A monomeric G protein-coupled receptor isolated in a high-density lipoprotein particle efficiently activates its G protein

Supporting Information

Files in this Data Supplement:

Fig. 6. Labeling efficiency of b2AR estimated by TIRF imaging and photobleach-step counting. (a) The percentage of receptors containing one, two, or three fluorescent labels was determined by manually counting the number of photobleach steps in the intensity time trace for at least 400 individual receptors or rHDL particles. In each case it was impossible to count photobleach steps for a fraction of molecules due to either low signal-to-noise or fluorophore blinking. (b) Illustrative time traces for molecules counted as containing one, two, or three fluorescent labels. To achieve quantitative labeling, preparations of b2AR were labeled with cysteine-reactive derivatives of Cy3 or Cy5 at a stoichiometry of 20 fluorophores to one receptor molecule. From absorption spectra we estimate that between 1.6 and 2.5 mol of fluorophore were incorporated per mole of b2AR under these conditions. Over 35% of Cy3-and Cy5-b2AR molecules were observed to contain multiple fluorophores, as determined by counting the number of photobleaching events for individual receptors imaged by TIRF (e.g., Cy3-b2AR, 28% double-labeled, 9% triple-labeled; Cy5-b2AR, 37% double-labeled, 5% triple-labeled). To estimate the fraction of unlabeled b2AR in such a sample, we simulated the labeling kinetics of the three most reactive cysteines using a mathematical model. This model was able to accurately reproduce the observed fractions of double-and triple-labeling. Based on our most conservative estimate from this model, a maximum of 0.4% of b2AR remains unlabeled (see SI Text).

Fig. 7. Purified b_2AR in DDM micelles is monomeric before reconstitution in vesicles. To ensure that the Cy3-or Cy5-labeled b_2AR is not oligomeric before rHDL reconstitution (i.e., in detergent micelles), cross-linking analysis with BS³ (11.4-Å spacer, Pierce) was performed. Although considerable intramolecular cross-linking (band spreading) was observed, no detectable intermolecular cross-linking was observed in Cy3 b_2A R in detergent micelles. These data suggest that there are insignificant levels of oligomeric receptor complexes in detergent micelles. In contrast, incubation with BS³ resulted in the concentration-dependent appearance of higher-order species, including dimers, and what appear to be aggregates near the top portion of the separating gel. Cy3-labeled b_2AR (1 mM) in (a) detergent micelles (0.1% DDM) or (b) reconstituted in phospholipid vesicles were incubated in the absence or presence of increasing concentrations of BS³ for 60 min on ice. Reaction was terminated with 20 mM Tris-HCl, pH 8.0. Samples were resolved by SDS/PAGE and imaged on a UV gel documentation station (FluorChem 8800, Alpha Innotech) using rhodamine/Texas red filters. Similar results were obtained using an amine cross-linker with a longer linker, Bis(NHS)PEO₅ (21.7-Å pegylated spacer, Pierce; data not shown). Phospholipid reconstitution into DOPC:cholesterol vesicles was performed as described above.

SI Text

Buffer A: 50 mM Tris-HCl pH 8, 50 mM NaCl, protease inhibitors (PIs: 34 mg/ml each of L-tosylamido-2-phenylethyl chloromethyl ketone, 1-chloro-3-tosylamido-7-amino-2-heptanone and phenylmethylsulfonyl fluoride, and 3 mg/ml each of leupeptin and lima bean trypsin inhibitor).

Buffer B: 50 mM Tris-HCl pH 8, 50 mM NaCl, PIs.

Buffer C: 20 mM Hepes pH 8, 1 mMEDTA, 0.1% DDM, PIs.

Buffer D: 50 mM Tris-HCl pH 8, 1 mM CaCl₂, 3 M NaCl, 5 m M EDTA.

Buffer E: 20 mM Tris-HCl pH 8, 1 mM CaCl₂, 5 mM EDTA, 0.1% Triton X 100.

Buffer F: 25 mM K-acetate pH 5, 1 mM EDTA, 0.1% Triton X-100.

Buffer G: 20 mM Hepes pH 8, 100 mM NaCl, 1 mM EDTA.

Buffer H: 20 mM Hepes pH 7.5, 100 mM NaCl.

Buffer I: 100 mM NaCl, 20 mM Hepes pH 7.5, 0.1% dodecylmaltoside (Anatrace).

Buffer J: Buffer I and 1 mM EDTA.

Buffer K: Buffer I with 300 mM alprenolol (Sigma) and 1 mM CaCl₂.

Buffer L: Buffer I with 1 mM CaCl₂.

Buffer M: Buffer I with 0.01% cholesterol hemisuccinate.

Receptor purification and labeling. b2AR (WT or CFP-fused) was expressed in Sf9 cells and solubilized using methods previously described (1).For CFP-b2AR: DDM-solubilized extract was applied to a metal-chelate affinity column (Talon, Clontech). Samples were washed with Buffer B + 0.1% DDM with 2.5 mM imidazole and then subsequently eluted with Buffer B with 100 mM imidazole, 0.1% DDM. Peak fractions were applied to a 1 ml Source Q anion exchange column (GE Healthcare) in Buffer C. CFP-b₂AR was eluted with a 15 ml 0-40% linear gradient with Buffer C + 1 M NaCl. Peak fractions were pooled and resolved on a Superdex 200 size exclusion column in Buffer C + 50 mM NaCl to resolve the CFP-b₂AR from the clipped CFP. The resultant CFP-b₂AR is greater than 95% pure and stored on ice until use. For wt-b₂AR: CaCl₂ was added to the DDM-solubilized extract to a final concentration of 1 mM and the detergent solubilized b2AR was purified by M1-Flag affinity chromatography (Sigma). The receptor was eluted from the M1-Flag resin in Buffer J. The concentration of functional, purified receptor was determined using a saturating concentration (10 nM) of [³H]dihydroalprenolol as previously described (1). Flag-purified receptor was then purified by alprenolol-Sepharose chromatography as described (1). The receptor eluted from alprenolol-Sepharose with Buffer K and loaded directly onto M1-Flag resin. The M1-Flag resin was washed with Buffer I to remove free alprenolol and eluted with Buffer J. Two liters of Sf9 cells typically yield 500 ml of a 5 mM solution of b2AR.

Purification of human apoA-I: WT human apoA-I was purified from human serum by a protocol adapted from Gan et al. (2); all procedures were performed at room temperature (RT) unless noted. Frozen serum (-20°C in 10 mM CaCl₂) was thawed at 37°C, strained through cheesecloth, and then centrifuged at 5,000 ´ g for 10 min to pellet any debris. This clarified serum was then made up to the following buffer condition: Buffer D. This solution was combined with equal serum volume of blue agarose resin (Cibacron blue F3GA-agarose, Sigma) equilibrated in Buffer D, and stirred for 30 min. Resin wasthen washed by filtering through a Whatman #1 filter in a Büchner funnel. The resin cake was then resuspended in 3 x resin volume Buffer D and then re-filtered. The resin was washed (3-4 times), until absorbance at 280 nm of the filtrate was less than 0.025.Then the resin is washed twice more in the same manner with Buffer D minus 3 M NaCl. After the last wash, the cake is resuspended in an equal volume of the same buffer and loaded onto an empty column. The remaining bound proteins were then eluted with same buffer + 5 mM cholate. ApoA-I was typically 80-90% pure at this stage. To delipidate the apoA-I, fractions were pooled and concentrated using an Amicon stirred ultrafiltration cell affixed with a 10,000 MWCO filter (Millipore) and then diluted 1:1 in 25 mM Tris-HCl pH 8, 1 mM CaCl₂, 5 mM EDTA, 0.2% Triton X-100. This was then applied to a 70 ml Q Sepharose (Amersham Pharmacia) column equilibrated in Buffer E and eluted with a shallow linear gradient with Buffer E + 1 M NaCl - apoA-I usually eluted around 10-15%. The remaining contaminants were removed using a SP Sepharose (Amersham Pharmacia) column (70 ml) equilibrated in Buffer F and eluted with a linear gradient against Buffer F + 1 M NaCl. To exchange the Triton X-100 for cholate, SP Sepharose fractions were applied to a Superdex 200 size exclusion column (Amersham Pharmacia) in Buffer G + 20 mM cholate, at 4°C. ApoA-I fractions were pooled and concentrated to at least 10 mg/ml, then dialyzed at 4°C against Buffer G + 5 mM cholate and stored at -80°C until further use.

In vitro reconstitution of HDL (3): Palmitoyl-oleoyl-phosphatidylcholine (POPC) and palmitoyl-oleoyl-phosphatidylglycerol (POPG) were used in combination at a 3:2 molar ratio, a mixture that mimics the zwitterionic environment of a cell membrane (4). Briefly, lipids were dried under argon (or nitrogen) from a chloroform solution and placed in a vacuum dessicator for 30-60 min to remove residual traces of chloroform. Lipids were then solubilized in Buffer G + 50 mM cholate for incorporating WT b₂AR or Cy3 (or Cy5)-labeled b₂AR. For incorporation of receptor, detergent-solubilized purified receptor was then added. Finally, a concentrated stock of purified apoA-I (above) was added such that the final concentration of the components was: 24 mM detergent, 8 mM lipids, and 100 mM apoA-I. This solution was incubated for 1-2 h at 4°C (Tm_POPC,/POPG = -2°C). The mixture was then added to BioBeads (BioRad), 0.5 mg/ml, to remove detergents. Samples were stored on ice until used. The amount of receptor added varied, depending on the use; however, to achieve high levels of incorporation efficiency, we kept apoA-I in at least 10-fold excess, and the amount of receptor added was no more than 10% of the final reconstitution volume.

Analytical size exclusion chromatography: Analytical size exclusion chromatography was performed on a HR10/30 column (Amersham Pharmacia) packed with »20 ml of Superdex 200 resin (Amersham Pharmacia) (V₀ = 7 ml). The column was calibrated with thyroglobulin [669 kDa, Stokes diameter (Sd) = 17.2 nm], apoferritin (432 kDa, Sd = 12.2 nm), alcoholdehydrogenase (150 kDa, Sd = 9.1 nm), BSA (66 kDa, Sd = 7.2 nm), carbonic anhydrase (29 kDa, Sd = 4 nm). Samples were loaded in 100-500 ml static loops and run at a flow rate of 0.7 ml/min using the BioLogic DuoFlow system (BioRad) at 4°C. 200 ml fractions from the column were collected in a 96-well plate for further analysis.

Saturation radioligand-binding assays: Binding reactions were prepared in 100 ml volumes in 96-well plates. Samples were incubated with various concentrations of b₂AR antagonist [³H]dihydroalprenolol ([³H]DHAP) (0.1-46 nM) in 50 mM Tris-HCl pH 8, 150 mM NaCl (TBS) [or TBS with 1% detergent (DDM or cholate) for detergent-solubilized binding]. Nonspecific binding was determined in the presence of 20 mM propranolol. Membrane samples were incubated for 90 min at RT. Detergent-solubilized, purified samples were incubated for 60 min at 30°C for saturation isotherms, or for 30 min at 30°C for single point saturation binding. For separating free [³H]DHAP from bound, membrane samples were filtered on glass fiber filter plates and detergent solubilized samples were filtered on gel filtration columns. For the rHDL samples, both methods were used successfully, although the glass fiber plates retained only about 80% of the binding seen on the gel filtration columns.

For glass fiber filtering, GF/B 96-well filter plates (Whatman) were used in conjunction with a vacuum manifold. Wells were prewet with TBS. Samples were applied and washed 3x with 200 ml of TBS. Scintillation mixture was added (Microscint 0, Packard) and plates were counted on a TopCount scintillation counter (Packard). For gel filtration, samples were applied to Sephadex G-50 columns equilibrated in TBS (or TBS + 0.05% DDM or 0.5% cholate). Scintillation mixture (Cytoscint, MP Biomedicals) was added to the flowthrough fractions (containing the bound receptor) and counted on a Beckman LS5000. Specific binding was determined by subtracting nonspecific binding from total binding.

G protein reconstitution: Purified Gs (5) (stored in 11 mM CHAPS) was reconstituted into preformed impure b₂AR•rHDL particles (containing excess empty rHDL particles) at an initial R:G ratio of 1:50. Concentrated Gs stocks were added such that the CHAPS was diluted at least 700-fold to reduce the CHAPS concentration to well below the CMC (610 mM). This had no effect on the integrity of the particles, as assessed by size exclusion chromatography (data not shown). Treatment of Gs-reconstituted samples with BioBeads, to remove trace amounts of CHAPS, before gel filtration chromatography had no effect on the results.

Agonist competition assays: Agonist competition assays were performed on G protein-reconstituted samples under similar conditions as used in the saturation binding assays except that a fixed concentration of [³H]DHAP (2 nM) was competed with various concentrations of isoproterenol (1 ´10^-12-1 ´ 10^-4 M), with or without the addition of 10 mM GTPgS. Binding reactions contained 0.02% ascorbic acid to prevent oxidation of the isoproterenol. Samples were incubated for 30 min at 30°C and then filtered on glass fiber plates as above. Normalized data were fit to a two-site competition binding model using Prism (GraphPad).

GTPgS-binding assay: G protein reconstitution was performed as above. Agonist-stimulated [³⁵S]GTPgS-binding assays were performed essentially as described by Asano et al. (6). Aliquots were combined with either 1 mM isoproterenol, 10 mM timolol, or 1 mM isoproterenol plus 3 mM propranolol in buffer comprised of Buffer H plus 2 mM MgCl₂ and 0.02% ascorbic acid. These were allowed to preincubate at 30°C for »5 min before being combined with 100 nM isotopically diluted [³⁵S]GTPgS. Fifty-microliter aliquots were removed at specific times and added to 100 ml of quench buffer (Buffer G + 10 mM MgCl₂, 100 mM GTPgS, 100 mM timolol) on ice. Samples were filtered on GF/B 96-well plates as above, except that filters were washed five times with ice-cold 20 mM Tris-HCl pH 8, 100 mM NaCl, 10 mM MgCl₂. The glass-fiber filtration method was selected over the more traditional nitrocellulose filter method (including detergent, Lubrol) to be consistent with the conditions selected for the radioligand-binding assays. In addition, to avoid disrupting the rHDL particles, we excluded detergent (Lubrol) from our binding and filtration steps, the inclusion of which could enhance the recovery of [³⁵S]GTPgS binding to G proteins on nitrocellulose filters by as much as 50%.

To quantitate total G protein in the reconstitution capable of binding [³⁵S]GTPgS, we did, however, use more traditional methods. Total [³⁵S]GTPgS binding was assessed in Buffer H plus 50 mM MgCl2, 10 mM [³⁵S]GTPgS, 0.05% Lubrol. After incubating for 30 min at 30°C, samples were rapidly filtered through BA-85 nitrocellulose filters (Whatman) and washed 4 ´ 2 ml with ice-cold 20 mM Tris-HCl pH 8, 100 mM NaCl, 10 mM MgCl_2. Using nitrocellulose filtration and inclusion of detergent (Lubrol) to determine the total [³⁵S]GTPgS binding to G protein, we obtain data that are consistent with our observations using glass-fiber filtration. Our data suggest that »20% of the [³⁵S]GTPgS-binding activity survives the reconstitution step, only a portion of which (»29 +/- 5% of the 20% [³⁵S]GTPgS binding, or 5.8% of the G protein added to the reconstitution) copurifies with the b2AR•rHDL on an anti-FLAG resin (the b2AR is FLAG-tagged). No binding was observed in the column flowthrough or wash fractions. Thus, the remaining two-thirds appears to get hung up on the anti-FLAG column, presumably in the form of an aggregate that is capable of binding [³⁵S]GTPgS. The remaining »14% of the total G protein will be present in the assay, but as an aggregate and unlikely to be accessible to the b2AR in rHDL.

On the surface, the presence of the aggregates that are capable of binding [³⁵S]GTPgS could confound the results and interpretation of the data in Fig. 5a. However, the conditions used in Fig. 5a (100 nM [³⁵S]GTPgS, 2 mM MgCl₂) are dramatically different from used to measure total [³⁵S]GTPgS binding (10 mM [³⁵S]GTPgS and 50 mM MgCl2). Such conditions of low nucleotide concentrations, such as those used in Fig. 5a, should display a preference for detection of receptor-stimulated G protein binding.

Cy3 and Cy5 labeling of b2R. Purified b2AR (5 mM) was labeled with cysteine-reactive Cy3-maleimide (100 mM) or Cy5-maleimide (100 mM, GE Healthcare) for 60 min at 25°C in Buffer H + 0.1% DDM, in the presence of Tris(2-carboxyethyl) phosphine hydrochloride (200 mM). The b2AR samples were incubated with iodoacetamide (2 mM) for 30 min to alkylate unlabeled cysteines and prevent the formation of disulfide-linked oligomers in detergent solution. The conjugation reactions were quenched by cysteine (2 mM). Labeled protein (Cy3-b2AR or Cy5-b2AR) was separated from free dye and iodoacetamide by gel filtration (Sephadex G-50 Fine). The efficiency of labeling (stoichiometry) was estimated from UV-vis absorption spectra, using the following extinction coefficients (7): e554, Cy3 = 150,000 M^-1 cm^-1, e652, Cy5 = 250,000 M^-1

cm ^-1, and e280, b2AR = 116 mM^-1 cm ^-1. The receptor extinction coefficient at 280 nm was estimated from a standard dilution curve of purified b2AR for which the concentration was determined by saturation radioligand binding. The contribution of fluorophore absorbance at 280 nm was subtracted (8% of the absorbance at 554 nm for Cy3, and 5% of the absorbance at 652 nm for Cy5, CyDye mono-reactive NHS esters handbook, GE Healthcare). To prepare the dual-labeled positive control for colocalization, a portion of Cy3-b2AR (2 mM) was labeled by incubation with the amine reactive dye Cy5-mono-NHS-ester (400 mM, GE Healthcare) for 3 h at 25°C. Labeled protein (Cy3-Cy5b2AR) was separated from free dye by gel filtration.

Reconstitution of Cy3-b2AR and Cy5-b2AR into DOPC:cholesterol vesicles. Lipid stock mixtures of DOPC (3 mg/ml, Avanti) and cholesterol hemisuccinate (0.3 mg/ml, Steraloids, Inc.) were prepared in Buffer H including 1% octyl glucoside (Anatrace). The lipid stocks were removed from storage, vortexed, and sonicated for 30 min in an ice bath. The reconstitution mixture was prepared in Buffer H + 0.1% DDM containing a 10fold dilution of the lipid stock and either Cy3-b2AR (1 mM) alone or a mixture of Cy3-b2AR (500 nM) plus Cy5-b2AR (500 nM). The final DOPC and cholesterol hemisuccinate concentrations in this mixture were 0.3 and 0.03 mg/ml, respectively. The reconstitution mixture was inverted several times and incubated for 2 h on ice. Vesicles were then formed by removing detergent by gel filtration on a Sephadex G-50 Fine column pre-equilibrated with Buffer H. The vesicle solution eluting from this column contained »250 nM b2AR, with a lipid:receptor ratio of »60:1 and was stored on ice until use.

Ensemble FRET measurements. Steady-state fluorescence spectroscopy was used to measure the association of receptors reconstituted into either phospholipid vesicles or rHDL by FRET. Fluorescence spectroscopy was performed on a Spex FluoroMax-3 spectrofluorometer (Horiba Jobin Yvon, Inc.) with photon-counting mode. All spectra were recorded at 25°C, using excitation and emission bandpasses of 3 nm and an integration time of 1.0 s nm^-1. The following spectra were recorded for each sample: Cy3 emission (excitation at 525 nm, emission from 535 to 751 nm), Cy5 emission (excitation at 625 nm, emission from 635 to 751 nm), Cy3 excitation (emission at 580 nm, excitation from 450 to 570 nm), and Cy5 excitation (emission at 690 nm, excitation from 450 to 680 nm). All spectra were corrected for the background fluorescence from buffer. The effect of cross-talk (direct excitation of Cy5 at 525 nm) was subtracted from all Cy3 emission spectra by measuring the contribution from Cy5 alone, and the spectra were normalized by area.

Fluorescence spectra were recorded for 1:1 mixtures of Cy3-b2AR plus Cy5-b2AR reconstituted into either lipid vesicles or rHDL (see above). Spectra were recorded in buffer H to measure the association of receptors in intact lipid bilayer environments. The reconstituted samples were also diluted into Buffer H plus 1.0% DDM, vortexed extensively, and incubated for 5 min to solubilize lipid vesicles or rHDL. Spectra were also obtained for two negative controls: Cy3-b2AR and Cy5-b2AR that had been reconstituted into rHDL separately and then mixed 1:1 (in Buffer H), and a 1:1 mixture of soluble Cy3-b2AR and Cy5-b2AR that had never been reconstituted (in Buffer H plus 0.1% or 1.0% DDM). The final receptor concentrations used for fluorescence spectroscopy were »10 nM for rHDL, 40 nM for vesicles, and 90 nM for soluble receptors.

By looking at the emission spectra and calculating the proximity ratio, I(acceptor)/[I(donor) + I(acceptor)], the limit of detection for FRET is a proximity ratio of 4.5-4.6 in detergent micelles. As such, the rHDL samples are slightly above threshold (proximity ratio = 6.4-6.5), and the rVesicles are very much above threshold (proximity ratio = 8.3-16.7). However, the proximity ratio is a "qualitative" measurement of the FRET efficiency.

Total internal reflectance fluorescence (TIRF) microscopy. Single molecule imaging was performed on a home-made TIRF microscope based on a Nikon TE2000-U, as described in Methods (8). An excitation filter (D640/20x, Chroma, Rockingham, VT) was inserted in the light path of the red laser to block undesired wavelengths. The two lasers are combined and coupled into a single-mode optical fiber (OZ Optics, Ottawa, Ontario, Canada). The fiber output is collimated and then focused to the back focal plane of a high numerical aperture objective (Plan Apo 100x Oil, NA = 1.4, Nikon). The emitted fluorescence is collected by the same objective, separated from the excitation laser by a multiband dichroic mirror (400-535-635 TBDR, Omega Optical, Brattleboro, VT), filtered by a band-pass filter (HQ595/50m for Cy3 imaging, HQ675/50m for Cy5 imaging, Chroma), and then imaged to an intensified CCD camera (I-PentaMAX, Roper Scientific, Tucson, AZ). The excitation lasers are controlled with mechanical shutters that are synchronized with the CCD image acquisition to minimize unnecessary photobleaching. The images are recorded by WinView (Roper Scientific) and analyzed with home-written programs.

Single-molecule imaging. Lab-Tek II chambered coverglasses (Nalge Nunc International, Rochester, NY) were used to contain and image all samples. Before each experiment, the coverglass was cleaned by sonicating in 1 M potassium hydroxide for 15 min, rinsing with MilliQ water, sonicating in MilliQ water for 5 min, rinsing, followed by drying with a nitrogen stream. To image a sample, 500 ml of »10 pM b_2AR reconstituted in rHDL was added to the chamber to allow nonspecific adsorption of rHDL to the glass surface. After a 5-min incubation, the protein solution was pipetted off and immediately replaced by 500 ml of PBS (Gibco) to stop adsorption of additional molecules from solution.

Images of Cy5 and Cy3 are acquired sequentially by switching the excitation laser and the emission band-pass filter. The excitation power density is 48.8 W/cm² for the green laser and 22.4 W/cm² for the red laser. The integration time of the CCD camera was 0.4 sec per frame. To avoid the possible masking of Cy3 fluorescence due to energy transfer from Cy3 to Cy5, 200 frames are acquired in the Cy5 channel before acquisition of the Cy3 image to ensure that all Cy5 molecules are photobleached.

Image analysis. In a fluorescent image, each local maximum exceeding a certain height threshold was fitted with a two-dimensional Gaussian function in a 7 pixel x 7 pixel area around it. A fluorescent molecule was identified if the fitted results satisfied both the height and width criteria. The fluorescence time trace of a molecule is obtained by identifying the molecule in the first frame of an image stack and then fitting each subsequent frame with a two-dimensional Gaussian function.

Photobleach step analysis. To assess the labeling stoichiometry of fluorescent receptors, the number of photobleaching steps was counted manually from the time traces of at least 400 molecules of each Cy3- b2AR, Cy5-b2AR, Cy3-b2AR•rHDL, and Cy5b2AR•rHDL. The number of photobleaching steps represents the number of detectable fluorophores bound to a receptor at the time of imaging. Receptors were classified as being single-, double-, or triple-labeled based on the observed number of photobleaching steps. In some cases, the precise number of photobleaching steps could not be discerned from the time trace (usually due to low signal-to-noise or fluorophore blinking). Such molecules were classified as "not countable". We used the most conservative estimate of the labeling efficiency, by assuming that all molecules in the "not countable" category are actually single-labeled.

To estimate the fraction of unlabeled receptor in each sample, we simulated the labeling kinetics of the three most reactive cysteines of the b2AR using a mathematical model. For WT b2AR, cysteine-265 is »10 times more reactive than the next most reactive cysteine. Given the reactivity of cysteine-265, it is intuitive that the vast majority of receptors should become singly-labeled before an appreciable fraction becomes multiply-labeled. The model confirmed this suspicion, and was able to accurately reproduce the observed fractions of double-and triple-labeling with fluorophore. The model consists of the following reactions that are equally valid for Cy3 or Cy5.

The model is described by a set of coupled differential equations:

[1]

We assumed that each reaction is irreversible and that only cysteine-265 can react with the first fluorophore (Eq. 1). The equations were solved numerically with Mathematica (Wolfram Research, Inc., Champaign, IL) using the following parameters: kCys-265 = 0.8 s^-1 mM^-1, k2 = 0.08 s^-1 mM^-1, initial [b2AR] = 5 mM. The rate constants are based on the reactivity of WT and C265A b2AR as measured by fluorescein maleimide labeling as previously described (9). The total concentration of Cy3 incorporated was arbitrarily chosen as 7.2 mM, because this value accurately reproduces the fractions of single-, double-and triple-labeling as measured by photobleach step counting (model results 64% single-labeled, 28% double-labeled, 8% triple-labeled, and 0.4% unlabeled at steady-state, compare with SI Fig. 10,assuming all "not countable" molecules are actually singly-labeled). This [Cy3] corresponds to a labeling efficiency of 1.4 mol flurophore per mol of b2AR, in good agreement with our absorbance measurements (1.6-2.5 fluorophores per protein; see text).

Colocalization analysis. For each molecule identified in the Cy3 image, the distance to each molecule identified in the Cy5 image was calculated. A Cy3/Cy5 pair was counted as colocalized if the distance between them was smaller than one CCD pixel. The percent of colocalized molecules was calculated as: 2 ´(no. colocalized spots) ¸ (no. Cy3 spots + no. Cy5 spots), assuming that a colocalized spot contains two receptors. The percent of colabeled molecules for the dual-labeled Cy3-Cy5-b₂AR sample was calculated as: (no. colocalized spots) ¸ (no. Cy3 spots + no. Cy5 spots - no. colocalized spots).

Cross-linking of Cy3-b2AR. The bifunctional amine-reactive cross-linkers bis(sulfosuccinimidyl) suberate (BS³, 11.4 Å spacer arm) or bis N-succinimidyl[pentaethylene glycol] ester (Bis(NHS)PEO5, 21.7-Å spacer arm, not shown) were purchased from Pierce Biotechnology (Rockford, IL). BS³ stock solutions of 100 mM were prepared from solid reagent in Buffer H + 0.1% DDM or dimethyl sulfoxide immediately before each use. Cross-linking reactions of Cy3-b_2AR in detergent micelles were performed in Buffer H plus 0.1% DDM at a final receptor concentration of 2 mM. Cross-linking reactions of Cy3-b_2AR reconstituted in lipid vesicles were performed in Buffer H at a final receptor concentration of »250 nM. The final concentrations of either Bis(NHS)PEO 5 or BS³ used in each case were 0, 0.25, 0.5, 1.0, 2.5, and 5.0 mM. After addition of cross-linker stock solution, the samples were mixed by inverting and incubated for 2 h on ice. The reaction was quenched by addition of 20 mM Tris. The cross-linked samples were then analyzed by SDS.PAGE (10% gels) with »0.5 mg of protein loaded in each well. Cy3 fluorescence was directly visualized using a FluorChem 8800 (Alpha Innotech) imaging system.

1. Kobilka BK (1995) Anal Biochem 231:269-271.

2. Gan KN, Smolen A, Eckerson HW, La Du BN (1991) Drug Metab Dispos 19:100-106.

3. Jonas A (1986) Methods Enzymol 128:553-582.

4. Cerione RA, Ross EM (1991) Methods Enzymol 195:329-342.

5. Kozasa T, Gilman AG (1995) J Biol Chem 270:1734-17341.

6. Asano T, Pedersen SE, Scott CW, Ross EM (1984) Biochemistry 23:5460-5467.

7. Mansoor SE, Palczewski K, Farrens DL (2006) Proc Natl Acad Sci USA 103:3060-3065.

8. Huang B, Perroud TD, Zare RN (2004) ChemPhysChem 5:1523-15231.

9. Yao X, Parnot C, Deupi X, Ratnala VR, Swaminath G, Farrens D, Kobilka B (2006) Nat Chem Biol 2:417-422.