Detection of low numbers of Bacillus anthracis spores in three soils using five commercial DNA extraction methods with and without an enrichment step


  • Present address
    Antonio J. Luna, CH2M Hill Inc., 4350 W. Cypress St., Suite 600, Tampa, FL 33607-4178, USA.

  • J.S. Gulledge and V.A. Luna contributed equally to this paper.

Vicki Ann Luna, Center for Biological Defense, University of South Florida, College of Public Health, 3602 Spectrum Blvd., Tampa, FL 33612, USA. E-mail:


Aims:  To (i) compare the limits of detection of Bacillus anthracis spores in three soils (one Florida, one Texas, and one a commercial Garden product) by PCR using DNA extracted with five commercial extraction kits and (ii) examine if removing organic acids or adding an enrichment step utilizing a growth medium will improve the detection limits.

Methods and Results: Bacillus anthracis spores were added to soil aliquots and used immediately with a DNA extraction kit or pretreated to remove organics or incubated overnight in a selective growth medium before the DNA extraction was performed. Using hybridization and PCR assays for capC, pag and lef genes, 105–106B. anthracis spores were detected in untreated Florida soil, 104–107 spores in untreated Texas soil and 106–107 in Garden soil. Pretreatment did not reliably improve detection. DNA from untreated and pretreated soils was suitable for hybridization but not always for PCR. When 101–102 spores were added to the soils and allowed to amplify in a growth medium selective for B. anthracis, DNA extracted using four methods reliably produced PCR acceptable DNA positive for the B. anthracis genes.

Conclusions:  The quality of DNA extracted with commercial kits appears to be influenced by the soil type and pretreatment. Yet, with an enrichment step added, four of five extraction methods produced PCR suitable DNA and detected ≤102 spores.

Significance and Impact of the Study:  The enrichment step could enhance the detection of B. anthracis spores in soils and small samples contaminated with soil.


Bacillus anthracis, long recognized as a deadly mammalian pathogen, realized its potential as a bio-weapon when spores were purposefully released and disseminated in 2001. Since then a number of molecular tests and sample preparation methods have been developed to aid in detecting B. anthracis spores in clinical specimens and environmental samples such as powders (Christensen et al. 2006; Drago et al. 2002; Luna et al. 2003). However, soil remains a challenging sample type for the detection of B. anthracis because it is a difficult matrix from which to extract DNA that is usable for molecular methods such as PCR. Previous nonculture methods of quick detection of B. anthracis in environmental samples have fallen short of promised detection limits (King et al. 2003). Soils and sediments vary greatly in physical characteristics and in chemical and organic composition (Ellis et al. 2003). Some compounds such as humic and fulvic acids inhibit restriction enzymes and Taq polymerases (Jacobsen and Rasmusssen 1992; Tebbe and Vahjen 1993; Tsai and Olson 1992). Other characteristics such as clay composition and amount of hydrophobic particles may sequester the spores and obviate their detection (Doyle et al. 1984). In addition, the bacterial, fungal, viral and plant DNA load in the soil can be very high and thus mask the presence of spore DNA in the sample (Kuske et al. 1998).

There are several methods that have been developed to extract DNA from soil, but they can be time-consuming, labour intensive, or use large volumes of soil (Felske et al. 2003; Jacobsen and Rasmusssen 1992; Kuske et al. 1998, 2002; Mergel et al. 2001). Some of these manual methods and principles have been incorporated into commercially available kits and automated systems because a simple procedure was needed that would provide a rapid and sensitive detection of pathogens in a small soil sample. These new commercial kits were developed specifically for soil and have been used in studying soil bacterial communities (Dunbar et al. 1999, 2002). However, it is not clear if the different methods can detect low amounts of B. anthracis spores in different soil types.

Because the soil samples for the commercial kits normally require much lower amounts of soil than the older methods, the probability of missing low numbers of B. anthracis spores increases. Intuitively it is clear that 0·5 g of soil would have many less spores than would 5 or 15 g. Thus, it is probable that the small sample size required by the commercial methods precludes the ability of the researcher to detect B. anthracis spores when the spores are extremely low in number. Yet even larger amounts of soil do not guarantee the detection of B. anthracis because of the increased amount of other bacteria in the soil. However, it is important to detect low numbers of spores in a sample. For example, when first responders take environmental and soil samples in the vicinity of a potential bio-threat, the samples taken are typically very small and usually obtained on swabs. Therefore, detection limits need to be improved.

Culture is considered by many to be the most sensitive method for the detection of B. anthracis in soils, yet it can be time-consuming and labour intense. This laboratory previously reported that the use of a modified selective medium improved the selection of B. anthracis in soil cultures (Luna et al. 2009). Low numbers of B. anthracis spores germinated and increased by 6–7 log10 in the broth version of the medium when incubated at 30°C for 24 h. Thus, we hoped that this medium could be useful in detecting very low numbers of B. anthracis spores in DNA extracted from soils.

The purpose of this study was to compare 1) the suitability of DNA extracted by five commercial extraction methods (manual and automated) for PCR assays and 2) the detection limits of the five methods in the detection of B. anthracis spores in three soils before and after an enrichment step utilizing a selective medium.

Materials and methods

Soils and sampling sites

Three different soils were used to examine the different DNA extraction methods. One soil sample was taken at a site located on the grounds of the University of South Florida at Tampa, Florida (GPS location: N28°03·475′; W82°25·203′). A top layer of soil covering 12 × 15 cm in area and 2·5 cm in depth was removed from the test site to avoid contaminants and results of ultraviolet (UV) action on the exposed soil. A second layer (10 × 12·5 cm) of soil of equal depth was removed and placed into a plastic Nalgene bottle (catalogue number 02-893C) (Fisher Scientific, Inc., Pittsburg, PA, USA). This layer was expected to have the highest concentration of DNA compared to deeper layers (Mergel et al. 2001). A second soil sample was obtained in Texas (GPS location: N32°7·404′; W102°27·776′). The third soil sample was obtained from a commercially prepared garden soil purchased from a local Tampa area home improvement store. The soils were vigorously shaken to mix them thoroughly. Each sample weighed between 70·8–73·5 g (average 72·1 ± 1·35). A portion 20–25 ml (35–36 g) of each soil was analysed for its physical characteristics. Samples were described utilizing the Unified Soil Classification System (USCS) (American Society for Testing and Materials, 1985). Sample testing included a visual inspection on a contrasting surface, and a hydrometer test to help in determining the approximate volumes of each grain size. Grain Size Analysis using sieves was not an option at the time of testing, so percentages given are close approximations using field expedient hydrometer methods (Gee and Bauder 1979).

Soil cultures

Aliquots of each of the three initial soils (1–5 g) were cultured as previously described (Dragon et al. 2001; Dunbar et al. 1999; Luna et al. 2009). Bacillus anthracis was never isolated in any culture attempts. Aliquots of soil (0·1–1 g) were tested with or without the pretreatment steps described later. Dot blots made from crude soil extracts were also consistently negative in DNA–DNA hybridization assays for target genes (lef, pag, and capC) of B. anthracis plasmids pX01 and pX02 (Luna et al. 2006). Therefore, the soils were considered negative for B. anthracis and safe for use in this study. The soil aliquots that later had B. anthracis spores added to them and were to be used on automated instruments were autoclaved before testing to preclude contamination of the instruments.

Bacterial spores

Bacillus anthracis Pasteur (CDC BC 3233) spores of known concentration (108 CFU ml−1) in saline were obtained from the Florida Department of Health (FDOH). For spore preparations of B. anthracis Sterne (BB001) also procured from FDOH, bacteria were grown on modified nutrient sporulation medium including phosphate (NSMP) (Atlas 1993) for 48–72 h at 30°C or until 90–99% phase-bright spores were observed by phase-contrast light microscopy. The balance of the cells seen were predominantly phase dark spores being in the early stages of germination or dead ‘ghost’ vegetative cells. The spores with germinating and vegetative cells were harvested into microcentrifuge tubes, heat shocked at 65–80°C for 30 min to kill the vegetative cells, washed with 1 ml saline, vortexed hard for five min, and centrifuged at 16 000 g for 15 min. Liquid was removed and the process repeated twice. Finally, 1 ml of 1X phosphate-buffered saline was added to the spore pellet and the mixture was vortexed hard three times in 5 min increments. Multiple serial dilutions were made to ascertain the estimated concentration and a portion of the spores were tested with hydrochloric acid following established protocols (Sagripanti et al. 2007). One ml of spores was sacrificed and checked for purity, approximate amount, acid survival, and continued ability to germinate every 2–4 months until the experiments were finished. The counts from later tests were compared with the initial and other earlier tests to see if there was a decrease in the number of germinating spores. Spores were treated with DNAse to remove exterior DNA before usage. After the nuclease treatment, PCR reactions described in the following paragraphs, using direct spores or supernatant as template, gave negative results for the different assays. All spores were stored at 4 or −70°C.

A known amount of spores was added to an aliquot of soil before either directly processing with one of the manual kits and automatic methods or before pretreatment and/or enrichment steps. First half of the soils was autoclaved after adding the spores, while the other half was not. Later, all of the soils except those used with the enrichment step below were autoclaved after adding the spores as previously described (Luna et al. 2003). Because we did not notice any difference in detection of gene targets in the DNA from autoclaved or unautoclaved spore/soil mixtures, we autoclaved all samples before DNA extraction. When the enrichment step was used, we autoclaved all of the bacterial pellets before the DNA extractions were performed. All tests were performed in triplicate. To be confident of the number of spores added to a soil aliquot, the spores were counted using a Rosenfuch’s hemacytometer or heat-shocked (60°C for 15 min), serially diluted ten-fold, and plated onto tryptic soy agar supplemented with 5% sheep red blood cells (BA) (Remel, Lenexa, KS, USA) to confirm the visually counted and /or calculated amount of spores added to the different tests. The plates were incubated at 30°C. At 24 h and each subsequent day, colonies were counted and the colony forming units (CFU) per ml were calculated. Plates were held up to 7 days. All spores were treated with DNAse to remove any free DNA from their surface before adding to the soil aliquots. The average number ± standard deviation of either Pasteur or Sterne spores added to the soil aliquots were: 1·33 ± 0·32 × 107; 1·58 ± 0·95 × 106; 2·5 ± 1·84 × 105; 1·84 ± 1·51 × 104; 3·8 ± 0·33 × 103; 2·4 ± 0·77 × 102; 6·7 ± 2·1 × 101; and 5·0 ± 3·1 × 100.

Soil pretreatment

In order to remove humic acids and free DNA from the test samples, spores were added to aliquots of soil (0·1–0·5 g) that were then suspended in 5 ml (10 ml for garden soil) of a Pretreatment Solution [0·1% sodium pyrophosphate tetrabasic (Na4P2O7), 1 mmol l−1EDTA, 10 mmol l−1 Tris-Cl]. The soil–spore-solution mixture was vortexed hard two to three min at least three times (Rösch et al. 2002). The mixture was allowed to settle for 10 min at room temperature and centrifuged for 10 min at 4500 g. The supernatant was discarded, and pellet processed with one of the five protocols.

Testing kits with sterilized soil treated with UV

To confirm that the B. anthracis target gene was not already in the soil samples and that the PCR was detecting the added spores, twelve aliquots (2–5 g) of the three soils were placed into 1·5 ml screw-topped autoclavable tubes (Fisher Scientific, Suwannee, GA, USA) moistened with 5–10 μl sterile water, sonicated for 30 min in a Branson 1500 sonicating water bath (Branson Ultrasonics Corp., Danbury, CT, USA), and autoclaved as previously described (Luna et al. 2003). The tubes were opened, placed into a Bio-Safety Cabinet and exposed to UV light (254 nm) overnight to destroy any DNA within the sample. The next day a small portion of each soil (up to one-half of soil aliquot) was cultured for any bacterial growth. The remainder of the sonicated and UV-exposed soil was placed into additional 1·5-ml screw-topped autoclavable tubes that were divided into two different groups. One group was processed as is, while the other group had 106 or 107B. anthracis spores added before processing. The amount of soil in each tube that was used for DNA extraction ranged from 0·1–1 g following the manufacturer’s requirements. DNA was extracted from all of the aliquots.

Soil enrichment in selective broth

Aliquots of soil (0·5–1 g) with the added B. anthracis spores were suspended into 25 ml of a selective medium polymyxin B-lysozyme-EDTA-thallium acetate broth (PLET) supplemented with sulfamethoxazole (38 μg ml−1) and trimethoprim (2 μg ml−1) (Luna et al. 2009). The soil mixture was vortexed hard for 10–20 s three times, heated to 65 or 80°C for 30 min and incubated overnight in a 30°C shaking water bath. The next day the bacterial growth was vortexed hard twice for 10–20 s, filtered through one layer of gauze (Fisher Scientific, Pittsburg, PA, USA) to remove large soil particles and centrifuged for 1 h at 5500 g. The supernatant was discarded, and the pellet washed with dH2O and then centrifuged again for 20 min at 5500 g. The bacterial pellet was immediately used in the DNA extraction steps with one of the five tested methods.

DNA extractions

Three commercially available soil kits for manual DNA extractions were examined: (i) MoBio UltraClean™ Soil DNA Isolation Kit (12800-50) (MO BIO Laboratories, Inc., Carlsbad, CA, USA); (ii) Epicentre SoilMaster™ DNA Extraction Kit (SM02050) (Epicentre® Biotechnologies, Madison, WI, USA); and (iii) Fast DNA® SPIN Kit for Soil (116560-200) (MP Biomedicals, Solon, OH, USA). The two automatic systems evaluated were MagNa Pure®LC (Roche Diagnostics Corp., Indianapolis, IN, USA) and Qiagen® BioRobot M48 Workstation (Qiagen, Valencia, CA, USA).

Whole cell DNA was extracted using the three manual kits and two automatic instruments following the various manufacturers’ instructions. The two automated kits had extra instructions for processing the soil. The Roche MagNA Pure®LC DNA Kit III was used for extracting DNA from soil with the MagNA Pure LC automated system. The MagNa Pure®LC directions required autoclaving of the soil aliquot and a prelysis step in the lysis buffer found in their MagNaPure®LC DNA Isolation Kit III before a 10 min centrifugation at 13 000 g. The entire amount of supernatant was then used with the rest of the kit solutions as directed by the manufacturer. The Qiagen® BioRobot M48 Workstation directions required the soil samples to be pretreated with components from two Qiagen® kits (QIAamp®DNA Stool Mini Kit and MagAttract DNA Mini 48 Kit). Water (600 ul) was added to each soil aliquot (up to 0·5 g), incubated for 10 min at 95°C and centrifuged for 2 min at 4000 g. The supernatant was transferred to a 2-ml tube and 190 ul of buffer G2 from the MagAttract DNA Mini M48 Kit was added, followed by one-half of an InhibitEX tablet from the QIAamp® DNA Stool Mini Kit. The supernatant–buffer–tablet solution was incubated for one min before being mixed and centrifuged for 2 min at 10 000 g. The resulting supernatant was loaded into sample tubes in the Qiagen® BioRobot M48 Workstation instrument and processed.

Following the manufacturer’s recommendations, untreated soil with and without spores were directly processed by each method and tested in parallel with both the pretreated soil pellets and the bacterial pellets obtained after enrichment. DNA was stored at −20°C until needed.


Primers for capC were previously designed using LaserGene® (DNAStar®, Madison, WI, USA) and B. anthracis pX02 sequence obtained from GenBank (accession number NC_002146) (Table 1) (Luna et al. 2006) and produced a 1535 bp product. Primers for lef and pag genes designed from the pX01 sequence obtained from GenBank (accession number NC_001496) yielded amplicons 432 and 539 bp, respectively (Table 1). Using MegAlign® (DNAStar®), an alignment of the 16S sequences of different Bacillus species, including Bacillus atrophaeus (X60607), B. anthracis (AE017334 and AE017225), Bacillus cereus (X55060 and X55063), Bacillus mycoides (X55061), Bacillus subtilis (EF422864) and Bacillus thuringiensis (X55062) was used to derive primers that would work with various Bacillus species. The resulting 16S assay 403-bp amplicon was used as a control to demonstrate that the PCR chemistry was not inhibited.

Table 1.   Oligonucleotides* used for PCR assays and DNA-DNA hybridizations for 16S rDNAand capC, lef and pag genes
GenePrimer/probeSequence (5′→3′)Location (bp)Size (bp)GenBank numberOrganismReference
  1. *Oligonucleotides were used as PCR primers and/or probes for Southerns of PCR amplicons and whole cell DNA dot blots. Oligonucleotide probes were internal to, and specific for, the target sequence as determined by N-blast with NCBI library. Probes were also used for DNA–DNA hybridization of dot blots made from either genomic or plasmid DNA.

16S16SBac FCGT GGG GAG CGA ACA GGA TTA GAT A757–781400X55060Bacillus cereusThis paper
16SBac RGTT TGT CAC CGC GAG TCA CCT TAG AG1131–1157 X55060B. cereusThis paper
capCcapB2FGGT CTT CCC AGA TAA TGC ATC GCT TG56 631–56 6561535NC_002146Bacillus anthracisLuna et al. 2006
capA1RAGT TGT TGT CTC CAC TGA TAC TTG ATT TTC55 121–55 150 NC_002146B. anthracisLuna et al. 2006
capA1FCAA CAT TTG CAA TCA TGA ATA TTT ATT ACT TAT55 625–55 659 NC_002146B. anthracisLuna et al. 2006
capB2RTTC TTT CTG TAA AAA TAA GGC TCA GTG TAA CTC CT56 005–56 039 NC_002146B. anthracisLuna et al. 2006
lefLF1FTTA GAT AAT GAG CGT TTG AAA TGG AGA A129 020–129 048432NC_001496B. anthracisThis paper
LF1RTAG GGA GAG TAA TAT CGG TAA AAA CAA ATC129 422–129 452 NC_001496B. anthracisThis paper
LF5TTA GAT AAT GAG CGT TTG AAA TGG AGA A129 176–129 199 NC_001496B. anthracisThis paper
pagPA1FAGT GCA TGC GTC GTT CTT TGA TA134 421–133 916529NC_001496B. anthracisThis paper
PA1RGAA TTT GCG GTA ACA CTT CAC TCC133 916–133 940 NC_001496B. anthracisThis paper
PA6AGG GGA AAG AAC TTG GGC TGA AA134 322–134 346 NC_001496B. anthracisThis paper
PA7CCT TAG CTT TAA TTG TCG CGA GTG TTT GAT134 177–134 206 NC_001496B. anthracisThis paper

PCR reactions of 10 μl [1·5 mmol l−1 MgCl2; 1X buffer; 2 μmol l−1 dNTP; 1·3 μmol l−1 primers; 0·025U Taq DNA polymerase (Takara, Madison, WI, USA)] were carried out using 0·5–2 ng of extracted DNA as template in a T1 Thermocycler (Biometra®, Horsham, PA, USA). For the 16S, pag and lef assays, the following parameters were followed: initial heating for 2 min at 94°C, followed by 40 cycles of 20 s at 94°C, 20 s at 58°C, and 3 min at 72°C, with a final extension at 72°C for 5 min. For the longer capC assay, the annealing temperature was raised to 60°C, while the denaturing and annealing times were lengthened to one min and the final extension time to 7 min. PCR assays were performed in triplicate. When assays produced only negative results for a target, total volume of reaction mixtures and DNA templates were increased up to 50 μl and 50 ng, respectively. Positive and negative controls were used for all assays (B. anthracis Pasteur CBD 63 (+); B. anthracis CBD 131(+); B. anthracis Sterne BB001 (+), B. cereus CBD 58 (−) and /or dH2O (−). In addition, positive control DNA was mixed with an equal amount of the extracted DNA obtained by each of the methods and used as template to show that the PCR reactions worked in that particular soil.

The PCR products and the ‘DNA’ from the UV-treated and sterilized soils were electrophoresed on a 1% agarose gel (0·5X TBE (44.5 mmol Tris-borate and 1 mmol EDTA, pH 8.3) with 0·05 μg ml−1 ethidium bromide) for 50 min at 80 V constant voltage. Staining with additional ethidium bromide and destaining were not necessary. The DNA bands were visualized with UV light, and photographs were made using the GelDoc® (Bio-Rad, Hercules, CA, USA).

DNA–DNA hybridization

DNA dot blots containing 10–100 μg extracted DNA and Southern blots from gels of the electrophoresed PCR amplicons were made using Immobilon-nYplus nylon membrane (Millipore, Bedford, MA, USA) or Roche nylon membrane (Roche Diagnostics) and prepared following standard protocols (Sambrook et al. 1989). The DNA was bound to the membrane by UV irradiation using the Spectrolinker XL1000 (Spectronics Corporation, Westbury, NY, USA). Oligonucleotide probes specific for internal portions of the targeted genes were designed with LaserGene® (DNAStar®) (Table 1) and labelled with digoxigenin using the DIG Oligonucleotide Tailing Kit (Roche) as per the manufacturer’s instructions. The different probes detected the gene targets in as low as 2 ng of whole cell DNA from the positive control. Positive and negative DNA was used with each hybridization test.

Statistical tests

All comparisons of the different results were analysed with anova statistical tests. The null hypotheses were that there were no differences in detection between the different methods or soils.


Soil characteristics

The Texas soil sample TX-A3 was alkaline, dry, very loose, tan medium-grained silica sand containing silt, organics and minor sub-rounded gravel. The composition of the soil was roughly 8% coarse-grained sand, 30% medium-grained sand, 55% fine-grained sand, 6% silt and clay and 1% organic compounds. Using the USCS method, the soil TX-A3 was classified as Poorly-Graded SAND with silt, having ≥50% of coarse fraction larger than 4·75 mm and >50% retained by a 75-μm sieve. The Florida soil sample FL-51 was acidic, dry, very loose, tan, fine-grained calcareous sand, with silt. It also contained minor coarse-grained sand and minor fine sub-rounded gravel. FL-51 was composed of roughly 87% fine-grained sand, 1% silt and clay, and 12% organics and by the USCS method was classified as Poorly-Graded SAND having ≥50% of coarse fraction less than 4·75 mm and >50% retained by a 75-μm sieve. The commercial garden soil sample Com-A was acidic, black, damp, and very loose peat with fine to medium-grained sand. The soil composition was roughly 4% fine to medium-grained sand, about 1% silt and clay, and 95% organics. The soil Com-A was classified as PEAT when following the USCS method.

Untreated and pretreated soils

The sterilized and UV-treated soil samples did not produce any bacterial growth after 7 days incubation. The soils were processed with the three manual kits (Epicentre SoilMaster™ DNA Extraction Kit, Fast DNA® SPIN Kit for Soil and MoBio UltraClean™ Soil DNA Isolation Kit). Spectroscopy of the potential DNA product extracts from the UV-treated soils produced readings equal to a blank sample, and DNA was not visually detectable on an agarose gel. When DNA from B. anthracis Pasteur or Sterne was added to these preparations, PCR assays were positive. In addition, all of the UV-sterilized soil aliquots that had B. anthracis Pasteur spores added after the UV exposure and then processed produced DNA that yielded positive PCR amplicons for both 16S and capC.

The different methods extracted a wide range of DNA concentrations from the test soil samples that were spiked with B. anthracis spores (Table 2). The MagNa Pure®LC instrument gave the highest concentrations of DNA, yet was the most inconsistent in the amount of DNA extracted from a single soil type. The extracted DNA from this instrument was consistently negative for the B. anthracis genes and positive for 16S when multiple PCR assays were performed from the same extracted DNA sample even when a high number of B. anthracis cells or spores were in the sample. There was statistical significance between the results obtained with the DNA extracted with the MagNaPure®LC and the results obtained with extracted DNA from the other four methods (P = 0·00000127). With the other methods, no pattern correlating the DNA concentrations and PCR results could be discerned. Surprisingly, the garden soil did not always yield more total DNA than the other two soils. For example, the Epicentre SoilMaster™ DNA Extraction Kit produced 300 μl of DNA from untreated Florida, untreated Texas and untreated garden soils, with average DNA concentration of 18 ± 1·9 ng ul−1 (range = 14–21 ng μl−1), 7·8 ± 4·5 ng μl−1 (range = 3–15 ng μl−1) and 10·5 ± 6·1 ng μl−1 (range = 4·3–22 ng μl−1), respectively (Table 2).

Table 2.   Concentration means, standard deviations, ranges and purity of extracted DNA from soils for five methods
Kit/methodSoil weight Soils
Untreated*PretreatedWith enrichmentUntreatedPretreatedWith enrichmentUntreatedPretreatedWith enrichment
  1. *‘Untreated’ soils were directly added to the components of the manufacturer’s DNA extraction kits as directed by the manufacturer; ‘Pretreated’ soils were treated to remove organic and humic acids before DNA extraction following manufacturer directions. ‘With enrichment’ soils were added to the modified selective growth medium and incubated overnight before being pelleted and then used in the DNA extraction kits.

  2. †‘SD’ denotes standard deviation. Each kit/method was tested a minimum of six times for each concentration of spores with each soil, treated, pretreated and after enrichment.

Epicentre SoilMaster0·1 gDNA conc.
mean ± SD
(ng μ−1)
18 ± 1·95·3 ± 2·638·7 ± 18·67·8 ± 4·44·1 ± 2·651·2 ± 11·410·5 ± 6·126·1 ± 2·436·2 ± 25·1
Range (ng μl−1)14–211·8–921·1–683–151·6–935·9–70·84–2216–308·2–79·8
DNA Purity A260/A280 ratio0·6–1·21·1–1·31·8–1·90·9–1·50·6–1·41·8–1·90·6–1·60·6–1·21·7–1·9
MoBio UltraClean1·0 gDNA conc.
mean ± SD
(ng μ−1)
7·5 ± 2·56·6 ± 5·56·6 ± 3·515·2 ± 4·612·0 ± 3·410·4 ± 9·553·4 ± 29·325·2 ± 5·25·4 ± 3·0
Range (ng μl−1)3–112·1–181·3–16·58–236–171·2–34·115–9716–301·3–11·4
DNA purity A260/A280 ratio1·3–1·71·2–1·71·7–1·91·1–1·61·2–1·71·9–2·01·2–1·31·2–1·41·7–2·0
Q-BIOgene Fast DNA® SPIN0·5 gDNA conc.
mean ± SD
(ng μ−1)
29·7 ± 13·657 ± 16·215·3 ± 7·421·7 ± 11·037·6 ± 14·120·3 ± 8·014·7 ± 10·655·6 ± 32·018·9 ± 9·2
Range (ng μl−1)11–4842–867·6–3611–4017–569·8–35·95·2–359·4–1006·4–39·2
DNA purity A260/A280 ratio1·2–1·71·3–1·71·9–2·01·2–1·71·2–1·61·8–1·91·2–1·51·2–1·31·8–1·9
MagNaPure®LC0·5 gDNA conc.
mean ± SD
(ng μ−1)
33·8 ± 31·172 ± 23·821·7 ± 12·132·5 ± 29·986·5 ± 3·914·3 ± 10·239·4 ± 36·841·8 ± 32·316·8 ± 13·9
Range (ng μL−1)4·6–8333–9113–49·31·4–8480–926·2–50·55–1009–753·3–43·6
DNA purity A260/A280 ratio1·2–1·51·2–1·41·7–1·91·2–1·51·3–1·41·8–1·91·2–1·41·2–1·51·8
Qiagen® BioRobot M48 Workstation0·5 gDNA conc.
mean ± SD
(ng μ−1)
13·1 ± 8·96·8 ± 2·74·0 ± 1·65·3 ± 5·37·0 ± 2·93·9 ± 1·910·5 ± 9·68·8 ± 2·26·6 ± 6·1
Range (ng μl−1)1·4–244–141·6–6·61–213·3–161·4–10·31–236–121·8–23
DNA purity A260/A280 ratio1·1–1·71·1–1·61·8–1·91·2–1·81·2–1·61·9–20·6–1·21·1–1·31·8–1·9

The 16S PCR assay was not specific for B. anthracis and gave positive results in reactions with template DNA from other Bacillus and Staphylococcus spp. Although B. anthracis was never cultured from any of the three soils, numerous other Bacillus spp, Staphylococcus and other Gram-positive bacteria were present in the soil cultures (results not shown). All of the methods and soils produced DNA that yielded positive results for the 16S PCR assay, and no significant difference between the soils and methods was noted (P = 8·279).

Although capC has been noted in other species of Bacillus isolates, repeated PCR assays performed on the soils without any added spores were consistently negative. As noted earlier, the three unspiked soils had in addition previously tested negative for the lef and pag genes found on the pX01 plasmid by both PCR and DNA–DNA hybridization studies. After the addition of B. anthracis spores to the soil, the capC gene (carried by the Pasteur strain) or the pag and lef genes (carried by Sterne) were detected in the Florida and Texas soil samples (both pretreated and untreated) using the DNA extracted by both the Epicentre SoilMaster™ DNA Extraction Kit and the automated Qiagen® BioRobot M48 Workstation method (Table 3). However, none of the B. anthracis genes were detected in the garden soil samples that were processed by either protocol. The MoBio UltraClean™ Soil DNA Isolation Kit extracted DNA in which B. anthracis genes were detected for the pretreated Florida soil (106 spores), untreated and pretreated Texas soil samples (105 and 104 spores, respectively). The DNA extracted with the Fast DNA® SPIN Kit for Soil produced positive PCR results for the B. anthracis genes only with high concentrations (107) of spores in the Texas and garden soil. However, the DNA produced by the same kit did not yield positive PCR results for capC, lef or pag in any of the Florida soil samples. In contrast, none of the DNA extracted by the MagNa Pure®LC from the three soil types produced positive PCR results for the B. anthracis genes even though the 16S assays were positive. Only the Qiagen® BioRobot M48 Workstation produced DNA that gave positive results for the target gene when ≥103 spores were present in the soil sample. Because the kits stipulated different optimal amounts of soil to use in their procedures, the limit of detection was calculated per gramme of soil for each method for comparison. The detection limit for the Epicentre SoilMaster™ DNA Extraction Kit increased by 1 log to 107 spores per gram of soil, while the detection limits for the other kits remained the same as found in Table 3. Yet applying statistical tests, we found no statistical difference among the four methods (minus the MagNa Pure®LC) with untreated soils (P = 0·0997).

Table 3.   The lowest concentration of Bacillus anthracis spores* detected by PCR† for capC gene in DNA extracted from three soils with and without enrichment step
Untreated‡PretreatedWith enrichmentUntreatedPretreatedWith enrichmentUntreatedPretreatedWith enrichment
  1. *Average numbers of spores added to the different soils were: 1·33 ± 0·32 × 107; 1·58 ± 0·95 × 106; 2·5 ± 1·84 × 105; 1·84 ± 1·31 × 104; and 1·6 ± 0·31 × 103.

  2. †PCR assays for the 16S target were consistently positive for the three soils and all five methods of DNA extraction.

  3. ‡‘Untreated’ soils were directly added to the components of the manufacturer’s DNA extraction kits as directed by the manufacturer; ‘Pretreated’ soils were treated to remove organic and humic acids before DNA extraction following manufacturer directions. ‘With enrichment’ soils were added to the modified selective growth medium and incubated overnight before being pelleted and then used in the DNA extraction kits.

  4. §‘N’ denotes ‘not detected’. The B. anthracis target genes were not detected in the sample after addition of 107 spores.

Epicentre SoilMaster106107102106105101N102
MoBio UltraCleanN106102105104102106N102
Fast DNA® SPINNN102107N101107107102
Qiagen® BioRobot M48 Workstation105106102104103102NN101

Initially the pretreatment step appeared to benefit the PCR testing. Using the DNA extracted with the MoBio UltraClean™ Soil DNA Isolation Kit, PCR could detect a lower concentration of spores in the pretreated Florida and Texas soils than was detected in the untreated samples. However, an examination of the table shows that the step is not always necessary when using the other extraction methods. Statistically, there was no difference between the untreated soils and pretreated soils (P = 0·654). In addition, the pretreatment was not necessary when the soil and spores were grown in the modified PLET and the pellet rinsed with dH2O (data not shown).

DNA purity and Bacillus anthracis detection after enrichment step

The extracted DNA produced by the five different methods after enrichment in the modified PLET medium had comparable purity (A260/A280 ratio range of 1·79–2·0) and was sufficient for PCR and DNA–DNA hybridization (Table 2). This was an improvement from the DNA purity results obtained when the enrichment step was not included in the protocols. Without the enrichment step, the A260/A280 ratio for extracted DNA from both untreated and pretreated soils were consistently low and ranged from 0·6 to 1·7 (Table 2). No one kit gave superior DNA. This revealed that the DNA had probable protein contamination and necessitated further purification steps before use for PCR. There appeared to be no correlation between DNA purity and positive PCR results as demonstrated by the MagNa Pure®LC higher A260/A280 ratio and consistent negative PCR results for the target B. anthracis genes.

The concentration of the extracted DNA from the different kits after being enriched in the modified medium were usually about the same or slightly lower than when the soil was used directly with the method or treated with the pretreatment solution (Table 2). The exception was the higher DNA concentrations after extraction by the Epicentre SoilMaster™ DNA Extraction Kit. This was seen in all three soil types. Even when cell pellets or soil samples underwent a freeze-thaw cycle, the total amount of DNA was not increased noticeably.

The PCR assays using the extracted DNA from the soils incubated in the modified PLET broth yielded positive results for four of the five methods (Table 3). Statistical testing showed that the difference in detection by PCR for samples having enrichment and for samples not having enrichment was significant (P = 0·0 000 783). Even samples that originally had very low numbers of spores (101–102 spores) added to them yielded DNA that was positive for the target genes by PCR assays. The detection limits of the four methods were lowered by several logs (2–6 log10) if the enrichment step was included in the protocol. This is not surprising because the enrichment step allows the germination of spores and multiplication of vegetative cells in the sample as demonstrated in a previous report (Luna et al. 2009).


Because DNA extraction from a soil sample results in the co-extraction of humic and folic acids that interfere with polymerases and restriction endonucleases, cumbersome and time-consuming methods were employed to remove the interfering organics. The introduction of manual kits and instrument protocols that incorporated the organics removal has attempted to improve the purity of the extracted DNA. In this study, the five examined DNA extraction methods did not consistently yield DNA suitable for PCR unless the DNA was purified again after the extraction process. This occurred many times if the soil sample was used directly or processed with the pretreatment solution to remove the organics from it. When the DNA was suitable for PCR, the PCR assays for the 16S target were consistently positive but the B. anthracis genes were not detected at all in most of the tests. When the target B. anthracis genes were detected, it was usually at levels ≥106 spores per sample. Only the Qiagen® BioRobot M48 Workstation protocol generated B. anthracis–positive DNA when only 104 and 103 spores were added to the untreated and pretreated Texas soils, respectively. This is similar to Qiagen’s stated detection limit of 104 cells. However, these results were not repeated in the other two soil types. It is interesting to note that none of the other systems offered detection limits for B. anthracis spores or cells.

One possible reason for the numerous negative PCR results for B. anthracis plasmid genes with the DNA extracted from the untreated and pretreated soils is that there may be too many micro-organisms in the sample. Therefore, the abundance of competing DNA is too great compared to the target DNA extracted from the spores. Thus, the target gene is lost when the sample is diluted to the optimal concentration for PCR. This was addressed by performing multiple PCR assays and performing PCR in larger volume mixtures for each sample. This overabundance of ‘other’ DNA could be why the target genes were not detected in the MagNa Pure®LC-extracted DNA. This platform had previously been useful in the detection of low numbers of spores in powders and dust. However, these matrices had minimal amounts of total DNA (‘other’ or target) with little inhibiting organics and the extracted DNA needed to be concentrated to get a consistent positive reaction (Luna et al. 2003). Yet, after the enrichment step was added to this protocol for the MagNa Pure®LC, the B. anthracis specific PCR assays remained consistently negative. When the PCR reaction volumes were increased and performed multiple times, the PCR assays for B. anthracis were still negative with the DNA from the MagNa Pure®LC, while the assays for 16S were positive.

Another possibility for repeated negative PCR reactions for both 16S and B. anthracis target genes is that soil characteristics may interfere with the protocols. There may be too much debris or amount of organics in the soil that a protocol might not be able to overcome. The pretreatment step was intended to aid in this problem, but this added step could potentially cause a loss of the spores before the DNA extraction. This is perhaps why B. anthracis was at times detected with lower numbers of spores in untreated soil than the number of spores needed for detection in the pretreated soils. Additionally, the pretreatment steps could also have interfered with the components of the extraction method. This was not possible to explore as many of the component recipes are proprietary information. Lastly, there can be compounds such as clay or organic or physical structures to which the spores adhere and the spores may not become detached from these during the extraction process. This could explain the negative results with the DNA extracts from the garden soil. Although Wawrik et al. (2005) reported that soil type was not important in identifying members of a bacterial community using PCR, this study shows that the combination of kit/method and soil type does affect the detection of B. anthracis spore DNA. Thus, a researcher should know the type of soil to be examined for the presence of B. anthracis spores and then choose the method accordingly. All three soils used in this study had very low amounts of clay in them. Thus, the detection of spores in the unprocessed soils was probably easier than in soils with high clay content as the clay can adhere to the spores and prevent them from being separated and the DNA extracted. Yet, even with these low clay content soils, the different methods did not detect B. anthracis spores readily when using untreated and pretreated soils.

In contrast, this study demonstrates the usefulness of using a heat-shock and enrichment step before extracting the DNA from the soil. This enrichment step utilized a selective medium that contained polymyxin B, sulfamethoxazole and trimethoprim, heat shock, incubation temperature and aeration by shaking to encourage the germination and multiplication of B. anthracis cells while inhibiting or slowing the growth of other bacteria in the sample. Although other researchers did not recommend PLET for recovery of B. anthracis spores from soil (Dragon and Rennie 2001), previous work in our laboratory demonstrated that PLET supplemented with trimethoprim/sulfamethoxazole aided in the isolation and identification of B. anthracis from soil if an enrichment step was included (Luna et al. 2009). This current study shows that after the enrichment step with the modified growth medium was added to the different DNA extraction protocols, all five methods produced DNA that was consistently purer than the DNA extracted from untreated or pretreated soils. The DNA could be used immediately for PCR. Other researchers have used broth media to increase DNA purity (Cheun et al. 2003), but the broth used did not select for B. anthracis by inhibiting the growth of other bacteria found in the soil samples. With the enrichment step added, the detection limits for four of the five methods were improved so that we could detect B. anthracis in the soil samples that originally had as low as 101–102 spores added to them. Of course, this detection was after the spores multiplied overnight to 107–1011 CFU ml−1 as seen previously (Luna et al. 2009).

Because different B. anthracis strains carry different copies of the two virulence plasmids (Coker et al. 2003), our laboratory estimated that our two test strains carried low numbers of both plasmids (1–3 copies of pX01 and 5–30 copies of pX02) (data not shown). These fall at the lower end of the ranges described by Coker’s work. Thus, it is possible that one or another extraction method could have detected lower numbers of spores from a strain with very high copy numbers of either plasmid when testing the soils directly or after pretreatment. We tried to overcome this by performing the PCR assays a minimum of three times for each PCR target and by increasing the reaction volume to 50 μl and the DNA amount to 50 ng. Yet, we never obtained positive reactions with more DNA template or larger reaction volumes that were previously negative with the original testing. However, the number of plasmids carried by cells and spores of a particular strain is not as critical when the enrichment step is included in the extraction process. Although we could not test our virulent strains with the automated instruments that are not in the BSL3 laboratory because of our Institutional Biosafety Committee concerns, we feel confident that we would be able to detect low numbers of them in the three soils. All of the virulent strains in our collection have been able to grow to a high density of cells in the enrichment medium. (Luna et al. 2009).

Although anthrax of cattle normally occurs in areas with alkaline soils and optimal mineral content, it is possible that a more acidic soil (as in much of the Eastern United States) could harbour the spores (Van Ness 1971). Fortuitously, it appears that B. anthracis is normally found in warm alkaline soils because they can grow in areas of soil that periodically go through wet and dry spells, called ‘incubator areas’ (Van Ness 1971). This is supported by work that has defined the proper pH and calcium content for B. anthracis to multiply in the soil (Smith et al. 2000; Vilain et al. 2006). Other researchers report that B. anthracis does not germinate and grow in soil, but rather, the spores remain indefinitely in the environment and the areas of high anthrax outbreaks should be renamed ‘storage areas’ (Dragon et al. 2001; Jensen et al. 2003). In this view, the B. anthracis spores germinate and grow only in an animal host where the vegetative cells multiply until the animal dies and the spores are released to the environment. Dissemination of the spores occurs by insect and scavenger activity and water movement over the storage areas (Braack and DeVos 1990; Dragon et al. 2001; Turell and Kundson 1987). In this scenario, it would not make any difference if the soil was alkaline or acidic.

Therefore, it was worrisome that the garden soil masked the presence of 107 spores in a tiny sample when the untreated or even pretreated soil was used directly in the five methods. Soil in many parts of the world is rich in humic acids and contains many organics and debris as seen in the commercial soil. Beef and dairy cattle and other ruminants that are most likely to be exposed to the spores are found in all parts of the world. Thus, in order to prevent future infections, it is important to be able to definitively verify if a potentially contaminated soil does indeed harbour B. anthracis spores. Soil is an important specimen for the detection of B. anthracis spores for both agronomic and forensic investigations. As seen in this study, the enrichment step aided in the detection of low numbers of spores in the garden soil for four of the five methods. The success of detecting low spore numbers in the garden soil implies that the enrichment step could potentially improve detection of spores in such humic soils,

The purpose of this study was to compare the five different DNA extraction methods without treatment, with organic removal and after an enrichment step utilizing a selective medium. Yet, although we are not attempting a cost analysis, the costs of the different methods can be mentioned. All five methods were easily executed, and little time was needed to teach personnel how to perform the manipulations. The cost to process one sample was $3·46 for MoBio UltraClean™ Soil DNA Isolation Kit, $3·60 for Epicentre SoilMaster™ DNA Extraction Kit, and $4·60 for Q-BIOgene Fast DNA® SPIN Kit for Soil. Not including the large initial cost for instrument procurement, the cost per sample using the two automated instrument systems was $7·44 and $11·93 for the MagNaPure®LC and Qiagen® BioRobot M48 Workstation, respectively. Processing time for one sample varied from 24 min (two automated instruments) to 60, 70 and 90 min for the Q-BIOgene Fast DNA® SPIN Kit for Soil, the Epicentre SoilMaster™ DNA Extraction Kit, and the MoBio UltraClean™ Soil DNA Isolation Kit, respectively. Batches of six samples could be easily handled for the manual kits, while the automated systems could process up to 24 samples at a time. The two automated instruments had many disposable components provided by the manufacturer, while the manual kits had few required disposables also provided by the manufacturer. The product volume for each method was 300 μl for the Epicentre SoilMaster™ DNA Extraction Kit, 200 μl for the Qiagen® BioRobot M48 Workstation, and 100 μl for the remaining methods.

In conclusion, this study has shown that the five examined kits or instruments for extracting DNA from soil produced DNA that was not always suitable for molecular testing such as PCR when the soil was used directly or was pretreated to remove organics. However, an enrichment step involving heat-shock and overnight incubation in a selective medium improved the purity of extracted DNA and improved the detection of low numbers of B. anthracis spores in the soil. This study also shows that although the type of soils being tested, the organic load of the soil sample (bacterial, eukaryote, and chemical) and the small sample size used in the different methods are important factors to consider, the addition of the enrichment step lowered the detection limits for four of the five methods in all three soil types. This simple step could potentially aid in future testing of soil or surfaces contaminated with soil that may contain B. anthracis spores.


This work was supported by US Army Research, Development and Engineering Command, contract. W911SR-07-C-0084. The B. anthracis spores and cultures were generously donated by Lea Heberlein-Larson and Frank Reeves at the Florida Department of Health Bureau of Laboratories, Tampa, FL.

Transparency declaration

None to declare.