Structural design principles for specific ultra-high affinity interactions between colicins/pyocins and immunity proteins.

The interactions of the antibiotic proteins colicins/pyocins with immunity proteins is a seminal model system for studying protein-protein interactions and specificity. Yet, a precise and quantitative determination of which structural elements and residues determine their binding affinity and specificity is still lacking. Here, we used comparative structure-based energy calculations to map residues that substantially contribute to interactions across native and engineered complexes of colicins/pyocins and immunity proteins. We show that the immunity protein α1-α2 motif is a unique structurally-dissimilar element that restricts interaction specificity towards all colicins/pyocins, in both engineered and native complexes. This motif combines with a diverse and extensive array of electrostatic/polar interactions that enable the exquisite specificity that characterizes these interactions while achieving ultra-high affinity. Surprisingly, the divergence of these contributing colicin residues is reciprocal to residue conservation in immunity proteins. The structurally-dissimilar immunity protein α1-α2 motif is recognized by divergent colicins similarly, while the conserved immunity protein α3 helix interacts with diverse colicin residues. Electrostatics thus plays a key role in setting interaction specificity across all colicins and immunity proteins. Our analysis and resulting residue-level maps illuminate the molecular basis for these protein-protein interactions, with implications for drug development and rational engineering of these interfaces.

Lipopolysaccharides promote binding and unfolding of the antibacterial colicin E3 rRNAse domain.

Nuclease colicins are antibacterial proteins produced by certain strains of E. coli to reduce competition from rival strains. These colicins are generally organized with an N-terminal transport (T)-domain, a central receptor binding (R)-domain, and a C-terminal cytotoxic nuclease domain. These colicins are always produced in complex with an inhibitory immunity protein, which dissociates prior entrance of the cytotoxic domain in the target cell. How exactly colicins traverse the cell envelope is not understood, yet this knowledge is important for the design of new antibacterial therapies. In this report, we find that the cytotoxic rRNAse domain of colicin E3, lacking both T- and R-domains, is sufficient to inhibit cell growth provided the immunity protein Im3 has been removed. Thus, while the T-domain is needed for dissociation of Im3, the rRNAse alone can associate to the cell surface without R-domain. Accordingly, we find a high affinity interaction (Kd ~1-2µM) between the rRNAse domain and lipopolysaccharides (LPS). Furthermore, we show that binding of ColE3 to LPS destabilizes the secondary structure of the toxin, which is expectedly crucial for transport through the narrow pore of the porin OmpF. The effect of LPS on binding and unfolding of ColE3 may be indicative of a broader role of LPS for transport of colicins in general.

A Chimeric protein of CFA/I, CS6 subunits and LTB/STa toxoid protects immunized mice against enterotoxigenic Escherichia coli.

Enterotoxigenic Escherichia Coli (ETEC) strains are the commonest bacteria causing diarrhea in children in developing countries and travelers to these areas. Colonization factors (CFs) and enterotoxins are the main virulence determinants in ETEC pathogenesis. Heterogeneity of CFs is commonly considered the bottleneck to developing an effective vaccine. It is believed that broad spectrum protection against ETEC would be achieved by induced anti-CF and anti-enterotoxin immunity simultaneously. Here, a fusion antigen strategy was used to construct a quadrivalent recombinant protein called 3CL and composed of CfaB, a structural subunit of CFA/I, and CS6 structural subunits, LTB and STa toxoid of ETEC. Its anti-CF and antitoxin immunogenicity was then assessed. To achieve high-level expression, the 3CL gene was synthesized using E. coli codon bias. Female BALB/C mice were immunized with purified recombinant 3CL. Immunized mice developed antibodies that were capable of detecting each recombinant subunit in addition to native CS6 protein and also protected the mice against ETEC challenge. Moreover, sera from immunized mice also neutralized STa toxin in a suckling mouse assay. These results indicate that 3CL can induce anti-CF and neutralizing antitoxin antibodies along with introducing CFA/I as a platform for epitope insertion.

Immunity protein release from a cell-bound nuclease colicin complex requires global conformational rearrangement.

Nuclease colicins bind their target receptor BtuB in the outer membrane of sensitive Escherichia coli cells in the form of a high-affinity complex with their cognate immunity proteins. The release of the immunity protein from the colicin complex is a prerequisite for cell entry of the colicin and occurs via a process that is still relatively poorly understood. We have previously shown that an energy input in the form of the cytoplasmic membrane proton motive force is required to promote immunity protein (Im9) release from the colicin E9/Im9 complex and colicin cell entry. We report here that engineering rigidity in the structured part of the colicin translocation domain via the introduction of disulfide bonds prevents immunity protein release from the colicin complex. Reduction of the disulfide bond by the addition of DTT leads to immunity protein release and resumption of activity. Similarly, the introduction of a disulfide bond in the DNase domain previously shown to abolish channel formation in planar bilayers also prevented immunity protein release. Importantly, all disulfide bonds, in the translocation as well as the DNase domain, also abolished the biological activity of the Im9-free colicin E9, the reduction of which led to a resumption of activity. Our results show, for the first time, that conformational flexibility in the structured translocation and DNase domains of a nuclease colicin is essential for immunity protein release, providing further evidence for the hypothesis that global structural rearrangement of the colicin molecule is required for disassembly of this high-affinity toxin-immunity protein complex prior to outer membrane translocation.

Modular structure of microcin H47 and colicin V.

Microcins are gene-encoded peptide antibiotics produced by enterobacteria that act on strains of gram-negative bacteria. In this work, we concentrated on higher-molecular-mass microcins, i.e., those possessing 60 or more amino acids. They can be subdivided into unmodified and posttranslationally modified peptides. In both cases, they exhibit conserved C-terminal sequences that appear to be characteristic of each subgroup. In the hypothesis that these sequences could correspond to domains, gene fusions between the activity genes for the unmodified microcin colicin V and the modified microcin H47 were constructed. These two microcins differ in their mode of synthesis, uptake, target, and specific immunity. Through this experimental approach, chimeric peptides with exchanged C-terminal sequences were encoded. Cells carrying the fusions in different genetic contexts were then assayed for antibiotic production. Many of them produced antibiotic activities with recombinant properties: the toxicity of one microcin and the mode of uptake of the other. The results led to the identification of a modular structure of colicin V and microcin H47, with the recognition of two domains in their peptide chains: a toxic N-terminal domain and an uptake C-terminal domain. This modular design would be shared by other microcins from each subgroup.

Cyclic peptides selected by phage display mimic the natural epitope recognized by a monoclonal anti-colicin A antibody.

A 10-mer random peptide library displayed on filamentous bacteriophage was used to determine the molecular basis of the interaction between the monoclonal anti-colicin A antibody 1C11 and its cognate epitope. Previous studies established that the putative epitope recognized by 1C11 antibody is composed of amino acid residues 19-25 (RGSGPEP) of colicin A. Using the phage display technique it was confirmed that the epitope of 1C11 antibody was indeed restricted to residues 19-25 and the consensus motif RXXXPEP was identified. Shorter consensus sequences (RXXPEP, RXXEP, KXXEP) were also selected. It was also demonstrated that the disulfide bond found in one group of the selected peptides was crucial for 1C11 antibody recognition. It was shown that cyclization of the peptides by disulfide bond formation could result in a structure that mimics the natural epitope of colicin A.

Involvement of colicin in the limited protection of the colicin producing cells against bacteriophage.

The restriction/modification system is considered to be the most common machinery of microorganisms for protection against bacteriophage infection. However, we found that mitomycin C induced Escherichia coli containing ColE7-K317 can confer limited protection against bacteriophage M13K07 and lambda infection. Our study showed that degree of protection is correlated with the expression level of the ColE7 operon, indicating that colicin E7 alone or the colicin E7-immunity protein complex is directly involved in this protection mechanism. It was also noted that the degree of protection is greater against the single-strand DNA bacteriophage M13K07 than the double-strand bacteriophage(lambda). Coincidently, the K(A) value of ColE7-Im either interacting with single-strand DNA (2.94x10(5)M(-1)) or double-strand DNA (1.75x10(5)M(-1)) reveals that the binding affinity of ColE7-Im with ssDNA is 1.68-fold stronger than that of the protein complex interacting with dsDNA. Interaction between colicin and the DNA may play a central role in this limited protection of the colicin-producing cell against bacteriophages. Based on these observations, we suggest that the colicin exporting pathway may interact to some extent with the bacteriophage infection pathway leading to a limited selective advantage for and limited protection of colicin-producing cells against different bacteriophages.

Structural inhibition of the colicin D tRNase by the tRNA-mimicking immunity protein.

Colicins are toxins secreted by Escherichia coli in order to kill their competitors. Colicin D is a 75 kDa protein that consists of a translocation domain, a receptor-binding domain and a cytotoxic domain, which specifically cleaves the anticodon loop of all four tRNA(Arg) isoacceptors, thereby inactivating protein synthesis and leading to cell death. Here we report the 2.0 A resolution crystal structure of the complex between the toxic domain and its immunity protein ImmD. Neither component shows structural homology to known RNases or their inhibitors. In contrast to other characterized colicin nuclease-Imm complexes, the colicin D active site pocket is completely blocked by ImmD, which, by bringing a negatively charged cluster in opposition to a positively charged cluster on the surface of colicin D, appears to mimic the tRNA substrate backbone. Site-directed mutations affecting either the catalytic domain or the ImmD protein have led to the identification of the residues vital for catalytic activity and for the tight colicin D/ImmD interaction that inhibits colicin D toxicity and tRNase catalytic activity.

Assembly of colicin A in the outer membrane of producing Escherichia coli cells requires both phospholipase A and one porin, but phospholipase A is sufficient for secretion.

Three oligomeric forms of colicin A with apparent molecular masses of about 95 to 98 kDa were detected on sodium dodecyl sulfate (SDS)-polyacrylamide gels loaded with unheated samples from colicin A-producing cells of Escherichia coli. These heat-labile forms, called colicins Au, were visualized both on immunoblots probed with monoclonal antibodies against colicin A and by radiolabeling. Cell fractionation studies show that these forms of colicin A were localized in the outer membrane whether or not the producing cells contained the cal gene, which encodes the colicin A lysis protein responsible for colicin A release in the medium. Pulse-chase experiments indicated that their assembly into the outer membrane, as measured by their heat modifiable migration in SDS gels, was an efficient process. Colicins Au were produced in various null mutant strains, each devoid of one major outer membrane protein, except in a mutant devoid of both OmpC and OmpF porins. In cells devoid of outer membrane phospholipase A (OMPLA), colicin A was not expressed. Colicins Au were detected on immunoblots of induced cells probed with either polyclonal antibodies to OmpF or monoclonal antibodies to OMPLA, indicating that they were associated with both OmpF and OMPLA. Similar heat-labile forms were obtained with various colicin A derivatives, demonstrating that the C-terminal domain of colicin A, but not the hydrophobic hairpin present in this domain, was involved in their formation.

Competition by allelopathy proceeds in traveling waves: colicin-immune strain aids colicin-sensitive strain.

Producing toxic chemicals to suppress both the growth and survivorship of local competitors is called allelopathy; some strains of the bacteria Escherichia coli produce a toxin (named colicin) which may kill colicin-sensitive neighbors while they themselves are immune. In a previous paper, the competitive outcome between colicin-producing and colicin-sensitive strains was shown to differ between a spatially structured and a completely mixed population. In this paper, we analyze the role of a third, "colicin-immune," strain, which does not produce colicin but is immune to it. Without spatial structure, the colicin-immune strain suppresses the colicin-producing strain and enables the colicin-sensitive strain to win. In a spatially structured population, modeled as a reaction-diffusion system, we examine the speed of boundaries between areas dominated by different strains in traveling waves and the events after the collision of two such boundaries. The colicin-immune strain passes through the area dominated by the colicin-sensitive strain and drives the colicin-producing strain to extinction. Subsequently the colicin-sensitive strain occupies the whole population.

The purified colicin S8 is a multimeric protein.

Bacteriocins have been isolated both as simple proteins and as proteins in association with carbohydrates, lipids, etc. Colicins are commonly inducible and extracellular. Their molecular masses range from 30 to 90 kDa. Pure colicin S8 was obtained in three steps from supernatant of induced cells: (i) Ammonium sulfate precipitation; (ii) anion exchange chromatography; and (iii) phenyl-Sepharose hydrophobic chromatography, either by preparative or fast performance liquid chromatography (FPLC) analytical purification procedure. In our hands, purified colicin S8 was an aggregation of extremely related polypeptides. Composition of those active fractions was the same: five polypeptides of molecular weight around 55 kDa. Behavior on molecular filtration indicated a molecular weight higher than 200 kDa. Similar results were obtained when purification was carried out through FPLC. Producing strains contain a single plasmid that encodes colicin S8; in minicells, this plasmid was shown to specify a 60 kDa polypeptide. We conclude that more than one form of colicin S8 exists. The forms are structurally related and can be recognized by antibodies raised against one of the polypeptides. Consistent with this conclusion, comparison of peptides produced after hydrolysis with chlorosuccinamide indicated that the active proteins contained both shared and unique components.

Hierarchical order of critical residues on the immunity-determining region of the Im7 protein which confer specific immunity to its cognate colicin.

The directed mutagenesis study of the Im7 protein of colicin E7 revealed that three residues, D31, D35, and E39, located in the loop 1 and helix 2 regions of the protein were critical for initiating the complex formation with its cognate colicin E7. Interestingly, the importance of these three critical residues in conferring specific immunity to its own colicin was exhibited in a hierarchical order, respectively. Moreover, we found that existence of the three critical residues was common among the DNase-type Im proteins. Most likely the three residues of the DNase-type immunity proteins are critical for initiating the unique protein-protein interactions with their cognate colicin. In addition, replacement of the helix 2 of Im7 by the corresponding region of Im8 produced a phenotype of the mutant protein very similar to that of Im8. This result suggests that the DNase-type Im proteins indeed share a "homologous-structural framework" and evolution of the Im proteins may be engendered by minor amino acid changes in this specific immunity-determining region without causing structural alteration of the proteins.

Resistência a drogas e produçäo de bacteriocinas em linhagens de Escherichia coli e Enterobacter isoladas de indivíduos hospitalizados / Drug resistance and colicin production by Escherichia coli and Enterobacter isolated from hospitalized humans

Escherichia coli e Enterobacter foram isolados de grupos de indivíduos hospitalizados (H), recém-hospitalizados (P), sendo estudados quanto ao perfil de resistência às drogas ampicilina, cefalotina, cloranfenicol, estreptomicina, tetraciclina, gentamicina, canamicina e ao bicloreto de mercúrio, por meio da técnica de diluiçäo em meio sólido, e quanto à produçäo de colicinas. A resistência aos antibióticos betalactâmicos foi maior nas amostras isoladas dos grupos de portadores näo-hospitalizados (P) (modelo ampicilina-cefalosporina), em especial para o gênero Enterobacter. Por outro lado, a freqüencia das amostras colicinogênicas descresceu no grupo H em comparaçäo com amostras colicinogênicas decresceu no grupo H em comparaçäo com as amostras do grupo P. É provável que, em ambientes seletivos pela presença de elevadas concentraçöes de antibióticos, as linhagens portadoras de fenótipos sejam selecionadas e a colicinogênese, deslocada ou substituída, sendo mantida em ambientes näo-seletivos, devido à competitividade as linhagens de uma mesma espécie

Colicin diversity: a result of eco-evolutionary dynamics.

Colicins are plasmids that are carried in Escherichia coli. They code for a toxic protein and for proteins that confer on the host immunity against this toxin. When bacteria carry plasmids their growth rate is reduced. At the same time, the production of toxins makes it possible for colicinogenic bacteria to invade bacterium strains that are not immune. In natural bacterium populations there is a high diversity of colicin types. The reason for the maintenance of this diversity has been the subject of much recent debate. We have studied a simple eco-evolutionary model of the interaction of bacteria with colicins and show that high diversity of colicins is to be expected. We find two different dynamical modes each with a high diversity: a hyperimmunity mode and a multitoxicity mode. Bacteria are immune to most toxins in the first mode but in fact produce very few toxins. In the second mode bacteria are immune only to those toxins that they actually produce. In the second mode toxin levels per bacterium are much higher, whereas immunity levels per bacterium are lower.

Integration of the colicin A pore-forming domain into the cytoplasmic membrane of Escherichia coli.

The pore-forming domain of colicin A (pfColA) fused to a prokaryotic signal peptide (sp-pfColA) inserted into the inner membrane of Escherichia coli and apparently formed a functional channel, when generated in vivo. We investigated pfColA functional activity in vivo by the PhoA gene fusion approach, combined with cell fractionation and protease susceptibility experiments. Alkaline phosphatase was fused to the carboxy-terminal end of each of the ten alpha-helices of sp-pfColA to form a series of differently sized fusion proteins. We suggest that the alpha-helices anchoring pfColA in the membrane are first translocated into the periplasm. We identify two domains that anchor pfColA to the membrane in vivo: domain 1, extending from helix 1 to helix 8, which contains the voltage-responsive segment and domain 2 consisting of the hydrophobic helices 8 and 9. These two domains function independently. Fusion proteins with a mutation inactivating the voltage-responsive segment or with a domain 1 lacking helix 8 were peripherally associated with the outside of the inner membrane, and were therefore digested by proteases added to spheroplasts. In contrast, fusion proteins with a functional domain 1 were protected from proteases, suggesting as expected that most of domain 1 is inserted into the membrane or is indeed translocated to the cytoplasm during pfColA channel opening.

The tip of the hydrophobic hairpin of colicin U is dispensable for colicin U activity but is important for interaction with the immunity protein.

The hydrophobic C terminus of pore-forming colicins associates with and inserts into the cytoplasmic membrane and is the target of the respective immunity protein. The hydrophobic region of colicin U of Shigella boydii was mutated to identify determinants responsible for recognition of colicin U by the colicin U immunity protein. Deletion of the tip of the hydrophobic hairpin of colicin U resulted in a fully active colicin that was no longer inactivated by the colicin U immunity protein. Replacement of eight amino acids at the tip of the colicin U hairpin by the corresponding amino acids of the related colicin B resulted in colicin U(575-582ColB), which was inactivated by the colicin U immunity protein to 10% of the level of inactivation of the wild-type colicin U. The colicin B immunity protein inactivated colicin U(575-582ColB) to the same degree. These results indicate that the tip of the hydrophobic hairpin of colicin U and of colicin B mainly determines the interaction with the corresponding immunity proteins and is not required for colicin activity. Comparison of these results with published data suggests that interhelical loops and not membrane helices of pore-forming colicins mainly interact with the cognate immunity proteins and that the loops are located in different regions of the A-type and E1-type colicins. The colicin U immunity protein forms four transmembrane segments in the cytoplasmic membrane, and the N and C termini face the cytoplasm.

Intracellular immunization of prokaryotic cells against a bacteriotoxin.

Intracellularly expressed antibodies have been designed to bind and inactivate target molecules inside eukaryotic cells. Here we report that an antibody fragment can be used to probe the periplasmic localization of the colicin A N-terminal domain. Colicins form voltage-gated ion channels in the inner membrane of Escherichia coli. To reach their target, they bind to a receptor located on the outer membrane and then are translocated through the envelope. The N-terminal domain of colicins is involved in the translocation step and therefore is thought to interact with proteins of the translocation system. To compete with this system, a single-chain variable fragment (scFv) directed against the N-terminal domain of the colicin A was synthesized and exported into the periplasmic space of E. coli. The periplasmic scFv inhibited the lethal activity of colicin A and had no effect on the lethal activity of other colicins. Moreover, the scFv was able to specifically inactivate hybrid colicins possessing the colicin A N-terminal domain without affecting their receptor binding. Hence, the periplasmic scFv prevents the translocation of colicin A and probably its interaction with import machinery. This indicates that the N-terminal domain of the toxin is accessible in the periplasm. Moreover, we show that production of antibody fragments to interfere with a biological function can be applied to prokaryotic systems.

The N-terminal domain of colicin E3 interacts with TolB which is involved in the colicin translocation step.

Colicins use two envelope multiprotein systems to reach their cellular target in susceptible cells of Escherichia coli: the Tol system for group A colicins and the TonB system for group B colicins. The N-terminal domain of colicins is involved in the translocation step. To determine whether it interacts in vivo with proteins of the translocation system, constructs were designed to produce and export to the cell periplasm the N-terminal domains of colicin E3 (group A) and colicin B (group B). Producing cells became specifically tolerant to entire extracellular colicins of the same group. The periplasmic N-terminal domains therefore compete with entire colicins for proteins of the translocation system and thus interact in situ with these proteins on the inner side of the outer membrane. In vivo cross-linking and co-immunoprecipitation experiments in cells producing the colicin E3 N-terminal domain demonstrated the existence of a 120 kDa complex containing the colicin domain and TolB. After in vitro cross-linking experiments with these two purified proteins, a 120 kDa complex was also obtained. This suggests that the complex obtained in vivo contains exclusively TolB and the colicin E3 domain. The N-terminal domain of a translocation-defective colicin E3 mutant was found to no longer interact with TolB. Hence, this interaction must play an important role in colicin E3 translocation.