INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

A number of gram-positive bacteria form spores when they encounter a nutrient shortage or are exposed to certain chemicals. The genera Bacillus and Clostridium are the classical genera that form endospores (7). In recent years the genus Bacillus has been divided into several groups or genera (for details see reference 11). Bacillus strains and, in particular, Bacillus species are not easily distinguished by spore analysis (14). Current technology has failed to yield immunological, biochemical, or nucleic acid-based methods for identifying spores of these organisms rapidly and definitively. Because of the resistance of spores (13), the approaches that are most frequently used to identify them require, as a first step, germination and culturing of the resulting vegetative bacteria. DNA- and RNA-based characterization also requires complex sample preparation. More rapid antibody-based analytical methods have been described, and these methods identify the genus Bacillus; however, they do not distinguish Bacillus spores at the species level (12). Since the genus Bacillus contains Bacillus thuringiensis, an industrially important nonpathogenic pesticide, Bacillus cereus, a noninfectious food pathogen, and Bacillus anthracis, a lethal infectious pathogenic bacterium, rapid discrimination of the spores from each other is necessary for effective intervention and treatment of human disease. In this study we evaluated matrix-assisted laser desorption ionization (MALDI)-mass spectrometry to determine whether it can be used to directly characterize Bacillus spores with speed, reliability, and sensitivity.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Microorganisms. Bacillus subtilis 168, Bacillus globigii, B. thuringiensis subsp. kurstaki HD-1, and B. cereus T, B33, and NCTC8035 spores were obtained from the U.S. Army Medical Research Institute of Infectious Diseases (USAMRIID), Frederick, Md., and were grown in chemically defined sporulation medium by using previously described procedures (4). B. globigii spores were also produced in new sporulation medium (NSM) (10). The spores were harvested by mild centrifugation at 10,000 × g for 10 min. The remaining vegetative cells were destroyed by treating the harvested material with lysozyme (50 µg/ml) in 50 mM Tris-HCl (pH 7.2). The purified spores were lyophilized and stored at 20°C. B. anthracis Sterne, a nonpathogenic strain, was prepared on plates containing NSM (10). Spore preparations were evaluated microscopically for the presence of dormant spores, germinated spores, and vegetative cells. Samples of B. globigii spores produced in 1966 and 1996 by using casein acid digest medium (CADM) were obtained from the Bioferm Corporation, Wasco, Calif. Montana soil was NIST (Gaithersburg, Md.) reference material 2710, and urban particulates were NIST reference material 1648.

Mass spectrometric study. MALDI mass spectra were obtained with a Kompact MALDI 4 time of flight instrument (Kratos Analytical Instruments, Chestnut Ridge, N.Y.) in the linear mode at an accelerating voltage of 20 kV by using a 0.3-µs delay time. The laser fluence was typically 10 mJ/cm². Each spectrum comprised the ions from 50 laser shots. Untreated spores were suspended (1 to 5 mg/ml) in acetonitrile-0.1% trifluoroacetic acid (70:30, vol/vol), and 0.3 µl was deposited in a well on a 20-well Kratos sample slide. The sample was covered with 0.3 µl of a 50 mM solution of sinapinic acid prepared in the same solvent mixture, unless another matrix is specified. Samples treated by corona plasma discharge (CPD) in a sample well were typically mixed with 0.3 µl of saturated sinapinic acid in acetonitrile-0.1% trifluoroacetic acid (70:30, vol/vol). Both internal and external mass markers were used to provide a mass accuracy of 1 part in 3,000.

CPD. A high-frequency, high-voltage generator (model BD-20A; MesoSystems Technology, Richland, Wash.) was used for CPD experiments (1). The original electrode was replaced with a 20-mm-diameter hollow cylinder with a sharp edge. This electrode was placed about 10 mm above each well in the sample slide in air in order to provide low-current CPD pulses with repetition rates of 120 pulses/s (1a).

Electron microscopy. Scanning electron micrographs of CPD-treated B. cereus T spores and untreated controls were obtained with a model 1820D microscope (Amray Co., Bedford, Mass.). The spores were suspended at a concentration of 5 mg/ml in acetonitrile-water (70:30, vol/vol). This stock preparation was diluted so that the final estimated concentration was 50 µg of spores/ml. Plasma discharge was performed with dried films of the spores directly on a microscope glass slide. Before introduction into the scanning electron microscope, the samples were coated with a platinum-gold alloy by using a vacuum deposition apparatus. Scanning electron micrographs were then obtained at an accelerating potential of 20 kV and a magnification of ×20,400.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Distinctive biomarkers were present in the MALDI spectra of the spores of the five Bacillus species studied (Fig. 1). Spectra of these biomarkers were qualitatively reproduced at the University of Maryland, at USAMRIID, and at the Johns Hopkins University Applied Physics Laboratory (1). Some variability in the molecular masses was observed, and this variability could be related to replacement of protons by sodium and/or potassium cations. The molecular mass assignments shown in Fig. 1 and 2 were supported by anion MALDI spectra and high-resolution Fourier-transform mass spectrometry (FTMS) measurements. The biomarkers could be washed off the spores with a variety of organic solvents; thus, they were characterized as lipophilic and were presumed to be present on the outsides of the spores. Their structures are being studied.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Linear-mode MALDI spectra of spores of five Bacillus species. (a) B. thuringiensis subsp. kustaki HD-1. (b) B. subtilis 168. (c) B. globigii. (d) B. cereus T. (e) B. anthracis Sterne.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Reflectron mode MALDI spectra of B. globigii spores grown in three different media, CADM (a), chemically defined sporulation medium (b), and NSM (c).

In order to determine whether the spectra characterized secreted biomarkers or chemicals adsorbed from the growth medium, B. globigii spores were grown in three different growth media used for MALDI analysis. The spectra of the three samples all contained the same family of peaks around m/z 1059 (Fig. 2). A second set of peaks around m/z 1496 was also characteristic of B. globigii spores. One spectrum (Fig. 2a) was obtained from spores grown in CADM in 1996. The MALDI spectrum obtained from a sample of the same species grown in CADM and stored since 1966 contained the same molecular ions, but the relative intensities were different (results not shown).

More importantly, the MALDI fingerprint was reproducible. Different cultures grown on different days provided similar MALDI spectra.

The limit of detection of this technique for characterization of B. thuringiensis was evaluated by using dilutions of a suspension whose spore concentration had been determined by using a Petroff-Hauser counting chamber. A spectrum obtained from about 5,000 spores contained the peak shown in Fig. 1a, which was characteristic of B. thuringiensis spores. In this study the sensitivity was determined by limitations related to sample handling rather than by the mass spectrometer.

The molecular masses of the lipophilic biomarkers of the spores of three strains of B. cereus (T, B33, and NCTC8035) are summarized in Table 1. The spectra suggest that some Bacillus spores can be distinguished at the strain level by MALDI-mass spectrometry. The values in Table 1 were obtained in reflectron mode (see above) in order to provide accurate mass values to address this critical question.

Workers in our laboratory have recently described the effectiveness of CPD for improving the accessibility of biomarkers for MALDI analysis of refractory samples (1a). The MALDI spectra of untreated B. cereus T spores and of spores treated by CPD for 5 and 30 s are shown in Fig. 3. The external biomarkers that dominated the spectrum of the untreated sample were also visible in the spectrum of spores treated for 5 s (Fig. 3a and b). However, a larger set of higher-mass ions was present in the MALDI spectra of the CPD-treated samples (Fig. 3c).

View larger version (13K): [in this window] [in a new window]   FIG. 3.   Linear-mode MALDI spectra of B. cereus T spores not treated by CPD (a) or treated by CPD for 5 s (b) or 30 s (c).

The extent and nature of spore disruption by CPD were examined by using several techniques. Electron micrographs of B. cereus T spores are shown in Fig. 4; these micrographs were obtained before and after CPD treatment for 120 s. Clearly, the outside spore surface was altered; however, the spores did not disintegrate.

View larger version (121K): [in this window] [in a new window]   FIG. 4.   Electron micrographs of B. cereus T spores before (A) and after (B) 120 s of treatment by CPD (magnification, ×20,400).

To assess the viability of the spores following CPD treatment, B. cereus T spores were enumerated on blood sheep agar plates as described above. Exposure of spores to CPD for up to 2 min reduced the viability by less than 1 log, while longer treatment times resulted in a greater than 6-log kill (Fig. 5). In addition, recovery of nucleic acids leaking from the spores treated by CPD was assayed independently. Triplicate measurements of RNA recovery revealed a 27% ± 4% increase following 2 min of CPD treatment, while a 436% increase was observed when spores were opened by using a classical method (15). The results of these experiments are consistent with the interpretation that the spore wall was modified during the short discharge times used with MALDI analysis but was not completely opened. Such resilience is appropriate for the candidate agent of universal panspermia (6).

View larger version (6K): [in this window] [in a new window]   FIG. 5.   Numbers of CFU following CPD treatment of 107 B. cereus T spores for different times.

Figure 6 shows MALDI spectra obtained after CPD treatment of the same five Bacillus species characterized in the experiment whose results are shown in Fig. 1. Clearly, the biomarker sets provided a unique and distinctive fingerprint for each of the samples studied. The external and internal biomarkers characterized in Fig. 1 and 6 could be combined to provide a larger number of biomarkers for distinguishing the species. Rule-based computer programs for this purpose are under development, and these programs do not involve library searching and do not require reproducible relative intensities of peaks. Identification based on searches of the proteome-genome database has also been proposed recently (3). When ions in the spectrum of B. subtilis 168 spores (Fig. 6b) were compared with proteins predicted on the basis of the genome of B. subtilis 168 (8), the observed molecular masses (6,846, 6,948, and 9,135 Da) matched within 1 Da the masses predicted for YvrF protein, YdiQ protein, and YuK [A-F, I-M] proteins, respectively. Entries for no other organism in the SwissProt/TrEMBL database matched more than one of the peaks.

View larger version (18K): [in this window] [in a new window]   FIG. 6.   Linear-mode MALDI spectra of spores of five Bacillus species obtained after 15 s of CPD treatment. (a) B. thuringiensis subsp. kurstaki HD-1. (b) B. subtilis 168. (c) B. globigii. (d) B. cereus T. (e) B. anthracis Sterne.

The molecular masses of biomarkers observed in MALDI spectra of samples of the three B. cereus strains obtained after 15 s of CPD treatment are summarized in Table 2. These strains are the same strains studied in the experiment whose results are shown in Table 1, and again the three strains can be readily distinguished.

Conclusions. MALDI-mass spectrometry provides reproducible biomarkers for characterization of spores of members of the genus Bacillus. Two classes of biomarkers have been observed. Lipophilic compounds present on the outside of the spore wall have been found to be uniquely associated with samples of different bacteria grown under a variety of conditions. Further support for the hypothesis that these compounds are true biomarkers was provided by independent characterization of B. globigii and B. thuringiensis spores at multiple sites under different sampling conditions. A larger selection of higher-mass biomarkers can be obtained when the spores are first treated by CPD, and many of these biomarkers have masses characteristic of proteins from the Bacillus proteome. These two classes of compounds can be combined to provide a larger set of biomarkers for computer-supported recognition of spores based on MALDI mass spectra.