Genetic Analysis and Complete Primary Structure of Microcin L

Abstract

Escherichia coli LR05, in addition to producing MccB17, J25, and D93, secretes microcin L, a newly discovered microcin that exhibits strong antibacterial activity against related Enterobacteriaceae, including Salmonella enterica serovars Typhimurium and Enteritidis. Microcin L was purified using a two-step procedure including solid-phase extraction and reverse-phase C₁₈ high-performance liquid chromatography. A 4,901-bp region of the DNA plasmid of E. coli LR05 was sequenced revealing that the microcin L cluster consists of four genes, mclC, mclI, mclA, and mclB. The structural gene mclC encoded a 105-amino-acid precursor with a 15-amino-acid N-terminal extension ending with a Gly-Ala motif upstream of the cleavage site. This motif is typical of the class II microcins and other gram-positive bacteriocins exported by ABC transporters. The mclI immunity gene was identified upstream of the mclC gene and encodes a 51-amino-acid protein with two potential transmembrane domains. Located on the reverse strand, two genes, mclA and mclB, encoded the proteins MclA and MclB, respectively. They bear strong relatedness with the ABC transporter proteins and accessory factors involved in the secretion of microcins H47, V, E492, and 24. The microcin L genetic system resembles the genetic organization of MccV. Furthermore the MccL primary structure has been determined. It is a 90-amino-acid peptide of 8,884 Da with two disulfide bridges. The N-terminal region has significant homologies with several gram-positive bacteriocins. The C-terminal 32-amino-acid sequence is 87.5% identical to that of MccV. Together, these results strongly indicate that microcin L is a gram-negative class II microcin.

As bacterial resistance to currently used antibiotics is increasing, new pathogenic agents are discovered, and traditional bacterial diseases reappear, increased efforts to search for new antibiotics are needed. Over the past 40 years, research has been restricted largely to improving those well-known compound classes that are active against a standard set of drug targets. As no new classes of antibiotic have been discovered, such insufficient chemical variability exists that there is a potential for serious escalation in clinical microbe resistance. Numerous living organisms are able to produce a variety of ribosomally synthesized antibacterial peptides or proteins involved in their innate defense against microorganisms. During the past 15 years, these compounds have attracted considerable attention, offering many exciting possibilities for the future of antibiotics, in the face of current declining efficacy of conventional treatment (20, 22). Of the bacteriocins produced by bacteria, many direct activity against pathogens and, in particular, food-borne microorganisms, such as the gram-positive bacterium Listeria monocytogenes (7, 10) and gram-negative bacteria Salmonella enterica and Escherichia coli (34, 40).

Microcins are secreted by members of the Enterobacteriaceae family, in particular strains of E. coli. They constitute a class of low-molecular-mass peptides (<10 kDa) that exhibit a narrow antimicrobial spectrum of activity directed against bacterial species phylogenetically related to the producing strains (33). To protect itself, the microcin-producing bacterium exhibits immunity to the action of its own microcin. Recent developments in the biochemical characterization and mode of action allowed us to propose a classification of these peptides into two classes (17, 36). Class I, which includes to date microcins B17, C7, D93, and J25, encompasses peptides with molecular mass below 5 kDa that are highly posttranslationally modified. These microcins display a range of unrelated chemical structures, which in turn results in a variety of action mechanisms (9). Class II includes microcins E492, H47, V, most likely microcin 24, and now microcin L. This second group is more homogeneous and shares several common structural properties with class IIa gram-positive bacteriocins: size ranging from 7 to 10 kDa, absence of modified amino acids, and presence of a consensus motif. Additionally, they are synthesized as precursor peptides with a double-glycine type leader peptide and their secretion is mediated by ATP binding cassette (ABC) transporters. Furthermore, several studies have indicated the cytoplasmic membrane as a target for the action of MccV (47) and MccE492 (8, 31). It is noteworthy that despite similarities between gram-positive bacteriocins and class II microcins, there is homology neither in amino acid sequence nor in immunity proteins. Moreover, the common cationic character of gram-positive bacteriocins is lacking in most of these microcins (36).

Class II microcins identified so far are synthesized as precursors and contain an N-terminal leader sequence which is removed upon externalization of the mature microcin. The synthesis, immunity, and export of these microcins rely upon a minimal genetic structure, generally consisting of four clustered genes grouped in one or two operons (33). Such a microcin gene cluster is composed of one structural gene, that encodes the precursor peptide and a dedicated immunity gene that is usually located next to and expressed with the microcin structural gene. Furthermore, the cluster includes two genetically linked genes that encode a large protein belonging to the ABC superfamily of transporters and a membrane fusion protein (MFP) that resides in the inner membrane. Both of these proteins are required for the efficient processing and secretion of the microcin. The N-terminal leader sequence of the microcin precursor serves as a recognition signal for the processing of the mature microcin by cleavage of the leader peptide by the proteolytic domain of the transporter protein (23, 24). Thus, MccV, E492, and H47 act independently of the general secretion pathway and are secreted instead by this dedicated transport system. The involved transporter genes have been identified and sequenced for MccV (cvaA and cvaB), MccE492 (mceG and mceH), MccH47 (mchE and mchF), and Mcc24 (mtfA and mtfB), and all display high homologies. The presence of an outer membrane export protein such as TolC encoded by a chromosomal gene would also be required for the secretion of these microcins (1, 12, 19, 26, 29). Additional genes may be present in the cluster encoding an array of proteins that could participate in the processing and maturation of active microcin (3, 30).

Upon screening poultry intestinal bacterial isolates displaying anti-Salmonella activity, we recently identified a microcinogenic strain, E. coli LR05, that possesses the hitherto unseen ability to produce four microcins, MccJ25, MccB17, MccD93, and MccL. This new microcin has been purified and biochemically characterized. Composition analysis originally estimated the peptide to be 90 amino acids in length, and protein sequencing identified the first 40 amino acids at the N terminus of the active microcin (17).

We report here the purification of microcin L by a novel and effective method and the determination of its complete amino acid sequence. We describe the nucleic acid sequence of the gene encoding microcin L (mclC) and of the surrounding regions. We demonstrate that mclI is devoted to the immunity function. The export function is assigned to the products of two genes located in the cluster, mclA and mclB. We propose that MccL is secreted by an ABC exporter constituted by MclA and MclB and the outer membrane protein TolC.

MATERIALS AND METHODS

Bacterial strains and media.

The bacterial strains and plasmids used in this study are listed in Table 1. Chemicals were purchased from Sigma. Unless otherwise stated, all strains were grown in brain heart infusion (Difco) at 37°C and maintained on nutrient agar. Minimal medium M63 was prepared as described by Miller (32) and supplemented with glucose (0.2%, wt/vol) and thiamine (0.01%, wt/vol). Soft and normal solid media prepared by adding, respectively, 0.6 or 1.2% (wt/vol) type E agar (Biokar) were used for the activity and immunity assays. Ampicillin, kanamycin, chloramphenicol, isopropyl-β-d-thiogalactopyranoside (IPTG), and 5-bromo-4-chloro-3-indolyl-β-d-galactoside (X-Gal) were used at concentrations of 100, 30, 30, 50, and 50 μg ml⁻¹, respectively.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristicsa	Reference or source
E. coli strains
LR05	Wild-type producer strain of MccL; MccL+ MccB17+ MccJ25+ MccD93+ ImmL+ ImmB17+ ImmJ25+ ImmD93+ ImmV+	17
LRM411	E. coli MC4100 pL102 MccL+ ImmL+ Apr	This study
TG1	F−traD36 lacI Δ(lacZ) M15 proA+B+/supE thi Δ(lac-proAB) Δ(hsdM-mcrB)5(rK− mK− McrBC−)	4
MC4100	F−araD139 Δ(org F−lac) rpsL150(Srr) relA1	5
MC4100 tolC::Tn5	Kmr transductant of MC4100 infected with P1 grown on GC7442	5
Plasmids
pL102	pUC19 carrying the MccL genetic system; Apr	41
pHK11	pBr322 carrying the ColV(MccV) genetic system; Apr	18
pEX4	MccH47+ ImmH47+ Apr	38
pJAM229	Cosmid pHC79 carrying the MccE492 genetic system; Apr	46
pGOB18	pBr 322 carrying the Mcc24 genetic system; Apr	35
pAX629	pACYC184 derivative carrying tolC; Cmr	25

^aAp^r, ampicillin resistant; Cm^r, chloramphenicol resistant; Km^r, kanamycin resistant.

DNA manipulations.

Plasmid DNA was isolated from E. coli using a Qiagen plasmid midi kit (Qiagen). Restriction enzymes were purchased from Promega. Competent E. coli cells were prepared and transformed by electroporation in 0.2-cm-diameter cuvettes in a Gene Pulser apparatus (settings: 25 μF, 2.5 kV, 200 Ω; Bio-Rad). PCR amplifications were performed in a final volume of 25 μl using 50 ng of plasmid DNA. Reaction mixtures contained an 0.8 μM concentration of each primer, 1.5 mM MgCl₂, a 200 μM concentration of each deoxynucleoside triphosphate, and 1 U of Taq polymerase (Promega) in the corresponding buffer. Denaturation at 94°C for 2.5 min was followed by 30 cycles of amplification (30 s at 92°C, 30 s at the annealing temperature, and from 30 s to 3.5 min at 72°C) and elongation for 2 min at 72°C. PCR products and DNA fragments were purified from agarose gels using the GencleanII kit (Bio 101, Quantum Biotechnologies).

DNA sequencing and analysis.

DNA sequencing was performed using Big Dye terminator sequencing kits. The sequencing reactions were loaded on a model 310 automated DNA sequencer (Applied Biosystems). Both strands were sequenced. DNA or protein homology searches (GenBank, EMBL, and Swiss-Prot) and sequence analysis were performed with the programs of the Genetics Computer Group sequence analysis software package (University of Wisconsin).

Plasmid constructions.

A number of strategies were used to clone and to analyze the genes involved in microcin L production, immunity, and export. The pL102 plasmid (Table 1) was used as a template for all the experiments. Restriction sites were added at the 5′ end of a number of primers to give the possibility to clone PCR fragments using the corresponding restriction enzymes. Sequences of the primers can be obtained from the authors. Open reading frame 3 (ORF3) and ORF4 were cloned to investigate their role in microcin L production and immunity. A 379-bp region containing ORF3 was amplified and cloned into pGEM-T easy (Promega), generating pL1. Plasmid pL2 was constructed by cloning into pUC19 (Stratagene) a DNA fragment of 686 bp (ORFs 3 and 4). All these constructs were checked by sequencing. The recombinant plasmids were introduced into E. coli TG1 strains. We tried to clone a DNA fragment carrying ORF5 and ORF6 into pGEM-T easy or into pUC19 vector after a SalI/SphI digestion. All the experiments were unsuccessful.

Microcin activity and immunity.

Microcin production was tested by patch test, which consisted of picking each strain under study with a sterile toothpick and stabbing it onto a lawn of MccL-sensitive cells. After incubation of the plates (37°C), halos of growth inhibition appeared around MccL-producing stabs. Microcin activity of the culture supernatant or purified peptide was assayed by the agar well diffusion method (40). For quantification of sample activity, twofold dilutions were tested. The reciprocal of the highest dilution that produced a clear zone of growth inhibition around the well was defined as the number of arbitrary units (AU) of microcin activity. Unless otherwise stated, E. coli MC4100 was used as the standard indicator strain. Microcin immunity was also assayed by patch test or by the cross-streaking assay as described previously (37).

Determination of MIC and MBC.

The MIC was determined by the standard macrodilution broth method as described by Sahm and Washington (42). We checked that the highest concentration of acetonitrile in broth was not affecting the growth of tested bacteria. The minimum bactericidal concentration (MBC) was determined from tubes showing complete inhibition. A nutrient broth agar plate was seeded in surface with 0.1 ml from each clear tube and incubated (24 h at 37°C). The MBC was defined as the lowest concentration in the tubes giving no growth on nutrient broth plate afterwards.

Presence of intracellular MccL.

Intracellular MccL was determined as described by Gaggero et al. (16). Potential producing strains were grown in M63 broth until they reached an optical density of about 0.5 at 600 nm. A 20-ml aliquot of culture was centrifuged (14,000 × g), and the cell pellet was resuspended with 100 μl of STET solution (0.1 M NaCl, 10 mM Tris-HCl [pH 8.0], 1 mM EDTA [pH 8.0], 5% Triton X-100). After addition of lysozyme (2 mg ml⁻¹), the tubes were incubated on ice for 10 min and transferred to a boiling water bath for 1 min. After centrifugation, the supernatants were assessed as described above.

Purification of MccL.

Flasks (1,000 ml) containing 200 ml of M63 broth were seeded with a 1% inoculum from an overnight culture of E. coli LR05 and incubated at 37°C for 6 h with shaking. The culture was centrifuged (13,000 × g) and heated for 10 min at 100°C. The supernatant was loaded onto Sep-Pak Plus Environmental C₁₈ cartridges (Waters, St. Quentin en Yvelines, France). After washing with water, three stepwise elutions were performed with, successively, 50% methanol and 40 and 45% acetonitrile in water. The 45% Sep-Pak fractions were subjected to reverse-phase (RP) chromatography on a column equilibrated in water acidified with 0.1% trifluoroacetic acid (TFA). Separation was performed with a quadriphasic gradient of 0 to 30% acetonitrile in acidified water (0.1% TFA) over 10 min, 30 to 54% acetonitrile over 26 min, a step at 54% acetonitrile over 2 min, and 54 to 65% acetonitrile over 12 min, at a flow rate of 0.7 ml min⁻¹.

Enzymatic cleavage.

Due to the low amount of purified MccL available, only a few picomoles, our strategy did not use the classical reduction-alkylation step previously used in the enzymatic digestion. The MccL protein was digested with endoproteinase Asp-N (Roche Diagnostics, Meylan, France). The RP high-performance liquid chromatography (RP-HPLC)-purified MccL was dissolved in 10 μl of water and 10 μl of acetonitrile. Asp-N endoproteinase (2 μg diluted in 50 μl of H₂O) was then added to the solution and incubated for 1 min at 37°C. The reaction was stopped by acidification with 0.1% TFA. To prevent disulfide bridge position exchange, MccL solutions were kept at pH 7, digestion was performed at pH 7, and HPLC separation was carried out at pH 3.5.

LC-MS.

The peptides generated by protease hydrolysis were separated by RP-HPLC on a Pepmap column (C₁₈ [3 μm by 1 mm by 150 mm]; LC Packings, Amsterdam, The Netherlands). Separation was performed with a linear gradient of 2 to 60% acetonitrile (0.08% TFA) in water (0.1% TFA) over 60 min at a flow rate of 70 μl min⁻¹ at 35°C. The RP-HPLC columns were coupled to a VG BioQ mass spectrometer (Bio Tech, Manchester, United Kingdom) fitted with an electrospray source (upgraded by the manufacturer so that the source had Quattro II performances) operating at atmospheric pressure and a triple quadrupole analyzer. The potential applied between the end of the capillary and the first electrode was 3.2 kV. The solvent was evaporated using a nitrogen flow heated at 80°C. The calibration was performed using a solution (2 pmol μl⁻¹) of horse myoglobin. For the liquid chromatography-mass spectrometry (LC-MS) experiments, the eluate from the RP-HPLC column (70 μl min⁻¹) was split such that 20 μl min⁻¹ was introduced in the ESMS device for on-line LC-MS analysis and the remaining 50 μl min⁻¹ was collected for off-line MS analysis after UV detection. Peptides were detected at 214 nm by a Waters 996 Photodiode Array Detector and each fraction was collected manually in Eppendorf tubes, analyzed by MS (ESMS-QTOF II) and sequenced by Edman degradation.

NanoES MS.

Intact protein and fractions collected from previous LC-MS experiments were analyzed with a Micromass QTOF II mass spectrometer fitted with a z-spray nanoelectrospray (nanoES) source. A 2-μl aliquot of each fraction was loaded into a metal coated nano-spray capillary (Protana, Odense, Denmark), previously rinsed with 2 μl of H₂O-CH₃CN-HCOOH (40:40:20, vol/vol). The mass spectrometer was operated with a capillary voltage of 1,000 to 1,200 V, and the sampling cone was set at 30 to 40 V. The scan time which defines the time period of signal accumulation for one mass spectrum from multiple orthogonally accelerated ion packets into the QTOF analyzer was set at 1 s. Calibration of the mass analyzer was performed under the same conditions with either the multiply-charged ions of a solution (2 pmol μl⁻¹) of horse myoglobin or poly-dl-alanine.

Edman sequencing.

Amino acid sequence analysis of intact protein or peptides was determined by automated Edman degradation on an Applied Biosystems model 477 protein sequencer.

Nucleotide sequence accession number.

The nucleotide and protein sequence for the microcin L region has been assigned GenBank accession number AY237108.

RESULTS

DNA sequence analysis of the microcin L locus.

Using the degenerate primers GGYGAYGTDAACTGGGTDGA and RCCHACNGCRAANGCRCCNG as forward and reverse primer, respectively, a 120-bp PCR product was amplified from the plasmid DNA preparation of E. coli LR05. This product corresponded to the DNA region encoding the 40 N-terminal amino acids of the microcin L.

To examine the possibility of clustered genes, DNA sequences surrounding the 120-bp product were determined. Using a primer-walking strategy, 4,901 bp of the LR05 plasmid DNA preparation were sequenced. Computer analysis of the sequence identified a number of ORFs, some of which were renamed if functions could be assigned to their encoded products. A schematic view of these ORFs and a partial restriction map of the 4,901-bp fragment are presented in Fig. 1. A more detailed analysis of the genes described in this work is listed in Table 2.

FIG. 1.

Schematic overview of the genetic organization of plasmid pL102. The upper part of the diagram displays a partial restriction map of the 13.5-kb plasmid pL102. The plasmids used in this study are located below. Open arrows represent the genes whose functions were defined. Abbreviations for restriction enzymes: B, BanII; E, EcoRI; K, KpnI; H, HindIII; S, SalI.

TABLE 2.

Characteristics of ORFs of the MccL cluster and features of the encoded putative proteins^a

ORF	Position in nucleotide	Length (aa)	Molecular mass (kDa)	Gene name	pI (calculated)	Similar amino acid sequence (s)b (%)
ORF1	235-471	78	9.17	cvi	8.47	Immunity MccV protein (98)
ORF2	449-532	27	2.9	cvaC′	4.68	Peptide leader of MccV precursor (100)
ORF3	529-684	51	5.69	mclI	11.26	NS
ORF4	688-1005	105	10.55	mclC	4.40	NS
ORF5c	4555-3281	424	48.70	mclA	7.72	E. coli membrane fusion proteins MchE (100), CvaA (98), MceH (92), and MtfA (72)
ORF6c	3288-1192	698	78.26	mclB	7.94	E. coli ABC transporters MchF (92), CvaB (95), MceG (89), and MtfB (74)

^aAbbreviations: aa, amino acids; NS, no similarity.

^bMchE/F, CvaA/B, MceH/G, MtfA/B are, respectively, microcin H47, V, E492, and 24 dedicated export proteins.

^cEncoded on the complementary strand.

Analysis of the gene products encoded by the microcin L locus.

Based on amino acid sequencing results, the gene encoding the MccL precursor corresponded to ORF4 and so was named mclC (Fig. 1). A putative ribosome binding site sequence (AGGGG) was located 10 nucleotides upstream of the start codon, which corresponds to the optimal spacing determined by Vellanoweth and Rabinowith (44). Analysis of the upstream region of mclC revealed the presence of a putative promoter. Indeed, potential −35 (TTGTTA) and −10 (ATTAAT) boxes separated by 16 nucleotides were located 34 nucleotides upstream of the start codon. Furthermore, an inverted repeat located 107 bp downstream of mclC (ΔG value of −22.4 kcal mol⁻¹) determined according to the method of Tinoco et al. (43) may function as a Rho-independent transcription terminator. mclC spans 317 bp, starts with the usual start codon ATG, and encodes a protein that consists of 105 amino acid residues. Comparison of this sequence with the sequence of the mature microcin L (see below) confirmed that mclC is the microcin L structural gene. The microcin is synthesized as a precursor consisting of a 90-residue C-terminal peptide, matching the amino acids of the purified microcin L, and a 15-residue N-terminal extension exhibiting a double-glycine type motif, characteristic of many antibacterial peptides secreted through ABC exporters. Homology analysis reveals that the N-terminal leader peptide has significant similarity to the leader peptides of nonlantibiotics, some lantibiotics found in gram-positive bacteria, and class II microcins. An alignment of the amino acid sequences of MccV, E492, H47, 24, and L leader peptides is shown in Fig. 2.

FIG. 2.

Sequence alignment of class II microcin leader peptides. Conserved amino acids are boldfaced; semiconserved amino acids are shaded.

ORFs upstream of mclC.

ORF1, with the same polarity as mclC, encodes a polypeptide of 78 residues with a theoretical molecular mass of 9,174 Da. Comparison of its deduced sequence with those in databases revealed that the protein encoded is nearly identical to the cvi gene product, the immunity MccV protein, differing by a single amino acid at position 34. This finding correlates with the experimental finding that the wild-type E. coli LR05 is not sensitive to an MccV-producing strain (41). Thus, the substitution of residue 34 (Gly→Glu) does not modify the immunity protein activity. Located directly adjacent to ORF1, now designated cvi, is a short ORF, ORF2. Homology searches reveal that the corresponding 15 N-terminal amino acids are fully identical with the 15 residues of the leader peptide of the MccV precursor (CvaC). However, the nucleotide sequence is truncated, and surprisingly no homology with MccV is found in the 12 residues next the leader peptide. We designated ORF2 as cvaC′.

ORF3 is located upstream and directly adjacent to the mclC structural gene, separated by only five nucleotides. ORF3 starts with an ATG codon, located at an appropriate distance (5 bp) from a potential ribosome binding site sequence, AGGAA. Putative −35 (TGGAGG) and −10 (TATAAC) boxes separated by 12 nucleotides were identified 18 nucleotides upstream of the start codon. Unlike mclC, no inverted repeat was observed downstream of ORF3. The predicted protein consists of 51 amino acids and contains two hydrophobic regions spanning residues 6 to 34 and 39 to 47 related to two transmembrane domains. Comparison of the sequence with those in the databases did not reveal any significant homology with previously described sequences. The location of the ORF in the cluster as well as the two potential transmembrane regions makes it the best candidate for immunity gene.

ORFs downstream of mclC.

Two large ORFs, ORFs 5 and 6, were identified downstream of mclC, on the opposite strand (Fig. 1). Computer analysis based on the TTG initiation codon reveals that the product of ORF5 presents 100% identity with MchE (accession number Q9EXN6) and is highly homologous to the MFPs CvaA (98%), MceH (92%), and MtfA (72%). These proteins correspond to the MccH47, V, E492, and 24 systems, respectively. So we designated ORF5 as mclA. In addition to encoding the full-length 48.7-kDa MclA product, the mclA gene also could encode a smaller protein, MclA* (24.4 kDa), which lacks the N terminus of the larger protein. The MclA* protein is the product of an in-frame translation restart within mclA. Similarly, we found strong homologies of ORF6 with microcin ABC transporters, MchF (92%), CvaB (95%), MceG (89%), and MtfB (74%) of the MccH47, V, E492, and 24 systems, respectively. Consequently we named ORF6 mclB.

All these results strongly suggested that MccL is secreted through an ABC export system, comprising an ABC protein, MclB, and a membrane fusion protein, MclA, both of which display the ABC transporter family signatures.

Immunity is provided by mclI.

For all class II microcin-encoding gene clusters, the immunity gene has been characterized and located upstream of the microcin structural gene. Thus, ORF3 was a good immunity protein candidate. Plasmid pL1 carrying ORF3 was introduced into the MccL-sensitive E. coli plasmidless strain TG1. TG1(pL1) transformants display full immunity to microcin L (3,000 AU ml⁻¹) as well as the wild-type MccL producer strain E. coli LR05 and the recombinant E. coli LRM411. Therefore, cloning and expression of ORF3 in a microcin L-sensitive strain rescued immunity to a level comparable to that of the wild-type producer. These results clearly demonstrated that ORF3 was necessary and sufficient to express full immunity toward MccL. The antagonistic activity of MccV, H47, E492, and 24 recombinant microcin-producing strains carrying, respectively, pHK11, pEX4, pJAM229, and pGOB18 (Table 1) was tested against E. coli TG1(pL1). In any case TG1(pL1) was sensitive, indicating that the immunity protein is highly specific for MccL, as is common among microcins. Consequently, ORF3 is designated mclI. The sequence predicts a protein of 51 amino acids with a theoretical molecular mass of 5.69 kDa. It is cationic with a calculated pI of 11.26 and is largely hydrophobic with the exception of six central residues. Hydropathy analysis predicts that MclI has two transmembrane fragments. Similar characteristics have been described for MccV, H47, E492, and 24 immunity systems (Table 3). We assume that MclI is most likely an integral inner membrane protein.

TABLE 3.

Immunity proteins of class II microcins

Microcin system	Immunity gene	Immunity protein					Reference or source
		Size (aa)a	pI (calculated)	TMb	Asp+Glu	Arg+Lys
Mcc L	mclI	51	11.26	2	0	4	This study
Mcc V	cvi	78	8.84	2	8	11	18
Mcc H47	mchI	69	9.60	2	3	6	39
Mcc E492	mceB	95	9.72	3	2	9	30
Mcc 24	mtfI	93	9.10	3	3	8	GenBank accession no. U47048

^aaa, amino acid

^bTM, transmembrane domain.

Cloning both immunity and precursor genes (mclI and mclC).

Recombinant strain TG1(pL2) was tested for its activity and immunity. Full immunity was found; however, no extracellular microcin was produced. Considering the absence of supposed export genes mclA and mclB, this result was expected. Conversely, lysates from this strain contained antibacterial activity (data not shown) similarly to that observed for MccV (48) and MccH47 (38).

Cloning mclA and mclB genes.

Several strategies were used to clone mclA and mclB genes in a single construct. All attempts were unsuccessful, suggesting that this construct is lethal for E. coli.

TolC is required for MccL secretion.

The MccV, H47, and E492 export systems require the outer membrane protein TolC to facilitate extracellular secretion of their respective protein substrates (1, 13, 19, 29). Owing to the similarity of the genetic system of microcin L, we suspected that TolC might be involved in the traverse of MccL across the E. coli outer membrane. Strains MC4100 and MC4100 tolC::Tn5 were used as hosts to transform with the plasmid pL102. Whereas MC4100(pL102) produced extracellular microcin L, as expected MC4100 tolC::Tn5(pL102) did not. Nevertheless, when the cells were disrupted, intracellular activity was detected, as observed with the recombinant strain TG₁(pL2). The MccL⁺ phenotype can be rescued by transformation with pAX629 (25), a pACYC184 derivative carrying the tolC gene. These results indicate that TolC is necessary for the export of MccL.

Purification.

MccL was purified to homogeneity from late-logarithmic- or early-stationary-phase supernatant cultures. In order to produce sufficient amounts of MccL for detailed studies on its biological activity and for elucidation of its complete structure, we attempted to improve our previously described protocol (17) and replace the last step, C₁₈ RP-HPLC, with cationic-exchange HPLC. The method used in this study, involving sequential C₁₈ solid-phase extraction and C₁₈ RP-HPLC, was found to be a suitable method of purification of MccL. This procedure suppresses the desalting step that was necessary before MS experiments and leads to a higher final yield of the microcin. From the final C₁₈ RP column a single peak at 215 nm eluted at 58% (vol/vol) acetonitrile-0.1%(vol/vol) TFA.

MS analysis of microcin L.

Sequencing of the DNA corresponding to MccL yielded the following amino acid sequence: G₁DVNWVDVGKTVATNGAGVIGGAFGAGLCGPVCAGAFAVGSSAAVAALYDAAGNSNSAKQKPEGLPPEAWNYAEGRMCNWSPNNLSDVCL₉₀. Mass measurements of RP-HPLC-purified MccL using a Micromass QTOF II, revealed a molecular mass of 8,884.04 ± 0.08 Da instead of the expected mass value of 8,887.91 Da. This 4-Da difference between the experimental mass measurement obtained by MS analysis and mass calculated from the DNA-deduced amino acid sequence was interpreted to be due to the presence of two disulfide bridges formed internally between the four cysteines present in the MccL sequence.

Digestion with Asp-N endoproteinase and assignment of disulfide bridges.

The purified protein cleaved with Asp-N endoproteinase and the resulting fragments were separated by RP-HPLC for on-line and off-line MS analysis. In order to assign the disulfide bridge positions, the MccL was digested without reduction and alkylation to preserve the disulfide connections. The digestion yielded a series of peptides, which were separated by RP-HPLC; the masses of these peptide were measured, and the peptides were collected for Edman degradation (Fig. 3). Four peptides were detected and identified by chromatography. Peptide 1 (A1) corresponds to the first six N-terminal amino acids when no cleavage was observed at position D₂ as expected, probably because of the short incubation time (1 min). Peptide 2 (A2) corresponds to amino acids 7 to 49, with a specific cleavage by Asp-N endoproteinase. Here the disulfide bridge engaged Cys₂₉ and Cys₃₃. This hypothesis was confirmed by mass measurements of peptides P2 (3,797.46 Da), which gave a mass 2 Da lower than the calculated mass (3,799.35 Da). Peptide 3 (A3) corresponds to amino acids 50 to 73. A nonspecific cleavage by Asp-N endoproteinase at position E₇₄ is therefore observed. Peptide 4 (A4/A5) reveals a mass of 1,952.74 Da, which did not correspond to any expected mass for the digested peptide. Its amino acid sequence as determined by Edman degradation on 13 cycles yielded a 2-amino-acid sequence for the first four cycles and then a sequence of amino acids 78 to 86 only. This peptide was therefore identified as two peptides attached by a disulfide bridge between Cys₇₈ and Cys₈₉. The calculated mass for this double chain peptide (1,954.83 − 2H⁺ Da) corresponds well to the measured mass (1,952.74 Da). The disulfide bridges are summarized in Fig. 3. The primary structure of MccL initially determined by DNA sequencing is then confirmed by LC-MS Asp-N mapping and Edman sequencing. Almost all of the MccL sequence (90%) has been determined by Edman sequencing. These results confirmed the predicted amino acid sequence of the A3, A4, and A5 fragments. Only a short sequence (SSAAVAALY) for the C-terminal end of a 43-amino-acid peptide (A2) was not confirmed by Edman degradation as expected. Nevertheless, the clear agreement between the mass predicted by DNA sequencing and the peptide mass measured by electrospray ionization-MS seems to be in favor of the predicted amino acid sequence.

FIG. 3.

Amino acid sequence of microcin L. Sites of Asp-N digestion (arrows indicate Asp-N cleavage) and intramolecular disulfide bridges: Cys₂₉-Cys₃₃ and Cys₇₈-ys₈₉. Footnote letters: a, ion masses correlated to monoisotopic molecular mass of neutral peptides; b, average mass of neutral peptide.

Biochemical characteristics of MccL.

Amino acid composition analysis indicates that MccL does not contain any modified amino acid. The glycine residues (15.6%) are regularly spread along its length, and the 90-amino-acid peptide is strongly hydrophobic (47% hydrophobic residues). Its predicted hydropathy profile calculated according to Kyte and Doolittle (28) indicated that the N-terminal half is largely hydrophobic, whereas the C-terminal half is mainly hydrophilic. The protein is negatively charged at neutral pH, the net charge being −3 with a predicted pI of 4.39. Purified microcin L was tested for sensitivity to heat and various enzymes. Its antibacterial activity is completely inactivated with pronase, trypsin, or papain. Pepsin treatment resulted in partial loss of the activity. Additionally, MccL can be classified as heat stable, since full biological activity was retained after heating at 100°C for 10 min. Its activity is not abolished following a 5 mM addition of dithiothreitol. Homology searches through Swiss-Prot reveal that the N-terminal amino acid sequence displays significant homologies with several gram-positive bacteriocins: 35.2% with gassericin T, 43.6% with subunit lafA of lactacin F, 28.6% with subunit lafX of lactacin F, and 38.5% with divercinV41. We found a striking similarity (87.5%) between the 32 C-terminal amino acids of microcin L and the C terminus of MccV. Comparison of the primary structure of MccL and V is presented in Fig. 4.

FIG. 4.

Sequence alignment of microcins L and V. Identical residues are indicated by asterisks, conserved residues are indicated by double dots, and semiconserved residues are indicated by single dots.

Antibacterial spectrum.

The quantities of MccL obtained in our previous studies (17) were insufficient to test its inhibitory activity against a wide variety of strains, and assays were performed against a limited number of strains, so we checked the activity of purified MccL (3,000 AU ml⁻¹) against numerous strains belonging to the same bacterial species (Table 4). MccL is active against Shigella sp., diarrheagenic E. coli, Pseudomonas sp., and all S. enterica strains tested, including serovars Enteritidis and Typhimurium. The MccL activity is bactericidal for E. coli MC4100 (MIC, 4 nM; MBC, 90 nM). In contrast, all the members of a set of gram-positive bacteria were resistant.

TABLE 4.

Inhibitory spectrum of microcin L

Bacterium species (no. of tested strains)	Origina	Sensitivityb
Gram-negative bacteria
Citrobacter freundii (1)	LC	−
Escherichia coli enteropathogen (9)	LC	++
Hafnia alvei (1)	CIP5731T	−
Hafnia alvei (1)	LC	−
Klebsiella oxytoca (1)	CIP666	+
Klebsiella pneumoniae (2)	LC	−
Salmonella enterica serovar Agona (1)	LC	+
Salmonella enterica serovar Braenderup (1)	LC	++
Salmonella enterica serovar Brandenburg (1)	LC	++
Salmonella enterica serovar Bredeney (4)	LC	+
Salmonella enterica serovar Derby (2)	LC	+
Salmonella enterica serovar Dublin (1)	LC	+
Salmonella enterica serovar Enteritidis (1)	ATCC 13076	+
Salmonella enterica serovar Enteritidis (5)	LC	++
Salmonella enterica serovar Hadar (1)	LC	+
Salmonella enterica serovar Heildeberg (1)	LC	+
Salmonella enterica serovar Indiana (1)	LC	+
Salmonella enterica serovar Kottbus (2)	LC	+
Salmonella enterica serovar Newport (1)	LC	+
Salmonella enterica serovar St. Paul (4)	LC	+
Salmonella enterica serovar Typhimurium (1)	CIP5858	++
Salmonella enterica serovar Typhimurium (2)	LC	+
Salmonella enterica serovar Virchow (3)	LC	+
Serratia marcescens (1)	CIP6755	−
Serratia marcescens (1)	LC	−
Shigella sonnei (1)	CIP5236	++
Shigella sonnei (1)	LC	++
Shigella flexneri (1)	CIP5236	++
Proteus mirabilis (1)	LC	−
Proteus morganii (1)	LC	−
Proteus vulgaris (2)	LC	−
Providencia stuartii (1)	LC	+
Pseudomonas aeruginosa (1)	CIP100720T	+
Pseudomonas aeruginosa (1)	ATCC 27853	+
Pseudomonas stutzeri (1)	LC	−
Gram-positive bacteria
Bacillus cereus (1)	LC	−
Bacillus subtilis (1)	CIP5262	−
Bacillus megaterium (1)	CIP6620T	−
Enterococcus cloacae (1)	CIP6585T	−
Enterococcus faecalis (1)	CIP103214	−
Listeria monocytogenes (1)	CIP80110T	−
Staphylococcus aureus (1)	CIP53154	−

^aAbbreviations: LC, laboratory collection; CIP, Institut Pasteur Collection; ATCC, American Type Culture Collection.

^bResistance is indicated by − (no inhibition zone); sensitivity is indicated from low (+) (inhibition zone 8 to 12 mm) to high (++) (inhibition zone from 12 to 18 mm).

DISCUSSION

We report here the complete primary structure of microcin L, one of four microcins produced by the wild-type E. coli LR05 isolated from a poultry intestinal tract. Previous genetic studies demonstrated that the MccL gene cluster is plasmid encoded. A 13.5-kb HindIII-SalI DNA fragment issued from DNA plasmid was cloned into pUC19, resulting in pL102. It directs the production of MccL and immunity to MccL and MccV (41).

Microcin L was purified to homogeneity and fully characterized at the level of its amino acid sequence, using a combination of RP chromatography, Edman degradation, and MS (electrospray MS and nanoES). MccL is a 90-amino-acid, hydrophobic peptide (46.7% nonpolar amino acids) without any posttranslational modification. It contains a high content of glycine (15.6%) and four cysteines engaged in two intramolecular disulfide bridges in the mature microcin, conferring to the peptide a high stability. MccL is an anionic peptide with three negative charges and a calculated isoelectric point of 4.39. Thus, MccL has many general characteristics in common with other antimicrobial peptides such as MccV, MccH47, and MccE492 (36). These results confirm that it belongs to the small heat-stable hydrophobic class II unmodified microcins characterized by a Gly-2 Ala-1 Xaa processing site in the microcin precursor. However, its overall structure is unique in that the primary structure of the N-terminal domain displays strong homologies with class II gram-positive bacteriocins such as gassericin T, lactacin F, and divercin V41, and the C-terminal hydrophilic region is nearly identical to that of MccV. It has been suggested that mature MccV contains a disulfide bond between the cysteine residues at positions 76 and 87 of the polypeptide (24). In a comparable way, our results demonstrate the presence of two disulfide bridges in MccL, one being located in the MccV homologous C-terminal part.

The spectrum of MccL activity is narrow as for the well-known microcins. The inhibitory activity is mostly directed toward, closely related to, and in competition with E. coli LR05. Of interest to human industries, the targeted bacteria are among the major bacterial strains responsible for human and animal food-borne diseases, generating medical costs and productivity losses. Faced with the increasing multiresistance of these pathogenic agents, including Pseudomonas, toward conventional antibiotics, microcin L presents an opportunity for the future of disease prevention and treatment.

MccV and E492 exert their lethal action by interaction with the cytoplasmic membrane, leading to the dissipation of the membrane potential (31, 47). Many studies suggested that the antimicrobial activity of the cationic peptides, such as those produced by gram-positive bacteria, is related to their interaction with the membrane interface via their net positive charges (11). Nevertheless, MccL and most of the class II microcins present an anionic character. A recent review has underlined the high variability in potential targets in bacteria and the necessity of considering all possibilities when studying any bactericidal antibiotic such as the cationic antimicrobial peptides (21). It is possible for antibacterial substances to have more than one toxic inhibitory activity. Individual peptides would tend to select one or more of these as their preferred target(s) (21). Further studies clearly must be developed to investigate the possible mechanism(s) of action of MccL.

Using reverse genetics, DNA encoding mclC, the MccL structural gene, as well as DNA of the surrounding region was sequenced. Analysis revealed the presence of a DNA fragment corresponding to the beginning of cvaC, the MccV structural gene. Nevertheless, the following nucleotides were not related to the amino acid sequence of MccV, nor were they related to a protein sequence described in the databases. Surprisingly, despite the absence of a functional MccV structural gene, we found the presence of the MccV immunity gene (cvi). In addition, we identified four ORFs belonging to the MccL genetic system, mclC, mclI, and most probably mclA and mclB. Comparison of the sequence of the mclC structural gene and primary structure of purified MccL showed that microcin L is first synthesized as a prepeptide which contains an N-terminal leader sequence. Subsequent cleavage of the precursor at the Gly-2 Ala-1 processing site removes the leader sequence. In most cases, especially for class II microcins, the structural gene encodes an N-terminal extension that is cleaved concomitantly with secretion in the extracellular medium. In addition to the position of conserved double glycine type residues, other consensus elements are revealed by the alignment of the amino acid sequences of the leader peptide of the class II microcins. This leader peptide consensus is found in numerous gram-positive bacteriocins as well (24).

The presence of DNA sequences homologous to MccV genes and located close to the MccH47 genetic system has been previously described (1). No antibacterial or immunity function, however, was encoded by those sequences. Our results display that the MccL genetic region and its adjacent sequences found in E. coli LR05 are unusual. We have ascribed this to a large genetic rearrangement organization resulting in a functional MccV immunity protein and the synthesis of MccL as a structural hybrid of MccV and a gram-positive bacteriocin.

Immunity genes described for the MccV, MccE492, and MccH47 systems are located upstream and adjacent to the respective microcin structural gene in an operon (3). Within the microcin L gene cluster, the structural gene mclC and the gene encoding the immunity protein mclI are located adjacent to one another and transcribed in the same orientation. Although putative promoters were found upstream of the start codon of the structural gene, it is likely that they form an operon. mRNA analyses are under way to verify this hypothesis.

Table 3 outlines the common characteristics of class II microcins for comparison with MclI. These proteins are comparable in size, charge, and hydrophobicity. The predicted proteins are small, from 51 to 95 amino acids; are cationic, with net charge varying from +3 to +7; and form two or three transmembrane helices that may insert into the membrane. Very little is known about the mode of action of these immunity proteins. Contrary to immunity proteins of gram-positive bacteriocins (11, 15), the known immunity proteins of microcins are completely specific, with one immunity gene giving protection toward a single microcin. Nonetheless, until now no significant sequence similarities between the microcins have been displayed. We have shown that MccL presents a high degree of sequence similarity with MccV, located within the C-terminal regions of the two microcins (87.5%). Despite this fact, the amino acid sequences of the immunity proteins of MccV and MccL show no similarity. Consistent with this, no cross-reactivity appeared between the two microcin systems. We can hypothesize that immunity is due to an interaction between the immunity protein and the N-terminal moiety of the corresponding microcin.

Studies performed by Zhang et al. (48) demonstrated that unprocessed MccV already possesses antibiotic activity similar to that of mature microcin. Moreover, MccV exhibited its activity only when presented from outside the inner membrane of the bacterium (48). MccL could act similarly. Furthermore, ColV-1, a Tn5 ColV (MccV) mutant form whose 21 carboxy-terminal amino acids of colicin V are replaced by eight heterologous residues, could be secreted normally but lost all biological activity (14). Point mutations in the C-terminal cysteine residues completely abolish MccV activity, but each mutant was processed and secreted normally from the cell in the presence of CvaAB. These results together support the hypothesis that the C-terminal region of MccV is critical for antibacterial activity but not secretion (48). We therefore propose that the C terminus is primarily responsible for the antibacterial activity of MccV and MccL.

The extraordinary similarity between the proteins encoded by the two genes mclA and mclB and the MFP and ABC transporter proteins involved in MccH47, MccV, and MccE492 export led us to speculate that microcin L utilizes a similar ABC export system. We then inferred that mclA and mclB are organized in an operon, as is the case for the above-mentioned microcins. We demonstrated that the MccL export system would require, in addition to MclA and MclB, the TolC outer membrane protein.

Multiple attempts to clone mclA and mclB were unsuccessful, suggesting that the construct was lethal for the bacteria. Preceding attempts (2) to overproduce CvaB have failed, indicating that CvaB may be detrimental to the cell if overexpressed. Mutations of several amino acids—i.e., Glu-248, Ala-262, Thr-274, Leu-285, Gly-313, Ala-322, or Val-335—of CvaA protein greatly reduced secretion of MccV (27). These residues are all conserved in the MclA amino acid sequence.

During the past decade, both genetic and biochemical studies have revealed that many gram-positive bacteriocin producers synthesize more than one bacteriocin (6). We demonstrated that it is the same for numerous microcin-producing bacteria (unpublished results). Thus, E. coli LR05 carries the genetic determinants of microcins L, B17, J25, and D93 as well as the immunity gene of MccV (41). The production of multiple microcins provides the microorganism an ecological advantage in a competitive bacterial environment, widening the target cell spectrum and increasing the effectiveness of target cell inhibition. Moreover, the MccV immunity protein protects the E coli LR05 strain from antagonistic activity of the MccV-producing strains, which are present in the intestinal environment.