Molecular Cloning, Sequencing, and Expression of omp-40, the Gene Coding for the Major Outer Membrane Protein from the Acidophilic Bacterium Thiobacillus ferrooxidans

Abstract

Thiobacillus ferrooxidans is one of the chemolithoautotrophic bacteria important in industrial biomining operations. Some of the surface components of this microorganism are probably involved in adaptation to their acidic environment and in bacterium-mineral interactions. We have isolated and characterized omp40, the gene coding for the major outer membrane protein from T. ferrooxidans. The deduced amino acid sequence of the Omp40 protein has 382 amino acids and a calculated molecular weight of 40,095.7. Omp40 forms an oligomeric structure of about 120 kDa that dissociates into the monomer (40 kDa) by heating in the presence of sodium dodecyl sulfate. The degree of identity of Omp40 amino acid sequence to porins from enterobacteria was only 22%. Nevertheless, multiple alignments of this sequence with those from several OmpC porins showed several important features conserved in the T. ferrooxidans surface protein, such as the approximate locations of 16 transmembrane beta strands, eight loops, including a large external L3 loop, and eight turns which allowed us to propose a putative 16-stranded beta-barrel porin structure for the protein. These results together with the previously known capacity of Omp40 to form ion channels in planar lipid bilayers strongly support its role as a porin in this chemolithoautotrophic acidophilic microorganism. Some characteristics of the Omp40 protein, such as the presence of a putative L3 loop with an estimated isoelectric point of 7.21 allow us to speculate that this can be the result of an adaptation of the acidophilic T. ferrooxidans to prevent free movement of protons across its outer membrane.

Thiobacillus ferrooxidans is a chemolithoautotrophic acidophilic bacterium with great industrial importance due to its applications in biomining (11, 27, 36). During bioleaching of minerals, the microorganisms have to adhere to the solid substrate (29, 31). The presence of lipopolysaccharides and other external structures have been described on the surface of the gram-negative T. ferrooxidans, and the possible role of these components during bacterial attachment to the ores has been considered (4, 10, 14, 23). A major outer membrane protein of 40 kDa (Omp40) has been previously described in T. ferrooxidans (16, 28), and a possible role for the protein in forming small pores was also reported (35). We have previously found that the expression of Omp40 and other proteins change with variations in the external medium of the bacterium, such as pH and phosphate starvation (3, 16, 25, 33). Also, depending on the oxidizable substrate employed, outer membrane protein changes have been observed in T. ferrooxidans cells grown in ferrous iron or sulfur (7, 18, 23).

When 50% or more of the lipopolysaccharide is removed from T. ferrooxidans cells, an increased exposure of Omp40 on the surface was observed (4) with a concomitant increase in the adherence of the microorganisms to solid sulfur particles, suggesting an important role for these outer membrane proteins during bacterial interaction with the substrate to be oxidized (4). Outer membrane proteins have also been implicated in adhesion mechanisms from other microorganisms. Thus, the major outer membrane protein (38 kDa) from Rahnella aquatilis was shown to be involved in the adhesion of this organism to wheat roots (1).

The major outer membrane proteins from gram-negative bacteria are organized in trimeric structures that form water-filled channels that allow diffusion of small nutrients through the outer membrane (19). Our previously determined N-terminal amino acid sequence of 26 residues from Omp40 from T. ferrooxidans (16) was not long enough to indicate if this protein was related to the enterobacterial porin family. The existence of a different type of outer membrane protein in acidophiles was also possible, since these microorganisms may require a different kind of molecular sieve in their outer membrane to control the passage of ions in the presence of very high concentrations of protons.

Due to the importance of Omp40 as a molecular pore and to find out if the protein presents some specific features that allow T. ferrooxidans to adapt to its acidic environment, in the present report we have employed reverse genetics to isolate and characterize the gene for this surface protein.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The T. ferrooxidans strain ATCC 19859 was used in these studies. Growth on ferrous iron was done in modified 9K medium as described before (3, 4). Escherichia coli JM109 was cultivated in Luria-Bertani (LB) medium (30) at 37°C.

Preparation of outer membrane proteins from T. ferrooxidans and E. coli.

The cells were harvested in the mid- to late-exponential-growth phase by centrifugation (15,000 × g for 15 min at 4°C). The cell pellet was washed three times with 0.01 N diluted sulfuric acid (pH 2), three times with 10 mM sodium citrate (pH 6.9), and one time with the sonication buffer (50 mM Tris-HCl; 10 mM EDTA, pH 8.15; 50 μg of RNase A per ml). All the solutions contained 50 μg of phenylmethylsulfonyl fluoride per ml. Finally, a 20-mg cell pellet was resuspended in 2 ml of sonication buffer. To obtain the outer membrane, a modification of the procedure of Booth and Curtiss (6, 16) was used. Unless indicated otherwise, all of the following operations were at 4°C. The cell suspension was sonicated (five times during 30 s). The lysate was centrifuged at low speed (11,500 × g for 20 min) to eliminate cellular debris. The supernatant was then centrifuged (100,000 × g for 2 h) to pellet the total membrane fraction. The total membrane pellet was washed with the sonication buffer in the presence of 50 mM NaCl, resuspended in 600 μl of 2% sodium laurylsarcosinate, and incubated for 1 h at 37°C. The suspension was centrifuged at low speed (11,500 × g for 20 min), and the supernatant was centrifuged (100,000 × g for 2 h at 4°C) to pellet the outer membrane fraction. The supernatant was discarded, and the pellet was washed with the sonication buffer in the presence of 50 mM NaCl and solubilized in a 7.3% Nonidet P-40, 0.18 M dithiothreitol, and 9% β-mercaptoethanol solution at 56°C for 30 min. The suspension was centrifuged at low speed (11,500 × g for 20 min), and the supernatant was employed as the outer membrane fraction. The E. coli outer membrane fraction was prepared following essentially the same procedure, except that cells were not washed in 0.01 N sulfuric acid and sodium citrate (pH 6.9).

Protein analysis.

Standard two-dimensional (2-D) polyacrylamide gel electrophoresis (PAGE) (pH 5 to 7 in the first dimension) (21) or nonequilibrium pH 2-D PAGE (2-D NEPHGE) (pH 3 to 10 in the first dimension) (22) was performed as described previously for T. ferrooxidans (3, 24, 37). Sodium dodecyl sulfate (SDS)-PAGE consisted of 7.5 to 15% polyacrylamide gradients.

Purification of Omp40 from 2-D gels, N-terminal amino acid sequencing, and production of antiserum against Omp40.

The outer membrane protein fraction from T. ferrooxidans was separated by 2-D PAGE and protein spots corresponding to Omp40 (16) were cut out from the dried Coomassie blue-stained gels by using a scalpel. After rehydration and concentration of the Omp40 spots by SDS-PAGE, the proteins were electroblotted onto a polyvinylidene difluoride (PVDF) Immobilon P (Millipore) membrane as described by Towbin (38), by employing the Trans-Blot Cell system (Bio-Rad) in transfer buffer and an application of 850 mA constant current for 72 min. These proteins were then used for microsequencing (37, 38) or as antigens for the preparation of antiserum. For the generation of internal peptides, the protein was subjected to partial proteolysis and, after separation of the peptides by high-pressure liquid chromatography (HPLC), some of them were subjected to N-terminal-end sequencing. Some sequences were performed by the Laboratoire de Microséquençage des Protéines of the Institut Pasteur Laboratory, and others were performed at the sequencing facilities of the Gesellschaft für Biologische Forschung (GBF), Germany, thanks to Bernd Hofer and Kenneth Timmis.

The antiserum against Omp40 was made by immunizing a BALB/c mouse intraperitoneally with approximately 50 μg of Omp40 (this corresponded to three 2-D gel pieces from the respective gels for each immunization). To prepare the samples for immunization, the gel pieces containing Omp40 were loaded in one well of a slab gel prepared with a meltable synthetic electrophoresis matrix (ProtoPrep; National Diagnostics) and were allowed to rehydrate for 1 h in 50 mM Tris-HCl–10 mM EDTA (pH 8.1). After electrophoresis the gel was stained with Coomassie blue, and the concentrated Omp40 band was excised and washed four times with distilled water. The slice containing Omp40 was weighed, and 1 volume (assuming 1 mg = 1 μl) of ProtoPrep Dissolution Reagent was added and incubated for 1 h at 65°C. The melted viscous ProtoPrep mixture was then directly mixed with 1 volume of Freund's complete adjuvant (Gibco BRL) and vortexed during 1 h to produce an injectable emulsion with a final volume of 800 μl. Immunization was done four times at 1-week intervals. At 2 days after the last injection, the blood was collected from the mouse, and the serum was obtained by centrifugation.

Western immunoblotting.

The proteins separated by SDS-PAGE were electrotransferred to a PVDF membrane as described above. For the antigen-antibody reaction, the membrane containing the transferred proteins was treated with the antiserum against Omp40 as the primary antibody (1:4,000 dilution), and monoclonal anti-mouse antibodies were conjugated with peroxidase (Amersham) as the secondary antibodies (1:2,500 dilution). The specificity of the mouse anti-Omp40 serum was tested with both the preimmune and the immune sera (1:4,000 dilution) against pure Omp40 from T. ferrooxidans and total proteins from T. ferrooxidans and E. coli BL21(DE3). No cross-reaction was observed with E. coli proteins, while only one reacting band was detected in Thiobacillus samples (results not shown).

DNA manipulations.

Standard procedures were used to manipulate T. ferrooxidans DNA (30). After separation of the restriction enzyme-digested DNA fragments by electrophoresis, they were denatured and transferred to a positively charged nylon membrane (Hybond-N⁺; Amersham) by the semidry capillary method (30). Prehybridization and hybridization were performed at 42°C with the DIG Easy Buffer (Roche). Digoxigenin-labeled probes were obtained by PCR as described by Roche with the nondegenerate primer Omp40NH2A-ND/P4023B-ND deduced after DOP-PCR of the T. ferrooxidans DNA fragment sequences. Detection was accomplished by using the DIG Luminescent Detection Kit as described by Roche.

The dideoxy chain termination method was employed to sequence DNA by using [γ-³³P]dATP and the dsDNA Cycle Sequencing System from Gibco BRL. The DNA sequences were compiled and analyzed with the University of Wisconsin GCG package (version 9.1; Genetics Computer Group, Madison, Wis.).

Primers and PCR conditions.

The oligonucleotide primers were purchased from the Fundación Para Estudios Biomédicos Avanzados and Genset Corporation. Taq polymerase and Pwo polymerase from Promega and Roche, respectively, were used according to the manufacturer's recommendations. The fragments were recovered from 1% agarose gels, purified with Wizard PCR Prep (Promega), and cloned in the pGEMT vector (Promega). Next, 20-mer degenerate oligonucleotides (DOPs) were designed on the basis of amino-terminal-end sequence determinations. A total of 60 pmol of each nucleotide and 25 ng of T. ferrooxidans total DNA were used in 50-μl reactions.

Amplification of flanking sequences was done by inverse PCR and SSP-PCR as described before by Ochman et al. (20) and by Shyamala et al. (34), respectively.

For DOP-PCR, the oligonucleotide primers and the amino acid sequences used to deduce them were Omp40NH2A (5′-GTNTTYGGNTAYGCNCARAT-3′) (VFGYAQI), P4017A (5′-TAYTAYATHCARGGNNCNTA-3′) (YYIQGAY), P4017B (5′-TANGCNCCYTGDATRTARTA-3′) (YYIQGAY), P4023A (5′-CAYGCNGAYGAYGTNATGGG-3′) (HADDVMG), and P4023B (5′-CCCATNACRTCRTCNGCRTG-3′) (HADDVMG). The DOP-PCR amplifications were as follows: 3 min at 95°C, followed by 25 cycles at 95°C for 25 s, 55°C for 30 s, and 72°C for 45 s, and then 3 min at 72°C.

For inverse PCR we used the Omp40-1B (5′-GCACCAAAAATGAGGCCATT-3′) and Omp40-2A (5′-GGCACCGCGGGTAATGAACT-3′) primers (see DNA sequence in Fig. 2).

FIG. 2.

Nucleotide and deduced amino acid sequences of the omp-40 gene. The black dot indicates the putative transcription initiation site. The signal peptide sequence recognized by a putative signal peptidase is underlined. The possible ribosomal binding site is shaded.

Inverse PCR reactions were performed on total T. ferrooxidans DNA digested with AvaI and then religated as follows: 3 min at 95°C, followed by 30 cycles at 95°C for 25 s, 67°C for 30 s, and 72°C for 1 min, and then 3 min at 72°C.

RNA manipulations.

T. ferrooxidans total RNA was prepared by the hot phenol method (2) from 600 ml of a culture grown on a medium containing ferrous iron. For the next steps, all the solutions were treated with dimethylpyrocarbonate (DMPC). The cellular pellet was resuspended in 500 μl of 0.02 M sodium acetate (pH 5.5)–0.5% SDS–1 mM EDTA. The lysed cells were extracted twice at 60°C with phenol equilibrated with 0.02 M sodium acetate (pH 5.5)–0.5% SDS–1 mM EDTA. Total RNA was precipitated with KCl (20 mM final) and 95% ethanol. The pellet was resuspended in 100 μl of DMPC-treated water. DNA-free RNA was finally obtained with the High Pure RNA isolation kit (Roche), omitting the lysozyme step.

Determination of a putative transcription initiation site.

We used the 5′ RACE (rapid amplification of cDNA ends) system according to the recommendations described by Gibco BRL. Total RNA (1 mg) was used to obtain a cDNA strand with the Omp40-1B primer (nucleotides 219 through 200; see Fig. 2). After purification, the cDNA was tailed with dCTP. The dC-tailed cDNA was amplified directly by PCR using the 5′ RACE abridged anchor primer and the Omp40-Ext1 primer (nucleotides 145 through 125; see Fig. 2). The amplified DNA fragment was purified from agarose gel by using Wizard PCR Prep (Promega) and sequenced by using the Omp40-Ext1 and Omp40-Ext2 primers (nucleotides 113 through 94; see Fig. 2).

omp40 gene cloning and expression.

We used the pET system from Novagen. The omp40 gene fragments were obtained by PCR using the Omp40PLNdeI–Omp40CHindIII primer pair and the Omp40PMNdeI–Omp40CHindIII primer pair, which allowed us to obtain the omp40 gene with or without, respectively, the coding sequence for the leader peptide. We used the Pwo polymerase (Roche) and a low number of amplification cycles to decrease the sequence error. After purification and digestion by the corresponding restriction enzymes, the two different DNA fragments were ligated to the pET21a vector digested with NdeI and HindIII. The ligation product was used to transform the E. coli BL21(DE3)/pLysS. The recombinant clones were selected on LB solid medium supplemented with ampicillin (100 μg/ml). The induction-expression analysis was done in LB liquid medium with or without 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside). Expression of Omp40 was determined by using total protein or outer membrane preparations from the different E. coli cells. After SDS-PAGE of these fractions, Western blotting was done using the Omp40 antiserum for detection.

Nucleotide sequence accession number.

The nucleotide sequence of the omp40 gene is available in the EMBL database under accession no. AJ012661.

RESULTS AND DISCUSSION

Solubilization properties of Omp40.

Although Omp40 from T. ferrooxidans was considered as a porin by its ability to form channels in planar lipid bilayers (35), only preliminary biochemical data is available for this protein. It solubilizes at 100°C as a single species and forms an oligomer of 90 kDa; this does not explain the possible formation of a trimeric structure (28, 35). We analyzed the behavior of Omp40 in SDS-PAGE at different temperatures, using a specific antibody against Omp40. Figure 1 shows that Omp40 was solubilized in Laemmli sample buffer after a 5-min incubation at temperatures of >56°C, since the 40-kDa monomer was present only at 75 and 100°C (arrow). These results indicate that Omp40 most likely forms a stable trimer of about 120 kDa (filled dot) which dissociates at high temperature, a behavior similar to that described for other porins.

FIG. 1.

Solubilization of Omp40 at different temperatures. The purified outer membrane fraction from T. ferrooxidans was solubilized in Laemmli buffer at 20°C (lanes a), 37°C (lanes b), 56°C (lanes c), 75°C (lanes d), and 100°C (lanes e). The proteins were then resolved by SDS-PAGE and were stained with Coomassie blue (A) or were transferred to a PVDF membrane, followed by reaction with antiserum against Omp40 and colorimetric development as described in Materials and Methods (B). Numbers to the left indicate molecular mass markers in kilodaltons. The migrating position of Omp40 is indicated by an arrow, and its trimeric form is indicated by a filled dot.

Purification of Omp40 and amino acid sequences of some of its peptides.

Omp40 was purified as a single spot by 2-D PAGE as we have described before (16). The isolated Omp40 was subjected to N-terminal-end sequencing, which confirmed our previously reported sequence (16): ADTSNADTGPVVFGYAQITGAQQFGT (amino acids 23 through 43; Fig. 2). We also generated the following internal peptide sequences from Omp40: GEAVPGVTYYIQGAY (amino acids 69 through 83) and SAGAMLHADDVMGTG (amino acids 170 through 184).

Isolation of the omp40 gene from T. ferrooxidans.

Based on the amino acid sequences obtained from Omp40, degenerate primers were designed as described in Materials and Methods and, after cloning of the amplified DNA fragments, a sequence of 1,301 nucleotides was obtained which contained an open reading frame (ORF) with the complete putative omp-40 gene. This nucleotide sequence and the deduced amino acid sequence obtained are shown in Fig. 2.

The ORF corresponding to Omp40 started with an AUG codon in nucleotide 40 and ended with a stop UAG codon in nucleotide 1,252. Identity searching in databases with the ALIGN program (http://vega.igh.cnrs.fr/bin/nph-align-qury.pl) indicated a similarity of this ORF to several bacterial porin genes. It was preceded by a plausible ribosome-binding site with a GAGGAG sequence located upstream from the initiating AUG codon (nucleotides 28 through 33).

As expected for an outer membrane protein, the omp-40 gene contained a signal peptide sequence corresponding to the 22 amino acids indicated in Fig. 2. Therefore, the deduced mature Omp40 protein had 382 amino acids (from nucleotides 97 through 1,252) and a molecular mass of 40,095.7 Da, with a theoretical isoelectric point of 4.73. These values correlate fairly well with the 40 kDa value that we previously obtained by 2-D PAGE analysis of Omp40 (16).

Determination of a putative transcription initiation site for omp-40.

To determine the transcription initiation site by using the primer extension procedure, we needed to know some of the omp-40 gene sequence upstream of the presumptive initiation site. However, the AvaI site chosen for the inverse PCR cloning experiment was too close to the front of the Omp40 ORF. Lacking this information, as an approach to obtain this data, we employed the 5′ RACE system. Although this method does not allow an exact determination of the transcription initiation site, we could obtain an estimation of a possible putative site of initiation of transcription, as shown in Fig. 3. If the mRNA does not start with one or more cytosines, the dC tail added to the cDNA strongly suggests that the adenine indicated with the asterisk (located 40 bp from the initiating translation codon) may be the transcription initiation site (this is the thymine indicated with a filled dot at the beginning of the sequence of the corresponding complementary strand shown in Fig. 2).

FIG. 3.

Determination of a putative transcription initiation site for omp40. RNA was purified from cells of T. ferrooxidans, and the cDNA was obtained with reverse transcriptase and a primer as described in Materials and Methods. The purified DNA was tailed with dCTP and, after amplification by PCR, was used for nucleotide sequencing. The Omp40-Ext1 and Omp40-Ext2 primers were used in combination with the 5′ RACE anchored primer to generate the sequence. Lanes A, C, G, and T show the sequence ladders generated. The relevant DNA sequence is shown on the right, and the position of the possible start site is indicated with an asterisk.

In vivo expression of omp-40 in E. coli.

The omp40 gene amplified by PCR was cloned in the plasmid pET21a and was used to study the expression of the T. ferrooxidans protein in E. coli. As Fig. 4A shows, there was an increased level of synthesis of a protein band of around 40 kDa (arrow) in cells containing a plasmid with or without the region coding for the signal peptide. This product was expressed under the control of the lac promoter of the cloning vector when the cells were induced by the presence of IPTG. To confirm that this 40-kDa protein band corresponded to Omp40, proteins of the same kinds of cells were subjected to Western blotting by using the antibodies against Omp40 (Fig. 4B). Omp40 was clearly expressed under the control of the lac promoter. No reaction of the antiserum against the E. coli outer membrane proteins was seen under the conditions of our experiments.

FIG. 4.

Expression of T. ferrooxidans omp-40 gene in E. coli. The omp-40 gene containing the signal peptide coding region (lanes 3, b, and e) or without this fragment (lanes 2, a, and d) were amplified by PCR using Pwo polymerase and a low number of cycles. The amplified fragments were then cloned in the pET21a plasmid which was employed to transform E. coli strain BL21(DE3)/pLysS. Control E. coli cells transformed with pET21a without the insert (lanes 1 and c) were also used. All of the strains were grown for 2 h in the presence (+) or in the absence (−) of 1 mM IPTG added at the half-logarithmic phase of growth as indicated. The total cell proteins (A and B) or the outer membrane fraction (C and D) from each bacterial strain were separated by SDS-PAGE and stained with Coomassie blue (A and C) or were subjected to Western blotting employing antiserum against Omp40 (B and D). Some of the samples (C and D) were denatured before electrophoresis at 45°C (lanes a and b) or at 95°C (lanes c, d, and e). The arrow indicates the monomeric form of Omp40, and the filled dot indicates its trimeric form.

When the outer membrane preparations of E. coli cells induced to express Omp40 in the presence of IPTG were used for SDS-PAGE analysis (Fig. 4C), a faint band of 40 kDa (arrow) was seen close to the intense bands of around 38 kDa (most probably corresponding to OmpC and OmpF from E. coli). When these samples were solubilized at 45°C, mainly 120-kDa bands were seen, and no 40-kDa bands were seen. These high-molecular-weight bands disappeared by solubilization of the samples at 95°C, with a concomitant appearance of bands in the range of 38 to 40 kDa. To identify Omp40 in these E. coli outer membrane samples, we used Western blotting analysis with the antiserum against Omp40 as seen in Fig. 4D. The results clearly confirm that Omp40 has monomeric (arrow) and trimeric (filled dot) states but, more importantly, these results also show that only when the signal peptide sequence was present in the omp-40 gene was the protein localized to the E. coli outer membrane fraction. The amount of Omp40 present in the outer membrane of E. coli was rather low, suggesting that the presence of the much higher amounts of OmpC and OmpF do not allow a higher incorporation of the heterologous Omp40 protein. Although the E. coli signal peptidase may recognize the T. ferrooxidans signal peptide present in omp40, the protein may be inserted in the E. coli membrane in a nonfunctional form. Due to the great difference in the growth pH (4 or 5 U) between these two microorganisms, it may not be possible to test the functionality of Omp40 in E. coli under different pH values.

Some characteristics of Omp40 compared with other porins.

When aligned with the amino acid sequence of other porins such as OmpC from several species, only about 22% identity was obtained (Fig. 5). However, some highly conserved sequences present in most porin sequences, such as RLGFKGE, were also present in Omp40 in the same approximate region. Omp40 also contained the N-terminal phenylalanine, which is important for the structure of the barrel (9) and which is present in all members of the porin superfamily (15). The Fig. 5 alignment also shows that several amino acid residues highly conserved in most porins are also present in Omp40. This suggests strongly that Omp40 is a porin, an idea further supported by previous studies that indicated that this protein has the capacity of forming pores in planar lipid bilayers (35).

FIG. 5.

Multiple amino acid sequence alignment of omp-40 with potential OmpC porin homologues. Organisms are indicated as follows: ST, Salmonella typhimurium (EMBL accession no. AF039309); STY, Salmonella typhi (SwissProt accession no. 052503); EC, E. coli (SwissProt accession no. P06996); KP, Klebsiella pneumoniae (SwissProt accession no. Q48473); RA, Rahnella aquatilis (SwissProt accession no. 033507); SM, Serratia marcescens (SwissProt accession no. Q54471); and TF, T. ferrooxidans. Conserved residues present in at least four of the sequences compared, including Omp40, are in boldface. The β strands, loops, and turns present in the previously known porins are indicated.

The Omp40 amino acid sequence analyzed by employing the GOR secondary structure prediction from Southampton Bioinformatics Data Server (http://molbiol.soton.ac.uk/cgi-bin/GOR.pl) fits well with the predicted β-strands, loops, and turns defined for several of the porins (Fig. 5). This is in spite of the rather low degree of amino acid sequence identity with enterobacterial porins, in which the predicted β-strands are highly conserved. On the basis of this comparison, we propose a working folding model for Omp40 as shown in Fig. 6. In this model most of the conserved amino acids present in the OmpC porins shown in Fig. 5 form part of the putative β-strands.

FIG. 6.

A proposed model for the predicted folding of Omp40 from T. ferrooxidans. The tentative β-strands array was located on the basis of the proposal by Paul and Rosenbusch (26) and by comparison with other known models for porins (8). The conserved amino acid residues indicated in boldface in Fig. 5 are indicated here in boldface and circled.

Porins form trimers, each monomer constituting a discrete pore. Within each pore a long polypeptide loop (L3) runs along one side of the barrel wall and narrows the pore to create an “eyelet” region (5, 8, 19). Across this region, these porins have a strong transverse electric field generated by basic residues in the barrel wall and acidic residues and peptide carbonyl groups on L3 (8, 17). This loop forms a constriction that determines pore characteristics such as channel size and ion selectivity (39). Thus, unusually large channels are produced in OmpG, an E. coli porin that lacks the large external loop L3 (13). The permeability is determined not only by the size of the penetrating molecule but also by its charges, which have to be oriented within the transverse electric field in the constriction zone (5, 17). Changes in the pore size have been reported when amino acids of the PEFGG sequence present in loop L3, which is highly conserved in the superfamily of bacterial porins (15) are replaced by mutation (5, 12). For example, the change of glutamic acid for cysteine altered the permeability of some charged molecules (12).

Although much more experimental work remains to be done with Omp40 from T. ferrooxidans, we think it interesting to advance the next speculation. Omp40 contains a putative loop L3 (AQAQLMDAWINFAPVPFAQLQVGKFKTPEGLEYTGTAGN) in which a PVPFAQ sequence exists instead of PEFGG. The isoelectric point calculated for this L3 sequence is 7.21. On the other hand, OmpC from E. coli has an L3 loop (NYGVVYDVTSWTDVLPEFGGDTYGSDNFMQQRGN GFA) with a calculated isoelectric point of 3.47. The net electric charge of this loop is important for the permeability of the molecules passing through the pore. At neutral pH, if one assigns to the cationic amino acids arginine and lysine each a charge of +1, to the cationic amino acid histidine a charge of +0.5, and to the anionic amino acids glutamic acid and aspartic acid each a charge of −1, one can calculate the net charges of the loops as the sum of the charges. This charge for E. coli OmpC L3 is (−4). For an acidophilic microorganism such as T. ferrooxidans, growing at pH 1.5 to 2.5, one can calculate for the putative L3 present in Omp40 L3 a net charge of (+2) at pH 2.5. Additionally, all of the putative loops of Omp40 would give a sum resulting in a net charge of +5.5 compared to a negative net charge for the sum of the charges present at pH 7.0 in all the loops of E. coli OmpC or OmpF. This difference in charge may represent a special adaptation of acidophilic microorganisms allowing them to somehow control the free passage of protons from the outside and thus avoid an excessive acidification of their periplasmic space (usually at pH 2.5 to 3.5). Being positively charged, the pore would restrict the diffusion of protons both from outside and from the periplasmic space toward the environment. Consequently, a small size pore in the outer membrane of T. ferrooxidans would be advantageous to survival at very low external pH. In agreement with this speculation, the channel formed by Omp40 has been described as a small pore and slightly anionic (35).