ABSTRACT
Microcin E492 uptake by FepA, Fiu, and Cir is cooperative, with FepA being the main receptor. No TonB-mediated interaction with the ferric catecholate receptors is needed for microcin to exert action at the cytoplasmic membrane. Microcin E492 uptake by the receptors is inhibited by the dimer and trimer of dihydroxybenzoylserine.
Microcin E492 (MccE492) is a channel-forming bacteriocin produced by Klebsiella pneumoniae RYC492 (4, 10). MccE492 uses as receptors FepA, Fiu, and Cir (1, 20), which are the receptors for the ferric form of enterochelin and its hydrolysis products (6). Enterochelin is an antagonist of MccE492 activity, in contrast to ferric enterochelin, which has no inhibitory effect (19). The receptor for ferric enterochelin is FepA, whereas the enterochelin hydrolysis products—a monomer, dimer, and trimer of dihydroxybenzoylserine (DHBS)—complexed with iron are recognized by FepA, by Fiu, and to a minor extent by Cir (6). FepA, a gated channel, is converted to an open state after a conformational change induced by the ferric siderophore binding and transmitted to TonB. TonB functions as a molecular gatekeeper that stimulates the opening of the channel (23), allowing the ferric enterochelin to pass through the channel. To exert its bactericidal activity, MccE492 needs TonB (22) at the level of the interaction with the cytoplasmic membrane of the target cells (12). However, it has not been established if the presence of the receptors is required for MccE492-TonB interaction.
We sought here to establish the contribution of FepA, Fiu, and Cir to MccE492 uptake and to determine whether the TonB-dependent step for MccE492 activity requires the presence of the receptors. In order to provide new insight on the interaction between MccE492 with its receptors, the effect of the precursor dihydroxybenzoic acid (DHBA) and the monomer, dimer, and trimer of DHBS that are natural ligands of these receptors was investigated. The characterization of MccE492 uptake is of great interest not only because there are other microcins that probably share a similar mechanism for receptor recognition but also because this particular bacteriocin has the unusual property of acting as a toxin in some eukaryotic cell lines by inducing apoptosis (7). Thus, MccE492 offers an interesting model in which a protein can recognize both eukaryotic and prokaryotic cells, probably by completely different mechanisms.
MccE492 uptake by FepA, Fiu, and Cir is cooperative, and these receptors contribute differently to MccE492 sensitivity.
The bacterial strains mutated in the ferric-catecholate receptors are described elsewhere (6), and some of them were kindly provided by K. Hantke. MccE492 was purified according to the method of de Lorenzo (4), starting from the recombinant overproducing strain Escherichia coli VCS257JEM15 (29) in the conditions described previously (3, 12), in which mainly a nonmodified microcin was isolated (5, 12, 21, 25). The activity was determined by the critical dilution method (16) and expressed either as arbitrary antibiotic units (AU) per milliliter according to the last inhibition halo observed in the dilutions or as a specific activity (AU/micrograms). The MccE492 concentration was determined spectrophotometrically as follows: concentration (in μg/ml) = 1,000 × [(A280 × 1.55) − (A260 × 0.76)].
Figure 1 A shows that cells with a single mutation on Fiu or Cir do not present a significant difference in the sensitivity to MccE492 at the different concentrations tested compared to the wild-type strain. The bacterial concentration necessary to produce 50% of cellular death (BC50) (Table 1) calculated were used only for comparative purposes because they depend on the specific activity of the microcin used, and for this reason the same pool of MccE492 preparation was used in the experiments presented in Fig. 1A and B. The BC50 obtained for the wild type and the single- and double-mutant strains that conserved the FepA receptor were similar (0.14 to 0.17 μM MccE492), indicating that most of MccE492 uptake is carried out by FepA. The contribution of the other receptors is clearly established in Fig. 1B. When mutants carrying Fiu or Cir as the only receptor were used as sensitive cells, the BC50 was 4 to 46 times higher, respectively, than the wild type (Table 1). When these values are compared to that obtained from cells carrying both receptors, Fiu and Cir, a 2.3-fold increase in the BC50 was observed. Hence, the contribution of receptors Fiu and Cir on MccE492 activity is more than that of the single addition. Although it is evident that FepA is the main receptor, in its absence the contribution of MccE492 uptake is highly enhanced when Fiu and Cir are combined than when they are presented as single receptors. This synergy has been observed in several receptor systems and can be explained through the model of “conformational spread” (2), in which coupling between adjacent receptors alters the activity of other receptors in their immediate vicinity. As a consequence, the response of a group of receptors would be greater than that of an equivalent set of isolated receptors, increasing the sensitivity. The synergy between Fiu and Cir and the preference of MccE492 for FepA were also observed when the cells were grown in DTPA, an iron chelator in which maximal expression of the receptors is achieved (8, 17).
Contribution of receptors FepA, Fiu, and Cir to microcin E492 uptake and its dispensability in outer membrane permeabilized cells. (A and B) Bacterial strains were grown in minimal medium to 5 × 108 cells/ml and treated with increasing concentrations of a MccE492 preparation with a specific activity of 180 AU/μg. The percentage of cell death was calculated relative to the viability of the cells without MccE492 treatment. E. coli H1443 is the parental strain with the three receptors (•), and E. coli H1876 was used as the triple-mutant strain (FepA− Cir− Fiu−) (○). (A) E. coli mutant strains in one receptor: H873 (FepA−) (▿), H1594 (Fiu−) (□), and H800 (Cir−) (⋄). (B) E. coli mutant strains in two receptors: H1728 (Fiu− Cir−) (▾), H1875 (FepA− Cir−) (▪), and H1877 (FepA− Fiu−) (♦). (C) The specific activity of MccE492 used was 39 AU/μg, and the concentration of spheroplasts used was 2 × 106 cells/ml. Symbols: E. coli H1876 (FepA− Cir− Fiu−) with (▵) or without (○) treatment to form spheroplasts; E. coli H1443 (wild type) with (▴) or without (•) treatment to form spheroplasts. Mcc, microcin E492. Error bars represent the standard deviations from three determinations.
Microcin concentration necessary to produce 50% of death on wild-type and mutant receptor strainsa
The curves in Fig. 1A and B show the cooperativity with a Hill coefficient between 2.4 and 3.5, indicating that conformational changes are involved. The transport data of ferric enterochelin by FepA also showed sigmoidal uptake kinetics with a Hill coefficient of approximately 3 (26) that correlates with the native FepA protein, a trimer that would have three interacting cooperative binding sites (15). A possible explanation for the specificity in the interaction between MccE492 and its receptors has been given by Braun et al. (1). There are three microcins that are recognized by FepA, Fiu, and Cir: microcin E492, H47, and M (20, 27). Although they are not homologs, the C termini of these three microcins are very similar, all of them being serine-rich, which in turn is similar to the backbone of catecholate siderophores. This region could be the binding determinant that would establish the receptor specificity.
MccE492 dependence on TonB is uncoupled from the outer membrane receptors.
In a previous study it was demonstrated that MccE492 needs TonB at the level of the cytoplasmic membrane of the target cells, because no bactericidal activity was observed when the outer membrane of TonB− cells were permeabilized (12). To determine whether an interaction between the receptors and TonB is necessary for MccE492 activity, an experiment with spheroplasts in which the requirement for the outer membrane receptors can be bypassed was performed. The outer membrane of mutant cells lacking the three receptors were permeabilized by using a procedure to form spheroplasts (12, 28), and the sensitivity to increasing concentrations of MccE492 was tested. Figure 1C shows that permeabilized triple-mutant cells are sensitive to MccE492, indicating that there is no need of an interaction between the receptors and the periplasmic portion of TonB for the bactericidal action of MccE492. The curves obtained for the treated cells, both wild type and triple mutant, were almost identical. The toxicity of MccE492 on the spheroplasts was 10 times higher than on the nontreated wild-type control, and the cooperativity observed in the untreated cells was almost completely lost, which is consistent with the notion that the interaction of different conformations is at the level of the receptors in the outer membrane. This result also indicates that the limiting step for TonB-mediated interaction of MccE492 with the cytoplasmic membrane is the internalization of microcin by the receptors. This is further supported by the fact that there are no differences in the BC50 observed in spheroplasts prepared from cells grown in the presence or absence of DTPA (data not shown). It is important to point out that these experiments were carried out with nonmodified MccE492 and, in the event that the modified form had a higher affinity for the receptors, a smaller difference in toxicity with spheroplasts versus nontreated cells as the indicator cells would be observed. The role of TonB in Mcc E492 activity is different from that of colicins B and D (1). The dependence on TonB of large bacteriocins that recognize catecholate receptors such as colicin B and D involves an interaction of the TonB box (24) of the receptor with the periplasm domain of TonB (9, 14). MccE492 does not have a distinguishable TonB box (11) and, consistent with this observation, treatment of sensitive cells with MccE492 combined with the pentapeptide ETVIV that corresponds to the consensus TonB box did not present any inhibitory effect on the bactericidal activity (not shown), as has been described for other proteins that are TonB dependent (14). The conclusion that MccE492 action is independent of the interaction of TonB with the receptors is further supported by the finding that a mutant on TonB that lacks the periplasmic domain necessary for such interaction is still sensitive to MccE492 bactericidal activity (M. Baeza and R. Lagos, unpublished results).
Dimer and trimer of DHBS inhibit MccE492 bactericidal action.
Enterochelin and the monomer, dimer, and trimer of DHBS used in the following experiments were purified by high-pressure liquid chromatography (30), and their purity and identification were assessed as described previously (18). To prevent the uptake of these siderophores that could lower the effective concentration of the DHBS derivatives during the inhibition experiments, E. coli AN260 (FepC−) (13), a strain that is unable to import the siderophores from the periplasm to the cytoplasm, was used as an indicator strain. Incubation with the precursor DHBA at up to 1 mM had no inhibitory effect, whereas incubation with the monomer of DHBS showed a slight inhibition at 0.5 mM. In contrast, enterochelin and the dimer and trimer of DHBS strongly inhibited MccE492 bactericidal activity (Fig. 2) with no significant variations in the concentration necessary to reach a 50% inhibition (IC50 values of 0.29, 0.17, and 0.15 μM, respectively). The observation that DHBA does not inhibit MccE492 uptake can be explained because there is no competition with MccE492. This assumption is based on the fact that this compound is not imported efficiently by FepA (6), probably due to a poor binding affinity. The small inhibition observed by the DHBS monomer could be explained also as a differential affinity between the monomer and the receptors or that the monomer can be internalized in the nonferric form. No inhibition of MccE492 bactericidal activity by enterochelin and the dimer and trimer of DHBS was observed when spheroplasts were used as sensitive cells (Table 2), indicating that the inhibition by these compounds is at the level of receptor uptake, probably through competition for the binding site between the siderophores and MccE492. The untreated cells incubated with MccE492 plus the respective siderophore presented a 80 to 90% protection against the bactericidal activity.
Effect of increasing concentrations of the dimer and trimer of DHBS on microcin E492 bactericidal activity. DHBS2 (DHBS dimer), DHBS3 (trimer), and Ent (enterochelin) were incubated for 15 min at 37°C with MccE492 (0.15 μM; specific activity, 49 AU/μg), further incubated with 107 cells of E. coli AN260 (FepC−) at 37°C for 15 min, and plated for survivors. The indicated symbols are used for DHBS2 (▴), DHBS3 (□), and enterochelin (○). The data were fit by using the program “standard curve, four-parameter logistic curve” equation (Sigma Plot 8.0). Error bars represent the standard deviation from three determinations.
Effect of enterochelin and its hydrolysis products on spheroplasts survivala
Conclusions.
The recognition of MccE492 by its receptors seems to involve interacting cooperative binding sites. The translocation to the periplasm that involves the formation of MccE492-receptor complexes in the outer membrane would be the limiting step in the bactericidal process. The interaction of MccE492 with the TonB system is independent on the presence of the receptors. MccE492 prefers FepA as the main receptor, followed by Fiu and to a minor extent by Cir. In the absence of FepA, a strong enhancement in MccE492 uptake by Fiu plus Cir was observed. The uptake by the receptors is prevented by enterochelin and the dimer and trimer of DHBS.
ACKNOWLEDGMENTS
We thank Roselyn Orellana for technical assistance.
E.S. was the recipient of predoctoral CONICYT and MECESUP fellowships. This work was supported by grant 1020757 from the Fondo Nacional de Desarrollo Científico y Tecnológico.
FOOTNOTES
- Received 8 February 2005.
- Returned for modification 11 March 2005.
- Accepted 18 March 2005.
- Copyright © 2005 American Society for Microbiology