INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The syphilis spirochete, Treponema pallidum subsp. pallidum, possesses a unique outer membrane containing 100-fold-less membrane-spanning outer membrane protein than typical gram-negative bacterial outer membranes (30, 38). The identification of these T. pallidum rare outer membrane proteins, collectively termed TROMPs (9), has been a major focus of syphilis research because they are surface exposed proteins of likely significance to both pathogenesis and immunity.

We previously reported a method to isolate the T. pallidum outer membrane (8) that resulted in the identification of two proteins markedly enriched in these preparations. One of these proteins, termed Tromp1, was found to have properties consistent with an outer membrane origin, including amphiphilicity and porin activity measured in planar lipid bilayers (6). We also found that recombinant Tromp1, expressed and exported in Escherichia coli, in part localized to the E. coli outer membrane, where porin activity similar to that of native Tromp1 was again demonstrated (7).

While our past findings have been consistent with Tromp1 being an outer membrane protein, two recent reports have challenged this conclusion. Hardham et al. (18) have found that the tromp1 gene, referred to as troA in their studies, is part of an operon with similarities to ABC transporter systems, and that TroA has 26 to 28% sequence identity to the periplasmic binding protein component of these operons. This observation has resulted in the alternative view that Tromp1 is a periplasmic binding protein (18). Akins et al. (1) reported that native Tromp1 has an uncleaved signal peptide which anchors Tromp1 to the cytoplasmic membrane and accounts for its demonstrated hydrophobicity. These investigators also reported that recombinant Tromp1, made without a signal peptide, lacked hydrophobic properties and porin activity, findings used to support their conclusion that Tromp1 is an inner membrane anchored periplasmic binding protein. Thus, whether Tromp1 is a bona fide outer membrane protein or periplasmic binding protein has been a topic of current debate.

To address the controversy surrounding Tromp1, we have recently engaged in a structural analysis of the native protein. We have conclusively demonstrated, by use of mass spectrometry, that native Tromp1 has a cleaved signal peptide (10) in contrast to the report by Akins et al. (1). Implicit from this finding is that the observed hydrophobicity of native Tromp1 is not due to an uncleaved signal peptide as reported but is rather a property of its conformation.

In the present study, we show that denaturation of native Tromp1 eliminates its hydrophobicity, a finding consistent with the conformation of Tromp1 being responsible for this property. In addition, we show that purified native Tromp1, which has porin activity, exists in a trimeric conformation. These properties of native Tromp1 are successfully recreated by using a signal-peptide-lacking form of recombinant Tromp1 which was renatured into a trimeric, hydrophobic, and porin-active conformation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Source of T. pallidum. T. pallidum subsp. pallidum, Nichols strain, was maintained by testicular passage in New Zealand White rabbits as previously described (24). T. pallidum used for all experiments was extracted from infected rabbit testicles in phosphate-buffered saline, pH 7.4 (PBS), centrifuged two times at 400 × g to pellet gross tissue debris, and then centrifuged at 10,000 × g to pellet the organisms. The treponemal pellet was washed once in PBS, recentrifuged to again pellet the organisms, and then resuspended in PBS to yield a final concentration of 10¹⁰ organisms/ml.

FPLC purification of native Tromp1. A total of 2 × 10¹¹ T. pallidum, extracted from 20 infected rabbits, was incubated in 40 ml of ice-cold 0.1% Triton X-100 (TX-100; Calbiochem) for 2 h in order to solubilize the outer membrane and completely release Tromp1 (approximately 5 µg) (6). The suspension was then centrifuged at 20,000 × g for 40 min at 4°C in order to pellet the protoplasmic cylinders. The supernatant, containing all of the Tromp1 and other detergent-extracted proteins, was subjected to anion-exchange chromatography by using a fast-performance liquid chromatography (FPLC) system (Pharmacia Co., Alameda, Calif.). The anion-exchange buffer consisted of 50 mM Tris-HCl (pH 8.0) with 0.2% hydrogenated Triton X-100, and the proteins were eluted by using a salt gradient of 0 to 1 M NaCl. Fractions enriched with Tromp1, determined by immunoblot analysis with monospecific anti-recombinant Tromp1 (rTromp1) serum generated as previously described (7), were collected, pooled, and rechromatographed under these conditions. The second round of Tromp1-enriched fractions was then subjected to chromatofocusing chromatography by using a MonoP column (Pharmacia). The starting buffer contained of 25 mM bis-Tris (pH 6.7) with 0.1% hydrogenated Triton X-100, and the elution buffer contained 10% Polybuffer, pH 5.0 (Pharmacia), containing 0.1% hydrogenated Triton X-100. After chromatofocusing, Tromp1 purity was demonstrated by silver stain (see Fig. 2) and immunoblot analysis by using syphilitic infection-derived immune rabbit serum (data not shown). The total amount of recovered native Tromp1 from 2 × 10¹¹ organisms was estimated by gel analysis to be 500 ng (10% recovery). A total of 10 isolations, each using 2 × 10¹¹ organisms (a total of 200 infected rabbits), were performed which resulted in the isolation of 5 µg of purified native Tromp1.

Expression of a signal-less form of rTromp1. A signal-less form of recombinant Tromp1 (rTromp1), having a calculated molecular mass of 31,109 Da, was generated as follows. An N-terminal primer of 5'-CGCCATATGAGCAAGGATGCCGCAGCAGAC-3' (underlined region indicates the tromp1 gene sequence) corresponding to signal peptide cleavage after alanine-phenylalanine-glycine (AFG) was generated containing an NdeI restriction endonuclease site at the 5' end (Gibco BRL, Gaithersburg, Md.). This construct results in a single methonine residue placed ahead of the site of signal peptide cleavage for the purpose of translation. A C-terminal primer consisting of 5'-CGCGGATCCCTAGCGAGCCAACGCAGCAA-3' and corresponding to the end of the tromp1 gene was generated containing a BamHI restriction endonuclease site at the 5' end. PCR was performed by using these primers as described previously (7). The tromp1 PCR product was ligated into pET17b (Novagen, Inc.), previously digested with NdeI and BamHI. The resulting construct was transformed into E. coli BL21(DE3)(pLysE) (Novagen, Inc.) by using cells made competent by CaCl₂. Expression and FPLC purification of rTromp1 was performed as described previously (7).

Renaturation of rTromp1. The protocol described by Qi et al. (27) for the renaturation of recombinant neisserial porin proteins was used for the renaturation of signal-less rTromp1 as follows. Solid urea was added to 1 ml of FPLC-purified rTromp1 (2 mg/ml) in 50 mM Tris-100 mM NaCl-0.05% Zwittergent-3,14 (Calbiochem-Novabiochem Corp., La Jolla, Calif.) at pH 8.0 to give a final concentration of 8 M urea. The suspension was then boiled for 15 min in order to completely denature rTromp1. After denaturation, the suspension was added to an equal volume of 1% Zwittergent-3,14 (Calbiochem-Novabiochem Corp.) and then dialyzed overnight at 4°C against a buffer containing 100 mM Tris, 200 mM NaCl, 10 mM EDTA, and 0.02% sodium azide at pH 8.0 in order to remove the urea and renature rTromp1. The renatured rTromp1 suspension was then applied to a Sephacryl S-300 column (1.5 by 75 cm; Pharmacia Biotech, Inc.), which was equilibrated in the same buffer used for dialysis and which had been previously calibrated with known molecular mass standards (gel filtration molecular weight markers; Sigma Chemical Co.). In some experiments, FPLC-purified rTromp1 was added directly to the column without prior denaturation and renaturation treatment. Fractions containing rTromp1 were identified by spectrophotometry (optical density at 280 nm) and by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblot analysis. Renatured rTromp1 samples were stored at 70°C.

Cross-linking experiments. Renatured and nonrenatured rTromp1 and native Tromp1 were dialyzed against PBS. Twenty-five microliters of each of the Tromp1 preparations containing 1 µg of protein was mixed with an equal volume of bis(sulfosuccinimidyl) suberate (BS³) (Pierce, Rockford, Ill.) at various concentrations. BS³ is a homobifunctional N-hydroxysuccinimide ester (NHS-ester) which can form noncleavable covalent bonds between N-terminal -amine groups and/or side chains of amino acids of peptides and proteins that are separated by less than 11 Å; -amine of lysine is the principle side chain target for this NHS-ester (25). After 30 min of incubation at room temperature, 1 M Tris (pH 7.5) was added to give a final concentration of 40 mM in the reaction mixture, which was then incubated for an additional 15 min at room temperature. The mixtures were then dried by using a speed vacuum, and the dried contents were resuspended in sample buffer containing 8 M urea. Samples were boiled for 10 min prior to SDS-PAGE, followed by transfer to polyvinylidene difluoride (PVDF) immunoblotting membranes. Immunoblotting for the detection of Tromp1 was conducted as described below.

TX-114 treatments. Triton X-114 (TX-114; Calbiochem) detergent extraction of T. pallidum and separation of the extracted material into detergent (hydrophobic) and aqueous (hydrophilic) phases was performed as follows. Ten microliters of treponemal suspension (10⁸ T. pallidum) was added to 500 µl of ice-cold 0.1% TX-114 in PBS. The suspension was incubated at 4°C for 2 h to selectively solubilize the outer membrane, followed by centrifugation at 13,000 × g to pellet the treponemal protoplasmic cylinders. This procedure, using 2% TX-114, has been shown not to solubilize the inner membrane of T. pallidum (12, 29). The protoplasmic cylinders were washed once in PBS and then recentrifuged prior to resuspension in SDS-PAGE sample buffer containing 2% SDS, 5% 2-mercaptoethanol, and 15% glycerol in 100 mM Tris (pH 6.8). To the extracted supernatant, 100 µl of 10% TX-114 was added (final concentration of 2.1% detergent) prior to warming at 37°C for 5 min to induce cloud formation. The suspension was then centrifuged at 13,000 × g for 5 min at room temperature in order to separate aqueous (top) and detergent (bottom) phases. Both aqueous and detergent phases were recovered and reextracted three times as described above with cold 2% Triton X-114 and PBS, respectively. In some experiments, urea treatment of T. pallidum was performed prior to TX-114 extraction and phase separation as follows. Ten microliters of treponemal suspension (10⁸ T. pallidum) was combined with 70 µl of 8 M urea (final concentration of 7 M urea), followed by heating at 90°C for 10 min. The suspension was then cooled and added to 500 µl of 0.1% TX-114. Extraction and phase separation was performed as described above.

Porin assays. The pore-forming ability of native and recombinant Tromp1 was examined by using planar lipid bilayers (5). Lipid bilayers made from 1.5% (wt/vol) oxidized cholesterol in n-decane were formed across a 0.2-mm² hole separating two compartments of a Teflon chamber containing a 1 M KCl bathing solution. Bilayer formation was recognized by the fact that the membrane appeared black when viewed by incident light. Calomel electrodes were implanted in each compartment, with one connected to a voltage source and one connected to a current amplifier. Conductance data were recorded on a strip chart recorder for further analysis. Native and rTromp1 samples were diluted in 1 M KCl containing 0.1% Triton X-100 to yield a final concentration of approximately 1 µg of Tromp1 protein per ml of solution. Ten microliters of diluted sample, containing approximately 10 ng of Tromp1 protein, was added to the 1 M KCl bathing solution. Pore-forming ability was then assessed by applying a voltage of 50 mV across the lipid bilayer and measuring the increases in conductance.

CD analysis. Circular dichroism (CD) analysis was performed on a JASCO J-600 spectropolarimeter by using a scan speed of 5 nm/min, a time constant of 8 s, and a bandwidth of 1.0 nm. Four scans were averaged for each spectrum. The baseline correction option was used to subtract a buffer baseline. Spectra were recorded from 240 to 190 nm in 1-mm pathlength cells with protein concentrations of 0.1 to 0.2 mg/ml. Spectra were analyzed for secondary structure by using a neural network algorithm (2, 23).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The hydrophobicity of native Tromp1 is conformation dependent. In order to investigate the basis of Tromp1's hydrophobicity, urea treatment was utilized to test whether denaturing conditions have an effect upon this property. As previously reported (6) and again shown in Fig. 1A, 0.1% TX-114 detergent extraction of whole organisms, known to completely solubilize the T. pallidum outer membrane (12, 29), resulted in the complete release of Tromp1 with no detection present in the protoplasmic cylinders (lane P). Upon phase separation of the released material, all of the Tromp1 was present in the detergent phase (lane D), once again demonstrating the hydrophobicity of native Tromp1. However, when whole organisms were first treated with 8 M urea at 90°C prior to TX-114 extraction and phase separation, all of the Tromp1 now phase separated into the hydrophilic aqueous phase (Fig. 1B, lane A), indicating the complete loss of its hydrophobic nature. By comparison, when this same immunoblot was reprobed with a monoclonal antibody against the hydrophobic 47-kDa lipoprotein (Fig. 1C), this protein was found to remain in the detergent phase, and no detection in the aqueous phase was observed. In addition, a significant amount of 47-kDa lipoprotein remained associated with the protoplasmic cylinders (lane P), in contrast to Tromp1 and consistent with the current belief that the 47-kDa lipoprotein is inner membrane anchored (28).

View larger version (35K): [in this window] [in a new window]   FIG. 1.   Hydrophobicity of native Tromp1 is conformation dependent. A total of 108 T. pallidum organisms were extracted in 0.1% TX-114 in the presence or absence of urea prior to phase separation and SDS-PAGE. Transferred immunoblots contain whole organisms (W), protoplasmic cylinders (P), aqueous-phase proteins (A), and detergent-phase (D) proteins. Panels: A, urea-untreated fractions probed with anti-Tromp1 serum (Tromp1); B, urea-treated fractions probed with Tromp1; C, the same blot as in panel B reprobed with a monoclonal antibody against the 47-kDa lipoprotein (47-kDa LP). Numbers on the left indicate the positions of the molecular weight standards in thousands.

Porin activity and trimer conformation of native Tromp1. To confirm the pore-forming activity of native Tromp1, isolated previously by isoelectric focusing (6), and to analyze Tromp1's structural characteristics, the native protein was purified from 0.1% TX-100 detergent extracts of whole organisms by FPLC. The small amount of Tromp1 present in T. pallidum and a 10% efficiency of recovery by using multiple steps of FPLC purification resulted in the isolation of only 5 µg of purified native Tromp1 from 2 × 10¹² organisms. The purity of the isolated material was confirmed by silver staining after separation by SDS-PAGE. As shown in Fig. 2, single stained bands in amounts of 0.25 and 1.0 µg of electrophoresed Tromp1 were the only proteins detected. These proteins reacted specifically with anti-Tromp1 antiserum by immunoblot analysis (data not shown).

View larger version (79K): [in this window] [in a new window]   FIG. 2.   Purity of FPLC-isolated native Tromp1. Samples containing 0.25 and 1.0 µg of FPLC-purified native Tromp1 were separated by SDS-PAGE and silver stained. Numbers on the left indicate the molecular weights (in thousands) of the standards separated.

View larger version (12K): [in this window] [in a new window]   FIG. 3.   Porin activity of native Tromp1. FPLC-purified native Tromp1 was added to the aqueous-phase (1 M KCl) bathing a lipid bilayer membrane. Histogram of single-channel conductance increases for over 136 membrane insertion events. Conductance increases showed a mean distribution ca. 0.96 ± 0.33 nS (± the standard deviation).

View larger version (49K): [in this window] [in a new window]   FIG. 4.   Molecular cross-linking of FPLC-purified native Tromp1. Samples containing 1 µg of native Tromp1 were incubated in the cross-linker BS3 at the concentrations indicated. Samples were separated by SDS-PAGE, immunoblotted, and probed with anti-Tromp1 serum. Arrows indicated the molecular mass position of monomeric Tromp1 and positions of putative dimeric and trimeric forms of Tromp1. Numbers on the left indicate the positions of the molecular weight standards (in thousands).

Renaturation of a signal-less form of rTromp1 into a trimeric, hydrophobic, and porin-active conformation. In an attempt to recreate the functional and structural properties observed for native Tromp1, a 31,109-Da signal-peptide-lacking form of rTromp1 was expressed, purified, and subjected to a renaturation procedure previously reported for recombinant neisserial porins (27, 32). As shown in Fig. 5, all renatured rTromp1 loaded onto a Sephacryl gel exclusion column, previously calibrated with known molecular mass standards, eluted at the position of the phosphorylase b standard, which is 97 kDa. By comparison, all nonrenatured rTromp1, loaded onto the same column, eluted at the position of the carbonic anhydrase standard, which is 29 kDa (Fig. 5, inset). This finding suggested that renaturation resulted in the complete trimerization of all rTromp1.

View larger version (14K): [in this window] [in a new window]   FIG. 5.   Sephacryl S-300 column elution profile of renatured rTromp1. Renatured rTromp1 was loaded onto an S-300 column previously calibrated with the following molecular mass standards: blue dextran (200 kDa), phosphorylase b (97 kDa), bovine serum albumin (BSA; 67 kDa), and carbonic anhydrase (29 kDa). The inset shows the elution profile as a function of the log molecular mass versus the eluted volume. Numbers in the inset indicate the following: 1, phosphorylase b; 2, renatured rTromp1; 3, bovine serum albumin; 4, nonrenatured rTromp1; and 5, carbonic anhydrase.

View larger version (49K): [in this window] [in a new window]   FIG. 6.   Molecular cross-linking of nonrenatured and renatured rTromp1. Samples containing 1 µg of nonrenatured and renatured Tromp1 were incubated in the cross-linker BS3 at the concentrations indicated. Samples were separated by SDS-PAGE, immunoblotted, and probed with anti-Tromp1 serum. Arrows indicate the molecular mass positions of monomeric rTromp1 and the positions of putative dimeric and trimeric forms of rTromp1. Numbers on the left indicate the positions of the molecular weight standards (in thousands).

View larger version (60K): [in this window] [in a new window]   FIG. 7.   Renatured rTromp1 is hydrophobic. Samples containing nonrenatured and renatured rTromp1 were combined with TX-114 and phase separated into aqueous (AQ) and detergent (DET) phases prior to SDS-PAGE and immunoblotting. The blot was probed with anti-Tromp1 serum. Numbers on the left indicate the positions of the molecular weight standards (in thousands).

View larger version (38K): [in this window] [in a new window]   FIG. 8.   Porin activity of renatured rTromp1. Renatured rTromp1 was added to the aqueous phase (1 M KCl) bathing a lipid bilayer membrane. (A) Step increases in conductance after the addition of renatured rTromp1. (B) Histogram of single-channel conductance increases for more than 104 membrane insertion events. Conductance increases showed a mean distribution about 0.65 ± 0.24 nS (± the standard deviation). No conductance activity was observed for nonrenatured rTromp1 added to the lipid bilayer membrane (data not shown).

Comparison of nonrenatured and renatured rTromp1 by CD spectroscopy. In order to determine whether renaturation had resulted in a measurable change in the secondary structure of rTromp1, both nonrenatured and renatured rTromp1 were analyzed by CD spectroscopy. The CD spectrum of nonrenatured rTromp1 displayed prominent double minima (negative peaks) at 208 and 222 nm (Fig. 9), findings typical for a protein with considerable alpha-helical secondary structure. By using a neural network algorithm (2, 23) nonrenatured rTromp1 was calculated to possess 48% alpha-helix, 18% beta-sheet, and 34% random coil. Similar findings have been reported by Akins et al. (1) for their signal-peptide-lacking form of recombinant Tromp1. In contrast, the CD spectrum of renatured rTromp1 reflected a loss of alpha-helical structure, with the magnitude of the 222- and 208-nm peaks reduced relative to nonrenatured rTromp1 and the 222-nm peak reduced relative to the 208-nm peak. Renatured rTromp1 was calculated to contain 26% alpha-helix, 17% beta-sheet, and 57% random coil. These findings provide quantitative evidence that renaturation of rTromp1 leads to changes in the secondary structure of the molecule, with the renatured form having less alpha-helix than the nonrenatured form.

View larger version (17K): [in this window] [in a new window]   FIG. 9.   Secondary structure analysis of nonrenatured and renatured rTromp1 by CD spectroscopy. The double minimum absorbances displayed by nonrenatured rTromp1 (solid line) is typical of proteins with a high percentage of alpha-helix. By comparison, renatured rTromp1 (dotted line) showed a distinct loss in absorbance at 222 nm, indicating a loss in alpha-helical secondary structure.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Protein conformation has been recognized for gram-negative membrane-spanning outer membrane porins to be a key factor to their structural and functional properties (11). Immunologically relevant epitopes involved in bactericidal activity have been attributed to porin conformation (14, 34, 39). Structural analysis of porins, including X-ray crystallography studies, has only been performed with purified native porins because native conformation of these proteins has not been faithfully replicated with recombinant forms. However, there has been progress in this area since it has been shown recently that recombinant forms of the gonococcal, meningococcal, and Haemophilus influenzae porins can be renatured into conformations which have several properties in common with their respective native proteins, including trimerization and porin function (26, 27, 32), suggesting that native conformation is attainable with a recombinant.

The importance of renaturation of a recombinant porin protein has particular significance to the study of porins of T. pallidum, an organism which can only be acquired in limited numbers from infected animals and whose membrane-spanning outer membrane protein content is 100-fold less than that of typical gram-negative bacteria (30, 38). These limitations have all but eliminated the use of native T. pallidum porin for structural studies which require milligram quantities of material, such as for X-ray crystallography, which will undoubtedly rely in the future on a renatured recombinant.

In this study, the ability to renature a signal-peptide-lacking form of recombinant Tromp1 into a hydrophobic, trimeric, and porin-functional conformation, properties which we also show for native Tromp1, provides strong evidence that the conformation of Tromp1 is a central issue to the native biological properties of this protein. Indeed, denaturation of native Tromp1 with the chaotropic agent urea abrogates the hydrophobicity of the native protein, a property which is consistent with Tromp1 being located in the outer membrane.

The significance of these findings has important implications in view of recent reports pertaining to Tromp1. Akins et al. reported that a signal-peptide-lacking construct of recombinant Tromp1 was not hydrophobic and showed no porin activity as demonstrated by the liposome swelling assay (1). These findings were used to support the conclusion by these investigators that Tromp1 is not an outer membrane protein. In a recent further study with recombinant Tromp1 (TroA) constructs containing and lacking a signal peptide, these investigators have again concluded that the hydrophobicity of native Tromp1 is the result of its hydrophobic signal peptide (13). However, our recent mass spectrometry analysis of the hydrophobic form of native Tromp1 has now clearly demonstrated that the signal peptide is cleaved (10). Recently, a crystal structure of a nonrenatured monomeric form of recombinant Tromp1 (TroA) has been reported and used to conclude that Tromp1 (TroA) is a periplasmic zinc-binding protein (13, 21). While the finding that recombinant Tromp1 monomers can bind zinc may have important implications in regard to its structure and function, our findings that native Tromp1 and renatured recombinant Tromp1 are trimeric in conformation have provided structural data which is inconsistent with Tromp1 being a typical metal-binding periplasmic protein, since these proteins have been shown to be single polypeptide chain monomers in their functional state (19, 37). In addition, the hydrophobicity and porin activity of both native and renatured recombinant Tromp1 are also inconsistent with its being a typical periplasmic binding protein.

One possible explanation for this paradox is that divergent evolution has enabled pathogenic bacteria to use homologues of ABC transporter operons for diverse purposes related, perhaps, to biological function and pathogenesis. Consistent with this possibility is the finding that a related ABC operon system is also found with the surface adhesin proteins of streptococci. Here the periplasmic binding protein homologue is replaced by a surface adhesin (20, 22), which interestingly, also shares 23 to 28% sequence identity to Tromp1 (6). Yet another example where a gene encoding a surface layer protein occupies the position analogous to that of tromp1 in an ABC transporter operon is found in Caulobacter crescentis (4). Thus, the observation that Tromp1 has biological properties consistent with an outer membrane porin protein but has genetic homology to ABC transporter components from gram-negative and gram-positive bacteria may not be inconsistent in the context of a divergent evolutionary process of genes from this spirochete.

Although Tromp1 has physical properties consistent with outer membrane porin proteins from gram-negative bacteria, including hydrophobicity and trimer conformation, it does not appear to share secondary structure characteristics with these proteins. Porins of gram-negative bacteria are known to possess significant amounts of beta-sheet secondary structure, usually greater than 50%, and very little alpha-helical secondary structure, usually less than 15% (36). This has also been found to be the case for the pathogenic spirochete L. kirschneri, whose porin protein OmpL1 has 61% beta-sheet and 19% alpha-helical secondary structure (17). By comparison, computer analysis of the amino acid sequence of Tromp1 has revealed only 34% beta-sheet secondary structure and 56% alpha-helix secondary structure. While CD analysis of renatured recombinant Tromp1 showed a significant loss of alpha-helical secondary structure compared to the nonrenatured recombinant, no corresponding increase in beta-sheet secondary structure was observed. One possible explanation for this finding is that renatured recombinant Tromp1 may not have faithfully replicated native conformation. An alternative explanation for these observations is that membrane-spanning outer membrane proteins of T. pallidum may utilize a secondary structure conformation other than the beta-sheet to span the outer membrane. It is pertinent to note that the T. pallidum outer membrane, unlike that of gram-negative bacteria or other spirochetes with the exception of Borrelia burgdorferi, does not contain lipopolysaccharide. It is also pertinent to note that analysis of the recently released genome sequences of T. pallidum and B. burgdorferi (15, 16) have not revealed any protein with typical gram negative porin protein secondary structure characteristics. Thus, the outer membrane-spanning proteins of these two spirochetes, while having functional properties similar to those of gram-negative outer membrane proteins, may have quite different membrane-spanning structural properties.

The renaturation of recombinant Tromp1 into a conformation which parallels that of the native protein has now provided for sufficient amounts of material to be used in studies designed to address issues of pathogenesis and host immunity. To date, no T. pallidum recombinant protein used for immunization has elicited complete protective immunity in experimental animals comparable to that achieved by infection-derived immunity.

In summary, the results presented in this study demonstrate that the properties of native Tromp1, an outer-membrane-associated protein that is hydrophobic, trimeric in structure, and porin active, can be successfully recreated by renaturation of purified recombinant Tromp1. We believe that native conformation of T. pallidum outer membrane proteins will be found to be a critical issue in determining the role of these proteins in syphilis pathogenesis and immunity.