- Email: [email protected]
The role of cholesterol on the activity and stability of neurotensin receptor 1
Biochimica et Biophysica Acta 1818 (2012) 2228–2233
Contents lists available at SciVerse ScienceDirect
Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbamem
The role of cholesterol on the activity and stability of neurotensin receptor 1 Joanne Oates, Belinda Faust, Helen Attrill, Peter Harding, Marcella Orwick, Anthony Watts ⁎ Biomembrane Structure Unit, Department of Biochemistry, University of Oxford, South Parks Road, Oxford, OX1 3QU, UK
a r t i c l e
i n f o
Article history: Received 10 November 2011 Received in revised form 15 March 2012 Accepted 12 April 2012 Available online 20 April 2012 Keywords: Neurotensin GPCR Cholesterol Reconstitution POPE FRET
a b s t r a c t Understanding the role of specific bilayer components in controlling the function of G-protein coupled receptors (GPCRs) will be a key factor in the development of novel pharmaceuticals. Cholesterol-dependence in particular has become an area of keen interest with respect to GPCR function; not least since the 2.6 Å crystal structure of the β2 adrenergic receptor revealed a putative cholesterol binding motif conserved throughout class-A GPCRs. Furthermore, experimental evidence for cholesterol-dependent GPCR function has been demonstrated in a limited number of cases. This modulation of receptor function has been attributed to both direct interactions between cholesterol and receptor, and indirect effects caused by the influence of cholesterol on bilayer order and lateral pressure. Despite the widespread occurrence of cholesterol binding motifs, available experimental data on the functional involvement of cholesterol on GPCRs are currently limited to a small number of receptors. Here we investigate the role of cholesterol in the function of the neurotensin receptor 1 (NTS1) a class‐A GPCR. Specifically we show how cholesterol, and the analogue cholesteryl hemisuccinate, influence activity, stability, and oligomerisation of both purified and reconstituted NTS1. The results caution against using such motifs as indicators of cholesterol‐dependent GPCR activity. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Found in a wide range of organisms and controlling many essential biological functions, the family of G‐protein-coupled receptors (GPCRs) forms one of the largest classes of membrane proteins. The vast numbers of GPCRs (over 800 in human cells [1]) and their central role in cell communication and signalling have made them the target of numerous pharmaceutical agents as a way of treating and controlling disease. Currently over 50% of pharmaceutical compounds on the market are designed to modulate GPCR function [2,3]. Their potential as therapeutic targets has been slowed due primarily to our poor understanding of GPCR structure and mode of action. The lipid environment of the cell membrane presents a challenge not only in the study of GPCRs but membrane proteins in general. Experimental approaches are typically hampered by low yields of purified protein and the need to maintain and/or replicate the natural lipid environment by the use of detergents. Understanding which components of the lipid bilayer are critical in maintaining GPCRs in the appropriate, biologically-relevant state is of paramount importance for structure–function studies, and therefore in the design of novel pharmaceuticals (for a recent review of membrane GPCR interactions see Ref. [4]. One particular component of the bilayer—cholesterol, is predicted to be important for the functional and structural integrity
⁎ Corresponding author. Tel.: + 44 1865 613219; fax: + 44 1865 275234/275259. E-mail address: [email protected] (A. Watts). 0005-2736/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2012.04.010
of GPCRs. Evidence for this comes from a variety of data including the recent crystal structure of the β2 adrenergic receptor [5,6] in which two cholesterol monomers were found in a cleft between helices I, II, III, and IV [5] and are proposed to interact with the receptor though a highly conserved cholesterol binding motif; more recently, the crystal structure of the β1 adrenergic receptor was also shown to contain two cholesterol monomers, one at an identical site and one at an alternative site [7]. A specific interaction with cholesterol has been determined for a number of other GPCRs including rhodopsin, and the oxytocin, galanin and serotonin1A receptors [8–13]. Cholesterol has also been demonstrated to indirectly affect GPCR function by altering bilayer properties such as phospholipid order and lateral pressure [14–19]. Evidence for the role of cholesterol comes primarily from experiments using cholesterol modulators such as methyl-β-cyclodextrin which can both deplete and restore membrane cholesterol or cholesterol analogues. The resulting effect of cholesterol on receptor activity can be to either up- or down-regulate ligand binding, or in the case of detergentsolubilised receptors stabilise the protein [20,21]. This latter observation has been critical for the production of suitably diffracting crystals including those of β2 adrenergic and adenosine A2a receptors [6,22]. Despite evidence pointing to a likely role for cholesterol in GPCR structure and function, no universal or generalised conclusions can be drawn from the available data, thus the specific effect of cholesterol on each receptor cannot readily be predicted. Furthermore a number of GPCRs can be functionally expressed in the E. coli membrane [20,21,23], which does not contain cholesterol (or other sterols) [24].
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
We sought to investigate the role of cholesterol on the activity and stability of a clinically important GPCR; neurotensin receptor 1 (NTS1). NTS1 responds to the binding of the peptide hormone neurotensin, and has roles in the central nervous system and gastrointestinal tract. Diseases such as Alzheimer's, Parkinson's, and certain cancers can be attributable to errors involving NTS1-dependent pathways (reviewed in Refs. [25,26]). The development of successful purification and reconstitution strategies for NTS1, along with assays for function, mean that we are now in a position to investigate the role of membrane components, and in particular cholesterol, in NTS1 activity and stability [27–29]. Early work has shown that the choice of detergent for NTS1 solubilisation and stability is critical for recovery of active receptors in cell extracts, in particular the presence of the cholesterol analogue cholesteryl hemisuccinate (CHS) is believed to have a stabilising effect [21]. Here we carry out the first investigation of the effect of cholesterol on purified and reconstituted NTS1. We find the presence of CHS to have a small role in stabilising purified NTS1 in the presence of detergents, particularly during ligand affinity chromatography. However, in a lipid environment we find cholesterol unnecessary for ligand binding and oligomerisation. 2. Methods 2.1. Purification of NTS1B and NTS1-eYFP/eCFP NTS1B and NTS1-eYFP/eCFP fusion proteins have been described previously [27,28,30]. Purification of NTS1 eYFP/eCFP was carried out as described [28]. NTS1B was purified using immobilised metal affinity chromatography (IMAC) and a neurotensin (NT) ligand affinity column two-step protocol similar to that described in Ref. [29], with the final buffer containing 50 mM Tris pH 7.4, 15% glycerol (v/v), 250 mM NaCl, 0.1% DDM (w/v), 0.01% CHS (w/v), and 1 mM EDTA. Purification of NTS1B in the absence of CHS was carried out in the same manner using buffers containing DDM-only throughout solubilisation and purification processes with the final buffer as that detailed above without CHS. Protein concentration was determined by absorbance at 280 nm using calculated extinction coefficients, or by gel densitometry. 2.2. Reconstitution of NTS1B and NTS1-eYFP/eCFP Phospholipids 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE) and porcine brain polar lipid extract (BPL) were purchased from Avanti Polar Lipids, cholesterol was purchased from Sigma. Phospholipids were prepared by mixing in equimolar amount. Cholesterol was included at a 25% molar content. Lipids were prepared and detergent-disrupted as described previously [28]. For FRET experiments, NTS1-eYFP and NTS1-eCFP were mixed at a 1:1 molar ratio, added to disrupted vesicles at a protein lipid molar ratio of 1:6000 and incubated for 1 h at 4 °C. To remove detergent, 120 mg/ ml Biobeads SM-2 (Bio-Rad) was added and the sample agitated overnight at 4 °C. Proteoliposomes were isolated using sucrose gradient density ultracentrifugation (100,000 ×g; 4 °C for 15 h in a swingout rotor). NTS1B was reconstituted in the same manner except a protein lipid molar ratio of 1:4000 was used, and proteoliposomes isolated by ultracentrifugation in a Beckman TLX table-top ultracentrifuge (100,000 ×g, for 3 h at 4 °C). 2.3. Cholesterol depletion from brain polar lipid (BPL) vesicles NTS1B BPL proteoliposomes were prepared as described above. To extract cholesterol, a stock solution of 40 mM methyl-β-cyclodextrin (MβCD) (Sigma) was freshly prepared in reconstitution buffer. The pelleted, reconstituted receptor (1.2 μM) was resuspended in 250 μl 0, 10, 20 or 40 mM methyl-β-cyclodextrin and stirred for 2–3 h at
2229
room temperature. The proteoliposomes were then pelleted and resuspended in 100 μl reconstitution buffer. Activity was measured as detailed below and cholesterol content assayed using Amplex® Red Cholesterol Assay Kit (Invitrogen). 2.4. Fluorescence resonance energy transfer Fluorescence resonance energy transfer (FRET) was measured using a LS-55 Spectrofluorimeter (Perkin Elmer). Co-reconstituted NTS-eYFP/eCFP (900 μl) was placed in a 1.5 ml quartz microcuvette (Hellma) and measurements conducted at 4 °C with stirring. Excitation and emission slits were both set to 2.5 nm. Apparent FRET efficiency (FRETapp) was calculated as described and values of 10 are considered to be indicative of dimer formation [28]. 2.5. Ligand-binding analysis of purified and reconstituted receptor To quantify the total amount of active (ligand binding) receptor in either detergent solution [27] or lipid bilayers [28], a radioligand saturation assay was conducted as described. The assay buffer for protein purified in the absence CHS also contained no CHS. The stability of ligand binding was assessed over an 8 h period by incubating samples at room temperature for the indicated times, followed by rapid freezing in liquid nitrogen and storage at – 80° C until required. Standard errors were calculated from experiments carried out in triplicate. 2.6. Membrane order Membrane order of POPC/POPE liposomes with and without cholesterol was monitored by continuous-wave electron paramagnetic resonance (CW-EPR). Liposomes were prepared as described above to a final concentration of 20 mg/ml of lipid. In addition ~ 1 mol% of 8-DOXYL-stearic acid was added to the samples as the reporter molecule. EPR measurements were performed at 9 GHz on a Bruker EMX spectrometer. The degree of order in the membrane reported by a stearic acid spin nitroxide can be defined by an order parameter S [31], (Eq. (1)), which is calculated from the observed hyperfine anisotropy (A∥–A⊥), measured from the spectrum, to a maximum hyperfine anisotropy (Azz–Axx), for a maximum rigidity (S = 1), calculated from a frozen spectrum of the same sample to correct for polarity effects. S¼
Ajj −A⊥ Azz −Axx
ð1Þ
3. Results 3.1. Cholesterol motifs in NTS1 The observed bound, or trapped, cholesterol in the crystal structure of the β2 adrenergic receptor has led to the proposal of a specific cholesterol binding motif, highly conserved throughout the primary sequences of GPCRs of this family. Known as the conserved cholesterol motif (CCM), this region comprises one residue from helix II, and three from IV. Tryptophan 4.50 (Ballesteros–Weinstein nomenclature [32]) is predicted to interact with the sterol ring, as well as isoleucine 4.46 on the same helix and tyrosine 2.41. A charged residue (arginine 4.43) at the intercellular face of helix IV is proposed to form the strongest interaction with the cholesterol in terms of binding energy, through an electrostatic interaction with the 3β hydroxyl moiety of the cholesterol. Fig. 1a shows the CCM motif identified in NTS1 [from our unpublished model] in comparison to the same motif in the β2 adrenergic receptor [5] with the residues involved highlighted. An alternative cholesterol binding motif, the rather generic CRAC
2230
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
motif (cholesterol recognition / interaction amino acid consensus), has been found in a number of membrane proteins known to specifically interact with cholesterol [33,34]. Recently this motif has been highlighted in GPCRs including the serotonin1A receptor and the β2 adrenergic receptor [35]. Here we also find this motif in two locations in the NTS1 sequence (Fig. 1b). Despite the presence of these motifs there is little experimental evidence for their importance in determining or influencing function, or whether they can be used as a predictor for specific cholesterol binding. 3.2. Purified NTS1 is active in the presence and absence of cholesteryl hemisuccinate (CHS) Purification of NTS1 was carried out in buffer containing micelles of DDM-only or a combination of DDM, CHAPS and CHS (addition of CHAPS promotes efficient solubilisation of CHS). The final step of purification utilises a neurotensin (NT) ligand affinity column. It was found that NTS1 prepared in the presence and absence of CHS was effectively isolated by ligand affinity chromatography, implying ligand binding is not dependent on CHS. In both cases the yield of receptor was similar (see gel in Fig. 2a); however, the receptor prepared without CHS was not as pure (Fig. 2a), possibly due to decreased stability during purification, with the contaminating bands on the gel representing degradation products and aggregation caused by inefficient solubilisation. Assessment of ligand binding using tritiated NT highlighted a ~75% reduction in activity in the DDM-only preparation compared to the CHS containing preparation, once normalised for protein concentration (Fig. 2b). One reason for the discrepancy between comparable receptor binding to the NT column and the 75% difference observed in the NT binding assay could be the reduced stability of the CHS-free receptor preparation after the NT chromatography step. To investigate this further the stability of ligand binding to both preparations CHS containing and CHS free was assessed over an 8 h period at room temperature. Fig. 2c shows 14% activity remaining in the CHS containing preparation while the CHS-free sample is reduced to 4%. Therefore CHS is contributing a small amount to receptor stability at room temperature; however, this doesn't fully explain the observed decrease in ligand binding of the CHS-free preparation (75%). It is possible that the high local concentration of receptor on the NT affinity column may lead to the isolation of inactive receptors in the absence of CHS. Indeed, it has been observed recently [36] that proximity to the NT resin during chromatography can destabilise NTS1, possibly
Fig. 2. Analysis of NTS1 purified in the presence and absence of cholesteryl hemisuccinate (CHS): a) Coomassie-stained SDS PAGE gel of NTS1B purified with (+) and without (−) CHS. The arrow indicates NTS1B; b) Difference in ligand binding activity of NTS1B in terms of specific and relative activity; c) Stability of NTS1B ligand binding activity over time at room temperature. In all cases the ligand used was tritiated neurotensin (Perkin Elmer), at a final concentration of 8.33 nM.
due to detergent-related perturbations [36], an effect which may be enhanced during purification without CHS. 3.3. Reconstituted NTS1 is ligand binding in the presence and absence of cholesterol Although micelles are useful tools for in vitro studies of membrane proteins, an investigation into the effect of cholesterol on NTS1 in a
Fig. 1. Cholesterol binding motifs in NTS1: a) Detail of the CCM binding motif from the crystal structure of β2 adrenergic receptor with bound cholesterol (PDB ID: 3D4S) (left) and a model of the corresponding region of NTS1. Residues forming the CCM motif are highlighted, numbered in Ballesteros–Weinstein nomenclature [32]. Images produced using Pymol (http://pymol.sourceforge.net/). b) The cholesterol recognition amino acid consensus motif (CRAC) and underneath the two identified sequences in NTS1, with the position of the first residue in brackets.
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
2231
ligand‐binding properties of NTS1, and thus a direct interaction of cholesterol with the receptor may not be required for activity. It can be concluded that other components of BPL must be required to regain full ligand-binding activity, either through a direct receptor interaction or due to alteration of bilayer properties such as curvature [37] or thickness [38]. 3.4. Dimerisation in the absence of cholesterol Previously we showed NTS1 forms constitutive homo-dimers in BPL whereas, in detergent, dimerisation could only be detected at high protein–detergent ratios [28,39]. Therefore we proposed that particular components in the bilayer may contribute directly to dimerisation. Using the fluorescently-tagged receptors NTS1Y and NTS1C described previously [28], dimerisation was probed in the minimal lipid environment identified above, both in the presence and absence of cholesterol. The results of FRET measurements (Fig. 3b) show that dimerisation reported by FRET efficiency, is unaffected by the presence of cholesterol in a POPC/POPE background. However, with POPC the addition of cholesterol does lead to an increase in FRET efficiency to a level comparable to POPC/POPE. Therefore either cholesterol or POPE facilitates dimerisation. 3.5. Cholesterol influences membrane order in POPC/POPE lipids
Fig. 3. Activity and dimerisation of NTS1 reconstituted in the presence and absence of cholesterol: a) NTS1B purified both with and without cholesteryl hemisuccinate (CHS) was reconstituted into five different lipid mixtures consisting of: brain polar lipid (BPL); 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC); 1-palmitoyl-2oleoyl-sn-glycero-3-phosphoethanolamine (POPE); cholesterol (Ch), the concentration of NTS1 was 1.2 μM. Ligand binding is reported relative to BPL; b) Dimerisation of NTS1 tagged with eCFP or eYFP was measured using FRET in the same lipid mixtures as above. Apparent FRET efficiency was calculated as previously reported [28] the protein. The total amount of NTS1C and NTS1Y used in each experiment was 175 pMol.
biologically-relevant lipid environment was carried out. We have previously reported successful reconstitution of NTS1 into liposomes and demonstrated ligand binding as well as dimerisation [28]. Here, NTS1 was reconstituted into various lipid mixtures in the presence and absence of 25 mol% cholesterol and ligand‐binding activity measured. We used brain polar lipid extract (BPL), which represents the native environment for the receptor, along with more simple mixtures composed of POPC/POPE. PC and PE lipids were chosen since together they comprise around half of the lipids in BPL [Avanti Polar Lipids, product information]. Fig. 3a shows the comparative ligand-binding activity of reconstituted NTS1. BPL represents the most effective lipid in supporting ligand binding. In lipid mixtures of POPC/POPE NTS1 ligand binding is only 40% as effective as in BPL; however ligand binding in the presence and absence of cholesterol is not affected (Fig. 3a), therefore suggesting that cholesterol is not required for ligand binding activity when NTS1 is reconstituted into a POPC/POPE environment. Furthermore, since ligand binding in POPC is below the level of detection, it is proposed that POPC/POPE provides a minimal mix for functional receptor reconstitution. The NTS1 used in these experiments was prepared in the presence of CHS, therefore it is possible that it was not removed during the reconstitution process (by adsorption on to Biobeads) and effectively substitutes for cholesterol. To address this possibility we repeated the experiment using receptor purified in CHSfree buffer (see Section 3.2). Fig. 3a shows little change in measured ligand‐binding activity, and it is therefore concluded that in a POPC/ POPE background the presence of cholesterol does not affect the
Changes in ordering of phospholipid acyl chains, caused by the presence of cholesterol, have been reported to influence function for some GPCRs [14,40]. From the functional studies presented above, NTS1 appears unaffected by the presence of cholesterol. To confirm cholesterol is affecting membrane order in our system continuous‐ wave electron spin resonance (CW-ESR) combined with a spinlabelled stearic acid was used to report on the order within the lipid acyl chains. From the resulting spectra (Fig. 4) the order parameter was determined. As predicted, the membranes became more ordered in the presence of cholesterol [41]. These data confirm that changes in membrane order are unlikely to be affecting NTS1 function, when compared with the functional data (Section 3.4). 3.6. Effect of cholesterol depletion Methyl-β-cyclodextrin (MβCD), a cyclic glucose oligomer, has been extensively employed in the study of GPCRs with cholesterol, and it has been shown to sequester hydrophobic molecules such as cholesterol out of the bilayer [42]. Although these experiments tend
S:
-
0.68
+
0.78
3300
3350
3400
Field [G] Fig. 4. Membrane order of POPC/POPE with and without 25 mol% cholesterol: EPR spectra were recorded at 294 K on bilayers in the presence (+) and absence (−) of cholesterol, and represent an average of 10 scans. The order parameter (S) was calculated as described in the Methods section. The dotted lines highlight the change in the parallel component of the spectra when cholesterol is present.
2232
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
Fig. 5. Cholesterol extraction using methyl beta cyclodextrin (MβCD): a) NTS1B reconstituted into bilayers of brain polar lipid was incubated with different concentrations of MβCD. Proteoliposomes were subsequently re-isolated by ultracentrifugation and ligand binding activity measured. Protein concentration was determined by gel densitometry using Image J and cholesterol concentration was calculated using the Amplex Red kit from Invitrogen. b) The same data plotted to show NTS1 specific activity as a function of protein concentration.
to be carried out in in vivo systems, here MβCD was used to remove cholesterol from NTS1 reconstituted into BPL liposomes as an alternative approach to that presented in Section 3.4 (in which cholesterol was added into the lipid bilayers). After incubation with increasing concentrations of MβCD, liposomes were recovered by ultracentrifugation, and ligand‐binding activity measured. The Amplex-Red cholesterol assay kit (Invitrogen) was used to confirm the reduction in cholesterol content in the bilayers. The results (Fig. 5a) show that although relative activity (recovery) is reduced by 40% as MβCD concentration increases, this correlates with a reduction of 40% in relative protein recovery. An additional plot of cholesterol concentration versus specific activity clearly demonstrates that the specific activity of NTS1 is unaffected by cholesterol concentration (Fig. 5b). Therefore, in agreement with the data above (Section 3.4), ligand binding to NTS1 is not influenced by the presence of cholesterol in bilayers. 4. Discussion The dependence of GPCR stability and function on the presence of CHS has been previously reported in a small number of cases [5,21,43,44]. Compared with NTS1, the closely related β2 adrenergic receptor is also active in the absence of CHS, although it is significantly more stable, given that a 5-fold increase in the half-life of protein denaturation is observed upon addition of CHS [5]. Furthermore, the stability of the adenosine A2a receptor has also been linked to the presence of CHS, although, unlike NTS1 and β2 adrenergic
receptor, this receptor appears completely inactive in DDM micelles without CHS [44]. It is proposed that the presence of CHS modifies the dimensions of the micelle environment to match those of the biological membrane for the A2a receptor [43]. However, it is important to point out that the adenosine A2a receptor along with NTS1 and the β2 adrenergic receptor are all capable of ligand binding in the E. coli membrane—which does not contain cholesterol [20,21,23]. Therefore inferring a biological significance to CHS dependent effects from this observation is difficult. Clearly there is a stabilisation advantage to the use of CHS in structural determination of GPCRs, but the inference that cholesterol is therefore important for activity must be investigated in a well-defined lipid environment. The data presented here show NTS1 is active and relatively stable when purified in the absence of CHS (Fig. 2b and c), indicating no significant effect of the cholesterol analogue on receptor function. However, NTS1 function does appear sensitive to the inclusion of CHS during the course of the purification process (Fig. 2a and b). These results are in contrast to previous experiments with NTS1 from detergent‐solubilised cell extracts, where the absence of CHS reduced receptor yield and half-life of ligand binding by over 50% [21]. These differences may be explained by the two distinct environments employed in these studies; detergent-solubilised cell extracts in the former work compared with purified protein investigated here. As we are interested specifically in purified systems in a well-defined lipid environment we focused on characterising the purified receptor in this environment, free of any other components. We also demonstrate that once reconstituted, the presence of cholesterol appears to have little influence on the activity and dimerisation of NTS1 in the lipid mixes used here (Fig. 3). Although activity was clearly reduced in POPC/POPE compared to the more native BPL mix, this could not be attributed specifically to the absence of cholesterol. The influence of cholesterol on the activity of other GPCRs is believed to arise in part from the ordering effect that cholesterol imparts on the lipid acyl chains [14,15]. We confirm this effect with EPR measurements on the POPC/POPE bilayers with and without cholesterol, and conclude that changes in phospholipid order are not influencing NTS1 activity or dimerisation. Dimerisation in terms of FRET efficiency was restored from a POPC background in the presence of either POPE or cholesterol, unlike activity which was not restored by cholesterol. These results therefore suggest that dimerisation and ligand binding are not linked processes. The role of particular lipids in GPCR structure and function remains poorly characterised at present, although POPE interactions have been linked to transitions between the MI and MII states of rhodopsin [45,46]. POPE is non‐ lamellar forming and as such alters the physical properties of the bilayer; it would therefore be interesting to further study how NTS1 is influenced by POPE. To further confirm cholesterol does not influence NTS1 ligand‐ binding function MCβD was used to deplete membrane cholesterol from NTS1 reconstituted into BPL. Despite having a non-polar cavity, MCβD has been shown to specifically remove cholesterol from bilayers [47,48]. Here we show that although cholesterol was removed and NTS1 ligand binding decreased, protein concentration also decreased at a rate closely linked to activity. The reason for this loss of protein is unclear but care must be taken in these types of experiments to confirm protein concentration. The overall conclusions from these depletion studies however, complement well with the cholesterol addition studies. Based on the results presented here we find no correlation between the presence of cholesterol binding motifs in NTS1 and cholesterol-dependent ligand-binding activity; in this case we are not able to comment on G‐protein activation. Recent data with other GPCRs have led to the proposal that cholesterol and GPCR function (both ligand binding and downstream signalling) are connected, which is clearly the case for the oxytocin receptor for example [14]. It has also been suggested that in vivo, cholesterol is required to
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
regulate signalling by locating the receptor to specific areas of the membrane, for example in lipid-microdomains or caveolae [49]. Specific membrane location has been reported with NTS1 in vivo [50]. Further work will be required to reveal the molecular details of cholesterol-regulated GPCR activity, as well as the influence of other components of the bilayer, but we conclude that cholesterol is not required for ligand-binding activity of purified and reconstituted NTS1 in a well-defined lipid environment. Acknowledgements Funding was provided by the Medical Research Council (MRC) (Grant no. G0900076 to A.W.), and the Engineering and Physical Sciences Research Council (EPSRC) (for support of CAESR (EP/D048559/ 1), A.W. co-applicant). References [1] R. Fredriksson, H.B. Schioth, The repertoire of G-protein-coupled receptors in fully sequenced genomes, Mol. Pharmacol. 67 (2005) 1414–1425. [2] A.D. Howard, G. McAllister, S.D. Feighner, Q. Liu, R.P. Nargund, L.H. Van der Ploeg, A.A. Patchett, Orphan G-protein-coupled receptors and natural ligand discovery, Trends Pharmacol. Sci. 22 (2001) 132–140. [3] T. Klabunde, G. Hessler, Drug design strategies for targeting G-protein-coupled receptors, Chembiochem 3 (2002) 928–944. [4] J. Oates, A. Watts, Uncovering the intimate relationship between lipids, cholesterol and GPCR activation, Curr. Opin. Struct. Biol. 21 (2011) 802–807. [5] M.A. Hanson, V. Cherezov, M.T. Griffith, C.B. Roth, V.P. Jaakola, E.Y. Chien, J. Velasquez, P. Kuhn, R.C. Stevens, A specific cholesterol binding site is established by the 2.8 A structure of the human beta2-adrenergic receptor, Structure 16 (2008) 897–905. [6] V. Cherezov, D.M. Rosenbaum, M.A. Hanson, S.G. Rasmussen, F.S. Thian, T.S. Kobilka, H.J. Choi, P. Kuhn, W.I. Weis, B.K. Kobilka, R.C. Stevens, High-resolution crystal structure of an engineered human beta2-adrenergic G protein-coupled receptor, Science 318 (2007) 1258–1265. [7] T. Warne, R. Moukhametzianov, J.G. Baker, R. Nehme, P.C. Edwards, A.G. Leslie, G.F. Schertler, C.G. Tate, The structural basis for agonist and partial agonist action on a beta(1)-adrenergic receptor, Nature 469 (2011) 241–244. [8] G. Gimpl, F. Fahrenholz, Cholesterol as stabilizer of the oxytocin receptor, Biochim. Biophys. Acta 1564 (2002) 384–392. [9] G. Gimpl, J. Reitz, S. Brauer, C. Trossen, Oxytocin receptors: ligand binding, signalling and cholesterol dependence, Prog. Brain Res. 170 (2008) 193–204. [10] G. Gimpl, V. Wiegand, K. Burger, F. Fahrenholz, Cholesterol and steroid hormones: modulators of oxytocin receptor function, Prog. Brain Res. 139 (2002) 43–55. [11] L. Pang, M. Graziano, S. Wang, Membrane cholesterol modulates galanin–GalR2 interaction, Biochemistry 38 (1999) 12003–12011. [12] T.J. Pucadyil, A. Chattopadhyay, Cholesterol modulates ligand binding and Gprotein coupling to serotonin(1A) receptors from bovine hippocampus, Biochim. Biophys. Acta 1663 (2004) 188–200. [13] A.D. Albert, J.E. Young, P.L. Yeagle, Rhodopsin–cholesterol interactions in bovine rod outer segment disk membranes, Biochim. Biophys. Acta 1285 (1996) 47–55. [14] G. Gimpl, K. Burger, F. Fahrenholz, Cholesterol as modulator of receptor function, Biochemistry 36 (1997) 10959–10974. [15] D.C. Mitchell, M. Straume, J.L. Miller, B.J. Litman, Modulation of metarhodopsin formation by cholesterol-induced ordering of bilayer lipids, Biochemistry 29 (1990) 9143–9149. [16] R. Prasad, P. Singh, A. Chattopadhyay, Effect of capsaicin on ligand binding activity of the hippocampal serotonin1A receptor, Glycoconj. J. 26 (2009) 733–738. [17] E.S. Levitt, M.J. Clark, P.M. Jenkins, J.R. Martens, J.R. Traynor, Differential effect of membrane cholesterol removal on mu- and delta-opioid receptors: a parallel comparison of acute and chronic signaling to adenylyl cyclase, J. Biol. Chem. 284 (2009) 22108–22122. [18] P. Huang, W. Xu, S.I. Yoon, C. Chen, P.L. Chong, L.Y. Liu-Chen, Cholesterol reduction by methyl-beta-cyclodextrin attenuates the delta opioid receptor-mediated signaling in neuronal cells but enhances it in non-neuronal cells, Biochem. Pharmacol. 73 (2007) 534–549. [19] A.D. Albert, K. Boesze-Battaglia, Z. Paw, A. Watts, R.M. Epand, Effect of cholesterol on rhodopsin stability in disk membranes, Biochim. Biophys. Acta 1297 (1996) 77–82. [20] H.M. Weiss, R. Grisshammer, Purification and characterization of the human adenosine A(2a) receptor functionally expressed in Escherichia coli, Eur. J. Biochem. 269 (2002) 82–92. [21] J. Tucker, R. Grisshammer, Purification of a rat neurotensin receptor expressed in Escherichia coli, Biochem. J. 317 (Pt 3) (1996) 891–899. [22] V.P. Jaakola, M.T. Griffith, M.A. Hanson, V. Cherezov, E.Y. Chien, J.R. Lane, A.P. Ijzerman, R.C. Stevens, The 2.6 angstrom crystal structure of a human A2A adenosine receptor bound to an antagonist, Science (2008) 1211–1217. [23] S. Marullo, C. Delavier-Klutchko, Y. Eshdat, A.D. Strosberg, L. Emorine, Human beta 2-adrenergic receptors expressed in Escherichia coli membranes retain their pharmacological properties, Proc. Natl. Acad. Sci. U. S. A. 85 (1988) 7551–7555.
2233
[24] S. Morein, A. Andersson, L. Rilfors, G. Lindblom, Wild-type Escherichia coli cells regulate the membrane lipid composition in a "window" between gel and nonlamellar structures, J. Biol. Chem. 271 (1996) 6801–6809. [25] R.M. Myers, J.W. Shearman, M.O. Kitching, A. Ramos-Montoya, D.E. Neal, S.V. Ley, Cancer, chemistry, and the cell: molecules that interact with the neurotensin receptors, ACS Chem. Biol. 4 (2009) 503–525. [26] S. Dupouy, N. Mourra, V.K. Doan, A. Gompel, M. Alifano, P. Forgez, The potential use of the neurotensin high affinity receptor 1 as a biomarker for cancer progression and as a component of personalized medicine in selective cancers, Biochimie 93 (9) (2011) 1369–1378. [27] H. Attrill, P.J. Harding, E. Smith, S. Ross, A. Watts, Improved yield of a ligandbinding GPCR expressed in E. coli for structural studies, Protein Expr. Purif. 64 (2009) 32–38. [28] P.J. Harding, H. Attrill, J. Boehringer, S. Ross, G.H. Wadhams, E. Smith, J.P. Armitage, A. Watts, Constitutive dimerization of the G-protein coupled receptor, neurotensin receptor 1, reconstituted into phospholipid bilayers, Biophys. J. 96 (2009) 964–973. [29] J.F. White, L.B. Trinh, J. Shiloach, R. Grisshammer, Automated large-scale purification of a G protein-coupled receptor for neurotensin, FEBS Lett. 564 (2004) 289–293. [30] R. Grisshammer, J.F. White, L.B. Trinh, J. Shiloach, Large-scale expression and purification of a G-protein-coupled receptor for structure determination—an overview, J. Struct. Funct. Genomics 6 (2005) 159–163. [31] P. Fretten, S.J. Morris, A. Watts, D. Marsh, Lipid–lipid and lipid–protein interactions in chromaffin granule membranes. A spin label ESR study, Biochim. Biophys. Acta 598 (1980) 247–259. [32] J.A. Ballesteros, H. Weinstein, C.S. Stuart, Integrated methods for the construction of three-dimensional models and computational probing of structure–function relations in G protein-coupled receptors, Meth. Neurosci., vol. 25, Academic Press, 1995, pp. 366–428. [33] R.M. Epand, A. Thomas, R. Brasseur, R.F. Epand, Cholesterol interaction with proteins that partition into membrane domains: an overview, Subcell. Biochem. 51 (2010) 253–278. [34] A.I. Greenwood, J. Pan, T.T. Mills, J.F. Nagle, R.M. Epand, S. Tristram-Nagle, CRAC motif peptide of the HIV-1 gp41 protein thins SOPC membranes and interacts with cholesterol, Biochim. Biophys. Acta 1778 (2008) 1120–1130. [35] M. Jafurulla, S. Tiwari, A. Chattopadhyay, Identification of cholesterol recognition amino acid consensus (CRAC) motif in G-protein coupled receptors, Biochem. Biophys. Res. Commun. 404 (2011) 569–573. [36] J.F. White, R. Grisshammer, Stability of the neurotensin receptor NTS1 free in detergent solution and immobilized to affinity resin, PLoS One 5 (2010) e12579. [37] K.L. Madsen, V.K. Bhatia, U. Gether, D. Stamou, BAR domains, amphipathic helices and membrane-anchored proteins use the same mechanism to sense membrane curvature, FEBS Lett. 584 (2010) 1848–1855. [38] T.M. Weiss, P.C. van der Wel, J.A. Killian, R.E. Koeppe 2nd, H.W. Huang, Hydrophobic mismatch between helices and lipid bilayers, Biophys. J. 84 (2003) 379–385. [39] J.F. White, J. Grodnitzky, J.M. Louis, L.B. Trinh, J. Shiloach, J. Gutierrez, J.K. Northup, R. Grisshammer, Dimerization of the class A G protein-coupled neurotensin receptor NTS1 alters G protein interaction, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 12199–12204. [40] K.G. Harikumar, V. Puri, R.D. Singh, K. Hanada, R.E. Pagano, L.J. Miller, Differential effects of modification of membrane cholesterol and sphingolipids on the conformation, function, and trafficking of the G protein-coupled cholecystokinin receptor, J. Biol. Chem. 280 (2005) 2176–2185. [41] J.L. Rubenstein, J.C. Owicki, H.M. McConnell, Dynamic properties of binary mixtures of phosphatidylcholines and cholesterol, Biochemistry 19 (1980) 569–573. [42] Y. Ohtani, T. Irie, K. Uekama, K. Fukunaga, J. Pitha, Differential effects of alpha-, beta- and gamma-cyclodextrins on human erythrocytes, Eur. J. Biochem. 186 (1989) 17–22. [43] M.A. O'Malley, M.E. Helgeson, N.J. Wagner, A.S. Robinson, The morphology and composition of cholesterol-rich micellar nanostructures determine transmembrane protein (GPCR) activity, Biophys. J. 100 (2011) L11–L13. [44] M.A. O'Malley, T. Lazarova, Z.T. Britton, A.S. Robinson, High-level expression in Saccharomyces cerevisiae enables isolation and spectroscopic characterization of functional human adenosine A2a receptor, J. Struct. Biol. 159 (2007) 166–178. [45] A.V. Botelho, N.J. Gibson, R.L. Thurmond, Y. Wang, M.F. Brown, Conformational energetics of rhodopsin modulated by nonlamellar-forming lipids, Biochemistry 41 (2002) 6354–6368. [46] O. Soubias, W.E. Teague, K. Gawrisch, Evidence for specificity in lipid–rhodopsin interactions, J. Biol. Chem. 281 (2006) 33233–33241. [47] G. Gimpl, Cholesterol–protein interaction: methods and cholesterol reporter molecules, Subcell. Biochem. 51 (2010) 1–45. [48] U. Klein, G. Gimpl, F. Fahrenholz, Alteration of the myometrial plasma membrane cholesterol content with beta-cyclodextrin modulates the binding affinity of the oxytocin receptor, Biochemistry 34 (1995) 13784–13793. [49] Y. Xiang, V.O. Rybin, S.F. Steinberg, B. Kobilka, Caveolar localization dictates physiologic signaling of beta 2-adrenoceptors in neonatal cardiac myocytes, J. Biol. Chem. 277 (2002) 34280–34286. [50] Y. Heakal, M. Kester, Nanoliposomal short-chain ceramide inhibits agonistdependent translocation of neurotensin receptor 1 to structured membrane microdomains in breast cancer cells, Mol. Cancer Res. 7 (2009) 724–734.
Contents lists available at SciVerse ScienceDirect
Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbamem
The role of cholesterol on the activity and stability of neurotensin receptor 1 Joanne Oates, Belinda Faust, Helen Attrill, Peter Harding, Marcella Orwick, Anthony Watts ⁎ Biomembrane Structure Unit, Department of Biochemistry, University of Oxford, South Parks Road, Oxford, OX1 3QU, UK
a r t i c l e
i n f o
Article history: Received 10 November 2011 Received in revised form 15 March 2012 Accepted 12 April 2012 Available online 20 April 2012 Keywords: Neurotensin GPCR Cholesterol Reconstitution POPE FRET
a b s t r a c t Understanding the role of specific bilayer components in controlling the function of G-protein coupled receptors (GPCRs) will be a key factor in the development of novel pharmaceuticals. Cholesterol-dependence in particular has become an area of keen interest with respect to GPCR function; not least since the 2.6 Å crystal structure of the β2 adrenergic receptor revealed a putative cholesterol binding motif conserved throughout class-A GPCRs. Furthermore, experimental evidence for cholesterol-dependent GPCR function has been demonstrated in a limited number of cases. This modulation of receptor function has been attributed to both direct interactions between cholesterol and receptor, and indirect effects caused by the influence of cholesterol on bilayer order and lateral pressure. Despite the widespread occurrence of cholesterol binding motifs, available experimental data on the functional involvement of cholesterol on GPCRs are currently limited to a small number of receptors. Here we investigate the role of cholesterol in the function of the neurotensin receptor 1 (NTS1) a class‐A GPCR. Specifically we show how cholesterol, and the analogue cholesteryl hemisuccinate, influence activity, stability, and oligomerisation of both purified and reconstituted NTS1. The results caution against using such motifs as indicators of cholesterol‐dependent GPCR activity. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Found in a wide range of organisms and controlling many essential biological functions, the family of G‐protein-coupled receptors (GPCRs) forms one of the largest classes of membrane proteins. The vast numbers of GPCRs (over 800 in human cells [1]) and their central role in cell communication and signalling have made them the target of numerous pharmaceutical agents as a way of treating and controlling disease. Currently over 50% of pharmaceutical compounds on the market are designed to modulate GPCR function [2,3]. Their potential as therapeutic targets has been slowed due primarily to our poor understanding of GPCR structure and mode of action. The lipid environment of the cell membrane presents a challenge not only in the study of GPCRs but membrane proteins in general. Experimental approaches are typically hampered by low yields of purified protein and the need to maintain and/or replicate the natural lipid environment by the use of detergents. Understanding which components of the lipid bilayer are critical in maintaining GPCRs in the appropriate, biologically-relevant state is of paramount importance for structure–function studies, and therefore in the design of novel pharmaceuticals (for a recent review of membrane GPCR interactions see Ref. [4]. One particular component of the bilayer—cholesterol, is predicted to be important for the functional and structural integrity
⁎ Corresponding author. Tel.: + 44 1865 613219; fax: + 44 1865 275234/275259. E-mail address: [email protected] (A. Watts). 0005-2736/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2012.04.010
of GPCRs. Evidence for this comes from a variety of data including the recent crystal structure of the β2 adrenergic receptor [5,6] in which two cholesterol monomers were found in a cleft between helices I, II, III, and IV [5] and are proposed to interact with the receptor though a highly conserved cholesterol binding motif; more recently, the crystal structure of the β1 adrenergic receptor was also shown to contain two cholesterol monomers, one at an identical site and one at an alternative site [7]. A specific interaction with cholesterol has been determined for a number of other GPCRs including rhodopsin, and the oxytocin, galanin and serotonin1A receptors [8–13]. Cholesterol has also been demonstrated to indirectly affect GPCR function by altering bilayer properties such as phospholipid order and lateral pressure [14–19]. Evidence for the role of cholesterol comes primarily from experiments using cholesterol modulators such as methyl-β-cyclodextrin which can both deplete and restore membrane cholesterol or cholesterol analogues. The resulting effect of cholesterol on receptor activity can be to either up- or down-regulate ligand binding, or in the case of detergentsolubilised receptors stabilise the protein [20,21]. This latter observation has been critical for the production of suitably diffracting crystals including those of β2 adrenergic and adenosine A2a receptors [6,22]. Despite evidence pointing to a likely role for cholesterol in GPCR structure and function, no universal or generalised conclusions can be drawn from the available data, thus the specific effect of cholesterol on each receptor cannot readily be predicted. Furthermore a number of GPCRs can be functionally expressed in the E. coli membrane [20,21,23], which does not contain cholesterol (or other sterols) [24].
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
We sought to investigate the role of cholesterol on the activity and stability of a clinically important GPCR; neurotensin receptor 1 (NTS1). NTS1 responds to the binding of the peptide hormone neurotensin, and has roles in the central nervous system and gastrointestinal tract. Diseases such as Alzheimer's, Parkinson's, and certain cancers can be attributable to errors involving NTS1-dependent pathways (reviewed in Refs. [25,26]). The development of successful purification and reconstitution strategies for NTS1, along with assays for function, mean that we are now in a position to investigate the role of membrane components, and in particular cholesterol, in NTS1 activity and stability [27–29]. Early work has shown that the choice of detergent for NTS1 solubilisation and stability is critical for recovery of active receptors in cell extracts, in particular the presence of the cholesterol analogue cholesteryl hemisuccinate (CHS) is believed to have a stabilising effect [21]. Here we carry out the first investigation of the effect of cholesterol on purified and reconstituted NTS1. We find the presence of CHS to have a small role in stabilising purified NTS1 in the presence of detergents, particularly during ligand affinity chromatography. However, in a lipid environment we find cholesterol unnecessary for ligand binding and oligomerisation. 2. Methods 2.1. Purification of NTS1B and NTS1-eYFP/eCFP NTS1B and NTS1-eYFP/eCFP fusion proteins have been described previously [27,28,30]. Purification of NTS1 eYFP/eCFP was carried out as described [28]. NTS1B was purified using immobilised metal affinity chromatography (IMAC) and a neurotensin (NT) ligand affinity column two-step protocol similar to that described in Ref. [29], with the final buffer containing 50 mM Tris pH 7.4, 15% glycerol (v/v), 250 mM NaCl, 0.1% DDM (w/v), 0.01% CHS (w/v), and 1 mM EDTA. Purification of NTS1B in the absence of CHS was carried out in the same manner using buffers containing DDM-only throughout solubilisation and purification processes with the final buffer as that detailed above without CHS. Protein concentration was determined by absorbance at 280 nm using calculated extinction coefficients, or by gel densitometry. 2.2. Reconstitution of NTS1B and NTS1-eYFP/eCFP Phospholipids 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE) and porcine brain polar lipid extract (BPL) were purchased from Avanti Polar Lipids, cholesterol was purchased from Sigma. Phospholipids were prepared by mixing in equimolar amount. Cholesterol was included at a 25% molar content. Lipids were prepared and detergent-disrupted as described previously [28]. For FRET experiments, NTS1-eYFP and NTS1-eCFP were mixed at a 1:1 molar ratio, added to disrupted vesicles at a protein lipid molar ratio of 1:6000 and incubated for 1 h at 4 °C. To remove detergent, 120 mg/ ml Biobeads SM-2 (Bio-Rad) was added and the sample agitated overnight at 4 °C. Proteoliposomes were isolated using sucrose gradient density ultracentrifugation (100,000 ×g; 4 °C for 15 h in a swingout rotor). NTS1B was reconstituted in the same manner except a protein lipid molar ratio of 1:4000 was used, and proteoliposomes isolated by ultracentrifugation in a Beckman TLX table-top ultracentrifuge (100,000 ×g, for 3 h at 4 °C). 2.3. Cholesterol depletion from brain polar lipid (BPL) vesicles NTS1B BPL proteoliposomes were prepared as described above. To extract cholesterol, a stock solution of 40 mM methyl-β-cyclodextrin (MβCD) (Sigma) was freshly prepared in reconstitution buffer. The pelleted, reconstituted receptor (1.2 μM) was resuspended in 250 μl 0, 10, 20 or 40 mM methyl-β-cyclodextrin and stirred for 2–3 h at
2229
room temperature. The proteoliposomes were then pelleted and resuspended in 100 μl reconstitution buffer. Activity was measured as detailed below and cholesterol content assayed using Amplex® Red Cholesterol Assay Kit (Invitrogen). 2.4. Fluorescence resonance energy transfer Fluorescence resonance energy transfer (FRET) was measured using a LS-55 Spectrofluorimeter (Perkin Elmer). Co-reconstituted NTS-eYFP/eCFP (900 μl) was placed in a 1.5 ml quartz microcuvette (Hellma) and measurements conducted at 4 °C with stirring. Excitation and emission slits were both set to 2.5 nm. Apparent FRET efficiency (FRETapp) was calculated as described and values of 10 are considered to be indicative of dimer formation [28]. 2.5. Ligand-binding analysis of purified and reconstituted receptor To quantify the total amount of active (ligand binding) receptor in either detergent solution [27] or lipid bilayers [28], a radioligand saturation assay was conducted as described. The assay buffer for protein purified in the absence CHS also contained no CHS. The stability of ligand binding was assessed over an 8 h period by incubating samples at room temperature for the indicated times, followed by rapid freezing in liquid nitrogen and storage at – 80° C until required. Standard errors were calculated from experiments carried out in triplicate. 2.6. Membrane order Membrane order of POPC/POPE liposomes with and without cholesterol was monitored by continuous-wave electron paramagnetic resonance (CW-EPR). Liposomes were prepared as described above to a final concentration of 20 mg/ml of lipid. In addition ~ 1 mol% of 8-DOXYL-stearic acid was added to the samples as the reporter molecule. EPR measurements were performed at 9 GHz on a Bruker EMX spectrometer. The degree of order in the membrane reported by a stearic acid spin nitroxide can be defined by an order parameter S [31], (Eq. (1)), which is calculated from the observed hyperfine anisotropy (A∥–A⊥), measured from the spectrum, to a maximum hyperfine anisotropy (Azz–Axx), for a maximum rigidity (S = 1), calculated from a frozen spectrum of the same sample to correct for polarity effects. S¼
Ajj −A⊥ Azz −Axx
ð1Þ
3. Results 3.1. Cholesterol motifs in NTS1 The observed bound, or trapped, cholesterol in the crystal structure of the β2 adrenergic receptor has led to the proposal of a specific cholesterol binding motif, highly conserved throughout the primary sequences of GPCRs of this family. Known as the conserved cholesterol motif (CCM), this region comprises one residue from helix II, and three from IV. Tryptophan 4.50 (Ballesteros–Weinstein nomenclature [32]) is predicted to interact with the sterol ring, as well as isoleucine 4.46 on the same helix and tyrosine 2.41. A charged residue (arginine 4.43) at the intercellular face of helix IV is proposed to form the strongest interaction with the cholesterol in terms of binding energy, through an electrostatic interaction with the 3β hydroxyl moiety of the cholesterol. Fig. 1a shows the CCM motif identified in NTS1 [from our unpublished model] in comparison to the same motif in the β2 adrenergic receptor [5] with the residues involved highlighted. An alternative cholesterol binding motif, the rather generic CRAC
2230
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
motif (cholesterol recognition / interaction amino acid consensus), has been found in a number of membrane proteins known to specifically interact with cholesterol [33,34]. Recently this motif has been highlighted in GPCRs including the serotonin1A receptor and the β2 adrenergic receptor [35]. Here we also find this motif in two locations in the NTS1 sequence (Fig. 1b). Despite the presence of these motifs there is little experimental evidence for their importance in determining or influencing function, or whether they can be used as a predictor for specific cholesterol binding. 3.2. Purified NTS1 is active in the presence and absence of cholesteryl hemisuccinate (CHS) Purification of NTS1 was carried out in buffer containing micelles of DDM-only or a combination of DDM, CHAPS and CHS (addition of CHAPS promotes efficient solubilisation of CHS). The final step of purification utilises a neurotensin (NT) ligand affinity column. It was found that NTS1 prepared in the presence and absence of CHS was effectively isolated by ligand affinity chromatography, implying ligand binding is not dependent on CHS. In both cases the yield of receptor was similar (see gel in Fig. 2a); however, the receptor prepared without CHS was not as pure (Fig. 2a), possibly due to decreased stability during purification, with the contaminating bands on the gel representing degradation products and aggregation caused by inefficient solubilisation. Assessment of ligand binding using tritiated NT highlighted a ~75% reduction in activity in the DDM-only preparation compared to the CHS containing preparation, once normalised for protein concentration (Fig. 2b). One reason for the discrepancy between comparable receptor binding to the NT column and the 75% difference observed in the NT binding assay could be the reduced stability of the CHS-free receptor preparation after the NT chromatography step. To investigate this further the stability of ligand binding to both preparations CHS containing and CHS free was assessed over an 8 h period at room temperature. Fig. 2c shows 14% activity remaining in the CHS containing preparation while the CHS-free sample is reduced to 4%. Therefore CHS is contributing a small amount to receptor stability at room temperature; however, this doesn't fully explain the observed decrease in ligand binding of the CHS-free preparation (75%). It is possible that the high local concentration of receptor on the NT affinity column may lead to the isolation of inactive receptors in the absence of CHS. Indeed, it has been observed recently [36] that proximity to the NT resin during chromatography can destabilise NTS1, possibly
Fig. 2. Analysis of NTS1 purified in the presence and absence of cholesteryl hemisuccinate (CHS): a) Coomassie-stained SDS PAGE gel of NTS1B purified with (+) and without (−) CHS. The arrow indicates NTS1B; b) Difference in ligand binding activity of NTS1B in terms of specific and relative activity; c) Stability of NTS1B ligand binding activity over time at room temperature. In all cases the ligand used was tritiated neurotensin (Perkin Elmer), at a final concentration of 8.33 nM.
due to detergent-related perturbations [36], an effect which may be enhanced during purification without CHS. 3.3. Reconstituted NTS1 is ligand binding in the presence and absence of cholesterol Although micelles are useful tools for in vitro studies of membrane proteins, an investigation into the effect of cholesterol on NTS1 in a
Fig. 1. Cholesterol binding motifs in NTS1: a) Detail of the CCM binding motif from the crystal structure of β2 adrenergic receptor with bound cholesterol (PDB ID: 3D4S) (left) and a model of the corresponding region of NTS1. Residues forming the CCM motif are highlighted, numbered in Ballesteros–Weinstein nomenclature [32]. Images produced using Pymol (http://pymol.sourceforge.net/). b) The cholesterol recognition amino acid consensus motif (CRAC) and underneath the two identified sequences in NTS1, with the position of the first residue in brackets.
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
2231
ligand‐binding properties of NTS1, and thus a direct interaction of cholesterol with the receptor may not be required for activity. It can be concluded that other components of BPL must be required to regain full ligand-binding activity, either through a direct receptor interaction or due to alteration of bilayer properties such as curvature [37] or thickness [38]. 3.4. Dimerisation in the absence of cholesterol Previously we showed NTS1 forms constitutive homo-dimers in BPL whereas, in detergent, dimerisation could only be detected at high protein–detergent ratios [28,39]. Therefore we proposed that particular components in the bilayer may contribute directly to dimerisation. Using the fluorescently-tagged receptors NTS1Y and NTS1C described previously [28], dimerisation was probed in the minimal lipid environment identified above, both in the presence and absence of cholesterol. The results of FRET measurements (Fig. 3b) show that dimerisation reported by FRET efficiency, is unaffected by the presence of cholesterol in a POPC/POPE background. However, with POPC the addition of cholesterol does lead to an increase in FRET efficiency to a level comparable to POPC/POPE. Therefore either cholesterol or POPE facilitates dimerisation. 3.5. Cholesterol influences membrane order in POPC/POPE lipids
Fig. 3. Activity and dimerisation of NTS1 reconstituted in the presence and absence of cholesterol: a) NTS1B purified both with and without cholesteryl hemisuccinate (CHS) was reconstituted into five different lipid mixtures consisting of: brain polar lipid (BPL); 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC); 1-palmitoyl-2oleoyl-sn-glycero-3-phosphoethanolamine (POPE); cholesterol (Ch), the concentration of NTS1 was 1.2 μM. Ligand binding is reported relative to BPL; b) Dimerisation of NTS1 tagged with eCFP or eYFP was measured using FRET in the same lipid mixtures as above. Apparent FRET efficiency was calculated as previously reported [28] the protein. The total amount of NTS1C and NTS1Y used in each experiment was 175 pMol.
biologically-relevant lipid environment was carried out. We have previously reported successful reconstitution of NTS1 into liposomes and demonstrated ligand binding as well as dimerisation [28]. Here, NTS1 was reconstituted into various lipid mixtures in the presence and absence of 25 mol% cholesterol and ligand‐binding activity measured. We used brain polar lipid extract (BPL), which represents the native environment for the receptor, along with more simple mixtures composed of POPC/POPE. PC and PE lipids were chosen since together they comprise around half of the lipids in BPL [Avanti Polar Lipids, product information]. Fig. 3a shows the comparative ligand-binding activity of reconstituted NTS1. BPL represents the most effective lipid in supporting ligand binding. In lipid mixtures of POPC/POPE NTS1 ligand binding is only 40% as effective as in BPL; however ligand binding in the presence and absence of cholesterol is not affected (Fig. 3a), therefore suggesting that cholesterol is not required for ligand binding activity when NTS1 is reconstituted into a POPC/POPE environment. Furthermore, since ligand binding in POPC is below the level of detection, it is proposed that POPC/POPE provides a minimal mix for functional receptor reconstitution. The NTS1 used in these experiments was prepared in the presence of CHS, therefore it is possible that it was not removed during the reconstitution process (by adsorption on to Biobeads) and effectively substitutes for cholesterol. To address this possibility we repeated the experiment using receptor purified in CHSfree buffer (see Section 3.2). Fig. 3a shows little change in measured ligand‐binding activity, and it is therefore concluded that in a POPC/ POPE background the presence of cholesterol does not affect the
Changes in ordering of phospholipid acyl chains, caused by the presence of cholesterol, have been reported to influence function for some GPCRs [14,40]. From the functional studies presented above, NTS1 appears unaffected by the presence of cholesterol. To confirm cholesterol is affecting membrane order in our system continuous‐ wave electron spin resonance (CW-ESR) combined with a spinlabelled stearic acid was used to report on the order within the lipid acyl chains. From the resulting spectra (Fig. 4) the order parameter was determined. As predicted, the membranes became more ordered in the presence of cholesterol [41]. These data confirm that changes in membrane order are unlikely to be affecting NTS1 function, when compared with the functional data (Section 3.4). 3.6. Effect of cholesterol depletion Methyl-β-cyclodextrin (MβCD), a cyclic glucose oligomer, has been extensively employed in the study of GPCRs with cholesterol, and it has been shown to sequester hydrophobic molecules such as cholesterol out of the bilayer [42]. Although these experiments tend
S:
-
0.68
+
0.78
3300
3350
3400
Field [G] Fig. 4. Membrane order of POPC/POPE with and without 25 mol% cholesterol: EPR spectra were recorded at 294 K on bilayers in the presence (+) and absence (−) of cholesterol, and represent an average of 10 scans. The order parameter (S) was calculated as described in the Methods section. The dotted lines highlight the change in the parallel component of the spectra when cholesterol is present.
2232
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
Fig. 5. Cholesterol extraction using methyl beta cyclodextrin (MβCD): a) NTS1B reconstituted into bilayers of brain polar lipid was incubated with different concentrations of MβCD. Proteoliposomes were subsequently re-isolated by ultracentrifugation and ligand binding activity measured. Protein concentration was determined by gel densitometry using Image J and cholesterol concentration was calculated using the Amplex Red kit from Invitrogen. b) The same data plotted to show NTS1 specific activity as a function of protein concentration.
to be carried out in in vivo systems, here MβCD was used to remove cholesterol from NTS1 reconstituted into BPL liposomes as an alternative approach to that presented in Section 3.4 (in which cholesterol was added into the lipid bilayers). After incubation with increasing concentrations of MβCD, liposomes were recovered by ultracentrifugation, and ligand‐binding activity measured. The Amplex-Red cholesterol assay kit (Invitrogen) was used to confirm the reduction in cholesterol content in the bilayers. The results (Fig. 5a) show that although relative activity (recovery) is reduced by 40% as MβCD concentration increases, this correlates with a reduction of 40% in relative protein recovery. An additional plot of cholesterol concentration versus specific activity clearly demonstrates that the specific activity of NTS1 is unaffected by cholesterol concentration (Fig. 5b). Therefore, in agreement with the data above (Section 3.4), ligand binding to NTS1 is not influenced by the presence of cholesterol in bilayers. 4. Discussion The dependence of GPCR stability and function on the presence of CHS has been previously reported in a small number of cases [5,21,43,44]. Compared with NTS1, the closely related β2 adrenergic receptor is also active in the absence of CHS, although it is significantly more stable, given that a 5-fold increase in the half-life of protein denaturation is observed upon addition of CHS [5]. Furthermore, the stability of the adenosine A2a receptor has also been linked to the presence of CHS, although, unlike NTS1 and β2 adrenergic
receptor, this receptor appears completely inactive in DDM micelles without CHS [44]. It is proposed that the presence of CHS modifies the dimensions of the micelle environment to match those of the biological membrane for the A2a receptor [43]. However, it is important to point out that the adenosine A2a receptor along with NTS1 and the β2 adrenergic receptor are all capable of ligand binding in the E. coli membrane—which does not contain cholesterol [20,21,23]. Therefore inferring a biological significance to CHS dependent effects from this observation is difficult. Clearly there is a stabilisation advantage to the use of CHS in structural determination of GPCRs, but the inference that cholesterol is therefore important for activity must be investigated in a well-defined lipid environment. The data presented here show NTS1 is active and relatively stable when purified in the absence of CHS (Fig. 2b and c), indicating no significant effect of the cholesterol analogue on receptor function. However, NTS1 function does appear sensitive to the inclusion of CHS during the course of the purification process (Fig. 2a and b). These results are in contrast to previous experiments with NTS1 from detergent‐solubilised cell extracts, where the absence of CHS reduced receptor yield and half-life of ligand binding by over 50% [21]. These differences may be explained by the two distinct environments employed in these studies; detergent-solubilised cell extracts in the former work compared with purified protein investigated here. As we are interested specifically in purified systems in a well-defined lipid environment we focused on characterising the purified receptor in this environment, free of any other components. We also demonstrate that once reconstituted, the presence of cholesterol appears to have little influence on the activity and dimerisation of NTS1 in the lipid mixes used here (Fig. 3). Although activity was clearly reduced in POPC/POPE compared to the more native BPL mix, this could not be attributed specifically to the absence of cholesterol. The influence of cholesterol on the activity of other GPCRs is believed to arise in part from the ordering effect that cholesterol imparts on the lipid acyl chains [14,15]. We confirm this effect with EPR measurements on the POPC/POPE bilayers with and without cholesterol, and conclude that changes in phospholipid order are not influencing NTS1 activity or dimerisation. Dimerisation in terms of FRET efficiency was restored from a POPC background in the presence of either POPE or cholesterol, unlike activity which was not restored by cholesterol. These results therefore suggest that dimerisation and ligand binding are not linked processes. The role of particular lipids in GPCR structure and function remains poorly characterised at present, although POPE interactions have been linked to transitions between the MI and MII states of rhodopsin [45,46]. POPE is non‐ lamellar forming and as such alters the physical properties of the bilayer; it would therefore be interesting to further study how NTS1 is influenced by POPE. To further confirm cholesterol does not influence NTS1 ligand‐ binding function MCβD was used to deplete membrane cholesterol from NTS1 reconstituted into BPL. Despite having a non-polar cavity, MCβD has been shown to specifically remove cholesterol from bilayers [47,48]. Here we show that although cholesterol was removed and NTS1 ligand binding decreased, protein concentration also decreased at a rate closely linked to activity. The reason for this loss of protein is unclear but care must be taken in these types of experiments to confirm protein concentration. The overall conclusions from these depletion studies however, complement well with the cholesterol addition studies. Based on the results presented here we find no correlation between the presence of cholesterol binding motifs in NTS1 and cholesterol-dependent ligand-binding activity; in this case we are not able to comment on G‐protein activation. Recent data with other GPCRs have led to the proposal that cholesterol and GPCR function (both ligand binding and downstream signalling) are connected, which is clearly the case for the oxytocin receptor for example [14]. It has also been suggested that in vivo, cholesterol is required to
J. Oates et al. / Biochimica et Biophysica Acta 1818 (2012) 2228–2233
regulate signalling by locating the receptor to specific areas of the membrane, for example in lipid-microdomains or caveolae [49]. Specific membrane location has been reported with NTS1 in vivo [50]. Further work will be required to reveal the molecular details of cholesterol-regulated GPCR activity, as well as the influence of other components of the bilayer, but we conclude that cholesterol is not required for ligand-binding activity of purified and reconstituted NTS1 in a well-defined lipid environment. Acknowledgements Funding was provided by the Medical Research Council (MRC) (Grant no. G0900076 to A.W.), and the Engineering and Physical Sciences Research Council (EPSRC) (for support of CAESR (EP/D048559/ 1), A.W. co-applicant). References [1] R. Fredriksson, H.B. Schioth, The repertoire of G-protein-coupled receptors in fully sequenced genomes, Mol. Pharmacol. 67 (2005) 1414–1425. [2] A.D. Howard, G. McAllister, S.D. Feighner, Q. Liu, R.P. Nargund, L.H. Van der Ploeg, A.A. Patchett, Orphan G-protein-coupled receptors and natural ligand discovery, Trends Pharmacol. Sci. 22 (2001) 132–140. [3] T. Klabunde, G. Hessler, Drug design strategies for targeting G-protein-coupled receptors, Chembiochem 3 (2002) 928–944. [4] J. Oates, A. Watts, Uncovering the intimate relationship between lipids, cholesterol and GPCR activation, Curr. Opin. Struct. Biol. 21 (2011) 802–807. [5] M.A. Hanson, V. Cherezov, M.T. Griffith, C.B. Roth, V.P. Jaakola, E.Y. Chien, J. Velasquez, P. Kuhn, R.C. Stevens, A specific cholesterol binding site is established by the 2.8 A structure of the human beta2-adrenergic receptor, Structure 16 (2008) 897–905. [6] V. Cherezov, D.M. Rosenbaum, M.A. Hanson, S.G. Rasmussen, F.S. Thian, T.S. Kobilka, H.J. Choi, P. Kuhn, W.I. Weis, B.K. Kobilka, R.C. Stevens, High-resolution crystal structure of an engineered human beta2-adrenergic G protein-coupled receptor, Science 318 (2007) 1258–1265. [7] T. Warne, R. Moukhametzianov, J.G. Baker, R. Nehme, P.C. Edwards, A.G. Leslie, G.F. Schertler, C.G. Tate, The structural basis for agonist and partial agonist action on a beta(1)-adrenergic receptor, Nature 469 (2011) 241–244. [8] G. Gimpl, F. Fahrenholz, Cholesterol as stabilizer of the oxytocin receptor, Biochim. Biophys. Acta 1564 (2002) 384–392. [9] G. Gimpl, J. Reitz, S. Brauer, C. Trossen, Oxytocin receptors: ligand binding, signalling and cholesterol dependence, Prog. Brain Res. 170 (2008) 193–204. [10] G. Gimpl, V. Wiegand, K. Burger, F. Fahrenholz, Cholesterol and steroid hormones: modulators of oxytocin receptor function, Prog. Brain Res. 139 (2002) 43–55. [11] L. Pang, M. Graziano, S. Wang, Membrane cholesterol modulates galanin–GalR2 interaction, Biochemistry 38 (1999) 12003–12011. [12] T.J. Pucadyil, A. Chattopadhyay, Cholesterol modulates ligand binding and Gprotein coupling to serotonin(1A) receptors from bovine hippocampus, Biochim. Biophys. Acta 1663 (2004) 188–200. [13] A.D. Albert, J.E. Young, P.L. Yeagle, Rhodopsin–cholesterol interactions in bovine rod outer segment disk membranes, Biochim. Biophys. Acta 1285 (1996) 47–55. [14] G. Gimpl, K. Burger, F. Fahrenholz, Cholesterol as modulator of receptor function, Biochemistry 36 (1997) 10959–10974. [15] D.C. Mitchell, M. Straume, J.L. Miller, B.J. Litman, Modulation of metarhodopsin formation by cholesterol-induced ordering of bilayer lipids, Biochemistry 29 (1990) 9143–9149. [16] R. Prasad, P. Singh, A. Chattopadhyay, Effect of capsaicin on ligand binding activity of the hippocampal serotonin1A receptor, Glycoconj. J. 26 (2009) 733–738. [17] E.S. Levitt, M.J. Clark, P.M. Jenkins, J.R. Martens, J.R. Traynor, Differential effect of membrane cholesterol removal on mu- and delta-opioid receptors: a parallel comparison of acute and chronic signaling to adenylyl cyclase, J. Biol. Chem. 284 (2009) 22108–22122. [18] P. Huang, W. Xu, S.I. Yoon, C. Chen, P.L. Chong, L.Y. Liu-Chen, Cholesterol reduction by methyl-beta-cyclodextrin attenuates the delta opioid receptor-mediated signaling in neuronal cells but enhances it in non-neuronal cells, Biochem. Pharmacol. 73 (2007) 534–549. [19] A.D. Albert, K. Boesze-Battaglia, Z. Paw, A. Watts, R.M. Epand, Effect of cholesterol on rhodopsin stability in disk membranes, Biochim. Biophys. Acta 1297 (1996) 77–82. [20] H.M. Weiss, R. Grisshammer, Purification and characterization of the human adenosine A(2a) receptor functionally expressed in Escherichia coli, Eur. J. Biochem. 269 (2002) 82–92. [21] J. Tucker, R. Grisshammer, Purification of a rat neurotensin receptor expressed in Escherichia coli, Biochem. J. 317 (Pt 3) (1996) 891–899. [22] V.P. Jaakola, M.T. Griffith, M.A. Hanson, V. Cherezov, E.Y. Chien, J.R. Lane, A.P. Ijzerman, R.C. Stevens, The 2.6 angstrom crystal structure of a human A2A adenosine receptor bound to an antagonist, Science (2008) 1211–1217. [23] S. Marullo, C. Delavier-Klutchko, Y. Eshdat, A.D. Strosberg, L. Emorine, Human beta 2-adrenergic receptors expressed in Escherichia coli membranes retain their pharmacological properties, Proc. Natl. Acad. Sci. U. S. A. 85 (1988) 7551–7555.
2233
[24] S. Morein, A. Andersson, L. Rilfors, G. Lindblom, Wild-type Escherichia coli cells regulate the membrane lipid composition in a "window" between gel and nonlamellar structures, J. Biol. Chem. 271 (1996) 6801–6809. [25] R.M. Myers, J.W. Shearman, M.O. Kitching, A. Ramos-Montoya, D.E. Neal, S.V. Ley, Cancer, chemistry, and the cell: molecules that interact with the neurotensin receptors, ACS Chem. Biol. 4 (2009) 503–525. [26] S. Dupouy, N. Mourra, V.K. Doan, A. Gompel, M. Alifano, P. Forgez, The potential use of the neurotensin high affinity receptor 1 as a biomarker for cancer progression and as a component of personalized medicine in selective cancers, Biochimie 93 (9) (2011) 1369–1378. [27] H. Attrill, P.J. Harding, E. Smith, S. Ross, A. Watts, Improved yield of a ligandbinding GPCR expressed in E. coli for structural studies, Protein Expr. Purif. 64 (2009) 32–38. [28] P.J. Harding, H. Attrill, J. Boehringer, S. Ross, G.H. Wadhams, E. Smith, J.P. Armitage, A. Watts, Constitutive dimerization of the G-protein coupled receptor, neurotensin receptor 1, reconstituted into phospholipid bilayers, Biophys. J. 96 (2009) 964–973. [29] J.F. White, L.B. Trinh, J. Shiloach, R. Grisshammer, Automated large-scale purification of a G protein-coupled receptor for neurotensin, FEBS Lett. 564 (2004) 289–293. [30] R. Grisshammer, J.F. White, L.B. Trinh, J. Shiloach, Large-scale expression and purification of a G-protein-coupled receptor for structure determination—an overview, J. Struct. Funct. Genomics 6 (2005) 159–163. [31] P. Fretten, S.J. Morris, A. Watts, D. Marsh, Lipid–lipid and lipid–protein interactions in chromaffin granule membranes. A spin label ESR study, Biochim. Biophys. Acta 598 (1980) 247–259. [32] J.A. Ballesteros, H. Weinstein, C.S. Stuart, Integrated methods for the construction of three-dimensional models and computational probing of structure–function relations in G protein-coupled receptors, Meth. Neurosci., vol. 25, Academic Press, 1995, pp. 366–428. [33] R.M. Epand, A. Thomas, R. Brasseur, R.F. Epand, Cholesterol interaction with proteins that partition into membrane domains: an overview, Subcell. Biochem. 51 (2010) 253–278. [34] A.I. Greenwood, J. Pan, T.T. Mills, J.F. Nagle, R.M. Epand, S. Tristram-Nagle, CRAC motif peptide of the HIV-1 gp41 protein thins SOPC membranes and interacts with cholesterol, Biochim. Biophys. Acta 1778 (2008) 1120–1130. [35] M. Jafurulla, S. Tiwari, A. Chattopadhyay, Identification of cholesterol recognition amino acid consensus (CRAC) motif in G-protein coupled receptors, Biochem. Biophys. Res. Commun. 404 (2011) 569–573. [36] J.F. White, R. Grisshammer, Stability of the neurotensin receptor NTS1 free in detergent solution and immobilized to affinity resin, PLoS One 5 (2010) e12579. [37] K.L. Madsen, V.K. Bhatia, U. Gether, D. Stamou, BAR domains, amphipathic helices and membrane-anchored proteins use the same mechanism to sense membrane curvature, FEBS Lett. 584 (2010) 1848–1855. [38] T.M. Weiss, P.C. van der Wel, J.A. Killian, R.E. Koeppe 2nd, H.W. Huang, Hydrophobic mismatch between helices and lipid bilayers, Biophys. J. 84 (2003) 379–385. [39] J.F. White, J. Grodnitzky, J.M. Louis, L.B. Trinh, J. Shiloach, J. Gutierrez, J.K. Northup, R. Grisshammer, Dimerization of the class A G protein-coupled neurotensin receptor NTS1 alters G protein interaction, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 12199–12204. [40] K.G. Harikumar, V. Puri, R.D. Singh, K. Hanada, R.E. Pagano, L.J. Miller, Differential effects of modification of membrane cholesterol and sphingolipids on the conformation, function, and trafficking of the G protein-coupled cholecystokinin receptor, J. Biol. Chem. 280 (2005) 2176–2185. [41] J.L. Rubenstein, J.C. Owicki, H.M. McConnell, Dynamic properties of binary mixtures of phosphatidylcholines and cholesterol, Biochemistry 19 (1980) 569–573. [42] Y. Ohtani, T. Irie, K. Uekama, K. Fukunaga, J. Pitha, Differential effects of alpha-, beta- and gamma-cyclodextrins on human erythrocytes, Eur. J. Biochem. 186 (1989) 17–22. [43] M.A. O'Malley, M.E. Helgeson, N.J. Wagner, A.S. Robinson, The morphology and composition of cholesterol-rich micellar nanostructures determine transmembrane protein (GPCR) activity, Biophys. J. 100 (2011) L11–L13. [44] M.A. O'Malley, T. Lazarova, Z.T. Britton, A.S. Robinson, High-level expression in Saccharomyces cerevisiae enables isolation and spectroscopic characterization of functional human adenosine A2a receptor, J. Struct. Biol. 159 (2007) 166–178. [45] A.V. Botelho, N.J. Gibson, R.L. Thurmond, Y. Wang, M.F. Brown, Conformational energetics of rhodopsin modulated by nonlamellar-forming lipids, Biochemistry 41 (2002) 6354–6368. [46] O. Soubias, W.E. Teague, K. Gawrisch, Evidence for specificity in lipid–rhodopsin interactions, J. Biol. Chem. 281 (2006) 33233–33241. [47] G. Gimpl, Cholesterol–protein interaction: methods and cholesterol reporter molecules, Subcell. Biochem. 51 (2010) 1–45. [48] U. Klein, G. Gimpl, F. Fahrenholz, Alteration of the myometrial plasma membrane cholesterol content with beta-cyclodextrin modulates the binding affinity of the oxytocin receptor, Biochemistry 34 (1995) 13784–13793. [49] Y. Xiang, V.O. Rybin, S.F. Steinberg, B. Kobilka, Caveolar localization dictates physiologic signaling of beta 2-adrenoceptors in neonatal cardiac myocytes, J. Biol. Chem. 277 (2002) 34280–34286. [50] Y. Heakal, M. Kester, Nanoliposomal short-chain ceramide inhibits agonistdependent translocation of neurotensin receptor 1 to structured membrane microdomains in breast cancer cells, Mol. Cancer Res. 7 (2009) 724–734.