rapid-viability pcr method for detection of live, virulent bacillus anthracis in environmental...

9
APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Sept. 2011, p. 6570–6578 Vol. 77, No. 18 0099-2240/11/$12.00 doi:10.1128/AEM.00623-11 Copyright © 2011, American Society for Microbiology. All Rights Reserved. Rapid-Viability PCR Method for Detection of Live, Virulent Bacillus anthracis in Environmental Samples Sonia E. Le ´tant, 1 * Gloria A. Murphy, 1 Teneile M. Alfaro, 1 Julie R. Avila, 1 Staci R. Kane, 1 Ellen Raber, 1 Thomas M. Bunt, 1 and Sanjiv R. Shah 2 Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, California 94550, 1 and U.S. Environmental Protection Agency, National Homeland Security Research Center, USEPA-8801R, 1200 Pennsylvania Avenue NW, Washington, DC 20460 2 Received 18 March 2011/Accepted 6 July 2011 In the event of a biothreat agent release, hundreds of samples would need to be rapidly processed to characterize the extent of contamination and determine the efficacy of remediation activities. Current biolog- ical agent identification and viability determination methods are both labor- and time-intensive such that turnaround time for confirmed results is typically several days. In order to alleviate this issue, automated, high-throughput sample processing methods were developed in which real-time PCR analysis is conducted on samples before and after incubation. The method, referred to as rapid-viability (RV)-PCR, uses the change in cycle threshold after incubation to detect the presence of live organisms. In this article, we report a novel RV-PCR method for detection of live, virulent Bacillus anthracis, in which the incubation time was reduced from 14 h to 9 h, bringing the total turnaround time for results below 15 h. The method incorporates a magnetic bead-based DNA extraction and purification step prior to PCR analysis, as well as specific real-time PCR assays for the B. anthracis chromosome and pXO1 and pXO2 plasmids. A single laboratory verification of the optimized method applied to the detection of virulent B. anthracis in environmental samples was conducted and showed a detection level of 10 to 99 CFU/sample with both manual and automated RV-PCR methods in the presence of various challenges. Experiments exploring the relationship between the incubation time and the limit of detection suggest that the method could be further shortened by an additional 2 to 3 h for relatively clean samples. If a biothreat agent was released, hundreds to thousands of environmental samples of diverse types would need to be rap- idly processed and analyzed in order to first characterize the contamination of the site and then assess the effectiveness of decontamination activities. Decision-makers also need rapid results for remobilizing disinfection equipment in the case of incomplete decontamination and for reopening facilities and areas based on results from clearance sampling (12–14). Current methods used by the Centers for Disease Control and Prevention (CDC) to assess the viability of spores on surfaces rely on culturing samples on solid media (5, 6). These methods involve several manual steps, including pipetting to prepare dilution series, plating of numerous replicates for a series of dilutions, and colony counting, which make it labor-, space-, and time-intensive. Typically, only 30 to 40 samples may be processed each day with confirmed results obtained days later (5, 6). Validated rapid-viability test protocols are therefore needed to ensure public safety and to help mitigate impacts due to facility closures following a biothreat agent release. This critical need was highlighted during the response to the 2001 anthrax attacks, in which clearance sampling and analysis required excessive time prior to facilities reopening. Because risk assessment after such an attack is determined on the basis of the presence of viable spore populations, and because DNA and antigenic materials remain after decontam- ination (3), it is critical to determine viability rather than sim- ply the presence of nucleic acid or protein from a pathogen. We leveraged the useful features of real-time PCR and ex- panded its capabilities by conducting PCR analysis before and after incubating samples and using the change in PCR re- sponse to indicate the presence of viable spores or cells. The approach, referred to as rapid-viability (RV)-PCR, uses ac- cepted methods including culturing and real-time PCR analysis (although in a different format) to allow more rapid and spe- cific analysis. High-throughput sample processing is enabled by commercial automation in combination with 96-well real-time PCR analysis, leading to the processing of hundreds of surface samples per day with results achieved in less than 24 h. Initial RV-PCR protocols were developed and tested with surrogate organisms including Bacillus atrophaeus and the non- virulent Bacillus anthracis Sterne strain. In these experiments, detection of low levels of viable spores (1 to 10 CFU/sample) was demonstrated for various sample types (wipes, swabs, and vacuum filters) in the presence of environmental backgrounds, high populations of live nontarget spores/microorganisms, and dead target spores killed by chlorine dioxide fumigation (8). Hundreds of samples were processed, demonstrating high- throughput analysis and detection limits and accuracy similar to those for traditional viability analysis. The most-probable- number rapid-viability (MPN RV)-PCR, a method variation in which replicates of dilution series are analyzed to provide a quantitative estimate of the spore levels, was also tested along- side the traditional culture method for the quantification of * Corresponding author. Mailing address: Lawrence Livermore Na- tional Laboratory, L-236, 7000 East Avenue, Livermore, CA 94550. Phone: (925) 423-9885. Fax: (925) 423-1026. E-mail: [email protected]. Published ahead of print on 15 July 2011. 6570

Upload: independent

Post on 21-Nov-2023

1 views

Category:

Documents


0 download

TRANSCRIPT

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Sept. 2011, p. 6570–6578 Vol. 77, No. 180099-2240/11/$12.00 doi:10.1128/AEM.00623-11Copyright © 2011, American Society for Microbiology. All Rights Reserved.

Rapid-Viability PCR Method for Detection of Live, VirulentBacillus anthracis in Environmental Samples�

Sonia E. Letant,1* Gloria A. Murphy,1 Teneile M. Alfaro,1 Julie R. Avila,1 Staci R. Kane,1Ellen Raber,1 Thomas M. Bunt,1 and Sanjiv R. Shah2

Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, California 94550,1 andU.S. Environmental Protection Agency, National Homeland Security Research Center,

USEPA-8801R, 1200 Pennsylvania Avenue NW, Washington, DC 204602

Received 18 March 2011/Accepted 6 July 2011

In the event of a biothreat agent release, hundreds of samples would need to be rapidly processed tocharacterize the extent of contamination and determine the efficacy of remediation activities. Current biolog-ical agent identification and viability determination methods are both labor- and time-intensive such thatturnaround time for confirmed results is typically several days. In order to alleviate this issue, automated,high-throughput sample processing methods were developed in which real-time PCR analysis is conducted onsamples before and after incubation. The method, referred to as rapid-viability (RV)-PCR, uses the change incycle threshold after incubation to detect the presence of live organisms. In this article, we report a novelRV-PCR method for detection of live, virulent Bacillus anthracis, in which the incubation time was reduced from14 h to 9 h, bringing the total turnaround time for results below 15 h. The method incorporates a magneticbead-based DNA extraction and purification step prior to PCR analysis, as well as specific real-time PCRassays for the B. anthracis chromosome and pXO1 and pXO2 plasmids. A single laboratory verification of theoptimized method applied to the detection of virulent B. anthracis in environmental samples was conducted andshowed a detection level of 10 to 99 CFU/sample with both manual and automated RV-PCR methods in thepresence of various challenges. Experiments exploring the relationship between the incubation time and thelimit of detection suggest that the method could be further shortened by an additional 2 to 3 h for relativelyclean samples.

If a biothreat agent was released, hundreds to thousands ofenvironmental samples of diverse types would need to be rap-idly processed and analyzed in order to first characterize thecontamination of the site and then assess the effectiveness ofdecontamination activities. Decision-makers also need rapidresults for remobilizing disinfection equipment in the case ofincomplete decontamination and for reopening facilities andareas based on results from clearance sampling (12–14).

Current methods used by the Centers for Disease Controland Prevention (CDC) to assess the viability of spores onsurfaces rely on culturing samples on solid media (5, 6). Thesemethods involve several manual steps, including pipetting toprepare dilution series, plating of numerous replicates for aseries of dilutions, and colony counting, which make it labor-,space-, and time-intensive. Typically, only 30 to 40 samplesmay be processed each day with confirmed results obtaineddays later (5, 6). Validated rapid-viability test protocols aretherefore needed to ensure public safety and to help mitigateimpacts due to facility closures following a biothreat agentrelease. This critical need was highlighted during the responseto the 2001 anthrax attacks, in which clearance sampling andanalysis required excessive time prior to facilities reopening.

Because risk assessment after such an attack is determinedon the basis of the presence of viable spore populations, and

because DNA and antigenic materials remain after decontam-ination (3), it is critical to determine viability rather than sim-ply the presence of nucleic acid or protein from a pathogen.We leveraged the useful features of real-time PCR and ex-panded its capabilities by conducting PCR analysis before andafter incubating samples and using the change in PCR re-sponse to indicate the presence of viable spores or cells. Theapproach, referred to as rapid-viability (RV)-PCR, uses ac-cepted methods including culturing and real-time PCR analysis(although in a different format) to allow more rapid and spe-cific analysis. High-throughput sample processing is enabled bycommercial automation in combination with 96-well real-timePCR analysis, leading to the processing of hundreds of surfacesamples per day with results achieved in less than 24 h.

Initial RV-PCR protocols were developed and tested withsurrogate organisms including Bacillus atrophaeus and the non-virulent Bacillus anthracis Sterne strain. In these experiments,detection of low levels of viable spores (1 to 10 CFU/sample)was demonstrated for various sample types (wipes, swabs, andvacuum filters) in the presence of environmental backgrounds,high populations of live nontarget spores/microorganisms, anddead target spores killed by chlorine dioxide fumigation (8).Hundreds of samples were processed, demonstrating high-throughput analysis and detection limits and accuracy similarto those for traditional viability analysis. The most-probable-number rapid-viability (MPN RV)-PCR, a method variation inwhich replicates of dilution series are analyzed to provide aquantitative estimate of the spore levels, was also tested along-side the traditional culture method for the quantification of

* Corresponding author. Mailing address: Lawrence Livermore Na-tional Laboratory, L-236, 7000 East Avenue, Livermore, CA 94550.Phone: (925) 423-9885. Fax: (925) 423-1026. E-mail: [email protected].

� Published ahead of print on 15 July 2011.

6570

B. anthracis Sterne spores in macrofoam swabs from a multi-center validation study conducted by the CDC (10). MPNRV-PCR provided correct identification for all samples ana-lyzed in this study in less than 24 h, and the estimation of thenumber of spores by MPN RV-PCR was within the order ofmagnitude of the values determined using the traditional culturemethod (6). An integrated culture and real-time PCR method toassess viability of disinfectant-treated Bacillus spores using ro-botics and the MPN approach was also tested by Varugheseet al. (19). Results showed no significant difference betweenbroth culture enrichment followed by PCR and filter plating onagar for untreated spores, but recoveries of chlorine-treatedspores were significantly improved when using the broth en-richment method.

This study describes optimized automated and manual RV-PCR methods for the detection of virulent B. anthracis Amesin wipes, air filters, and water samples. Real-time PCR assaystargeting the chromosome and both plasmids were selected insilico, experimentally optimized, and evaluated for selectivity,sensitivity, and robustness in the presence of growth mediumand cell debris before being integrated into the method. Themethod endpoint was shortened from its initial overnight in-cubation (14 h) to 9 h by performing a magnetic bead-basedDNA extraction and purification procedure before PCR anal-ysis, and a single laboratory verification demonstrated that aspore level of 10 to 99 CFU/sample was detected for all sampletypes in the presence of challenges including high levels of dirt,high levels of autoclaved B. anthracis Ames spores, and highlevels of live environmental background including spores andvegetative cells.

MATERIALS AND METHODS

Bacterial strain and culture conditions. The pathogenic B. anthracis Amesstrain was obtained from the Lawrence Livermore National Laboratory (LLNL)strain collection and cultivated in brain heart infusion (BHI) medium as well asBHI agar plates. Spore stocks were stored in a 70% water and 30% ethanolsolution at 4°C for the duration of the study (1 year). All work performed in thisstudy was approved by the LLNL Institutional Biosafety Committee and con-ducted in the LLNL Select Agent Facility, with required permits from the CDC.

B. anthracis Ames spore preparation. B. anthracis was streaked for growth ontoBHI agar and incubated overnight at 36°C (6). The organism was then streakedand incubated a second time for isolation. A 108-CFU/ml suspension of the 24-hgrowth was prepared in phosphate buffer (25 mM KH2PO4, pH 7.4), plated ontosoil extract beef peptone agar (1), and incubated at 36°C for 7 days until 99%sporulation was achieved. Sporulation was evaluated using phase-contrast mi-croscopy of a wet mount (vegetative cells are phase dark, while spores are phasebright, due to their high refractive index). Plates were then scraped and rinsedusing sterile water and a cell scraper (the content of each plate was transferredto a 50-ml centrifuge tube in a total of 30 ml of water). The spore preparation wascleaned using vortexing (2 min), centrifugation (4,000 rpm for 15 min), removalof the supernatant, and addition of sterile water. This cleanup procedure wasrepeated 4 times. Twenty milliliters of a 1:1 (ethanol-water) solution was thenadded to the centrifuge tubes, which were vortexed for 2 min to resuspend thespore pellets. Tubes were then placed on a shaker platform for 1 h at 80 rpm.After this step, the spore suspension was washed again 7 consecutive times usingthe vortexing, centrifugation, and supernatant exchange technique describedabove. The suspension titer after these washing steps was 109 CFU/ml, as mea-sured by plating, and the fraction of dead spores, measured by phase-contrastmicroscopy, was �1%. The final spore resuspension was performed using amixture of 70% water and 30% ethanol in order to generate a spore stock forstorage at 4°C for the duration of the study (1 year).

Sample spiking. Prior to each RV-PCR experiment, a B. anthracis workingspore stock (104 CFU/ml) was vortexed on a platform vortexer (model VX-2500,VWR International) for 20 min. Two successive 10-fold dilutions of the sporestock were prepared in phosphate buffer (70% of 0.25 mM KH2PO4/0.1% Tween

80 [pH 7.4], and 30% ethanol). Three replicates each of the first and seconddilutions were cultured on agar plates using the spread plate method describedbelow. Typically, 100 �l of the second dilution (102 CFU/ml) and 50 �l of the firstdilution (103 CFU/ml) were plated in triplicate, in order to determine the actualinoculation levels. Water (sterile, deionized water), air filter (catalog number[Cat. No.] FSLW04700, Millipore) and wipe samples (Cat. No. 8052, VWRInternational) were placed in 30-ml conical tubes, and a mesh support (Cat. No.93185 T17; 2.75-inch [6.98-cm] diameter, 0.033-inch [�0.084-cm] openings;McMaster-Carr Inc.) was added to maintain wipe and air filter samples to theside of the tube, clear of pipetting. Samples were then inoculated using 100 �l ofthe second dilution (final concentration, 10 CFU/sample level), 100 �l of the firstdilution (final concentration, 100 CFU/sample level), or 100 �l of the workingstock (final concentration, 1,000 CFU/sample level) pipetted into the samplematerial.

The targeted levels for this study were the 10-spore level (10 to 99 CFU/sample), the 100-spore level (100 to 999 CFU/sample), and the 1,000-spore level(1,000 to 9,999). Although the goal was to test the lowest spore numbers for eachlevel, variability with pipetting, vortexing, and potential surface binding of thespores to stock tubes generated slightly different CFU values for each experi-ment, which were quantified by systematic plating of the spiking solutions priorto sample inoculation using the spread plate method.

Preparation of dirty wipes. The well-characterized Arizona Fine Test Dust(Powder Technology Inc., Burnsville, MN) was used for this study. The materialconsists of Arizona sand including Arizona Road Dust, Arizona Silica, AirCleaner Fine and Air Cleaner Coarse Test Dusts, Society of Automotive Engi-neers Fine and Coarse Test Dusts, J726 Test Dusts, ISO Ultrafine, ISO Fine, ISOMedium, and ISO Coarse Test Dusts, and MIL STD 810 Blowing Dust (11).Analysis of chemical composition performed by the manufacturer indicates thatthe material consists of the following: SiO2 (68 to 76%), Al2O3 (10 to 15%),Fe2O3 (2 to 5%), Na2O (2 to 4%), CaO (2 to 5%), MgO (1 to 2%), TiO2 (0.5 to1.0%), and K2O (2 to 5%). Additional characterization of the test dust wasperformed by the CDC (6). The Arizona Test Dust was autoclaved at 126°C and15 lb/in2 for 30 min in order to differentiate chemical challenges from biologicalchallenges. Two hundred fifty milligrams of Arizona Test Dust was added to eachwipe sample tube and weighed directly in the tube to avoid any loss of materialduring transfers (referred to as dirty wipe sample).

Preparation of dirty air filters. Air filters (Cat. No. FSLW04700, Millipore)collected from portable air sampling units that operated for 24 h at approxi-mately 200 liters/min were used as challenges for this study. One unit was locatedin a subway, and another unit was located outdoors. Filters were spiked asreceived.

Preparation of chemically spiked water samples. The water used in this studywas filtered through a Milli-Q water system (Millipore Co., Billerica, MA), whichincluded a 0.22-�m filter rated for bacterial removal of �0.1 CFU/ml andparticulate (diameter, �0.22 �m) removal (�1 particulate/ml). Challenge sam-ples were prepared by adding ferrous sulfate and humic acid (solid materialpurchased from Sigma-Aldrich, Co., Saint Louis, MO; Cat. Nos. 53680 andF8048, respectively) at levels of 10 mg/liter.

Addition of autoclaved B. anthracis Ames spore background. A stock of B.anthracis Ames spores (106 CFU/ml confirmed by the spread plate method) waskilled by autoclaving three consecutive times at 126°C and 15 lb/in2 for 20 min,for a total autoclaving time of 1 h. Six 100-�l aliquots were cultured on solid BHImedium and incubated for 48 h at 37°C to confirm nonviability of the stock afterthe autoclave treatment. One milliliter of the autoclaved spore stock solution(106 spores/sample) was added to each sample type as a challenge.

Addition of live nontarget background. A Bacillus atrophaeus (ATCC No.9372) spore preparation was acquired from Apex Laboratories Inc., Apex, NC.The spore preparation was diluted to a spore level of 104 CFU/ml, confirmed byplating using the spread plate method. One hundred microliters was then inoc-ulated on each sample (103 CFU/sample) type as a live challenge. Pseudomonasaeruginosa (ATCC 10145) cells were grown overnight in a shake flask and dilutedto a concentration of 107 CFU/ml using optical density measurements at 620 nmto assess the cell concentration. One hundred microliters (106 CFU/sample) ofthis diluted culture was inoculated on each sample. The final live background foreach sample was a combination of 103 B. atrophaeus spores and 106 P. aeruginosacells.

Traditional viability. For traditional viability analysis using the spread platemethod, 2 or 3 successive 10-fold dilutions were cultured on BHI agar andincubated overnight at 30°C. Three 100-�l replicates were plated for each sam-ple. Colony counts were obtained the next day and corrected for dilution.

Rapid-viability PCR. The experimental protocol outline is provided in Fig. 1.Twenty milliliters of extraction buffer (70% of 0.25 mM KH2PO4/0.1% Tween 80[pH 7.4] and 30% ethanol in order to lyse vegetative cells; final pH �9.5) was

VOL. 77, 2011 RV-PCR METHOD TO DETECT LIVE, VIRULENT B. ANTHRACIS 6571

added to each sample tube (for wipes and filters), and the tubes were vortexedfor 20 min on a platform vortexer to remove spores from the sample matrix.Thirteen milliliters was then transferred from each sample tube to a filter cup(when performing the protocol manually, 15 ml was pipetted out of the sampletube and 13 ml was dispensed into the filter cup, to minimize the probability ofaerosol formation with select agents; the same volume was transferred withrobotics, to provide consistency). Spores suspended in the extraction buffer werethen collected with the 0.45-�m filter of the filter cups by using a vacuummanifold and a vacuum pump (0.45-�m filters provided faster filtration in thepresence of dirt and debris while generating results that were not statisticallydifferent from those for 0.22-�m filters) (data not shown). Filters were washedwith 7 ml of filter-sterilized 210 mM KH2PO4 buffer (pH 6.0) followed by 3 mlof filter-sterilized 25 mM KH2PO4 buffer (pH 7.4). Filter cups were then sealedon the bottom using a custom capping plate containing quick-turn fittings (Cat.No. 51525K372, McMaster-Carr), 2.5 ml of BHI growth medium was added, andthen filter cups were sealed on the top using push-in caps (Cat. No. 94075K56,McMaster-Carr). After vortexing for 10 min on a rack vortexer, 100-�l aliquotswere taken from each filter cup and transferred to Eppendorf tubes (T0 aliquots)for PCR analysis (after dilution in 900 �l of BHI growth medium followed bymagnetic bead-based DNA extraction and purification, as described below). Thecups were resealed on the top and incubated for 9 h at 37°C and 230 rpm.One-milliliter aliquots were taken from each filter cup at the endpoint (T9

aliquots), after vortexing the filter cup manifold on a platform vortexer for10 min.

When the RV-PCR protocol was performed manually, all liquid handling waseffectuated with serological pipettes and micropipettes. In the automated versionof the protocol, a robotic platform (Janus workstation, Perkin-Elmer) performedall the liquid handling steps required to implement the RV-PCR method (mixingand transferring buffer from sample extracts to filtration cups for spore collec-

tion, as well as performing washes on the filters, adding growth medium to thefilter cups for culturing, and sampling cultures for PCR analysis), with theexception of the initial sample spiking and magnetic bead-based DNA extractionand purification.

Magnetic bead-based DNA extraction and purification. The 1-ml sample ali-quots taken after 9 h of incubation (T9 aliquots) were manually processed usingthe Promega Magnesil paramagnetic particle (PMP) DNA extraction and puri-fication kit (Cat. No. MD1360, Technical Bulletin 312, Promega Co.). In this kit,DNA separation does not rely on any column or centrifugation steps, which arechallenging to automate. In addition, DNA binding occurs in solution, increasingbinding efficiency, and PMPs are completely resuspended during wash steps,enhancing the removal of contaminants in complex environmental samples (2).Briefly, 1 ml of each sample was transferred from the filter cup into a 2 mlEppendorf tube, followed by addition of 600 �l of Promega bead mix and 360 �lof Promega lysis buffer. Sample, buffer, and PMPs were mixed by pipetting andtubes were mounted on a tube rack interfaced with a magnet (DynaMag-2magnet, Cat. No. 123-21D, Invitrogen). Beads with attached DNA were attractedto the magnet and the supernatant was removed by pipetting. An additional lysiswith 360 �l of lysis buffer was conducted with mixing by vortexing and removalof the supernatant. Two washes ith 360 �l of Promega salt wash solution werethen performed, followed by mixing by vortexing and removal of the supernatant.Finally, two washes with 360 �l of Promega alcohol wash solution were per-formed with mixing and supernatant removal. Beads were allowed to air dry for2 min, followed by transfer of the tube rack from the magnetic support to a heatblock and heating at 80°C until samples were totally dried (typically 30 to 40min). DNA elution/concentration was then performed by adding 200 �l ofPromega elution buffer after letting samples cool for 5 min. The sample withbuffer was mixed and transferred to the magnetic support, and the supernatantwith eluted DNA was recovered (typically 80 �l). A 10-fold dilution of the eluted

FIG. 1. Summary of RV-PCR sample processing steps, with corresponding sample volumes.

6572 LETANT ET AL. APPL. ENVIRON. MICROBIOL.

sample in PCR-grade water was systematically performed using PCR-grade wa-ter and a 96-well Bioblock (Cat. No. 662000, E&K Scientific) prior to runningPCR, in order to counter inhibition from the environmental background (allsamples were processed with a 1:10 dilution for consistency).

It should be noted that sample volumes aliquoted at T0 and T9 are different:100 �l was aliquoted at T0 and diluted 10-fold in BHI medium prior to DNAextraction, while 1 ml was aliquoted at T9 and then diluted 10-fold prior toperforming PCR. Although the aliquot volumes at T0 and T9 were different, thetotal sample volume processed for DNA extraction and purification remainedunchanged (1 ml), and the final dilution factor remained identical (1:10 dilution).The reason for aliquoting only 100 �l of sample at T0 was to avoid the withdrawalof a significant portion of the sample from the filter cup, which contains only 2.5ml of growth medium. A large-volume withdrawal at T0 would have negativelyimpacted the limit of detection of the method.

Extracted DNA controls. DNA controls were generated for the B. anthracisAmes strain. DNA was extracted from cultured cells by using a complete DNAand RNA purification kit (MasterPure, Cat. No. MC85200, Epicentre Biotech-nologies) and the corresponding DNA purification procedure including the ad-dition of RNase A, according to the manufacturer’s protocol. DNA concentra-tion was measured with a Qubit fluorometer (Cat. No. Q32857, Invitrogen) usingthe PicoGreen assay (Cat. No. Q32851, Invitrogen). Standard concentrationsranging from 1 ng/�l to 1 fg/�l were prepared in PCR-grade water. Seven 10-folddilutions, ranging from 5 ng per 25-�l PCR to 5 fg per 25-�l PCR, were run witheach set of PCR plates.

PCR. Five-microliter sample aliquots were transferred to a 96-well PCR platewith 20 �l of PCR mix. PCR mix was prepared for each of the 3 primer-probesets using 12.5 �l of TaqMan 2� universal master mix (Cat. No. 4305719,Applied Biosystems) and 1 �l of 2 �M probe solution. For the chromosome andpXO1 assays, 1 �l of a 25 �M solution was used for each primer. For the pXO2assays, 0.3 �l of the 25 �M solution was used for each primer. The volume ofPCR mix was completed to 20 �l using PCR-grade water. After mixing andcentrifugation, PCR was run using the ABI 7500 Fast platform (Applied Bio-systems) in fast mode. Thermal cycling parameters were as follows: 2 min at 50°Cfor uracil-N-glycosylase incubation, 10 min at 95°C for AmpliTaq gold activation,followed by 45 amplification cycles (5 s at 95°C for denaturation and 20 s at 60°Cfor annealing/extension). Each sample was analyzed against each of the 3 primer/probe sets. A minimum of 4 sample replicates were analyzed for each set ofexperimental conditions. The ROX reference dye contained in the ABI universalmaster mix was used to normalize the fluorescent reporter signal. The fluores-cence baseline was set from 5 cycles to 2 cycles prior to the cycle with the earliestvisible amplification. The threshold was set manually above the baseline andwithin the exponential growth region of a group of amplification plots includingall samples analyzed in a given experiment and corresponding extracted DNAcontrols (7500/7500 fast real-time PCR system standard curve experiments man-ual, part number 4387779; Applied Biosystems). The cycle threshold (CT) wasconsistent with being the third consecutive cycle which had an increase in fluo-rescence for DNA standards.

RV-PCR controls at T0. Control experiments performed by aliquoting 1 mlfrom each filter cup immediately after addition of BHI growth medium (T0)followed by vortexing and processing aliquots with the magnetic bead-basedDNA extraction and purification method described above showed no detectablePCR signal for any of the 3 B. anthracis assays for spiking levels up to 105

CFU/sample. These results confirmed that the DNA extraction and purificationprocedure does not lyse B. anthracis spores and therefore that no DNA isreleased from spores at T0. Similar controls were performed on samples inocu-lated with 106 autoclaved B. anthracis spores, again showing an absence of anymeasurable amplification at T0. While the analysis of T0 samples is superfluouswhen working at low spore levels with a highly purified spore suspension, itbecomes critical when working with environmental samples, which might containhigh concentrations of indigenous DNA due to the presence of dead vegetativecells.

It should be noted that the RV-PCR method detects a change in DNAconcentration (CT value) after enrichment, which can be generated only by liveorganisms. Any indigenous DNA detected at T0 would also be detected at T9 atthe same level and would therefore not contribute to a change in CT value.

RV-PCR result interpretation. Criteria for positive, viable B. anthracis sporedetection with the RV-PCR method are as follows: endpoint PCR CT after 9-hincubation of �36.0 for all 3 assays, and �CT (CT[T0] � CT[T9]) of �9.0 for all3 assays (to represent at least a 3-log increase in DNA concentration after 9 h ofincubation). It should be noted that most laboratories use the PCR CT value of�40.0 as a cutoff value for a positive detection of live or dead B. anthracis spores(when 45 PCR cycles are used). We set a stringent CT cutoff value to test thereliability and robustness of the RV-PCR method. Depending upon the end user

requirement and the phase of response during an event, a lower �CT and ahigher endpoint CT could be set.

Since no CT values were detected at T0 for any sample, these values were setat 45 (the number of amplification cycles performed). All results from this studywere therefore presented as CT values after 9 h of incubation (T9). Each CT

represents the average of 4 replicate samples. Standard deviations were reportedfor each CT value.

The limit of detection of the RV-PCR method was defined as the lowest sporelevel at which the positive, viable B. anthracis spore detection criteria definedabove were met for all 3 assays.

PCR assay selectivity study. PCR assay selectivity was tested using a panel of13 B. anthracis strains (Turkey 32, A0149, A0248, V770-NP-1R, Ba1015, SK-102,Ba1035, K3, PAK-1, RA3, Vollum 1B, Sterne, Ames) and 15 B. anthracis nearneighbors (Bacillus cereus [S2-8, 3A, E33L, D17, FM1, 03BB102, 03BB108],Bacillus thuringiensis [HD1011, 97-27, HD682, HD571], B. thuringiensis subsp.israelensis, B. thuringiensis subsp. kurstaki, B. thuringiensis subsp. morrisoni, B.thuringiensis strain Al Hakam) acquired from the LLNL strain collection. Con-centrations of extracted DNA template stocks were measured with a Qubitfluorometer (Invitrogen, Carlsbad, CA) using the PicoGreen assay (Invitrogen,Carlsbad, CA), and DNA stocks were then diluted to appropriate concentrationsusing PCR-grade water. PCR plate layouts were prepared to run triplicatereactions for each assay against each target and near-neighbor DNA template(two concentrations were run for each template: 500 fg per reaction and 50 fg perreaction for targets, and 5 pg and 500 fg per reaction for near neighbors). Threepositive controls (extracted B. anthracis Ames DNA) and three negative controls(PCR-grade water) were run for each assay on each PCR plate.

RESULTS AND DISCUSSION

PCR assays. Potential assays were ranked and selectedusing in silico analysis performed with LLNL’s KPATH sys-tem (16). Selection criteria included signature specificityagainst all available sequences in GenBank, virulence geneassociation, availability of prior assay screening data, and am-plicon characteristics. The output of this analysis was a com-putational prediction of virulent B. anthracis strain detectionfor 44 candidate assays ranked for predicted selectivity, ampli-con size, and gene target. Ten assays (3 for the chromosome, 4for the pXO1 plasmid, and 3 for the pXO2 plasmid) wereselected based on the in silico analysis and then optimized forreal-time PCR. Three assays (1 for the chromosome targetinga hypothetical protein [NC_005945.1], 1 for the pXO1 plasmid[4] targeting a hypothetical protein [NC_001496.1], and 1 forthe pXO2 plasmid [4] targeting the capsule biosynthesis genecapB [NC_007323.3]) were ultimately selected for RV-PCRbased on sensitivity, selectivity, and robustness in the presenceof growth medium and cell debris (sequence for primers andprobes and amplicon length are provided in Table 1 for eachassay).

Assay sensitivity with extracted B. anthracis Ames DNA wasshown to be below 10 genome copies for the 3 selected assays(Table 2). Assay specificity was tested using a panel of 13 B.anthracis strains and 15 B. anthracis near-neighbor strains de-scribed in Materials and Methods. As seen in Table 3, the 3assays detected all 13 B. anthracis strains tested. The pXO1 andpXO2 assays exhibited cross-reactivity with just 1 of 15 (B. cereusO3BB102) and 2 of 15 (B. cereus O3BB102 and O3BB108) near-neighbor strains, respectively. By BLAST analysis, the 101-bppXO1 amplicon shows 100% homology with the B. cereusO3BB102 sequence, whereas the 77-bp pXO2 amplicon con-tains 2 nonadjacent mismatches within the probe sequence.The B. cereus O3BB108 whole-genome shotgun sequencecould not be compared in silico; however, the strain DNA wasshown to be positive for B. anthracis pXO2 encapsulation pro-

VOL. 77, 2011 RV-PCR METHOD TO DETECT LIVE, VIRULENT B. ANTHRACIS 6573

tein genes by PCR (7). No amplification was observed with anyother near-neighbor organism tested. Finally, the robustness ofthe 3 assays in the presence of growth medium and cell debriswas tested by diluting extracted B. anthracis Ames DNA in alysed culture of B. atrophaeus (109 CFU/ml). No statisticaldifference in average CT value was induced by the presence ofgrowth medium and cell debris compared to PCR-grade water,confirming that the assay performance was adequate for use inthe RV-PCR method (data not shown).

Detection of live B. anthracis Ames spores on clean samples.Manual and automated RV-PCR experiments were performedon clean wipe, air filter, and water samples spiked with live B.anthracis Ames spore levels of 10 and 100 CFU/sample (Fig.2). Average CT values at T9 were below 36.0 for all samples andall assays, indicating a limit of detection at or below the levelof 10 to 99 CFU/sample for clean samples. It should be notedthat at this low spiking level (for example, 10 spores/sample,which becomes 10 spores/20 ml of buffer in the sample tube),plating of the samples directly from the sample tube or fromthe filter cup after addition of growth medium did not lead toany detectable colonies after 24 h of incubation on BHI agar.This result exemplifies the advantage of the RV-PCR methodover the plating method (where the entire sample volume isnot typically plated) for the detection of low levels of live B.anthracis spores. Culturing of sample aliquots drawn from filtercups after 9 h of incubation on BHI agar plates showed thatsamples went from being undetectable to a level of 106

CFU/ml in 9 h (data not shown). This observation matchesestimates based on a doubling time on the order of 30 min anda germination time on the order of 30 min, which lead to apredicted number of spores of 8.5 � 105 in filter cups at T9,

when spiking samples with 10 spores/sample. This qualitativeagreement between the experimental results and the sporenumber estimates suggests reasonably good growth conditionsfor B. anthracis in this high-throughput format (filter cups) andconfirms that the CT values recorded at T9 originated fromviable spores.

From the germination and doubling time estimates above,the number of genome copies per PCR can be projected forsamples spiked with 31 CFU/sample to be on the order of 2,500gene copies after the 1:10 dilution. According to results pre-sented in Table 2, such a number should generate CT values onthe order of 29 for the chromosome assay, which is indeed thecase (see Fig. 2), confirming optimal growth as well as anabsence of PCR inhibition.

The plasmid assays consistently generated slightly higher CT

values than the chromosome assay, which may be attributed toa combination of a higher number of plasmid copies relative tothe chromosome (15) and a lower efficiency for the chromo-some assay relative the plasmid assays (Table 2).

Results provided by the manual and automated methodswere not statistically different (results from two-tailed Stu-dent’s t test for wipe samples are as follows: P � 0.070 for thechromosome assay, P � 0.728 for the pXO1 assay, and P �0.848 for the pXO2 assay), which suggests that automation maybe used to reduce labor and personnel exposure to contami-nated samples without compromising the limit of detection ofthe RV-PCR method.

Detection of live B. anthracis Ames spores in the presence ofdirt and/or debris. Manual and automated RV-PCR experi-ments were performed on dirty wipe, air filter, and water sam-ples spiked with live B. anthracis Ames spores at levels of 10

TABLE 2. CT values and corresponding standard deviationsa for down-selected PCR assays, tested with extractedBacillus anthracis Ames DNA

DNA amt (pg)b Genome copy no.cCT (SD) or % efficiency for indicated PCR assay:

Chromosome pXO1 pXO2

5,000 829,000 19.4 (0.2) 20.7 (0.2) 18.5 (0.1)500 82,900 22.6 (0.1) 23.1 (0.1) 21.4 (0.1)50 8,290 26.5 (0.1) 26.4 (0.1) 24.7 (0.1)5 829 30.6 (0.1) 29.9 (0.2) 28.4 (0.1)0.5 82.9 34.6 (0.1) 33.4 (0.1) 31.9 (0.1)0.05 8.29 38.5 (0.4) 37.2 (0.6) 35.4 (0.3)0 0 NDTd NDTd NDTd

PCR efficiency 81% (R2 � 0.99) 100% (R2 � 0.99) 96% (R2 � 0.99)

a CT values are averages of 3 replicate reactions.b DNA was quantified by fluorimetry using the PicoGreen assay.c Genome copies are based on estimated genome size of 5.5 Mb.d NDT indicates that no CT value was measured.

TABLE 1. Sequences of primers and probes used to detect the Bacillus anthracis chromosome and plasmids

Assay Forward primer Reverse primer Probe Ampliconlength (bp)

Chromosome TTTCGATGATTTGCAATGCC

TCCAAGTTACAGTGTCGGCATATT

ACATCAAGTCATGGCGTGACTACCCAGACTT

105

pXO1 plasmid GCGGATAGCGGCGGTTA TCGGTTCGTTAAATCCAAATGC ACGACTAAACCGGATATGACATTAAAAGAAGCCCTTAA

101

pXO2 plasmid TGCGCGAATGATATATTGGTTT

GCTCACCGATATTAGGACCTTCTTTA

TGACGAGGAGCAACCGATTAAGCGC 77

6574 LETANT ET AL. APPL. ENVIRON. MICROBIOL.

and 100 CFU/sample (Fig. 3). As described in Materials andMethods, dirty wipes were prepared by adding 250 mg of Ar-izona Fine Test Dust, dirty air filters came from air samplingunits (operated at 85 to 100 liters/min for 24 h), and dirty waterwas prepared by spiking Milli-Q-filtered water with 10 mg/literof humic acid and 10 mg/liter of ferrous sulfate, which areknown PCR inhibitors (9, 17, 18). Although high CT valueswere obtained for dirty samples relative to clean samples, allspiked samples were detected, with average T9 CT values below

36.0 for all assays, indicating that the limit of detection ob-tained for clean samples was maintained in the presence of dirtand/or debris. It should be noted that all samples are diluted bya factor of 10 in PCR-grade water after the magnetic bead-based DNA extraction and purification step, as described inMaterials and Methods. This additional step was added intothe method in order to counter PCR inhibition from environ-mental compounds, such as humic acid, ferrous sulfate, andmetal oxides present in water and dust/soil (typical average CT

TABLE 3. CT values and corresponding standard deviationsa for down-selected PCR assays, tested with extracted DNAfrom various strains of Bacillus anthracis

Strain

CT (SD) for PCR assays at indicated DNA template concnb (fg):

Chromosome pXO1 pXO2

50 500 50 500 50 500

Bacillus anthracis Turkey 32 36.5 (0.7) 32.7 (0.2) 34.7 (0.4) 31.3 (0.1) 34.0 (0.3) 30.2 (0.1)Bacillus anthracis A0149 37.2 (0.6) 32.7 (0.5) 35.6 (0.3) 31.9 (0) 35.8 (0.5) 31.9 (0.3)Bacillus anthracis A0248 36.6 (0.7) 32.9 (0.3) 35.6 (0.8) 31.7 (0.2) 35.9 (1.0) 31.5 (0.2)Bacillus anthracis V770-NP-1Rc 36.4 (0.4) 33.0 (0.3) 35.1 (0.8) 31.2 (0.1) NDTd NDTd

Bacillus anthracis Ba1015 38.1 (1.5) 33.7 (0.2) 36.2 (1.0) 32.5 (0.2) 35.5 (0.5) 32.6 (0.2)Bacillus anthracis SK-102 37.1 (0.7) 32.9 (0.2) 35.7 (0.2) 32.2 (0.2) 35.0 (0.2) 30.7 (0.4)Bacillus anthracis Ba1035 37.4 (1.2) 33.3 (0.5) 35.0 (0.1) 31.4 (0.1) 34.6 (0.7) 30.3 (0.1)Bacillus anthracis K3 37.1 (0.3) 33.4 (0.2) 35.5 (0.5) 31.7 (0.1) 35.9 (1.5) 30.7 (0.2)Bacillus anthracis Ames 35.9 (0.5) 33.0 (0.3) 35.1 (0.1) 31.4 (0.1) 35.1 (0.9) 31.4 (0.3)Bacillus anthracis PAK-1 36.8 (1.7) 33.0 (0.2) 34.8 (0.2) 31.7 (0.1) 34.0 (0.2) 30.6 (0.2)Bacillus anthracis RA3 38.4 (0.5) 33.9 (0.3) 35.6 (0.3) 32.5 (0.1) 35.5 (0.8) 31.9 (0.3)Bacillus anthracis Vollum 1B 38.8 (0.4) 34.3 (0.3) 37.1 (0.9) 33.3 (0.1) 35.0 (0.2) 31.5 (0.3)Bacillus anthracis Sternec 36.7 (0.6) 33.1 (0.1) 35.0 (0.2) 31.5 (0.3) NDTd NDTd

a CT values are averages of 3 replicate reactions.b DNA was quantified by fluorimetry using the PicoGreen assay.c pXO2� strain.d NDT indicates that no CT value was measured.

FIG. 2. Manual and automated RV-PCR results obtained for cleanwipe, air filter, and water samples spiked with Bacillus anthracis Amesspores (samples processed with the manual protocol were spiked with31 2 CFU/sample and samples processed with the automated pro-tocol were spiked with 26 1 CFU/sample, as measured by plating).Each CT value is the average of 4 replicate samples. Error bars rep-resent 1 standard deviation above and below the average CT value.

FIG. 3. Manual and automated RV-PCR on dirty wipe, air filter,and water samples spiked with Bacillus anthracis Ames spores (samplesprocessed with the manual protocol were spiked with 49 3 CFU/sample and samples processed with the automated protocol werespiked with 40 2 CFU/sample, as measured by plating). Each CTvalue is the average of 4 replicate samples. Error bars represent 1standard deviation above and below the average CT value.

VOL. 77, 2011 RV-PCR METHOD TO DETECT LIVE, VIRULENT B. ANTHRACIS 6575

values for the 10-spore level without dilution are in the 37-to-45 [45 � nondetected] range for dirty samples, while typicalCT values after 1:10 dilution are in the 30-to-36 range for thesame sample extracts). The results provided by the manual andautomated methods were not statistically different, as previ-ously observed for clean samples, which suggests that the useof high-throughput automation is appropriate for environmen-tal samples. In addition, filtration times for dust-containingsamples were not significantly different from filtration times forclean samples, since the filter cups have a large filtration area(4.7 cm2).

Detection of live B. anthracis Ames spores in the presence ofa high background of autoclaved B. anthracis Ames spores.Overcoming the challenge posed by high levels of dead B.anthracis spores is critical for remediation activities involvingpost-decontamination clearance. In this application, very lowlevels of live spores must be detected in a high background ofkilled spores. Manual and automated RV-PCR experimentswere performed on clean wipe, air filter, and water samplesspiked with live B. anthracis Ames spores at levels of 10 and100 CFU/sample in a background of 106 B. anthracis Amesspores/sample killed by autoclaving. Figure 4 summarizes theT9 CT values obtained with 10 to 99 B. anthracis Ames CFU/sample with each of the 3 assays. For all sample types, averageCT values were below 36.0 for all assays, indicating that a levelof 10 live B. anthracis Ames spores per sample was detected ina background of 106 autoclaved spores. Interferences of thedecontamination method (fumigants or liquid disinfectants)with the RV-PCR method were not tested in this study; how-ever, earlier studies with B. anthracis surrogates showed noimpact of residual fumigant (8). Typical effects of decontami-

nation including delayed germination and growth and PCRinhibition can usually be overcome by increasing incubationtime and sample dilution (8). The RV-PCR protocol includeswashes of the spores in the filter cups in order to remove anyresidual disinfectant.

Detection of live B. anthracis Ames spores in the presence ofa high background of live nontarget organisms. Manual andautomated RV-PCR experiments were performed on cleanwipe, air filter, and water samples spiked with live B. anthracisAmes spores at levels of 10 and 100 CFU/sample in a com-bined background of 103 live B. atrophaeus spores and 106 P.aeruginosa cells/sample. B. atrophaeus is a Gram-positive, aer-obic, endospore-forming bacterium which is used in industry asa sterilization control strain and within the biodefense researchcommunity as a nonpathogenic surrogate for B. anthracis (8).P. aeruginosa is a Gram-negative, aerobic bacterium which isfound in soil, water, skin flora, and most man-made environ-ments throughout the world. These two organisms were chosento create a mixture of live nontarget spores and vegetative cellswhich could compete with B. anthracis for growth and affectthe efficiency of the RV-PCR method.

Figure 5 summarizes the T9 CT values obtained with 10 to 99B. anthracis Ames CFU/sample for each of the 3 assays. For allsample types, average CT values were below 36.0, indicatingthat the presence of a high background of live organisms didnot negatively impact the limit of detection of the method. Theability to detect live target organisms in a complex environ-mental background is provided by the selectivity of the PCRassays and constitutes a critical advantage of the RV-PCRmethod over the standard plating method, in which agar plates

FIG. 4. Manual and automated RV-PCR results obtained for cleanwipe, air filter, and water samples spiked with Bacillus anthracis Amesspores (samples processed with the manual protocol were spiked with14 1 CFU/sample and samples processed with the automated pro-tocol were spiked with 38 2 CFU/sample, as measured by plating) ina background of 106 autoclaved Bacillus anthracis Ames spores/sample.Each CT value is the average of 4 replicate samples. Error bars rep-resent 1 standard deviation above and below the average CT value.

FIG. 5. Manual and automated RV-PCR results obtained for cleanwipe, air filter, and water samples spiked with Bacillus anthracis Amesspores (samples processed with the manual protocol were spiked with33 2 CFU/sample and samples processed with the automated pro-tocol were spiked with 21 1 CFU/sample, as measured by plating) ina combined background of 103 live Bacillus atrophaeus and 106 Pseu-domonas aeruginosa CFU/sample. Each CT value is the average of 4replicate samples. Error bars represent 1 standard deviation above andbelow the average CT value.

6576 LETANT ET AL. APPL. ENVIRON. MICROBIOL.

may be overwhelmed by the growth of live nontarget organ-isms. The sample extraction buffer contains 30% ethanol inorder to lyse vegetative cells (while spores were shown toremain intact and viable in this buffer) and to reduce compe-tition for growth in the filter cups.

Relationship between incubation time and limit of detection.Three additional RV-PCR experiments were performed in or-der to further explore the relationship between the limit ofdetection of the RV-PCR method and the incubation time.Three spore levels (103, 102, and 101 CFU/sample) were spikedon clean wipes, and samples were processed with the RV-PCRmethod using incubation times ranging from 6 to 8 h. Sevenreplicate samples were analyzed for each set of experimentalconditions. Results from these experiments are summarized inFig. 6. As expected, the shorter the incubation time, the lowerthe DNA concentration and the higher the CT value. The onlyaverage CT values above the detection threshold of 36.0 werethe chromosome assay for the T7 experiment, with the 10-sporelevel; the chromosome and pXO1 assays for the T6 experiment,with the 10-spore level; and the chromosome assay for the T6

experiment, with the 100-spore level. These results stress theneed for enrichment but suggest that depending on the criteriaused to determine whether a sample is positive for viable B.anthracis spores (number of assays and CT threshold) and thedesired limit of detection, the incubation time could potentiallybe reduced to 7 or 6 h when processing relatively clean sam-ples, reducing the total turnaround time of the method to 12 to13 h for a batch of 24 samples. It should be stressed thatgermination and growth characteristics can be affected by de-

contamination treatments (8), and strains may differ in thesecharacteristics; therefore, the longer incubation time of 9 hmight be better suited for unknown strains and environmentalconditions (particularly for post-decontamination sampling)when low limits of detection are required.

It also should be noted that the RV-PCR method describedin this study was developed for qualitative and not quantitativeanalysis. Although experimental results show a qualitative de-crease in CT values with increasing spore levels and a qualita-tive decrease in CT values with increasing incubation times, theCT difference between two consecutive 10-fold dilutions, ortwo consecutive 1-h incubation times does not strictly followtheoretical values based on projected differences in the num-ber of spores in the sample. The lack of quantitation of themethod is tied to sample-to-sample variations in sample spik-ing, spore recovery, and DNA extraction efficiency, as well asto biological variations in germination kinetics, growth kinet-ics, and a potential slight lag in one replicate relative to theother.

The RV-PCR method was tested with B. atrophaeus, andsimilar limits of detection were obtained with clean and dirtysamples (data not shown). Application of the method to veg-etative cells would require changes in the protocol bypassingthe vacuum filtration step and is the object of a separate study.

Conclusions. A verified RV-PCR method was presented fordetection of live, virulent B. anthracis spores in wipe, air filter,and water samples. The method endpoint was shortened fromits initial overnight incubation (14 h) to 9 h by performing amagnetic bead-based DNA extraction and purification proce-

FIG. 6. Manual RV-PCR results obtained for clean wipe samples spiked with 3 levels of Bacillus anthracis Ames (54 9 CFU/sample, 813 43 CFU/sample, and 8,130 430 CFU/sample). The incubation time was varied from 8 to 7 and then 6 h (noted as T8, T7, and T6, respectively).Each CT value is the average of 7 replicate samples. Error bars represent 1 standard deviation above and below the average CT value.

VOL. 77, 2011 RV-PCR METHOD TO DETECT LIVE, VIRULENT B. ANTHRACIS 6577

dure before PCR analysis. Using this method, the total pro-cessing time from start to finish for 24 samples was reduced to15 h (approximately 3 h of processing time should be added foreach additional set of 24 samples when processing samples inseries), which is significantly shorter than the standard platingmethod, which requires up to several days to obtain confirmedresults and requires multiple steps to evaluate the sample ex-tract volume (analysis of a small fraction with the spread platemethod, concentration of approximately one-third of the sam-ple with microfunnel filtration prior to culturing, and perfor-mance of enrichment culturing with the remaining volume),using only a small fraction of the enrichment culture for PCRanalysis if B. anthracis colonies are not observed. In addition,we report the extension of the RV-PCR method to virulentB. anthracis by using 3 specific assays, targeting the chromo-some and each of the two plasmids. Manual and automatedversions of the method showed limits of detection at the 10-spore level with and without debris for all three sample types.The 10-spore level was consistently detected for both manualand automated methods in autoclaved B. anthracis spore back-grounds of 106 spores/samples and live, combined nontargetbackgrounds of 103 B. atrophaeus and 106 P. aeruginosa CFU/sample.

In addition to the shorter turnaround time and lower detec-tion limit, the RV-PCR method also presents operational ad-vantages over the plating method, including a smaller footprint(one tabletop incubator is sufficient to incubate up to 96 sam-ples), a reduction in labor per sample, and a higher throughput(�96 samples may be analyzed in 24 h with a single robot andpersonnel working in successive 8-h shifts).

Based on the advantages described above, the RV-PCRmethod may constitute a new capability to ensure public safetyfor reoccupancy following a biothreat agent release. Otherpotential applications of the method may lie in surveillance,public health, animal health, and food safety, which have sim-ilar needs for accuracy, low limits of detection, high-through-put capacity, and rapid turnaround times.

ACKNOWLEDGMENTS

We thank Elizabeth Vitalis and Marissa Lam for performing the insilico analysis of the PCR assays for this study. We are also grateful toGene Rice, from the EPA’s National Homeland Security ResearchCenter, and Stephanie Harris, from U.S. EPA Region 10, for theircritical review of the manuscript.

This work was performed under the auspices of the U.S. Depart-ment of Energy by Lawrence Livermore National Laboratory underContract DE-AC52-07NA27344.

The U.S. Environmental Protection Agency through its Office ofResearch and Development funded and managed the research de-

scribed here. It has been subjected to the Agency’s administrativereview and approved for publication.

Mention of trade names or commercial products does not constituteendorsement or recommendation for use.

REFERENCES

1. Atlas, R. M. 1996. Handbook of microbiological media, 2nd edition, p. 1267.CRC Press, New York, NY.

2. Bailey, A. M., L. Pajak, I. R. Fruchey, C. A. Cowan, and P. A. Emanuel. 2003.Robotic nucleic acid isolation using a magnetic bead resin and an automatedliquid handler for biological agent stimulants. J. Lab. Autom. 8:113–120.

3. Buttner, M. P., et al. 2004. Determination of the efficacy of two buildingdecontamination strategies by surface sampling with culture and quantitativePCR analysis. Appl. Environ. Microbiol. 70:4740–4747.

4. Francy, D. S., et al. 2009. Comparison of traditional and molecular analyticalmethods for detecting biological agents in raw and drinking water followingultrafiltration. J. Appl. Microbiol. 107:1479–1491.

5. Hodges, L. R., L. J. Rose, A. Peterson, J. Noble-Wang, and M. J. Arduino.2006. Evaluation of a macrofoam swab protocol for the recovery of Bacillusanthracis spores from a steel surface. Appl. Environ. Microbiol. 72:4429–4430.

6. Hodges, L. R., L. J. Rose, H. O’Connell, and M. J. Arduino. 2010. Nationalvalidation study of a swab protocol for the recovery of Bacillus anthracisspores from surfaces. J. Microbiol. Methods 81:141–146.

7. Hoffmaster, A. R. et al. 2006. Characterization of Bacillus cereus isolatesassociated with fatal pneumonias: strains are closely related to Bacillusanthracis and harbor B. anthracis virulence genes. J. Clin. Microbiol. 44:3352–3360.

8. Kane, S. R., et al. 2009. Rapid, high-throughput, culture-based PCR methodsto analyze samples for viable spores of Bacillus anthracis and its surrogates.J. Microbiol. Methods 76:278–284.

9. Kreader, C. A. 1996. Relief of amplification inhibition in PCR with bovineserum albumin or T4 gene 32 protein. Appl. Environ. Microbiol. 62:1102–1106.

10. Letant, S. E., et al. 2010. Most probable number rapid viability PCR methodto detect viable spores of Bacillus anthracis in swab samples. J. Microbiol.Methods 81:200–202.

11. Powder Technology Inc. Materials safety data sheet. Arizona test dust, 2006.Powder Technology Inc., Burnsville, MN. http://www.powdertechnologyinc.com/pages/english/PTI_MSDS.html.

12. Raber, E., A. Jin, K. Noonan, R. McGuire, and R. D. Kirvel. 2001. Decon-tamination issues for chemical and biological warfare agents: how clean isclean enough? Int. J. Environ. Health Res. 11:128–148.

13. Raber, E., et al. 2004. How clean is clean enough? Recent developments inresponse to threats posed by chemical and biological warfare agents. Int. J.Environ. Health Res. 14:31–41.

14. Raber, E., et al. 2002. Chemical and biological agent incident response anddecision process for civilian and public sector facilities. Risk Anal. 22:195–202.

15. Rasko, D. A., et al. 2007. Complete sequence analysis of novel plasmids fromemetic and periodontal Bacillus cereus isolates reveals a common evolution-ary history among the B. cereus-group plasmids, including Bacillus anthracispXO1. J. Bacteriol. 189:52–64.

16. Slezak, T., et al. 2003. Comparative genomics tools applied to bioterrorismdefense. Brief. Bioinform. 4:133–149.

17. Teng, F., G. Yuntao, and Z. Wanpeng. 2008. A simple and effective methodto overcome the inhibition of Fe to PCR. J. Microbiol. Methods 75:362–364.

18. Tsai, Y.-L., and B. Olson. 1992. Rapid method for separation of bacterialDNA from humic substances in sediments for PCR. Appl. Environ. Micro-biol. 58:2292–2295.

19. Varughese, E. A., L. J. Wymer, and R. A. Haugland. 2007. An integratedculture and real-time PCR method to assess viability of disinfectant treatedBacillus spores using robotics and the MPN quantification method. J. Mi-crobiol. Methods 71:66–70.

6578 LETANT ET AL. APPL. ENVIRON. MICROBIOL.