Cloning and Characterization of the DNA Region Responsible for Megacin A-216 Production in Bacillus megaterium 216

Abstract

Upon induction, Bacillus megaterium 216 produces the bacteriocin megacin A-216, which leads to lysis of the producer cell and kills B. megaterium and a few other bacterial species. The DNA region responsible for megacinogeny was cloned in B. megaterium. The nucleotide sequence of a 5,494-bp-long subfragment was determined, and the function of the genes on this fragment was studied by generating deletions and analyzing their effects on MegA phenotypes. An open reading frame (ORF) encoding a 293-amino-acid protein was identified as the gene (megA) coding for megacin A-216. BLAST searches detected sequence similarity between megacin A-216 and proteins with phospholipase A2 activity. Purified biologically active megacin A-216 preparations contained three proteins. Mass spectrometry analysis showed that the largest protein is the full-length translation product of the megA gene, whereas the two shorter proteins are fragments of the long protein created by cleavage between Gln-185 and Val-186. The molecular masses of the three polypeptides are 32,855, 21,018, and 11,855 Da, respectively. Comparison of different megacin preparations suggests that the intact chain as well as the two combined fragments can form biologically active megacin. An ORF located next to the megA gene and encoding a 91-amino-acid protein was shown to be responsible for the relative immunity displayed by the producer strain against megacin A-216. Besides the megA gene, at least two other genes, including a gene encoding a 188-amino-acid protein sharing high sequence similarity with RNA polymerase sigma factors, were shown to be required for induction of megacin A-216 expression.

Megacins are bacteriocins produced by certain strains of the gram-positive spore-forming bacterium Bacillus megaterium (13, 17, 36, 39). Megacinogeny was first described in 1954 (17). It was observed that some cells in a growing culture of B. megaterium 216 spontaneously produced a substance that lysed cells of other B. megaterium strains but did not affect most other bacterial species. Production of this substance, later termed megacin A-216, is inducible by UV light (17), N-methyl-N′-nitro-N-nitrosoguanidine (28), or mitomycin C (23). Induction of megacin A-216 synthesis leads to lysis of the culture in 2 to 3 h (17). Cell-free supernatants of an induced culture of B. megaterium 216 kill sensitive cells even in 10⁴-fold dilutions, whereas the producer strain 216 exhibits relative immunity to the bacteriocin (17). Like bacteriocins in general, megacin A-216 has a rather narrow antibacterial spectrum: besides B. megaterium, it is active against some strains of Bacillus subtilis, Bacillus anthracis, Micrococcus aurantiacus, and M. cinnabareus (17, 18, 28).

Many observations indicate that megacin A-216 acts by impairing the cell membrane integrity of sensitive bacteria. Microscopic pictures show that the cell content is partially released from the killed bacteria, while the cell wall appears to stay intact and the shape of the cells is retained (19). Megacin A-216 is active against protoplasts of megacin-sensitive strains (13, 28). In isosmotic medium, the cell material of the protoplasts becomes less dense after megacin treatment, but the cell contours remain visible for a long time. These observations suggest that megacin A-216 somehow damages the membrane permeability barrier but does not destroy the membrane completely. The bacteriocidic effect of megacin A-216 is inhibited by low temperature, suggesting involvement of enzymatic activity in megacin A-216 action (13).

Megacin A-216, first purified by Holland, was found to be an acidic protein precipitating around pH 4.0 (14). Later studies determined that it has a native molecular mass of ∼66 kDa, contains two different subunits with molecular masses of 30 and 15 kDa, respectively, and displays phospholipase A2 (PLA2) activity (27, 28, 39).

A specific inhibitor of megacin A-216, purified from cultures of B. megaterium 216, was suggested to confer immunity to the self-produced bacteriocin. The partially purified substance, whose chemical nature was not identified, inhibited the toxic effect as well as the phospholipase activity of megacin A-216 (26).

After recognition of bacteriocin production in other B. megaterium strains, a classification of megacins was proposed (types A, B, and C). Based on its inducibility by a low level of UV irradiation or mitomycin C treatment, megacin A-216 was classified as type A (15). In a broader context, its physical properties and enzymatic activity place megacin A-216 in class III of bacteriocins produced by gram-positive bacteria (12, 20). There are only a few bacteriocins (megacin A-216, megacin A-19213, and thuricin) that have been shown to have phospholipase activity (9, 28, 39). Interestingly, these bacteriocins, although clearly different proteins, have similar producer hosts (closely related gram-positive bacteria) and similar genetic localizations (encoded by plasmids) (9, 39).

A study using protoplast fusion and protoplast transformation demonstrated that the genetic determinants of megacin A-216 production, immunity to megacin A-216, and inducibility of megacin A-216 production were transferable into nonmegacinogenic B. megaterium strains, and acquisition of the MegA phenotypes was associated with the transfer of an ∼47-kb plasmid (pBM309), one of the 10 indigenous plasmids present in B. megaterium 216 (32). Fragments of pBM309 were cloned in B. megaterium, and the genes responsible for megacin production were localized to an approximately 10-kb region of pBM309; however, this DNA region was not analyzed beyond establishing a restriction map (30, 41).

The renewed interest in bacteriocins as antimicrobial agents (5, 8, 10) and in B. megaterium as an industrial microorganism and cloning host (38) prompted us to revisit megacinogeny of B. megaterium. The interest in B. megaterium is highlighted by the ongoing whole-genome sequencing project (http://www.bios.niu.edu/b_megaterium). Our goal is to study the molecular mechanisms underlying the megacin A-216 action and the relative immunity displayed by the producer strain. The PLA2 activity and its associated natural inhibitor (26) make the megacin A-216 system particularly interesting because PLA2 enzymes play important roles in a wide range of physiological and pathophysiological processes (3), as well as in bacterial virulence (35). Because the recombinant plasmids constructed previously (30, 41) were not available, as a first step toward characterization of the megacin A-216 system, we recloned the megacin A-216 region, determined its DNA sequence, and used deletion mapping to identify the genes controlling megacin A-216 production and immunity. Assignment of the megA gene was also confirmed by mass spectrometry (MS) analysis of purified megacin A-216. We report data showing that the two subunits previously detected in megacin A-216 preparations (39) are proteolytic cleavage products of the protein encoded by the megA gene.

MATERIALS AND METHODS

Strains, plasmids, and growth conditions.

The B. megaterium strains 216 (MegA⁺ MegC⁺), KM (Spo⁻; megacin sensitive), AL (MegA⁺ Arg⁻ Leu⁻ Str^r; pBM309), and THT (KM Thr⁻ His⁻ Trp⁻ Str^r) were described previously (17, 32). The kanamycin-resistant mutant of the THT strain was obtained from K. Fodor. Escherichia coli ER1821 F⁻ glnV44 e14 (McrA⁻) endA1 thi-1 Δ(mcrC-mrr)114::IS10 obtained from New England Biolabs was used as the E. coli cloning host.

To obtain an E. coli-B. megaterium shuttle vector with a single PstI site, part of the polycloning site of the 4.7-kb plasmid vector pHY300PLK (16) was deleted by digestion with EcoRI and XbaI, filling-in the ends with Klenow polymerase, and ligation of the blunted ends to yield pHY301. All recombinant plasmids reported here are based on pHY301. Bacteria were routinely grown in LB medium at 30°C (B. megaterium) or at 37°C (E. coli). Ampicillin was used at 100 μg/ml, tetracycline at 10 μg/ml (B. megaterium) or 12.5 μg/ml (E. coli), and kanamycin at 50 μg/ml.

DNA techniques.

Plasmid DNA was introduced into B. megaterium by protoplast transformation (42). Plasmid preparation, restriction digestion, agarose gel electrophoresis, and cloning of DNA fragments were carried out using standard procedures (33). For plasmid preparation from B. megaterium, cells were first converted to protoplasts by incubation with lysozyme to facilitate lysis. DNA sequence was determined by an automated sequencer.

Megacin assay.

Five-microliter aliquots of overnight cultures were pipetted and allowed to dry onto the surface of LB agar plates. After incubation at 30°C for 6 h, the plates were overlaid with LB-kanamycin soft agar (0.4 to 0.5%) containing 1/100 vol of a dense culture of the Kn^r mutant of B. megaterium strain THT and incubated at 30°C overnight. Kanamycin was added to the soft agar to inhibit growth of the colonies tested. A clear zone surrounding the colony identified clones producing megacin.

To study UV light induction of megacin A-216 production, the agar plates with the dried test cultures were incubated at 30°C for 2 h and then irradiated with UV light (germicid lamp) for 15 s. After the UV treatment, the plates were returned to 30°C for 4 h and then megacin production was assayed as described above.

Immunity to megacin A-216 was estimated by titrating a cell-free lysate of a mitomycin C-induced B. megaterium 216 culture on the clone tested and, as a control, on the sensitive strain B. megaterium KM(pHY301). Five-microliter aliquots of serial dilutions of the lysate were pipetted onto the surface of a 3-ml soft agar layer containing 10 μl of a dense B. megaterium culture. After an overnight incubation at 30°C, the level of resistance was assessed by determining the highest dilution that caused a clear spot in the bacterial lawn. Under the conditions of the assay, the 10⁵-fold dilution still caused lysis of the sensitive KM strain, whereas clones carrying the intact megA region were not affected even by the undiluted lysate.

Protein purification.

Megacin A-216 was purified by a modification of methods described previously (14, 39). An overnight culture of B. megaterium 216 was diluted 1:100 into fresh LB medium. After shaking at 30°C for 2 h, mitomycin C was added to a concentration of 0.5 μg/ml and shaking was continued for 3 to 4 h until lysis occurred.

All steps of purification were performed at 4°C. Cell debris was removed by centrifugation. Proteins were precipitated from the supernatant by adding ammonium sulfate to 80% saturation. The precipitated material was dissolved in PC buffer (20 mM sodium phosphate [pH 6.0], 10 mM 2-mercaptoethanol, 0.1 mM EDTA, 5% glycerol) and dialyzed against the same buffer. Proteins precipitated during dialysis were removed by centrifugation. The solution was treated with 2% streptomycin sulfate to remove nucleic acids. After centrifugation, the supernatant was dialyzed against PC buffer and loaded onto a DE52 (Whatman) anion-exchange column equilibrated with the same buffer. Proteins were eluted with a 0 to 0.5 M linear NaCl gradient in PC buffer. Fractions were assayed for megacin A-216 activity by titration on B. megaterium KM strain. Pooled active fractions were dialyzed against PC buffer and loaded onto a hydroxyapatite column equilibrated with PC buffer. Proteins were eluted with a 20 to 400 mM sodium phosphate (pH 6.0) gradient containing 10 mM 2-mercaptoethanol, 0.1 mM EDTA, and 5% glycerol. Peak megacin-containing fractions were pooled, and the protein was precipitated with ammonium sulfate (80% saturation). The precipitate was dissolved in a minimal volume of buffer S (20 mM sodium phosphate [pH 7.4], 150 mM NaCl, 10 mM 2-mercaptoethanol, 0.1 mM EDTA) and further purified by gel filtration on a Sephacryl S-200 column in buffer S.

Purified megacin preparations were dialyzed against PC buffer and stored at −80°C for several months with only slight loss of biological activity. Protein concentration was determined by the Bradford reaction, using Bio-Rad protein assay reagent with a bovine serum albumin calibration curve.

Estimation of native molecular weight by gel filtration.

A sample of purified megacin A-216 (450 μl, 7.5 mg/ml) was loaded onto a 1-by-50-cm Sephacryl S-200 column in S buffer. Dextran blue 2000 was used for determination of the exclusion volume, and aldolase, bovine serum albumin, chymotrypsinogen A, and RNase A (with molecular masses of 158, 67, 25, and 13.7 kDa, respectively) were used for calibration.

Electrophoresis of proteins.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of proteins was performed in 12% gels using Coomassie brilliant blue R-250 staining.

For electrophoresis under nondenaturing conditions, SDS and 2-mercaptoethanol were omitted. Megacin samples migrated toward the anode. Proteins were extracted from nondenaturing gels by electroelution. Gel pieces containing unstained bands were cut out using adjacent gel slices with stained megacin bands and prestained molecular weight markers as guide. Electroelution was performed in Little Blue Tank gel electrophoresis equipment (ISCO) using 1× Tris-glycine (25 mM Tris, 192 mM glycine [pH 8.3]) in the tank and 0.2× Tris-glycine in the sample-holding chamber. Two hundred volts (1 to 2 mA per sample holder) was applied for 7 h.

MS.

Samples were prepared for MS by in-gel digestion (http://ms-facility.ucsf.edu/ingel.html). Briefly, disulfide bridges were reduced by dithiothreitol, and then the free Cys residues were alkylated by iodoacetamide. Samples were incubated with side-chain-protected porcine trypsin (Promega) for 4 h at 37°C. Unfractionated digests were analyzed on a Bruker Reflex III matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometer in reflectron mode. One microliter of 10 μl of digest solution was loaded with 1 μl 2,5-dihydroxybenzoic acid solution (Sigma-Aldrich) as a matrix for the MS analysis. External calibration was applied. Post-source decay (PSD) analysis was performed in order to obtain sequence information for selected peptides.

Bioinformatic methods.

DNA and protein sequence similarity searches were performed using the BLAST program package (1, 2), pairwise comparison of sequences by the BLAST 2 Sequences program (37), conserved domain searches by the CD-Search program (22), and PROSITE signature searches by the ScanProsite program (6). In all cases, default settings were used.

Nucleotide sequence accession number.

Nucleotide sequence data have been deposited in GenBank under accession no. EU014074.

RESULTS

Cloning of the DNA region responsible for megacinogeny.

Plasmid pBM309 carrying the genetic determinants of megacin A-216 production and immunity (32) was purified from B. megaterium AL, a strain harboring only this plasmid. We chose to clone the ∼10-kb PstI fragment, which, in a previous study (30), was shown to carry the megA genes. To obtain a plasmid shuttle vector with a unique PstI site, pHY300PLK (Ap^r Tc^r), originally developed as a B. subtilis-E. coli shuttle vector (16), was modified to yield pHY301, as described in Materials and Methods. PstI fragments of pBM309 were ligated to PstI-digested pHY301, and the ligated DNA was used to transform protoplasts of the plasmid-less B. megaterium THT strain. Transformed cells were regenerated in a tetracycline-containing batch culture. It was assumed that the acquired capacity of megacin A-216 production would endow clones carrying the megA genes with a very strong selective advantage. Restriction analysis of the plasmids purified from Tc^r clones revealed that all contained the same ∼10-kb PstI fragment. One plasmid (pMGC1) was selected for further analysis. In the overlay assay, B. megaterium THT colonies harboring pMGC1 produced a narrow clear zone around the colony, which became wider when the colonies were briefly exposed to UV irradiation; this indicated that the cloned PstI fragment contained the genetic elements associated with the MegA phenotypes (Fig. 1). Interestingly, pMGC1 conferred the MegA⁺ phenotype also to E. coli ER1821 cells, although the zone of growth inhibition was smaller than around MegA⁺ B. megaterium colonies (not shown).

FIG. 1.

Megacin A-216 production by B. megaterium THT cells carrying different sections of the megacin A-216 region and by the parental B. megaterium 216 strain. Megacin A-216 production was tested by the overlay assay as described in Materials and Methods. Colonies in the bottom row were irradiated by UV light. Numbers refer to recombinant pMGC plasmids (see text and Fig. 2). B.m., B. megaterium 216.

A 5.5-kb HindIII fragment of the pMGC1 insert was subcloned in pHY301 to yield pMGC3 and was subjected to further analysis.

Sequence analysis of the DNA region responsible for megacinogeny.

The nucleotide sequence (5,494 bp) of the cloned HindIII fragment was determined (GenBank accession no. EU014074). Nine open reading frames (ORFs) starting with ATG, encoding a protein longer than 50 amino acids and nonoverlapping with other ORFs, were identified (Fig. 2 and Table 1). Translated sequences of all ORFs were used to search the GenBank database for similar amino acid sequences. Start and stop coordinates of the ORFs and results of the BLASTP search are listed in Table 1. Going from left to right on the map of Fig. 2, the translated product of ORF59 showed a medium level of similarity to a hypothetical protein of Bacillus cereus AH1134 and weaker similarity to a number of other hypothetical bacterial proteins in the database. The predicted protein (P91) encoded by the next ORF showed marginal sequence similarity to a single protein, a hypothetical protein of Enterococcus faecalis. Conceptual translation of ORF293A yielded a protein with moderate level of similarity to proteins with PLA2 activity. Moreover, the BLAST search identified conserved phospholipase domains (PLA2_plant, PLA2_bee_venom_like, Phospholip_A2_2, PLA2c, and PLA2_like) matching the C-terminal 30% of P293A. In the light of previous results demonstrating phospholipase activity of purified megacin A-216 (28, 39), detection of sequence similarity to proteins with phospholipase activity suggested that ORF293A is the gene encoding megacin A-216. P62, determined by the next ORF, showed very low level of similarity to hypothetical proteins from diverse sources and a medium level of similarity to a hypothetical protein of the taxonomically closely related B. thuringiensis. The neighboring gene (ORF85) encodes a protein with medium level of similarity to glutaredoxins and related proteins. The BLAST search also identified an NrdH conserved domain in P85. The predicted product of the next ORF (ORF293B) shares high sequence identity with a large number of serine proteases. The putative ORF188 product is very similar to bacterial RNA polymerase sigma factors containing the Sigma70_r4 conserved domain. The translational start point of P188 is somewhat ambiguous, because there is an in-frame ATG 9 bp upstream of the ATG at 3361. However, there is no appropriately positioned ribosomal binding site in front of the upstream start codon; thus, ATG₃₃₆₁ (Table 1) is more likely to be the initiator codon for this protein. Hypothetical translation of the next ORF (ORF73) yields a protein displaying a low level of similarity to hypothetical proteins of different Bacillus species. Finally, the predicted ORF185 product is highly similar to several bacterial cell wall hydrolases with N-acetylmuramoyl-l-alanine amidase activity.

FIG. 2.

Functional mapping of the megA region. Horizontal bars indicate DNA segments present in the plasmids. Restriction sites used to delete specific DNA segments are shown. ORFs present in the pMGC3 insert and left intact in the deletion derivatives are indicated by filled horizontal arrows, and the putative tRNA^Cys gene is indicated by an empty arrow. Numbers below the arrows in the top row indicate the length of the ORF in amino acids, and names above the arrows indicate the putative function derived from BLAST search and functional analysis. Gltrdx, ORF encoding a glutaredoxin-like protein; reg, ORF encoding a putative regulatory protein playing a role in megacin expression. The capacity to produce megacin A-216 (MegA column), immunity to megacin A-216 (MegA^imm column), and inducibility of megacin production by UV light (UVind. column) is indicated by +, whereas the lack of these features is indicated by −. A low level of megacin production is indicated by ±. pMGC24*, insert in the opposite orientation.

TABLE 1.

Results of amino acid sequence similarity search with proteins encoded by the megA region^a

ORF no.	Start coordinate	Stop coordinate	Conserved domain with highest similarityb	Characteristic of protein with highest sequence similaritya
				Protein	Database ID	Organism	E value
59	259	80	None	Hypothetical protein BcerAH1_14544	167937449	Bacillus cereus AH1134	9e-13
91	396	671	None	Hypothetical protein EF0543	29375161	Enterococcus faecalis	1.2
293A	1593	712	PLA2_plant	Hypothetical protein BcerW_26065	167956469	Bacillus cereus W	1e-05
62	1885	1697	None	Hypothetical protein BT9727_0866	49477018	Bacillus thuringiensis serovar konkukian strain 97-27	5e-06
85	2145	1888	NrdH	Glutaredoxin	169824245	Finegoldia magna ATCC 29328	4e-11
293B	3158	2277	Peptidase_S8	Probable intracellular serine protease	167937446	Bacillus cereus AH1134	3e-68
188	3361	3927	Sigma70_r4	Possible DNA-directed RNA polymerase sigma subunit FliA	167937451	Bacillus cereus AH1134	6e-36
73	3971	4192	None	Hypothetical protein BSG1_14318	149183037	Bacillus sp. strain SG-1	9e-06
185	4381	4938	Amidase_3	Putative N-acetylmuramoyl-l-alanine amidase	148747751	Geobacillus virus E2	9e-42

^aORFs starting with ATG, longer than 50 amino acids, and nonoverlapping with other ORFs are shown. ORFs are identified by the number of the encoded amino acids and are listed in their order of location from left to right as shown on Fig. 2. Start and stop coordinates define nucleotide positions within the sequenced 5,494-bp fragment. Proteins encoded by ORF91 through ORF185 have GenBank identifiers ABS44966 through ABS44973.

^bIdentified by BLAST search of the nonredundant protein database (March 2008).

Gram-positive bacteria are characterized by strong ribosomal binding sites (24, 25, 31). Sequences located immediately upstream of the ORFs were screened for putative Shine-Dalgarno sequences. All ORFs have preceding sequences with similarity to consensus Bacillus ribosomal binding sites (34) and to a consensus sequence deduced from the 3′ ends of B. megaterium 16S rRNAs (Table 2). The potentially strongest ribosomal binding sites (showing the highest similarity to the consensus sequence) were detected before ORF293A, ORF91, and ORF293B (Table 2). A consensus Shine-Dalgarno sequence was detected before ORF59, but the distance between the “core” GGAGG motif and the start codon is probably too long for efficient function (24, 25).

TABLE 2.

Putative ribosomal binding sites

ORF no.	Shine-Dalgarno sequencea
59	AAGGAGGTGATcgtactgATG
91	GGAGGaGAtttttaATG
293A	atAGGAGGaGAaatATG
62	ttGGAGtattaagtATG
85	AagGGtGaaGtctaATG
293B	gaGAGGTGAaaaagATG
188	AAGGAatgtATgtaATG
73	tgGGAGGaaAagttATG
185	agAAGGtGaaagatATG

^aThe ATG start codons are shown in boldface. Nucleotides matching the consensus B. megaterium sequence (5′-AAGGAGGTGAT) derived from two GenBank entries (DQ660362 and AY180964) for 16S rRNA are printed in capital letters.

A BLASTN search identified a short, approximately 70-bp region, located between nucleotide positions 4279 and 4345, that shows strong similarity to tRNA^Cys genes of gram-positive bacteria. The highest sequence similarity (94%) was detected with the tRNA₂^Cys gene of Bacillus thuringiensis serovar konkukian strain 97-27 (NC_005957, positions 969811 to 969884) of the B. thuringiensis genome. An inverted repeat was found between nucleotides 4488 and 4521 (within ORF185).

The GC content of the sequenced region is 30.85%, lower than the average GC content of B. megaterium DNA (37%) (29) and lower than that of two B. megaterium plasmids, pBM300 (35.2%) and pBM400 (36.5%), whose sequences have become available recently (21, 34).

To our knowledge, transcription signals of B. megaterium have not been systematically analyzed, making promoter assignment difficult. A sequence resembling the B. subtilis sigma A promoter consensus sequence TTGACA (−35), TATAAT (−10) (11), was found to precede ORF91.

Functional mapping of the megA region.

Detection of a conserved PLA2 domain strongly suggested that ORF293A is the structural gene of megacin A-216. To corroborate this assignment and identify other genes that might play a role in megacin A-216 production and immunity, deletion derivatives of pMGC3 were constructed by removing restriction fragments or by subcloning fragments of the original HindIII insert in pHY301. The plasmids were constructed in E. coli and subsequently introduced into B. megaterium THT to test MegA phenotypes. Extension of the deletions and their phenotypic effects are shown in Fig. 1 and 2.

With regard to megacin production, pMGC1 and its deletion derivatives can be classified into four types. (i) The B. megaterium THT clone harboring pMGC1 with the ∼10-kb PstI insert produced a narrow zone of growth inhibition, which increased substantially upon UV light irradiation. This phenotype was similar to that of the parental B. megaterium 216 strain (Fig. 1). (ii) pMGC3 carrying the sequenced 5.5-kb HindIII fragment and its deletion derivatives (pMGC21, pMGC20, and pMGC23), in which only ORF293B or ORF185 was inactivated either alone or in combination, produced a wider inhibitory zone around uninduced colonies than pMGC1 (Fig. 1 and 2). The diameter of the inhibitory zone around UV-irradiated colonies was similar to that around induced colonies harboring pMGC1. The effect of UV light was especially evident when samples from young cultures were tested and the cells, dried onto the surface of the agar plate, were allowed to grow for only a short period of time (2 h) before the UV treatment. (iii) Clones belonging to the third class in some experiments produced a narrow clear zone around the colony. This was more often detected with pMGC6 and pMGC10, than with pMGC24, but the phenomenon was, in general, not reproducible. UV irradiation did not increase megacin production (Fig. 1). These plasmids (pMGC6, pMGC10, and pMGC24) carried shorter or larger deletions affecting several genes located on the 5.5-kb HindIII fragment, but ORF293A was always intact (Fig. 2). (iv) Clones categorized in the fourth class did not show megacin production. Some of the plasmids belonging to this class (pMGC26, pMGC17, and pMGC30) lacked intact ORF293A, while others (pMGC28, pMGC4, and pMGC29) had intact ORF293A but carried deletions affecting ORF73 and the tRNA₂^Cys gene (pMGC28, pMGC4, and pMGC29) (Fig. 2).

Megacin immunity was tested by a plate assay using a high-titer cell extract prepared from B. megaterium 216. A large deletion removing all ORFs but ORF91 and ORF59 (pMGC30) did not impair megacin immunity, whereas deletions inactivating ORF91 (pMGC24, pMGC17, and pMGC40) led to loss of megacin immunity, suggesting that the putative 91-amino-acid protein is responsible for the relative resistance of the producer strain to megacin A-216.

Analysis of the phenotypes of the deletion derivatives allowed us to test whether, in addition to ORF293A and ORF91, other genes have a role in megacinogeny. The lack of phenotypic change associated with the deletion of ORF293B and ORF185 (see above) suggested that neither the putative protease nor the cell wall amidase plays a role in megacin production. A deletion extending into ORF188 led to severe reduction (pMGC6), whereas deletion of ORF73 and the tRNA₂^Cys gene alone (pMGC28), or in combination with ORF188 (pMGC4 and pMGC29), abolished megacin production. All deletions affecting ORF188 or ORF73/tRNA₂^Cys led to loss of UV inducibility of megacin production. These observations point to the crucial role of the predicted sigma factor-like protein (P188) in megacin expression and especially in high-level expression after UV treatment. ORF73 and/or the tRNA₂^Cys gene seems to be even more important, because their inactivation had more drastic effect than deletion of ORF188. The available deletions did not allow separate testing of ORF73 and the tRNA₂^Cys gene. Puzzlingly, pMGC10 and, in some assays, pMGC24, which lack both ORF188 and ORF73/tRNA₂^Cys, showed signs of weak megacin production, but inducibility was lost also in these cases.

Purification and analysis of the megacin A-216 protein in vitro.

DNA sequence analysis and deletion mapping described above identified a single gene (ORF293A) that encodes megacin A-216. This finding was surprising in the light of previous results, which showed the presence of two different subunits in purified megacin A-216 preparations (39). To address this question, megacin A-216 was purified from lysates of mitomycin C-induced cultures of B. megaterium 216 and B. megaterium THT(pMGC3). Estimations using the biological assay indicated that megacin activity in lysates of the native host was approximately 100-fold higher than that in lysates of the recombinant clone. Megacin preparations were analyzed by SDS-PAGE. B. megaterium 216 lysates contained three dominant bands with mobilities corresponding to ∼40, 30, and 15 kDa (Fig. 3). The 30- and 15-kDa polypeptides appeared to be present in equal molar amounts, whereas the amount of the 40-kDa species varied in different lysates. During the same purification, the relative intensities of the bands did not change. The 30- and 15-kDa polypeptides corresponded to the α and β subunits described previously (39). In the same study, the authors also found a 40-kDa polypeptide, which copurified with megacin A-216, but this longer polypeptide was a minor component of their purified megacin preparations and was not considered to be part of the megacin A-216 system (39). In our hands, the 40-kDa protein, designated γ polypeptide, was a major component of the purified preparation, when purification started from B. megaterium 216 extract. Moreover, this protein was the single dominant protein in lysates of B. megaterium THT(pMGC3) and the main component of purified megacin obtained from the recombinant clone (Fig. 3).

FIG. 3.

Purification of megacin A-216 from B. megaterium THT(pMGC3) and B. megaterium 216. Shown are the results of SDS gel electrophoresis of protein samples obtained in different stages of purification. M, molecular mass markers (Fermentas); lane 1, lysate; lane 2, after DEAE-cellulose and Sephacryl-S 200 chromatography; lane 3, lysate; lane 4, after ammonium sulfate precipitation, streptomycin-sulfate precipitation of nucleic acids, and dialysis; lane 5, after DEAE-cellulose chromatography; lane 6, after hydroxyapatite chromatography.

To study their relationship, unfractionated tryptic digests of the three proteins purified from the B. megaterium 216 lysate were subjected to MALDI-TOF MS analysis. Measured masses were compared with calculated masses of predicted tryptic peptides deduced from the sequences of ORF293A, -293B, -191(188), and -185. Peptides detected in the 30-kDa sample matched parts of the N-terminal half of the P293A sequence, whereas peptides of the 15-kDa sample matched the C-terminal part of P293A (Fig. 4). The unfractionated tryptic digest of the protein with the apparent molecular mass of ∼40 kDa featured tryptic peptides characteristic for both shorter proteins. However, an abundant component with MH⁺ at m/z 2,501.7, not matching any predicted peptide, was detected only in the 15-kDa sample (Fig. 4). These data indicate that all three proteins of the purified megacin preparation are products of the same gene (ORF293A) and suggest that the α and β polypeptide chains are cleavage products of the full-length γ chain.

FIG. 4.

MALDI-TOF MS spectra of the unfractionated tryptic digests of 40-kDa (a), 30-kDa (b), and 15-kDa (c) proteins. Peptides confirmed by PSD analysis are labeled with asterisks. Mass of the predicted N-terminal peptide is underlined.

In the MALDI-TOF mass spectra of the tryptic digests of the 40- and 30-kDa protein, masses representing peptides [1-14] and [2-14] of P293A were detected, which indicated that the initiating methionine residue was posttranslationally excised from a fraction of the molecules. MALDI-TOF measurements yielded an average molecular mass of 11,832 Da for the intact β chain (data not shown). Assuming that the C terminus of the peptide was intact, this value suggested that the β chain starts at Val-186. The calculated mass for the 108-amino-acid polypeptide beginning with Val-186 and ending with Met-293 is 11,855, which differs only by 0.2% from the experimental value. The predicted MH⁺ (VKLPVPCFNNSTGCCTFSNNGK) of the unique tryptic peptide (m/z 2,501.7, with carbamidomethyl Cys) was indeed detected in the tryptic digest of the β chain. Its identity was confirmed by PSD analysis of the peptide (Fig. 5). These data indicated that the α and β polypeptides were products of proteolytic cleavage between Gln-185 and Val-186 of the full-length γ chain.

FIG. 5.

PSD spectrum of the predicted N-terminal peptide (MH⁺ = 2,501.7) of the β subunit. PSD segments were smoothed and baseline fitted before merging. For simplicity, only the y and b ions are labeled (nomenclature according to reference 4). Cys residues are carbamidomethylated.

The calculated molecular masses (32,855, 21,018 and 11,855 Da for the γ, α, and β chains, respectively) are lower than the values derived from SDS-PAGE (∼40, ∼30, and ∼15 kDa, respectively). There are no signs of posttranslational modifications other than the removal of the N-terminal methionine. Analysis of the amino acid sequence shows that megacin A-216 is a hydrophilic molecule and that the γ chain probably consists of two different domains, with their boundary at the hypothetical cleavage site between the α and β chains. The α chain is rich in acidic residues, resulting in a calculated pI of 3.92, while the β chain contains many Lys and Arg residues and has a theoretical pI of 9.05.

The native molecular weight of megacin A-216 was estimated by gel filtration on a Sephacryl S-200 column. Megacin A-216 eluted in fractions corresponding to 66 to 68 kDa (not shown), which is in accordance with earlier estimates (39). To test the composition of native megacin, purified megacin was electrophoresed in a 12% polyacrylamide gel under nondenaturing and nonreducing conditions. Two major bands, a more intense, slow-moving band and a less intense, fast-moving band, were detected (Fig. 6A, lanes 1 and 2). Proteins from both bands were electroeluted, and analyzed by SDS-PAGE (Fig. 6B). The upper (slow) band contained all three chains detected in the purified megacin preparation, whereas the lower band contained only the α chain. In biological activity tests, only the sample extracted from the slow band had megacin activity. If 2-mercaptoethanol was added to megacin samples prior to electrophoresis in the nondenaturing gel, the slow band disappeared (Fig. 6A, lanes 3 and 4), suggesting that native megacin contains intrachain disulfide bonds characteristic for phospholipase 2 enzymes, and these bonds may be important for maintaining its structure. The α chain does not contain Cys residues, which can explain why reduction did not have a big effect on its mobility. The β chain, dissociated by 2-mercaptoethanol treatment from the rest of the megacin molecule, probably did not enter the gel because its pI is higher (9.05) than the pH of the gel buffer. It is less clear why the γ chain disappeared after treatment with the reducing agent. A possible explanation is that disruption of the disulfide bonds yielded a heterogeneous population of molecules with different tertiary structures and thus with different electrophoretic mobilities.

FIG. 6.

(A) Electrophoresis of megacin A-216 in 12% nondenaturing polyacrylamide gel at pH 8.8. Lanes 1 and 2, two fractions of purified megacin A-216 after hydroxyapatite chromatography; lanes 3 and 4, samples of the same fractions after reduction with 2-mercaptoethanol; S, slow-migrating band; F, fast-migrating band. (B) SDS-PAGE analysis of proteins electroeluted from the slow and fast bands in a nondenaturing gel. Lane M, molecular weight marker; lanes P, purified megacin A-216 preparation after Sephacryl-S 200 chromatography; lane S, proteins extracted from the slow band; lane F, proteins extracted from the fast band. (C) Scheme illustrating some important features of the megacin A-216 protein. Filled horizontal bar, segment 1-185 corresponding to the α chain; empty bar, segment 186-293 corresponding to the β chain with the conserved PLA2-like superfamily domain. The protease cleavage site between residues Gln-185 and Val-186 is indicated by an arrow, and positions of cysteine residues are indicated by lollipops.

The purified megacin preparation was active on protoplasts of the sensitive strain B. megaterium KM, whereas B. megaterium 216 was immune against our preparation (data not shown). Biological activity of the preparation did not decrease after 1 h of incubation at 65°C, in agreement with previous observations demonstrating heat stability of megacin A-216 (14, 28).

DISCUSSION

Previous studies described several aspects of megacinogeny in the B. megaterium 216 strain, including the microbiological phenomenon of megacin A-216 production and location of the megA genes on one of the large plasmids present in the host strain (14, 17, 19, 28, 30, 32, 39-41). The main goal of the present work was the cloning and characterization of the genes involved in megacin A-216 production and immunity. Phenotypes (megacin production and effect of UV irradiation) of the primary clone carrying the recombinant plasmid pMGC1 were similar to those observed with the parental B. megaterium 216 strain (Fig. 1). The B. megaterium clone (pMGC3) carrying the 5,494-bp HindIII subfragment of the original 10 kb insert exhibited a wider inhibitory zone around noninduced colonies. It remains to be determined whether this difference is due to a gene present on the PstI fragment but missing from the shorter HindIII fragment or to the higher copy number of pMGC3. The observation that both pMGC1 and pMGC3 expressed megacin also in E. coli supports the interpretation that all genes playing a role in megacinogeny are present on the HindIII fragment. An ORF encoding a 293-amino-acid protein (ORF293A) was identified as the gene encoding megacin A-216. This assignment was based on three lines of evidence: (i) deletion mapping, which showed that deletion of ORF293A invariably resulted in a MegA⁻ phenotype; (ii) amino acid sequence similarity between the protein encoded by ORF293A and several known proteins with PLA2 activity; and (iii) MS analysis of tryptic peptides obtained from purified megacin A-216.

Megacin preparations purified from B. megaterium 216 contained two main proteins with apparent molecular masses of 30 and 15 kDa and a copurifying 40-kDa protein. These results were in accordance with previous observations (39). MS analysis proved that all three polypeptides are encoded by ORF293A and that there is a precursor-product relationship between them. The largest protein (termed γ chain) is the full-length product of ORF293A, whereas the shorter α and β chains appear to be products of proteolytic cleavage between Gln-185 and Val-186 of the γ chain. The calculated molecular masses of the γ, α, and β chains (32,855, 21,018, and 11,855 Da, respectively) are substantially lower than the values deduced from electrophoretic mobility in denaturing gels. The likely reason of the anomalous migration, at least in the cases of the α and γ chains, is the highly acidic composition (pI 3.92 and 4.51, respectively).

Our purified megacin preparations contained all three chains. The α and β subunits were always present in approximately equal amounts, whereas the amount of the γ chain was variable. In preparations purified from the parental strain B. megaterium 216, the two shorter chains were more abundant, whereas preparations purified from B. megaterium THT(pMGC3) were dominated by the γ chain (Fig. 3). Biologically active megacin ran in one band in native polyacrylamide gel and contained all three chains. The most likely interpretation of these results is that the full-length protein, as well as the cleavage products, can form active megacin A-216. The native molecular mass of megacin A-216 estimated by gel filtration (∼66 kDa) is consistent with an α₂β₂, αβγ, or γ₂ composition. The hypothetical protease cleaving the primary product has not been identified. It also remains to be determined what role the protease cleavage has in megacin A-216 activity. Preliminary observations suggest that the specific activity of megacin purified from B. megaterium 216 is higher than that of megacin purified from B. megaterium THT(pMGC3), suggesting that the specific protease cleavage leads to an increase in specific activity.

In the C-terminal third of P293A, approximately corresponding to the β chain, a PLA2 domain was predicted by conserved domain search. A PLA2 catalytic motif, CCCnHDkC, was recognized by PROSITE search. The predicted PLA2 domain, with the presence of all conserved catalytic residues, is in agreement with the PLA2 activity measured in megacin A-216 preparations (28, 39) and with observations showing that megacin A-216 interferes with membrane integrity (13, 19, 28). A calcium-binding motif characteristic for Ca-dependent phospholipases or a signal sequence characteristic for secreted phospholipases has not been detected in the megacin A-216 sequence.

BLAST searches of nonredundant databases of all organisms yielded proteins with low-level similarities to megacin A-216. Most of them were phospholipases or hypothetical proteins, and the similarities were mostly restricted to the C-terminal region of P293A, where the PLA2 domain is located.

This is the first case, to our knowledge, in which the sequence of a bacterial PLA2 with antibacterial activity has become known. It is known that some eukaryotic PLA2 enzymes display bactericidal activity. Also, there are bacterial PLA2 proteins that have a role in virulence (prophage-associated Sla protein in Streptococcus pyogenes strains and ExoU in Pseudomonas strains) (see the references in reference 35). In bacteria, only a few proteins with PLA2 activity have been found so far. Recently, many hypothetical protein sequences in bacterial genome projects have been predicted to contain the PLA2 domain. P293A gave the highest BLAST scores with such hypothetical proteins and with the group A Streptococcus prophage-associated PLA2 proteins.

The hypothetical bacterial proteins with the best BLAST scores share the general organization of megacin A-216: the C-terminal part containing the predicted PLA2 domain is preceded by a longer section rich in acidic residues. One of them is the 248-residue-long hypothetical protein BT9727-0864 of the Bacillus thuringiensis serovar konkukian 97-27 strain. Interestingly, in this case, the similarity extends to several proteins encoded by the megacin region (Fig. 7). Six other predicted proteins encoded by the same, approximately 5-kb region of the B. thuringiensis genome share sequence similarity with respective proteins encoded by the B. megaterium megacin region. In addition, the region contains a tRNA^Cys gene which has 94% identity with the predicted tRNA^Cys gene in the megacin region. However, the arrangements of these genes are different in B. megaterium and in B. thuringiensis. Comparison of the two genomic regions suggests that this group of genes probably moved via lateral transfer between the two species and later evolved by recombinational events.

FIG. 7.

Comparison of the megacin region of B. megaterium 216 and a region of the B. thuringiensis serovar konkukian strain 97-27 genome (NC_005957). ORFs with predicted protein products, which in a pairwise comparison share amino acid sequence similarity between the two strains, are shown by filled arrows. Corresponding ORF pairs are indicated by the same number below the arrows. Numbers in parentheses refer to the length of the protein in amino acids. RefSeq gi numbers for the B. thuringiensis hypothetical proteins are 49477011 through 49477019 and apply to the ORFs on the map from left to right. The genomic location of the B. thuringiensis tRNA₂^Cys gene is 969811 to 969884.

Deletion mapping of the cloned fragment also identified the gene responsible for megacin immunity. This gene (ORF91 [megA^imm]), located adjacently to the megA gene, encodes a short, 91-amino-acid protein. The viability of cells harboring pMGC24, which has the megacin gene (ORF293A) but lacks the immunity gene (ORF91), is somewhat surprising and probably indicates the very low level of megacin expression in the absence of ORF188 and ORF73 (see below). Previous studies found in B. megaterium 216 a specific inhibitor of megacin A-216, which inhibited both the bacteriocidic action and the phospholipase activity and which was suggested to mediate immunity to megacin A-216 (26). Although Ochi et al. did not determine whether the inhibitory substance was a protein, we assume that it was identical to P91. The sequence of the ribosomal binding site preceding ORF91 is more similar to the consensus than that of ORF293A (Table 2), suggesting that the immunity protein is expressed at higher level than megacin A-216.

Although the original goal of this work was cloning and characterization of the megA and megA^imm genes, deletion mapping revealed that at least two of three other genes located on the cloned 5.5-kb fragment are also essential for megacin production (Fig. 1 and 2). Of the three adjacent genes, the role of ORF188 was unambiguously demonstrated. It remains to be determined which of the two other genes (ORF73 or tRNA₂^Cys) is required for efficient expression and induction of megacin A-216. ORF188 and ORF73 are preceded by good ribosomal binding sites (especially ORF73 [Table 2]), suggesting that both encode proteins. The predicted P188 protein shares very high sequence similarity with some sigma factors of gram-positive bacteria; thus, its role might be the modification of RNA polymerase initiation specificity, leading to transcription of the megacin gene. The weak similarity detected between P73 and a few, mostly hypothetical, bacterial proteins does not allow a similar prediction for its function. Although ORF73 is more likely to be required for megacin expression than the tRNA₂^Cys gene, the role of the latter gene cannot be excluded, given the high Cys content (14/293) of the megacin A-216 protein. In this context, it is worth mentioning that the subtle phenotypic difference observed between pMGC10, which usually produced some growth inhibition around the colony, and pMGC29, which always tested negative, suggests that the glutaredoxin-like P85 might also have some role in megacin A-216 expression. It is tempting to speculate that this effect relates to the putative disulfide reductase activity of P85 and the large number of cysteines in the megacin protein.

A sequence (AAAAAAAGTTGCCATTTGGTCAACTTTTTTTCTTTT) resembling rho-independent transcriptional terminators (7) was found downstream of ORF73, between positions 4191 and 4226. The lack of a similar sequence between ORF188 and ORF73 suggests that the two genes might form an operon. Studies focusing on the genetic background of megacinogeny and on the mechanism of megacin action and immunity are under way in our laboratories.