Characterisation of a mobile protein-binding epitope in the translocation domain of colicin E9

The 61 kDa colicin E9 protein toxin enters the cytoplasm of susceptible cells by interacting with outer membrane and periplasmic helper proteins, and kills them by hydrolysing their DNA. The membrane translocation function is located in the N-terminal domain of the colicin, with a key signal sequence being a pentapeptide region that governs the interaction with the helper protein TolB (the TolB box). Previous NMR studies (Collins et al., 2002 J. Mol. Biol. 318, 787-804) have shown that the N-terminal 83 residues of colicin E9, which includes the TolB box, is largely unstructured and highly flexible. In order to further define the properties of this region we have studied a fusion protein containing residues 1-61 of colicin E9 connected to the N-terminus of the E9 DNase by an eight-residue linking sequence. 53 of the expected 58 backbone NH resonances for the first 61 residues and all of the expected 7 backbone NH resonances of the linking sequence were assigned with 3D (1)H-(13)C-(15)N NMR experiments, and the backbone dynamics of these regions investigated through measurement of (1)H-(15)N relaxation properties. Reduced spectral density mapping, extended Lipari-Szabo modelling, and fitting backbone R(2) relaxation rates to a polymer dynamics model identifies three clusters of interacting residues, each containing a tryptophan. Each of these clusters is perturbed by TolB binding to the intact colicin, showing that the significant region for TolB binding extends beyond the recognized five amino acids of the TolB box and demonstrating that the binding epitope for TolB involves a considerable degree of order within an otherwise disordered and flexible domain. Abbreviations : Im9, the immunity protein for colicin E9; E9 DNase, the endonuclease domain of colicin E9; HSQC, heteronuclear single quantum coherence; ppm, parts per million; DSS, 2,2-(dimethylsilyl)propanesulfonic acid; TSP, sodium 3-trimethylsilypropionate; T(1 - 61)-DNase fusion protein, residues 1-61 of colicin E9 connected to the N-terminus of the E9 DNase by an eight residue thrombin cleavage sequence.     

Interactions of TolB with the translocation domain of colicin E9 require an extended TolB box

The mechanism by which enzymatic E colicins such as colicin E3 (ColE3) and ColE9 cross the outer membrane, periplasm, and cytoplasmic membrane to reach the cytoplasm and thus kill Escherichia coli cells is unique in prokaryotic biology but is poorly understood. This requires an interaction between TolB in the periplasm and three essential residues, D35, S37, and W39, of a pentapeptide sequence called the TolB box located in the N-terminal translocation domain of the enzymatic E colicins. Here we used site-directed mutagenesis to demonstrate that the TolB box sequence in ColE9 is actually larger than the pentapeptide and extends from residues 34 to 46. The affinity of the TolB box mutants for TolB was determined by surface plasmon resonance to confirm that the loss of biological activity in all except one (N44A) of the extended TolB box mutants correlates with a reduced affinity of binding to TolB. We used a PCR mutagenesis protocol to isolate residues that restored activity to the inactive ColE9 D35A, S37A, and W39A mutants. A serine residue at position 35, a threonine residue at position 37, and phenylalanine or tyrosine residues at position 39 restored biological activity of the mutant ColE9. The average area predicted to be buried upon folding (AABUF) was correlated with the activity of the variants at positions 35, 37, and 39 of the TolB box. All active variants had AABUF profiles that were similar to the wild-type residues at those positions and provided information on the size, stereochemistry, and potential folding pattern of the residues of the TolB Box.     

Gating of a voltage-dependent channel (colicin E1) in planar lipid bilayers: translocation of regions outside the channel-forming domain

Summary C-terminal fragments of colicin E1, ranging in mol wt from 14.5 to 20kD, form channels with voltage dependence and ion selectivity qualitatively similar to those of whole E1, placing an upper limit on the channel-forming domain. Under certain conditions, however, the gating kinetics and ion selectivity of channels formed by these different E1 peptides can be distinguished. The differences in channel behavior appear to be correlated with peptide length. Enzymatic digestion with trypsin of membrane-bound E1 peptides converts channel behavior of longer peptides to that characteristic of channels formed by shorter fragments. Apparently trypsin removes segments of protein N-terminal to the channel-forming region, since gating behavior of the shortest fragment is little affected by the enzyme. The success of this conversion depends on the side of the membrane to which trypsin is added and on the state, open or closed, of the channel. Trypsin modifies only closed channels from thecis side (the side to which protein has been added) and only open channels from thetrans side. These results suggest that regions outside the channel-forming domain affect ion selectivity and gating, and they also provide evidence that large protein segments outside the channel-forming domain are translocated across the membrane with channel gating.     

High-resolution crystal structure of a truncated ColE7 translocation domain: implications for colicin transport across membranes

ColE7 is a nuclease-type colicin released from Escherichia coli to kill sensitive bacterial cells by degrading the nucleic acid molecules in their cytoplasm. ColE7 is classified as one of the group A colicins, since the N-terminal translocation domain (T-domain) of the nuclease-type colicins interact with specific membrane-bound or periplasmic Tol proteins during protein import. Here, we show that if the N-terminal tail of ColE7 is deleted, ColE7 (residues 63-576) loses its bactericidal activity against E.coli. Moreover, TolB protein interacts directly with the T-domain of ColE7 (residues 1-316), but not with the N-terminal deleted T-domain (residues 60-316), as detected by co-immunoprecipitation experiments, confirming that the N-terminal tail is required for ColE7 interactions with TolB. The crystal structure of the N-terminal tail deleted ColE7 T-domain was determined by the multi-wavelength anomalous dispersion method at a resolution of 1.7 angstroms. The structure of the ColE7 T-domain superimposes well with the T-domain of ColE3 and TR-domain of ColB, a group A Tol-dependent colicin and a group B TonB-dependent colicin, respectively. The structural resemblance of group A and B colicins implies that the two groups of colicins may share a mechanistic connection during cellular import.     

The structure of TolB, an essential component of the tol-dependent translocation system, and its protein-protein interaction with the translocation domain of colicin E9

BACKGROUND: E colicin proteins have three functional domains, each of which is implicated in one of the stages of killing Escherichia coli cells: receptor binding, translocation and cytotoxicity. The central (R) domain is responsible for receptor-binding activity whereas the N-terminal (T) domain mediates translocation, the process by which the C-terminal cytotoxic domain is transported from the receptor to the site of its cytotoxicity. The translocation of enzymatic E colicins like colicin E9 is dependent upon TolB but the details of the process are not known. RESULTS: We have demonstrated a protein-protein interaction between the T domain of colicin E9 and TolB, an essential component of the tol-dependent translocation system in E. coli, using the yeast two-hybrid system. The crystal structure of TolB, a procaryotic tryptophan-aspartate (WD) repeat protein, reveals an N-terminal alpha + beta domain based on a five-stranded mixed beta sheet and a C-terminal six-bladed beta-propeller domain. CONCLUSIONS: The results suggest that the TolB-box residues of the T domain of colicin E9 interact with the beta-propeller domain of TolB. The protein-protein interactions of other beta-propeller-containing proteins, the yeast yPrp4 protein and G proteins, are mediated by the loops or outer sheets of the propeller blades. The determination of the three-dimensional structure of the T domain-TolB complex and the isolation of mutations in TolB that abolish the interaction with the T domain will reveal fine details of the protein-protein interaction of TolB and the T domain of E colicins.     

NMR study of Ni2+ binding to the H-N-H endonuclease domain of colicin E9.

Ni2+ affinity columns are widely used for protein purification, but they carry the risk that Ni2+ ions may bind to the protein, either adventitiously or at a physiologically important site. Dialysis against ethylenediaminetetraacetic acid (EDTA) is normally used to remove metal ions bound adventitiously to proteins; however, this approach does not always work. Here we report that a bacterial endonuclease, the DNase domain of colicin E9, binds Ni2+ acquired from Ni2+ affinity columns, and appears to bind [Ni(EDTA)(H2O)n]2- at low ionic strength. NMR was used to detect the presence of both Ni2+ coordinated to amino acid side chains and [Ni(EDTA)(H2O)N]2-. Dialysis against > or =0.2 M NaCl was required to remove the [Ni(EDTA)(H2O)n]2-. The NMR procedure we have used to characterize the presence of Ni2+ and [Ni(EDTA)(H2O)n]2- should be applicable to other proteins where there is the possibility of binding paramagnetic metal ions that are present to expedite protein purification. In the present case, the binding of Ni2+ seems likely to be physiologically relevant, and the NMR data complement recent X-ray crystallographic evidence concerning the number of histidine ligands to bound Ni2+.     

Slow conformational dynamics of an endonuclease persist in its complex with its natural protein inhibitor

The bacterial toxin colicin E9 is secreted by producing Escherichia coli cells with its 9.5 kDa inhibitor protein Im9 bound tightly to its 14.5 kDa C-terminal DNase domain. Double- and triple-resonance NMR spectra of the isolated DNase domain uniformly labeled with 13C/15N bound to unlabeled Im9 contain more signals than expected for a single DNase conformer, consistent with the bound DNase being present in more than one form. The presence of chemical exchange cross peaks in 750 MHz 15N-1H-15N HSQC-NOESY-HSQC spectra for backbone NH groups of Asp20, Lys21, Trp22, Leu23, Lys69, and Asn70 showed that the bound DNase was in dynamic exchange. The rate of exchange from the major to the minor form was determined to be 1.1 +/- 0.2 s(-1) at 298 K. Previous NMR studies have shown that the free DNase interchanges between two conformers with a forward rate constant of 1.61 +/- 0.11 s(-1) at 288 K, and that the bound Im9 is fixed in one conformation. The NMR studies of the bound DNase show that Im9 binds similarly to both conformers of the DNase and that the buried Trp22 is involved in the dynamic process. For the free DNase, all NH groups within a 9 A radius of any point of the Trp22 ring exhibit heterogeneity suggesting that a rearrangement of the position of this side chain is connected with the conformational interchange. The possible functional significance of this feature of the DNase is discussed.     

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.     

Kinetic basis for the competitive recruitment of TolB by the intrinsically disordered translocation domain of colicin E9

TolB and Pal are members of the Tol-Pal system that spans the cell envelope of Gram-negative bacteria and contributes to the stability and integrity of the bacterial outer membrane (OM). Lipoylated Pal is tethered to the OM and binds the β-propeller domain of periplasmic TolB, which, as recent evidence suggests, disengages TolB from its interaction with other components of the Tol system in the inner membrane. Antibacterial nuclease colicins such as colicin E9 (ColE9) also bind the β-propeller domain of TolB in order to catalyze their translocation across the bacterial OM. In contrast to Pal, however, colicin binding to TolB promotes its interaction with other components of the Tol system. Here, through a series of pre-steady-state kinetic experiments utilizing fluorescence resonance energy transfer pairs within the individual protein-protein complexes, we establish the kinetic basis for such 'competitive recruitment' by the TolB-binding epitope (TBE) of ColE9. Surprisingly, the 16-residue disordered ColE9 TBE associates more rapidly with TolB than Pal, a folded 13-kDa protein. Moreover, we demonstrate that calcium ions, which bind within the confines of the TolB β-propeller domain tunnel and are known to increase the affinity of the TolB-ColE9 complex, do not exert their influence through long-range electrostatic effects, as had been predicted, but through short-range effects that slow the dissociation rate of ColE9 TBE from its complex with TolB. Our study demonstrates that an intrinsically disordered protein undergoing binding-induced folding can compete effectively with a globular protein for a common target by associating more rapidly than the globular protein.     

Chemical modification of the two histidine and single cysteine residues in the channel-forming domain of colicin E1

Summary The two histidine residues of COOH-terminal channel-forming peptides of colicin E1 were modified by addition of a carbethoxy group through pretreatment with diethylpyrocarbonate. The consequences of the modification were examined by the action of the altered product on both phospholipid vesicles and planar membranes. At pH 6, where activity is low, histidine modification resulted in a decrease of the single channel conductance from 20 pS to approximately 9 pS and a decrease in the selectivity for sodium relative to chloride, showing that histidine modification affected the permeability properties of the channel. At pH 4, where activity is high, the single channel conductance and ion selectivity were not significantly altered by histidine modification. The histidine modification assayed at pH 4 resulted in a threefold increase in the rate of Cl^– efflux from asolectin vesicles, and a similar increase in conductance assayed with planar membranes. This conductance increase was inferred to arise from an increase in the fraction of bound histidine-modified colicin molecules forming channels at pH 4, since the increase in activity was not due to (i) an increase in binding of the modified peptide, (ii) a change in ion selectivity, (iii) a change of single channel conductance, or (iv) a change in the pH dependence of binding. The sole cysteine in the colicin molecule was modified in 6m urea with 5,5-dithiobis(2-nitrobenzoic acid). The activities of the colicin and its COOH-terminal tryptic peptide were found to be unaffected by cysteine modification, arguing against a role of (-SH) groups in protein insertion and/or channel formation.     

Solid-state NMR studies of the membrane-bound closed state of the colicin E1 channel domain in lipid bilayers.

The colicin E1 channel polypeptide was shown to be organized anisotropically in membranes by solid-state NMR analysis of samples of uniformly 15N-labeled protein in oriented planar phospholipid bilayers. The 190 residue C-terminal colicin E1 channel domain is the largest polypeptide to have been characterized by 15N solid-state NMR spectroscopy in oriented membrane bilayers. The 15N-NMR spectra of the colicin E1 show that: (1) the structure and dynamics are independent of anionic lipid content in both oriented and unoriented samples; (2) assuming the secondary structure of the polypeptide is helical, there are both trans-membrane and in-plane helical segments; (3) trans-membrane helices account for approximately 20-25% of the channel polypeptide, which is equivalent to 38-48 residues of the 190-residue polypeptide. The results of the two-dimensional PISEMA spectrum are interpreted in terms of a single trans-membrane helical hairpin inserted into the bilayer from each channel molecule. These data are also consistent with this helical hairpin being derived from the 38-residue hydrophobic segment near the C-terminus of the colicin E1 channel polypeptide.     

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.     

A very short peptide makes a voltage-dependent ion channel: the critical length of the channel domain of colicin E1

Cleavage of colicin E1 molecules with a variety of proteases or with cyanogen bromide (CNBr) generates COOH-terminal fragments which have channel-forming activity similar to that of intact colicin in planar lipid bilayer membranes. The smallest channel-forming fragment obtained by CNBr cleavage of the wild-type molecule consists of the C-terminal 152 amino acids. By the use of oligonucleotide-directed mutagenesis, we have made nine mutants along this 152 amino acid peptide, in which an amino acid was replaced by methionine in order to create a new CNBr cleavage site. The smallest of the CNBr-cleaved C-terminal fragments with channel-forming activity, in planar bilayer membranes, was generated by cleavage at new Met position 428 and has 94 amino acids, whereas a 75 amino acid peptide produced by cleavage of a new Met at position 447 did not have channel activity. The NH2-terminus of the channel-forming domain of colicin E1 appears therefore to lie between residues 428 and 447. Since, however, the last six C-terminal residues of the colicin can be removed without changing activity, the number of amino acids necessary to form the channel is 88 or less. In addition, the unique Cys residue in colicin E1 was replaced by Gly, and nine mutants were then made with Cys placed at sequential locations along the peptide for eventual use as sulfhydryl attachment sites to determine the local environment of the replaced amino acid. In the course of making 21 mutants, eight charged residues have been replaced by uncharged Met or Cys without changing the biological activity of the intact molecule. It has been proposed previously that the conformation of the colicin E1 channel is a barrel formed from five or six alpha-helices, each having 20 amino acids spanning the membrane and two to four residues making the turn at the boundary of the membrane. Our finding that 88 amino acids can make an active channel, combined with recently reported stoichiometric evidence that the channel is a monomer excludes this model and adds significant constraints which can be used in building a molecular model of the channel.     

Destabilization of the colicin E9 Endonuclease domain by interaction with negatively charged phospholipids: implications for colicin translocation into bacteria

We have shown previously that the 134-residue endonuclease domain of the bacterial cytotoxin colicin E9 (E9 DNase) forms channels in planar lipid bilayers (Mosbahi, K., Lema?tre, C., Keeble, A. H., Mobasheri, H., Morel, B., James, R., Moore, G. R., Lea, E. J., and Kleanthous, C. (2002) Nat. Struct. Biol. 9, 476-484). It was proposed that the E9 DNase mediates its own translocation across the cytoplasmic membrane and that the formation of ion channels is essential to this process. Here we describe changes to the structure and stability of the E9 DNase that accompany interaction of the protein with phospholipid vesicles. Formation of the protein-lipid complex at pH 7.5 resulted in a red-shift of the intrinsic protein fluorescence emission maximum (lambda(max)) from 333 to 346 nm. At pH 4.0, where the E9 DNase lacks tertiary structure but retains secondary structure, DOPG induced a blue-shift in lambda(max), from 354 to 342 nm. Changes in lambda(max) were specific for anionic phospholipid vesicles at both pHs, suggesting electrostatics play a role in this association. The effects of phospholipid were negated by Im9 binding, the high affinity, acidic, exosite inhibitor protein, but not by zinc, which binds at the active site. Fluorescence-quenching experiments further demonstrated that similar protein-phospholipid complexes are formed regardless of whether the E9 DNase is initially in its native conformation. Consistent with these observations, chemical and thermal denaturation data as well as proteolytic susceptibility experiments showed that association with negatively charged phospholipids destabilize the E9 DNase. We suggest that formation of a destabilizing protein-lipid complex pre-empts channel formation by the E9 DNase and constitutes the initial step in its translocation across the Escherichia coli inner membrane.     

Structure in the channel forming domain of colicin E1 bound to membranes: the 402-424 sequence

To explore the structure of the pore-forming fragment of colicin E1 in membranes, a series of 23 consecutive single cysteine substitution mutants was prepared in the sequence 402-424. Each mutant was reacted with a sulfhydryl-specific reagent to generate a nitroxide labeled side chain, and the mobility of the side chain and its accessibility to collision with paramagnetic reagents was determined from the electron paramagnetic resonance spectrum. Individual values of these quantities were used to identify tertiary contact sites and the nature of the surrounding solvent, while their periodic dependence on sequence position was used to identify secondary structure. In solution, the data revealed a regular helix of 11 residues in the region 406-416, consistent with helix IV of the crystal structure. Upon binding to negatively charged membranes at pH 4.0, helix IV apparently grows to a length of 19 residues, extending from 402-420. One face of the helix is solvated by the lipid bilayer, and the other by an environment of a polar nature. Surprisingly, a conserved charged pair, D408-R409, is located on the lipid-exposed face. Evidence is presented to suggest a transmembrane orientation of this new helix, although other topographies may exist in equilibrium.     

Association and dissociation kinetics of colicin E3 and immunity protein 3: convergence of theory and experiment

The rapid binding of cytotoxic colicin E3 by its cognate immunity protein Im3 is essential in safeguarding the producing cell. The X-ray structure of the E3/Im3 complex shows that the Im3 molecule interfaces with both the C-terminal ribonuclease (RNase) domain and the N-terminal translocation domain of E3. The association and dissociation rates of the RNase domain and Im3 show drastically different sensitivities to ionic strength, as previously rationalized for electrostatically enhanced diffusion-limited protein-protein associations. Relative to binding to the RNase domain, binding to full-length E3 shows a comparable association rate but a significantly lower dissociation rate. This outcome is just what was anticipated by a theory for the binding of two linked domains to a protein. The E3/Im3 system thus provides a powerful paradigm for the interplay of theory and experiment.     

On the nature of the unfolded intermediate in the in vitro transition of the colicin E1 channel domain from the aqueous to the membrane phase.

The transition of the colicin E1 channel polypeptide from a water-soluble to membrane-bound state occurs in vitro at acid pH values that are associated with an unfolded channel structure whose properties qualitatively resemble those of a "molten globule," or "compact unfolded," intermediate state. The role of such a state for activity was tested by comparing the pH dependence of channel-induced solute efflux and the amplitude of the near-UV CD spectrum. The requirement of a partly unfolded state for activity was shown by the coincidence of the onset of channel activity measured for 4 different lipid compositions with the decrease in near-UV CD amplitude as a function of pH. Tertiary constraints on the 3 tryptophans of the colicin channel, assayed by the amplitude of the near-UV CD spectrum, are retained over the pH range 3-4 where channel activity could be measured and, as well, at pH 2. In addition, the tryptophan fluorescence emission spectrum is virtually unchanged over the pH range 2-6. The temperature independence of the near-UV spectrum at pH 3-6 up to 70 degrees C implies that the colicin E1 channel polypeptide is more stable than that of colicin A. A transition between 53 and 58 degrees C in the amplitude of the near-UV CD is consistent with preservation of part of the hydrophobic core in a destabilized state at pH 2. Thus, the unfolded state associated with colicin activity at acidic pH has the properties of a "compact unfolded" state, having some, but not all of the properties of a "molten globule."(ABSTRACT TRUNCATED AT 250 WORDS)     

Change of thermal stability of colicin E7 triggered by acidic pH suggests the existence of unfolded intermediate during the membrane-translocation phase

Purified colicin E7 was analyzed by CD spectrum and gel filtration chromatography in a mimicking membrane-translocation phase. It was found that the CD spectra of colicin E7 at pH 7 and pH 2.5 were similar. Although the melting temperature of the protein shifted from 54.5°C to 34°C at low pH, the thermal denaturation curves of colicin E7 at different pH conditions still fit a two-state model. These experimental results imply that a minor structural change, triggered by acidic pH, for instance, may reduce the energy required for protein melting. In contrast to the minor change in secondary structure at different pH conditions, we observed that, in vitro, all monomeric colicin E7s converted into multimer-like conformations after recovering from the partial unfolding process. This multimeric form of colicin can only be dissociated by formamide and guanidine hydrochloride, indicating that this protein complex is indeed formed by aggregation of the monomeric colicins. Most interestingly, the aggregated colicins still perform in vivo bacteriocidal activity. We suggest that in a partial unfolding state the colicin is prepared for binding to the specific targets for translocation through the membrane. However, in the absence of specific targets in vitro these unfold intermediates may therefore aggregate into the multimeric form of colicins. Proteins 32:17–25, 1998. © 1998 Wiley-Liss, Inc.