Bacillus nealsonii sp. nov., isolated from a spacecraft-assembly facility, whose spores are {gamma}-radiation resistant

The main focus of the National Aeronautics and Space Administration's planetary-protection efforts is the development of cleaning and sterilization technologies for spacecraft preparation prior to launch. Knowledge of the microbial diversity of spacecraft-assembly facilities, as well as any extreme characteristics these microbes might possess, is essential to the development of these technologies. The spacecraft-assembly facilities can be considered extreme environments created by the controlled air circulation, low humidity and low-nutrient conditions found in these clean-rooms. A wide variety of micro-organisms can survive under such conditions (Puleo et al., 1973, 1975, 1977; Venkateswaran et al., 2001).

In on-going investigations to determine and document possible microbial contamination on representative spacecraft components and accessories, several physiologically and phylogenetically novel micro-organisms were encountered (Venkateswaran et al., 2001). Witness plates made of spacecraft-quality stainless steel were exposed for ∼9 months at a Jet Propulsion Laboratory Spacecraft Assembly Facility (JPL-SAF) and the particulate materials collected revealed the presence of novel Bacillus species. Micro-organisms that exhibit resistance to an assortment of free radicals and conditions employed in emergent technologies for sterilization of spacecraft components are significant. Here, we describe Bacillus nealsonii, whose spores are resistant to UV, γ-radiation, H₂O₂ and desiccation.

Sample preparation and isolation of microbes from a spacecraft-assembly facility.
The dimensions of the JPL-SAF are 25 m wide, 36 m long and 15 m high. Relative humidity was controlled at 40±5 % with a cap at 45 % and the mean temperature was maintained at 20±5 °C. This JPL-SAF was maintained by qualified contamination control personnel with periodic checks to ensure a class 100 000 (the maximum number of particles of the size >0·5 µm per cubic foot of air) clean-room level. Stainless steel witness plates (type 304, no. 4 finish, 0·050·08 cm thick; size, 2·5x5 cm; Mechanical Workshop, JPL) were ultrasonically cleaned in acetone (510 min) followed by 2-propanol (510 min). After air drying, the plates were sterilized by heating at 175 °C for 2 h. The pre-sterilized witness plates were exposed in JPL-SAF on stands about 2 m high. This minimized contamination from human exhalation and sweat and ensured collection of dust particles that were naturally falling onto the witness plates. After a 9-month exposure, all 20 witness plates were individually placed into 50 ml polypropylene disposable sterile centrifuge tubes.

Microbial examination.
Each retrieved witness plate was placed into 30 ml of sterile phosphate-buffered (pH 7·2) rinse solution (Anonymous, 1980). The plate and rinse solution were sonicated for 2 min (25 kHz, 0·35 W cm^-2). The rinse solution was aseptically divided into two 15 ml aliquots. One aliquot of the rinse solution, along with the witness plate, was subjected to heat-shock (80 °C for 15 min), while the other aliquot was not heated. Total aerobic counts in appropriate aliquots of samples were determined by the pour plate technique using tryptic soy agar (TSA; Difco) as the growth medium (32 °C for 37 days). Type strains of different Bacillus species were procured from established culture collections and used as controls when necessary to validate the procedures.

Sporulation.
Bacillus endospores were purified using the following two procedures. Cells of an overnight TSA culture were harvested, washed in sterile water and heat-shocked at 80 °C for 15 min. The heat-shock procedure killed vegetative cells but not mature spores. The heat-shocked samples were grown overnight on Difco nutrient agar supplemented with 5 p.p.m. MnSO₄ (MN agar), which triggers sporulation of the test microbe. About 200 µl of the heat-shocked samples was spread onto multiple MN agar plates to harvest sufficient quantities of the test isolate. The cells grown on agar were washed in sterile water and the heat-shock procedure, followed by growth on MN agar, was repeated until 99 % spores were obtained. The percentage of spores was determined by viewing the spore preparations using phase-contrast microscopy. Spores appear as bright bodies when viewed with a phase-contrast microscope. Purification of spores using this MN agar method resulted in the retention of a loosely attached extraneous layer around the spore coat. In addition, a nutrient broth sporulation medium (NSM) was used to produce spores (Nicholson & Setlow, 1990; Schaeffer et al., 1965). A single purified colony of the strain to be sporulated was inoculated into liquid NSM. After 23 days of growth at 32 °C, the cultures were examined in wet mounts to determine the level of sporulation. Once the number of free spores in the culture was greater than the number of vegetative cells, the culture was harvested and the spores were purified. Spore purification was performed by treating the spores with lysozyme and washing with salt and detergent (Nicholson & Setlow, 1990). The chemical treatments used in this method removed the extraneous layer surrounding the spore coat. The purified spores were resuspended in sterile deionized water, heat-shocked (80 °C for 15 min) and stored at 4 °C in glass tubes.

Microscopy.
The refractile nature of the spores was examined by phase-contrast microscopy using an Olympus microscope (BX-60). A Field-Emission Environmental Scanning Electron Microscope (ESEM; Philips XL30) was also used. Very high resolution/magnification and an excellent signal to noise ratio in regular high vacuum was achieved due to the field-emission electron source. Non-destructive examination of spores and vegetative cells was possible using this microscope. Specimen preparation procedures, which usually lead to sample artifacts, are not necessary when using the ESEM. In addition, standard scanning and transmission electron microscopy were used to examine the surface details and cross-sections, respectively, as per established methods (Cole & Popkin, 1981).

Characterization of spores for various physical and chemical conditions.
Radiation dosimetry at the Co⁶⁰ source was performed using an ion chamber with accuracy to the US Bureau of Standards (Coss, 1999) standard. All irradiations were carried out in glass vials using spore samples in water. The spores (10⁸ spores ml^-1) were exposed to both 1 Mrad (50 rad s^-1 for 330 min) and 0·5 Mrad (25 rad s^-1 for 330 min.) and survival was quantitatively verified by growing the γ-radiation treated samples in TSA at 32 °C.

Purified spores were diluted in PBS (pH 7·2), placed into an uncovered Petri dish and exposed to UV radiation (254 nm; UV Products). At appropriate intervals, samples of spores were removed, diluted serially 10-fold in PBS and plated onto NSM agar. Plates were incubated at 37 °C for up to 5 days and colonies were counted.

A liquid H₂O₂ protocol, developed by Riesenman & Nicholson (2000), was modified and used to examine H₂O₂ resistance in spores. Suitable aliquots of spore suspensions prepared in PBS were treated with H₂O₂ (5 % final concentration) and incubated at room temperature (∼25 °C) with gentle mixing. After 60 min incubation, a 100 µl sample was removed and diluted in a solution of bovine catalase (100 µg ml^-1 in PBS). Serial 1:10 dilutions of the catalase-treated suspension were prepared in tryptic soy broth (TSB; Difco) to check viability and spread onto TSA for quantitative measurement of the H₂O₂-resistant spores.

For desiccation resistance, the spore suspension (20 µl) was dispensed onto pre-sterilized metals and glass-fibre discs (10³ spores per disc; Millipore). After removing most of the water content by drying at room temperature (∼4050 % humidity in Pasadena, CA, USA) for 1 or 2 days, the colonies were counted on TSA medium. Briefly, the desiccated sample was placed in sterile PBS, mixed thoroughly and sonicated for 2 min before plating onto TSA medium. Plates were incubated at 32 °C for 2 days and the number of spores that survived was counted.

Identification
Phenotypic characterization and fatty acid analysis.
Routine biochemical tests were carried out according to established procedures (Claus & Berkeley, 1986; Priest, 1993). The ability to grow at a NaCl concentration of 110 % was determined in T₁N₁ liquid medium (1 % Bacto tryptone and appropriate amount of NaCl) and the ability to grow without NaCl was determined in 1 % sterile tryptone water. The API CHB 50 kit and API 20E (bioMérieux) were used (75 biochemical tests). Identification of the test isolate was carried out by computing and comparing biochemical test results from the bioMérieux database. In addition, the commercially available Biolog identification system was also used, according to manufacturer's specifications. Fatty acid methyl ester (FAME) profiles were examined from overnight cultures grown at 32 °C in TSB, as described previously (Ringelberg et al., 1994).

16S rDNA sequencing.
Purified genomic DNA (Johnson, 1981) from liquid cultures was quantified and ∼10 ng of DNA was used as the template for PCR amplification. Universal primers (Bact 11 and 1,492) were used to amplify the 1·5 kb PCR fragment by protocols established by Ruimy et al. (1994). Amplicons were sequenced directly following purification on Qiagen columns. The identity of a given PCR product was verified by sequencing using the dideoxy chain termination method with the Sequenase DNA sequencing kit (United States Biochemical) and an ABI 373A automated sequencer (Perkin-Elmer). The phylogenetic relationships of organisms covered in this study were determined by comparison of individual 16S rDNA sequences to other existing sequences in GenBank. Evolutionary trees were constructed using PAUP (Swofford, 1990).

DNADNA hybridization.
Cells were suspended in 0·1 M EDTA (pH 8·0) and digestion of the cell wall was carried out by treating the cells with lysozyme (final concentration, 2 mg ml^-1). DNA was isolated by standard procedures (Johnson, 1981). DNADNA homology was studied by microplate hybridization methods (Ezaki et al., 1989) with photobiotin labelling and colorimetric detection, using 1,2-phenylenediamine (Sigma) as the substrate and streptavidine-peroxidase conjugate (Boehringer Mannheim) as the colorimetric enzyme (Satomi et al., 1997).

Microbial and particle contamination of JPL-SAF
Particles of the size 11150 µm were collected on witness plates (Anonymous, 1989). The stainless steel witness plates accumulated mid-range size (26100 µm) particles and the abundance of particles decreased when the particle size decreased (data not shown). Microbial contamination transferred through particulate materials was not high, in terms of microbial load, in this well-controlled facility. The particles trapped on stainless steel witness plates harboured an equivalent number of both vegetative (5 c.f.u. cm^-2) and spore-forming (6±1 c.f.u. cm^-2) microbes. When the isolated colonies were exposed to harsh conditions, such as UV, γ-radiation, H₂O₂ and desiccation, some spore-formers showed resistance. Among these spore-formers, a strain, designated as FO-92^T, exhibited distinct spore morphology and was further characterized for its phylogenetic affiliation.

Morphological and physiological characteristics
Strain FO-92^T is a Gram-positive, facultatively anaerobic, rod-shaped, spore-forming bacterium. Cells are 45 µm in length, 1 µm in diameter and are motile. On TSA medium incubated at 32 °C, young colonies are beige, irregular, with a diameter of 34 mm, rough, umbonate with undulate or lobate edges. Endospores of strain FO-92^T are oval (1x0·5 µm; Fig. 1a), with one spore per cell. Spores purified using the MN agar procedure contain a distinctive extraneous layer (Fig. 1b). Cross-sections of the MN agar-purified spores clearly show a loosely arranged layer outside the spore coat (Fig. 1c, d). This structure resembles the exosporium of the Bacillus cereus group (data not shown). This extraneous layer can be removed from the FO-92^T spores by washing with detergents and salts using the Nicholson & Setlow (1990) protocol. Spores of Bacillus subtilis ATCC 6633^T, Bacillus pumilus ATCC 7061^T and Bacillus megaterium IAM 13418^T did not show an extraneous layer when purified from MN agar. The extra layer (exosporium) was retained in Bacillus cereus JCM 1252^T and B. sphaericus 34hs1 even after the chemical treatments (Nicholson & Setlow, 1990) used to purify the spores (data not shown). The characterization and the physiological role of this extraneous layer of strain FO-92^T spores is not discussed in this paper. However, the resistance of the spores with and without extraneous layers against various treatments was measured.

(121K): Fig. 1. Environmental scanning electron micrograph (a, b) and transmission electron micrograph (c, d) of B. nealsonii FO-92T spores. Spores (bd) were purified on MN agar (see Methods). (a) Purified spores as per the protocol of Nicholson & Setlow (1990); (b) MN agar-purified spores retaining the extraneous layer (EL); (c) longitudinal section of a spore where the extraneous layer (EL), spore coat (SC), cortex and spore core are shown; (d) cross-section of a spore where the loosely attached extraneous layer is observed. The extraneous layer was removed by salt and detergent washes (a). Cross-sections of spores purified according to the method of Nicholson & Setlow (1990) did not possess the extraneous layer (data not shown).

Resistance of FO-92^T spores to various physical and chemical conditions
The resistance of Bacillus spores to a variety of conditions is common as seen in our control experiments (data not shown) and in other studies (for a review, see Nicholson et al., 2000). The spores of FO-92^T exhibited resistance to 0·5 Mrad (5 kGy) γ-radiation (Co⁶⁰), 200 J m^-2 UV (254 nm), 5 % liquid H₂O₂ and desiccation conditions. However, 1 Mrad γ-radiation was lethal and no spore germination was observed. Spores with the extraneous layer showed a 4-log reduction whereas spores without the extraneous layer showed a 5-log reduction at 0·5 Mrad γ-radiation. Although preliminary experiments suggest a protective role of the extraneous layer against γ-radiation, more detailed studies are warranted. The FO-92^T spores exhibited classic UV inactivation kinetics, with a characteristic shoulder extending to ∼100 J m^-2, followed by strict exponential inactivation thereafter. FO-92^T spores exhibited an LD₉₀ value (the 90 % lethal dose) of ∼200 J m^-2 (Fig. 2), in good agreement with UV resistance values obtained for spores of the model organism, B. subtilis strain 168 (Nicholson et al., 2000). The vegetative cells of strain FO-92^T were resistant to 5 % liquid H₂O₂ (data not shown). Purified spores that were exposed to 5 % liquid H₂O₂ for 3060 min showed resistance, but prolonged incubation to 90 min eliminated the viability (data not shown).

(13K): Fig. 2. Resistance of FO-92T spores to 254 nm UV radiation. Results shown are the means and standard deviations of three experiments. Spores purified by the Nicholson & Setlow (1990) protocol were used in this experiment.

Optimum growth conditions
Strain FO-92^T grew at 2560 °C, with optimum growth at 3035 °C and over the pH range of 610 (optimum 67). This strain did not require Na⁺ for growth and was as desiccation resistant as other spore-formers. However, it is interesting to note that the centre of an overnight colony on TSA (at 32 °C) predominantly consisted of spores when compared to the periphery of the colony (data not shown). Such an immediate response in triggering sporulation during nutrient-depleted conditions is common in Bacillus species. But, when compared to B. subtilis ATCC 6633^T, where spores were formed in 34 days on TSA (data not shown), strain FO-92^T produced spores in 1 day.

Phenotypic characterization
The biochemical characterization of strain FO-92^T is presented in Table 1. In addition to the characters shown, strain FO-92^T produced catalase but hydrogen sulfide was not produced from thiosulfite. The carbon substrate profile of FO-92^T, as measured by the BioLog system, showed an identification match for Bacillus. Phenotypically, as measured by the API system, this strain resembles B. circulans ATCC 4513^T.

Table 1. Biochemical characteristics of B. nealsonii FO-92T and related species Strain: 1, B. licheniformis ATCC 14580T; 2, B. subtilis IAM 1026T; 3, B. pumilus ATCC 7061T; 4, B. mycoides ATCC 6462T; 5, B. circulans ATCC 4513T; 6, B. firmus ATCC 14575T; 7, B. nealsonii FO-92T. All strains are Gram-positive rods, facultatively anaerobic and do not denitrify. None produces lysine or ornithine decarboxylases, urease, tryptophan deaminase, hydrogen sulfide or indole. None utilizes glucose, inositol, sucrose, citrate, sorbitol, rhamnose, melibiose or arabinose as sole carbon source. All ferment L-arabinose, D-glucose, D-fructose, D-mannose, mannitol, methyl α-D-glucoside, amygdalin, arbutin, aesculin, salicin, cellobiose, maltose, sucrose and trehalose. None ferments methyl α-D-mannoside, inulin, xylitol, L-fucose, erythritol, L-xylose, methyl β-xyloside, L-sorbose, dulcitol, D-fucose, L-arabitol or 5-ketogluconate.

Bacillus species that produce acid from a variety of sugars, including glucose, are classified under rRNA group 1 (Priest, 1993). Most of these species were able to grow at least weakly in the absence of oxygen. Spores of these species were ellipsoidal and did not swell the mother cell. These species are considered the subtilis group because of their similar physiological properties (Priest, 1993). Strain FO-92^T, isolated from JPL-SAF, exhibited the characteristics necessary to place it into the rRNA group 1.

Cellular fatty acid composition
Strain FO-92^T contained straight-chain and terminally branched saturated and mono-unsaturated fatty acids with a composition of 18, 73 and 9 %, respectively (Table 2). Among the fatty acids measured, tetradecanoic acid (14 : 0), 13-methyl pentadecanoic acid (15 : 0 iso) and 12-methyl tetradecanoic acid (15 : 0 anteiso) were the major fatty acids in FO-92^T. This FAME profile identified strain FO-92^T as Bacillus circulans DSM 11^T. FAME analysis of other Bacillus species showed distinct profiles. For example, Bacillus licheniformis ATCC 14580^T contained ∼90 % terminally branched saturated fatty acids, whereas Bacillus mycoides ATCC 6462^T showed more monosaturated fatty acids. Although both B. subtilis IAM 1026^T and strain FO-92^T exhibited high levels of straight-chain saturated fatty acids, B. subtilis IAM 1026^T contained high levels of pentadecanoic acid (15 : 0). Unfortunately, different culture conditions can result in high variability within FAME profiles (Venkateswaran et al., 1999). FAME analysis is ambiguous because type strains of some of the Bacillus species could not be identified correctly (Table 2). Because of these uncertainties, the identification of strain FO-92^T could not be conclusively determined by fatty acid profiles.

Table 2. Fatty acid methyl ester composition (%) of B. nealsonii FO-92T and related species Strain: 1, B. circulans DSM 11T; 2, B. firmus DSM 12T; 3, B. megaterium DSM 32T; 4, B. simplex DSM 1321; 5, B. pumilus DSM 27T; 6, B. subtilis DSM 10T; 7, B. subtilis IAM 1026T; 8, B. licheniformis ATCC 14580T; 9, B. mycoides ATCC 6462T; 10, B. nealsonii FO-092T.

16S rDNA sequence analysis
Molecular methods are less susceptible to artifactual misinterpretation than culture-based approaches. Studies have revealed that organisms with less than 97 % similarity over the 16S rRNA gene do not yield DNA reassociation values of more than 60 % (Stackebrandt & Goebel, 1994). While the gene sequence of the small subunit of the 16S rRNA molecule is acceptable for defining phylogenetic relationships between distinctly related organisms (Woese, 1987), this molecule at times lacks the specificity required for the differentiation of close relatives (Fox et al., 1992; Venkateswaran et al., 1998, 1999; Yamada et al., 1999). Strain FO-92^T closely resembled B. circulans by conventional phenotypic characterization and FAME profiles. In order to confirm the species identity, molecular phylogeny was carried out on this strain.

The 16S rDNA sequences of all known Firmicutes were compared with that of FO-92^T. All phylogenetic analyses, based on 16S rDNA sequence, unambiguously demonstrated that FO-92^T belonged to the low G+C Gram-positive bacteria. The 16S rDNA sequences of all known members of the Gram-positive bacteria were compared with that of FO-92^T. Their phylogenetic relationships were then analysed and the study was repeated with several different subdomains of the 16S rDNA sequence. Bootstrapping (500 replicates) analysis was performed to avoid sampling artifacts. The resulting analyses indicated that FO-92^T shares a close phylogenic relationship with Bacillus species. Neighbour-joining, parsimony and maximum-likelihood analyses were undertaken on this subset of bacteria, using several subdomains of the 16S rDNA. In all analyses, FO-92^T was most closely associated with members of the genus Bacillus.

The similarities in the 16S rDNA nucleotide sequences between FO-92^T and the top 17 closely related Bacillus species, recognized by GenBank BLAST searches, were between 95 and 98·7 %. A sequence variation of ∼1 % was found between FO-92^T and B. circulans ATCC 4513^T and 2 % between FO-92^T and Bacillus benzoevorans DSM 5391^T as well as Bacillus firmus IAM 12464. A very high sequence variation (5 %) was noticed between FO-92^T and both B. subtilis ATCC 6633^T and B. pumilus OM-F6. Such a high degree of dissimilarity within a well-described genus is not uncommon.

A phylogenetic tree based on 16S rDNA sequences is shown in Fig. 3. The branching order of this tree showed two distinct clusters in which one clade consisted of the B. subtilis group and another stock formed with 12 other species, including strain FO-92^T. These 12 other species exhibited five subclusters in which three major clades each contained at least three species. The first clade comprised FO-92^T, B. circulans ATCC 4513^T and B. benzoevorans DSM 5391^T. The second clade contained Bacillus macroides strain dhr2, Bacillus fumarioli LMG 17492 and Bacillus niacini IFO 15566^T, and the third clade included Bacillus simplex DSM 1321^T, Bacillus flexus IFO 15715^T and Bacillus megaterium IAM 13418. Because of the inadequacy of 16S rDNA analysis for species differentiation, DNADNA hybridization was performed.

(31K): Fig. 3. Phylogenetic tree of various species of Bacillus based on 16S rDNA nucleotide sequences. The numbers after the name of the bacteria are the GenBank nucleotide accession numbers and the numbers above the lines are the percentage bootstrap values of that branch of the tree.

DNADNA hybridization
DNADNA hybridization was performed between FO-92^T and 18 strains, comprising 12 Bacillus species. None of the Bacillus species that showed very high similarities with the 16S rDNA sequences (∼97 %) exhibited >70 % DNADNA reassociation values that would place the strain within the same species. Particularly, the similarity between FO-92^T and B. circulans ATCC 4513^T was only 16 %. This pair showed 98·7 % similarity in their 16S rDNA sequences. Similarly, FO-92^T and B. benzoevorans ATCC 49005^T showed only 15 % DNADNA hybridization values whereas this pair exhibited ∼98 % similarity in their 16S rDNA sequence. Based on the DNADNA reassociation values, FO-92^T is a novel Bacillus species.

Description of Bacillus nealsonii sp. nov
Bacillus nealsonii (neal'son.i.i. N.L. gen. n. nealsonii referring to Kenneth H. Nealson, a well-known American microbiologist).

Cells are rod-shaped, 45 µm in length, 1 µm in diameter and motile. Gram-positive, facultatively anaerobic and forms endospores. Spores show an additional extraneous layer similar to an exosporium. Colonies on TSA are irregular, rough, umbonate with undulate or lobate edges and beige in colour. Sodium ions are not essential and it grows at 08 % NaCl. Grows at pH 610, optimum pH 7. Grows at 2560 °C, optimum 3035 °C. Catalase and β-galactosidase are produced, but gelatinase, arginine dihydrolase, lysine and ornithine decarboxylases, lipase, amylase and alginase are not. It neither produces H₂S from thiosulfite nor denitrifies. Based on 16S rDNA nucleotide sequences, this bacterium belongs to the class Firmicutes and is a member of the genus Bacillus. The type strain, FO-92^T (=ATCC BAA-519^T =DSM 15077^T), was isolated from dust particles collected at the Jet Propulsion Laboratory Spacecraft Assembly Facility.

We thank S. Chung, C. Echeverria, R. Koukol, L. Link, M. Musick, J. Sawyer and A. Vu, for technical assistance. Our thanks are due to K. Buxbaum, T. Luchik and R. Manvi for valuable advice and encouragement. We acknowledge J. Edens, P. Koen and J. Kulleck, for assistance performing the electron microscopy, and M. Wiedeman for γ-radiation analyses.