Conformational changes of colicin Ia channel-forming domain upon membrane binding: a solid-state NMR study

Channel-forming colicins are bactericidal proteins that spontaneously insert into hydrophobic lipid bilayers. We have used magic-angle spinning solid-state nuclear magnetic resonance spectroscopy to examine the conformational differences between the water-soluble and the membrane-bound states of colicin Ia channel domain, and to study the effect of bound colicin on lipid bilayer structure and dynamics. We detected (13)C and (15)N isotropic chemical shift differences between the two forms of the protein, which indicate structural changes of the protein due to membrane binding. The Val C(alpha) signal, unambiguously assigned by double-quantum experiments, gave a 0.6 ppm downfield shift in the isotropic position and a 4 ppm reduction in the anisotropic chemical shift span after membrane binding. These suggest that the alpha-helices in the membrane-bound colicin adopt more ideal helical torsion angles as they spread onto the membrane. Colicin binding significantly reduced the lipid chain order, as manifested by (2)H quadrupolar couplings. These results are consistent with the model that colicin Ia channel domain forms an extended helical array at the membrane-water interface upon membrane binding.

The zinc ion in the HNH motif of the endonuclease domain of colicin E7 is not required for DNA binding but is essential for DNA hydrolysis

The HNH motif was originally identified in the subfamily of HNH homing endonucleases, which initiate the process of the insertion of mobile genetic elements into specific sites. Several bacteria toxins, including colicin E7 (ColE7), also contain the 30 amino acid HNH motif in their nuclease domains. In this work, we found that the nuclease domain of ColE7 (nuclease-ColE7) purified from Escherichia coli contains a one-to-one stoichiometry of zinc ion and that this zinc-containing enzyme hydrolyzes DNA without externally added divalent metal ions. The apo-enzyme, in which the indigenous zinc ion was removed from nuclease-ColE7, had no DNase activity. Several divalent metal ions, including Ni2+, Mg2+, Co2+, Mn2+, Ca2+, Sr2+, Cu2+ and Zn2+, re-activated the DNase activity of the apo-enzyme to various degrees, however higher concentrations of zinc ion inhibited this DNase activity. Two charged residues located at positions close to the zinc-binding site were mutated to alanine. The single-site mutants, R538A and E542A, showed reduced DNase activity, whereas the double-point mutant, R538A + E542A, had no observable DNase activity. A gel retardation assay further demonstrated that the nuclease-ColE7 hydrolyzed DNA in the presence of zinc ions, but only bound to DNA in the absence of zinc ions. These results demonstrate that the zinc ion in the HNH motif of nuclease-ColE7 is not required for DNA binding, but is essential for DNA hydrolysis, suggesting that the zinc ion not only stabilizes the folding of the enzyme, but is also likely to be involved in DNA hydrolysis.

Membrane protein topology probed by (1)H spin diffusion from lipids using solid-state NMR spectroscopy

We describe a two-dimensional solid-state NMR technique to investigate membrane protein topology under magic-angle spinning conditions. The experiment detects the rate of (1)H spin diffusion from the mobile lipids to the rigid protein. While spin diffusion within the rigid protein is fast, magnetization transfer in the mobile lipids is an inefficient and slow process. Qualitative analysis of (1)H spin-diffusion build-up curves from the lipid chain-end methyl groups to the protein allows the identification of membrane-embedded domains in the protein. Numerical simulations of spin-diffusion build-up curves yield the approximate insertion depth of protein segments in the membrane. The experiment is demonstrated on the selectively (13)C labeled colicin Ia channel domain, known to have a membrane-embedded domain, and on DNA/cationic lipid complexes where the DNA rods are bound to the membrane surface. The experiment is designed for X-nucleus detection, which could be (13)C or (15)N in the protein and (31)P for the DNA. Finally, we show that a qualitative distinction between membrane proteins with and without a membrane-embedded domain can be made even by using an unlabeled protein, by detection of lipid signals. This spin-diffusion experiment is simple to perform and requires no oriented bilayer preparations and only standard NMR hardware.

Translocation of a functional protein by a voltage-dependent ion channel

The voltage-dependent gating of the colicin channel involves a substantial structural rearrangement that results in the transfer of about 35% of the 200 residues in its pore-forming domain across the membrane. This transfer appears to represent an unusual type of protein translocation that does not depend on a large, multimeric, protein pore. To investigate the ability of this system to transport arbitrary proteins, we made use of a pair of strongly interacting proteins, either of which could serve as a translocated cargo or as a probe to detect the other. Here we show that both an 86-residue and a 134-residue hydrophilic protein inserted into the translocated segment of colicin A are themselves translocated and are functional on the trans side of the bilayer. The disparate features of these proteins suggest that the colicin channel has a general protein translocation mechanism.

Surface aspartate residues are essential for the stability of colicin A P-domain: a mechanism for the formation of an acidic molten-globule

The pore-forming domains of members of a family of bacterial toxins, colicins N and A, share > 50% sequence identity, identical folds and yet display strikingly different behavior in acid conditions. At low pH colicin A forms a molten globule state while colicin N retains a native fold. This is relevant to in vivo activity since colicin A requires acidic phospholipids for its toxic activity but colicin N does not. The pI of colicin A (5.25) is far lower than that of colicin N (10.2) because colicin A contains seven extra aspartate residues. We first introduced separately each of these acidic amino acids into homologous sites in colicin N, but none caused destabilization at low pH. However, in the reverse experiment, the sequential replacement of these acidic side chains of colicin A by alanine revealed six sites where this change destabilized the protein at neutral pH. Some of these residues, which each contribute less than 4% to the total negative charge, appear to stabilize the protein via a network of hydrogen bonds and charge pairs which are sensitive to protonation. Other residues have no clear interactions that explain their importance. The colicin A is thus a protein that relies upon acid sensitive interactions for its stability at neutral pH and its in vivo activity.

Microcin E492 is an unmodified peptide related in structure to colicin V

The pore-forming microcin E492 was purified by solid-phase extraction and reversed-phase high-pressure liquid chromatography. Its molecular mass was 7,886 Da. The entire 84-amino-acid sequence was determined. There is no postranslational modification in the secreted microcin, and the sequence has homologies with the sequence of the microcin colicin V.

Mechanism and cleavage specificity of the H-N-H endonuclease colicin E9

Colicin endonucleases and the H-N-H family of homing enzymes share a common active site structural motif that has similarities to the active sites of a variety of other nucleases such as the non-specific endonuclease from Serratia and the sequence-specific His-Cys box homing enzyme I-PpoI. In contrast to these latter enzymes, however, it remains unclear how H-N-H enzymes cleave nucleic acid substrates. Here, we show that the H-N-H enzyme from colicin E9 (the E9 DNase) shares many of the same basic enzymological characteristics as sequence-specific H-N-H enzymes including a dependence for high concentrations of Mg2+ or Ca2+ with double-stranded substrates, a high pH optimum (pH 8-9) and inhibition by monovalent cations. We also show that this seemingly non-specific enzyme preferentially nicks double-stranded DNA at thymine bases producing 3'-hydroxy and 5'-phosphate termini, and that the enzyme does not cleave small substrates, such as dinucleotides or nucleotide analogues, which has implications for its mode of inhibition in bacteria by immunity proteins. The E9 DNase will also bind single-stranded DNA above a certain length and in a sequence-independent manner, with transition metals such as Ni2+ optimal for cleavage but Mg2+ a poor cofactor. Ironically, the H-N-H motif of the E9 DNase although resembling the zinc binding site of a metalloenzyme does not support zinc-mediated hydrolysis of any DNA substrate. Finally, we demonstrate that the E9 DNase also degrades RNA in the absence of metal ions. In the context of current structural information, our data show that the H-N-H motif is an adaptable catalytic centre able to hydrolyse nucleic acid by different mechanisms depending on the substrate and metal ion regime.

Crystal structure of colicin E3: implications for cell entry and ribosome inactivation

Colicins kill E. coli by a process that involves binding to a surface receptor, entering the cell, and, finally, intoxicating it. The lethal action of colicin E3 is a specific cleavage in the ribosomal decoding A site. The crystal structure of colicin E3, reported here in a binary complex with its immunity protein (IP), reveals a Y-shaped molecule with the receptor binding domain forming a 100 A long stalk and the two globular heads of the translocation domain (T) and the catalytic domain (C) comprising the two arms. Active site residues are D510, H513, E517, and R545. IP is buried between T and C. Rather than blocking the active site, IP prevents access of the active site to the ribosome.

Colicin A immunity protein interacts with the hydrophobic helical hairpin of the colicin A channel domain in the Escherichia coli inner membrane

The colicin A pore-forming domain (pfColA) was fused to a bacterial signal peptide (sp-pfColA). This was inserted into the Escherichia coli inner membrane in functional form and could be coimmunoprecipitated with epitope-tagged immunity protein (EpCai). We constructed a series of fusion proteins in which various numbers of sp-pfColA alpha-helices were fused to alkaline phosphatase (AP). We showed that a fusion protein made up of the hydrophobic alpha-helices 8 and 9 of sp-pfColA fused to AP was specifically coimmunoprecipitated with EpCai produced in the same cells. This is the first biochemical evidence that Cai recognizes and interacts with the colicin A hydrophobic helical hairpin.

Import of colicins across the outer membrane of Escherichia coli involves multiple protein interactions in the periplasm

Several proteins of the Tol/Pal system are required for group A colicin import into Escherichia coli. Colicin A interacts with TolA and TolB via distinct regions of its N-terminal domain. Both interactions are required for colicin translocation. Using in vivo and in vitro approaches, we show in this study that colicin A also interacts with a third component of the Tol/Pal system required for colicin import, TolR. This interaction is specific to colicins dependent on TolR for their translocation, strongly suggesting a direct involvement of the interaction in the colicin translocation step. TolR is anchored to the inner membrane by a single transmembrane segment and protrudes into the periplasm. The interaction involves part of the periplasmic domain of TolR and a small region of the colicin A N-terminal domain. This region and the other regions responsible for the interaction with TolA and TolB have been mapped precisely within the colicin A N-terminal domain and appear to be arranged linearly in the colicin sequence. Multiple contacts with periplasmic-exposed Tol proteins are therefore a general principle required for group A colicin translocation.

Immunity proteins: enzyme inhibitors that avoid the active site

Immunity proteins are high affinity inhibitors of colicins--SOS-induced toxins released by bacteria during times of stress. Recent work has shown that nuclease-specific immunity proteins are exosite inhibitors, binding adjacent to the enzyme active site and inhibiting colicin activity indirectly. Unusually, their binding sites comprise a near contiguous sequence that lies N-terminal to active site sequences, raising the possibility that immunity proteins bind colicins co-translationally. Exosite binding accounts for the extensive sequence diversity seen at the interfaces of colicin-immunity protein complexes, which is not only a selective advantage to colicin-producing bacteria, but also represents a powerful model system for studying specificity in protein-protein recognition.

Tryptophan-dependent sensitized photoinactivation of colicin E1 channels in bilayer lipid membranes

The bacterial toxin colicin E1 is known to induce voltage-gated currents across a planar bilayer lipid membrane. In the present study, it is shown that the colicin-induced current decreased substantially upon illumination of the membrane in the presence of the photosensitizer, aluminum phthalocyanine. This effect was almost completely abolished by the singlet oxygen quencher, sodium azide. Using single tryptophan mutants of colicin E1, Trp495 was identified as the amino acid residue responsible for the sensitized photodamage of the colicin channel activity. Thus, the distinct participation of a specific amino acid residue in the sensitized photoinactivation of a defined protein function was demonstrated. It is suggested that Trp495 is critical for the translocation and/or anchoring of the colicin channel domain in the membrane.

Assembly of pore proteins on gold electrodes

Protein-protein interactions at the cell surface are important in the activity of bacterial toxins such as colicins. We have developed methods to study these events using tethered lipid bilayers, which can be probed by impedance spectroscopy and surface plasmon resonance. Recently we have attached the receptor proteins directly to gold electrodes and this offers new possibilities for measuring protein-protein interactions on solid supports.

Pore-forming colicins and their relatives

The pore-forming colicins, the first proteins that were capable of forming voltage-dependent ion channels to be sequenced, have turned out to be both less tractable and more mysterious than imagined; yet they have proved interesting at every step of their short journey from producing cell to vanquished target cell. Starting out as a remarkably extended water-soluble protein, the colicin molecule is designed to interact simultaneously with several components of the complex membrane of the target cell, transform itself into a membrane protein, and become an ion channel with inscrutable properties. Unraveling how it does all this appears to be leading us into the dark recesses of protein/protein and protein/membrane interaction, where lurk fundamental processes reluctantly waiting to be revealed.

Cleavage of colicin D is necessary for cell killing and requires the inner membrane peptidase LepB

Colicin D is known to kill target cells by cleaving tRNA(Arg). A colicin D-resistant mutant was selected that was altered in the inner membrane leader peptidase, LepB. The substituted residue (Asn274Lys) is located close to the catalytic site. The mutation abolishes colicin D cleavage but not the processing of exported proteins. LepB is required for colicin D cleavage, releasing a small C-terminal fragment that retains full tRNase activity. The immunity protein was found to prevent colicin D processing and furthermore masks tRNase activity, thus protecting colicin D against LepB-mediated cleavage during export. Catalytic colicins share a consensus sequence at their putative processing site. Mutations affecting normal processing of colicin D abolish cytotoxicity without affecting the in vitro tRNase activity.

Dipolar filtered 1H-13C heteronuclear correlation spectroscopy for resonance assignment of proteins

Resonance assignment is necessary for the comprehensive structure determination of insoluble proteins by solid-state NMR spectroscopy. While various 2D and 3D correlation techniques involving 13C and 15N spins have been developed for this purpose, H chemical shift has not been exploited sufficiently. We demonstrate the combination of the regular 1H-13C heteronuclear correlation (HETCOR) experiment and a dipolar filtered HETCOR technique to obtain better resolved 1H chemical shift spectra. The dipolar filtered experiment, MELODI-HETCOR. simplifies the 1H spectra by suppressing the directly bonded C-H correlation peaks and retaining only the medium- and long-range cross peaks. We apply this MELODI-HETCOR technique to several amino acids and proteins with various isotopic labeling patterns. The enhanced 1H chemical shift resolution allows the assignment of overlapping H alpha and H beta resonances in Ser, identifies the 1H chemical shift differences between neutral and cationic imidazole rings of His, and permits the assignment of residues with side chain nitrogen atoms in ubiquitin. The potential utility of this dipolar filtered HETCOR technique to resonance assignment of extensively labeled proteins is discussed.

Processing of DNase domain during translocation of colicin E7 across the membrane of Escherichia coli

Translocation of colicin across the membrane of sensitive cells has been studied extensively. However, processing of the toxicity domain of colicin during translocation has been the subject of much controversy. To investigate the final translocation product of colicin across the membrane of Escherichia coli, an endogenously expressed His-tagged Im7 protein was constructed to detect any translocation product containing the DNase domain traversed the inner membrane into cytoplasm of the E. coli cells. As a result, a final processed DNase domain of ColE7 was identified in the intracellular space of the cells treated with Col-Im complex. In the presence of periplasmic extracts, in vitro processing of DNase domain of ColE7 was also observed. These results suggest that the processing of ColE7 has occurred for translocation of the DNase-type colicin across the membrane and the process is probably taking place in the periplasmic space of the membrane.

Sensitivity-enhanced static 15N NMR of solids by 1h indirect detection

A method for enhancing the sensitivity of 15N spectra of nonspinning solids through 1H indirect detection is introduced. By sampling the 1H signals in the windows of a pulsed spin-lock sequence, high-sensitivity 1H spectra can be obtained in two-dimensional (2D) spectra whose indirect dimension yields the 15N chemical shift pattern. By sacrificing the 1H chemical shift information, sensitivity gains of 1.8 to 2.5 for the 15N spectra were achieved experimentally. A similar sensitivity enhancement was also obtained for 2D (15)N-(1)H dipolar and 15N chemical shift correlation spectroscopy, by means of a 3D 1H/15N-1H/15N correlation experiment. We demonstrate this technique, termed PRINS for proton indirectly detected nitrogen static NMR, on a crystalline model compound with long 1H T(1rho) and on a 25-kDa protein with short 1H T(1rho). This 1H indirect detection approach should be useful for enhancing the sensitivity of 15N NMR of oriented membrane peptides. It can also be used to facilitate the empirical optimization of 15N-detected experiments where the inherent sensitivity of the sample is low.

Calorimetric investigations of the structural stability and interactions of colicin B domains in aqueous solution and in the presence of phospholipid bilayers

The effects of pH and temperature on the stability of interdomain interactions of colicin B have been studied by differential-scanning calorimetry, circular dichroism, and fluorescence spectroscopy. The calorimetric properties were compared with those of the isolated pore-forming fragment. The unfolding profile of the full-length toxin is consistent with two endothermic transitions. Whereas peak A (T(m) = 55 degrees C) most likely corresponds to the receptor/translocation domain, peak B (T(m) = 59 degrees C) is associated with the pore-forming domain. By lowering the pH from 7 to 3.5, the transition temperature of peaks A and B are reduced by 25 and 18 degrees C, respectively, due to proton exchange upon denaturation. The isolated pore-forming fragment unfolds at much higher temperatures (T(m) = 65 degrees C) and is stable throughout a wide pH range, indicating that intramolecular interactions between the different colicin B domains result in a less stable protein conformation. In aqueous solution circular dichroism spectra have been used to estimate the content of helical secondary structure of colicin B ( approximately 40%) or its pore-forming fragment ( approximately 80%). Upon heating, the ellipticities at 222 nm strongly decrease at the transition temperature. In the presence of lipid vesicles the differential-scanning calorimetry profiles of the pore-forming fragment exhibit a low heat of transition multicomponent structure. The heat of transition of membrane-associated colicin B (T(m) = 54 degrees C at pH 3.5) is reduced and its secondary structure is conserved even at intermediate temperatures indicating incomplete unfolding due to strong protein-lipid interactions.

The C-terminal half of the colicin A pore-forming domain is active in vivo and in vitro

The pore-forming domain of colicin A (pfColA) fused to a prokaryotic signal peptide (sp-pfColA) is transported across and inserts into the inner membrane of Escherichia coli from the periplasmic side and forms a functional channel. The soluble structure of pfColA consists of a ten-helix bundle containing a hydrophobic helical hairpin. Here, we generated a series of mutants in which an increasing number of sp-pfColA alpha-helices was deleted. These peptides were tested for their ability to form ion channels in vivo and in vitro. We found that the shortest sp-pfColA mutant protein that killed Escherichia coli was composed of the five last alpha-helices of sp-pfColA, whereas the shortest peptide that formed a channel in planar lipid bilayer membranes similar to that of intact pfColA was the protein composed of the last six alpha-helices. The peptide composed of the last five alpha-helices of pfColA generated a voltage-independent conductance in planar lipid bilayer with properties very different from that of intact pfColA. Thus, helices 1 to 4 are unnecessary for channel formation, while helix 5, or some part of it, is important but not absolutely necessary. Voltage-dependence of colicin is evidently controlled by the first four alpha-helices of pfColA.

Using chimeric immunity proteins to explore the energy landscape for alpha-helical protein folding

To address the role of sequence in the folding of homologous proteins, the folding and unfolding kinetics of the all-helical bacterial immunity proteins Im2 and Im9 were characterised, together with six chimeric derivatives of these proteins. We show that both Im2 and Im9 fold rapidly (k(UN)(H(2)O)) approximately 2000 s(-1) at pH 7.0, 25 degrees C) in apparent two-state transitions, through rate-limiting transition states that are highly compact (beta(TS)0.93 and 0.96, respectively). Whilst the folding and unfolding properties of three of the chimeras (Im2 (1-44)(Im9), Im2 (1-64)(Im9 )and Im2 (25-44)(Im9)) are similar to their parental counterparts, in other chimeric proteins the introduced sequence variation results in altered kinetic behaviour. At low urea concentrations, Im2 (1-29)(Im9) and Im2 (56-64)(Im9) fold in two-state transitions via transition states that are significantly less compact (beta(TS) approximately 0.7) than those characterised for the other immunity proteins presented here. At higher urea concentrations, however, the rate-limiting transition state for these two chimeras switches or moves to a more compact species (beta(TS) approximately 0.9). Surprisingly, Im2 (30-64)(Im9) populates a highly collapsed species (beta(I)=0.87) in the dead-time (2.5 ms) of stopped flow measurements. These data indicate that whilst topology may place significant constraints on the folding process, specific inter-residue interactions, revealed here through multiple sequence changes, can modulate the ruggedness of the folding energy landscape.

The structure, function, and origin of the microcin H47 ATP-binding cassette exporter indicate its relatedness to that of colicin V

Microcin H47, a gene-encoded peptide antibiotic produced by a natural Escherichia coli strain, was shown to be secreted by a three-component ATP-binding cassette exporter which was revealed to be strongly related to that of colicin V. The results of sequence and gene fusion analyses, as well as heterologous complementation assays, are presented.

Identification of specific residues in colicin E1 involved in immunity protein recognition

The basis of specificity between pore-forming colicins and immunity proteins was explored by interchanging residues between colicins E1 (ColE1) and 10 (Col10) and testing for altered recognition by their respective immunity proteins, Imm and Cti. A total of 34 divergent residues in the pore-forming domain of ColE1 between residues 419 and 501, a region previously shown to contain the specificity determinants for Imm, were mutagenized to the corresponding Col10 sequences. The residue changes most effective in converting ColE1 to the Col10 phenotype are residue 448 at the N terminus of helix VI and residues 470, 472, and 474 at the C terminus of helix VII. Mutagenesis of helix VI residues 416 to 419 in Col10 to the corresponding ColE1 sequence resulted in increased recognition by Imm and loss of recognition by Cti.