INTRODUCTION

Top Abstract Introduction Materials & Methods Results & Discussion References

Microcin E492 is a low-molecular-weight channel-forming bacteriocin (21) produced by Klebsiella pneumoniae that is active on strains of the family Enterobacteriaceae (7, 8). All of the other low-molecular-weight channel-forming bacteriocins that have been described are from gram-positive bacteria, among them the lantibiotic nisin and lactococcins A and B, which are nonlantibiotic heat-stable bacteriocins (reviewed in references 1, 19, and 33). Recently, there has been a renewed interest in the study of the mechanism of action of different bacteriocins and the effect of chemical modifications because understanding the strategies used by bacteriocins to kill specific bacteria may be useful in the design of antibiotics for pharmaceutical use (3). In order to better understand the mechanism of action, export, processing, and possible posttranslational modification of microcin E492, the genes needed for immunity protein and microcin production, which encompass a 13-kb DNA fragment, were cloned and expressed in Escherichia coli (35). Studies of the genetic determinants of all microcin systems, with the exception of microcins H47 and E492, have shown that these determinants are encoded on plasmids encompassing DNA segments of less than 6 kb. Microcins H47 (15) and E492 (35), however, are both chromosomally encoded on DNA fragments of more than 10 kb. Preliminary studies of random transposition mutagenesis of the cloned microcin E492 system suggest that most of the 13-kb fragment is needed for the production of active microcin, indicating that this system may involve more components for the production of active microcin than other systems already described. In previous work, we demonstrated that the microcin E492 immunity protein is located inside a 3-kb ClaI DNA fragment (35). Here we report that, in addition to the immunity protein gene, the microcin structural gene is also encoded in this region. The deduced amino acids sequence indicates that no similarity exits between this sequence and that of any known channel-forming bacteriocin, and thus, it appears that this microcin defines a novel class of low-molecular-weight pore-forming bacteriocin which is not related to colicin-like channel-forming bacteriocins from gram-negative microorganisms or to channel-forming bacteriocins from gram-positive bacteria. On the other hand, it was found that microcin E492 shares some characteristic cleavage site features with lactococcins and related bacteriocins found in gram-positive bacteria, as in the case of colicin V (12, 17). Insight into how the immunity protein- and microcin-encoding genes are organized will enrich our understanding of the coordinated regulation of these genes during stationary phase (35).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results & Discussion References

Bacterial strains and plasmids. The strains, vectors, and recombinant plasmids used in this work are described in Table 1. E. coli K38pGP1-2 and pT7-7 were kindly provided by S. Tabor (Harvard Medical School).

Microcin E492 assay on plates. To determine the capacity of the strains used to produce microcin E492, 2 × 10⁷ sensitive E. coli BL21(DE3)/p112 bacteria were overlaid onto Luria broth-(LB)-ampicillin (100 µg/ml) plates with 3 ml of soft agar. Single colonies to be checked were transferred with a toothpick onto the seeded plates. After overnight incubation, the colonies producing microcin E492 were surrounded by clear growth inhibition zones.

Microcin immunity test. Aliquots (50 to 300 µl) of cultures to be tested for immunity were grown in LB-ampicillin, mixed with 3 ml of top agar, and overlaid onto LB plates. Five-microliter volumes of serial dilutions of purified microcin were tested (23). Immune cells did not show any inhibition halo. The purified microcin used for this test was prepared as described by de Lorenzo (7).

DNA isolation and manipulation. Standard techniques were used for the cloning, transformation, preparation, and restriction analysis of plasmid DNA from E. coli (2, 28).

Nucleotide sequencing and analysis. Both strands were entirely sequenced from plasmid pBSC-47 (Bluescript vector) by using automated sequencing technology from ABI, in accordance with the manufacturer's directions, at Chiron Corporation (Emeryville, Calif.). Sequence analysis was performed with the University of Wisconsin Genetics Computer Group package, version 9.1 (10), at CEM, Facultad de Ciencias, Universidad de Chile. Sequence comparisons were performed with the TBLASTN, BLASTN, FASTA, and GAP programs.

Expression and localization of immunity protein MceB. The mceB gene was amplified by PCR from pBSC-47. The forward primer (5'CGGATAAAAC*ATATGACATTACTTTCATTTGG3') generated a NdeI restriction site at the mceB initiation codon (underlined). The reverse primer (5'AAAGCAAGAA*TTC*AGTCCTTTTGACTAATTTC3') was designated to be 15 bp downstream of the mceB stop codon and contained an EcoRI site (underlined). Nucleotides with asterisks indicate the changes made to create the restriction sites. The resulting 0.3-kb product was inserted into the corresponding sites of expression vector pT7-7 to give p157 with the correct in-frame reading phase to produce the 95-amino-acid product. The recombinant vector was introduced into E. coli K38pGP1-2, grown to mid-exponential phase in LB plus kanamycin-ampicillin, transferred to M9 minimal medium (24) minus methionine, and incubated for 30 min. The immunity protein was induced by shifting the temperature to 42°C for 20 min. Rifampin was added to a final concentration of 200 µg/ml, and the culture was incubated further for 10 min at 42°C. [³⁵S]methionine (50 µCi; ICN Translabel) was then added, and the incubation was continued for 15 min. The bacterial cells were collected by centrifugation, washed once with TEN buffer (50 mM Tris-HCl [pH 8.0], 1 mM EDTA, 100 mM NaCl), and immediately processed.

N-terminal microsequencing. Microcin E492 was transferred from polyacrylamide gels onto a polyvinylidene difluoride sequencing membrane (Bio-Rad) for N-terminal sequencing by following the manufacturer's instructions. The portion of the polyvinylidene difluoride membrane containing the microcin band (without staining) was cut out and microsequenced. N-terminal Edman microsequencing was done with an Applied Biosystems 477A sequencer equipped for on-line high-performance liquid chromatography analysis of the phenylthiohydantoin derivatives at the Protein Chemistry Service, Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Científicas, Madrid, Spain.

Nucleotide sequence accession number. The nucleotide sequence reported here has been submitted to the GenBank database under accession no. AF063590.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials & Methods Results & Discussion References

Cloning and sequence analysis of mceA and mceB genes. In a previous study, we used deletion analysis to demonstrate that the genetic determinant(s) required for immunity to microcin E492 are encoded in a 3-kb ClaI fragment (35). This fragment, obtained from pJAM434 (35), was cloned into the ClaI site of pBluescript (pBSC-47). This DNA segment could be cloned only in one orientation and conferred immunity to microcin E492. The DNA sequence between the EcoNI and BamHI sites of the cloned ClaI fragment is shown in Fig. 1. Two overlapping open reading frames (ORFs) were identified. The first, designated mceB, encoded a 95-amino-acid polypeptide with a theoretical molecular weight of 10,943 and corresponds to the microcin E492 immunity protein. The second, mceA, encoded a polypeptide of either 103 amino acids with a theoretical molecular weight of 10,091 if the first methionine is used as the start codon or a 99-amino-acid polypeptide with a theoretical molecular weight of 9,562 if the methionine at position 5 is used as the start codon. One of these polypeptide (or both) would correspond to nonprocessed microcin E492. The assignment of these genes (see below) was done experimentally. A putative promoter region with characteristic features, such as AT-rich stretches and two likely 10 (AATAAT) and 35 (TTGACG) regions with 17 bp between them, was identified in the 5' region. The first ORF, corresponding to mceB, has a potential ribosome binding site with a Shine-Dalgarno sequence (AACGGAT) 7 bp upstream of the ATG initiation codon. The two genes (mceA and -B) overlap by either 23 or 11 nucleotides (depending on the start codon), with a +1 shift for the mceA gene. No independent promoter or ribosome binding site for mceA could be identified if the translation starts in the first methionine, but if translation initiation occurs at the methionine in position 5, a good consensus for a Shine-Dalgarno sequence (TGGAGC) 5 bp upstream of this methionine can be found. Thus, it is more likely that translation initiation occurs at the methionine in position 5. It is possible that these genes are transcriptionally and translationally coupled, and this is consistent with the in vivo studies reported by Wilkens et al. (35) showing that microcin and the immunity protein are expressed only during the exponential phase of growth. An exhaustive analysis of the DNA sequence upstream of the putative promoter by using the computer programs FASTA and BLASTN was performed to search for sequence similarities. No obvious known regulation sites were found, although some similarities to E. coli and Bacillus subtilis genomic sequences with unknown functions were observed. However, it is interesting that there are two direct repeats that overlap the putative 35 region indicated between arrowheads in Fig. 1.

View larger version (41K): [in this window] [in a new window]   FIG. 1.   Nucleotide sequence of the mceAB-containing region. The deduced amino acid sequences for the mceA (microcin E492) and mceB (immunity protein) genes from K. pneumoniae RYC492 are given below the DNA sequences. The putative transcriptional start is indicated by the curved arrow. Deduced 10 and 35 regions in DNA are in boldface, and the proposed consensus Shine-Dalgarno sequences are underlined. Two direct repeats are indicated between the arrowheads. The inferred cleavage site in MceA is represented by a vertical arrow between amino acids 19 and 20. The amino acids in boldface are those confirmed by N-terminal sequencing of the protein.

The genes encoding microcin E492 and its immunity protein are homologous to those of the microcin 24 system. Analysis of the sequences of mceA and mceB by using the computer programs FASTA and TBLASTN revealed that the two corresponding ORFs belong to a novel family of bacteriocins and that each presents significant similarities to only one protein in the GenBank database. mceA turned out to be homologous to mtfS, the structural gene of microcin 24, isolated from an E. coli uropathogenic strain (25), while mceB is a homolog of mtfI, a gene encoding the microcin 24 immunity protein (25). Information concerning microcin 24 is only available in GenBank. The deduced amino acid sequences were compared by using the GAP program (Fig. 2). The MtfS and MceA proteins are 90 and 103 amino acids (for comparison, we will use the sequence that starts at the first methionine), respectively, with an identity of 52% and a similarity of 59%. On the other hand, MtfI and MceB are very similar in size (93 and 95 amino acids, respectively), with an identity of 39% and a similarity of 56%. These results strongly suggest that both microcins belong to the same family and probably have an ancestor in common. No information is available about the mechanism of action of microcin 24, but due to its similarity to microcin E492, it is possible to speculate that the target of microcin 24 is probably also the cytoplasmic membrane. The five determinants needed for the production of microcin 24 and its immunity protein are encoded in a 5.3-kb DNA segment (25), while the microcin E492 system is encoded in a 13-kb segment. The difference in size may reflect the complexity of the microcin E492 system, an idea supported by preliminary results with random Tn5 mutagenesis that suggest that between six and eight cistrons (besides the structural gene) encoded in the 13-kb DNA fragment are needed for the production of active microcin (4). In this respect, microcin E492 is also different from colicin V, which mechanism of action at the inner membrane level could be related to microcin E492. Only four plasmid genes, spanning 4.5 kb of DNA, are required for colicin V synthesis, export, and immunity (14, 19). The similarity between the sequences of the microcin protein and the immunity protein of microcins 24 and E492 that suggests that these proteins have an ancestor in common is also present at the level of gene organization. In both cases, the immunity protein- and microcin-encoding genes seem to be under the control of the same promoter, and the C-terminal coding region of the immunity protein overlaps the N terminus of microcin, with a frameshift in the ORF of the microcin protein. As with the transcription of the two converging operons in colicin V, transcription of the mtfI/mtfS operon of microcin 24 also could be under the control of the Fur protein, which binds to DNA at a specific sequence called the Fur box (9), because a sequence similar to a Fur box was found in the mtfI/mtfS promoter region (25). Although microcin E492 expression is regulated by the presence of iron in the growth medium (26), no site for the Fur repressor was found in an extensive region upstream (240 bp) of the mceB gene.

View larger version (33K): [in this window] [in a new window]   FIG. 2.   Comparison of the deduced amino acid sequences of nonprocessed microcin MceA with MtfS (A) and immunity protein MceB with MtfI (B). Identical matches (vertical lines) and strongly (colons) and weakly (dots) conserved amino acid residues are indicated. The comparison was performed with the GAP program. MceA and MtfS displayed an identity of 52% and a similarity of 59%, while MceB and MtfI presented an identity of 39% and a similarity of 56%.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Hydrophobicity profiles of MceA (A) and MceB (B). Hydropathy was calculated by the method of Kyte and Doolittle (20) with a span length of 13 amino acid residues and by using a linear weight variation model. The dashed line in panel A shows the inferred processing site for microcin E492.

N-terminal amino acid sequencing of microcin E492. The amino acid sequence determined by microsequencing of the N-terminal region of microcin E492 cloned in E. coli was X-X-Thr-Asp-Pro-Asn-Thr-Glu-Leu-Leu-Asn-Asp-Leu. The two first amino acids could not be identified due to interference, so it was not until cycle 3 that the amino acids were clearly identified. The same interference was observed with microcin E492 from K. pneumoniae purified by high-performance liquid chromatography (34). This result indicates that the primary translation product of mceA is a precursor that is processed to remove the first 15 or 19 amino acids. Thus, the processed microcin E492 molecule would be an 84-amino-acid polypeptide with a predicted molecular weight of 7,887.

Similarity between cleavage sites of microcin E492 and those of colicin V and lactococcins. Microcin E492 was found to share many cleavage site features with colicin V and lactococcins and related bacteriocins found in gram-positive bacteria. These features (12, 17) include the presence of Gly residues in position 2 or 1 relative to the processing site. Most of the sequences contain the double-glycine motif, with the exception of lactococcin DR, which contains a glycine-alanine motif (17). Analysis of several leader peptides of the double-glycine motif showed that seven amino acids are found to be conserved in more than half of the aligned leader peptides (17): glycine in position 1, glycine in position 2, isoleucine in position 4, leucine in position 7, glutamic acid in position 8, serine in position 11, and leucine in position 12. From these amino acids, microcin E492 presents glycine in position 2, leucine in position 7, and serine in position 11. The aspartic acid in position 8 is a conservative amino acid substitution for glutamic acid, and isoleucine in position 12 is a conservative amino acid substitution for leucine. Other similarities were found among the cleavage sites of microcin E492, colicin V, and lactococcins and related bacteriocins. The first is a cleaved leader sequence of 15 to 20 amino acids (microcin E492 has an either 15- or 19-amino-acid leader peptide). The second is a predicted -helical structure over most of the leader sequence. The prediction was made by using the Chou-and-Fasman (5) and Levitt (22) methods (data not shown). This analysis suggested that leader sequence residues 1 to 16 are mostly -helical and are followed by a -turn that begins at amino acid 17 (data not shown). It is necessary to point out that these studies were done only for comparison, as these methods are frequently used when comparing these properties of bacteriocins. The third is a predicted -turn that begins one to three amino acids before the cleavage site (amino acid 3, in this case), which would expose the cleavage site to allow cutting by the specific peptidase. The fourth is the fact that the processed products, in all of these cases, are small, heat-stable protein bacteriocins (43 to 88 amino acids in length) which function by increasing membrane permeability in target cells.

Genetic identification and localization of immunity protein MceB. To demonstrate unequivocally that mceB is the gene that encodes a protein that confers immunity to microcin E492, the DNA fragment containing only the ORF coding for the 95-amino-acid immunity protein was cloned into an expression vector. For this purpose, the pT7-7 expression vector under the control of T7 RNA polymerase was employed. The DNA segment coding only for the immunity protein was PCR amplified and cloned into pT7-7 (see Materials and Methods). Cells transformed with this recombinant vector (p157) acquired immunity to microcin E492, in contrast with the control transformed with pT7-7, which was not immune to microcin, demonstrating that this gene is enough to confer immunity to a high-titer microcin suspension. This construct was first introduced in E. coli XL1-Blue, in order to screen for the cloned immunity gene in a background where no expression was expected, but surprisingly, it was found that these cells acquired immunity to microcin E492. Although expression under the control of a T7 promoter is thought to be completely restricted to cells that express T7 RNA polymerase, there are reports (30) demonstrating that it is possible to find basal T7 RNA polymerase-independent expression of genes situated downstream from a T7 promoter. This very low level of expression would be enough to confer immunity to microcin E492. The pT7 system allows specific radioactive labeling of the protein under the control of the T7 RNA polymerase promoter (32; see Materials and Methods), and we took advantage of this characteristic to determine the localization of the immunity protein. For these experiments, the host strain E. coli BL21(DE3) was not suitable, because these cells segregated the immunity character. This strain does not tightly regulate the expression of T7 RNA polymerase (which is under the control of the lac promoter), resulting in a basal expression of 20 to 30% with respect to that of the fully induced protein. Overexpression of this protein seems to be harmful for the host cells, which, in turn, develop a mechanism to bypass its expression. This interpretation is consistent with the fact that the 3-kb ClaI DNA fragment could be cloned in pBluescript in one only orientation, in which the immunity and microcin genes are opposite to the lac promoter, probably because the cells could not tolerate high-level expression of these proteins. For this reason, we used the tightly regulated system of E. coli K38pGP1-2, in which the T7 RNA polymerase gene is under the control of promoter P_L, and thus permitted induction by a temperature shift just before protein labeling (32). E. coli K38pGP1-2 harboring p157 was immune to a microcin suspension with the highest titer available (10⁶ U/ml). Figure 4A shows SDS-PAGE of the soluble and particulate fractions of a bacterial extract in which the immunity protein has been induced and the corresponding autoradiograph (Fig. 4B). This protein was clearly identified in the autoradiograph as the only labeled protein, and as expected, it was located in the particulate fraction, which contains the inner and outer membrane fractions.

View larger version (85K): [in this window] [in a new window]   FIG. 4.   Expression of the immunity protein from p157. (A) Expression from the T7 promoter of pT7-7 and p157 was assessed by using E. coli K-38pGP1-2 as the bacterial host. Induction of T7 RNA polymerase was performed by incubation at 42°C. Specific transcription from the T7 promoter was performed by addition of rifampin, and proteins were labeled with [35S]methionine. SDS-16% PAGE samples corresponding to E. coli K-38pGP1-2/pT7-7 are in lanes 1, 2, and 3, and those corresponding to E. coli K-38pGP1-2/p157 are in lanes 4, 5, and 6. Total (lanes 1 and 4), soluble (2 and 5), and particulate (3 and 6) fractions of the bacterial extracts were analyzed separately. The arrow indicates the band corresponding to the immunity protein. Molecular weight markers in lane 7 were myoglobin fragments with Mrs of 16,950, 14,440, 8,160, 6,210, and 3,460. (B) Autoradiography of the SDS-PAGE shown in panel A.

View larger version (130K): [in this window] [in a new window]   FIG. 5.   Localization of the immunity protein in the inner membrane. Induction of the T7 RNA polymerase of E. coli K-38pGP1-2 harboring pT7-7 or p157 was performed by incubation at 42°C for 5 h, and the separation of inner and outer membrane fractions was carried out as described in Materials and Methods. SDS-14% PAGE samples corresponding to E. coli K-38pGP1-2/pT7-7 are in lanes 1, 2, and 3, and those corresponding to E. coli K-38pGP1-2/p157 are in lanes 5, 6, and 7. Total (lanes 1 and 5), inner membrane (2 and 6), and outer membrane (3 and 7) fractions of the bacterial extracts were analyzed separately. The arrowhead indicates the band corresponding to the immunity protein. Molecular weight markers in lane 4 were myoglobin fragments with Mrs of 16,950, 14,440, 8,160, 6,210, and 3,460.