Development of Three Real-Time PCR assays to Detect Bacillus anthracis and Assessment of Diagnostic Utility

Tanya M. Parsons, Victoria Cox, Angela Essex-Lopresti, Margaret G. Hartley, Roman A. Lukaszewski, Phillip A. Rachwal, Helen L. Stapleton and Simon A. Weller^*

Defence Science and Technology Laboratory, Dstl Porton Down, Salisbury, UK

*Corresponding Author:

Simon A. Weller
Defence Science and Technology Laboratory
Dstl Porton Down, Salisbury, UK
Tel: +44 (0) 1980617404
E-mail: sweller@mail.dstl.gov.uk

Received Date: December 01, 2012; Accepted Date: January 18, 2013; Published Date: January 23, 2013

Citation: Parsons TM, Cox V, Essex-Lopresti A, Hartley MG, Lukaszewski RA, et al. (2013) Development of Three Real-Time PCR assays to Detect Bacillus anthracis and Assessment of Diagnostic Utility. J Bioterr Biodef S3:009. doi: 10.4172/2157-2526.S3-009

Copyright: © 2013 Parsons TM, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Visit for more related articles at Journal of Bioterrorism & Biodefense

Abstract

Three real-time PCR assays to detect Bacillus anthracis genetic targets (pXO1; pXO2 and chromosome) were developed. Two of the PCR assays (pXO1-MGB and Ba chr-MGB) were tested against DNA extracts produced from whole blood samples obtained from a replicated B. anthracis murine infection model. Across all three models 45 samples were tested in total, within which a subset of 41 samples were shown to contain B. anthracis by either PCR or microbiological culture. Using microbiological culture as an analogue of conventional blood culture (as used in clinical settings) the detection rates of PCR and blood culture were compared. In two of the murine models blood culture had a significantly higher detection rate than PCR (BA1, p=0.004; BA3, p=0.013). In the BA2 model there was no significant difference between the detection rates of PCR and blood culture.

Keywords

Anthrax; BWA; Medical countermeasures

Introduction

Anthrax is a zoonotic disease caused by Bacillus anthracis. It is a Gram-positive, non-motile, facultative anaerobic spore-forming bacillus. In humans, three types of anthrax infection have been recorded according to the route of infection: cutaneous, gastro-intestinal and inhalation. Inhalational anthrax is rare and usually induced by exposure to spores, either from environmental sources [1] or by bioterrorism [2]. A recent outbreak amongst drug users in the UK has been attributed to heroin contaminated with anthrax spores [3].

The potential for B. anthracis to be used as a biological weapon means there is a requirement for rapid and sensitive identification. Rapid diagnosis of the agent causing a disease enables implementation of suitable medical countermeasures. Delayed application of appropriate antimicrobial therapy [4] is a contributing factor to increased patient mortality [5]. Diagnosis is usually based on bacteriological or serological analysis which can be time consuming. In a recent case of inhalational anthrax [1] a preliminary diagnosis of B. anthracis using blood culture techniques took some 72 hours after the patient presented to hospital.

The Polymerase Chain Reaction (PCR) is commonly used to detect many bacterial pathogens from different sample types [6,7] and can achieve a result within 2 hours. Detection of B. anthracis is often facilitated by PCR assays designed against genetic targets on the virulence associated pXO1 and pXO2 plasmids. However, these plasmids (or elements originating from these plasmids) have been found in other bacillus spp. [8-10] so the provision of a chromosomal assay, specific to B. anthracis, can help distinguish classical B. anthracis strains from other pXO1 and pXO2 harbouring species. In this paper we describe the development of three real-time PCRs to detect different genetic regions of B. anthracis (pXO1, pXO2 and chromosomal targets) and assess the diagnostic utility of two of these PCRs from samples obtained from three B. anthracis murine infection models.

Materials and Methods

Bacterial strains and DNA extraction

Bacillus anthracis and other bacterial strains evaluated in this study are listed in table 1. DNA extracts were either obtained from external collections or bacterial strains were grown on L-agar at 37°C. DNA was then extracted from cultures, using a Qiagen DNAeasy Kit (Qiagen), and retained for PCR analysis. DNA was quantified using the Qubit^® 2.0 Fluorometer (Life Technologies).

Table 1: B. anthracis strains, and close relatives, tested by the real-time PCRs developed in this study.

Identification of unique DNA sequences and assay design

Putative unique gene targets on the pXO1 and pXO2 plasmids were identified in silico by screening individual gene sequences from the B. anthracis A2012 pXO1 and pXO2 plasmid sequences (GenBank Accession Numbers: AE011190.1 and AE011191.1) using the BLAST search facility (https://www.ncbi.nlm.nih.gov/). The target sequence for the chromosomal PCR assay was designed by using the in silico Insignia (https://insignia.cbcb.umd.edu/results.php) genome comparison tool. Unique and conserved B. anthracis chromosomal “signature” sequences identified by Insignia were subjected to an additional BLAST search to confirm specificity. Primer and probe sequences were designed (within pXO1, pXO2 and chromosomal gene sequences) using PrimerExpress Version 2.0 (Applied Biosystems, Foster City, CA). Amplicons for each assay were again BLAST searched to help ensure specificity. PCR primer and probe sequences are summarized in table 2.

Table 2: Nucleotide sequences of PCR assays developed in this study.

Real time PCR

PCR primers and probes were purchased (Life Technologies Corporation). Probes were covalently labelled at the 5`-end with the fluorescent reporter dye FAM and a minor groove binder (MGB) at the 3`-end. PCR names, primer and probe sequences and optimised primer concentrations are listed in table 2. Real-time PCRs (25 μL reaction volume) comprised forward and reverse primers, 1×PCR mastermix (containing 5×10^-2 mol/L Tris-HCl, 5×10^-5 mol/L EGTA, 1 μg/μL BSA, 4×10^-3 mol/L MgCl₂, 0.04 U/μL JumpStart Taq polymerase and 5×10^-4 mol/L dNTPs), sterile water and template DNA or sample extract (10 μL). All PCR reactions were performed on the Smartcycler platform (Cepheid) using a thermocycling profile consisting of 95°C for 3 minutes, followed by 45 cycles of 95°C for 15 s and 60°C for 30 s. Real time PCR results were analysed in terms of cycle threshold value (C_T), the cycle number the background corrected fluorescence exceeds the threshold. A cycle threshold setting of 30 fluorescent units was used for all experiments.

Specificity and performance of real-time PCRs

To confirm specificity each assay was tested against 100 pg of DNA from B. anthracis and other closely related bacterial strains. To evaluate PCR performance each assay was challenged with two putative amounts of DNA (100 fg and 10 fg) obtained from a pure culture of B. anthracis Ames strain. Each assay was challenged with ten replicates of each amount and the same DNA stock was used for each PCR.

B. anthracis murine models

A mouse model of infection was used to evaluate the performance of the PCR assays in clinical samples. Briefly, AJ strain female mice (Harlan Laboratories) were challenged with B. anthracis Russian vaccine strain (STI) which does not harbour the pXO2 plasmid. Challenge inoculums were prepared from a master stock of B. anthracis STI and used to challenge mice via an intraperitoneal (IP) route. Independent 1:60 dilutions of master stock were prepared (to obtain suspensions in a putative range of 1×10⁷ cfu·mL^-1). To determine bacterial load for each challenge, a dilution series from each challenge inoculum was prepared. Each dilution was plated (100 μL aliquots) onto Nutrient Agar (NA) plates and incubated overnight at 37°C. Counts were performed from the most appropriate dilution according to bacterial load. The inoculums (per mouse) for each of the models were; BA1 (18 mice)- 1.72×10⁶ cfu; BA2 (13 mice)–1.76×10⁶ cfu; BA3 (14 mice)-5.45×10⁵ cfu.

Blood samples were harvested from mice culled at time-points 24, 48 and 72 h post-exposure. Humane end points were strictly observed; the BA1 Model mice assigned for a 96 h time-point were culled after 72 h, due to the severity of clinical signs. Blood from each mouse was collected into BD Vacutainer^® collection tubes (Becton Dickinson, Oxford, UK) containing 0.105 molar buffered sodium citrate solution. DNA was extracted from 100 μL volumes of collected blood using a Qiagen DNeasy Blood and Tissue Kit (Qiagen), as per the specific Animal Blood protocol. 10 μL aliquots of the resulting eluents (eluted into 100 μL of AE buffer) were tested by the pXO1-MGB and Ba chr- MGB PCRs described above. Aliquots that tested negative were diluted 1:10 and re-tested to mitigate the effects of potential PCR inhibitors in the extracts.

Direct viability and bacterial load determination from blood samples

Blood samples obtained from each mouse in each model were also analysed by microbiological culture. To indicate the presence of viable B. anthracis cells 2×100 μL aliquots of each blood sample were plated out onto Nutrient Agar (Oxoid) and incubated at 37°C. Bacterial loads were estimated by serially diluting each blood sample (in 1×PBS) followed by cultivation of 3×100 μL aliquots of each dilution.

Results

Specificity and performance of each PCR

Results from each PCR when tested against panels of B. anthracis and other bacterial strains are summarised in table 1. Experimental testing indicated that each assay was specific to its target, i.e. the pXO1 PCR only detected B. anthracis strains harbouring the pXO1 plasmid, and displayed no cross-reaction with any other Bacillus strain. In silico analysis of the pXO1-MGB and pXO2-MGB PCR sequences did indicate the potential for identification of pXO1 or pXO2 like plasmids harboured by non- B. anthracis strains. The implications of this are considered in the Discussion.

Results from performance studies are summarised in table 3. The addition of 100 fg of B. anthracis DNA to each PCR resulted in 100% detection rates (from 10 replicates) allowing C_T values to be analysed. A single-factor ANOVA of mean C_T values obtained when each PCR assay (10 replicates) was challenged with 100 fg of DNA indicated that the means were significantly different (p=8.3×10^-11). Tukey’s Honest Significant Difference (HSD) test was used to evaluate where means were significantly different. A HSD value of 0.61 was calculated (3.354×√(0.344/10)) indicating that the mean C_T value of the pX01 PCR was significantly lower than the pXO2 and chr PCR assays (Difference in means 1.1 and 2.9 respectively), and the pXO2 PCR (with a difference of 1.7) had a significantly lower mean C_T value than the chr PCR (at 95% confidence).

Table 3: Result from limit of detection experiments (purified B. anthracis Ames strain DNA) of each B. anthracis PCR.

At the 10 fg level results were inconsistent with negative results being returned from some, but not all, of the replicates tested by each assay. As there was a ratio of positive to negative results, the C_T values could not be analysed, so the data was treated as binary and a probit model fitted. The probit model showed that there was evidence that both the pX01 PCR and pX02 PCR method gave significantly more positive results than the chr PCR method at the 90% confidence level (p=0.07, both analyses). There was no evidence of a significant difference between the pX01 PCR and pX02 PCR methods.

PCR performance against murine model samples

PCR results from each murine model are summarised in table 4. In all twelve samples with estimated bacterial loads of 10³ cfu•mL^-1 or higher, both the pXO1 and chr PCRs returned positive results from each sample. Samples with estimated bacterial loads below 10³ cfu•mL^-1 returned variable PCR results with at least one PCR positive from 8 of 20 such samples. One sample from the BA3 model (72 h, M3) returned both pXO1 and chr PCR positive results where no viable bacterial cells were detected by blood culture.

Table 4: Blood culture, PCR, and ELISA data generated from blood samples taken
from a replicated murine infection model.

Results generated from bacterial load determination and PCR are summarised in table 5 in a binary format, with a positive result being defined as either, a) at least one bacterial colony observed after blood culture of neat blood aliquots or b) at least one PCR positive replicate, from any assay, per sample. Using this definition, the presence of B. anthracis was determined in 41 of the 45 murine model samples tested in this study, by at least one of the technologies employed.

Table 5: Summary of results from the murine model experiments.

To compare the detection rates of bacterial load determination (as an analogue of conventional blood culture) and PCR, probit models were fitted to the data. The three models (BA1-3) were treated separately. For the BA1 murine model the probit analysis showed evidence of a significant difference between the methods at the 99% confidence level with the bacterial load method giving significantly higher number of positives than the PCR method (p=0.004). The percentage of positives was 83% and 35% respectively. For the BA2 model the probit analysis showed no evidence of a significant difference between the methods. The percentage of positives was 100% and 77% for the bacterial load determination and PCR method respectively. This caveat to this result is that one method (bacterial load determination) had a detection rate of 100% which can affect the robustness of a probit model. For the BA3 model the probit model showed evidence of a significant difference between the methods at the 95% confidence level with the bacterial load determination giving significantly higher number of positives than the PCR method (p=0.013). The percentage of positives was 86% and 38% respectively.

Discussion

In this study we present three real-time PCRs developed for the high confidence detection of Bacillus anthracis. For detection of pXO1 and pXO2 plasmids we purposely tried to design PCR assays away from the lef (pXO1) and cap (pXO2) gene targets often used in PCR detection of this pathogen [11,12], in order to broaden the assay targets available. However, in silico screening (BLAST) of many potential gene targets from both plasmids (as taken from the AE011190.1 and AE011191.1 GenBank Accessions) often showed close homology with other sequenced members of the bacillus genus (data not shown). The design of the chromosomal assay was facilitated by the Insignia genome comparison tool [13] which identified unique chromosomal gene targets present in B. anthracis from all sequenced bacterial genomes. From a selection of identified unique “signature” gene targets the Ba chr- MGB PCR was designed. In silico analysis (BLAST) and experimental testing indicated the specificity of the Ba chr PCR assay. We were not able to train the Insignia tool to perform a similar search using pXO1 and pXO2 gene targets as reference sequences, though a more expert user of this tool might be able to achieve this.

The provision of three separate PCR assays to detect different targets within B. anthracis allows more information to potentially be obtained from a test matrix. pXO1 and pXO2 plasmids and/or elements have been identified within the genomes of other members of the genus bacillus [10]. In silico (BLAST) analysis of the pXO1 and pXO2 PCR amplicon sequences indicated that if tested against the B. cereus G9241 strain [8] then a pXO1 positive would result, as would a pXO2 PCR if tested against the B. cereus (var.) anthracis strain [9]. Both these strains have entire pXO1 and pXO2 like plasmids and have been associated with anthrax like disease symptoms. It has been suggested that such strains be classified as B. cereus/B. anthracis sensu lato [14]. Although we have not been able to test our assays against any of these strains it is possible that the chromosomal assay we have developed may provide an initial or tentative identification of such a strain (if negative) and at least one of the pXO1 or pXO2 PCR assays returned a positive. The chromosomal assay was tested against 14 B. cereus strains (all negative) and 24 B. anthracis strains (all positive) indicating that it does have good specificity within the current classification of B. cereus and B. anthracis. In addition in silico (BLAST) analysis of the chr-MGB PCR amplicon sequence did not indicate the potential for cross reaction with the B. cereus G9241 or B. cereus (var.) anthracis strains (or any other B. cereus/B. anthracis sensu lato member).

From DNA purified from culture of the B. anthracis Ames strain, all PCR assays exhibited limits of detection expected from the PCR technology. With a B. anthracis Ames genome estimated to weigh 5.72 fg [14] 100 fg of DNA per PCR would equate to 17.5 Genome Equivalents (GEs). At this level of DNA (per PCR) all ten replicates of each PCR assay produced positive results. At the 10 fg level (per PCR) each PCR produced inconsistent results. Although PCR has a theoretical limit of detection of 1 target per assay, in reality the stochastic distribution of DNA at low concentrations [15] makes it impossible to define a reproducible Limit of Detection below 3 GEs [16]. Our assays would appear to have reproducible limits of detection somewhere in the range of 100 fg-17 fg (17.5 GEs to 3GEs) per PCR, consistent with other B. anthracis PCRs [17].

From the B. anthracis Ames purified DNA at the 100 fg per PCR level, the pXO1 PCR exhibited significantly lower C_T values than the pXO2 PCR which in turn exhibited significantly lower C_T values than the BA chr MGB- PCR. Unless these PCRs have decreasing efficiencies (from pXO1 through BA chr) then this is most likely due to increased copy numbers per genome of the pXO1 plasmid over copy numbers of the pXO2 plasmid. A previous study [18] has indicated that copy number of the pXO1 plasmid can vary between 33-243 per bacterial cell, with pXO2 copy number varying between 1-32. The results of our study support these findings within the Ames DNA extract we used, also indicating that the pXO2 gene target has a higher copy number than that of the Ba chr PCR target.

Murine model samples were tested to evaluate the clinical utility of two of the PCR assays (pXO1-MGB and Ba chr MGB). We used microbiological culture to both measure the presence of viable bacterial cells within our samples and also as an analogue of blood culture, commonly regarded as a “gold-standard” approach in clinical microbiology, to benchmark the performance of the PCRs. PCR detected all samples with a bacterial load of 10³ cfu•mL^-1 or higher with sporadic positives at the 10² cfu•mL^-1 [μ1] level and one PCR positive where no bacterial load was determined by culture (BA3, 72hr, M3). A bacterial load of 1×10³ cfu•mL^-1 would imply that 1×10² cfu was processed by the DNA extraction method used (from 100 μL starting volume of whole blood). A 10 μL aliquot of the resulting DNA extract introduced into the PCR would therefore contain 10 GEs-assuming all DNA would have been processed and eluted into the extract. Therefore these results are comparable with the Limits of Detection indicated by the sensitivity experiments conducted on DNA extracted from pure bacterial cultures.

In recent UK cases of anthrax [1,3] preliminary indication of B. anthracis infection was obtained after blood culture rather than direct molecular analysis of whole blood. Blood culture can take 72 hours to deliver a preliminary identification of B. anthracis [1]. The murine model samples used in our study were obtained after an intraperitoneal (IP) route of infection with a high initial dose per mouse (10⁵ or 10⁶ cfu). Mice culled at 24 hours had bloods with bacterial loads varying between 10⁴ cfu•mL^-1 [μ2] or below. One mouse culled at 24 hr (BA3, 24hr, M4) did not produce any observable colonies on an agar plate. However, care must be taken when extrapolating results from a murine model of infection to those found in human disease progression. Although we have not been able to find any reported data on titres of B. anthracis in infected human cases, the study of Rossi et al. [19] examined the blood titres of nine African Green Monkeys infected, via an inhalation route, with B. anthracis. The first detected levels of bacteraemia from this primate model were >10³ cfu•mL^-1 (2 cases), 10² cfu•mL^-1 (4 cases) and 10¹ cfu•mL^-1 (2 cases). Bloods were taken at 12 hour intervals after inoculation and in some cases the onset of bacteraemia was not immediately associated with observable disease symptoms.

The results of our study, and those of Rossi et al. [19] suggest that a direct PCR approach could have diagnostic utility in the rapid identification of B. anthracis in whole blood samples and be able to detect all samples with a bacterial titre in the range of 10³ cfu•mL^-1 and above with also some positives at ranges below this level. However, statistical analysis showed that the detection rates of microbiological culture were higher than PCR, when testing the murine model samples. It is possible that with an improved DNA extraction methodology the sensitivities of the B. anthracis PCRs could be improved. We used a silica spin-column method from blood volumes of 0.1 ml. Although a robust limit of detection was provided with this method, there was some evidence of residual PCR inhibitors within some extracts (1:10 dilutions were required to be tested in some instances before a positive result was obtained). A recent report has indicated that an automated, magnetic bead based, DNA extraction system can improve the sensitivity (over the silica column approach) of a real-time PCR based method to detect bacteria associated with sepsis in blood samples [20]. Such an approach, or other methods where DNA is processed from larger sample volumes [21,22], could well increase the amount of agent DNA presented to each assay, and help increase the sensitivity of B. anthracis PCRs. This would also help bridge the gap in sensitivity between blood culture and PCR and allow the best chance of a rapid diagnosis, without a requirement for extended incubation periods.

The practical application of the B. anthracis PCRs, as discussed above, supposes that a clinician would have a clinical indication to order tests specific for B. anthracis. In certain cases, for example a symptomatic animal hide worker [1], enough ancillary information may be present for a clinician to select a B. anthracis test. However, many acute diseases have similar initial “flu-like” symptoms which can hinder diagnosis and also make it difficult to select the appropriate diagnostic tests. Recently the performance of the pXO1 and pXO2 assays described in this paper have been evaluated when deposited within a microfludic array card architecture able to allow one DNA extract to be analysed by multiple PCR assays [23]. Although PCR assays in this format are generally a log unit less sensitive than when used in singleplex format with a magnetic bead DNA extraction approach, a recent study [24] has achieved a 89% detection rate when using Array Card PCRs, compared to the same PCRs in singleplex format, when testing 292 clinical samples for respiratory agents. This study therefore indicates that, with an appropriate DNA extraction method, the gap in sensitivities between Array Card and singleplex PCRs can be bridged. Such cards could therefore allow the B. anthracis PCRs described in this paper to be used in conjunction with assays to multiple other agents capable of causing acute diseases, mitigating the inherent bias against PCR testing and reducing the burden on clinicians in having to decide which assays to select.

Acknowledgements

This study was funded by the UK Ministry of Defence, Programme Office. ^©Crown Copyright 2012. Published with permission of the Defence Science and Technology Laboratory on behalf of the Controller of HMSO.

References

1. Anaraki S, Addiman S, Nixon G, Krahé D, Ghosh R, et al. (2008) Investigations and control measures following a case of inhalation anthrax in East London in a drum maker and drummer, October 2008. Euro Surveill 13.
2. Inglesby TV, Henderson DA, Bartlett JG, Ascher MS, Eitzen E, et al. (1999) Anthrax as a biological weapon: medical and public health management. Working Group on Civilian Biodefense. JAMA 281: 1735-1745.
3. Ramsay CN, Stirling A, Smith J, Hawkins G, Brooks T, et al. (2010) An outbreak of infection with Bacillus anthracis in injecting drug users in Scotland. Euro Surveill 15.
4. Steward J, Lever MS, Simpson AJ, Sefton AM, Brooks TJ (2004) Post-exposure prophylaxis of systemic anthrax in mice and treatment with fluoroquinolones. J Antimicrob Chemother 54: 95-99.
5. Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, et al. (2006) Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 34: 1589-1596.
6. Mitchell JL, Chatwell N, Christensen D, Diaper H, Minogue TD, et al. (2010) Development of real-time PCR assays for the specific detection of Francisella tularensis ssp. tularensis, holarctica and mediaasiatica. Mol Cell Probes 24: 72-76.
7. Rose HL, Dewey CA, Ely MS, Willoughby SL, Parsons TM, et al. (2011) Comparison of eight methods for the extraction of Bacillus atrophaeus spore DNA from eleven common interferents and a common swab. PLoS One 6: e22668.
8. Hoffmaster AR, Ravel J, Rasko DA, Chapman GD, Chute MD, et al. (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.
9. Klee SR, Brzuszkiewicz EB, Nattermann H, Brüggemann H, Dupke S, et al. (2010) The genome of a Bacillus isolate causing anthrax in chimpanzees combines chromosomal properties of B. cereus with B. anthracis virulence plasmids. PLoS One 5: e10986.
10. Luna VA, King DS, Peak KK, Reeves F, Heberlein-Larson L, et al. (2006) Bacillus anthracis Virulent Plasmid pX02 Genes Found in Large Plasmids of Two Other Bacillus Species. J Clin Microbiol 44: 2367-2377.
11. Hadjinicolaou AV, Demetriou VL, Hezka J, Beyer W, Hadfield TL, et al. (2009) Use of molecular beacons and multi-allelic real-time PCR for detection of and discrimination between virulent Bacillus anthracis and other Bacillus isolates. J Microbiol Methods 78: 45-53.
12. Ramisse V, Patra G, Garrigue H, Guesdon JL, Mock M (1996) Identification and characterization of Bacillus anthracis by multiplex PCR analysis of sequences on plasmids pXO1 and pXO2 and chromosomal DNA. FEMS Microbiol Lett 145: 9-16.
13. Phillippy AM, Mason JA, Ayanbule K, Sommer DD, Taviani E, et al. (2007) Comprehensive DNA signature discovery and validation. PLoS Comput Biol 3: e98.
14. Okinaka R, Pearson T, Keim P (2006) Anthrax, but Not Bacillus anthracis? PLoS Pathog 2: e122.
15. Taberlet P, Griffin S, Goossens B, Questiau S, Manceau V, et al. (1996) Reliable genotyping of samples with very low DNA quantities using PCR. Nucleic Acids Res 24: 3189-3194.
16. Wittwer CT, Kusakawa N (2004) Real-time PCR. In: Persing DH, Tenover FC, Versalovic J, Tang JW, Unger ER, Relman DA, White TJ, editors. Molecular microbiology: diagnostic principles and practice: ASM Press Washington 71-84.
17. Bode E, Hurtle W, Norwood D (2004) Real-time PCR assay for a unique chromosomal sequence of Bacillus anthracis. J Clin Microbiol 42: 5825-5831.
18. Coker PR, Smith KL, Fellows PF, Rybachuck G, Kousoulas KG, et al. (2003) Bacillus anthracis virulence in Guinea pigs vaccinated with anthrax vaccine adsorbed is linked to plasmid quantities and clonality. J Clin Microbiol 41: 1212-1218.
19. Rossi CA, Ulrich M, Norris S, Reed DS, Pitt LM, et al. (2008) Identification of a surrogate marker for infection in the African green monkey model of inhalation anthrax. Infect Immun 76: 5790-5801.
20. Regueiro BJ, Varela-Ledo E, Martinez-Lamas L, Rodriguez-Calviño J, Aguilera A, et al. (2010) Automated extraction improves multiplex molecular detection of infection in septic patients. PLoS One 5: e13387.
21. Pel J, Broemeling D, Mai L, Poon HL, Tropini G, et al. (2009) Nonlinear electrophoretic response yields a unique parameter for separation of biomolecules. Proc Natl Acad Sci USA 106: 14796-14801.
22. Störmer M, Kleesiek K, Dreier J (2007) High-volume extraction of nucleic acids by magnetic bead technology for ultrasensitive detection of bacteria in blood components. Clin Chem 53: 104-110.
23. Rachwal PA, Rose HL, Cox V, Lukaszewski RA, Murch AL, et al. (2012) The potential of TaqMan Array Cards for detection of multiple biological agents by real-time PCR. PLoS One 7: e35971.
24. Kodani M, Yang G, Conklin LM, Travis TC, Whitney CG, et al. (2011) Application of TaqMan low-density arrays for simultaneous detection of multiple respiratory pathogens. J Clin Microbiol 49: 2175-2182.