Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Phys Chem B. Author manuscript; available in PMC 2011 April 8.
Published in final edited form as:
PMCID: PMC2858569

Tryptophan probes at the α-synuclein and membrane interface


Understanding how environmental factors affect the conformational dynamics of α-synuclein (α-syn) is of great importance because the accumulation and deposit of aggregated α-syn in the brain are intimately connected to Parkinson’s disease etiology. Measurements of steady-state and time-resolved fluorescence of single tryptophan-containing α-syn variants have revealed distinct phospholipid vesicle and micelle interactions at residues 4, 39, 94, and 125. Our circular dichroism (CD) data confirm that Trp mutations do not affect α-syn membrane binding properties (apparent association constant equation M1 for all synucleins) saturating at an estimated lipid-to-protein molar ratio of 380 or approximately 120 proteins covering ~7% of the surface area of an 80 nm diameter vesicle. Fluorophores at positions 4 and 94 are the most sensitive to the lipid bilayer with pronounced spectral blue-shifts (W4: Δλmax ~23 nm; W94: Δλmax ~10 nm) and quantum yield increases (W4, W94: ~3 fold) while W39 and W125 remain primarily water-exposed. Time-resolved fluorescence data show that all sites (except W125) have subpopulations that interact with the membrane.

Keywords: Parkinson’s disease, time-resolved fluorescence, membrane binding, phospholipid vesicles, SDS micelles


Parkinson’s disease (PD) is a prevalent age-related neurodegenerative disease with an estimated 1.5% lifetime risk for developing the disease.1 PD patients exhibit resting tremors, bradykinesia, rigidity, and impaired balance caused by the loss of dopamine-producing cells in a midbrain region called the substantia nigra.1 The origin of dopaminergic cell death is unknown; however, biochemical, histological, and genetic studies have implicated a neuronal protein, α-synuclein (α-syn), in PD pathogenesis. For example, α-syn is the primary component of intracellular proteinacious aggregates called Lewy bodies and neurites found in PD patients.1 Age-related increases in α-syn concentration also are observed in nigral brain regions.2 Furthermore, genetic findings link early-onset PD to the gene triplication of α-syn and three missense α-syn mutations (A30P, E46K, A53T).3 Whether sporadic or genetic in origin, it is not simply the presence of α-syn that is associated with PD, but also its corresponding conformational state. While soluble α-syn is characterized as natively unfolded in vitro,4 it is the aggregated β-sheet form (amyloid fibrils) that is found in Lewy bodies and neurites.5,6 Protofibrillar (oligomeric) species have been implicated as pathogenic agents;7,8 however, the exact role of α-syn oligomers in cytotoxic and aggregation pathways is not clear. Oligomers have been shown to both accelerate9 and inhibit α-syn fibril formation.10 It is these features that consequently place PD in the category of a protein misfolding disease.6,11

Membrane interactions are of particular interest because α-syn localizes near synaptic vesicles and mitochondrial membranes in vivo.1214 Specifically, the protein undergoes disordered-to-helical structural changes with the addition of membrane mimics such as SDS micelles and upon binding to anionic phospholipid vesicles of varying size and composition.15 The ability of α-syn to bind to membranes is anticipated by its amino acid sequence: of its 140 total residues, the first 89 residues contain seven imperfect eleven residue repeats (XKTKEGVXXXX) reminiscent of membrane binding apolipoproteins (apoPs).16,17 The hydrophobic central peptide fragment, or non-amyloid β component (NAC), is comprised of residues 61–95. While this region, first identified in amyloid plaques of Alzheimer’s disease patients,18 is involved in membrane association, it is implicated mainly as an initiation site for protein aggregation.19 In contrast, the C-terminal domain of α-syn is highly acidic (15 carboxylates) and generally does not associate with lipids; however, upon binding calcium ions (charge neutralization), even this region interacts with membranes.20 Notably, vesicle leakage assays reveal that protofibrillar α-syn can permeabilize membranes suggesting yet another possible mechanism for PD pathogenesis.8,21,22 The presence of anionic phospholipids or detergents indeed, have been shown to both stimulate23 and retard24,25 amyloid formation. Because of the strong relationship between membranes and α-syn aggregation behavior, measurements of protein conformation and dynamics on the membrane surface are necessary to gain insight into how this protein converts from a benign to a pathogenic form.

Currently, two membrane-bound α-syn conformations have been derived from spectroscopic methods. Structures assigned from NMR spectroscopy on SDS-micelle-bound protein2632 and site-directed spin labeling EPR on vesicle-bound α-syn30 are characterized by two N- and C-terminal, antiparallel, α-helices (residues 3–37 and 45–92) separated by a short linker (residues 38–44). However, several additional EPR3335 and single molecule fluorescence3638 studies show α-syn conformation is dominated by one continuous, extended α-helix when bound to small unilamellar vesicles (SUVs). A recent NMR study suggests that there are multiple distinct α-syn membrane binding modes that are dependent on α-syn-to-lipid stoichiometry.39 Despite numerous work on membrane composition and size effects on α-syn membrane binding properties, a consensus on the conditions required to promote α-syn into one, both, or other variants of the proposed membrane-bound structures remains elusive. It is not clear whether or not these conformers are modulated by membrane surface availability (i.e. surface area available for protein binding), detergent or phospholipid headgroup composition, or simply solution conditions. Furthermore, given the ability for this protein to undergo large conformational rearrangements, it is even more likely that bound structures are not mutually exclusive and can interchange.

To develop a detailed understanding of how membranes influence α-syn conformation, site-specific probes of protein conformational heterogeneity and polypeptide-membrane interactions are necessary. Fluorescence spectroscopy is particularly suited for this application because of the availability of environmentally sensitive fluorophores and the ease of performing experiments near physiological temperatures and concentrations even down to a single molecule. In prior work, we have exploited tryptophan40 and 5-fluorotryptophan41 as fluorescent amino acids of α-syn conformation and dynamics42,43 in solution. The emission properties of the indole side chain are exquisitely responsive to local environment and conformation providing remarkably useful probes of protein-lipid interactions.4449 In this study, we have employed anionic SUVs and SDS micelles as membrane mimics to investigate membrane-induced conformational changes by fluorescence as well as circular dichroism (CD) spectroscopy. Tryptophan was substituted at four different aromatic residues (F4W, Y39W, F94W, and Y125W) to report information on local polypeptide environment and conformational heterogeneity between SUV- and SDS-micelles-bound α-syn. Furthermore, insights into the role of surface coverage (total number of proteins bound and respective surface area occupied) were extracted from saturable equilibrium binding curves for all α-synucleins in the presence of SUVs.



Lipids in chloroform were purchased from Avanti Polar Lipids (Alabaster, AL), N-acetyl-tryptophanamide (NATA), and sodium dodecyl sulfate (SDS) were purchased from Sigma (St. Louis, MO) and used as received.

Lipid Vesicle Preparation

Lipid vesicles were made from a 1:1 molar ratio of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphate (POPA) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC). Solutions of 1:1 POPA and POPC dissolved in chloroform were dehydrated under a nitrogen stream for 15 minutes followed by vacuum desiccation for 45 minutes to ensure complete removal of organic solvent. Dehydrated samples were resuspended in 10 mM NaPi, 100 mM NaCl buffer, pH 7.4 to a final concentration of 5 mg/mL. Complete dissolution was achieved by bath sonication for 5 minutes. SUVs were prepared via ultrasonication in a water bath (45 minutes, 50% duty cycle, microtip limit, Branson 415 Sonifier). SUV solutions were diluted to a final concentration of 2.5 mg/mL and equilibrated overnight (14–24 h) at 30°C. Vesicles were prepared freshly for each experiment. Stock lipid concentrations were confirmed using the Bartlett method.50 An average vesicle hydrodynamic radius of 40 nm was measured using an 18-angle photometer dynamic light scattering instrument with QELS detection (Wyatt EOS). Vesicle sizes ranged from 20–100 nm as visualized by electron microscopy (JEOL-1200EX).

Recombinant Protein Expression and Purification

Wild-type and single Trp-containing α-synucleins were expressed51,52 and purified40 as previously reported with minor modifications. All purification steps are now performed at 4 °C and a HiPrep 16/10 DEAE FF column (GE Healthcare) is used for the first chromatographic step. The protein was eluted with a linear gradient from 100–300 mM NaCl in 20 mM Tris buffer (pH 7.4–8.0). Protein concentrations were determined using a molar extinction coefficient estimated on the basis of amino-acid content: ε280 nm = 5,120 M−1cm−1 (wild-type); ε280 nm = 9,970 M−1cm−1 (Y39W and Y125W); ε280 nm = 10,810 M−1cm−1 (F4W and F94W). The purity of all protein samples was assessed by native- and SDS-PAGE on a Pharmacia Phastsystem (Amersham Biosciences) visualized by silver-staining methods. The protein molecular weights were confirmed by ESI-MS. All purified proteins were concentrated using Centriprep YM-3 (MWCO 3kD, Millipore) and stored at −80 °C.

Steady-state Spectroscopic Measurements

Spectroscopic measurements were made as follows: absorption, Cary 300 Bio spectrophotometer; CD, Jasco J-715 spectropolarimeter (198–260 nm, 1 nm steps, 1 nm bandwidth, 0.5 s integration time, and 50 nm/min); and luminescence, Fluorolog-3 spectrofluorimeter (λex = 295 nm, λobs = 300–500 nm, 0.20–0.25 s integration time, 1 nm excitation and emission slit widths). All measurements were collected at 25 °C using temperature controlled cuvette holders except for absorbance experiments. Prior to experiments, all protein samples were exchanged into the appropriate buffer (10 mM NaPi, 100 mM NaCl, pH 7.4) using gel filtration chromatography (PD-10 column, GE Healthcare). Protein-vesicle and protein-SDS solutions were incubated for at least 30 minutes at room temperature prior to measurement. Steady-state fluorescence spectra were measured before and after laser experiment to verify minimal photodamage (<10%).

Time-resolved Fluorescence Measurements

Tryptophan fluorescence decay kinetics were measured using the fourth harmonic (295 nm) of a regeneratively-amplified femtosecond Ti:sapphire (Clark-MXR) pumped optical parametric amplifier laser (Light Conversion) as an excitation source (60–260 μW, 1 kHz) and a picosecond streak camera (Hamamatsu C5680) in photon counting mode for detection. Tryptophan emission between 325 and 400 nm was selected through edge (REF-325) and short-pass (UG-11) filters (CVI Laser). Protein samples (3–5 μM in 10 mM NaPi, 100 mM NaCl buffer, pH 7.4) were deoxygenated on a Schlenk line by 3 sets of 5 repeated evacuation/Ar fill cycles over 30 minutes. Buffer solutions were filtered (0.22 μm membrane) to remove any particulate matter. A collection temperature of 25 °C was maintained using a temperature controlled cuvette holder. To ensure good signal to noise, all kinetics data were collected such that at least 10,000 counts were achieved in highest channel.

Data Analysis

Circular Dichroism

Mean residue ellipticity, [Θ] (deg cm2 dmol−1), was calculated according to the equation equation M2, where θ is the measured ellipticity (mdeg), c, the sample concentration (mM), l, the path length (cm), and N, the number of amino acids. Percent α-helicity was calculated using the equation equation M3. Reference values of [Θ]222 nm = 0 and −36,000 (deg cm2 dmol−1) were used for 0% and 100% helicity respectively.53

Protein-membrane Equilibrium

Binding equilibrium data were modeled according to the simple law of mass action for identical and independent binding sites.54,55 Estimation of apparent association constants were extracted for the following two-state equilibrium:

equation M4

where Pf is the free α-syn concentration, Bf is the free binding sites, and PBb is the bound α-syn concentration (occupied binding sites). The total binding sites (Bt) is defined as Bmax L, where Bmax is the maximum binding sites per lipid molecule, and L is the total lipid concentration. Substituting Pt = Pf + PBb and Bt = Bf + PBb into the expression for equation M5 and using the definition of Bt, we obtain:

equation M6

Rearranging this equation yields:

equation M7

Since our CD data (Δ[Θ]222 nm) is proportional to equation M8, Eq. 3 can be solved for our fitting equation (4):

equation M9

Fitting was performed using IGOR Pro 6.01 (Wavemetrics).

Tryptophan Fluorescence Decays

Fluorescence decay kinetics were logarithmically compressed (100 points per time decade) and normalized (I(t=0) = 1). Kinetics were fit using a MATLAB (The Mathworks, Inc.) algorithm (LSQNONNEG; hereafter referred to as NNLS) that minimizes the sum of the squared deviations (χ2) between the observed and calculated values of I(t), subject to a nonnegativity constraint on the probability distribution of rate constants, P(k). NNLS fitting produces the narrowest P(k) distributions.40,56

Protein Packing on Lipid Vesicles

The number of lipid molecules per vesicle was estimated by dividing total vesicle surface area by the area per lipid molecule parallel to the bilayer surface. A total of 56,795 lipid molecules per vesicle was calculated using 40 nm for outer vesicle radius, 0.643 nm2 for the lipid area (DPPC), and 3.86 nm for the bilayer thickness.57,58 This number is consistent with that predicted from the molecular weight of a hydrated vesicle.59 Estimation of the maximum number of α-synucleins binding to vesicles was calculated by modeling the proteins as rigid bodies of dimensions consistent with α-syn in continuous extended-35 (14.97 nm) and bent-helical29,30 (approximate segment lengths: helix 1 = 5.5 nm, linker = 1.4 nm, helix 2 = 6.8 nm) conformations. Using a value of 0.75 nm for the protein-helix diameter, we estimate an area of 10.3 nm2 and 11.3 nm2 is required for α-syn to bind in bent-helical and extended-helical conformations, respectively. Maximum α-synuclein packing number on a vesicle using Bmax (maximum number of protein binding sites per lipid molecule) were calculated considering that only the lipid molecules on the outer leaflet of the vesicle (~31,300) are available for protein binding. Percent surface coverage was calculated using the maximum total of protein sites extracted from Bmax.


Membrane-mediated α-helical Structure Formation

To monitor site specific interactions of α-syn membrane binding, four single tryptophan-containing proteins (F4W, Y39W, F94W, and Y125W) were produced. To assess the effects of the Trp mutations, CD spectroscopy was employed to characterize vesicle-induced secondary structural changes for all α-synucleins. In pH 7.4 buffer solutions (10 mM NaPi, 100 mM NaCl, 25 °C), CD spectra for wild-type and Trp variants (3–5 μM) exhibit a dominant minimum near 200 nm that is consistent with random coil configurations. With increasing amounts of anionic phospholipid vesicles (equal molar 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphate (POPA, anionic −1 per headgroup) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC, zwitterionic)), the CD spectra show prototypical features (negative maxima at 208 and 222 nm) for α-helical structure (Figure 1). The spectra for the Trp mutants were indistinguishable from the wild-type protein (data not shown), indicating that this naturally occurring fluorophore does not change the secondary structure of vesicle-bound protein under these solution conditions. An isodichroic point (204 nm) suggests that binding is consistent with an unstructured to α-helical conformation transition with a maximum total α-helical content of ~67% (mean residue ellipticity at 222nm, [Θ]222 nm, = −23,130 deg cm2 dmol−1), consistent with previously reported values.15 Similarly, all α-syn variants exhibit an α-helical content of ~67% in the presence of SDS micelles, demonstrating not only that substitution of Trp residues does not perturb micelle-bound structure, but also that α-helical content is not a distinguishing feature between vesicle- and SDS micelles-bound α-syn (Figure S1). As α-syn has substantially reduced affinity for zwitterionic vesicles,36 100% POPC vesicles (at highest lipid concentration ~2 mg/mL) were employed to assess macromolecular crowding effects. No apparent conformational transitions were observed for wild-type and Trp mutants verifying the involvement of electrostatic interactions in membrane binding affinity (Figure S2).

Figure 1
Phospholipid vesicles induced secondary structure in α-synuclein. Representative far UV CD spectra of Y39W α-syn (5 μM) as a function of increased lipid vesicles (0–2 mM 1:1 POPA:POPC in 10 mM NaPi, 100 mM NaCl buffer at ...

Membrane binding curves were generated for all proteins using mean residue ellipticity (Figure 2). All isotherms are nearly identical and reversible, with similar midpoint transitions (lipid-to-protein molar ratio of 120–160) approaching saturation at a lipid-to-protein molar ratio of ~380. Interestingly, while α-synuclein membrane binding curves generated from calorimetric60 and single molecule fluorescence36 techniques are not well described by membrane-partition models, we achieved adequate fits using a law of mass action for identical independent binding sites.54,55 An apparent association constant ( equation M10) and a total number of binding sites per lipid molecule (Bmax ~0.004) were extracted for all α-synucleins. To ensure that equation M11 values derived from CD data were independent of bound polypeptide concentration and to assess nonideality in the low lipid-to-protein regimes, inverse fluorescence titrations also were performed (varying protein concentrations (1–20 μM) while maintaining a constant lipid concentration (1 mM), Figure S3).44,61 Because α-syn membrane binding likely is affected by electrostatic repulsion of the acidic C-terminal tail, our equation M12 values are unexpectedly similar to several full length apoPs (apoA-I and apoE4 (egg PC), equation M13 and 8.3 × 106 M−1, respectively).55,62 While equation M14 values for α-syn suggest comparable affinity as the apoPs, we would anticipate an isolated N-terminal amphiphilic α-syn polypeptide would bind to membranes even more tightly. Furthermore, the full length α-syn membrane affinity would be expected to increase in solution conditions favoring charge neutralization (acidic pH) and the binding of biomolecules or metal ions such as polyamines63 and calcium ions20 to the C-terminus.

Figure 2
Comparison of vesicle binding properties of wild-type and fluorescent single Trp α-synucleins. (Top) Schematic representation of α-synuclein primary sequence with the tryptophan mutations (W4, W39, W94, and W125) and the non-amyloid β ...

Estimation of α-Synuclein Vesicle Surface Coverage

Using CD binding curves we assessed the possible role of membrane surface coverage in modulating α-syn conformation. Assuming that only outer leaflet lipids bind proteins and using Bmax obtained from fits, we find that the vesicle surface contains ~120 protein binding sites with ~250 lipids per binding site. If all binding sites are occupied, only ~7% of the total vesicle surface would be covered by proteins (See Materials and Methods). Simple geometrical estimates of the total possible number of α-syn proteins bound to SUVs in both bent- and extended-helical conformations are higher than our experimentally derived value (~1000 and ~650 respectively; Supporting Information). These disparities likely are attributable to repulsive electrostatic protein-protein and even phospholipid-protein interactions with the acidic C-terminus. Our data suggest that α-syn proteins are not closely packed on the membrane surface at binding saturation; this result is consistent with electron spin resonance data (36–100 lipids per protein; DMPG).64 Interestingly, the surface coverage determined for α-syn is low compared to other peripheral membrane proteins such as apocytochrome c (2 lipids per protein; DMPG)64,65 and myelin basic protein (18 lipids per protein; DMPG).64,66

Notably, α-syn fibril formation in the presence of low lipids (lipid-to-protein molar ratio ~4) is ~3.5 times faster than protein alone.23 For this case, we estimate a staggering ~14,000 proteins-per-vesicle; a value significantly higher than the maximum binding sites per vesicle. While our data suggest that α-syn proteins are not closely packed at saturation, it is plausible that changing the lipid-to-protein ratio could result in conformational rearrangement at the vesicle surface, a phenomena observed in apocytochrome c.65 Because intracellular synaptic vesicles contain heterogeneous distributions of phospholipid headgroups and membrane-associated proteins, the biological relevance of the conformational rearrangement of α-syn due to protein packing remains to be elucidated.

Site Specific α-Synuclein Membrane Interactions Revealed by Tryptophan Probes

Site specific interactions of α-syn with SUVs were monitored by steady-state and time-resolved Trp fluorescence as a function of POPA:POPC vesicle concentration (0 to 2.5 mM in 10 mM NaPi, 100 mM NaCl buffer at pH 7.4) (Figure 3). In solution, all four Trp residues exhibit identical fluorescence maxima at 348 nm consistent with water-exposed side chains and similar local protein conformations. The model complex, N-acetyl-tryptophanamide also has a maximum at 348 nm under similar solution conditions. However, upon additions of phospholipid vesicles, Trp photophysical properties become site-dependent with differences in the emission maxima (W4 > W94 ~W39 [dbl greater-than sign] W125, Figures 3 and S4) and quantum yields (W4 ~ W94 > W39 [dbl greater-than sign] W125, Figure 3). In particular, W4 exhibits a pronounced spectral blue shift (Δλmax = 23 nm) reminiscent of integral (E. Coli outer membrane protein A (OmpA): Δλmax = 25–28 nm, DMPC)67,68 rather than surface-bound membrane proteins (mimics of the class A1 helices in exchangeable apoPs: Δλmax = 14 nm, POPC).69 In addition, the three fold increase in W4 quantum yield is consistent with this side chain partitioning into the bilayer. While this residue was initially predicted to be membrane exposed using SDS micelles studies,29 it was not characterized in the α-syn vesicle-bound EPR structure (residues 9–89).35

Figure 3
Site specific tryptophan fluorescence of α-synucleins. Steady-state emission spectra of W4, W39, W94, and W125 (top to bottom) α-syn mutants in the presence of lipid vesicles (0–2.5 mM 1:1 POPA:POPC in 10mM NaPi, 100 mM NaCl buffer, ...

In contrast, W39, which is located in the loop region of the micelle-bound NMR structure29 and water-exposed in the SUV-bound conformation,35 exhibits a modest spectral blue shift (λmax ~339 nm, Δλmax ~10 nm) and quantum yield increase (1.2 fold) in the presence of vesicles. This reduced W39 fluorescence changes compared to that of W4 point to greater water exposure, consistent with proposed micelle and liposome bound conformations. Located in the C-terminal end of the NAC region, W94 emission exhibits a similar spectral shift as W39 (λmax ~338 nm, Δλmax ~9 nm), but a significant quantum yield increase (comparable to W4) upon vesicle binding suggesting that this indole is lipid-exposed or in a conformation that alleviates nonradiative pathways (i.e. restriction of mobility and electron transfer). The pronounced spectral response of W94 to membranes is particularly interesting and somewhat unexpected because this residue is predicted to be solvent exposed28,32,35 and, while micelle facing, is unstructured when bound to SDS.29 It is important to note that Trp fluorescence quantum yields are modulated by local charges as well as fluorophore mobility.70 If the indole ring is in close proximity to a positive (negative) charge then its quantum yield could increase (decrease). It is plausible as the protein binds to the membrane surface, W94 may become closer to K96, the closest intramolecular positive charge. Alternatively, W94 may reside near the trimethylammonium cation of the POPC headgroup. Not surprisingly, W125 fluorescence remains unchanged in the presence of phospholipid vesicles confirming that the C-terminal tail remains in solution. Although W125 is unresponsive, it provides a control to ensure that our spectroscopic changes are solely due to protein-membrane interactions and not the mere presence of the membranes themselves. To further verify that the emission features were not a result of macromolecular crowding, the spectra of all Trp variants in the presence of zwitterionic POPC vesicles were obtained. No changes were observed even at the highest vesicle concentrations (Figure S2).

Because secondary structural characterization does not provide a distinguishing marker of micelle- vs. vesicle-bound proteins (Figure S1), Trp fluorescence data also were collected for SDS micelles-bound α-synuclein (Figure 4). In all positions except W125, Trp data reveal significant differences for liposome- and SDS micelles-bound conformations. The W4 spectrum in the presence of SDS micelles (40 mM, critical micelle concentration = 1.33 mM71 (100 mM NaCl, 25°C)) exhibits a quantum yield decrease (20% decrease compared to in solution) and the spectral shift is greatly reduced (Δλmax = 23 (vesicles) → 15 (micelle) nm). The reduced quantum yield suggests considerable differences in the local polypeptide conformation near the vicinity of residue 4. The reduced spectral shift also supports different protein-lipid interactions (head group vs. hydrocarbon chain) as well as variations in micellar vs. bilayer microenvironment hydration. Interestingly, in the presence of SDS micelles, W39 also exhibits a decrease in quantum yield which may be due to its proximity to a Trp quencher, such as His50, the most likely side chain in the amino acid sequence to participate in excited-state deactivation processes (there is no native Cys and only one His).45 Examination of the SDS structure (PDB 1XQ8)29 reveals that the nearest negative charges (E35) is over 15 Å away and the nearest positive charge (K45) is much closer (8 Å) to Y39. Because these side chains are relatively distant, likely it is the proximity of W39 to the sulfonate groups that is responsible for the reduced quantum yield (vide supra). In contrast, W94 appears to be the most sensitive probe for micelle interaction with both a spectral shift and an increase in quantum yield, qualitatively similar to that of the bilayer. Not surprisingly, the W125 spectrum remains unchanged, indicating that both micelle- and vesicle-bound conformers have C terminal regions that remain solvent exposed.

Figure 4
Comparison of tryptophan fluorescence for micelles- and liposome-bound α-synuclein. Steady-state emission spectra for W4, W39, and W94 (top to bottom) in buffer (10mM NaPi, 100 mM NaCl, pH 7.4) (light blue), bound to SDS micelles (40 mM) (red), ...

Our steady-state fluorescence data clearly demonstrate site specific differences for the SDS micelle and liposome-bound α-syn conformers. While the trends in spectral shifts are consistent between micelle- and liposome-bound conformations, the quantum yields reveal distinct behavior, particularly highlighting the local polypeptide environmental changes of W4. Notably, W4 and W94 were not characterized in the continuous α-helical (residues 9–89) membrane-associated structure reported by Langen and coworkers.35 If we use the periodicity of 3.67 amino acids per turn from Langen35 and previous NMR studies on micelles,32 then W4 and W94 are predicted to be lipid- and solvent-exposed, respectively. Since our data demonstrate that both W4 and W94 partition into the bilayer, we suggest that this α-helical periodicity (3.67 aa per turn) spans from residues 4 to 89 and is not extended to position 94. However, the strong spectral response we observe for W94 does suggests that this residue interacts directly with the micelle. Taken together, our vesicle and micelle data support that there are more protein-membrane interaction sites beyond helical regions.

Conformational Heterogeneity of SUV-bound α-Synuclein Monitored by Tryptophan Kinetics

To characterize membrane-bound protein conformational heterogeneity we measured tryptophan decay kinetics for W4, W39, W94, and W125 α-synucleins (Figure 5). For all variants and solution conditions, decay kinetics are not well described by single exponential functions as it is typically observed for Trp-containing proteins attributable to side chain local environments and conformations.4548 Accordingly, we used a nonnegative linear least squares method to analyze our data which produces the narrowest probability distribution of lifetimes (P(τ)) required to fit the decay kinetics and distinguish protein subpopulations (See Figure S5 for residuals).40 Our time-resolved data are consistent with steady-state observations (W4 and W94: increases in average fluorescence lifetime (<τ>); W39: modest changes in <τ>; W125: no changes, Figure 5). Notably, the P(τ) extracted from our fits provided insights in the evolution of subpopulations in the protein ensemble and revealed local conformational heterogeneity that was not apparent through steady-state analyses.

Figure 5
Time-resolved Trp fluorescence decay kinetics of W4, W39, W94, and W125 (top to bottom) α-syn mutants in the presence of lipid vesicles (1:1 POPA:POPC (midpoint and saturated concentrations: 0.5 mM and 2.0 mM respectively) in 10 mM NaPi, 100 mM ...

In solution (10 mM NaPi, 100 mM NaCl, pH 7.4), two (<τ1> = 3.0 ns, P1) = 57%; <τ2> = 1.0 ns, P2) = 43%) and three (<τ1> = 4.2 ns, P1) = 24%; <τ2> = 2.1 ns, P2) = 59%; <τ3> = 0.5 ns, P3) = 17%) distinct populations were observed for W4 and W94, respectively (Figure 6). For both W4 and W94, vesicle-bound decays generally are characterized by increases in average lifetimes as a function of added vesicles; however, the evolution of the individual lifetimes is distinct for W4 and W94. Upon vesicle addition, W4 exhibits characteristic <τ> increases (<τ1>: 3.0 ns → 5.7 ns; <τ2>: 1.0 ns → 1.7 ns) with only modest changes in relative amplitudes (P1): 57% → 41%; P2): 43% → 59%). These average lifetime increases are consistent with those that have been observed for folding of single-Trp containing OmpA into DMPC SUVs (τ1: 4.5–5.5 → 5.3–6.3 ns; τ2 = 1.5–1.8 → 2.1–2.7 ns (urea → DMPC)).68 The small relative amplitude changes in P(τ) coupled to increases in both <τ1> and <τ2> suggest that all W4 subpopulations are sensitive to the presence of the bilayer. Similar to W4, two W94 subpopulations change with vesicle addition; however, W94 kinetics reflect an increased conformational heterogeneity not observed in W4. While both τ1 and τ2 increase in average lifetime (<τ1>: 4.2 ns → 7.7 ns; <τ1>: 2.1 ns → 2.7 ns) indicative of a membrane-bound conformations, the relative amplitude of τ2 is appreciably decreased in favor of τ1 (P1): 24% → 40%; P2) 59% → 41%) suggesting a conformational transition that is dependent on lipid-to-protein ratio. The presence of a third component τ3 that exhibits no significant changes in <τ> or P(τ) even at saturating lipid concentrations in addition to the change in P2) is suggestive of a unbound subpopulation. However, because τ3 represents <25% of the total distribution and τ2 exhibits only a modest lipid induced lifetime increase, the exact nature of this heterogeneity is not clear. The presence of an unbound subpopulation would not be surprising because of the electrostatic repulsion due to the acidic C-terminus in this region. Alternatively, because the Trp excited state captures nanosecond snapshots of the protein ensemble, heterogeneity observed in W94 could reflect side chain dynamics.

Figure 6
Analysis of time-resolved Trp fluorescence decay kinetics for α-synucleins. Representative probability distribution of fluorescence lifetimes (P(τ)) extracted from NNLS fits for W4 (top) and W94 (bottom) α-synucleins in buffer ...

While assignment of distinct populations is challenging due to the smaller dynamic range for changes in W39, inspection of the kinetics do reflect a conformational heterogeneity similar to that of W94 (data not shown). Although the W39 decays accelerate only modestly, there is a consistent development of shorter lifetime components from buffer to vesicle-containing solutions which is atypical for lipid-associated changes. It appears that upon membrane binding, there may be quenchers in closer proximity to W39 as a result of local conformational changes (vide supra). Finally, consistent with steady-state results, W125 decay kinetics are not affected to the presence of vesicles even at saturating lipid conditions.

Time-resolved analyses demonstrate that all sites examined (except W125) have interactions with the membrane. Particularly, our data suggest that W4 is a high affinity membrane binding site with increases in average lifetimes for all components; it appears that W4 kinetics are sensitive to the presence of vesicles even at low lipid concentrations (100 μM, [lipid]midpoint = 500 μM). Though exhibiting only modest changes, reside W39 does exhibit conformational heterogeneity with increases in average lifetime with vesicle addition. Our W39 data agree with both micelle- and liposome-bound structures in that this residue is likely both solvent-exposed and flexible at the interface of the lipid headgroups and the surrounding solvent. While W94 steady-state results show a spectral response comparable to that of W4, W94 kinetics data also reveal conformational heterogeneity for this position with vesicle addition. The local conformational change between the solvent-to-lipid or different lipid-bound states is evidenced in the lipid concentration dependent decrease in P2) in favor of P1). W94 data could indicate that protein-membrane stoichiometry is crucial in modulating membrane interactions in this region of the protein. Moreover, W4 could act as an anchor with all populations sensitive to the membrane at all lipid concentrations, while W94 has both lipid and solvent exposed conformers that are dependent on lipid-to-protein ratio. It is noteworthy that the hydrophobic NAC region has been proposed to be secluded from intermolecular contacts by the intramolecular N- and C-terminal interactions.72 Similarly, it is feasible for membrane stimulated aggregation to arise because as the local protein concentration increases at the surface, polypeptide conformational rearrangement ensues leading to exposure of potential protein-protein interaction sites promoting amyloid formation.


We have demonstrated that Trp fluorescence is a sensitive, site-specific probe of α-syn interactions with both liposomes (quantum yield: W4 ~ W94 > W39 [dbl greater-than sign] W125; <λ>: W4 > W94 ~W39 [dbl greater-than sign] W125) and micelles (quantum yield: W94 > W4 > W125 > W39; <λ>: W4 > W94 ~W39 [dbl greater-than sign] W125). Vesicle-binding curves generated from steady-state CD and fluorescence data can be described by a two-state equilibrium partition model with equation M15. We estimate that α-syn binds to POPA:POPC vesicles with a relatively low surface coverage ~7% (~120 proteins for a 80 nm diameter vesicle) at saturation. Furthermore, measurements of fluorescence decay kinetics reveal the presence of protein conformational heterogeneity in the bilayer, suggesting that the both W4 and W94 exhibit high membrane affinity. Notably, both of these sites have not been characterized previously in the vesicle-bound α-syn structure.35 With this approach, we can determine the crucial protein-to-membrane conditions and key sites of interaction that promote protein aggregation and ultimately, monitor membrane-mediated amyloid formation processes.

Supplementary Material



Supported by the Intramural Research Program of the National Institutes of Health, National Heart, Lung, and Blood Institute. We thank Dr. Hank Fales (Laboratory of Applied Mass Spectrometry) for technical assistance with ESI-MS, Dr. Myoung-soon Hong (Laboratory of Cell Biology), Dr. Mathew Daniels, and Patricia Connelly (Electron Microscopy Core Facility) for technical assistance with EM, Dr. Grzegorz Piszczek (Biophysical Facility) for assistance with the dynamic light scattering measurements, and Michel de Messieres for assistance with computer modeling.



Figures S1–S4. Computer simulation details of random packing geometries on a sphere. This information is available free of charge via the Internet at


1. Lees A, Hardy J, Revesz T. Lancet. 2009;373:2055. [PubMed]
2. Chu YP, Kordower JH. Neurobiol Dis. 2007;25:134. [PubMed]
3. Bisaglia M, Mammi S, Bubacco L. FASEB J. 2009;23:329. [PubMed]
4. Weinreb PH, Zhen WG, Poon AW, Conway KA, Lansbury PT. Biochemistry. 1996;35:13709. [PubMed]
5. Conway KA, Harper JD, Lansbury PT. Biochemistry. 2000;39:2552. [PubMed]
6. Chiti F, Dobson CM. Annu Rev Biochem. 2006;75:333. [PubMed]
7. Conway KA, Lee SJ, Rochet JC, Ding TT, Williamson RE, Lansbury PT. Proc Natl Acad Sci U S A. 2000;97:571. [PubMed]
8. Ding TT, Lee SJ, Rochet JC, Lansbury PT. Biochemistry. 2002;41:10209. [PubMed]
9. Uversky VN, Li J, Fink AL. J Biol Chem. 2001;276:10737. [PubMed]
10. Hong DP, Fink AL, Uversky VN. J Mol Biol. 2008;383:214. [PMC free article] [PubMed]
11. Dauer W, Przedborski S. Neuron. 2003;39:889. [PubMed]
12. Clayton DF, George JM. Trends Neurosci. 1998;21:249. [PubMed]
13. Clayton DF, George JM. J Neurosci Res. 1999;58:120. [PubMed]
14. Maroteaux L, Campanelli JT, Scheller RH. J Neurosci. 1988;8:2804. [PubMed]
15. Davidson WS, Jonas A, Clayton DF, George JM. J Biol Chem. 1998;273:9443. [PubMed]
16. George JM, Jin H, Woods WS, Clayton DF. Neuron. 1995;15:361. [PubMed]
17. Segrest JP, Jackson RL, Morriset JD, Gotto AM. FEBS Lett. 1974;38:247. [PubMed]
18. Ueda K, Fukushima H, Masliah E, Xia Y, Iwai A, Yoshimoto M, Otero DAC, Kondo J, Ihara Y, Saitoh T. Proc Natl Acad Sci U S A. 1993;90:11282. [PubMed]
19. El-Agnaf OMA, Bodles AM, Guthrie DJS, Harriott P, Irvine GB. Eur J Biochem. 1998;258:157. [PubMed]
20. Tamamizu-Kato S, Kosaraju MG, Kato H, Raussens V, Ruysschaert JM, Narayanaswami V. Biochemistry. 2006;45:10947. [PubMed]
21. Volles MJ, Lansbury PT. Biochemistry. 2002;41:4595. [PubMed]
22. Volles MJ, Lee SJ, Rochet JC, Shtilerman MD, Ding TT, Kessler JC, Lansbury PT. Biochemistry. 2001;40:7812. [PubMed]
23. Zhu M, Li J, Fink AL. J Biol Chem. 2003;278:40186. [PubMed]
24. Necula M, Chirita CN, Kuret J. J Biol Chem. 2003;278:46674. [PubMed]
25. Naayanan V, Scarlata S. Biochemistry. 2001;40:9927. [PubMed]
26. Bisaglia M, Tessari I, Pinato L, Bellanda M, Giraudo S, Fasano M, Bergantino E, Bubacco L, Mammi S. Biochemistry. 2005;44:329. [PubMed]
27. Eliezer D, Kutluay E, Bussell R, Browne G. J Mol Biol. 2001;307:1061. [PubMed]
28. Chandra S, Chen XC, Rizo J, Jahn R, Sudhof TC. J Biol Chem. 2003;278:15313. [PubMed]
29. Ulmer TS, Bax A, Cole NB, Nussbaum RL. J Biol Chem. 2005;280:9595. [PubMed]
30. Drescher M, Veldhuis G, van Rooijen BD, Milikisyants S, Subramaniam V, Huber M. J Am Chem Soc. 2008;130:7796. [PubMed]
31. Borbat P, Ramlall TF, Freed JH, Eliezer D. J Am Chem Soc. 2006;128:10004–10005. [PubMed]
32. Bussell R, Eliezer D. J Mol Biol. 2003;329:763. [PubMed]
33. Georgieva ER, Ramlall TF, Borbat PP, Freed JH, Eliezer D. J Am Chem Soc. 2008;130:12856. [PMC free article] [PubMed]
34. Jao CC, Der-Sarkissian A, Chen J, Langen R. Proc Natl Acad Sci U S A. 2004;101:8331. [PubMed]
35. Jao CC, Hegde BG, Chen J, Haworth IS, Langen R. Proc Natl Acad Sci U S A. 2008;105:19666. [PubMed]
36. Rhoades E, Ramlall TF, Webb WW, Eliezer D. Biophys J. 2006;90:4692. [PubMed]
37. Ferreon ACM, Gambin Y, Lemke EA, Deniz AA. Proc Natl Acad Sci U S A. 2009;106:5645. [PubMed]
38. Trexler AJ, Rhoades E. Biochemistry. 2009;48:2304. [PMC free article] [PubMed]
39. Bodner C, Dobson C, Bax A. J Mol Biol. 2009;390:775. [PMC free article] [PubMed]
40. Lee JC, Langen R, Hummel PA, Gray HB, Winkler JR. Proc Natl Acad Sci U S A. 2004;101:16466. [PubMed]
41. Winkler GR, Harkins SB, Lee JC, Gray HB. J Phys Chem B. 2006;110:7058. [PubMed]
42. Lee JC, Gray HB, Winkler JR. J Am Chem Soc. 2005;127:16388. [PubMed]
43. Lee JC, Lai BT, Kozak JJ, Gray HB, Winkler JR. J Phys Chem B. 2007;111:2107. [PMC free article] [PubMed]
44. Ladokhin AS, Jayasinghe S, White SH. Analytical Biochemistry. 2000;285:235. [PubMed]
45. Chen Y, Barkley MD. Biochemistry. 1998;37:9976. [PubMed]
46. Eftink MR. Methods Biochem Anal. 1991;35:127. [PubMed]
47. Beechem JM, Brand L. Annu Rev Biochem. 1985;54:43. [PubMed]
48. Yu HT, Vela MA, Fronczek FR, McLaughlin ML, Barkley MD. J Am Chem Soc. 1995;117:348.
49. Clayton AHA, Sawyer WH. Eur Biophys J Biophys Lett. 2002;31:9. [PubMed]
50. Bartlett GR. J Biol Chem. 1959;234:466. [PubMed]
51. Jakes R, Spillantini MG, Goedert M. FEBS Lett. 1994;345:27. [PubMed]
52. Lee JC, Gray HB, Winkler JR. J Am Chem Soc. 2008;130:6898. [PMC free article] [PubMed]
53. Chen YH, Yang JT, Chau KH. Biochemistry. 1974;13:3350. [PubMed]
54. Tamm LK. Biochim Biophys Acta. 1991;1071:123. [PubMed]
55. Saito H, Dhanasekaran P, Baldwin F, Weisgraber KH, Lund-Katz S, Phillips MC. J Biol Chem. 2001;276:40949. [PubMed]
56. Lee JC, Engman KC, Tezcan FA, Gray HB, Winkler JR. Proc Natl Acad Sci U S A. 2002;99:14778. [PubMed]
57. Nagle JF, Wiener MC. Biochim Biophys Acta. 1988;942:1. [PubMed]
58. Kucerka N, Tristram-Nagle S, Nagle JF. Biophys J. 2006;90:L83. [PubMed]
59. Gazzara JA, Phillips MC, Lund-Katz S, Palgunachari MN, Segrest JP, Anantharamaiah GM, Rodrigueza WV, Snow JW. J Lipid Res. 1997;38:2147. [PubMed]
60. Nuscher B, Kamp F, Mehnert T, Odoy S, Haass C, Kahle PJ, Beyer K. J Biol Chem. 2004;279:21966. [PubMed]
61. White SH, Wimley WC, Ladokhin AS, Hristova K. Methods Enzymol. 1998;295:62. [PubMed]
62. Schulthess G, Compassi S, Werder M, Han CH, Phillips MC, Hauser H. Biochemistry. 2000;39:12623. [PubMed]
63. Fernández CO, Hoyer W, Zweckstetter M, Jares-Erijman EA, Subramaniam V, Griesinger C, Jovin TM. EMBO J. 2004;23:2039. [PubMed]
64. Ramakrishnan M, Jensen PH, Marsh D. Biochemistry. 2003;42:12919. [PubMed]
65. Gorrissen H, Marsh D, Rietveld A, Dekruijff B. Biochemistry. 1986;25:2904. [PubMed]
66. Sankaram MB, Brophy PJ, Marsh D. Biochemistry. 1989;28:9685. [PubMed]
67. Kleinschmidt JH, den Blaauwen T, Driessen AJM, Tamm LK. Biochemistry. 1999;38:5006. [PubMed]
68. Kim JE, Arjara G, Richards JH, Gray HB, Winkler JR. J Phys Chem B. 2006;110:17656. [PMC free article] [PubMed]
69. Mishra VK, Palgunachari MN. Biochemistry. 1996;35:11210. [PubMed]
70. Callis PR, Liu TQ. J Phys Chem B. 2004;108:4248.
71. Helenius A, McCaslin DR, Fries E, Tanford C, Sidney F, Lester P. Methods Enzymol. 1979;56:734. [PubMed]
72. Bertoncini CW, Jung YS, Fernández CO, Hoyer W, Griesinger C, Jovin TM, Zweckstetter M. Proc Natl Acad Sci U S A. 2005;102:1430. [PubMed]