This Article Abstract Full Text (PDF) All Versions of this Article: biophysj.105.079251v1 90/12/4692    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rhoades, E. Articles by Eliezer, D. Search for Related Content PubMed PubMed Citation Articles by Rhoades, E. Articles by Eliezer, D.

Quantification of -Synuclein Binding to Lipid Vesicles Using Fluorescence Correlation Spectroscopy

Elizabeth Rhoades ^*, Trudy F. Ramlall , Watt W. Webb ^* and David Eliezer 

^* Applied and Engineering Physics, Cornell University, Ithaca, New York; and Department of Biochemistry and Program in Structural Biology, Weill Medical College of Cornell University, New York, New York

Correspondence: Address reprint requests to David Eliezer, Weill Medical College of Cornell University, Dept. of Biochemistry, 1300 York Ave., New York, NY 10021. Tel.: 212-746-6557; E-mail: dae2005{at}med.cornell.edu; or to Watt W. Webb, Cornell University, School of Applied & Engineering Physics, 212 Clark Hall, Ithaca, NY 14853-2501. Tel.: 607-255-3331; E-mail: www2{at}cornell.edu.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
-Synuclein (S) is a soluble synaptic protein that is the major proteinaceous component of insoluble fibrillar Lewy body deposits that are the hallmark of Parkinson's disease. The interaction of S with synaptic vesicles is thought to be critical both to its normal function as well as to its pathological role in Parkinson's disease. We demonstrate the use of fluorescence correlation spectroscopy as a tool for rapid and quantitative analysis of the binding of S to large unilamellar vesicles of various lipid compositions. We find that S binds preferentially to vesicles containing acidic lipids, and that this interaction can be blocked by increasing the concentration of NaCl in solution. Negative charge is not the only factor determining binding, as we clearly observe binding to vesicles composed entirely of zwitterionic lipids. Additionally, we find enhanced binding to lipids with less bulky headgroups. Quantification of the protein-to-lipid ratio required for binding to different lipid compositions, combined with other data in the literature, yields an upper bound estimate for the number of lipid molecules required to bind each individual molecule of S. Our results demonstrate that fluorescence correlation spectroscopy provides a powerful tool for the quantitative characterization of S-lipid interactions.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
-Synuclein (S) is the major fibrillar component of the Lewy body deposits in the substantia nigra that are characteristic of Parkinson's disease (PD) (1,2). Though the precise role of S in the degeneration of dopaminergic neurons in PD remains unclear, the identification of two autosomal dominant mutations in S that result in early onset PD established that it is critical to the progression of the disease state (3,4). More recently, a third S mutation, as well as triplication of the S gene, have also been linked to disease (5–8). -Synuclein is a small, 14.5 kDa, soluble protein that is found abundantly in presynaptic nerve terminals in various regions of the brain (9). The native function of S is poorly understood, though it has been shown to affect dopamine synthesis (10–12) and transport (13), and it is thought that S may play a role both in maintaining neuronal plasticity (14) and in the regulation of synaptic vesicle recycling (15). Despite these proposed functions, S is only loosely associated with either the synaptic membrane or synaptic vesicles, often behaving as a soluble protein in brain tissue extracts (1,14).

Although S is intrinsically unstructured in solution (16), it undergoes a conformational change to a predominantly -helical structure upon association with lipid membranes (17,18). Sequence similarities between the N-terminus of S and the lipid-binding domains of some classes of apolipoproteins led to the proposal that the N-terminus mediates S interactions with membranes (16,18). This has been confirmed by NMR and EPR studies of S bound to detergent micelles and small unilamellar vesicles (SUVs) (19–21), which show that the N-terminus forms an amphipathic helix (with an unusual 11/3 periodicity) upon association with micelles or vesicles, while the C-terminus remains largely free and unstructured (20,22). It has been suggested that the interaction of S with membranes plays a role not only in the normal function of the protein but also in the pathology of PD. Specifically, studies have identified an oligomeric form of S that is toxic to cells in culture (23), and shown that S is capable of forming stable, porelike oligomers in vitro (24), leading to the postulation that S oligomers may be responsible for permeabilization of cell membranes in the disease state (25–27).

There have been a number of in vitro studies aimed at elucidating the nature of the interaction of S with membranes, using model synthetic lipid vesicles as well as extracted brain tissue lipids. However, much of the existing literature regarding these interactions is contradictory. The majority of studies indicate that S binds preferentially to acidic phospholipids (17,18,28,29), though different affinities for specific phospholipid headgroups have been reported. In contrast, binding of S to zwitterionic large unilamellar vesicles (LUVs) (30) and zwitterionic gel-phase SUVs (31) has also been described. Adding to the debate is the question as to whether or not association with lipids may inhibit (29,32) or promote (29,33–35) fibrillar S formation.

The lack of general consensus regarding the nature of S-lipid interactions may in part be due to experimental factors, including the long timescales needed for the commonly used techniques of gel filtration and gel electrophoresis, or the technical constraints of thin layer chromatography (17). Fluorescence correlation spectroscopy (FCS) allows for rapid, equilibrium characterization of solutions of fluorescently-labeled molecules. The potential for FCS as a powerful technique for studying the binding of peptides and proteins to LUVs has been demonstrated by a handful of studies (36–38). The work of Rusu et al. (37) studied the binding of a model peptide to LUVs, to allow for comparison between the binding parameters determined by FCS with those measured by other techniques. Using a well-characterized peptide-lipid system allowed them to evaluate the efficacy of FCS as a means of quantifying protein and peptide binding to phospholipid vesicles.

Here we report our studies of fluorescently-labeled S binding to LUVs using FCS. We characterize the effects of vesicle composition and size, as well as the presence of monovalent and divalent ions, on the binding of S to vesicles. In particular, we screen for affinity of S for specific lipid headgroups to investigate the role of charge and packing density in determining binding.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Protein expression and labeling
A plasmid containing a construct for wild-type S (S-wt) under control of the T7 promoter was generously provided by Dr. Peter Lansbury (Harvard Medical School). A serine-to-cysteine mutation was introduced at position 9 to allow for specific labeling of the protein (S-S9C) using a QuikChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA). Both S-wt and S-S9C were expressed and purified as described previously (19). The protein S-S9C was labeled with AlexaFluor 488 C₅ maleimide dye (Molecular Probes, Invitrogen, Carlsbad, CA) generally according to the protocol provided by the manufacturer. Alexa 488 was chosen for the fluorophore based on its photophysical properties, including photostability, high quantum yield upon conjugation to S, and large two-photon cross section (as characterized by the authors and colleagues). Briefly, the S-S9C protein was dissolved in pH 6.4 phosphate-buffered saline and a 25 mM stock solution of the dye was prepared in dimethylsulfoxide and added to the protein solution to a final molar ratio of 5:1 dye/protein. The protein-dye solution was incubated in the dark at room temperature for a minimum of 2 h and unreacted free dye was removed using Micro Bio-Spin 6 Chromatography columns (Bio-Rad, Hercules, CA). The labeled protein, S-S9C-AL488, was passed through an additional spin column just before measurement to remove any residual free dye. The final protein concentrations and labeling efficiency were determined by absorption.

Preparation of vesicles
All vesicle components—1-palmitoyl-2-oleoyl-phosphatidylcholine (POPC), 1-palmitoyl-2-oleoyl-phosphatidylserine (POPS), 1-palmitoyl-2-oleoyl-phosphatidic acid (POPA), 1-palmitoyl-2-oleoyl-phosphoethanolamine (POPE), cholesterol, egg sphingomyelin, 1,2-dioleoyl-phosphocholine (DOPC), and 1,2-dioleoyl-phosphoethanolamine lissamine rhodamine (DOPE-rhod)—were purchased in lyophilized form from Avanti Polar Lipids (Birmingham, AL). Stock solutions of 10 mg/ml were prepared in chloroform and stored in sealed glass vials at –20°C. Stock solutions were assayed to determine actual concentrations before preparation of vesicles (39,40). Aliquots of the appropriate amounts of the stock solutions were mixed in clean glass vials and the chloroform was removed under a gentle stream of nitrogen. The lipids were placed under vacuum for at least 4 h to ensure complete removal of residual chloroform and then rehydrated in 50 mM sodium phosphate buffer, pH 7.4, to a final concentration of 2.5 or 5 mM lipid. Large unilamellar vesicles (LUVs) were prepared by repetitive extrusion through two layers of polycarbonate films in a handheld extruder (Avestin, Ottawa, Canada) (41). Vesicles were used within 2 days of preparation but were found to be structurally stable for at least two weeks when stored at 4°C. For vesicles composed of more than one type of lipid, all percentages of lipids are mole percentages. Unless specifically noted, LUVs refers to vesicles formed using 100 nm polycarbonate films.

Fluorescent vesicles were prepared by the addition of a small amount (1:20,000 molar ratio labeled-lipid/total lipid) of labeled lipid, DOPE-rhod, to the stock solution.

Preparation of samples
Stock solutions of 10 µM protein were prepared of S-S9C-AL488 and S-wt. The stock LUV solutions were first diluted to the desired concentration in 50 mM sodium phosphate buffer, pH 7.4, then S, S9C-AL488 or wt, was added. The samples were mixed by gently pipeting the solutions several times and then immediately transferred to sample chambers for measurements.

Measurements were made in eight-chambered Lab-Tek cover-glass slides (Nunc, Rochester, NY). The chambers were treated by plasma cleaning and then incubated with polylysine-conjugated polyethylene glycol to prevent adhesion of S to the surfaces (42). The chambers were then incubated with low concentrations of aqueous POPC solutions. The chambers were rinsed thoroughly with HPLC H₂O before placing the sample in the chambers. Chambers that were not used immediately were filled with HPLC water, which was removed just before filling chambers with sample. There was no evidence of either vesicles or S adsorbing to the chamber surfaces.

Principles of fluorescence correlation spectroscopy
Although FCS was invented more than 30 years ago (43–45), relatively recent improvements in computation, optical, and detection technologies (46) have made it a versatile technique that is widely used for a vast array of biochemical and biophysical characterization. The measured parameter in FCS is the spectrum of spontaneous fluctuations in fluorescence intensity caused by the changes in the emission of fluorescently-labeled molecules in the optically defined focal volume due to diffusion or chemical kinetics (Fig. 1 A). The concentration and the diffusion time of the diffusing species can be calculated from the autocorrelation function of the fluctuations. The diffusion time-scales linearly with the hydrodynamic radius of the diffusing species.

View larger version (24K): [in this window] [in a new window]   FIGURE 1  (A) Cartoon depicting the optically defined focal volume of an FCS instrument. Fluctuations in the intensity of the fluorescence signal result from the diffusion of the fluorescent molecules—shown here as free protein and LUV-bound protein—in and out of the focal volume. (B) Autocorrelation curves of S only (black) and LUV only (light gray) are fit to Eq. 3 (fits in red); the autocorrelation curve of a mixture of free S and LUV-bound S (dark gray) is fit to Eq. 4 (fit in blue) to yield 60% LUV-bound S and 40% free S. The free S has a diffusion time of 250 µs and the 120 nm LUVs have a diffusion time of 3.8 ms, 15 times longer than that of the free protein. This difference in diffusion times allows for separation of the contribution of each type of molecule in a mixed solution to the autocorrelation curve.

(1)

where F() is the fluorescence obtained from the volume at delay time and F() = F() – F(). The autocorrelation curves are fit with the standard formula for a three-dimensional multicomponent fit (47),

(2)

where N_i is the number of molecules of species i, _Di the characteristic diffusion time of species i, and is the ratio of the axial/radial dimensions of the observation volume. For a single diffusing species, i = 1, the equation is simply

(3)

For a two-component system, such as that consisting of free S and LUV-bound S, the correlation function becomes

(4)

where f is the fraction of the signal that comes from free S, _S is the characteristic diffusion time of free S, and _LUV is the characteristic diffusion time of the vesicles. Because of the considerable difference in size between free S and LUV-bound S (Fig. 1 B), the contribution of each species to the correlation function can easily be distinguished by fitting the experimental data to Eq. 4.

Two-photon excitation fluorescence correlation spectroscopy
FCS measurements with two-photon excitation were made on an instrument assembled in our lab as described previously (48). Briefly, a pulsed-IR titanium-sapphire laser (Tsunami, Spectra-Physics, Palo Alto, CA) with 80 MHz, 100-fs pulse-width was used as the excitation source, = 895 nm. The laser was directed into a 1.2 NA, 63x objective (Zeiss, Jena, Germany) mounted in an inverted microscope (Axiovert 35, Zeiss). The excitation power at the objective was between 1 and 4 mW, low enough to ensure that no photobleaching of the fluorophores occurred during their residence in the focal volume. Fluorescence emission was collected through the objective, separated from the excitation beam with a 670-nm longpass dichroic filter (Chroma, Rockingham, VT), and focused onto the face of a GaAsP photomultiplier tube (model No. H-7421, Hamamatsu, Somerville, NJ). The signal from the photomultiplier was autocorrelated by a digital correlator (ALV-6000, ALV, Langen, Germany). A typical data set consisted of 10 autocorrelation curves of 10 s each. These 10 curves were averaged and the average curve was fit according to Eqs. 3 or 4 using the standard deviation as a weighting factor.

Rhodamine green, a dye whose diffusion parameters have been well characterized by FCS (summarized in (49)), was measured as a calibration standard to determine , in Eqs. 2–4 above. Fitting autocorrelation curves of rhodamine green gives rise to = 3.4. This value was then fixed for all subsequent fitting. Steady-state fluorescence (data not shown) was used to determine a quantum yield ratio of 1 for free S-S9C-AL488 relative to that of LUV-bound S-S9C-AL488, and thus there was no need for correction of the fractions of free and LUV-bound protein as determined by Eq. 4.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
FCS measurements of free S and dye-labeled vesicles in separate experiments were used to determine their individual diffusion times. Fig. 1 B shows the autocorrelation curves of S-S9C-AL488 and DOPE-rhod-labeled vesicles, as well as a curve with both free and LUV-bound S-S9C-AL488 components. The curves for free S and the LUVs were fit to Eq. 3, resulting in diffusion times of 250 µs for the protein and 3.8 ms for the vesicles. Both curves were well fit to single-component equations indicating monodisperse solutions. The corresponding hydrodynamic radius of 60 nm, or diameter of 120 nm, calculated for the vesicles is within the reported range of diameters for vesicles formed by extrusion through 100 nm pores. We saw some small variations in _LUV, depending upon the composition of the vesicles and the amount of protein bound; thus for the fitting of autocorrelation curves of solutions containing both free S and LUV-bound S using Eq. 4, _S was fixed to 250 µS, while _LUV was allowed to float to obtain the best fit for the data.

The placement of the Alexa 488 label was carefully chosen based on NMR-derived models of S-micelle and S-SUV interactions (20) to minimize possible disruption of protein-lipid interactions, and prior NMR studies confirmed that the S9C mutation does not significantly perturb the structure of detergent micelle-bound S (22). Nevertheless, the possibility remained that the dye would affect—either negatively or positively—the binding of S to the vesicles. To control for this, we made three samples containing 100% POPS LUVs with a fixed overall concentration of S, but with varying relative amounts of S-S9C-AL488 and S-wt. The normalized autocorrelation curves (Fig. 2) are virtually indistinguishable, and fitting of the curves yields a fraction LUV-bound protein of 0.8 for all three cases (Table 1). Therefore, the wild-type and dye-labeled proteins appear to bind to the vesicles equally well. We also measured solutions of free Alexa 488 dye mixed with high (mM) concentrations of POPC and POPS LUVs and found no evidence of interactions of the free dye with the LUVs, as seen by autocorrelation curves that fit to a single diffusing component identical to that of free Alexa 488 dye in buffer (data not shown).

View larger version (15K): [in this window] [in a new window]   FIGURE 2  Autocorrelation curves of solutions with three different ratios of S-S9C-AL488 and S-wt for the same total concentration S. (A) 200 nM S-S9C-AL488 and 200 nM S-wt. (B) 100 nM S-S9C-AL488 and 300 nM S-wt. (C) 50 nM S-S9C-AL488 and 350 nM S-wt. The fits to Eq. 4 of the three curves are shown in red. The curves are normalized to a G(0) = 1 to allow for comparison of the data.

View this table: [in this window] [in a new window]   TABLE 1  Fraction of bound S for each of the S combinations shown in Fig. 2

View larger version (16K): [in this window] [in a new window]   FIGURE 3  (A) Autocorrelation curves of solutions with increasing concentration of S with 25 µM 100% POPS LUVs. The concentrations of S shown are: 50 nM (squares), 200 nM (circles), 400 nM (triangles), 500 nM (inverted triangles), 750 nM (diamonds), 1 µM (cross-hairs), and 1.5 µM (pentagons) (the last three concentrations are difficult to differentiate on this plot). (Inset) The curves in panel A were fit with Eq. 4 to determine the fraction of free and bound S. Increasing the S concentration beyond saturation results in a decrease in the fraction of bound protein. (B). The data in Fig. 4 A, inset, is replotted as LUV-bound S as a function of total S. Above 500 nM total S, the total amount of bound protein begins to decrease instead of remaining constant as may be expected.

View larger version (17K): [in this window] [in a new window]   FIGURE 4  (A) Fraction S bound as a function of lipid concentration for five different vesicle compositions. The concentration of S was maintained at 200 nM, whereas the concentration of lipid was increased. The fraction S bound was calculated from fitting Eq. 4 to the autocorrelation curves for each concentration of lipid: 100% POPS (circles), 70% POPS/30% POPC (triangles), 50% POPS/50% POPC (squares), 30% POPS/70% POPC (diamonds), and 100% POPC (inverted triangles). Also shown is the binding curve for 60 nm diameter 100% POPS vesicles (solid circles). The fits shown in shaded representation are to the Hill equation, which yields a Hill coefficient 2.4 and a KD for each composition of lipids. (B) The log of KD as determined from the fits of the Hill equation to binding curves shown in panel A as a function of molar percentage POPS.

To investigate the role of the negatively charged lipids in binding, LUVs composed of various ratios of zwitterionic POPC and acidic POPS were prepared. For each LUV composition, FCS data were collected at increasing LUV concentrations, while maintaining a constant S concentration (200 nM), and the autocorrelation curves were fit to Eq. 4 to yield the relative amounts of LUV-bound and free S. The fraction of LUV-bound protein calculated from FCS data was plotted as a function of the accessible lipid concentration, which is estimated to be half of the total lipid concentration, because it is assumed that S cannot access the lipid of the interior of the vesicles. The data (Fig. 4 A) show that S binds much more readily to LUVs containing a higher percentage of negatively charged lipids, in agreement with the majority of other in vitro lipid-binding studies. The binding curves are not hyperbolic and could not be well fit based on a simple partition equilibrium model. Data points collected at low lipid concentrations may be affected by the high protein/lipid ratio, as observed above, and may be partially responsible for the apparent lag in protein binding. This effect has also been noted in a recent study using calorimetry to monitor the binding of S to SUVs (31). This same study also noted that electrostatic interactions between the protein and the lipid vesicles lead to deviations from a simple partition equilibrium. Both of these effects might be expected to be mitigated at lower protein concentrations, but binding curves from data collected at a 10-fold lower concentration of S (20 nM) showed the same shape and were also poorly fit by the partition equilibrium model (data not shown).

Reasonable fits to the binding curves could be obtained using the Hill equation and yielded apparent disassociation constants, K_D, for the various lipid compositions (Fig. 4 B). Similar K_D values were obtained either from best fits to a simple partition equilibrium or from 50% binding values extracted from best fits to a dose-response model. The apparently exponential relationship between K_D and the fraction of acidic lipid is predicted by Boltzmann and electrostatic theory (50). The value of the Hill coefficient obtained from the binding curves was 2.4, but it is not clear how to interpret this value.

To further examine the role of charge-charge interactions in the binding of S to lipids, increasing amounts of NaCl were added to a fixed amount of LUV and S-S9C-AL488. As the concentration of NaCl was increased from 0 to 400 mM, the fraction of LUV-bound S decreased (Fig. 5), showing that high salt concentrations inhibit lipid binding by S. We also investigated the ability of the divalent ion calcium to disrupt S-vesicle interactions. Due to the mutual insolubility of phosphate and calcium, the LUVs were made in 20 mM MOPS buffer, pH 7.4, rather than the standard phosphate buffer. Amounts of up to 20 mM CaCl₂, well above the physiologically relevant concentrations, had no effect on the binding of S to the vesicles (data not shown).

View larger version (8K): [in this window] [in a new window]   FIGURE 5  Fraction S bound to 100% POPS LUVs as a function of NaCl concentration, with 200 nM S. Increasing concentration of NaCl results in screening of the favorable electrostatic interactions between S and the vesicles, and thus a decreasing fraction of bound protein.

View larger version (8K): [in this window] [in a new window]   FIGURE 6  Comparison of fraction S bound for three different combinations of lipids, with 200 nM S. Each combination consists of a zwitterionic lipid and an acidic lipid in a 1:1 molar ratio: POPA/POPC (triangles), POPS/POPE (squares), and POPS/POPC (circles). The protein shows much stronger binding to the POPA/POPC and slightly stronger binding to POPS/POPE when compared to the POPS/POPC LUVs.

Lastly, synaptic vesicles contain a high proportion of sphingomyelin and cholesterol, suggesting a propensity for phase heterogeneity in the membrane, a phenomenon that is thought to underlie the formation of raft domains in cellular membranes. Several recent articles have indicated localization of S to lipid rafts in cells (51) or binding of S with LUVs containing raft-associated components (52,53). Thus, we prepared LUVs corresponding to the canonical raft mixture of 1:1:1 molar ratio DOPC/sphingomyelin/cholesterol, which mimics the composition of the outer leaflet of the plasma membrane (54), as well as 100% DOPC LUVs. In both cases, the vesicles are electrostatically neutral. The fraction of LUV-bound S for both types of LUVs was statistically indistinguishable from that found for 100% POPC vesicles (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Using FCS to monitor S-vesicle interactions
Although several other studies have established that FCS can be a powerful method for analyzing peptide and protein binding to LUVs (36–38), a primary goal of this study was to demonstrate the utility of FCS for quantifying binding of S to lipid vesicles. FCS has several advantages over the techniques that have commonly been used to characterize S-lipid interactions. A series of autocorrelation curves that allows for statistical analysis of the binding behavior can be collected within 2–3 min of first mixing the protein and vesicles together. This is in direct contrast to more traditional methods such as chromatography, where the protein and vesicles may spend several hours traversing a column or gel, or ultracentrifugation, which also requires several hours.

FCS is also applicable over a wide range of fluorophore concentrations, from nM to µM. For most of the experiments presented in this article, a fixed concentration of 200 nM S-S9C-AL488 was used. However, the data shown in Fig. 3 were measured with 25 nM S-S9C-AL488, and binding measurements using as little as 5 nM S-S9C-AL488 were also performed (data not shown). This is between 100 and 1000 times lower protein concentration than is required for other techniques, such as circular dichroism (CD), chromatography, and calorimetry. Furthermore, as demonstrated in Fig. 3, the upper end of the range of protein concentrations can be extended almost indefinitely by maintaining the amount of fluorescently-labeled protein within the desired nM–µM range and adding unlabeled protein to achieve the desired final protein concentration, so long as the binding properties of the labeled and unlabeled proteins are the same (as we show to be the case here in Fig. 2).

CD is used to detect the binding of S to lipids via the appearance of spectral lines associated with -helical secondary structure (18), but quantitative interpretation of the results may not be straightforward. For example, a study by Jo et al. (17) found that although thin layer chromatography indicated binding of S to POPE, POPE did not induce -helical structure in S as detected by CD unless an acidic lipid was also present. In another example, Narayanan and Scarlata (30) reported that interaction of S with 100% POPS LUVs did not induce significant amounts of -helical structure in S according to their CD data. However, at S concentrations above the saturating concentration, the primary component of the CD spectra comes from random-coil free S with only a minor contribution arising from the small population of -helical LUV-bound S. In contrast, interpretation of binding data from FCS is straightforward, because the signals from the free S and LUV-bound S are both detectable and separable, as well as being independent of secondary structure formation.

Lipid-binding properties of S
Our data reveal a marked preference by S for binding to acidic phospholipids (Fig. 4), consistent with a number of published studies. We also confirm that electrostatic interactions are important for S-lipid binding, because physiological extracellular NaCl concentrations result in a considerable decrease, although not complete elimination, of the amount of S bound to LUVs (Fig. 5). However, the data indicates a significantly higher affinity for vesicles composed of some portion of POPA over those containing an equivalent amount of POPS, as well as a slightly higher affinity for POPE over POPC (Fig. 6). Because both POPA and POPS are negatively charged and both POPE and POPC are zwitterionic, it appears that S is sensitive to lipid headgroup properties beyond total charge. At pH 7.4, some percentage of POPA headgroups will have a valence of –2 (average relative charge –1.2 for POPA versus –1 for POPS at pH 7.4 (55)), which may partially account for the greater affinity of S for POPA. In addition, POPA is lacking in the bulky amino-serine group of POPS, and might therefore be able to pack more closely together in a lipid bilayer, producing a higher charge density. Alternatively, the bulkier POPS headgroup may interfere sterically with S binding. Several other studies have noted a preference for binding to POPA over POPS (18,28,56). This preference may be biologically relevant, because S has been shown to regulate the activity of phospholipase D, an enzyme that produces PA from PC, and furthermore PA is a regulator of synaptic vesicle biogenesis, a process that may be affected by S. However, it has also been pointed out that PS comprises 7–12% of the total lipid in synaptic vesicles where S localizes, whereas PA is only 0–2% (18).

Our data provide a direct quantitation of the binding of S to lipid vesicles. Fig. 3 B suggests that for POPS, the maximum achievable molar ratio of bound protein to total lipid is 1:85, corresponding to a weight ratio of bound protein to total lipid of 1:5. The POPS data in Fig. 4 A also show that complete protein binding can occur at a protein/total lipid molar ratio of 1:100. For 1:1 POPA/POPC vesicles, Fig. 6 shows that complete binding can occur at a protein/lipid molar ratio of 1:250. For 1:1 POPS/POPC, however, Fig. 4 A indicates that complete protein binding occurs at protein/lipid molar ratio of 1:2000, corresponding to a weight ratio of bound protein to total lipid of 1:100. However, it is important to note that this number merely provides an upper bound on the number of lipid molecules required to bind each S molecule, since, in these experiments, the protein concentration is held fixed and might not be high enough to saturate every binding site. It is instructive to compare these results with other published work. An early study of S lipid binding showed, using ultracentrifugation, that for 1:1 POPA/POPC vesicles, a 1:20 protein/lipid weight ratio (corresponding to a molar ratio of 1:370) sufficed for complete protein binding, but the same ratio of protein to 1:1 POPS/POPC vesicles was insufficient for complete binding (18). These results are in good agreement with our own. Another early study reported that for 1:1 POPS/POPC vesicles, a protein/lipid weight ratio of 1:10 (corresponding to a molar ratio of 1:185) was sufficient for complete protein binding as determined by CD data (17). A similar result was obtained for 1:1 POPG/POPC vesicles by another group (31). These studies were performed at higher protein concentrations (10 µM) than our own (200 nM), and this may account for the lower observed number of lipid molecules per bound protein molecule. In total, our results combined with those of others suggest that each molecule of S can bind to a lipid bilayer patch composed of 85 acidic lipid molecules. Dilution of acidic lipids with neutral lipids significantly reduces the affinity of S for the lipid surface but appears unlikely to alter the number of lipids bound per S molecule.

These conclusions can be evaluated in the context of our knowledge of the structure of lipid-bound S. The protein binds to lipid surfaces through a highly helical conformation adopted by the N-terminal 100 residues, while the acidic C-terminal tail of the protein remains free. A 100-residue helix would be 15-nm long and 1-nm wide, and would therefore cover a surface area of 15 nm². In comparison, an individual lipid molecule in a bilayer occupies 0.5 nm² of surface area. Therefore, a minimum of 30 surface-accessible lipid molecules, or a 60-molecule bilayer patch, would be required to bind a single S molecule. Our data indicate that 1.5 times this minimum number of lipid molecules is actually required per S molecule and that trying to force a higher S/lipid ratio may result in destabilization of the lipid bilayer. Interestingly, S oligomers have been reported to permeabilize phospholipid vesicles (26), and this has been proposed as a general mechanism of toxicity for amyloid-forming proteins (57). It is possible that in the absence of sufficient lipid surface area, S oligomerization may be facilitated (this has been referred to as the parking-problem) leading to lipid bilayer disruption. Such a mechanism might also provide a partial explanation for the cooperativity observed in the binding curves of Fig. 4. However, we do not have any direct evidence for such a mechanism at present.

Finally, our preliminary examination of the interaction of S with vesicles containing cholesterol and sphingomyelin shows that S exhibits the same binding affinity for vesicles composed of a raft mixture of lipids as for 100% DOPC vesicles, suggesting that it is the presence of negatively charged components—either protein/peptide or phospholipids—in the raft domain of the membrane, rather than any physical property of the lipid phase itself that influences the binding of the protein. Nuscher et al. (31) characterized the binding of S to neutral gel phase SUVs and attributed the binding affinity both to the high curvature of the SUVs and to the formation of lipid phase domains within the vesicles. It is therefore possible that further binding studies using gel phase or liquid-ordered phase SUVs, instead of the LUVs examined here, would reveal a higher propensity for S binding. Because of the possible physiological relevance of the raft membrane composition, this remains an avenue for further exploration. It is possible that small changes in the proportions of the lipid components may result in dramatic differences in membrane heterogeneity and thus drastically affect S binding. FCS provides a rapid, quantitative means of characterizing these interactions.

	ACKNOWLEDGEMENTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
This work was supported in part by a CMBSTD grant of the W. M. Keck Foundation (to E.R.), National Institute of Biomedical Imaging and BioEngineering, National Institutes of Health grant No. 9 P41 EB001976 (to W.W.W.), National Institute on Aging, National Institutes of Health grant No. 1 R21 AG26650 (to W.W.W.), National Institute on Aging, National Institutes of Health grant No. AG19391 (to D.E.), and by a gift from Herbert and Ann Siegel (to D.E.).

Submitted on December 8, 2005; accepted for publication March 13, 2006.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
1. Ueda, K., H. Fukushima, E. Masliah, Y. Xia, A. Iwai, M. Yoshimoto, D. A. C. Otero, J. Kondo, Y. Ihara, and T. Saitoh. 1993. Molecular cloning of CDNA—encoding an unrecognized component of amyloid in Alzheimer disease. Proc. Natl. Acad. Sci. USA. 90:11282–11286.[Abstract/Free Full Text]

2. Goedert, M. 2001. Alpha-synuclein and neurodegenerative diseases. Nat. Rev. Neurosci. 2:492–501.[CrossRef][Medline]

3. Kruger, R., W. Kuhn, T. Muller, D. Woitalla, M. Graeber, S. Kosel, H. Przuntek, J. T. Epplen, L. Schols, and O. Riess. 1998. Ala₃₀Pro mutation in the gene encoding -synuclein in Parkinson's disease. Nat. Genet. 18:106–108.[CrossRef][Medline]

4. Polymeropoulos, M. H., C. Lavedan, E. Leroy, S. E. Ide, A. Dehejia, A. Dutra, B. Pike, H. Root, J. Rubenstein, R. Boyer, E. S. Stenroos, S. Chandrasekharappa, A. Athanassiadou, T. Papapetropoulos, W. G. Johnson, A. M. Lazzarini, R. C. Duvoisin, G. DiIorio, L. I. Golbe, and R. L. Nussbaum. 1997. Mutation in the -synuclein gene identified in families with Parkinson's disease. Science. 276:2045–2047.[Abstract/Free Full Text]

5. Zarranz, J. J., J. Alegre, J. C. Gomez-Esteban, E. Lezcano, R. Ros, I. Ampuero, L. Vidal, J. Hoenicka, O. Rodriguez, B. Atares, V. Llorens, E. G. Tortosa, T. del Ser, D. G. Munoz, and J. G. de Yebenes. 2004. The new mutation, E46K, of -synuclein causes Parkinson and Lewy body dementia. Ann. Neurol. 55:164–173.[CrossRef][Medline]

6. Bradbury, J. 2003. Alpha-synuclein gene triplication discovered in Parkinson's disease. Lancet Neurol. 2:715.[Medline]

7. Singleton, A. B., M. Farrer, J. Johnson, A. Singleton, S. Hague, J. Kachergus, M. Hulihan, T. Peuralinna, A. Dutra, R. Nussbaum, S. Lincoln, A. Crawley, M. Hanson, D. Maraganore, C. Adler, M. R. Cookson, M. Muenter, M. Baptista, D. Miller, J. Blancato, J. Hardy, and K. Gwinn-Hardy. 2003. Alpha-synuclein locus triplication causes Parkinson's disease. Science. 302:841.[Free Full Text]

8. Farrer, M., J. Kachergus, L. Forno, S. Lincoln, D. S. Wang, M. Hulihan, D. Maraganore, K. Gwinn-Hardy, Z. Wszolek, D. Dickson, and J. W. Langston. 2004. Comparison of kindreds with Parkinsonism and -synuclein genomic multiplications. Ann. Neurol. 55:174–179.[CrossRef][Medline]

9. Clayton, D. F., and J. M. George. 1998. The synucleins: a family of proteins involved in synaptic function, plasticity, neurodegeneration and disease. Trends Neurosci. 21:249–254.[CrossRef][Medline]

10. Lee, F. J. S., F. Liu, Z. B. Pristupa, and H. B. Niznik. 2001. Direct binding and functional coupling of -synuclein to the dopamine transporters accelerate dopamine-induced apoptosis. FASEB J. 15:916–926.[Abstract/Free Full Text]

11. Wersinger, C., D. Prou, P. Vernier, and A. Sidhu. 2003. Modulation of dopamine transporter function by -synuclein is altered by impairment of cell adhesion and by induction of oxidative stress. FASEB J. 17:2151–2153.[Abstract/Free Full Text]

12. Wersinger, C., and A. Sidhu. 2003. Attenuation of dopamine transporter activity by -synuclein. Neurosci. Lett. 340:189–192.[CrossRef][Medline]

13. Lotharius, J., S. Barg, P. Wiekop, C. Lundberg, H. K. Raymon, and P. Brundin. 2002. Effect of mutant -synuclein on dopamine homeostasis in a new human mesencephalic cell line. J. Biol. Chem. 277:38884–38894.[Abstract/Free Full Text]

14. George, J. M., H. Jin, W. S. Woods, and D. F. Clayton. 1995. Characterization of a novel protein regulated during the critical period for song learning in the Zebra Finch. Neuron. 15:361–372.[CrossRef][Medline]

15. Lotharius, J., and P. Brundin. 2002. Impaired dopamine storage resulting from -synuclein mutations may contribute to the pathogenesis of Parkinson's disease. Hum. Mol. Genet. 11:2395–2407.[Abstract/Free Full Text]

16. Weinreb, P. H., W. G. Zhen, A. W. Poon, K. A. Conway, and P. T. Lansbury. 1996. NACP, a protein implicated in Alzheimer's disease and learning, is natively unfolded. Biochemistry. 35:13709–13715.[CrossRef][Medline]

17. Jo, E. J., J. McLaurin, C. M. Yip, P. St George-Hyslop, and P. E. Fraser. 2000. Alpha-synuclein membrane interactions and lipid specificity. J. Biol. Chem. 275:34328–34334.[Abstract/Free Full Text]

18. Davidson, W. S., A. Jonas, D. F. Clayton, and J. M. George. 1998. Stabilization of -synuclein secondary structure upon binding to synthetic membranes. J. Biol. Chem. 273:9443–9449.[Abstract/Free Full Text]

19. Eliezer, D., E. Kutluay, R. Bussell, and G. Browne. 2001. Conformational properties of -synuclein in its free and lipid-associated states. J. Mol. Biol. 307:1061–1073.[CrossRef][Medline]

20. Bussell, R., and D. Eliezer. 2003. A structural and functional role for 11-mer repeats in -synuclein and other exchangeable lipid binding proteins. J. Mol. Biol. 329:763–778.[CrossRef][Medline]

21. Jao, C. C., A. Der-Sarkissian, J. Chen, and R. Langen. 2004. Structure of membrane-bound -synuclein studied by site-directed spin labeling. Proc. Natl. Acad. Sci. USA. 101:8331–8336.[Abstract/Free Full Text]

22. Bussell, R., T. F. Ramlall, and D. Eliezer. 2005. Helix periodicity, topology, and dynamics of membrane-associated -synuclein. Protein Sci. 14:862–872.[Abstract/Free Full Text]

23. Rochet, J. C., T. F. Outeiro, K. A. Conway, T. T. Ding, M. J. Volles, H. A. Lashuel, R. M. Bieganski, S. L. Lindquist, and P. T. Lansbury. 2004. Interactions among -synuclein, dopamine, and biomembranes—some clues for understanding neurodegeneration in Parkinson's disease. J. Mol. Neurosci. 23:23–33.[CrossRef][Medline]

24. Volles, M. J., S. J. Lee, J. C. Rochet, M. D. Shtilerman, T. T. Ding, J. C. Kessler, and P. T. Lansbury. 2001. Vesicle permeabilization by protofibrillar -synuclein: implications for the pathogenesis and treatment of Parkinson's disease. Biochemistry. 40:7812–7819.[CrossRef][Medline]

25. Lashuel, H. A., B. M. Petre, J. Wall, M. Simon, R. J. Nowak, T. Walz, and P. T. Lansbury. 2002. Alpha-synuclein, especially the Parkinson's disease-associated mutants, forms pore-like annular and tubular protofibrils. J. Mol. Biol. 322:1089–1102.[CrossRef][Medline]

26. Lashuel, H. A., D. Hartley, B. M. Petre, T. Walz, and P. T. Lansbury. 2002. Neurodegenerative disease—amyloid pores from pathogenic mutations. Nature. 418:291.

27. Conway, K. A., S. J. Lee, J. C. Rochet, T. T. Ding, R. E. Williamson, and P. T. Lansbury. 2000. Acceleration of oligomerization, not fibrillization, is a shared property of both -synuclein mutations linked to early-onset Parkinson's disease: implications for pathogenesis and therapy. Proc. Natl. Acad. Sci. USA. 97:571–576.[Abstract/Free Full Text]

28. Perrin, R. J., W. S. Woods, D. F. Clayton, and J. M. George. 2000. Interaction of human -synuclein and Parkinson's disease variants with phospholipids—structural analysis using site-directed mutagenesis. J. Biol. Chem. 275:34393–34398.[Abstract/Free Full Text]

29. Zhu, M., J. Li, and A. L. Fink. 2003. The association of -synuclein with membranes affects bilayer structure, stability, and fibril formation. J. Biol. Chem. 278:40186–40197.[Abstract/Free Full Text]

30. Narayanan, V., and S. Scarlata. 2001. Membrane binding and self-association of -synucleins. Biochemistry. 40:9927–9934.[CrossRef][Medline]

31. Nuscher, B., F. Kamp, T. Mehnert, S. Odoy, C. Haass, P. J. Kahle, and K. Beyer. 2004. Alpha-synuclein has a high affinity for packing defects in a bilayer membrane—a thermodynamics study. J. Biol. Chem. 279:21966–21975.[Abstract/Free Full Text]

32. Zhu, M., and A. L. Fink. 2003. Lipid binding inhibits -synuclein fibril formation. J. Biol. Chem. 278:16873–16877.[Abstract/Free Full Text]

33. Zhao, H. X., E. K. J. Tuominen, and P. K. J. Kinnunen. 2004. Formation of amyloid fibers triggered by phosphatidylserine-containing membranes. Biochemistry. 43:10302–10307.[CrossRef][Medline]

34. Cole, N. B., D. D. Murphy, T. Grider, S. Rueter, D. Brasaemle, and R. L. Nussbaum. 2002. Lipid droplet binding and oligomerization properties of the Parkinson's disease protein -synuclein. J. Biol. Chem. 277:6344–6352.[Abstract/Free Full Text]

35. Lee, H. J., C. Choi, and S. J. Lee. 2002. Membrane-bound -synuclein has a high aggregation propensity and the ability to seed the aggregation of the cytosolic form. J. Biol. Chem. 277:671–678.[Abstract/Free Full Text]

36. Dorn, I. T., K. R. Neumaier, and R. Tampe. 1998. Molecular recognition of histidine-tagged molecules by metal-chelating lipids monitored by fluorescence energy transfer and correlation spectroscopy. J. Am. Chem. Soc. 120:2753–2763.[CrossRef]

37. Rusu, L., A. Gambhir, S. McLaughlin, and J. Radler. 2004. Fluorescence correlation spectroscopy studies of peptide and protein binding to phospholipid vesicles. Biophys. J. 87:1044–1053.[Abstract/Free Full Text]

38. Takakuwa, Y., C. G. Pack, X. L. An, S. Manno, E. Ito, and M. Kinjo. 1999. Fluorescence correlation spectroscopy analysis of the hydrophobic interactions of protein 4.1 with phosphatidyl serine liposomes. Biophys. Chem. 82:149–155.[CrossRef][Medline]

39. Fiske, C. H., and Y. Subbarow. 1925. The colorimetric determination of phosphorus. J. Biol. Chem. 66:375–400.[Free Full Text]

40. Chen, P. S., T. Y. Toribara, and H. Warner. 1956. Microdetermination of phosphorus. Anal. Chem. 28:1756–1758.

41. Macdonald, R. C., R. I. Macdonald, B. P. M. Menco, K. Takeshita, N. K. Subbarao, and L. R. Hu. 1991. Small-volume extrusion apparatus for preparation of large unilamellar vesicles. Biochim. Biophys. Acta. 1061:297–303.[Medline]

42. Kenausis, G. L., J. Voros, D. L. Elbert, N. P. Huang, R. Hofer, L. Ruiz-Taylor, M. Textor, J. A. Hubbell, and N. D. Spencer. 2000. Poly(L-lysine)-g-poly(ethylene glycol) layers on metal oxide surfaces: attachment mechanism and effects of polymer architecture on resistance to protein adsorption. J. Phys. Chem. B. 104:3298–3309.

43. Magde, D., W. W. Webb, and E. Elson. 1972. Thermodynamic fluctuations in a reacting system—measurement by fluorescence correlation spectroscopy. Phys. Rev. Lett. 29:705–708.[CrossRef]

44. Magde, D., E. L. Elson, and W. W. Webb. 1974. Fluorescence correlation spectroscopy. 2. Experimental realization. Biopolymers. 13:29–61.[CrossRef][Medline]

45. Magde, D., and E. L. Elson. 1978. Fluorescence correlation spectroscopy. 3. Uniform translation and laminar-flow. Biopolymers. 17:361–376.[CrossRef]

46. Eigen, M., and R. Rigler. 1994. Sorting single molecules—application to diagnostics and evolutionary biotechnology. Proc. Natl. Acad. Sci. USA. 91:5740–5747.[Abstract/Free Full Text]

47. Thompson, N. L. 1991. Fluorescence correlation spectroscopy. In Topics in Fluorescence Microscopy. J. Lakowicz, editor. Plenum Press, New York. 337–378

48. Larson, D. R., Y. M. Ma, V. M. Vogt, and W. W. Webb. 2003. Direct measurement of Gag-Gag interaction during retrovirus assembly with FRET and fluorescence correlation spectroscopy. J. Cell Biol. 162:1233–1244.[Abstract/Free Full Text]

49. Hess, S. T., and W. W. Webb. 2002. Focal volume optics and experimental artifacts in confocal fluorescence correlation spectroscopy. Biophys. J. 83:2300–2317.[Abstract/Free Full Text]

50. Arbuzova, A., L. B. Wang, J. Y. Wang, G. Hangyas-Mihalyne, D. Murray, B. Honig, and S. McLaughlin. 2000. Membrane binding of peptides containing both basic and aromatic residues. Experimental studies with peptides corresponding to the scaffolding region of caveolin and the effector region of MARCKS. Biochemistry. 39:10330–10339.[CrossRef][Medline]

51. Fortin, D. L., M. D. Troyer, K. Nakamura, S. Kubo, M. D. Anthony, and R. H. Edwards. 2004. Lipid rafts mediate the synaptic localization of -synuclein. J. Neurosci. 24:6715–6723.[Abstract/Free Full Text]

52. Kubo, S., V. M. Nemani, R. J. Chalkey, M. D. Anthony, N. Hattori, Y. Mizuno, R. H. Edwards, and D. L. Fortin. 2005. A combinatorial code for the interaction of -synuclein with membranes. J. Biol. Chem. 280:31664–31672.[Abstract/Free Full Text]

53. Narayanan, V., Y. J. Guo, and S. Scarlata. 2005. Fluorescence studies suggest a role for -synuclein in the phosphatidylinositol lipid signaling pathway. Biochemistry. 44:462–470.[CrossRef][Medline]

54. Veatch, S. L., and S. L. Keller. 2002. Organization in lipid membranes containing cholesterol. Phys. Rev. Lett. 89:268101-1–268101-4.

55. Marsh, D. 1990. CRC Handbook of Lipid Bilayers. CRC Press, Boca Raton, FL.

56. Bussell, R., and D. Eliezer. 2004. Effects of Parkinson's disease-linked mutations on the structure of lipid-associated -synuclein. Biochemistry. 43:4810–4818.[CrossRef][Medline]

57. Bucciantini, M., E. Giannoni, F. Chiti, F. Baroni, L. Formigli, J. S. Zurdo, N. Taddei, G. Ramponi, C. M. Dobson, and M. Stefani. 2002. Inherent toxicity of aggregates implies a common mechanism for protein misfolding diseases. Nature. 416:507–511.[CrossRef][Medline]

This article has been cited by other articles:

				B. Wu, Y. Chen, and J. D. Muller Fluorescence Correlation Spectroscopy of Finite-Sized Particles Biophys. J., April 1, 2008; 94(7): 2800 - 2808. [Abstract] [Full Text] [PDF]