Structural characterization of the fusobacterial non-specific porin FomA suggests a 14-stranded topology, unlike the classical porins

Pål Puntervoll1, Morten Ruud1, Live J. Bruseth1, Hans Kleivdal1, Bente T. Høgh1, Roland Benza,1 and Harald B. Jensen1

Department of Molecular Biology, University of Bergen, HiB, Thormøhlensgate 55, N-5020 Bergen, Norway1

Author for correspondence: Pl Puntervoll. Tel: +47 55584500. Fax: +47 55589683. e-mail: pal.puntervoll{at}mbi.uib.no


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Native and recombinant FomA proteins were extracted by detergent from the cell envelopes of Fusobacterium nucleatum and Escherichia coli, and purified to near homogeneity by chromatography. Circular dichroism analysis revealed that the FomA protein consists predominantly of ß-sheets, in line with the previously proposed 16-stranded ß-barrel topology model. Results obtained by trypsin treatment of intact cells and cell envelopes of F. nucleatum, and from limited proteolysis of purified FomA protein, indicated that the N-terminal part of the FomA protein is not an integral part of the ß-barrel, but forms a periplasmic domain. Based on these results a new topology model is proposed for the FomA protein, where the C-terminal part forms a 14-stranded ß-barrel separate from the periplasmic N-terminal domain.

Keywords: Fusobacterium nucleatum, porins, outer-membrane proteins, circular dichroism

Abbreviations: CD, circular dichroism; FomA, fusobacterial outer-membrane protein A; OBG, octyl ß-glucopyranoside; OM, outer membrane; OMP, outer-membrane protein

a Present address: Lehrstuhl für Biotechnologie, Theodor-Boveri-Institut (Biozentrum) der Universität Würzburg, Germany.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Porins from several Gram-negative bacteria have been studied to variable extents over the last two decades. However, the majority of studies have been performed on porins from proteobacterial species. These studies have shown that porins are outer-membrane proteins (OMPs) that allow non-specific diffusion of solutes usually smaller than 650 Da (Nakae, 1976 ). The structures of several proteobacterial porins have been resolved, revealing that they exist as trimers, where each monomer forms a separate channel across the outer membrane (OM) (Cowan et al., 1992 ; Dutzler et al., 1999 ; Hirsch et al., 1997 ; Kreusch & Schulz, 1994 ; Weiss & Schulz, 1992 ; Zeth et al., 2000 ). Three of the four proteobacterial subgroups now have at least one solved porin structure, and despite substantial sequence variability, the structures show remarkably similar features. The monomer consists of 16 transmembrane antiparallel ß-strands that hydrogen bond to form a ß-barrel. The ß-strands are connected by turns or short loops at the periplasmic side, and longer loops exposed to the outside. In general, hydrophobic residues face the lipids and the trimer interface, whereas hydrophilic residues face the channel, thus allowing the channel to be water-filled. Due to the stability of the ß-barrel structure, most ß-barrel proteins, including the porins, are resistant to denaturation by SDS at lower temperatures (Heller, 1978 ; Koebnik et al., 2000 ). This property is termed heat modifiability, since the compactly folded native proteins migrate differently in SDS-PAGE compared to the denatured proteins. Also common to all porins with resolved structure is that the third surface-exposed loop actually folds back into the pore lumen, where it constricts the pore and plays a major role in determining the permeability properties of the porin (Cowan, 1993 ).

The Gram-negative anaerobe Fusobacterium nucleatum is an oral bacterium that exists as a part of the normal oral microflora (Bolstad et al., 1996 ). It has a pathogenic potential and is implicated in periodontal diseases as well as being the most common periodontal bacterium in clinical infections of other body sites (Moore & Moore, 1994 ). The major OMP of F. nucleatum, FomA, has been shown to function as a non-specific porin in lipid bilayer membranes (Kleivdal et al., 1995 ), and to function as a porin in vivo when expressed recombinantly in Escherichia coli (Kleivdal et al., 1999 ). A topology model, suggesting that the FomA protein possesses the general topology of the non-specific porins, has been proposed on the basis of structural principles derived for OMPs of E. coli (Bolstad et al., 1994 ). This model was largely confirmed by the demonstration of the surface exposure of loops L3 to L7 by epitope insertion mutagenesis (Puntervoll et al., 2000 ), together with the fact that deletions in loops L1 to L7 did not impede the pore function (Kleivdal et al., 2001 ). In addition, deletions made in loop L6 led to a marked increase in the uptake of antibiotics through the FomA porin (Kleivdal et al., 2001 ), suggesting that this loop may function analogously to the pore constriction loop L3 of the structured non-specific porins of E. coli and Rhodobacter capsulatus (Cowan, 1993 ). Furthermore, a cluster of arginines that contributes to the constriction of the pore has been identified, similar to the cluster observed in the matrix porin OmpF of E. coli, but positioned closer to the C terminus (Kleivdal et al., 1999 ). We recently proposed that these structural deviations from the well-known porins may imply that FomA belongs to a separate subgroup of non-specific porins, reflecting the evolutionary distance between the fusobacteria and the proteobacteria (Kleivdal et al., 2001 ), but further structural information is clearly needed to establish whether this actually is the case. In the present study, we present the first physicochemical evidence that FomA is a ß-barrel, and examine the role of the proposed loop L1. The results presented here indicate that the proposed loop L1 is located in the periplasm, and that the FomA protein is a 14-stranded ß-barrel.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains and growth conditions.
Fusobacterium nucleatum ATCC 10953 was used as a source for wild-type FomA, and was grown anaerobically at 37 °C in a liquid Bacto-tryptone/yeast extract/ascorbic acid/glucose/0·5% NaCl medium (Bolstad et al., 1991 ).

Escherichia coli strain DH5{alpha} (Hanahan, 1983 ) was used for cloning purposes. The porin-deficient E. coli K-12 strain CE1224 (Tommassen et al., 1983 ) was used to express recombinant FomA protein as previously described (Puntervoll et al., 2000 ). The E. coli strains AK101 and DH5{alpha} mutS were used with the double-primer method for site-directed mutagenesis (Ohmori, 1994 ). E. coli was grown aerobically at 37 °C in LB medium, or, to impose phosphate limitation, in a low-phosphate medium described by Levinthal et al. (1962) . For large-scale growth under low-phosphate conditions, E. coli was grown in 200 ml LB medium overnight; the cells were harvested and washed twice in a low-phosphate medium and used to inoculate a 1 l low-phosphate medium culture, which was incubated overnight. The antibiotics chloramphenicol (35 µg ml-1), ampicillin (50 µg ml-1) and tetracycline (10 µg ml-1) were added for selective purposes when necessary.

Plasmids and constructs.
The plasmid pHB14 (Jensen et al., 1996 ) contains the part of the fomA gene encoding the mature part of the protein, immediately downstream of the phoE promoter and signal sequence encoding part of phoE, and was used to express the FomA protein recombinantly in E. coli in a functional manner (Jensen et al., 1996 ). Plasmid pHB14E2C was constructed by site-directed mutagenesis using the double-primer method (Ohmori, 1994 ). The pTZ19Urrh-derived phagemid pHKUrrh containing the chimeric phoE/fomA gene from pHB14 and a defective ori region (Kleivdal et al., 1999 ) was used to produce single-stranded DNA (Sambrook et al., 1989 ). The E2C (5'-GGTGCAGGCATAACACATGCAGCCTGTACAGATGC-3') and ori repair (5'-GGGAAGCGTGGCGCTTTCTCATAG-3') primers were used, and the site-directed mutagenesis was performed as described by Kleivdal et al. (1999) . The pHB14E2C plasmid sequence was confirmed by using the ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction Kit and the ABI PRISM 377 DNA Sequencer (Applied Biosystems).

Isolation of FomA from the outer membrane by detergent extraction.
The procedure described below was used to isolate FomA protein from the outer membrane of both F. nucleatum and E. coli. Bacterial cells from 1 l of the desired culture were harvested in the late exponential phase of growth by centrifugation, washed twice with 100 ml 50 mM Tris/HCl (pH 7·7), resuspended in 25 ml of the same buffer, and disrupted by passage through a French pressure cell three times at 900 p.s.i. (6·2 MPa). The samples were collected in pre-cooled tubes, and intact cells were removed by a low-speed centrifugation at 1000 g for 20 min at 4 °C. All succeeding centrifugations were performed at 20000 g for 45 min. The cell envelopes were isolated by centrifugation at 4 °C. The cell envelopes of F. nucleatum were resuspended in 20 ml 2% Triton X-100, 10 mM MgCl2, 10 mM HEPES (pH 7·4), sonicated for 30 min in a water bath sonicator, incubated for 45 min at room temperature, and subjected to centrifugation at 4 °C. The cell envelopes of E. coli were resuspended in 20 ml 0·05% CHAPS, 50 mM Tris/HCl, 20 mM EDTA (pH 8·0), incubated at room temperature for 10 min on a rotating wheel, and subjected to centrifugation at 14 °C. The supernatant was discarded, and the pellet was resuspended in 10 ml 1% octyl ß-glucopyranoside (OBG), 50 mM Tris/HCl, 20 mM EDTA (pH 8·0), and incubated with vigorous shaking (250 r.p.m.) at 37 °C for 30 min. The residual cell envelope material was pelleted by centrifugation at 24 °C, and the supernatant containing the extracted proteins was collected and stored at 4 °C.

Purification of FomA by chromatography.
Anion-exchange chromatography was used as the first step in the purification of FomA. A 20 ml Q-Sepharose column (1·5 cm diameter) was equilibrated with 20 mM Tris/HCl buffer (pH 8·0), and the flow-through of protein was monitored by measuring A280. When the applied sample contained a detergent other than OBG the column was washed with 20 ml 1 % OBG, 10 mM Tris/HCl (pH 8·0) to exchange the detergent. The protein was eluted from the Q-Sepharose column by using a LiCl gradient. A total volume of 50 ml was used, and the gradient ranged from 0 to 300 mM LiCl with fixed concentrations of 1% OBG and 10 mM Tris/HCl (pH 8·0). Fractions of 1·5 ml were collected, A280 was measured, and fractions of interest were analysed by SDS-PAGE.

Further purification of the FomA protein was achieved by applying size-exclusion chromatography. Fractions enriched in FomA were pooled, and the volume was reduced to less than 2 ml by using a concentration cell (Amicon) with a cut-off filter of 30 kDa. This sample was applied to a Sephacryl 200 column pre-equilibrated with 1% OBG, 10 mM Tris/HCl (pH 8·0), 200 mM LiCl, and size-exclusion chromatography was conducted in the same buffer at a constant flow of 0·5 ml min-1. Fractions of 1·0 ml were collected, A280 was measured, and fractions of interest were analysed by SDS-PAGE.

SDS-PAGE and Western blotting.
SDS-PAGE was performed as described by Lugtenberg et al. (1975) , and the samples were incubated in sample buffer at room temperature or 95 °C for 5 min prior to electrophoresis. The SDS-PA gels were either stained with Coomassie brilliant blue R250 (Sigma) or subjected to Western blot analysis as described by Towbin et al. (1979) . Anti-FomA antiserum {phi}239 (Bakken et al., 1989b ) and horseradish-peroxidase-coupled goat anti-rabbit antibody (Bio-Rad) were used as primary and secondary antibodies, respectively, and the immunoblots were developed with 4-chloro-1-naphthol (Bio-Rad).

Analysis of secondary structure content by circular dichroism.
Secondary structure content of the FomA protein was evaluated by circular dichroism (CD) spectroscopy using a Jasco J-810 spectropolarimeter equipped with a Jasco PTC-423S Peltier element for temperature control. Prior to analysis, the samples were dialysed against a 1% OBG, 20 mM Tris/HCl (pH 8·0) buffer to reduce the LiCl concentration to less than 5 mM. The CD spectra were recorded at 25 °C with samples containing 0·1–0·3 mg protein ml-1 by four consecutive scans at 0·5 nm wavelength intervals, and the mean spectra were stored. The temperature-dependent stability of secondary structure was monitored by heating the samples at a programmed rate of 1 °C min-1.

Trypsin treatment of intact cells and cell envelopes containing FomA, and of purified FomA.
Trypsin treatment of intact cells and cell envelopes was in essence performed as described by Merck et al. (1997) . In short, cells from an overnight culture were pelleted and resuspended in a buffer containing 0·1 M Tris/HCl (pH 8·0), 0·25 M sucrose and 10 mM MgCl2, and the samples were incubated on ice for 60 min with 0·1 mg trypsin ml-1. The trypsin reaction was stopped by adding soybean trypsin inhibitor (Sigma) to a final concentration of 500 µg ml-1, and the samples were subsequently washed twice with the above-described buffer containing trypsin inhibitor. Cell envelopes to be trypsin treated were isolated by passage through a French pressure cell and subsequent centrifugation as described by Jensen et al. (1996) , and subjected to trypsin digestion as described for intact cells. Purified FomA protein was subjected to trypsin digestion by adding trypsin to final concentrations of 10, 1 or 0·1 µg ml-1 and incubating on ice for 60 min; the reaction was stopped as described above. All trypsin-treated samples were immediately analysed by SDS-PAGE and Western blotting to avoid unwanted potential activity of (residual) trypsin.

N-terminal amino acid sequence determination.
Samples containing the FomA protein were subjected to SDS-PAGE, and the Coomassie-stained FomA bands were excised. The protein was extracted from the gel by incubation in 70% formic acid at room temperature overnight (Tsugita et al., 1987 ), and submitted to the protein sequencing facilities at the University of Oslo for N-terminal sequencing.

Single-channel conductance experiments of FomA reconstituted in lipid bilayer membranes.
FomA protein was prepared for single-channel conductance experiments by extracting the protein from OMs with 1% OBG, followed by SDS-PAGE. Samples were incubated with sample buffer at room temperature prior to SDS-PAGE, and run in parallel in all lanes of a 15-lane SDS-PA gel. The outer lanes were stained with Coomassie blue to guide the excision of non-stained native FomA proteins from the gel, and the protein was eluted with 0·5% OBG in 20 mM Tris/HCl (pH 8·0). The reconstitution of pore-forming activity in artificial membranes was performed by using black lipid membranes as described previously (Kleivdal et al., 1995 ).


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Purification of FomA from the cell envelopes of F. nucleatum and E. coli
For structural analysis, the FomA protein was purified to near homogeneity from both cell envelopes of F. nucleatum ATCC 10953 (FomA10953) and the OM of E. coli CE1224 cells expressing FomA (rFomA10953). Purification was performed by ion-exchange and size-exclusion chromatography as described in Methods. In addition, a truncated 37 kDa FomA protein (FomA10953TRU) was purified in the same manner, after treating OM fractions of F. nucleatum ATCC 10953 with trypsin as described in Methods. N-terminal sequencing yielded mixed signals and demonstrated that the FomA10953TRU protein had either D27 or E29 as a new N-terminus, giving a theoretical molecular mass of approximately 37 kDa. All the purified FomA proteins were heat-modifiable, since samples prepared at room temperature prior to SDS-PAGE migrated with a higher mobility than boiled samples (Fig. 1). This strongly suggests that the non-boiled purified FomA proteins, including the truncated FomA10953TRU, migrate as folded monomers in SDS-PAGE. These observations are in line with previous results obtained by our group (Bakken et al., 1989a ; Jensen et al., 1996 ).



View larger version (62K):
[in this window]
[in a new window]
 
Fig. 1. SDS-PAGE analysis of the purified FomA proteins used in this study. FomA10953, FomA purified from the cell envelope of F. nucleatum ATCC 10953; FomA10953TRU, a truncated 37 kDa FomA protein purified from trypsin-treated cell envelopes of F. nucleatum ATCC 10953. rFomA10953, FomA purified from the cell envelope of E. coli CE1224. Samples were either boiled for 5 min (+) or incubated at room temperature (-) prior to SDS-PAGE.

 
Secondary structure content analysis by circular dichroism
The CD analysis was performed on the FomA10953, FomA10953TRU and rFomA10953 proteins, as described in Methods. The CD spectra of the FomA proteins (Fig. 2, left panels), obtained at 25 °C (full lines), suggest that the FomA protein is composed predominantly of ß-pleated sheet structures, as the topology model suggests. Furthermore, the similarity between the spectra of FomA10953 and FomA10953TRU strongly indicates that the truncated FomA10953TRU protein possesses essentially the same secondary structure content as the native FomA10953 protein. The spectrum obtained with the rFomA10953 protein differed from the other two spectra (Fig. 2), suggesting that the solubilized rFomA10953 protein has a somewhat different secondary structure content. The FomA proteins were subjected to a thermal denaturation analysis by monitoring at 214 nm the disruption of secondary structure (Fig. 2, right panels). The thermal melting profiles of FomA10953 and rFomA10953 show cooperative unfolding transitions at 92 °C. However, FomA10953TRU does not seem to unfold cooperatively. The rFomA10953 protein shows an additional transition step at 68 °C, possibly reflecting two (or more) different conformational populations. CD spectra were also recorded for the samples after heating to 100 °C and immediate subsequent cooling to 25 °C (Fig. 2, left panels, dotted lines). The CD spectra of the heat-treated FomA10953 and rFomA10953 proteins suggest partial refolding but also significant loss of material, possibly due to aggregation. Unlike the FomA10953 and rFomA10953 proteins, the truncated FomA10953TRU protein does not seem to aggregate upon heating, thus displaying similar secondary structure content after cooling to 25 °C.



View larger version (37K):
[in this window]
[in a new window]
 
Fig. 2. Evaluation of secondary structure content of the FomA protein by CD analysis. CD analysis of the FomA10953, FomA10953TRU and rFomA10953 proteins was done in a 20 mM Tris/HCl buffer (pH 8·0) containing 1% OBG. The CD spectra were recorded at 25 °C (left-hand panels, full lines), and the lowest mean residue ellipticity ({theta}) value was determined (214 nm). Subsequently, thermal denaturation of the protein was monitored by increasing the temperature from 25 °C to 100 °C at a rate of 1 °C min-1, using 214 nm as fixed wavelength (right-hand panels). After heating the sample at 100 °C, the temperature was immediately lowered to 25 °C, and another CD spectrum was recorded (left-hand panels, dotted lines).

 
Assessment of the topology model of FomA by trypsin treatment of intact cells and cell envelopes, and purified FomA proteins
The FomA protein has been shown to be highly resistant to trypsin digestion when situated in the outer membrane of both F. nucleatum and E. coli (Bakken et al., 1989a ; Jensen et al., 1996 ), despite having several R and K residues distributed evenly in the entire primary structure. It was however reported that a small N-terminal segment was cleaved off when intact cells or cell envelopes of F. nucleatum were treated with trypsin, and that E29 became the new N-terminal amino acid, indicating surface exposure of the postulated loop L1 (Bakken et al., 1989a ). However, when these experiments were repeated using fresh cells and the conditions described in Methods, the FomA protein appeared to be completely resistant to trypsin (Fig. 3a, lanes 1–4), in the same manner as when present in the OM of intact E. coli cells (Jensen et al., 1996 ). In contrast, when cell envelopes were treated with trypsin, a small fragment was cleaved off (Fig. 3a, compare lanes 5 and 8), indicating that the residues R26 and R28 were accessible to trypsin and a periplasmic location of these residues. If the N-terminus of the FomA protein was an integral part of the ß-barrel of the FomA protein, as the proposed topology model suggests (Bolstad et al., 1994 ), one would expect that at least the ß-strand part of the N-terminus would remain as part of the folded ß-barrel, when migrating as a folded monomer in SDS-PAGE. This does not seem to be the case, since the folded monomer migrates faster after trypsin digestion than the undigested one (Fig. 3a, compare lanes 6 and 7). Furthermore, N-terminal sequencing of the truncated folded monomer FomA10953TRU isolated by excision from an SDS-PA gel (corresponding to the one shown in Fig. 1, lane 3), revealed that the original N-terminus was not present, and that the new N-terminus started from either D27 or E29. Since the trimers of the FomA10953 protein cannot be observed in SDS-PAGE, we also performed this analysis on the FomA protein from F. nucleatum T18 (FomAT18) (not shown), which displays stable trimers (apparent molecular mass ~70 kDa) in SDS-PAGE (Haake & Wang, 1997 ). The FomAT18 trimer was not disrupted by the trypsin treatment, since the migration in SDS-PAGE only increased slightly compared to full-length trimers, and the truncated FomAT18 protein had new N-termini identical to those observed with FomA10953TRU.



View larger version (54K):
[in this window]
[in a new window]
 
Fig. 3. Trypsin accessibility of FomA protein in intact cells, in cell walls and in solution. (a) Western blot analysis of FomA in whole cells and cell walls. Samples were treated with trypsin as described in Methods where indicated, and either boiled for 5 min (+) or incubated at room temperature (-) prior to SDS-PAGE. (b) Western blot analysis of FomAFev1 protein isolated from the OM of F. nucleatum Fev1 (lower blot), and rFomA10953 isolated from the OM of E. coli CE1224 (upper blot), following treatment with trypsin. Three trypsin concentrations were used (0·1, 1 and 10 µg ml-1), and the experiment was performed as described in Methods. All samples were applied to SDS-PA gels in duplicate: one was boiled (+) in sample buffer, whereas the other was incubated at room temperature (-). Molecular mass standards are given in kDa to the left.

 
Several porins are completely resistant to tryptic hydrolysis also when solubilized (Minetti et al., 1998 ; Rosenbusch, 1974 ). Since the 37 kDa C-terminal fragment of the FomA protein was completely trypsin resistant when situated in the OM (Bakken et al., 1989a ; Jensen et al., 1996 ), trypsin digestion experiments were applied to the purified FomA proteins to assess the tertiary structure. Since the FomA protein from F. nucleatum ATCC 10953 only displays weak trimeric bands on SDS-PAGE (Fig. 1, lane 1), the FomA protein from F. nucleatum Fev1 (FomAFev1), which has stable trimers in SDS-PAGE (Kleivdal et al., 1995 ), was purified and included in this analysis. The initial analysis using 10 µg trypsin ml-1 revealed extensive digestion, indicating that several trypsin sites, in addition to R26/R28, were susceptible in the FomA proteins in solution. Fig. 3(b) shows a more detailed analysis that was performed with the rFomA10953 and FomAFev1 proteins. Three different trypsin concentrations were used, 0·1, 1 and 10 µg ml-1, and the samples were applied to an SDS-PA gel in duplicate; one sample was incubated at room temperature and the other was boiled prior to electrophoresis. It is evident that both populations of OM-extracted FomA protein were much more susceptible to trypsin when in solution than when present in the membrane, since the protein is degraded to fragments smaller than the 37 kDa OM-protected fragment (Fig. 3b, lanes 6–8). Furthermore, the unheated samples show that trypsin digestion apparently leads neither to a dissociation of the trimers of the FomAFev1 protein, nor to the disintegration of the ß-barrel structure of the folded monomers (Fig. 3b, lanes 3–5). This is in line with the very stable structure revealed by the CD experiments. Finally, trypsin digestion leads to a shift in mobility of both trimers and folded monomer (Fig. 3b, compare lanes 2 and 3), suggesting that the N-terminus is not an integral part of the ß-barrel.

In order to pursue the hypothesis that the N-terminus is not part of the ß-barrel, the 37 kDa truncated FomA10953TRU was incorporated into artificial bilayer membranes, as described in Methods. Fig. 4 shows that FomA10953TRU indeed functions as a pore, and a mean single-channel conductance of 1·22 nS was found, similar to that observed with full-length FomA10953 (1·17 nS) (Kleivdal et al., 1995 ).



View larger version (7K):
[in this window]
[in a new window]
 
Fig. 4. Analysis of single-channel events due to incorporation of FomA10953TRU into artificial bilayer membranes.

 
Role of the N-terminus of FomA in trimer stability
The strong sequence conservation of the N-terminus among the different FomA proteins suggests an important structural and/or functional role (Bolstad et al., 1994 ). In a parallel work, an rFomA10953 mutant carry ing an E2C substitution was constructed (named rFomA10953E2C), in an attempt to identify the trimeric interface of the FomA protein (L. J. Bruseth, unpublished). This E2C modification was found to have a trimer-stabilizing effect when the protein was recombinantly expressed in E. coli, leading to trimers in a non-reducing SDS-PAGE analysis (Fig. 5, e.g. lane 1), unlike the rFomA10953 protein (e.g. Fig. 1, lane 5). The observed trimer-stabilizing effect is the result of the cysteine of one monomer of the rFomA10953E2C protein forming a disulphide bond with the cysteine of one of the neighbouring monomers within the trimer, since no trimers were visible after reduction of the disulphide bond (not shown). This disulphide bond apparently stabilizes the entire trimeric structure, implying the direct involvement of the N-terminus in trimer stability. In order to analyse this E2C mutant more closely, the rFomA10953E2C protein was isolated from the OM of E. coli CE1224 cells as described in Methods, and solubilized rFomA10953E2C was subjected to a denaturation analysis (Fig. 5). This analysis shows that the rFomA10953E2C protein is highly resistant to denaturation by temperature, and that, in the absence of a reducing agent, the protein denatures into a monomer and a covalently linked dimer, when heated at 100 °C for 20 min prior to SDS-PAGE. The effects of exposure to various temperatures were followed both in 0·5% OBG, 10 mM Tris/HCl (pH 8·0) solution (Fig. 5, lanes 6–9) and in the same solution also containing 4 M urea (Fig. 5, lanes 1–5). The results clearly show that when the rFomA10953E2C protein was heated to 90 °C in the presence of urea, a portion of the protein denatured completely into the 40 kDa monomer and the 80 kDa dimer. But, in the absence of urea, several additional protein bands can be observed on the SDS-PA gel, including bands corresponding to folded monomers (38 kDa) and folded dimers (54 kDa, 62 kDa). This demonstrates that the trimers of the FomA protein can disintegrate into folded monomers (and in the case of the E2C mutant, folded dimers), and that the monomers and dimers are less stable than the intact trimer in vitro (Fig. 5, compare lanes 4 and 9).



View larger version (58K):
[in this window]
[in a new window]
 
Fig. 5. Western blot analysis of the rFomA10953E2C protein: trimer dissociation and denaturation in vitro. The rFomA10953E2C protein extracted from the OM of E. coli CE1224 was incubated at different temperatures in the OBG buffer with and without urea for 20 min, as indicated. The samples were subsequently incubated in a non-reducing sample buffer at room temperature for 5 min prior to SDS-PAGE. The analysis allowed the observation of intact trimers (T) (lanes 1–4 and 6–9), folded dimers (D) and monomers (M) (lanes 8 and 9), denatured dimers (D*) and monomers (M*) (lanes 4, 5 and 9). Molecular mass standards are given in kDa to the left.

 

   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The CD spectra strongly suggest that the FomA protein is rich in ß-sheet structure and are similar to those obtained with other porins from E. coli (Rosenbusch, 1974 ) and from Neisseria meningitidis (Minetti et al., 1997 , 1998 ). Furthermore, the thermal melting profiles suggest that the FomA protein possesses a ß-barrel architecture typical for OMPs (Schulz, 2000 ), since the cooperative unfolding transition occurs at high temperatures (>80 °C), reflecting the very stable three-dimensional structures of these proteins. The evidence for a ß-barrel architecture is in line with the published topology model for the FomA protein suggesting that the FomA protein is a 16-stranded ß-barrel (Bolstad et al., 1994 , 1995 ). The apparent lower content of secondary structure observed with the rFomA10953 protein in combination with its melting profile suggests that at least a fraction of this protein has a different conformation in solution than the FomA10953 protein. This may indicate either that a population of the rFomA10953 protein expressed in E. coli CE1224 does not attain the native conformation, or that a change in conformation is induced as a consequence of the detergent extraction. Since the rFomA10953 protein in the OM possesses other characteristics indicating proper folding, such as heat modifiability and trypsin resistance, and the isolated protein has a mean single-channel conductance very similar to that of the FomA10953 protein (Jensen et al., 1996 ), the latter explanation is favoured. Trypsin analysis of the purified FomA proteins strongly suggests that both the quaternary and tertiary structures are intact in solution, in spite of increased sensitivity to trypsin compared to the FomA proteins present in the OM, since neither folded monomers nor trimers are (completely) disrupted by proteolysis (Fig. 3b, lanes 3–5).

The trypsin analysis of the FomA protein present in the OM and in solution also provided new structural information. In contrast to earlier published work (Bakken et al., 1989a ), no R or K residues seem to be accessible to trypsin on the surface of intact cells. Instead, the results presented here suggest that the trypsin site reported to be exposed on the cell surface is in fact exposed to the periplasm, since it is fully accessible to trypsin only when cell envelopes are treated with trypsin. Thus residues 1–28 do not seem to be part of the ß-barrel, since the N-terminal fragment containing these residues apparently dissociates from the rest of the protein, as evident from the increased mobility in SDS-PAGE and the fact that none of these residues can be found by N-terminal sequencing. Furthermore, the first putative membrane-spanning segment of the topology model of the FomA protein contains several alternating proline residues, and hence this segment cannot exist as a classical ß-strand (Bolstad et al., 1994 ). The 28 aa N-terminal fragment also seems to dissociate from the FomA protein when subjected to trypsin treatment in solution, since there is a shift in the mobility of both the folded trimers of FomAFev1 and the folded monomers of rFomA10953 in SDS-PAGE (Fig. 3b, compare lanes 2 and 4). Finally, the fact that several additional trypsin sites were cleaved in solubilized FomA protein (Fig. 3b, lanes 6 and 7), with no apparent effect on the migration of the folded barrel (Fig. 3b, lanes 3 and 4), together with the fact that FomA10953TRU functions as a pore (Fig. 4) with properties similar to those of the FomA10953 protein, strongly supports the hypothesis that the N-terminus is not part of the ß-barrel.

Earlier work has demonstrated that it is possible to replace residues 8 to 23 with the residues RS without the loss of pore function (Kleivdal et al., 2001 ), suggesting that this segment is not critical for maintaining the ß-barrel structure of the FomA protein. In support of this is also the fact that the 37 kDa trypsin fragment of the FomA (FomA10953TRU) protein functioned as an intact pore in artificial bilayer membranes (Fig. 4), with a mean single-channel conductance similar to that of the full-length FomA10953 protein. On the other hand, the results obtained with the E2C FomA mutant protein suggest that the N-terminus plays a role in stabilizing the quaternary structure, since the rFomA10953E2C protein displays more stable trimers than the wild-type FomA10953 protein under non-reducing conditions, due to a disulphide bond between two monomers within the trimer.

In the light of the results presented here, we propose an alternative topology model where the C-terminal part of the protein forms a 14-stranded ß-barrel (Fig. 6). Although the results do not directly implicate the second putative ß-strand of the original model, the constriction rules for transmembrane ß-barrels formulated by Schultz (2000) imply an even number of ß-strands, thus leading us to propose that the first two strands and loop L1 of the original model are located in the periplasm. The 14-stranded ß-barrel topology model not only conforms to the results presented here, but is also in accordance with previously published results for the FomA porin (Kleivdal et al., 1999 , 2001 ; Puntervoll et al., 2000 ). The insertion of a foreign epitope after E19 did not result in the surface exposure of this epitope, in contrast to similar insertions in loops L2 to L6 (named according to the current topology model – Fig. 6) (Puntervoll et al., 2000 ), in line with a periplasmic location of this part of the protein. Furthermore, neither the insertion of a foreign epitope after E19 (Puntervoll et al., 2000 ), nor the deletion of residues 8–23 (Kleivdal et al., 2001 ), severely affected the function of the FomA porin, in line with the N-terminal part not being an integral part of the ß-barrel. Recently, the projection structure of the E. coli OmpG porin was published at 6 (0·6 nm) resolution (Behlau et al., 2001 ), and it was suggested that this porin is a 14-stranded ß-barrel. Compared to the classical non-specific proteobacterial porins, the OmpG porin is also atypical in that it does not seem to form trimers, in contrast to the FomA porin (Kleivdal et al., 1995 ). In conclusion, the FomA porin, along with the OmpG porin from E. coli, seems to deviate from the common understanding that non-specific porins are 16-stranded ß-barrels. In order to assess the accuracy of this novel 14-stranded topology model, obtaining the crystal structure of FomA would be of utmost value, and this is our current goal. The purification procedures presented here form a solid base for obtaining large amounts of pure protein for crystallization.



View larger version (51K):
[in this window]
[in a new window]
 
Fig. 6. The 14-stranded topology model of the FomA protein of F. nucleatum ATCC 10953. The top of the model shows the putative surface-exposed loops and the central part represents the presumed transmembrane segments. Amino acid residues, shown in one-letter code, are indicated by boxes when they are supposed to form ß-strands (shaded box if proposed to face the lipids or the subunit interface) and circles for turns (T), loops (L) and the putative periplasmic N-terminal domain. Amino acid residues referred to in the text are numbered, and the arrows indicate the sites that were accessible to trypsin when cell envelopes containing FomA were treated. The GenBank accession number for the FomA10953 sequence is X72583.

 

   ACKNOWLEDGEMENTS
 
This work was supported by the Norwegian Research Council. We are grateful to M. Thorolfsson and Dr A. Martinez, University of Bergen, and Dr H. H. Hauge, University of Oslo, for competent assistance with the circular dichroism experiments. We also appreciate the assistance from Dr V. Bakken, University of Bergen, in cultivating the F. nucleatum T18 cells. Finally, we acknowledge Nina Glomnes for excellent technical assistance.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bakken, V., Aaro, S. & Jensen, H. B. (1989a). Purification and partial characterization of a major outer-membrane protein of Fusobacterium nucleatum. J Gen Microbiol 135, 3253-3262.[Medline]

Bakken, V., Aaro, S., Hofstad, T. & Vasstrand, E. N. (1989b). Outer membrane proteins as major antigens of Fusobacterium nucleatum. FEMS Microbiol Immunol 1, 473-483.[Medline]

Behlau, M., Mills, D. J., Quader, H., Kuhlbrandt, W. & Vonck, J. (2001). Projection structure of the monomeric porin OmpG at 6 resolution. J Mol Biol 305, 71-77.[Medline]

Bolstad, A. I., Skaug, N. & Jensen, H. B. (1991). Use of synthetic oligonucleotide DNA probes for the identification of different strains of Fusobacterium nucleatum. J Periodontal Res 26, 519-526.[Medline]

Bolstad, A. I., Tommassen, J. & Jensen, H. B. (1994). Sequence variability of the 40-kDa outer membrane proteins of Fusobacterium nucleatum strains and a model for the topology of the proteins. Mol Gen Genet 244, 104-110.[Medline]

Bolstad, A. I., Høgh, B. T. & Jensen, H. B. (1995). Molecular characterization of a 40-kDa outer membrane protein, FomA, of Fusobacterium periodonticum and comparison with Fusobacterium nucleatum. Oral Microbiol Immunol 10, 257-264.[Medline]

Bolstad, A. I., Jensen, H. B. & Bakken, V. (1996). Taxonomy, biology, and periodontal aspects of Fusobacterium nucleatum. Clin Microbiol Rev 9, 55-71.[Abstract]

Cowan, S. W. (1993). Bacterial porins: lessons from three high-resolution structures. Curr Opin Struct Biol 3, 501-507.

Cowan, S. W., Schirmer, T., Rummel, G., Steiert, M., Ghosh, R., Pauptit, R. A., Jansonius, J. N. & Rosenbusch, J. P. (1992). Crystal structures explain functional properties of two E. coli porins. Nature 358, 727-733.[Medline]

Dutzler, R., Rummel, G., Alberti, S., Hernandez-Alles, S., Phale, P., Rosenbusch, J., Benedi, V. & Schirmer, T. (1999). Crystal structure and functional characterization of OmpK36, the osmoporin of Klebsiella pneumoniae. Struct Fold Des 7, 425-434.[Medline]

Haake, S. K. & Wang, X. (1997). Cloning and expression of FomA, the major outer-membrane protein gene from Fusobacterium nucleatum T18. Arch Oral Biol 42, 19-24.[Medline]

Hanahan, D. (1983). Studies on transformation of Escherichia coli with plasmids. J Mol Biol 166, 557-580.[Medline]

Heller, K. B. (1978). Apparent molecular weights of a heat-modifiable protein from the outer membrane of Escherichia coli in gels with different acrylamide concentrations. J Bacteriol 134, 1181-1183.[Medline]

Hirsch, A., Breed, J., Saxena, K., Richter, O. M., Ludwig, B., Diederichs, K. & Welte, W. (1997). The structure of porin from Paracoccus denitrificans at 3·1 resolution. FEBS Lett 404, 208-210.[Medline]

Jensen, H. B., Skeidsvoll, J., Fjellbirkeland, A., Høgh, B., Puntervoll, P., Kleivdal, H. & Tommassen, J. (1996). Cloning of the fomA gene, encoding the major outer membrane porin of Fusobacterium nucleatum ATCC10953. Microb Pathog 21, 331-342.[Medline]

Kleivdal, H., Benz, R. & Jensen, H. B. (1995). The Fusobacterium nucleatum major outer-membrane protein (FomA) forms trimeric, water-filled channels in lipid bilayer membranes. Eur J Biochem 233, 310-316.[Abstract]

Kleivdal, H., Benz, R., Tommassen, J. & Jensen, H. B. (1999). Identification of positively charged residues of FomA porin of Fusobacterium nucleatum which are important for pore function. Eur J Biochem 260, 818-824.[Abstract/Free Full Text]

Kleivdal, H., Puntervoll, P. & Jensen, H. B. (2001). Topological investigations of the FomA porin from Fusobacterium nucleatum and identification of the constriction loop L6. Microbiology 147, 1059-1067.[Abstract/Free Full Text]

Koebnik, R., Locher, K. P. & Van Gelder, P. (2000). Structure and function of bacterial outer membrane proteins: barrels in a nutshell. Mol Microbiol 37, 239-253.[Medline]

Kreusch, A. & Schulz, G. E. (1994). Refined structure of the porin from Rhodopseudomonas blastica. Comparison with the porin from Rhodobacter capsulatus. J Mol Biol 243, 891-905.[Medline]

Levinthal, C., Signer, E. R. & Fetherolf, K. (1962). Reactivation and hybridization of reduced alkaline phosphatase. Proc Natl Acad Sci USA 48, 1230-1237.[Medline]

Lugtenberg, B., Meijers, J., Peters, R., van der Hoek, P. & van Alphen, L. (1975). Electrophoretic resolution of the ‘‘major outer membrane protein’’ of Escherichia coli K12 into four bands. FEBS Lett 58, 254-258.[Medline]

Merck, K. B., de Cock, H., Verheij, H. M. & Tommassen, J. (1997). Topology of the outer membrane phospholipase A of Salmonella typhimurium. J Bacteriol 179, 3443-3450.[Abstract]

Minetti, C., Tai, J. Y., Blake, M. S., Pullen, J. K., Liang, S. M. & Remeta, D. P. (1997). Structural and functional characterization of a recombinant PorB class 2 protein from Neisseria meningitidis. Conformational stability and porin activity. J Biol Chem 272, 10710-10720.[Abstract/Free Full Text]

Minetti, C. A., Blake, M. S. & Remeta, D. P. (1998). Characterization of the structure, function, and conformational stability of PorB class 3 protein from Neisseria meningitidis. A porin with unusual physicochemical properties. J Biol Chem 273, 25329-25338.[Abstract/Free Full Text]

Moore, W. E. & Moore, L. V. (1994). The bacteria of periodontal diseases. Periodontol 2000 5, 66-77.

Nakae, T. (1976). Outer membrane of Salmonella. Isolation of protein complex that produces transmembrane channels. J Biol Chem 251, 2176-2178.[Abstract]

Ohmori, H. (1994). A new method for strand discrimination in sequence-directed mutagenesis. Nucleic Acids Res 22, 884-885.[Medline]

Puntervoll, P., Kleivdal, H., Dahl, K. O., Bitter, W., Tommassen, J. & Jensen, H. B. (2000). The Fusobacterium nucleatum porin FomA possesses the general topology of the non-specific porins. Microbiology 146, 1437-1445.[Abstract/Free Full Text]

Rosenbusch, J. P. (1974). Characterization of the major envelope protein from Escherichia coli. J Biol Chem 248, 8019-8029.

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Schulz, G. E. (2000). Beta-barrel membrane proteins. Curr Opin Struct Biol 10, 443-447.[Medline]

Tommassen, J., van Tol, H. & Lugtenberg, B. (1983). The ultimate localization of an outer membrane protein of Escherichia coli K-12 is not determined by the signal sequence. EMBO J 2, 1275-1279.[Medline]

Towbin, H., Staehelin, T. & Gordon, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 76, 4350-4354.[Abstract]

Tsugita, A., Ataka, T. & Uchida, T. (1987). Approaches for submicrosequencing. J Protein Chem 6, 121-130.

Weiss, M. S. & Schulz, G. E. (1992). Structure of porin refined at 1·8 resolution. J Mol Biol 227, 493-509.[Medline]

Zeth, K., Diederichs, K., Welte, W. & Engelhardt, H. (2000). Crystal structure of Omp32, the anion-selective porin from Comamonas acidovorans, in complex with a periplasmic peptide at 2·1 resolution. Structure Fold Des 8, 981-992.[Medline]

Received 16 July 2002; accepted 18 July 2002.