HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 282, Issue 8, 5944-5958, February 23, 2007

This Article Abstract Full Text (PDF) All Versions of this Article: 282/8/5944    most recent M610215200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Deka, R. K. Articles by Norgard, M. V. Search for Related Content PubMed PubMed Citation Articles by Deka, R. K. Articles by Norgard, M. V. Social Bookmarking                      What's this?

Crystal Structure of the Tp34 (TP0971) Lipoprotein of Treponema pallidum

IMPLICATIONS OF ITS METAL-BOUND STATE AND AFFINITY FOR HUMAN LACTOFERRIN^*

Ranjit K. Deka¹, Chad A. Brautigam¹, Farol L. Tomson, Sarah B. Lumpkins^¶, Diana R. Tomchick, Mischa Machius, and Michael V. Norgard²

From the Departments of Microbiology and Biochemistry, University of Texas Southwestern Medical Center, Dallas, Texas 75390 and the ^¶University of Oklahoma, Norman, Oklahoma 73019

Received for publication, November 1, 2006 , and in revised form, December 7, 2006.

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

 
The Tp34 (TP0971) membrane lipoprotein of Treponema pallidum, an obligate human pathogen and the agent of syphilis, was previously reported to have lactoferrin binding properties. Given the non-cultivatable nature of T. pallidum, a structure-to-function approach was pursued to clarify further potential relationships between the Tp34 structural and biochemical properties and its propensity to bind human lactoferrin. The crystal structure of a nonacylated, recombinant form of Tp34 (rTp34), solved to a resolution of 1.9Å_, revealed two metaloccupied binding sites within a dimer; the identity of the ion most likely was zinc. Residues from both of the monomers contributed to the interfacial metal-binding sites; a novel feature was that the -sulfur of methionine coordinated the zinc ion. Analytical ultracentrifugation showed that, in solution, rTp34 formed a metal-stabilized dimer and that rTp34 bound human lactoferrin with a stoichiometry of 2:1. Isothermal titration calorimetry further revealed that rTp34 bound human lactoferrin at high (submicromolar) affinity. Finally, membrane topology studies revealed that native Tp34 is not located on the outer surface (outer membrane) of T. pallidum but, rather, is periplasmic. How propensity of Tp34 to bind zinc and the iron-sequestering lactoferrin may relate overall to the biology of T. pallidum infection in humans is discussed.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

 
Syphilis is a chronic, complex sexually transmitted disease of humans caused by the noncultivatable spirochete Treponema pallidum (1, 2). Despite the availability of effective antimicrobials, syphilis remains a significant threat to global health and can facilitate the transmission of human immunodeficiency virus (3-5). After cutaneous inoculation, the bacteria become blood-borne and invade various organ systems, including the liver, heart, and nervous system (1, 2). Humans are the only known reservoir for T. pallidum, and although syphilis is one of the oldest recognized sexually transmitted diseases, it remains very poorly understood (1, 2). This largely is a consequence of the fact that T. pallidum cannot be cultivated continuously in vitro (1), thereby hampering studies on treponemal virulence and syphilis pathogenesis. Approaches for discerning T. pallidum membrane biology, which could reveal key aspects of the enigmatic parasitic strategy of the bacterium, thus, have been substantially limited (6-8).

In an attempt to elucidate molecules that could mediate interactions of T. pallidum with its human host, solubilization with the nonionic detergent Triton X-114 was employed to identify a group of integral membrane lipoproteins (9, 10). This approach identified at least five lipoproteins, most of them representing dominant immunogens of T. pallidum (10). This finding constituted a major advance in the treponemal field, because there is increasing evidence that the lipoproteins serve functions vital to the membrane biology of T. pallidum (11). In other bacteria, membrane lipoproteins have importance as virulence factors, modular components of ATP binding cassette (ABC)³ transporters, enzymes, receptors, protective immune targets, and proinflammatory agonists that contribute significantly to innate immune responses (11-18). Whereas early experiments revealed only five treponemal lipoproteins (10), more recent analysis of T. pallidum genome sequence information has predicted at least 45 lipoprotein genes (19). Most of these predicted that open reading frames encode proteins with no known homologies to other structurally or functionally characterized bacterial proteins. In addition, the absence of techniques for in vitro cultivation and genetic manipulation of T. pallidum has greatly hindered the ability to characterize the functional aspects of T. pallidum membrane lipoproteins.

As an alternative approach to investigating the peculiar membrane biology of T. pallidum, we have adopted a structure-to-function approach to formulate new testable hypotheses regarding the potential function(s) of a number of the treponemal lipoproteins. That a structural biology approach for analyzing T. pallidum lipoproteins is a fruitful avenue of investigation is exemplified in at least three recent studies on Tp47, Tp32, and PnrA (20-22).

The current structural study focuses on the Tp34 membrane lipoprotein (also known as Tp0971) of T. pallidum. Early studies by Staggs et al. (23) denoted a protein, most likely Tp34, as having lactoferrin binding properties. Lactoferrin is a mammalian iron-binding glycoprotein found on mucosal surfaces and within biological fluids, including milk, saliva, and seminal fluid (24-26). Despite the preliminary functional characterization of Tp34, NCBI Clusters of Orthologous Groups data base includes Tp34 with five other bacterial proteins as COG3470, which is denoted as "uncharacterized proteins probably involved in high affinity Fe²⁺ transport." However, the structural and functional aspects of this group of proteins remain undetermined. To investigate further the potential of Tp34 to serve as a ligand-binding protein (LBP), we determined the three-dimensional structure of Tp34 using x-ray crystallography. The relevance of our findings to the notion that Tp34 is involved in lactoferrin binding as well as the potential implications of this interaction for iron utilization by T. pallidum are discussed.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

 
Construction of His₆-tagged rTp34—To produce a nonlipidated, recombinant version of Tp34 (rTp34) in Escherichia coli, a fragment encoding amino acid residues 2-185 (cloned without the post-translationally modified N-terminal Cys residue) was amplified by PCR from treponemal genomic DNA using primer pairs encoding the predicted 5' and 3' termini. The PCR products were ligated into pProExHTa vector (Invitrogen) using BamHI and PstI restriction sites such that the resultant construct encoded a fusion protein with a His₆ tag at its N terminus; this His₆ tag region is cleavable by tobacco etch virus protease. Ligation mixtures were transformed into E. coli XL1-Blue (Stratagene), and the sequences of plasmids isolated from colonies that tested positive by restriction digest were verified using standard DNA sequencing techniques.

Expression and Purification of rTp34—E. coli XL1-Blue was grown at 37 °C in LB medium containing 100 µg/ml ampicillin until the cell density reached an A₆₀₀ of 0.6. The culture was then induced for 4 h with 600 µM isopropyl 1-thio--D-galactopyranoside. Cells derived from 1 liter of culture were harvested by centrifugation and lysed at room temperature with gentle rocking for 30 min using 50 ml of B-PER II (Pierce). The resulting suspension was centrifuged at 25,000 x g for 15 min to remove cell debris. rTp34 was isolated from the supernatant by nickel-nitrilotriacetic acid-agarose (Qiagen) using a standard method (22).

The N-terminal His₆ tag was removed via cleavage with tobacco etch virus protease (Invitrogen) at room temperature for 3 h, and then the protease was removed according to the manufacturer's protocol. Removal of the fusion segment was verified by SDS-PAGE. rTp34 was concentrated and bufferexchanged with 20 mM Hepes, 20 mM NaCl, pH 7.5 (buffer A) using an Amicon ultracentrifuge filter device (Millipore) with a 10,000 molecular weight exclusion limit. The concentrated protein was applied to a HiLoad 16/60 Superdex 75 prep grade column and purified on anÄ_kta fast performance liquid chromatography system (GE Healthcare) using buffer A. Subsequent to elution, peak fractions were analyzed by SDS-PAGE. At this stage, the protein was pure to apparent homogeneity (i.e. >95%). Fractions containing purified rTp34 were pooled and concentrated to 10 mg/ml in buffer A for protein crystallization. Protein concentration was determined spectrophotometrically using an absorption coefficient of 1.491 mg^-1ml^-1cm^-1 at 280 nm (27). To express protein for nuclear magnetic resonance (NMR) characterization, bacteria were grown in M9 minimal medium supplemented with ¹⁵NH₄Cl as the sole nitrogen source. The resultant ¹⁵N-labeled protein was purified as described above. For the production of a selenomethionine (SeMet) variant of rTp34, the expression vector was transformed into the E. coli methionine auxotroph B834 cells (Novagen) and grown in the presence of M9 medium containing 5% LB supplemented with 125 mg/liter each of adenine, uracil, thymine, and guanosine, 2.5 mg/liter thiamine, 4 mg/liter D-biotin, 20 mM glucose, 2 mM MgSO₄, 50 mg/liter each of 19 L amino acids (excluding methionine), and 50 mg/liter L-selenomethionine (Sigma). Labeled protein was prepared as described above with the addition of 15 mM -mercaptoethanol to the purification buffers.

NMR Spectroscopic Screening—Suitability of rTp34 for structure determination was evaluated by NMR screening of ¹⁵N-labeled protein (28). All ¹H,¹⁵N heteronuclear single quantum coherence (HSQC) spectra of labeled protein were acquired at 25 °C by using a Varian INOVA 500-MHz spectrometer with NMRPipe (29) used for data processing and NMRView (30) for analysis. Data were collected on 500 µM samples in 50 mM K₃PO₄, 50 mM NaCl, pH 7.5.

Preparation of PnrA, Tp32, Human Milk Lactoferrin (hLF), Bovine Milk Lactoferrin (bLF), and Human Serum Transferrin (hTF)—The purification procedures for PnrA and Tp32 have been described previously (21, 22). As a final step, both proteins were dialyzed against buffer B (20 mM Hepes, 100 mM NaCl, 2 mM octyl -glucoside, pH 7.5). Lyophilized hLF, bLF, and hTF were obtained from Sigma and reconstituted in buffer A at protein concentrations of 10 mg/ml. Proteins were then dialyzed overnight into buffer B. ApohLF was prepared in the same buffer after exhaustive dialysis against 0.1 M citric acid (31).

Crystallization, Data Collection, and Structure Determinations—Crystals were grown at 20 °C using the vapor diffusion method in hanging drop mode by mixing 4 µl of protein (10 mg/ml in buffer A) with 4 µl of reservoir solution (2.4 M ammonium sulfate, 0.1 M Bicine, pH 9.0) and equilibrating against 500 µl of reservoir solution. Crystals appeared within 24 h and grew to average dimensions of 0.2 x 0.6 x 0.1 mm in about 3 days. The crystals were cryo-protected in 2.8 M ammonium sulfate, 15% (v/v) ethylene glycol, 0.1 M Bicine, pH 9.0, and then flash-cooled in liquid propane. Crystals exhibited the symmetry of space group P2₁2₁2₁ with cell dimensions of a = 34 Å, b = 66 Å, c = 151 Å and contained two molecules per asymmetric unit and 41% solvent. Crystals for the selenomethionine variant of rTp34 (rTp34-SeMet) were produced under the same conditions.

Diffraction data were collected at beamlines 19-ID and 19-BM (SBC-CAT) at the Advanced Photon Source (Argonne National Laboratory, Argonne, IL) and processed with HKL2000 (32). Crystals diffracted to a minimum Bragg spacing (d_min) of about 1.63 Å.

Phases for rTp34-SeMet were obtained from a single anomalous diffraction experiment using x-rays with an energy near the selenium K absorption edge. Using data to 2.5 Å, 6 of 12 possible selenium sites were located and refined with the program CNS (33), resulting in a figure of merit of 0.35. Phases were extended to 1.7 Å and improved by density modification (CNS), resulting in a final overall figure of merit of 0.92.

The resulting electron density map was of sufficient quality to automatically construct an initial model using the program ARP/wARP (34). Manual rebuilding was carried out with the program Coot (35). Refinement was accomplished using the program REFMAC5 (36) from the CCP4 package (37).

The rTp34-SeMet model was used as a starting model for the refinement of native rTp34. The electron density revealed strong, spherical features near His-124 in each monomer, indicative of metal ions. These features were not present in rTp34-SeMet. An anomalous difference Fourier electron-density map showed peaks of about 6.4 and 4.7 that coincided with the features. Subsequently, a crystal of rTp34 was incubated with 10 mM zinc acetate (rTp34-Zn), and diffraction data were collected using x-rays with an energy near the zinc K absorption edge. The anomalous difference Fourier electron-density map revealed very strong peaks of about 32 and 33 at the previously observed positions. Zinc ions were, therefore, included in the model. An additional six zinc ions were located, but they had rather high B values and are, thus, minor sites. Although the nature of the metal ions in native rTp34 was not unambiguously clear, the peaks were interpreted as zinc ions at this stage. These metal-binding sites were not fully occupied as indicated by the comparatively weak electron density and high B factors. Furthermore, a second conformation appeared for some of the zinc-ligand residues, whereas these residues adopted only a single conformation in rTp34-Zn, where the main zinc sites were presumably fully occupied. As a control, a dataset was collected from a crystal of rTp34 that was grown in the presence of 1 mM EDTA (rTp34-EDTA). Density observed at the zinc sites was weak and was interpreted as water molecules.

Final models had an R_work and R_free of 18.4 and 21.5% for rTp34-SeMet, 18.8 and 23.5% for native rTp34, 19.0 and 23.3% for rTp34-Zn, and 18.9 and 22.0% for rTp34-EDTA, respectively. About 31 N-terminal residues could not be located in all cases.

Figure Generation—Figures that feature structural details of rTp34 were generated using PyMOL. They were rendered using POV-Ray.

Isothermal Titration Calorimetry—The binding of rTp34 or other T. pallidum lipoproteins to various target proteins was analyzed by injecting a solution of each lipoprotein into a solution of target protein and monitoring the resultant heat changes upon binding using a VP-ITC titration microcalorimeter (MicroCal, Inc.). All proteins were exhaustively exchanged into buffer B before the titration experiment. Protein concentrations were obtained by measuring the absorbance of proteincontaining solutions spectrophotometrically followed by the calculation of the protein concentration using Beer's Law (A = cl, where A is absorbance, is the molar extinction coefficient, c is the protein concentration, and l is the path length). In the case of rTp34, hLF, and bLF, the extinction coefficients were experimentally determined following the method of Pace et al. (27). For other proteins was calculated by the ProtParam tool available at the ExPASy Proteomics Server. The titration was carried out using 31 serial injections of 8 µl of lipoprotein at concentrations of 500-600 µM into a stirred reaction cell (1.4 ml) containing solutions of the target protein at 30 µM.To study the potential binding of hTF to rTp34, a solution of hTF (126 µM) was injected into a solution of rTp34 (13.6 µM) as described above. The heats of interaction resulting from the titration were monitored by the VP-ITC instrument until the target protein was saturated with the lipoprotein. Integration of the resultant thermogram yielded the binding isotherm, which was analyzed using the Microcal-ORIGIN software package to yield the association constants and the thermodynamic parameters associated with binding. Titrations using rTp34 as the injectant and hLF as the target protein were carried out in triplicate; others were performed only once.

Analytical Ultracentrifugation—Sedimentation velocity experiments were performed in a Beckman XL-I analytical ultracentrifuge. For rTp34 sedimentation, samples of 390 µl of reference buffer C (20 mM Tris-Cl, 20 mM NaCl, pH 7.5) or rTp34 (19 µM) diluted in reference buffer were loaded into a dualsector charcoal-filled epon centerpiece. In experiments that contained divalent metal ions, they were included at a concentration of 3 mM unless otherwise noted. The samples were centrifuged at 50,000 rpm in an An60-Ti rotor, and sedimentation was monitored using either laser interferometry or absorbance spectrophotometry at a wavelength of 280 nm. Data were analyzed using the program SEDFIT, which generates a continuous c(s) distribution for the sedimenting species (38). For experiments designed to monitor the binding of rTp34 to another protein, the same buffer was used. The volumes, rotor, centrifugation speed, and instrumentation were as described above. Three simultaneous sedimentation experiments were performed in these cases, with the solution contents being (i) rTp34 (15 µM) alone, (ii) target protein (5 µM) alone, and (iii) a mixture of rTp34 (7.5 µM) and the target protein (2.5 µM). For such experiments, two signals from the proteins were concurrently used to follow their sedimentation. For some experiments, the two signals were absorbance at 280 nm and refractive index detection by laser interferometry; for others, the signals were absorbances at 280 and at 250 nm. Such data were analyzed according to the multisignal c_k(s) methodology of Balbo et al. (39), as implemented in the program SEDPHAT. Briefly, the sedimentation profiles obtained from experiments i and ii were used to calculate the unknown extinction coefficient of the protein at 250 nm or the unknown interferometric extinction coefficient. This is possible because the extinction coefficient of the proteins at 280 nm is known. Once the extinction coefficient at both signals is known, a characteristic signal increment _k of each protein component k at each signal can be defined that allows the spectral decomposition of the c(s) distribution into two c_k(s) distributions, one for each of the protein components. Integration of a c_k(s) peak yields the concentration of a given component in that peak. The concentrations of co-sedimenting proteins may, thus, be derived by analyzing the two c_k(s) distributions; comparison of the two concentrations yields the stoichiometry of the proteins in that peak. The program SEDNTERP was used to estimate the partial specific volume of the proteins as well as the density and viscosity of the buffer solutions. All sedimentation coefficients quoted in this study have been corrected to values that would be observed under standard conditions (i.e. s_20,w).

Cultivation and Isolation of Treponemes—T. pallidum subspecies pallidum (Nichols strain) was maintained and passaged by intratesticular inoculation of adult male New Zealand White rabbits as previously described (40). Spirochetes were harvested in T. pallidum medium (41) and flushed with a reduced oxygen atmosphere of 3% O₂, 5%CO₂, and the balance of nitrogen. The medium consisted of Earle's balanced salt solution, minimum essential medium (MEM) (Hyclone) amino acids, MEM nonessential amino acids, MEM vitamin, 2 mM L-glutamine, 550 µM D-mannitol, 320 µM L-histidine, 24 mM sodium bicarbonate, 25 mM MOPS, 900 µM sodium pyruvate, 14 mM D-(+)glucose, 650 µM dithiothreitol, and 20% (v/v) fetal bovine serum (Mediatech, Inc.) with a final pH of 7.5. Rabbit testicular debris was removed from treponemal suspensions by two successive rounds of slow-speed centrifugation (200 x g for 8 min). Spirochetes remaining suspended were enumerated by darkfield microscopy and diluted to a concentration of 10⁸ ml^-1 in T. pallidum medium.

Indirect Immunofluorescence Detection of Antigens in T. pallidum—The agarose gel microdroplet method (42) was utilized to assess whether lipoproteins were exposed on the surface of T. pallidum. T. pallidum was encapsulated within agarose gel microdroplets as previously described (42). The encapsulation method stabilizes the fragile outer membrane of T. pallidum during antibody exposure and washing treatments. Mouse monoclonal anti-Tp47 (clone 11E3) (43) and mouse monoclonal anti-Tp34 (44) were diluted 1:20 and added directly to 1 ml of aliquots of microdroplets with or without 0.15% (v/v) Triton X-100. The microdroplets were incubated overnight at 4 °C. The microdroplets were collected by centrifugation at (500 x g, 5 min) and washed 5 times with fresh T. pallidum medium. Two µg of goat anti-mouse Alexa Fluor 488 (Molecular Probes) were added to the microdroplets for 2 h before being re-washed 5 times and viewed on glass slides with a Olympus BH2 microscope (darkfield lens and BP490 (green) filter (Olympus America Inc.)). Images were obtained using a SPOT Pursuit camera (Diagnostic Instruments, Inc.). Slides were prepared from each of 3 independent microdroplet preparations, and 100 bacteria were counted per slide. Any spirochete with a fluorescent signal, which was often punctated, was considered positive and compared with the total number of spirochetes observed via darkfield microscopy.

SDS-PAGE and Immunoblotting—Samples for protein analysis were separated on 12.5% polyacrylamide resolving gels. Proteins were transferred electrophoretically to a 0.45-µm-pore-size nitrocellulose filter (Schleicher & Schuell) for immunoblotting. Immunoblots were incubated with a 1:100 dilution of mouse monoclonal antibody 4A4. This was followed by incubations with a dilution of 1:10,000 of goat anti-mouse immunoglobulin G (heavy and light chain-specific) peroxidase conjugates (Jackson ImmunoResearch). Immunoblots were colorimetrically developed with 4-chloro-1-naphthol as the substrate.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

 
Production of rTp34 and Screening by ¹⁵N HSQC NMR Spectroscopy—Recombinant Tp34 was created as a nonacylated fusion protein with an N-terminal His₆ tag followed by a tobacco etch virus protease cleavage site. The expression of this fusion protein in E. coli resulted in the production of non-lipidated, soluble protein. Typical yields were 20-25 mg of protein per liter of culture (>95% pure protein). Unless otherwise noted, hereafter "rTp34" shall refer to the recombinant Tp34 that has had its N-terminal His tag removed by the action of tobacco etch virus protease. To assess the suitability of rTp34 for structure determination, ¹⁵N-labeled rTp34 was subjected to ¹H,¹⁵N HSQC NMR spectroscopy. HSQC spectra of ¹⁵N-labeled rTp34 revealed the appropriate number of peaks according to the protein sequence with good peak dispersion and roughly equal intensity (data not shown). These results indicated that rTp34 was well folded, conformationally homogeneous, and, thus, likely suitable for crystallization.

The Crystal Structure of rTp34—The crystal structure of rTp34 was determined using single-wavelength anomalous diffraction of a crystal containing selenomethionyl-substituted protein (Table 1). The native structure (Fig. 1) was refined using data to a resolution of 1.9 Å. Residues 2-27 of rTp34 were not visible in electron-density maps. The asymmetric unit of rTp34 crystals contained two rTp34 monomers. Monomeric rTp34 contains mainly two antiparallel -sheets that are packed on one another (Fig. 1A). The structure and topology of these -sheets is similar to the classical immunoglobulin fold (Ig-fold (45)). Accordingly, the topological nomenclature of Ig-folds has been adopted for rTp34, with the -strands consecutively lettered A, B, C, C', D, E, F, and G; the -sheets shall hereafter be referred to as the "front" and "back" -sheets according to their orientation in Fig. 1A. The fold of rTp34 is in the "hybrid" subcategory of Ig-folds, inasmuch as strands C' and D appear to form a single, interrupted -strand, but C' participates in the front -sheet, whereas D is part of the back -sheet.

A dimer of rTp34 (Fig. 1B) is generated by a 2-fold rotation axis relating the two monomers present in the asymmetric unit of the rTp34 crystals. The total surface area buried at the dimer interface is 3300 Ų, with about 18% of the total surface area of a monomer contributing to the dimer interface. The two monomers are structurally very similar; a superposition of the two results in a root mean square deviation of 0.4 Šbetween the 157 equivalent C atoms. In the dimer, the aforementioned -strands ab and fg extend the back -sheet by two strands; interdigitated between -strands D and ab of one monomer is -strand fg of the neighboring molecule. The antiparallel hydrogen-bonding pattern of the back -sheet is broken by the parallel interaction between -strands D and fg.

In electron-density maps derived from a native structure of rTp34 determined at 1.9 Å (Table 1), density (not shown) is present near to the side chains of residues His-70, Glu-72, Met-117, His-124, and His-155' (the ' denotes that this residue is from the neighboring molecule in the dimer; see Fig. 1B). The identity of the atom in this density was assigned as the divalent metal cation Zn²⁺, based on several lines of evidence. First, all atoms near to the cation are chemically compatible ligands of Zn²⁺, and the ligation geometry is commensurate with the fact that Zn²⁺ may have a coordination number of four, five, or six. In addition, a crystal of rTp34 was soaked with 10 mM zinc acetate. Diffraction data were collected from this crystal at an x-ray wavelength near to the K absorption edge for zinc; any zinc atoms present in the structure would, therefore, scatter such x-rays anomalously. A 33- peak is present at the putative metal ion-binding site in an anomalous difference Fourier map calculated from these data, demonstrating bound Zn²⁺ (Fig. 2A). These data were also used to refine a structure of rTp34 with Zn²⁺ included in the model (Fig. 2A; rTp34-Zn in Table 1); the atom at the metal ion-binding site refined well as a Zn²⁺, with B factors that were comparable with those of the surrounding atoms. As a final piece of evidence, diffraction data were collected from a crystal of rTp34 that was grown in the presence of the divalent metal ion chelator EDTA. Because these crystals are sufficiently isomorphous to the native crystal, an F_o(native) - F_o(EDTA) difference Fourier map could be calculated. This map has an 8- peak present at the metal ion-binding site (Fig. 2A), indicating that density present there in the native structure is not present when free divalent metal ions are sequestered from solution by EDTA. The structure derived from these data (rTp34-EDTA in Table 1), refined using data to a resolution of 1.6 Å, exhibits almost no electron density at the metal ion-binding site, and the liganding side chains in general move away from the site (not shown). All of these crystallographic data indicate that there is a Zn²⁺-binding site present at the dimeric interface of rTp34 and support the assignment of the aforementioned electron density in the native rTp34 structure as Zn²⁺. Because two monomers contribute to this zinc-binding site, it is a "protein interface zinc site," as characterized by Auld (46).

The B factors of the Zn²⁺ ions in the native structure are higher than those of the surrounding atoms, indicating that the cations may have a lower occupancy. Another factor that indicates low occupancy of the Zn²⁺ in the native structure is that the side chain of His-124 adopts multiple conformations (not shown), but in the Zn²⁺-soaked structure, the side chain is locked into a single, Zn²⁺-proximal conformation (Fig. 2A). The native crystals of rTp34 were not grown in the presence of added metal ions; any such ions present in this structure must have co-purified with rTp34 (from E. coli) or been scavenged from the crystallization medium. This fact may account for the apparently low occupancy of the metal ions in the rTp34 native structure. It also indicates a high affinity for metal ions by the observed interfacial metal-binding sites.

View larger version (55K): [in this window] [in a new window]   FIGURE 1. The crystal structure of rTp34. A, monomeric structure. A ribbon model of the crystal structure is shown with secondary structural elements colored purple for -strands, green for the 310 helix, and blue for regions with no regular secondary structure. Letters of the elements are indicated, as are the N and C termini of the model. B, dimeric structure. The bottom monomer is colored as in A and is rotated about 90° to the left around the vertical axis. The other monomer is colored tan. Residues that form inner-sphere contacts with the two interdimeric Zn2+ ions (orange spheres) are shown as balls and sticks and are colored according to the monomer from which they originate; purple for the bottom monomer, and tan for the top. The liganding interactions are shown as black dashed lines.

View larger version (25K): [in this window] [in a new window]   FIGURE 2. Zn2+ binding by rTp34. A, a stereo diagram of the Zn2+ binding site. The zinc(II) ion and side chains are colored as described in Fig. 1 except nitrogen atoms are blue, oxygen atoms are red, and the sulfur atom is yellow. Superposed on this final model are three electron density maps, a 1.9-Å 2Fo - Fc map from the rTp34-Zn structure contoured at the 1- level (blue), an anomalous difference map from the rTp34-Zn data calculated using reflections to 2.5 Å and contoured at the 20- level (green), and an Fo(native) - Fo(rTp34-EDTA) map using reflections to 1.7 Å and contoured at the 5- level (red). B, further details of the binding site. Distances are given in Å.

View larger version (13K): [in this window] [in a new window]   FIGURE 3. Solution properties of rTp34. A, rTp34 c(s) distributions. Shown are c(s) distributions for three different concentrations of rTp34. Black is 5 µM, red is 10 µM, and cyan is 50 µM. Note that the cyan curve has a different y axis, shown on the right of the plot. B, the effect of metal ions on the sedimentation. The rTp34 protein was sedimented in the presence of 20 mM NaCl (black distribution), 3 mM NiSO4 (green), 300 µM ZnCl2 (gray), 3 mM MgCl2 (orange), or 100 mM NaCl (cyan). The resultant c(s) profiles are shown. AU, absorbance units.

The c(s) protocol used fits a single frictional coefficient ratio (f/f₀) for the entire range of sedimentation values considered; most compact, globular proteins have a value of about 1.2 for f/f₀. In these experiments, f/f₀ was about 1.5, which indicates that rTp34 has a moderately elongated structure in solution. Such a ratio comports with the known disorder and presumed flexibility of the N terminus of rTp34. Large f/f₀ values have been observed in proteins with disordered and apparently flexible N termini (e.g. see Ref. 52).

The presence of Zn²⁺ at the dimeric interface of rTp34 (Figs. 1B and 2) prompted the examination of the solution behavior of rTp34 in the presence of divalent metal ions. SV experiments demonstrated that the dimeric form of rTp34 is indeed stabilized when divalent cations are included in solution (Fig. 3B). All divalent metal ions tested, i.e. Mg²⁺, Ca²⁺, Co²⁺, Ni²⁺, Mn²⁺, Cu²⁺, and Zn²⁺, displayed an ability to stabilize the dimeric form of the protein relative to the monomeric form at a concentration of 3 mM (300 µM for Zn²⁺, pH 7.5, and Cu²⁺, pH 5.5, because higher concentrations caused the precipitation of rTp34; see Table 2). To examine the ability of iron to dimerize rTp34, an SV experiment was conducted in the presence of ferric citrate (i.e. chelated Fe³⁺). Inclusion of ferric citrate caused 9% of the monomers in solution to form dimers (Table 2). SV experiments in the presence of unchelated Fe²⁺ and Fe³⁺ were attempted at pH 5.5 because of the low solubility of those cations at pH 7.5. SV in the presence of Fe²⁺ exhibited artifacts that were possibly because of oxidation of the cation during the course of the sedimentation. Inclusion of Fe³⁺ in the buffer caused rTp34 to aggregate (not shown), whereas Zn²⁺, Ni²⁺, and Cu²⁺ encouraged dimerization at pH 5.5 (Table 2). It is noteworthy that lowering the pH to 5.5 from 7.5 had a deleterious effect on the ability of most metal ions to induce dimerization of rTp34 (Table 2). The lone exception is Cu²⁺; even at only 300 µM concentration, this metal ion induced the complete dimerization of rTp34 (Table 2).

The inclusion of more sodium ions (100 mM) caused about 60% of the protein to form dimers (Fig. 3B; Table 2). It is not known whether this latter effect is due to Na⁺ binding at the metal ion-binding site or to a more generalized effect whereby mutual repulsion of the rTp34 monomers is overcome by the increased ionic strength of the solution. Given that the normal concentration of sodium and chloride ions in human plasma is above 100 mM, it is plausible that Tp34, a lipoprotein believed to be periplasmic in T. pallidum (see below), is mostly dimeric in vivo. The highly flexible N-terminal portion of the protein would allow the dimerization to occur even when the protein ostensibly is tethered to the inner membrane of T. pallidum by its acyl moiety, which occurs at Cys-1 in the native protein. Thus, rTp34 appears to be a metal ion-stabilized dimer in vitro and probably in vivo as well.

It is important to note that the SV assay does not directly measure metal ion binding; it simply measures the efficacy of metal ions in inducing the dimerization of rTp34. However, two pieces of evidence indicate that the observed dimerization phenomenon is a direct result of the cations binding at the known metal ion site in rTp34. First, the metal ion-binding site occurs at the dimeric interface of rTp34, with residues from both monomers participating in metal ligation (Figs. 1B and 2). Second, as noted above, lowering the pH has a negative effect on the efficiency with which metal ions induce dimerization (Table 2). For example, in the presence of Co²⁺, c(s) distributions show a significant peak for dimeric rTp34 at pH 7.5 but almost none at pH 5.5. This indicates that an ionizable group must contribute to the metal ion-binding site. The pK_a of His is about 6.5, and 3 His residues participate in the interdimeric metal ion-binding site of rTp34 (Fig. 2). To approximate the pK_a values of these histidines in the context of rTp34, the structure of monomeric, metal-free rTp34 was subjected to electrostatic calculations (89). These calculations were performed on both monomers of the rTp34-EDTA structure independently, and the two pK_a values thus obtained for each His were averaged. The results demonstrate that the pK_a values of His-70 (7.3) and His-155 (7.7) are probably titratable in the pH range used for the SV experiments (5.5-7.5). The tendency for less metal-induced dimerization at pH 5.5, thus, buttresses the notion that the observed metal ion-binding site is responsible for the metal ion-induced dimerization of rTp34.

Lactoferrin Binding—Previous reports (23, 53) had indicated that Tp34 associates with hLF, a protein that serves as a ferric ion carrier within mammalian mucosal tissues. Isothermal titration calorimetry (ITC) and analytical ultracentrifugation were used to verify the previous reports of Tp34 binding to hLF. The ITC experiments showed that rTp34 specifically and avidly bound to iron-free hLF (apo-hLF; Fig. 4A). The thermogram obtained from this experiment is complex; it clearly displays two different modes of binding. The first mode is endothermic and starts to saturate before the second, which is exothermic. Control experiments established that the heat of dilution of rTp34 is minimal. Satisfactory fits to the isotherm could be obtained using a two-independent-site model (Fig. 4A) or a sequential binding model (not shown). Fitting the data to the former model, the dissociation constant obtained for the endothermic binding site is 66 ± 7nM; for the second, exothermic site, it is 540 ± 10 nM. For both sites, the stoichiometry is 1:1; therefore, the overall stoichiometry of binding is two rTp34 proteins per apo-hLF molecule. Significantly, the structure of hLF is bilobed, with both lobes exhibiting very similar structures (54). Further experiments are warranted to examine the facile interpretation that one rTp34 monomer binds one lobe of hLF, resulting in the observed 2:1 stoichiometry. Titrations carried out using iron-loaded hLF (Fe-hLF) were qualitatively similar to those using apo-hLF (not shown). Notably, the buffer in which these experiments were conducted contained 100 mM NaCl, which enables rTp34 dimerization (Fig. 3B). The potential, thus, exists for multiple heat-producing reactions to occur upon injection of concentrated solutions of rTp34 into apo-hLF, e.g. dimer dissociation, monomer association, dimer-apo-hLF association, and monomer-apo-hLF association. Given the complexity of this system, proposing a definitive model for the rTp34-hLF interaction must await further biochemical and structural characterization of these proteins.

The rTp34-hLF interaction was also studied using SV. Alone, apo-hLF exhibited a sedimentation coefficient of 5.3 S, which is in agreement with its status as a monomer in solution (Fig. 5A). Upon the addition of a 3-fold molar excess of rTp34, this peak in the c(s) profile is fully shifted to 6.3 S. The calculated molecular mass of the 6.3 S peak (133,000 Da) is consistent with 2 molecules of rTp34 (23,000 Da) binding to one molecules of apohLF (76,000 Da). By using multi-signal c_k(s) analysis (39), the approximate concentrations of the individual components making up the 6.3 S peak can be calculated (Fig. 5B). The calculated concentration of rTp34 in the 6.3 S peak is 5.7 µM, whereas it is 2.7 µM for apohLF. The calculated molar ratio of rTp34:hLF is, therefore, 2.1:1. Hence, two independent biophysical methods show that two molecules of rTp34 bind to one molecule of apohLF. Such experiments were also performed to assess the interaction between Fe-hLF and rTp34. Similar results to those presented above were obtained (not shown).

View larger version (30K): [in this window] [in a new window]   FIGURE 4. Calorimetric detection of rTp34 binding to hLF. The upper panels of A and B and all of C show the power (y axis) that the calorimeter applies to the reaction cell at a given time (x axis) to maintain the cell at a constant temperature. The lower panels in A and B show the binding isotherm. Black squares represent the area under each peak in the upper panels, and the line through them represents the calculated fit of the binding model to the data. A, rTp34 binding to apo-hLF. A two-site binding model was used for the fit. B, rTp34 binding to apo-bLF. The data were fit to a one-site binding model. C, control experiments. Shown are the time course of the titration of rTp32 (green), PnrA (red), and hTF (black). The y axis was kept on approximately the same scale as those in A and B for facile comparison.

Given that the rTp34 dimer is stabilized by divalent metal ions, experiments were conducted to study the rTp34-apo-hLF interaction with such cations present in solution. Unfortunately, at the high protein concentrations necessary for ITC, both proteins were subject to precipitation and other aberrant behaviors in the presence of Zn²⁺. Nonetheless, SV could be employed at lower protein concentrations and with 300 µM Zn²⁺. The results (supplemental Fig. 1) demonstrated that rTp34 binds to apo-hLF with a stoichiometry of 2:1, even in the presence of the divalent metal ion, implying that dimeric rTp34 binds to apohLF. Further experiments are planned to explore the dimeric rTp34:apohLF association.

Interestingly, an earlier report demonstrated that bLF bound to rTp34 (23). We also found that apobLF bound to rTp34, but the stoichiometry was 1:1, and only the endothermic phase of the binding was observed with a dissociation constant of 630 nM (Fig. 4B). Because bLF and hLF are structurally similar (55), it is not surprising that one of the binding sites is preserved between the two proteins. The 1:1 stoichiometry of the rTp34-bLF interaction could illustrate a lack of conservation of the second exothermic binding site.

Several control experiments were performed to confirm the specificity of the observed rTp34-hLF interaction. We first examined whether hTF, a serum iron-transport protein with structural homology to hLF (56), could bind to rTp34. In neither ITC (Fig. 4C) nor analytical ultracentrifugation experiments (not shown) could an interaction between hTF and rTp34 be demonstrated. The lack of an rTp34:hTF interaction is in accord with earlier results demonstrating that hTF does not interact with proteins from T. pallidum (23, 53). Other T. pallidum lipoproteins (PnrA and rTp32) failed to bind to hLF specifically in the ITC assay (Fig. 4C). From these control experiments, we conclude that the avid binding of rTp34 to hLF is specific.

The simplest deduction from the combined data is that rTp34 is a treponemal receptor for hLF. As such, it may be responsible for iron uptake by T. pallidum. Other pathogenic bacterial species are known to utilize hLF to acquire iron from their mammalian hosts. Specifically, under conditions of iron starvation, certain Neisseria express outer-membrane hLF receptors (LbpAs) that specifically bind Fe-hLF (57). Once the LbpAs bind to Fe-hLF, accessory proteins are recruited to the complex, and ferric iron is stripped from Fe-hLF. The cations are then transported into the periplasm. The process is dependent on energy transduction via a TonB system present in the bacterial inner membrane. There are several facts that belie the operation of a Neisseria-like iron import system in T. pallidum. LbpA, an integral protein of the outer membrane, has no detectable sequence homology to Tp34, a putative periplasmic lipoprotein (see below). Also, there is no evidence that the T. pallidum genome encodes the requisite accessory proteins akin to those used for iron transport into the periplasm of Neisseria.If T. pallidum uses rTp34 as a receptor for Fe-hLF with the purpose of acquiring iron, it must carry this out via an uncharacterized mechanism.

View larger version (18K): [in this window] [in a new window]   FIGURE 5. SV studies of the rTp34-hLF interaction. A, the c(s) profiles of two separate experiments are shown. In one (black line), only hLF was sedimented, and it sedimented as a monomer. The second experiment (gray line) was performed with a 3-fold excess of rTp34. The hLF peak was completely shifted by about 1 S. B, shown are the spectrally decomposed ck(s) distributions that correspond to the experiment represented by the gray line in A. The black line denotes the signal present from hLF. The gray line denotes the signal for rTp34. The rTp34 protein is present in the 6.2 S peak at twice the concentration of hLF.

Sequence analysis of complete genomes has allowed the placement of Tp34 into a family of sequentially similar proteins called HBG261060 (see the Hogenprot data base (60)). The only annotation for this family is "membrane antigen precursor." A representation of the sequence comparison among these proteins is found in Fig. 7. The overall sequence similarity of these proteins is 78%. This family contains proteins from 23 bacterial species. Many of these species are known human pathogens, including T. pallidum, Yersinia pestis, and Bordetella pertussis. The residues whose side chains contact Zn²⁺, as detailed above, are identical in all members of this family (Fig. 7, white bars), suggesting that the metal-binding site is important for the proper functioning of the HGB261060 proteins. Also, residues interacting with a neighboring monomer at the dimer interface tend to be in conserved regions of the family; the mean similarity of the residues at the interface is 91% (Fig. 7, dark gray bars). Such a conservation pattern implies that the other proteins in the family are also capable of forming dimers. The high sequence conservation of the family allows for the hypothesis that all of these proteins may bind to lactoferrin-like proteins in their respective hosts. In support of this notion is the fact that 90% of the bacterial species that express HBG261060 family members are mammalian pathogens or commensals.

Membrane Topology Assessment of Tp34 in T. pallidum—That rTp34 contained metal ions and bound the iron-sequestering protein hLF in receptor-like fashion(s) implied that native Tp34 might be exposed on the outer membrane of T. pallidum, where it would be accessible to such nutrients within human body fluids. However, despite decades of intensive efforts and some claims of success, not a single protein has yet been shown definitively to be located on the outer leaflet of the T. pallidum outer membrane (7, 8). A previous study by Cox et al. (42) using polyclonal antibodies provided initial evidence that Tp34 was not located on the surface of the T. pallidum outer membrane. In that study (42), an agarose gel microdroplet method, the current reference method for assessing membrane surface topology in the pathogenic spirochetes T. pallidum and Borrelia burgdorferi (61) was utilized. Encapsulation of spirochetes in agarose gel microdroplets protects the fragile T. pallidum outer membrane from physical disruption (during reactivity with antibodies) by providing a supporting matrix but one that is sufficiently porous to allow reagent (e.g. antibody probes, secondary antibody conjugates) access and removal (42, 61). The location of proteins beneath the surface is inferred by observing immunofluorescence of the bacteria upon treatment of encapsulated T. pallidum with low levels of Triton X-100 (0.15%) (which partially disrupts outer membrane integrity) in the presence of the primary antibody probes (42, 61). In our study the anti-Tp34 monoclonal antibody (4A4) was used to further investigate Tp34 membrane topology using the gel micro-droplet technique. Of note, in immunoblots, monoclonal antibody 4A4 reacted with rTp34 as well as with a broader-migrating 34-kDa protein band in T. pallidum whole cell lysates (Fig. 8A); this broadly migrating pattern for Tp34 has been well documented (44, 62). For these studies, we also employed an antibody directed against Tp47, a subsurface penicillin binding lipoprotein of T. pallidum (63), as a known marker for subsurface localization. When monoclonal antibody 4A4 or anti-Tp47 antibodies were used to probe intact T. pallidum encapsulated within agarose gel microdroplets (in the absence of detergent), spirochetes were not labeled (Fig. 8B). In contrast, when spirochetes were treated with a low concentration of Triton X-100, they fluoresced when probed with the same antibody preparations (Fig. 8B). Quantitatively, less than 6% of spirochetes fluoresced when exposed to anti-Tp34 or anti-Tp47 antibodies, whereas in the presence of Triton X-100, immunolabeling of treponemes was observed at 92 and 93% when reacted with anti-Tp47 or anti-Tp34 antibodies, respectively (Fig. 8C). These results as well as those previously reported by Cox et al. (42) suggest strongly that Tp34 is not surfaceexposed in T. pallidum.

View larger version (48K): [in this window] [in a new window]   FIGURE 6. Superposition of IL2R D1 to rTp34. A stereo diagram of a smoothed trace through the C atoms of IL2R D1 (blue) and rTp34 (red) is shown. The orientation of rTp34 is similar to that shown in Fig. 1A.

View larger version (26K): [in this window] [in a new window]   FIGURE 7. Sequence similarity among members of the HBG261060 family of proteins. The x axis is the sequence number of rTp34. The sequence was aligned to the 22 other members of the family. The bars (y axis) represent the percentage of family members that have an amino acid residue similar to the predominant residue at that position. Residues were classified into four groups: hydrophobic, polar, acidic, or basic. Dark gray bars show the positions of residues that are involved in interdimeric contacts in rTp34. Only residues that contacted the other monomer in both copies of rTp34 in the asymmetric unit were included in the dimericcontact analysis. White, annotated bars are residues that are inner-sphere ligands to the bound Zn2+ ions in the rTp34-Zn structure.

View larger version (26K): [in this window] [in a new window]   FIGURE 8. Membrane topology assessment of Tp34 in T. pallidum. A, immunoblot (using monoclonal antibody 4A4) of either 1 µg of rTp34 or 2.5 x 107 T. pallidum whole cells. B, T. pallidum bacteria encapsulated in agarose gel microdroplets were exposed to either a mouse monoclonal antibodies anti-Tp34 (4A4) or anti-Tp47 (11E3) in the presence or absence of 0.15% (v/v) Triton X-100 (TX). After secondary antibody exposure (anti-mouse Alexa Fluor 488), spirochetes were examined by either darkfield (DF) or fluorescent microscopy. C, the number of fluorescent treponemes were quantified from at least three independent microdroplet preparations and plotted. The error bars represent S.D.

Although Cu²⁺ binding to rTp34 has not yet been directly observed, the high efficiency with which that metal induces the dimerization of rTp34 (Table 2) suggests that it binds to the interfacial metal ion site. The suitability of the rTp34 metal ion site (Fig. 2) as a copper binding motif is established by the fact that His nitrogens, Glu carboxylate oxygens, and sulfurs of Met are known ligands for copper ions (48). The physiological relevance of the putative Cu²⁺ binding to rTp34 is unknown; Cu²⁺ is thought to be almost wholly sequestered in vivo by metal-binding proteins (69), and no other treponemal proteins are known to require copper. However, examination of the genomes of organisms that carry tp34-like genes shows a striking correlation; in all cases but one (Chromobacter violaceum), an open reading frame within two genes of the tp34-like gene is a putative Fe²⁺/Pb²⁺ permease belonging the FTR1 family. This finding is significant because in yeast, FTR1 works in concert with a copper-containing protein (FET3) to import iron (70). There is, thus, the intriguing possibility that Tp34 acts either as a treponemal FTR1 accessory protein or as a coppersequestering protein that supplies the cation to another protein. The similarity of the metal-binding site in rTp34 to a treponemal iron-binding site in the superoxide reductase (Tp0823) of T. pallidum (71), however, suggests that Tp34 could be involved in binding and transporting iron.

The proteins of the Tp34-containing Hogenprot family 261060 are encoded by the genomes of 23 species of bacteria; in 9 of these organisms, the tp34-like genes are clustered with an ABC-type permease and ATPase. We classify the proteins from these nine organisms (Bifidobacterium longum, Erwinia carotovora, Pasteurella multocida, Treponema denticola, Wolinella succinogenes, Y. pestis KIM 5, Y. pestis CO-92, Y. pestis 91001, and Yersinia pseudotuberculosis) as belonging to subfamily 261060ABC. The genome of T. pallidum does not encode a member of 261060ABC, although that of the related spirochete T. denticola, does. ABC-type operons usually encode a protein (a periplasmic LBP) that binds the targeted ligand (72, 73). However, in the genomes of the members of subfamily 261060ABC, such a protein is not encoded near tp34-like genes. It can be speculated that Tp34 assumes the role of the binding protein for an ABC-type metal ion-transport system. Although Tp34 is not structurally homologous to LBPs, the two proteins apparently employ similar strategies for binding ligands. LBPs utilize a "clamshell" motion to enclose their ligands, effectively surrounding the ligand with protein. The Tp34 dimer is stabilized upon metal ion binding, and this event also shields the metal from solvent. That the genome of T. pallidum does not include either an ABC permease or an ATPase in proximity to the tp34 gene does not preclude, however, Tp34 from acting as a receptor for such a metal ion-transport system; the permease and ATPase may be encoded at distant genomic loci. In fact, in the case of the methionine transporter (Tp32) in T. pallidum, its hypothetical ATPase (tp0120) and permease (tp0119) are together but distant from tp32 (tp0821) (21, 74). Whether such non-operonic ABC transport systems are yet another unusual feature of T. pallidum remains to be more fully explored via additional functional studies on other treponemal ABC transport systems in this organism.

Although iron generally is an essential element for microorganisms, it is not readily available in most biological systems. This poses a formidable metabolic obstacle when a virulent bacterium, like T. pallidum, invades the human host. In humans, the amount of free iron is extremely low (10^-18 M) and usually insufficient to sustain bacterial growth (75). The majority of iron is found intracellularly as myoglobin, ferritin, or hemoglobin, and smaller quantities of extracellular iron are complexed with transferrin (TF) or lactoferrin (LF) (76-78). TF is present at very high levels (25-44 µM) in serum but is only a minor component (0.2-1.3 µM) of mucosal secretions. On the other hand, LF is present at very low levels in serum (3.8-8.7 nM) but is in relatively high concentrations (6-13 µM) (79) in the mucosa, sites where T. pallidum typically establishes its initial infection. Thus, in order for iron to be accessible to the bacterium, it must be extracted from the human host carrier molecules by specialized uptake mechanisms (77, 80, 81). The major role for iron is its involvement in enzymatic redox reactions, but iron in proteins can also play a structural role. T. pallidum lacks major iron-containing proteins typically found in other bacterial pathogens, including the oxidative defense enzymes superoxide dismutase, catalase, and peroxidase (65). However, the spirochete is believed to contain superoxide reductase (Tp0823) and rubredoxin (Tp0991), both of which are likely to participate in oxygen detoxification (50, 82-84). This is consistent with the prevailing view that T. pallidum is microaerophilic (1). Genome analysis has further revealed that there likely are at least nine T. pallidum proteins (Tp0053, Tp0080, Tp0735, Tp0823, Tp0842, Tp0939, Tp0991, Tp1008, and Tp1038) that require iron as a cofactor (50, 65, 71, 85). Thus, it is plausible that T. pallidum requires iron.

The potential role of Tp34 as a hLF receptor is intriguing in the light of the belief that T. pallidum lacks an outer membrane receptor for iron acquisition and probably does not produce siderophores to trap extracellular iron. Nonetheless, it is tempting to speculate that Tp34 is involved in iron acquisition via its propensity to bind lactoferrin. This, however, poses an interesting paradox. In early studies it was reported that a molecule likely to be Tp34 was capable of binding mammalian iron transport proteins (23). However, those findings were difficult to reconcile, given the fact that there was no evidence that Tp34 was on the surface of T. pallidum (where Tp34 could interact with lactoferrin). Our structural studies herein as well as the fact that T. pallidum likely utilizes iron now prompt a reevaluation of the earlier work of Staggs et al. (23). First, Tp34 contains a dimeric Ig-fold that is stabilized by divalent metal ions. The ability of the Ig-fold to serve as a receptor for a proteinaceous ligand is well known (e.g. see Refs. 59 and 86). Second, we have garnered additional biochemical evidence via ITC and SV that Tp34 indeed binds avidly to human lactoferrin. Third, although agarose gel microdroplet encapsulation and immunofluorescence microscopy indicated that Tp34 is not on the outer surface of T. pallidum, Hazlett et al. (87) recently provided compelling evidence that the outer membrane of T. pallidum can be transiently perturbed by TP0453, a membrane lipoprotein that has several membrane-interacting amphipathic -helices. In this scenario TP0453 is believed to be tethered via its acyl chains to the inner leaflet of the treponemal outer membrane whereby transient interactions of the -helices of the polypeptide with the inner leaflet of the outer membrane then perturbs the lipid bilayer. This transient disruption of outer membrane integrity, however, is not thought to be sufficient to allow large proteins, such as antibodies, to traverse the outer membrane. On the oth