Intramolecular VersusIntermolecular Disulfide Bonds in Prion Proteins
2002; Elsevier BV; Volume: 277; Issue: 36 Linguagem: Inglês
10.1074/jbc.m204273200
ISSN1083-351X
AutoresErvin Welker, Lynne D. Raymond, Harold A. Scheraga, Byron Caughey,
Tópico(s)Porphyrin Metabolism and Disorders
ResumoPrion protein (PrP) is the major component of the partially protease-resistant aggregate that accumulates in mammals with transmissible spongiform encephalopathies. The two cysteines of the scrapie form, PrPSc, were found to be in their oxidized (i.e. disulfide) form (Turk, E., Teplow, D. B., Hood, L. E., and Prusiner, S. B. (1988)Eur. J. Biochem. 176, 21–30); however, uncertainty remains as to whether the disulfide bonds are intra- or intermolecular. It is demonstrated here that the monomers of PrPSc are not linked by intermolecular disulfide bonds. Furthermore, evidence is provided that PrPSc can induce the conversion of the oxidized, disulfide-intact form of the monomeric cellular prion protein to its protease-resistant form without the temporary breakage and subsequent re-formation of the disulfide bonds in cell-free reactions. Prion protein (PrP) is the major component of the partially protease-resistant aggregate that accumulates in mammals with transmissible spongiform encephalopathies. The two cysteines of the scrapie form, PrPSc, were found to be in their oxidized (i.e. disulfide) form (Turk, E., Teplow, D. B., Hood, L. E., and Prusiner, S. B. (1988)Eur. J. Biochem. 176, 21–30); however, uncertainty remains as to whether the disulfide bonds are intra- or intermolecular. It is demonstrated here that the monomers of PrPSc are not linked by intermolecular disulfide bonds. Furthermore, evidence is provided that PrPSc can induce the conversion of the oxidized, disulfide-intact form of the monomeric cellular prion protein to its protease-resistant form without the temporary breakage and subsequent re-formation of the disulfide bonds in cell-free reactions. transmissible spongiform encephalopathy prion protein cellular form of the PrP protease-sensitive PrP protease-resistant PrP disease-associated scrapie form of the PrP 2-aminoethyl-methanethiosulfonate 1,4 dithiothreitol reduced DTT proteinase K N-ethylmaleimide indicates the absence of the glycosyl-phosphatidyl-inositol anchor that attaches the protein to the cell membrane Transmissible spongiform encephalopathies (TSE)1 are fatal neurodegenerative diseases that include sporadic and familial Creutzfeldt-Jakob disease, kuru, and Gerstmann-Straussler-Scheinker syndrome in humans, scrapie in sheep and goats, and bovine spongiform encephalopathy in cattle (1Prusiner S.B. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 13363-13383Crossref PubMed Scopus (5168) Google Scholar, 2Hope J. Curr. Opin. Genet. Dev. 2000; 10: 568-574Crossref PubMed Scopus (18) Google Scholar). TSE is associated with the accumulation of an abnormal isoform of prion protein (PrP) (3Caughey B. Trends Biochem. Sci. 2001; 26: 235-242Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). The conformational transition of the normal form of the protein (PrPC), which contains both an α-helical and a flexibly disordered portion (4Zahn R. Liu A. Lu¨hrs T. Riek R. von Schroetter C. Garcıáa F.L. Billeter M. Calzolai L. Wider G. Wu¨thrich K. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 145-150Crossref PubMed Scopus (952) Google Scholar), to a highly β-sheet conformation (PrPSc) (5Caughey B. Raymond G., J. Bessen R.A. J. Biol. Chem. 1998; 273: 32230-32235Abstract Full Text Full Text PDF PubMed Scopus (284) Google Scholar) is the central event of the disease that seems to be responsible for its neuropathology. Whereas PrPC is protease sensitive (PrPsen), PrPSc is partially resistant to proteolysis (PrPres) and prone to form high-molecular-weight aggregates (6Horiuchi M. Caughey B. Structure Fold. Des. 1999; 7: R231-R240Abstract Full Text Full Text PDF Scopus (81) Google Scholar). This latter feature makes it difficult to analyze the conformational transition that leads to the formation of PrPres or to characterize it. In PrPsen, a disulfide bond stabilizes the overall fold of the protein by connecting the two long C-terminal helices (4Zahn R. Liu A. Lu¨hrs T. Riek R. von Schroetter C. Garcıáa F.L. Billeter M. Calzolai L. Wider G. Wu¨thrich K. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 145-150Crossref PubMed Scopus (952) Google Scholar). Some experiments have suggested that the presence of this sole intramolecular disulfide bond in PrPsen is required for conversion to PrPres (7Muramoto T. Scott M. Cohen F.E. Prusiner S.B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 15457-15462Crossref PubMed Scopus (176) Google Scholar, 8Herrmann L.M. Caughey B. Neuroreport. 1998; 9: 2457-2461Crossref PubMed Scopus (86) Google Scholar, 9Maiti N.R. Surewicz W.K. J. Biol. Chem. 2001; 276: 2427-2431Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar). By contrast, several communications have discussed the possibility that the disulfide bond of PrPC is reduced during the conversion (6Horiuchi M. Caughey B. Structure Fold. Des. 1999; 7: R231-R240Abstract Full Text Full Text PDF Scopus (81) Google Scholar, 10Mehlhorn I. Groth D. Sto¨ckel J. Moffat B. Reilly D. Yansura D. Willett W.S. Baldwin M. Fletterick R. Cohen F.E. Vandlen R. Henner D. Prusiner S.B. Biochemistry. 1996; 35: 5528-5537Crossref PubMed Scopus (194) Google Scholar, 11Jackson G.S. Hosszu L.L.P. Power A. Hill A.F. Kenney J. Saibil H. Craven C.J. Waltho J.P. Clarke A.R. Collinge J. Science. 1999; 283: 1935-1937Crossref PubMed Scopus (365) Google Scholar, 12Feughelman M. Willi B.K. J. Theor. Biol. 2000; 206: 313-315Crossref PubMed Scopus (18) Google Scholar, 13Welker E. Wedemeyer W.J. Scheraga H.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 4334-4336Crossref PubMed Scopus (62) Google Scholar). The reduction of the disulfide bond may provide flexibility for the protein molecule, which may be necessary for the conversion to the β conformation. The cysteines of the prion protein in the scrapie form were assumed to be involved in a disulfide bond because denaturant-solubilized PrPres reacts with thiol-specific reagents only after treatment with reducing agents (14Turk E. Teplow D.B. Hood L.E. Prusiner S.B. Eur. J. Biochem. 1988; 176: 21-30Crossref PubMed Scopus (270) Google Scholar), suggesting that the possible reduction of this disulfide bond may be just a temporary event. Whether the (re-formed) disulfide bonds in PrPres are intermolecular (between monomers) or intramolecular (within monomers) is a critical issue that has important consequences for understanding the molecular mechanism of the conformational transition. The extreme stability of PrPres, as well as other features of the protein, and the PrPsen to PrPresconversion would be more easily understandable if PrPreswere a covalently linked polymer. However, this possibility has been discounted for a long time because it seems to contradict the fact that, upon denaturation, PrP monomers with an intact (intramolecular) disulfide bond dissociate from an infectious PrPrespreparation (14Turk E. Teplow D.B. Hood L.E. Prusiner S.B. Eur. J. Biochem. 1988; 176: 21-30Crossref PubMed Scopus (270) Google Scholar, 15Safar J. Wang W. Padgett M.P. Ceroni M. Piccardo P. Zopf D. Gajdusek D.C. Gibbs C.J., Jr. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 6373-6377Crossref PubMed Scopus (83) Google Scholar). On the other hand, it was shown recently how, in theory, a putative disulfide-linked polymer of the prion protein could depolymerize into monomers without any added reducing agent (13Welker E. Wedemeyer W.J. Scheraga H.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 4334-4336Crossref PubMed Scopus (62) Google Scholar). This is because PrP has only two cysteines; thus, one of the two cysteines at the end of a putative disulfide-linked PrPres polymer would not be involved in a disulfide bond. A similar situation frequently arises when a particular protein has both a disulfide bond(s) and a thiol(s). The protein is stable when folded. The three-dimensional structure inhibits thiol-disulfide exchange by sequestering the disulfide bond(s) from the intramolecular attack by the thiol(s). However, when the folded structure of the protein is denatured the thiol attacks the now-exposed disulfide bond, forming a new disulfide bond and freeing the thiol of another cysteine residue. This fast intramolecular reshuffling persists in the unfolded protein, resulting in a dynamic equilibrium involving different disulfide-bonded isomers. This is a well characterized phenomenon (16Welker E. Narayan M. Wedemeyer W.J. Scheraga H.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 2312-2316Crossref PubMed Scopus (91) Google Scholar, 17Saito K. Welker E. Scheraga H.A. Biochemistry. 2001; 40: 15002-15008Crossref PubMed Scopus (19) Google Scholar). In this manner, this reshuffling reaction could take place in a linear disulfide-bonded polymer of PrPres upon denaturation. However, in such a polymer the attack by the thiol on the disulfide leads to dissociation of a monomer with an intramolecular disulfide bond, leaving another free thiol at the end of the polymer. The next reshuffling reaction releases another monomer giving rise to a self-depolymerization chain-reaction in the unfolded polymer. Therefore, this depolymerization reaction would lead such a disulfide-linked linear polymer to dissociate preferentially into monomers with intramolecular disulfide bonds (13Welker E. Wedemeyer W.J. Scheraga H.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 4334-4336Crossref PubMed Scopus (62) Google Scholar). Consequently, PrPres will appear as a monomer upon solubilization with a denaturating agent regardless of whether it has an intra- or intermolecular disulfide bond. Thus, to distinguish between intra- and intermolecular disulfide bonds, this putative depolymerization through thiol/disulfide exchange must be prevented by blocking the terminal thiols when PrPres is denatured. In the early work of Turk et al. (14Turk E. Teplow D.B. Hood L.E. Prusiner S.B. Eur. J. Biochem. 1988; 176: 21-30Crossref PubMed Scopus (270) Google Scholar) PrPres was solubilized by heating in 2% SDS, and the solubilized fraction that might have contained the depolymerized monomers was analyzed for intra- or intermolecular disulfide bonds by SDS-PAGE. It is worth noting that solubilization in SDS loading buffer alone could lead a putative PrPres disulfide-bonded polymer to dissociate preferentially into monomers with an apparently intact disulfide bond. The available biochemical data do not distinguish between intra- or intermolecular disulfide bonds (14Turk E. Teplow D.B. Hood L.E. Prusiner S.B. Eur. J. Biochem. 1988; 176: 21-30Crossref PubMed Scopus (270) Google Scholar, 15Safar J. Wang W. Padgett M.P. Ceroni M. Piccardo P. Zopf D. Gajdusek D.C. Gibbs C.J., Jr. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 6373-6377Crossref PubMed Scopus (83) Google Scholar), raising the interesting possibility that PrPres may have intermolecular disulfide bonds (13Welker E. Wedemeyer W.J. Scheraga H.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 4334-4336Crossref PubMed Scopus (62) Google Scholar). The experiments in the present study were therefore undertaken to gain more information about the possible involvement of disulfide exchange reactions in the conformational transition of the prion protein. Specifically, we asked two questions: Is the disulfide bond in PrPres intra- or intermolecular, and does any disulfide exchange reaction occur during the transition of PrPsen to its protease-resistant form in cell-free conversion reactions? Samples of [35S]methionine-labeled hamster GPINEG PrPsen and PrPres derived from hamsters infected with scrapie strain 263K were prepared as described previously (18Caughey B.W. Dong A. Bhat K.S. Ernst D. Hayes S.F. Caughey W.S. Biochemistry. 1991; 30: 7672-7680Crossref PubMed Scopus (744) Google Scholar, 19Caughey B. Horiuchi M. Demaimay R. Raymond G.J. Methods Enzymol. 1999; 309: 122-133Crossref PubMed Scopus (31) Google Scholar). 500 ng of PrPres was suspended in 30 μl of 2.5 m Gdn·HCl/phosphate-buffered saline, pH 7.5 (sample I), 2.5 mGdn·HCl/phosphate-buffered saline, pH 7.5, supplemented with 80 mm NEM (Pierce) (sample II), or in 2.5 mGdn·HCl/100 mm acetic acid (sample III). After a 1-h incubation of all three PrPres samples at 37 °C the samples were centrifuged, and the pellets were used in the subsequent denaturation experiment. Two 6-μl aliquots were taken from each supernatant for confirming the effectiveness of the thiol-blocking, and then the remaining supernatants were prepared for SDS-PAGE analyses by diluting out the Gdn·HCl with 200 μl of 0.2 m Tris buffer, pH 8, adding 4 μl of 5 mg/ml thyroglobulin (Sigma) as a carrier protein, and precipitating the proteins with 950 μl of methanol (19Caughey B. Horiuchi M. Demaimay R. Raymond G.J. Methods Enzymol. 1999; 309: 122-133Crossref PubMed Scopus (31) Google Scholar). For the samples in acetic acid (sample III) the 0.2m Tris buffer contained 0.3 m AEMTS (20Bruice T.W. Kenyon G.L. J. Prot. Chem. 1982; 1: 47-58Crossref Scopus (95) Google Scholar) as well. The pellets from the prewash step were resuspended in 30 μl of the same solution as in the prewash step, containing the same concentration of thiol-blocking reagent except that 6 m instead of 2.5 m Gdn·HCl was used. After a 1-h additional incubation at 37 °C, the samples were prepared for SDS-PAGE analyses as described for the supernatants of the prewash step above. The methanol precipitates were resuspended in loading buffer (8Herrmann L.M. Caughey B. Neuroreport. 1998; 9: 2457-2461Crossref PubMed Scopus (86) Google Scholar) containing no reducing agent. One-half of each sample was mixed with an equal volume of loading buffer containing β-mercaptoethanol; the other half was mixed with an equal volume of loading buffer containing no reducing agent. The samples were then run on 12% precast Tris-glycine SDS gel or 10% NuPage (Invitrogen). Proteins were electroblotted onto Immobilon-P polyvinylidene difluoride (Millipore). PrP was detected with 3F4 anti-PrP monoclonal antibody (21Kascsak R.J. Rubenstein R. Merz P.A. Tonna-DeMasi M. Fersko R. Carp R.I. Wisniewski H.M. Diringer H. J. Virol. 1987; 61: 3688-3693Crossref PubMed Google Scholar) as described previously (19Caughey B. Horiuchi M. Demaimay R. Raymond G.J. Methods Enzymol. 1999; 309: 122-133Crossref PubMed Scopus (31) Google Scholar). The 6-μl aliquots that were removed from the supernatants of the prewash step were mixed with 6 μl of 6m Gdn·HCl/phosphate-buffered saline, pH 7.5 (samples I and II) or 6 μl of 100 mm acetic acid buffer (sample III) with or without 5 mm reduced DTT. After 15 min of incubation at 37 °C, the samples were prepared for SDS-PAGE gel analyses as described above, except that the methanol-precipitated samples were resuspended in loading buffer containing 1.8 mm biotinylated-maleimide (Sigma), run on a 10% NuPage SDS gel, electroblotted as before, and stained with a streptavidin conjugate (Bio-Rad). The rationale for this experiment is according to the following: 5 mm DTT reduces the disulfide bonds in the sample, and after removing the excess DTTred by methanol precipitation biotinylated maleimide reacts with the free cysteines of the protein. The streptavidin staining visualizes the reacted biotinylated protein that had free thiols after DTT reduction. Thyroglobulin, the carrier protein, has a great number of cysteines and gave strong staining in the DTTred-treated sample. However, samples without DTTred or samples that have thiol-blocking agent did not give strong staining, indicating that the thiol blocks in these experiments were effective in blocking millimolar concentrations of DTTred. This reaction was carried out as described previously (19Caughey B. Horiuchi M. Demaimay R. Raymond G.J. Methods Enzymol. 1999; 309: 122-133Crossref PubMed Scopus (31) Google Scholar), except that some of the samples were treated with the maleimide blocking agent. Briefly, 250 ng of hamster PrPres was preincubated in 2.5m Gdn·HCl, pH 6, at 37 °C for 1 h (19Caughey B. Horiuchi M. Demaimay R. Raymond G.J. Methods Enzymol. 1999; 309: 122-133Crossref PubMed Scopus (31) Google Scholar). Maleimide (Sigma) in 2.5 m Gdn·HCl was added to some of the samples (to a final concentration of 300 mm) at the end of the preincubation. After 15 min, 35S-labeled hamster GPINEG PrPsen (25,000 cpm) and buffer were added to bring the solution to final concentrations of 1 mGdn·HCl, 50 mm citrate, pH 6, 5 mm cetyl pyridinium chloride, and 1% N-lauryl sarcosine. The final concentration for maleimide was 50 mm. The conversion was allowed to proceed for 2 days at 37 °C; 10% of each sample was then removed (for PK− samples), and the remaining 90% was digested with 20 μg/ml proteinase K (giving a 2:1 PK/PrPres weight ratio for PK+ samples) in Tris-saline (50 mm Tris-HCl, pH 7.5, 150 mm NaCl in 50 μl) at 37 °C for 1 h. Digestion by PK was stopped by adding 1 μl of 0.1 m Pefabloc (Roche Molecular Biochemicals), 4 μl of 5 μg/ml thyroglobulin, and four volumes of methanol. The PK− samples were similarly precipitated. Methanol-precipitated pellets were subjected to SDS-PAGE on 10% precast NuPage gel (Invitrogen), and radioactive proteins were visualized with a PhosphorImager (Molecular Dynamics). The conversion was carried out by mixing 500 ng of hamster PrPres and 35S-labeled GPINEG PrPsen (50,000 cpm), pH 7, in 50 mm imidazole and 200 mm KCl, 0.2 mg/ml heparan sulfate, but with or without 70 mm AEMTS, at 0 °C. The samples were kept at 0 °C for 1 h and subsequently incubated at 37 °C for 3 h. During the 0 °C incubation, some PrPsen bound to PrPres, but there was no detectable conversion of the labeled PrPsen (19Caughey B. Horiuchi M. Demaimay R. Raymond G.J. Methods Enzymol. 1999; 309: 122-133Crossref PubMed Scopus (31) Google Scholar). Proteinase K+ digestion was carried out as in the Gdn·HCl conversion with 4:1 and 20:1 PK/PrPres weight ratios, and the samples were analyzed on SDS-PAGE (precast 14% Tris-glycine gel) as described for the Gdn·HCl conversion. To confirm the effectiveness of the thiol-blocking methods, a control experiment was carried out as in the previous assay for intramolecular disulfides bonds. Aliquots were removed from the conversion samples before the proteinase K treatment and were mixed with an equal volume of 10 mm DTTred in 6 mGdn·HCl/400 mm Tris buffer, pH 9. Control samples without DTTred were prepared, and all samples were quenched with 1 μl of 250 mm AEMTS and precipitated with methanol, and radioactive proteins were analyzed on SDS-PAGE as described above. The purpose of the following experiments is to determine whether or not the monomeric subunits of PrPres aggregates are linked by intermolecular disulfide bonds. For this purpose, any existing free thiols on PrPres were blocked with the protein being simultaneously denatured with Gdn·HCl; if the thiols were not blocked, a linear disulfide-bonded polymer could self-depolymerize mostly into monomers with intramolecular disulfide bonds upon denaturation (13Welker E. Wedemeyer W.J. Scheraga H.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 4334-4336Crossref PubMed Scopus (62) Google Scholar). However, by blocking the thiols, such a disulfide-bonded polymer would remain intact and could easily be distinguished from PrP monomers by SDS-PAGE. To block the possible self-depolymerization reaction of a disulfide-linked linear polymer fully during its denaturation, the terminal thiols of such a self-depolymerizing polymer have to be blocked faster than the disulfide-reshuffling reaction can occur. Although the terminal thiol of such a polymer would have to attack the disulfide bond between two monomers in the polymer, it is a unimolecular reaction in kinetic terms and thus is very fast. The full blocking of this otherwise fast self-depolymerization reaction is a critical technical issue in this experiment. To understand the underlying experimental rationale, it is useful to consider the three commonly used protein thiol-blocking methods: (i) irreversible blocking by alkylation; (ii) reversible blocking by forming a mixed disulfide with the protein-thiolate; (iii) reversible blocking by a pH quench, which protonates the thiolate. Alkylation is applicable for permanently blocking thiols that are easily available without denaturing the protein (22Rothwarf D.M. Scheraga H.A. J. Am. Chem. Soc. 1991; 113: 6294-6296Crossref Scopus (47) Google Scholar, 23Weissman J.S. Kim P.S. Science. 1991; 253: 1386-1393Crossref PubMed Scopus (492) Google Scholar). However, it is a relatively slow reaction; hence, much faster unimolecular disulfide-reshuffling reactions are prevented only when the alkylating agent is applied at high concentrations. Such high concentrations of alkylating agent frequently lead to covalent modification of other non-cysteine amino acids (24Jocelyn P.C. Methods Enzymol. 1987; 143: 44-67Crossref PubMed Scopus (301) Google Scholar). By contrast, reversible blocking with thiosulfonates such as AEMTS or pH quench is sufficiently fast to block the unimolecular disulfide-reshuffling reaction fully, although reversible blocking requires extra caution to avoid rearrangements during subsequent analysis steps (22Rothwarf D.M. Scheraga H.A. J. Am. Chem. Soc. 1991; 113: 6294-6296Crossref Scopus (47) Google Scholar, 23Weissman J.S. Kim P.S. Science. 1991; 253: 1386-1393Crossref PubMed Scopus (492) Google Scholar, 25Rothwarf D.M. Scheraga H.A. Biochemistry. 1993; 32: 2671-2679Crossref PubMed Scopus (140) Google Scholar). To avoid possible artifacts in any particular blocking procedure, two different thiol-blocking approaches were employed. In an irreversible thiol-blocking approach 80 mm NEM was used. In the other reversible approach, AEMTS blocking was combined with the pH quench method; experiments were carried out at low pH that blocked the disulfide reactions efficiently by protonating the thiolates that are the reactive species. At the end of the experiments, the pH was raised to 8 in the presence of 70 mm AEMTS that forms a disulfide with the now-deprotonated thiols faster than any other thiol-disulfide exchange reaction could occur. Thus, the subsequent analysis step (methanol precipitation and SDS-PAGE) could be carried out at ambient pH. To decrease the background due to PrPres-associated PrPsenmolecules, a prewash step in 2.5 m Gdn·HCl was carried out to strip off the associated PrPsen molecules (26Caughey B. Kocisko D.A. Raymond G.J. Lansbury P.T., Jr. Chem. Biol. 1995; 2: 807-817Abstract Full Text PDF PubMed Scopus (172) Google Scholar). The remaining pellet contains the PrPres molecules and infectivity (27Caughey B. Raymond G.J. Kocisko D.A. Lansbury P.T., Jr. J. Virol. 1997; 71: 4107-4110Crossref PubMed Google Scholar). The various thiol-reactive reagents were added to both the prewash and pellet denaturation steps. In one experiment (I), PrPSc was incubated in 2.5 m Gdn·HCl, pH 7.5 without any thiol-blocking reagent. In this experiment, a possible thiol-induced self-depolymerization of the putative disulfide-linked PrP polymer was not blocked. Thus, denaturation of the PrPres would result in mostly disulfide-intact monomers, regardless of whether the monomers were linked by intermolecular disulfide bonds or the disulfide-intact monomers were associated only by non-covalent enthalpic interactions. In other experiments, PrPres was incubated with NEM (II), or the low pH approach in combination with AEMTS blocking was used (III). In II and III, the possible thiol-induced self-depolymerization would not take place during the denaturation of the protein. The samples were analyzed for self-depolymerization on Western blots as follows. Half of the samples had β-mercaptoethanol in the loading buffer to reduce all disulfide bonds in the sample to distinguish disulfide-bonded oligomers from oligomers joined by other types of covalent linkages (Figs.1 and 2,lanes 1-3). The other half of the samples did not have the β-mercaptoethanol to keep all disulfide bonds intact during the analyses (Figs. 1 and 2, lanes 4-6).Figure 2Intra- versus intermolecular disulfide bonds in PrPSc. To examine whether thiol-blocked PrPSc was dissociated into monomers or remained polymerized during the 6 m Gdn·HCl denaturation, the methanol-precipitated proteins were analyzed on immunoblot (10% NuPage SDS gel) as described under "Experimental Procedures."Lane 1, control without any thiol-blocking reagent;lane 2, with 80 mm NEM; lane 3, with 100 mm acetic acid, subsequently blocked with 0.3m AEMTS. β-mercaptoethanol was present in the samples oflanes 1-3. Lanes 4-6 are the same samples as lanes 1-3, respectively, except that β-mercaptoethanol was omitted. Neither thiol-blocking nor β-mercaptoethanol caused significant differences in the ratio of the monomeric and multimeric forms of the prion protein on this gel, showing that no prion protein molecules linked by intermolecular disulfide bonds were detectable. The numbers on the left-hand side indicate molecular mass in kilodaltons.View Large Image Figure ViewerDownload Hi-res image Download (PPT) To overcome the difficulty associated with the inefficient transfer of high-molecular-weight material from the gels to the membrane during electroblotting, the gels in Figs. 1 and 2 were overloaded to provide better detection in the higher-molecular-weight region. In Fig. 1, the supernatants of the 2.5 m Gdn·HCl prewash step are shown. Fig. 2 shows the protein in the pellets that were solubilized in 6m Gdn·HCl. If the disulfide bonds in PrPres were intermolecular, and free thiols could initiate a disulfide-reshuffling-induced self-depolymerization upon denaturation (13Welker E. Wedemeyer W.J. Scheraga H.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 4334-4336Crossref PubMed Scopus (62) Google Scholar), then blocking free thiols should have resulted in less PrP monomer and more high-molecular-weight material toward the top of the gel in lanes 5 and6 in Figs. 1 and 2, compared with the unblocked control samples (lane 4). By contrast, if the disulfide bond were intramolecular within the PrPres monomers, then neither the thiol-blocking reagents nor β-mercaptoethanol would affect the ratio of monomeric and multimeric forms of PrP observed on the gel. Figs. 1 and 2 show that there is no effect of the thiol blocking on PrPres depolymerization. Although there may be a smaller amount of PrPres in lanes 2 and 5 in Fig. 2, this difference is not affected by β-mercaptoethanol in the loading buffer, indicating that it does not originate from an intermolecular disulfide-bonded population in the samples. Therefore, there was no evidence for intermolecular disulfide-linked PrPres aggregates in Figs. 1 and 2. The experiment was carried out several times and with other, slightly different thiol-blocking approaches; the gels were loaded with different amounts of samples, which also made it possible to compare the monomer regions of the lanes. No effect of the thiol blocking on either monomeric or multimeric forms of PrP was evident in β-mercaptoethanol (−) lanes compared with the respective β-mercaptoethanol (+) lanes on any of these gels. From a comparison of the β-mercaptoethanol (−) lanes to the respective β-mercaptoethanol (+) lanes, no band from a disulfide-bonded dimer was observed on these gels. The capability of the thiol-blocking methods to quench 5 mmDTTred was confirmed in aliquots taken from each sample before methanol precipitation (data not shown; see "Experimental Procedures"). Thus, these results show that the molecules in the bulk of the PrPres aggregate are not linked by intermolecular disulfide bonds. Although the disulfide bond in PrPres is therefore intramolecular within the PrPres monomers, experiments on recombinant prion protein have led to the hypothesis that the intramolecular disulfide bond does not persist during the conversion of PrPsen to PrPres (10Mehlhorn I. Groth D. Sto¨ckel J. Moffat B. Reilly D. Yansura D. Willett W.S. Baldwin M. Fletterick R. Cohen F.E. Vandlen R. Henner D. Prusiner S.B. Biochemistry. 1996; 35: 5528-5537Crossref PubMed Scopus (194) Google Scholar, 11Jackson G.S. Hosszu L.L.P. Power A. Hill A.F. Kenney J. Saibil H. Craven C.J. Waltho J.P. Clarke A.R. Collinge J. Science. 1999; 283: 1935-1937Crossref PubMed Scopus (365) Google Scholar). A possible temporary reduction of the intramolecular disulfide bond of PrPsen may facilitate the conformational transition, providing higher flexibility for the protein and destabilizing the α-helical fold. The subsequent re-formation of the original intramolecular disulfide bond may stabilize a new β conformation. The conversion of the cellular prion protein to the protease-resistant form on a PrPres template was shown to occur in the absence of intact cells (28Kocisko D.A. Come J.H. Priola S.A. Chesebro B. Raymond G.J. Lansbury P.T. Caughey B. Nature. 1994; 370: 471-474Crossref PubMed Scopus (792) Google Scholar). The cell-free systems have reproduced features of the in vivo conversion (29Bessen R.A. Kocisko D.A. Raymond G.J. Nandan S. Lansbury P.T. Caughey B. Nature. 1995; 375: 698-700Crossref PubMed Scopus (461) Google Scholar, 30Bossers A. Belt P.B.G.M. Raymond G.J. Caughey B. de Vries R. Smits M.A. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 4931-4936Crossref PubMed Scopus (160) Google Scholar, 31Raymond G.J. Hope J. Kocisko D.A. Priola S.A. Raymond L.D. Bossers A. Ironside J. Will R.G. Chen S.G. Petersen R.B. Gambetti P. Rubenstein R. Smits M.A. Lansbury P.T., Jr. Caughey B. Nature. 1997; 388: 285-288Crossref PubMed Scopus (239) Google Scholar); however, generation of infectivity has not been demonstrated in these systems (32Hill A.F. Antoniou M. Collinge J. J. Gen. Virol. 1999; 80: 11-14Crossref PubMed Scopus (176) Google Scholar). The experiments in this sub-section were carried out to determine whether thiol-disulfide exchange reactions are involved in the cell-free conversion reactions. The presence of a thiol (strictly speaking the deprotonated form, the thiolate, which is the reactive species involved in disulfide reactions) is needed for any disulfide exchange reaction to occur. PrPres preparations may contain co-purified thiols that can initiate the reduction of the disulfide bonds during the cell-free conversion reaction. Thus, by blocking the thiols we can learn whether the conformational transition involves a disulfide exchange reaction or not. If thiol-blocking reagents do not inhibit the cell-free conversion reaction, it would indicate that no thiol-disulfide exchange reaction is involved, and thus, that the disulfide bond is maintained intact during the conversion. Based on the above considerations, a Gdn·HCl conversion reaction was carried out as described previously (19Caughey B. Horiuchi M. Demaimay R. Raymond G.J. Methods Enzymol. 1999; 309: 122-133Crossref PubMed Scopus (31) Google Scholar) with or without 50 mm maleimide, another thiol-alkylating reagent. The detection of the converted PrPsen is based on the fact that converted PrPsen molecules are partially resistant to proteinase K digestion. Although proteinase K fully digests PrPsen, it digests PrPres only partially, leaving the core of the protein intact. [35S]methionine labeling of PrPsen makes it possible, even in the presence of excess unlabeled PrPres template, to detect the newly converted PrPres. Fig. 3 shows the Gdn·HCl conversion reaction. The non-glycosylated and the monoglycosylated bands of PrPsen in all PK− lanes show approximately the same intensity, indicating that all of the samples have the same amount of35S-labeled PrPsen substrate. As observed previously, 35S-labeled PrPsen was converted to the expected partially protease-resistant forms (∼7 kDa smaller after PK digestion than the original [35S]PrPsen) in the presence of a PrPres template (compare lanes 4 and 3) but not in its absence (compare lanes 1 and 2). Maleimide (25 mm) had a subtle effect on the pattern of lower [35S]PrPresbands but did not block conversion overall (compare lanes 4and 6). A lower [35S]PrPres band was less intense in the PrPres+ PK+ samples, and there was a slight upward shift in the other [35S]PrPres bands (Fig. 3, lane 6). One interpretation of these results is that maleimide reactions with free sulfhydryls affect the conversion of a subset of PrP molecules. However, the slight mobility shift in the [35S]PrPres bands led us to suspect that these changes in the pattern of conversion products without overall inhibition of conversion may be due to the covalent modification of some non-cysteine amino acids of the PrP molecules. This potential problem led us to test for the involvement of disulfide exchange in PrP conversions using different conditions and an alternative thiol-blocking reagent, which is less prone to side reactions. Gdn·HCl-free conversions were carried out with or without AEMTS (Fig. 4) using a shorter incubation time (3–6 h versus 1–2 days). AEMTS was chosen after testing its applicability on RNaseA to this unusually long blocking period that was required with PrP. The 70 mm AEMTS used in the experiment should provide efficient blocking of any known disulfide reaction (22Rothwarf D.M. Scheraga H.A. J. Am. Chem. Soc. 1991; 113: 6294-6296Crossref Scopus (47) Google Scholar, 25Rothwarf D.M. Scheraga H.A. Biochemistry. 1993; 32: 2671-2679Crossref PubMed Scopus (140) Google Scholar). To simplify the interpretation of the conversion product bands, the experiment was also carried out with GPINEG PrPsen purified from tunicamycin-treated cells, which had only the non-glycosylated form of the protein. As shown in Fig. 4 (lanes 5–7, 12, and13), AEMTS has little effect on the conversion reaction. Compared with the maleimide experiment (Fig. 3), a higher PK/PrPres ratio was applied to overcome the interference of AEMTS with PK digestion. The efficacy of AEMTS blocking was demonstrated by the quenching of 10 mm DTTredby aliquots taken from conversion reactions at the end of the 3-h incubation (see "Experimental Procedures"). Without AEMTS, DTT reduced PrP under the applied conditions, judging from the slower mobility of the protein (8Herrmann L.M. Caughey B. Neuroreport. 1998; 9: 2457-2461Crossref PubMed Scopus (86) Google Scholar) in lane 1 compared withlane 2 in Fig. 4. However, when AEMTS was present (lane 3) DTTred was quenched, and PrP remained oxidized (lane 3 compared with lane 2 in Fig.4.). Thus, the observations that distinct [35S]PrPres conversion products were generated in the presence of either maleimide or AEMTS provides evidence that PrPsen can be induced by PrPresto convert to the protease-resistant form without thiol-disulfide exchange reactions. A previous study reported the full inhibition of the Gdn·HCl-based conversion by pretreatment of the PrPSc template with 100 mm DTT followed by the addition of 300 mmN-ethyl-maleimide (8Herrmann L.M. Caughey B. Neuroreport. 1998; 9: 2457-2461Crossref PubMed Scopus (86) Google Scholar). Because the 100 mm DTT pretreatment did not appear to reduce the bulk of the PrPSctemplate, it was supposed that maleimide might have inhibited the conversion by blocking some free thiols that were present in the PrPSc preparation (13Welker E. Wedemeyer W.J. Scheraga H.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 4334-4336Crossref PubMed Scopus (62) Google Scholar). However, the present results agree with the interpretation given in the former study (8Herrmann L.M. Caughey B. Neuroreport. 1998; 9: 2457-2461Crossref PubMed Scopus (86) Google Scholar) that the DTT pretreatment was also necessary for the observed inhibition by maleimide. DTT might have reduced the disulfide bonds in a critical fraction of PrPSc, which forms the conversion sites where PrPC binds to the PrPSc aggregate to be converted to a protease-resistant form. The reduction of this critical fraction of PrPSc with the subsequent NEM treatment that may have blocked the now-free protein thiols may have destroyed these conversion sites irreversibly, leading to the observed full inhibition of the conversion. Alternatively, NEM might have modified some non-cysteine amino acid(s) of the prion protein whose modification interferes with the conversion reaction. PrPSc appears to have intramolecular disulfide bonds within its monomeric subunits. A similar conclusion was obtained in earlier work by Turk et al. (14Turk E. Teplow D.B. Hood L.E. Prusiner S.B. Eur. J. Biochem. 1988; 176: 21-30Crossref PubMed Scopus (270) Google Scholar). However, the possible spontaneous depolymerization of PrPSc that would occur upon denaturation of a PrPSc polymer linked by intermolecular disulfide bonds without any added reducing reagent was not blocked in these earlier experiments; this made the distinction between intra- and intermolecular disulfide bonds uncertain. The mechanism of the prion protein conversion in vivo may be different from the cell-free conversion reaction; however, our data demonstrate that the conformational transition can occur without disulfide exchange.
Referência(s)