes of

the sbo-alb locus of Bacillus subtilis are required for

production of the antilisterial bacteriocin subtilosin.

J. Bacteriol. 181, 7346–7355.