The Spatial Organization of Apolipoprotein A-I on the Edge of Discoidal High Density Lipoprotein Particles
2003; Elsevier BV; Volume: 278; Issue: 29 Linguagem: Inglês
10.1074/jbc.m302764200
ISSN1083-351X
AutoresW. Sean Davidson, George M. Hilliard,
Tópico(s)Diabetes, Cardiovascular Risks, and Lipoproteins
ResumoThe three-dimensional structure of human apoA-I on nascent, discoidal HDL particles has been debated extensively over the past 25 years. Recent evidence has demonstrated that the α-helical domains of apoA-I are arranged in a belt-like orientation with the long axis of the helices perpendicular to the phospholipid acyl chains on the disc edge. However, experimental information on the spatial relationships between apoA-I molecules on the disc is lacking. To address this issue, we have taken advantage of recent advances in mass spectrometry technology combined with cleavable cross-linking chemistry to derive a set of distance constraints suitable for testing apoA-I structural models. We generated highly homogeneous, reconstituted HDL particles containing two molecules of apoA-I. These were treated with a thiol-cleavable cross-linking agent, which covalently joined Lys residues in close proximity within or between molecules of apoA-I in the disc. The cross-linked discs were then exhaustively trypsinized to generate a discrete population of peptides. The resulting peptides were analyzed by liquid chromatography/mass spectrometry before and after cleavage of the cross-links, and resulting peaks were identified based on the theoretical tryptic cleavage of apoA-I. We identified at least 8 intramolecular and 7 intermolecular cross-links in the particle. The distance constraints are used to analyze three current models of apoA-I structure. The results strongly support the presence of the salt-bridge interactions that were predicted to occur in the "double belt" model of apoA-I, but a helical hairpin model containing the same salt-bridge docking interface is also consistent with the data. The three-dimensional structure of human apoA-I on nascent, discoidal HDL particles has been debated extensively over the past 25 years. Recent evidence has demonstrated that the α-helical domains of apoA-I are arranged in a belt-like orientation with the long axis of the helices perpendicular to the phospholipid acyl chains on the disc edge. However, experimental information on the spatial relationships between apoA-I molecules on the disc is lacking. To address this issue, we have taken advantage of recent advances in mass spectrometry technology combined with cleavable cross-linking chemistry to derive a set of distance constraints suitable for testing apoA-I structural models. We generated highly homogeneous, reconstituted HDL particles containing two molecules of apoA-I. These were treated with a thiol-cleavable cross-linking agent, which covalently joined Lys residues in close proximity within or between molecules of apoA-I in the disc. The cross-linked discs were then exhaustively trypsinized to generate a discrete population of peptides. The resulting peptides were analyzed by liquid chromatography/mass spectrometry before and after cleavage of the cross-links, and resulting peaks were identified based on the theoretical tryptic cleavage of apoA-I. We identified at least 8 intramolecular and 7 intermolecular cross-links in the particle. The distance constraints are used to analyze three current models of apoA-I structure. The results strongly support the presence of the salt-bridge interactions that were predicted to occur in the "double belt" model of apoA-I, but a helical hairpin model containing the same salt-bridge docking interface is also consistent with the data. High plasma levels of high density lipoprotein (HDL) 1The abbreviations used are: HDL, high density lipoprotein; apoA-I, apolipoprotein A-I; DSP, dithiobis(succinimidyl propionate); DTT, dithiothreitol; HPLC, high pressure liquid chromatography; MS, mass spectroscopy; LCMS, liquid chromatography mass spectrometry; SPB, standard phosphate buffer; TOF, time of flight; TIC, total ion chromatogram; POPC, 1-palmitoyl 2-oleoyl phosphatidylcholine; rHDL, reconstituted HDL; DTT, dithiothreitol. are widely thought to be protective against human cardiovascular disease. Apolipoprotein (apo)A-I, a 243-amino acid, 28-kDa protein is a key mediator of HDL function. It is required for lecithin:cholesterol acyl transferase-mediated maturation of HDL and may be a major ligand by which cholesteryl esters are delivered to the liver via the scavenger receptor type B class 1 receptor (1Xu S. Laccotripe M. Huang X. Rigotti A. Zannis V.I. Krieger M. J. Lipid Res. 1997; 38: 1289-1298Abstract Full Text PDF PubMed Google Scholar). In addition, the recent discovery of the importance of the ATP binding cassette protein in lipid transport (2Rust S. Rosier M. Funke H. Real J. Amoura Z. Piette J.C. Deleuze J.F. Brewer H.B. Duverger N. Denefle P. Assmann G. Nat. Genet. 1999; 22: 352-355Crossref PubMed Scopus (1275) Google Scholar, 3Bodzioch M. Orso E. Klucken J. Langmann T. Bottcher A. Diederich W. Drobnik W. Barlage S. Buchler C. Porsch-Ozcurumez M. Kaminski W.E. Hahmann H.W. Oette K. Rothe G. Aslanidis C. Lackner K.J. Schmitz G. Nat. Genet. 1999; 22: 347-351Crossref PubMed Scopus (1355) Google Scholar, 4Brooks-Wilson A. Marcil M. Clee S.M. Zhang L.H. Roomp K. van Dam M. Yu L. Brewer C. Collins J.A. Molhuizen H.O. Loubser O. Ouelette B.F. Fichter K. Ashbourne-Excoffon K.J. Sensen C.W. Scherer S. Mott S. Denis M. Martindale D. Frohlich J. Morgan K. Koop B. Pimstone S. Kastelein J.J. Hayden M.R. Nat. Genet. 1999; 22: 336-345Crossref PubMed Scopus (1517) Google Scholar) indicates that an important apoA-I to cell surface interaction may occur during HDL formation and reverse cholesterol transport. One of the major obstacles to a better understanding of these interactions has been the paucity of detailed structural information for apoA-I in its various states of lipid association. Homogeneous discoidal forms of HDL are easily reconstituted from purified protein and lipids in vitro (5Jonas A. Methods Enzymol. 1986; 128: 553-582Crossref PubMed Scopus (299) Google Scholar), and these reconstituted HDL (rHDL) analogs have been used heavily for structural studies. The discs likely exist as a phospholipid/cholesterol bilayer surrounded at its edges by the hydrophobic regions of the amphipathic helices of apoA-I (for a review see Ref. 6Brouillette C.G. Anantharamaiah G.M. Engler J.A. Borhani D.W. Biochim. Biophys. Acta. 2001; 1531: 4-46Crossref PubMed Scopus (233) Google Scholar). Segrest et al. and others (7Segrest J.P. Chem. Phys. Lipids. 1977; 18: 7-22Crossref PubMed Scopus (95) Google Scholar, 8Wlodawer A. Segrest J.P. Chung B.H. Chiovetti R.J. Weinstein J.N. FEBS Lett. 1979; 104: 231-235Crossref PubMed Scopus (61) Google Scholar) proposed in the late 1970s that the α-helices of apoA-I were arranged around the disc circumference with the long axis of the helices perpendicular to the acyl chains. This became known as the "belt" or "bicycle wheel" model. Alternatively, other investigators theorized that the 22-amino acid helical repeats could traverse the bilayer edge with the helices parallel to the acyl chains (9Tall A.R. Small D.M. Deckelbaum R.J. Shipley G.G. J. Biol. Chem. 1977; 252: 4701-4711Abstract Full Text PDF PubMed Google Scholar). This "picket fence" model was challenged by the first successful x-ray crystal structure of a lipid-free fragment of apoA-I by Borhani et al. (10Borhani D.W. Rogers D.P. Engler J.A. Brouillette C.G. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12291-12296Crossref PubMed Scopus (417) Google Scholar). The crystal structure showed a tetramer of highly α-helical apoA-I molecules arranged in a ring-shaped complex, with no evidence of hairpin turns. Borhani et al. hypothesized that the ring motif in the crystal structure could be applied to the case of lipid-bound apoA-I on a disc. Since then, a belt-like orientation for the helices of apoA-I has been supported by polarized infrared spectroscopy experiments performed by Koppaka et al. (11Koppaka V. Silvestro L. Engler J.A. Brouillette C.G. Axelsen P.H. J. Biol. Chem. 1999; 274: 14541-14544Abstract Full Text Full Text PDF PubMed Scopus (126) Google Scholar). In addition, we published a series of studies in which various single tryptophan mutants of apoA-I were analyzed in discoidal HDL particles containing phospholipids with quenching groups at various positions along the acyl chain. The results clearly showed that all eight 22-amino acid helices in apoA-I were oriented perpendicular to the phospholipid acyl chains (12Panagotopulos S.E. Horace E.M. Maiorano J.N. Davidson W.S. J. Biol. Chem. 2001; 276: 42965-42970Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 13Maiorano J.N. Davidson W.S. J. Biol. Chem. 2000; 275: 17374-17380Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). With the question of apoA-I helical orientation addressed, attention has focused on determining spatial relationships between two molecules of apoA-I on a disc. Segrest et al. (14Segrest J.P. Jones M.K. Klon A.E. Sheldahl C.J. Hellinger M. De Loof H. Harvey S.C. J. Biol. Chem. 1999; 274: 31755-31758Abstract Full Text Full Text PDF PubMed Scopus (304) Google Scholar) recently published a computer model referred to as the "double belt" model for a reconstituted HDL particle containing two molecules of apoA-I. In this model, two ring-shaped molecules of apoA-I are stacked on top of each other with both molecules forming an almost continuous helix that wraps around the perimeter of the phospholipid disc in an anti-parallel orientation. Computer analysis of the model predicted a particular registry between the monomers resulting in the greatest potential for salt bridge connections between the two molecules. An alternative belt-like model, initially suggested by Brouillette (15Brouillette C.G. Anantharamaiah G.M. Biochim. Biophys. Acta. 1995; 1256: 103-129Crossref PubMed Scopus (168) Google Scholar), predicts that apoA-I molecules are arranged in a hairpin orientation. In this model, about half of the molecule interacts with one leaflet, there is a turn, and the other half runs anti-parallel to the first on the opposing leaflet. This idea was supported by fluorescence energy transfer experiments performed by Tricerri et al. (16Tricerri M.A. Behling Agree A.K. Sanchez S.A. Bronski J. Jonas A. Biochemistry. 2001; 40: 5065-5074Crossref PubMed Scopus (84) Google Scholar). The model preserves the potential for stabilizing salt bridge interactions between the same residues that were proposed for the double belt, although these must occur intramolecularly in the hairpin model. We have proposed a third model termed the "Z" belt orientation (13Maiorano J.N. Davidson W.S. J. Biol. Chem. 2000; 275: 17374-17380Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). This arrangement is similar to the hairpin except that, instead of traversing back along itself, the molecule proceeds in the same direction on the opposing leaflet, giving the potential for interlocking interactions between the molecules. Although traditional spectroscopic methods such as fluorescence and circular dichroism have proven useful for studying the generalities of apoA-I structure in rHDL, data from these approaches are not suitable for high resolution modeling. Nuclear magnetic resonance (NMR) and x-ray crystallography data would be very useful, but these techniques have not yet been successfully applied to native rHDL particles. However, Bennett et al. (17Bennett K.L. Kussmann M. Bjork P. Godzwon M. Mikkelsen M. Sorensen P. Roepstorff P. Protein Sci. 2000; 9: 1503-1518Crossref PubMed Scopus (134) Google Scholar) recently reported an elegant study that demonstrated the power of combining high precision mass spectrometry/peptide analysis with cross-linking chemistry to identify sites of interaction between two protein molecules. Young et al. (18Young M.M. Tang N. Hempel J.C. Oshiro C.M. Taylor E.W. Kuntz I.D. Gibson B.W. Dollinger G. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 5802-5806Crossref PubMed Scopus (395) Google Scholar) used a similar approach in combination with a sequence threading technique to generate a structure of monomeric human fibroblast growth factor that matched well with the known NMR structure of the protein. In this work, we report the successful adaptation of this approach to the problem the spatial relationships of two molecules of apoA-I on the edge of a discoidal HDL particle. The results provide the most comprehensive determinations of distance constraints within an rHDL particle and strongly confirm the presence of the salt bridge interactions predicted by Segrest et al. (14Segrest J.P. Jones M.K. Klon A.E. Sheldahl C.J. Hellinger M. De Loof H. Harvey S.C. J. Biol. Chem. 1999; 274: 31755-31758Abstract Full Text Full Text PDF PubMed Scopus (304) Google Scholar) present in both the double belt and hairpin models. ApoA-I Purification—Purified human plasma apoA-I was obtained from human HDL (1.21 < density > 1.062 g/ml) isolated as reported (19Lund-Katz S. Phillips M.C. Biochemistry. 1986; 25: 1562-1568Crossref PubMed Scopus (95) Google Scholar). Briefly, HDL was freeze-dried and extracted with chloroform/methanol. The pellet was suspended in 10 mm Tris HCl with 6 m urea and applied to a Q-Sepharose column (XK 2.6/40, Amersham Biosciences) pre-equilibrated and eluted at 4 ml/min at room temperature in the same buffer. Fractions containing apoA-I as determined by SDS-PAGE electrophoresis were dialyzed into 5 mm ammonium bicarbonate buffer and freeze-dried. Proteins were solubilized in 3 m guanidine for 1 h and then dialyzed into standard phosphate buffer (SPB) (20 mm sodium phosphate, 0.15 m NaCl, pH. 7.8) prior to use in reconstitution experiments. Preparation of rHDL Particles—Reconstituted HDL (rHDL) particles were prepared using 1-palmitoyl 2-oleoyl phosphatidylcholine (POPC) (Avanti Polar Lipids, Alabaster, AL) at lipid to protein molar ratios of 90:1 according to the method of Jonas (5Jonas A. Methods Enzymol. 1986; 128: 553-582Crossref PubMed Scopus (299) Google Scholar). Lipids were dried under nitrogen and resuspended in SPB. Deoxycholate (Fisher, deoxycholate: lipid, 1.3:1, w/w) was added and incubated at 37 °C for 1.5 h with mild vortexing every 15 min. The protein was added and incubated a 37 °C for 1 h. The cholate was removed by dialysis against SPB (5 changes of 2 liters for at least 4 h each at 4 °C). The particles were analyzed on a non-denaturing, native polyacrylamide Phast gel (Amersham Biosciences, Piscataway, NJ) (20Davidson W.S. Gillotte K.L. Lund-Katz S. Johnson W.J. Rothblat G.H. Phillips M.C. J. Biol. Chem. 1995; 270: 5882-5890Abstract Full Text Full Text PDF PubMed Scopus (139) Google Scholar). Before cross-linking, the particles were passed down a Superdex 200 gel filtration column (Amersham Biosciences) to remove unreacted protein and lipid. Fractions corresponding to the 96-Å diameter complex were pooled and concentrated by filtration. The phosphorus method of Sokolof and Rothblat (21Sokoloff L. Rothblat G.H. Proc. Soc. Exp. Biol. Med. 1974; 146: 1166-1172Crossref PubMed Scopus (109) Google Scholar) and the Markwell modification of the Lowry assay (22Markwell M.A. Haas S.M. Bieber L.L. Tolbert N.E. Anal. Biochem. 1978; 87: 206-210Crossref PubMed Scopus (5369) Google Scholar) determined the final phospholipid and protein concentrations, respectively. The atomic phosphorus standard was obtained from Sigma (St, Louis, MO). Cross-linking, Reduction, and Generation of Tryptic Peptides—DSP (dithiobis(succinimidyl propionate)) (Pierce) was weighed out and dissolved in ice-cold Me2SO to a concentration of 6.5 mg/ml and used within 5 min. A 7:1 molar ratio of DSP to apoA-I protein was added to a solution containing human apoA-I/POPC discs in SPB on ice at a concentration between 0.5 and 1.0 mg/ml. The reaction was incubated at 4 °C for 24 h with periodic vortexing. The reaction was quenched by adding a stock of 1 m Tris, pH.7.8, to a final Tris concentration of 100 mm. The samples were dialyzed into 5 mm ammonium bicarbonate to remove any unreacted cross-linker and were lyophilized to dryness. The lipids were extracted with chloroform/methanol, and the protein fraction was solubilized in SPB with 3 m guanidine HCl. In some experiments, the monomeric protein was separated from the dimeric protein after cross-linking by passage down a Superdex 200 column equilibrated in the same buffer. Fractions corresponding to the dimer and monomer were collected, concentrated, and dialyzed into SPB. Each cross-linked sample was split in two equal fractions. One was labeled "reduced" and incubated with 25 mm dithiothreitol (DTT) from a stock solution in water for 2 h at 37 °C. The other fraction was labeled "x-linked" and treated identically except receiving the same volume of water instead of DTT. Both samples were dialyzed against 5 mm ammonium bicarbonate (in separate containers). 5% (weight of trypsin to apoA-I) sequencing grade trypsin (Promega) was allowed to digest the protein at 37 °C for 2 h. The samples were lyophilized to dryness in a microcentrifuge tube in 100-μg aliquots. The samples were stored at -20 °C until used. Mass Spectrometry—Liquid chromatography mass spectrometry (LCMS) experiments were carried out on a Sciex QSTAR DE (quadrupole time-of-flight (TOF)) mass spectrometer fitted with an atmospheric electrospray ionizer controlled by using Analyst QS 1.1 software (Applied Biosystems). The spectrometer was programmed for TOF-MS scans from 100 to 2800 atomic mass units at a 1.0-s accumulation time. An Agilent 1100 capillary HPLC with an Agilent ZORBAX SB-C18 0.5-mm × 15-cm reverse phase column at a flow rate of 7.5 μl per min was used to separate tryptic peptides prior to introduction into the mass spectrometer by complex peptide gradient chromatography. Lyophilized samples were solubilized in mobile phase A (distilled water with 0.1% trifluoroacetic acid) and eluted with a 0–100% gradient of mobile phase B (95% acetonitrile in water with 0.085% trifluoroacetic acid). 160 pmol of protein was injected per run. The mass spectra were internally calibrated by a three-point linear method based on the monoisotopic masses of the following peptides derived from apoA-I: 154–160 (mass of 780.4242 Da), 161–171 (1300.6412 Da), and 62–77 (1931.9265 Da). These peptides were used because they exhibited prominent peaks in all samples and covered much of the expected mass range. Data Analysis—Using the Analyst QS software, individual mass spectra were generated for each peak in the total ion chromatograph (TIC) for each sample. Masses that exhibited an intensity of at least 25 detector counts were recorded and identified in terms of ion type (M+H, M+2H, etc.). Monoisotopic masses were determined by averaging the ion series for each mass. The resulting list of masses was analyzed by the software GPMAW (ChemSW, Inc.) to assign a putative amino acid sequence identity as either an unmodified peptide of apoA-I or one or more peptides containing one or more DSP modifications (see Table I). To identify cross-links occurring between two peptides, a spreadsheet was used to sum the masses of all possible combinations of peptides and one or more intervening cross-links. The resulting data base of theoretical cross-linked peptide masses was then searched for a given experimental mass. Mass identity assignments were made using the following criteria: 1) An assignment was made only if the experimental mass matched the theoretical mass within 40 ppm. This cutoff was sufficient to identify all single cleavage peptides in control experiments using unmodified apoA-I that had been completely trypsinized. 2) All peptides present in a cross-link must contain a Lys residue (not at the C terminus). 3) It was assumed that trypsin does not cleave on C-terminal side of a modified Lys residue (17Bennett K.L. Kussmann M. Bjork P. Godzwon M. Mikkelsen M. Sorensen P. Roepstorff P. Protein Sci. 2000; 9: 1503-1518Crossref PubMed Scopus (134) Google Scholar). This feature of trypsin was advantageous because it reduced the complexity of the peptide mixture in the cross-linked sample. 4) Putative cross-links were assigned to masses present in the cross-linked sample but only if they completely disappeared when the sample was reduced with DTT. 5) Identities were assigned for putative intermolecular or intramolecular cross-links only if all peptide components were recovered with the appropriate number of reduced cross-links on eligible Lys residues after DTT reduction. 6) It was assumed that trypsin fully cleaved the protein at every opportunity, with no partials.Table IPossible peptide modifications resulting from DSP cross-linkingType of modificationaSee the text for a description of how cross-links were assigned as intra- or intermolecular. All masses are expressed as the monoisotopic massPeptide component(s)Mass addition due to DSPDa per eventIntrapeptide cross-link1 peptide only173.9809Intramolecular cross-link2+ peptides in same apoA-1 molecule separated by at least one trypsin cleavage site173.9809Intermolecular cross-link2+ peptides on different molecules of apoA-1 on the same rHDL particle173.9809Hydrolyzed cross-link1 peptide191.9915Reduced cross-link1 peptide87.9983a See the text for a description of how cross-links were assigned as intra- or intermolecular. All masses are expressed as the monoisotopic mass Open table in a new tab Distinguishing Intraversus Intermolecular Cross-links—Intra-versus intermolecular cross-links were assigned by comparing the intensity of a given mass between spectra generated from dimeric and monomeric forms of apoA-I isolated from a cross-linked rHDL particle by gel filtration. The maximal detector count intensities for the entire ion series for a given mass were summed for the dimeric and monomeric cross-linked samples, respectively (23Kapron J.T. Hilliard G.M. Lakins J.N. Tenniswood M.P. West K.A. Carr S.A. Crabb J.W. Protein Sci. 1997; 6: 2120-2133Crossref PubMed Scopus (102) Google Scholar). The ratio of the dimeric/monomeric intensities was used to determine if a mass was substantially less prevalent in the monomeric versus the dimeric cross-linked protein. For a particular mass, an intensity ratio below 1.7 was identified as a putative intramolecular cross-link. By contrast, an intermolecular cross-link was proposed for ratios higher than 1.7. The appearance of small amounts of intermolecular cross-links in the monomeric sample was due to slight contamination of the dimeric form in the sample (see Fig. 3). The Approach—The case of purified apoA-I on a well-defined rHDL particle with two molecules of apoA-I is essentially a homodimeric non-covalent interaction similar to that studied by Bennett et al. (17Bennett K.L. Kussmann M. Bjork P. Godzwon M. Mikkelsen M. Sorensen P. Roepstorff P. Protein Sci. 2000; 9: 1503-1518Crossref PubMed Scopus (134) Google Scholar) using mass spectrometry. ApoA-I contains 21 Lys residues that are generally evenly spread throughout the molecule, making it a manageable candidate for this approach. The homodimer is first incubated with the homobifunctional cross-linking reagent DSP, which reacts with the ϵ-amine group of lysine residues. Numerous cross-links randomly form both intra- and intermolecularly depending on the number Lys residues within the reagent's spacer arm length of 12 Å. In addition, the cross-linker may bind to one Lys residue but fail to cross-link to a second Lys residue before spontaneous hydrolysis of the cross-linker (17Bennett K.L. Kussmann M. Bjork P. Godzwon M. Mikkelsen M. Sorensen P. Roepstorff P. Protein Sci. 2000; 9: 1503-1518Crossref PubMed Scopus (134) Google Scholar). Once cross-linked, trypsin is used to cleave after Arg and Lys residues to generate a population of peptides, some of which are unmodified whereas others are cross-linked. An aliquot of the peptide mixture is treated with the reducing agent DTT to cleave the disulfide linkage within DSP (Fig. 1) to liberate any joined peptides. A second aliquot is left untreated. Both the cross-linked and reduced peptide mixtures are then separated by reverse phase HPLC and analyzed immediately upon elution by electrospray MS. Highly accurate mass spectra are taken for each peak as they elute from the column. If a particular mass is present in both chromatograms, it represents a peptide that was not modified by DSP. However, masses appearing in the cross-linked chromatogram, but not in the reduced chromatogram, indicate the presence of cross-linked peptides that were cleaved by DTT. Masses appearing in the reduced chromatogram are due to the newly freed peptides. The mass data are used to assign a sequence identity to each peptide by comparing the experimentally derived peptide mass to a data base of theoretical peptide masses generated from the known protein sequence and the known cleavage sites of trypsin. Unmodified peptides exhibit masses equal to the sum of the amino acids in the peptide sequence. The mass of a cross-linked peptide complex is the sum of each component peptide mass plus an intact cross-link (see Table I). Similarly, the mass of peptides containing cleaved cross-links is the sum of the peptide mass plus a reduced cross-link. After peptide identification, one can deduce that a cross-link was formed between peptides X and Y in the native dimer. If each peptide had a single Lys residue in the sequence, then one can conclude that the two Lys residues were within about 12 Å in the native protein structure. rHDL Particle Reconstitution, Characterization, and Cross-linking Optimization—We generated a simple discoidal reconstituted rHDL particle with a diameter of 96 Å that is well known to contain two molecules of apoA-I and about 160 molecules of POPC (24Jonas A. Toohill K.L. Krul E.S. Wald J.H. Kezdy K.E. J. Biol. Chem. 1990; 265: 22123-22129Abstract Full Text PDF PubMed Google Scholar, 25McGuire K.A. Davidson W.S. Jonas A. J. Lipid Res. 1996; 37: 1519-1528Abstract Full Text PDF PubMed Google Scholar). This particle was selected because it is stable, easily produced in high yield in vitro, and has been used for computer simulation studies testing various models for apoA-I structure (14Segrest J.P. Jones M.K. Klon A.E. Sheldahl C.J. Hellinger M. De Loof H. Harvey S.C. J. Biol. Chem. 1999; 274: 31755-31758Abstract Full Text Full Text PDF PubMed Scopus (304) Google Scholar, 26Phillips J.C. Wriggers W. Li Z. Jonas A. Schulten K. Biophys. J. 1997; 73: 2337-2346Abstract Full Text PDF PubMed Scopus (118) Google Scholar, 27Brasseur R. Lins L. Vanloo B. Ruysschaert J.M. Rosseneu M. Proteins. 1992; 13: 246-257Crossref PubMed Scopus (60) Google Scholar). Fig. 2A shows a non-denaturing PAGE analysis demonstrating the homogeneity of this particle. Its diameter measured 96 ± 3 Å, and it contained a final POPC to apoA-I ratio of 79 ± 4:1. To work out the conditions for cross-linking the apoA-I molecules on the complex with DSP, pilot experiments were performed in which the molar ratio of DSP to apoA-I was varied from 2.5:1 to 50:1 (data not shown). Ratios above 10:1 were sufficient to cross-link about 97% of the apoA-I to a dimeric form as visualized by SDS-PAGE. Higher ratios could drive the reaction completely to dimers. We chose a ratio of 7:1 for further experiments to minimize the chances of perturbing the particle structure and to give an opportunity to study the monomeric form of the cross-linked apoA-I (see below). Cross-linking at 4 °C was found to reduce the heterogeneity of the dimeric band versus incubations at room temperature probably by minimizing thermal motions within the particle. Fig. 2B shows the results of a typical cross-linking experiment under these conditions. ApoA-I (28 kDa) in the rHDL particle was cross-linked to a dimer (56 kDa), which could be mostly reduced back to a monomer by cleavage of the cross-link with DTT. We consistently observed a small amount of trimer (84 kDa) in these particles (Fig. 2, lane 3, top band) of about 5% of the total staining. We believe that this arose from a slight contamination of the 108-Å rHDL particles in our preparation, a complex with three molecules of apoA-I (24Jonas A. Toohill K.L. Krul E.S. Wald J.H. Kezdy K.E. J. Biol. Chem. 1990; 265: 22123-22129Abstract Full Text PDF PubMed Google Scholar). A small percentage (about 5%) of the apoA-I remained as a monomer under these conditions. The appearance of the dimer upon DSP cross-linking was independent of the rHDL concentration between 0.5 to 5.0 mg/ml (data not shown), indicating that the cross-links were formed within rHDL particles and not between two rHDL particles (28Swaney J.B. O'Brien K. J. Biol. Chem. 1978; 253: 7069-7077Abstract Full Text PDF PubMed Google Scholar). The presence of the cross-links did not change the average helical content of apoA-I from that of the unmodified form as measured by circular dichroism (both were about 75% helical), arguing that the cross-links did not significantly perturb the structure of the protein. Furthermore, we found that the ratio of cross-linker to apoA-I did not significantly affect the cross-links identifiable by MS (data not shown). To get a sense of which experimental mass values originate from intermolecular versus intramolecular cross-links, a sample of cross-linked rHDL particles (i.e. the sample in Fig. 2, lane 3) was delipidated and then separated into the component monomeric and dimeric species by gel filtration chromatography (see "Discussion"). Fig. 3 shows that cross-linked apoA-I was resolved into two fractions, which will hereafter be referred to as the "cross-linked dimer" sample and the "cross-linked monomer" sample, respectively. Despite our best efforts, there was a slight contamination of <5% of the dimeric form in the cross-linked monomer fraction. Mass Spectrometry—Four different delipidated apoA-I samples were prepared from rHDL particle preparations. They were the unmodified monomer, DSP cross-linked dimer, DSP cross-linked monomer, or cross-linked then reduced by DTT (reduce
Referência(s)