Title: Structure and Dynamics of Micelle-bound Human α-Synuclein
Abstract: Misfolding of the protein α-synuclein (aS), which associates with presynaptic vesicles, has been implicated in the molecular chain of events leading to Parkinson's disease. Here, the structure and dynamics of micelle-bound aS are reported. Val3-Val37 and Lys45-Thr92 form curved α-helices, connected by a well ordered, extended linker in an unexpected anti-parallel arrangement, followed by another short extended region (Gly93-Lys97), overlapping the recently identified chaperone-mediated autophagy recognition motif and a highly mobile tail (Asp98-Ala140). Helix curvature is significantly less than predicted based on the native micelle shape, indicating a deformation of the micelle by aS. Structural and dynamic parameters show a reduced helical content for Ala30-Val37. A dynamic variation in interhelical distance on the microsecond timescale is complemented by enhanced sub-nanosecond timescale dynamics, particularly in the remarkably glycine-rich segments of the helices. These unusually rich dynamics may serve to mitigate the effect of aS binding on membrane fluidity. The well ordered conformation of the helix-helix connector indicates a defined interaction with lipidic surfaces, suggesting that, when bound to larger diameter synaptic vesicles, it can act as a switch between this structure and a previously proposed uninterrupted helix. Misfolding of the protein α-synuclein (aS), which associates with presynaptic vesicles, has been implicated in the molecular chain of events leading to Parkinson's disease. Here, the structure and dynamics of micelle-bound aS are reported. Val3-Val37 and Lys45-Thr92 form curved α-helices, connected by a well ordered, extended linker in an unexpected anti-parallel arrangement, followed by another short extended region (Gly93-Lys97), overlapping the recently identified chaperone-mediated autophagy recognition motif and a highly mobile tail (Asp98-Ala140). Helix curvature is significantly less than predicted based on the native micelle shape, indicating a deformation of the micelle by aS. Structural and dynamic parameters show a reduced helical content for Ala30-Val37. A dynamic variation in interhelical distance on the microsecond timescale is complemented by enhanced sub-nanosecond timescale dynamics, particularly in the remarkably glycine-rich segments of the helices. These unusually rich dynamics may serve to mitigate the effect of aS binding on membrane fluidity. The well ordered conformation of the helix-helix connector indicates a defined interaction with lipidic surfaces, suggesting that, when bound to larger diameter synaptic vesicles, it can act as a switch between this structure and a previously proposed uninterrupted helix. The protein α-synuclein (aS) 1The abbreviations used are: aS, human α-synuclein; EPR, electron paramagnetic resonance; PRE, paramagnetic relaxation enhancement; RDC, residual dipolar coupling; SUV, small unilamellar vesicle; TROSY, transverse relaxation optimized spectroscopy; NOE, nuclear Overhauser effect; NOESY, NOE spectroscopy; TOCSY, total correlation spectroscopy; TEMED, N,N,N′,N′-tetramethylethylenediamine; HSQC, heteronuclear single quantum coherence; CT, constant time.1The abbreviations used are: aS, human α-synuclein; EPR, electron paramagnetic resonance; PRE, paramagnetic relaxation enhancement; RDC, residual dipolar coupling; SUV, small unilamellar vesicle; TROSY, transverse relaxation optimized spectroscopy; NOE, nuclear Overhauser effect; NOESY, NOE spectroscopy; TOCSY, total correlation spectroscopy; TEMED, N,N,N′,N′-tetramethylethylenediamine; HSQC, heteronuclear single quantum coherence; CT, constant time. has been implicated in the molecular chain of events leading to Parkinson's disease, the second most common neurodegenerative disorder in humans (1Goedert M. Nat. Rev. Neurosci. 2001; 2: 492-501Crossref PubMed Scopus (1088) Google Scholar, 2Dev K.K. Hofele K. Barbieri S. Buchman V.L. van der Putten H. Neuropharmacology. 2003; 45: 14-44Crossref PubMed Scopus (229) Google Scholar, 3Nussbaum R.L. Ellis C.E. N. Engl. J. Med. 2003; 348: 1356-1364Crossref PubMed Scopus (1010) Google Scholar). Lewy bodies and Lewy neurites found in Parkinson's disease in general contain aggregates of aS (4Spillantini M.G. Schmidt M.L. Lee V.M.Y. Trojanowski J.Q. Jakes R. Goedert M. Nature. 1997; 388: 839-840Crossref PubMed Scopus (6111) Google Scholar, 5Mezey E. Dehejia A.M. Harta G. Suchy S.F. Nussbaum R.L. Brownstein M.J. Polymeropoulos M.H. Mol. Psychiatry. 1998; 3: 493-499Crossref PubMed Scopus (106) Google Scholar), which is prone to form aggregates in vitro (6Conway K.A. Lee S.J. Rochet J.C. Ding T.T. Williamson R.E. Lansbury P.T. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 571-576Crossref PubMed Scopus (1333) Google Scholar, 7Fernandez C.O. Hoyer W. Zweckstetter M. Jares-Erijman E.A. Subramaniam V. Griesinger C. Jovin T.M. EMBO J. 2004; 23: 2039-2046Crossref PubMed Scopus (209) Google Scholar). Moreover, three missense mutations (A30P, E46K, and A53T) in the gene encoding aS cause familial Parkinson's disease as well as aS gene triplication (3Nussbaum R.L. Ellis C.E. N. Engl. J. Med. 2003; 348: 1356-1364Crossref PubMed Scopus (1010) Google Scholar, 8Kahle P.J. Haass C. Kretzschmar H.A. Neumann M. J. Neurochem. 2002; 82: 449-457Crossref PubMed Scopus (77) Google Scholar, 9Vekrellis K. Rideout H.J. Stefanis L. Mol. Neurobiol. 2004; 30: 1-21Crossref PubMed Google Scholar). The amino acid sequence of aS consists of 140 residues with 7 copies of an unusual 11-residue repeat, followed by a hydrophilic tail (Fig. 1). Based on sequence analysis, it was suggested that aS interacts with lipid membranes through its repeat region (10Davidson W.S. Jonas A. Clayton D.F. George J.M. J. Biol. Chem. 1998; 273: 9443-9449Abstract Full Text Full Text PDF PubMed Scopus (1238) Google Scholar, 11George J.M. Jin H. Woods W.S. Clayton D.F. Neuron. 1995; 15: 361-372Abstract Full Text PDF PubMed Scopus (725) Google Scholar) and interactions with small unilamellar vesicles (SUVs) and micelles preferentially containing negatively charged head groups have been documented in vitro (10Davidson W.S. Jonas A. Clayton D.F. George J.M. J. Biol. Chem. 1998; 273: 9443-9449Abstract Full Text Full Text PDF PubMed Scopus (1238) Google Scholar, 12Chandra S. Chen X.C. Rizo J. Jahn R. Sudhof T.C. J. Biol. Chem. 2003; 278: 15313-15318Abstract Full Text Full Text PDF PubMed Scopus (418) Google Scholar, 13Eliezer D. Kutluay E. Bussell R. Browne G. J. Mol. Biol. 2001; 307: 1061-1073Crossref PubMed Scopus (850) Google Scholar, 14Narayanan V. Scarlata S. Biochemistry. 2001; 40: 9927-9934Crossref PubMed Scopus (158) Google Scholar). In accordance with those properties, aS localizes in a physiological environment at nerve termini in close proximity to synaptic vesicles (15Jensen P.H. Nielsen M.S. Jakes R. Dotti G. Goedert M. J. Biol. Chem. 1998; 273: 26292-26294Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar, 16Iwai A. Masliah E. Yoshimoto M. Ge N.F. Flanagan L. Desilva H.A.R. Kittel A. Saitoh T. Neuron. 1995; 14: 467-475Abstract Full Text PDF PubMed Scopus (1122) Google Scholar) and may be implicated in synaptic plasticity and neurotransmitter release (8Kahle P.J. Haass C. Kretzschmar H.A. Neumann M. J. Neurochem. 2002; 82: 449-457Crossref PubMed Scopus (77) Google Scholar, 9Vekrellis K. Rideout H.J. Stefanis L. Mol. Neurobiol. 2004; 30: 1-21Crossref PubMed Google Scholar). As evidenced by a number of biophysical techniques, aS is predominantly a random coil in aqueous solution but has been shown to adopt secondary structure of mostly helical nature upon association with negatively charged SUV or detergent micelle surfaces (10Davidson W.S. Jonas A. Clayton D.F. George J.M. J. Biol. Chem. 1998; 273: 9443-9449Abstract Full Text Full Text PDF PubMed Scopus (1238) Google Scholar, 12Chandra S. Chen X.C. Rizo J. Jahn R. Sudhof T.C. J. Biol. Chem. 2003; 278: 15313-15318Abstract Full Text Full Text PDF PubMed Scopus (418) Google Scholar, 13Eliezer D. Kutluay E. Bussell R. Browne G. J. Mol. Biol. 2001; 307: 1061-1073Crossref PubMed Scopus (850) Google Scholar, 17Weinreb P.H. Zhen W.G. Poon A.W. Conway K.A. Lansbury P.T. Biochemistry. 1996; 35: 13709-13715Crossref PubMed Scopus (1308) Google Scholar). In every case, the repeat region mediates lipid or detergent interactions, whereas the hydrophilic tail remains free in solution. In the presence of SUV the large majority of aS molecules exist vesicle-bound (10Davidson W.S. Jonas A. Clayton D.F. George J.M. J. Biol. Chem. 1998; 273: 9443-9449Abstract Full Text Full Text PDF PubMed Scopus (1238) Google Scholar). Based on the backbone 1H-1H NOE and secondary chemical shift pattern, helical secondary structure has been attributed to the entire repeat region in the micelle-bound state with the exception of a short stretch near Ser42-Thr44 (12Chandra S. Chen X.C. Rizo J. Jahn R. Sudhof T.C. J. Biol. Chem. 2003; 278: 15313-15318Abstract Full Text Full Text PDF PubMed Scopus (418) Google Scholar, 18Bussell R. Eliezer D. J. Mol. Biol. 2003; 329: 763-778Crossref PubMed Scopus (338) Google Scholar). When associated with SUVs of 300–400 Å diameter, electron paramagnetic resonance (EPR) data were interpreted as evidence for an uninterrupted helix extending throughout the entire repeat region (19Jao C.C. Der-Sarkissian A. Chen J. Langen R. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 8331-8336Crossref PubMed Scopus (309) Google Scholar). Helical wheel models have been proposed to describe the interaction of the helix side chains with the membrane surface (12Chandra S. Chen X.C. Rizo J. Jahn R. Sudhof T.C. J. Biol. Chem. 2003; 278: 15313-15318Abstract Full Text Full Text PDF PubMed Scopus (418) Google Scholar, 18Bussell R. Eliezer D. J. Mol. Biol. 2003; 329: 763-778Crossref PubMed Scopus (338) Google Scholar, 19Jao C.C. Der-Sarkissian A. Chen J. Langen R. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 8331-8336Crossref PubMed Scopus (309) Google Scholar). However, besides those qualitative properties no further structural information of the lipid-bound state of aS is available. The dynamic properties of aS along the helix and information on site-specific interactions of aS with lipidic surfaces are also unknown. Moreover, although aS adopts β-stranded conformation in the aggregated, fibrillar forms characteristic of Lewy bodies (20Der-Sarkissian A. Jao C.C. Chen J. Langen R. J. Biol. Chem. 2003; 278: 37530-37535Abstract Full Text Full Text PDF PubMed Scopus (291) Google Scholar), aggregation of aS into dimers and multimers is promoted by lipid environments that induce an α-helical conformation (14Narayanan V. Scarlata S. Biochemistry. 2001; 40: 9927-9934Crossref PubMed Scopus (158) Google Scholar, 21Cole N.B. Murphy D.D. Grider T. Rueter S. Brasaemle D. Nussbaum R.L. J. Biol. Chem. 2002; 277: 6344-6352Abstract Full Text Full Text PDF PubMed Scopus (354) Google Scholar, 22Lee H.J. Choi C. Lee S.J. J. Biol. Chem. 2002; 277: 671-678Abstract Full Text Full Text PDF PubMed Scopus (376) Google Scholar). Thus, high resolution structural and dynamic information of aS in its helical conformation appears pertinent to developing a better understanding of the physiological role of aS, as well as possible structural features relevant to aS misfolding. Here, the structure and dynamics of aS in the micelle-bound form, determined by solution NMR spectroscopy, are presented and related to the vesicle-bound state. As solution NMR is limited by particle size, the direct study of vesicle-bound aS is not feasible. We therefore elected to resort to smaller diameter micelles, which elicit a similar aS helical content as SUV (12Chandra S. Chen X.C. Rizo J. Jahn R. Sudhof T.C. J. Biol. Chem. 2003; 278: 15313-15318Abstract Full Text Full Text PDF PubMed Scopus (418) Google Scholar). Protein Production—The human α-synuclein (aS) gene was expressed from the kanamycin restricted, T7lac promoter controlled pET-41 vector (Novagen, Inc.). A mutant, aS(S87C), was constructed from this vector following the QuikChange protocol (Stratagene, Inc.). Both genes were overexpressed in Escherichia coli BL21(DE3) cells, cultured in M9 minimal medium (6 g/liter Na2HPO4·7H2O, 3 g/liter KH2PO4, 0.5 g/liter NaC1, 1 mm MgSO4, 100 μm CaC12·2H2O, and d-glucose/NH4Cl as specified below). 1 g/liter 98% 15NH4Cl was employed to achieve 15N enrichment. 2 g/liter 99% [13C]glucose, 1 g/liter 98% 15ND4Cl, and 1 g/liter 98% D/99% 13C/98% 15N-rich growth medium supplement (Isogro, Isotec, Inc.) in D2O solution were employed to label with 2H/13C/15N. Uniformly 2H/13C/15N-enriched protein with selectively protonated Val-γ1,2, Leu-δ1,2, and Ile-δ1 methyl groups was prepared by using 2 g/liter 97% D, 99% [13C]glucose, 1 g/liter 98% 15ND4Cl in D2O solution, and, added 45 min before protein induction, 50 mg/liter 3,3-2H2-, α-[13C]ketobutyrate, and 100 mg/liter 3-2H-, α-[13C]ketoisovalerate (23Goto N.K. Gardner K.H. Mueller G.A. Willis R.C. Kay L.E. J. Biomol. NMR. 1999; 13: 369-374Crossref PubMed Scopus (439) Google Scholar). Protein expression was induced at A600 = 0.8 for 2 h in H2 Oor 4 h in D2O. Purification involved heat precipitation (10 min, 80 °C) of the cells in 50 mm Tris·HCl, pH 7.5, 500 mm NaCl (2 mm β-mercaptoethanol for S87C). Subsequent ion-exchange chromatography on Q-Sepharose (Amersham Biosciences) yielded >98% pure protein of the correct mass as judged by SDS-PAGE and electrospray mass spectrometry. NMR Sample Preparation—All samples were prepared in H2O to contain aS at a concentration of 0.5 mm (ϵ280 = 5120 m-1cm-1), 75 mm SDS, 6% D2O, and 0.02% (w/v) NaN3 in a total volume of 270 μl. Nine samples were prepared. Sample A was enriched only in 15N, sample B was uniformly enriched in 2H/13C/15N, sample C with 2H/13C/15N, but selectively protonated at Val-γ1,2, Leu-δ1,2, and Ile-δ1 methyl groups and also employed deuterated SDS. Samples D and E used 2H/13C/15N-enriched aS in the presence of charged stretched polyacrylamide gels (see below). Samples F–H employed 2H/13C/15N-enriched aS(S87C) tagged with cysteaminyl-EDTA, complexed with Ca2+,Mn2+, and Co2+, respectively. Samples A–E were buffered by 20 mm NaH2PO4/Na2HPO4, pH 7.4, and samples F–H by 20 mm HEPES·NaOH, pH 7.4. The last sample, I, contained 2H/13C/15N-labeled aS, buffered by 20 mm Tris·HCl, pH 8.4. The gel for sample D was polymerized from a 4.6% w/v solution of acrylamide (AA), 2-acrylamido-2-methyl-1-propanesulfonate (AMPS), and bisacrylamide with a monomer to cross-linker ratio of 39:1 (w/w) and a molar ratio of 96:4 of AA to AMPS. To allow easy comparison between gels containing different amounts of charged monomer, AMPS is counted here as AA and the AMPS:AA ratio is specified separately. Gel for sample E was polymerized from a 5.2% w/v solution of AA, AMPS, and piperazine diacrylamide as cross-linker with a monomer to cross-linker ratio of 39:1 (w/w) and a molar ratio of 98:2 of AA to AMPS. For the reported monomer to cross-linker ratio, piperazine diacrylamide was counted as bisacrylamide to again allow for direct comparison between gels. Polymerization was initiated by the addition of ammonium persulfate and TEMED at 0.1% w/v and v/v, respectively, and allowed to proceed in a cylinder of 5.4-mm diameter for 3 h at 23 °C. Subsequently, gels were dialyzed, dried, soaked with protein solution for 2 days, and transferred to NMR tubes as described (24Ulmer T.S. Ramirez B.E. Delaglio F. Bax A. J. Am. Chem. Soc. 2003; 125: 9179-9191Crossref PubMed Scopus (248) Google Scholar, 25Chou J.J. Gaemers S. Howder B. Louis J.M. Bax A. J. Biomol. NMR. 2001; 21: 377-382Crossref PubMed Scopus (220) Google Scholar). 2H quadrupolar splittings of 1.2 and 1.7 Hz were measured for samples D and E, respectively. A 20 mm HEPES·NaOH, pH 8.0, 200 mm NaCl, 25 mm EDTA, 0.02% NaN3 solution of 50 μm aS(S87C) containing 1 mm dithiothreitol was exchanged into the dithiothreitol-free solution containing 1 mmN-[S-(2-pyridylthio)cysteaminyl]-EDTA (Toronto Research Chemicals, Inc.) by ultrafiltration and allowed to react overnight at room temperature. Small amounts of homodimeric aS(S87C) were removed by ion-exchange chromatography, and aS(S87C)-cysteaminyl-EDTA was exchanged into 20 mm HEPES·NaOH, pH 7.4, 50 mm NaCl by ultrafiltration. The EDTA tag was then charged by adding a 1.2 times excess of the desired ion over the protein, washed extensively with 20 mm HEPES·NaOH, pH 7.4, 700 mm NaCl to remove non-specifically bound ions, and exchanged into 20 mm HEPES·NaOH, pH 7.4, 0.02% NaN3 by ultrafiltration. Samples were completed by the addition of SDS and D2O to 75 mm and 6%, respectively. NMR Spectroscopy—All experiments were carried out at 25 °C on Bruker spectrometers operating at 1H frequencies of 600, 750, and 800 MHz equipped with cryogenic (600 MHz) or room temperature (750 and 800 MHz) probes. Data were processed and analyzed with the nmrPipe package (26Delaglio F. Grzesiek S. Vuister G.W. Zhu G. Pfeifer J. Bax A. J. Biomol. NMR. 1995; 6: 277-293Crossref PubMed Scopus (11533) Google Scholar). Throughout most experiments, the TROSY N-H component was selected (27Pervushin K. Riek R. Wider G. Wuthrich K. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12366-12371Crossref PubMed Scopus (2075) Google Scholar, 28Weigelt J. J. Am. Chem. Soc. 1998; 120: 10778-10779Crossref Scopus (174) Google Scholar). HN, N, Cα, and C′ assignments were made from HNCA, HN(CO)CA, HNCO, and HN(CA)CO experiments using sample B and the reported HN, N, and Cα assignments (12Chandra S. Chen X.C. Rizo J. Jahn R. Sudhof T.C. J. Biol. Chem. 2003; 278: 15313-15318Abstract Full Text Full Text PDF PubMed Scopus (418) Google Scholar). Subsequently, Hα and partial Hβ assignments were obtained from TOCSY-TROSY experiments (sample A, τmix = 36 ms). CH3 assignments of Val-γ1,2, Leu-δ1,2, and Ile-δ1 moieties of sample C were based on (H)CCONH and H(C)CONH experiments (29Grzesiek S. Anglister J. Bax A. J. Magn. Reson. Ser. B. 1993; 101: 114-119Crossref Scopus (585) Google Scholar). HN-HN NOEs were measured from HSQC-NOESY-TROSY experiments (sample B, τmix = 170 ms), HN-Hα NOEs from NOESY-TROSY experiments (sample A, τmix = 100 ms), HN-Hγ,δ (Val/Leu/Ile) NOEs from NOESY-TROSY experiments (sample C, τmix = 150 ms), and CH3-CH3 NOEs from CT-HSQC-NOESY-CT-HSQC experiments (sample C, τmix = 150 ms). 3JC′Cγ and 3JNCγ couplings for aromatic and aliphatic residues were obtained from quantitative J-correlation spectroscopy (30Hu J.S. Bax A. J. Biomol. NMR. 1997; 9: 323-328Crossref PubMed Scopus (62) Google Scholar, 31Hu J.S. Grzesiek S. Bax A. J. Am. Chem. Soc. 1997; 119: 1803-1804Crossref Scopus (101) Google Scholar) with dephasing times of 60 and 160 ms, respectively. 1JNH, 1JCαC′, 1JC′N, and 1JNH + 1DNH, 1JCαC′ + 1DCαC′, 1JC′N + 1DC′N couplings were determined from mixed-constant time, HN-coupled HNCO, Cα-coupled HNCO, or quantitative J-correlation HN(CA)CO (32Jaroniec C.P. Ulmer T.S. Bax A. J. Biomol. NMR. 2004; 30: 181-194Crossref PubMed Scopus (25) Google Scholar) and quantitative J-correlation HNCO experiments (33Chou J.J. Delaglio F. Bax A. J. Biomol. NMR. 2000; 18: 101-105Crossref PubMed Scopus (50) Google Scholar) of samples B, D, and E, respectively. The 15N relaxation parameters R1, R2, and {1H}-15N NOE were determined at 60.8-MHz for sample B (34Zhu G. Xia Y.L. Nicholson L.K. Sze K.H. J. Magn. Reson. 2000; 143: 423-426Crossref PubMed Scopus (157) Google Scholar, 35Farrow N.A. Muhandiram R. Singer A.U. Pascal S.M. Kay C.M. Gish G. Shoelson S.E. Pawson T. Formankay J.D. Kay L.E. Biochemistry. 1994; 33: 5984-6003Crossref PubMed Scopus (2012) Google Scholar). For the {1H}-15N NOE measurement, 5 s of presaturation preceded by a recycling delay of 4 s were used for the NOE experiment and a 9-s recycle delay for the reference experiment. Mn2+-induced paramagnetic relaxation enhancements (PRE) of 1HN transverse magnetizations, Γ2, were determined from the difference in R2(1HN) of the Ca2+- and Mn2+-loaded samples, employing the pulse sequence described in Ref. 36Donaldson L.W. Skrynnikov N.R. Choy W.Y. Muhandiram D.R. Sarkar B. Forman-Kay J.D. Kay L.E. J. Am. Chem. Soc. 2001; 123: 9843-9847Crossref PubMed Scopus (151) Google Scholar. Co2+-induced pseudo-contact shifts were found to result in peak doubling of resonances close to the metal ion, indicating the presence of two diastereomers for the cysteaminyl·EDTA·Co2+ complex and, for this reason, were not pursued further. Structure Calculation—Due to the pronounced differences in dynamics experienced by the RDC (see “Results”), the local geometry of Val3-Lys97, exhibiting general order parameters (S2) above 0.5, was generated by molecular fragment replacement (37Delaglio F. Kontaxis G. Bax A. J. Am. Chem. Soc. 2000; 122: 2142-2143Crossref Scopus (218) Google Scholar). The remaining residues are represented by random-coil conformations. A fragment length of 7 residues, 372 Hα, N, Cα, and C′ chemical shifts, and 520 1DNH, 1DCαC′, and 1DC′N residual dipolar couplings (RDCs) were used for molecular fragment replacement, allowing the alignment tensor for each fragment to be an adjustable parameter. For each fragment ten candidates in best agreement with the chemical shifts and RDCs were selected from the molecular fragment replacement data base (37Delaglio F. Kontaxis G. Bax A. J. Am. Chem. Soc. 2000; 122: 2142-2143Crossref Scopus (218) Google Scholar). Backbone dihedral angles were then determined by averaging over all selected candidates of all (overlapping) fragments containing any given angle, but excluding fragments where this angle is located in its N- or C-terminal residue. The alignment tensor magnitudes, Da, calculated for this average structure varies little with fragment lengths ranging from 7 to 11 residues (Supplementary Fig. 1), indicating that dynamics is not significantly impacting the fragment structures over this length range. Standard deviations for the angles were typically between 2 and 10° for most residues, and between 10 and 25° for Ala30-Thr44. The local geometry thus defined was implemented as backbone dihedral restraints in subsequent structure calculations with the program Xplor-NIH 2.9.5 (38Schwieters C.D. Kuszewski J.J. Tjandra N. Clore G.M. J. Magn. Reson. 2003; 160: 65-73Crossref PubMed Scopus (1861) Google Scholar). The backbone HN-HN,α NOEs were found to be in good agreement with this generated structure, and side-chain conformations were restricted by HN-Hγ(Ile/Thr), HN-Hδ(Val/Leu) and CH3-CH3 NOEs. For residues with non-rotamer-averaged 3JC′Cγ and 3JNCγ couplings, the χ1 rotamer state was selected based on those couplings (30Hu J.S. Bax A. J. Biomol. NMR. 1997; 9: 323-328Crossref PubMed Scopus (62) Google Scholar, 31Hu J.S. Grzesiek S. Bax A. J. Am. Chem. Soc. 1997; 119: 1803-1804Crossref Scopus (101) Google Scholar). For the translation of PRE Γ2 values into distances, the cysteaminyl-EDTA-Mn2+ group was represented by an ensemble of three structures during simulated annealing, and the model-free Solomon-Bloembergen equations were applied (39Iwahara J. Schwieters C.D. Clore G.M. J. Am. Chem. Soc. 2004; 126: 5879-5896Crossref PubMed Scopus (283) Google Scholar). The Γ2 values of repeat region residues exhibited a constant, reproducible background (Supplementary Fig. 2), which was subtracted (6 Hz) to achieve agreement between the position of the paramagnetic tag and helix-C, i.e. internal referencing was performed. For helix-N the HN intensity ratios between dia- and paramagnetic samples, IMn2+/ICa2+, followed the Γ2 values except for the first nine residues (Supplementary Fig. 2), suggesting some variation in distance between helix-N and -C with more distant conformers contributing overproportionally to the ensemble-averaged R2(Mn2+) rate, i.e. the calculated interhelical distances will be somewhat biased to more remote distances than the actual average. In order not to introduce opposing restraints, the structure was calculated by only considering the Γ2 values of Lys10-Lys21 of helix-N, besides the values for helix-C. It is also noted that the paramagnetic tag is clearly oriented to one side of helix-C (Supplementary Fig. 3). To aid in the placement of the helices relative to each other, global alignment tensors were used to describe RDC for residues with S2 > 0.8 and with fragment number exhibiting |Da| > 6 Hz (residues 9–23, 54–63, and 70–81 for sample D and residues 9–23, 57–63, and 70–81 for sample E). Simulated annealing calculations were carried out from 500 to0Kin the presence of an empirical backbone-backbone hydrogen-bonding potential (40Grishaev A. Bax A. J. Am. Chem. Soc. 2004; 126: 7281-7292Crossref PubMed Scopus (100) Google Scholar) and torsion angle potentials of mean force (41Kuszewski J.J. Clore G.M. J. Magn. Reson. 2000; 146: 249-254Crossref PubMed Scopus (93) Google Scholar). In order not to distort helix-C by the relatively strong forces acting on the paramagnetic tag, the dihedral angle restraints of residues 85–88 were tightened. To allow a wider range of helix-helix orientations, the dihedral angle restraints of residues 39–43 were relaxed. The energy-minimized average structure, calculated from the ensemble of 20 lowest energy structures, was deposited in the protein data base under accession code 1XQ8. SDS Titration—Titrations were performed in 20 mm NaH2PO4/Na2HPO4 (pH 7.4)/6% D2O at 25 °C and 800 MHz. The aS starting concentration was 0.2 mm, and 11 titrations steps were performed from aS:SDS molar ratios of 0:422 (Fig. 2). At each titration point a TROSYHSQC spectrum was recorded (t15N,max = 144.9 ms, t1HN,max = 113.9 ms). In the absence of SDS selected aS assignments were made based on an HNCA experiment and comparison of aS wild-type spectra with spectra of aS mutants A30P and A53T. Between aS:SDS ratios of 14 to 43 the assignment of a few residues with characteristic chemical shift was made. For aS:SDS ratios ranging from 58 to 422 peaks shift relatively little and assignments followed directly from the assignments of sample B (cf. above). Association of aS with SDS Micelles Is Quantitative and Not Strongly Dependent on the Chemical Makeup of the Micelle— The transition of aS from the free to the micelle-bound state has been examined by NMR spectroscopy to monitor the structuring of aS and to assess the tightness of aS binding to a micelle composed of SDS molecules. A tight interaction is mandatory for structure determination by NMR. Subsequently, mixed-micelles in which 70% of the SDS molecules have been replaced by the lipid-like detergent dodecyl phosphocholine (DPC) were compared with SDS-only micelles to assess the influence of micelle chemical makeup on aS structural properties. In the presence of aS, a SDS micelle was estimated to contain ∼70 SDS molecules (12Chandra S. Chen X.C. Rizo J. Jahn R. Sudhof T.C. J. Biol. Chem. 2003; 278: 15313-15318Abstract Full Text Full Text PDF PubMed Scopus (418) Google Scholar). Consequently, up to a molar aS:SDS ratio of ∼1:70, multiple aS molecules may bind to a single micelle. A titration with SDS was carried out to cover molar aS:SDS ratios ranging from 0 to 422 at an initial aS concentration of 0.2 mm. Chemical shift changes of the backbone 1HN and 15N nuclei of each residue were monitored during the titration, and the 1HN chemical shift changes of four residues, Ala29, Thr54, Ala56, and Met116, are presented here (see Fig. 2). Changes in chemical shifts are sensitive indicators of changes in local structure and, thus, excellent probes of the transition of aS from the free to the micelle-bound state. The first titration step, to an aS:SDS ratio of 1:14, broadened the resonances of the entire repeat region of aS (Fig. 1) essentially beyond detection. The broadening may arise from the binding of multiple copies of aS, resulting in large particle sizes, or exchange kinetics that is intermediate on the NMR timescale. As the titration progresses (aS:SDS = 1:29), the missing resonances reappear at new positions. For residues of the C-terminal tail region (Fig. 1), no significant change in signal intensity was observed during these two titration steps, but essentially all of their very small shift changes take place there (Fig. 2D). At an aS:SDS ratio of 1:43, two species arise for each repeat region residue, one resonance that often, but not always, appears close to the previous titration step and one that is headed toward the final micelle-bound state (Fig. 2C). At aS:SDS ratios of 1:58 and higher only one resonance was observed for all residues, which asymptotically approached the final micelle-saturated state (Fig. 2, A–C). In good agreement with the previous estimate of the SDS micelle size (12Chandra S. Chen X.C. Rizo J. Jahn R. Sudhof T.C. J. Biol. Chem. 2003; 278: 15313-15318Abstract Full Text Full Text PDF PubMed Scopus (418) Google Scholar), the chemical shift changes for repeat region residues leveled off rapidly after the titration point at an aS:SDS ratio of 1:73 was reached (Fig. 2, A–C), i.e. after each micelle had at most a single aS molecule bound, showing that quantitative aS·micelle complex formation had occurred. Remarkably, for most residues of the repeat region, if not all, the chemical shift in the absence of SDS does not fall on the smooth curve toward the final micelle-saturated state (Fig. 2, A and B), indicating that the repeat region progresses through an intermediate state to reach the micelle-saturated state. Given the excess of aS over micelles, this intermediate state could involve two or more micelle-bound aS molecules. However, an analogous titration monitored by circular dichroism spectroscopy has found the aS helical content to increase until a aS:SDS ratio of ∼1:70 is reached (12Chandra S. Chen X.C. Rizo J. Jahn R. Sudhof T.C. J. Biol. Chem. 2003; 278: 15313-15318Abstract Full Text Full Text PDF PubMed Scopus (418) Google Scholar), indicating that below this ratio the free aS state remains populated and/or that, within such higher order complexes,