Inactivation of Bacillus Spores by Ultraviolet or Gamma ... - CiteSeerX

Inactivation of Bacillus Spores by Ultraviolet or Gamma Radiation Ernest R. Blatchley III, A.M.ASCE1; Anne Meeusen2; Arthur I. Aronson3; and Lindsay Brewster4 Abstract: Bacillus anthracis spores represent an important bioterrorism agent that can be dispersed in air or water. Existing decontamination practices based on these spores have focused on chemical disinfectants; however, the basic characteristics of radiation-based disinfectants suggest potential advantages in their application for control of Bacillus spores. Experiments were conducted to examine the effectiveness of ultraviolet 共UV254兲 radiation and ␥ radiation for inactivation of Bacillus spores. Spores of Bacillus cereus were used for most experiments because of their similarity to B. anthracis. A limited number of experiments were also conducted using B. anthracis Sterne spores. In aqueous suspension, B. anthracis Sterne spores were observed to be slightly more resistant to UV254 than the spores of B. cereus. For the conditions of culture and assay used in these experiments, both spore types were more sensitive to UV254 radiation in aqueous suspension than the spores of B. subtilis, which are commonly used to characterize the performance of UV disinfection systems for water. Dried spores on surfaces were observed to be more resistant to UV254 than the same spores in aqueous suspension; it is likely that the increased resistance to UV of the dried spores was attributable to surface characteristics 共porosity and texture兲 of the solid materials. ␥ radiation was shown to accomplish similar rates of inactivation for spores in aqueous suspension and for dried spores on surfaces. Collectively, these results suggest that the application of UV or ionizing radiation may hold promise for decontamination following bioterrorism events. DOI: 10.1061/共ASCE兲0733-9372共2005兲131:9共1245兲 CE Database subject headings: Ultraviolet radiation; Terrorism; Contamination; Disinfection.

Introduction Real and suspected events within the last several years have demonstrated the vulnerability of infrastructure systems to terrorist activities. Among these activities have been intentional exposures of employees and patrons of the U.S. Postal Service and other government entities to biological warfare agents. Some individuals have become seriously ill and a few have died as a result of these exposures 共USGAO 2003兲. The methods used for decontamination of the media that have been the targets of bioterrorism 共e.g., parcels, office buildings兲 are crude, as are the methods used to ensure adequate delivery of the agents required for decontamination. The endospores of Bacillus anthracis, the causative agent of 1

Professor, School of Civil Engineering, Purdue Univ., 550 Stadium Mall Dr., West Lafayette, IN 47907-2051 共corresponding author兲. E-mail: [email protected] 2 Intercommunale Ontwikkelingsmaatschappij der Kempen, Antwerpseweg 1, 2440 Geel, Belgium; formerly, Graduate Student, School of Civil Engineering, Purdue Univ., 550 Stadium Mall Dr., West Lafayette, IN 47907-2051. 3 Professor, Dept. of Biological Sciences, Purdue Univ., West Lafayette, IN 47907. E-mail: [email protected] 4 Student, Dept. of Biological Sciences, Purdue Univ., West Lafayette, IN 47907. Note. Discussion open until February 1, 2006. Separate discussions must be submitted for individual papers. To extend the closing date by one month, a written request must be filed with the ASCE Managing Editor. The manuscript for this paper was submitted for review and possible publication on September 10, 2004; approved on January 26, 2005. This paper is part of the Journal of Environmental Engineering, Vol. 131, No. 9, September 1, 2005. ©ASCE, ISSN 0733-9372/2005/9-1245– 1252/$25.00.

the disease anthrax, represent a noteworthy bioterrorism agent. The most serious threat to human health by B. anthracis spores is via the inhalation pathway 共Dixon et al. 1999兲. Therefore, transmission of the spores in dry form on or in parcels presents an important risk. Other methods of bioterrorist attack with B. anthracis spores may also be possible. For example, airborne spores could be introduced to the HVAC systems of public buildings. It may also be possible to conduct a bioterrorist attack with these organisms through public water supplies. Although this method of attack will result in substantial public exposure through direct ingestion, the inhalation pathway may also be important for people in close proximity to water sprays 共e.g., showers兲 as a result of spore aerolosolization. Clearly, other mechanisms of attack are possible, and attacks could also come through the use of other biological warfare agents. Regardless of the mechanism of attack, bioterrorist attacks will require a decontamination response. Literature information indicates that Bacillus spores are fairly resistant to inactivation by conventional disinfection strategies 共Knudson 1986; Nicholson et al. 2000兲. In general, the disinfectant doses required to accomplish inactivation of B. anthracis spores are substantially larger than those required for the vegetative cell forms of most bacterial pathogens. However, the database of reliable dose–response behavior for these organisms and common disinfectants is quite limited. Bacillus spores have been reported to be 10–75 times more resistant to inactivation than their corresponding vegetative cells 共Knudson 1986; Setlow 1992a,b; Nicholson et al. 2000兲. Dry Bacillus spores subjected to ultraviolet 共UV254兲 irradiation have been reported to be inactivated at similar rates to those in aqueous suspension 共Lindberg and Horneck 1991; Nicholson et al. 2000兲. Bacillus spore resistance to UV254 is largely attributable

JOURNAL OF ENVIRONMENTAL ENGINEERING © ASCE / SEPTEMBER 2005 / 1245

to dehydration of the cytoplasm and protection of the DNA, which is in an altered conformation due to dehydration, by a group of small acid soluble proteins 共SASPs兲共Setlow 1992a,b兲. Because of the altered DNA conformation, the predominant photoproduct in spores is 5-thyminyl-5,6-dihydrothymine 共SP兲, for which an efficient, specific repair mechanism exists 共Nicholson et al. 2000兲 which also contributes to the greater UV resistance of the spore. By contrast, the primary photoproducts formed in vegetative Bacillus cells when subjected to UV254 are cis, syn-cyclobutane dimers, which are not repaired as efficiently as SP and may be more damaging to the organism 共Lindberg and Horneck 1991兲. Ionizing radiation can also be used to accomplish inactivation of bacterial spores. Relevant mechanisms of microbial inactivation involving ionizing radiation include so-called “direct” and “indirect” processes. Direct processes involve absorption of photon energy by a target molecule, resulting in damage to the target. Indirect processes occur when photon energy is absorbed by a nearby molecule 共e.g., water兲 resulting in the formation of highly reactive species 共e.g., radicals兲, which in turn react with target molecules in the microorganism. For the source of ionizing radiation used in this research 共 60Co兲, photon energy is sufficiently high that it can be used as a surrogate for other sources of ionizing radiation, including other ␥ emitters 共e.g., 137Cs兲 and electron beams 共Thompson and Blatchley 2000兲. As with most disinfectants, Bacillus spores are more resistant to inactivation by ␥ irradiation than their corresponding vegetative cells. Spore water content is lower than in the corresponding vegetative cell 共Nicholson et al. 2000兲. It is possible that indirect reactions involving production of radicals 共and other reactive intermediates兲 within the core of an organism could play an important role in inactivation. If so, then the low water content of the spore core could influence spore resistance to ionizing radiation. Although SASP content is known to be important in the responses of spores to oxidants, it apparently does not influence spore sensitivity to ␥ radiation 共Setlow 1995兲. The research described herein was conducted to examine the feasibility of UV and gamma 共␥兲 radiation-based disinfection and decontamination strategies for water supplies and dry surfaces that have been contaminated with Bacillus anthracis spores. Experiments were conducted to provide quantitative information regarding the responses of Bacillus spores to UV and ␥ radiation. Bacillus cereus spores were used as a surrogate for Bacillus anthracis spores for most experiments involved in this research. A small number of experiments were also conducted using B. anthracis Sterne spores as the target.

Materials and Methods Bacterial Strains: Propagation, Spore Production, Surface Attachment, Recovery from Surfaces, and Viability Assays Bacillus cereus T was grown and sporulated in G-tris medium 共liquid and plates兲 and the spores harvested and washed as previously described 共Aronson and Pandey 1978兲. Spores were stored at 4 ° C in de-ionized water and checked for refractility under a phase microscope prior to use. Total spore counts were determined in a Petroff–Hauser chamber and confirmed by plating dilutions of preparations after heating at 65° C for 30 min. The spores were subjected to irradiation in aqueous solution and on surfaces. For experiments involving exposure of waterborne

B. cereus spores to UV254 or ␥ radiation, spores were suspended in a sterile aqueous 0.01 M NaHCO3 solution 共pH 8.0兲. Suspensions used in experiments involving exposure to UV254 radiation had a spore concentration of roughly 106 cfu/ mL. Suspensions used in experiments involving exposure to ␥ radiation had a spore concentration of approximately 105 cfu/ mL. Spores were applied to planar surfaces by first cutting 6 mm diameter disks of each material surface 共aluminum foil, plastic, paper, painted paper, cellulose acetate membrane兲 with a hole punch. A small volume 共5.0 ␮L兲 of an aqueous stock solution of spores 共spore concentration ⬃5 ⫻ 108 cfu/ mL兲 was applied to each material disk and allowed to air dry for approximately 1 h. Taking into account the area of the membrane 共28.3 mm2兲 and the average area of one spore 共0.5 ␮m2兲, this concentration resulted in roughly 3.5% occupation of the dry surface area by the spores. It was assumed that this amount of surface coverage was sufficiently low to limit spore clumping or multilayer accumulation on surfaces. This approach and the resulting spore surface coverage were similar to those reported in previous research involving irradiation of dry spores on surface 共e.g., Lindberg and Horneck 1991兲. Following disinfectant exposure, spores were removed from surfaces by submerging each disk in 1.0 mL of 0.1% 共v:v兲 Nonidet P-40 prepared in 0.01 M NaHCO3 buffer. Spore suspensions were vortexed two times for 1 min with a 30 s idle interval after each vortexing treatment, followed by addition of another 1.0 mL of sterile sodium bicarbonate buffer and another 60 s of vortexing. Thus, in total the membranes were vortexed 3 min with two 30 s idle intervals. Spore recovery was quantified in separate experiments by application and removal of spores from the surfaces without disinfectant exposure, and comparison with recoveries from the original aqueous suspensions. Spore recovery by this method was consistently 90–100%. Once the spores were removed from a surface, quantitation of spore viability was conducted as described below. The concentration of viable B. cereus spores present in aqueous suspensions and detergent suspensions 共spores recovered from surfaces兲 was measured by spread plating. Samples were serially diluted in sterile NaHCO3 buffer solution. For each dilution examined in the assay, 100 ␮L of the diluted sample was applied to a G-tris agar plate using a cell spreader. Plates were inverted and incubated for 16– 20 h at 30° C. Plates with at least ten colonies, but not more than 300 colonies, were used for analysis of dose–response behavior. The lower limit was selected to maintain statistical reliability in the data, while the upper limit was chosen to ensure accurate counting of bacterial colonies. Triplicate plating was performed for all samples. Bacillus anthracis 共Sterne兲 were grown in liquid nutrient sporulation medium 共NSM兲共Aronson and Pandey 1978兲 at 37° C for 12– 14 h. For spore preparation, cells were spread on the surface of NSM agar plates which were incubated for approximately 4 days at 37° C. Spores from the NSM agar plates were harvested and washed as described for B. cereus. B. anthracis Sterne spore preparations usually contained 5–10% vegetative cells; these were removed by centrifugation through a gradient of Renografin 共Aronson and Pandey 1978兲. The spore pellet was washed two additional times with de-ionized water and was ⬎99% pure, as judged by examination of wet mounts in a phase microscope. Ultraviolet Exposure Ultraviolet irradiation of aqueous spore suspensions and dried spores on surfaces was accomplished using a flat-plate collimated

1246 / JOURNAL OF ENVIRONMENTAL ENGINEERING © ASCE / SEPTEMBER 2005

beam 共Blatchley 1997兲, which provided an incident beam of nearly uniform intensity 共⬃140 ␮W / cm2兲. The UV source for the collimated beam was a low-pressure mercury lamp, which was characterized by nearly monochromatic output at a characteristic wavelength of 254 nm. The beam diameter 共60 mm兲 was sufficiently large to flood the entire surface of the object being irradiated, such that uniform exposure of the entire object 共and its contents兲 was ensured. Incident radiation intensity 共I0兲 was measured with a radiometer 共International Light Model 1400A radiometer, SEL 240 detector, T2NS254 optical filter兲, which had been calibrated against a NIST standard. When aqueous spore suspensions were subjected to UV irradiation, it was necessary to account for absorbance within the sample. Transmittance 共␭ = 254 nm; path length= 1.0 cm兲 measurements of the aqueous spore suspensions were conducted using a Perkin–Elmer 共Lambda 20兲 UV-visible scanning spectrophotometer. The spatial average of radiation intensity was used to characterize the exposure Iavg = I0

1 − exp共− ␣H兲 ␣H

共1兲

where Iavg = average UV254 intensity within aqueous spore suspension 共mW/ cm2兲; I0 = incident UV254 intensity 共mW/ cm2兲; ␣ = absorbance coefficient at 254 nm 共cm−1兲 = ln共T兲 / ᐉ; T = transmittance 共254 nm兲 of spore suspension; ᐉ = optical path length for transmittance measurement 共cm兲; and H = depth of aqueous spore suspension 共cm兲. Transmittance values of spore suspensions were 92–95% through a 1.0 cm optical path. The spore suspension used under the collimated beam had a depth of less than 5 mm. Aqueous spore suspensions were always mixed during exposure by the use of a small magnetic stir bar. In combination with the uniformity of the incident UV intensity field, these characteristics promoted uniformity of dose delivery. For aqueous spore suspensions and air-dried spores, multiple samples were used to allow exposure to a wide range of UV doses. Dose was defined as the product of Iavg and exposure time. For UV exposures of air-dried spores, Iavg was defined as the incident radiation intensity. In most experiments, doses ranged from 0 to 100 mJ/ cm2, as this is representative of the range of doses that are likely to be applied in treatment applications. Following UV exposure, samples were extracted from surfaces 共if necessary兲, serially diluted as required, and subjected to triplicate analysis by plating, as described above. ␥ Exposure ␥ irradiation was performed by placing a beaker with several plastic test tubes 共which served as batch reactors兲 on the stage of a Nordion International Gammacell 220 Irradiator, which was constructed with a series of pencil-shaped 60Co sources arranged in a radial pattern to promote uniform exposure within the irradiated zone. The stage was lowered into the irradiated zone of the device to allow controlled exposure to the radiation field. Test tubes containing either aqueous spore suspensions or dried surfaces were removed from the irradiator following predefined exposure times to allow examination of a range of relevant ␥ doses. Dose delivered to each sample was estimated as the product of dose rate and exposure time. Dose rate was estimated based on an original dose rate measurement for the system, elapsed time since the dose rate measurement, and the half-life of the 60Co source. The original measurement of dose rate delivered by the system

Fig. 1. Ultraviolet254 dose–response behavior of B. cereus spores suspended in aqueous solution of sodium bicarbonate 共pH= 8.0兲. Fit of series-event model 共k = 0.953 cm2 / mJ; n = 13兲 is included for data up through inflection point.

was conducted by Fricke 共ferrous sulfate兲 dosimetry. Fricke dosimetry has been demonstrated to provide precise, accurate measurements of the rate of energy absorption from sources of ionizing radiation, including 60Co 共Ross et al. 1989; Klassen et al. 1999兲. Dose rates applied in this research were 14– 15 Gy/ min. Doses of up to 8,000 Gy were applied to samples. Following ␥ exposure, samples were extracted from surfaces 共if necessary兲, serially diluted as required, and plated in triplicate, as described above.

Results Irradiation of Aqueous Spore Suspensions The spores of Bacillus cereus were the focal point of this portion of the research. Keim et al. 共1997兲 determined that B. cereus and B. thuringiensis are the closest relatives to B. anthracis. Helgason et al. 共2000兲 claimed that B. anthracis was “…genetically indistinguishable from members of the B. cereus-B. thuringiensis group.” The primary differences between B. cereus and B. anthracis appear to be the presence of the plasmids pX01 and pX02 in B. anthracis, and their absence in B. cereus. Fig. 1 illustrates the UV254 dose–response behavior of B. cereus spores in an aqueous suspension. Dose–response experiments were conducted on spores generated from two separate batches 共A and B兲, with multiple experiments being conducted with spores from each batch, and triplicate plating of several sample dilutions at each UV dose. As described above, B. cereus was chosen as a model for Bacillus anthracis because of genetic and physiological similarities. Also, previous research had indicated similar UV254 dose– response behavior between the vegetative cells of B. anthracis 共Sterne and Vollum strains兲 and B. cereus 共Knudson 1986兲. Although no experiments were conducted in this research using virulent forms of B. anthracis, a limited number of experiments were conducted with the spores of B. anthracis Sterne, an avirulent veterinary vaccine strain 共see Fig. 2兲. Experiments involving exposure of B. cereus spores in aqueous suspension to ␥ radiation were conducted under several different conditions 共see Fig. 3兲. The first two experiments were


Fig. 2. Ultraviolet254 dose–response behavior of B. anthracis Sterne spores suspended in aqueous solution of sodium bicarbonate 共pH= 8.0兲. Fit of series-event model 共k = 0.766 cm2 / mJ; n = 13兲 is included for data up through inflection point.

conducted with aqueous spore suspensions that were prepared under conditions in which no attempts were made to include or exclude any dissolved constituents 共i.e., atmospheric gases兲 from the suspension that could influence the responses of the spores to irradiation. These were followed by experiments in which the concentration of dissolved oxygen was intentionally manipulated by equilibration of the aqueous suspension with either N2 gas or air. Irradiation of Dried Spores Air-dried B. cereus spores were subjected to UV254 irradiation on several planar surfaces. A summary of the data from these experiments is presented in Fig. 4. The materials used in these experiments included aluminum foil, plastic 共3 M transparency sheet, PP-2200兲, envelope paper 共Annapolis Premium paper, plain white兲, and the same paper painted with spray paint 共blue Krylon

Fig. 3. ␥ dose–response behavior of B. cereus spores suspended in aqueous solution of sodium bicarbonate 共pH= 8.0兲. Fits of seriesevent model to data sets are included.

Fig. 4. Ultraviolet254 dose–response behavior of air-dried B. cereus spores on material surfaces

high gloss paint兲. Aluminum foil was selected because it is an essentially nonporous, UV-reflective material that may be representative of building construction materials that could be subjected to spore contamination 共e.g., sheet aluminum used in air ducts兲. Clearly, a broad range of plastics exist that could be subjected to spore contamination. The specific plastic chosen for these experiments was selected because it is easily applied to a planar surface so that a well-defined UV dose could be delivered to the spores. Envelope paper was chosen because of previous reports of B. anthracis contamination of postage parcels. The pore structure of the paper surface, although poorly defined, could potentially provide shelter for spores from UV radiation. Finally, the painted paper surface was examined because it was expected that the application of paint would alter the surface texture and pore structure of the paper. Paint could also affect the wetting characteristics of the surface, which in turn could alter the distribution of spores when applied in an aqueous suspension and allowed to air dry. Air-dried spores were also subjected to ␥ irradiation. A summary of the data from this experiment is provided in Fig. 5. The surface examined in these experiments was a nitrocellulose membrane filter 共pore size 0.45 ␮m兲.

Fig. 5. ␥ dose–response behavior of air-dried B. cereus spores on nitrocellulose membrane surface


Discussion Ultraviolet Irradiation of Aqueous Spore Suspensions B. cereus UV254 dose–response behavior for spores in aqueous suspension was characterized by a lag in inactivation for doses up to 15– 20 mJ/ cm2, followed by roughly first-order inactivation. At doses above approximately 30 mJ/ cm2, corresponding to inactivation responses of roughly 4 log10 units, inactivation consistently tailed off. The data presented in Fig. 1 represent only those samples 共and their respective dilutions兲 that yielded an acceptable colony count in the plating assay. In particular, samples with fewer than ten colonies were not considered to provide adequate statistical reliability, and were therefore excluded from analysis; samples that yielded more than 300 colonies per plate were also excluded from the analysis because of difficulties in discerning individual colonies. It is important to note that many samples from high-dose experiments 共i.e., in the “tailing” region兲 yielded acceptable colony counts in the plating assays. Therefore, the onset of tailing behavior cannot be explained simply by an approach to the limit of detection. The series-event model represents a mathematical tool for description of the response of a microbial population to disinfectant exposure 共Severin et al. 1983兲. Mathematical descriptions of microbial dose–response behavior are necessary in the prediction of disinfection process efficiency in treatment operations, and for comparisons of dose–response behavior among microorganisms or disinfectants. For physical disinfectants, such as UV or ␥ radiation, the series-event model is defined mathematically as follows: n−1

N 共kD兲i = exp共− kD兲 N0 i! i=0

兺

共2兲

where N = concentration of viable organisms 共cfu/mL兲; N0 = concentration of viable organisms prior to disinfectant exposure 共cfu/mL兲; k = inactivation rate constant 共cm2 / mJ or Gy−1兲; D = dose of UV or ␥ radiation 共mJ/ cm2 or Gy兲; n = threshold number of damage events required for microbial inactivation. The series-event model has been used to describe the dose– response behavior of several simple microorganisms, including some viruses and bacteria. It has also been used to describe the behavior of larger, more complex organisms which often display a lag in their dose–response behavior. The lag is presumed to correspond to the ability of microorganisms to accumulate a finite amount of damage before showing any signs of losing viability within the population. Like most microbial “dose–response” relationships, the series-event model does not account for tailing behavior. Included in Fig. 1 is a fit of the series-event model to the data for inactivation of B. cereus by UV254 radiation. This fit was performed by nonlinear regression of the data up through the inflection point in the data set, where tailing became evident. For this data set, tailing was evident at UV254 doses above 30 mJ/ cm2. Several mechanisms have been identified as possible causes of tailing in dose–response relationships, including: population heterogeneity, the presence of particles, and disinfectant-induced resistance. The methods used in this research did not allow for identification of the mechanism共s兲 of tailing. However, spore purity was checked prior to individual experiments, and the mutation frequency to be expected in these experiments would have been too low to be observed in this type of experiment.

Fig. 6. Comparison of ultraviolet254 dose–response behavior of B. anthracis Sterne spores with ultraviolet254 dose–response behavior of B. cereus spores. Limits defined for ultraviolet254 dose–response behavior of B. subtilis spores, as defined in Environmental Protection Agency Ultraviolet Disinfection Guidance Manual 共USEPA 2003兲 are included for comparison. Also included are data from Nicholson and Galeano 共2003兲 regarding measured ultraviolet dose–response behavior of B. anthracis Sterne.

The lower limit of 10 colonies per plate would account for a standard deviation in the estimates of viable spore concentration 共for each individual sample兲 of approximately 30% of the mean. As such, some variation in measurements of dose–response behavior is to be expected in the plate counts corresponding to doses in the tailing region. However, most samples in the tailing region yielded more than one replicate with 10 or more colonies; this source of error was not large enough to account for the observed tailing behavior. Therefore, it appears that at least some of the tailing behavior was real 共i.e., not a statistical artifact兲. Based on the arguments presented above, the most likely cause of tailing behavior was the presence of particles or spore aggregates. Microscopic examination of spore suspensions pre- and postirradiation indicated that the spores existed in suspension as discrete particles. However, if particle-association or spore aggregation did occur in these experiments, and it was the cause of tailing behavior, it would be difficult to prove by microscopic examination. At 4 log10 units of inactivation, only 1 in 10,000 spores would retain viability. If aggregation or particle association were the cause, it would be difficult to find such an aggregate among the spore population by microscopic examination. The spores of B. anthracis Sterne displayed UV254 dose– response behavior that was qualitatively similar to that of B. cereus, in that it was characterized by a lag in response for doses up to 15– 20 mJ/ cm2, followed by relatively rapid inactivation and tailing. However, B. anthracis Sterne spores appeared to be somewhat more resistant to UV254 irradiation than B. cereus spores 共Fig. 6兲. If aggregation or particle association were contributors to the observed tailing behavior, then it is possible that the conditions employed in this research may have contributed to this behavior. Specifically, spore concentrations used in the UV experiments were roughly 106 cfu/ mL. These high spore concentrations, which were used to improve the limit of detection for the method and thereby increase the range of doses that could be applied, are


orders of magnitude higher than could reasonably be expected for a potable water supply or a municipal wastewater effluent, even if intentionally contaminated. The observed UV254 dose–response behaviors of these two spore types were similar, but not identical. In general, B. anthracis Sterne spores were more resistant to UV254 radiation than were the spores of B. cereus. However, this generalization must be qualified because spore dose–response behavior is known to be affected by several factors 共see below兲. The results of these comparisons illustrate important differences in the dose–response behavior of Bacillus spores. It is interesting to note that Knudson 共1986兲 observed the vegetative cells of B. cereus to be slightly more resistant to UV irradiation than those of B. anthracis 共Sterne or Vollum兲. The data illustrated in Fig. 6 suggest the opposite trend to be true for the spores of B. cereus and B. anthracis 共Sterne兲. Conditions of culture and assay 共e.g., media, temperature兲 can influence spore composition and properties. The observed differences in UV sensitivity of B. cereus versus B. anthracis may have been attributable to these factors, at least in part. Both were grown in “enriched” media, but of different composition, and B. cereus was sporulated at 30 versus 37° C for B. anthracis. Factors that are intrinsic to target organisms can also influence their sensitivity to UV irradiation. For example, Benoit et al. 共1990兲 demonstrated that some plasmids increase UV sensitivity in B. thuringiensis spores 共probably due to induction by UV of lysogenic phage兲. In this research, B. cereus spores were chosen as a surrogate for B. anthracis on the basis of their genetic similarity. The primary difference between these species is believed to be their plasmid content. In particular, B. anthracis contains the plasmids pX01 and pX02, which are responsible for its virulence. B. anthracis Sterne lacks the pX02 plasmid, while B. cereus contains neither of these plasmids. Knudson 共1986兲 reported essentially no difference in the UV sensitivity of B. anthracis Sterne and Vollum vegetative cells; however, B. cereus cells were observed to be slightly more resistant to UV254 irradiation than B. anthracis Sterne or Vollum. In this research, B. cereus spores were observed to be more sensitive to UV254 radiation than the spores of B. anthracis Sterne. Therefore, it is possible that the virulence plasmids may influence UV sensitivity of Bacillus anthracis spores. Nicholson and Galeano 共2003兲 reported that B. anthracis Sterne and B. subtilis yielded essentially the same resistance to UV inactivation, when grown and sporulated under the same conditions. However, their conclusions were based on data sets that were much smaller than those reported herein. Moreover, their data yielded some evidence of tailing in the limit of high doses for the B. anthracis Sterne spores, but not in the data from their experiments with B. subtilis. It is also important to recognize that their experiments were limited to a smaller dose range than reported herein. At present, the standard practice for characterization and validation of UV disinfection systems used for treatment of potable water is to quantify the “reduction equivalent dose” 共RED兲 for the system based on biodosimetry 共USEPA 2003兲. In biodosimetry, a challenge organism is imposed on a UV reactor system under tightly controlled conditions. Biodosimetry tests are generally conducted over the range of operating conditions that are expected for the system. Influent and effluent samples are collected to allow quantification of the fraction of challenge organisms that are inactivated for each operating condition. This fraction is compared with the measured dose–response behavior of the challenge organism, as quantified through a collimated-beam experiment,

similar to the experiments described herein. The RED for each operating condition is defined as the dose from the collimatedbeam experiment corresponding to the observed extent of challenge organism inactivation from the biodosimetry test on the flowthrough reactor system. The two most commonly applied challenge organisms used in biodosimetry are coliphage MS-2 and Bacillus subtilis spores. Well-defined methods have been developed for the application of both challenge organisms 共e.g., USEPA 2003兲. Part of the process of standardizing these methods is the definition of acceptable limits on dose–response behavior for challenge organisms. For purposes of illustration, USEPA proposed limits for UV dose–response behavior of B. subtilis spores to be used in biodosimetry testing are included in Fig. 6. The spores of B. cereus and B. anthracis Sterne appear to be more sensitive to UV254 radiation than the spores of B. subtilis over the relevant dose range. Similarly, B. cereus and B. anthracis Sterne spores are generally more sensitive to UV254 radiation than coliphage MS-2 共USEPA limits not shown兲; however, in the limit of low UV254 dose 共i.e., in the “lag” region of the Bacillus dose–response curves兲, the opposite trend is observed. Given these observations, it appears that the results of biodosimetry based on the application of B. subtilis spores as the challenge organism will be preferred in systems where an assessment of B. anthracis inactivation is sought. Although it is not possible to use the results of biodosimetry to make quantitative predictions of the responses of nonchallenge organisms, it is sometimes possible to use these results to define the lower limit of inactivation to be expected in a system. In particular, if B. subtilis spores were used as a challenge organism for a system, then it would be possible to definitively state that the extent of B. cereus or B. anthracis Sterne spore inactivation 共as measured using the methods and spores described herein兲 to be expected in the system would be at least as great as that of the challenge organism for the given set of operating conditions. As mentioned above, Nicholson and Galeano 共2003兲 conducted a limited set of experiments that were similar to those described herein; however, they reached a slightly different conclusion regarding the relative sensitivities of B. subtilis spores and B. anthracis Sterne spores to UV radiation. In particular, they suggested that the two spore types illustrated essentially identical UV254-dose–response behavior. An important attribute of the study by Nicholson and Galeano 共2003兲 was their use of identical conditions of growth, sporulation, and assay for the spores. These conditions were used to allow direct comparisons of the data from the two spore types. Although the conditions of sporulation and assay used for B. anthracis Sterne in this study were somewhat different than those used by Nicholson and Galeano 共2003兲, the observed dose–response behavior was similar, but substantially more detailed. For visual comparison, the data of Nicholoson and Galeano 共2003兲 are also included in Fig. 6; the magnitude of the lag and tailing behavior from Nicholson and Galeano 共2003兲 was similar to that observed in our data sets for the dose range reported in their study. ␥ Irradiation of Aqueous Spore Suspensions Four separate dose–response experiments were conducted with aqueous suspensions of B. cereus spores. The results of the first two experiments, in which no attempts were made to control the concentration of dissolved oxygen, suggest that the ␥ dose– response behavior of B. cereus spores is reproducible. Previous research 共Thompson and Blatchley 2000兲 had indicated that


dissolved oxygen concentration had a much larger effect on microbial inactivation rates under ␥ irradiation than the concentrations of common radical scavengers, such as the carbonate system. Therefore, experiments were not conducted to examine the effects of carbonate alkalinity variations in inactivation kinetics in this research. Elimination of dissolved oxygen by equilibration with N2 gas resulted in little change in the observed dose–response behavior. Bubbling of air through samples prior to irradiation, resulting in equilibration of oxygen with the ambient atmosphere, yielded a slight increase in inactivation efficiency. In general, the concentration of dissolved oxygen had only a minor effect on the sensitivity of B. cereus spores to ionizing radiation. The inactivation responses of the spores when subjected to ␥ irradiation demonstrated a slight lag, followed by a steady, nearly first-order decline in viable spore numbers 共Fig. 3兲. In contrast with the results of experiments involving UV254 irradiation, no evidence of tailing was found in the data from the ␥ irradiation experiments. In the UV experiments, tailing became evident at UV doses corresponding to 3 – 4 log10 units of spore inactivation. Spore inactivation responses in the ␥ experiments were less than 4 log10 units, so it is possible that these experiments were not carried out over a sufficiently wide dose range to detect tailing behavior. However, the range of ␥ doses applied in these experiments was believed to be relevant for disinfection system designs based on ionizing radiation. Given the greater penetrating power of ␥ radiation as compared to UV254, it is reasonable to expect that a system based on ionizing radiation would be less susceptible to particle-induced tailing than a UV system. Ultraviolet Irradiation of Dried Spores on Surfaces Spore inactivation on air-dried surfaces by UV254 irradiation was substantially different than for an aqueous suspension 共Fig. 4兲. No evidence of a lag was present in any of the data sets, but there was much more scatter than in the experiments involving UV irradiation of aqueous spore suspensions. Presumably, this variability was attributable to the characteristics of the surfaces, including porosity and texture. All tests conducted with dried spores involved “off the shelf” surface materials; no attempts were made to control surface characteristics in the samples used in this research. Although variability in the data from the experiments involving UV irradiation of air-dried spores on surfaces may mask some of the trends in spore UV254 dose–response behavior, it appears that the general characteristics of these dose–response results were different than those of the experiments involving spores in aqueous suspension. In general, the air-dried spores were more resistant to inactivation by UV254 irradiation than were the spores in aqueous suspension. Spore inactivation was most effectively accomplished on the aluminum foil surfaces. Spores on the plastic surface appeared to be slightly more resistant to UV254 than those on the aluminum foil surfaces, possibly due to the lower reflectivity of the surface as compared with aluminum foil. Spores were poorly inactivated on the paper surface. This was attributed to the porous nature of the paper surface. The fact that spores were inactivated slightly more effectively on the painted paper surface than on the unaltered paper surface supports the hypothesis that the pore structure of the surface provided shelter from UV radiation; presumably, the paint smoothed the paper surface and eliminated 共blocked兲 some pore spaces that could have provided shelter to spores from UV.

␥ Irradiation of Dried Spores on Surfaces It has been hypothesized that water content in the spore core plays an important role in the ␥ resistance of Bacillus spores 共Nicholson et al. 2000兲, presumably because of its role in the generation of 共hydroxyl兲 free radicals. More generally, it is known that ␥ irradiation can lead to microbial inactivation through direct ionization of microbial constituents, or through so-called indirect damage, wherein reactive chemical species are generated within or near the organism that ultimately lead to microbial inactivation. It was not possible to differentiate the effects of these mechanisms on Bacillus cereus spores based on the results of this work. However, it does appear that the resistance of B. cereus spores was not significantly altered by the differences in spore water content between the experiments involving spores in aqueous suspension and the air-dried spores on surfaces. Collectively, the results of experiments based on 60Co as a ␥ source indicate that inactivation of Bacillus spores by ionizing radiation will take place at similar rates for aqueous media and surfaces. Disinfection or decontamination strategies based on ionizing radiation will be reliable and effective for inactivation of Bacillus spores if dose delivery can be monitored or verified.

Acknowledgment This material is based on work supported by the National Science Foundation under Grant Number BES 0210350.

References Aronson, A. I., and Pandey, N. K. 共1978兲. “Comparative structural and functional aspects of spore coats.” Spores VII, G. Chambliss and J. C. Vary, eds., American Society for Microbiology, Washington, D.C., 54–61. Benoit, T. G., Wilson, G. R., Bull, D. L., and Aronson, A. I. 共1990兲. “Plasmid-associated sensitivity of Bacillus-thuringiensis to UV-light.” Appl. Environ. Microbiol., 56共8兲, 2282–2286. Blatchley, E. R., III. 共1997兲. “Numerical modelling of UV intensity: Application to collimated-beam reactors and continuous-flow systems.” Water Res., 31共9兲, 2205–2218. Dixon, T. C., Meselson, M., Guillemin, J., and Hanna, P. C. 共1999兲. “Anthrax.” N. Engl. J. Med., 341共11兲, 815–826. Helgason, E., Okstad, O. A., Caugant, D. A., Johansen, H. A., Fouet, A., Mock, M., Hegna, I., and Kolsto, A. B. 共2000兲. “Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis—One species on the basis of genetic evidence.” Appl. Environ. Microbiol., 66共6兲, 2627–2630. Keim, P., Kalif, A., Schupp, J., Hill, K., Travis, S. E., Richmond, K., Adair, D. M., HughJones, M., Kuske, C. R., and Jackson, P. 共1997兲. “Molecular evolution and diversity in Bacillus anthracis as detected by amplified fragment length polymorphism markers.” J. Bacteriol., 179共3兲, 818–824. Klassen, N. V., Shortt, K. R., Seuntjens, J., and Ross, C. K. 共1999兲. “Fricke dosimetry: the difference between G共Fe3+兲 for Co-60 gammarays and high-energy x-rays.” Phys. Med. Biol., 44共7兲, 1609–1624. Knudson, G. B. 共1986兲. “Photoreactivation of ultraviolet-irradiated, plasmid-bearing, and plasmid-free strains of Bacillus-anthracis.” Appl. Environ. Microbiol., 52共3兲, 444–449. Lindberg, C., and Horneck, G. 共1991兲. “Action spectra for survival and spore photoproduct formation of Bacillus-anthracis irradiated with short-wavelength 共200– 300 nm兲 UV at atmospheric-pressure and in vacuo.” J. Photochem. Photobiol., B, 11共1兲, 69–80. Nicholson, W. L., and Galeano, B. 共2003兲. “UV resistance of Bacillus anthracis spores revisited: Validation of Bacillus subtilis spores as UV surrogates for spores of B-anthracis Sterne.” Appl. Environ. Micro-


biol., 69共2兲, 1327–1330. Nicholson, W. L., Munakata, N., Horneck, G., Melosh, H. J., and Setlow, P. 共2000兲. “Resistance of Bacillus endospores to extreme terrestrial and extraterrestrial environments.” Microbiol. Mol. Biol. Rev., 64共3兲, 548. Ross, C. K., Klassen, N. V., Shortt, K. R., and Smith, G. D. 共1989兲. “A direct comparison of water calorimetry and Fricke dosimetry.” Phys. Med. Biol., 34共1兲, 23–42. Setlow, P. 共1992a兲. “DNA in dormant spores of Bacillus species is in an A-like conformation.” Mol. Microbiol., 6共5兲, 563–567. Setlow, P. 共1992b兲. “I will survive—Protecting and repairing spore DNA.” J. Bacteriol., 174共9兲, 2737–2741. Setlow, P. 共1995兲. “Mechanisms for the prevention of damage to DNA in

spores of Bacillus species.” Annu. Rev. Microbiol., 49, 29–54. Severin, B. F., Suidan, M. T., and Engelbrecht, R. S. 共1983兲. “Kinetic modeling of UV disinfection of water.” Water Res., 17共11兲, 1669– 1678. Thompson, J. E., and Blatchley, E. R., III. 共2000兲. “Gamma irradiation for inactivation of C-parvum, E-coli, and coliphage MS-2.” J. Environ. Eng., 126共8兲, 761–768. United States Environmental Protection Agency 共USEPA兲. 共2003兲. Ultraviolet disinfection guidance manual, Office of Water, USEPA, Washington, D.C. United States Government Accounting Office 共USGAO兲. 共2003兲. “Issues associated with anthrax testing at the Wallingford Facility.”GAO-03787T, USGAO, Washington, D.C.
