Fast Sterility Assessment by Germinable-Endospore Biodosimetry

Abstract

The increased demand for sterile products has created the need for rapid technologies capable of validating the hygiene of industrial production processes. Bacillus endospores are in standard use as biological indicators for evaluating the effectiveness of sterilization processes. Currently, culture-based methods, requiring more than 2 days before results become available, are employed to verify endospore inactivation. We describe a rapid, microscopy-based endospore viability assay (μEVA) capable of enumerating germinable endospores in less than 15 min. μEVA employs time-gated luminescence microscopy to enumerate single germinable endospores via terbium-dipicolinate (Tb-DPA) luminescence, which is triggered under UV excitation as 10⁸ DPA molecules are released during germination on agarose containing Tb³⁺ and a germinant (e.g., l-alanine). Inactivation of endospore populations to sterility was monitored with μEVA as a function of thermal and UV dosage. A comparison of culturing results yielded nearly identical decimal reduction values, thus validating μEVA as a rapid biodosimetry method for monitoring sterilization processes. The simple Tb-DPA chemical test for germinability is envisioned to enable fully automated instrumentation for in-line monitoring of hygiene in industrial production processes.

The era of modern microbiology began in the 1870s when the life cycle of an endospore-forming pathogen, Bacillus anthracis, was elucidated using new methods for isolating pure cultures from single-cell clones on solid growth media. Bacterial endospores are dormant microbial structures that are highly resistant to chemical, physical, and radiation sterilization processes (2, 6, 19, 26). In fact, Bacillus subtilis endospores have survived for 6 years in space while exposed to high-vacuum conditions, temperature extremes, and intense solar and galactic radiation (12, 18). Bacterial endospores are routinely employed as biological indicators (i.e., biodosimeters) to validate the effectiveness of sterilization methods (e.g., autoclaves) used in the medical device (15, 16), pharmaceutical, health care (5), food preparation, wastewater remediation (23), and biodefense (29) industries.

The effectiveness of sterilization processes is measured and reported in terms of sterility assurance levels (SALs), which are defined as the expected probability that a product remains contaminated with viable microorganisms after exposure to a validated sterilization process. A sterilization process that yields predictable SALs is considered to be validated. Confidence in achieving a required SAL is obtained by the use of biological indicators that present a considerably greater population and resistance challenge than the expected bioburden (21), and the most effective way to test the efficiency of a sterilization process is to place biological indicators within and on test products of interest.

Currently, endospore inactivation is quantified by measuring the log reduction in CFU. This method, however, requires several days of incubation, during which 20 cycles of cell replication ultimately yield visible colonies that can then be enumerated. In contrast, endospore germination can be initiated and monitored on a timescale of minutes. Germination of Bacillus endospores can be triggered by simple biomolecules, such as l-alanine, l-asparagine, or glucose (8, 24, 27), which cause the release of approximately 10⁸ molecules of dipicolinic acid (DPA) from the core of the endospore during the first stage of germination. DPA exists in all bacterial endospores as 5 to 15% of the cellular dry weight and is a unique, defining constituent of the cellular dry weight of endospores (10, 14, 22).

Here we report details of a rapid endospore viability assay (EVA) in which Bacillus atrophaeus endospores were immobilized on terbium ion (Tb³⁺)/l-alanine-doped agarose. The l-alanine serves to trigger germination, during which DPA is released from endospores. The Tb³⁺ subsequently binds DPA, resulting in green luminescent spots under UV excitation in a microscope field of view, which were enumerated as germinable endospores using time-gated Tb-DPA luminescence microscopy (i.e., μEVA). Here we validate μEVA against culturing as a method for rapid endospore viability assessment and evaluate its application for monitoring endospore inactivation by thermal and UV sterilization regimens.

MATERIALS AND METHODS

Chemicals.

Terbium(III) chloride hexahydrate, 99.999%, l-alanine, and other salts were purchased from Sigma (St. Louis, MO) and were used as received. Ultrapure agarose (>90%) was purchased from Invitrogen (Carlsbad, CA). Tryptic soy agar (TSA), nutrient broth, and agar were obtained from Becton Dickinson and Company (Sparks, MD).

Preparation of endospore stock suspension.

B. atrophaeus (ATCC 9372) endospores were purchased from Raven Biological Laboratories. B. subtilis (ATCC 27370) vegetative cells were grown on TSA and inoculated onto a sporulation medium after reaching exponential growth phase. The sporulation medium contained 1.6% nutrient broth, 1.6% agar, 0.2% KCl, 0.05% MgSO₄, 1 mM Ca(NO₃)₂, 100 μM MnCl₂·4H₂O, 1 μM FeSO₄, and 0.1% glucose (pH 7.0) (20). After incubation at 37°C for 1 week, cells were suspended into sterile deionized water. With phase-contrast microscopy, 95% of the cells formed endospores free of sporangia. Endospores were harvested and separated from vegetative cells and debris by being centrifuged at 6,300 × g, washed 10 times, and sonicated (25 kHz) for 5 min. The endospore suspension was incubated in lysozyme (0.2 mg/ml) and trypsin (0.1 mg/ml) at 30°C with constant stirring overnight to lyse and degrade vegetative cells. Endospores were purified by eight cycles of centrifugation (6,300 × g) and washed with sterile deionized water until >99.9% of the cells were fully refractile with no noticeable cellular debris. Endospore suspensions were stored at 4°C in the dark before use. Total endospore concentrations were determined using a Petroff-Hausser hemocytometer, and CFU concentrations were determined using TSA spread plating in triplicate measurements.

Sample preparation for μEVA experiments.

Endospore suspensions were filtered onto 0.2-μm polycarbonate membrane filters (Whatman, Florham Park, NJ) using filtration manifolds of different diameters depending on the desired concentration factor such that there were less than 300 endospores per microscopic field of view. To ensure that the endospore surface density was optimal for a given initial endospore concentration, suspensions of >10⁶ spores/ml were filtered onto 25-mm-diameter spots using glass filtration funnels, and suspensions of <10⁶ spores/ml were filtered onto 1.5-mm² spots using a 96-well microsample filtration manifold (Schleicher & Schuell, Keene, NH). Endospores concentrated on the filter were transferred to a ∼0.5-mm-thick, 9-mm-diameter slab of 1.5% agarose substrate containing 100 μM TbCl₃ and 20 mM l-alanine mounted in a silicone isolator (Molecular Probes, Eugene, OR) on a quartz microscope slide. After endospore transfer, the agarose surface was covered with a piece of 0.2-mm-thick polydimethylsiloxane (PDMS).

PDMS was prepared by mixing the polymer base and curing agent (Sylgard; Dow Corning) in a 10-to-1 ratio. After degassing, the mixture was cast over a 0.2-mm-thick stainless steel mold and cured in an oven for 2 h at 65°C. Agarose, silicone isolator, and PDMS were autoclaved at 121°C for 15 min before use. A piece of PDMS was peeled off and attached to the top of an endospore-laden agarose surface for sealing.

μEVA instrument.

The instrument consists of a time-gated camera (Photonics Research Systems, Salford, United Kingdom) mounted on a Nikon SMZ800 stereoscopic microscope (large working distance for xenon lamp), a xenon flash lamp (PerkinElmer, Waltham, MA) mounted at 45° with respect to the sample, and a temperature-controlled microscope slide holder (Thermal) (Fig. 1). The slide holder enabled endospores to germinate at 37°C. The charge-coupled-device camera has a resolution of 752 by 582 pixels at 14 bits with a chip size of 2/3 in. The camera has 50% sensitivity between 430 and 730 nm, with peak sensitivity at 550 nm. It was Peltier cooled to 40°C below the ambient temperature and was synchronized to the xenon lamp via transistor-transistor logic (TTL) pulses (300 Hz, with a tail time of up to 50 μs). A high-pass filter (03FCG067; Melles Griot) centered at 500 nm was placed along the light path on the emission side before reaching the microscope objective. We collected time stacks of time-gated images by real-time streaming with a delay of 100 μs and an exposure time of 5 s in each frame.

FIG. 1.

(a) Configuration of the μEVA instrument used in this investigation, consisting of a stereomicroscope mounted with a time-gated camera and a xenon flash lamp for UV excitation. (b) Sample well on quartz microscope slide containing Tb³⁺/l-alanine-doped agarose. (c) Schematic representation of the sample slide, consisting of a quartz slide on which Tb³⁺/l-alanine-doped agarose is confined by a red rubber gasket well. Endospores (brown circles) are inoculated onto agarose substrate and subsequently covered with a thin layer of PDMS. (d) Inoculated endospores germinate due to l-alanine, causing the release of ∼10⁸ molecules of DPA and subsequent formation of highly luminescent Tb-DPA complexes that appear as discrete bright spots in the microscope field of view. (e) Absorption-energy transfer-emission photophysics of the Tb-DPA luminescence assay. DPA acts as a light harvester that transfers excitation energy to luminescent terbium ion. (f) Energy (Jablonski) diagram of the Tb-DPA photophysics.

Endospore germination and germinable-endospore assignment.

Endospore germination on the agarose surface followed the reported germination dynamics (11). DPA released from single B. atrophaeus and B. subtilis endospores manifested as individual bright spots in 15 min under time-gated microscopy due to local formation of Tb-DPA. We performed image analysis in Matlab to obtain a background-subtracted stack of time-gated images. Assignment of germinable endospores was made based on intensity and size. Adaptive thresholding was applied to segment pixels that were three times brighter than the background, with a characteristic rising intensity. Each bright spot must exhibit a continuous rising intensity over the course of germination in order to register an endospore count. This criterion eliminated false positives by not counting sporadic bright spots and long-lived luminescent interference. The eight connected adjacent pixels were analyzed to screen for endospore clumps. The number of endospores present was calculated by dividing the squared sum of neighboring pixel brightness by the mean brightness of an individual endospore determined empirically. This was done in a recursive way until all of the pixels were counted and marked. Germinable-endospore units would be reported as too numerous to count if they exceeded 300 in a field of view. Further dilution would be carried out in those cases.

Phase-contrast microscopy for measuring total endospore concentration.

The total number of endospores was determined by direct enumeration using a hemocytometer counting chamber under phase-contrast microscopy. At high concentrations (i.e., >10⁶ spores/ml), populations were calculated based on an average of 10 random microscopic fields of view. At low concentrations (i.e., <10⁶ spores/ml), an air-dried smear of a known volume of endospore suspension was imaged. The entire area of the smear would be counted due to low numbers and uneven spatial distribution. The motorized stage and automatic counting algorithm expedited the capture and analysis of more than 100 images per sample.

Inactivation experiments.

Heat resistance and UV resistance were measured by the germinability (DPA release) and culturability (colony formation on TSA) of water-suspended endospores. In the heat inactivation experiment, 2 ml of B. atrophaeus endospores in sealed glass vials were heat treated in a water bath at 95°C. Vials were removed and placed in ice at different time intervals for μEVA and CFU enumeration. Lethality of the process includes only the holding period at 95°C. In the UV inactivation experiment, a 2-ml aliquot of B. atrophaeus endospores contained in a 6-cm-diameter glass petri dish was exposed to 254 nm UV from a mercury lamp (UVP, Upland, CA) coupled with a 0.7-neutral-density filter. Uniform UV exposure was achieved by agitating the petri dish on an orbital shaker (60 rpm). The area of the aliquot spread on the dish and, therefore, the amount of energy delivered in millijoules was determined to be 22.9 μW/cm². After various lengths of irradiation, endospore suspensions were transferred into vials under ice for subsequent μEVA and CFU enumeration.

Statistical analysis.

In this study, bacterial counts could be categorized into two regimes of distribution. High-count regimes were defined as containing >10 counts per field of view for μEVA and >10 counts per growth plate for culturing experiments, which followed Gaussian distribution. A square root transformation was performed on some of the data to homogenize the variance and normalize the negative skewness. Normality was tested using the Shapiro-Wilk test by looking at the skewness and kurtosis of the distribution. Parametric analyses, such as Student's t test, were used to determine the confidence interval, and the F test was used to examine differences between the sample variances. Low-count regimes were defined as containing <10 counts per field of view for μEVA and <10 counts per growth plate for culturing experiments, which followed the Poisson distribution. Nonparametric analyses, such as the Wilcoxon test, were used to determine the confidence interval, and the Kruskal-Wallis test was used for variance analysis. We reported the 95% confidence interval based on a 1.96 standard deviation of the normally distributed datasets. Data falling within the 5 and 95 percentile ranking constituted the reported 95% confidence interval for the Poisson-distributed datasets.

RESULTS AND DISCUSSION

We validated μEVA for analysis of germinable endospores in water suspension using a three-pronged approach. First, we confirmed the assignment that μEVA luminescent spots are germinable endospores by comparing single-endospore-germination dynamics observed with phase-contrast microscopy and μEVA. Second, we evaluated the μEVA sensitivity, dynamic range, and false-positive rate in comparison to standard TSA culturing. Third, we applied μEVA versus TSA culturing to monitor thermal and UV inactivation of B. atrophaeus.

Monitoring single-endospore-germination dynamics.

Rapid endospore viability assessment is achieved by measuring early observables in the germination-replication pathway. In particular, DPA release with subsequent water uptake can be observed during stage I germination, which occurs within minutes of the germinant addition. In our investigation, we have demonstrated that individual germinable endospores can be enumerated on a timescale of 10 min via Tb-DPA luminescence with μEVA. A comparison of phase-contrast microscopy and μEVA time-lapse images is shown in Fig. 2. After a brief microlag period, DPA release, followed by water influx, takes place during the first stage of germination (8, 25). DPA release was observed with μEVA via DPA complexation with the Tb³⁺ doped into the agarose matrix. The water uptake can be observed with phase-contrast microscopy as phase-bright endospores transition into phase-dark-germinated endospores. The time course data clearly show the coincidence of DPA release and water uptake going to completion in approximately 15 min, which is consistent with stage I germination (11). The microlag times reported by μEVA and phase-contrast microscopy were 3 and 8 min, respectively, which are also consistent with the sequence of germination (27). In addition, time-lapse excitation spectra observed during germination show characteristic Tb-DPA excitation spectra (a maximum λ of between 271 and 279 nm) (see Fig. S1 in the supplemental material), confirming the release of DPA under μEVA conditions. In combination, these data establish that μEVA observables are germinating endospores.

FIG. 2.

Germination time courses of single B. atrophaeus spores at 22°C monitored by phase transition from bright to dark as observed under phase-contrast microscopy (a) and Tb-DPA luminescence using μEVA (b).

Sensitivity, dynamic range, and false-positive rate.

To further validate μEVA, we performed parallel germination and culturing experiments over 7 orders of magnitude in endospore concentrations. Figure 3 shows the germinable-endospore concentrations measured with μEVA, and culturable endospore concentrations measured with TSA plate counting plotted against total endospore concentrations as determined with phase-contrast microscopy in the trace concentration regime of 0 to 52 spores/ml. Sterile samples did not yield false-positive counts, which enabled us to achieve the ultimate sensitivity of one germinable endospore per μEVA field of view. Of the total endospore concentrations, μEVA revealed that 56.7% ± 4.4% were germinable endospores and TSA culturing determined that 38.4% ± 3.5% were culturable endospores. The ratio of germinable/culturable endospores was 1.48, which is consistent with the fact that a subset of the total endospore population is germinable but not culturable (30). Similar to plate counting, where dilution factors are applied until the concentration yields 30 to 300 CFU per plate, μEVA requires concentrations that are less than 200 germinable endospores per field of view. Fig. S2 in the supplemental material shows μEVA versus TSA plate-counting results after application of appropriate dilution factors to samples in the 10¹ to 10⁶ spores/ml concentration range.

FIG. 3.

Endospore concentration dependence showing a comparison of μEVA (solid line) and heterotrophic plate (dashed line) measurements as a function of total endospore concentration as determined by phase-contrast microscopy.

Monitoring thermal and UV sterilization of Bacillus atrophaeus endospores.

The inactivation of B. atrophaeus endospores using thermal treatment at 95°C and UV irradiation at 254 nm with a power of 22.9 μW/cm² was monitored from an initial inoculum of 10⁷ phase-bright spores/ml to sterility with μEVA and TSA plate counting (Fig. 4). The endospore inactivation followed a first-order decay (31) reaction that was preceded by a shoulder and followed by a tail. The log endospore survivor curve can be simulated by a model described by Geeraerd et al. using the following system of differential equations (7):

where N denotes the endospore population, k_max is the maximum inactivation rate, and C_c is the number of hypothetical critical components inside endospores that induce a shoulder behavior associated with the inactivation regimen. The decimal reduction (D) value is defined as the time in minutes at a particular constant temperature or UV irradiation power to reduce the viable population by a factor of 10 in the log-linear regime (1). Heat inactivation D values were found to be 4.74 and 4.80 min by using μEVA and TSA plate counting, respectively. UV inactivation D values were calculated to be 30.52 and 30.43 min by using μEVA and TSA plating, respectively. The similarity in μEVA and TSA plate counting D values and inactivation time courses demonstrated that μEVA is a rapid alternative to monitoring endospore inactivation. For a given inactivation dosage, the germinable-endospore population remaining was always greater than the culturable population (Fig. 3), making germinable endospores a more conservative biological indicator and consequently yielding increased confidence for achieving a desired SAL.

FIG. 4.

Inactivation of B. atrophaeus spores showing μEVA (solid line) and heterotrophic plate (dashed line) counts as a function of inactivation dose for heat inactivation at 95°C (a) and UV inactivation with a mercury lamp irradiating samples at 254 nm with a power of 22.9 μW/cm² (b). The inactivation data were fit to a semiempirical model reported by Geeraerd et al. (7).

In μEVA experiments, individual germinable endospores are counted in a microscope field of view after germinant addition. As endospores germinate, ∼10⁸ DPA molecules are released into the immediate area surrounding the endospore. DPA then combines with Tb³⁺ in the agarose matrix to form the Tb-DPA luminescence halos under UV excitation (9). The germinating endospores manifest as bright spots that grow in intensity over a period of 3 to 5 min and are enumerated in a microscope field of view. The characteristic germination time course allows unambiguous assignments of germinating endospores. The duration of germination depends on a number of factors, such as species, inoculum size, germinants, and temperature. The reported phase transition for individual bacterial endospores ranges from 75 s to approximately an hour (11). This is manifested in the observed μEVA time course overlays for different species, with germination times ranging from 7 to 22 min (Fig. 5).

FIG. 5.

Germination time course plots of pure laboratory strain endospores (solid line) and environmental samples (dotted line) measured by μEVA. Each of the curves represents an average of 10 endospore germination time courses. B. atrophaeus, circles and solid line; B. cereus, squares and solid line; B. subtilis, diamonds and solid line; Geobacillus stearothermophilus, triangles and solid line; Atacama Desert (Chile) extract, circles and dotted line; Lake Vida (Antarctica) extract, squares and dotted line; Greenland ice core (GISP2) extract, diamonds and dotted line; Alaskan permafrost extract, triangles and dotted line.

Comparison to spectroEVA.

Previously, we reported a related method where germinating endospores were enumerated in bulk suspension by luminescence spectroscopy (i.e., spectroEVA), where Tb-DPA luminescence intensities were tabulated against a B. atrophaeus endospore calibration curve (28). Results obtained by a comparison of μEVA to spectroEVA showed that the two methods are in good agreement (see Fig. S3 in the supplemental material). The μEVA approach is superior to spectroEVA because μEVA is capable of enumerating single endospores, while the limit of detection of spectroEVA is 1,000 spores/ml. This advantage is gained because in μEVA experiments, the millimolar DPA halos surrounding single germinated endospores are readily imaged with a high contrast, whereas germination of single endospores in bulk suspension (∼1 ml) gives rise to mere femtomole DPA concentrations, which are far below the limit of detection for spectroEVA.

Application to environmental samples.

The unique photophysical and chemical characteristics of the Tb-DPA luminescence (3) endospore viability assay make μEVA a powerful instrument tool for endospore viability assessment and validation of sterilization. With μEVA, we take advantage of the long luminescence lifetime (0.5 to ∼2 ms) of Tb-DPA (13)_, enabling the use of time gating to effectively remove background fluorescence (i.e., interferent fluorophores with nanosecond lifetimes). Time gating eliminates potential false-positive-causing features and renders the image background dark. Elimination of this background enables a striking increase in image contrast and detection sensitivity even for the most challenging environmental extracts (Fig. 5), including soil extracts from the nearly sterile Atacama Desert, Chile (4, 17), and extracts from Greenland ice cores (32), Arctic permafrost, and Antarctic lakes (Lake Vida).

Automation.

μEVA is not only much more rapid than culture-dependent methods (10 to 15 min versus 2 to 3 days), but the simple chemistry, instrumentation, and image analyses are all amenable for automation. Automated viability assessment of endospores will have the potential to find application in many areas where microbial inactivation needs to be monitored and assured, including health care, food, and pharmaceutical industries. Specific examples for applications include automated performance testing for autoclaves, milk powder production lines, wastewater treatment facilities, and validation of bioagent inactivation after a biological attack. In the case of an anthrax attack, rapid viability assessment technology will aid field personnel to rapidly determine the viability of anthrax endospores before and after countermeasures. In the case of biological attacks with other agents (e.g., Yersinia pestis, Francisella tularensis, Brucella, Burkholderia species, and variola and foot-and-mouth disease viruses), Bacillus endospores can be used as a biological indicator for monitoring decontamination efficiency.