Evaluation of two methods for detection of viable Bacillus anthracis simulant spores in maritime environmental samples

Abstract

Analytical methods exist to detect biothreat agents in environmental samples during a response to biological contamination incidents. However, the coastal zone facilities and assets of the US Coast Guard (USCG), including response boats in diverse geographical areas and maritime environmental conditions, can pose complex and unique challenges for adapting existing analytical detection methods. The traditional culture (TC) and the rapid viability polymerase chain reaction (RV-PCR) methods were evaluated for their compatibility for maritime environmental surface and grab sample analysis to detect spores of Bacillus thuringiensis subspecies kurstaki (Btk), a surrogate for Bacillus anthracis. The representative samples collected from a USCG installation included surfaces, such as aluminum on boats, nonskid tread on decks of watercraft, computer touchscreens, and concrete piers, and grab samples of boat washdown water, soil, vegetation, and gravel from surrounding areas. Replicate samples were spiked with Btk spores at two to three tenfold increasing levels and analyzed. Out of a total of 150 samples collected and analyzed, the TC method gave 10 false-positive and 19 false-negative results, while the RV-PCR method-based analysis resulted in 0 false-positive and 26 false-negative results. An abundance of microbial background and particulates in some samples interfered with true results, while both methods gave similar results for samples with low microbial background and particulates. Improved and high-throughput sample processing methods are needed for analysis of complex environmental samples.

Introduction

Following a biothreat agent contamination incident such as the release of spores of Bacillus anthracis, the causative bacterial agent for anthrax, accurate sample analysis can help determine the extent and magnitude of contamination, which informs responders for selection of decontamination strategies and helps determine the success of decontamination. Sample analysis results help inform the responsible authorities in making reoccupancy decisions. Extensive protocols for sampling and analysis have been developed to respond to environmental contamination involving biological threat agents (the US Environmental Protection Agency’s (EPA) Environmental Sampling & Analytical Methods [ESAM] Program [US EPA, 2022]). However, response to any contamination incident is specific to the affected site and surrounding environment. The USA has coastlines on multiple oceans, gulfs, and great lakes and, therefore, has thirty coastal states. Coastal cities in those states are ports of entry, shipping ports, and popular tourist destinations and, hence, have a significant portion of the US population. The coastal zone facilities and assets, including US Coast Guard (USCG) boats, in diverse geographical locations and maritime environmental conditions, can pose complex and unique challenges for adapting existing methods or developing new methods for sampling, analysis, and decontamination to respond to such contamination incidents. The performance of the analytical methods may, in part, depend on the coastal zone surfaces and materials being sampled and analyzed.

The traditional microbiological culture-based method remains the gold-standard for analysis of environmental samples collected during a response to bioterrorism incidents such as anthrax. The rapid viability polymerase chain reaction, or RV-PCR (Létant et al., 2011; Shah, 2017) method, which integrates high throughput sample processing, short-incubation broth culture, and highly sensitive and specific real-time PCR assays, can also be used to detect live biothreat agents from environmental samples. In the present study, these two analytical methods were evaluated for their compatibility for maritime environmental surface and grab sample analysis to detect spores of Bacillus thuringiensis subspecies kurstaki (Btk), a surrogate for Bacillus anthracis.

Surface and grab samples were collected from a USCG installation. The samples were transported to the lab, spiked with the Btk spores at different levels, processed to recover spores, and analyzed using the traditional culture and the RV-PCR methods. This study provides data and information that can be used to inform sampling and analysis operations and strategies following a biothreat contamination incident impacting coastal zone facilities and assets.

Materials and methods

Sampling methods and Bacillus thuringiensis subspecies kurstaki endospores

Multiple surfaces were sampled at a USCG base in Portsmouth, VA in two campaigns. Campaign 1 occurred on a clear, sunny day on 04 November 2020, from approximately 0900 to 1800 h. Early morning temperature and relative humidity (RH) were 20 °C and 50%, to a mid-day high temperature of 23 °C and 43%, and an end-of-day condition of 17 °C and 65%, respectively. Sampling campaign 2 occurred on a mostly cloudy morning from 0800 to 1200 h on 26 March 2021. Early morning conditions were 22 °C and 80% RH, with the temperature rising to 25 °C and RH dropping to 66% by noon. Rain occurred on 25 March 2021. In all sampling events, no dew or unevaporated rain was present, and surfaces or materials were dry when sampled. Surface samples were collected using sponge sticks (SS, Item #SSL10NB, 3 M, St. Paul, MN) and 37-mm vacuum filter cassettes (VFC, Item # 225–3-01, SKC Inc., Eighty Four, PA). Sampling included surfaces such as aluminum, nonskid tread, and computer touchscreens on boats, and concrete piers. Grab samples included gravel, soil, vegetation (grass), and washdown water of a small response boat.

Surface samples were collected using SSs and VFCs following previously described collection methods (Calfee et al., 2013; Rose et al., 2011). Gravel was collected by filling a 1-L Nalgene bottle (Item # 02-893D, Fisher Scientific, Hampton, NH) to the one-half full mark, which is ~ 900 g, as described previously (Boehm et al., 2009; Serre & Oudejans, 2017). Soil was collected into a 1-L Nalgene bottle using a garden hand spade to scrape and scoop the top 1 to 2 inches of soil. Vegetation was collected by cutting the grass just above the soil, then folding the grass to fit into 1-L Nalgene bottles as previously described (Mikelonis et al., 2020). Vessel washdown water samples were generated by spraying freshwater from a USCG dock water supply using a nozzle similar to a heavy-duty adjustable brass spray nozzle (Item # 800, Dramm, Manitowoc, WI), which generated an estimated ^< 30-cm-diameter cone at a 1-m distance. An estimated area of 4 m² of the exterior surface of a Response Boat Small II (RBS), including the glass windows, aluminum roof and deck, nonskid tread, and some exposed aluminum was washed in a ^< 5-min period with a nozzle-to-surface distance of 0.5 to 1.5 m. The water stream was collected into 1-L Nalgene bottles from the scupper of the RBS and/or side drains of the medium response boat with an average fill rate of approximately 2 L/min at peak flow.

Bacillus thuringiensis subspecies kurstaki with T1B2 (Btk T1B2) genetic barcode (Buckley et al., 2012) was used as a simulant for Bacillus anthracis. The T1B2 barcode makes it distinguishable from wild-type/naturally occurring Btk at the molecular level, enabling real-time polymerase chain reaction (PCR) (hereafter referred to as PCR) detection of the test strain. Btk T1B2 endospores (spores) were prepared by inoculating 500 mL of sporulation medium with 50 mL of an overnight culture of Btk T1B2 grown in Brain Heart Infusion Broth (BHIB; Item # DF0037-17–8, Fisher Scientific, Hampton, NH) in baffled 3-L Fernbach flasks at 30 ± 2 °C and 200 rpm for a minimum of 72 h until > 90% of spores appeared phase bright under phase contrast microscopy. The sporulation medium was modified G medium (Kim & Goepfert, 1974). The sporulated culture was pelleted, suspended in sterile deionized (DI) water, heat shocked for 1 h at 60 °C, then washed twice using sterile DI water before being suspended in sterile DI water. All centrifugation steps were performed at 10,000 × g for 12 min. Spores were spread plated and streaked on Trypticase^™ Soy Agar (TSA; Item # B21283X, Fisher Scientific, Hampton, NH) pre- and post-heat shock to assess spore maturity (resistance to heat) and purity (single morphology). Spore stocks were stored at 2 to 8 °C for the duration of the study. The stock titer was 2 × 10⁸ CFU/mL.

Samples collected in the field were spiked by pipetting 100 μL of the appropriate Btk T1B2 spore stock dilution onto SSs, VFCs, gravel, and soil in a dropwise fashion (5 μL per droplet); washdown water samples were spiked with 100 μL volume; and vegetation samples were spiked with 500 μL volume and a 25-μL droplet size to enable downward movement and distribution of the liquid into the grass grab sample. For each experiment, each spore spike level was separately spread plated on TSA and incubated to determine actual spiked spore load as shown in Supplemental Table 1.

Spore recovery from Btk T1B2 spiked SSs and VFCs was performed following the processing procedures described in the RV-PCR section of the Protocol for Detection of Bacillus anthracis in Environmental Samples During the Remediation Phase of an Anthrax Incident, 2nd edition (Shah, 2017) (also referred to as the EPA Protocol). The resultant spore recovery suspension was portioned in half for traditional culture (TC) and RV-PCR analytical methods. This portioning of samples is a modification to the established analytical methods described in the EPA Protocol (Shah, 2017).

To recover spores from Btk T1B2 spiked gravel and vegetation, 500 mL of phosphate-buffered saline with 0.05% Tween-20 (PBST) and pH 7.4 (Item # P0201, Teknova, Hollister, CA) was added, and the 1-L Nalgene bottle was shaken vigorously for 2 min by hand. The sample was then allowed to settle for 30 s, and the spore recovery suspension was poured into a clean container. The spore recovery suspension (up to 250 mL for gravel; up to 500 mL for vegetation) was then filtered through a mixed cellulose ester (MCE) membrane (Item # 4800, MicroFunnel^™ Filter Funnel, Pall Corporation, Washington, NY), the spores were recovered from the MCE membrane following the RV-PCR spore recovery method (MCE placed in a 50-mL centrifuge tube), followed by two sequential 10-mL vortex steps using PBST with 30% ethanol (PBSTE), and finally split in half (10 mL each) for TC and RV-PCR analytical methods. The vessel washdown water samples were extracted by manual shaking as described for gravel and vegetation with up to 500 mL suspension was then filtered through the MCE membrane. The spores were recovered from the MCE membrane in PBSTE as described for gravel and vegetation and then split in half (10 mL each) for TC and RV-PCR analytical methods.

To recover spores from soil, 10 g of soil was weighed in a 50-mL centrifuge tube, hydrated with 40 mL of PBST, vortex-mixed for 30 s, bath sonicated for 10 min to disaggregate particles, manually shook for 2 min, and centrifuged for 5 min in a swinging bucket (1,000 × g). The supernatant (spore recovery suspension) was then transferred to a clean tube, and the spore recovery suspension and pellet (concentrated debris and microorganisms post-centrifugation) were heat shocked for 1 h at 70 °C to reduce background microbial load. The resultant suspension was split in half (20 mL each) for TC and RV-PCR analytical methods.

Test matrix

The maritime sample type (surface or material sampled) and number of sample replicates analyzed are shown in Table 1. In total, 150 samples were collected and analyzed, comprising 57 SS samples, 48 VFC samples, and 45 grab samples. Additionally, a subset of controls was analyzed alongside the collected samples that consisted of reagents that were only handled in the laboratory or opened temporarily in the field, but not used to sample a surface.

Table 1.

Test matrix for collected real-world maritime samples

Maritime surface/material sampled	Sample ID	Sampling method	Replicate samples per spore load	Target spore load colony-forming units (CFU)
RBS marine grade aluminum	SBMGAL	SS	6, 6, 6	0, 300, 3,000
Nonskid tread	NSKID	SS	7, 7, 7	0, 300, 3,000
Touch screen	TCHSCRN	SS	6, 6, 6	0, 300, 3,000
Nonskid tread	NSKID	VFC	8, 8, 8	0, 300, 3,000
Concrete pier	CONPIER	VFC	8, 8, 8	0, 300, 3,000
RBS washdown water	SBWASH	Grab	3, 3, 3, 3	0, 300, 3,000, 30,000
Gravel parking lot	GRAVEL	Grab	3, 3, 3	0, 300, 3,000
Loam soil	SOIL	Grab	3, 3, 3, 3	0, 3,000, 30,000, 300,000
Grass vegetation	VEGE	Grab	3, 3, 3	0, 300, 3,000, 30,000

Analytical methods

Traditional culture method

Spore recovery suspension was vortex-mixed on high for 30 s, spread plated or concentrated onto MicroFunnel filter membranes, and incubated overnight at 30 ± 2 °C. For spread plating, 0.1 mL of neat and/or 1:10 serial dilutions in PBST were spread plated onto TSA. For MicroFunnel filter plating, ≥ 1-mL volumes were concentrated onto 0.45-μm Metricel^® membranes (Item # 4804 or 4805, Pall Corporation, Washington, NY) by prewetting the membrane with 5 mL of PBST and then adding ≥ 1 mL of spore recovery suspension to 10 mL of PBST within the MicroFunnel, followed by a 10 mL PBST rinse. Metricel membranes were removed using sterile forceps and applied to TSA. Btk T1B2 colonies appear flat or slightly convex, are 2 to 5 mm in diameter with edges that are slightly irregular, and have an appearance described as ground-glass. All culture counts are reported as presumptive Btk T1B2 colonies with spore recovery values calculated using spread plates containing 25 to 250 colony-forming units (CFU) per plate and 20 to 80 CFU per membrane plate, when possible. For TC, spore recovery suspension volumes plated were 2 mL and 8 mL for SS samples; 1 mL and 3 mL for VFC samples; 0.1 mL, 2 mL, and 8 mL for washdown water samples; 1 mL and 4 mL for gravel samples; 0.1 mL for soil samples; and 0.1, 2 mL, and 8 mL for vegetation samples.

Isolated presumptive Btk T1B2 colonies were screened using PCR by transfer of a portion of the presumptive colony with an inoculating loop or needle into 100 μL of PCR-grade water in a 1.5-mL tube. The colony suspension was heated for 5 min at 95 ± 2 °C to lyse the cells, centrifuged at 18,407 × g for 2 min, and the supernatant was analyzed using PCR. If presumptive Btk T1B2 colonies were not confirmed PCR positive from spread or membrane plates, then PCR analysis of the enriched SS, MCE membrane from VFC or MicroFunnel, or pellet was performed.

The extracted SS, MCE membrane from VFC or MicroFunnel, or pellet was enriched by adding 25 mL of BHIB and incubating at 30 ± 2 °C for 24 to 48 h. If media appeared turbid and colonies were not confirmed PCR positive from the spore recovery suspension, 10 μL of BHIB enrichment culture were streaked for isolation of colonies in triplicate and incubated overnight at 30 ± 2 °C. If presumptive Btk T1B2 colonies were isolated, colony PCR was performed. If presumptive Btk T1B2 colonies were not isolated, and the BHIB was turbid from incubation of SS or MCE membrane, then 50 μL of turbid BHIB was centrifuged at 12,000 × g for 2 min, the supernatant discarded, and the pellet suspended in 100 μL of PCR-grade water followed by heating for 5 min at 95 ± 2 °C to lyse cells, centrifuged at 18,407 × g for 2 min, and the supernatant was processed using PCR.

RV-PCR method

The spore recovery suspension was vortex-mixed on high for 30 s and then allowed 30 s of settle time before transferring a volume (up to 12 mL at a time) to a Cytiva Whatman^™ Autovial Syringeless Filter (filter vial) (Fisher Scientific Cat. AV125NP-UAQU) until full volume was concentrated by filtration or a maximum filtration time of 1 h. Two buffer washes were then performed according to the EPA Protocol (Shah, 2017). The first wash was 12.5 mL of cold (4 °C) high salt buffer (10X PBS, Item # BP3994, Fisher Scientific, Hampton, NH) followed by 12.5 mL of cold (4 °C) low salt wash buffer (1X PBS, Item # BP24384, Fisher Scientific, Hampton, NH). If filtration of full volume of spore recovery suspension took longer than 15 min, reduced volume of 10X PBS and 1X PBS (5 mL) was applied. If the spore recovery suspension took ≥ 1 h for filtration, the remaining volume was removed, and the 10X PBS and 1X PBS washes were omitted. All filtration steps were performed using vacuum manifold (Item # 1,701,232–1 and 1,701,232–2, Pacon Manufacturing, Inc., Livermore, CA) operating at 5 to 10 psi. The top portion of the manifold was then removed and placed into a capping tray (Item # 1,701,233, Pacon Manufacturing, Inc., Livermore, CA) with Luer lock caps to seal the filter vials. Five (5) mL of cold BHIB was then added to each filter vial, the vials were capped and then vortex-mixed for 10 min on a setting of seven on a platform vortex (Item # 58,816–115, VWR, Radnor, PA). Following the vortex step, the broth culture was mixed by pipetting up and down ~ 10 times, and a 1-mL aliquot was removed and stored at −20 °C as the time zero (T₀) aliquot. The capped filter vials were then incubated overnight for ~ 16 h (time final, T_f) in an incubator shaker set to 30 ± 2 °C and 230 rpm. Following overnight incubation, a T_f aliquot was removed following the same procedure described for the T₀ aliquot. DNA was extracted from T₀ and T_f aliquots as described in the EPA Protocol using a manual Magnesil Blood Genomic, Max Yield System Kit (Item # MD1360, Promega, Madison, WI) prior to PCR analysis (Shah, 2017).

Real-time PCR assay

Real-time PCR was performed on an Applied Biosystems 7500 Fast Real-Time PCR Instrument (Waltham, MA). The PCR reaction targeted the barcode T1B2 utilizing the specific tag 2 primer sequences previously described (Buckley et al., 2012), coupled with a TaqMan^® probe designed using the PrimerQuest Tool (Integrated DNA Technologies, Coralville, IA). Each PCR reaction consisted of 12.5 μL of TaqMan Fast Advanced 2 × PCR Mix (Item # 4,444,556, Applied Biosystems, Waltham, MA), 5 μL of DNA extract, 0.1 μL of Platinum Taq Polymerase (Item # 10,966,034, Invitrogen, Waltham, MA), 1 μL each of 25 μM forward and reverse primers (GGTACAAGCAACGATCTCCAGAAT and TGAAGGTTAATTAGCGCATTTGAA), 2 μM probe (FAM-CGCCGACGCTTTACATACTATGAGAGG-MGBNFQ), and 4.4 μL of PCR-grade water. Thermal cycling conditions were as follows: 2 min at 50 °C, 2 min at 95 °C, and 45 cycles of 3 s at 95 °C and 30 s at 60 °C. All analyses used a ΔRn (magnitude of reporter dye signal) threshold to generate cycle threshold (Ct) values. Samples were analyzed in 96-well plates in triplicate reactions and contained PCR controls, four positive control reactions containing 50 pg of DNA extracted from Btk T1B2, and four no template control (NTC) reactions. Acceptance criteria for PCR were that all positive controls needed to cross the threshold to generate a Ct value and no NTCs could cross the threshold.

Data interpretation

For TC PCR analysis, a sample was considered positive if a presumptive Btk T1B2 colony or BHIB enrichment PCR reaction generated an average Ct of ≤ 40 for the PCR target. For RV-PCR, a delta Ct (ΔCt) value was calculated by subtracting the average Ct value generated by the T_f aliquot from the average Ct value generated by the T₀ aliquot. A ΔCt value of ≥ 9 for the PCR target indicated viable Btk T1B2 spores were detected in the sample and were considered a positive result.

Each sample spiked with Btk T1B2 spores resulted in a true positive or false-negative value. A true positive was defined as a sample spiked with Btk T1B2 spores that had a PCR positive analytical method result, and a false negative as a sample that was spiked with Btk T1B2 spores, but the PCR analytical method was negative. Samples that were not spiked with Btk T1B2 spores resulted as a true negative or false positive. A true negative was defined as a zero-spiked sample that had a PCR negative analytical method result and a false positive as a zero-spiked sample that had a PCR positive analytical method result.

Results and discussion

Surface samples

Surface samples from a USCG coastal zone facility in Portsmouth, VA, were collected from response boats (SBMGAL, NSKID, and TCHSCRN) that were likely to be close personnel contact zones, and the surrounding maintained grounds (gravel parking lot, grass vegetation from a lawn that is regularly mowed, and soil that supports the growth of the grass lawn) were sampled to understand the impact of background microorganisms and inert contaminants that may interfere with the TC and RV-PCR analytical methods. Surface sampling utilized SS wipes on smooth surfaces (SBMGAL and TCHSCRN) and one rough surface (NSKID). Additionally, VFCs were utilized to sample rough surfaces (NSKID and CONPIER). Samples were recovered from active response boats that were used for excursions and were on a standard maintenance and usage schedule. Surface samples from two SBMGAL, NSKID, and CONPIER surfaces with three replicates from each unit to diversify potential background interferents was collected. The small size of TCHSCRN, typically 12 × 12 inches, necessitated the sampling from three or four boats to collect a minimum of 18 replicates. The surrounding grounds were sampled using a grab technique to collect composite samples of GRAVEL, VEGE, SBWASH, and SOIL that were mixed and parsed into sample replicates to represent the grounds of this specific base.

Sponge stick samples

Table 2 gives the positive results for each surface sampled by environmental sampling technique and at each nominal spore load per sample. For each of the surface types (SBMGAL, NSKID, and TCHSCRN) sampled using an SS, which was subsequently spiked in the laboratory, Btk T1B2 spores were confirmed to be present by PCR analysis (positive) at the 300 and 3000 CFU target (nominal) spore load per sample for both TC and RV-PCR analytical methods. For the TC method, positive samples were confirmed by PCR analysis of presumptive Btk colonies isolated from the spore recovery suspension on TSA membrane plates or PCR analysis of the BHIB enrichment culture of the extracted sponge by either presumptive Btk colonies isolated from streak plates of the culture or of the liquid culture. The threshold for a positive PCR reaction was an average Ct value of ≤ 40. For the RV-PCR method, samples were confirmed positive with a ΔCt value of ≥ 9, which is calculated from the average pre- and post-enrichment (T₀ and T_f) Ct values.

Table 2.

Percent positive results for surface samples using TC and RV-PCR analytical methods

Maritime surface/ material sampled	Sample ID	Sampling method	Nominal spore load (CFU)	TC			RV-PCR
				Samples positive	Total samples	% Samples positive	Samples positive	Total samples	% Samples positive
RBS marine grade aluminum	SBMGAL	SS	0	2	6	33	0	6	0
			300	5	6	83	6	6	100
			3,000	6	6	100	6	6	100
Nonskid tread	NSKID		0	3	7	43	0	7	0
			300	5	7	71	6	7	86
			3,000	7	7	100	7	7	100
Touch screen	TCHSCRN		0	1	6	17	0	6	0
			300	6	6	100	6	6	100
			3,000	6	6	100	6	6	100
All surfaces sampled	All surfaces sampled		0	6	19	32	0	19	0
			300	16	19	84	18	19	95
			3,000	19	19	100	19	19	100
N/A	Lab/field controls		0	0	11	0	1	11	9
			300	4	4	100	4	4	100
			3,000	4	4	100	4	4	100
Nonskid tread	NSKID	VFC	0	1	8	13	0	8	0
			300	5	8	63	5	8	63
			3,000	8	8	100	7	8	88
Concrete pier	CONPIER		0	3	8	38	0	8	0
			300	7	8	88	6	8	75
			3,000	8	8	100	7	8	88
All surfaces sampled	All surfaces sampled		0	4	16	25	0	16	0
			300	12	16	75	11	16	69
			3,000	16	16	100	14	16	88
N/A	Lab/field controls		0	1	11	9	0	11	0
			300	6	6	100	6	6	100
			3,000	6	6	100	6	6	100

For all surfaces sampled at the 0 CFU spore load, 6 of 19 (32%) samples were positive (false positive) possibly due to a low level of cross-contamination or nonspecific PCR amplification for the TC method and 0 of 19 (0%) were positive for RV-PCR. At the 300 CFU spore load, 16 of 19 (84%) samples were positive for the TC analytical method and 18 of 19 (95%) were positive for the RV-PCR analytical method. At the 3000 CFU spore load, 19 of 19 (100%) were positive for both TC and RV-PCR analytical methods (Table 2). The “Analytical assay performance” section includes a discussion of the overall impact of the higher frequency of false-positive results (when no spores were spiked) for the TC method than the RV-PCR method.

Vacuum filter cassette samples

As shown in Table 2, for each of the surface types (NSKID and CONPIER) sampled using a VFC, which was subsequently spiked in the laboratory, Btk T1B2 spores were confirmed positive at the 300 and 3000 CFU spore load for both TC and RV-PCR analytical methods. For all surfaces sampled at the 0 CFU spore load, 4 of 16 (25%) samples were false positive for the TC method possibly due to a low level of cross-contamination or nonspecific PCR amplification, and 0 of 16 (0%) were positive for RV-PCR. At the 300 CFU spore load, 12 of 16 (75%) samples were positive for TC analytical method, and 11 of 16 (69%) samples were positive for the RV-PCR analytical method. At the 3,000 CFU spore load, 16 of 16 (100%) were positive for TC analytical methods, and 14 of 16 (88%) samples were positive for the RV-PCR analytical method (Table 2). The “Analytical assay performance” section includes a discussion of the overall impact of the higher frequency of false-positive results (when no spores were spiked) in the TC samples than the RV-PCR samples.

Grab samples

As shown in Table 3, for grab samples, the lowest level of detection varied depending on the surface sampled. For all surfaces with 0 CFU spore load, zero samples were positive. At the 300 CFU spore load, the GRAVEL sample type had 1 in 3 (33%) positive samples for both TC and RV-PCR analytical methods, and all other grab samples had 0 of 3 (0%, soil was not tested at 300 CFU load). At the 3000 CFU spore load, SBWASH, GRAVEL, SOIL, and VEGE had positive sample replicates at 100%, 67%, 67%, and 33% for the TC analytical method and 0%, 33%, and 0% for the RV-PCR analytical method. At the 30,000 CFU spore load, SBWASH, SOIL, and VEGE were positive in 100% of samples for both TC and RV-PCR analytical methods (GRAVEL was not tested at this spore level). At the 300,000 CFU spore load, 100% of SOIL samples were positive in 100% of samples for both TC and RV-PCR analytical methods (no other grab samples were analyzed at this spore load). The “Analytical assay performance” section includes a discussion of the overall impact of the higher frequency of false-negative results (when spores were spiked) in the RV-PCR samples than the TC samples.

Table 3.

Percent positive results for grab samples using TC and RV-PCR analytical methods

Maritime surface/ material sampled	Sample ID	Nominal spore load (CFU)	TC			RV-PCR
			Samples positive	Total samples	% Samples positive	Samples positive	Total samples	% Samples positive
RBS washdown water	SBWASH	0	0	3	0	0	3	0
		300	0	3	0	0	3	0
		3000	3	3	100	0	3	0
		30,000	3	3	100	3	3	100
Gravel parking lot	GRAVEL	0	0	3	0	0	3	0
		300	1	3	33	1	3	33
		3000	2	3	67	1	3	33
Loam soil	SOIL	0	0	3	0	0	3	0
		3000	2	3	67	1	3	33
		30,000	3	3	100	3	3	100
		300,000	3	3	100	3	3	100
Grass vegetation	VEGE	0	0	3	0	0	3	0
		300	0	3	0	0	3	0
		3000	1	3	33	0	3	0
		30,000	3	3	100	3	3	100

Analytical assay performance

The percentage of samples that were true positive or negative, false positive, and false negative for samples with 0 CFU, 300 CFU, and 3000 CFU spore load are presented in Figs. 1 and 2 for both analytical methods. A true positive or negative sample means that the analytical method result matched the expected outcome (i.e., a sample spiked with Btk T1B2 resulted in detection of Btk T1B2 spores). For SS samples (Fig. 1a), 15 of 18 (83%) SBMGAL samples, 16 of 21 (76%) NSKID samples, and 17 of 18 (94%) TCHSCRN samples were true positive or negative for the TC analytical method compared to 18 of 18 (100%) SBMGAL samples, 20 of 21 (95%) NSKID samples, and 18 of 18 (100%) TCHSRN samples for the RV-PCR analytical method. For VFCs (Fig. 1b), 20 of 24 (83%) NSKID and CONPIER samples were true positive or negative for the TC analytical method compared to 20 of 24 (83%) NSKID samples and 21 of 24 (88%) CONPIER samples for the RV-PCR analytical method. For grab samples (Fig. 2), 6 of 9 (67%) SBWASH, 6 of 9 (67%) gravel, and 4 of 9 (44%) vegetation samples were true positive or negative for the TC analytical method compared to 3 of 9 (33%) SBWASH, 5 of 9 (56%) gravel, and 3 of 9 (33%) vegetation samples for the RV-PCR analytical method. Control samples, which consisted of clean material with low to no background microorganisms or inert material from sampling an environmental surface, only resulted in two false-positive results—a SS sample that had a ΔCt of 9.8 for the RV-PCR method (Fig. 1a) and a VFC sample had a PCR Ct value of 38.3 ± 6.3 when analyzing the BHIB enriched MCE membrane in the TC method (Fig. 1b). Both of these low level false-positive results are from potential cross-contamination during sample processing. Grab samples of soil were also collected and analyzed in this study; however, the soil sample results are not presented in Fig. 2 because samples were not analyzed at the 300 CFU spore load. The lowest Btk T1B2 loading level was 3000 CFU. Consequently, a direct comparison to the other samples was not feasible. At spike loads less than 3000 CFU, the background microbial load overwhelmed the Btk T1B2 spores loaded onto the grab samples.

Fig. 1.

Analytical assay comparison of surface samples. Traditional culture method results (light gray bars) and RV-PCR method results (dark gray bars) for each surface sampled for SS (a) and VFC (b). True-positive or true-negative results represented by solid fill, false-positive results represented by diagonal stripe pattern, and false-negative results represented by zig zag pattern. RV-PCR data for CONPIER results in a total percentage of 101 due to rounding. SBMGAL, small boat marine grade aluminum; NSKID, nonskid tread; TCHSCRN, touch screen; CONPIER, concrete pier. Controls are reagent or field blank samplers processed with samples

Fig. 2.

Analytical assay comparison of grab samples. Traditional culture method results (light gray bars) and RV-PCR method results (dark gray bars). True-positive or true-negative results represented by solid fill, false-positive results represented by diagonal stripe pattern, and false-negative results represented by zig zag pattern. SBWASH, small boat washdown water; GRAVEL, gravel parking lot; VEGE, grass vegetation. Controls are reagent or field blanks processed with samples

The data described above and shown in Figs. 1 and 2 demonstrate three main points. First, the laboratory and field control samples (i.e., samples with low or no background microorganisms) resulted in near 100% true-positive or true-negative results across both analytical techniques (TC and RV-PCR) and all three sampling approaches (SS, VFC, and grab samples). Second, the sample results shown in Fig. 1a (SS), b (VFC) demonstrate true results (positive or negative) at a much higher percentage (76% to 94%) than for the grab sample results in Fig. 2 (33 to 67%) regardless of surface sampled. The SS and VFC samples were used to collect samples from surfaces (e.g., marine grade aluminum/nonskid tread/touchscreen of a boat or a concrete pier) that are more frequently cleaned or used, rather than a natural environment like the grab samples (e.g., wash water, gravel, or vegetation). Therefore, one possible reason for more true results in the SS and VFC samples compared to the grab samples is due to a lower background microbial load.

Figure 3 is a scatter plot of the percentage of true results for each analytical method as a function of the background microbial load. The level of total background microbial load collected from a sample has an impact on the detection level of target Btk T1B2 spores and thus the performance, for both TC and RV-PCR analytical assays. A higher background microbial load leads to growth competition for nutrients with the target Btk T1B2 spores during BHIB enrichment for TC and RV-PCR, as well as overcrowding on TSA plates for the TC method. The SS samples had the lowest total microbial load and highest true-positive or true-negative results, and the grab samples had the highest background with the lowest true-positive or true-negative results.

Fig. 3.

Percent true results per sample type and analytical method. Total background is the average number of CFU/mL recovered from zero-spike samples on TSA plates. Traditional culture results are represented by circles, SS (black circle), VFC (gray circle, overlaps with gray square), and grab (patterned circle). RV-PCR results are represented by squares, SS (black square), VFC (gray square), and grab (patterned square). Pos, positive; Neg, negative

Figure 4 shows the average presumptive Btk and total microbial load recovered from each of the sample types, SS, VFC, and grab with 0 CFU Btk spore load. For the TC method, the amount of background microorganisms that have the same morphology as the target organism can lead to a higher percentage of false-negative results. For example, Btk is a common biopesticide and is naturally found in the environment, which is indistinguishable visually from Btk T1B2 colonies; therefore, it is expected that Btk will be recovered from some environmental samples. For SS, VFC, and grab sample types, presumptive Btk colonies were seen in 14 of 19 (74%), 14 of 16 (88%) and 12 of 12 (100%) 0 CFU spore load samples, respectively. These results led to the selection of presumptive Btk colonies for PCR from 0 CFU spore load samples and also led to false-negative results when an inadequate number of colonies were screened using PCR from samples that were spiked. For example, three SS samples at the 300 CFU spore load were false negative when only one presumptive Btk colony was PCR-screened. Considering the 0 CFU spore load samples collected from these surfaces (SBMGAL and NSKID) had 1.2 ± 0.7 and 6.8 ± 11.3 CFU/mL of background presumptive Btk colonies isolated, more presumptive Btk colonies should have been screened to reduce the probability of false-negative results. Increasing the number of presumptive colonies for PCR confirmation could be accomplished by pooling colonies into a single PCR reaction without increasing the number of colony PCR reactions, although it is not feasible to screen all colonies isolated in samples with a high microbial load. The level of presumptive Btk isolated from each maritime surface and the PCR results from the TC method are reported in Supplemental Tables 1 and 2.

Fig. 4.

Presumptive and total background microbial load recovered from 0 CFU Btk spore load samples. Recovered CFU recovered from zero-spiked samples with presumptive Btk morphology (light gray bars) and total background microorganisms, all morphologies (dark gray bars) for each sample type

The problem of isolating presumptive colonies is not unique to this study or Btk, as presumptive B. anthracis colonies can also be isolated from environmental samples, even when a differential media (TSA with 5% Sheep’s Blood) is utilized, as plate crowding can interfere with hemolysis patterns. Both TC and RV-PCR methods are adversely impacted by high total microbial load. High background microbial load leads to the suppression of Btk T1B2 growth from competition for nutrients with background organisms. The RV-PCR results and average total microbial load are reported in Supplemental Table 3.

Conclusions

The traditional culture and RV-PCR analytical methods and associated sample processing procedures were evaluated for their compatibility with detection of spores of genetically barcoded B. thuringiensis subsp. kurstaki, a surrogate for B. anthracis, in the surface and grab samples collected from the USCG installation assets and their immediate surroundings in Portsmouth, VA. The representative samples collected included surfaces, such as aluminum on boats, nonskid tread on decks of watercraft, computer touchscreens, and concrete piers, and grab samples of boat washdown water, soil, vegetation, and gravel. These samples were spiked with Btk spores, processed for spore recovery, and analyzed by the two methods. Based on the results of this study, the following conclusions have been derived:

• Both the traditional culture and the RV-PCR methods are valuable methods and can give similar results for relatively clean maritime environmental samples. Typically, sample analysis by the culture method includes multiple way splitting of the recovered target biothreat agent suspension for dilutions plating, MicroFunnel Filter membrane plating, and enrichment of the remainder of sample followed by streak plating and/or PCR-based confirmation, especially, for samples containing a low level of target agent. This may affect time-to-results for sample analysis if sample enrichment is required.

• The traditional culture method can potentially provide semi-quantitative analytical results based on typical colony characteristics of the target bacterium on solid growth medium plates in the absence of high microbial background in the samples. The RV-PCR method provides qualitative analytical results.

• Maritime asset surface samples containing high microbial background load can mask the colonies of the target bacterium on solid growth medium plates and interfere with the culture method. Additionally, some other environmental bacilli exhibiting colony characteristics similar or identical to those of the target bacterium can be present in the samples, triggering PCR screening of more colonies, and possibly, repeated PCR screening (to minimize risk of false negatives) if presumptive morphology is present in large numbers. High microbial background can also affect sample analysis results by the RV-PCR method, but only when the results algorithm (ΔCt value of ≥ 9) is not satisfied. Particulates within samples (especially vegetation and soil grab samples) can reduce the amount of sample volume processed by the RV-PCR method and, therefore, increase overall sample processing time and could also affect the results.

• In general, the culture method was found to be labor- and reagent/material-intensive, but very straightforward for laboratory staff to perform sample analysis. The RV-PCR method requires custom parts and, due to integrated DNA extraction-purification, has many steps, but has the potential to be much more streamlined, with less labor and fewer laboratory consumables.

• It would be valuable to assess ways to simplify both analytical methods to improve turnaround time and reduce the amount of biohazardous waste generated.

• For the RV-PCR method, a reduction in processing steps could be evaluated in a sample complexity-dependent manner to determine which steps are crucial for detection. Additionally, throughput and accuracy of the RV-PCR method could be improved through use of automated sample processing and nucleic acid extraction.

• Complex environmental samples such as soil, grass, and other grab samples are difficult to analyze using the current sample processing methods to effectively recover spores. To mitigate this problem, a major emphasis needs to be placed on development of improved and high-throughput sample processing methods for such sample types.

Data availability

All data will be publicly available at data.gov.