Partitioning of Differently Sized Poly(ethylene glycol)s into OmpF Porin

To understand the physics of polymer equilibrium and dynamics in the confines of ion channel pores, we study partitioning of poly(ethylene glycol)s (PEGs) of different molecular weights into the bacterial porin, OmpF. Thermodynamic and kinetic parameters of partitioning are deduced from the effects of polymer addition on ion currents through single OmpF channels reconstituted into planar lipid bilayer membranes. The equilibrium partition coefficient is inferred from the average reduction of channel conductance in the presence of PEG; rates of polymer exchange between the pore and the bulk are estimated from PEG-induced conductance noise. Partition coefficient as a function of polymer weight is best fitted by a “compressed exponential” with the compression factor of 1.65. This finding demonstrates that PEG partitioning into the OmpF channel pore has sharper dependence on polymer molecular weight than predictions of hard-sphere, random-flight, or scaling models. A 1360-Da polymer separates regimes of partitioning and exclusion. Comparison of its characteristic size with the size of a 2200-Da polymer previously found to separate these regimes for the α-toxin shows good agreement with the x-ray structural data for these channels. The PEG-induced conductance noise is compatible with the polymer mobility reduced inside the OmpF pore by an order of magnitude relatively to its value in bulk solution.     

Ionic current blockades from DNA and RNA molecules in the alpha-hemolysin nanopore

We characterize the substate structure of current blockades produced when single-stranded polynucleotide molecules were electrophoretically driven into the alpha-hemolysin protein pore. We frequently observe substates where the ionic current is reduced by approximately 50%. Most of these substates can be associated with a molecular configuration where a polymer occupies only the vestibule region of the pore, though a few appear related to a polymer occupying only the transmembrane beta-barrel region of the pore. The duration of the vestibule configuration depends on polymer composition and on which end of the polymer, 3' or 5', subsequently threads into the narrowest constriction and initiates translocation. Below approximately 140 mV a polymer is more likely to escape from the vestibule against the applied voltage gradient, while at higher voltages a polymer is more likely to follow the voltage gradient by threading through the narrowest constriction and translocating through the pore. Increasing the applied voltage also increases the duration of the vestibule configuration. A semiquantitative model of these trends suggests that escape has stronger voltage dependence than threading, and that threading is sensitive to polymer orientation while escape is not. These results emphasize the utility of alpha-hemolysin as a model system to study biologically relevant physical and chemical processes at the single-molecule level.     

Polymeric nonelectrolytes to probe pore geometry: application to the alpha-toxin transmembrane channel

Asymmetrical (one-sided) application of penetrating water-soluble polymers, polyethylene glycols (PEGs), to a well-defined channel formed by Staphylococcus aureus alpha-toxin is shown to probe channel pore geometry in more detail than their symmetrical (two-sided) application. Polymers added to the cis side of the planar lipid membrane (the side of protein addition) affect channel conductance differently than polymers added to the trans side. Because a satisfactory theory quantitatively describing PEG partitioning into a channel pore does not exist, we apply the simple empirical rules proposed previously (, J. Membr. Biol. 161:83-92) to gauge the size of pore openings as well as the size and position of constrictions along the pore axis. We estimate the radii of the two openings of the channel to be practically identical and equal to 1. 2-1.3 nm. Two apparent constrictions with radii of approximately 0. 9 nm and approximately 0.6-0.7 nm are inferred to be present in the channel lumen, the larger one being closer to the cis side. These structural findings agree well with crystallographic data on the channel structure (, Science. 274:1859-1866) and verify the practicality of polymer probing. The general features of PEG partitioning are examined using available theoretical considerations, assuming there is no attraction between PEG and the channel lumen. It is shown that the sharp dependence of the partition coefficient on polymer molecular weight found under both symmetrical and asymmetrical polymer application can be rationalized within a "hard sphere nonideal solution model." This finding is rather surprising because PEG forms highly flexible coils in water with a Kuhn length of only several Angstroms.     

Probing distance and electrical potential within a protein pore with tethered DNA

DNA molecules tethered inside a protein pore can be used as a tool to probe distance and electrical potential. The approach and its limitations were tested with alpha-hemolysin, a pore of known structure. A single oligonucleotide was attached to an engineered cysteine to allow the binding of complementary DNA strands inside the wide internal cavity of the extramembranous domain of the pore. The reversible binding of individual oligonucleotides produced transient current blockades in single channel current recordings. To probe the internal structure of the pore, oligonucleotides with 5' overhangs of deoxyadenosines and deoxythymidines up to nine bases in length were used. The characteristics of the blockades produced by the oligonucleotides indicated that single-stranded overhangs of increasing length first approach and then thread into the transmembrane beta-barrel. The distance from the point at which the DNA was attached and the internal entrance to the barrel is 43 A, consistent with the lengths of the DNA probes and the signals produced by them. In addition, the tethered DNAs were used to probe the electrical potential within the protein pore. Binding events of oligonucleotides with an overhang of five bases or more, which threaded into the beta-barrel, exhibited shorter residence times at higher applied potentials. This finding is consistent with the idea that the main potential drop is across the alpha-hemolysin transmembrane beta-barrel, rather than the entire length of the lumen of the pore. It therefore explains why the kinetics and thermodynamics of formation of short duplexes within the extramembranous cavity of the pore are similar to those measured in solution, and bolsters the idea that a "DNA nanopore" provides a useful means for examining duplex formation at the single molecule level.     

A functional protein pore with a "retro" transmembrane domain.

Extended retro (reversed) peptide sequences have not previously been accommodated within functional proteins. Here, we show that the entire transmembrane portion of the beta-barrel of the pore-forming protein alpha-hemolysin can be formed by retrosequences comprising a total of 175 amino acid residues, 25 contributed by the central sequence of each subunit of the heptameric pore. The properties of wild-type and retro heptamers in planar bilayers are similar. The single-channel conductance of the retro pore is 15% less than that of the wild-type heptamer and its current-voltage relationship denotes close to ohmic behavior, while the wild-type pore is weakly rectifying. Both wild-type and retro pores are very weakly anion selective. These results and the examination of molecular models suggest that beta-barrels may be especially accepting of retro sequences compared to other protein folds. Indeed, the ability to form a retro domain could be diagnostic of a beta-barrel, explaining, for example, the activity of the retro forms of many membrane-permeabilizing peptides. By contrast with the wild-type subunits, monomeric retro subunits undergo premature assembly in the absence of membranes, most likely because the altered central sequence fails to interact with the remainder of the subunit, thereby initiating assembly. Despite this difficulty, a technique was devised for obtaining heteromeric pores containing both wild-type and retro subunits. Most probably as a consequence of unfavorable interstrand side-chain interactions, the heteromeric pores are less stable than either the wild-type or retro homoheptamers, as judged by the presence of subconductance states in single-channel recordings. Knowledge about the extraordinary plasticity of the transmembrane beta-barrel of alpha-hemolysin will be very useful in the de novo design of functional membrane proteins based on the beta-barrel motif.     

Location of a constriction in the lumen of a transmembrane pore by targeted covalent attachment of polymer molecules

Few methods exist for obtaining the internal dimensions of transmembrane pores for which 3-D structures are lacking or for showing that structures determined by crystallography reflect the internal dimensions of pores in lipid bilayers. Several approaches, involving polymer penetration and transport, have revealed limiting diameters for various pores. But, in general, these approaches do not indicate the locations of constrictions in the channel lumen. Here, we combine cysteine mutagenesis and chemical modification with sulfhydryl-reactive polymers to locate the constriction in the lumen of the staphylococcal alpha-hemolysin pore, a model protein of known structure. The rates of reaction of each of four polymeric reagents (MePEG-OPSS) of different masses towards individual single cysteine mutants, comprising a set with cysteines distributed over the length of the lumen of the pore, were determined by macroscopic current recording. The rates for the three larger polymers (1.8, 2.5, and 5.0 kD) were normalized with respect to the rates of reaction with a 1.0-kD polymer for each of the seven positions in the lumen. The rate of reaction of the 5.0-kD polymer dropped dramatically at the centrally located Cys-111 residue and positions distal to Cys-111, whether the reagent was applied from the trans or the cis side of the bilayer. This semi-quantitative analysis sufficed to demonstrate that a constriction is located at the midpoint of the pore lumen, as predicted by the crystal structure, and although the constriction allows a 2.5-kD polymer to pass, transport of a 5.0-kD molecule is greatly restricted. In addition, PEG chains gave greater reductions in pore conductance when covalently attached to the narrower regions of the lumen, permitting further definition of the interior of the pore. The procedures described here should be applicable to other pores and to related structures such as the vestibules of ion channels.     

Transport of alpha-helical peptides through alpha-hemolysin and aerolysin pores

A series of negatively charged alpha-helical peptides of the general formula fluorenylmethoxycarbonyl (Fmoc)-D(x)A(y)K(z) were synthesized, where x and z were 1, 2, or 3 and y was 10, 14, 18, or 22. The translocation of the peptides through single pores, which were self-assembled into lipid membranes, was analyzed by measuring the current blockade i(block) and the duration t(block). The pores were either alpha-hemolysin, which has a wide vestibule leading into the pore, or aerolysin, which has no vestibule but has a longer pore of a similar diameter. Many thousands of events were measured for each peptide with each pore, and they could be assigned to two types: bumping events (type I) have a small i(block) and long t(block), and translocation events (type II) have a larger i(block) and shorter t(block). For type-II events, both i(block) and t(block) increase with the length of the peptides on both pores tested. The dipole moment and the net charge of each peptide has a major effect on the transport characteristics. The ratio of type-II/type-I events increases as the dipole moment increases, and uncharged peptides gave mostly type-I events. The structural differences between the two nanopores were reflected in the characteristic values of i(block), and in particular, the vestibule of alpha-hemolysin helps to orient the peptides for translocation. Overall, the results demonstrate that the nanopore technology can provide useful structural information but peptide sequencing will require further improvements in the design of the pores.     

Subunit composition of a bicomponent toxin: staphylococcal leukocidin forms an octameric transmembrane pore

Staphylococcal leukocidin pores are formed by the obligatory interaction of two distinct polypeptides, one of class F and one of class S, making them unique in the family of beta-barrel pore-forming toxins (beta-PFTs). By contrast, other beta-PFTs form homo-oligomeric pores; for example, the staphylococcal alpha-hemolysin (alpha HL) pore is a homoheptamer. Here, we deduce the subunit composition of a leukocidin pore by two independent methods: gel shift electrophoresis and site-specific chemical modification during single-channel recording. Four LukF and four LukS subunits coassemble to form an octamer. This result in part explains properties of the leukocidin pore, such as its high conductance compared to the alpha HL pore. It is also pertinent to the mechanism of assembly of beta-PFT pores and suggests new possibilities for engineering these proteins.     

The leukocidin pore: evidence for an octamer with four LukF subunits and four LukS subunits alternating around a central axis

The staphylococcal alpha-hemolysin (alphaHL) and leukocidin (Luk) polypeptides are members of a family of related beta-barrel pore-forming toxins. Upon binding to susceptible cells, alphaHL forms water-filled homoheptameric transmembrane pores. By contrast, Luk pores are formed by two classes of subunit, F and S, rendering a heptameric structure displeasing on symmetry grounds at least. Both the subunit stoichiometry and arrangement within the Luk pore have been contentious issues. Here we use chemical and genetic approaches to show that (1) the predominant, or perhaps the only, form of the Luk pore is an octamer; (2) the subunit stoichiometry is 1:1; and (3) the subunits are arranged in an alternating fashion about a central axis of symmetry, at least when a fused LukS-LukF construct is used. The experimental approaches we have used also open up new avenues for engineering the arrangement of the subunits of beta-barrel pore-forming toxins.     

Association of hydrophobically-modified poly(ethylene glycol) with fusogenic liposomes

We present results on using cooperative interactions to shield liposomes by incorporating multiple hydrophobic anchoring sites on polyethylene glycol (PEG) polymers. The hydrophobically-modified PEGs (HMPEGs) are comb-graft polymers with strictly alternating monodisperse PEG blocks (M(w)=6, 12, or 35 kDa) bonded to C18 stearylamide hydrophobes. Cooperativity is varied by changing the degree of oligomerization at a constant ratio of PEG to stearylamide. Fusogenic liposomes prepared from N-C12-DOPE:DOPC 7:3 (mol:mol) were equilibrated with HMPEGs. Affinity for polymer association to liposomes increases with the degree of oligomerization; equilibrium constants (given as surface coverage per equilibrium concentration of free polymer) for 6 kDa PEG increased from 6.1+/-0.8 (mg/m(2))/(mg/ml) for 2.5 loops to 78.1+/-12.2 (mg/m(2))/(mg/ml) for 13 loops. In contrast, the equilibrium constant for distearoylphosphatidylethanolamine-poly(ethylene glycol) (DSPE-PEG5k) was 0.4+/-0.1 (mg/m(2))/(mg/ml).The multi-loop HMPEGs demonstrate higher levels of protection from complement binding than DSPE-PEG5k. Greater protection does not correlate with binding strength alone. The best shielding was by HMPEG6k-DP3 (with three 6 kDa PEG loops), suggesting that PEG chains with adequate surface mobility provide optimal protection from complement opsonization. Complement binding at 30 min and 12 h demonstrates that protection by multi-looped PEGs is constant whereas DSPE-PEG5k initially protects but presumably partitions off of the surface at longer times.     

The internal cavity of the staphylococcal alpha-hemolysin pore accommodates approximately 175 exogenous amino acid residues

The cavity within the cap domain of the transmembrane staphylococcal alpha-hemolysin (alphaHL) pore is roughly a sphere of diameter approximately 45 A (molecular surface volume approximately 39,500 A(3)). We tested the ability of the cavity to accommodate exogenous polypeptide chains. Concatemerized Gly/Ser-containing sequences ("loops", L; number of repeats = n; number of residues = 10n + 5, n = 0-21) were inserted at a position located within the cavity of the fully assembled heptameric alphaHL pore. Homomeric pores containing 25 or less residues in each loop (n or= 7, only one L subunit was incorporated. As the inserted loop was lengthened, transient closures were observed in planar bilayer experiments with single pores. However, L(1)W(6) pores with very long loops (n = 14 and 21) had unitary conductance values close to those of W(7), suggesting that the loop is extruded through the opening in the cap of the pore into the external medium. Further analysis of bilayer recordings and electrophoretic migration patterns indicates that the upper capacity of the cavity is approximately 175 amino acids. The findings suggest that small functional peptides or proteins might be assembled within the alphaHL pore.     

Prolonged residence time of a noncovalent molecular adapter, beta-cyclodextrin, within the lumen of mutant alpha-hemolysin pores.

Noncovalent molecular adapters, such as cyclodextrins, act as binding sites for channel blockers when lodged in the lumen of the alpha-hemolysin (alphaHL) pore, thereby offering a basis for the detection of a variety of organic molecules with alphaHL as a sensor element. beta-Cyclodextrin (betaCD) resides in the wild-type alphaHL pore for several hundred microseconds. The residence time can be extended to several milliseconds by the manipulation of pH and transmembrane potential. Here, we describe mutant homoheptameric alphaHL pores that are capable of accommodating betaCD for tens of seconds. The mutants were obtained by site-directed mutagenesis at position 113, which is a residue that lies near a constriction in the lumen of the transmembrane beta barrel, and fall into two classes. Members of the tight-binding class, M113D, M113N, M113V, M113H, M113F and M113Y, bind betaCD approximately 10(4)-fold more avidly than the remaining alphaHL pores, including WT-alphaHL. The lower K(d) values of these mutants are dominated by reduced values of k(off). The major effect of the mutations is most likely a remodeling of the binding site for betaCD in the vicinity of position 113. In addition, there is a smaller voltage-sensitive component of the binding, which is also affected by the residue at 113 and may result from transport of the neutral betaCD molecule by electroosmotic flow. The mutant pores for which the dwell time of betaCD is prolonged can serve as improved components for stochastic sensors.     

Probing protein nanopores with poly(ethylene glycol)s

Neutral water-soluble poly(ethylene glycol)s (PEGs) have been extensively explored in protein nanopore research for the past several decades. The principal use of PEGs is to investigate the membrane protein ion channel physical characteristics and transport properties. In addition, protein nanopores are used to study polymer–protein interactions and polymer physicochemical properties. In this review, we focus on the biophysical studies on probing protein ion channels with PEGs, specifically on nanopore sizing by PEG partitioning. We discuss the fluctuation analysis of ion channel currents in response to the PEGs moving within their confined geometries. The advantages, limitations, and recent developments of the approach are also addressed.     

Association of hydrophobically-modified poly(ethylene glycol) with fusogenic liposomes

We present results on using cooperative interactions to shield liposomes by incorporating multiple hydrophobic anchoring sites on polyethylene glycol (PEG) polymers. The hydrophobically-modified PEGs (HMPEGs) are comb-graft polymers with strictly alternating monodisperse PEG blocks (M_w=6, 12, or 35 kDa) bonded to C18 stearylamide hydrophobes. Cooperativity is varied by changing the degree of oligomerization at a constant ratio of PEG to stearylamide. Fusogenic liposomes prepared from N-C12-DOPE:DOPC 7:3 (mol:mol) were equilibrated with HMPEGs. Affinity for polymer association to liposomes increases with the degree of oligomerization; equilibrium constants (given as surface coverage per equilibrium concentration of free polymer) for 6 kDa PEG increased from 6.1±0.8 (mg/m²)/(mg/ml) for 2.5 loops to 78.1±12.2 (mg/m²)/(mg/ml) for 13 loops. In contrast, the equilibrium constant for distearoylphosphatidylethanolamine-poly(ethylene glycol) (DSPE-PEG5k) was 0.4±0.1 (mg/m²)/(mg/ml).The multi-loop HMPEGs demonstrate higher levels of protection from complement binding than DSPE-PEG5k. Greater protection does not correlate with binding strength alone. The best shielding was by HMPEG6k-DP3 (with three 6 kDa PEG loops), suggesting that PEG chains with adequate surface mobility provide optimal protection from complement opsonization. Complement binding at 30 min and 12 h demonstrates that protection by multi-looped PEGs is constant whereas DSPE-PEG5k initially protects but presumably partitions off of the surface at longer times.     

Through pore diameter in the cell wall of Chara corallina.

Determination of pore size of the cell wall of Chara corallina has been made by using the polyethylene glycol (PEG) series as the hydrophilic probing molecules. In these experiments, the polydispersity of commercial preparation of PEGs was allowed for. The mass share (gamma(p)) of polyethylene glycol preparation fractions penetrating through the pores was determined using a cellular 'ghost', i.e. fragments of internodal cell walls filled with a 25% solution of non-penetrating PEG 6000 and tied up at the ends. In water, such a 'ghost' developed a hydrostatic pressure close to the cell turgor which persisted for several days. The determination of gamma(p), for polydisperse polyethylene glycols with different average molecular mass (M) was calculated from the degree of pressure restoration after water was replaced by a 5-10% polymer solution. Pressure was recorded using a dynamometer, which measures, in the quasi-isometric mode, the force necessary for the partial compression of the 'ghost' in its small fragment. By utilizing the data on the distribution of PEG 1000, 1450, 2000, and 3350 fractions over molecular mass (M), it was found that gamma(p), for these polyethylene glycols corresponded to the upper limit of ML=800-1100 D (hydrodynamic radius of molecules, r(h)=0.85-1.05 nm). Thus, the effective diameter of the pores in the cell wall of Chara did not exceed 2.1 nm.     

Passage times for polymer translocation pulled through a narrow pore

We study the passage times of a translocating polymer of length N in three dimensions, while it is pulled through a narrow pore with a constant force F applied to one end of the polymer. At small to moderate forces, satisfying the condition FN(nu)/k(B)T less, similar 1, where nu approximately 0.588 is the Flory exponent for the polymer, we find that tau(N), the mean time the polymer takes to leave the pore, scales as N(2+nu) independent of F, in agreement with our earlier result for F = 0. At strong forces, i.e., for, FN(nu)/k(B)T > 1, the behavior of the passage time crosses over to tau(N) approximately N(2)/F. We show here that these behaviors stem from the polymer dynamics at the immediate vicinity of the pore-in particular, the memory effects in the polymer chain tension imbalance across the pore.     

Molecular dynamics simulations of water within models of ion channels.

The transbilayer pores formed by ion channel proteins contain extended columns of water molecules. The dynamic properties of such waters have been suggested to differ from those of water in its bulk state. Molecular dynamics simulations of ion channel models solvated within and at the mouths of their pores are used to investigate the dynamics and structure of intra-pore water. Three classes of channel model are investigated: a) parallel bundles of hydrophobic (Ala20) alpha-helices; b) eight-stranded hydrophobic (Ala10) antiparallel beta-barrels; and c) parallel bundles of amphipathic alpha-helices (namely, delta-toxin, alamethicin, and nicotinic acetylcholine receptor M2 helix). The self-diffusion coefficients of water molecules within the pores are reduced significantly relative to bulk water in all of the models. Water rotational reorientation rates are also reduced within the pores, particularly in those pores formed by alpha-helix bundles. In the narrowest pore (that of the Ala20 pentameric helix bundle) self-diffusion coefficients and reorientation rates of intra-pore waters are reduced by approximately an order of magnitude relative to bulk solvent. In Ala20 helix bundles the water dipoles orient antiparallel to the helix dipoles. Such dipole/dipole interaction between water and pore may explain how water-filled ion channels may be formed by hydrophobic helices. In the bundles of amphipathic helices the orientation of water dipoles is modulated by the presence of charged side chains. No preferential orientation of water dipoles relative to the pore axis is observed in the hydrophobic beta-barrel models.     

Size selectivity between gap junction channels composed of different connexins

Gap junction channels are traditionally viewed as large, nonspecific pores connecting cells. Recently the diversity in the connexin family has drawn more attention to their permeability characteristics. Several studies have shown that both size and charge contribute to the permeability of gap junctional channels. We have used a graded series of neutral polyethylene glycol probes (PEGs), which eliminate charge contribution completely, to specifically assess the physical exclusion limits of gap junction channels formed by different connexins. Cx 26, 32 and 37 were expressed in paired Xenopus oocytes to form homotypic gap junctional channels. PEG probes were perfused intracellularly into one side of the oocyte pair. A reversible drop in conductance of the gap juctional channels indicated that the probe was small enough to enter the pore and hinder ion flux. Our data suggest that Cx32 channels have a size cut-off between PEG 400 (11.2 A) and PEG 300 (9.6 A) despite their relatively small single channel conductance (approximately 55 pS). Cx26 channels (approximately 130 pS single channel conductance) have a size exclusion limit around PEG 200 (8.0 A), while Cx37 channels show the most restricted size cut-off between PEG 200 (8.0 A) and TriEG (6.8 A), despite having the largest unitary conductance (approximately 300 pS).     

Field-dependent effect of crown ether (18-crown-6) on ionic conductance of alpha-hemolysin channels

Closing linear poly(ethylene glycol) (PEG) into a circular "crown" dramatically changes its dynamics in the alpha-hemolysin channel. In the electrically neutral crown ether (C2H4O)6, six ethylene oxide monomers are linked into a circle that gives the molecule ion-complexing capacity and increases its rigidity. As with linear PEG, addition of the crown to the membrane-bathing solution decreases the ionic conductance of the channel and generates additional conductance noise. However, in contrast to linear PEG, both the conductance reduction (reporting on crown partitioning into the channel pore) and the noise (reporting on crown dynamics in the pore) now depend on voltage strongly and nonmonotonically. Within the whole frequency range accessible in channel reconstitution experiments, the noise power spectrum is "white", showing that crown exchange between the channel and the bulk solution is fast. Analyzing these data in the framework of a Markovian two-state model, we are able to characterize the process quantitatively. We show that the lifetime of the crown in the channel reaches its maximum (a few microseconds) at about the same voltage (approximately 100 mV, negative from the side of protein addition) where the crown's reduction of the channel conductance is most pronounced. Our interpretation is that, because of its rigidity, the crown feels an effective steric barrier in the narrowest part of the channel pore. This barrier together with crown-ion complexing and resultant interaction with the applied field leads to behavior usually associated with voltage-dependent binding in the channel pore.     

Crystal structure of the monomeric porin OmpG

The outer membrane (OM) of Gram-negative bacteria contains a large number of channel proteins that mediate the uptake of ions and nutrients necessary for growth and functioning of the cell. An important group of OM channel proteins are the porins, which mediate the non-specific, diffusion-based passage of small (<600 Da) polar molecules. All porins of Gram-negative bacteria that have been crystallized to date form stable trimers, with each monomer composed of a 16-stranded beta-barrel with a relatively narrow central pore. In contrast, the OmpG porin is unique, as it appears to function as a monomer. We have determined the X-ray crystal structure of OmpG from Escherichia coli to a resolution of 2.3 A. The structure shows a 14-stranded beta-barrel with a relatively simple architecture. Due to the absence of loops that fold back into the channel, OmpG has a large ( approximately 13 A) central pore that is considerably wider than those of other E. coli porins, and very similar in size to that of the toxin alpha-hemolysin. The architecture of the channel, together with previous biochemical and other data, suggests that OmpG may form a non-specific channel for the transport of larger oligosaccharides. The structure of OmpG provides the starting point for engineering studies aiming to generate selective channels and for the development of biosensors.