Title: Crystal Structure of the Dimeric C-terminal Domain of TonB Reveals a Novel Fold
Abstract: The TonB-dependent complex of Gram-negative bacteria couples the inner membrane proton motive force to the active transport of iron·siderophore and vitamin B12 across the outer membrane. The structural basis of that process has not been described so far in full detail. The crystal structure of the C-terminal domain of TonB fromEscherichia coli has now been solved by multiwavelength anomalous diffraction and refined at 1.55-Å resolution, providing the first evidence that this region of TonB (residues 164–239) dimerizes. Moreover, the structure shows a novel architecture that has no structural homologs among any known proteins. The dimer of the C-terminal domain of TonB is cylinder-shaped with a length of 65 Å and a diameter of 25 Å. Each monomer contains three β strands and a single α helix. The two monomers are intertwined with each other, and all six β-strands of the dimer make a large antiparallel β-sheet. We propose a plausible model of binding of TonB to FhuA and FepA, two TonB-dependent outer-membrane receptors. The TonB-dependent complex of Gram-negative bacteria couples the inner membrane proton motive force to the active transport of iron·siderophore and vitamin B12 across the outer membrane. The structural basis of that process has not been described so far in full detail. The crystal structure of the C-terminal domain of TonB fromEscherichia coli has now been solved by multiwavelength anomalous diffraction and refined at 1.55-Å resolution, providing the first evidence that this region of TonB (residues 164–239) dimerizes. Moreover, the structure shows a novel architecture that has no structural homologs among any known proteins. The dimer of the C-terminal domain of TonB is cylinder-shaped with a length of 65 Å and a diameter of 25 Å. Each monomer contains three β strands and a single α helix. The two monomers are intertwined with each other, and all six β-strands of the dimer make a large antiparallel β-sheet. We propose a plausible model of binding of TonB to FhuA and FepA, two TonB-dependent outer-membrane receptors. outer membrane proton motive force 4-morpholinoethanesulfonic acid root mean square ion exchange chromatography immobilized metal ion affinity chromatography multiwavelength anomalous diffraction The outer membrane (OM1) of Gram-negative bacteria constitutes a permeability barrier, protecting the cell against a variety of toxic agents. The lipopolysaccharides located in the outer leaflet of the OM confer to the bacteria a polar and negatively charged surface, restricting the cellular uptake of toxic organic molecules and detergents such as bile salts, the detergents in the gut. However, although the OM is an effective protective barrier against harmful environmental components, it also represents an additional obstacle for the uptake of nutrients, which can be circumvented in three ways. While small hydrophilic nutrients (<600 Da) enter the periplasm by simple diffusion through porins in a non-selective manner (1Schirmer T. J. Struct. Biol. 1998; 121: 101-109Crossref PubMed Scopus (211) Google Scholar), larger molecules are taken up by pores with an internal binding site for the ligand (such as LamB) in a stereospecific and selective manner (2Ehrmann M. Ehrle R. Hofmann E. Boos W. Schlosser A. Mol. Microbiol. 1998; 29: 685-694Crossref PubMed Scopus (109) Google Scholar) and can subsequently enter the cytoplasm by a variety of transporters located in the inner membrane (3Buchanan S.K. Trends Biochem. Sci. 2001; 26: 3-6Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar). A few nutrients, notably iron and vitamin B12, need to be taken up into the periplasm against their concentration gradients. For this purpose, a complex consisting of TonB, ExbB, and ExbD couples the inner membrane proton motive force (pmf) to the active transport of iron siderophores and vitamin B12 across the OM through specialized porins. Recently, crystal structures were solved for two TonB-dependent receptors, FepA and FhuA (4Buchanan S.K. Smith B.S. Venkatramani L. Xia D. Esser L. Palnitkar M. Chakraborty R. van der Helm D. Deisenhofer J. Nat. Struct. Biol. 1999; 6: 56-63Crossref PubMed Scopus (492) Google Scholar, 5Ferguson A.D. Hofmann E. Coulton J.W. Diederichs K. Welte W. Science. 1998; 282: 2215-2220Crossref PubMed Scopus (670) Google Scholar, 6Locher K.P. Rees B. Koebnik R. Mitschler A. Moulinier L. Rosenbusch J.P. Moras D. Cell. 1998; 95: 771-778Abstract Full Text Full Text PDF PubMed Scopus (459) Google Scholar). Like all other known porins, they are β-barrels, but unlike the other porin structures they have much larger interiors, which are almost completely obscured by a protein domain sitting inside the barrel (termed the “cork” or “hatch region”), which is encoded within the N-terminal segment of either protein. Iron uptake into bacteria is initiated by the binding of the iron·siderophore complex to the high affinity OM receptor. The dissociation constant is around 100–200 nm (7Newton S.M. Allen J.S. Cao Z. Qi Z. Jiang X. Sprencel C. Igo J.D. Foster S.B. Payne M.A. Klebba P.E. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 4560-4565Crossref PubMed Scopus (52) Google Scholar, 8Nikaido H. Saier Jr., M.H. Science. 1992; 258: 936-942Crossref PubMed Scopus (162) Google Scholar). An electron spin resonance study (9Jiang X. Payne M.A. Cao Z. Foster S.B. Feix J.B. Newton S.M. Klebba P.E. Science. 1997; 276: 1261-1264Crossref PubMed Scopus (92) Google Scholar), later rationalized by three-dimensional structural models (4Buchanan S.K. Smith B.S. Venkatramani L. Xia D. Esser L. Palnitkar M. Chakraborty R. van der Helm D. Deisenhofer J. Nat. Struct. Biol. 1999; 6: 56-63Crossref PubMed Scopus (492) Google Scholar, 5Ferguson A.D. Hofmann E. Coulton J.W. Diederichs K. Welte W. Science. 1998; 282: 2215-2220Crossref PubMed Scopus (670) Google Scholar, 6Locher K.P. Rees B. Koebnik R. Mitschler A. Moulinier L. Rosenbusch J.P. Moras D. Cell. 1998; 95: 771-778Abstract Full Text Full Text PDF PubMed Scopus (459) Google Scholar), has shown that this event triggers conformational changes in the OM receptor. This might allow TonB to contact specific regions on the receptor. It appears that “energized” TonB is then able to deliver its energy to the receptor, resulting in ligand translocation into the periplasm (10Reynolds P.R. Mottur G.P. Bradbeer C. J. Biol. Chem. 1980; 255: 4313-4319Abstract Full Text PDF PubMed Google Scholar,11Wooldridge K.G. Morrissey J.A. Williams P.H. J. Gen. Microbiol. 1992; 138: 597-603Crossref PubMed Scopus (19) Google Scholar). ExbB·ExbD are implicated in the recycling of TonB, from its high affinity OM receptor association to a high affinity inner membrane association (12Larsen R.A. Thomas M.G. Postle K. Mol. Microbiol. 1999; 31: 1809-1824Crossref PubMed Scopus (155) Google Scholar, 13Letain T.E. Postle K. Mol. Microbiol. 1997; 24: 271-283Crossref PubMed Scopus (133) Google Scholar). The structural changes in this whole process have remained almost completely unclear. TonB of Escherichia coli is a protein consisting of 239 amino acids. Homologs of TonB have been found in several Gram-negative species (14Larsen R.A. Myers P.S. Skare J.T. Seachord C.L. Darveau R.P. Postle K. J. Bacteriol. 1996; 178: 1363-1373Crossref PubMed Google Scholar). The N terminus is in the cytoplasm; the protein is anchored in the inner membrane by its uncleaved N-terminal signal sequence (15Postle K. Skare J.T. J. Biol. Chem. 1988; 263: 11000-11007Abstract Full Text PDF PubMed Google Scholar, 16Karlsson M. Hannavy K. Higgins C.F. Mol. Microbiol. 1993; 8: 379-388Crossref PubMed Scopus (74) Google Scholar), and most of the protein extends into the periplasm. The membrane anchor sequence contains a set of highly conserved residues located on one face of the α-helix (SHLS motif). The sequence SXXXH (where X is any amino acid) has been defined as the minimal structural requirement for the coupling of TonB to the electrochemical gradient of the inner membrane (17Larsen R.A. Postle K. J. Biol. Chem. 2001; 276: 8111-8117Abstract Full Text Full Text PDF PubMed Scopus (49) Google Scholar). The amino acid sequence of TonB contains a long central region with a high percentage of proline residues between residues 70 and 102 (17%), which is thought to confer to TonB the conformational rigidity and extended shape necessary to span the periplasm, and thereby to allow the C-terminal domain to contact the receptor embedded in the OM (18Skare J.T. Ahmer B.M. Seachord C.L. Darveau R.P. Postle K. J. Biol. Chem. 1993; 268: 16302-16308Abstract Full Text PDF PubMed Google Scholar). Mutational studies have defined the last 48 residues as being essential to make contact with the OM receptor (19Larsen R.A. Foster-Hartnett D. McIntosh M.A. Postle K. J. Bacteriol. 1997; 179: 3213-3221Crossref PubMed Google Scholar). TonB forms a complex in the inner membrane with ExbB and ExbD (13Letain T.E. Postle K. Mol. Microbiol. 1997; 24: 271-283Crossref PubMed Scopus (133) Google Scholar), two membrane proteins that could potentially act as proton translocators. ExbB is homologous to the protein MotA, and ExbD has a similar topology as MotB, both of which are thought to exploit the proton gradient to drive the bacterial flagellum. ExbB has been proposed to modulate the conformation of TonB (20Larsen R.A. Thomas M.G. Wood G.E. Postle K. Mol. Microbiol. 1994; 13: 627-640Crossref PubMed Scopus (85) Google Scholar), as well as mediate its recycling (12Larsen R.A. Thomas M.G. Postle K. Mol. Microbiol. 1999; 31: 1809-1824Crossref PubMed Scopus (155) Google Scholar, 13Letain T.E. Postle K. Mol. Microbiol. 1997; 24: 271-283Crossref PubMed Scopus (133) Google Scholar), but it has remained an enigma as to what these structural changes might be. Cross-linking studies have suggested the regions through which TonB might interact with its partners in the inner membrane: The contact with ExbB is mediated by the signal anchor (20Larsen R.A. Thomas M.G. Wood G.E. Postle K. Mol. Microbiol. 1994; 13: 627-640Crossref PubMed Scopus (85) Google Scholar), whereas the residues responsible for the interaction with ExbD are unknown (21Higgs P.I. Myers P.S. Postle K. J. Bacteriol. 1998; 180: 6031-6038Crossref PubMed Google Scholar). TonB and its associated proteins ExbB·ExbD thus play the role of an energy-transducing complex, coupling the electrochemical proton gradient of the inner membrane to active import processes across the OM (13Letain T.E. Postle K. Mol. Microbiol. 1997; 24: 271-283Crossref PubMed Scopus (133) Google Scholar, 22Braun V. FEMS Microbiol. Rev. 1995; 16: 295-307Crossref PubMed Scopus (287) Google Scholar). The energy is provided by the proton motive force (10Reynolds P.R. Mottur G.P. Bradbeer C. J. Biol. Chem. 1980; 255: 4313-4319Abstract Full Text PDF PubMed Google Scholar, 23Hancock R.W. Braun V. J. Bacteriol. 1976; 125: 409-415Crossref PubMed Google Scholar,24Bradbeer C. J. Bacteriol. 1993; 175: 3146-3150Crossref PubMed Scopus (149) Google Scholar). For the transduction process to occur, the C-terminal domain of TonB must contact the OM receptor. Based on genetic (25Schramm E. Mende J. Braun V. Kamp R.M. J. Bacteriol. 1987; 169: 3350-3357Crossref PubMed Google Scholar, 26Gudmundsdottir A. Bell P.E. Lundrigan M.D. Bradbeer C. Kadner R.J. J. Bacteriol. 1989; 171: 6526-6533Crossref PubMed Scopus (88) Google Scholar), cross-linking (19Larsen R.A. Foster-Hartnett D. McIntosh M.A. Postle K. J. Bacteriol. 1997; 179: 3213-3221Crossref PubMed Google Scholar, 27Cadieux N. Kadner R.J. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 10673-10678Crossref PubMed Scopus (145) Google Scholar, 28Merianos H.J. Cadieux N. Lin C.H. Kadner R.J. Cafiso D.S. Nat. Struct. Biol. 2000; 7: 205-209Crossref PubMed Scopus (98) Google Scholar, 29Cadieux N. Bradbeer C. Kadner R.J. J. Bacteriol. 2000; 182: 5954-5961Crossref PubMed Scopus (72) Google Scholar), and affinity chromatography (30Moeck G.S. Coulton J.W. Postle K. J. Biol. Chem. 1997; 272: 28391-28397Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar) studies, a recognition site has been suggested on the receptor, the TonB box, a hydrophobic stretch of seven amino acids, which is highly conserved in all the TonB-dependent OM receptors (31Postle K. J. Bioenerg. Biomembr. 1993; 25: 591-601PubMed Google Scholar). A recent study resulted in the proposal that the conformation rather than the sequence of the TonB box is important for the recognition process between TonB and the receptor (19Larsen R.A. Foster-Hartnett D. McIntosh M.A. Postle K. J. Bacteriol. 1997; 179: 3213-3221Crossref PubMed Google Scholar). Moreover, it has been hypothesized that other regions of both interacting partners are also involved (27Cadieux N. Kadner R.J. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 10673-10678Crossref PubMed Scopus (145) Google Scholar). Most strikingly, TonB dependence is maintained if the complete cork domain is deleted (32Braun M. Killmann H. Braun V. Mol. Microbiol. 1999; 33: 1037-1049Crossref PubMed Scopus (81) Google Scholar, 33Scott D.C. Cao Z. Qi Z. Bauler M. Igo J.D. Newton S.M. Klebba P.E. J. Biol. Chem. 2001; 276: 13025-13033Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar), including the TonB box. It follows that the recognition site cannot be limited to the TonB box. A number of phages and colicins have exploited the TonB·ExbB·ExbD system for gaining entry into bacteria (34Lazdunski C.J. Bouveret E. Rigal A. Journet L. Lloubes R. Benedetti H. J. Bacteriol. 1998; 180: 4993-5002Crossref PubMed Google Scholar). A similar system, TolQRA, has also been described as allowing entry for other phages and colicins (34Lazdunski C.J. Bouveret E. Rigal A. Journet L. Lloubes R. Benedetti H. J. Bacteriol. 1998; 180: 4993-5002Crossref PubMed Google Scholar). The cellular function of the TolQRA system has remained enigmatic, and its deletion leads to a leaky phenotype (although no such effect is caused by the deletion of TonB·ExbB·ExbD). Nevertheless, both systems can partially complement each other (35Eick-Helmerich K. Braun V. J. Bacteriol. 1989; 171: 5117-5126Crossref PubMed Google Scholar). We have recently described the crystal structure of the C-terminal domain of TolA (36Lubkowski J. Hennecke F. Plückthun A. Wlodawer A. Structure. 1999; 7: 711-722Abstract Full Text Full Text PDF PubMed Scopus (136) Google Scholar), and we became interested in finding out whether any structural similarity might exist between the C-terminal domains of both TonB and TolA. In this paper, we present the crystal structure of the C-terminal domain of TonB at 1.55-Å resolution and show that this protein exhibits a novel fold that is without homology to any known structures. Moreover, we provide the first evidence that the C-terminal domain of TonB forms a tightly intertwined dimer. The sequence encoding residues 155–239 of tonB from E. coli strain JM83 was polymerase chain reaction-amplified and cloned into the plasmid pAT37 (based on pQE30 from Qiagen). pAT37 codes for protein D (gpD) from bacteriophage λ with an N-terminal His-tag, under control of the T5 promoter (37Forrer P. Jaussi R. Gene. 1998; 224: 45-52Crossref PubMed Scopus (65) Google Scholar). The tonB gene was fused to the C terminus of gpD, with an enterokinase cleavage site engineered by the polymerase chain reaction in between the two proteins. Recombinant bovine enterokinase (purchased from Invitrogen) cleaves after the sequence Asp-Asp-Asp-Asp-Lys. The recombinant protein was expressed overnight at 30 °C in E. coli XL1-Blue, after induction with 1 mmisopropyl-β-d-thio-galactopyranoside. Cells were lysed with a French press and, after centrifugation, the gpD-TonB fusion protein remained in the soluble fraction. The undigested fusion was purified at pH 8.0, using the coupled IMAC-IEX (cation exchange) protocol (38Plückthun A. Krebber A. Krebber C. Horn U. Knüpfer U. Wenderoth R. Nieba L. Proba K. Riesenberg D. McCafferty J. Hoogenboom H.R. Antibody Engineering: A Practical Approach. IRL Press, Oxford1996: 203-252Google Scholar) on a BIOCAD-60 workstation. After dialysis against 50 mm Tris, pH 8.0, 1 mm CaCl2, 0.1% Tween 20, the cleavage reaction was performed at room temperature for 4 h, using 1 unit per mg of fusion of the recombinant bovine enterokinase (Invitrogen). The solution was then dialyzed against 50 mm Mes/Hepes/acetate buffer, pH 8.0. Removal of gpD and enterokinase was again achieved with the coupled IMAC-IEX (cation exchange) protocol, based on the different pI of TonB, gpD, and enterokinase. The C-terminal domain of TonB was dialyzed against 20 mm Tris buffer at pH 7.5 and was concentrated to 15 mg/ml. Crystallization was performed by the hanging-drop vapor diffusion method at 22 °C. Crystal screen I (Hampton Research) was used for the initial screening. Small, rod-shaped crystals were found under conditions 6, 19, 27, and 36. The final refined crystallization conditions were 28–30% polyethylene glycol 3350, 0.1 m Tris buffer at pH 7.5, 50–100 mm CaCl2. After refinement of the conditions, crystals were grown to the size of 0.3–0.5 mm. When a crystal was picked up from a droplet, the diffraction pattern showed split spots or high mosaicity. To improve their quality, crystals were moved from the droplet to a well containing mother liquor and stored for more than 1 day. Such treatment both increased the resolution of diffraction and lowered the mosaicity. TonB crystals were found to belong to the orthorhombic space group P21212 with the unit cell parameters a = 63.78 Å, b = 86.34 Å, c = 26.56 Å. The asymmetric unit contains two molecules, and the V M value is 1.89 Å3/Da (solvent content 35%). The structure of TonB was solved by derivatization with Br− ions (39Dauter Z. Dauter M. Rajashankar K.R. Acta Crystallogr. Sect. D Biol. Crystallogr. 2000; 56: 232-237Crossref PubMed Scopus (285) Google Scholar, 40Dauter Z. Li M. Wlodawer A. Acta Crystallogr. Sect. D Biol. Crystallogr. 2001; 57: 239-249Crossref PubMed Scopus (62) Google Scholar). To prepare a crystal for this procedure, it was soaked for 50 s in a solution containing 1.0m KBr in addition to the crystallization buffer. Subsequently, the crystal was picked up with a nylon loop (Hampton Research) and was flash-frozen in a nitrogen stream. All data sets were collected at 100 K using the ADSC Quantum 4 charge-coupled device detector on the synchrotron beamline X9B at the National Synchrotron Light Source, Brookhaven National Laboratory, Upton, NY. The bromine fluorescence edge was scanned to determine the energy of the inflection, peak, and remote points. Three data sets were measured at 2.0-Å resolution to provide all information necessary for a multiwavelength anomalous diffraction (MAD) experiment. In addition, a data set extending to 1.55 Å was obtained for the purpose of structure refinement. Data were integrated and scaled using the HKL2000 program suite (41Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref PubMed Scopus (38703) Google Scholar). Data collection statistics are summarized in TableI.Table IData collection statisticsRemote 2Remote 1PeakInflectionWavelength (Å)0.98000.91630.91960.9199Resolution (Å)20–1.5520–2.020–2.020–2.0Total reflections138,98780,45181,13358,622Unique reflections21,51819,195 (F+ and F− separated)19,199 (F+ and F− separated)19,128 (F+ and F− separated)Completeness (%) (last shell)96.8 (80.2)99.6 (98.3)99.6 (97.6)99.1 (96.5)R merge2.7 (11.0)2.3 (4.1)2.5 (4.3)1.7 (4.2) Open table in a new tab Four Br− sites were found by both direct and Patterson methods and were refined using the program SOLVE (42Terwilliger T.C. Berendzen J. Acta Crystallogr. Sect. D Biol. Crystallogr. 1996; 52: 749-757Crossref PubMed Scopus (76) Google Scholar), utilizing three data sets corresponding to the peak, inflection, and remote wavelengths, in the resolution range 10–2.5 Å. These sites were also confirmed with the program SHELXD (43Sheldrick G.M. Methods Enzymol. 1997; 276: 628-641Crossref PubMed Scopus (129) Google Scholar). The phases from SOLVE were modified and extended to 1.55 Å using the program DM (44Cowtan K. Main P. Acta Crystallogr. Sect. D Biol. Crystallogr. 1998; 54: 487-493Crossref PubMed Scopus (309) Google Scholar) in the CCP4 program suite (45Collaborative Computational Project Number 4 Acta Crystallogr. Sect. D Biol. Crystallogr. 1994; 50: 760-763Crossref PubMed Scopus (19825) Google Scholar), with the solvent content set at 25%. The mean figure of merit of the phase set was 0.608 for the 10–2.5 Å data after SOLVE, and 0.489 for 20–1.55 Å after DM (0.780 for 20–2.5 Å). The initial model was built using the automatic model-building option of the program ARP/wARP (46Perrakis A. Morris R. Lamzin V.S. Nat. Struct. Biol. 1999; 6: 458-463Crossref PubMed Scopus (2566) Google Scholar) with the full-DM phase set as input. The model was rebuilt with the program O (47Jones T.A. Kieldgaard M. Methods Enzymol. 1997; 277: 173-208Crossref PubMed Scopus (505) Google Scholar) using either electron density maps based on the combination of the MAD and model phases, or straight 2F o − F c maps. The combined phase set was obtained using SIGMAA in the CCP4 program package. After each cycle of rebuilding, the model was refined using SHELXL (48Sheldrick G.M. Schneider T.R. Methods Enzymol. 1997; 277: 319-343Crossref PubMed Scopus (1897) Google Scholar) at the resolution range of 20 to 1.55 Å, without applying any non-crystallographic (NCS) restraints, as the latter prevented proper convergence. Eight full cycles of remodeling and refinement were performed, with the refinement of individual anisotropicB-factors for all atoms initiated in cycle six. In addition to protein atoms, 219 water molecules and four bromide ions have been added to the model. The R-value for all reflections between 20 and 1.55 Å is 16.0% (R free 23.0%). The geometrical properties of the model were assessed by the program PROCHECK (49Laskowski R.A. MacArthur M.W. Moss D.S. Thornton J.M. J. Appl. Crystallogr. 1993; 26: 283-291Crossref Google Scholar), and the secondary structure elements were assigned by the program PROMOTIF (50Hutchinson E.G. Thornton J.M. Protein Sci. 1996; 5: 212-220Crossref PubMed Scopus (1001) Google Scholar). The figures were prepared with Molscript (51Kraulis P.J. J. Appl. Crystallogr. 1991; 24: 946-950Crossref Google Scholar) or Bobscript (52Esnouf R.M. Acta Crystallogr. Sect. D Biol. Crystallogr. 1999; 55: 938-940Crossref PubMed Scopus (851) Google Scholar) and rendered with Raster3D (53Meritt E.A. Murphy M.E.P. Acta Crystallogr. Sect. D Biol. Crystallogr. 1994; 50: 869-873Crossref PubMed Scopus (2859) Google Scholar). The crystal structure of the C-terminal domain of TonB has been determined by multiwavelength anomalous diffraction, and has been refined using SHELXL at 20–1.55 Å, yielding a model with lowR-factor and excellent stereochemistry. The refinement statistics and the indicators of model quality are listed in TableII. The electron density maps (both the combined map utilizing the phases of the MAD data and of the model, and the final 2F o − F c map) are generally of excellent quality (Fig. 1). However, both maps are poorly defined in the neighborhood of residues 194–201. In this region, B-factors of all atoms are relatively high, indicating extensive flexibility of the polypeptide chain. Some disorder is also present at both termini of each molecule. Residues that are not visible in the maps include the first ten N-terminal amino acids of our construct (residues 155–164), as well as the last one or two residues on the C terminus (residues 238 and 239 of molecule A, and 239 of molecule B). The electron density of the remaining parts of the protein is very well defined. The mean positional error in atomic coordinates as estimated by the Luzzati plot is 0.16 Å. All non-glycine and non-proline residues of the model lie either in the most favorable region or in the additionally allowed region of the Ramachandran plot.Table IIRefinement statistics for the final coordinates of the C-terminal domain of TonBResolution range20.0–1.55 ÅUnique reflections used20,365R cryst16.0%R free23.0% (5% test set)r.m.s. deviations from idealityBond lengths0.009 ÅAngles0.028 ÅNon-zero chiral volumes0.054 Å3Zero chiral volumes0.047 Å3Number of amino acid residues73 + 74Number of protein atoms578 + 586Number of heteroatoms4Number of solvent atoms219 Open table in a new tab The C-terminal domain of TonB is cylinder-shaped with the length of 65 Å, and the diameter of 25 Å, with two protein chains forming a single compact entity. Each chain is rich in β-strands (∼50% of the secondary structure) with much more limited extent (∼15%) of residues found in helical conformation (either α-helix or 310 helix). Each monomer contains three β-strands (strand S1, residues 169–182; S2, 188–194; and S3, 221–236) and one α-helix (residues 200–210 in molecule A, and 200–211 of molecule B). In addition, a short 310 helix includes residues 211–213 of molecule A. All six β-strands make a large antiparallel β-sheet. The β-strand S3 of each monomer is swapped between the monomers (Fig. 2). Four β-strands, S1 and S3 of both molecules, are located on one side, whereas the two short β-strands S2 and the helices are located on the other side (Fig. 3).Figure 3Stereo ribbon diagram of the C-terminal domain of TonB, showing the intertwined dimer. The color scheme is the same as in Fig. 2. The atomic coordinates have been deposited in the Protein Data Bank (accession code 1IHR).View Large Image Figure ViewerDownload Hi-res image Download (PPT) The two monomers differ slightly from each other. At any refinement stage, application of non-crystallographic restraints between the two molecules resulted in significantly worse behavior than if such restraints were not utilized. The r.m.s. deviation between the Cα atoms of the two monomers is 0.42 Å, whereas the r.m.s. deviation between the side-chain atoms of the two monomers is 1.15 Å. In the case of the main chain of the protein, the largest differences are found at the N terminus. As judged by their high temperature factors, both termini are located in highly flexible regions of the structure. The difference between Cα positions of Ala-165, the first visible residue on the N terminus, exceeds 2 Å (residues preceding Ala-165 were not visible in either molecule). In the case of the side chains, several residues show significantly different values of the χ1 angle. These residues include Leu-170, Arg-171, Glu-173, Asn-200, Lys-219, and Lys-231. The only hydrophobic amino acid among them is Leu-170, and different orientation of its side chain leads to the presence of more hydrophobic contacts in molecule B. The other residues are all polar or charged, and all are solvent-exposed. The orientations of the Cγ atoms in the residues belonging to the β-sheet (Arg-171, Glu-173, Lys-231) closely coincide. For positively charged residues (Arg-171 and Lys-231 of both molecules) Cγ atoms point toward the N terminus of molecule A. The Cγ atoms of the negatively charged residue Glu-173 point toward the N terminus of molecule B. Asn-200 is located just after the highly flexible loop and is itself flexible, judged by its high B-factor. Lys-219 is located at the end of the molecule and is also flexible. The interactions between the two protein chains that form a single compact molecule of the C-terminal domain of TonB are unusually extensive. The dimeric interface area covers 41% of the surface of each monomer, thus the individual chains are unlikely to be able to exist independently and the protein becomes stable only as a dimer. The region of the β-sheet shows tight dimeric interactions, whereas the interactions on the opposite side of the molecule are not as close. Although the single antiparallel β-sheet present in the dimer is composed of strands originating from different molecules, the hydrogen bonding pattern is close to ideal. The loop between β-strand S2 and the α-helix is very flexible, as indicated by its highB-factor. The average B-factors of the main-chain atoms in this loop are 71 Å2 and 60 Å2 for molecules A and B, respectively, as compared with the respective averages for other areas of 20.6 Å2 and 21.7 Å2. Crystal packing in the vicinity of these loops is rather loose, resulting in the formation of clefts or channels on the molecular surface. The channels are made by residues 195–200 and 172–176 in both molecules, the former belonging to the loop, and the latter to the β-strand S1. The binding of a nutrient, such as vitamin B12 or an iron·siderophore complex, to the external face of an outer membrane receptor triggers a series of conformational changes: The N-terminally located TonB box, which is hidden within the barrel of the unliganded receptor, is made to project in an extended form into the periplasm and, thus, becomes freely accessible for interaction with the C-terminal domain of TonB (28Merianos H.J. Cadieux N. Lin C.H. Kadner R.J. Cafiso D.S. Nat. Struct. Biol. 2000; 7: 205-209Crossref PubMed Scopus (98) Google Scholar). Additionally, subtle structural changes of the receptor observed crystallographically, such as the upward translation of selected loops of the cork domain (also termed “hatch region”), may disrupt hydrophobic interactions between the so-called switch helix (residues 24–29 in FhuA) and the internal wall of the barrel, and the helix unfolds (6Locher K.P. Rees B. Koebnik R. Mitschler A. Moulinier L. Rosenbusch J.P. Moras D. Cell. 1998; 95: 771-778Abstract Full Text Full Text PDF PubMed Scopus (459) Google Scholar). It thus constitutes a candidate for signaling the occupancy of the outer binding site by a