|Home | About | Journals | Submit | Contact Us | Français|
Eleven Bacillus isolates from the surface and subsurface waters of the Gulf of Mexico were examined for their capacity to sporulate and harbor prophages. Occurrence of sporulation in each isolate was assessed through decoyinine induction, and putative lysogens were identified by prophage induction by mitomycin C treatment. No obvious correlation between ability to sporulate and prophage induction was found. Four strains that contained inducible virus-like particles (VLPs) were shown to sporulate. Four strains did not produce spores upon induction by decoyinine but contained inducible VLPs. Two of the strains did not produce virus-like particles or sporulate significantly upon induction. Isolate B14905 had a high level of virus-like particle production and a high occurrence of sporulation and was further examined by genomic sequencing in an attempt to shed light on the relationship between sporulation and lysogeny. In silico analysis of the B14905 genome revealed four prophage-like regions, one of which was independently sequenced from a mitomycin C-induced lysate. Based on PCR and transmission electron microscopy (TEM) analysis of an induced phage lysate, one is a noninducible phage remnant, one may be a defective phage-like bacteriocin, and two were inducible prophages. One of the inducible phages contained four putative transcriptional regulators, one of which was a SinR-like regulator that may be involved in the regulation of host sporulation. Isolates that both possess the capacity to sporulate and contain temperate phage may be well adapted for survival in the oligotrophic ocean.
Lysogeny and sporulation are strategies for phage and host survival, respectively, under adverse conditions. During both processes, the genome of the phage or bacterium is replicated into a form, prophage or endospore, that increases its survival (14). The initiation of both lysogeny and sporulation involves the repression and activation of promoters that are regulated by feedback from their gene products. In coliphage λ, cI binds to the lytic promoter PR during lysogeny while cro binds to PR during a virulent infection (39). SinR binds to the vegetative promoter PV in the Sin operon under normal cell conditions, and SpoA activates the sporulation promoter (27). The tertiary structures of the DNA-binding domain of SinR from Bacillus subtilis and CI and Cro from the Escherichia coli phage 434 are nearly identical (25). Recent work has also demonstrated that CIII and sporulation control protein SpoVM are both inhibited by the FtsH protease (21). These structural and functional similarities indicate a possible evolutionary relationship between prophage induction and sporulation.
Previous work has indicated a possible link between phages and sporulation. Meijer et al. (28) found that lytic development of the virulent Bacillus phage 29 was repressed in sporulating cells through inhibition of transcription of early phage genes by SpoA. A phage-encoded sigma factor in the B. anthracis virulent phage Fah was negatively controlled by a sporulation anti-sigma factor (29). Lytic development was also found to be suppressed in temperate phages during sporulation, with no virus production occurring during UV induction (35). This phenomenon could be a survival mechanism that allows the phage to be protected in the spore under conditions that are unfavorable to infection, such as low host abundance under nutrient-depleted conditions. Survival would be increased if the prophage contained genes that allow the host to sporulate at a higher frequency than noninfected bacteria.
Phages that can perform this function are known as spore-converting phages, and they have been found in Bacillus and Clostridium species (7, 47). Bacillus phages PMB12 and SP10 are believed to induce sporulation by activating sporulation initiation signals (46). Unfortunately, most of these studies of Bacillus phages predate current sequencing technologies. However, the genome of Clostridium perfringens temperate phage Φ3226 contained a sporulation-associated sigma factor homolog and a sporulation-dependent transcriptional regulator that was believed to augment host sporulation (53). Although spore-converting phages have been characterized for terrestrial Bacillus species, the phages infecting marine Bacillus strains have not been studied.
In this study, a collection of 11 Bacillus isolates from the Gulf of Mexico were investigated for spore and phage production through traditional induction experiments. Genomic and molecular techniques were used to study the prophages of one of the isolates, B14905, in depth. These studies provide a glimpse into host-prophage relationships of a marine Bacillus strain and serve as a basis for functional studies to investigate the influence of prophages on host sporulation.
Eleven pigmented marine Bacillus isolates were collected from different sites and depths during a 1992 research cruise to the Gulf of Mexico, as outlined in Table Table1.1. The collection methods and the 16S rRNA gene sequence for each isolate were previously described by Siefert et al. (45). These strains were maintained in pure culture by monthly plating onto artificial seawater nutrient (ASWJP+PY [38a]) plates and Trypticase soy agar (TSA) plates (Becton Dickinson, Cockeysville, MD).
Before the induction experiments, a smear from a colony of each isolate on a month-old plate was observed for the presence or absence of spores using phase-contrast microscopy. This indicated the capacity of each strain to sporulate.
Sporulation was initially induced in the isolates by use of a protocol modified from Mitani et al. (30). Decoyinine is an adenine-ketose antibiotic that inhibits GMP synthetase. Starvation of GMP can initiate sporulation in the presence of ample nutrients, serving as a quick and effective means to study sporulation (30, 49). For each Bacillus strain, 1 ml of overnight culture was diluted into 9 ml of fresh ASWJP+PY and incubated at 28°C with shaking. When the optical density at 600 nm reached 0.4, 2-ml aliquots were placed in separate 15-ml tubes, one for a control and one for treatment. Sterile filtered (0.22-μm filter size) decoyinine (0.5 mg ml−1) (Sigma, St. Louis, MO) was added to the treatment tube. After a 24-h incubation, 0.02 μm filtered formalin (1.0%) was added to the control and treatment tubes to stop bacterial growth. Wet mounts of these cultures were prepared in triplicate and immediately viewed. Vegetative cells and spores have different refractive indices that allow differentiation of the two bodies under phase-contrast microscopy. The different cell types were counted on 10 field grids per slide to determine the number of spores (Cs) and vegetative cells (Cv) by use of the equation Cg/Vg = Cs,v, where Cg is the cell (spore or vegetative-cell) direct count per grid and Vg is the volume (in ml) per grid. Vg was calculated using the equation (Vs/Acs) × Ag = Vg, where Vs is the volume of sample, Acs is the area of coverslip, and Ag is the area of the grid.
The percent population of cells that sporulated (% spore) was calculated by use of the equation (Cs/Cv) × 100 = % spore.
Isolates that did not have significant sporulation following decoyinine treatment were subjected to heat and ethanol treatments to induce sporulation. For the heat treatments, 1 ml of log-phase cultures (optical density at 600 nm, 0.4) was incubated at 80°C for 10 min, and 1 ml of the culture left at room temperature served as a control. Serial dilutions from the treatment and control cultures were plated in duplicate on ASWJP+PY plates. Ethanol treatment was performed essentially as described by Koransky et al. (22). One ml of log-phase cultures (optical density at 600 nm, 0.4) was diluted 1:1 with 0.22 μm filtered absolute ethanol for 1 h, while cultures diluted 1:1 with 0.22 μm deionized (DI) water served as controls. Serial dilutions from the treatment and control cultures were plated in duplicate on ASWJP+PY plates. The plates were incubated for 24 to 72 h until colonies appeared.
The presence of prophage in these Bacillus isolates was determined by mitomycin C treatment. Five milliliters of overnight culture was diluted into 45 ml of fresh ASWJP+PY medium in separate flasks, one for a control and one for a treatment. When the optical density at 600 nm reached 0.4, mitomycin C (1 μg ml−1) was added to the treatment flask and returned to the shaker. After 24 h, the absorbance measurements were taken and 5 ml was removed from each flask for phage enumeration. Samples were stained with SYBR gold (Molecular Probes, Eugene, OR) as described by Chen et al. (12), with a staining time of 12 min. Virus-like particles (VLPs) were enumerated using blue light excitation with an Olympus BH-2 epifluorescence microscope. At least 200 phage particles in 10 fields per slide were counted. The difference in prophage production between treated and control cultures was calculated. Statistical analysis was done using the pooled Student t test for equal variances (α = 0.05).
Two different methods were used to address the cross-infectivity of mitomycin C-induced lysates from the 11 Bacillus isolates within the collection. Phage lysate was produced for each of the Bacillus strains that had significant prophage production as determined by the initial screening described above.
Host sensitivity to the phage lysates was determined using a cross-streaking test (20). Approximately 30 μl of each phage lysate was streaked on an ASWJP+PY plate and allowed to dry before a loopful of an overnight culture of the test Bacillus isolate was cross-streaked against the lysate. Plates were incubated at 28°C for 24 to 74 h and then examined for any lysis of bacterial growth.
A spot lysis assay was performed in duplicate by spotting approximately 2 × 107 VLPs of each phage lysate on ASWJP+PY soft agar overlays containing 1 ml of overnight Bacillus culture (19). Plates were incubated at 28°C for 24 to 74 h and then examined for any lysis of bacterial growth.
Phage production in B14905 was monitored over a 24-h time series following induction with mitomycin C. Prophage induction was carried out as described above except that absorbance readings were taken every 2 h until 8 h and at 24 h after induction, and 5-ml subsamples were removed from each flask to enumerate bacteria and viruses. Samples were stained with SYBR gold and VLPs were enumerated as described above. The direct-counting method was also used for bacterial direct counts (BDC). Statistical analysis was done using the pooled Student t test for equal variances (α = 0.05).
The genome of Bacillus sp. NRRL B14905 was sequenced using whole-genome shotgun sequencing by the J. Craig Venter Institute as part of the Gordon and Betty Moore Foundation Marine Microbial Genome Sequencing Project. The genome was partially assembled and annotated by using the subsystems approach in the SEED databank (36).
B14905 phage particles were isolated by use of a large-scale mitomycin C induction procedure, and phage DNA was extracted as described by Mobberley et al. (31). The B14905 phage DNA extract was cloned using linker-amplified shotgun libraries (LASL) as previously described (38). Clones from these libraries were sequenced using an ABI 3100 capillary sequencer (Applied Biosystems, Foster City, CA). The software program Sequencher 3.0 (Genecodes, Ann Arbor, MI) was used to assemble contigs. Gaps between the contigs were resolved using multiplex PCR and by primer walking using phage DNA as the template.
The gene annotations of B14905 were searched in the SEED database using the term “phage,” which generated a list of all open reading frames (ORFs) including the term. When these ORFs were clustered together, the region of the contig was investigated for prophage-related genes, including integrases, repressors/antirepressors, packaging protein genes, and lysis-associated genes. Potential prophage-encoded regions were exported into KODON for further analysis. ORFs were analyzed more in depth by BLASTP analysis of the nonredundant database (2).
Phage lysate samples were separated on a one-dimensional 12% sodium dodecyl sulfate-polyacrylamide gel run at 4 mA for 16.5 h by use of the Laemmli method (24). Gels were fixed in methanol-acetic acid-water (50:10:40), followed by overnight staining with SYPRO ruby (Bio-Rad, Hercules, CA). Protein-containing bands were numbered and excised from the gel with a scalpel. The excised samples were sent to the Proteomics Core Facility at the Moffitt Cancer Research Center (Tampa, FL) for analysis by matrix-assisted laser desorption ionization-time-of-flight-time-of-flight (MALDI-TOF-TOF) mass spectroscopy as described by Paul et al. (31, 38). The tandem mass analysis produces a collision-induced dissociation (CID) spectrum that can be used for protein identification based on the peptide sequence (52).
Phage genomic DNA was extracted from a large-scale mitomycin C lysate as described previously (31). B14905 chromosomal DNA was extracted using a Wizard genomic DNA purification kit (Promega, Madison, WI) according to the manufacturer's protocol. The quantity and quality of the DNA preparations were assessed by agarose gel electrophoresis and by use of a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE).
Unique primer sets were created for each potential prophage genome and for the host, as listed in Table Table2.2. In order to account for the differing amounts of phage DNA present in the preparations, 10 ng of host chromosomal DNA and 1 ng of phage DNA were used in the reactions. GoTaq green mastermix (Promega, Madison, WI) was used as recommended, and the primer sets were added to a final concentration of 0.5 μM per primer (Table (Table2).2). All PCRs were performed with a My Cycler thermal cycler (Bio-Rad, Hercules, CA). For phage primer sets 1, 2, and 4 and the host primer set, the following PCR program was used: initial denaturation for 2 min at 95°C; 30 cycles of 95°C for 1 min, 58°C for 1 min, and 72°C for 2 min; and a final extension at 72°C for 10 min. The PCR program used for phage primer set 3 was as follows: initial annealing for 2 min at 95°C; 30 cycles of 95°C for 1 min, 56°C for 1 min, and 72°C for 2 min; and a final extension at 72°C for 10 min. The amplicons were loaded on a 1% agarose gel with 0.5 μg μl−1 of ethidium bromide and run at 89 V for 45 min. The gel was imaged using an AlphaImager 2200 imaging system (Alpha Innotech, San Leandro, CA).
Transmission electron microscopy (TEM) was used to determine the presence and morphologies of multiple phages from a polyethylene glycol (PEG)-purified phage lysate. Twelve ml of a mitomycin C-induced culture was ultracentrifuged at 38,000 × g for 2 h at 4°C. The phage pellet was resuspended in 100 μl of 0.02 μm filtered deionized water, and the VLPs were enumerated with epifluorescence as described above. A Formvar carbon-coated grid (Electron Microscopy Sciences, Hatfield, PA) was floated on a large drop of lysate for 2 h. The grid was air dried and then negatively stained with 2% uranyl acetate for 1 min. The phage particles were visualized with a Hitachi 7100 transmission electron microscope.
The unfinished genome of Bacillus sp. NRRL B14905 was submitted to GenBank under accession number AAXV00000000.
Month-old colonies of all Bacillus strains were observed to contain spores whose morphology was similar to that described by Siefert et al. (45). Sporulation was induced in several of the Bacillus isolates after a 24-h treatment with decoyinine (Fig. (Fig.1).1). Isolates B14905, B14906, B14910, B14911, B14850, and B14851 had a significantly higher level of spore production with decoyinine than did the controls. Spore production for decoyinine-treated isolates B14904, B14907, B14908, B14909, and B14912 was not significantly different from that for the controls. Within this group, B14904, B14908, B14910, and B14912 had frequencies of sporulation that were less than 1%, while B14907 had a high frequency of spontaneous sporulation. The difference between spontaneous spore production for the controls and decoyinine-induced spore production for each isolate is displayed in Table Table33.
The five Bacillus isolates that did not sporulate with decoyinine treatment, as well as B14850, which had high levels of spontaneous sporulation, were subjected to both heat and ethanol treatments to verify that they produced spore spontaneously. There was no significant sporulation in B14904, B14908, B14909, and B14912 with either of the treatments (data not shown). B14907 had a 74% frequency of sporulation with heat treatment, and ethanol treatment resulted in a 7% frequency of sporulation in B14850.
Prophage induction was investigated for 11 Bacillus isolates after 24 h of incubation with mitomycin C (Fig. (Fig.2).2). Virus-like particles produced by strains B14851, B14904, and B14911 were not significantly different from those produced by the controls (P > 0.05). Isolate B14912 was significantly different from the controls, but there was a high level of spontaneous induction in the control culture. Isolates B14908 and B14910 had significant prophage induction, but it was less than an order of magnitude greater than that for the controls. Phage production in isolates B14905, B14906, B14907, B14909, and B14850 was more than a magnitude greater than that in the controls. Table Table33 shows the difference between levels of spontaneous phage production in the control and mitomycin C-induced cultures.
The percent difference in prophage production between the mitomycin C-treated and control cultures and the percent difference in spore production between the decoyinine-treated and control cultures for each isolate listed in Table Table33 were plotted against each other, as shown in Fig. Fig.3.3. This shows that while there is no correlation between sporulation and prophage induction, the isolates appear to cluster into 3 groups: those with a high level of prophage production but a low incidence of sporulation, those with a low level of prophage production and a high incidence of sporulation, and those that both sporulated and produced elevated levels of prophage.
Neither cross-streaking nor spot lysis produced visible lysis in isolates from our Bacillus collection.
A 24-h time series prophage induction with mitomycin C that was performed on isolate B14905 indicated that isolate B14905 contained inducible phages (data not shown). After 24 h, prophage induction in the mitomycin C-treated culture (3.11 × 1010 VLPs ml−1) was significantly higher than that in the control culture (2.51 × 108 VLPs ml−1). As expected, bacterial growth monitored by absorbance decreased in the mitomycin C-treated culture, indicative of host lysis.
The draft B14905 genome is 4.4 Mbp in length, with a G+C content of 37%. The genome was assembled into 99 contigs and contained 4,626 protein coding regions and 126 RNA genes. Fifty-five percent of the coding regions could be assigned a putative functional assignment.
Sequencing of the mitomycin C-induced lysate resulted in a single double-stranded phage genome, ΦB05-1, which is 18,118 bp in length and has a G+C content of 33%. This assembly was verified by comparison to the prophage sequence in the B14905 host genome contained in contig 1000022. Sixteen of the ORFs (61%) shared significant similarity (e < 10−4) at the protein level with other sequences in GenBank. Twelve of the ORFs were similar to genes found in other Bacillus species and Bacillus phages. Significant similarities to other sequences in GenBank and putative functional assignments are listed in Table Table44 and Fig. Fig.4.4. ΦB05-1 is mitomycin C inducible, based on PCR amplification from a purified host-free lysate (Fig. (Fig.5,5, lane 7).
ΦB05-1 contained several genes that were similar to those found in temperate phages. ORF 1 was similar to an integrase and ORF 6 was similar to the replication protein from the Bacillus cereus phage BC6A52. ORF 8 shared similarity with ORF 13 of B. subtilis phage Φ-105 and ORF 40 of Clostridium phage 39-O. ORF 12 of ΦB05-1 was similar to the antirepressor protein of Streptococcus phage 10750.4. No phage structural or lytic genes were identified through similarity searching.
The ΦB05-1 genome contained four putative transcriptional regulators, ORFs 2, 3, 17, and 20, based on the presence of helix-turn-helix DNA binding domains. ORF 2 is similar to the DNA binding domains of post-exponential-phase transcriptional regulators from Bacillus halodurans C-125 and Bacillus sp. NRRL B-14911. Post-exponential-phase transcriptional regulators are involved in expression of genes that allow Bacillus to adapt to suboptimal conditions in the transition from active to stationary phase (48). The ORF also shared a weak similarity to SinR-like proteins in Lysinibacillus sphaericus C3-41 and Bacillus cereus ATCC B4264. SinR is a transition-state regulator that represses stage II sporulation genes and spo0A expression during vegetative growth. ORF 3 is a weak hit to a transcriptional regulator found in the genomes of both Lysinibacillus sphaericus C3-41 and Staphylococcus aureus. It may be that repression of sporulation by phage repressors ensures better prophage survival and that it is more difficult for prophage induction to occur in a spore.
ORF 17 was a very strong hit to a MerR-type transcriptional regulator from B. cereus E33L. Transcriptional regulators in this family have been found in bacteria and phages and are known to respond to a variety of stressful stimuli, including metals and antibiotics (8). ORF 18 is part of a group (ORF 17 to 20) that contains 2 ORFs similar to choloylglycine hydrolases from B. cereus ATCC 14579, B. cereus E33L, and Clostridium botulinum. Choloylglycine hydrolases are bacterial products, usually associated with intestinal microflora, that catalyze the degradation of bile salts and result in the production of free amino acids (5). ORF 20 is similar to a transcriptional regulator found in Streptococcus pyogenes (3e−8, 32% identity). It also shares similarity to LlaI, part of a phage-encoded restriction modification system in Lactococcus lactis (15).
ΦB05-2 is a prophage-like region 17,159 bp in length with a G+C content of 35.5% that was found on contig 1000019 of the B14905 host genome. This segment is composed of 24 ORFs, 18 (75%) of which shared significant similarity (e < 10−4) at the protein level with other sequences in GenBank (Fig. (Fig.4).4). Similarities to other sequences and putative functional assignments are listed in Table Table55 and Fig. Fig.4.4. Replication of ΦB05-2 was not induced by mitomycin C treatment (Fig. (Fig.5,5, lane 8).
The ΦB05-2 segment contained phage-related proteins that are found in other Bacillus strains and prophages. ORFs 3 and 4 are similar to proteins from Streptococcus pyogenes phage 315.3 and Pneumococcus prophage EJ-1, respectively. Two phage replication-associated proteins were found in the genome: ORF 5 is similar to a replication protein found in the lytic Bacillus phage Fah, and ORF 7 is similar to the single-stranded DNA binding protein of Staphylococcus aureus prophage PVL. ORF 18 putatively encodes a terminase small subunit that is similar to the protein in temperate Bacillus clarkii phage BCJA1c. The large terminase gene encoded by ORF 19 is similar to those in Listeria phage B054, temperate Haemophilus phage Aaphi23, and Lactococcus phage TP901-1. The final ORF of ΦB05-2 is similar to transposases found in Bacillus sp. SG-1 and Streptococcus pyogenes. No integrase, lysogeny-related, or structural genes were identified by similarity searching.
The prophage-like region of ΦB05-3 is found on contig 1000024 and is 25,898 bp in length. The segment is composed of 43 ORFs, 27 (62%) of which were similar to other proteins in GenBank (Fig. (Fig.4).4). Table Table66 shows the significant similarities to other proteins and putative functional assignments. ΦB05-3 was present in a bacterium-free lysate made from an induced host (Fig. (Fig.5,5, lane 9).
The ΦB05-3 genome contained lysogeny-related and phage replication genes that were similar to those of other temperate phages and prophage regions in bacterial genomes (Table (Table6).6). ORF 1 is similar to integrase proteins found in temperate Listeria phage A500 and Mycobacterium phage Ms6. ORF 3 encodes a phage repressor that is similar to those in Geobacillus phage GBSV1 and B. cereus phage BC6A51. ORF 7 is similar to the phage antirepressor proteins from S. aureus prophage PV83 and temperate coliphage P1. The protein encoded by ORF 10, which has low similarity to a transcriptional regulator from Geobacillus kaustophilus, was expressed and detected in a cesium chloride-purified phage lysate (Fig. (Fig.6).6). The estimated size of the protein was similar to the observed molecular size of the protein. ORF 20 and ORF 21 are similar to phage replication genes from several Bacillus phages and Staphylococcus phage PVL108.
Late phage genes involved in packaging and lysis are found on the right side of the ΦB05-3 genome (Fig. (Fig.4).4). ORF 35 was similar to the terminase small subunits from Clostridium prophage C2 and B. clarkii temperate phage BCJA1c. ORF 36 encodes the terminase large subunit that shares similarity with genes in Bacillus species and Staphylococcus phages ETA2 and C2. ORF 37 is similar to portal proteins from Staphylococcus phages NM4 and ETA3. ORFs 41 and 42 are similar to a holin and a lysis protein, respectively, found in temperate Bacillus phages. No capsid or tail genes were found based on similarity to known protein sequences.
The ΦB05-4 17,991-bp prophage-like region was found on the host contig 1000003. The genome contained 24 ORFs, 22 (92%) of which shared similarity with other proteins in GenBank (Fig. (Fig.4;4; Table Table7)7) . ΦB05-4 shared significant homology (22 ORFs) to a prophage-like region found in Lysinibacillus sphaericus C3-41 (18). Twelve of the ORFS were similar to genes from the temperate myoviruses Clostridium phage C2 and Streptococcus phage EJ-1. Ten ΦB05-4 ORFs also shared similarity to genes of the defective Bacillus phage PBSX, a defective prophage-like element known for packaging random portions of host DNA. Based on PCR evidence, ΦB05-4 was not mitomycin C inducible (Fig. (Fig.5,5, lane 10).
ORF 24 was similar to the integrase XkdA from B. subtilis, and ORF 23 was similar to a phage repressor protein from Clostridium phage CD119. ORFs 7 to 11 and 14 to 21 encode gene products that are similar to tail proteins from the temperate myoviruses C2 and EJ-1 and the defective phage PBSX found on B. subtilis (Table (Table7).7). Two of these proteins, the tape tail measure (ORF 15) and the tail sheath (ORF 19), were expressed in a PEG-precipitated lysate (Fig. (Fig.6).6). ORF 2, which is similar to a protein from Lysinibacillus sphaericus C3-41, was also found to be expressed in the phage proteome. ORFs 3, 13, and 14 are similar to cell wall hydrolases of C2 and PBSX, and ORF 4 is similar to a holin protein found in C2. No identifiable replication, capsid, or DNA packaging genes were found in ΦB05-4.
Electron micrographs of negatively stained B14905 lysates showed two different phage morphologies (Fig. (Fig.7).7). A possible myovirus-like particle is seen in panel A, with a capsid diameter of 138 nm and a tail length of 307 nm and width of 23 nm. Panels B and C show smaller particles with icosahedral capsids (103-nm diameter) with thick tails (210-nm length and 35-nm width), consistent with Myoviridae. There were also particles that resembled incomplete phage tail components, as seen in panels D, E, and F.
Members of the Bacillus genus are ubiquitous in nature due to their ability to form spores that are resistant to adverse conditions. The Gulf of Mexico Bacillus isolates displayed a range of spore production frequencies when sporulation was induced by artificial means during exponential growth. The isolates that had significant levels of sporulation with decoyinine treatment indicate that this chemical was effective in inhibiting GMP synthesis in these isolates. Strains B14850 and B14907 are interesting because of their high levels of sporulation with both the control and decoyinine treatments after 24 h. This could be due to poorly controlled sporulation initiation caused by a mutation resulting in repression of catabolite sensing (6, 44). Alternatively, these isolates may have salvage pathways that allow them to obtain GMP by other means, so that decoyinine treatment is ineffective in inducing sporulation (40). With the exception of B14907, the five isolates that did not produce spores during chemical or heat treatment had frequencies of sporulation below 1% under both control and treatment conditions. This indicates that these isolates may have tightly regulated initiation of sporulation that prevents any sporulation occurring when nutrients are present. Further work is needed to elucidate the signals that control sporulation in these isolates.
Eight (66%) of the isolates may have contained temperate phages, since virus-like particles were inducible upon mitomycin C addition. The high incidence of lysogeny in this isolate collection is not unexpected given the prevalence of inducible prophages in the Bacillus genus (17). Based on the absence of lytic activity during the cross-infection experiments with induced lysates, the phages from these isolates were either host specific or already lysogenized with a homoimmune prophage. It is also possible (but relatively rare) for phages to lysogenize a host without producing a lytic infection. Isolate B14912 may have a pseudotemperate phage, since the prophage was not inducible with mitomycin C but the strain displayed a high level of spontaneous prophage induction (50). Either the three Bacillus isolates that did not produce significant levels of phage do not contain a temperate phage or the prophages are not mitomycin C inducible. The absence of prophage has been confirmed from in silico investigation of the draft genome of B14911 (GenBank accession number AAOX00000000). The range of prophage induction in these Bacillus isolates indicates that the phages in these Bacillus isolates have different relationships with their hosts.
Isolates B14905, B14906, B14910, and B14850 had significant prophage induction and high levels of sporulation (Fig. (Fig.3;3; Table Table3).3). It is possible that these isolates may contain temperate phages that enhance host sporulation, such as temperate phages PMB12, SP10, and Φ3226 (46, 53). Prophage-encoded transcriptional factors may be the mechanism for sporulation enhancement in our Bacillus isolates, as was seen for Clostridium phage Φ3226 (53).
The lysogen B14905, which produced high levels of spores under decoyinine induction, was selected for sequencing in order to investigate the influence of prophage on sporulation. Sequencing of a mitomycin C-induced lysate resulted in a single phage genome, ΦB05-1. Parallel sequencing of the B14905 chromosomal DNA confirmed the presence of ΦB05-1 as well as three additional prophage-like regions. Genomic analysis of these four regions along with transmission electron microscopy and PCR analysis was performed in order to elucidate which of these regions were inducible prophages.
The genome sizes of the four prophages, ranging from 17,991 bp to 25,898 bp, were much smaller than the sizes of typical tailed prophages (1, 10). Smaller phage genome sizes are usually associated with lytic phages, defective phages, or phage remnants (26). However, a survey of the genomes of 113 marine bacterial isolates indicated that most prophage regions were under 30 kb (37). Gene content, although it can be related to size, is a more important consideration for determining prophages. The genomes of ΦB05-1, ΦB05-3, and ΦB05-4 contained integrases and phage repressor proteins, while that of ΦB05-2 contained only phage replication, terminase, and transposase genes (Fig. (Fig.4).4). Since ΦB05-2 was not amplified by PCR from the mitomycin C-induced phage DNA, we believe this region to be a prophage remnant. Prophage remnants, which can contain functional genes, are common in bacterial genomes and are believed to be a result of decay processes (11).
ΦB05-4 was the only putative prophage of B14905 that contained several identifiable structural genes. These ORFs were similar to tail genes from the temperate myoviruses from Gram-positive bacteria, C2 and EJ-1, and to those of the B. subtilis defective phage PBSX. The presence of lysogeny-like genes, along with the absence of any replicative or packaging genes, on ΦB05-4 implies that this segment may encode a defective phage. Defective phages can be mitomycin C inducible, and some are able to form phage particles that have bactericidal activity but are noninfectious (17). In the case of PBSX, random 13-kb portions of the host chromosomal genome are packaged but are not injected into other host cells (3, 51). In this respect, ΦB05-4 may be similar to PBSX since it did not produce a PCR amplicon in the induced phage DNA. A notable difference between the two phages is the lack of identifiable terminase, capsid, and packaging proteins in ΦB05-4. In PBSX, this group of genes, which is about 6,000 bp in length, is located between the xre repressor and tail genes (23); this region is not seen in the gene map of ΦB05-4 (Fig. (Fig.4).4). Given the absence of injectable phage DNA and the lack of capsid proteins, ΦB05-4 may be a tail-like bacteriocin. Bacteriocins are usually proteinaceous particles that have bactericidal activity toward closely related strains (13). Some high-molecular-weight bacteriocins resemble phage tails, including the F- and R-type pyocins of Pseudomonas aeruginosa, carotovoricin Er from Erwinia carotovora, and a bacteriocin from Bacillus azotofixans (32, 33, 42). When B14905 mitomycin C-induced lysates were examined by TEM, many myovirus-like tail particles were observed (Fig. 7D to F). The tail tape measure and tail sheath proteins were also expressed in a phage lysate (Fig. (Fig.7).7). Based on the evidence from our experiments, we believe that ΦB05-4 is a tail-like bacteriocin, although additional experiments to identify in situ production of tail proteins in an induced lysate are needed to confirm this.
ΦB05-1 and ΦB05-3 are inducible temperate phages of isolate B14905, based on their amplification from induced phage DNA. Given the lack of identifiable structural genes in these two prophages, it was not possible to identify these phages with the tailed phage structures observed by TEM examination of the lysates. Besides the presence of a lysogeny module, the genomic architectures of these two prophages differ substantially. The genome of ΦB05-1 is smaller and does not have the functional modules that are associated with tailed prophages. An interesting feature of the ΦB05-1 phage genome is the presence of four transcriptional regulators. The one encoded by ORF 2 was not similar to phage-encoded gene products but to the Xre-like DNA binding domains in other Bacillus transition-state regulators. These transcriptional regulators, which include AbrB and SinR, redirect cellular metabolic activity to use the available source and regulate the expression of sporulation genes during the beginning of the stationary phase (43, 48). SinR specifically inhibits transcription of spo0A and stage II sporulation genes until sufficient levels of phosphorylated Spo0A accumulate. Possession of transition-state regulators by prophages may provide an additional level of sporulation control in marine Bacillus isolates. The lack of similarity of the N terminus of ORF 2 to any known sequences may indicate that this gene is a novel transcriptional regulator. Further experiments are needed to assess the function of this potential prophage-encoded transcriptional regulator in B14905 sporulation.
ORF 17 is another transcriptional regulator found on the genome of ΦB05-1. This gene is located upstream from two putative choloylglycine hydrolase genes; all three genes are similar to those found in two B. cereus genomes (16). Choloylglycine hydrolases are bacterial proteins that degrade bile salts in the mammalian intestine (5). Since B. cereus is an opportunistic intestinal pathogen, these genes might serve as a survival mechanism for this bacterium. The presence of the genes in an inducible prophage of a marine Bacillus isolate might be indicative of a horizontal gene transfer event mediated by transduction.
The genome of ΦB05-1 also contained two other ORFs that had weak hits to transcriptional regulators found in bacterial genomes. It is possible that ORF 3 is involved in lytic gene expression, given that its location and orientation are similar to those of cro repressors in other temperate phages (9). Given its similarity to experimentally characterized protein LlaI, ORF 20 may be involved in phage-encoded resistance. The location of this ORF in the genome is similar to those of methyltransferase proteins found in lytic DNA modification modules encoded by other temperate phages, such as VHML and N15 (34, 41). The variety of transcriptional regulators in ΦB05-1 may be indicative of the role of these proteins in regulating host and phage functions. For example, Chen et al. (12) clearly showed that the cI repressor of coliphage λ not only bound to the operator regions of the lytic λ genes but also could bind to the operators of host metabolic genes. There were multiple cI binding sites in the operator of the host gene pckA (phosphoenol carboxykinase), the first gene involved in gluconeogenesis. Downregulating metabolically expensive or wasteful pathways could provide lysogens an advantage during starvation survival, as occurs in most of the oligotrophic oceans (37).
The genome of ΦB05-3 is the largest of the prophage-like regions of B14905 and is the closest in genome content to that of a classic tailed phage (11). Genes involved in phage replication, phage particle assembly, and lysis are similar to those found on temperate phages. Even though the genomic and PCR evidence suggests that ΦB05-3 is an inducible prophage, the genome of this phage was not recovered by sequencing of an induced phage lysate as ΦB05-1 was. It is possible that this phage was induced at a low copy number relative to that of ΦB05-1, so that it was at too low of a concentration to be sequenced. Further functional studies are needed to determine the relationship this phage has with the host strain.
The combination of in silico analysis of prophage in the B14905 genome and in vivo molecular studies of the induced phage lysate provided information about host-phage interactions that would not be possible using either approach alone. The results from this study indicate that polylysogeny occurs in marine Bacillus strains, as has been observed for terrestrial strains (17, 23). The diversity of the two inducible prophage and two prophage-like elements found on this single bacterial genome supports metagenomic studies that have determined that the diversity of phages in the ocean is vast (4). This work provides further suggestive evidence that prophages and hosts have coevolved advantageous adaptations to survive under adverse conditions.
This research was supported by NSF Biocomplexity award OCE0221763 to J.H.P. and A.M.S. and a Florida Sea grant and Aylesworth Foundation Old Salts scholarship to J.M.
We are indebted to the Proteomics Core Facility of the Moffitt Cancer Center for performing MALDI-TOF-TOF analysis.
Published ahead of print on 11 December 2009.
†The authors have paid a fee to allow immediate free access to this article.