An Extracellular Loop of the Mannose Phosphotransferase System Component IIC Is Responsible for Specific Targeting by Class IIa Bacteriocins

Bacterial strain and growth conditions. Lactococcus lactis B488 (10), a man-PTS null mutant derived from L. lactis IL1403, was used as an expression host in this study. Lactococcal clones were grown at 30°C in M17 medium (Oxoid) supplemented with 0.4% (wt vol⁻¹) galactose and with 5 μg ml⁻¹ erythromycin and 5 μg ml⁻¹ chloramphenicol, when appropriate.

Construction of hybrid man-PTS genes. ptnCD of L. lactis IL1403 and mptCD of Listeria monocytogenes EGD-e were used as sources for the construction of hybrid man-PTS genes. Different combinations of the man-PTS IIC (mptC and ptnC) and IID (mptD and ptnD) genes were fused using a two-step PCR approach (21). In this procedure, two separate fragments were amplified in the first step by using one outer primer and one inner primer for each fragment. Overlapping sequences were introduced by the inner primers, and the two fragments were fused in a second PCR using the outer primers. The primers and template DNA used to construct hybrids H1 to H16, H2X1, H2X2, and H2X3 are outlined in Table 1, and primer sequences are given in Table 2. Plasmid pH1, a pNZ8037 derivative containing mptACD with an XmaI site inserted between mptA and mptC, was constructed first and was used as a cloning cassette to introduce different versions of man-PTS IIC and IID genes downstream of mptA between the XmaI and XhoI restriction sites. All constructs were verified by sequencing. An overview of the plasmids used in this study is found in Table 3.

Site-directed mutagenesis.Site-directed mutagenesis was performed using the mptC gene with plasmid pH1 and a two-step PCR procedure. Mutations were introduced by the use of mutagenic inner primers in combination with outer primers mk236 and mk65. Mutagenic primers mk270 and mk271 were used for mutation G86S, mk272 and mk273 for G87N, mk180 and mk181 for Q88F, mk266 and mk267 for G89H, mk277 and mk278 for G89A, and mk279 and mk280 for G92A. All primer sequences are given in Table 2.

Transformation and heterologous expression.The nisin-inducible two-plasmid system based on pNZ9530 (25) and pNZ8037 (8) was used to express various hybrid combinations of man-PTS genes. pNZ8037-derived plasmids were propagated in Escherichia coli (34) prior to electroporation into L. lactis B488 (22), which is an L. lactis IL1403 ptnABCD deletion mutant carrying plasmid pNZ9530, which contains genes necessary for nisin-induced gene expression. Expression of man-PTS genes was induced by the addition of 0.1 ng ml⁻¹ nisin to the growth medium. The expression and functionality of the man-PTS hybrids in terms of sugar transport were assessed by growing cells in M17 medium (Oxoid) with and without 1% (wt vol⁻¹) 2-deoxy-d-glucose, a nonmetabolizable glucose analogue. The M17 complex medium (containing 5 g of tryptone, 5 g of soya peptone, 5 g of meat digest, 2.5 g of yeast extract, 0.5 g of ascorbic acid, 0.25 g of magnesium sulfate, and 19 g of disodium β-glycerophosphate per liter) supports growth of L. lactis even without the addition of sugar. Growth inhibition by 2-deoxy-d-glucose provides evidence for the presence of a functional sugar transporter (36).

Bacteriocins and bacteriocin assay.All bacteriocins were concentrated from spent supernatants by precipitation with 30% ammonium sulfate (see Table 4 for a list of the producer strains), except for curvacin A, leucocin A, and leucocin C, which were obtained as purified fractions (kindly provided by Helen S. Haugen and Jon Nissen-Meyer). Bacteriocin sensitivity was measured using microtiter plate assays. Stationary-phase cultures of the indicator strains (10⁷ CFU ml⁻¹) were diluted 50-fold and exposed to 2-fold dilutions of the bacteriocins in a total volume of 200 μl in each well. The plates were incubated for 7 to 8 h at 30°C before the growth inhibition was scored spectrophotometrically at 600 nm. The MIC was defined as the amount of bacteriocin required to produce a 50% growth inhibition.

The IIC protein is the major specificity determinant for class IIa bacteriocins.The individual subunits within the man-PTSs are well conserved across the different bacterial phyla. Nevertheless, our recent study revealed that subtle differences within their primary sequences can group IIC and IID proteins into phylogenetically defined subgroups and that, more importantly, this subgrouping corresponds well with their relative levels of sensitivity to class IIa bacteriocins (24). Thus, the man-PTSs of Listeria monocytogenes (mpt) and E. faecalis, which belong to the same phylogenetic subgroup, are both highly potent receptors for class IIa bacteriocins, while the corresponding man-PTS of Lactococcus lactis (ptn), which belongs to a more distantly related subgroup, confers no sensitivity to class IIa bacteriocins. To localize the part(s) of IIC and/or IID that is responsible for specific recognition by class IIa bacteriocins, we constructed hybrid man-PTSs with components derived from the highly sensitive mpt and the nonsensitive ptn genes (Table 3) and then assessed their sensitivities to a panel of 10 different class IIa bacteriocins. Different combinations of the class IIC (mptC and ptnC) and IID (mptD and ptnD) genes were fused and coexpressed with mptA (encoding the IIAB subunits from Listeria monocytogenes) in the L. lactis B488 host strain (10), from which the endogenous man-PTS genes ptnABCD have been deleted to prevent background interference.

As expected, expression of the wild-type mptCD (H1) genes of Listeria conferred sensitivity to all class IIa bacteriocins tested whereas expression of the lactococcal ptnCD (H2) genes did not (Fig. 1A). Combinations of mptC with ptnD (H3) and ptnC with mptD (H4) revealed that only the former could confer strong sensitivity. H1 and H3 displayed similar degrees of sensitivity for most bacteriocins, with the exception of sakacin P, which was over 30-fold less active against H3 (Fig. 1A). This suggests that the listerial class IIC protein (MptC) is the main determinant for specific targeting by these bacteriocins. This result also holds for the man-PTS genes manM and manN (encoding IIC and IID, respectively) of Lactobacillus sakei, which is known to be sensitive to class IIa bacteriocins (10). Here, only the combination of manM (encoding IIC) with ptnD (IID) rendered cells sensitive to class IIa bacteriocins whereas the reciprocal combination, i.e., ptnC (IIC) with manN (IID), did not (data not shown), confirming that IIC is the main specificity determinant.

• Open in new tab
• Download powerpoint

FIG. 1.

Relative bacteriocin sensitivities of L. lactis B488 clones expressing different combinations of man-PTS genes of the listerial system (mpt) and the lactococcal system (ptn). (A) Wild-type IIC and IID genes (H1 and H2) and intergenic hybrids (H3 and H4); (B) intragenic IIC hybrids (H5 to H8); (C) six different clones with point mutations in mptC (G86S, G87N, Q88F, G89H, G89A, and G92A); (D) two clones with an exchange of one of the extracellular loops (H2X1 with 11 aa exchanged and H2X3 with 40 aa exchanged). Ptn proteins (PtnC and PtnD) are represented by black boxes and Mpt proteins (MptC and MptD) by white boxes. The different constructs were coexpressed with mptA (encoding a IIAB subunit in Listeria monocytogenes) to form a complete man-PTS complex (Table 3). Ten different class IIa bacteriocins were tested: pediocin PA-1 (Ped PA-1 [20, 29]), sakacin P (SakP [38]), leucocin C (LeuC [15]), avicin A (AviA [3]), leucocin A (LeuA [18]), enterocin P (EntP [4]), curvacin A (CurA [37]), bacteriocin RC714 (Bac RC714 [7]), hiracin JM79 (Hir JM79 [35]), and penocin A (PenA [9]). One arbitrary unit (AU) was defined as representing the MIC of clone H1, and the MICs of the other test clones were determined relative to this value. MIC values were determined at least three times. Clones were defined as not inhibited (NI) when the MIC values increased more than 600-fold (>600 AU). Growth inhibition by 2-deoxy-d-glucose (2-DG) is indicated with a plus sign; absence of inhibition is indicated with a minus sign.

In a previous study, Ramnath et al. (32) showed that heterologous expression of mptC alone was sufficient to render resistant L. lactis strains sensitive to class IIa bacteriocins; this result appeared contradictory to both previous and later studies showing that both IIC and IID are necessary for the receptor function (6, 10). However, given that MptC can form a potent receptor together with PtnD (H3; Fig. 1A), this discrepancy can be resolved, since it is likely that the heterologous MptC protein interacted with the endogenous PtnD protein in L. lactis and that the two together formed a receptor complex that resulted in sensitivity to class IIa bacteriocins.

A defined N-terminal region of IIC is involved in specific interactions with class IIa bacteriocins.MptC and PtnC are 268 amino acids (aa) and 270 aa long, respectively, and have significant sequence similarity over the entire lengths of the sequences (59% identity and 75% similarity). Both proteins are predicted to contain seven transmembrane segments (TMSs) and four extracellular loops. To determine the part(s) of MptC that are directly involved in the specific interaction with class IIa bacteriocins, two reciprocal chimeric IIC genes were constructed from mptC and ptnC. An N-terminal sequence of either MptC (151 aa) or PtnC (155 aa) was genetically fused with the remaining sequence of the counterpart (resulting in MptC-PtnC or PtnC-MptC). The adjoining region was thus selected to be in the middle of the protein, within transmembrane segment 5 (TMS5), so that the topologies of the loops or TMSs potentially involved in bacteriocin interaction were kept largely intact. The resulting hybrid IIC gene was coexpressed with either mptD or ptnD, giving rise to clones H5 to H8. As shown in Fig. 1B, only clones producing a class IIC protein with the N terminus from MptC (H5 and H7) became sensitive to class IIa bacteriocins, while the other clones producing the reciprocal hybrids (H6 and H8) did not. Although there were some variations found in comparisons of H5 and H7 with respect to receptor potency for some of the bacteriocins, e.g., H7 was less sensitive than H5 to sakacin P and avicin A, the source of the coexpressed IID gene in H5 and H7 (mptD or ptnD) did not have any major effect on the receptor specificity. Together, these findings indicate that the specificity for class IIa bacteriocins relies mainly on the N-terminal part of the IIC protein.

As the poor receptor activity exhibited by some of the man-PTS hybrids tested could have been due to poor expression of the cloned genes, the functionality of the resulting man-PTSs in terms of sugar import was examined. When cells were grown in a medium containing 2-deoxy-d-glucose, a nonmetabolizable glucose analogue, growth inhibition has provided evidence for the presence of a functional sugar transporter (36). Indeed, all clones expressing the hybrid man-PTS species (H1 to H8) were shown to be sensitive to 2-deoxy-d-glucose (data summarized in Fig. 1), which implies that all the hybrid genes were expressed and their gene products correctly structured into functional sugar permeases.

An extracellular loop of MptC determines the specificity for class IIa bacteriocins.Next, we wished to reveal the feature in the N-terminal half of MptC that is involved in specific bacteriocin recognition. Amino acid sequence alignment of the N-terminal halves of PtnC and MptC revealed a region with a markedly higher heterogeneity than the rest of the sequence (Fig. 2A). This region of 14 to 16 amino acids, corresponding to residues 85 to 99 in MptC and 87 to 103 in PtnC, is predicted to constitute an extracellular loop. Such marked differences in this region could also be seen when the alignment was extended to include other IIC proteins from a set of man-PTSs previously shown to display various receptor activities for class IIa bacteriocins (e.g., no, poor, low, or high receptor potency) (24) (Fig. 2B). In general, the sequences of man-PTSs with medium or high receptor activity are more similar to each other in this region than to those of the man-PTSs with low or no receptor activity.

• Open in new tab
• Download powerpoint

FIG. 2.

(A) Alignment of MptC from Listeria monocytogenes and PtnC from L. lactis. Only the N-terminal part of the alignment is shown. An asterisk, two dots, and one dot (in that order) indicate decreasing degrees of conservation. Predicted transmembrane helix residues (H), intracellular loop residues (i), and extracellular loop residues (o) determined by the use of TMHMM version 2.0 software (27) are indicated for MptC. The bullets (•) designate residues subjected to site-directed mutagenesis. The cloned region in H2X3 is underlined; the cloned region in H2X1 is shown in italics. (B) Multiple sequence alignment of the region containing the predicted extracellular loop in eight different mannose-PTS IIC proteins: MptC (Lm) from Listeria monocytogenes, MptC (Ef) from E. faecalis, ManM (Ls) from Lactobacillus sakei, Pts9C from Lactobacillus plantarum, ManM (St) from Streptococcus thermophilus, PtnC (Ll) from L. lactis, ManY from E. coli, and MpoC from Listeria monocytogenes. The potencies of these proteins as receptors for class IIa bacteriocins were determined in a previous work (24). The alignments were constructed using MUSCLE version 3.7 software (12).

To investigate whether this region is important for specific interaction with the class IIa bacteriocins, six site-directed mutations in five residues situated in the predicted extracellular loop of MptC were performed (Fig. 2A and Table 3). The MptC residues were changed to the amino acids found at similar positions in PtnC (G86S, G87N, and Q88F) (Fig. 2A), with the exception of G92, which was mutated to alanine (G92A), and G89, which was mutated to two different amino acids: histidine (G89H), which is found in a similar position in PtnC, and alanine (G89A), which is a relatively conserved mutation. By and large, four of the mutations (G86S, G87N, G89A, and G92A) had little or no effect on receptor activity (Fig. 1C). The remaining two mutations (Q88F and G89H), on the other hand, caused a significant reduction in receptor activity for all the bacteriocins tested. Q88F caused MIC values to increase 6- to 500-fold, with avicin A being the most affected; most drastically, the G89H mutation totally disrupted the receptor function for all the bacteriocins tested. Surprisingly, three of the mutations (G86S, Q88F, and G89H) led to a compromised sugar transport function of the man-PTSs (Fig. 1C). The molecular nature of this adverse effect on sugar import is unknown, but it might indicate that these mutations somehow interfered with the substrate binding and/or caused major structural changes in the protein that impaired sugar uptake and possibly also bacteriocin interaction (for Q88F and G89H). Interestingly, G86, Q88, and G89 in MptC (together with G87) constitute a sequence (GGQG) which highly resembles the loop-located GGXG sequence motifs that are known to be important for the transport function of other transmembrane proteins (11). Residues in the extracellular loop might therefore have an important role in sugar uptake, in addition to being a potential target site for bacteriocins. It should be noted that, since the G89H mutant is inactive both as a bacteriocin receptor and as a sugar transporter, whether the proteins encoded in this construct are stably expressed and correctly folded is unknown. Importantly, however, a functioning man-PTS in terms of sugar uptake is not a requirement for a potent bacteriocin receptor, as demonstrated with the G86S mutant. Furthermore, our previous study (10) also showed that heterologous coexpression of mptC and mptD without mptAB could confer bacteriocin sensitivity even though the man-PTS itself was not a functional sugar transporter.

As the site-directed mutagenesis results provided indications that some residues located in the extracellular loop of MptC are critical for the interaction with class IIa bacteriocins, we further investigated whether this region is entirely responsible for the specific interaction with class IIa bacteriocins by replacing 13 putative loop residues in PtnC (Q87 to T99) with the corresponding residues from MptC (L85 to S95). However, the resulting clone (H2X1) did not become sensitive to class IIa bacteriocins (Fig. 1D), and this chimeric system was also unable to transport sugars, suggesting that the exchange of these loop residues could have a severe impact on the total structuring of PtnC or that the construct could not be stably expressed. As flanking regions could play a role in the stability and/or structuring of the loop structure, a longer region of 40 aa (V77 to T116) from MptC, including the residues of the flanking transmembrane helices, was used to replace the corresponding region in PtnC. Remarkably, the resulting chimera became a highly potent receptor for most of the class IIa bacteriocins as well as being a functional sugar transporter (Fig. 1D).

Together, these results show that the predicted extracellular loop and the flanking regions involved in transmembrane segments (40 aa in total) are essential for the specific recognition by class IIa bacteriocins. We cannot exclude the possibility that other residues located elsewhere are important for the bacteriocin-receptor interaction; however, these interactions are not species-specific, because the bacteriocin-sensitive clone H2X3 contains both the lactococcal IIC and IID genes except for the cloned region (encoding 40 aa) derived from the listerial man-PTS.

Lactococcin A interacts with its receptor in a complex manner.The availability of the various mpt-ptn hybrid constructs allowed us to assess whether the Lactococcus-targeting bacteriocin lactococcin A also recognizes a specific region(s) on its man-PTS receptor. Overall, it can be seen that, unlike the class IIa bacteriocins, lactococcin A required both IIC and IID from the same source, namely, the lactococcal man-PTS PtnCD, to kill target cells (clone H2), as any interchange with a noncognate ptn component (clones H1, H3, and H4) had a detrimental effect on the receptor activity (Fig. 3A). To further examine whether lactococcin A recognizes discrete regions on IIC and/or IID clones, additional intragenic hybrid clones were constructed and analyzed for their sensitivity. Besides the clones producing intragenic hybrids of IIC (H5 to H8, H2X1, and H2X3; Fig. 1), four more clones containing intragenic hybrids of IID were made (MptD/PtnD and PtnD/MptD in H9 to H12; Table 3). Combinations of the two sets of hybrid genes gave rise to another four clones containing both IIC and IID hybrids (H13 to H16; Table 3). Bacteriocin sensitivity assays revealed that only three of them (H5, H11, and H2X3) could display some receptor activity (Fig. 3B), with clone H11 being the most sensitive (only a 4-fold increase in MIC compared to the wild-type receptor results). Clone H11 contains an intact PtnC and a hybrid IID composed of an N-terminal part (113 aa) from MptD and a C-terminal (193 aa) part from PtnD, indicating that the N-terminal part of PtnD is less important for specific lactococcin A interaction. Intriguingly, although the sequence of H2X1 is more similar to those of the wild-type genes (H2) than to that of H2X3, H2X1 exhibited much less receptor activity than H2X3. The nature of the factors behind these rather contrasting results is unknown; however, it is possible that a larger portion of MptC in H2X3 could have stabilizing effects on the structure that somehow allowed lactococcin A to form a stronger interaction with the receptor or that H2X1 was not stably expressed. It should be underlined here that the H2X3 hybrid and (to some extent) H5 are the first known (genetically modified) receptors being targeted by both lactococcin A and the class IIa bacteriocins and that such a receptor is yet to be found in nature.

• Open in new tab
• Download powerpoint

FIG. 3.

Sensitivity of various L. lactis B488 clones to lactococcin A (LcnA [23]). (A) Wild-type IIC and IID genes (H1 and H2) and intergenic hybrids (H3 and H4); (B) intragenic hybrids H5 to H16, H2X1, and H2X3 (Table 3). For lactococcin A, 1 AU was defined as representing the MIC of clone H2; the MICs of the other test clones were determined relative to this value. MIC values were determined at least three times. Clones were defined as not inhibited (NI) when the corresponding MIC showed an increase of at least 4,000-fold. Growth inhibition and no growth inhibition induced by 2-deoxy-d-glucose (2-DG) are indicated with a plus sign and a minus sign, respectively.

Taken together, these results indicate a rather complex bacteriocin-receptor interaction for lactococcin A compared to what has been observed for class IIa bacteriocins; this might also explain why lactococcin A has an extremely narrow inhibitory spectrum, killing only lactococci (23).

Conclusions.Man-PTSs in both Gram-positive and Gram-negative bacteria function as receptors for several antibacterial macromolecules, including peptides and bacteriophages (2, 6, 10, 14, 19). Here we demonstrate that the members of one group of antimicrobial peptides, the class IIa bacteriocins, require a short region containing an extracellular loop in the N-terminal part of the IIC protein to specifically target sensitive cells. These results support the previously proposed idea (13, 32) that IIC is important for the specific interaction with bacteriocins whereas IID is essential for the stability and structuring of the protein complex. However, IIC is not the major specificity determinant for all man-PTS targeting bacteriocins, as we also demonstrated here that parts of both IIC and IID appeared to be essential for the specific recognition by the class IIc bacteriocin lactococcin A.

Lactococcin A has several physiochemical properties different from those of class IIa bacteriocins: it is 54 aa in length whereas class IIa bacteriocins are in the range of 37 to 48 aa, and it shares little or no sequence similarity with class IIa bacteriocins whereas the latter share significant sequence similarity with each other (30). Such divergent features might suggest different mechanisms for targeting and/or killing sensitive cells, and this notion is in fact supported by the results of the present study, which show a much more complex receptor targeting for lactococcin A than for class IIa bacteriocins. Despite these marked differences, producer cells for both class IIa bacteriocins and lactococcin A somehow employ a common immunity mechanism. It has been shown for both types of bacteriocins that their cognate immunity proteins act by locking the bacteriocins on man-PTSs in a tripartite complex (bacteriocin-immunity protein-man-PTS), thereby preventing the bacteriocins from forming lethal pores (10). In terms of structural biology, it would have been interesting to discover how two such different classes of bacteriocins have evolved divergently in some aspects (different sequences and different degrees of species specificity) and at the same time have converged in some other aspects (use by both of a man-PTS as a receptor and a common mechanism of immunity). Future work aimed at resolving the structure of the tripartite complex is likely to provide detailed insights into these fascinating systems.