Characterization of Bacillus cereus Isolates Associated with Fatal Pne
http://www.100md.com
《微生物临床杂志》
Epidemiologic Investigations Laboratory, Meningitis and Special Pathogens Branch, MS G34
Office of the Director
Epidemic Intelligence Service, MS E92, Centers for Disease Control and Prevention, 1600 Clifton Road, Atlanta, Georgia 30333
Infectious Disease Epidemiology and Surveillance Division, T-801, Texas Department of State Health Services, Austin, Texas 78756
Bioscience
Decision Applications Divisions, Los Alamos National Laboratory, Los Alamos, New Mexico 87545
Defense Biology Division, Lawrence Livermore National Laboratory, Livermore, California 94551
ABSTRACT
Bacillus cereus is ubiquitous in nature, and while most isolates appear to be harmless, some are associated with food-borne illnesses, periodontal diseases, and other more serious infections. In one such infection, B. cereus G9241 was identified as the causative agent of a severe pneumonia in a Louisiana welder in 1994. This isolate was found to harbor most of the B. anthracis virulence plasmid pXO1 (13). Here we report the characterization of two clinical and one environmental B. cereus isolate collected during an investigation of two fatal pneumonia cases in Texas metal workers. Molecular subtyping revealed that the two cases were not caused by the same strain. However, one of the three isolates was indistinguishable from B. cereus G9241. PCR analysis demonstrated that both clinical isolates contained B. anthracis pXO1 toxin genes. One clinical isolate and the environmental isolate collected from that victim's worksite contained the cap A, B, and C genes required for capsule biosynthesis in B. anthracis. Both clinical isolates expressed a capsule; however, neither was composed of poly-D-glutamic acid. Although most B. cereus isolates are not opportunistic pathogens and only a limited number cause food-borne illnesses, these results demonstrate that some B. cereus strains can cause severe and even fatal infections in patients who appear to be otherwise healthy.
INTRODUCTION
Bacillus cereus is ubiquitous in nature and a common cause of emetic and diarrheal food poisoning. Most B. cereus isolates appear to be harmless, but some are considered opportunistic pathogens. In immunocompromised individuals or patients recovering from surgery, B. cereus can cause a variety of infections, including endophthalmitis, bacteremia, septicemia, endocarditis, salpingitis, cutaneous infections, pneumonia, and meningitis (19). Some strains are known to cause periodontal disease (8). Recently there have been reports of severe and sometimes fatal cases of pneumonia caused by B. cereus in apparently healthy welders (13, 20). The severity of these cases was unusual for B. cereus infections, and the patients were neither immunocompromised nor had any known underlying conditions causing susceptibility to these infections.
B. cereus G9241, which was associated with severe pneumonia in a welder from Louisiana in 1994, has been well characterized, and its genome has been sequenced and analyzed (13). Genomic analysis and multilocus sequence typing (MLST) of this isolate revealed it to be closely related to B. anthracis. Several methods have shown that B. cereus isolates closely related to B. anthracis tend to be of clinical rather than environmental origin (9, 10, 11). However, it is very uncommon for B. cereus isolates, even those that are closely related to B. anthracis, to carry B. anthracis virulence plasmids (23). B. cereus G9241 carries an almost complete pXO1 plasmid, designated pBCXO1. This isolate also harbors a 218-kb circular plasmid (pBC218) and a cryptic bacteriophage (pBClin29) (13). Another unique feature of G9241 relative to other B. cereus isolates is the presence of a capsule. However, this capsule is not composed of D-glutamyl polypeptides and is not encoded by the B. anthracis cap genes normally located on the B. anthracis pXO2 plasmid. Instead, it has been hypothesized to be a polysaccharide and be encoded by a putative polysaccharide capsule biosynthetic operon located on pBC218 (13). While the presence of these plasmids in an isolate that causes severe disease similar to inhalation anthrax is intriguing, their roles, if any, in the virulence of the isolate or the presentation of disease has not yet been determined.
In October and November 2003, two fatal cases of B. cereus pneumonia occurred in metal workers (a welder and a muller operator) at different locations in Texas. A detailed report of these cases and the epidemiologic investigation is in preparation (S. B. Avashia, submitted for publication). In this report, we describe the initial molecular genetic characterization of two clinical and one environmental B. cereus isolate from the investigation of these two fatal cases. We also describe how these isolates are related to each other, to B. anthracis, and to other previously characterized B. cereus and B. thuringiensis isolates.
MATERIALS AND METHODS
Origin of B. cereus isolates. The two clinical B. cereus isolates described in this paper were collected from two fatal pneumonia cases that occurred in Texas metal workers within a 3-week period in 2003. The first isolate, B. cereus 03BB102, was collected from a 39-year-old white male welder, while the second isolate, B. cereus 03BB87, was cultured from a 56-year-old black male muller operator. Both B. cereus isolates were cultured from blood. In an effort to identify the source of the infections, environmental samples were cultured for Bacillus spp. and isolates were screened by PCR for the presence of pXO1 or pXO2. This screen resulted in the identification of a single B. cereus isolate from settled dust at the welder's worksite; the isolate, 03BB108, was PCR positive for the pXO2 cap genes and thus further characterized in this study. A summary of the strains and their common identifiers used throughout the paper is included in Table 1. The B. cereus, B. anthracis, and B. thuringiensis isolates used for comparison in phylogenetic studies were provided by the U.S. Department of Homeland Security Microbial Strain Archive maintained at Los Alamos National Laboratory, Brigham Young University, and the Centers for Disease Control and Prevention.
Biochemical and phenotypic characterization. The B. cereus isolates were characterized by standard microbiological methods (19). Motility was determined by microscopic observation of wet mounts of cells grown in heart infusion broth. Testing to determine susceptibility to gamma phage (3) was performed by adding 5 μl of gamma phage (3.8 x 108 PFU/ml) on the first and second quadrants of isolation streaks on Trypticase soy agar plates containing 5% (vol/vol) sheep blood (Becton Dickinson Microbiology Systems, Cockeysville, MD) and incubating at 37°C overnight. The B. anthracis-specific, two-component direct fluorescent assay (DFA) for the detection of the B. anthracis cell wall antigen and capsule antigen was performed as previously described by De et al. (6). Capsule staining of cells was performed on cells incubated at 30 and 37°C in heart infusion broth supplemented with 0.8% sodium bicarbonate and 50% horse serum as used for the capsule DFA (6). Capsules were visualized using India ink (REMEL, Lenexa, KS) and a 100x oil immersion objective as described previously (13).
DNA isolation and purification. Five milliliters of nutrient broth was inoculated with bacteria from a single colony of each isolate, and each culture was incubated overnight with shaking at 28°C. Four colonies from each isolate were used to produce four different cultures. Bacterial cells were harvested by centrifugation at 1,000 x g for 15 min. The resulting pellets were subjected to three freeze-thaw cycles. DNA was isolated from the disrupted cells using a QIAamp tissue kit (QIAGEN, Inc., Valencia, CA) following the protocol provided by the manufacturer. The quantity and quality of the isolated DNA were determined by gel electrophoresis of a small amount of the sample through a 1.0% agarose gel dissolved in a solution containing 10 mM Tris borate (pH 8.3) and 1 mM EDTA for 1 h at 80 V (26). Gels were stained for 20 min with a solution containing 1 μg ethidium bromide/ml, destained in distilled water, and visualized and photographed under UV light. Known concentrations of DNA molecular-weight standards were compared to the isolated bacterial DNA to determine the size and approximate concentration of the bacterial DNA.
PCR amplification of DNA. DNA oligomers that functioned as PCR primers to amplify representative B. anthracis gene fragments are listed in Table S1 in the supplemental material. Primers designed to amplify a portion of all bacterial 16S rRNA genes were used as positive PCR controls. Initial PCR contained 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 0.001% (wt/vol) gelatin, 0.2 mM of each dNTP, 20 pmol of each primer, 2.5 U of AmpliTaq DNA polymerase, and approximately 1 ng template DNA in a 100-μl total reaction volume. PerkinElmer reagents (Foster City, CA) were used for all reactions. Template DNA was initially denatured by heating at 94°C for 2 min. This was followed by 35 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 1 min, and primer extension at 72°C for 1 min. Incubation for 5 min at 72°C followed to complete the extension. PCR was conducted in a PerkinElmer GeneAmp PCR system 9600 thermal cycler. Control reactions containing B. anthracis 91-429C-2 DNA demonstrated whether each PCR primer set functioned as expected to amplify the appropriate B. anthracis plasmid or genomic amplicon.
Analysis of PCR amplicons. PCR amplicons were analyzed by electrophoresis through a 3.0% agarose gel dissolved in a solution containing 10 mM Tris borate (pH 8.3) and 1 mM EDTA. Electrophoresis was for 1 h at 80 V. Gels were stained for 20 min with a solution containing 1 μg of ethidium bromide/ml, destained in distilled water, and then visualized and photographed under UV light. Images were captured electronically using a Stratagene Eagle Eye II still video system (Stratagene).
Protective antigen (PA) gene sequencing and multilocus sequence typing. The pagA gene was amplified and sequenced as described previously (12, 24). However, additional primers amplifying a 1,076-nucleotide fragment between pXO1 nucleotide 134,396 and 135,472 (see Table S1 in the supplemental material) were used in an attempt to amplify the 5' end of pagA in B. cereus 03BB102 because this gene failed to amplify using previously described primers. MLST was based on the analysis of seven housekeeping gene partial sequences (glpF, gmk, ilvD, pta, pur, pycA, and tpi) and is described online at http://pubmlst.org/bcereus (25). Phylogenetic and molecular evolutionary analyses were conducted using MEGA, version 2.1 (18).
AFLP analysis of DNA samples. Amplified fragment length polymorphism (AFLP) analysis was accomplished as previously described (11, 14, 15, 26). Briefly, 100 ng of DNA was digested with the restriction endonucleases EcoRI and MseI and the resulting fragments were ligated to double-stranded adapters. The digested and ligated DNA was then amplified by PCR using EcoRI and MseI +0/+0 primers (5'-GTAGACTGCGTACCAATTC-3' and 5'-GACGATGAGTCCTGAGTAA-3', respectively). The + 0/+0 PCR product was analyzed by agarose gel electrophoresis to determine the size range of amplified fragments. Three microliters of each product was used as a template in subsequent selective amplifications using the +1/+1 primer combination of 6-carboxyfluorescein-labeled EcoRI-C (5'-GTAGACTGCGTACCAATTCC-3') and MseI-G (5'-GACGATGAGTCCTGAGTAAG-3'). Selective amplifications were performed in 20-μl reaction mixtures. The resulting products (0.5 to 1.0 μl) were mixed with a solution containing a mixture of DNA size standards (GeneScan-500 [Applied Biosystems Inc., Foster City, CA] and MapMarker 500 [BioVentures, Inc., Murfreesboro, TN]) labeled with N,N,N,N-tetramethyl-6-carboxyrhodamine. Following a 2-min heat denaturation at 90°C, the reactions were loaded onto 5% Long Ranger DNA sequencing gels (Cambrex Bio Science, Rockland, ME) and visualized on an ABI 377 automated fluorescent sequencer. To make sure that the DNA assayed in these experiments was representative of each of the isolates, four individual colonies from each isolate were inoculated into medium and DNA purified. The DNA preparations were then analyzed using AFLP. Forty DNA fragments generated from the AFLP experiments were used to represent each of the DNA preparations. These DNA preparations from individual colonies of the same isolate were indistinguishable from each other. Therefore, only one DNA preparation was used in the other experiments. Each set of reactions also contained an AFLP reaction using B. anthracis Vollum DNA as an experimental control. The inclusion of such a reaction in each analysis set allowed a comparison of results from earlier archived analysis sets run at different times or on different gels. GeneScan analysis software (Applied Biosystems, Inc.) was used to determine the lengths of the sample fragments by comparisons to the DNA fragment length size standards included with each sample.
To minimize gel electrophoresis artifacts, each labeling reaction was run in triplicate. Samples were loaded on three different gels in a random order. AFLP data analysis was performed as previously described by Ticknor et al. (26). Sample fragments of between 100 and 500 bp, with fluorescence above 50 arbitrary units in all three runs on the ABI sequencer, were used in the analysis.
Similarities among samples were determined by the Jaccard coefficient. The 40 tallest peaks for each sample fingerprint were used to calculate the Jaccard coefficient among samples. Dendrograms were produced by using the similarity matrix of Jaccard coefficients and the unweighted pair-group mean average method (F. J. Rohlf, NTSYS-PC numerical taxonomy and multivariate analysis system, version 1.8; Exeter Software, Setauket, NY). All statistical data manipulations were performed using codes developed in S-Plus (Data Analysis Products Division, MathSoft, Seattle, WA).
RESULTS
Phenotypic analysis. The B. cereus isolates were hemolytic, motile, and resistant to lysis by gamma-phage and had phenotypes typical of the species (Table 1). Each of the three isolates had distinct colony morphologies. B. cereus 03BB102 (isolated from the Texas welder) produced small, convex, smooth colonies, demonstrating a greater zone of hemolysis on sheep blood agar than that of B. cereus 03BB87 (isolated from the Texas muller operator). This latter isolate produced large, flat, tan, granular colonies. The B. cereus 03BB108 (collected from the welder's worksite) was intermediate in size, umbonate (having a raised center), tan, and granular.
The two-component B. anthracis-specific DFA (6) was performed on all three isolates and produced positive results for the cell wall antigen but was negative for the B. anthracis capsule antigen (poly--D-glutamic acid) when cells were incubated under several conditions, including conditions used for B. anthracis capsule expression. B. cereus clinical isolates are sometimes positive for the cell wall antigen. B. cereus G9241, isolated from a Louisiana welder (13), also tested positive for this cell wall antigen. However, unlike the results obtained for B. cereus G9241, India ink staining of the two Texas clinical isolates did not initially reveal any type of capsule when grown on sheep blood agar at 37°C in ambient atmosphere due to a very small percentage of the cells being encapsulated (Table 1). Adjusting the medium and temperature resulted in increased capsule formation, but not to the level exhibited by B. cereus G9241. A capsule was not detected on the environmental isolate 03BB108.
Plasmid PCR analysis. PCR analysis was performed using primers designed to amplify sequences from pXO1, pXO2, and pBC218. Table S1 in the supplemental material shows the sequences and targets for a battery of PCR primers used to assay for the presence of different pXO1, pXO2, and pBC218 open reading frames (ORFs). Table 2 shows the results of PCR assays to detect pXO1 and pXO2 sequences. All primers generated the expected amplicon when the control strain (B. anthracis 91-429C-2) DNA was used as the template.
The results obtained for B. cereus 03BB87 were identical to those obtained for B. cereus G9241, which was collected from the Louisiana welder almost 10 years earlier. Both were either missing pXO1-ORF 1, pXO1-ORF 135, and pXO1-ORF 143 or the sequences in the priming sites within these ORFs were sufficiently different so that the primers used did not amplify a product. Sequences complementary to these primer sequences were identical in the four B. anthracis isolates so far sequenced. A search of the G9241 plasmid pBC218 did not show homology to any of these primer sequences. In contrast, neither B. cereus 03BB87 nor B. cereus G9241 contained sequences with sufficient homology to any of the pXO2 sequences tested to allow amplification of the different pXO2-specific ORF fragments.
B. cereus 03BB102 and 03BB108, isolated from the Texas welder and his worksite, respectively, did not contain sequences of sufficient homology to generate PCR amplicons from many of the pXO1 ORFs tested (Table 2). The amplification profile of B. cereus 03BB108 was a subset of that for B. cereus 03BB102. This latter isolate contained pXO1 ORFs that were not present in B. cereus 03BB108. A possible explanation of this result is a loss of DNA from B. cereus 03BB102 to produce the profile seen in B. cereus 03BB108, but the pattern of ORFs missing from B. cereus 03BB108 relative to B. cereus 03BB102 is complex relative to the order of these ORFs in the B. anthracis pXO1 plasmid. While B. cereus 03BB102 tested positive for several genes located within the pXO1 pathogenicity island (22), including the three toxin genes (pagA, lef, and cya), the environmental isolate (B. cereus 03BB108) was missing the majority of the pathogenicity island ORFs.
Analyses to detect pXO2-specific ORFs revealed that, while little if any of this plasmid was present in either B. cereus 03BB102 or 03BB108, the capA, capC, and capB genes, which are required for biosynthesis of the B. anthracis capsule, were detected in both isolates (Table 2). Genomic analysis of B. cereus G9241 identified a 218-kb plasmid, pBC218, which includes a polysaccharide capsule operon that was hypothesized to encode the observed capsule for this isolate (13). PCR analysis using primers that are complementary to pBC218 (Table 3) showed that B. cereus 03BB87 has an identical PCR profile to B. cereus G9241 for 32 different pBC218 markers scattered around the plasmid, strongly suggesting the presence of pBC218 or a very similar plasmid in B. cereus 03BB87. In particular, B. cereus 03BB87 possessed the putative polysaccharide polymerase gene (ORF pBC218-0073) thought to participate in the synthesis of a polysaccharide capsule, while B. cereus 03BB102 and 03BB108 did not possess this gene.
PCR analysis using primers that amplify sequences dispersed throughout the B. anthracis genome and shown to be very specific for B. anthracis chromosomal sequences showed that none of these sequences were present in B. cereus 03BB87, 03BB102, 03BB108, or G9241. (Table 2). In contrast, these sequences are present in all 79 B. anthracis isolates tested to date.
pagA sequencing. The B. anthracis protective antigen gene pagA was detected by PCR in B. cereus 03BB87 and 03BB102. To determine similarities between the B. anthracis pagA gene, G9241 pagA, and the pagA sequences from these isolates, we attempted to amplify and sequence pagA from each of the Texas clinical isolates. The pagA sequence in B. cereus 03BB87 was identical to the pagA sequence from B. cereus G9241 (13) and has been assigned the designation of PA genotype 9 (Table 1). Attempts to amplify the extreme 5' end of pagA from B. cereus 03BB102 were not successful, despite using additional primers. We were therefore able to generate only a partial pagA sequence that was missing the first 15 bases relative to the published B. anthracis pagA sequence (see reference 24 and other GenBank entries for this gene). A comparison of this partial sequence to the pagA gene sequence amplified from B. cereus 03BB87 and G9241 revealed two additional mutations at positions 1862 and 1898, which further distinguishes them from the other two isolates and previously sequenced B. anthracis isolates. These two mutations result in serine-to-asparagine and lysine-to-threonine changes, respectively, and are located in the N-terminal 20-kDa PA fragment (PA20) that is cleaved from PA63 (63-kDa C-terminal fragment) by a class of furin proteases (21).
Molecular subtyping. All three Texas isolates and the Louisiana isolate (G9241) were analyzed using AFLP to understand their genetic relationship to one another, to B. anthracis, and to other previously characterized B. cereus and B. thuringiensis isolates. AFLP is a multilocus sampling method that relies on the generation of a fingerprint or signature that is created by restriction enzyme digestion of the entire genome and any extrachromosomal material that might be present. After digestion with restriction enzymes and further selective PCR amplification, the AFLP fingerprint for an isolate was represented by approximately 40 fragments between 100 and 500 bp in length that were present in all three replicates of each isolate. These fingerprint fragment sizes were then used to generate a phylogenetic tree that illustrates the genetic relationships among the isolates analyzed. The fingerprints generated from the three Texas isolates were compared to a fingerprint generated from total DNA isolated from the B. anthracis Vollum strain that is included in every AFLP analysis as a control. They were also compared to fingerprints generated for a battery of B. cereus and B. thuringiensis isolates that are known to be closely related to B. anthracis. Figure 1 shows that, relative to the majority of B. cereus and B. thuringiensis isolates studied, all three Texas isolates (B. cereus 03BB87, 03BB102, and 03BB108) and the Louisiana isolate (B. cereus G9241) are very closely related to B. anthracis (11, 26). However, they are clearly distinct from this pathogen. They mapped to a branch of the phylogenetic tree that includes many of the known pathogenic and toxigenic B. cereus and B. thuringiensis isolates so far characterized (11), while most B. cereus and B. thuringiensis isolates map to other branches of the tree. AFLP analysis could not differentiate B. cereus 03BB87, isolated from the Texas muller operator, from B. cereus G9241, isolated from the Louisiana welder in 1994. In contrast, B. cereus 03BB102 is phylogenetically significantly different from these two clinical isolates. Its AFLP fragment profile was similar to, but distinct from, the profile generated for B. cereus 03BB108, the isolate collected from the working environment of the B. cereus 03BB102-infected victim (Texas welder). Figure 2 shows a direct comparison of a major portion of the AFLP DNA fragment profiles for the two clinical isolates and that of G9241. A direct comparison of the AFLP profiles of B. cereus G9241, isolated in 1994 from a Louisiana welder, and B. cereus 03BB87, isolated in 2003 from a Texas muller operator (Fig 2A), showed that, within the resolution of the method, the isolates were virtually identical. In contrast, Fig. 2B shows the B. cereus 03BB102 profile. There were distinct differences between this profile and the two profiles shown in Fig. 2A. The AFLP profile of B. cereus 03BB102 from the Texas welder, was similar to but distinct from the profile for B. cereus 03BB108, which was collected from the welder's work environment (Fig. 1).
In addition to AFLP, we applied a MLST scheme that was originally used to subtype B. cereus G9241 (13) to subtype the three Texas isolates. This method compares partial nucleotide sequences of seven housekeeping gene sequences (glpF, gmk, ilvD, pta, pur, pycA, and tpi) (25). B. cereus 03BB87 and G9241 were indistinguishable by this MLST method (Fig. 3) (MLST type 78). B. cereus 03BB102 and 03BB108 were closely related to each other and more closely related to B. anthracis than were B. cereus 03BB87 and G9241, in agreement with the AFLP analyses (compare results shown in Fig. 1 and 3). B. cereus 03BB102 and 03BB108 were also closely related to another clinical B. cereus isolate (R_3039/03) in the MLST database (MLST genotype 75), originally collected in England (Fig. 3).
DISCUSSION
B. cereus isolates do not frequently cause pneumonia or other serious diseases. However, there are several reports of severe and often fatal pneumonia in welders caused by B. cereus (1, 2, 14, 20). The genome sequence of B. cereus G9241, isolated from a Louisiana welder, shares significant homology with B. anthracis genome sequences and harbors almost the entire pXO1 anthrax virulence plasmid. Here we report the characterization of three B. cereus isolates associated with two fatal pneumonia cases that occurred in 2003 within a 3-week period in Texas. The two victims were from geographically distinct locations in Texas, but both were metal workers (a welder and a muller operator). Although it is highly unusual for two fatal cases of B. cereus pneumonia to occur in Texas, both in metal workers and within a few weeks of each other, there was no identified epidemiological link between the cases (S. B. Avashia, submitted for publication). The molecular methods employed here to characterize the isolates also show that the two Texas clinical isolates are clearly distinct from one another and thus support the epidemiological assessment. However, B. cereus 03BB87, isolated from the muller operator, is indistinguishable from B. cereus G9241, isolated in 1994 from a Louisiana welder, by all molecular methods applied to the isolates in this study. The only difference detected between these two isolates was variation in the degree of capsule expression. PCR analysis suggests that B. cereus 03BB87, like B. cereus G9241, carries the pBC218 operon, which is speculated to be responsible for capsule formation in the latter isolate (13). Further studies of these two isolates will be required to understand differences in capsule expression in these pathogens.
B. cereus 03BB102 and 03BB108, from the Texas welder and his workplace, respectively, are distinctly different from B. cereus 03BB87, collected from the Texas muller operator, based on AFLP and MLST analyses (Fig. 1, 2, and 3). These two isolates are similar to, but distinct from, one another. AFLP analysis places them in a small subbranch of the phylogenetic tree (Fig. 1). MLST results (Fig. 3) are consistent with this finding, showing the two isolates to be similar but distinct. PCR analysis (Table 2) also reveals differences in their pXO1 gene profiles. The isolates also manifest different colony morphologies. It is clear from these results that B. cereus 03BB108 was not derived from B. cereus 03BB102 by a simple loss of plasmid DNA material since the two isolates are genetically distinct from one another beyond this level. A comparison of AFLP and MLST profiles for B. cereus 03BB102 and 03BB108 to those generated from B. anthracis and other B. cereus and B. thuringiensis isolates (Fig. 1 and 3) demonstrates that, while these isolates may contain fewer sequences homologous to B. anthracis pXO1 and pXO2 sequences, they are phylogenetically more similar to B. anthracis than are B. cereus 03BB87 and G9241. They are also closely related to two other pathogenic Bacillus isolates (B. thuringiensis 97-27 and B. cereus E33L) and two toxigenic B. cereus isolates (D17 and 3a) (Fig. 1).
The presence of the capA, B, and C genes in B. cereus 03BB102 and 03BB108 suggests that these genes may play a role in virulence as they do in B. anthracis. However, we did not observe the expression of the poly--D-glutamic acid capsule in these isolates under in vitro conditions where it is normally expressed in B. anthracis and we do not yet know whether the entire genes are present or under what conditions they might be expressed. The presence of the cap gene sequences in the absence of any other detectable pXO2 sequences is unique relative to other isolates so far characterized. Most of the differences in the detection of pXO1 genes between these isolates are within the pathogenicity island (22).
Despite the fact that B. anthracis toxin genes are found in several of the B. cereus isolates associated with severe pneumonia in metal workers, we do not know whether these genes are appropriately expressed or whether any expressed toxins are functional or play a role in virulence. At least one such isolate, described by Miller et al. (20), did not contain the pagA gene yet still caused a fatal infection in a previously healthy individual (13). However, relative to the general population, professional exposure to a variety of environmental factors may place welders at additional risk of acquiring such infections. Reports suggest welders experience higher frequencies of pulmonary infections with increased severity and duration (1, 2). The majority of data suggest that while welders may not be at increased risk for infections overall, certain conditions and practices may impact susceptibility to pneumonias (1, 4, 7). It has also been suggested that increased exposure to these types of B. cereus isolates and other respiratory pathogens may play a role in increased pneumonia in foundry workers, including welders (5, 17).
All of the isolates clustered in a branch of the AFLP-based phylogenetic tree occupied by B. anthracis and several other pathogenic B. cereus and B. thuringiensis isolates so far characterized (Fig. 1). It is also intriguing that they were collected in a relatively small geographic area of the United States (Texas and Louisiana) and that this area coincides with regions where anthrax occurs naturally in herbivores. If these isolates acquired these plasmids or plasmid sequences via horizontal transfer from B. anthracis in the environment, it is tempting to speculate that such isolates may be restricted to or at least be more common in areas where anthrax is endemic.
B. cereus is ubiquitous in the environment, which can make it difficult to link clinical cases to their environmental sources. Although the source of the clinical isolate was not identified for either case, the isolation of 03BB108 from the welder's worksite illustrates that closely related B. cereus isolates were present. Increased sampling and analysis of isolates may have allowed for the detection of isolates in the environment that matched the clinical isolate. It is also possible that the source of infection was present in material to which both victims were occupationally exposed because they shared similar work environments. This scenario would be more likely if the same strain infected both Texas victims. However, B. cereus 03BB87, isolated from the Texas muller operator, is very similar or identical to B. cereus G9241, isolated from the Louisiana welder 10 years earlier. Therefore, the possibility of a common source is very intriguing, even though these cases were separated temporally by 10 years.
Taken together with previous reports, these results clearly demonstrate the need for increased awareness of the potential for B. cereus to cause serious systemic infections, even in patients who appear to be otherwise healthy. They also demonstrate the need for a better understanding of environmental and occupational risk factors that may increase the susceptibility of individuals to infection by B. cereus.
ACKNOWLEDGMENTS
The work described in the manuscript (review number UCRL-JRNL-219753) was partially performed under the auspices of the U.S. Department of Energy by Los Alamos National Laboratory under contract W-7405-ENG-36 and Lawrence Livermore National Laboratory under contract W-7405-ENG-48.
Funding for a portion of this work was provided by the United States Department of Homeland Security.
FOOTNOTES
Corresponding author. Mailing address: 1600 Clifton Road, MS G34, Atlanta, GA 30333. Phone: (404) 639-0852. Fax: (404) 639-3023. E-mail: amh9@cdc.gov.
Supplemental material for this article may be found at http://jcm.asm.org/.
Present address: Department of Microbiology and Immunology, Emory University School of Medicine, 3004 Rollins Research Center, Atlanta, GA 30322.
Present address: Internal Medicine/Pediatrics, Seton Kozmetsky Clinic, 3706 South First Street Austin, TX 78704.
REFERENCES
Antonini, J. M., A. B. Lewi, J. R. Roberts, and D. A. Whaley. 2003. Pulmonary effects of welding fumes: review of worker and experimental animal studies. Am. J. Ind. Med. 43:350-360.
Antonini, J. M., M. D. Taylor, L. Millecchia, A. R. Bebout, and J. R. Roberts. 2004. Suppression in lung defense responses after bacterial infection in rats pretreated with different welding fumes. Toxicol. Appl. Pharmacol. 200:206-218.
Brown, E. R., and W. B. Cherry. 1955. Specific identification of Bacillus anthracis by means of a variant bacteriophage. J. Infect. Dis. 96:34-39.
Coggon, D., H. Inskip, P. Winter, and B. Pannett. 1994. Lobar pneumonia: an occupational disease in welders. Lancet 344:4-5.
Cordes, L. G., E. W. Brink, P. J. Checko, A. Lentnek, R. W. Lyons, P. S. Hayes, T. C. Wu, D. G. Tharr, and D. W. Fraser. 1981. A cluster of Acinetobacter pneumonia in foundry workers. Ann. Intern. Med. 95:688-693.
De, B. K., S. L. Bragg, G. N. Sanden, K. E Wilson, L. A. Diem, C. K. Marston, A. R. Hoffmaster, G. A. Barnett, R. S. Weyant, T. G. Abshire, J. W. Ezzell, and T. Popovic. 2002. A two-component direct fluorescent-antibody assay for rapid identification of Bacillus anthracis. Emerg. Infect. Dis. 8:1060-1065.
Doig, A. T., and P. J. Challen. 1964. Respiratory hazards in welding. Ann. Occup. Hyg. 111:223-231.
Helgason, E., D. A. Caugant, I. Olsen, and A.-B. Kolst. 2000. Genetic structure of population of Bacillus cereus and B. thuringiensis isolates associated with periodontitis and other human infections. J. Clin. Microbiol. 38:1615-1622.
Helgason, E., O. A. kstad, D. A. Caugant, H. A. Johansen, A. Fouet, M. Mock, I. Hegna, and A.-B. Kolst. 2000. Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis—one species on the basis of genetic evidence. Appl. Environ. Microbiol. 66:2627-2630.
Helgason, E., N. J. Tourasse, R. Meisal, D. A. Caugant, and A.-B. Kolsto. 2004. Multilocus sequence typing scheme for bacteria of the Bacillus cereus group. Appl. Environ. Microbiol. 70:191-201.
Hill, K. K., L. O. Ticknor, R. T. Okinaka, M. Asay, H. Blair, K. A. Bliss, M. Laker, P. E. Pardington, A. P. Richardson, M. Tonks, D. J. Beecher, J. D. Kemp, A.-B. Kolsto, A. C. Lee Wong, P. Keim, and P. J. Jackson. 2004. Fluorescent amplified fragment length polymorphism analysis of Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis isolates. Appl. Environ. Microbiol. 70:1068-1080.
Hoffmaster, A. R., C. C. Fitzgerald, E. Ribot, L. W. Mayer, and T. Popovic. 2002. Molecular subtyping of Bacillus anthracis and the 2001 bioterrorism-associated anthrax outbreak, United States. Emerg. Infect. Dis. 8:1111-1116.
Hoffmaster, A. R., J. Ravel, D. A. Rasko, G. D. Chapman, M. D. Chute, C. K. Marston, B. K. De, C. T. Sacchi, C. Fitzgerald, L. W. Mayer, M. C. J. Maiden, F. G. Priest, M. Barker, L. Jiang, R. Z. Cer, J. Rilsone, S. N. Peterson, R. S. Weyant, D. R. Galloway, T. D. Read, T. Popovic, and C. M. Fraser. 2004. Identification of anthrax toxin genes in a Bacillus cereus associated with an illness resembling inhalation anthrax. Proc. Natl. Acad. Sci. USA 101:8449-8454.
Jackson, P. J., K. K. Hill, M. T. Laker, L. O. Ticknor, and P. Keim. 1999. Genetic comparison of Bacillus anthracis and its close relatives using amplified fragment length polymorphism and polymerase chain reaction analysis. J. Appl. Microbiol. 87:263-269.
Keim, P., A. Kalif, J. Schupp, K. Hill, S. E. Travis, K. Richmond, D. M. Adair, M. Hugh-Jones, C. R. Kuske, and P. Jackson. 1997. Molecular evolution and diversity in Bacillus anthracis as detected by amplified fragment length polymorphism markers. J. Bacteriol. 179:818-824.
Reference deleted.
Kennedy, S. M. 1994. When is a disease occupational Lancet 344:41-43.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17:1244-1255.
Logan, N. A., and P. C. Turnbull. 1999. Bacillus and recently derived genera, p. 357-369. In P. R. Murray (ed.), Manual of clinical microbiology. American Society for Microbiology, Washington, D.C.
Miller, J. M., J. G. Hair, M. Hebert, L. Hebert, F. J. Roberts, Jr., and R. S. Weyant. 1997. Fulminating bacteremia and pneumonia due to Bacillus cereus. J. Clin. Microbiol. 35:504-507.
Mock, M., and T. Mignot. 2003. Anthrax toxins and the host: a story of intimacy. Cell. Microbiol. 5:15-23.
Okinaka, R. T., K. Cloud, O. Hampton, A. R. Hoffmaster, K. K. Hill, P. Keim, T. M. Koehler, G. Lamke, S. Kumano. J. Mahillon, D. Manter, Y. Martinez, D. Ricke, R. Svensson, and P. J. Jackson. 1999. Sequence and organization of pXO1, the large Bacillus anthracis plasmid harboring the anthrax toxin genes. J. Bacteriol. 181:6509-6515.
Pannucci, J., R. T. Okinaka, R. Sabin, and C. R. Kuske. 2002. Bacillus anthracis pXO1 plasmid sequence conservation among closely related bacterial species. J. Bacteriol. 184:134-141.
Price, L. B., M. Hugh-Jones, P. J. Jackson, and P. Keim. 1999. Genetic diversity in the protective antigen gene of Bacillus anthracis. J. Bacteriol. 181:2358-2362.
Priest, F. G., M. Barker, L. W. Baillie, E. C. Holmes, and M. C. Maiden. 2004. Population structure and evolution of the Bacillus cereus group. J. Bacteriol. 186:7959-7970.
Ticknor, L. O., A.-B. Kolst, K. K. Hill, P. Keim, M. T. Laker, M. Tonks, and P. J. Jackson. 2001. Fluorescent amplified fragment length polymorphism analysis of Norwegian Bacillus cereus and Bacillus thuringiensis soil isolates. Appl. Environ. Microbiol. 67:4863-4873.(Alex R. Hoffmaster, Karen K. Hill, Jay E)
Office of the Director
Epidemic Intelligence Service, MS E92, Centers for Disease Control and Prevention, 1600 Clifton Road, Atlanta, Georgia 30333
Infectious Disease Epidemiology and Surveillance Division, T-801, Texas Department of State Health Services, Austin, Texas 78756
Bioscience
Decision Applications Divisions, Los Alamos National Laboratory, Los Alamos, New Mexico 87545
Defense Biology Division, Lawrence Livermore National Laboratory, Livermore, California 94551
ABSTRACT
Bacillus cereus is ubiquitous in nature, and while most isolates appear to be harmless, some are associated with food-borne illnesses, periodontal diseases, and other more serious infections. In one such infection, B. cereus G9241 was identified as the causative agent of a severe pneumonia in a Louisiana welder in 1994. This isolate was found to harbor most of the B. anthracis virulence plasmid pXO1 (13). Here we report the characterization of two clinical and one environmental B. cereus isolate collected during an investigation of two fatal pneumonia cases in Texas metal workers. Molecular subtyping revealed that the two cases were not caused by the same strain. However, one of the three isolates was indistinguishable from B. cereus G9241. PCR analysis demonstrated that both clinical isolates contained B. anthracis pXO1 toxin genes. One clinical isolate and the environmental isolate collected from that victim's worksite contained the cap A, B, and C genes required for capsule biosynthesis in B. anthracis. Both clinical isolates expressed a capsule; however, neither was composed of poly-D-glutamic acid. Although most B. cereus isolates are not opportunistic pathogens and only a limited number cause food-borne illnesses, these results demonstrate that some B. cereus strains can cause severe and even fatal infections in patients who appear to be otherwise healthy.
INTRODUCTION
Bacillus cereus is ubiquitous in nature and a common cause of emetic and diarrheal food poisoning. Most B. cereus isolates appear to be harmless, but some are considered opportunistic pathogens. In immunocompromised individuals or patients recovering from surgery, B. cereus can cause a variety of infections, including endophthalmitis, bacteremia, septicemia, endocarditis, salpingitis, cutaneous infections, pneumonia, and meningitis (19). Some strains are known to cause periodontal disease (8). Recently there have been reports of severe and sometimes fatal cases of pneumonia caused by B. cereus in apparently healthy welders (13, 20). The severity of these cases was unusual for B. cereus infections, and the patients were neither immunocompromised nor had any known underlying conditions causing susceptibility to these infections.
B. cereus G9241, which was associated with severe pneumonia in a welder from Louisiana in 1994, has been well characterized, and its genome has been sequenced and analyzed (13). Genomic analysis and multilocus sequence typing (MLST) of this isolate revealed it to be closely related to B. anthracis. Several methods have shown that B. cereus isolates closely related to B. anthracis tend to be of clinical rather than environmental origin (9, 10, 11). However, it is very uncommon for B. cereus isolates, even those that are closely related to B. anthracis, to carry B. anthracis virulence plasmids (23). B. cereus G9241 carries an almost complete pXO1 plasmid, designated pBCXO1. This isolate also harbors a 218-kb circular plasmid (pBC218) and a cryptic bacteriophage (pBClin29) (13). Another unique feature of G9241 relative to other B. cereus isolates is the presence of a capsule. However, this capsule is not composed of D-glutamyl polypeptides and is not encoded by the B. anthracis cap genes normally located on the B. anthracis pXO2 plasmid. Instead, it has been hypothesized to be a polysaccharide and be encoded by a putative polysaccharide capsule biosynthetic operon located on pBC218 (13). While the presence of these plasmids in an isolate that causes severe disease similar to inhalation anthrax is intriguing, their roles, if any, in the virulence of the isolate or the presentation of disease has not yet been determined.
In October and November 2003, two fatal cases of B. cereus pneumonia occurred in metal workers (a welder and a muller operator) at different locations in Texas. A detailed report of these cases and the epidemiologic investigation is in preparation (S. B. Avashia, submitted for publication). In this report, we describe the initial molecular genetic characterization of two clinical and one environmental B. cereus isolate from the investigation of these two fatal cases. We also describe how these isolates are related to each other, to B. anthracis, and to other previously characterized B. cereus and B. thuringiensis isolates.
MATERIALS AND METHODS
Origin of B. cereus isolates. The two clinical B. cereus isolates described in this paper were collected from two fatal pneumonia cases that occurred in Texas metal workers within a 3-week period in 2003. The first isolate, B. cereus 03BB102, was collected from a 39-year-old white male welder, while the second isolate, B. cereus 03BB87, was cultured from a 56-year-old black male muller operator. Both B. cereus isolates were cultured from blood. In an effort to identify the source of the infections, environmental samples were cultured for Bacillus spp. and isolates were screened by PCR for the presence of pXO1 or pXO2. This screen resulted in the identification of a single B. cereus isolate from settled dust at the welder's worksite; the isolate, 03BB108, was PCR positive for the pXO2 cap genes and thus further characterized in this study. A summary of the strains and their common identifiers used throughout the paper is included in Table 1. The B. cereus, B. anthracis, and B. thuringiensis isolates used for comparison in phylogenetic studies were provided by the U.S. Department of Homeland Security Microbial Strain Archive maintained at Los Alamos National Laboratory, Brigham Young University, and the Centers for Disease Control and Prevention.
Biochemical and phenotypic characterization. The B. cereus isolates were characterized by standard microbiological methods (19). Motility was determined by microscopic observation of wet mounts of cells grown in heart infusion broth. Testing to determine susceptibility to gamma phage (3) was performed by adding 5 μl of gamma phage (3.8 x 108 PFU/ml) on the first and second quadrants of isolation streaks on Trypticase soy agar plates containing 5% (vol/vol) sheep blood (Becton Dickinson Microbiology Systems, Cockeysville, MD) and incubating at 37°C overnight. The B. anthracis-specific, two-component direct fluorescent assay (DFA) for the detection of the B. anthracis cell wall antigen and capsule antigen was performed as previously described by De et al. (6). Capsule staining of cells was performed on cells incubated at 30 and 37°C in heart infusion broth supplemented with 0.8% sodium bicarbonate and 50% horse serum as used for the capsule DFA (6). Capsules were visualized using India ink (REMEL, Lenexa, KS) and a 100x oil immersion objective as described previously (13).
DNA isolation and purification. Five milliliters of nutrient broth was inoculated with bacteria from a single colony of each isolate, and each culture was incubated overnight with shaking at 28°C. Four colonies from each isolate were used to produce four different cultures. Bacterial cells were harvested by centrifugation at 1,000 x g for 15 min. The resulting pellets were subjected to three freeze-thaw cycles. DNA was isolated from the disrupted cells using a QIAamp tissue kit (QIAGEN, Inc., Valencia, CA) following the protocol provided by the manufacturer. The quantity and quality of the isolated DNA were determined by gel electrophoresis of a small amount of the sample through a 1.0% agarose gel dissolved in a solution containing 10 mM Tris borate (pH 8.3) and 1 mM EDTA for 1 h at 80 V (26). Gels were stained for 20 min with a solution containing 1 μg ethidium bromide/ml, destained in distilled water, and visualized and photographed under UV light. Known concentrations of DNA molecular-weight standards were compared to the isolated bacterial DNA to determine the size and approximate concentration of the bacterial DNA.
PCR amplification of DNA. DNA oligomers that functioned as PCR primers to amplify representative B. anthracis gene fragments are listed in Table S1 in the supplemental material. Primers designed to amplify a portion of all bacterial 16S rRNA genes were used as positive PCR controls. Initial PCR contained 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 0.001% (wt/vol) gelatin, 0.2 mM of each dNTP, 20 pmol of each primer, 2.5 U of AmpliTaq DNA polymerase, and approximately 1 ng template DNA in a 100-μl total reaction volume. PerkinElmer reagents (Foster City, CA) were used for all reactions. Template DNA was initially denatured by heating at 94°C for 2 min. This was followed by 35 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 1 min, and primer extension at 72°C for 1 min. Incubation for 5 min at 72°C followed to complete the extension. PCR was conducted in a PerkinElmer GeneAmp PCR system 9600 thermal cycler. Control reactions containing B. anthracis 91-429C-2 DNA demonstrated whether each PCR primer set functioned as expected to amplify the appropriate B. anthracis plasmid or genomic amplicon.
Analysis of PCR amplicons. PCR amplicons were analyzed by electrophoresis through a 3.0% agarose gel dissolved in a solution containing 10 mM Tris borate (pH 8.3) and 1 mM EDTA. Electrophoresis was for 1 h at 80 V. Gels were stained for 20 min with a solution containing 1 μg of ethidium bromide/ml, destained in distilled water, and then visualized and photographed under UV light. Images were captured electronically using a Stratagene Eagle Eye II still video system (Stratagene).
Protective antigen (PA) gene sequencing and multilocus sequence typing. The pagA gene was amplified and sequenced as described previously (12, 24). However, additional primers amplifying a 1,076-nucleotide fragment between pXO1 nucleotide 134,396 and 135,472 (see Table S1 in the supplemental material) were used in an attempt to amplify the 5' end of pagA in B. cereus 03BB102 because this gene failed to amplify using previously described primers. MLST was based on the analysis of seven housekeeping gene partial sequences (glpF, gmk, ilvD, pta, pur, pycA, and tpi) and is described online at http://pubmlst.org/bcereus (25). Phylogenetic and molecular evolutionary analyses were conducted using MEGA, version 2.1 (18).
AFLP analysis of DNA samples. Amplified fragment length polymorphism (AFLP) analysis was accomplished as previously described (11, 14, 15, 26). Briefly, 100 ng of DNA was digested with the restriction endonucleases EcoRI and MseI and the resulting fragments were ligated to double-stranded adapters. The digested and ligated DNA was then amplified by PCR using EcoRI and MseI +0/+0 primers (5'-GTAGACTGCGTACCAATTC-3' and 5'-GACGATGAGTCCTGAGTAA-3', respectively). The + 0/+0 PCR product was analyzed by agarose gel electrophoresis to determine the size range of amplified fragments. Three microliters of each product was used as a template in subsequent selective amplifications using the +1/+1 primer combination of 6-carboxyfluorescein-labeled EcoRI-C (5'-GTAGACTGCGTACCAATTCC-3') and MseI-G (5'-GACGATGAGTCCTGAGTAAG-3'). Selective amplifications were performed in 20-μl reaction mixtures. The resulting products (0.5 to 1.0 μl) were mixed with a solution containing a mixture of DNA size standards (GeneScan-500 [Applied Biosystems Inc., Foster City, CA] and MapMarker 500 [BioVentures, Inc., Murfreesboro, TN]) labeled with N,N,N,N-tetramethyl-6-carboxyrhodamine. Following a 2-min heat denaturation at 90°C, the reactions were loaded onto 5% Long Ranger DNA sequencing gels (Cambrex Bio Science, Rockland, ME) and visualized on an ABI 377 automated fluorescent sequencer. To make sure that the DNA assayed in these experiments was representative of each of the isolates, four individual colonies from each isolate were inoculated into medium and DNA purified. The DNA preparations were then analyzed using AFLP. Forty DNA fragments generated from the AFLP experiments were used to represent each of the DNA preparations. These DNA preparations from individual colonies of the same isolate were indistinguishable from each other. Therefore, only one DNA preparation was used in the other experiments. Each set of reactions also contained an AFLP reaction using B. anthracis Vollum DNA as an experimental control. The inclusion of such a reaction in each analysis set allowed a comparison of results from earlier archived analysis sets run at different times or on different gels. GeneScan analysis software (Applied Biosystems, Inc.) was used to determine the lengths of the sample fragments by comparisons to the DNA fragment length size standards included with each sample.
To minimize gel electrophoresis artifacts, each labeling reaction was run in triplicate. Samples were loaded on three different gels in a random order. AFLP data analysis was performed as previously described by Ticknor et al. (26). Sample fragments of between 100 and 500 bp, with fluorescence above 50 arbitrary units in all three runs on the ABI sequencer, were used in the analysis.
Similarities among samples were determined by the Jaccard coefficient. The 40 tallest peaks for each sample fingerprint were used to calculate the Jaccard coefficient among samples. Dendrograms were produced by using the similarity matrix of Jaccard coefficients and the unweighted pair-group mean average method (F. J. Rohlf, NTSYS-PC numerical taxonomy and multivariate analysis system, version 1.8; Exeter Software, Setauket, NY). All statistical data manipulations were performed using codes developed in S-Plus (Data Analysis Products Division, MathSoft, Seattle, WA).
RESULTS
Phenotypic analysis. The B. cereus isolates were hemolytic, motile, and resistant to lysis by gamma-phage and had phenotypes typical of the species (Table 1). Each of the three isolates had distinct colony morphologies. B. cereus 03BB102 (isolated from the Texas welder) produced small, convex, smooth colonies, demonstrating a greater zone of hemolysis on sheep blood agar than that of B. cereus 03BB87 (isolated from the Texas muller operator). This latter isolate produced large, flat, tan, granular colonies. The B. cereus 03BB108 (collected from the welder's worksite) was intermediate in size, umbonate (having a raised center), tan, and granular.
The two-component B. anthracis-specific DFA (6) was performed on all three isolates and produced positive results for the cell wall antigen but was negative for the B. anthracis capsule antigen (poly--D-glutamic acid) when cells were incubated under several conditions, including conditions used for B. anthracis capsule expression. B. cereus clinical isolates are sometimes positive for the cell wall antigen. B. cereus G9241, isolated from a Louisiana welder (13), also tested positive for this cell wall antigen. However, unlike the results obtained for B. cereus G9241, India ink staining of the two Texas clinical isolates did not initially reveal any type of capsule when grown on sheep blood agar at 37°C in ambient atmosphere due to a very small percentage of the cells being encapsulated (Table 1). Adjusting the medium and temperature resulted in increased capsule formation, but not to the level exhibited by B. cereus G9241. A capsule was not detected on the environmental isolate 03BB108.
Plasmid PCR analysis. PCR analysis was performed using primers designed to amplify sequences from pXO1, pXO2, and pBC218. Table S1 in the supplemental material shows the sequences and targets for a battery of PCR primers used to assay for the presence of different pXO1, pXO2, and pBC218 open reading frames (ORFs). Table 2 shows the results of PCR assays to detect pXO1 and pXO2 sequences. All primers generated the expected amplicon when the control strain (B. anthracis 91-429C-2) DNA was used as the template.
The results obtained for B. cereus 03BB87 were identical to those obtained for B. cereus G9241, which was collected from the Louisiana welder almost 10 years earlier. Both were either missing pXO1-ORF 1, pXO1-ORF 135, and pXO1-ORF 143 or the sequences in the priming sites within these ORFs were sufficiently different so that the primers used did not amplify a product. Sequences complementary to these primer sequences were identical in the four B. anthracis isolates so far sequenced. A search of the G9241 plasmid pBC218 did not show homology to any of these primer sequences. In contrast, neither B. cereus 03BB87 nor B. cereus G9241 contained sequences with sufficient homology to any of the pXO2 sequences tested to allow amplification of the different pXO2-specific ORF fragments.
B. cereus 03BB102 and 03BB108, isolated from the Texas welder and his worksite, respectively, did not contain sequences of sufficient homology to generate PCR amplicons from many of the pXO1 ORFs tested (Table 2). The amplification profile of B. cereus 03BB108 was a subset of that for B. cereus 03BB102. This latter isolate contained pXO1 ORFs that were not present in B. cereus 03BB108. A possible explanation of this result is a loss of DNA from B. cereus 03BB102 to produce the profile seen in B. cereus 03BB108, but the pattern of ORFs missing from B. cereus 03BB108 relative to B. cereus 03BB102 is complex relative to the order of these ORFs in the B. anthracis pXO1 plasmid. While B. cereus 03BB102 tested positive for several genes located within the pXO1 pathogenicity island (22), including the three toxin genes (pagA, lef, and cya), the environmental isolate (B. cereus 03BB108) was missing the majority of the pathogenicity island ORFs.
Analyses to detect pXO2-specific ORFs revealed that, while little if any of this plasmid was present in either B. cereus 03BB102 or 03BB108, the capA, capC, and capB genes, which are required for biosynthesis of the B. anthracis capsule, were detected in both isolates (Table 2). Genomic analysis of B. cereus G9241 identified a 218-kb plasmid, pBC218, which includes a polysaccharide capsule operon that was hypothesized to encode the observed capsule for this isolate (13). PCR analysis using primers that are complementary to pBC218 (Table 3) showed that B. cereus 03BB87 has an identical PCR profile to B. cereus G9241 for 32 different pBC218 markers scattered around the plasmid, strongly suggesting the presence of pBC218 or a very similar plasmid in B. cereus 03BB87. In particular, B. cereus 03BB87 possessed the putative polysaccharide polymerase gene (ORF pBC218-0073) thought to participate in the synthesis of a polysaccharide capsule, while B. cereus 03BB102 and 03BB108 did not possess this gene.
PCR analysis using primers that amplify sequences dispersed throughout the B. anthracis genome and shown to be very specific for B. anthracis chromosomal sequences showed that none of these sequences were present in B. cereus 03BB87, 03BB102, 03BB108, or G9241. (Table 2). In contrast, these sequences are present in all 79 B. anthracis isolates tested to date.
pagA sequencing. The B. anthracis protective antigen gene pagA was detected by PCR in B. cereus 03BB87 and 03BB102. To determine similarities between the B. anthracis pagA gene, G9241 pagA, and the pagA sequences from these isolates, we attempted to amplify and sequence pagA from each of the Texas clinical isolates. The pagA sequence in B. cereus 03BB87 was identical to the pagA sequence from B. cereus G9241 (13) and has been assigned the designation of PA genotype 9 (Table 1). Attempts to amplify the extreme 5' end of pagA from B. cereus 03BB102 were not successful, despite using additional primers. We were therefore able to generate only a partial pagA sequence that was missing the first 15 bases relative to the published B. anthracis pagA sequence (see reference 24 and other GenBank entries for this gene). A comparison of this partial sequence to the pagA gene sequence amplified from B. cereus 03BB87 and G9241 revealed two additional mutations at positions 1862 and 1898, which further distinguishes them from the other two isolates and previously sequenced B. anthracis isolates. These two mutations result in serine-to-asparagine and lysine-to-threonine changes, respectively, and are located in the N-terminal 20-kDa PA fragment (PA20) that is cleaved from PA63 (63-kDa C-terminal fragment) by a class of furin proteases (21).
Molecular subtyping. All three Texas isolates and the Louisiana isolate (G9241) were analyzed using AFLP to understand their genetic relationship to one another, to B. anthracis, and to other previously characterized B. cereus and B. thuringiensis isolates. AFLP is a multilocus sampling method that relies on the generation of a fingerprint or signature that is created by restriction enzyme digestion of the entire genome and any extrachromosomal material that might be present. After digestion with restriction enzymes and further selective PCR amplification, the AFLP fingerprint for an isolate was represented by approximately 40 fragments between 100 and 500 bp in length that were present in all three replicates of each isolate. These fingerprint fragment sizes were then used to generate a phylogenetic tree that illustrates the genetic relationships among the isolates analyzed. The fingerprints generated from the three Texas isolates were compared to a fingerprint generated from total DNA isolated from the B. anthracis Vollum strain that is included in every AFLP analysis as a control. They were also compared to fingerprints generated for a battery of B. cereus and B. thuringiensis isolates that are known to be closely related to B. anthracis. Figure 1 shows that, relative to the majority of B. cereus and B. thuringiensis isolates studied, all three Texas isolates (B. cereus 03BB87, 03BB102, and 03BB108) and the Louisiana isolate (B. cereus G9241) are very closely related to B. anthracis (11, 26). However, they are clearly distinct from this pathogen. They mapped to a branch of the phylogenetic tree that includes many of the known pathogenic and toxigenic B. cereus and B. thuringiensis isolates so far characterized (11), while most B. cereus and B. thuringiensis isolates map to other branches of the tree. AFLP analysis could not differentiate B. cereus 03BB87, isolated from the Texas muller operator, from B. cereus G9241, isolated from the Louisiana welder in 1994. In contrast, B. cereus 03BB102 is phylogenetically significantly different from these two clinical isolates. Its AFLP fragment profile was similar to, but distinct from, the profile generated for B. cereus 03BB108, the isolate collected from the working environment of the B. cereus 03BB102-infected victim (Texas welder). Figure 2 shows a direct comparison of a major portion of the AFLP DNA fragment profiles for the two clinical isolates and that of G9241. A direct comparison of the AFLP profiles of B. cereus G9241, isolated in 1994 from a Louisiana welder, and B. cereus 03BB87, isolated in 2003 from a Texas muller operator (Fig 2A), showed that, within the resolution of the method, the isolates were virtually identical. In contrast, Fig. 2B shows the B. cereus 03BB102 profile. There were distinct differences between this profile and the two profiles shown in Fig. 2A. The AFLP profile of B. cereus 03BB102 from the Texas welder, was similar to but distinct from the profile for B. cereus 03BB108, which was collected from the welder's work environment (Fig. 1).
In addition to AFLP, we applied a MLST scheme that was originally used to subtype B. cereus G9241 (13) to subtype the three Texas isolates. This method compares partial nucleotide sequences of seven housekeeping gene sequences (glpF, gmk, ilvD, pta, pur, pycA, and tpi) (25). B. cereus 03BB87 and G9241 were indistinguishable by this MLST method (Fig. 3) (MLST type 78). B. cereus 03BB102 and 03BB108 were closely related to each other and more closely related to B. anthracis than were B. cereus 03BB87 and G9241, in agreement with the AFLP analyses (compare results shown in Fig. 1 and 3). B. cereus 03BB102 and 03BB108 were also closely related to another clinical B. cereus isolate (R_3039/03) in the MLST database (MLST genotype 75), originally collected in England (Fig. 3).
DISCUSSION
B. cereus isolates do not frequently cause pneumonia or other serious diseases. However, there are several reports of severe and often fatal pneumonia in welders caused by B. cereus (1, 2, 14, 20). The genome sequence of B. cereus G9241, isolated from a Louisiana welder, shares significant homology with B. anthracis genome sequences and harbors almost the entire pXO1 anthrax virulence plasmid. Here we report the characterization of three B. cereus isolates associated with two fatal pneumonia cases that occurred in 2003 within a 3-week period in Texas. The two victims were from geographically distinct locations in Texas, but both were metal workers (a welder and a muller operator). Although it is highly unusual for two fatal cases of B. cereus pneumonia to occur in Texas, both in metal workers and within a few weeks of each other, there was no identified epidemiological link between the cases (S. B. Avashia, submitted for publication). The molecular methods employed here to characterize the isolates also show that the two Texas clinical isolates are clearly distinct from one another and thus support the epidemiological assessment. However, B. cereus 03BB87, isolated from the muller operator, is indistinguishable from B. cereus G9241, isolated in 1994 from a Louisiana welder, by all molecular methods applied to the isolates in this study. The only difference detected between these two isolates was variation in the degree of capsule expression. PCR analysis suggests that B. cereus 03BB87, like B. cereus G9241, carries the pBC218 operon, which is speculated to be responsible for capsule formation in the latter isolate (13). Further studies of these two isolates will be required to understand differences in capsule expression in these pathogens.
B. cereus 03BB102 and 03BB108, from the Texas welder and his workplace, respectively, are distinctly different from B. cereus 03BB87, collected from the Texas muller operator, based on AFLP and MLST analyses (Fig. 1, 2, and 3). These two isolates are similar to, but distinct from, one another. AFLP analysis places them in a small subbranch of the phylogenetic tree (Fig. 1). MLST results (Fig. 3) are consistent with this finding, showing the two isolates to be similar but distinct. PCR analysis (Table 2) also reveals differences in their pXO1 gene profiles. The isolates also manifest different colony morphologies. It is clear from these results that B. cereus 03BB108 was not derived from B. cereus 03BB102 by a simple loss of plasmid DNA material since the two isolates are genetically distinct from one another beyond this level. A comparison of AFLP and MLST profiles for B. cereus 03BB102 and 03BB108 to those generated from B. anthracis and other B. cereus and B. thuringiensis isolates (Fig. 1 and 3) demonstrates that, while these isolates may contain fewer sequences homologous to B. anthracis pXO1 and pXO2 sequences, they are phylogenetically more similar to B. anthracis than are B. cereus 03BB87 and G9241. They are also closely related to two other pathogenic Bacillus isolates (B. thuringiensis 97-27 and B. cereus E33L) and two toxigenic B. cereus isolates (D17 and 3a) (Fig. 1).
The presence of the capA, B, and C genes in B. cereus 03BB102 and 03BB108 suggests that these genes may play a role in virulence as they do in B. anthracis. However, we did not observe the expression of the poly--D-glutamic acid capsule in these isolates under in vitro conditions where it is normally expressed in B. anthracis and we do not yet know whether the entire genes are present or under what conditions they might be expressed. The presence of the cap gene sequences in the absence of any other detectable pXO2 sequences is unique relative to other isolates so far characterized. Most of the differences in the detection of pXO1 genes between these isolates are within the pathogenicity island (22).
Despite the fact that B. anthracis toxin genes are found in several of the B. cereus isolates associated with severe pneumonia in metal workers, we do not know whether these genes are appropriately expressed or whether any expressed toxins are functional or play a role in virulence. At least one such isolate, described by Miller et al. (20), did not contain the pagA gene yet still caused a fatal infection in a previously healthy individual (13). However, relative to the general population, professional exposure to a variety of environmental factors may place welders at additional risk of acquiring such infections. Reports suggest welders experience higher frequencies of pulmonary infections with increased severity and duration (1, 2). The majority of data suggest that while welders may not be at increased risk for infections overall, certain conditions and practices may impact susceptibility to pneumonias (1, 4, 7). It has also been suggested that increased exposure to these types of B. cereus isolates and other respiratory pathogens may play a role in increased pneumonia in foundry workers, including welders (5, 17).
All of the isolates clustered in a branch of the AFLP-based phylogenetic tree occupied by B. anthracis and several other pathogenic B. cereus and B. thuringiensis isolates so far characterized (Fig. 1). It is also intriguing that they were collected in a relatively small geographic area of the United States (Texas and Louisiana) and that this area coincides with regions where anthrax occurs naturally in herbivores. If these isolates acquired these plasmids or plasmid sequences via horizontal transfer from B. anthracis in the environment, it is tempting to speculate that such isolates may be restricted to or at least be more common in areas where anthrax is endemic.
B. cereus is ubiquitous in the environment, which can make it difficult to link clinical cases to their environmental sources. Although the source of the clinical isolate was not identified for either case, the isolation of 03BB108 from the welder's worksite illustrates that closely related B. cereus isolates were present. Increased sampling and analysis of isolates may have allowed for the detection of isolates in the environment that matched the clinical isolate. It is also possible that the source of infection was present in material to which both victims were occupationally exposed because they shared similar work environments. This scenario would be more likely if the same strain infected both Texas victims. However, B. cereus 03BB87, isolated from the Texas muller operator, is very similar or identical to B. cereus G9241, isolated from the Louisiana welder 10 years earlier. Therefore, the possibility of a common source is very intriguing, even though these cases were separated temporally by 10 years.
Taken together with previous reports, these results clearly demonstrate the need for increased awareness of the potential for B. cereus to cause serious systemic infections, even in patients who appear to be otherwise healthy. They also demonstrate the need for a better understanding of environmental and occupational risk factors that may increase the susceptibility of individuals to infection by B. cereus.
ACKNOWLEDGMENTS
The work described in the manuscript (review number UCRL-JRNL-219753) was partially performed under the auspices of the U.S. Department of Energy by Los Alamos National Laboratory under contract W-7405-ENG-36 and Lawrence Livermore National Laboratory under contract W-7405-ENG-48.
Funding for a portion of this work was provided by the United States Department of Homeland Security.
FOOTNOTES
Corresponding author. Mailing address: 1600 Clifton Road, MS G34, Atlanta, GA 30333. Phone: (404) 639-0852. Fax: (404) 639-3023. E-mail: amh9@cdc.gov.
Supplemental material for this article may be found at http://jcm.asm.org/.
Present address: Department of Microbiology and Immunology, Emory University School of Medicine, 3004 Rollins Research Center, Atlanta, GA 30322.
Present address: Internal Medicine/Pediatrics, Seton Kozmetsky Clinic, 3706 South First Street Austin, TX 78704.
REFERENCES
Antonini, J. M., A. B. Lewi, J. R. Roberts, and D. A. Whaley. 2003. Pulmonary effects of welding fumes: review of worker and experimental animal studies. Am. J. Ind. Med. 43:350-360.
Antonini, J. M., M. D. Taylor, L. Millecchia, A. R. Bebout, and J. R. Roberts. 2004. Suppression in lung defense responses after bacterial infection in rats pretreated with different welding fumes. Toxicol. Appl. Pharmacol. 200:206-218.
Brown, E. R., and W. B. Cherry. 1955. Specific identification of Bacillus anthracis by means of a variant bacteriophage. J. Infect. Dis. 96:34-39.
Coggon, D., H. Inskip, P. Winter, and B. Pannett. 1994. Lobar pneumonia: an occupational disease in welders. Lancet 344:4-5.
Cordes, L. G., E. W. Brink, P. J. Checko, A. Lentnek, R. W. Lyons, P. S. Hayes, T. C. Wu, D. G. Tharr, and D. W. Fraser. 1981. A cluster of Acinetobacter pneumonia in foundry workers. Ann. Intern. Med. 95:688-693.
De, B. K., S. L. Bragg, G. N. Sanden, K. E Wilson, L. A. Diem, C. K. Marston, A. R. Hoffmaster, G. A. Barnett, R. S. Weyant, T. G. Abshire, J. W. Ezzell, and T. Popovic. 2002. A two-component direct fluorescent-antibody assay for rapid identification of Bacillus anthracis. Emerg. Infect. Dis. 8:1060-1065.
Doig, A. T., and P. J. Challen. 1964. Respiratory hazards in welding. Ann. Occup. Hyg. 111:223-231.
Helgason, E., D. A. Caugant, I. Olsen, and A.-B. Kolst. 2000. Genetic structure of population of Bacillus cereus and B. thuringiensis isolates associated with periodontitis and other human infections. J. Clin. Microbiol. 38:1615-1622.
Helgason, E., O. A. kstad, D. A. Caugant, H. A. Johansen, A. Fouet, M. Mock, I. Hegna, and A.-B. Kolst. 2000. Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis—one species on the basis of genetic evidence. Appl. Environ. Microbiol. 66:2627-2630.
Helgason, E., N. J. Tourasse, R. Meisal, D. A. Caugant, and A.-B. Kolsto. 2004. Multilocus sequence typing scheme for bacteria of the Bacillus cereus group. Appl. Environ. Microbiol. 70:191-201.
Hill, K. K., L. O. Ticknor, R. T. Okinaka, M. Asay, H. Blair, K. A. Bliss, M. Laker, P. E. Pardington, A. P. Richardson, M. Tonks, D. J. Beecher, J. D. Kemp, A.-B. Kolsto, A. C. Lee Wong, P. Keim, and P. J. Jackson. 2004. Fluorescent amplified fragment length polymorphism analysis of Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis isolates. Appl. Environ. Microbiol. 70:1068-1080.
Hoffmaster, A. R., C. C. Fitzgerald, E. Ribot, L. W. Mayer, and T. Popovic. 2002. Molecular subtyping of Bacillus anthracis and the 2001 bioterrorism-associated anthrax outbreak, United States. Emerg. Infect. Dis. 8:1111-1116.
Hoffmaster, A. R., J. Ravel, D. A. Rasko, G. D. Chapman, M. D. Chute, C. K. Marston, B. K. De, C. T. Sacchi, C. Fitzgerald, L. W. Mayer, M. C. J. Maiden, F. G. Priest, M. Barker, L. Jiang, R. Z. Cer, J. Rilsone, S. N. Peterson, R. S. Weyant, D. R. Galloway, T. D. Read, T. Popovic, and C. M. Fraser. 2004. Identification of anthrax toxin genes in a Bacillus cereus associated with an illness resembling inhalation anthrax. Proc. Natl. Acad. Sci. USA 101:8449-8454.
Jackson, P. J., K. K. Hill, M. T. Laker, L. O. Ticknor, and P. Keim. 1999. Genetic comparison of Bacillus anthracis and its close relatives using amplified fragment length polymorphism and polymerase chain reaction analysis. J. Appl. Microbiol. 87:263-269.
Keim, P., A. Kalif, J. Schupp, K. Hill, S. E. Travis, K. Richmond, D. M. Adair, M. Hugh-Jones, C. R. Kuske, and P. Jackson. 1997. Molecular evolution and diversity in Bacillus anthracis as detected by amplified fragment length polymorphism markers. J. Bacteriol. 179:818-824.
Reference deleted.
Kennedy, S. M. 1994. When is a disease occupational Lancet 344:41-43.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17:1244-1255.
Logan, N. A., and P. C. Turnbull. 1999. Bacillus and recently derived genera, p. 357-369. In P. R. Murray (ed.), Manual of clinical microbiology. American Society for Microbiology, Washington, D.C.
Miller, J. M., J. G. Hair, M. Hebert, L. Hebert, F. J. Roberts, Jr., and R. S. Weyant. 1997. Fulminating bacteremia and pneumonia due to Bacillus cereus. J. Clin. Microbiol. 35:504-507.
Mock, M., and T. Mignot. 2003. Anthrax toxins and the host: a story of intimacy. Cell. Microbiol. 5:15-23.
Okinaka, R. T., K. Cloud, O. Hampton, A. R. Hoffmaster, K. K. Hill, P. Keim, T. M. Koehler, G. Lamke, S. Kumano. J. Mahillon, D. Manter, Y. Martinez, D. Ricke, R. Svensson, and P. J. Jackson. 1999. Sequence and organization of pXO1, the large Bacillus anthracis plasmid harboring the anthrax toxin genes. J. Bacteriol. 181:6509-6515.
Pannucci, J., R. T. Okinaka, R. Sabin, and C. R. Kuske. 2002. Bacillus anthracis pXO1 plasmid sequence conservation among closely related bacterial species. J. Bacteriol. 184:134-141.
Price, L. B., M. Hugh-Jones, P. J. Jackson, and P. Keim. 1999. Genetic diversity in the protective antigen gene of Bacillus anthracis. J. Bacteriol. 181:2358-2362.
Priest, F. G., M. Barker, L. W. Baillie, E. C. Holmes, and M. C. Maiden. 2004. Population structure and evolution of the Bacillus cereus group. J. Bacteriol. 186:7959-7970.
Ticknor, L. O., A.-B. Kolst, K. K. Hill, P. Keim, M. T. Laker, M. Tonks, and P. J. Jackson. 2001. Fluorescent amplified fragment length polymorphism analysis of Norwegian Bacillus cereus and Bacillus thuringiensis soil isolates. Appl. Environ. Microbiol. 67:4863-4873.(Alex R. Hoffmaster, Karen K. Hill, Jay E)