Biological membrane fusion is catalyzed by specialized fusion proteins. Measuring the fusogenic properties of proteins can be achieved by lipid mixing assays. We present a method for purifying recombinant Drosophila atlastin, a protein that mediates homotypic fusion of the ER, reconstituting it to preformed liposomes, and testing for fusion capacity.
Membrane fusion is a crucial process in the eukaryotic cell. Specialized proteins are necessary to catalyze fusion. Atlastins are endoplasmic reticulum (ER) resident proteins implicated in homotypic fusion of the ER. We detail here a method for purifying a glutathione S-transferase (GST) and poly-histidine tagged Drosophila atlastin by two rounds of affinity chromatography. Studying fusion reactions in vitro requires purified fusion proteins to be inserted into a lipid bilayer. Liposomes are ideal model membranes, as lipid composition and size may be adjusted. To this end, we describe a reconstitution method by detergent removal for Drosophila atlastin into preformed liposomes. While several reconstitution methods are available, reconstitution by detergent removal has several advantages that make it suitable for atlastins and other similar proteins. The advantage of this method includes a high reconstitution yield and correct orientation of the reconstituted protein. This method can be extended to other membrane proteins and for other applications that require proteoliposomes. Additionally, we describe a FRET based lipid mixing assay of proteoliposomes used as a measurement of membrane fusion.
Membrane fusion is a critical process in many biological reactions. Under biological conditions, membrane fusion is not spontaneous and requires specialized fusion proteins to catalyze such reactions1. ER homotypic membrane fusion is mediated in animals by the dynamin related GTPase atlastin2. Atlastin’s role in homotypic fusion is fundamental for three-way junctions in peripheral ER, which constitutes a large interconnected network of tubules that extend throughout the cell. Atlastins have a conserved domain morphology consisting of a large GTPase, a three helix bundle middle domain, a hydrophobic membrane anchor, and a short cytoplasmic C-terminal tail3. In vitro studies with recombinant Drosophila atlastin have shown that when reconstituted to liposomes, it maintains its fusogenic properties. Other atlastins, including human homologs have not been able to recapitulate fusion in vitro. We describe here a methodology for purifying a GST and poly-histidine tagged recombinant Drosophila atlastin, reconstituting it to liposomes, and assaying fusion.
Studying membrane fusion in vitro presents a challenge as fusogenic proteins usually have a membrane anchor. In order to study them, it is necessary to reconstitute them into model lipid bilayers. Large unilamellar vesicles (LUV) are a useful tool to study lipid protein interactions. We present here a system to make LUVs of different lipid compositions for protein reconstitution and fusion assays. Reconstitution of integral proteins into LUVs can be achieved by a variety of methods including, organic solvent-mediated reconstitution, mechanical mechanisms, or detergent assisted reconstitution4. We present here a method for reconstituting Drosophila atlastin into preformed liposomes by detergent removal. Advantages of this reconstitution method include high reconstitution yields and proper orientation of atlastin in the lipid bilayer. Additionally, through this method, the protein is not dried or exposed to organic solvents thereby maintaining structure and function. Among its disadvantages, the presence of detergents may not be ideal for all proteins and the final proteoliposomes may have some incorporated detergent in the lipid bilayer. Further dialysis may be used to eliminate more of the detergent. However, dialysis may take a long time and can therefore lead to loss of protein activity.
Assessing atlastin’s fusion activity can be determined by lipid mixing assays as previously described2. Here, we delineate a method for measuring atlastin mediated fusion through N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD)/Lissamine rhodamine-B sulfonyl (rhodamine) labeled lipids. This assay requires fusion of donor (labeled) proteoliposomes and acceptor (unlabeled) proteoliposomes. A FRET release can be measured during the reaction as the dilution of a donor–acceptor pair from “labeled” liposomes to “unlabeled” liposomes as a result of lipid mixing during membrane fusion (Figure 1)5. While this assay serves as a proxy for membrane fusion, it is limited in distinguishing between membrane fusion and hemifusion, a state where only the outer leaflets mix. To address this issue, an alternative is outer leaflet quenching of NBD by dithionite. Following the same methodology as NBD/rhodamine lipid mixing assays, upon quenching the outer leaflet any NBD FRET release by fusion will be due to inner leaflet mixing8.
Alternative fusion assays by inner aqueous content mixing address full fusion only5. Examples of this are terbium (Tb)/dipicolinic acid (DPA) assays and aminonaphthalene trisulfonic acid (ANTS)/p-xylene bis(pyridinium) bromide (DPX) assays. In Tb/DPA assays, a pool of liposomes with encapsulated Tb are mixed and fused with liposomes with encapsulated DPA; upon fusion, fluorescence is increased via internal energy transfer from DPA to Tb within the [Tb(DPA)3]3- chelation complex6. In contrast, for ANTS/DPX assays, ANTS fluorescence is quenched by DPX7. While these systems address inner content mixing, more in-depth preparation of liposomes is required for removal of non-encapsulated reagents, as well as unintended interaction of the fluorophores.
1. Purification of GST-DAtl-His8
2. Reconstitution of Recombinant Atlastin into Liposomes
3. Lipid Mixing Assays
4. Liposome Floatation on Iohexol Discontinuous Gradient12
5. Analysis of Orientation of Reconstituted Protein by Thrombin Proteolysis
NOTE: The atlastin construct reported here has a thrombin cut site between the end of the N-terminal GST tag and the beginning of atlastin. Atlastin in the correct orientation will have this cut site accessible to the protease, while protein in the wrong orientation will be protected by the lipid bilayer.
The efficiency of atlastin reconstitution is presented in Figure 2. Reconstituted proteoliposomes were floated in an iohexol discontinuous gradient. Unincorporated protein was sedimented in the bottom layer (B) or in the middle layer (M). Reconstituted protein would float to the top layer (T). Samples of the gradient were harvested and analyzed by SDS-PAGE and Coomassie staining. The quantification of the gel by densitometry shows a very high efficiency of reconstitution with negligible loses; 96% of the total protein was found as proteoliposomes that floated to the top layer (T). Less than 1% of protein was unreconstituted and found in the middle layer (M), and only 3% was unreconstituted or aggregated and sedimented to the bottom layer (B).
In addition to describing the extent of reconstitution, analyzing the orientation of atlastin after reconstitution was quantified by thrombin cleavage assays14. Reconstituted protein could potentially be in the wrong orientation, that is, facing the lumenal space of the liposome. Protein in the wrong orientation should be protected from proteolysis by the lipid bilayer. A thrombin cleavage site is coded between the end of the N-terminal GST tag and the beginning of atlastin. Floated proteoliposomes were incubated with thrombin for 1 h at 37 °C, followed by a 30 min inactivation with EDTA-free protease inhibitor. Samples were analyzed by SDS-PAGE and Coomassie stain and quantified by densitometry Figure 3. As a negative control, a sample of untreated proteoliposomes is shown in the left lane. Detergent solubilized proteoliposomes (right lane) served as a positive control and to show the extent of thrombin cleavage, with only 1% left un-cleaved. In all, this assay shows that most of the reconstituted protein was cleaved with only 7% being protected from the protease (middle lane). It is worth noting that the remainder of uncut protein may be a result of reconstituted protein aggregates that may not have the cut site accessible. In all, these results describe a robust system for reconstituting atlastin.
The kinetics and the extent of atlastin-mediated proteoliposome fusion were analyzed by lipid mixing assays (Figure 1). A sample of atlastin fusion kinetics and quantification are shown in Figure 4. The full kinetic run is depicted in Figure 4A. A 5 min incubation was done before inducing fusion with GTP at the zero timepoint and after 1 h, n-Dodecyl β-D-maltoside was added to solubilize the proteoliposomes and get the maximum FRET release. The fusion maximum in the run was of 11% of maximum fluorescence (Figure 4B,C). Uninduced (No GTP) controls can be used to determine the background baseline.
Figure 1: Fusion assay model of liposome fusion. A population of labeled liposomes with fluorescently tagged fluorescent donor lipids NBD-phosphatidylethanolamine (represented as green spheres) and the acceptor rhodamine- phosphatidylethanolamine (represented as red spheres) is mixed with an unlabeled pool of liposomes. Before fusion, NBD’s fluorescence is low due to the proximity with the acceptor fluorophore, rhodamine. Upon fusion with unlabeled liposomes a surface area increase leads to dilution of the probes and a FRET release of NBD can be measured. Please click here to view a larger version of this figure.
Figure 2: Reconstitution efficiency analyzed by floatation assays. Reconstitution efficiency of atlastin can be measured by floatation of proteoliposomes in an iohexol discontinuous gradient. Reconstituted protein should float as proteoliposomes to the top layer (T), while unreconstituted and aggregated protein should sediment to the bottom layer (B) or the middle layer (M). A Coomassie-stained SDS-PAGE gel was quantified by densitometry and measured 96% of the total protein floated as proteoliposomes with negligible loses (~4%).
Figure 3: Analysis of reconstituted atlastin in proteoliposomes by protease digestion. Recombinant atlastin has a N-terminal GST tag followed by a thrombin cut sequence. Proteoliposomes of recombinant atlastin were treated with the serine protease thrombin to analyze the orientation of the reconstituted protein in respect to the lipid bilayer. After thrombin treatment the protease is inhibited and the samples are analyzed by SDS-PAGE, Coomassie stain, and quantified by densitometry. The presented proteoliposomes have a low proportion of protected protein, with only 4% that remained undigested (middle lane). As a positive control some proteoliposomes were solubilized with detergent (0.5% TX100) (right lane); the negative control, untreated proteoliposomes, was not cleaved (left lane).
Figure 4: Lipid mixing assays of atlastin proteoliposomes. (A) Sample kinetic trace of fusion reaction with a 5 min incubation time before inducing fusion with GTP at 0 min. After 1 h run, detergent solubilization (black arrow) was used to determine the maximum FRET release of NBD. (B) Zoomed in view of the trace from timepoint 0 to 1 h with and (C) average fusion (n = 3) of 11.3%. Please click here to view a larger version of this figure.
The methods here delineate an efficient method for purifying, reconstituting, and measuring fusion activity of recombinant atlastin. To ensure high yields of functional atlastin some critical steps must be considered. Expression of atlastin must be done at low temperatures (16 °C) to avoid aggregation and one should aim for a final concentration between 0.4–1.5 mg/mL. Very dilute protein will not be reconstituted optimally at a 1:400 protein to lipid ratio. Reconstitution efficiency can be optionally analyzed by floatation assays described here (Figure 2). One of the advantages of liposomes floatation assays is that they can be extended to analyze soluble proteins that associate with lipid bilayers13. Additionally, orientation of the reconstituted protein may be also analyzed by protease digestion14, however, this may not differentiate between reconstituted misfolded or aggregated protein or reconstitution in the wrong orientation (Figure 3). The proteoliposomes developed by this method can be applied to assay for fusion or for studies of atlastin in a model lipid bilayer. While making the reconstitution procedure is relatively simple, small deviations from optimal detergent and lipid concentrations may reduce efficiencies of reconstitution. For example, an excessive amount of detergent may lead to liposome solubilization.
Lipid mixing assays require a pool of donor and acceptor liposomes (Figure 1). This method uses radioactivity by tritiated lipid for quantification of lipid concentration at every step. However, alternative quantification methods have been reported using the intrinsic fluorescence of donor liposomes and acceptor liposomes with dansyl-phosphoethanolamine15,16. Sizing of liposomes can also be modified during extrusion, as the polycarbonate membrane pore size can be adjusted accordingly.
While lipid mixing assays are an efficient way to analyze fusion, it may not differentiate between fusion and hemifusion. Several studies visualizing atlastin proteoliposomes after fusion by electron microscopy15 and by lipid fluorescence16,17 further support full fusion of atlastin. However, when analyzing specific atlastin mutants or other fusion proteins, it is of interest to ensure full fusion. Additional steps to troubleshoot this could be done by quenching the outer leaflet with dithionite and measuring inner leaflet lipid mixing. Alternative fusion assays, such as inner aqueous content mixing may be employed for this, however, additional steps and dialysis of liposomes will be necessary.
It is of interest that this method can be applied to other membrane proteins for in vitro studies of proteins in model lipid bilayers. Other membrane proteins have been reported to reconstitute with this method16,17. This method can therefore be applied to a variety of membrane and fusion proteins.
The authors have nothing to disclose.
We thank Dr. Michael Stern and his lab for their insights and feedback on atlastin related projects. This work was supported by the National Institute of General Medical Sciences [R01GM101377] and the National Institute of Neurological Disorders and Stroke [R01NS102676].
10 mL poly-prep chromatography columns | Biored | 731-1550 | |
10 x 75 mm Flint glass tubes | VWR | 608225-402 | |
47 mm diameter, 0.45um pore whatman sterile membrane filters | Whatman | 7141 104 | |
96 well white plate | NUNC | 437796 | |
Anapoe X-100 | Anatrace | 9002-931-1 | |
Cell disrupter | Avestin | Avestin Emulsiflex C3 | |
DOPS (1,2-dioleoyl-sn-glycero-3-phospho-L-serine (sodium salt)) | Avanti | 840035P-10mg | DOPS |
EDTA | Research organics inc. | 6381-92-6 | Ethylenediaminetetraacetic acid |
EDTA-free protease inhibitor cocktail | Roche | 11873580001 | Complete protease inhibitor |
Extruder | Sigma Aldrich | Z373400 | Liposofast Basic Extruder |
GE Akta Prime liquid chromatography system | GE Pharmacia | 8149-30-0006 | |
Glutathione agarose beads | Sigma aldrich | G4510-50ml | |
Glycerol | EMD | GX0185-5 | |
GTP | Sigma Aldrich | 36051-31-7 | Guanosine 5' triphosphate sodium salt hydrate |
HEPES, acid free | Omnipur | 5330 | |
Imidazole | fluka | 5670 | |
Immobilized metal affinity chromatography (IMAC) resin column | GE Healthcare | 17040801 | 1 mL HiTrap Chelating HP immobilized metal affinity chromatography columns |
Iohexol | Accurate chemical and scientific corporation | AN 7050 BLK | Accudenz/Nycodenz |
IPTG | Research products international corp. | I56000-100.0 | IPTG, dioxane free |
L-Glutathione reduced | Sigma-Aldrich | G4251-5g | |
Magnesium chloride | Fisher | 7791-18-6 | |
Methanol | Omnisolv | MX0488-1 | |
NBD-DPPE (1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(7-nitro-2-1,3-benzoxadiazol-4-yl) (ammonium salt)) | Avanti | 236495 | NBD-DPPE |
n-Dodecyl β-D-maltoside | Chem-Impex International | 21950 | |
Nonpolar polystyrene adsorbent beads | BioRad | 152-3920 | SM2 Biobeads |
Nuclepore track-etch polycarbonate 19 mm 0.1 um pore membrane | Whatman | 800309 | |
Optima LE80K Ultra centrifuge | Beckman Coulter | ||
Phosphatidylcholine, L-α-dipalmitoyl [choline methyl-3H] | ARC | ART0284 | Titriated lipids |
Plate reader | TECAN | TECAN infinite M200 plate reader | |
POPC (1-palmitoyl-2-oleoyl-glycero-3-phosphocholine) | Avanti | 850457C-25mg | POPC |
Potassium chloride | MP | 151944 | |
Rh-DPPE (1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (ammonium salt)) | Avanti | 236495 | Rh-DPPE |
Scintillation Cocktail | National Diagnostics | LS-272 | Ecoscint XR Scintillation solution for aqueous or non-aqueous samples |
Scintillation vials | Beckman | 592928 | Fast turn cap Mini Poly-Q Vial |
Thrombin | Sigma | T1063-1kU | Thrombin from human plasma |
Triton X-100 | Fisher | BP151-500 | |
Ultra-clear centrifuge tubes 5 x 41 mm | Beckman | 344090 | |
Vortex 9 to 13mm Tube Holder | VWR | 58816-138 | Insert for vortexing flint glass tubes |
Vortex Insert Retainer | VWR | 58816-132 | Retainer needed for vortex tube holder |
Vortexer | VWR | 2.235074 | Vortex Genie 2 model G560 |
β-mercaptoethanol molecular biology grade | Calbiochem | 444203 |