Kinetic description of structural changes linked to membrane import of the colicin E1 channel protein.

Upon binding to membranes, the 178-residue colicin E1 C-terminal channel protein forms a steady-state closed-channel intermediate that is a flexible extended two-dimensional helical array [Zakharov et al. (1998) Proc. Natl. Acad. Sci. U.S.A. 95, 4282-4287]. Analysis of the kinetics of binding-insertion to liposome membranes of the channel protein, P178, and of changes of spectral parameters associated with structure transitions allowed a correlation of the sequence of tertiary and secondary structure changes with binding-insertion. Binding and insertion were distinguished by use of lipids modified with quenchers of Trp fluorescence attached to lipid headgroups or acyl chains. Secondary and tertiary structure changes were inferred, respectively, from changes in far-UV circular dichroism and relative changes of interresidue distances by fluorescence resonance energy transfer (FRET). "Single Trp" mutants were used in FRET analysis, with the background Tyr contribution determined through use of a "zero Trp" mutant. The sequence of distinguishable events and the pseudo-first-order rate constants under "standard" conditions (large unilamellar vesicles, pH 4.0, I = 0.1 M) was binding (30 +/- 5 s(-)(1)) --> unfolding (12.6 +/- 0.5 s(-)(1)) --> helix elongation (9.0 +/- 1.0 s(-)(1)) --> insertion (6. 6 +/- 0.5 s(-)(1)). Thus, helix elongation on the surface of the membrane can occur after unfolding and does not require insertion. Binding-insertion and structural transitions of P178 occur significantly faster with small unilamellar vesicles. The relevance to general mechanisms of protein import of the structural changes associated with import of the colicin channel is discussed.     

Voltage-dependent,monomeric channel activity of colicin E1 in artificial membrane vesicles

Summary The dependence of colicin channel activity on membrane potential and peptide concentration was studied in large unilamellar vesicles using colicin E1, its COOH-terminal thermolytic peptide and other channel-forming colicins. Channel activity was assayed by release of vesicle-entrapped chloride, and could be detected at a peptide: lipid molar ratio as low as 10^–7. The channel activity was dependent on the magnitude of atrans-negative potassium diffusion potential, with larger potentials yielding faster rates of solute efflux. For membrane potentials greater than –60mV (K _in ⁺ /K _out ⁺ 10), addition of valinomycin resulted in a 10-fold increase in the rate of Cl^– efflux. A delay in Cl^– efflux observed when the peptide was added to vesicles in the presence of a membrane potential implied a potential-independent binding-insertion mechanism. The initial rate of Cl^– efflux was about 1% of the single-channel conductance, implying that only a small fraction of channels were initially open, due to the delay or latency of channel formation known to occur in planar bilayers.The amount of Cl^– released as a function of added peptide increased monotonically to a concentration of 0.7 ng peptide/ml, corresponding to release of 75% of the entrapped chloride. It was estimated from this high activity and consideration of vesicle number that 50–100% of the peptide molecules were active. The dependence of the initial rate of Cl^– efflux on peptide concentration was linear to approximately the same concentration, implying that the active channel consists of a monomeric unit.     

Dynamic properties of the colicin E1 ion channel

Abstract The mechanism of channel formation and action of channel-forming colicins is a paradigm for the study of dynamic aspects of membrane-protein interactions. The following experimental results concerning interaction of the colicin E1 channel domain with target membranes, in vitro and in vivo, are discussed: (1) the nature of the translocation-competent state of the channel-forming domain; (2) unfolding of the colicin channel peptide during in vitro binding and anchoring of the channel to liposome membranes at acidic pH; (3) reversal of channel peptide binding to liposomes by an alkaline-directed pH shift; (4) voltage-driven translocation and gating of the ion channel, discussed in the context of a four-helix model for a monomeric channel; (5) rescue of colicin-treated cells by high levels of external K⁺; (6) trypsin rescue of cells depolarized by the colicin ion channel; and (7) interaction of the channel domain with its immunity protein.     

Toward elucidating the membrane topology of helix two of the colicin E1 channel domain

The membrane-bound closed state of the colicin E1 channel domain was investigated by site-directed fluorescence labeling using a bimane fluorophore attached to each single cysteine residue within helix 2 of each mutant protein. The fluorescence properties of the bimane fluorophore were measured for the membrane-associated form of the closed channel and included fluorescence emission maximum, fluorescence anisotropy, apparent polarity, surface accessibility, and membrane bilayer penetration depth. The fluorescence data show that helix 2 is an amphipathic alpha-helix that is situated parallel to the membrane surface, but it is less deeply embedded within the bilayer interfacial region than is helix 1 in the closed channel. A least squares fit of the various data sets to a harmonic wave function indicated that the periodicity and angular frequency for helix 2 in the membrane-bound state are typical for an amphipathic alpha-helix (3.8 +/- 0.1 residues per turn and 94 +/- 4 degrees, respectively) that is located at an interfacial region of a membrane bilayer. Dual quencher analysis also revealed that helix 2 is peripherally membrane associated, with one face of the helix dipping into the interfacial region of the lipid bilayer and the other face projecting outwardly into the aqueous solvent. Finally, our data show that helices 1 and 2 remain independent helices upon membrane association with a short connector link (Tyr(363)-Gly(364)) and that short amphipathic alpha-helices participate in the formation of a lipid-dependent, toroidal pore for this colicin.     

Structure and dynamics of the colicin E1 channel

The toxin-like and bactericidal colicin E1 molecule is of interest for problems of toxin action, polypeptide translocation across membranes, voltage-gated channels, and receptor function. Colicin E1 binds to a receptor in the outer membrane and is translocated across the cell envelope to the inner membrane. Import of the colicin channel-forming domain into the inner membrane involves a translocation-competent intermediate state and a membrane potential-dependent movement of one third to one half of the channel peptide into the membrane bilayer. The voltage-gated channel has a conductance sufficiently large to depolarize the Escherichia coli cytoplasmic membrane. Amino acid residues that affect the channel ion selectivity have been identified by site-directed mutagenesis. The colicin E1 channel is one of a few membrane proteins whose secondary structures in the membrane, predominantly alpha-helix, have been determined by physico-chemical techniques. Hypothesis for the identity of the trans-membrane helices, and the mechanism of binding to the membrane, are influenced by the solved crystal structure of the soluble colicin A channel peptide. The protective action of immunity protein is a unique aspect of the colicin problem, and information has been obtained, by genetic techniques, about the probable membrane topography of the imm gene product.     

Determination of membrane protein topology by red-edge excitation shift analysis: application to the membrane-bound colicin E1 channel peptide

A new approach for the determination of the bilayer location of Trp residues in proteins has been applied to the study of the membrane topology of the channel-forming bacteriocin, colicin E1. This method, red-edge excitation shift (REES) analysis, was initially applied to the study of 12 single Trp-containing channel peptides of colicin E1 in the soluble state in aqueous medium. Notably, REES was observed for most of the channel peptides in aqueous solution upon low pH activation. The extent of REES was subsequently characterized using a model membrane system composed of the tripeptide, Lys-Trp-Lys, bound to dimyristoyl-sn-glycerol-3-phosphatidylserine liposomes. Subsequently, data accrued from the model peptide-lipid system was used to interpret information obtained on the channel peptides when bound to dioleoyl-sn-glycerol-3-phosphatidylcholine/dioleoyl-sn-glycerol-3-phosphatidylglycerol membrane vesicles. The single Trp mutant peptides were divided into three categories based on the change in the REES values observed for the Trp residues when the peptides were bound to liposomes as compared to the REES values measured for the soluble peptides. F-404W, F-413W, F-443W, F-484W, and W-495 peptides exhibited small and/or insignificant REES changes (ΔREES) whereas W-424, F-431W, and Y-507W channel peptides possessed modest REES changes (3 nm≤ΔREES≤7 nm). In contrast, wild-type, Y-367W, W-460, Y-478W, and I-499W channel peptides showed large ΔREES values upon membrane binding (7 nm<ΔREES≤12 nm). The REES data for the membrane-bound structure of the colicin E1 channel peptide proved consistent with previous data for the topology of the closed channel state, which lends further credence to the currently proposed channel model. In conclusion, the REES method provides another source of topological data for assignment of the bilayer location for Trp residues within membrane-associated proteins; however, it also requires careful interpretation of spectral data in combination with structural information on the proteins being investigated.     

Determination of membrane protein topology by red-edge excitation shift analysis: application to the membrane-bound colicin E1 channel peptide

A new approach for the determination of the bilayer location of Trp residues in proteins has been applied to the study of the membrane topology of the channel-forming bacteriocin, colicin E1. This method, red-edge excitation shift (REES) analysis, was initially applied to the study of 12 single Trp-containing channel peptides of colicin E1 in the soluble state in aqueous medium. Notably, REES was observed for most of the channel peptides in aqueous solution upon low pH activation. The extent of REES was subsequently characterized using a model membrane system composed of the tripeptide, Lys-Trp-Lys, bound to dimyristoyl-sn-glycerol-3-phosphatidylserine liposomes. Subsequently, data accrued from the model peptide-lipid system was used to interpret information obtained on the channel peptides when bound to dioleoyl-sn-glycerol-3-phosphatidylcholine/dioleoyl-sn-glycerol-3-phosphatidylglycerol membrane vesicles. The single Trp mutant peptides were divided into three categories based on the change in the REES values observed for the Trp residues when the peptides were bound to liposomes as compared to the REES values measured for the soluble peptides. F-404 W, F-413 W, F-443 W, F-484 W, and W-495 peptides exhibited small and/or insignificant REES changes (Delta REES) whereas W-424, F-431 W, and Y-507 W channel peptides possessed modest REES changes (3 nm< or = Delta REES< or = 7 nm). In contrast, wild-type, Y-367 W, W-460, Y-478 W, and I-499 W channel peptides showed large Delta REES values upon membrane binding (7 nm< Delta REES< or =12 nm). The REES data for the membrane-bound structure of the colicin E1 channel peptide proved consistent with previous data for the topology of the closed channel state, which lends further credence to the currently proposed channel model. In conclusion, the REES method provides another source of topological data for assignment of the bilayer location for Trp residues within membrane-associated proteins; however, it also requires careful interpretation of spectral data in combination with structural information on the proteins being investigated.     

Constraints imposed by protease accessibility on the trans-membrane and surface topography of the colicin E1 ion channel.

The surface topography of a 190-residue COOH-terminal colicin E1 channel peptide (NH2-Met 333-Ile 522-COOH) bound to uniformly sized 0.2-micron liposomes was probed by accessibility of the peptide to proteases in order (1) to determine whether the channel structure contains trans-membrane segments in addition to the four alpha-helices previously identified and (2) to discriminate between different topographical possibilities for the surface-bound state. An unfolded surface-bound state is indicated by increased trypsin susceptibility of the bound peptide relative to that of the peptide in aqueous solution. The peptide is bound tightly to the membrane surface with Kd < 10(-7) M. The NH2-terminal 50 residues of the membrane-bound peptide are unbound or loosely bound as indicated by their accessibility to proteases, in contrast with the COOH-terminal 140 residues, which are almost protease inaccessible. The general protease accessibility of the NH2-terminal segment Ala 336-Lys 382 excludes any model for the closed channel state that would include trans-membrane helices on the NH2-terminal side of Lys 382. Lys 381-Lys 382 is a major site for protease cleavage of the surface-bound channel peptide. A site for proteinase K cleavage just upstream of the amphiphilic gating hairpin (K420-K461) implies the presence of a surface-exposed segment in this region. These protease accessibility data indicate that it is unlikely that there are any alpha-helices on the NH2-terminal side of the gating hairpin K420-K461 that are inserted into the membrane in the absence of a membrane potential. A model for the topography of an unfolded monomeric surface-bound intermediate of the colicin channel domain, including a trans-membrane hydrophobic helical hairpin and two or three long surface-bound helices, is proposed.     

Identification of a voltage-responsive segment of the potential-gated colicin E1 ion channel.

The voltage dependence of channel activity of the bactericidal protein colicin E1 was found to be correlated with insertion into the membrane bilayer of a specific segment of the 178-residue COOH-terminal thermolytic colicin channel peptide. The insertion into the bilayer was detected by an increase in labeling by one of two different lipophilic photoaffinity probes or by a decrease in iodination of peptide tyrosines from the external solution. Imposition of a potassium diffusion potential of -100 mV resulted in an increase of 35-60% in the labeling of the peptide by the lipophilic probe in the bilayer and a concomitant decrease in labeling of Tyr residues in the peptide by the iodination reagent in the external solution. The change in labeling decreased upon dissipation of the membrane potential with a half-time of about 1 min. The labeling change was localized to a 36-residue peptide segment bounded by alanine-425 and by tryptophan-460. This segment containing seven positively charged residues at low pH is a voltage-sensitive region that inserts into the membrane bilayer when the channel is turned on by the potential and is extruded from it when the voltage is removed and the channel is turned off.     

Lipid dependence of the channel properties of a colicin E1-lipid toroidal pore

Colicin E1 belongs to a group of bacteriocins whose cytotoxicity toward Escherichia coli is exerted through formation of ion channels that depolarize the cytoplasmic membrane. The lipid dependence of colicin single-channel conductance demonstrated intimate involvement of lipid in the structure of this channel. The colicin formed "small" conductance 60-picosiemens (pS) channels, with properties similar to those previously characterized, in 1,2-dieicosenoyl-sn-glycero-3-phosphocholine (C20) or thinner membranes, whereas it formed a novel "large" conductance 600-pS state in thicker 1,2-dierucoyl-sn-glycero-3-phosphocholine (C22) bilayers. Both channel states were anion-selective and voltage-gated and displayed a requirement for acidic pH. Lipids having negative spontaneous curvature inhibited the formation of both channels but increased the ratio of open 600 pS to 60 pS conductance states. Different diameters of small and large channels, 12 and 16 A, were determined from the dependence of single-channel conductance on the size of nonelectrolyte solute probes. Colicin-induced lipid "flip-flop" and the decrease in anion selectivity of the channel in the presence of negatively charged lipids implied a significant contribution of lipid to the structure of the channel, most readily described as toroidal organization of lipid and protein to form the channel pore.     

Folded state of the integral membrane colicin E1 immunity protein in solvents of mixed polarity

The colicin E1 immunity protein (ImmE1), a 13.2-kDa hydrophobic integral membrane protein localized in the Escherichia coli cytoplasmic membrane, protects the cell from the lethal, channel-forming activity of the bacteriocin, colicin E1. Utilizing its solubility in organic solvents, ImmE1 was purified by 1-butanol extraction of isolated membranes, followed by gel filtration and ion-exchange chromatography in a chloroform/methanol/H(2)O (4:4:1) solvent system. Circular dichroism analysis indicated that the alpha-helical content of ImmE1 is approximately 80% in 1-butanol or 2,2,2-trifluoroethanol, consistent with a previous membrane-folding model with three extended hydrophobic transmembrane helical domains, H1-H3. Each of these extended hydrophobic domains contains a centrally located single Cys residue that could be used as a probe of protein structure. The presence of tertiary structure of purified ImmE1 in a solvent of mixed polarity, chloroform/methanol/H(2)O (4:4:1) was demonstrated by (i) the constraints on Tyr residues shown by the amplitude of near-UV circular dichroism spectra in the wavelength interval, 270-285 nm; (ii) the correlation between the near-UV Tyr CD spectrum of single and double Cys-to-X mutants of the Imm protein and their in vivo activity; (iii) the upfield shift of methyl groups in a 1D NMR spectrum, a 2D- HSQC NMR spectrum of ImmE1 in the mixed polarity solvent mixture, and a broadening and disappearance of the indole (1)H proton resonance from Trp94 in H3 by a spin label attached to Cys16 in the H2 hydrophobic domain; (iv) near-UV circular dichroism spectra with a prominent ellipticity band centered at 290 nm from a single Trp inserted into the extended hydrophobic domains. It was concluded that the colicin E1 immunity protein adopts a folded conformation in chloroform/methanol/H(2)O (4:4:1) that is stabilized by helix-helix interactions. Analysis of the probable membrane folding topology indicated that several Tyr residues in the bilayer region of the three transmembrane helices could contribute to the near-UV CD spectrum through helix-helix interactions.     

Membrane topology of the colicin E1 channel using genetically encoded fluorescence

The membrane topology of the colicin E1 channel domain was studied by fluorescence resonance energy transfer (FRET). The FRET involved a genetically encoded fluorescent amino acid (coumarin) as the donor and a selectively labeled cysteine residue tethered with DABMI (4-(dimethylamino)phenylazophenyl-4'-maleimide) as the FRET acceptor. The fluorescent coumarin residue was incorporated into the protein via an orthogonal tRNA/aminoacyl-tRNA synthetase pair that allowed selective incorporation into any site within the colicin channel domain. Each variant harbored a stop (TAG) mutation for coumarin incorporation and a cysteine (TGT) mutation for DABMI attachment. Six interhelical distances within helices 1-6 were determined using FRET analysis for both the soluble and membrane-bound states. The FRET data showed large changes in the interhelical distances among helices 3-6 upon membrane association providing new insight into the membrane-bound structure of the channel domain. In general, the coumarin-DABMI FRET interhelical efficiencies decreased upon membrane binding, building upon the umbrella model for the colicin channel. A tentative model for the closed state of the channel domain was developed based on current and previously published FRET data. The model suggests circular arrangement of helices 1-7 in a clockwise direction from the extracellular side and membrane interfacial association of helices 1, 6, 7, and 10 around the central transmembrane hairpin formed by helices 8 and 9.     

Identification of the plasmid-encoded immunity protein for colicin E1 in the inner membrane of Escherichia coli

A set of plasmids containing portions of the Col El plasmid were transformed into recA⁻ cells. These cells, after UV irradiation, only incorporate labelled amino acids into plasmid-encoded proteins. UV-irradiated cells label a 14.5 kDa band if they are phenotypically immune to colicin E1, and do not contain this band if they are sensitive to colicin E1. We conclude that the 14.5 kDa protein is the colicin E1 immunity protein. When the inner and outer membranes of these cells are fractionated, the labelled band appears in the inner membrane. The immunity protein must be an intrinsic inner membrane protein, confirming the predictions made by hydrophobicity calculations from primary sequence data.

Maxicell Col El plasmid Immunity protein Hydrophobicity calculation     

Gating of a voltage-dependent channel (colicin E1) in planar lipid bilayers: the role of protein translocation

Summary The voltage-dependent channel formed in planar lipid bilayers by colicin E1, or its channel-forming C-terminal fragments, is susceptible to destruction by the nonspecific protease pepsin under well-defined conditions. In particular, pepsin acts only from thecis side (the side to which colicin has been added) and only upon channels in the closed state. Channels in the open state are refractory to destruction bycis pepsin, and neither open nor closed channels are destroyed bytrans pepsin. Colicin E1 channels are normally turned on bycis positive voltages and turned off bycis negative voltages. For large (>80 mV) positive voltages, however, channels inactivate subsequent to opening. Associated with the inactivated state, some channels become capable of being turned on bycis negative voltages and turned off bycis positive voltages, as if the channel-forming region of the molecule has been translocated across the membrane. Consistent with this interpretation is the ability now oftrans pepsin to destroy these reversed channels when they are closed, but not when they are open, whereascis pepsin has no effect on them in either the open or closed state. Our results indicate that voltage gating of the E1 channel involves translocation of parts of the protein across the membrane, exposing different domains to thecis andtrans solutions in the different channel states.     

Adventures in membrane protein topology. A study of the membrane-bound state of colicin E1.

The molecular aggregate size of the closed state of the colicin E1 channel was determined by fluorescence resonance energy transfer experiments involving a fluorescence donor (three tryptophans, wild-type protein) and a fluorescence acceptor (5-(((acetyl)amino)ethyl)aminonaphthalene-1-sulfonic acid (AEDANS), Trp-deficient protein). There was no evidence of energy transfer between the donor and acceptor species when bound to membrane large unilamellar vesicles. These experiments led to the conclusion that the colicin E1 channel is monomeric in the membrane-bound closed channel state. Experiments were also conducted to study the membrane topology of the closed colicin channel in membrane large unilamellar vesicles using acrylamide as the membrane-impermeant, nonionic quencher of tryptophan fluorescence in a battery of single tryptophan mutant proteins. Furthermore, additional fluorescence parameters, including fluorescence emission maximum, fluorescence quantum yield, and fluorescence decay times, were used to assist in mapping the topology of the closed channel. Results suggest that the closed channel comprises most of the polypeptide of the channel domain and that the hydrophobic anchor domain does not transverse the membrane bilayer but nonetheless is deeply embedded within the hydrocarbon core of the membrane. Finally, a model is proposed which features at least two states that are in rapid equilibrium with each other and in which one state is more heavily populated than the other.     

Synthesis and functioning of the colicin E1 lysis protein: comparison with the colicin A lysis protein.

The colicin E1 lysis protein, CelA, was identified as a 3-kDa protein in induced cells of Escherichia coli K-12 carrying pColE1 by pulse-chase labeling with either [35S]cysteine or [3H]lysine. This 3-kDa protein was acylated, as shown by [2-3H]glycerol labeling, and seemed to correspond to the mature CelA protein. The rate of modification and processing of CelA was different from that observed for Cal, the colicin A lysis protein. In contrast to Cal, no intermediate form was detected for CelA, no signal peptide accumulated, and no modified precursor form was observed after globomycin treatment. Thus, the rate of synthesis would not be specific to lysis proteins. Solubilization in sodium dodecyl sulfate of the mature forms of both CelA and Cal varied similarly at the time of colicin release, indicating a change in lysis protein structure. This particular property would play a role in the mechanism of colicin export. The accumulation of the signal peptide seems to be a factor determining the toxicity of the lysis proteins since CelA provoked less cell damage than Cal. Quasi-lysis and killing due to CelA were higher in degP mutants than in wild-type cells. They were minimal in pldA mutants.     

Fourier transform infrared evidence for a predominantly alpha-helical structure of the membrane bound channel forming COOH-terminal peptide of colicin E1.

The structure of the membrane bound state of the 178-residue thermolytic COOH-terminal channel forming peptide of colicin E1 was studied by polarized Fourier transform infrared (FTIR) spectroscopy. This fragment was reconstituted into DMPC liposomes at varying peptide/lipid ratios ranging from 1/25-1/500. The amide I band frequency of the protein indicated a dominant alpha-helical secondary structure with limited beta- and random structures. The amide I and II frequencies are at 1,656 and 1,546 cm-1, close to the frequency of the amide I and II bands of rhodopsin, bacteriorhodopsin and other alpha-helical proteins. Polarized FTIR of oriented membranes revealed that the alpha-helices have an average orientation less than the magic angle, 54.6 degrees, relative to the membrane normal. Almost all of the peptide groups in the membrane-bound channel protein undergo rapid hydrogen/deuterium (H/D) exchange. These results are contrasted to the alpha-helical membrane proteins, bacteriorhodopsin, and rhodopsin.     

Unfolding study of a trimeric membrane protein AcrB

The folding of a multi‐domain trimeric α‐helical membrane protein, Escherichia coli inner membrane protein AcrB, was investigated. AcrB contains both a transmembrane domain and a large periplasmic domain. Protein unfolding in sodium dodecyl sulfate (SDS) and urea was monitored using the intrinsic fluorescence and circular dichroism spectroscopy. The SDS denaturation curve displayed a sigmoidal profile, which could be fitted with a two‐state unfolding model. To investigate the unfolding of separate domains, a triple mutant was created, in which all three Trp residues in the transmembrane domain were replaced with Phe. The SDS unfolding profile of the mutant was comparable to that of the wild type AcrB, suggesting that the observed signal change was largely originated from the unfolding of the soluble domain. Strengthening of trimer association through the introduction of an inter‐subunit disulfide bond had little effect on the unfolding profile, suggesting that trimer dissociation was not the rate‐limiting step in unfolding monitored by fluorescence emission. Under our experimental condition, AcrB unfolding was not reversible. Furthermore, we experimented with the refolding of a monomeric mutant, AcrB_Δloop, from the SDS unfolded state. The CD spectrum of the refolded AcrB_Δloop superimposed well onto the spectra of the original folded protein, while the fluorescence spectrum was not fully recovered. In summary, our results suggested that the unfolding of the trimeric AcrB started with a local structural rearrangement. While the refolding of secondary structure in individual monomers could be achieved, the re‐association of the trimer might be the limiting factor to obtain folded wild‐type AcrB.