|Home | About | Journals | Submit | Contact Us | Français|
Spore germination of 17 Bacillus cereus food isolates and reference strains was evaluated using flow cytometry analysis in combination with fluorescent staining at a single-spore level. This approach allowed for rapid collection of germination data under more than 20 conditions, including heat activation of spores, germination in complex media (brain heart infusion [BHI] and tryptone soy broth [TSB]), and exposure to saturating concentrations of single amino acids and the combination of alanine and inosine. Whole-genome sequence comparison revealed a total of 11 clusters of operons encoding germinant receptors (GRs): GerK, GerI, and GerL were present in all strains, whereas GerR, GerS, GerG, GerQ, GerX, GerF, GerW, and GerZ (sub)clusters showed a more diverse presence/absence in different strains. The spores of tested strains displayed high diversity with regard to their sensitivity and responsiveness to selected germinants and heat activation. The two laboratory strains, B. cereus ATCC 14579 and ATCC 10987, and 11 food isolates showed a good germination response under a range of conditions, whereas four other strains (B. cereus B4085, B4086, B4116, and B4153) belonging to phylogenetic group IIIA showed a very weak germination response even in BHI and TSB media. Germination responses could not be linked to specific (combinations of) GRs, but it was noted that the four group IIIA strains contained pseudogenes or variants of subunit C in their gerL cluster. Additionally, two of those strains (B4086 and B4153) carried pseudogenes in the gerK and gerRI (sub)clusters that possibly affected the functionality of these GRs.
IMPORTANCE Germination of bacterial spores is a critical step before vegetative growth can resume. Food products may contain nutrient germinants that trigger germination and outgrowth of Bacillus species spores, possibly leading to food spoilage or foodborne illness. Prediction of spore germination behavior is, however, very challenging, especially for spores of natural isolates that tend to show more diverse germination responses than laboratory strains. The approach used has provided information on the genetic diversity in GRs and corresponding subclusters encoded by B. cereus strains, as well as their germination behavior and possible associations with GRs, and it provides a basis for further extension of knowledge on the role of GRs in B. cereus (group member) ecology and transmission to the host.
Dormant bacterial spores are able to monitor the environment for conditions that favor growth, indicated by the presence of specific nutrients, such as amino acids, which can trigger spore germination. Germination is a relatively fast biophysical process required to resume vegetative growth that can be initiated by activation of the germinant receptors (GRs) located in the inner membrane of the spore (1, 2). The majority of sporeformers, with Clostridium difficile being one of the exceptions (3, 4), contain at least one and usually several GRs that may differ in their specificities for different nutrients (5, 6). The model sporeformer Bacillus subtilis 168 carries three functional GRs, of which GerA responds specifically to l-alanine, whereas GerB and GerK cooperate to respond to the mixture of asparagine, glucose, fructose, and K+ (AGFK) (1). Notably, spores of the toxin-producing foodborne human pathogen Bacillus cereus ATCC 14579 germinate most efficiently in response to a mixture of alanine and inosine and are equipped with seven GRs (GerG, GerI, GerK, GerL, GerQ, GerR, and GerS) that show limited similarity to the GRs present in B. subtilis (7). B. cereus ATCC 14579 GerR plays a dominant role in germination, as its disruption affected germination in response to many amino acids, purine ribosides, and food products (7, 8). GerG appears to be specifically required for germination with glutamine (7). Additionally, it has been suggested that GRs can respond to more than one germinant and that cooperation of multiple GRs could enhance the germination response with specific individual germinants (6, 9,–12). Despite the attempts of Ross and Abel-Santos (13) to standardize the nomenclature used for the GRs, the annotation and naming of GRs are inconsistent across and within spore-forming species; this complicates comparative analysis and prediction of GR specificity.
GRs are usually composed of three subunits (A, B, and C), and genes encoding these subunits are typically arranged in tricistronic operons (14, 15), with some exceptions more frequently found in anaerobic strains of the genus Clostridium. The spore GR subunits A and B are integral membrane proteins composed of 5 to 8 and 10 to 12 predicted membrane-spanning domains, respectively (6, 12). Subunit B belongs to a subfamily of single-component membrane transporters and is speculated to be involved in germinant recognition (6). On the other hand, subunit C is membrane associated and conceivably bound to the A and B subunits (6). So far, the exact functions of the individual GR subunits and their interactions have not been elucidated (12). GRs are thought to cluster in complexes, so-called germinosomes, involving the GerD lipoprotein that influences GR-dependent germination rates (16, 17) and possibly the SpoVA channels (located in the spore inner membrane) that are involved in the release of small molecules, mainly dipicolinic acid (DPA) and monovalent cations, and uptake of water during germination (12). Recently, it was shown that heat resistance and the germination rate of B. subtilis spores could be attributed to the number of spoVA2mob copies on the genome, with a higher number of copies (up to three) correlating with increased heat resistance and reduced germination rate (18, 19).
The spore germination process follows well-described sequential steps (11, 12, 20). Spore swelling of germinating spores and full rehydration of the core require hydrolysis of the peptidoglycan cortex layer. In B. subtilis, two germination-specific cortex-lytic enzymes, namely, CwlJ and SleB, are responsible for cortex peptidoglycan degradation (1, 15). After full rehydration, metabolic activity is regained and spore outgrowth is initiated, followed by vegetative growth. Heat activation is commonly applied to enhance fast and homogeneous spore germination; however, the processes involved remain unknown. Spore germination can also be initiated by nonnutrient germinants, including chemical triggers (Ca-DPA and the cationic surfactant dodecylamine), mechanical triggers (high hydrostatic pressure), enzymatic treatment (lysozyme), or bacterial cell wall fragments (muropeptides). These nonnutrient triggers usually bypass GRs and either directly target the release of ions and Ca-DPA or activate cortex-lytic enzymes (1, 2).
Traditionally, germination is monitored by measurement of the optical density at 600 nm (OD600) of spore suspensions, with the percent decrease in OD600 correlating with germination efficacy assessed by phase-contrast microscopy and/or plate counts (7, 21). Additional methods include measurement of Ca-DPA content by Raman spectroscopy, sometimes combined with laser tweezers (22,–24), automated phase-contrast or differential interference contrast (DIC) microscopy, and time-lapse microscopy (for a review, see the paper by Wells-Bennik et al. ). Alternatively, spore-staining approaches can be applied in combination with high-throughput analysis of individual spores using flow cytometry (FCM) (25,–27). Germinated spores, but not dormant spores, can be stained by DNA fluorescent dyes, such as SYTO-9, since spores lose their structural integrity upon germination, which allows for access of the dye into the spore core and subsequent binding to DNA (25,–28).
In the current study, we applied FCM combined with SYTO-9 fluorescent dye staining of germinated spores to evaluate the germination responses of 15 whole-genome-sequenced B. cereus food isolates and of two well-studied sequenced laboratory strains, B. cereus ATCC 14579 and ATCC 10987 (29, 30), and correlated the germination responses with the presence/absence of Ger clusters in the corresponding genomes. The approach used provided information on the genetic diversity in GRs and corresponding subclusters encoded by B. cereus strains, as well as their germination behavior and possible associations with GRs.
The germination behavior of heat-activated spores from B. cereus food isolates and laboratory strains was evaluated at the single-spore level under 20 conditions representing saturating concentrations of single amino acids and mixtures of germinants that can be found in food matrices (Table 1; see also Table S2 in the supplemental material).
With the exception of strain B4153, heat-activated spores of all strains showed good germination in response to a mixture of alanine (Ala) and inosine (Ino) (i.e., AlaIno). Brain heart infusion (BHI) and tryptone soy broth (TSB) supported germination for most strains. However, a subset of strains belonging to phylogenetic group IIIA (B4085, B4116, B4153, and B4086) showed poor germination in those tested media (BHI and TSB) and in response to individual amino acids. Also, non-heat-activated spores of these strains showed a poor germination response to AlaIno, next to that of B4080 and B4087 (Table S2).
The data presented in Table 1 clearly show diversity in the sensitivity and responsiveness to single amino acids and complex media. We previously showed that the strains used in this study are representative of the diversity found among B. cereus strains with respect to carbohydrate utilization and capacity to occupy different environmental niches (soil, food products, and intestinal tract) (31). The poorly germinating group IIIA strains (B4085, B4086, B4116, and B4153) were previously shown to lack specific carbohydrate utilization clusters (those for starch, glycogen, and the aryl beta-glucosides salicin, arbutin, and esculin), suggesting a reduced capacity to utilize plant-associated carbohydrates for growth (31). Since B. cereus subgroup IIIA representatives are able to utilize host-associated carbohydrates and carry corresponding carbohydrate utilization gene clusters (31) and a subset of unique Ger (sub)clusters, additional studies using host-derived compounds as germinants may provide further insights into the germination efficacy of their spores.
Notably, all B. cereus strains tested carry spoVA1 and spoVA2 operons (18). Interestingly, only two members of the poorly germinating group IIIA strains, namely, B4085 and B4116, carry a spoVA2mob operon, the occurrence of which was recently shown to correlate with slow germination and increased heat resistance of B. subtilis spores (18, 19).
Without heat activation, exposure to Ala resulted in very efficient germination of spores of strains B4078 and B4088 and less-efficient germination of spores of strains B4155 and B4158. A study by Broussolle et al. (32) showed significant germination (measured by OD600 drop) of spores of a laboratory strain, i.e., B. cereus ATCC 14579, upon exposure to 1 mM Ala, while wild B. cereus isolates typically required concentrations above 1 mM and, for some, even 200 mM did not result in a maximal OD600 drop (32). When the maximum germination rate was reached, a further increase in germinant concentration did not improve germination. Both the maximum germination rate and the minimal concentration to reach this were shown to be strain and germinant dependent (32). Notably, the l-alanine-induced germination response may be affected by the presence of alanine racemase (encoded by the alr gene) that is able to convert germination-stimulating l-alanine into germination-inhibiting d-alanine that competitively binds to l-alanine GRs (33, 34). Differences in the l-alanine response (and in the combination with inosine) between strains may be due to differences in Alr activities in the spore cortex.
Cysteine (Cys) has been reported as a potent germinant in B. cereus strains (7, 35), and significant Cys-induced germination was indeed observed for heat-activated spores of the majority of the strains tested. Despite the heat activation step, the spores of seven strains (B4085, B4086, B4116, B4117, B4118, B4153, and B4158) stayed insensitive or responded only weakly to one of six individual amino acids: Ala, Cys, glutamine (Glu), glycine (Gly), valine (Val), and isoleucine (Ile). All other strains tested showed enhanced spore germination in response to two or more amino acids tested. The applied 20 or 100 mM concentration of individual amino acids (Ala, Val, Glu, Iso, Cys, and Gly) and Ino used in this study is unlikely to be encountered in the environment; however, it allows fast germination responses. The poorest germination was observed in response to Val and Iso, suggesting that despite an approximately 10 times lower concentration of free amino acids in BHI and TSB (Table S1), their broader composition (Table S1) most likely resulted in a cumulative germination response.
GRs are responsible for recognition and binding of the nutrient germinants; however, the exact mechanism in which GR subunits interact and proceed upon germinant binding to the downstream germination pathway remains to be elucidated. Recognizing the fact that annotation of GRs is not a trivial issue and is not consistent across species and within species (6, 13), we present an overview of known GRs and their putative germinant specificity for nine sequenced and well-studied B. cereus group reference strains, shown in Table 2.
To compare and assign GRs present in the food isolates to the GRs present in well-studied strains, phylogenetic trees were composed for each GR's A, B, and C subunits of the reference strains and B. cereus food isolates (Fig. S2 to S4). Based on these trees, we have identified consistent clustering of corresponding subunits A, B, and C of given types of GRs, illustrating their coevolution, as has been observed previously (2, 13). The clusters were named, when possible, after the GR name/letter of the B. cereus type strain, ATCC 14579 (Table 2). The GRs of 29 strains could be allocated to 11 main clusters and their subclusters (Table 3).
B. cereus strains typically harbor a relatively high number of GRs compared with other spore-forming species. Until now, B. cereus strains were believed to encompass a core group of five GRs, namely, GerR, GerL, GerK, GerS, and GerI, plus a selection of five additional GRs (6, 36). Indeed, an analysis of 29 genomes revealed that all strains contain ger genes belonging to clusters K, I, L, R, and S, albeit that clusters R and S can be differentiated into subclusters RI and RII and subclusters SI, SII, and SIII, respectively, and that some strains contained one or more pseudogenes encoding putative individual Ger subunits, conceivably resulting in altered functionality of the respective GRs (Table 3). The majority of pseudogenes and/or variants in gerK and gerL are found in strains belonging to phylogenetic group IIIA, which contains mainly poorly germinating strains. In fact, within both clusters K and L, pseudogenes and/or variants were identified in the genomes of strains B4086 and B4153, which are two poorly germinating strains (Tables S2 and S3). At the same time, two other strains, B4085 and B4116, also belonging to the phylogenetic group IIIA and producing poorly germinating spores (Table S2), carried a variant of the cluster L (subunit C) in combination with GerXIII. The combination of those two GRs that are possibly nonfunctional variants could have a negative effect on the germination ability. Both variants of GerL and GerXIII are also present in B. cereus AH187; spores of this strain were not included in the current germination assays but were previously shown to germinate poorly in response to amino acids or food (36).
Interestingly, GerR, which is a dominant GR in B. cereus ATCC 14579 that responds to most amino acids and to model foods (7, 8), belongs to subcluster GerRI. Differentiation between subclusters GerRI and GerRII might also reflect functional differences between the two subclusters. In fact, comparative analysis of GerY of B. anthracis with unknown function (10) located this Ger receptor in cluster RII, although its subunit C is encoded by a pseudogene and might be a nonfunctional version of the GR variant (Table 3). Moreover, genes encoding subunits C of GerRI of strains B4086 and B4153, which produce poorly germinating spores, are pseudogenes. Complementation experiments could provide insights into the role of gerR in spore germination of other strains. Despite the differences in GR presence/absence, the organization of the GR operons was conserved within the clusters with dominating ABC order, while three clusters (R, F, and Z) followed an ACB organization, clusters XI and XII revealed a BAC organization, and cluster XIII had a unique ACC organization.
In six of the B. cereus food isolates and in B. cereus AH187, a tricistronic ger operon composed of one truncated subunit A and two subunits C was found, creating cluster XIII. In B. subtilis, the incorporation of subunit C of GerA into the membrane was shown to depend on subunits A (37) and B (38). This suggests either that a different mechanism to form stable receptor assembly (e.g., interaction with subunits of other GRs in the germinosome) is in place in B. cereus or that the GR is not functional. Furthermore, subunit B was previously suggested to play a role in nutrient recognition and specificity of the germination response in Bacillus megaterium (39, 40). However, the assembly and function of GRs within cluster XIII remain to be elucidated.
Spores of strain B4117 (B. mycoides) germinated well, even without heat activation, in response to AlaIno and Ino alone. Strain B4117 encodes two unique (among tested strains) GRs representing clusters QII and SIII. GerQ is known to be involved in Ino-induced germination in B. cereus (7), and it could be speculated that the GerQII version present in B4117 is more responsive than more commonly encountered GerQI. Besides B4117, only Bacillus weihenstephanensis KBAB4 carries a GR belonging to cluster SIII. In fact, both strains also carry GRs from cluster SII, which has been previously referred to as GerS2 (36), while most of the GRs within cluster S belonged to cluster SI. Notably, van der Voort et al. (36) previously showed that germination of B. weihenstephanensis KBAB4 spores with combinations of selected amino acids and inosine was far more efficient than that of spores from tested B. cereus ATCC 14579 and ATCC 10987 strains.
Our study identified additional putative GRs, a group of GRs creating cluster W, including a presumptive gerT gene of B. cereus AH187 (36), and a putative GR found only in B4079 comprising the one-item cluster Z. Those putative GRs share the tricistronic architecture and homology with known GRs; however, their functionality would require the testing of directed deletion mutants.
Interestingly, the presence/absence of GRs seems to be related to the phylogenetic clustering we reported previously (31). All strains encoding GRs belonging to cluster GII represented phylogenetic group II (Table 3). This group is also characterized by high prevalence (75%) of GRs from clusters SII and QI. A phylogenetic tree based on core genes of tested strains (reported by Warda et al. ) and phenotypic differences among the strains of group III suggest that two subgroups can be distinguished as IIIA and IIIB. In fact, the majority of strains belonging to group IIIA (and only those) encode GRs from cluster FII. The high prevalence of FII might be compensating for the lack of GRs from cluster Q and high numbers of pseudogenes within clusters K and L. Based on the encoded GRs, experimentally tested strains within group IIIA seem more comparable to strains within group IV than IIIB (Table S3), possibly suggesting similar niche requirements. This may also be reflected by the presence of carbohydrate utilization genes that show similarities for members within those groups. All the strains within group IIIB encode GRs from clusters K, I, L, and SI. GRs from all the clusters can be found only in strains belonging to group IV, indicating the diversity of this group, with half of the strains encoding GRs from cluster FI (not found in other strains tested). Finally, strains belonging to group VI often encode alternative GRs; in fact, GRs from clusters SIII and QII are present only in this group, while RII and XI are found only in one other group. Moreover, strains representing group VI do not encode GRs belonging to cluster G or cluster F (Table 3).
Our study further supports a high degree of diversity in GRs and nutrient-induced germination in spores of different strains of B. cereus, generating leads for further studies. However, despite the different germination conditions tested, we could not directly link the presence (or absence) of given GRs to the germination responses such that germination behavior could be predicted. This may be due to the requirement for different types or concentrations of certain germinants or combinations of germinants. A number of factors that can affect the GR-dependent germination have been previously discussed (12, 20, 41), including accessibility and the number of GRs or the downstream germination mechanisms. Moreover, genome-based studies may be affected by the quality of draft genomes, e.g., contig length, the presence of GR subunits (i.e., C subunits) on short contigs, and/or location of GRs or subunits (i.e., those in cluster XIII) at the contig's edge. Nevertheless, despite those limitations, the approach presented in the current study allows for the inclusion of rapidly increasing numbers of draft genomes.
Previous studies could not link the diversity in spore germination responses with the presence and/or similarity of the GRs, most likely due to the large number of factors affecting GR-dependent germination (42). Similarly, others (A. O. Krawczyk, A. de Jong, J. Omony, S. Holsappel, M. H. Wells-Bennik, R. T. Eijlander, and O. P. Kuipers, unpublished data) found no correlation between poor germination of B. subtilis spores with Ala and the sequences of their GerA subunits. However, the authors did show that “modest germination” with AGFK correlated with the presence of several common amino acid substitutions in the subunits of GerB and GerK (A. O. Krawczyk, unpublished data). Similarly, diverse germination responses to Ala of spores of 46 strains of Bacillus licheniformis, which is a close relative of B. subtilis, could not be linked with the clusters formed by their gerAB-gerAC sequences (44). Nevertheless, complementation of a gerAA disruption mutant with gerA operons of slow- and fast-germinating B. licheniformis revealed that differences in gerA family operons are partly responsible for the differences in germination efficiency in response to Ala (44). The greater diversity and complexity of GRs and germination responses among B. cereus strains and B. cereus group strains compared with, for example, B. subtilis, create a challenge for future studies.
In conclusion, our comparative genotyping and phenotyping approach showed that four B. cereus strains, B4085, B4086, B4116, and B4153, had poor germination in response to many different germinants and that these strains cluster in phylogenetic group IIIA. These IIIA group strains contain either pseudogenes or variants of genes encoding subunit C in their GerL cluster combined with pseudogenes in cluster K and subcluster RI (B4086 and B4153) or combined with the presence of a spoVA2mob transposon (B4085 and B4116), which leads to high-level heat resistance with concomitant reduced germination responses in B. subtilis spores (18, 19). The approach used has provided information on the genetic diversity in GRs and corresponding subclusters encoded by B. cereus strains, as well as their germination behavior and possible associations with GRs, and it provides a basis for further extension of knowledge on the role of GRs in B. cereus (group member) ecology and transmission to the host.
Two laboratory strains, B. cereus ATCC 14579 and B. cereus ATCC 10987, were obtained from the American Type Culture Collection (ATCC) and the culture collection of the Laboratory of Food Microbiology, respectively. In addition, we used 15 sequenced B. cereus strains (45, 46) isolated from food products and the food-processing environment (Table 4). Strain B4117 (GenBank accession no. LJKG00000000.1) was initially included as a B. cereus strain but was recently reclassified by the NCBI as Bacillus mycoides based on criteria of average nucleotide identity (ANI) typing (47). Strains were cultured in Bacto brain heart infusion (BHI) broth (Becton Dickinson, France) at 30°C with aeration at 200 rpm.
Spores were prepared on a nutrient-rich chemically defined sporulation medium, designated MSM medium, which was previously described (48). Ten milliliters of sporulation medium was inoculated with 100 μl of an overnight-grown preculture in 100-ml flasks and incubated at 30°C with aeration at 200 rpm. When the mid-exponential-growth phase was reached (corresponding to an optical density at 600 nm [OD600] of ~0.5 [Novaspec II; Pharmacia Biotech, United Kingdom]), 200 μl of culture was spread on MSM plates (solidified with 1.5% agarose) and incubated at 30°C within plastic bags to prevent drying. After 7 days, 100 mM chilled phosphate buffer (pH 7.4) containing 0.1% Tween 80 was added to the plates, and spores were scraped off the surface and harvested by a 15-min centrifugation at 3,438 × g and 4°C (5804R; Eppendorf, Germany). Spores were washed in decreasing concentrations of Tween 80 and prepared for use as described previously (49). A single spore crop per strain was used for all the experiments.
Fifty microliters of a spore suspension containing approximately 108 to 109 spores/ml in 100 mM phosphate buffer (pH 7.4) with 0.01% Tween 80 (here referred to as suspension buffer) was exposed to germinants by mixing with 50 μl of concentrated stock solution of either individual or mixed germinants. The final germinant concentrations used were 20 mM for valine (Val), isoleucine (Ile), glutamine (Glu), glycine (Gly), cysteine (Cys), and inosine (Ino) and 100 mM for l-alanine (Ala), and final concentrations were 10 mM Ala and 2 mM Ino for the mixture of Ala and Ino (AlaIno). Amino acid stock solutions were prepared in distilled water and filter sterilized. Control experiments were performed using the complex media BHI and tryptone soy broth (TSB; Becton Dickinson, France) and 25 mM HEPES buffer (pH 7.4). Spores were exposed to germinants for 30 min at room temperature, followed by a centrifugation step for 1 min at 13,000 × g to remove the germinant. The resulting spore pellet was resuspended in 100 μl of water containing 1 μM SYTO-9 (Invitrogen, The Netherlands) and incubated in the dark for 10 min at room temperature to stain permeabilized spores. Next, the unbound dye was removed by centrifugation, and the resulting spores were resuspended in HEPES buffer to obtain approximately 1,000 events per second after loading to a FACSAria III flow cytometer (BD, USA). Each experimental run included control spores that were not exposed to germinants. For each sample, 10,000 events were evaluated. For heat activation, spores were heated for 10 min at 80°C in a thermal cycler (Verity; Applied Biosystems, Singapore).
Flow cytometer data were analyzed using the data analysis software FlowJo (version X.0.7; LCC, USA); Fig. S1 in the supplemental material contains the flow chart for data analysis. An approach using SYTO-9 staining and FCM for germination evaluation has been previously validated by Cronin and Wilkinson (25). Scatterplots of forward scatter area (FSC-A) versus a green fluorescence intensity (fluorescein isothiocyanate area [FITC-A]) were used to exclude atypical data sets indicating the presence of spore clumps or strain-specific fluorescence, i.e., high autofluorescence due to the presence of DNA on the spore surface. Events with fluorescence intensity lower than the spore autofluorescence (control sample, FITC-A of <101) and higher than FITC-A of >105 were excluded from the analysis.
For each strain individually, the maximum fluorescence of dormant spores was determined based on the cutoff between dormant (unstained) and germinating (stained) spores in control samples (not exposed to germinant) and samples exposed to germinant. The control samples showed dormancy rates between 81.8 and 99.4%, while microscopic observations prior to the experiment showed a minimum of 95% dormancy. Spores with higher fluorescence than the maximum value for dormant spores were considered stained/germinated. Next, a percentage of germinated spores in the population was calculated as a percentage of stained spores among all events considered (unstained and stained).
Per strain, the average percentage of germinated spores in the control samples and its standard deviations were calculated based on three independent measurements, with and without a heat activation step. Three categories of germination performance were defined: spores with <15% germination (poor germination), spores with 15 to 50% germination (intermediate germination), and spores with >50% germination (good germination).
The free amino acid content in BHI and TSB (same lot as used in germination experiments) was determined using liquid chromatography as described previously (50).
To investigate whether the presence of GR operons correlated with experimentally tested germinant-induced spore germination, genomes were mined as described previously (31). The analysis included 20 newly sequenced B. cereus food isolates (45, 46), with nine B. cereus group strains with publically available genome sequences and with experimentally determined germination responses (6, 36, 48). The strains used and their isolation sources are listed in Table 4. To enhance genome comparisons, additional genomes previously used for B. cereus genome comparison were used (31). To improve the comparative analysis, all genomes were (re)annotated using RAST (51), and for the resulting annotated genomes, orthologous groups (OGs; i.e., genes that are descended from the same gene in the last common ancestor of the strains studied, putatively sharing similar functionality) were defined using Ortho-MCL (52). The OGs containing sequences of known germination receptor subunits were extracted from the data set, and additional germination receptor subunits were identified manually by keyword searches and inspection of genome context. Multiple-sequence alignment (MSA) files were made with MUSCLE (53), aligning the protein sequences within specific OGs to facilitate the identification of pseudogenes (encoding incomplete proteins) and manual correction of inaccurate autoannotation using Artemis (54) and Jalview (55). Next, phylogenetic trees of individual GR subunit A, B, or C based on the aligned amino acid sequences of the 29 B. cereus group strains and B. subtilis 168 were constructed using Clustal X (56) and visualized using LOFT (57) (Fig. S2 to S4). When subunits A, B, and C of a given type of GRs clustered consistently in one of the 11 clusters (and their subclusters), the candidate GR was assigned the GR name/letter of the well-studied strain, preferentially B. cereus ATCC 14579 (Table 2). Newly identified and known GR operons are summarized in Table 3 using the updated GR naming system (Table 2).
Supplemental material for this article may be found at https://doi.org/10.1128/AEM.02490-16.