Functional assay of Salmonella typhi OmpC using

Biochimie 88 (2006) 1419–1424 www.elsevier.com/locate/biochi

Functional assay of Salmonella typhi OmpC using reconstituted large unilamellar vesicles: a general method for characterization of outer membrane proteins N. Sundara Baalaji a, M.K. Mathew b, S. Krishnaswamy a,* a

Center of Excellence in Bioinformatics, School of Biotechnology, Madurai-Kamaraj University, Palkalainagar, Madurai 625021, India b National Center for Biological Sciences, Bangalore 560065, India Received 17 March 2006; accepted 11 May 2006 Available online 24 May 2006

Abstract The immunodominant trimeric β-barrel outer membrane protein OmpC from Salmonella typhi, the causative agent of typhoid, has been functionally characterized here. The activity in the vesicle environment was studied in vitro using OmpC reconstituted into proteoliposomes. Passage of polysaccharides and polyethyleneglycols through OmpC has been examined to determine the permeability properties. The relative rate of neutral solute flux yields a radius of 1.1 nm for the S. typhi OmpC pore. This is almost double the pore size of Escherichia coli. This provides an example of large pore size present in the porins that form trimers as in the general bacterial porin family. The method used in this study provides a good membrane model for functional studies of porins. © 2006 Elsevier Masson SAS. All rights reserved. Keywords: Salmonella typhi; Membrane porin; OmpC; PEG; Liposome swelling assay; LUV; Pore size

1. Introduction The protective outer membrane of gram-negative bacteria provides an efficient permeability barrier against high molecular weight solutes. Outer membrane proteins called porins allow passage of nutrients and antibiotics. Porins can be either general diffusion pores or substrate specific pores. A few wellknown porins are osmoporin (OmpC, OmpF), phosphoporin (PhoE) from the general bacterial porin family and sucrose porin (ScrY), maltoporin (LamB) from the sugar porin family (www.tcdb.org). These proteins assemble as trimers of beveled β-barreled cylinders, the short sides of the bevel facing the interior of the trimer [1]. Osmotic shifts and other changes in the environment influence the structures and properties of biological membranes (including liposomes and proteoliposomes) and can be used to characterize the activities of membraneassociated proteins. In cells, membrane protein mediated transAbbreviations: LUV, large unilamellar vesicle; MLV, multi lamellar vesicle; OmpC, outer membrane protein C; PEG, polyethyleneglycol. * Corresponding author. E-mail address: [email protected] (S. Krishnaswamy). 0300-9084/$ - see front matter © 2006 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.biochi.2006.05.010

port may be confounded by physiological processes that modulate transport either directly or indirectly. Reconstitution into artificial membrane systems such as liposomes or planar bilayers, allows the study of membrane protein function in isolation. Rapid swelling of liposomes following influx of solutes through reconstituted pores was used to demonstrate that LamB preferentially transports malto-oligosaccharides [2]. Nikaido and Rosenberg [3] used the same liposome swelling assay (LSA) to estimate the effective size of transmembrane pores of Escherichia coli outer membrane using multi lamellar vesicles (MLV). They further modified the reconstitution method [4] to study the influence of size, charge, and hydrophobicity on solute flux through E. coli K12 porins OmpC, OmpF and PhoE. Subsequently, polyethyleneglycols (PEGs) were also used as neutral solutes to study channel characteristics and pore insertion into liposomes [5]. However, the Salmonella porins have so far not been characterized. OmpC and OmpF are relatively non-specific, water-filled pores that act as passive diffusion channels with slightly different conductance and exclusion properties. The EnvZ-OmpR

1420

N. Sundara Baalaji et al. / Biochimie 88 (2006) 1419–1424

osmoregulation system regulates the differential expression of OmpC and OmpF depending on the osmolarity conditions. In E. coli, under conditions of high osmolarity, OmpC expression is enhanced while OmpF expression is repressed and in low osmolarity conditions OmpF expression is enhanced [6]. However, unlike in the case of E. coli, in Salmonella typhi, the causative organism for typhoid, the immunodominant [7] OmpC is expressed in both low and high osmolarity conditions [8], possibly because Salmonella encounters different osmolarity conditions in the host during the life cycle. This work characterizes the pore properties of S. typhi OmpC through osmotic shifts in the solution environment by reconstitution of OmpC into model membranes. The study is a first report on characterization of a bacterial porin reconstituted in large unilamellar vesicles (LUVs) using PEGs as neutral solutes.

MLVs were extruded using a hand held extrusion device [12] obtained from Avanti Polar Lipids Inc. (Alabaster, AL, USA) to generate large unilamellar liposomes (LUVs). Lipid dispersions were passed 21 times through a 19 mm diameter Nucleopore Polycarbonate Filter (Avanti Polar Lipids Inc.) with 200 nm diameter pores. This suspension was then subjected to three cycles of freeze–thaw: immersion in liquid nitrogen for 30 s followed by thawing at 37 °C. Control liposomes were prepared in a similar fashion without added protein. The liposome preparation was kept at room temperature for at least 30 min after the last thaw to attain complete sealing of LUVs. 2.4. Permeability assay by liposome swelling

Soybean lecithin (Avanti Polar Lipids, Alabaster, AL) was stored at –20 °C under an atmosphere of argon to prevent lipid peroxidation. Polysaccharides and PEG were from Sigma (St. Louis, MI). Dextran (MW 10,000 Da) was used as impermeant (Molecular Probes, Carlsbad, CA). The liposome solution was buffered with 5 mM MgCl2, 5 mM 4-(2-hydroxyethyl)-1-piperazine ethane sulfonic acid (HEPES). Test solutes were prepared in the same buffer for isotonic experiments.

The relative rate of permeation of non-electrolytes was determined by the initial rate of swelling of vesicles upon dilution into isotonic solutions (see below) of test compounds [3, 4]. The PEGs used along with their hydrodynamic radius RH [13,14] given in parentheses are: PEG200 (0.40 nm); PEG300 (0.48 nm); PEG400 (0.56 nm); PEG600 (0.69 nm); PEG900 (0.85 nm); PEG1000 (0.89 nm); and PEG1450 (1.1 nm). Ten microliters of a liposome suspension was diluted into 990 μl of an isotonic solution of test solute in buffer A in a 1-ml microcuvette and mixed manually. The change in absorbance at 520 nm was recorded at 2 s intervals for 3–10 min. All absorbance measurements were made on a Cary-Varian 1 UV–VIS spectrophotometer, operated in the kinetic measurement mode. The time to the first reading was 7–10 s.

2.2. Protein purification

2.5. Determination of isotonic concentration for test solutes

OmpC was extracted and purified from S. typhi Ty21a, a galE mutant [9] and from E. coli HB101, by a modified salt extraction method involving successive steps where the amount of SDS is reduced while the NaCl concentration is increased [10]. OmpC was passed through Q-sepharose twice to remove bound LPS to a large extent. The porin sample was concentrated and exchanged into a buffer containing 50 mM Tris–HCl pH 7.2, 1% octyl PoE and 3 mM NaN3 using 70 kDa molecular weight cut-off Amicon ultrafiltration devices. The purity of OmpC samples was determined on 13% SDS-PAGE by silver staining.

The isotonic concentration of each solute was determined by diluting control liposomes (without protein) made in 3 mM dextran into different concentrations of test solutes while monitoring changes in light scattering at 520 nm. The concentration of solute at which no change in absorbance was observed, was considered the isotonic concentration [2].

2. Materials and methods 2.1. Reagents and buffers

2.3. Reconstitution of proteoliposomes Purified OmpC from S. typhi Ty21a, a galE mutant, was reconstituted into LUV by a modification of the sonication freeze–thaw procedure [11]. Briefly, a film of lipid was deposited in a round-bottomed flask by swirling a chloroform solution of soybean lecithin (50 mg lecithin in 1 ml chloroform) under a stream of argon gas. The flask was then placed in a vacuum desiccator overnight. The dried lipid film was hydrated in 1 ml of buffer A (5 mM HEPES-NaOH, pH 7.4 and 5 mM MgCl2) containing 3 mM dextran by vortexing for 30 s followed by a 30 s standing phase to generate MLVs. Detergent solubilized OmpC protein was added in a 1:25 w/w protein/ lipid ratio to generate multilamellar proteoliposomes.

2.6. Data analysis and pore size estimation All data were analyzed and graphical representations made using Origin 6.1 (OriginLab Corporation, Northampton, MA). Relative rates of diffusion were determined from the initial slopes of changes in optical density over the first 30 s, normalized with respect to the slope obtained for a reference solute. The pore diameter was estimated by fitting to the Renkin [15] equation. The Renkin formalism provides an estimate of the total restriction to diffusion, due to the combined effects of steric hindrance at the entrance to the pores and frictional resistance within the pores. This is given by: A=Ao ¼ ½1
ða=rÞ2 ½1
2:104ða=rÞ þ 2:09ða=rÞ3
0:95ða=rÞ5  where ‘A’ is the effective area of the opening, ‘Ao’ the total cross-sectional area of the pore, ‘r’ radius of the pore and ‘a’

N. Sundara Baalaji et al. / Biochimie 88 (2006) 1419–1424

1421

is radius of the molecule. The ratio (A/Ao) is equivalent to the permeability of the membrane.

general diffusion porins and other outer membrane porins like mitochondrial porins.

3. Results and discussion

3.1. Diffusion through S. typhi and E. coli OmpC channels

Outer membrane proteins from gram-negative bacteria have so far been found to be made up of β-barrels [16]. Two to three percent of the sequences in bacterial genomes were identified to encode β-barrel outer membrane proteins [17,18]. Functional characterization of these ubiquitous β-barrel structures is vital in understanding processes like transport, signal transduction and to link them with structural features. Purification from the native membrane and incorporation of outer membrane protein into an artificial membrane continue to be crucial steps in studying functional properties. Reproducible methods of reconstitution allow molecular information on the channel lumen and details of functional events to be obtained. Here we report on pore size of the immunodominant OmpC from the human pathogen S. typhi using LUVs. Subsequently, our group has also used this method for characterizing S. typhi general diffusion porins OmpF, recombinant OmpC and insertion/ deletion mutants of OmpC. In our present studies we have modified and combined existing methods of liposome preparation: freezing, extrusion, sonication to characterize the functional activity of OmpC with a unilamellar system. Initially we dispersed a uniform lipid film into an aqueous phase by thorough vortexing, leading to a suspension of MLVs. The disadvantages of this system, a heterogeneous size distribution and low entrapment capacity, was overcome by using extrusion in subsequent steps. The heterogeneous population of fairly large liposomes (with high turbidity) was passed through polycarbonate membranes under hand pressure. This simple technique has been reported to result in a relatively homogeneous suspension of vesicles. EM and 31P NMR studies report that sequential extrusion produces a population of vesicles, 90% of which are unilamellar with a homogenous size distribution around the membrane cutoff [19]. Extrusion was followed by freeze thawing of liposomes. Dehydration and rehydration of liposomes by the freezing and thawing phases facilitates tight sealing of lipid membranes and complete equilibration of the impermeant solute, dextran, between the interior of the LUVs and the exterior aqueous phase. Non-uniform distribution of entrapped solute between bulk water and interlamellar water in MLVs results in complicated swelling behavior [20]. At the beginning of the swelling phase, some layers of the MLVs may swell and some others shrink hence giving only a qualitative demonstration of solute penetration through the channel. The apparent isotonic concentration of different solutes can thus vary over a significant range. With unilamellar liposomes, the isotonic concentration range determined with control liposomes for various solutes is very narrow between 15 and 18 mM for saccharides and 10 mM for PEGs. This, in turn, means that swelling assays are performed with very similar concentration gradients for the entire gamut of solutes analyzed. The LSA developed here using reconstituted LUVs can be suitable for evaluation of

OmpC was extracted from outer membranes of S. typhi or E. coli and purified into the non-ionic detergent octyl polyoxy ethylene. The purified OmpC was reconstituted into unilamellar soy lecithin vesicles by freeze–thaw cycles and extrusion (see Section 2). The reconstituted liposomes contain dextran (MW 10,000 Da) as a solute too large to pass through OmpC. The intravesicular dextran ensures the absence of an osmotic gradient when suspended in an isosmotic solution of the test solute. Proteoliposomes showed rapid influx of uncharged molecules like pentoses and hexoses. The passage through the pore is reflected by the rapid decrease in absorbance of liposomes diluted into isosmotic solutions of these sugars (Fig. 1A, B). The decrease in absorbance as a consequence of reduced light scattering is indicative of liposome swelling. Control liposomes prepared by this method without reconstituted protein were not permeable to solutes (Fig. 1A). The swelling assay shows that OmpC is not denatured by SDS treatment during the purification step. Thus, the functional unit is preserved after detergent solubilization from outer membranes followed by several rounds of detergent exchange. S. typhi OmpC and E. coli OmpC were reconstituted in a similar manner and at the same protein, impermeant concentration (3 mM dextran) and protein/lipid ratio (1:25 w/w). Stachyose was able to pass through the S. typhi OmpC channel (Fig. 1A) however it was found to be impermeable through E. coli OmpC [4]. Resuspending the E. coli OmpC carrying proteoliposomes in a stachyose solution resulted in no appreciable swelling, indicating that the pore could not accommodate stachyose (Fig. 1B). The radius of the pore estimated from our data for E. coli HB101 OmpC is 0.62 nm that is comparable to the radius of 0.54 nm reported earlier from E. coli K12 OmpC [4]. The rate of swelling decreases gradually as the molecular weight of the solute increases, clearly demonstrating passage of solutes through a pore of defined size, comparable to the size of the solutes tested. Estimating the fraction of protein that retains native structure in the proteoliposomes is not straightforward using the swelling assay. It may be noted that the freeze–thaw extrusion method avoids harsh conditions such as sonication and reverse phase evaporation thereby limiting possibilities of denaturation. Reconstitution by reverse phase evaporation, of alkaline phosphatase, for example, results in the retention of only 40% enzymatic activity [21]. 3.2. Diffusion of PEGs through S. typhi OmpC Facile passage of stachyose through the S. typhi OmpC pore posed the requirement of still larger solutes to be used in the assay. Since the Renkin formalism assumes spherical particles passing through a cylindrical pore, it is critical to ensure that the solutes used be approximately spherical. Long oligosaccharides are increasingly non-spherical, so recourse was taken to using PEGs, which have a short persistence length of 3.8 Å

1422

N. Sundara Baalaji et al. / Biochimie 88 (2006) 1419–1424

Fig. 1. Passage of oligosaccharides through (A) S. typhi OmpC (B) E. coli OmpC was examined by monitoring decrease in absorbance at 520 nm. Solutes were used at isotonic concentration of 15–18 mM. The solutes used were arabinose (ARA), fructose (FRC), sucrose (SCR), raffinose (RAF) and stachyose (STAC). Open symbols represent proteoliposomes and filled symbols represent control liposomes.

[22]. With a persistence length small compared to the contour length, this polymer approximates a sphere more closely than do oligosaccharides. PEGs have been used for probing pore geometry of OmpF and toxin channels in electrophysiological studies [23,24]. The hydrodynamic radius of the pore formed by sticholysin I in RBC was determined to be around 0.96 nm using PEGs of different sizes as the test solutes [5]. In our assay the isotonic concentration determined from dextran containing control liposomes was found to be 10 mM for PEGs. The diffusion rates of PEGs were also dependent on their molecular weight. The kinetics shows that the S. typhi OmpC pore exhibits a cut-off around PEG 1450 (Fig. 2A). 3.3. Hydrodynamic radius of S. typhi OmpC The Renkin equation relates the relative mobility of spherical particles passing through a cylindrical pore to the hydrody-

Fig. 2. (A) Passage of PEGs through S. typhi OmpC was examined by monitoring decrease in absorbance at 520 nm. Solutes were used at isotonic concentration of 10 mM. The solutes used were PEG 300, 400, 600, 1000 and 1450. (B) The relative permeability of PEGs passing through the S. typhi OmpC is plotted against hydrodynamic radius (RH) of the solutes. PEG 300 was used as reference. Values are average of three determinations from two different experiments.

namic radii of the particles. Particles of size comparable to the pore are retarded, with the extent of retardation dependent on the difference in size between particle and conduit. It is thus possible to make a finer estimate of pore diameter from the relative rates of permeation of solutes than from sieving experiments that yield an exclusion limit. The relative permeability of PEGs passing through the S. typhi OmpC is plotted against hydrodynamic radius (RH) in Fig. 2B. The pore radius estimated for S. typhi OmpC is 1.1 nm, which is significantly larger than that for E. coli [4]. Since the relative permeability data for the pore size calculation uses data from one set of liposomes at a time, the density of pores would not affect the pore size estimate. Moreover, the larger pore size was qualitatively evident from the observation that S. typhi OmpC permits the passage of stachyose (Fig. 1A) in contrast to E. coli OmpC (Fig. 1B). The pore radius for the immunodominant porin from

N. Sundara Baalaji et al. / Biochimie 88 (2006) 1419–1424

the pathogenic S. typhi has not been reported earlier. However, a homology model of OmpC from S. typhi has been reported earlier [25]. The resolution of the model is not sufficient to deduce conformational changes of sidechains lining the pore that could account for the larger pore size. The multiple sequence alignment of E. coli OmpC, S. typhi OmpC and E. coli OmpF is not sufficient to understand the pore difference. A structure based sequence alignment would be more useful [26]. However, the structures of S. typhi OmpC and E. coli OmpC are not yet available [27,28]. Outer membrane porins have been shown to be more than 106 molecules per cell. They form a conspicuous presence on the membrane. Gram-negative bacteria have multiple species of OMPs with specific roles. Their synthesis is regulated by environmental conditions as in case of PhoE by phosphate availability. Thus the general diffusion porins are major determinants for the passage of nutrients, antibiotics and other antibacterial molecules into the cell. Due to rare chromosomal mutations E. coli strains express monomeric β-barrel porin OmpG of larger diameter [29,30]. Andersen et al., [31] report plasmid pRSD2 encoded porin, RafY allows E. coli to grow on the unusual substrate raffinose. This porin forms pore of large diameter allowing diffusion of high molecular-mass carbohydrates like stachyose and maltoheptaose. The estimated radius for S. typhi OmpC of 1.1 nm is larger than that reported for other general diffusion porins from E. coli and the non-pathogenic organism K. pneumoniae. Since S. typhi encounters varying osmolarity conditions during its lifecycle, the determination of larger pore size for a membrane protein channel from this human pathogenic gram-negative bacterium assumes significance in the physiological context. Further comparative structural and functional analysis of porins could possibly throw light on the suggestive relationship between pathogenicity and pore size.

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14] [15] [16]

[17]

[18]

[19]

Acknowledgements [20]

We thank Professor Jayant. B. Udgaonkar for use of facilities. NSB thanks CSIR for fellowship and Dr. Pitchumani, School of Mathematics, MKU for data analysis. The use of UGC-CPSGS, UGC-SAP and DBT-BTIS facility at SBT, MKU are acknowledged.

[21]

[22]

References [1]

[2]

[3]

[4]

[5]

R. Koebnik, K.P. Locher, P. Van Gelder, Structure and function of bacterial outer membrane proteins: barrels in a nutshell, Mol. Microbiol. 37 (2000) 239–253. M. Luckey, H. Nikaido, Specificity of diffusion channels produced by λ phage receptor protein of Escherichia coli, Proc. Natl. Acad. Sci. USA 77 (1980) 167–171. H. Nikaido, E.Y. Rosenberg, Effect of solute size on diffusion rates through the transmembrane pores of the outer membrane of Escherichia coli, J. Gen. Physiol. 77 (1981) 121–135. H. Nikaido, E.Y. Rosenberg, Porin channels in E. coli: studies with liposomes reconstituted from purified proteins, J. Bacteriol. 153 (1983) 241– 252. M. Tejuca, M. Dalla Serra, C. Alvarez, C. Potrich, G. Menestrina, Sizing the radius of the pore formed in erythrocytes and lipid vesicles by the

[23] [24]

[25]

[26]

[27]

1423

toxin sticholysin I from the sea anemone Stichodactyla helianthus, J. Membr. Biol. 183 (2001) 125–135. K.V. Srividhya, S. Krishnaswamy, A simulation model of Escherichia coli osmoregulatory switch using E-CELL system, BMC Microbiol. 4 (2004) 44 (http://www.biomedcentral.com/1471-2180/4/44). J.L. Puente, D. Juarez, M. Bobadilla, C.F. Arias, E. Calva, The Salmonella OmpC gene: structure and use as a carrier for heterologous sequences, Gene 156 (1995) 1–9. J.L. Puente, A. Verdugo-Rodriguez, E. Calva, Expression of Salmonella typhi and Escherichia coli OmpC is influenced differently by medium osmolarity; dependence on Escherichia coli OmpR, Mol. Microbiol. 5 (1991) 1205–1210. R. Germanier, E. Furer, Isolation and characterization of galE mutant Ty21a of Salmonella typhi: a candidate strain for a live, oral typhoid vaccine, J. Infect. Dis. 131 (1975) 553–558. A. Arockiasamy, S. Krishnaswamy, Purification of integral outer-membrane protein OmpC, a surface antigen from Salmonella typhi for structure–function studies: a method applicable to enterobacterial major outermembrane protein, Anal. Biochem. 283 (2000) 64–70. M. Kasahara, C.P. Hinkle, Reconstitution and purification of the D-glucose transporter from human erythrocytes, J. Biol. Chem. 252 (1977) 7384–7390. R.C. MacDonald, R.I. MacDonald, B.P.M. Menco, K. Takeshita, M.K. Subbarao, L. Hu, Small-volume extrusion apparatus for preparation of large unilamellar vesicles, Biochim. Biophys. Acta 1061 (1991) 297– 303. S.G. Schultz, A.K. Solomon, Determination of the effective hydrodynamic radii of small molecules by viscometry, J. Gen. Physiol. 44 (1961) 1189–1199. S. Kuga, Pore size distribution analysis of gel substances by size exclusion chromatography, J. Chromatog. 206 (1981) 449–461. E.M. Renkin, Filtration, diffusion, and molecular sieving through porous cellulose membranes, J. Gen. Physiol. 38 (1954) 225–243. A.V. Katta, R. Marikkannu, R.V. Basaiawmoit, S. Krishnaswamy, Consensus based validation of membrane porins, In Silico Biol. 4 (2004) 549–561. W.C. Wimley, Toward genomic identification of β-barrel membrane proteins: composition and architecture of known structures, Prot. Sci. 11 (2002) 301–312. Y. Zhai, M.H. Saier Jr., The β-barrel finder (BBF) program, allowing identification of outer membrane β-barrel proteins encoded within prokaryotic genomes, Prot. Sci. 11 (2000) 2196–2207. L.D. Mayer, M.J. Hope, P.R. Cullis, Vesicles of variable sizes produced by a rapid extrusion procedure, Biochim. Biophys. Acta 858 (1986) 161– 168. T. Kaasgaard, O.G. Mouritsen, K. Jørgensen, Freeze/thaw effects on lipid-bilayer vesicles investigated by differential scanning calorimetry, Biochim. Biophys. Acta 1615 (2003) 77–83. F. Szoka, D. Papahadjopoulos, Procedure for preparation of liposomes with large internal aqueous space and high capture by reverse-phase evaporation, Proc. Natl. Acad. Sci. USA 75 (1978) 4194–4198. F. Kienberger, V.P. Pastushenko, G. Kada, H.J. Gruber, C. Riener, H. Schindler, P. Hinterdorfer, Static and dynamical properties of single poly ethylene glycol molecules investigated by force spectroscopy, Single Mol. 2 (2000) 123–128. S.M. Bezrukov, I. Vodyanoy, V.A. Parsegian, Counting polymers moving through a single ion channel, Nature 370 (1994) 279–281. O.V. Krasilnikov, J.B. Da Cruz, L.N. Yuldasheva, W.A. Varanda, R.A. Nogueira, A novel approach to study the geometry of the water lumen of ion channels: colicin Ia channels in planar lipid bilayers, J. Membr. Biol. 161 (1998) 83–92. A. Arockiasamy, S. Krishnaswamy, Homology model of surface antigen OmpC from S. typhi and its functional implications, J. Biomol Str Dyn. 18 (2000) 261–271. T.V. Gnanasekaran, S. Peri, A. Arockiasamy, S. Krishnaswamy, Profiles from structure based sequence alignment of porins can identify beta stranded integral membrane proteins, Bioinformatics 16 (2000) 839–842. A. Arockiasamy, S. Krishnaswamy, Crystallisation of the immunologically important outer membrane protein OmpC; the first protein crystals from the human pathogen S. typhi, FEBS Lett. 453 (1999) 380–382.

1424

N. Sundara Baalaji et al. / Biochimie 88 (2006) 1419–1424

[28] N. Sundara Baalaji, R. Acharya, T.P. Singh, S. Krishnaswamy, Highresolution diffraction from crystals of a membrane protein complex: bacterial outer membrane protein, OmpC with antibacterial eukaryotic protein, lactoferrin, Acta Cryst. F 61 (2005) 773–775. [29] D.A. Fajardo, J. Cheung, C. Ito, E. Sugawara, H. Nikaido, R. Misra, Biochemistry and regulation of a novel Escherichia coli K-12 porin protein, OmpG, which produces unusually large channels, J. Bacteriol. 180 (1998) 4452–4459.

[30] S. Conlan, H. Bayley, Folding of a monomeric porin, OmpG, in detergent solution, Biochemistry 42 (2003) 9453–9465. [31] C. Andersen, D. Krones, C. Ulmke, K. Schmid, R. Benz, The porin RafY encoded by the raffinose plasmid pRSD2 of Escherichia coli forms a general diffusion pore and not a carbohydrate-specific porin, Eur. J. Biochem. 254 (1998) 679–684.