Combining Non-canonical Amino Acid Mutagenesis and Native Chemical Ligation for Multiply Modifying Proteins: A case study of α-synuclein post-translational modifications

Abstract

The Parkinson’s disease associated protein α-synuclein ($\text{αS}$) has been found to contain numerous post-translational modifications (PTMs), in both physiological and pathological states. One PTM site of particular interest is serine 87, which is subject to both $O$-linked $\text{β}$-$N$-acetylglucosamine ($\text{gS}$) modification and phosphorylation ($\text{pS}$), with $\text{αS}-{\text{pS}}_{87}$ enriched in Parkinson’s disease. An often-overlooked aspect of these PTMs is their effect on the membrane-binding properties of $\text{αS}$, which are important to its role in regulating neurotransmitter release. Here, we show how one can study these effects by synthesizing $\text{αS}$ constructs containing authentic PTMs and labels for single molecule fluorescence correlation spectroscopy measurements. We synthesize $\text{αS}-{\text{gS}}_{87}$ and $\text{αS}-{\text{pS}}_{87}$ by combining native chemical ligation with genetic code expansion approaches. We introduce the fluorophore by a click reaction with a non-canonical amino acid. Beyond the specific problem of PTM effects on $\text{αS}$, our studies highlight the value of this combination of methods for multiply modifying proteins.

1. Introduction

The protein $\text{α}$-synuclein ($\text{αS}$) is the primary component of Lewy Bodies, the fibrillar aggregates that are the pathological hallmarks of Parkinson’s disease[1]. Although the role of $\text{αS}$ in disease has been widely studied[2], its native function is less understood, with many studies focusing on synaptic vesicle maintenance[3]. It has also been suggested that disruption of $\text{αS}$ interactions with membranes may contribute to the mechanism of toxicity in Parkinson’s disease[4]. While soluble $\text{αS}$ is highly disordered[5], it can adopt a helical conformation upon interaction with membranes[6] and has been observed to associate with membranes in neurons[7]. $\text{αS}$ contains 140 residues, with multiple domains, including an N-terminus with 7 imperfect repeats of a KTKEGV sequence that can form an amphipathic helix[8], followed by the NAC domain responsible for its aggregation[9], and a highly acidic and unstructured C-terminus[6] (Figure 1). Adding to its complexity, $\text{αS}$ is also subject to extensive post-translational modifications (PTMs), in both physiological and pathological contexts[10–12]. We and others have a long-standing interest in biochemically isolating and characterizing the effects of these PTMs with the goal of identifying those PTMs of greatest physiological relevance and which might be targeted therapeutically[11, 13–18]. For example, if a PTM increases the aggregation of $\text{αS}$, inhibition of the enzyme responsible for the installation of the PTM might slow disease progression. Alternatively, if a PTM slows aggregation, one might target the enzyme that removes the modification.

Figure 1.

Conformational states and aggregation pathway of $\text{αS}$. $\text{αS}$ rainbow-colored by sequence number with purple/blue N-terminus, green/yellow NAC, and red C-terminus. Imperfect KTKEGV repeats are underlined. Serine 87 O-GlcNAc ($\text{gS}$) and phosphorylation ($\text{pS}$) indicated by blue square and orange sphere, respectively.

One PTM that fits the second scenario is $O$-linked $\text{β}$-$N$-acetylglucosamine (O-GlcNAc) modification[19, 20]. This type of glycosylation is found intracellularly and is added/removed by O-GlcNAc transferase and O-GlcNAcase enzymes, respectively. The overall levels of O-GlcNAc are decreased in the brains of Alzheimer’s disease patients[21–25]. The Pratt laboratory previously demonstrated that O-GlcNAc also slows the aggregation of $\text{αS}$ [26–28]. These results catalyzed the development of O-GlcNAcase inhibitors[29], and application of several of these compounds in preclinical models of Parkinson’s disease have demonstrated that they slow the formation of aggregates and improve symptoms[30].

Serine 87 is a particularly interesting site of O-GlcNAc modification on $\text{αS}$ for multiple reasons. Mouse and human $\text{αS}$ are very similar in sequence but contain some amino acid changes including a Ser87 to Asn substitution in mice, and these differences can a have significant impact on the aggregation and pathogenicity of $\text{αS}$ fibrils[31]. Additionally, the Pratt laboratory recently found that O-GlcNAc modification at this position forces $\text{αS}$ to form amyloid fibrils with a unique structure that do not induce pathology in mice[32]. Serine 87 can also be phosphorylated by the kinase CKI, and this PTM is enriched in Parkinson’s disease[11, 33]. This sets up an interesting interplay between O-GlcNAcylation and phosphorylation, in which they compete for the same site of modification (Figure 1). Notably, in vitro analysis of $\text{αS}$ bearing phosphate at Serine 87 found that this PTM also inhibits protein aggregation, and this result was supported by cell-based studies with Ser87-to-Glu “phosphomimic” mutations.[11, 34]

An often-overlooked aspect of these PTMs is their effect on the membrane-binding properties of $\text{αS}$, which may detrimentally affect the endogenous functions of $\text{αS}$ [35]. This raises the possibility of unwanted site-effects from the O-GlcNAcase inhibitors. Typically, in vitro study of membrane binding is accomplished using circular dichroism (CD) spectroscopy to measure the transition of $\text{αS}$ from unstructured monomer to an extended $\text{α}$-helix in the presence of lipid vesicles. The Pratt laboratory previously showed that O-GlcNAc at serine 87 does not completely block this transition[27]. However, this type of experiment does not give information on any subtle differences in the affinity of the protein to the lipid vesicles. Fluorescence correlation spectroscopy (FCS) can be used to more precisely measure affinities with respect to lipid composition and membrane curvature between labeled $\text{αS}$ and lipid vesicles[36]. Here, we describe FCS studies using $\text{αS}$ constructs containing authentic PTMs and labels for lipid binding studies. Generating such multiply-modified protein constructs is a non-trivial endeavor requiring the development of new protein synthesis approaches[37, 38].

While PTMs are often studied using mimetics like Ser-to-Glu mutations, these mimetics do not always recapitulate the true effect of the selected modification, as the Petersson and Rhoades laboratories have demonstrated in the case of phosphotyrosine mimicry with Glu[39]. For O-GlcNAc, there is no similar canonical amino acid. One versatile method for generating modifications on proteins is native chemical ligation (NCL)[40] which allows for the ligation of multiple peptide fragments while retaining the native peptide bond. NCL is a particularly apt technique for the introduction of PTMs as well as biophysical probes into $\text{αS}$. Building on pioneering work by Brik and Lashuel[41], the Petersson laboratory has incorporated single atom modifications like thioamides[42], as well as fluorophores[43, 44], using traditional cysteine-mediated NCL and non-cysteine ligations enabled by chemoenzymatic modification of the N-terminus[45, 46]. With the Rhoades laboratory, they have incorporated phosphorylated tyrosine[39] and arginylated glutamate[14] using NCL. The Pratt laboratory has similarly used NCL and chemoselective reactions to incorporate diverse modifications, including those with synthetically challenging monomer units such as O-GlcNAc and ubiquitination.[47–56]

Here, we describe methods for generating site-specifically phosphorylated and glycosylated $\text{αS}$ that also contains a fluorescent probe for observation by FCS. Since $\text{αS}$ contains no native cysteines, artificially introduced cysteines have frequently been used for fluorescent labeling of $\text{αS}$ [12, 38]. However, using Cys also as a handle for NCL can become challenging since one may not want to attach the fluorophore at the same Cys sites used in ligating protein fragments to incorporate the PTM. Dawson and Deniz have published an approach to generating multiply-modified $\text{αS}$ using differential protection of Cys residues in the protein fragments[57], but we felt that a more versatile approach could be achieved by combining NCL with genetic code expansion (GCE) to introduce the fluorophore. The amber suppression GCE method can incorporate a non-canonical amino acid (nCAA) at a site specified by a UAG “amber” stop codon through the action of an evolved orthogonal tRNA/aminoacyl tRNA synthetase (aaRS) pair[58]. The nCAA can either be intrinsically fluorescent, such as acridonylalanine[59, 60], or it can be used to introduce an external probe through a biorthogonal or “click chemistry” reaction[61]. While the Petersson laboratory has used both approaches extensively in their prior work, the bright fluorophores necessary for single molecule measurements make the biorthogonal approach more attractive. Indeed, the Petersson and Rhoades laboratories have previously used NCL for PTM (Tyr phosphorylation, Glu arginylation) installation in combination with incorporation of the nCAA propargyl tyrosine (Ppy or $\text{π}$) for adding fluorescent probes by Cu(I)-catalyzed cycloaddition[14, 39]. Here, we combine that approach with the Pratt laboratory’s strategy for the synthesis of O-GlcNAc $\text{αS}$ to generate constructs that allow us to study the effect of Ser 87 glycosylation on membrane binding through FCS and compare those results to the effects of phosphorylation at that site.

2. N- and C-terminal fragment expression

2.1. Expression of $\text{αS}$ C-terminal fragment ($\mathbf{\text{α}}{\mathbf{\text{S}}}_{\mathbf{91}–\mathbf{114}}-{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}$)

nCAA incorporation via amber codon suppression was used to produce the C-terminal fragments for labeling with Atto 488 at position 114 as previously described[14]. Plasmid containing the desired $\text{αS}$ construct fused to a polyhistidine-tagged GyrA intein from Mycobacterium xenopi (Mxe) was transformed by heat shock at 42 °C into E. coli cells BL-21 cells with pre-transformed with a pDULE-pXF plasmid encoding the propargyltyrosine (Ppy or $\text{π}$) aaRS and tRNA. Cells were grown on ampicillin (Amp)/streptomycin (Strep) LB agar plates, and single colonies were picked to inoculate primary cultures in SOC (Super Optimal broth with Catabolite repression) media (2% w/v tryptone, 0.5% w/v sodium chloride, 0.5% w/v yeast extract, 250 mM potassium chloride, 100 mM magnesium chloride, 0.1% glucose v/v) supplemented with 0.1 mg/mL Amp and Strep. Secondary cultures were grown in M9 minimal media ($42\text{mM}{\text{Na}}_2{\text{HPO}}_4$, $22\text{mM}{\text{KH}}_2{\text{PO}}_4$, $8.6\text{mM}\text{NaCl}$, $19\text{mM}{\text{NH}}_4\text{Cl}$, $2\text{mM}{\text{MgSO}}_4$, $100\text{nM}{\text{CaCl}}_2$, $15\text{μg}/\text{L}{\text{FeCl}}_2$, $15\text{μg}/\text{L}{\text{ZnCl}}_2$, 0.02% w/v yeast extract, 0.5% w/v glucose) supplemented with 0.1 mg/mL Amp and Strep incubated at 37 °C in a shaker at 250 rpm until optical density (OD) reached ~0.8. Ppy (220 mg/L) was added to the culture with a 10 min incubation prior to inducing expression with isopropyl $\text{β}$-D-1-thiogalactopyranoside (IPTG). Cells were then grown in the shaker incubator at 18 °C overnight. After centrifugation (2800 xg, 20 min, 4 °C), cell pellets were resuspended in buffer (20 mM Tris pH 8.3, 1 Roche protease inhibitor tablet) and sonicated in a cup in an ice bath (5 min, 1 s ON, 1 s OFF). The resulting lysate was centrifuged (14,000 rpm, 25 min, 4 °C), and supernatant containing the protein of interest (POI) was purified over a Ni-NTA affinity column. The supernatant was incubated with a Ni-ATA Agarose resin (Gold Bio), which was previously equilibrated with the wash buffer (50 mM HEPES, pH 7.5), and the mixture was incubated for 1 h with gentle agitation at 4 °C. The matrix was washed with wash buffer (~5 column volumes) and then washed with ~ 5 column volumes of wash buffer 2 (50 mM HEPES, 5 mM imidazole, pH 7.5). The $\text{α}{\text{S}}_{91-140}-{\text{C}}_{91}$-intein was eluted with elution buffer (50 mM HEPES, 300 mM imidazole pH 7.5). Intein cleavage was carried out by incubation with 200 mM $\text{β}$-mercaptoethanol ($\text{β}$ ME) on a rotisserie overnight at room temperature. Cleaved $\mathbf{\text{α}}{\mathbf{\text{S}}}_{91-\mathbf{140}}-{\mathbf{\text{C}}}_{\mathbf{91}}-\mathbf{\text{C}}{\mathbf{\text{O}}}_{\mathbf{2}}\mathbf{\text{H}}$ was dialyzed into 20 mM Tris, pH 8 buffer before purification over a second Ni-NTA column to remove the free intein from the sample. $\mathbf{\text{α}}{\mathbf{\text{S}}}_{91-\mathbf{140}}-{\mathbf{\text{C}}}_{\mathbf{91}}-\mathbf{\text{C}}{\mathbf{\text{O}}}_{\mathbf{2}}\mathbf{\text{H}}$ was purified by reverse phase high-performance liquid chromatography (RP-HPLC) using a C4 column.

2.2. Expression of $\text{αS}$ N-terminal thioester ($\mathbf{\text{α}}{\mathbf{\text{S}}}_{\mathbf{1}-\mathbf{84}}-\mathbf{\text{SR}}$)

The N-terminal fragment of $\text{αS}$ (1–84) in a pTXB1 construct containing the Ava-DNAE N137A intein has been previously described[28]. BL21(DE3) chemically competent E. coli cells (VWR) were transformed with this plasmid by heat shock, plated on selective LB agar plates containing 100 μg/mL Amp, and incubated at 37 °C for 16 h. Single colonies were selected and used to inoculate starter cultures, which were grown at 37 °C with shaking at 250 rpm for 16 h. A starter culture was inoculated a 300 mL TB media culture (100 μg/mL Amp) and then grown to an OD of 0.6–0.7 at 37 °C with shaking at 250 rpm, and expression was induced with 0.5 mM IPTG at 25 °C with shaking at 250 rpm for 18 h. Bacteria were harvested by centrifugation (8,000 x g, 30 min, 4 °C), and the cell pellets were lysed by three freeze thaw cycles, using liquid ${\text{N}}_2$ and a 37 °C water bath. After harvesting bacteria by centrifugation (8,000 x g, 30 min, 4 °C), the cell pellet was resuspended on ice in 10 mL (per 300 mL of TB culture) of cold lysis buffer ($50\text{mM}{\text{NaH}}_2{\text{PO}}_4$, $300\text{mM}\text{NaCl}$, 5 mM imidazole, 2 mM TCEP HCl, pH 8.0) plus protease inhibitor cocktail and lysed by tip sonication (35% amplitude, 30 sec pulse duration, 30 sec rest for 12 min) while on ice. The crude cell lysate was cleared by centrifugation (42,000 x g, 30 min, 4 °C) and the supernatant was incubated with a Ni-ATA Agarose resin (Qiagen), which was previously washed with the wash buffer ($50\text{mM}{\text{NaH}}_2{\text{PO}}_4$, $300\text{mM}\text{NaCl}$, 20 mM imidazole, 2 mM TCEP HCl, pH 7.4), and the mixture was incubated for 1 h with shaking at 4 °C. The resin was washed with wash buffer (~5 column volumes), and the protein was eluted with elution buffer ($50\text{mM}{\text{NaH}}_2{\text{PO}}_4$, $300\text{mM}\text{NaCl}$, 250 mM imidazole, 2 mM TCEP HCl, pH=7.8). Elution fractions were concentrated by using spin-column concentrators (Amicon Ultra 3 kDa MW cut-off, Millipore) to a volume of 1 mL and then buffer exchanged into thiolysis buffer ($100\text{mM}{\text{NaH}}_2{\text{PO}}_4$, $150\text{mM}\text{NaCl}$, 1 mM EDTA, 1 mM TCEP HCl, pH=7.4). Sodium mercaptoethane sulfonate (MESNa) was added to a final concentration of 250 mM, and the thiolysis reaction was incubated at room temperature to generate the protein thioester. Upon completion, the thiolysis reaction was purified over a C4 semi-prep column (Higgins Analytical) and stored as a lyophilized solid. Pure protein thioesters were characterized by analytical RP-HPLC and electrospray ionization mass spectrometry (ESI-MS).

3. Synthesis of the $\mathbf{\text{αS}}$ central peptide thioesters ($\mathbf{\text{α}}{\mathbf{\text{S}}}_{\mathbf{85}-\mathbf{90}}-{\mathbf{\text{Z}}}_{\mathbf{85}}\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}-\mathbf{\text{SR}}$ and $\mathbf{\text{α}}{\mathbf{\text{S}}}_{\mathbf{85}-\mathbf{90}}-{\mathbf{\text{Z}}}_{\mathbf{85}}\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}-\mathbf{\text{SR}}$)

We synthesized both modified peptides manually using Dawson Dbz AM resin (Novabiochem) on a 0.1 mmol scale. In each coupling cycle (45 min), we used commercially available $N$-fluorenylmethoxycarbonyl (Fmoc) and side chain protected amino acids (5 eq, Novabiochem) by activation for 5 min with 5 eq of 2-(1 $H$-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU, Novabiochem) and 10 eq of $N$, $N^{\prime }$-diisopropylethylamine (DIEA, Sigma) in $N$, $N^{\prime }$-dimethylformamide (DMF, 3 mL). Following coupling, we removed the N-terminal Fmoc group by treatment with 20% v/v piperidine in DMF (4 mL) for 15 min. For the incorporation of O-GlcNAc modified serine, we prepared pentafluorophenyl (PFP) activated O-GlcNAc Fmoc-serine as described previously[62]. We purchased PFP-activated Fmoc-phospho-serine (Chem Impex). To couple these modified amino acids, we directly coupled them (2 eq) in DMF (3 mL) for 16 h. Upon completion of peptide synthesis, in the case of the O-GlcNAc modified peptide, we subjected the resin to hydrazine hydrate (80% v/v in MeOH, 3 mL) twice to remove the $O$-acetyl protecting groups from the sugar. We then activated the Dawson linker by treating with $p$-nitrophenyl chloroformate (5 eq) in ${\text{CH}}_2{\text{Cl}}_2$ (3 mL) for 1 h, followed by the addition of DIEA (5 eq) in DMF (3 mL) for 30 min. We cleaved the peptides from the resin and removed the amino acid protecting-groups by addition of our cleavage cocktail (95:2.5:2.5 trifluoroacetic acid/water/triisopropylsilane, 7 mL) for 4 h. We then precipitated the crude peptides by addition of 10 volumes of cold (−20 °C) diethyl ether and incubation of the resulting mixture at −80 °C for 16 h. We centrifuged the resulting suspension (5,000 x g, 30 min, 4 °C) and removed the supernatant by decanting before resuspending the peptide pellets in ${\text{H}}_2\text{O}$ and concentrating them by lyophilization. We then subjected the crude, dried material to 5 mL of thiolysis buffer ($150\text{mM}{\text{NaH}}_2{\text{PO}}_4$, 150 mM MESNa, pH 7.4) at room temperature for 2 h. Finally, we purified the resulting peptide thioesters using RP-HPLC and characterized them using ESI-MS.

4. Protein ligations

After generation of the three protein fragments, 1) an N-terminal thioester fragment (residues 1–84), 2) a central N-terminally protected peptide thioester (residues 85–90) containing either phosphorylated or $O$-GlcNAc at serine 87, and 3) a C-terminal fragment consisting of residues 91–140 with Ppy incorporated at residue 114, we first individually subjected the C-terminal fragment (4 mM, 1 eq) and either modified peptide (2 eq) in Expressed Protein Ligation buffer (3 M guanindine-HCl, 300 mM phosphate, 30 mM TCEP, 30 mM 4-mercaptophenylacetic acid (MPAA), pH 7.5). After 2 h at room temperature, we found that the reaction was complete by RP-HPLC (Figure 2, 1^st NCL) and the reaction mixture was treated with methoxylamine hydrochloride (150 mM) for 16 h. This results in the removal of the N-terminal thiazolindine protecting group from the newly formed ligation products (residues 85–140). We purified these protein products using RP-HPLC and then reacted them (Figure 2, 2^nd NCL) with the N-terminal thioester in the same ligation buffer. The mixture was placed on a rocker for 16 h at room temperature, and we purified the resulting full-length products by RP-HPLC. Both $\text{αS}$ constructs were characterized by matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) after the second NCL (Figures 3,5).

Figure 2.

RP-HPLC chromatograms after 16 h of the 1^st NCL (top) and 2^nd NCL (bottom) reactions for the synthesis of $\mathbf{\text{αS}}-{\mathbf{\text{C}}}_{\mathbf{85}}\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}$ (left) and $\mathbf{\text{αS}}-{\mathbf{\text{C}}}_{\mathbf{85}}\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}$ (right).

Figure 3.

MALDI-MS of $\mathbf{\text{αS}}-{\mathbf{\text{C}}}_{\mathbf{85}}\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}$ after 2^nd NCL.

Figure 5.

MALDI-MS of $\mathbf{\text{αS}}-{\mathbf{\text{C}}}_{\mathbf{85}}\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}$ after 2^nd NCL.

5. Fluorescent labeling

After purification, the ligated protein was re-dissolved in 20 mM Tris pH 8 and labeled with Atto 488-azide (Atto488-${\text{N}}_3$) via copper-catalyzed azide-alkyne cyclization. Catalytic mixture consisting of 2 equiv ${\text{CuSO}}_4$, 10 equiv Tris(3- hydroxypropyltriazolylmethyl)amine (THPTA), and 20 equiv sodium ascorbate was let sit for 10 min and added to the protein along with 2 equiv fluorophore. Labeling was monitored by MALDI-MS (Figures 3–6) at 30 min intervals, and was usually complete after 30–60 min. Labeled proteins were purified by RP-HPLC over a C4 column.

Figure 6.

MALDI-MS of $\mathbf{\text{αS}}-{\mathbf{\text{C}}}_{\mathbf{85}}\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ after click labeling with Atto-488 for 30 min.

6. Desulfurization

After labeling, to convert the cysteines used in ligation to the respective native alanines (${\text{A}}_{85}$ and ${\text{A}}_{91}$), the purified full-length product from above was redissolved in pH 7 Native Chemical Ligation (NCL) buffer (6 M Gdn•HCl, $200\text{mM}{\text{NaH}}_2{\text{PO}}_4$, pH 3) and incubated with 20 mM radical initiator 2,2′-azobis[2-(2-imidazolin-2-yl)propane (VA-044), 100 mM glutathione (GSH), and 250 mM TCEP in an argon-purged tube at 37 °C overnight. The desulfurized full-length products were purified by RP-HPLC over a C4 column and characterized by MALDI-MS (Figures 7,8).

Figure 7.

MALDI-MS of $\mathbf{\text{αS}}-\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ after desulfurization.

Figure 8.

MALDI-MS of $\mathbf{\text{αS}}-\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ after desulfurization.

7. Measuring lipid membrane binding using Fluorescence Correlation Spectroscopy (FCS)

7.1. FCS Measurements to determine the diffusion time of WT protein

Once we site-specifically modified protein constructs of interest, we can study the effect of each PTM on protein function. To this end, we take advantage of the single molecule technique Fluorescence Correlation Spectroscopy (FCS). FCS measures fluorescence intensity fluctuations of a fluorescent molecule as a function of time[63]. A small number of fluorescence molecules diffuse through a small observation volume, and the fluctuations in fluorescence intensity are analyzed through an autocorrelation algorithm to obtain autocorrelation curves as a function of time[64]. Requiring only nanomolar concentrations of samples, the binding of each protein (phosphorylation and glycosylation at ${\text{S}}_{87}$) was assayed by FCS to vesicles when compared to unmodified $\text{αS}$. Chambered cover slips (Nunc, ThermoFisher) were prepared by plasma cleaning followed by incubation overnight with polylysine-conjugated polyethylene glycol (PLL-PEG) using a modified Pierce PEGylation protocol (Pierce, Rockford, IL). Coated chambers were rinsed with and stored in Milli-Q water until use. Chambers are pre-incubated with PLL-PEG to minimize non-specific adsorption of the protein. FCS measurements were carried out on our home-built system as described previously[65] on an Olympus IX71 microscope with a continuous emission 488 nm DPSS 50 mW laser (Spectra-Physics, Santa Clara, CA). All measurements were made at 20°C. The laser power entering the microscope was adjusted to ~5 μW. Fluorescence emission collected through the objective was separated from the excitation signal through a Z488rdc long pass dichroic filter and an HQ600/200m bandpass filter (Chroma, Bellows Falls, VT). Emission signal was focused onto the aperture of a 50 μm optical fiber. Signal was amplified by an avalanche photodiode (Perkin Elmer, S13 Waltham, MA) coupled to the fiber. A digital correlator (FLEX03LQ-12, http://correlator.com) was used to collect 10 autocorrelation curves of 10 seconds for each measurement of WT free protein in buffer without lipids. Fitting was done using lab-written scripts[65] in MATLAB (The MathWorks, Natick, MA). To determine the diffusion time of the WT protein construct, WT $\text{αS}$ labeled with Atto 488 was measured in buffer without lipid. The average of 10 autocorrelation curves was fit to a 1- component autocorrelation function:

$$G(\tau )=\frac{1}{N}\times \frac{1}{1+\frac{\tau }{{\tau }_D}}\times \sqrt{\frac{1}{1+\frac{s^2\tau }{{\tau }_D}}}$$

where $G(\tau )$ is the autocorrelation function, $N$ is the number of molecules in the focal volume, ${\tau }_D$ is the diffusion time of $\text{αS}$, and $s$ is the ratio of radial to axial dimensions of the focal volume of the instrument. The counts per molecule (CPM) for each sample was calculated by dividing the average intensity (Hz) of the measured signal by the number of molecules $N$. The normalized CPM of each $\text{αS}$ was calculated by dividing by the CPM of freely diffusing fluorescent standard Atto 488. The average of 10 autocorrelation curves for each construct in the presence of vesicles was plotted, along with the fit to the above equation.

7.2. Preparation of synthetic lipid vesicles

Lipid vesicles were prepared by extrusion through porous membranes. A mixture in 50:50 molar ratio of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoserine (POPS) and 1-palmitoyl-2-oleoyl-sn-glycero-3- phosphocholine (POPC) were drawn from chloroform stock concentrations of 20 mg/mL as determined by total phosphorous assay[66, 67] (protocol adapted from Avanti Polar Lipids), and dried under nitrogen gas to form a film inside a glass vial. Films were desiccated under vacuum and re-hydrated in 3-($N$-morpholino)propanesulfonic acid (MOPS) buffer. 10 freeze-thaw cycles consisting of cooling in liquid nitrogen for 30 s and warming in a 50 °C water bath for 2 min were performed to aid the formation of uniformly sized vesicles. Using Hamilton syringes, vesicles were pushed 31 times through 50 nm pore membranes held in place inside an extruder set-up (Avanti Polar Lipids). Vesicles were determined by Dynamic Light Scattering (DLS) to be monodisperse and distributed uniformly around 80 nm in diameter, consistent across different concentrations. All lipid vesicles were prepared fresh and used within 48 h of extrusion.

7.3. Measuring vesicle binding

αS constructs labeled with Atto 488 were examined in the presence of 0.05 mM or 0.5 mM lipid vesicles consisting of 50:50 POPS/PC. The same digital correlator as described above was used to collect 30 autocorrelation curves of 30 seconds for each measurement in the presence of lipid vesicles. The average of 30 autocorrelation curves was fit to a 2-component equation:

$$G(\tau )=\frac{1}{N}\times \left(A\times \frac{1}{1+\frac{\tau }{{\tau }_1}}\times \sqrt{\frac{1}{1+\frac{s^2\tau }{{\tau }_1}}}+Q\times (1-A)\times \frac{1}{1+\frac{\tau }{{\tau }_2}}\times \sqrt{\frac{1}{1+\frac{s^2\tau }{{\tau }_2}}}\right)$$

where $G(\tau )$ is the autocorrelation function, N is the number of molecules in the focal volume, ${\tau }_1$ is the characteristic diffusion time of $\text{αS}$, ${\tau }_2$ is the characteristic diffusion time of the vesicles, s is the ratio of radial to axial dimensions of the focal volume of the instrument, Q is the brightness factor, and A is the fraction of bound $\text{αS}$. When fitting the autocorrelation curves for $\text{αS}$ in the presence of lipid vesicles, the diffusion time of unbound WT $\text{αS}$, ${\tau }_1$, was fixed to the diffusion time of the vesicles determined above; ${\tau }_2$ was fixed to the previously determined diffusion time of WT $\text{αS}$ in the presence of saturating 2 mM 50:50 POPS/PC vesicles.[36] The average of 30 autocorrelation curves for each construct in the presence of vesicles was plotted, along with the fit to the above equation.

8. Results

8.1. Synthesis of $\text{αS}$ constructs

Full-length doubly-modified $\text{αS}$ was successfully synthesized with either $\text{gS}$ or $\text{pS}$ at position 87 and Alexafluor 488 attached at position 114. The 85–90 segments were produced through solid phase peptide synthesis (SPPS) as N-terminally thiazolidine-protected C-terminal 2-mercaptoethane sulfonate (MES) thioesters in 28% ($\text{gS}$) or 30% ($\text{pS}$) yields. The 91–140 segment was recombinantly expressed in E. coli (21 mg/L) with Ppy inserted at position 114 using a M. jannaschii aaRS that is promiscuous for para-substituted phenylalanine derivatives. This segment also included an Ala-to-Cys mutation at position 91 for ligation, an appended N-terminal methionine, and a C-terminal intein with a His₆ purification tag. The N-terminal Met is removed in cells by methionine aminopeptidase so that the isolated protein contains an N-terminal thiazolidine, formed by reaction of the resulting N-terminal Cys with intracellular aldehydes. For the 91–140 segment, the intein is not needed for production of a C-terminal thioester, but we find that this tag is useful for separation from truncated protein since the intein-His₆ fusion is only produced upon successful ncAA insertion. After Ni-column purification using the His₆ tag, cleavage of the intein-His₆ using $\text{β}$-mercaptoethanol (βME), and deprotection of the N-terminal thiazolidine with methoxyamine, the 91–140 segment was ligated to either of the 85–90 segments (1^st NCL). Although the conversion of the 1^st NCL for the $\text{gS}$ protein was lower (59%), our result with the $\text{pS}$ protein shows that this ligation can be quantitative. After deprotection of the thiazolidine group, the 85–140 segments were then ligated (2^nd NCL) to the recombinantly expressed 1–84 segment, which was generated from a recombinantly expressed C-terminal intein-His₆ fusion as a MES thioester (12 mg/L). 4-Mercaptophenylacetic acid (MPAA) was added to accelerate the 2^nd NCL, which again proceeded better for the $\text{pS}$ protein, achieving 87% conversion of the 85–140 segment to full length $\mathbf{\text{αS}}-{\mathbf{\text{C}}}_{\mathbf{85}}\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}$ after 16 h. Conversion in the $\text{gS}$ reaction was again lower, achieving 28% yield with respect to the 1–84 segment, which was the limiting reagent. The yields of $\mathbf{\text{αS}}-{\mathbf{\text{C}}}_{\mathbf{85}}{\mathbf{\text{gS}}}_{\mathbf{87}}{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}$ and $\mathbf{\text{αS}}-{\mathbf{\text{C}}}_{\mathbf{85}}{\mathbf{\text{pS}}}_{\mathbf{87}}{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}$ after isolation by HPLC, were 1.3 mg and 1.4 mg, respectively. Next, the Ppy residues were labeled with Alexafluor 488 azide in a copper-catalyzed click reaction, and then the Cys ligation points were desulfurized to native Ala residues using the radical initiator VA-044 with glutathione. Both of these conversions were quantitative, but RP-HPLC purification was used to ensure complete removal of the small molecule reagents. The final yields, of purified $\mathbf{\text{αS}}-\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ and $\mathbf{\text{αS}}-\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ were <0.1 mg, indicating significant losses during purification in spite of quantitative conversion in the labeling and desulfurization reactions. While higher isolated yields could potentially be obtained by substituting dialysis for RP-HPLC purification after click labeling and desulfurization, for applications in single molecule fluorescence such as FCS, high purity is more important than the quantity of labeled protein.

8.2. Reduced binding of modified $\text{αS}$ to vesicles

Using the labeled $\mathbf{\text{αS}}-\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ and $\mathbf{\text{αS}}-\mathbf{\text{p}}{\mathbf{\text{S}}}_{87}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ proteins, we made FCS measurements in the presence of 80 nm vesicles composed of 50:50 1-palmitoyl-2-oleoyl-sn-glycero-3phosphoserine (POPS) and 1-palmitoyl-2-oleoyl-sn-glycero-3- phosphocholine (POPC). As shown in Figure 9, the $\text{pS}$ and $\text{gS}$ constructs both show shifts in their autocorrelation curves in the presence of 50 μM POPS/PC vesicles, but these are smaller than the shifts seen for WT $\text{αS}$ ($\mathbf{\text{αS}}-{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$). Previous experiments have shown that WT $\text{αS}$ binding to 80 nm POPS/PC vesicles is already saturated at 50 μM,[14, 39] so the smaller shifts for ${\text{pS}}_{87}$ and ${\text{gS}}_{87}$ could indicate that these variants have a weaker affinity. At 500 μM vesicle concentrations, there is no change in the autocorrelation curve for WT $\text{αS}$, which is as expected since binding should be saturated. However, there is also little change in the autocorrelation curves for pS₈₇ and ${\text{gS}}_{87}$ (Fig. 9), indicating that binding is either much weaker than WT $\text{αS}$ or that these proteins have lower apparent diffusion times when bound to vesicles. In either case, phosphorylation and glycosylation at serine 87 seem to have similar effects on $\text{αS}$ membrane binding. Further investigations will be necessary to determine the reason for the persistently lower diffusion times of the $\mathbf{\text{αS}}-\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ and $\mathbf{\text{αS}}-\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ constructs when bound to vesicles.

Figure 9.

Averaged FCS autocorrelation curves with fits for WT ($\mathbf{\text{αS}}-{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$), $\text{gS}$ ($\mathbf{\text{αS}}-\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$) and ${\text{pS}}_{87}$ ($\mathbf{\text{αS}}-\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$) either 50 μM or 500 μM POPS/PC vesicles. WT in the absence of vesicles is also plotted for comparison.

9. Discussion

In this manuscript, we have provided detailed experimental procedures for the construction of site-specifically phosphorylated or glycosylated $\text{αS}$ that also contains a fluorescent probe. This was accomplished using a combination of nCAA incorporation by GCE and click chemistry to easily introduce the fluorescent label, along with a three-part NCL strategy for $\text{gS}$ or $\text{pS}$ introduction. Using a synthetic strategy to incorporate the glycosylated or phosphorylated amino acid at site 87 allowed us to study the authentic PTM, offering an advantage over site-directed mutagenesis to introduce a natural amino acid PTM-mimic, because these do not always recapitulate the effects of the authentic PTM. Our study represents an efficient synthetic strategy for incorporating these selective PTMs concurrently with fluorescent probes into proteins, providing simultaneous insight into the potential neuroprotective effects of both PTMs. The results here, combined with our previous publications, provide support for the hypothesis that these modifications maintain the levels of soluble $\text{αS}$ as observed by the altered vesicle binding. However, additional study is needed to explain why the diffusion times do not reach those of WT $\text{αS}$ even at high concentrations.

More broadly, the protocols for the preparation of these $\text{αS}$ constructs can be readily adapted to the preparation and labeling of other proteins with a broad range of applications. Our work highlights the value of combining GCE and NCL approaches, which in spite of the vast numbers of publications using each method for protein modification, have seen only a few instances of their combined use. To our knowledge, the only examples of NCL use together with GCE, outside of our work on multiply labeling proteins with fluorophores and PTMs or biophysical probes, are the following: Chin has used the NCL + GCE approach to attach ubiquitin using a thiol-functionalized lysine derivative nCAA,[68] Schultz and others have used GCE incorporation of a thioester nCAA for protein sidechain functionalization via NCL,[69] and Rozovsky and Wang have site-specifically incorporated selenocysteine via GCE,[70] which has been used extensively in NCL by Rozovsky and others. NCL using SPPS affords the opportunity to incorporate the many modifications that cannot be accessed through GCE. Expressed protein ligation using GCE enables the production of large protein segments with a single modification, saving significant synthetic labor. We hope that the methods described here show how NCL and GCE can be combined relatively easy, making access to multiply modified proteins much more straightforward than with either method alone.

Figure 4.

MALDI-MS of $\mathbf{\text{αS}}-{\mathbf{\text{C}}}_{\mathbf{85}}\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{C}}}_{\mathbf{91}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ after click labeling with Atto-488 for 30 min.

Scheme 1.

Semi-synthesis of $\mathbf{\text{αS}}-\mathbf{\text{g}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$ and $\mathbf{\text{αS}}-\mathbf{\text{p}}{\mathbf{\text{S}}}_{\mathbf{87}}{\mathbf{\text{π}}}_{\mathbf{114}}^{\mathbf{488}}$. The N-terminal $\text{α}{\text{S}}_{1-84}$ fragment (navy) is expressed as a C-terminal intein fusion (Int-${\text{H}}_6$) for thioester functionalization and ligation. The central synthetic $\text{α}{\text{S}}_{85-90}-{\text{C}}_{85}$ fragment (orange) with O-GlcNAc or pSer at position 87 (both represented by S*) is made by solid phase peptide synthesis (SPPS) with an N-terminal thiazolidine protecting group and a thioester for ligation. The C-terminal $\text{α}{\text{S}}_{91-140}-{\text{C}}_{91}$ fragment (red) is expressed with an N-terminal cysteine for ligation and $\text{π}$ incorporated at position 114 by nCAA mutagenesis for labeling with Atto-488 azide.