Equilibrium-Fluctuation Analysis for Interaction Studies between

Sep 7, 2015 - G protein-coupled receptors (GPCRs) constitute the most versatile family of cell-membrane receptors and have been increasingly identifie...
3 downloads 13 Views 2MB Size
Download PDF
Article pubs.acs.org/Langmuir

Equilibrium-Fluctuation Analysis for Interaction Studies between Natural Ligands and Single G Protein-Coupled Receptors in Native Lipid Vesicles Olov Wahlsten,† Anders Gunnarsson,‡ Lisa Simonsson Nyström,† Hudson Pace,† Stefan Geschwindner,‡ and Fredrik Höök*,† †

Department of Applied Physics, Chalmers University of Technology, SE 41296 Gothenburg, Sweden Discovery Sciences, AstraZeneca R&D Mölndal, S-43183 Mölndal, Sweden

‡

ABSTRACT: G protein-coupled receptors (GPCRs) constitute the most versatile family of cell-membrane receptors and have been increasingly identified as important mediators of many physiological functions. They also belong to one of the most central drug target classes, but current screening technologies are limited by the requirements of overexpression and stabilization of GPCRs. This calls for sensitivity-increased detection strategies preferably meeting single-molecule detection limits. This challenge is here addressed by employing total internal reflection fluorescence microscopy to characterize the interaction kinetics between CXCR3, a GPCR involved in inflammatory responses, and two of its chemokine ligands, CXCL10 and CXCL11. Fluorescence labeling of the lipid membrane, rather than the membrane protein itself, of GPCR-containing native vesicles, and immobilization of the corresponding ligand on the surface, enabled determination of the interaction kinetics using singlemolecule equilibrium-fluctuation analysis. With a limit of detection of GPCR-containing vesicles in the low picomolar concentration regime, the results demonstrate the possibility to use inhibition in solution screening of high affinity ligands/drug candidates, which due to target-binding depletion of the inhibiting compounds is demanding using assays with more moderate detection limits.

■

INTRODUCTION G protein-coupled receptors (GPCRs) account for nearly 4% of the protein encoding genome and are involved in a variety of physiological processes such as sensory transduction, cell−cell communication, neuronal transmission, and hormonal signaling.1 GPCRs are therefore considered one of the most pertinent targets for design and development of novel therapeutic compounds. Indeed, more than 50% of the therapeutic agents currently on the market target GPCRs.1 The chemokine receptor CXCR3 (also known as GPR9 or CD183) is a GPCR primarily expressed on activated T lymphocytes but also on resting T lymphocytes, B cells, monocytes, and granulocytes.2,3 The complex interplay between this and other chemokine receptors and their ligands is essential for many physiological processes, such as immunecell trafficking, promotion of angiogenesis, and inflammatory responses. There are three known chemokine ligands that bind the CXCR3 receptor: CXCL9 (Mig), CXCL10 (IP-10), and CXCL11 (I-TAC). These chemokines are predominantly produced by activated monocytes and macrophages stimulated by the cytokine IFN-γ,4,5 and binding of these chemokines to CXCR3 has been observed to induce both calcium flux and chemotactic responses.6 Further, the interactions between CXCR3 and its chemokine ligands are involved in many human diseases, e.g. multiple sclerosis, arthritis, inflammatory © XXXX American Chemical Society

bowel disease (IBD), asthma, chronic obstructive pulmonary disease (COPD), and transplant rejection,7 and since prevention or inhibition of ligand−receptor interactions may mitigate inflammatory conditions, CXCR3 is considered a pharmaceutically highly relevant target.4 Direct receptor-binding assays (i.e., cell-free methods which do not rely upon downstream signaling), which are commonly utilized for investigating GPCR−ligand interactions, are challenged by both low GPCR expression levels and the sensitivity of GPCRs to environmental changes.8,9 Signal enhancement aided by detergent solubilization, protein purification, and subsequent reconstitution into suitable membrane mimics are common strategies to overcome these challenges.10−12 Further, a strategy to overcome their sensitivity to environmental variations has been to increase the stability of GPCRs by using directed mutations to lock them in specific conformations.13 However, such conformations may not necessarily be compatible with the investigation of the diverse modes of ligand binding, including agonists, antagonists, partial or inverse antagonists/agonists, etc. Similarly, attempts to increase the signal by modifying the ligand and/or the receptor Received: July 3, 2015 Revised: August 19, 2015

A

DOI: 10.1021/acs.langmuir.5b02463 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir via fluorescence or radioactive agents might also influence their modes of action.14,15 These approaches have in common that they are potentially adverse for this sensitive class of proteins and require careful evaluation to ensure that the protein stability, function, and activity have not been altered.16 Additionally, traditional techniques used to investigate ligand−receptor interactions are based on ensemble averaging, which means that compositional heterogeneities possible to obtain from single-molecule studies cannot be unraveled. In fact, it has recently been shown that such heterogeneities might alter conclusions drawn from proteoliposome measurements based on ensemble averaging,17 stressing the relevance of single-molecule approaches. A singlemolecule approach may also improve the sensitivity enough to eliminate the need for membrane protein overexpression, enabling studies of endogenous cell membranes. In this work we have addressed these challenges by using total internal reflection fluorescence (TIRF) microscopy to visualize individual GPCR-containing native membrane vesicles18,19 to determine the kinetic parameters representing the interaction between the chemokine receptor CXCR3 and two of its natural ligands: CXCL10 and CXCL11. The assay has been simplified to operate in small liquid volumes (∼20 μL) under stagnant conditions, and thanks to the single-molecule read-out scheme, the possibility of extracting kinetic parameters is preserved despite the lack of advanced liquid handling. To probe the ligand−receptor interaction, one of the chemokine ligands (CXCL10) was immobilized on a supported lipid bilayer (SLB), and the CXCR3 receptors were presented using native membrane vesicles produced via cell homogenization and centrifugation. By avoiding both detergent solubilization and membrane-protein reconstitution processes, the native membrane vesicle format maintains a near-natural environment for the GPCRs, which means that the GPCRs remains surrounded by essentially the same membrane components (e.g., proteins and lipids) as in their native cell. Additionally, the native-membrane-vesicle format offers a possibility to visualize individual interaction events between the receptors and surfaceimmobilized ligands without directly labeling either of the interacting entities (GPCR or ligand), but by instead label the lipid membrane in which the GPCR is embedded. The accompanied high sensitivity (low picomolar regime) originating from single-molecule readout is explored in the context of inhibition in solution using high affinity ligands, which due to target-binding depletion of the inhibiting compounds is demanding using assays with more moderate detection limits.

■

RH) were purchased from Avanti Polar Lipids Inc., Alabaster, AL. Phosphate buffered saline (PBS), sodium acetate (NaAc), 4-(2hydroxyethyl)piperazine-1-ethanesulfonic acid sodium salt (HEPES), magnesium chloride hexahydrate (MgCl2·6H2O), sodium chloride (NaCl), and 2-iminothiolane hydrochloride (Traut’s reagent) were purchased from Sigma-Aldrich, Germany. The buffer used for the inhibition studies with CXCL11 was 25 mM HEPES, 10 mM MgCl2, and 100 mM NaCl (pH 7.4). Preparation of Synthetic Lipid Vesicles. The lipid material, dissolved in methanol, was evaporated at the bottom of a round flask under a gentle stream of N2. Residual solvent was removed by exposing the interiors of the flask with the dried lipid film under vacuum for >1 h. The film was then hydrated with 1 mL of PBS buffer and vortexed until a turbid whitish solution was obtained (>1 min). The lipid solution was then extruded 11 times through a 30 nm polycarbonate membrane (Whatman, UK) using a mini extruder (Avanti Polar Lipids Inc.). The liposomes formed had an average diameter of around 120 nm, according to a nanoparticle tracking analysis (NTA) (NanoSight LM10, NanoSight, Wiltshire, UK) estimation. Surface Cleaning Protocol (QCM-D and TIRF). QCM-D measurements were performed using a Q-Sense E4 system (BiolinScientific/Q-Sense, Göteborg, Sweden) using AT-cut SiO2coated quartz crystals with a fundamental frequency of 5 MHz (QSense, Göteborg, Sweden) as sensors. The crystals were cleaned by sonication (Elmasonic S40H) in SDS (sodium dodecyl sulfate, SigmaAldrich, Germany) for 5 min followed by thorough rinsing in Milli-Q water (Millipore, France) for at least 1 min prior to 30 min UV-ozone cleaning. TIRF measurements were performed using rectangular microscope cover glasses No. 1 (24 × 40 mm, Fisherbrand). These were boiled for >15 min in a solution of Milli-Q water and 2% Liqui-Nox (Alconox, Whiteplaines, NY) purchased from Sigma-Aldrich, Germany. The glass slides were then rinsed with copious amounts of Milli-Q water to remove all residual detergent. The cleaned glass slide was blown dry with N2, and a thin rectangular piece of polydimethylsiloxane (PDMS), made from a mixture of 10:1 Sylgard 184 and curing agent (Dow Corning, Midland, MI), was put on top of it. The PDMS piece, measuring 25 × 15 × 1 mm, with eight punched circular holes (2 by 4) with diameters of 3 mm and a volume of ∼20 μL, was prior to use kept between two strips of regular office tape (with their sticky side to the PDMS) to enhance the adhesion to the glass slide. The eight circular holes form together with the glass slide separate compartments referred to herein as wells. Surface Modification Protocol (QCM-D and TIRF). For both the QCM-D and TIRF experiments chemokine ligands (CXCL10) were immobilized on a supported lipid bilayer (SLB) using a thiol− maleimide coupling chemistry. In the QCM-D experiments a supported lipid bilayer (SLB) was first formed on the sensor surface by exposing SiO2-coated QCM-D crystals to lipid vesicles containing POPC + 10 wt % DOPE-MCC (0.1 mg/mL). Primary amines of the CXCL10 ligands were modified to thiol groups by mixing the ligands with Traut’s reagent at a concentration of 50 μg/mL for both components for 15 min, corresponding to a molar ratio of 1:60 ligand to Traut’s reagent. Prior to addition of the modified CXCL10 ligands (100 μL at 10 μg/mL) to the SLB, the running buffer was exchanged to a 10 mM NaAc buffer (pH 5.0). After completed reaction and rinsing with NaAc buffer and exchange to PBS, (unlabeled) CXCR3containing native membrane vesicles (100 μL at ∼0.2 mg/mL) were added. (The protein concentration of the native membrane vesicles was given by the provider. The total mass concentration was estimated by assuming equal lipid and protein masses.) For the TIRF experiments the glass surfaces inside the wells were modified with supported lipid bilayers (SLB) consisting of POPC + 0.1 mol % DSPE-PEG2000-maleimide via vesicle fusion by incubating the corresponding vesicle solution (∼0.3 mg/mL in PBS) in each well for >10 min. Primary amines of CXCL10 were modified with Traut’s reagent as described above. Exchange of buffer in the wells to a 10 mM NaAc buffer (pH 5.0) after SLB formation was followed by addition of the modified CXCL10 ligand (∼5 μg/mL). The wells were

EXPERIMENTAL SECTION

Native Membrane Vesicles, Ligands, Lipids, Reagents, and Buffers. Native membrane vesicles, containing the human chemokine receptor CXCR3, were purchased from PerkinElmer, Boston, MA. These native membrane vesicles were derived from a CHO-K1 cell line via mechanical homogenization and two centrifugation steps (1000 rpm to remove nuclei and unbroken cells followed by 18 000 rpm on the collected supernatant). Recombinant human chemokine ligands CXCL10 (IP-10) and CXCL11 (I-TAC) were purchased from Peprotech, London, UK. The lipids 1-palmitoyl-2-oleoyl-sn-glycero-3phosphocholine (POPC), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-[4-(p-maleimidomethyl)cyclohexanecarboxamide] (DOPEMCC), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N[maleimide(poly(ethylene glycol))-2000] (DSPE-PEG2000-Maleimide), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(7-nitro-21,3-benzoxadiazol-4-yl) (DOPE-NBD), and 1,2-dioleoyl-sn-glycero-3phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (DOPEB

DOI: 10.1021/acs.langmuir.5b02463 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir subsequently rinsed after 15 min with NaAc buffer and PBS. Addition of fluorescently labeled CXCR3-containing native membrane vesicles (see below) to each well (∼0.5 μg/mL vesicles in the wells) was done for at least 2 min prior to each TIRF measurement. Fluorescence Labeling of Native Membrane Vesicles. Fluorescent DOPE-RH lipids were introduced into the CXCR3containing native membrane vesicles via sonication, as described by Pace et al.20 In brief, liposomes of POPC + 1 mol % DOPE-RH were mixed with the native membrane vesicles at a 1:1 mass ratio. The liposome mixture (50 μL at 0.8 mg/mL in a 0.5 mL Eppendorf tube) was then sonicated (Elmasonic S40H) for 10 min at 10 °C. Lipid mixing was, as detailed in the Results and Discussion section, confirmed with a FRET assay using a spectrofluorometer (QM-4/2005 spectrofluorometer, Photon Technology International Inc., Birmingham, NJ) and size determination measurements using NanoSight (NanoSight, Wiltshire, UK). Off note, a number of alternative membrane labeling protocols were investigated, including diI, PKH, and FM dyes. However, incomplete removal of excess dye and issues with dye transfer from the vesicles to the lipid bilayer complicated the data analysis of longer measurements. TIRF Microscopy and Image Analysis. TIRF measurements were carried out using an inverted Nikon Eclipse Ti-E microscope (Nikon Corporation, Tokyo, Japan) equipped with an Andor Ixon+ camera (Andor Technology, Belfast, Northern Ireland) with a 0.7× demagnification lens, a 60× magnification (NA = 1.49) oil immersion objective (Nikon Corporation), a TRITC (rhodamine-DHPE) filter cube (Nikon Corporation), and a perfect focus system (PFS). A mercury lamp (Intensilight C-HGFIE; Nikon Corporation) connected to the microscope via an optical fiber was used as illumination source. Micrographs containing 512 × 512 pixels (0.41 μm/pixel) were acquired in time-lapse mode with an exposure time of 100 ms. In TIRF mode only fluorescently labeled liposomes residing in close proximity to the surface (∼100 nm) are excited by the exponentially decaying evanescent field, thus effectively suppressing the signal from fluorescent vesicles in bulk.21 Processing of the TIRF images to detect vesicles attaching to the sensor surface as well as determination of their residence times was done using Matlab R2010a (The MathWorks, Inc., Natick, MA). Details about the script used for analysis have been described previously by Gunnarsson et al.22

Figure 1. Verification of ligand−receptor specificity. Quartz crystal microbalance with dissipation monitoring (QCM-D) plot showing the specific interaction between CXCR3 (receptor) and immobilized CXCL10 (ligand). (i) Injection of vesicles containing POPC + 10 wt % DOPE-MCC (0.1 mg/mL) and subsequent bilayer formation (emphasized in the upper right inset). (ii) NaAc (10 mM at pH 5.0) buffer step. (iii) Addition of CXCL10 ligands preincubated with Traut’s reagent (10 μg/mL) and subsequent rinsing with NaAc buffer, followed by (iv) a PBS buffer step and (v) injection of (unlabeled) CXCR3-containing native membrane vesicles (∼0.2 mg/mL). The red and blue curves represent SLBs with and without immobilized CXCL10, respectively.

CXCR3-containing native membrane vesicles in comparison to negative control SLBs lacking CXCL10 (∼−15 Hz vs ∼−3 Hz). While the native membrane vesicle binding in the negative control is not negligible, the 5-fold increase in frequency shift validates the specificity of the interaction. To further suppress nonspecific interactions in single vesicle experiments in TIRF mode, the maleimide-presenting lipids were exchanged for homologous lipids which instead utilized a 2 kDa PEG spacer between the lipid headgroup and the maleimide moietya strategy adapted from reports of grafted PEG on lipid surfaces preventing nonspecific protein adsorption.25 Additionally, we speculated that the length and flexibility of the PEG spacer could enhance the maleimideligand conjugation efficiency via increased accessibility. Such an increase in accessibility of the maleimide for ligand coupling might also allow the ligands to be more accessible for the receptor-containing native membrane vesicles, as indicated previously.26 However, since SLB formation is inhibited already using ∼3 mol % PEG 2 kDa−PE lipids27 and such low surface coverages produce binding responses below the limit of detection of the QCM-D, the binding specificity was instead verified using TIRF microscopy. Fluorescence Labeling of Native Membrane Vesicles. In order to visualize the ligand−receptor interaction using TIRF, the native membrane vesicles were fluorescently labeled. Several different membrane inserting dyes, including diI, PKH, and FM dyes, were initially tested without success. Although the labeling protocols worked, dye transfer to the SLB (partly due to difficulties to remove excess dye from the vesicle solution) upon addition of the native membrane vesicles resulted in a gradual increase in background fluorescence from the SLB, which complicated single vesicle detection in longer experiments. Instead, fluorescence labeling was achieved by

■

RESULTS AND DISCUSSION Ligand Immobilization on Supported Lipid Bilayer. Chemokine ligands (CXCL10) were immobilized on a supported lipid bilayer (SLB) using a thiol−maleimide coupling chemistry. Thiol-modified ligands were coupled to lipidanchored maleimide groups in SLBs of POPC. The ligandcoupling conditions were optimized using quartz crystal microbalance with dissipation monitoring (QCM-D), as shown in Figure 1. To verify specific binding of CXCR3containing vesicles to SLB-immobilized CXCL10 ligands, the SLB was formed from POPC lipid vesicles containing 10 wt % DOPE-MCC lipids. The relatively high concentration of maleimide groups was required to yield a sufficiently high signal to detect both CXCL10 coupling and subsequent vesicle binding using QCM-D (Figure 1). After complete SLB formation (step i), as verified by characteristic changes in the frequency response,23 the DOPEMCC-containing SLB was exposed to NaAc buffer (10 mM at pH 5.0) (step ii) followed by addition of CXCL10 (step iii), which after rinsing (step iv) resulted in a frequency shift of ∼−1 Hz. With a molecular weight of ∼9 kDa, saturated binding of CXCL10 is expected to form a film with an effective thickness of ∼1 nm and thus a frequency shift of ∼−5 Hz. The difference between the theoretical and observed frequency shifts is attributed to limited availability of the MCC units.24 However, sufficient CXCL10 coupling was confirmed by the relatively large frequency shift observed upon introduction of C

DOI: 10.1021/acs.langmuir.5b02463 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 2. Verification of lipid mixing. (A) Fluorescence emission for different vesicle suspensions when excited at 460 nm, corresponding to the excitation wavelength of NBD (donor). Intact and detergent-ruptured vesicles of POPC + 1 mol % DOPE-NBD + 1 mol % DOPE-RH (called FRET vesicles) give maximum and minimum FRET, respectively. The inset illustrates the distribution of fluorophores (green stars and pink squares) between the two vesicle entities before and after sonication. (B) Size distributions collected with NTA of fluorophore-containing POPC vesicles (blue), unlabeled CXCR3-containing native membrane vesicles (red), 1:1 mass ratio mixture of the two entities nonsonicated (green), and sonicated (black). Presented here is an average of three measurements.

Figure 3. Illustration of single native membrane vesicle imaging. Vesicles derived from membranes of cells expressing a GPCR (blue) were fluorescently labeled and added to a surface with the corresponding chemokine ligand (red) immobilized. Total internal reflection fluorescence (TIRF) microscopy was used to monitor (right) and characterize the ligand−receptor interaction with single native membrane vesicle / GPCR sensitivity.

Receptor−Ligand Interaction Studies. The interaction between CXCR3 and its natural ligand, CXCL10, was studied under stagnant conditions in a homemade microwell plate. CXCL10 ligands were immobilized on the SLBs, and fluorescently labeled CXCR3-containing native membrane vesicles were added to the wells prior to TIRF measurements, as schematically illustrated in Figure 3. Binding of labeled native membrane vesicles was monitored in time-lapse mode with an acquisition rate of 0.1 frames per second (fps). Inhibition of the interaction between the CXCR3-containing native membrane vesicles and the CXCL10 ligands immobilized on the surface was performed by adding free CXCL10 ligands to the solution. With a free ligand concentration of 1 μM (which is at least 2 orders of magnitude above the expected Kd of the ligand−GPCR interaction29−31), the number of binding events on the surface was reduced by more than 1 order of magnitude (Figure 4A). This strong reduction indicates a significantly lower degree of nonspecific binding than obtained in the QCM-D measurements (Figure 1), being attributed to the PEG moiety providing increased ligand-immobilization efficiency as well as a more inert surface. Further, both the dilution of the native membrane vesicles with POPC lipids and the reduction in

inducing lipid mixing between the CXCR3-containing native membrane vesicles and synthetic vesicles containing fluorescently labeled lipids (POPC + 1 mol % DOPE-RH) using gentle bath sonication. The lipid mixing was confirmed using (i) FRET measurements by including two different lipidattached fluorophores28 (DOPE-RH and DOPE-NBD) in the synthetic vesicles (Figure 2A) and (ii) size determination measurements using nanoparticle tracking analysis (NTA) prior to and after sonication (Figure 2B). A clear reduction in FRET was observed for the sonicated mixture of CXCR3-containing native membrane vesicles and vesicles containing POPC + 1 mol % DOPE-NBD + 1 mol % DOPE-RH, being consistent with efficient lipid mixing, attributed to multiple vesicle-fusion and rupture events. This interpretation is further supported by the fact that the vesicles emerging from the sonication procedure are on average significantly smaller than the native membrane vesicles alone, yet larger than the presonication mixture. The influence of GPCR functionality on the sonication step was addressed below by investigating their ligand-binding capacity. However, it is worth noting that the conservation of enzymatic activity of other membrane proteins subjected to similar sonication conditions has been previously confirmed.20 D

DOI: 10.1021/acs.langmuir.5b02463 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 4. Results from the equilibrium fluctuation analysis. (A) Cumulative number of new vesicles bound to the surface plotted versus time. Each curve corresponds to a specific bulk concentration of CXCL10 as inhibitor. (B) Binding rates from (A) make up the corresponding dose−response curves. Equation 1 was fitted (black line) to the data to extract the Kd for the interaction. Data from three individual experiments were compiled. (C) A dose−response curve for CXCL11 as an inhibitor. Equation 1 was fitted to the data to extract the Kd for the interaction. (D) A representative dissociation plot constructed from the residence times of vesicles that bind and release during the experiment. Also shown is a single-exponential curve fitted to the data.

same inhibition-in-solution approach was also applied to CXCL11, a second natural ligand of CXCR3 (Figure 4C). The geometric mean and the multiplicative standard deviation for three individual experiments yield an equilibrium dissociation constant of Kd = 13.4 ×/ 4.3 nM (n = 3), a value that is comparable to values previously obtained using radioligand assays.29 Further confirmation of representative determination of interaction kinetics of this type was obtained by also monitoring the fraction of native membrane vesicles that were released during the time frame of the experiments, from which statistics of the residence time of each individual interaction event can be used to determine the dissociation constant, koff.22 Thanks to operation under equilibrium fluctuation condition, this was possible without, as typically done, following the rate of biomolecular dissociation upon rinsing, and from statistics of the residence time distribution, a representation analogous to a conventional dissociation curve could be produced (Figure 4D). Fitting of the data to a singleexponential curve yielded a geometric mean value for koff and multiplicative standard deviation of 1.0 × 10−2 ×/ 1.4 s−1 (n = 3).

vesicle size upon sonication (Figure 2) are likely to reduce nonspecific interactions with a lipid membrane dominated by POPC lipids. To investigate the strength of the ligand binding, the cumulative number of newly arrived vesicles was measured for different concentrations of suspended (and inhibiting) CXCL10 (10 pM−1 μM) (Figure 4A). The equilibrium dissociation constant, Kd, was determined by a sigmoidal fit of eq 1: ⎛ ⎛ ⎞⎞ dN+ ⟨α⟩ ⎟⎟⎟ + B = A⎜⎜1 − exp⎜⎜ − ⎟ dt ⎝ 1 + C ligand /Kd ⎠⎠ ⎝

(1)

to the vesicle-binding rate versus concentration of inhibiting ligand (Figure 4B).18 Here N+ is the cumulative number of vesicles adsorbed to the surface, ⟨α⟩ the average number of receptors per vesicle (estimated to be 0.2, based on CXCR3 expression level stated by the manufacturer, average vesicle size, and by assuming equal lipid and protein mass in the native membrane vesicles), Cligand the concentration of suspended ligand, and Kd the equilibrium dissociation constant between the ligand and the receptor. Both A and B are constants independent of the other variables. The geometric mean and the multiplicative standard deviation for three individual experiments yield an equilibrium dissociation constant for the CXCR3/CXCL10 interaction of Kd = 6.9 ×/ 4.4 nM (n = 3), which is in good agreement with results reported in radioligand assays.29−31 The corresponding Kd value with ⟨α⟩ = 1 would be somewhat lower (4.6 ×/ 4.4 nM). (Both off-rates and dose− response data produce log-normally distributed variables. This makes a geometrical mean value with a variance based on a times-divided term a more appropriate choice for communicating the results, rather than the more commonly used arithmetic standard variation expressed with the plus−minus term.32) The

■

CONCLUSION These results demonstrate that this single native membrane vesicle/single GPCR-molecule approach enables kinetic parameter determination for the interaction between a GPCR (the chemokine receptor CXCR3) and its natural ligands (the chemokines CXCL10 and CXCL11) under stagnant liquid conditions, with each measurement taking less than around 10 min and that in total not more than around 100 GPCR−ligand interactions were required to extract each individual kinetic data point. The extracted values are in good agreement with previously reported values,29,31 although here obtained in a E

DOI: 10.1021/acs.langmuir.5b02463 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

(10) Harding, P. J.; Hadingham, T. C.; McDonnell, J. M.; Watts, A. Direct analysis of a GPCR-agonist interaction by surface plasmon resonance. Eur. Biophys. J. 2006, 35, 709−712. (11) Locatelli-Hoops, S.; Yeliseev, A. A.; Gawrisch, K.; Gorshkova, I. Surface plasmon resonance applied to G protein-coupled receptors. Biomed. Spectrosc. Imaging 2013, 2, 155−181. (12) Vega, B.; Calle, A.; Sanchez, A.; Lechuga, L. M.; Ortiz, A. M.; Armelles, G.; Rodriguez-Frade, J. M.; Mellado, M. Real-time detection of the chemokine CXCL12 in urine samples by surface plasmon resonance. Talanta 2013, 109, 209−215. (13) Bertheleme, N.; Singh, S.; Dowell, S. J.; Hubbard, J.; Byrne, B. Loss of constitutive activity is correlated with increased thermostability of the human adenosine A2A receptor. Br. J. Pharmacol. 2013, 169, 988−998. (14) Sridharan, R.; Zuber, J.; Connelly, S. M.; Mathew, E.; Dumont, M. E. Fluorescent approaches for understanding interactions of ligands with G protein coupled receptors. Biochim. Biophys. Acta, Biomembr. 2014, 1838, 15−33. (15) Sun, Y. S.; Landry, J. P.; Fei, Y. Y.; Zhu, X. D. Effect of Fluorescently Labeling Protein Probes on Kinetics of Protein-Ligand Reactions. Langmuir 2008, 24, 13399−13405. (16) Patching, S. G. Surface plasmon resonance spectroscopy for characterisation of membrane protein-ligand interactions and its potential for drug discovery. Biochim. Biophys. Acta, Biomembr. 2014, 1838, 43−55. (17) Mathiasen, S.; Christensen, S. M.; Fung, J. J.; Rasmussen, S. G. F.; Fay, J. F.; Jorgensen, S. K.; Veshaguri, S.; Farrens, D. L.; Kiskowski, M.; Kobilka, B.; Stamou, D. Nanoscale high-content analysis using compositional heterogeneities of single proteoliposomes. Nat. Methods 2014, 11, 931−934. (18) Gunnarsson, A.; Dexlin, L.; Wallin, P.; Svedhem, S.; Jonsson, P.; Wingren, C.; Hook, F. Kinetics of Ligand Binding to Membrane Receptors from Equilibrium Fluctuation Analysis of Single Binding Events. J. Am. Chem. Soc. 2011, 133, 14852−14855. (19) Moonschi, F. H.; Effinger, A. K.; Zhang, X. L.; Martin, W. E.; Fox, A. M.; Heidary, D. K.; DeRouchey, J. E.; Richards, C. I. CellDerived Vesicles for Single-Molecule Imaging of Membrane Proteins. Angew. Chem., Int. Ed. 2015, 54, 481−484. (20) Pace, H. P.; Simonsson Nystrom, L.; Gunnarsson, A.; Eck, E.; Monson, C. F.; Geschwindner, S.; Snijder, A.; Hook, F. Preserved Transmembrane Protein Mobility in Polymer-Supported Lipid Bilayers Derived from Cell Membranes. Anal. Chem. 2015, DOI: 10.1021/acs.analchem.5b01449. (21) Axelrod, D.; Burghardt, T. P.; Thompson, N. L. Total InternalReflection Fluorescence. Annu. Rev. Biophys. Bioeng. 1984, 13, 247− 268. (22) Gunnarsson, A.; Jonsson, P.; Marie, R.; Tegenfeldt, J. O.; Hook, F. Single-molecule detection and mismatch discrimination of unlabeled DNA targets. Nano Lett. 2008, 8, 183−188. (23) Keller, C. A.; Kasemo, B. Surface specific kinetics of lipid vesicle adsorption measured with a quartz crystal microbalance. Biophys. J. 1998, 75, 1397−1402. (24) Thid, D.; Bally, M.; Holm, K.; Chessari, S.; Tosatti, S.; Textor, M.; Gold, J. Issues of ligand accessibility and mobility in initial cell attachment. Langmuir 2007, 23, 11693−11704. (25) Du, H.; Chandaroy, P.; Hui, S. W. Grafted poly-(ethylene glycol) on lipid surfaces inhibits protein adsorption and cell adhesion. Biochim. Biophys. Acta, Biomembr. 1997, 1326, 236−248. (26) Tong, Z. H.; Mikheikin, A.; Krasnoslobodtsev, A.; Lv, Z. J.; Lyubchenko, Y. L. Novel polymer linkers for single molecule AFM force spectroscopy. Methods 2013, 60, 161−168. (27) Kaufmann, S.; Papastavrou, G.; Kumar, K.; Textor, M.; Reimhult, E. A detailed investigation of the formation kinetics and layer structure of poly(ethylene glycol) tether supported lipid bilayers. Soft Matter 2009, 5, 2804−2814. (28) Struck, D. K.; Hoekstra, D.; Pagano, R. E. Use of Resonance Energy-Transfer to Monitor Membrane-Fusion. Biochemistry 1981, 20, 4093−4099.

near-native environment of the GPCR and without direct labeling of either the receptor or the ligands. The approach has thus the potential of being used as an inhibition in solution assay (ISA) for drug-candidate screening, by simply replacing the suspended ligand for drug candidates of interest. This is particularly relevant since ISAs are known to suffer from not being applicable to highly potent binders, as used here, since the lowest detectable concentration of the membrane protein defines the highest affinity that can be unambiguously distinguished. In analogy with recent reports on the use of single membrane-protein sensitivity to screen high potent binders,33 our results thus demonstrate that this limitation should be possible to overcome also for GPCRs, possibly even at endogenous expression levels.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (F.H.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by the Göran Gustafsson foundation, the Swedish Foundation for Strategic Research (#RMA11-0104) and VR (#2014-5557). The authors also acknowledge helpful discussions with Dr. A. Lundgren and L. M. Carlred.

■

REFERENCES

(1) Zhang, R.; Xie, X. Tools for GPCR drug discovery. Acta Pharmacol. Sin. 2012, 33, 372−384. (2) Loetscher, M.; Gerber, B.; Loetscher, P.; Jones, S. A.; Piali, L.; ClarkLewis, I.; Baggiolini, M.; Moser, B. Chemokine receptor specific for IP10 and Mig: Structure, function, and expression in activated Tlymphocytes. J. Exp. Med. 1996, 184, 963−969. (3) Trentin, L.; Agostini, C.; Facco, M.; Piazza, F.; Perin, A.; Siviero, M.; Gurrieri, C.; Galvan, S.; Adami, F.; Zambello, R.; Semenzato, G. The chemokine receptor CXCR3 is expressed on malignant B cells and mediates chemotaxis. J. Clin. Invest. 1999, 104, 115−121. (4) Heise, C. E.; Pahuja, A.; Hudson, S. C.; Mistry, M. S.; Putnam, A. L.; Gross, M. M.; Gottlieb, P. A.; Wade, W. S.; Kiankarimi, M.; Schwarz, D.; Crowe, P.; Zlotnik, A.; Alleva, D. G. Pharmacological characterization of CXC chemokine receptor 3 ligands and a small molecule antagonist. J. Pharmacol. Exp. Ther. 2005, 313, 1263−1271. (5) Loetscher, M.; Loetscher, P.; Brass, N.; Meese, E.; Moser, B. Lymphocyte-specific chemokine receptor CXCR3: regulation, chemokine binding and gene localization. Eur. J. Immunol. 1998, 28, 3696− 3705. (6) Singh, A. K.; Arya, R. K.; Trivedi, A. K.; Sanyal, S.; Baral, R.; Dormond, O.; Briscoe, D. M.; Datta, D. Chemokine receptor trio: CXCR3, CXCR4 and CXCR7 crosstalk via CXCL11 and CXCL12. Cytokine Growth Factor Rev. 2013, 24, 41−49. (7) Watson, R. J.; Allen, D. R.; Birch, H. L.; Chapman, G. A.; Galvin, F. C.; Jopling, L. A.; Knight, R. L.; Meier, D.; Oliver, K.; Meissner, J. W. G.; Owen, D. A.; Thomas, E. J.; Tremayne, N.; Williams, S. C. Development of CXCR3 antagonists. Part 3: Tropenyl and homotropenyl-piperidine urea derivatives. Bioorg. Med. Chem. Lett. 2008, 18, 147−151. (8) Guerfal, M.; Claes, K.; Knittelfelder, O.; De Rycke, R.; Kohlwein, S. D.; Callewaert, N. Enhanced membrane protein expression by engineering increased intracellular membrane production. Microb. Cell Fact. 2013, 12, 122. (9) Muller, G. Towards 3D structures of G protein-coupled receptors: A multidisciplinary approach. Curr. Med. Chem. 2000, 7, 861−888. F

DOI: 10.1021/acs.langmuir.5b02463 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir (29) Clark-Lewis, I.; Mattioli, I.; Gong, J. H.; Loetscher, P. Structurefunction relationship between the human chemokine receptor CXCR3 and its ligands. J. Biol. Chem. 2002, 278, 289−295. (30) Nedjai, B.; Li, H.; Stroke, I. L.; Wise, E. L.; Webb, M. L.; Merritt, J. R.; Henderson, I.; Klon, A. E.; Cole, A. G.; Horuk, R.; Vaidehi, N.; Pease, J. E. Small molecule chemokine mimetics suggest a molecular basis for the observation that CXCL10 and CXCL11 are allosteric ligands of CXCR3. Br. J. Pharmacol. 2012, 166, 912−923. (31) Wang, X. K.; Li, X.; Schmidt, D. B.; Foley, J. J.; Barone, F. C.; Ames, R. S.; Sarau, H. M. Identification and molecular characterization of rat CXCR3: Receptor expression and interferon-inducible protein10 binding are increased in focal stroke. Mol. Pharmacol. 2000, 57, 1190−1198. (32) Limpert, E.; Stahel, W. A.; Abbt, M. Log-normal distributions across the sciences: Keys and clues. BioScience 2001, 51, 341−352. (33) Gunnarsson, A.; Snijder, A.; Hicks, J.; Gunnarsson, J.; Höök, F.; Geschwindner, S. Drug Discovery at the Single Molecule Level: Inhibition-in-Solution Assay of Membrane-Reconstituted β-Secretase Using Single-Molecule Imaging. Anal. Chem. 2015, 87, 4100−4103.

G

DOI: 10.1021/acs.langmuir.5b02463 Langmuir XXXX, XXX, XXX−XXX