Here, we describe protocols using fluorescent lipid sensors and liposomes to determine whether a protein extracts and transports phosphatidylserine or phosphatidylinositol 4-phosphate in vitro.
Several members of the evolutionarily conserved oxysterol-binding protein (OSBP)-related proteins(ORP)/OSBP homologs (Osh) family have recently been found to represent a novel lipid transfer protein (LTP) group in yeast and human cells. They transfer phosphatidylserine (PS) from the endoplasmic reticulum (ER) to the plasma membrane (PM) via PS/phosphatidylinositol 4-phosphate (PI(4)P) exchange cycles. This finding allows a better understanding of how PS, which is critical for signaling processes, is distributed throughout the cell and the investigation of the link between this process and phosphoinositide (PIP) metabolism. The development of new fluorescence-based protocols has been instrumental in the discovery and characterization of this new cellular mechanism in vitro at the molecular level. This paper describes the production and the use of two fluorescently labelled lipid sensors, NBD-C2Lact and NBD-PHFAPP, to measure the ability of a protein to extract PS or PI(4)P and to transfer these lipids between artificial membranes. First, the protocol describes how to produce, label, and obtain high-purity samples of these two constructs. Secondly, this paper explains how to use these sensors with a fluorescence microplate reader to determine whether a protein can extract PS or PI(4)P from liposomes, using Osh6p as a case study. Finally, this protocol shows how to accurately measure the kinetics of PS/PI(4)P exchange between liposomes of defined lipid composition and to determine lipid transfer rates by fluorescence resonance energy transfer (FRET) using a standard fluorometer.
The precise distribution of lipids between different membranes and within the membranes of eukaryotic cells1,2 has profound biological implications. Decrypting how LTPs function is an important issue in cell biology3,4,5,6, and in vitro approaches are of great value in addressing this issue7,8,9,10,11. Here, an in vitro, fluorescence-based strategy is presented that has been instrumental in establishing that several ORP/Osh proteins effect PS/PI(4)P exchange between cell membranes12 and thereby constitute a new class of LTPs. PS is an anionic glycerophospholipid that represents 2-10% of total membrane lipids in eukaryotic cells13,14,16. It is distributed along a gradient between the ER and the PM, where it represents 5-7% and up to 30% of glycerophospholipids, respectively17,18,19. Moreover, PS is essentially concentrated in the cytosolic leaflet of the PM. This build-up and the uneven partition of PS in the PM are critical for cellular signaling processes19. Owing to the negative charge of PS molecules, the cytosolic leaflet of the PM is much more anionic than the cytosolic leaflet of other organelles1,2,19,20. This enables the recruitment, via electrostatic forces, of signaling proteins such as myristoylated alanine-rich C-kinase substrate (MARCKS)21, sarcoma (Src)22, Kirsten-rat sarcoma viral oncogene (K-Ras)23, and Ras-related C3 botulinum toxin substrate 1 (Rac1)24 that contain a stretch of positively charged amino acids and a lipidic tail.
PS is also recognized by conventional protein kinase C in a stereoselective manner via a C2 domain25. However, PS is synthesized in the ER26, indicating that it must be exported to the PM before it can play its role. It was not known how this was accomplished19 until the finding that, in yeast, Osh6p and Osh7p transfer PS from the ER to the PM27. These LTPs belong to an evolutionarily conserved family in eukaryotes whose founding member is OSBP and that contains proteins (ORPs in human, Osh proteins in yeast) integrating an OSBP-related domain (ORD) with a pocket to host a lipid molecule. Osh6p and Osh7p consist only of an ORD whose structural features are adapted to specifically bind PS and transfer it between membranes. Nevertheless, how these proteins directionally transferred PS from the ER to the PM was unclear. Osh6p and Osh7p can trap PI(4)P as an alternative lipid ligand12. In yeast, PI(4)P is synthesized from phosphatidylinositol (PI) in the Golgi and the PM by PI 4-kinases, Pik1p and Stt4p, respectively. In contrast, there is no PI(4)P in the ER membrane, as this lipid is hydrolyzed to PI by the Sac1p phosphatase. Hence, a PI(4)P gradient exists at both the ER/Golgi and ER/PM interfaces. Osh6p and Osh7p transfer PS from the ER to the PM via PS/PI(4)P exchange cycles using the PI(4)P gradient that exists between these two membranes12.
Within one cycle, Osh6p extracts PS from the ER, exchanges PS for PI(4)P at the PM and transfers PI(4)P back to the ER to extract another PS molecule. Osh6p/Osh7p interact with Ist2p28, one of the few proteins that connect and bring the ER membrane and the PM into close proximity with each other to create ER-PM contact sites in yeast29,30,31. In addition, the association of Osh6p with negatively charged membranes becomes weak as soon as the protein extracts one of its lipid ligands due to a conformational change that modifies its electrostatic features32. This aids Osh6p by shortening its membrane dwell time, thereby maintaining the efficiency of its lipid transfer activity. Combined with the binding to Ist2p, this mechanism could allow Osh6p/7p to both quickly and accurately execute lipid exchange at the ER/PM interface. In human cells, ORP5 and ORP8 proteins execute PS/PI(4)P exchange at ER-PM contact sites via distinct mechanisms33. They have a central ORD, akin to Osh6p, but are directly anchored to the ER via a C-terminal transmembrane segment33 and dock into the PM via an N-terminal Pleckstrin homology (PH) domain that recognizes PI(4)P and PI(4,5)P233,34,35. ORP5/8 use PI(4)P to transfer PS, and it has been shown that ORP5/8 additionally regulate PM PI(4,5)P2 levels and presumably modulate signaling pathways. In turn, a decrease in PI(4)P and PI(4,5)P2 levels lowers ORP5/ORP8 activity as these proteins associate with the PM in a PIP-dependent fashion. Abnormally high PS synthesis, which leads to Lenz-Majewski syndrome, impacts PI(4)P levels through ORP5/836. When the activity of both proteins is blocked, PS becomes less abundant at the PM, lowering the oncogenic capability of signaling proteins37.
Conversely, ORP5 overexpression seems to promote cancer cell invasion and metastatic processes38. Thus, alterations to ORP5/8 activity can severely modify cellular behavior through changes in lipid homeostasis. Further, ORP5 and ORP8 occupy ER-mitochondria contact sites and preserve some mitochondrial functions, possibly by supplying PS39. Additionally, ORP5 localizes to ER-lipid droplet contact sites to deliver PS to lipid droplets by PS/PI(4)P exchange40. The strategy described herein to measure (i) PS and PI(4)P extraction from liposomes and (ii) PS and PI(4)P transport between liposomes has been devised to establish and analyze the PS/PI(4)P exchange activity of Osh6p/Osh7p12,32 and used by other groups to analyze the activity of ORP5/ORP835 and other LTPs10,41. It is based on the use of a fluorescence plate reader, a standard L-format spectrofluorometer, and two fluorescent sensors, NBD-C2Lact and NBD-PHFAPP, that can detect PS and PI(4)P, respectively.
NBD-C2Lact corresponds to the C2 domain of the glycoprotein, lactadherin, that was reengineered to include a unique solvent-exposed cysteine near the presumed PS binding site; a polarity-sensitive NBD (7-nitrobenz-2-oxa-1,3-diazol) fluorophore is covalently linked to this residue (Figure 1A)12. To be more precise, the C2 domain of lactadherin (Bos taurus, UniProt: Q95114,residues 270-427) was cloned into a pGEX-4T3 vector to be expressed in fusion with glutathione S-transferase (GST) in Escherichia coli. The C2Lact sequence was then mutated to substitute two solvent-accessible cysteine residues (C270, C427) with alanine residues and to introduce a cysteine residue into a region near the putative PS-binding site (H352C mutation) that can be subsequently labeled with N,N'-dimethyl-N-(iodoacetyl)-N'-(7-nitrobenz-2-oxa-1,3-diazol-4-yl) ethylene diamine (IANBD) 12. A cleavage site for thrombin is present between the GST protein and the N-terminus of the C2 domain. A major advantage is that this domain selectively recognizes PS in a Ca2+-independent manner contrary to other known C2 domains or Annexin A542. NBD-PHFAPP is derived from the PH domain of the human four-phosphate-adaptor protein 1 (FAPP1), which was reengineered to include a single solvent-exposed cysteine that can be labeled with an NBD group near the PI(4)P binding site (Figure 1A)43. The nucleotide sequence of the PH domain of the human FAPP protein (UniProt: Q9HB20, segment [1-100]) has been cloned into a pGEX-4T3 vector to be expressed in tandem with a GST tag. The PHFAPP sequence has been modified to insert a unique cysteine residue within the membrane-binding interface of the protein43. Moreover, a nine-residue linker has been introduced between the thrombin cleavage site and the N-terminus of the PH domain to ensure accessibility to the protease.
To measure PS extraction from liposomes, NBD-C2Lact is mixed with liposomes made of phosphatidylcholine (PC) containing trace amounts of PS. Owing to its affinity for PS, this construct binds to the liposomes, and the NBD fluorophore experiences a change in polarity as it comes into contact with the hydrophobic environment of the membrane, which elicits a blue-shift and an increase in fluorescence. If PS is extracted almost completely by a stoichiometric amount of LTP, the probe does not associate with liposomes, and the NBD signal is lower (Figure 1B)32. This difference in signal is used to determine whether an LTP (e.g., Osh6p) extracts PS. A similar strategy is used with NBD-PHFAPP to measure PI(4)P extraction (Figure 1B), as described previously12,32. Two FRET-based assays were designed to (i) measure PS transport from LA to LB liposomes, which mimic the ER membrane and the PM, respectively, and (ii) PI(4)P transport in the reverse direction. These assays are performed under the same conditions (i.e., same buffer, temperature, and lipid concentration) to measure PS/PI(4)P exchange. To measure PS transport, NBD-C2Lact is mixed with LA liposomes composed of PC and doped with 5 mol% PS and 2 mol% of a fluorescent rhodamine-labelled phosphatidylethanolamine (Rhod-PE)-and LB liposomes incorporating 5 mol% PI(4)P.
At time zero, FRET with Rhod-PE quenches the NBD fluorescence. If PS is transported from LA to LB liposomes (e.g., upon injecting Osh6p), a fast dequenching occurs due to the translocation of NBD-C2Lact molecules from LA to LB liposomes (Figure 1C). Given the amount of accessible PS, NBD-C2Lact remains essentially in a membrane-bound state over the course of the experiment12. Thus, the intensity of the NBD signal directly correlates with the distribution of NBD-C2Lact between LA and LB liposomes and can be easily normalized to determine how much PS is transferred. To measure the transfer of PI(4)P in the opposite direction, NBD-PHFAPP is mixed with LA and LB liposomes; given that it only binds to LB liposomes that contain PI(4)P, but not Rhod-PE, its fluorescence is high. If PI(4)P is transferred to LA liposomes, it translocates to these liposomes, and the signal decreases due to FRET with Rhod-PE (Figure 1C). The signal is normalized to determine how much PI(4)P is transferred43.
1. Purification of NBD-C2Lact
NOTE: Although this protocol details the use of a cell disruptor to break bacteria, it can be modified to use other lysis strategies (e.g., a French press). At the beginning of the purification, it is mandatory to use buffer that is freshly degassed, filtered, and supplemented with 2 mM dithiothreitol (DTT) to prevent the oxidation of cysteine. However, for the protein labelling step, it is crucial to completely remove DTT. Many steps must be carried out on ice or in a cold room to avoid any protein degradation. Samples of 30 µL volume must be collected at different steps of the protocol to perform an analysis by sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using a 15% acrylamide gel to check the progress of the purification. Mix enough denaturing Laemmli sample buffer with each aliquot, and heat the mixture at 95 °C. Freeze and store the tubes at -20 °C until analysis.
2. Purification of NBD-PHFAPP
NOTE: The procedure to produce and label PHFAPP is identical to that of NBD-C2Lact until the transfer of NBD-C2Lact solution to a centrifugal filter unit in step 1.3.4. From this step onwards, follow the protocol that is described below.
3. Preparation of liposomes for PS and PI(4)P extraction or transfer assays
NOTE: Perform all the steps at room temperature unless otherwise specified. Handle organic solvents, rotavapor, and liquid nitrogen with caution.
Lipid composition (mol/mol) | Lipid | |||||
Liposome name | DOPC (25 mg/mL) |
POPS (10 mg/mL) |
16:0 Liss Rhod-PE (1 mg/mL) |
C16:0/C16:0-PI(4)P (1 mg/mL) |
||
Extraction assays | Liposome 2 mol% PS | PC/PS 98/2 | 247 µL | 12.5 µL | ||
Liposome 2 mol% PI(4)P | PC/PI(4)P 98/2 | 247 µL | 153 µL | |||
PC liposome | PC 100 | 252 µL | ||||
Transport assays | LA | PC/PS/Rhod-PE 93/5/2 | 234 µL | 31.4 µL | 200 µL | |
LA without PS | PC/Rhod-PE 98/2 | 247 µL | 200 µL | |||
LB | PC/PI(4)P 95/5 | 237 µL | 383 µL | |||
LB without PI(4)P | PC 100 | 252 µL | ||||
LA-Eq | PC/PS/PI(4)P/Rhod-PE 93/2.5/2.5/2 | 234 µL | 15.7 µL | 200 µL | 191 µL | |
LB-Eq | PC/PS/PI(4)P 95/2.5/2.5 | 239 µL | 15.7 µL | 191 µL |
Table 1: Volumes of lipid stock solutions to be mixed for liposome preparation. Abbreviations: PS= phosphatidylserine; PC = phosphatidylcholine; PI(4)P = phosphatidylinositol 4-phosphate; Rhod-PE = rhodamine-labelled phosphatidylethanolamine; DOPC = dioleoylphosphatidylcholine; POPS= 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine; 16:0 Liss Rhod-PE = 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl).
4. Measurement of PS or PI(4)P extraction
NOTE: Measurements must be conducted using a black 96-well plate and a fluorescence plate reader equipped with monochromators: one for fluorescence excitation and one for emission, with a variable bandwidth.
5. Real-time measurement of PS transport
NOTE: A standard fluorimeter (90° format) equipped with a temperature-controlled cell holder and a magnetic stirrer is used to record lipid transfer kinetics. To accurately acquire data, it is key to permanently maintain the sample at the same temperature (set between 25 and 37 °C depending on the origin of the protein (e.g., yeast or human)) and to constantly stir it. The protocol described below is for the measurement of lipid transport in a 600 µL sample contained in a cylindrical quartz cell.
6. Real-time measurement of PI(4)P transport
7. Analysis of kinetics curves
Figure 1: Description of the fluorescent lipid sensors and in vitro assays. (A) Three-dimensional models of NBD-C2Lact and NBD-PHFAPP based on the crystal structure of the C2 domain of bovine lactadherin (PDB ID: 3BN648) and the NMR structure of the PH domain of the human FAPP1 protein (PDB ID: 2KCJ46). An N,N'-dimethyl-N-(thioacetyl)-N'-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)ethylenediamine moiety, built manually and energetically minimized, was grafted onto the thiol function of C352 (NBD-C2Lact) and C13 (NBD-PHFAPP) residues (represented as spheres, with carbon in green, nitrogen in blue, oxygen in red, sulfur in yellow, and hydrogen in white). The surface of the lipid-binding site of each probe was colored in blue. (B) Extraction assays. In the PS extraction assay, PC/PS liposomes (98/2 mol/mol) were incubated with 250 nM NBD-C2Lact. In the absence of extraction, the probe strongly bound to the liposomes, resulting in a blue shift of NBD fluorescence and an increase of its emission intensity. If PS extraction occurred in the presence of an LTP (e.g., Osh6p), the probe dissociated from the liposomes, and its fluorescence was lower. In the PI(4)P extraction assay, liposomes were doped with 2 mol% PI(4)P, and 250 nM NBD-PHFAPP was used. (C) FRET-based lipid transport assays. In the PS transport assay, PC/PS/Rhod-PE liposomes (93/5/2 mol/mol, LA) were incubated with 250 nM NBD-C2Lact. PC liposomes (LB), doped or not with 5 mol% PI(4)P, and Osh6p were sequentially added at t = 1 min and t = 4 min, respectively. If PS transport occurs, this elicits a dequenching of the NBD signal corresponding to the translocation of NBD-C2Lact from LA to LB liposomes. In the PI(4)P transport assay, LB liposomes doped with 5 mol% PI(4)P were incubated with 250 nM NBD-PHFAPP. PC liposomes (LA) doped or not with 5 mol% PS were added. If PI(4)P transport occurs, this causes a quenching of the NBD signal due to the translocation of NBD-PHFAPP from LB to LA liposomes. Abbreviations: NBD = 7-nitrobenz-2-oxa-1,3-diazol fluorophore; NMR = nuclear magnetic resonance; FAPP1 = four-phosphate-adaptor protein 1; PDB = Protein Data Bank; PS= phosphatidylserine; PC = phosphatidylcholine; LTP = lipid transfer protein; Osh6p = oxysterol-binding protein (OSBP) homolog 6 protein; PI(4)P = phosphatidylinositol 4-phosphate; FRET = fluorescence resonance energy transfer; Rhod-PE = rhodamine-labelled phosphatidylethanolamine; LA liposomes = liposomes composed of PC, doped with 2 mol% Rhod-PE, and containing or not 5 mol% PS; LB liposomes = liposomes incorporating 5 mol% PI(4)P. Please click here to view a larger version of this figure.
Figure 2A shows an SDS-PAGE analysis of the products of different steps leading to the purification of C2Lact. Lane 1 shows the protein profile of the lysed bacteria expressing GST-C2Lact (~44.8 kDa), whereas lanes 2 and 3 respectively show the protein profiles of the supernatant and bacterial debris after ultracentrifugation. The comparison of these lanes shows that GST-C2Lact has been recovered in the supernatant and thus can be isolated using glutathione-linked agarose beads. Lanes 4 and 5 show the protein profiles of the supernatant after incubation with beads and washes recovered by gravity flow, whereas lane 6 shows the profile of proteins that have been retained on the beads. An analysis of these lanes indicates that almost all GST-C2Lact has been recovered from the beads.
Lanes 8-12 show the presence of a major band corresponding to C2Lact (~17.9 kDa) in the supernatants recovered through successive washes of the beads after thrombin treatment. Lane 13 indicates that non-cleaved GST-C2Lact, along with GST (~26.9 kDa), remained bound to beads after this treatment. The comparison of these lanes indicates that the cleavage procedure, albeit not 100% efficient, did yield C2Lact that was then fluorescently labelled. Figure 2B shows an ultraviolet (UV)-visible absorbance spectrum of C2Lact labelled with NBD. As the construct is 100% pure, these results confirm that all C2Lact molecules were labelled with an NBD group based on the optical density measured at 280 nm (Trp) and 495 nm (NBD). The purity of NBD-C2Lact and its fluorescence were determined by SDS-PAGE analysis (Figure 2C).
Figure 2: NBD-C2Lact purification. (A) SDS-PAGE analysis was used to check the presence of the protein at different steps of the purification procedure before labeling. The arrowheads indicate the position of the C2Lact domain (red arrowhead), of the GST alone (grey arrowhead), and of the GST-C2Lact construct (black arrowhead). (B) UV-visible absorbance spectrum of NBD-C2Lact. (C) SDS-PAGE analysis of the purified NBD-C2Lact. The first image was acquired under UV illumination without staining and reveals the presence of the NBD-C2Lact construct as it emits fluorescence. The second image shows the same gel after a protein staining procedure (see the Table of Materials). Abbreviations: NBD = 7-nitrobenz-2-oxa-1,3-diazol fluorophore; NBD-C2Lact = N,N'-dimethyl-N-(thioacetyl)-N'-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)ethylenediamine moiety linked to the thiol function of C352 residue of a reengineered version of the C2 domain of bovine lactadherin (PDB ID: 3BN648); PDB = Protein Data Bank; SDS-PAGE = sodium dodecylsulfate polyacrylamide gel electrophoresis; GST = glutathione S-transferase; MW = molecular-weight size marker; UV = ultraviolet. Please click here to view a larger version of this figure.
Figure 3 shows the results from PS and PI(4)P extraction assays using Osh6p. When only incubated with liposomes containing 2 mol% PS, the fluorescence of NBD-C2Lact was maximal as the NBD fluorophore was inserted in the membrane (i.e., a hydrophobic context), indicating that the sensor was membrane-bound. In the presence of Osh6p, the fluorescence was lower and comparable to that measured with pure PC liposomes (Figure 3A). The normalization of intensity values at 536 nm indicated that ~75% of accessible PS was extracted. In the second assay, NBD-PHFAPP was mixed with liposomes containing 2 mol% PI(4)P. The NBD signal was high without Osh6p, but low when Osh6p was present and was similar to that measured with PI(4)P-free liposomes (Figure 3B). An analysis of the intensity revealed that ~100 % of accessible PI(4)P was extracted by the LTP.
Figure 3: Extraction assays. (A) Fluorescence spectra of NBD-C2Lact (250 nM) measured upon excitation at 460 nm in the presence of liposomes (80 µM, 2 mol% PS) in the absence or presence of 3 µM Osh6p. Reference spectra were recorded with pure PC liposomes incubated or not with NBD-C2Lact (left panel). Several spectra were recorded from different series of wells, corrected by subtracting the background scattering signal from DOPC liposomes alone and averaged (n=4, ± SEM). The percentage of accessible PS that was extracted is indicated (right panel). (B) Fluorescence spectra of NBD-PHFAPP (250 nM) mixed with liposomes (80 µM, 2 mol% PI(4)P) in the absence or presence of 3 µM Osh6p. Reference spectra recorded with PC liposomes in the presence and absence of the sensor are shown (left panel). Several spectra were recorded from different series of wells, corrected by subtracting the background scattering signal from DOPC liposomes alone and averaged (n=4, ± SEM). The percentage of accessible PI(4)P that was extracted is indicated (right panel). Abbreviations: NBD = 7-nitrobenz-2-oxa-1,3-diazol fluorophore; NBD-C2Lact = N,N'-dimethyl-N-(thioacetyl)-N'-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)ethylenediamine moiety linked to the thiol function of C352 residue of a reengineered version of the C2 domain of bovine lactadherin (PDB ID: 3BN648); PDB = Protein Data Bank; PS= phosphatidylserine; PC = phosphatidylcholine; DOPC = dioleoylphosphatidylcholine; Osh6p = oxysterol-binding protein (OSBP) homolog 6 protein; PI(4)P = phosphatidylinositol 4-phosphate; SEM = standard error of the mean; a.u. = arbitrary units. Please click here to view a larger version of this figure.
Figure 4A shows typical results from a PS transfer assay using Osh6p as an LTP. At time zero, NBD-C2Lact was mixed with LA liposomes containing 5 mol% 16:0/18:1-PS (POPS) and 2 mol% Rhod-PE in a volume of 570 µL of HKM buffer at 30 °C. As the probe bound to LA liposomes, its signal was quenched due to FRET with Rhod-PE present in these liposomes. After one minute, Rhod-PE-free LB liposomes (30 µL) were added; this was expected to only elicit a slight change in the signal due to light diffusion by this second liposome population and/or a dilution effect. The signal intensity after the addition of the LB liposomes corresponds to F0. The injection of a few µL of a stock solution of Osh6p (typically 40 µM) to dilute 200 nM of the protein in the reaction mix elicited a slow increase in the NBD signal due to the dequenching of the fluorophore as PS was transported from LA to LB liposomes, thereby promoting the translocation of NBD-C2Lact.
When LB liposomes contained 5 mol% PI(4)P, the dequenching was much faster, as PS was transferred more rapidly by Osh6p to LB liposomes due to the counterexchange of PS with PI(4)P by the LTP (second curve). The third curve corresponds to an experiment in which NBD-C2Lact was mixed with equal amounts of LA-Eq and LB-Eq liposomes. The signal was higher than F0 and corresponded to a situation where the probe was evenly bound to LA and LB liposomes, thus reflecting a situation where PS was fully equilibrated between the two populations of liposomes. FEq was calculated by averaging the value of the signal measured in the last 5 min of the experiment.
Figure 4: Typical PS transport kinetics measured with Osh6p and reference curve. (A) LB liposomes devoid of PI(4)P (200 µM total lipids) and Osh6p (200 nM) were sequentially added to a cuvette containing NBD-C2Lact (250 nM) and LA liposomes (200 µM) doped with 5 mol% PS and 2 mol% Rhod-PE (left curve). The same experiment was performed with LB liposomes doped with 5 mol% PI(4)P (middle curve). To determine FEq, NBD-C2Lact (250 nM) was premixed with LA-Eq liposomes; then, LB-Eq liposomes were added (right curve). (B) Averaged PS transport curves determined after the normalization of several measurements done using LB liposomes with or without PI(4)P (mean ± SEM, n=3). (C) Initial PS transfer rates (mean ± SEM, n=3). Abbreviations: NBD = 7-nitrobenz-2-oxa-1,3-diazol fluorophore; NBD-C2Lact = N,N'-dimethyl-N-(thioacetyl)-N'-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)ethylenediamine moiety linked to the thiol function of C352 residue of a reengineered version of the C2 domain of bovine lactadherin (PDB ID: 3BN648); PDB = Protein Data Bank; PS= phosphatidylserine; Osh6p = oxysterol-binding protein (OSBP) homolog 6 protein; PI(4)P = phosphatidylinositol 4-phosphate; Rhod-PE = rhodamine-labelled phosphatidylethanolamine; LA liposomes = liposomes composed of phosphatidylcholine, doped with 2 mol% Rhod-PE, and containing or not 5 mol% PS; LB liposomes = liposomes incorporating 5 mol% PI(4)P; F = fluorescence; F0 = fluorescence corresponding to NBD before addition of Osh6p; FEq = fluorescence signal if PS is fully equilibrated between LA and LB liposomes by a transfer process; SEM = standard error of the mean. Please click here to view a larger version of this figure.
Figure 4B shows average kinetic curves of PS transfer from LA liposomes to LB liposomes, doped or not doped with PI(4)P after the normalization of F data using F0 and FEq as reference values. The initial transport rate for each experiment was calculated by fitting the initial data points measured after the injection of the protein with a linear function. Figure 4C shows the mean initial PS transport rate determined from three distinct experiments using LB liposomes with or without PI(4)P. When LB liposomes contained 0 and 5 mol% PI(4)P, the rates were respectively equal to 1.4 and 15.8 PS.min-1 per Osh6p molecule. Figure 5A shows typical results from a PI(4)P transfer assay using Osh6p as an LTP, which was carried out with the same materials and conditions as those used for the PS transport assay.
Figure 5: Typical PI(4)P transport kinetics measured with Osh6p and reference curve. (A) PS-free liposomes containing 2 mol% Rhod-PE (LA, 200 µM total lipids) and Osh6p (200 nM) were sequentially added to a cuvette containing NBD-PHFAPP (250 nM) and LB liposomes (200 µM) doped with 5 mol% PI(4)P (left curve). The same experiment was performed with LA liposomes doped with 5 mol% PS (middle curve). To determine FEq, NBD-PHFAPP (250 nM) was premixed with LB-Eq liposomes; then, LA-Eq liposomes were added (right curve). (B) PI(4)P transfer kinetics determined after normalization of several measurements of LA liposomes with or without PS (mean ± SEM, n=3). (C) Initial PI(4)P transfer rates (mean ± SEM, n=3). Abbreviations: NBD = 7-nitrobenz-2-oxa-1,3-diazol fluorophore; NBD-PHFAPP = NBD-labeled Pleckstrin homology domain of the human four-phosphate-adaptor protein 1 (FAPP1, UniProt: Q9HB20, segment [1-100]); PS= phosphatidylserine; Osh6p = oxysterol-binding protein (OSBP) homolog 6 protein; PI(4)P = phosphatidylinositol 4-phosphate; Rhod-PE = rhodamine-labelled phosphatidylethanolamine; LA liposomes = liposomes composed of phosphatidylcholine, doped with 2 mol% Rhod-PE, and containing or not 5 mol% PS; LB liposomes = liposomes incorporating 5 mol% PI(4)P; F = fluorescence; F0 = fluorescence corresponding to NBD before addition of Osh6p; FEq = fluorescence signal if PI(4)P is fully equilibrated between LA and LB liposomes by a transfer process; SEM = standard error of the mean. Please click here to view a larger version of this figure.
At time zero, NBD-PHFAPP was mixed with LB liposomes containing 5 mol% PI(4)P in a volume of 570 µL of HKM buffer. Because NBD-PHFAPP was bound to LB liposomes, its signal was high. After one minute, LA liposomes (30 µL) were added, which was expected to only elicit a slight change in the signal. The intensity then corresponded to F0. Injecting Osh6p (200 nM final concentration) into the reaction mix triggered a quenching of the NBD signal due to the translocation of NBD-PHFAPP molecules to LA liposomes as PI(4)P was transferred from LB liposomes to LA liposomes. When LA liposomes contained 5 mol% POPS, the dequenching was much faster owing to a faster PI(4)P transfer resulting from PS/PI(4)P exchange. The third curve corresponds to an experiment in which NBD-PHFAPP was mixed with equal amounts of LA-Eq and LB-Eq liposomes.
The signal was lower than F0 as it corresponded to a situation where the probe was uniformly bound to LA and LB liposomes, thus indicative of a full equilibration of PI(4)P between the liposomes. FEq was calculated by averaging the value of the signal measured in the last 5 min of the experiment. Figure 5B,C show averaged kinetics curves obtained after signal normalization and mean PI(4)P initial transfer rates measured with LA liposomes that were doped or not doped with 5 mol% PS. Here, it is worth noting that both lipid ligands were transported faster and with similar velocities when each lipid was initially present in LA and LB membranes, respectively, which indicates that Osh6p is a PS/PI(4)P exchanger.
The outcomes of these assays directly rely on the signals of the fluorescent lipid sensors. Thus, the purification of these probes labelled at a 1:1 ratio with NBD and without free NBD fluorophore contamination is a critical step in this protocol. It is also mandatory to check whether the LTP under examination is properly folded and not aggregated. The amount of LTP tested in the extraction assays must be equal to or higher than that of accessible PS or PI(4)P molecules to properly measure whether this LTP efficiently extracts these lipids. Indeed, NBD-C2Lact and NBD-PHFAPP bind to PS and PI(4)P, respectively, following a classical saturation binding curve. Given their respective affinities for PS and PI(4)P, these probes remain largely bound to the liposomes even if the liposomes contain residual traces of these ligands. This can lead to the erroneous conclusion that an LTP does not efficiently extract these lipids. The protocol relies on standard and commercially available PC, PS, and PI(4)P subspecies that are 18:1/18:1-PC, 16:0/18:1-PS, and 16:0/16:0-PI(4)P, respectively. Performing experiments using lipid species with other acyl chains can give different results, as previously reported12,35.
In the transfer assays, if the bulk lipid composition (not considering PS or PI(4)P) of LA and LB liposomes has to be modified, it is key to also perform the control experiments using LA-Eq and LB-Eq liposomes with a bulk composition similar to that of LA and LB liposomes, respectively. The same principle applies for the extraction assays. Genetically encoded fluorescent lipid-binding domains are broadly used to analyze lipid distribution inside cells44. Caution is recommended when analyzing such experiments as the association of these lipid probes with the membrane, although primarily driven by the presence of the targeted lipid, can be influenced by other parameters (density of anionic lipids, lipid-packing) that differ between cell compartments. In these in vitro assays, the composition of liposomes is much simpler than that of cellular membranes: they are mostly made of PC, a zwitterionic lipid, and thereby expose a rather inert surface that does not impact how PS and PI(4)P are recognized by their corresponding sensor. PHFAPP can be used in the presence of anionic lipids, such as PS, phosphatidic acid (PA), and PI32, and C2Lact does not recognize PI(4)P12; both these observations allow unbiased measurements of PS/PI(4)P exchange. NBD-PHFAPP is not influenced by sterol43.
However, it is not known whether these probes, in particular, NBD-C2Lact, can be used with liposomes with extreme features (e.g., very low or high lipid-packing, highly negatively charged surface, presence of lipid domains). Despite this potential limitation with respect to liposome composition, these transfer assays based on fluorescent sensors have tremendous advantages compared to other methods. First, they do not rely on fluorescently labeled PS and PI(4)P that bear an extra bulky group and are likely not properly accommodated by the binding pocket of LTPs. Second, these assays offer far better time resolution than methods based on liposome separation (radioactivity-based assays45 or mass spectrometry33), and it should be noted that radiolabeled PI(4)P and PS are not commercially available. The capacity of a protein to transfer PS or PI(4)P is not impacted by the association of the fluorescent sensors with liposomes. Indeed, if the amount of NBD-C2Lact or NBD-PHFAPP and that of PS and PI(4)P in the outer leaflet of liposomes that is accessible to the LTP is considered, only 5% of PS or PI(4)P are associated with probe during a kinetics measurement.
Moreover, only 0.55-0.86% of the membrane surface is covered by the probe, considering the total surface of liposomes (LA+LB liposomes; 5.1 × 1016 nm2 with an area of 0.7 nm2 per lipid), the membrane surface that one individual C2 or PH molecule can occupy (≈3.1 and 4.8 nm2, respectively, estimated from references46,47,48) and their concentration (250 nM). Thus, the principles underlying these assays are adapted to properly analyze the kinetic and mechanistic aspects of LTPs. As mentioned in the introduction, alterations in the activity of ORP5/8 can lead to cellular dysfunctions49. For example, the invasiveness of pancreatic cancer cells seems to rely on the level of ORP5 expression. Moreover, a causality exists between high levels of ORP5 expression and the poor prognosis of human pancreatic cancer. A high level of ORP5 expression is also detected in lung tumor tissues, and more particularly, in metastatic cases. Interestingly, the ability of ORP5 to transfer lipids might explain why it promotes cell proliferation and migration.
ORP5 upregulates the mammalian target of rapamycin complex 1 (mTORC1) complex, which plays a central role in cell survival and proliferation, probably because it enhances the activity of Akt, an upstream activator of mTORC1, by supplying the PM with PS. Overall, these observations suggest that ORP5 might be an interesting pharmacological target, and that these in vitro assays might serve to screen molecules that are able to inhibit its activity. This assay can also help to better define whether other members of the ORP/Osh family are PS/PI(4)P exchangers. Notably, ORP10 was found to encapsulate PS in vitro and transfer PS in ER-Golgi contact sites27,50, but it is still unknown whether it acts as a PS/PI(4)P exchanger. Furthermore, these protocols can serve to explore the ability of LTPs belonging to other families to transport PI(4)P, as was recently shown with steroidogenic acute regulatory transfer-like proteins10, or PS. Additionally, NBD-PHFAPP has been used to follow the ability of LTP to transport PI(4,5)P2 between liposomes, as it recognizes this PIP10,35. Finally, this strategy could be adapted to measure other extraction or transport processes in vitro by using other lipid-binding domains (e.g., a non-catalytic PI-PLC to detect PI51, sporulation-specific protein 20 fragment to detect PA52, domain 4 (D4) of perfringolysin O to detect sterol53).
The authors have nothing to disclose.
We are grateful to Dr. A. Cuttriss for her careful proofreading of the manuscript. This work is funded by the French National Research Agency grant ExCHANGE (ANR-16-CE13-0006) and by the CNRS.
L-cysteine ≥97 % (FG) | Sigma | W326305-100G | Prepare a 10 mM L-cysteine stock solution in water. Aliquots are stored at -20 °C |
2 mL Amber Vial, PTFE/Rub Lnr, for lipids storage in CHCL3 | Wheaton | W224681 | |
4 mm-diameter glass beads | Sigma | Z265934-1EA | |
50 mL conical centrifuge tube | Falcon | ||
ÄKTA purifier | GE healthcare | FPLC | |
Aluminium foil | |||
Amicon Ultra-15 with a MWCO of 3 and 10 kDa | Merck | UFC900324, UFC901024 | |
Amicon Ultra-4 with a MWCO of 3 and 10 kDa | Merck | UFC800324, UFC801024 | |
Ampicillin | Prepare a 50 mg/mL stock solution with filtered and sterilized water and store it at -20 °C. | ||
Bestatin | Sigma | B8385-10mg | |
BL21 Gold Competent Cells | Agilent | ||
C16:0 Liss (Rhod-PE) in CHCl3 (1 mg/mL) | Avanti Polar Lipids | 810158C-5MG | |
C16:0/C16:0-PI(4)P | Echelon Lipids | P-4016-3 | Dissolve 1 mg of C16:0/C16:0-PI(4)P powder in 250 µL of MeOH and 250 µL of CHCl3. Then complete with CHCl3 to 1 mL. The solution must become clear. |
C16:0/C18:1-PS (POPS) in CHCl3 (10 mg/mL) | Avanti Polar Lipids | 840034C-25mg | |
C18:1/C18:1-PC (DOPC) in CHCl3 (25 mg/mL) | Avanti Polar Lipids | 850375C-500mg | |
CaCl2 | Sigma | Prepare 10 mM CaCl2 stock solution in water. | |
Cell Disruptor | Constant Dynamics | ||
Chloroform (CHCl3) RPE-ISO | Carlo Erba | 438601 | |
Complete EDTA-free protease inhibitor cocktail | Roche | 5056489001 | |
Deionized (Milli-Q) water | |||
Dimethylformamide (DMF), anhydrous, >99% pure | |||
DNAse I Recombinant, RNAse free, in powder | Roche | 10104159001 | |
DTT | Euromedex | EU0006-B | Prepare 1 M DTT stock solution in Milli-Q water. Prepare 1 mL aliquots and store them at -20 °C. |
Econo-Pac chromatography columns (1.5 × 12 cm). | Biorad | 7321010 | |
Electroporation cuvette 2 mm | Ozyme | EP102 | |
Electroporator Eppendorf 2510 | Eppendorf | ||
Fixed-Angle Rotor Ti45 and Ti45 tubes | Beckman | Spinning the batcerial lysates | |
Glass-syringes (10, 25, and 50 µL) for fluorescence experiment | Hamilton | ||
Glass-syringes (25 , 100, 250, 500, and 1000 µL) to handle lipid stock solutions | Hamilton | 1702RNR, 1710RNR, 1725RNR, 1750RN type3, 1001RN | |
Glutathione Sepharose 4B beads | GE Healthcare | 17-0756-05 | |
Glycerol (99% pure) | Sigma | G5516-500ML | |
Hemolysis tubes with a cap | |||
HEPES , >99 % pure | Sigma | H3375-500G | |
Illustra NAP 10 desalting column | GE healthcare | GE17-0854-02 | |
Isopropyl β-D-1-thiogalactopyranoside (IPTG) | Euromedex | EU0008-B | Prepare 1 M IPTG stock solution in Milli-Qwater. Prepare 1 mL aliquots and store them at -20 °C. |
K-Acetate | Prolabo | 26664.293 | |
Lennox LB Broth medium without glucose | Prepared with milli-Q water and autoclaved. | ||
Liquid nitrogen | Linde | ||
Methanol (MeOH) ≥99.8% | VWR | 20847.24 | |
MgCl2 | Sigma | Prepare a 2 M MgCl2 solution. Filter the solution using a 0.45 µm filter. | |
Microplate 96 Well PS F-Botom Black Non-Binding | Greiner Bio-one | 655900 | |
Mini-Extruder with two 1 mL gas-tight Hamilton syringes | Avanti Polar Lipids | 610023 | |
Monochromator-based fluorescence plate reader | TECAN | M1000 Pro | |
N,N'-Dimethyl-N-(Iodoacetyl)-N'-(7-Nitrobenz-2-Oxa-1,3-Diazol-4-yl)Ethylenediamine) (IANBD Amide) | Molecular Probes | Dissolve 25 mg of IANBD in 2.5 mL of dimethylsulfoxide (DMSO) and prepare 25 aliquot of 100 µL in 1.5 mL screw-cap tubes. Do not completely screw the cap. Then, remove DMSO in a freeze-dryer to obtain 1 mg of dry IANBD per tube. Tubes are closed and stored at -20 °C in the dark. | |
NaCl | Sigma | S3014-1KG | |
PBS | 137 mM NaCl, 2.7 mM KCl, 10 mM NaH2PO4, 1.8 mM KH2PO4, autoclaved and stored at 4 °C. | ||
Pear-shaped glass flasks (25 mL, 14/23, Duran glass) | Duran Group | ||
Pepstatin | Sigma | p5318-25mg | |
pGEX-C2LACT plasmid | Available on request from our lab | ||
pGEX-PHFAPP plasmid | Available on request from our lab | ||
Phenylmethylsulfonyl fluoride (PMSF) ≥98.5% (GC) | Sigma | P7626-25g | Prepare a 200 mM PMSF stock solution in isopropanol |
Phosphoramidon | Sigma | R7385-10mg | |
Polycarbonate filters (19 mm in diameter) with pore size of 0.2 µm | Avanti Polar Lipids | 610006 | |
Poly-Prep chromatography column (with a 0-2 mL bed volume and a 10 mL reservoir) | Biorad | 7311550 | |
Prefilters (10 mm in diameter). | Avanti Polar Lipids | 610014 | |
PyMOL | http://pymol.org/ | Construction of the 3D models of the proteins (Figure 1A) | |
Quartz cuvette for UV/visible fluorescence (minimum volume of 600 µL) | Hellma | ||
Quartz cuvettes | Hellma | ||
Refrigerated centrifuge Eppendorf 5427R | Eppendorf | ||
Rotary evaporator | Buchi | B-100 | |
Screw-cap microcentriguge tubes (1.5 mL) | Sarsted | ||
Small magnetic PFTE stirring bar (5 × 2 mm) | |||
Snap-cap microcentriguge tubes (0.5, 1, and 2 mL) | Eppendorf | ||
SYPRO orange | fluorescent stain to detect protein in SDS-PAGE gel | ||
Thermomixer | Starlab | ||
THROMBIN, FROM HUMAN PLASMA | Sigma | 10602400001 | Dissolve 20 units in 1 mL of milli-Q water and prepare 25 µL aliquots in 0.5 mL Eppendorf tubes. Then freeze and store at -80 °C. |
Tris, ultra pure | MP | 819623 | |
Ultracentrifuge L90K | Beckman | ||
UV/Visible absorbance spectrophotometer | SAFAS | ||
UV/visible spectrofluorometer with a temperature-controlled cell holder and stirring device | Jasco or Shimadzu | Jasco FP-8300 or Shimadzu RF-5301PC | |
Vacuum chamber | |||
Water bath | Julabo | ||
XK 16/70 column packed with Sephacryl S200HR | GE healthcare |