Solution Structure of the Pore-forming Protein of Entamoeba histolytica
2004; Elsevier BV; Volume: 279; Issue: 17 Linguagem: Inglês
10.1074/jbc.m312978200
ISSN1083-351X
AutoresOliver Hecht, N.A. van Nuland, Karin Schleinkofer, Andrew J. Dingley, H. D. Bruhn, Matthias Leippe, Joachim Grötzinger,
Tópico(s)Heme Oxygenase-1 and Carbon Monoxide
ResumoAmoebapore A is a 77-residue protein from the protozoan parasite and human pathogen Entamoeba histolytica. Amoebapores lyse both bacteria and eukaryotic cells by pore formation and play a pivotal role in the destruction of host tissues during amoebiasis, one of the most life-threatening parasitic diseases. Amoebapore A belongs to the superfamily of saposin-like proteins that are characterized by a conserved disulfide bond pattern and a fold consisting of five helices. Membrane-permeabilizing effector molecules of mammalian lymphocytes such as porcine NK-lysin and the human granulysin share these structural attributes. Several mechanisms have been proposed to explain how saposin-like proteins form membrane pores. All mechanisms indicate that the surface charge distribution of these proteins is the basis of their membrane binding capacity and pore formation. Here, we have solved the structure of amoebapore A by NMR spectroscopy. We demonstrate that the specific activation step of amoebapore A depends on a pH-dependent dimerization event and is modulated by a surface-exposed histidine residue. Thus, histidine-mediated dimerization is the molecular switch for pore formation and reveals a novel activation mechanism of pore-forming toxins. Amoebapore A is a 77-residue protein from the protozoan parasite and human pathogen Entamoeba histolytica. Amoebapores lyse both bacteria and eukaryotic cells by pore formation and play a pivotal role in the destruction of host tissues during amoebiasis, one of the most life-threatening parasitic diseases. Amoebapore A belongs to the superfamily of saposin-like proteins that are characterized by a conserved disulfide bond pattern and a fold consisting of five helices. Membrane-permeabilizing effector molecules of mammalian lymphocytes such as porcine NK-lysin and the human granulysin share these structural attributes. Several mechanisms have been proposed to explain how saposin-like proteins form membrane pores. All mechanisms indicate that the surface charge distribution of these proteins is the basis of their membrane binding capacity and pore formation. Here, we have solved the structure of amoebapore A by NMR spectroscopy. We demonstrate that the specific activation step of amoebapore A depends on a pH-dependent dimerization event and is modulated by a surface-exposed histidine residue. Thus, histidine-mediated dimerization is the molecular switch for pore formation and reveals a novel activation mechanism of pore-forming toxins. The protozoan parasite, Entamoeba histolytica, inhabits the colon of infected humans. It is the causative agent of human amoebiasis that often leads to tissue damage, colitis, and extraintestinal abscesses (1Ravdin J.I. Clin. Infect. Dis. 1995; 20 (quiz, 1465-1466): 1453-1464Crossref PubMed Scopus (199) Google Scholar). Amoebiasis is the second leading cause of death from parasitic diseases worldwide (2Stanley Jr., S.L. Lancet. 2003; 361: 1025-1034Abstract Full Text Full Text PDF PubMed Scopus (1038) Google Scholar). About 50 million people suffer from invasive amoebiasis, of whom up to 100,000 die annually (3Walsh J.A. Rev. Infect. Dis. 1986; 8: 228-238Crossref PubMed Scopus (526) Google Scholar). In the amoebic trophozoite, several factors have been identified that are involved in pathogenesis. In addition to a galactose-/N-acetylgalactosamine-specific lectin on the amoebic surface that mediates adhesion to colonic mucus and host cells (4Petri Jr., W.A. Haque R. Mann B.J. Annu. Rev. Microbiol. 2002; 56: 39-64Crossref PubMed Scopus (278) Google Scholar) and secreted cysteine proteinases that disintegrate tissues by cleaving extracellular matrix proteins (5Bruchhaus I. Jacobs T. Leippe M. Tannich E. Mol. Microbiol. 1996; 22: 255-263Crossref PubMed Scopus (156) Google Scholar), a family of membrane-active polypeptides have been discovered (6Leippe M. Parasitol. Today. 1997; 13: 178-183Abstract Full Text PDF PubMed Scopus (137) Google Scholar). These polypeptides exist as three isoforms and are named amoebapore A, B, and C, respectively. They are capable of lysing a broad spectrum of target cells, including human host cells and bacteria. It has been recently shown that trophozoites of E. histolytica lacking amoebapore A, due to transcriptional silencing of the encoding gene, became avirulent (7Bracha R. Nuchamowitz Y. Mirelman D. Eukaryot. Cell. 2003; 2: 295-305Crossref PubMed Scopus (110) Google Scholar), demonstrating that this protein is a key pathogenicity factor of the parasite. All three amoebapore isoforms have been isolated and biochemically characterized, and their primary structure has been elucidated by molecular cloning of the genes of their precursors (8Leippe M. Ebel S. Schoenberger O.L. Horstmann R.D. Müller-Eberhard H.J. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 7659-7663Crossref PubMed Scopus (183) Google Scholar, 9Leippe M. Tannich E. Nickel R. van der Goot G. Pattus F. Horstmann R.D. Müller-Eberhard H.J. EMBO J. 1992; 11: 3501-3506Crossref PubMed Scopus (100) Google Scholar, 10Leippe M. Andrä J. Nickel R. Tannich E. Müller-Eberhard H.J. Mol. Microbiol. 1994; 14: 895-904Crossref PubMed Scopus (137) Google Scholar). The mature proteins consist of 77 amino acid residues each and are localized within cytoplasmic granules. The overall sequence identity between the three amoebapores is between 35 and 57% (10Leippe M. Andrä J. Nickel R. Tannich E. Müller-Eberhard H.J. Mol. Microbiol. 1994; 14: 895-904Crossref PubMed Scopus (137) Google Scholar). Despite the substantial sequence divergence, they possess a characteristic disulfide bond pattern and a single conserved C-terminal histidine residue. The secondary structure of the major isoform A has been determined to be exclusively α-helical (9Leippe M. Tannich E. Nickel R. van der Goot G. Pattus F. Horstmann R.D. Müller-Eberhard H.J. EMBO J. 1992; 11: 3501-3506Crossref PubMed Scopus (100) Google Scholar), and this has also been predicted for amoebapore B and C (10Leippe M. Andrä J. Nickel R. Tannich E. Müller-Eberhard H.J. Mol. Microbiol. 1994; 14: 895-904Crossref PubMed Scopus (137) Google Scholar). All amoebapores are able to form pores in membranes by oligomerization and thus affect the integrity of target cell membranes (10Leippe M. Andrä J. Nickel R. Tannich E. Müller-Eberhard H.J. Mol. Microbiol. 1994; 14: 895-904Crossref PubMed Scopus (137) Google Scholar, 11Gutsmann T. Rieckens B. Bruhn H. Wiese A. Seydel U. Leippe M. Biochemistry. 2003; 42: 9804-9812Crossref PubMed Scopus (32) Google Scholar). By sequence similarity, amoebapores have been grouped into the family of saposin-like proteins (SAPLIP). 1The abbreviations used are: SAPLIP, saposin-like proteins; HPLC, high pressure liquid chromatography; NOE, nuclear Overhauser effect; NOESY, NOE spectroscopy; DEPC, diethylpyrocarbonate; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine. Although the members of the SAPLIP family have different biological functions, they are all able to interact with lipids. With one exception, all of them possess six conserved cysteine residues that are involved in forming the disulfide bond pattern characteristic of this protein family (12Munford R.S. Sheppard P.O. O'Hara P.J. J. Lipid Res. 1995; 36: 1653-1663Abstract Full Text PDF PubMed Google Scholar). Despite the enormous evolutionary distance, amoebapores reveal a substantial sequence similarity with membrane-permeabilizing effector molecules of mammalian lymphocytes such as porcine NK-lysin and human granulysin. Like their amoebic counterparts, the mammalian proteins reside in intracellular granules and are able to lyse bacteria and eukaryotic cells (10Leippe M. Andrä J. Nickel R. Tannich E. Müller-Eberhard H.J. Mol. Microbiol. 1994; 14: 895-904Crossref PubMed Scopus (137) Google Scholar, 13Andersson M. Gunne H. Agerberth B. Boman A. Bergman T. Sillard R. Jornvall H. Mutt V. Olsson B. Wigzell H. Dagerlind A. Boman H.G. Gudmundsson G.H. EMBO J. 1995; 14: 1615-1625Crossref PubMed Scopus (270) Google Scholar, 14Pena S.V. Hanson D.A. Carr B.A. Goralski T.J. Krensky A.M. J. Immunol. 1997; 158: 2680-2688PubMed Google Scholar). Based on structural information and biochemical data, several mechanisms have been suggested to explain how saposin-like proteins may form membrane pores (15Bruhn H. Leippe M. Biol. Chem. 1999; 380: 1001-1007Crossref PubMed Scopus (32) Google Scholar, 16Miteva M. Andersson M. Karshikoff A. Otting G. FEBS Lett. 1999; 462: 155-158Crossref PubMed Scopus (103) Google Scholar). For both NK-lysin and granulysin, it has been suggested that binding of the monomeric proteins to the membrane via their charged epitopes results in destabilization of the lipid bilayer (17Ruysschaert J.M. Goormaghtigh E. Homble F. Andersson M. Liepinsh E. Otting G. FEBS Lett. 1998; 425: 341-344Crossref PubMed Scopus (58) Google Scholar, 18Anderson D.H. Sawaya M.R. Cascio D. Ernst W. Modlin R. Krensky A. Eisenberg D. J. Mol. Biol. 2003; 325: 355-365Crossref PubMed Scopus (133) Google Scholar), whereas amoebapore A has been proposed to act via the barrel stave mechanism (19Keller F. Hanke W. Trissl D. Bakker-Grunwald T. Biochim. Biophys. Acta. 1989; 982: 89-93Crossref PubMed Scopus (19) Google Scholar). Here, we present the high resolution three-dimensional structure of monomeric amoebapore A solved by NMR spectroscopy. The structure reveals that the charge distribution on the surface of amoebapore A differs markedly from that of other members of the saposin-like family of proteins, namely the porcine NK-lysin (20Liepinsh E. Andersson M. Ruysschaert J.M. Otting G. Nat. Struct. Biol. 1997; 4: 793-795Crossref PubMed Scopus (223) Google Scholar) and human granulysin (18Anderson D.H. Sawaya M.R. Cascio D. Ernst W. Modlin R. Krensky A. Eisenberg D. J. Mol. Biol. 2003; 325: 355-365Crossref PubMed Scopus (133) Google Scholar). In contrast to NK-lysin and granulysin with their positively charged epitopes that are suggested to be responsible for membrane interaction, amoebapore A shows an extended hydrophobic surface area. Moreover, we report that dimerization of amoebapore A is a prerequisite for membrane binding and pore formation. This dimerization mechanism is a pH-dependent activation step of cell lysis by amoebapore A, which is mediated by a key histidine residue located at the dimer interface. Protein Purification—Frozen trophozoites of E. histolytica HM-1: IMSS were extracted twice in 10 volumes of 1 n HCl, 5% acetic acid, 1% trifluoroacetic acid, 1% NaCl overnight at 4 °C under constant shaking. The acid extracts were centrifuged at 150,000 × g at 4 °C for 1 h. The supernatants were combined and passed through a tC18 12-cc (2 g) Sep-Pak cartridge (Waters). The cartridge was washed with 12 ml of 0.1% trifluoroacetic acid, 35% acetonitrile, and elution of adsorbed material was achieved with 15 ml of 0.1% trifluoroacetic acid, 80% acetonitrile. The eluate was lyophilized, redissolved in 4 ml of 0.1% trifluoroacetic acid, and subsequently applied in portions to reversed-phase HPLC using an Aquapore Butyl 300 column (2.1 × 220 mm; Brownlee Laboratories, San Jose, CA) connected with a 130 A separation system (Applied Biosystems). Peptides were eluted with a nonlinear gradient of 0% (3 min), 40-50% (10 min), 50-70% (22 min), and 70% (10 min) acetonitrile in 0.1% trifluoroacetic acid at 50 °C and at a flow rate of 0.2 ml/min. All fractions were immediately tested for pore-forming activity. Active fractions were lyophilized, resuspended in 0.1% trifluoroacetic acid, and subjected to rechromatography using the same column and a linear gradient of 0-70% acetonitrile in 0.1% trifluoroacetic acid and otherwise the same conditions. Homogeneity of purified amoebapore A, which elutes at 55% acetonitrile, has been verified by N-terminal sequencing and mass spectrometry and was routinely confirmed by analytical reversed-phase HPLC and Tricine-SDS-PAGE using 13% separation gels (21Schägger H. von Jagow G. Anal. Biochem. 1987; 166: 368-379Crossref PubMed Scopus (10499) Google Scholar). Amoebapore A was repeatedly lyophilized and redissolved in 10 mm HCl to evaporate residual trifluoroacetic acid bound to the peptide and stored at -20 °C. NMR Spectroscopy—Lyophilized protein was dissolved in 93% H2O, 7% D2O, and the pH was adjusted to 3.5 The protein concentration was 1.5 mm. Sets of two-dimensional homonuclear 1H-NMR total correlation spectroscopy and NOESY spectra were recorded at 600 or 750 MHz and 24 °C on either a Varian INOVA 750 or Bruker Avance 600/750 spectrometers. The mixing times for the six NOESY spectra were 60, 100, 140, 180, 250, and 300 ms, respectively. The spectral width was 8200 Hz in both dimensions, and the water signal was suppressed by a weak radio frequency pulse during the relaxation delay. For each spectrum, 400 t1 increments were acquired, each with 2048 complex points using the time-proportional phase incrementation scheme. Prior to Fourier transformation, a 60° shifted sine bell window function was applied in both dimensions, and the spectra were zero-filled in t1 so that 1024 × 1024 data points were obtained. Finally, the spectra were baseline-corrected with a polynomial function. To verify the amide proton resonance assignment, we recorded a natural abundance 1H-15N-heteronuclear single quantum coherence spectrum at 750 MHz with spectral widths of 9000 and 2500 Hz for 1H and 15N, respectively. For the spectrum, 180 t1 increments were acquired each with 1600 complex points. For the H/D exchange experiments, the protein was lyophilized again and dissolved in 100% D2O. The pH and concentration were the same as in the above mentioned experiments. NOESY spectra were recorded with a mixing time of 150 ms at 500 MHz on a Bruker AVANCE 500. The spectral width was 6009 Hz in both dimensions, and the water signal was suppressed by a weak radio frequency pulse during the relaxation delay. Because of the slowly exchanging amide protons, partly due to the low pH, the temperature was increased from 24 °C in the first recordings to 50 °C in the last recording. All data processing was performed on an SGI Indigo work station using the program nmrDraw (22Delagio F. Grzesiek S. Vuister G.W. Zhu G. Pfeifer J. Bax A. J. Biomol. NMR. 1995; 6: 277-293PubMed Google Scholar). All subsequent procedures, such as spectral assignment, cross-peak integration, and distance determination, were performed by using the NMRview program (23Johnson B.A. Blevins R.A. J. Biomol. NMR. 1994; 4: 603-614Crossref PubMed Scopus (2685) Google Scholar). CD Spectropolarimetry—CD measurements were carried out with a Jasco J-720 spectropolarimeter (Japan Spectroscopic Co., Ltd., Tokyo, Japan), calibrated according to Chen and Yang (24Chen G.C. Yang J.T. Anal. Lett. 1977; 10: 1195-1207Crossref Scopus (397) Google Scholar). The spectral bandwidth was 1 nm. The measurements were carried out at 24 °C. Protein Concentration—Protein concentrations were calculated from absorption spectra in the range of 240-320 nm using the method of Waxman et al. (25Waxman E. Rusinova E. Hasselbacher C.A. Schwartz G.P. Laws W.R. Ross J.B. Anal. Biochem. 1993; 210: 425-428Crossref PubMed Scopus (36) Google Scholar). Assay for Pore-forming Activity—Pore-forming activity was determined by monitoring the dissipation of a valinomycin-induced diffusion potential in liposomes as described previously (8Leippe M. Ebel S. Schoenberger O.L. Horstmann R.D. Müller-Eberhard H.J. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 7659-7663Crossref PubMed Scopus (183) Google Scholar). Briefly, liposomes prepared from crude soybean phosphatidylcholine IIS (Sigma) (40 mg/ml) in 50 mm K2SO4, 0.5 mm EDTA, 50 mm Tris maleate, pH 5.2, were diluted 1: 4000 in a similar buffer containing Na+ instead of K+ and adjusted to pH 5.2. The addition of 1 nm valinomycin induced a diffusion potential that was monitored by the fluorescence quenching of 3,3′-diethylthiodicarbocyanine iodide dye (1 μm; Eastman Kodak Co.) using a fluorescence spectrophotometer (LS50B; PerkinElmer Life Sciences) with excitation and emission wavelengths of 620 and 670 nm, respectively. Pore-forming activity was measured as the initial change of fluorescence over time after the addition of sample. One unit is defined as a fluorescence increase to 5% of the prevalinomycin level in 1 min at 25 °C. Structure Calculation—Structure calculations were performed using the software program DYANA (26Guntert P. Mumenthaler C. Wuthrich K. J. Mol. Biol. 1997; 273: 283-298Crossref PubMed Scopus (2558) Google Scholar). The structure calculation was based on a set of 1418 NOE distance restraints. In addition to the NOE constraints, 2 × 58 hydrogen bond distance restraints, resulting from the H/D exchange experiments, and 3 × 12 distances, representing the spaces within a disulfide bond, were introduced. NOE cross-peak intensities were classified as strong, medium, and weak and assigned to restraints of 1.8-2.8, 1.8-3.4, and 1.8-5.0 Å., respectively, with appropriate pseudo-atom corrections. Hydrogen bond restraints were 1.8-2.4 and 2.8-3.4 Å for H/O and N/O distances, respectively. Disulfide bond restraints were 2.03-2.15, 3.03-3.13, and 2.97-4.49 Å for SG/SG, SG/CB, and SG/CA distances, respectively. In the final structure calculation, all 1570 distance restraints were used to calculate 75 structures. Of these structures, 20 structures with the lowest target functions were selected. The structural statistics of these structures are listed in Table I. The coordinates have been deposited in the Protein Data Bank (accession code 1OF9).Table IStructural statistics for the 20 conformers of amoebapore ADistance restraintsIntraresidue (i − j = 0)606Sequential (|i − j| =1)406Medium range (2 ≤|i − j| = ≤ 4)246Long range (|i − j| ≥ 5)160Hydrogen bonds2 × 58Disulfide bonds3 × 12All1570Pairwise r.m.s. deviation for residues 6-73 in ÅMean global backbone r.m.s.d.ar.m.s.d., root mean square deviation.0.71 ± 0.15Mean global heavy atom r.m.s.d.1.20 ± 0.16Pairwise r.m.s. deviation for secondary structures in ÅMean global backbone r.m.s.d.0.25 ± 0.11Mean global heavy atom r.m.s.d.0.78 ± 0.16Ramachandran plotbMean values for all 20 conformers.Most favored regions (%)75Additional allowed regions (%)18Generously allowed regions (%)7Disallowed regions (%)0a r.m.s.d., root mean square deviation.b Mean values for all 20 conformers. Open table in a new tab Size Exclusion Chromatography—Freeze dried amoebapore A was solved in 50 mm sodium citrate buffer (pH 3.5), 50 mm sodium phosphate buffer (pH 5.2), and 50 mm Tris/HCl buffer (pH 8.0), respectively, incubated for 1-2 h at room temperature, and then subjected to size exclusion chromatography with an equilibrated Superdex 75 (16/60) column (Amersham Biosciences) at a constant flow rate of 1 ml/min at 4 °C. The column was calibrated using a mixture of four proteins of known molecular mass, i.e. albumin (67 kDa), ovalbumin (43 kDa), chymotrypsinogen A (25 kDa), and ribonuclease A (13.7 kDa). The column was equilibrated with 50 mm sodium citrate buffer (pH 3.5), 50 mm sodium phosphate buffer (pH 5.2), and 50 mm Tris/HCl buffer (pH 8.0), respectively, and loaded with 1.0 ml of protein solution. Fractions of 3 ml were collected. Cross-linking—For cross-linking in solution, 0.5 μg of amoebapore A were incubated with 100 μm 1-ethyl-3-(dimethylaminopropyl)-carbodiimide in 1 m glycine methylester (pH 4.5) at 20 °C for 2 h (27Hoare D.G. Koshland Jr., D.E. J. Biol. Chem. 1967; 242: 2447-2453Abstract Full Text PDF PubMed Google Scholar, 28Carraway K.L. Koshland Jr., D.E. Biochim. Biophys. Acta. 1968; 160: 272-274Crossref PubMed Scopus (183) Google Scholar). Alternatively, dithiobis(succinimidyl propionate) was added from a freshly prepared stock solution in dimethyl sulfoxide to the same amount of amoebapore A in 50 mm sodium phosphate (pH 7.0) to reach a final concentration of 100 μm (29Lomant A.J. Fairbanks G. J. Mol. Biol. 1976; 104: 243-261Crossref PubMed Scopus (360) Google Scholar). After an incubation period of 30 min at 20 °C, the reaction was stopped by adding an excess of Tris. The samples were subjected to Tricine-SDS-PAGE using 13% gels (21Schägger H. von Jagow G. Anal. Biochem. 1987; 166: 368-379Crossref PubMed Scopus (10499) Google Scholar). DEPC Modification—Diethylpyrocarbonate (DEPC) modification was performed according to Andrä and Leippe (see Ref. 32Andrä J. Leippe M. FEBS Lett. 1994; 354: 97-102Crossref PubMed Scopus (39) Google Scholar). Briefly, 0.58 mg of amoebapore A were dissolved in 4.0 ml of a 50 mm sodium phosphate (pH 6.0) buffer. After incubation with a 200-fold excess of DEPC for 1.5 h at 4 °C and another 1 h at room temperature, the solution was subjected to size exclusion chromatography and eluted with a 50 mm sodium phosphate buffer (pH 5.2) at 4 °C. Structure of Amoebapore A—Since the protein could not be recombinantly expressed in Escherichia coli and was purified from its natural source E. histolytica, structure elucidation of the amoebapore A relied on conventional two-dimensional 1H-NMR experiments. The helical secondary structure elements in amoebapore A were identified by the presence of characteristic dαN(i + 3) and dβN(i + 3) sequential connectivities in the two-dimensional NOESY spectra of this protein. Fig. 1 shows these patterns of connectivities derived from the NOESY spectra. This specific pattern is missing in the loop regions and incomplete for helix 1 due to partial spectral overlap. Additional information about the secondary structure elements was derived from H/D exchange experiments. Even after increasing the temperature to 50 °C for 24 h, several amide protons were still observable (Fig. 1) and therefore included in the structure calculation as hydrogen bonds (see "Experimental Procedures"). The structure of amoebapore A (for statistics, see Table I) consists of five α-helices (Fig. 2, A and B), which comprise residues 4-16 (helix I), 25-35 (helix II), 42-52 (helix III), 55-63 (helix IV), and 67-73 (helix V). A single disulfide bridge connects helices 2 and 3, whereas two disulfide bridges connect helices 1 and 5. According to structural classification of proteins (SCOP) (30Murzin A.G. Brenner S.E. Hubbard T. C C. J. Mol. Biol. 1995; 247: 536-540Crossref PubMed Scopus (5606) Google Scholar), the structure can be described as a folded leaf in which helices 1 and 2 are one-half of the leaf and are folded against helices 3, 4, and 5 as the second half. Helices 1 and 2 are connected by a loop consisting of eight residues (17Ruysschaert J.M. Goormaghtigh E. Homble F. Andersson M. Liepinsh E. Otting G. FEBS Lett. 1998; 425: 341-344Crossref PubMed Scopus (58) Google Scholar, 18Anderson D.H. Sawaya M.R. Cascio D. Ernst W. Modlin R. Krensky A. Eisenberg D. J. Mol. Biol. 2003; 325: 355-365Crossref PubMed Scopus (133) Google Scholar, 19Keller F. Hanke W. Trissl D. Bakker-Grunwald T. Biochim. Biophys. Acta. 1989; 982: 89-93Crossref PubMed Scopus (19) Google Scholar, 20Liepinsh E. Andersson M. Ruysschaert J.M. Otting G. Nat. Struct. Biol. 1997; 4: 793-795Crossref PubMed Scopus (223) Google Scholar, 21Schägger H. von Jagow G. Anal. Biochem. 1987; 166: 368-379Crossref PubMed Scopus (10499) Google Scholar, 22Delagio F. Grzesiek S. Vuister G.W. Zhu G. Pfeifer J. Bax A. J. Biomol. NMR. 1995; 6: 277-293PubMed Google Scholar, 23Johnson B.A. Blevins R.A. J. Biomol. NMR. 1994; 4: 603-614Crossref PubMed Scopus (2685) Google Scholar, 24Chen G.C. Yang J.T. Anal. Lett. 1977; 10: 1195-1207Crossref Scopus (397) Google Scholar), whereas helices 3 and 4 virtually merge and are separated by only two residues (53 and 54). Residues 64-66 link helices 4 and 5. The two parts of the folded leaf are connected by a loop between helices 2 and 3 (residues 36-41). The superposition of 20 independently calculated structures demonstrates the coherence of the generated ensemble, which is shown as a schematic representation in Fig. 2A. In particular, within the helical regions, the root mean square deviations for the backbone atoms are 0.25 Å. Approximately 75% of the backbone Φ/ψ-dihedral angle pairs were in the most favorable region of a Ramachandran plot, and none were observed in disallowed regions. On the basis of this ensemble, an average structure of amoebapore A was calculated and resembles the typical fold of the saposin-like proteins (Fig. 2B). A comparison of this structure with the structures of porcine NK-lysin and human granulysin is shown in Fig. 3. Although the three structures show a similar global fold, they differ substantially in the spatial arrangement of the helices. The orientations of helices 2, 3, and 4 relative to each other are similar in the three structures. However, the orientation of helices 1 and 5 relative to 2, 3, and 4 diverges considerably in amoebapore A when compared with NK-lysin and granulysin (Fig. 3). In amoebapore A, helix 1 runs almost parallel to helices 2 and 3, whereas in NK-lysin and granulysin, helix 1 is kinked by ∼90°. Remarkably, the molecular surface character of amoebapore A constituted by residues from helices 1 and 2 is predominantly hydrophobic (Fig. 3). The length of this hydrophobic area of about 36 Å is sufficient to span a lipid bilayer (31Bransburg-Zabary S. Kessel A. Gutman M. Ben-Tal N. Biochemistry. 2002; 41: 6946-6954Crossref PubMed Scopus (24) Google Scholar). In contrast, both mammalian molecules, NK-lysin and granulysin, have clusters of positively charged amino acids, which are proposed to be responsible for the initial contact with the membrane and ultimately lead to its destruction (16Miteva M. Andersson M. Karshikoff A. Otting G. FEBS Lett. 1999; 462: 155-158Crossref PubMed Scopus (103) Google Scholar, 18Anderson D.H. Sawaya M.R. Cascio D. Ernst W. Modlin R. Krensky A. Eisenberg D. J. Mol. Biol. 2003; 325: 355-365Crossref PubMed Scopus (133) Google Scholar). As such cationic clusters are not present on the surface of amoebapore A (Fig. 3), a different mode of interaction with the lipid bilayer must exist.Fig. 3Ribbon (44Carson M. J. Appl. Crystallogr. 1991; 24: 946-950Crossref Scopus (784) Google Scholar) representations of amoebapore A (left), NK-lysin (middle), and granulysin (right) are shown in the upper panel. The C and N termini are depicted in black. All molecules were superimposed onto Cα atoms of residues 30-60. Helices are numbered by roman numerals. The electrostatic potential of the molecular surface is shown in the lower panel in the same orientation as in the upper panel, indicating positive potential in blue and negative potential in red. The representation of the surface was generated using GRASP (43Nicholls A. Sharp K.A. Honig B. Proteins. 1991; 11: 281-296Crossref PubMed Scopus (5318) Google Scholar).View Large Image Figure ViewerDownload Hi-res image Download (PPT) Amoebapore A Forms Stable Dimers—As amoebapore A has been shown to be most active at pH 5.2 (32Andrä J. Leippe M. FEBS Lett. 1994; 354: 97-102Crossref PubMed Scopus (39) Google Scholar), initial NMR experiments were performed at this pH. Unfortunately, the proton T2 values measured under these conditions were indicative of protein oligomerization. Stepwise lowering of the pH until 3.5 resulted in proton T2 values corresponding to a monomeric protein species (data not shown). Apparently, the oligomerization of amoebapore A, which pI has been calculated to 5.9, is driven by electrostatic interactions that are disrupted by protonation of acidic residues upon lowering the pH. To examine whether the global conformation of amoebapore A changes with pH, CD spectra of amoebapore A were recorded at pH values of 3, 4, 5, 6, and 7. The identical CD spectra observed in Fig. 4A clearly indicate that the overall conformation of amoebapore A is not influenced by shifting the pH between 3 and 7. Although CD spectroscopy is a very sensitive method to observe changes in secondary structure, we cannot rule out changes in tertiary structure. As the conformation of amoebapore A did not change over the pH range tested, NMR experiments were performed at a pH value (i.e. 3.5) at which the protein does not self-associate. The observed pH-dependent oligomerization of amoebapore A raised the question about the stoichiometry of the functional protein complexes. To determine the stoichiometry and the stability of the oligomers, we performed size exclusion chromatography at different pH values. In accordance with the T2 measurements, chromatography at pH 3.5 revealed that the protein exists as a stable monomer (Fig. 4B). At pH 5.2, the protein eluted from the column as a single peak at a retention time corresponding to the molecular weight of the dimeric species (Fig. 4B). No higher order oligomers were detectable. The fact that amoebapore A is most active at this pH (32Andrä J. Leippe M. FEBS Lett. 1994; 354: 97-102Crossref PubMed Scopus (39) Google Scholar) raised the possibility that dimerization is a prerequisite for pore formation. Fractions containing the amoebapore A dimer were examined for pore-forming activity by monitoring the dissipation of a valinomycin-induced diffusion potential in liposomes. All of the pore-forming activity (54,375 units) of the material loaded onto the column was recovered in fractions containing the dimer. To further examine the relationship between pH-dependent activity and the oligomerization state, we performed size exclusion chromatography experiments at pH 8, at which amoebapore A has been described to be inactive (32Andrä J. Leippe M. FEBS Lett. 1994; 354: 97-102Crossref PubMed Scopus (39) Google Scholar). Under these conditions, the protein elutes predominantly as a monomer (Fig. 4B). Although some entity was detected near the position of the dimer, it is known that amoebapore does not exert pore-formin
Referência(s)