Lipid-Coated Gold Nanoparticles and FRET Allow Sensitive

Article pubs.acs.org/Langmuir

Lipid-Coated Gold Nanoparticles and FRET Allow Sensitive Monitoring of Liposome Clustering Mediated by the Synaptotagmin‑7 C2A Domain Desmond J. Hamilton, Matthew D. Coffman, Jefferson D. Knight,* and Scott M. Reed* Department of Chemistry, University of Colorado Denver, Denver, Colorado 80217, United States ABSTRACT: Synaptotagmin (Syt) family proteins contain tandem C2 domains, C2A and C2B, which insert into anionic membranes in response to increased cytosolic Ca2+ concentration and facilitate exocytosis in neuronal and endocrine cells. The C2A domain from Syt7 binds lipid membranes much more tightly than the corresponding domain from Syt1, but the implications of this difference for protein function are not yet clear. In particular, the ability of the isolated Syt7 C2A domain to initiate membrane apposition and/or aggregation has been previously unexplored. Here, we demonstrate that Syt7 C2A induces apposition and aggregation of liposomes using Förster resonance energy transfer (FRET) assays, dynamic light scattering, and spectroscopic techniques involving lipid-coated gold nanoparticles (LCAuNPs). Protein− membrane binding, membrane apposition, and macroscopic aggregation are three separate phenomena with distinct Ca2+ requirements: the threshold Ca2+ concentration for membrane binding is lowest, followed by apposition and aggregation. However, aggregation is highly sensitive to protein concentration and can occur even at submicromolar Syt7 C2A; thus, highly sensitive assays are needed for measuring apposition without complications arising from aggregation. Notably, the localized surface plasmon resonance of the LCAuNP is sensitive to ≤10 nM Syt7 C2A concentrations. Furthermore, when the LCAuNPs were added into a FRET-based liposome apposition assay, the resultant energy transfer increased; possible explanations are discussed. Overall, LCAuNP-based methods allow for highly sensitive detection of protein-induced membrane apposition under conditions that miminize large-scale aggregation.

■

INTRODUCTION Synaptotagmin (Syt) proteins serve as a primary Ca2+ sensor for regulated exocytosis. Seventeen human isoforms of Syt exist, each of which contains two C-terminal C2 domains and a transmembrane helix near the N-terminus.1,2 In eight of these isoforms, the C2 domains dock to anionic membranes in a Ca2+-dependent fashion.3,4 These Ca2+-sensitive Syt isoforms accelerate SNARE-mediated membrane fusion in vivo as well as in reconstituted vesicle fusion assays via a mechanism that has remained elusive.5−7 One model posits that C2 domains bind to membranes to facilitate bridging, apposition, and ultimately fusion of the secretory granule membrane with the target plasma membrane; correspondingly, some Syt C2 domains are known to induce apposition and aggregation of secretory vesicles on their own in vitro.7,8 Correspondingly, both the C2AB fragment and the individual C2B domain of Syt1 can initiate liposome aggregation in vitro.9,10 Syt1 is the most extensively studied Syt isoform and is the primary isoform responsible for rapid exocytosis of neurotransmitter secretory vesicles.11−14 Conflicting reports exist on the ability of the Syt1 C2A domain to initiate liposome clustering and aggregation; when observed, it requires a much higher protein concentration than Syt1 C2B.10,15,16 The most Ca2+-sensitive Syt isoform, Syt7, facilitates slow asynchronous neurotransmitter release as well as exocytosis of large dense-core vesicles.11,14,17−19 The C2A domain of Syt7 has a much stronger membrane affinity than Syt1 C2A due to a © 2017 American Chemical Society

combination of deeper membrane penetration and electrostatic effects.20−23 The ability of Syt7 C2A to initiate liposome aggregation has been reported anecdotally but has not been studied systematically.20,23 In extremes, aggregation can lead to uncontrolled flocculation which is difficult to monitor quantitatively. Because all aggregation phenomena are highly concentration dependent, sensitive assays are helpful for detecting clustering and aggregation using low concentrations of protein and lipid in order to slow aggregation and/or limit the size of clusters. Gold nanoparticles (AuNPs) are particularly useful as biosensors due to their high extinction coefficients, increased photostability compared to organic fluorophores, relatively inert behavior, and ability to be coated with biological molecules such as lipids to create mimics of cellular membranes.24−29 Recently, the localized surface plasmon resonance (LSPR) of lipid-coated gold nanoparticles (LCAuNPs) has been used to monitor protein−membrane interactions.30,31 The LSPR wavelength of AuNPs is sensitive not only to changes in the refractive index near the nanoparticle surface but also to aggregation which produces a red-shift in the plasmon due to interparticle coupling.32−37 Furthermore, the LSPR of AuNPs has been shown theoretically and experReceived: April 24, 2017 Revised: July 26, 2017 Published: August 29, 2017 9222

DOI: 10.1021/acs.langmuir.7b01397 Langmuir 2017, 33, 9222−9230

Article

Langmuir

24 h. The vials were individually lowered into a 197 °C oil bath and rapidly stirred for 30 min. Caution: once lowered into the hot oil bath the vials are under increased pressure, and extra caution should be taken when handling them! Conversion to Highly Uniform Spherical AuNPs. For every 1 mL of octahedral AuNP colloid, 1.5 μL of 0.6 M NaCN in water was added and vigorously stirred for 20 h. The spherical AuNPs were purified using four rounds of centrifugation (Centrifuge 5424, Eppendorf) followed by sonication and resuspension in Milli-Q H2O similar to our previous report.45 Transmission Electron Microscopy (TEM). TEM images of the purified spherical AuNPs were obtained by adding 5 μL of AuNP colloid onto a copper grid (FCF300-Cu Formvar carbon film on 300 mesh copper grids, Electron Microscopy Sciences) and allowed to airdry. A FEI Tecnai G2 TEM with a Gatan Ultrascan 1000 digital camera was used to obtain the images. ImageJ measurement functions were used to determine the diameter of the AuNPs. Construction of LCAuNPs. To 2.4 mL of purified spherical AuNPs at an optical density (OD) of 0.67, 314 nmol of PC:PS (3:1) liposomes was added and stirred for 30 min. Propanethiol (PT) (1 μL, neat) was added to 10 mL of Milli-Q H2O, degassed with N2, and subsequently added (360 nmol PT) to the AuNPs within 24 h and allowed to stir for 1 h. The LCAuNPs were then purified to remove as much excess phospholipid as possible while maintaining LCAuNP solubility: the LCAuNP suspension was centrifuged in low-adhesion Eppendorf tubes once, followed by the removal of 90% of the supernatant (decreasing the liposome concentration 10-fold) and resuspension in Milli-Q water via sonication for 10 min. The LCAuNP colloids were then diluted to either 0.24 OD for LSPR or 0.9 OD for FRET assays. The complete membrane encapsulation of LCAuNPs was verified by adding 26 μL of 0.1 M NaCN in water to 1 mL of the LCAuNP colloid and stirring for at least 16 h. As a control, 26 μL of water was added to a sample of LCAuNPs and stirred for at least 16 h. UV−vis spectra (PerkinElmer 650) of the treated LCAuNPs were obtained and compared to the control. Observation of LSPR. LSPR OD, centroid, and full width at halfmaximum (fwhm) were monitored as previously reported.31 An Ocean Optics DH-2000 deuterium−tungsten halogen light source and an Ocean Optics HR4000 spectrometer were used to obtain extinction spectra. Sucrose Sensitivity Assays. LCAuNP colloid, 1 mL of 0.24 OD, was added to a polystyrene cuvette, and the LSPR was monitored for 5 min. Then, 5% w/v sucrose was added serially with 5 min between each addition up to 15% sucrose. The difference in centroid wavelength for each addition was obtained by averaging the final 110 s (100 data points) immediately before addition of sucrose and compared to the final 110 s (100 data points) after the addition of sucrose. The sensitivity was determined by performing linear regression on the change in centroid with respect to the known refractive index (RI) difference between sucrose solutions and water. Syt7 C2A LCAuNP LSPR Sensitivity. The LCAuNP colloid was diluted to 1 mL of approximately 0.24 OD in a 3 mL polystyrene cuvette with a Teflon stir bar, and the LSPR was monitored for 15 min. All stir bars were rinsed in freshly prepared aqua regia and rinsed once with Milli-Q H2O, once with ethanol, and five times with Milli-Q H2O prior to use. Buffer A was added and monitored for 30 min. Four additions of 9.3 nM Syt7 C2A were performed in 15 min intervals. For experiments with LCAuNPs, the quantity of lipid in P:L ratios is the amount assumed to be accessible for protein binding: the total phospholipid upon LCAuNPs plus one-half the phospholipid in liposomes. Finally, 2.6 mM EDTA was added, and the LSPR was monitored for an additional 15 min. DLS of Liposome Clustering. A suspension of 90 μM PC:PS (3:1) liposomes was prepared in Buffer A with 100 μM Ca2+. The zaverage diameter was obtained prior to Syt7 C2A addition. Syt7 C2A was added sequentially as nine equal additions up to a final concentration of 463 nM Syt7 C2A. A total of 30 min elapsed between each addition, and data were acquired 15 min after each addition. EDTA (2.9 mM) was added 30 min after the last Syt7 C2A addition, and the z-average diameter was measured.

imentally to enhance energy transfer between donor/acceptor pairs in proximity to the surface of AuNPs.38−40 FRET assays between organic donor/acceptor pairs have commonly been used as molecular rulers to monitor protein membrane interactions and membrane fusion events;41 nanoparticles have the potential to extend the range of such measurements.42 In this study we develop novel inter-liposome energy transfer assays with and without nanoparticles that allow for sensitive detection of Syt7 C2A-mediated liposome apposition and clustering. In contrast to commonly used turbidity assays which monitor changes in the scattering of light as aggregrates form,43 our methods allow detection of membrane apposition even under conditions where large-scale aggregation is not observed. Spherical AuNPs were prepared and then coated with a hybrid bilayer composed of an inner leaflet with a hydrophobic anchor, propanethiol, and an outer leaflet of phospholipids to which Syt7 C2A is known to bind. Changes in the LSPR, through interparticle coupling of LCAuNPs, were used to detect Syt7 C2A-mediated LCAuNP aggregation with as little as 9.3 nM Syt7 C2A. An interliposomal FRET assay with donor and acceptor fluorophores on separate liposomes allowed for further investigation of Syt7 C2A-induced liposome apposition and aggregation. Additionally, our results show that Syt7 C2A is able to mediate liposome clustering in a manner dependent on the protein:accessible lipid (P:L) ratio, Ca2+ concentration, and Syt7 C2A concentration. Finally, the interliposomal FRET assay in the presence of LCAuNPs led to an observed increase in energy transfer, suggesting the possibility of nanoparticle enhanced energy transfer (NEET).

■

EXPERIMENTAL SECTION

The synthetic fluorescent lipid 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(5-dimethylamino-1-naphthalenesulfonyl) (dansylPE) was from NOF America. All other lipids were synthetic and obtained from Avanti Polar Lipids: 1-palmitoyl-2-oleoyl-sn-glycero-3phosphocholine (phosphatidylcholine, PC), 1,2-dioleoyl-sn-glycero-3phosphoserine (phosphatidylserine, PS), 1,2-dipalmitoyl-sn-glycero-3phosphoethanolamine-N-(7-nitro-2−1,3-benzoxadiazol-4-yl) (NBD), and 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (RHO). Water (18.2 MΩ cm) was purified with a Milli-Q Academic Purification System (EMD Millipore). Propanethiol (Acros Organics), phosphoric acid (J.T. Baker), polydiallyldimethylammonium chloride (Sigma-Aldrich), ethylene glycol (Oakwood Chemicals), hydrogen tetrachloroaurate (Strem Chemicals), sodium cyanide (Alfa Aesar), N-2-hydroxyethylpiperazineN-2-ethanesulfonic acid (HEPES) (BDH), ethylenediaminetetraacetic acid (EDTA) (Sigma-Aldrich), sodium chloride (Mallinckrodt), and potassium chloride (Amresco) were used without further purification. Syt7 C2A was expressed and purified as described previously.20 Preparation of Liposomes and Membrane Components. PC and PS were mixed in a 3:1 molar ratio in chloroform and stored at −20 °C. The lipid solutions were dried into thin films with a flow of N2 and left under vacuum for at least 4 h. The films were resuspended in Milli-Q H2O or Buffer A (140 mM KCl, 15 mM NaCl, 25 mM HEPES pH 7.48), sonicated for 10 min (Branson 1510, Branson Ultrasonics, Danbury, CT), and extruded 13 times using a Mini Extruder (Avanti Polar Lipids, Alabaster, AL) with 100 nm pore size polycarbonate filters (Whatman). The z-average diameter of the liposomes was determined by dynamic light scattering (DLS) to be ∼145 nm. Synthesis of Octahedral AuNPs. Octahedral AuNPs were synthesized by modification of an existing procedure.44 First, 200 mL of ethylene glycol, 4 g of polydiallyldimethylammonium chloride, and 8 mL 1 M H3PO4 were stirred for 15 min. Then 200 μL of 0.5 M HAuCl4 in H2O was added and stirred for 15 min. The suspension was split into 20 10 mL aliquots in sealed scintillation vials and used within 9223

DOI: 10.1021/acs.langmuir.7b01397 Langmuir 2017, 33, 9222−9230

Article

Langmuir FRET Monitoring of Inter-Liposome Interactions. FRET experiments were completed using a Photon Technology International QM-2000-6SE fluorescence spectrometer at room temperature with slit widths of 1 nm (excitation) and 8 nm (emission). Separately labeled liposomes were suspended in Buffer A: 50 μM PC:PS (3:1) with 1.5% NBD and 50 μM PC:PS (3:1) with 1.5% RHO. Either Ca2+ or Syt7 C2A was added stepwise while stirring. For the LCAuNP assays, either 150 μL of 0.9 OD LCAuNPs or 150 μL of 11.5 μM unlabeled PC:PS (3:1) was added to 1 mL of premixed RHO/NBD liposomes prior to addition of Syt7 C2A. The donor, NBD, was excited at 460 nm, and the emission intensities of NBD and the acceptor, RHO, were collected at 530 and 583 nm, respectively. Experiments were completed in triplicate and averaged, and the controls shown are representative. Data were collected with integration times of 2 s, and emission intensity ratios (583 nm/530 nm) were time averaged for each minute collected and normalized by the average of the first 10 min for the experiment. EDTA was added at the end of each assay to chelate Ca2+, and the increase in emission ratio (RHO/NBD) was determined from the normalized baseline. FRET Monitoring of Syt7 C2A Binding. To monitor Syt7 C2A− PC:PS (3:1) membrane interactions, a Syt7 C2A Trp and dansyl-PE FRET pair was used. 1 mL of 100 μM PC:PS (3:1) with 5% dansyl-PE in Buffer A without added Ca2+ was monitored for 10 min. Syt7 C2A, 488 nM, was added, followed by the stepwise addition of 0.33, 1, 3, 10, 30, and 100 μM Ca2+ with 15 min in between additions. Finally, EDTA was added to chelate Ca2+ ions. Excitation was at 284 nm, and emission was collected at 512 nm. This assay was run in triplicate and averaged, and the control shown is representative.

To verify whether the lipid membranes completely encapsulate the AuNPs, NaCN was added to the LCAuNP colloids and the LSPR was measured in the 400−750 nm spectral range. Cyanide assists in the oxidation of Au to dissolve uncoated AuNP and form Au(CN)2 which does not absorb visible light (Figure 2, dashed line). LCAuNPs prepared as described above were resistant to NaCN treatment, indicating the presence of an ion-impermeable membrane surrounding the AuNPs (Figure 2).

■

Figure 2. LCAuNPs resistance to NaCN. Representative absorbance spectra of 40 nm diameter LCAuNPs used to determine efficiency of encapsulation: (gray) before addition of NaCN to LCAuNPs, (black) 16 h after NaCN addition, and (dashed black) 40 nm diameter AuNPs without lipid membrane 16 h after NaCN. The percentage of OD remaining after 16 h in the LCAuNP samples was 93.0 ± 3.9. N = 3.

RESULTS AND DISCUSSION Synthesis of Spherical AuNPs. In order to obtain highly uniform spherical AuNPs, octahedral AuNPs were first synthesized using a previously published protocol44 that was modified for larger batch synthesis.45 The octahedral AuNPs were subsequently converted to spherical AuNPs by addition of sodium cyanide,45 then purified via centrifugation, and redispersed in H2O with sonication. TEM analysis showed that the resulting AuNPs were spherical with an average diameter of 40 ± 7 nm (Figure 1). Construction of Lipid Membrane. A lipid membrane was assembled on the spherical AuNPs via the sequential addition of PC:PS (3:1) liposomes and propanethiol. Excess liposomes and propanethiol were removed via centrifugation, and the resultant LCAuNPs were then redispersed via sonication.

LCAuNPs Sensitivity to Changes in Refractive Index. The wavelength profile of the LSPR of AuNPs depends on the particle size and the RI of the surrounding solution.46 In order to determine the sensitivity of the LCAuNPs to solution RI, we monitored the centroid of the LSPR peak from extinction spectra of LCAuNPs upon serial addition of sucrose. Increasing sucrose concentrations provided an increased solution RI, which caused a red-shift in the centroid. The amount of redshift was greater upon the first addition than on subsequent additions, possibly due to direct interactions between sucrose and LCAuNPs (Figure 3A). After this point, the change in centroid per 5% sucrose addition was linear, with a LCAuNP RI sensitivity of 26 ± 1 nm/(RI unit) (Figure 3B). LSPR Centroid Shift Induced by Syt7 C2A. Proteins can induce changes in RI immediately adjacent to LCAuNP both by binding to surface lipids (if the protein RI differs from bulk solution) and by mediating association or clustering of LCAuNPs with liposomes or each other, as shown in our previous study.31 In principle, a protein that binds tightly and/ or induces clustering may produce centroid shifts that are detectable even at very low protein concentrations due to the strong extinction of AuNPs. In order to probe the sensitivity of the LSPR measurement, we chose a protein domain with a very strong membrane binding affinity and an ability to initiate clustering. The C2A domain of Syt7 has a strong affinity for PC:PS (3:1) liposomes in the presence of Ca2+.20 The qualitative detection limit of Syt7 C2A effects on LCAuNPs with PC:PS (3:1) was determined using serial additions of Syt7 C2A to colloidal suspensions of 40 nm LCAuNPs in the presence of 100 μM Ca2+. A red-shift in the

Figure 1. Transmission electron microscope images of 40 ± 7 nm spherical AuNPs. Scale bar: 200 nm. 9224

DOI: 10.1021/acs.langmuir.7b01397 Langmuir 2017, 33, 9222−9230

Article

Langmuir

Figure 4. LSPR response to serial addition of Syt7 C2A to LCAuNPs. Additions to 40 nm diameter LCAuNPs were completed with (black) Syt7 C2A and (gray) Buffer A added containing no Syt7 C2A. The sequence of the experiment was (1) LCAuNPs, (2) addition of Buffer A with a final Ca2+ concentration of 100 μM, (3−6) additions of 9.3 nM Syt7 C2A (P:L upon additions 3−6: 0.0023, 0.0045, 0.0068, and 0.0091), and (7) addition of 2.6 mM EDTA.

Figure 3. LCAuNP sensitivity to changes in RI. (A) Representative time series of serial sucrose additions of 5% w/v (1, 2, and 3) to 40 nm diameter LCAuNPs. The initial spike and decay in centroid is likely due to mixing artifacts. (B) The difference in centroid wavelength from water for each sucrose concentration. The sensitivity to changes in RI was 26 ± 1 nm/(RI unit). Error bars are reported as standard deviation. N = 3.

LSPR centroid was observed upon addition of 9.3 nM Syt7 C2A, and each subsequent 9.3 nM addition of the protein produced a further red-shift, although the final addition led to a blue-shift after a few minutes (Figure 4). On the basis of the concentration of the AuNPs,47 there was an effective concentration of 460 nM total accessible lipid on the LCAuNP surface in these experiments. It is important to note that in order to maintain solubility of LCAuNPs in these experiments, each LCAuNP suspension contained residual liposomes at approximately a 9:1 ratio of liposome surface-accessible lipid to LCAuNP surface-accessible lipid. The resultant dilution should decrease the fraction of LCAuNPs that undergo a red-shift. Assuming that the protein binds equally well to PC:PS (3:1) liposomes and LCAuNPs, the effective concentration of LCAuNP-bound protein detected in these experiments is