Camilla Oppegård

The Two-Peptide Class II Bacteriocins: Structure, Production, and Mode of Action

The two-peptide class II bacteriocins consist of two different unmodified peptides, both of which must be present in about equal amounts in order for these bacteriocins to exert optimal antimicrobial activity. These bacteriocins render the membrane of target cells permeable to various small molecules. The genes encoding the two peptides of two-peptide bacteriocins are adjacent to each other in the same operon and they are near the genes encoding (i) the immunity protein that protects the bacteriocin-producing bacteria from being killed by their own bacteriocin, (ii) a dedicated ABC transporter that transports the bacteriocin out of the bacteriocin-producing bacteria, and (iii) an accessory protein whose specific role is not known, but which also appears to be required for secretion of the bacteriocin. The production of some two-peptide bacteriocins is transcriptionally regulated through a three-component regulatory system that consists of a membrane-interacting peptide pheromone, a membrane-associated histidine protein kinase, and response regulators. Structure analysis of three two-peptide bacteriocins (plantaricin E/F, plantaricin J/K, and lactococcin G) by CD (and in part by NMR) spectroscopy reveal that these bacteriocins contain long amphiphilic α-helical stretches and that the two complementary peptides interact and structure each other when exposed to membrane-like entities. Lactococcin G shares about 55% sequence identity with enterocin 1071, but these two bacteriocins nevertheless kill different types of bacteria. The target-cell specificity of lactococcin G-enterocin 1071 hybrid bacteriocins that have been constructed by site-directed mutagenesis suggests that the β-peptide is important for determining the target-cell specificity.

Structure and Mode-of-Action of the Two-Peptide (Class-IIb) Bacteriocins

This review focuses on the structure and mode-of-action of the two-peptide (class-IIb) bacteriocins that consist of two different peptides whose genes are next to each other in the same operon. Optimal antibacterial activity requires the presence of both peptides in about equal amounts. The two peptides are synthesized as preforms that contain a 15–30 residue double-glycine-type N-terminal leader sequence that is cleaved off at the C-terminal side of two glycine residues by a dedicated ABC-transporter that concomitantly transfers the bacteriocin peptides across cell membranes. Two-peptide bacteriocins render the membrane of sensitive bacteria permeable to a selected group of ions, indicating that the bacteriocins form or induce the formation of pores that display specificity with respect to the transport of molecules. Based on structure–function studies, it has been proposed that the two peptides of two-peptide bacteriocins form a membrane-penetrating helix–helix structure involving helix–helix-interacting GxxxG-motifs that are present in all characterized two-peptide bacteriocins. It has also been suggested that the membrane-penetrating helix–helix structure interacts with an integrated membrane protein, thereby triggering a conformational alteration in the protein, which in turn causes membrane-leakage. This proposed mode-of-action is similar to the mode-of-action of the pediocin-like (class-IIa) bacteriocins and lactococcin A (a class-IId bacteriocin), which bind to a membrane-embedded part of the mannose phosphotransferase permease in a manner that causes membrane-leakage and cell death.

Sensitivity to the two-peptide bacteriocin lactococcin G is dependent on UppP, an enzyme involved in cell-wall synthesis

Most bacterially produced antimicrobial peptides (bacteriocins) are thought to kill target cells by a receptor‐mediated mechanism. However, for most bacteriocins the receptor is unknown. For instance, no target receptor has been identified for the two‐peptide bacteriocins (class IIb), whose activity requires the combined action of two individual peptides. To identify the receptor for the class IIb bacteriocin lactococcin G, which targets strains of Lactococcus lactis, we generated 12 lactococcin G‐resistant mutants and performed whole‐genome sequencing to identify mutations causing the resistant phenotype. Remarkably, all had a mutation in or near the gene uppP (bacA), encoding an undecaprenyl pyrophosphate phosphatase; a membrane protein involved in peptidoglycan synthesis. Nine mutants had stop codons or frameshifts in the uppP gene, two had point mutations in putative regulatory regions and one caused an amino acid substitution in UppP. To verify the receptor function of UppP, it was shown that growth of non‐sensitive Streptococcus pneumoniae could be inhibited by lactococcin G when L. lactis uppP was expressed in this bacterium. Furthermore, we show that the related class IIb bacteriocin enterocin 1071 also uses UppP as receptor. The approach used here should be broadly applicable to identify receptors for other bacteriocins as well.

Mutational analysis of putative helix-helix interacting GxxxG-motifs and tryptophan residues in the two-peptide bacteriocin lactococcin G.

The membrane-permeabilizing two-peptide bacteriocin lactococcin G consists of two different peptides, LcnG-alpha and LcnG-beta. The bacteriocin contains several tryptophan and tyrosine residues and three putative helix-helix interacting GxxxG-motifs, G 7xxxG 11 and G 18xxxG 22 in LcnG-alpha and G 18xxxG 22 in LcnG-beta. The tryptophan and tyrosine residues and residues in the GxxxG-motifs were altered by site-directed mutagenesis to analyze the structure and membrane-orientation of lactococcin G. Substituting the glycine residues at position 7 or 11 in the G 7xxxG 11-motif in LcnG-alpha with large hydrophobic or hydrophilic residues was highly detrimental, whereas small residues were tolerated. Qualitatively similar results were obtained for the G 18xxxG 22-motif in LcnG-beta. In contrast, replacement of the glycine residues in the middle of these two motifs with large hydrophilic residues was tolerated. All mutations in the G 18xxxG 22-motif in LcnG-alpha were relatively well-tolerated, indicating that this motif is not involved in helix-helix interactions. The four aromatic residues in the N-terminal part of LcnG-beta could individually be replaced by other aromatic residues, a hydrophilic positive residue, and a hydrophobic residue without a marked reduced activity, indicating that this region is structurally flexible and not embedded in a strictly hydrophobic or hydrophilic environment. The results are in accordance with a structural model where the G 7xxxG 11-motif in LcnG-alpha and the G 18xxxG 22-motif in LcnG-beta interact and allow the two peptides to form a parallel transmembrane helix-helix structure, with the tryptophan-rich N-terminal part of LcnG-beta positioned in the outer membrane interface and the cationic C-terminal end of LcnG-alpha inside the cell.

Analysis of the Two-Peptide Bacteriocins Lactococcin G and Enterocin 1071 by Site-Directed Mutagenesis

ABSTRACT The two peptides (Lcn-α and Lcn-β) of the two-peptide bacteriocin lactococcin G (Lcn) were changed by stepwise site-directed mutagenesis into the corresponding peptides (Ent-α and Ent-β) of the two-peptide bacteriocin enterocin 1071 (Ent), and the potencies and specificities of the various hybrid constructs were determined. Both Lcn and, to a lesser extent, Ent were active against all the tested lactococcal strains, but only Ent was active against the tested enterococcal strains. The two bacteriocins thus differed in their relative potencies to various target cells, despite their sequence similarities. The hybrid combination Lcn-α+Ent-β had low potency against all strains tested, indicating that these two peptides do not interact optimally. The reciprocal hybrid combination (i.e., Ent-α+Lcn-β), in contrast, was highly potent, indicating that these two peptides may form a functional antimicrobial unit. In fact, this hybrid combination (Ent-α+Lcn-β) was more potent against lactococcal strains than wild-type Ent was (i.e., Ent-α+Ent-β), but it was inactive against enterococcal strains (in contrast to Ent but similar to Lcn). The observation that Ent-α is more active against lactococci in combination with Lcn-β and more active against enterococci in combination with Ent-β suggests that the β peptide is an important determinant of target cell specificity. Especially the N-terminal residues of the β peptide seem to be important for specificity, since Ent-α combined with an Ent-β variant with Ent-to-Lcn mutations at positions 1 to 4, 7, 9, and 10 was >150-fold less active against enterococcal strains but one to four times more active against lactococcal strains than Ent-α+Ent-β. Moreover, Ent-to-Lcn single-residue mutations in the region spanning residues 1 to 7 in Ent-β had a more detrimental effect on the activity against enterococci than on that against lactococcal strains. Of the single-residue mutations made in the N-terminal region of the α peptide, the Ent-to-Lcn mutations N8Q and P12R in Ent-α influenced specificity, as follows: the N8Q mutation had no effect on activity against tested enterococcal strains but increased the activity 2- to 4-fold against the tested lactococcal strains, and the P12R mutation reduced the activity >150-fold and only ∼2-fold against enterococcal and lactococcal strains, respectively. Changing residues in the C-terminal half/part of the Lcn peptides (residues 20 to 39 and 25 to 35 in Lcn-α and Lcn-β, respectively) to those found in the corresponding Ent peptides did not have a marked effect on the activity, but there was an ∼10-fold or greater reduction in the activity upon also introducing Lcn-to-Ent mutations in the mid-region (residues 8 to 19 and 9 to 24 in Lcn-α and Lcn-β, respectively). Interestingly, the Lcn-to-Ent F19L+G20A mutation in an Lcn-Ent-β hybrid peptide was more detrimental when the altered peptide was combined with Lcn-α (>10-fold reduction) than when it was combined with Ent-α (∼2-fold reduction), suggesting that residues 19 and 20 (which are part of a GXXXG motif) in the β peptide may be involved in a specific interaction with the cognate α peptide. It is also noteworthy that the K2P and A7P mutations in Lcn-β reduced the activity only ∼2-fold, suggesting that the first seven residues in the β peptides do not form an α-helix.

Plantaricin A, a peptide pheromone produced by Lactobacillus plantarum, permeabilizes the cell membrane of both normal and cancerous lymphocytes and neuronal cells

Antimicrobial peptides produced by multicellular organisms protect against pathogenic microorganisms, whereas such peptides produced by bacteria provide an ecological advantage over competitors. Certain antimicrobial peptides of metazoan origin are also toxic to eukaryotic cells, with preference for a variety of cancerous cells. Plantaricin A (PlnA) is a peptide pheromone with membrane permeabilizing strain-specific antibacterial activity, produced by Lactobacillus plantarum C11. Recently, we have reported that PlnA also permeabilizes cancerous rat pituitary cells (GH(4) cells), whereas normal rat anterior pituitary cells are resistant. To investigate if preferential effect on cancerous cells is a general feature of PlnA, we have studied effects of the peptide on normal and cancerous lymphocytes and neuronal cells. Normal human B and T cells, Reh cells (from human B cell leukemia), and Jurkat cells (from human T cell leukemia) were studied by flow cytometry to detect morphological changes (scatter) and viability (propidium iodide uptake), and by patch clamp recordings to monitor membrane conductance. Ca(2+) imaging based on a combination of fluo-4 and fura-red was used to monitor PlnA-induced membrane permeabilization in normal rat cortical neurons and glial cells, PC12 cells (from a rat adrenal chromaffin tumor), and murine N2A cells (from a spinal cord tumor). All the tested cell types were affected by 10-100 microM PlnA, whereas concentrations below 10 microM had no significant effect. We conclude that normal and cancerous lymphocytes and neuronal cells show similar sensitivity to PlnA.

The lactococcin G immunity protein recognizes specific regions in both peptides constituting the two-peptide bacteriocin lactococcin G.

ABSTRACT Lactococcin G and enterocin 1071 are two homologous two-peptide bacteriocins. Expression vectors containing the gene encoding the putative lactococcin G immunity protein (lagC) or the gene encoding the enterocin 1071 immunity protein (entI) were constructed and introduced into strains sensitive to one or both of the bacteriocins. Strains that were sensitive to lactococcin G became immune to lactococcin G when expressing the putative lactococcin G immunity protein, indicating that the lagC gene in fact encodes a protein involved in lactococcin G immunity. To determine which peptide or parts of the peptide(s) of each bacteriocin that are recognized by the cognate immunity protein, combinations of wild-type peptides and hybrid peptides from the two bacteriocins were assayed against strains expressing either of the two immunity proteins. The lactococcin G immunity protein rendered the enterococcus strain but not the lactococcus strains resistant to enterocin 1071, indicating that the functionality of the immunity protein depends on a cellular component. Moreover, regions important for recognition by the immunity protein were identified in both peptides (Lcn-α and Lcn-β) constituting lactococcin G. These regions include the N-terminal end of Lcn-α (residues 1 to 13) and the C-terminal part of Lcn-β (residues 14 to 24). According to a previously proposed structural model of lactococcin G, these regions will be positioned adjacent to each other in the transmembrane helix-helix structure, and the model thus accommodates the present results.

Structure analysis of the two-peptide bacteriocin lactococcin G by introducing D-amino acid residues

The importance of 3D structuring in the N- and C-terminal ends of the two peptides (39-mer LcnG-alpha and 35-mer LcnG-beta) that constitute the two-peptide bacteriocin lactococcin G was analysed by replacing residues in the end regions with the corresponding D-isomeric residues. When assayed for antibacterial activity in combination with the complementary wild-type peptide, LcnG-alpha with four D-residues in its C-terminal region and LcnG-beta with four d-residues in either its N- or its C-terminal region were relatively active (two- to 20-fold reduction in activity). 3D structuring of the C-terminal region in LcnG-alpha and the C- and N-terminal regions in LcnG-beta is thus not particularly critical for retaining antibacterial activity, indicating that the 3D structure of these regions is not vital for interpeptide interactions or for interactions between the peptides and cellular components. The 3D structure of the N-terminal region in LcnG-alpha may be more important, as LcnG-alpha with four N-terminal D-residues was the least active of these four peptides (10- to 100-fold reduction in activity). The results are consistent with a proposed structural model of lactococcin G in which LcnG-alpha and -beta form a transmembrane parallel helix-helix structure involving approximately 20 residues in each peptide, starting near the N terminus of LcnG-alpha and at about residue 13 in LcnG-beta. Upon expressing the lactococcin G immunity protein, sensitive target cells became resistant to all of these D-residue-containing peptides. The end regions of the two lactococcin G peptides are consequently not involved in essential structure-dependent interactions with the immunity protein. The relatively high activity of most of the D-residue-containing peptides suggests that bacteriocins with increased resistance to exopeptidases may be generated by replacing their N- and C-terminal residues with d-residues.

A putative amino acid transporter determines sensitivity to the two-peptide bacteriocin plantaricin JK

Lactobacillus plantarum produces a number of antimicrobial peptides (bacteriocins) that mostly target closely related bacteria. Although bacteriocins are important for the ecology of these bacteria, very little is known about how the peptides target sensitive cells. In this work, a putative membrane protein receptor of the two‐peptide bacteriocin plantaricin JK was identified by comparing Illumina sequence reads from plantaricin JK‐resistant mutants to a crude assembly of the sensitive wild‐type Weissella viridescens genome using the polymorphism discovery tool VAAL. Ten resistant mutants harbored altogether seven independent mutations in a gene encoding an APC superfamily protein with 12 transmembrane helices. The APC superfamily transporter thus is likely to serve as a target for plantaricin JK on sensitive cells.

The Pediocin PA-1 Accessory Protein Ensures Correct Disulfide Bond Formation in the Antimicrobial Peptide Pediocin PA-1.

Peptides, in contrast to proteins, are generally not large enough to form stable and well-defined three-dimensional structures. However, peptides are still able to form correct disulfide bonds. Using pediocin-like bacteriocins, we have examined how this may be achieved. Some pediocin-like bacteriocins, such as pediocin PA-1 and sakacin P[N24C+44C], have four cysteines. There are three possible ways by which the four cysteines may combine to form two disulfide bonds, and the three variants are expected to be produced in approximately equal amounts if their formation is random. Pediocin PA-1 and sakacin P[N24C+44C] with correct disulfide bonds were the main products when they were secreted by the pediocin PA-1 ABC transporter and accessory protein, but when they were secreted by the corresponding secretion machinery for sakacin A, a pediocin-like bacteriocin with one disulfide bond (two cysteines), peptides with all three possible disulfide bonds were produced in approximately equal amounts. All five cysteines in the pediocin PA-1 ABC transporter and the two cysteines (that form a CxxC motif) in the accessory protein were individually replaced with serines to examine their involvement in disulfide bond formation in pediocin PA-1. The Cys86Ser mutation in the accessory protein caused a 2-fold decrease in the amount of pediocin PA-1 with correct disulfide bonds, while the Cys83Ser mutation nearly abolished the production of pediocin PA-1 and resulted in the production of all three disufide bond variants in equal amounts. The Cys19Ser mutation in the ABC transporter completely abolished secretion of pediocin PA-1, suggesting that Cys19 is in the proteolytic active site and involved in cleaving the prebacteriocin. Replacing the other four cysteines in the ABC transporter with serines caused a slight reduction in the overall amount of secreted pediocin PA-1, but the relative amount with the correct disulfide bonds remained large. These results indicate that the pediocin PA-1 accessory protein has a chaperone-like activity in that it ensures the formation of the correct disulfide bond in pediocin PA-1.