β2 Adrenergic Receptor Fluorescent Protein Fusions Traffic to the Plasma Membrane and Retain Functionality

Green fluorescent protein (GFP) has proven useful for the study of protein interactions and dynamics for the last twenty years. A variety of new fluorescent proteins have been developed that expand the use of available excitation spectra. We have undertaken an analysis of seven of the most useful fluorescent proteins (XFPs), Cerulean (and mCerulean3), Teal, GFP, Venus, mCherry and TagRFP657, as fusions to the archetypal G-protein coupled receptor, the β2 adrenergic receptor (β2AR). We have characterized these β2AR::XFP fusions in respect to membrane trafficking and G-protein activation. We noticed that in the mouse neural cell line, OP 6, that membrane bound β2AR::XFP fusions robustly localized in the filopodia identical to gap::XFP fusions. All β2AR::XFP fusions show responses indistinguishable from each other and the non-fused form after isoprenaline exposure. Our results provide a platform by which G-protein coupled receptors can be dissected for their functionality.


Introduction
G-protein coupled receptors (GPCRs) represent the largest family of eukaryotic membrane proteins [1]. The GPCR superfamily is natively expressed in many cell types; 90% of GPCRs are expressed in the brain. They are activated by almost every known neurotransmitter, peptide, and chemokine [2]. Mutations to GPCRs are known to cause a cascade of neurological and neurodegenerative diseases [3,4]. Common to all GPCRs is their seven transmembrane structure: seven alpha helical transmembrane domains connected by intra and extracellular loops with an extracellular amino terminus and an intracellular carboxyl terminus. The intracellular regions regulate G-protein binding, while ligand-binding properties exist in the extracellular and transmembrane domains [5]. GPCR's are the target of over 50% of therapeutic drugs, and a quarter of the top marketed drugs target GPCRs [6].
GPCRs function at the plasma membrane where they bind ligand and are subsequently regulated by ligand dependent endocytosis. In order to study behavior of GPCRs, it is necessary to visually track their subcellular localization throughout the process of ligand activation, internalization, recycling, and degradation [7].
The β 2 adrenergic receptor (β 2 AR) has served as the archetype of GPCRs. It binds ligands such as catecholamines and isoprenaline, which in turn induce a signal transduction cascade, catalyzing the exchange of GDP for GTP in the alpha subunit of the heterotrimeric G-protein. This reaction results in the activation of adenylyl cyclase leading to an increase in the concentrations of cyclic AMP in the cell [8]. The distribution of receptor in the cell upon ligand activation and the process of receptor internalization and resensitization have been well characterized [9,10].
One of the most useful methods for studying β 2 AR trafficking, distribution, internalization, and resensitization has been through the use of green fluorescent protein (GFP) fusions [11,12,13]. The overexpression of β 2 AR::GFP in cell lines does not affect its cellular activities and outcomes are readily quantifiable. β 2 AR::GFP fusions have provided a tool for studying its subcellular localization. In the absence of ligands, it is found on the plasma membrane of multiple cell types. After agonist exposure, β 2 AR::GFP internalization is easily visualized within the endosomal compartments. The possibility of GPCRs existing as quaternary structures has been studied through fusions with GFP and snap tags, but the increased availability of photostable fluorescent proteins may provide another reliable method [7]. The β 2 AR has also been commonly fused to two GFP derivatives, the yellow fluorescent protein (YFP) and the cyan fluorescent protein (CFP), however, both new fluorescent proteins and improved versions of these proteins now exist that make visualization more efficient [14,15,16]. The availability of a wide range of fluorescent proteins that span the visible spectrum of light could make it possible to simultaneously compare multiple β 2 AR mutants or multiple GPCRs more effectively.
The ability to monitor subcellular biochemical processes has been revolutionized by the availability of fluorescent proteins from multiple species capable of self-synthesizing chromophore [17]. Fluorescent proteins are commonly used as molecular tags in analyses of protein function, protein-protein interactions, subcellular trafficking, and degradation [7,18]. Ever since the discovery of GFP from Aequorea victoriae, the development of fluorescent proteins in a variety of excitation spectra have been a major goal in the field. The toolbox of fluorescent tags has been proven invaluable to protein studies [19].
While many fluorescent proteins have been developed, a handful have proven to be the most useful for their photostability, detectability, narrow emission spectra, and low cellular toxicity. GFP, Venus, Cerulean, Teal, mCherry, and TagRFP657 are some of the most useful fluorescent proteins for the above reasons and have proven especially useful for colocalization and Förster resonance energy transfer (FRET) studies [19].
As an Aequorea GFP variant, Venus is one of the easiest to detect of fluorescent proteins. It is one of the most useful probes in the yellow class and can be used as an acceptor for FRET analysis with cyan fluorescent protein donors. It is highly photostable and contains a mutation, which allows it to mature more quickly and efficiently than GFP at 37°C [20]. Venus has been previously fused to the β 2 AR; however, full functional analyses have not been compiled [21,22].
Cerulean, another Aequorea derivative, is a cyan fluorescent protein with higher photostability and detectability compared with eBFP [23]. mCerulean3 is an improved version of Cerulean, which decreases susceptibility to photo bleaching and increases brightness by 67% [24]. Cerulean was considered the best choice in the cyan fluorescent protein family until the development of Teal (mTFP1) from the Clavularia sp. coral [25]. Teal is green shifted, but more easily detectable with higher photostability and acid insensitivity when compared with Cerulean and mCerulean3. Teal is a better choice of the cyan fluorescent proteins for colocalization studies; it narrows the emission spectra from 60nm with other cyan fluorescent proteins to only 30 nm. This makes it a more suitable donor for FRET analysis with Venus as the acceptor [26].
Good red-emitting fluorescent proteins (RFPs) have been difficult to develop [27]. They're quite useful in that they have longer excitation, which provides less photo toxicity and can probe deeper into biological tissue. One of the most successful of the monomeric RFPs is mCherry from the Discosoma sp. coral. Its brightness is 50% of GFP, which is high for RFPs. It is also the most photostable of the reds [28].
Even more difficult to develop then the RFPs have been the far-reds. Recently, the red fluorescent protein mKate from the sea anemone Entacmaea quadricolor was shifted far-red with 10 amino acid substitutions. This variant has been named TagRFP657. The creation of this far-red protein now allows the use of three reds at once (TagRFP, mKate, and TagRFP657) for colocalization studies [29]. Recently, a more easily detectable version of TagRFP657 with a single lysine substitution, T10K, has been identified and is hereafter named auburn fluorescent protein (AFP).
We have tested a set of seven structurally different fluorescent proteins (Cerulean, mCerulean3, Teal, GFP, Venus mCherry, and AFP) derived from four distinct organisms as fusions to the mouse β 2 AR. We have characterized these chimeric proteins with respect to plasma membrane trafficking and response to ligand in order to determine if these fusions affect GPCR behavior. Our data show that trafficking and function of the β 2 AR is not affected by fluorescent protein fusions, making them useful tools for GPCR analyses. Our goal is to increase the resources for studying GPCR trafficking in order to elucidate structure function information.

Plasmid Construction
To create the in vitro expression vector (D346) for the β 2 AR fusion constructs, the peGFP-N1 (ClonTech) was digested with EcoRI and NotI to remove part of the multiple cloning site and GFP and an AscI site was added with linkers for cloning of other fluorescent protein fusion constructs (L34: 5' AATTCGGCGCGCCAAAAGC 3'; L35: 5' GGCCGCTTTTGGCGCGCCG 3'). ( Figure S1.) To create the gap fluorescent protein fusion constructs, each fluorescent protein was amplified by PCR with primers containing the gap sequence at the 5' end and EcoRI sites at the 5' and 3' ends (U196 : 5' CATTCAAATATGTATCCGCTCATGAGACAATAACCCTGATA AATGCTTCAATAATATTGAAAAAGGAAGAGTGCCACCATG CTGTGCTGCATCAGAAGAACTAAGCCGGTTGAGAAGAATG AAGAGGCCGATCAGGAGATGGTGAGCAAGGGCGAGG 3' and P422: 5' TTACTTGTACAGCTCGTCCATG 3'). The PCR product was shuttled into the pGEM-T Easy (Promega) vector system. Each pGEM-T/gap fluorescent protein construct was then digested with EcoRI and shuttled into D346 (Figure S1 B).
To create the β 2 AR fusion constructs, the mouse (129) β 2 AR coding sequence was amplified by PCR to remove the stop codon and to add a 5' AscI site and a 3' PacI site (File S1 for sequences). The PCR product was shuttled into the pGEM-T Easy Vector system (Promega) followed by a digest with EcoRI (from pGEM-T)/PacI and shuttled into a modified D346 vector. The D346 vector already contained a GPCR fusion protein scaffold as follows: EcoRI-AscI-GPCR-PacI-::GFP-AscI-NotI. This created the parent construct D357, β 2 AR::GFP for all other XFP insertions. The other six fluorescent proteins were (5' TTACTTGTACAGCTCGTCCATG 3') using the NotI site from pGEM-T. The β 2 AR /peGFP-N1 construct and the fluorescent protein/pGEM-T constructs were doubled digested with PacI/ NotI and each fluorescent protein was shuttled into the β 2 AR / peGFP-N1 parent construct (Figure S1 C). TagRFP657 was modified with a T10K mutation and GFP ends, as found in addgene, and confirmed by Vladislav Verkhusha as being more stable (personal communication). In our analyses, TagRFP657 was virtually undetectable as compared to AFP. To create the untagged fluorescent protein constructs, each fluorescent protein (P463/P196 or P422) was digested with EcoRI or EcoRI/NotI out of pGEM-T and shuttled into the in vitro expression vector (Figure S1 A). Templates for mCerluean3 and TagRFP657 were derived from minigenes (Genewiz). See File S1 for sequences.
Plasmid DNA constructs were transiently transfected using the Amaxa Nucleofector (Lonza) with PBS at 60%-70% confluency according to the manufacturer's protocol. Transfected cells were allowed to recover for 24 hours at 33°C and express the plasmid DNA.
For ligand activation experiments, OP 6 cells were exposed to 10 µM of isoprenaline (Sigma Aldrich Cat. # 15627) at 33°C. Live imaging of OP 6 cells was performed in culture media, except for labeling of the early endosomal compartments, which was performed in Opti-MEM (Gibco, Cat. #10569) with 20µg/ml of Alexa Fluor labeled transferrin (488 or 647, Life Technologies Cat. #s T-13342, T-23366) for 30 minutes before imaging.

Immunostaining
After co-transfection with β2AR::XFP and a myc-tagged β-arrestin2 (OriGene), OP 6 cells were washed once with PBS (Millipore) and either exposed to 10 µM of isoprenaline for 20 minutes and then fixed in 10% formalin (Fisher Scientific) for 20 minutes at room temperature (RT), or fixed immediately without stimulation of isoprenaline. Cells were then washed three times with RT PBS and blocked in blocking buffer containing 0.3% Triton X-100 (Roche) and 5% horse serum (Gibco) in PBS for 1h at RT. Subsequently, an anti-myc (rabbit) polyclonal antibody (Santa Cruz Cat. # sc-789) was added to the cells in a 1:100 dilution and incubated overnight at 4°C. Cells were then washed three times with PBST buffer (0.3% Triton X-100 in PBS), and blocking buffer with 1: 500 goat anti-rabbit Alexa Fluor 546 or 488 (Life Technologies) was added for 1 h at RT.
The cells were then imaged in PBS immediately after immunostaining.
For detection of wild type β 2 AR, antibody against β 2 AR (goat; Santa Cruz Cat# sc-570) was used at a 1:100 dilution followed by a 1:500 dilution of goat anti-rabbit Alexa Fluor 488 secondary antibody.

Laser Scanning Confocal Microscopy
OP 6 cells were imaged on a Zeiss LSM 510 microscope using a Zeiss, Plan-APO 25X water immersion objective. Cerulean, mCerulean3, and Teal were excited at 458 nm and collected at BP465-510 nm, GFP was excited at 488 nm and collected at BP500-545 nm. Venus was excited at 514 nm collected at LP520 nm, mCherry was excited at 561 nm and collected at LP575 nm, and AFP was excited at 633 nm collected at LP650 nm. All images were acquired using the multi-tracking feature. The time-lapse for β 2 AR::AFP was done on the LSM510 nm using excitation at 630 nm and collected at LP650 nm.

Spinning Disk Confocal Microscopy
For time-lapse experiments, OP 6 cells expressing the β 2 AR fusion constructs were exposed to 10µM isoprenaline and single plane images were taken every 5 minutes for 20 minutes on one of two spinning disk microscopes. For β 2 AR::GFP, β 2 AR::mCherry, and β 2 AR::Venus, an inverted LEICA CTR4000 microscope with Hamamatsu EM-CCD camera and CSU10 Yokogawa confocal unit was used with a LEICA 1.4 NA 60x oil immersion lens. β 2 AR::GFP and β 2 AR::Venus were excited at 491 nm and collected at BP(FF01-520/35 nm). β 2 AR::mCherry was excited at 561 nm and collected at BP620/660 nm. Images were obtained using the Volocity software. β 2 AR::Cerulean, β 2 AR::mCerulean3, and β 2 AR::Teal were imaged using the PerkinElmer ultraVIEW ERS system, which includes an inverted Nikon ECLIPSE TE2000-U microscope with Hamamatsu ORCA-EM camera and PerkinElmer ultraVIEW PhotoKinesis confocal unit, and a LEICA 1.4 NA 60x oil immersion lens. β 2 AR::Cerulean, β 2 AR::mCerulean3, and β 2 AR::Teal were excited at 440 nm and collected at LP485 nm. Images were obtained using the Volocity software. hours, and cells were allowed to recover for six hours in culture media before the FLIPR assay.

Statistical tests
To determine the significance of filopodia in cells expressing β 2 AR::XFPs or gap::XFPs in comparison to untagged fluorescent proteins, a two-tailed Fishers exact test using the method of summing small P values was performed for each fluorescent protein using the online Graph Pad calculator. This test was also used to determine the significance of the increase in Intensity Correlation Quotient (ICQ) values in colocalization experiments.
For transferrin and β-Arrestin2 experiments, images were processed in ImageJ for Intensity Colocalization Analysis (ICA) as described by Li et al [31]. The protocol for ICA in ImageJ was obtained from the online manual for the WCIF-ImageJ collection through the Wright Cell Imaging Facility (http:// www.uhnresearch.ca/facilities/wcif/imagej/ colour_analysis.htm). The Intensity Correlation Quotient (ICQ) values were obtained for cells prior to isoprenaline exposure and after isoprenaline exposure. Colocalization analysis for each β 2 AR::XFP combined visualization of fluorescent markers with change in ICQ values after isoprenaline stimulation.

Localization of Untagged Fluorescent Proteins (XFPs)
Do our set of untagged fluorescent proteins contain any internal signals for trafficking to the plasma membrane? We developed a transient transfection system using an olfactory placode cell line (OP 6) to test the fluorescent proteins expressed under a CMV promoter (Figure S1 A). OP 6 cells exhibit several morphological features common to neurons including long axonal like extensions and numerous filopodia. After 24 hours, each fluorescent protein was diffusely distributed in the cytoplasm ( Figure S2 A-G, Ai-Gi). GFP, Venus, mCherry, and Teal were the best detected with the highest stability while Cerulean, mCerulean3, and AFP were more difficult to detect and sustained higher instances of photobleaching.

Plasma Membrane Expression of gap::XFPs and β 2 AR::XFPs
To define plasma membrane expression of each fluorescent protein in OP 6 cells, a 20 amino acid sequence (gap) from Zebrafish gap43 protein containing an amino terminal palmitoylation site was fused to Cerulean, mCerulean3, Teal, GFP, Venus, mCherry, and AFP [32,33]. The gap fusion constructs were transiently transfected into OP 6 cells and imaged after 24 hours. All gap fusions (gap::XFPs) were diffusely distributed at the plasma membrane and localized to the filopodia of the cell (Figure 1 A-G). OP 6 cells have a distinct morphology and display long filopodia, which are highly membranous (Figure 1 Ai-Gi). The fluorescence of Venus, GFP, mCherry, and Teal again were easiest to detect, while Cerulean, mCerulean3, and AFP were harder to detect and prone to photobleaching, respectively. Since expression of protein in the filopodia was unique to the gap::XFP constructs, this feature of OP 6 cells was used to define plasma membrane expression. Of note, the fluorescence of untagged XFPs appeared ill defined compared to the discreet gap::XFPs fluorescence in the OP 6 cells.
To determine if GPCR fusions to Cerulean, mCerulean3, Teal, GFP, Venus, mCherry, and AFP affect trafficking and ligand binding, constructs were created in which each fluorescent protein is fused to the carboxyl terminus of the 2 adrenergic receptor (β 2 AR). At a minimum, the β 2 AR::XFP fusions should localize to the plasma membrane where the β 2 AR normally functions. To determine if each β 2 AR fluorescent protein fusion can traffic to the plasma membrane, OP 6 cells were transfected with each β 2 AR fusion and assayed for localization (Figure 1 H-N). All cells expressing the β 2 AR fluorescent protein fusions showed fluorescence indistinguishable from gap::XFPs including localization to the filopodia (Figure 1 Hi-Ni). Like the gap::XFP fusions, Venus was the easiest to detect with highest photostability followed by GFP, mCherry, Teal, mCerulean3, Cerulean, and AFP.

Filopodia Counts for Untagged Fluorescent Proteins, gap::XFPs, and β 2 AR::XFPs
To determine if protein localization at the filopodia was a means to identify plasma membrane expression in cells, each untagged fluorescent protein, gap::XFP, and β 2 AR::XFP were expressed, and filopodia of 10 random cells were counted. The total filopodia counts for gap::XFPs were 78-141; β 2 AR::XFPs were 97-167; as compared to untagged fluorescent protein were 4-16 ( Figure 2). The numbers of filopodia present in β 2 AR::XFPs and gap::XFPs were indistinguishable and both were significantly different from untagged XFPs (Fishers exact test P<0.0001).

Internalization of β 2 AR::XFPs
Membrane trafficking is only one measure of the functionality of a GPCR. Internalization of the β 2 AR::XFPs was investigated after acute stimulation by isoprenaline in OP 6 cells. Prior to stimulation, β 2 AR::XFPs are found in the plasma membrane with some perinuclear punctate expression, most likely due to over expression of the protein construct. After stimulation with 10µM isoprenaline, rapid internalization of the β 2 AR fusions into a punctate pattern was observed. Both the size and number of punctate aggregations of the β 2 AR fusions increase until the end of the 20-minute time course movies (Movies S1-S8). This pattern was obvious for all β 2 AR::XFP constructs, indicating that there is no adverse affect to internalization experienced by the fusion of the seven tested fluorescent proteins to the β 2 AR. By contrast, in the absence of ligand β 2 AR::GFP localization was unchanged after a 20-minute time course (Movie S8). Internalization was mediated by the β 2 AR::XFP fusions whereas gap::Teal coexpressed with untagged Venus was not internalized by isoprenaline and no punctate intracellular expression was observed (Figure S3   confirmed that OP 6 cells express the Gs α or Golf α G-protein subunit (data not shown).

Colocalization of β 2 AR::XFP with Transferrin after Ligand Stimulation
A second measure of β 2 AR::XFPs functionality after stimulation by ligand is its uptake into early endosomes for recycling. The pathway for transferrin endocytosis is known to be identical to that of the β 2 AR. Live cells were incubated with Alexa Fluor 647 labeled transferrin or Alexa Fluor 488 labeled transferrin for 30 minutes, followed by stimulation with 10µM isoprenaline for 20 minutes. Single cells were imaged prior to the addition of isoprenaline, and again after 20 minutes of isoprenaline exposure. Prior to ligand stimulation, β 2 AR::XFPs are diffusely distributed at the plasma membrane, and transferrin is punctate and localized perinuclearly with little colocalization with any of the β 2 AR fluorescent protein fusions as shown by low dependence ICQ values (Figure 3 A-G, 0 minutes). After stimulation with isoprenaline for 20 minutes, each fusion becomes punctate and colocalizes with transferrin ( Figure 3 A-G, 20 minutes). The intensity correlation quotients (ICQ) revealed that all β 2 AR::XFPs colocalized with transferrin (P<0.0001; see figure legend for ICQ values).

Colocalization of β 2 AR::XFPs with β-Arrestin2 after Ligand Stimulation
β-Arrestin2 is a regulatory protein involved in desensitization and sequestration of β 2 AR after agonist binding and phosphorylation [34]. After ligand stimulation, it is necessary for β 2 AR::XFPs to be regulated by β-Arrestin2 for proper functionality. β-Arrestin2 was not detectable in OP 6 cells by immunostaining (data not shown), so cells were cotransfected with a β 2 AR fusion and a myc-tagged β-Arrestin2 construct. In the absence of isoprenaline, all β 2 AR::XFPs were diffusely  distributed at the plasma membrane, and β-Arrestin2 (detected by a polyclonal myc antibody) was diffusely distributed in the cytoplasm with little colocalization with the β 2 AR fusions as shown by low dependence ICQ values (Figure 4 A-G, 0 minutes). After exposure to isoprenaline, all β 2 AR fusions colocalized with β-Arrestin2 with high dependence described by ICQ values (Figure 4 A-G, 20 minutes; see figure legend for ICQ values). The internalization was marked by punctate perinuclear expression with some expression remaining at the plasma membrane.

Time Course Colocalization Profiles of β 2 AR::Teal and β 2 AR::Venus
Teal and Venus have been previously characterized as excellent pairs for FRET analysis. To assure these fluorescent proteins did not affect the time course of β 2 AR internalization and could therefore be used as a FRET pair with GPCRS, colocalization of β 2 AR::Teal and β 2 AR::Venus constructs with transferrin and β-Arrestin2 after isoprenaline exposure were determined every 4 minutes for a total of 20 minutes using ICQ values (Figure 5 A,B). Both β 2 AR::Teal and β 2 AR::Venus were steadily internalized into the early endosomes as shown by corresponding increases in ICQ values over the 20 minute time course (Figure 5 A). Both β 2 AR::Teal and β 2 AR::Venus were able to interact indistinguishably with β-Arrestin2 showing sharp increases in dependency (ICQ) with β-Arrestin2 that plateau by the 8 minute time point ( Figure 5B).

Dose Response Profiles of β 2 AR::XFPs to Isoprenaline
Full functionality was finally assessed by the degree to which β 2 AR::XFPs were able to activate G-proteins. The FLIPR calcium assay was used to measure the release of internal calcium stores mediated by the Gq pathway (i.e. Gα15) in response to isoprenaline exposure. Response to increasing concentrations of ligand by β 2 AR::XFPs and subsequent human Gα15 activation was measured in relative fluorescent units (RFUs). Normalized dose response curves for all β 2 AR::XFPs had virtually identical EC50's, ranging from 2.6-5.2 x 10 -9 M similar to the unfused mouse β 2 AR and to the human β 2 AR::GFP, which had EC50s of 2.6 x 10 -9 and 3.1 x 10 -9 , respectively (Figure 6 A, Figure S4). Unnormalized dose response curves for each fluorescent protein fusion compared to unfused mouse β 2 AR revealed a decrease in maximum RFUs (Figure 6 B, Figure S4). β 2 AR::GFP with a 90 amino acid carboxyl terminal truncation to arginine 328 (β 2 AR::GFP truncated), did not traffic to the plasma membrane despite containing all seven transmembrane domains, and did not show a response to ligand (Figure 6 A-C). β 2 AR::GFP truncated localized perinuclearly and was most likely retained in the endoplasmic reticulum. This is in contrast to human β 2 AR::GFP, which did traffic to the filopodia and was stimulated by isoprenaline (Figure 6 A, B, D).

Discussion
Here, we have determined that some of the most useful fluorescent proteins do not affect the mouse β 2 AR trafficking and function when fused to its carboxyl terminus, consistent with the previous studies of human β 2 AR::GFP [11,12,13]. Since the β 2 AR is the archetype of GPCRs, it is likely that these fluorescent proteins do not affect trafficking and ligand response for other canonical GPCRs, providing a tool that makes the live visualization of protein localization, trafficking, and protein-protein interaction easier and efficient.
To determine the effects of XFPs on GPCRs, the ability of β 2 AR::XFPs were analyzed in the following ways: trafficking to the plasma membrane, internalization and trafficking to the early endosomes after ligand stimulation, regulation by β-Arrestin2 after ligand stimulation, and concentration dependent response to ligand. Our results confirm and extend previous findings. All of the β 2 AR::XFPs show functionality identical to that of β 2 AR::GFP whose behavior has been previously shown to be indistinguishable from β 2 AR WT [11,12,13].
We show that our mouse β 2 AR::XFPs and the untagged mouse β 2 AR WT have equivalent dose response profiles to that of human β 2 AR::GFP, which has been previously described. Importantly, the EC50 values, an indirect measure of the affinity of the receptor to ligand, were not affected and suggest a proper folding and integration of the fusion proteins into the cell membrane. All β 2 AR::XFPs showed lower maximum RFUs as compared to β 2 AR WT. This result could be due to slightly less protein being trafficked to the plasma membrane or slight steric hindrance caused by the intracellular XFPs, which could reduce the transduction efficacy without affecting the ligand affinity.
Although each fluorescent protein did not affect β 2 AR function, there were notable differences between fluorescent proteins that could affect their application. Differences were within detectability, photostability, and aggregation within intracellular compartments. Venus was the best fluorescent protein, both untagged and as fusions to the gap peptide and β 2 AR. After Venus, in order of detection ease, were GFP, Teal, mCherry, mCerulean3, Cerulean, and AFP. This order remained true whether fused or unfused. mCerulean3, an improved version of Cerulean, was more easily detected with lower instances of photobleaching, as previously described. However, Teal was considerably easier to detect and the most photostable of the cyans, which is also consistent with the literature [26]. Cerulean was the most vulnerable to photobleaching. When fused to β 2 AR and the gap peptide, mCherry and AFP were the most susceptible to aggregation in subcellular compartments, likely due to cellular toxicity.
Our protein trafficking analyses made use of OP 6 cells (an olfactory neuronal precursor) [30], which vary in morphology, but all share the feature of many filopodia. When the untagged β2adrenergic Receptor Fluorescent Protein Fusions PLOS ONE | www.plosone.org fluorescent proteins are expressed in OP 6 cells, they localize to the cytoplasm with few to no filopodia visible ( Figure S2 Ai-Gi). However, when each fluorescent protein is tagged with the gap peptide, all of the gap::XFPs localize to the plasma membrane, and are expressed at the filopodia (Figure 1 Ai-Gi). Localization of untagged fluorescent proteins were rarely found in filopodia (~1 per OP6 cell) as compared to the gap tagged fluorescent proteins, which marked ~12 filopodia in nearly all OP 6 cells ( Figure 2). This is consistent with the membrane association of the palmitoylated gap peptide, whose localization has been previously described in the filopodia of neuronal and non-neuronal cells [32,33].
Using this evidence, we have defined protein expression at the plasma membrane by its presence within filopodia. OP 6 cells are an efficient platform to measure plasma membrane trafficking efficacy of a protein. All seven β 2 AR fluorescent protein fusions were robustly found in the filopodia at ~14 filopodia in nearly all OP 6 cells. Thus, localization of the β 2 AR::XFPs were indistinguishable from that of gap::XFPs. In the future, OP 6 cells can be used to analyze the mechanisms of plasma membrane trafficking of β 2 AR::XFPs and other GPCR::XFPs.
A large number of GPCRs cannot be expressed in heterologous cell lines. This failure has prevented both structure-function and crystallographic analyses. Expression of mutant GPCRs::XFPs in OP 6 cells provides a high-throughput platform for analysis of plasma membrane trafficking using filopodia counts as an indicator of trafficking success. Moreover, OP 6 cells can be used to study the regulation of GPCRs during ligand activation, recycling after ligand activation, and G-protein coupling. Multiple fluorescent proteins can be coexpressed, providing visualization of subcellular localization of mutants in comparison to wild type receptors. This system is ideal for applications of live cell imaging, and utilizing time course analyses of subcellular processes, eliminating the need for antibodies in these applications.

Supporting Information
Live imaging of β 2 AR::Cerulean rapid internalization after exposure to isoprenaline. A single cell expressing β 2 AR::Cerulean exhibits diffuse expression at the plasma membrane. After 20 minutes of isoprenaline exposure, β 2 AR::Cerulean becomes punctate and internalized, with some protein remaining at the plasma membrane. This one-hour time course for β 2 AR::Cerulean reveals dynamic remodeling of the cell. (MOV)

Movie S2.
Live imaging of β 2 AR::mCerulean3 rapid internalization after exposure to isoprenaline. A single cell expressing β 2 AR::mCerulean3 exhibits diffuse expression at the plasma membrane. Twenty minute time-lapse imaging after isoprenaline exposure shows β 2 AR::mCerulean3 becoming punctate and internalized, with some protein remaining at the plasma membrane. (MOV) Movie S3. Live imaging of β 2 AR::Teal rapid internalization after exposure to isoprenaline. A single cell expressing β 2 AR::Teal exhibits diffuse expression at the plasma membrane. Twenty minute time-lapse imaging after isoprenaline exposure shows β 2 AR::Teal becoming punctate and internalized, with some protein remaining at the plasma membrane. (MOV) Movie S4. Live imaging of β 2 AR::GFP rapid internalization after exposure to isoprenaline. A single cell expressing β 2 AR::GFP exhibits diffuse expression at the plasma membrane. Twenty minute time-lapse imaging after isoprenaline exposure shows β 2 AR::GFP becoming punctate and internalized, with some protein remaining at the plasma membrane. (MOV)

Movie S5.
Live imaging of β 2 AR::Venus rapid internalization after exposure to isoprenaline. A single cell expressing β 2 AR::Venus exhibits diffuse expression at the plasma membrane. Twenty minute time-lapse imaging after isoprenaline exposure shows β 2 AR::Venus becoming punctate and internalized, with some protein remaining at the plasma membrane. (MOV)

Movie S6.
Live imaging of β 2 AR::mCherry rapid internalization after exposure to isoprenaline. A single cell expressing β 2 AR::mCherry exhibits diffuse expression at the plasma membrane. Twenty minute time-lapse imaging after isoprenaline exposure shows β 2 AR::mCherry becoming punctate and internalized, with some protein remaining at the plasma membrane. (MOV) Movie S7. Live imaging of β 2 AR::AFP rapid internalization after exposure to isoprenaline. A single cell expressing β 2 AR::AFP exhibits diffuse expression at the plasma membrane. Twenty minute time-lapse imaging after isoprenaline exposure shows β 2 AR::AFP becoming punctate and internalized, with some protein remaining at the plasma membrane. (MOV) Movie S8.
Live imaging of β 2 AR::GFP shows no internalization in the absence of isoprenaline. A single cell expressing β 2 AR::GFP exhibits diffuse expression at the plasma membrane. β 2 AR::GFP remains in the membrane and is not internalized after 20 minute time-lapse imaging in the absence of isoprenaline. (MOV)