Portal de Periódicos da CAPES

Application of Bacillus anthracis PCR to simulated clinical samples

2003; Elsevier BV; Volume: 9; Issue: 10 Linguagem: Inglês

10.1046/j.1469-0691.2003.00736.x

1469-0691

Kaisu Rantakokko‐Jalava, Matti K. Viljanen,

Bacteriophages and microbial interactions

We evaluated PCR for the detection of Bacillus anthracis DNA from simulated clinical specimens relevant for the microbiological diagnosis of anthrax or exposure to B. anthracis spores. In simulated blood specimens, the lowest limit of detection was 400 CFU per mL of blood, which may be sufficient for samples from patients with septic anthrax. Screening nasal swabs by PCR may not be sensitive enough to rule out dangerous exposure to anthrax spores, as a minimum of 2000 spores per sample was required for detectable amplification. As spores survived some standard DNA purification methods, special attention should be paid to laboratory safety when preparing samples possibly containing live spores. We evaluated PCR for the detection of Bacillus anthracis DNA from simulated clinical specimens relevant for the microbiological diagnosis of anthrax or exposure to B. anthracis spores. In simulated blood specimens, the lowest limit of detection was 400 CFU per mL of blood, which may be sufficient for samples from patients with septic anthrax. Screening nasal swabs by PCR may not be sensitive enough to rule out dangerous exposure to anthrax spores, as a minimum of 2000 spores per sample was required for detectable amplification. As spores survived some standard DNA purification methods, special attention should be paid to laboratory safety when preparing samples possibly containing live spores. The tragic incidents in the USA in October 2001 highlighted the need for rapid diagnostic tests for the detection of bioweapons, such as the anthrax-causing bacterium Bacillus anthracis. Because of their rapidity, specificity and relative ease of use for various targets and sample types, PCR-based methods are among the most important for this purpose. PCRs have mainly been used in differentiating B. anthracis strains from the closely related species of the B. cereus group. The most important distinctive features of fully virulent B. anthracis strains are two plasmids (pXO1 and pXO2) containing the genes encoding anthrax toxin (lef, cya, pag; pXO1) and structures of the capsule (cap; pXO2). These plasmids have been used as B. anthracis-specific PCR targets [1Ramisse V Patra G Garrigue H Guesdon JL Mock M Identification and characterization of Bacillus anthracis by multiplex PCR analysis of sequences on plasmids pXO1 and pXO2 and chromosomal DNA.FEMS Microbiol Lett. 1996; 145: 9-16Crossref PubMed Google Scholar, 2Makino SI Cheun HI Watarai M Uchida I Takeshi K Detection of anthrax spores from the air by real-time PCR.Lett Appl Microbiol. 2001; 33: 237-240Crossref PubMed Scopus (92) Google Scholar, 3Bell CA Uhl JR Hadfield TL et al.Detection of Bacillus anthracis DNA by LightCycler PCR.J Clin Microbiol. 2002; 40: 2897-2902Crossref PubMed Scopus (130) Google Scholar]. We evaluated some published PCR assays for the detection of B. anthracis DNA from simulated clinical specimens relevant to the microbiological diagnosis of anthrax or exposure to B. anthracis spores. Special attention was paid to real-time PCR methods employing the Lightcycler (LC) instrument (Roche Diagnostics, Mannheim, Germany), as these methods provide results within 1–2 h. All experiments were done with B. anthracis strains ATCC (American Type Culture Collection, Manassas, Virginia, USA) 4229 (pXO1–/pXO2+) and Sterne 7702 (pXO1+/pXO2–). These nonpathogenic strains were kindly provided by the Swedish Institute for Infectious Disease Control, Stockholm. Normal BSL2 practices were followed in handling the strains [4Klietmann WF Ruoff KL Bioterrorism. implications for the clinical microbiologist.Clin Microbiol Rev. 2001; 14: 364-381Crossref PubMed Scopus (134) Google Scholar]. Cells were suspended in NaCl, and colony-forming units (CFUs) were enumerated by the standard dilution plating method. Spore suspensions were prepared by culturing the Sterne strain on sporulating medium [5Williams DJ Franklin JG Chapman HR Clegg LFL Methods of assessing the sporidical efficiency of an ultra-high-temperature milk sterilizing plant. I. Experiments with suspensions of spores in water.J Appl Bacteriol. 1957; 20: 43-49Crossref Scopus (17) Google Scholar] for 3 days, harvesting the spores in 20 mL of sterile distilled water, and syringing the suspension through a 0.8-μm filter (Minisart, Sartorius, Göttingen, Germany) to remove any vegetative cells. The spores were collected by centrifugation (10 000g, 20 min), heated at 56 °C for 60 min, washed with sterile distilled water, and suspended in 0.5% sodium dodecylsulfate (SDS). Microscopy was used to confirm the purity of the spore preparation. The number of heat-resistant CFUs was determined as for the vegetative cells. DNA was isolated by standard phenol extraction [6Rantakokko-Jalava K Jalava J Development of conventional and real-time PCR assays for detection of Legionella DNA in respiratory specimens.J Clin Microbiol. 2001; 39: 2904-2910Crossref PubMed Scopus (78) Google Scholar], by use of the High Pure PCR Template preparation kit (Roche Diagnostics) or the guanidine thiocyanate–glassmilk method [7Ibrahim A Norlander L Macellaro A Sjostedt A Specific detection of Coxiella burnetii through partial amplification of 23S rDNA.Eur J Epidemiol. 1997; 13: 329-334Crossref PubMed Scopus (17) Google Scholar]. Some experiments with the spore suspensions included additional lysis steps carried out using the Mini-beadbeater and 0.1-mm-diameter glass beads (Biospec Products, Bartlesville, Oklahoma, USA) before DNA extraction. Limited germination by incubation of the spore suspensions in fetal calf serum (FCS) at 37 °C for 1 h was also used to improve the sensitivity of spore detection [8Titball RW Manchee RJ Factors affecting the germination of spores of Bacillus anthracis.J Appl Bacteriol. 1987; 62: 269-273Crossref PubMed Scopus (50) Google Scholar]. The ability of the DNA isolation methods to inactivate spores was studied by plating 1 μL of the PCR-ready DNA preparations from the two most concentrated suspensions (initially containing 20 000 spores/μL or 2000 spores/μL) on blood agar and counting bacillus-like colonies after overnight cultivation at 37 °C. Table 1 contains a list of the primers used in the PCRs. The conventional PCRs were performed as simplex 25-μL reactions with each primer at 0.8 μm, MgCl2 at 3 mm, each dNTP (Promega) at 0.2 mm and 1 U of DNA polymerase (Dynazyme, Finnzymes, Espoo, Finland, or AmpliTaq Gold, Applied Biosystems, Foster City, California, USA) with appropriate buffer, and 1 μL of the template DNA. Initial denaturation at 95 °C for 2 min (or 10 min for AmpliTaq Gold) was followed by 30 thermal cycles of 30 s at 94 °C, 30 s at 60 °C and 60 s at 72 °C, and a final extension step at 72 °C for 10 min. Five microliters were run on 2% agarose gel for the detection of PCR products. The 20-μL LC PCR reactions contained each primer at 0.5 μm, hybridization probes at 0.15 μm when used, MgCl2 at 3 mm, 1 U of uracil DNA glycosylase, and 2 μL of Lightcycler FastStart Reaction mix for SYBR Green I or hybridization probe detection (Roche Diagnostics). The hybridization probes and reaction parameters for rpoB PCR were as described in Qi et al [9Qi Y Patra G Liang X et al.Utilization of the rpoB gene as a specific chromosomal marker for real-time PCR detection of Bacillus anthracis.Appl Environ Microbiol. 2001; 67: 3720-3727Crossref PubMed Scopus (140) Google Scholar]. The hybridization probes for the detection of Ba 813 (BantFL 5′-TGA GTC TTT TAA TGC CAG GTT CTA TAC CGT-X and BantLC 640-TCA GCA AGC TAT TCC AAG TCA GAA AGA AA-p) were obtained from Tibmolbiol (Berlin, Germany), where they had been designed previously. Amplification was performed with standard FastStart parameters with 50 cycles and an annealing temperature of 60 °C. All runs included 50 ng of purified B. anthracis DNA as a positive control, and sterile distilled water as negative isolation and reagent controls.Table 1Detection limits for PCR assays evaluated in this workLimit of detection (CFU/μL)bLimits of detection were determined as the lowest number of CFU/μL in 200-μL specimens (PBS) with detectable amplification on at least two separate runs. The amount of DNA used in each reaction was 1 μL. As the phenol and High Pure extraction protocols include no concentration, the figures also represent the number of cells from which DNA was amplified in each reaction. In the glassmilk procedure, DNA was eluted in 1/10 of the original sample volume. The variation shown is due to differences between the results obtained from the two strains used, with the Sterne strain usually giving a 10-fold detection limit.TargetPrimersLength of amplicon (bp)PCRaConv., conventional PCR; LC-SYBR, Lightcycler-PCR with use of SYBR Green I; LC-hybr, Lightcycler-PCR with use of hybridization probes.DNA extraction methodCellsSporesReferenceBa813R1/R2152Conv.Phenol20–200≥2000[1Ramisse V Patra G Garrigue H Guesdon JL Mock M Identification and characterization of Bacillus anthracis by multiplex PCR analysis of sequences on plasmids pXO1 and pXO2 and chromosomal DNA.FEMS Microbiol Lett. 1996; 145: 9-16Crossref PubMed Google Scholar]High Pure200–200020 000Glassmilk2–202000LC-SYBRPhenol2000≥2000High Pure0.2–2≥2000Glassmilk0.2–20200LC-hybrPhenolNot done20This workHigh Pure0.2–22000Glassmilk0.02–0.220rpoBrpoBF1a/BR1a175LC-hybrPhenolNot done200[8Titball RW Manchee RJ Factors affecting the germination of spores of Bacillus anthracis.J Appl Bacteriol. 1987; 62: 269-273Crossref PubMed Scopus (50) Google Scholar]High Pure0.2–22000Glassmilk0.02–0.220pagPA6/PA7210Conv.Phenol2020 000[2Makino SI Cheun HI Watarai M Uchida I Takeshi K Detection of anthrax spores from the air by real-time PCR.Lett Appl Microbiol. 2001; 33: 237-240Crossref PubMed Scopus (92) Google Scholar]High Pure20020 000Glassmilk202000lefax-59/60cThe sequence of primer 59 used in this study differs from that published in Ramisse et al. [1], The first 'G' in the original primer sequence has been omitted, and a 'T' is included as the last base.992Conv.Phenol2000Not done[1Ramisse V Patra G Garrigue H Guesdon JL Mock M Identification and characterization of Bacillus anthracis by multiplex PCR analysis of sequences on plasmids pXO1 and pXO2 and chromosomal DNA.FEMS Microbiol Lett. 1996; 145: 9-16Crossref PubMed Google Scholar]High Pure2000Not doneGlassmilk200Not donecapMO11/M012572Conv.Phenol20Not done[2Makino SI Cheun HI Watarai M Uchida I Takeshi K Detection of anthrax spores from the air by real-time PCR.Lett Appl Microbiol. 2001; 33: 237-240Crossref PubMed Scopus (92) Google Scholar]High Pure20Not doneGlassmilk20Not donecapax17/ax20872Conv.Phenol200Not done[1Ramisse V Patra G Garrigue H Guesdon JL Mock M Identification and characterization of Bacillus anthracis by multiplex PCR analysis of sequences on plasmids pXO1 and pXO2 and chromosomal DNA.FEMS Microbiol Lett. 1996; 145: 9-16Crossref PubMed Google Scholar]High Pure200Not doneGlassmilk20Not donea Conv., conventional PCR; LC-SYBR, Lightcycler-PCR with use of SYBR Green I; LC-hybr, Lightcycler-PCR with use of hybridization probes.b Limits of detection were determined as the lowest number of CFU/μL in 200-μL specimens (PBS) with detectable amplification on at least two separate runs. The amount of DNA used in each reaction was 1 μL. As the phenol and High Pure extraction protocols include no concentration, the figures also represent the number of cells from which DNA was amplified in each reaction. In the glassmilk procedure, DNA was eluted in 1/10 of the original sample volume. The variation shown is due to differences between the results obtained from the two strains used, with the Sterne strain usually giving a 10-fold detection limit.c The sequence of primer 59 used in this study differs from that published in Ramisse et al. [1Ramisse V Patra G Garrigue H Guesdon JL Mock M Identification and characterization of Bacillus anthracis by multiplex PCR analysis of sequences on plasmids pXO1 and pXO2 and chromosomal DNA.FEMS Microbiol Lett. 1996; 145: 9-16Crossref PubMed Google Scholar], The first 'G' in the original primer sequence has been omitted, and a 'T' is included as the last base. Open table in a new tab The analytic sensitivities of the PCRs were first determined with 10-fold dilution series of B. anthracis cells and spores in phosphate-buffered saline (PBS). DNA was isolated from 200 μL of the suspensions containing a total of 4 to 4 × 106 CFU according to the protocols described above. The detection limit was determined as the most dilute specimen reproducibly yielding an appropriate PCR product, as judged by the presence of a clear-cut band of expected size on the 2% agarose gel (conventional PCR), detectable amplification and melting temperature of the amplified product within 1 °C of that of the positive control (LC with SYBR Green I), or amplification visible as fluorescence resonance energy transfer (FRET) signals (LC with hybridization probes). Simulated blood specimens were prepared by spiking 500-μL aliquots of normal human whole blood with 100 μL of B. anthracis cell suspension containing 20 to 2 × 106 CFU [10Loeffler J Henke N Hebart H et al.Quantification of fungal DNA by using fluorescence resonance energy transfer and the light cycler system.J Clin Microbiol. 2000; 38: 586-590PubMed Google Scholar]. Of the mixture, 500 μL was treated according to the High Pure protocol, and 100 μL according to the glassmilk protocol. Nasal swabs were obtained with a cotton swab from a healthy volunteer, and inserted in 200 μL of B. anthracis spore suspensions containing 2 to 2 × 106 heat-resistant CFUs. After incubation for 15 min at room temperature, the swabs were removed from the spore suspension, and DNA was isolated from them according to the phenol protocol, the glassmilk protocol with or without beadbeating, or, the High Pure protocol. The specificity of amplification was evaluated with 100 ng of DNA from B. cereus ATCC 11778, B. thuringiensis CCUG (Culture Collection of University of Gothenburg, Gothenburg, Sweden) 22499, and 11 other Bacillus strains (B. circulans ATCC 4314, B. coagulans ATCC 7050, B. megaterium ATCC 14581, B. pumilus ATCC 14884, B. sphaericus ATCC 14577, B. macerans ATCC 8244, and five local Bacillus sp. clinical isolates). In addition, DNA from 24 other bacterial species possibly present in the human respiratory tract was amplified [6Rantakokko-Jalava K Jalava J Development of conventional and real-time PCR assays for detection of Legionella DNA in respiratory specimens.J Clin Microbiol. 2001; 39: 2904-2910Crossref PubMed Scopus (78) Google Scholar]. The specificity test panel included DNA isolated from 72 human clinical specimens. Of these, seven originated from the skin, 27 from the nasopharynx or pharynx, and 38 from the lower respiratory tract, four being sputa, 30 bronchoalveolar lavage (BAL) specimens, and four lung biopsy specimens. Table 1 shows the detection limits obtained for suspensions of B. anthracis cells and spores in NaCl. Although PCR targeting the B. anthracis cap gene with primers MO11 and MO12 was originally described as an LC assay [2Makino SI Cheun HI Watarai M Uchida I Takeshi K Detection of anthrax spores from the air by real-time PCR.Lett Appl Microbiol. 2001; 33: 237-240Crossref PubMed Scopus (92) Google Scholar], we were unable to amplify DNA of B. anthracis ATCC 4229 reproducibly with this PCR. The plasmid-targeted PCRs with primers Ax59 and Ax60, and Ax17 and Ax20, were not applicable to the LC either, probably because of the long amplicons. For chromosomal targets, LC-PCR with FRET detection showed the lowest detection limits. Glassmilk purification was superior in sensitivity to High Pure and phenol extractions, perhaps because of the 10-fold concentration of DNA during this process. In comparison with vegetative cells, 10–1000 times more spores were required for a positive PCR result. Beadbeating for 1 min before DNA extraction from spores improved the detection limit by one dilution in glassmilk extraction, but not in phenol extraction. Germination did not improve the sensitivity of detection, although it reduced by 70% the number of heat-resistant CFUs. This was also true of the glassmilk and High Pure procedures, both of which should effectively wash out the inhibitory factors possibly present in FCS. Table 2 shows the detection limits obtained from simulated clinical specimens. In human blood, 400–4000 CFU/mL could be detected after High Pure purification and glassmilk purification, when the target was rpoB. With conventional PCR, 4 × 106 CFU per mL of blood could be detected by PCRs targeting Ba813 or lef (but not pag). The sensitivity of the LC assay may be sufficient for samples from patients with septic anthrax, as the bacterial burden in the blood of these patients can be as high as 108 CFU/mL [11Inglesby TV O'Toole T Henderson DA et al.Anthrax as a biological weapon, 2002: updated recommendations for management.JAMA. 2002; 287: 2236-2252Crossref PubMed Scopus (849) Google Scholar]. Gram-staining remains the most rapid means for a presumptive diagnosis, but PCR may be used to confirm the presence of B. anthracis in the circulation, especially if antibiotics have been administered before sampling.Table 2Sensitivity of evaluated Bacillus anthracis PCR assays in simulated clinical specimensLimit of detection inTargetPrimersPCRaConv., conventional PCR; LC-hybr, Lightcycler-PCR with use of hybridization probes.DNA extraction methodSimulated blood samples (CFU per mL of blood)Simulated nasal swabs (spores per sample)bNumber of spores determined as number of heat-resistant CFUs.Ba813R1/R2Conv.Phenol–etherNot tested2 × 106High Pure4 × 1062 × 106Glassmilk4 × 1062 × 104Beadbeating + glassmilkNot tested2 × 105rpoBrpoBF1a/BR1aLC-hybrPhenol–etherNot tested2 × 104High Pure4002 × 103Glassmilk400–4000cThe variation shown is due to differences between the results obtained from the two strains used, with the Sterne strain giving a 10-fold detection limit.2 × 103Beadbeating + glassmilkNot tested2 × 103lefax-59/60dSee footnote in Table 1 for modifications in the primer sequence.Conv.Phenol–etherNot tested>2 × 106High Pure>4 × 1062 × 106Glassmilk4 × 1062 × 104Beadbeating + glassmilkNot tested2 × 105pagPA6/PA7Conv.Phenol–etherNot tested2 × 106High Pure>4 × 1062 × 106Glassmilk>4 × 1062 × 104Beadbeating + glassmilkNot tested2 × 105a Conv., conventional PCR; LC-hybr, Lightcycler-PCR with use of hybridization probes.b Number of spores determined as number of heat-resistant CFUs.c The variation shown is due to differences between the results obtained from the two strains used, with the Sterne strain giving a 10-fold detection limit.d See footnote in Table 1 for modifications in the primer sequence. Open table in a new tab In the simulated nasal swabs, 2000 anthrax spores were detectable by the LC assay with hybridization probes after glassmilk and High Pure purifications, and 2 × 104 CFU after phenol purification. With conventional PCR, 2 × 104 CFU could be detected after glassmilk purification, and 2 × 106 CFU after phenol and High Pure purifications. Beadbeating did not improve detection in these experiments. The sensitivity of real-time PCR in detecting spores is consistent with that of a recent report [12Oggioni MR Meacci F Carattoli A et al.Protocol for real-time PCR identification of anthrax spores from nasal swabs after broth enrichment.J Clin Microbiol. 2002; 40: 3956-3963Crossref PubMed Scopus (39) Google Scholar]. Since the infective inhalation dose for humans is estimated to range from hundreds to thousands of spores [13Watson A Keir D Information on which to base assessments of risk from environments contaminated with anthrax spores.Epidemiol Infect. 1994; 113: 479-490Crossref PubMed Scopus (79) Google Scholar] or even less [14Peters CJ Hartley DM Anthrax inhalation and lethal human infection.Lancet. 2002; 359: 710-711Abstract Full Text Full Text PDF PubMed Scopus (28) Google Scholar], the screening of nasal swabs by PCR may not be sensitive enough to rule out dangerous exposure to anthrax spores. Recently, a protocol for the broth enrichment of anthrax spores before real-time PCR detection was described and may help overcome the problem [12Oggioni MR Meacci F Carattoli A et al.Protocol for real-time PCR identification of anthrax spores from nasal swabs after broth enrichment.J Clin Microbiol. 2002; 40: 3956-3963Crossref PubMed Scopus (39) Google Scholar]. The estimated infective dose of cutaneous anthrax is approximately 10 spores [13Watson A Keir D Information on which to base assessments of risk from environments contaminated with anthrax spores.Epidemiol Infect. 1994; 113: 479-490Crossref PubMed Scopus (79) Google Scholar], which is far below the detection limits of the PCR assays tested here. All tested purification methods effectively inactivated the vegetative cells. However, phenol extraction was the only DNA purification procedure that constantly inactivated B. anthracis spores, with cultures of all eight tested preparations from four separate purifications remaining negative. Cultures of DNA preparations from three separate High Pure purifications yielded no growth, but after the fourth purification, a 1-μL culture of PCR-ready DNA from the most concentrated suspension (originally containing 20 000 spores/μL) yielded 15 colonies, and that from the 10-fold dilution yielded two colonies. Inactivation of spores was least effective with guanidine thiocyanate–glassmilk purification, since more than 100 CFUs were cultured from seven of eight DNA preparations from the strongest suspension, the DNA preparation from the next dilution growing about 50 colonies. Heating the sample (95 °C, 5 min) in the lysis buffer inactivated the spores in all three experiments performed. These findings have important implications for laboratory safety, as purified DNA should be handled as potentially infectious until a laboratory has confirmed the inactivation of spores during DNA preparation. Both LC assays with FRET probe detection proved very specific. False-positive results were obtained from only one BAL specimen and two nasopharyngeal samples with assays targeting Ba813 and rpoB, respectively. In these cases, the threshold cycles were >40, and a positive result was never reproducible after new extraction. Thus, cross-contamination is a possible explanation for false-positive results. Ba813-targeted LC-PCR with SYBR Green I detection yielded non-reproducible false-positive amplification from four BAL specimens, with the melting peak differing from that of the positive control by less than 1 °C in two samples and by more than 1 °C in two samples. Nonspecific amplification with primers PA6 and PA7 was detected by SYBR Green I in 12 of 21 nasopharyngeal samples, all producing several melting peaks, making the LC-PCR assay with these primers unsuitable for the screening of B. anthracis DNA in this kind of sample. The conventional PCR amplification of DNA from several bacterial species and clinical samples by means of primers PA6 and PA7 and Dynazyme enzyme yielded several bands of different sizes, which did not appear when AmpliTaq Gold enzyme 'hot start' was used. Similarly, the amplification of DNA from several Bacillus species with primers MO11 and MO12 yielded multiple short bands, if PCR was performed without the 'hot' start procedure. These experiments suggest that the amplification of chromosomal targets rpoB or Ba813 by LC with FRET detection can be applied to the search for B. anthracis DNA in the blood of severely ill patients, as well as in other specimens in which the bacilli are in a vegetative form. Neither of these targets is absolutely specific for B. anthracis, and the closely related B. cereus isolates at least may give amplification signals in these assays [9Qi Y Patra G Liang X et al.Utilization of the rpoB gene as a specific chromosomal marker for real-time PCR detection of Bacillus anthracis.Appl Environ Microbiol. 2001; 67: 3720-3727Crossref PubMed Scopus (140) Google Scholar]. Confirmation of the results by the plasmid-targeted conventional PCR may raise problems associated with its inferior sensitivity. The failure to confirm an rpoB- or Ba813-positive result may be due to the absence of virulence factors in the Bacillus strain yielding a positive signal in LC-PCR, or too low an amount of B. anthracis DNA in the sample. Better sensitivity may be achieved using the LC Bacillus anthracis detection kit (Roche Diagnostics) [3Bell CA Uhl JR Hadfield TL et al.Detection of Bacillus anthracis DNA by LightCycler PCR.J Clin Microbiol. 2002; 40: 2897-2902Crossref PubMed Scopus (130) Google Scholar]. The assay is based on the amplification of pagA and capB in separate reactions monitored with FRET probes. Unfortunately, we did not have the opportunity to test this assay in our experiments. As for the potential exposure of the public to anthrax spores, analysis of surface swabs, etc. is probably the most useful approach to the assessment of risks and determination of required interventions. Recommended protocols for sampling have recently been published [15Sanderson WT Hein MJ Taylor L et al.Surface sampling methods for Bacillus anthracis spore contamination.Emerg Infect Dis. 2002; 8: 1145-1151Crossref PubMed Scopus (71) Google Scholar]. Such samples were not studied here. In our experience, less than 0.5% of clinical samples remain inhibitory to PCR after High Pure purification [6Rantakokko-Jalava K Jalava J Development of conventional and real-time PCR assays for detection of Legionella DNA in respiratory specimens.J Clin Microbiol. 2001; 39: 2904-2910Crossref PubMed Scopus (78) Google Scholar,16Rantakokko-Jalava K Jalava J Optimal DNA isolation method for detection of bacteria in clinical specimens by broad-range PCR.J Clin Microbiol. 2002; 40: 4211-4217Crossref PubMed Scopus (100) Google Scholar]. However, for reliable negative results, an inhibition control should be included in the diagnostic protocol. Once more, our findings underline the need for extensive validation before applying 'old' molecular assays to new settings. In the context of preparedness for bioterrorism, it may be frustrating to expend effort, time and money in developing tests that, we hope, will never be needed. The data included in this manuscript were presented in part in an abstract for the 19th Meeting of the Scandinavian Society for Antimicrobial Chemotherapy in Turku, Finland, 30 August to 1 September 2002. Ulla Eriksson at the Swedish Defence Research Agency and the members of 'Turku Task Force' (Heikki Arvilommi, Erkki Eerola, Jari Jalava, Raija Manninen, Mikael Skurnik and Hanna Soini) are thanked for useful discussions about the methods applied in this work. We thank Kaisa Leppänen and Tiina Haarala for excellent technical assistance. Simo Merne, MA, is acknowledged for language revision of this manuscript. This work was financed by a grant from the National Public Health Institute of Finland.