Evaluation of DNA extraction methods for Bacillus anthracis spores isolated from spiked food samples
Kingsley K. Amoako, Canadian Food Inspection Agency, National Centers for Animal Disease, Lethbridge Laboratory, PO Box 640, Township Road 9-1, Lethbridge, AB, Canada T1J 3Z4.
Nine commercial DNA extraction kits were evaluated for the isolation of DNA from 10-fold serial dilutions of Bacillus anthracis spores using quantitative real-time PCR (qPCR). The three kits determined by qPCR to yield the most sensitive and consistent detection (Epicenter MasterPure Gram Positive; MoBio PowerFood; ABI PrepSeq) were subsequently tested for their ability to isolate DNA from trace amounts of B. anthracis spores (approx. 6·5 × 101 and 1·3 × 102 CFU in 25 ml or 50 g of food sample) spiked into complex food samples including apple juice, ham, whole milk and bagged salad and recovered with immunomagnetic separation (IMS).
Methods and Results
The MasterPure kit effectively and consistently isolated DNA from low amounts of B. anthracis spores captured from food samples. Detection was achieved from apple juice, ham, whole milk and bagged salad from as few as 65 ± 14, 68 ± 8, 66 ± 4 and 52 ± 16 CFU, respectively, and IMS samples were demonstrated to be free of PCR inhibitors.
Detection of B. anthracis spores isolated from food by IMS differs substantially between commercial DNA extraction kits; however, sensitive results can be obtained with the MasterPure Gram Positive kit.
Significance and Impact of the Study
The extraction protocol identified herein combined with IMS is novel for B. anthracis and allows detection of low levels of B. anthracis spores from contaminated food samples.
Bacillus anthracis, a Gram-positive bacterium and the causative agent of anthrax, is one of five organisms classified as category A bioterrorism agents by the Centers for Disease Control. This is due to the fact that anthrax has a high mortality rate, potential to cause panic and social disruption, is easily disseminated and requires specialized response measures (CDC; http://www.bt.cdc.gov/agent/agentlist-category.asp). Bacillus anthracis is further regarded as the most environmentally stable biothreat agent (Sinclair et al. 2008). This is attributed to formation of endospores (hereafter referred to as spores), a metabolically inactive morphotype that can withstand harsh environmental conditions for decades (Liu et al. 2004). Bacillus anthracis spores are the most common aetiologic morphotype of anthrax and have been used previously in bioterrorist attacks (Jernigan et al. 2002).
Food and feed chains are well-recognized targets for accidental and intentional contamination with biothreat agents, and such an act would have widespread social and economic consequences (Knutsson et al. 2011). The food supply is now more globalized than ever, and food transport has been increasing exponentially since the 1960s (Ercsey-Ravasz et al. 2012). Agri-food and seafood imports into Canada totalled nearly $30 billion in 2010, and imports from 142 countries totalled $1 million or more per country (Agriculture and Agri-Food Canada; http://www.ats-sea.agr.gc.ca/stats/stats-eng.htm). Consequently, research needs have been identified that deal with the detection, surveillance, prevention and treatment of biothreat agents (Knutsson et al. 2011).
Detection of food borne bioterrorism agents is a challenging task given the complex nature of food. Molecular-based approaches have been developed to detect DNA signatures of pathogens, but many foods contain PCR inhibitors such as fats, glycogen, organic and phenolic compounds and humic acids, which can cause false-negative results (Wilson 1997). Enrichment is commonly used to overcome these technical challenges despite the fact that adding an enrichment step increases detection time (Nugen and Baeumner 2008). Direct DNA extraction from a small subsample has also been used (Wielinga et al. 2011); however, the use of a small sample reduces detection sensitivity. In recent years, immunomagnetic separation (IMS) has become more common for the concentration of pathogens from food (Allegra et al. 2011; Fedio et al. 2011), as it allows larger sample volumes to be processed and is able to exclude much of the natural microflora associated with food samples provided a suitable antibody is available for pathogen capture. While IMS has recently been applied to food samples for the isolation of B. anthracis spores (Shields et al. 2012) and there have been numerous investigations into DNA extraction from spores (Dauphin et al. 2009; Gulledge et al. 2010; Wielinga et al. 2011), it is uncertain how DNA extraction methods perform in the presence of immunomagnetic beads and debris from complex food samples.
The objectives of this study were to evaluate nine commercial DNA extraction kits for their ability to extract quantitative PCR (qPCR) quality DNA from B. anthracis spores and identify the most robust method for DNA extraction from low amounts of spores immunomagnetically separated from complex food samples including whole milk, apple juice, ham and bagged salad.
Materials and methods
Culture conditions and spore preparation
Bacillus anthracis Sterne strain spores were prepared as described previously (Shields et al. 2012). Sporulation was verified by malachite green staining, and spores were adjusted to a final concentration of 109–1011 spores ml−1 in PBS supplemented with 1% BSA (pH 7·0) and stored at −20°C.
Down selection of DNA extraction kits
A total of nine commercial DNA extraction kits were chosen for comparison of their ability to recover genomic DNA from B. anthracis spores (Table 1). DNA extractions were performed as per the manufacturers’ protocol (details below), and the minimum recommended elution/resuspension volumes for each kit were used to enhance detection sensitivity.
Table 1. Comparison of commercial DNA extraction kits used in the down-selection process including cost per extraction, time and additional equipment required for use
|UltraClean||C/M||Silica column||50||2·60||2 h||Vortex adapter||No|
|PowerFood||C/M||Silica column||100||2·94||2 h||Vortex adapter||No|
|FastDNA Spin Kit||C/M||Silica column||100||5·03||1 h||FastPrep Instrument||No|
|MasterPure Gram Positive||C||Precipitation||35||3·44||2·5 h||Not applicable||No|
|TruTip XT SPT||C/M||Silica pipette tip||100||5·5||1·6 h|| |
Rainin 200–2000 μl EDP3-Plus
Advanced Electronic Pipette
|DNeasy Blood and Tissue||C||Silica column||100||3·44||2·25 h||Not applicable||Yes|
|Wizard||C||Precipitation||100||1·64||3 h||Not applicable||No|
|PrepSEQ||B/C||Magnetic beads||50||5·00||3·5 h||Magnetic stand for 1·5-ml tubes||Yes|
|IT 1-2-3 Platinum Path||B/C/M||Magnetic beads||150||13·75||4 h|| |
PickPen 1-M Magnetic Wand
TruTip XT SPT kit (Akonii Biosystems, Frederick, MD, USA): the bead protocol for liquid cultures was followed. PrepSeq kit (Applied Biosystems, Foster City, CA, USA): the proteinase K lysis protocol was used. The lysate was mixed with beads using the Revolver Rotator tube mixer with rotisserie (Denville Scientific Inc., Metuchen, NJ, USA), and magnetic separation of DNA bound to magnetic beads was accomplished with the Biomag tube separator (Bangs Laboratories Inc., Fishers, IN, USA). FastDNA SPIN kit (Bio101, Vista, CA, USA): CLS-TC lysis solution was used, and two 40-s homogenization steps were performed at a speed setting of 6. Samples were mixed with the binding matrix using a vortex adapter at a speed setting of 4. MasterPure Gram Positive kit (Epicenter, Madison, WI, USA): prior to the addition of TE buffer following the ethanol wash, samples were heated at 37°C for 15 min, and residual ethanol was pipetted off the tube lid. UltraClean and PowerFood kits (MoBio Laboratories Inc, Carlsbad, CA, USA): the optional heating step at 65°C for 10 min prior to bead beating was incorporated. 1-2-3 Platinum Path (Idaho Technologies, Salt Lake City, UT, USA): the culture (DNA/RNA) protocol was performed with bead beating using the Mini-Beadbeater-16 (Biospec Products, Bartlesville, OK, USA). Wizard kit (Promega, Madison, WI, USA): DNA was extracted according to the Isolating Genomic DNA from Gram-positive bacteria protocol. DNeasy Blood and Tissue kit (Qiagen, Valencia, CA, USA): the pretreatment for Gram-positive bacteria protocol was followed.
Extraction time and cost per extraction were computed for comparison purposes (Table 1). DNA extractions were performed in triplicate on 10-fold serial dilutions of B. anthracis spores with final concentrations ranging from 1 × 102 to 1 × 106 viable spores per 150 μl of TE buffer. Starting concentrations were verified by spread plating on Tryptic Soy Agar (Difco, BD, Sparks, MD, USA) supplemented with 5% sheep blood (BAP). A negative extraction control was included, which consisted of 150 μl of TE buffer. All spore samples were heated at 95°C for 30 min prior to DNA extraction to inactivate viable spores. Following DNA extraction, samples were stored at −20°C prior to analysis.
qPCR conditions for commercial DNA extraction kit down selection
Down selection of DNA extraction kits involved performing qPCR on extracted samples to determine total DNA yield for each kit. The primers and probe were designed in-house targeting the cpn60 gene of the Bacillus cereus group. qPCR was performed in 25-μl reaction volumes containing 12·5 μl of Roche probe mastermix (Roche, Mississauga, ON, Canada) with 0·5 μmol l−1 of each cpn60_F (5′-GGTATCGAAAAAGCTGTAGTT-3′) and cpn60_R (5′-TCAGCAGCAGAAATAGCAG-3′), 0·2 μmol l−1 of the probe cpn60_P (5′-FAM-AATCGAAGGTAAATCTTCTATCGCA-BHQ1-3′) and 5 μl of DNA template in the LightCycler 480 II (Roche). Thermocycling conditions were as follows: 5 min at 95°C for enzyme activation followed by 50 cycles of denaturation for 10 s at 95°C, annealing for 15 s at 52°C and extension for 10 s at 72°C. Bacillus anthracis Sterne strain DNA was quantified using the Nanodrop ND-8000 spectrophotometer (ThermoScientific, Waltham, MA, USA) and adjusted to 5·7 ng μl−1 representing 1 × 106 genomic equivalents based on the following calculation for total genomic DNA in a single cell: amount of DNA (fg) = 5·4 × 106 bp × 660 Da bp−1 × 1·6 × 10−27 kg Da−1 × 1 × 1018 fg kg−1 (Rodríguez-Lázaro et al. 2005) assuming a 5·4-Mbp genome (Kolsto et al. 2009). A standard curve was generated for comparison between qPCR plates from tenfold serial dilutions (from 1 × 106 to 1 × 101) of the B. anthracis Sterne strain stock. The LightCycler 480 software (Roche) was used for crossing point (Cp) determination using the second derivative method. Total DNA yield for all kits, expressed in genomic equivalents (GE), was calculated by multiplying the amount of DNA estimated in the qPCR by a factor based on the minimum elution/resuspension volume (e.g. if the final volume was 50 and 2 μl was used as template, the factor to estimate total yield is 25; Fig. 1).
Preparation of spiked food samples
Apple juice, whole milk (3·25% milk fat), processed meat (black forest ham) and prewashed bagged salad (romaine lettuce) were purchased from a local grocery store and used for the food spiking experiments as previously described (Shields et al. 2012). Briefly, B. anthracis spores were added to 25 ml of liquid foods or 50 g of solid foods to achieve an inoculation of approx. 375 and 750 CFU for the low and high spiking amounts, respectively. Spiking amounts were selected based on preliminary data that showed recovery near the lower detection limit for the three kits. The milk samples were diluted with 25 ml of buffered peptone water (BPW) containing 1% Tween-20 (pH 7·2; BPWT), and the apple juice sample was diluted with 25 ml of BPW containing 0·2 mol l−1 Na2HPO4 and 1% Tween-20 (pH 8·0). For spore capture in solid foods, 50 g of sliced black forest ham or 50 g of bagged salad was placed into a stomacher bag and inoculated with B. anthracis spores and gently massaged to mix. Fifty millilitres of BPWT (pH 7·2) was added, and the mixture stomached for 2 min with the Stomacher 400 Circulator (Seaward Ltd., Port Saint Lucie, FL, USA) at 230 rpm. The liquid was further passed through a sponge filter, and a 50-μm stainless steel mesh filter using a vacuum pump and the filtrate was collected for analysis.
Following the preparation of spiked food samples, 50 ml of the prepared food sample was mixed with 1 mg (50 μl) of Pathatrix beads (Matrix MicroScience Ltd., Newmarket, UK) functionalized with rabbit anti-B. anthracis polyclonal antibody (Tetracore Inc., Rockville, MD, USA), which were previously shown to be highly specific for B. anthracis (Shields et al. 2012). The beads were mixed, captured and washed using the iCropTheBug protocol described by Shields et al. (2012). Following the capture and wash steps, samples were resuspended in 250 μl of PBS and divided into five 50-μl aliquots. Two of the aliquots were plated on BAPs to determine the approximate number of CFUs per aliquot and the remaining three aliquots used for DNA extraction (one aliquot per kit). Experiments involving each food matrix and B. anthracis concentration were performed in triplicate. Samples were allotted such that each kit received an aliquot of each replicate to reduce variation between DNA extraction kits resulting from variance between IMS replicates.
The three DNA extraction kits from the down-selection phase exhibiting the highest DNA yield were selected for evaluation of DNA extraction from B. anthracis Sterne spores following IMS from food matrices.
Evaluation of PCR inhibition in IMS extractions
To investigate the potential of PCR inhibition, Yersinia pestis gDNA was used as an internal amplification control with primers and probe developed previously (Amoako et al. 2010). Approximately 800 genomic equivalents of Y. pestis, an amount previously determined to yield a Cp of approx. 33 cycles in the absence of inhibition, were spiked into each qPCR containing 0·5 μmol l−1 internal amplification control (IAC) primers ADRI_Yp5F (5′-CGTTATCGCAGCGGTAGAAG-3′) and ADRI_Yp5R (5′- CGCCAACGGTTGAGTCGG-3′), 0·35 μmol l−1 IAC probe ADRI_Yp5P (5′- FAM-TGAAAAAACTGTCTGTACCTTGCTCTGATTCCA-BHQ1-3′) and 3 μl of extracted DNA or TE buffer for the control. qPCR was performed on the LightCycler 480 II (Roche) with a 5-min preheating step at 95°C to activate the Taq polymerase followed by 50 cycles of denaturation for 10 s at 95°C, primer annealing for 15 s at 58°C and extension for 20 s at 72°C.
Comparison of DNA extraction kits with IMS samples by qPCR
Due to the possible presence of Bacillus spp. in the commercial food samples tested, a multiplex qPCR assay specific for B. anthracis (Janse et al. 2010) was utilized for the evaluation of DNA extraction kits from food samples. qPCRs in 20 μl volumes were performed as described previously with 2X Probes Master mastermix (Roche), and data for the pXO1 target (cya) were used for evaluating the different extraction methods. qPCR was performed on the LightCycler 480 II (Roche) with a 5-min preheating step at 95°C followed by 50 cycles of denaturation for 5 s at 95°C and primer annealing and extension for 15 s at 60°C. Colour compensation was performed as per the LightCycler protocol (Roche). Data for the pXO1 target were compared between extraction kits due to enhanced detection sensitivity compared with the chromosomal target.
Down selection of commercial DNA extraction kits
The nine kits tested varied substantially in the amount of time required to process 18 samples, ranging from 1 h for the FastDNA spin kit to 4 h for the Platinum Path (Table 1). Although automation is possible for the most time intensive kits (Platinum Path and PrepSEQ), specialized equipment is required and was not practical for this study. The Platinum Path, Wizard and FastDNA SPIN kits performed the worst as they did not yield detectable DNA for three or more of the spore dilutions tested (Fig. 1). The MasterPure Gram Positive, PowerFood and PrepSeq kits yielded the most consistent results in extracting DNA from B. anthracis spores across the spore amounts tested (Fig. 1). There was no difference between the three kits at low spore numbers (1 × 102, 1 × 103 and 1 × 104); however, the MasterPure and PowerFood kits yielded the most DNA from the highest two amounts of spores tested (1 × 105 and 1 × 106). These kits were selected for comparison of their ability to extract DNA from B. anthracis spores following IMS from spiked food matrices.
IMS of Bacillus anthracis spores from food
The average recovery of B. anthracis spores from spiked food samples using IMS ranged from 52 to 68 CFU for the lower to 123–138 CFU per aliquot for the upper amount of spores tested (Table 2). Based on these results, the predicted total CFU captured from food samples for the low and high amounts of spores is 260–340 and 615–690 CFU, respectively. There was no difference in recovery of B. anthracis spores from the different food matrices (Table 2).
Table 2. Average number of Bacillus anthracis spores (Average CFU) per aliquot for DNA extraction and corresponding qPCR results (Cp) for pXO1 target detected from different DNA extraction methods for B. anthracis spores isolated from food by immunomagnetic separation
|Apple juice||65 ± 14||37·7 ± 0·50 b||37·61 ± 1·44||ND||40·87c|
|133 ± 9||36·5 ± 0·76||37·30 ± 0·65||ND||41·85 ± 0·06b|
|Ham||68 ± 8||35·7 ± 0·30||ND||ND||ND|
|138 ± 13||34·5 ± 0·50||ND||ND||ND|
|Milk||66 ± 4||37·8 ± 2·06||39·75 ± 1·14||ND||ND|
|131 ± 4||35·3 ± 0·67||37·30 ± 1·71||ND||ND|
|Salad||52 ± 16||36·6 ± 0·31||38·70 ± 1·40||ND||ND|
|123 ± 6||35·2 ± 0·46||36·55 ± 0·45||ND||ND|
Evaluation of PCR inhibition in DNA extractions after IMS
Samples tested for the presence of PCR inhibitors by performing qPCR on samples subsequently spiked with Y. pestis chromosomal DNA yielded nearly identical Cps in relation to the control condition (TE buffer in place of a DNA extract) indicating that there was no PCR inhibition. The Epicenter MasterPure Gram Positive extractions had an average Cp of 33·54 ± 0·12; the MoBio PowerFood, 33·53 ± 0·15; ABI PrepSeq, 33.59 ± 0·07; and the control yielded an average Cp of 33·56 ± 0·14 (Table 2).
DNA extraction and detection by qPCR following IMS
The negative extraction controls (food samples not spiked with B. anthracis but ran through the IMS protocol) for all three kits yielded no qPCR signals. The MasterPure Gram Positive kit was the only kit that yielded enough DNA for detection of B. anthracis via qPCR across all spore numbers and matrices tested and consistently detected as low as approx. 60 CFU with an average Cp of 37·7 ± 0·50, 35·7 ± 0·30, 37·8 ± 2·06 and 36·6 ± 0·31 for IMS from apple juice, ham, milk and salad, respectively. The PowerFood kit did not yield enough DNA for detection by qPCR for any of the samples. Similar results were observed with the PrepSeq kit, with the exception of apple juice, which was detected in one of three replicates at a Cp of 40·87 for the low amount of B. anthracis spores and two of three replicates for the high amount at an average Cp of 41·85. Diluted MasterPure extractions, conducted to simulate a 100-μl elution volume, were detected at similar Cps as the undiluted samples (Table 2).
A total of nine commercially available DNA extraction kits were evaluated using the manufacturers’ recommended protocols for their ability to isolate DNA from B. anthracis spores. DNA extraction kits are often designed for extracting DNA from pure cultures consisting of highly concentrated cells and are not necessarily optimized for scenarios requiring high sensitivity. A previous investigation into DNA extraction from low numbers of B. anthracis spores has been undertaken in the past; however, the yields for low spore concentrations, for example 200–480 ng of DNA from 10 spores, are problematic as they are orders of magnitude higher than the theoretical yield (51 fg) (Dauphin et al. 2009), which is also outside the detection limit for the instrument used for quantification (Nanodrop ND-8000). In the present study, DNA extractions were not quantified due to the fact that theoretical yields for all dilutions but the 1 × 106 concentration are below the detection threshold of the Nanodrop ND-8000 (Nanodrop). Furthermore, the 1 × 106 concentration would yield DNA at the lower limit of the instrument's range and have expected variance near 50% (assuming 100% yield) according to the Nanodrop ND-8000 Manual. As we were primarily interested in kit sensitivity, qPCR was relied upon for estimation of DNA extraction yields. In a previous study, the UltraClean kit (MoBio) was identified as being the most efficient among five kits tested for inactivating B. anthracis spores and was also the most sensitive and consistent kit for detection of B. anthracis spores from swabs and powders (Dauphin et al. 2009). In the present study, this kit performed well at the lowest spore concentration tested, with results similar to the top performers, but it did not perform well at higher concentrations.
The presence of potential PCR inhibitors in complex foods such as ham and salad is usually a concern, even with IMS samples. To ensure qPCR detection results were representative of the actual amount of spiked spores and not influenced by the presence of PCR inhibitors copurified during IMS, extractions from food samples were tested for the presence of PCR inhibitors. Samples spiked with known amounts of Y. pestis DNA yielded identical qPCR results between spiked extractions for all three kits as well as the control, indicating PCR inhibition was not a factor and the qPCR results represent actual yields.
The MasterPure Gram Positive kit was the only kit that consistently yielded enough DNA from IMS samples for detection of B. anthracis via qPCR for all spore numbers and matrices tested. The PowerFood and PrepSeq kits did not perform well with IMS samples, and both kits recommend larger elution volumes (100 and 50 μl, respectively) than the MasterPure Gram Positive kit (35 μl), which would result in lower DNA concentrations assuming 100% recovery of DNA. To ensure the differences in detection were not due to higher elution volumes diluting DNA to below the detection threshold, aliquots of the MasterPure Gram Positive extractions were diluted in TE to simulate a 100-μl resuspension, and qPCR was repeated with the diluted samples. Diluted samples were detected equally as well, indicating that the differences in detection were not the result of different elution volumes (Table 2). The lack of detection of IMS samples with the PowerFood and PrepSEQ kits may be the result of reduced extraction efficiency due to particulate matter retained from the food matrix or interference caused by the presence of immunomagnetic beads. Alternatively, it is possible that mother cells (Tocheva et al. 2011) or vegetative cells that failed to sporulate were present in the spore preparation. The IMS antibodies used in this study have been shown to be highly specific (Shields et al. 2012); in addition, we have observed that they do not capture vegetative cells efficiently nor do they bind DNA. As such, DNA from these sources would have been detected without IMS leading to higher perceived extraction yields during the down-selection phase.
A previous study using the Epicenter MasterPure kit reported detection limits as low as 100 B. anthracis spores when a preheating step designed to induce germination was employed (Luna et al. 2003). This kit differs slightly from the MasterPure Gram Positive kit, as there is no ready-lyse lysozyme step, and it appears to use a different lysis buffer (T&C lysis solution vs Gram-positive lysis solution – constituents unknown). Both kits employ ethanol precipitation of DNA, and thus, the final volume used to resuspend DNA can be decreased to enhance downstream detection capabilities. Kits using silica columns (PowerFood) or magnetite/silica-based magnetic beads often require a large volume of buffer (50–200 μl) for efficient elution of DNA. A second elution is recommended with such kits to improve total yield. Ethanol precipitation can be performed following elution to concentrate DNA, but adding such a step increases the extraction time.
The inoculum levels for B. anthracis spores in food used in this study are closer to the LD50 for subcutaneous infection (a few to a few hundred spores) than inhalation or oral infection (tens of thousands to millions of spores; Mogridge et al. 2010); however, B. anthracis spores have been shown to germinate and grow in some foods, including ground beef (Tamplin et al. 2008) and processed liquid eggs (Khan et al. 2009), and thus, a small contamination can have significant consequences. Furthermore, having the ability to detect genomic signatures with great sensitivity is important for early warning.
In conclusion, a comprehensive down selection of extraction kits yielded three candidates (Applied Biosystems PrepSeq Nucleic Acid Extraction Kit; Epicenter MasterPure Gram Positive DNA Purification Kit; MoBio PowerFood Microbial DNA Isolation Kit) for evaluation with IMS samples from food. Although the three kits had similar yields with spore preparations, only the Epicenter MasterPure Gram Positive kit facilitated detection of approx. 60 CFU isolated from 25 ml or 50 g of food matrix by IMS. The work described here demonstrates the Epicenter MasterPure Gram Positive kit as a highly sensitive method for extraction of DNA from B. anthracis spores isolated from food by IMS and facilitates highly sensitive downstream DNA-based detection technologies. This work is novel for B. anthracis detection in food and has potential application in food borne bioterrorism response. Future work on additional food matrices is warranted.
We thank Dr Elizabeth Golsteyn Thomas, OIE Reference Laboratory, for Anthrax, Canadian Food Inspection Agency for providing the Bacillus anthracis Sterne strain used in this study. The authors acknowledge the useful suggestions of Kym Antonation and Drs Soren Alexandersen, Elizabeth Golsteyn Thomas, and Matthew Gilmour regarding the manuscript. This work was funded by the Defence Research Development Canada Centre for Security Science, Chemical, Biological, Radiological, Nuclear and Explosive Research Technology Initiative (CRTI) funding from CRTI 08-0203RD.