The evolving small‐molecule fluorescent‐conjugate toolbox for Class A GPCRs

The past decade has witnessed fluorescently tagged drug molecules gaining significant attraction in their use as pharmacological tools with which to visualize and interrogate receptor targets at the single‐cell level. Additionally, one can generate detailed pharmacological information, such as affinity measurements, down to almost single‐molecule detection limits. The now accepted utilization of fluorescence‐based readouts in high‐throughput/high‐content screening provides further evidence that fluorescent molecules offer a safer and more adaptable substitute to radioligands in molecular pharmacology and drug discovery. One such drug‐target family that has received considerable attention are the GPCRs; this review therefore summarizes the most recent developments in the area of fluorescent ligand design for this important drug target. We assess recently reported fluorescent conjugates by adopting a receptor‐family‐based approach, highlighting some of the strengths and weaknesses of the individual molecules and their subsequent use. This review adds further strength to the arguments that fluorescent ligand design and synthesis requires careful planning and execution; providing examples illustrating that selection of the correct fluorescent dye, linker length/composition and geographic attachment point to the drug scaffold can all influence the ultimate selectivity and potency of the final conjugate when compared with its unlabelled precursor. When optimized appropriately, the resultant fluorescent conjugates have been successfully employed in an array of assay formats, including flow cytometry, fluorescence microscopy, FRET and scanning confocal microscopy. It is clear that fluorescently labelled GPCR ligands remain a developing and dynamic research arena.


Introduction
Recent years have witnessed a rapid expansion in the use of fluorescence-based techniques with which to interrogate biological processes and receptors of physiological and pharmacological importance. There are a number of methods by which a fluorescent probe can be generated, including genetic manipulation to label a protein with a fluorophore (Dedecker et al., 2013), measurement of the inherent fluorescence of a compound (Beltran et al., 2011;Burchak et al., 2011) or by using synthetic chemistry to covalently link a biologically active compound to a fluorophore of choice, creating a fluorescent ligand conjugate (Daly and McGrath, 2003). Fluorescent ligands can be designed to interact with different entities, for example, as reaction-based probes, which offer a powerful technique for detecting and studying small molecules and/or metal ions of interest in living systems (Chan et al., 2012). However, the most prevalent use of fluorescent ligands has been the study of protein-protein interactions (Kale et al., 2012) or ligand-receptor interactions (Leopoldo et al., 2009) in biological systems.
One of the most important human receptor families, from a drug discovery and development perspective, is the GPCRs (Alexander et al., 2013). GPCRs are 7-transmembrane spanning receptors, which account for nearly 4% of the proteinencoding human genome (Bjarnadóttir et al., 2006) and are the target of approximately 30% of all marketed drugs (Overington et al., 2006). GPCRs have been classified into five different classes (http://www.gpcr.org/7tm/proteinfamily), of which Class A is the largest and generally regarded as the most understood. GPCRs are signalling powerhouses and can regulate various intracellular biological cascades via the binding of extracellular endogenous ligands, such as peptides, hormones and neurotransmitters. There is significant interest surrounding the development of fluorescent ligands with which to study GPCRs, and research reports of fluorescent GPCR ligands have been previously reviewed (Middleton and Kellam, 2005;Kuder and Kieć-Kononowicz, 2008;Böhme and Beck-Sickinger, 2009). This article summarizes the smallmolecule (non-peptide) fluorescent conjugates for Class A GPCRs that have been reported subsequent to these aforementioned reviews.
Fluorescent ligands are powerful tools to study GPCRs as they can be employed in many varied experiments to reveal insight into receptor structure and function in native, live cells . High-affinity fluorescent antagonists can be used to label the target GPCR, and fluorescently tagged agonists can provide a means to monitor dynamic processes such as receptor internalization and trafficking (cf. examples within this review). A fluorescent ligand can be used as the competing probe in a competition-based binding assay (Cottet et al., 2011;Sexton et al., 2011;Stoddart et al., 2012) instead of a radiolabelled ligand, thereby avoiding the inherent safety risks, legal issues and disposal costs associated with the latter. In addition to measuring the direct displacement of a competing fluorescent ligand from a GPCR orthosteric site, fluorescent ligands have enormous potential for revealing elaborate and intricate details about receptor oligomerization through the use of FRET and BRET assays (Albizu et al., 2010;Cottet et al., 2011;. Kinetic measurements of the fluorescent ligand-receptor interaction can reveal insight into receptor allosterism (Hill et al., 2014, this issue) and receptor dimerization (May et al., 2011), while ligand-receptor diffusion times measured using techniques such as fluorescence correlation spectroscopy (FCS) can be used to distinguish different receptor complexes (Briddon and Hill, 2007;Jakobs et al., 2012). This brief precis highlights just some of the possible applications of fluorescent ligands in what is becoming a rapidly expanding field.

Fluorescent ligands for GPCRs
The design of a small-molecule-based fluorescent probe begins with selecting an amenable parent pharmacophore, surveying where to append the linker, determining what linker to use and, lastly, deciding what fluorophore to covalently tether (Jacobson, 2009). The linker position on the parent ligand must be tolerant to chemical change, which is often driven by existing structure activity relationship (SAR) data where available. There is an increasingly diverse range of commercially available fluorophores, and often the commercial fluorophore can be purchased as the pre-activated N-hydroxysuccinimidyl (NHS) ester primed for coupling to an amine on the pharmacophore-linker congener. Properties to consider when selecting the appropriate fluorophore to append to a congener include the absorption and emission profile of the fluorophore, lipophilicity (which can influence the conjugate's ability to diffuse across the cell membrane) and whether the fluorophore is quenched in certain environments. From a GPCR imaging perspective, the fluorescent ligand will ideally not enter the cells (unless bound to the internalized receptor), show very low levels of non-specific membrane binding, be quenched when not bound to the receptor and/or cell membrane and, for most applications, be displaceable using higher concentrations of a known nonfluorescent ligand that targets the same receptor. Once assembled, the fluorescent conjugate must be rigorously pharmacologically characterized, as its profile in terms of affinity and/or efficacy may be very different from that of the parent ligand. In the following sections, we review the small-molecule-based fluorescent conjugates that have been developed for Class A receptor families since publication of earlier reviews of this subject area (Middleton and Kellam, 2005;Kuder and Kieć-Kononowicz, 2008;Böhme and Beck-Sickinger, 2009).

Adenosine receptor
The use of fluorescent probes for studying the adenosine receptor has recently been comprehensively reviewed by Kozma et al. (2013b). The fluorescent ligand toolbox for the adenosine receptor family is relatively advanced compared with other Class A GPCRs, with many reports of both antagonist and agonist-based probes built around different pharmacophores (predominately for the A1and A3-adenosine receptor subtypes) by the research groups of Jacobson and Hill/Kellam (refer to references within Kozma et al., 2013b). Use of fluorescent antagonists for the adenosine A1-and A3-receptors is now at a stage where they can be used in place of radioligand-binding studies for screening purposes. A good example of this is a recent report from our laboratories of high-content screening of a fragment library to identify new synthetic scaffolds for the human A1-and A3-adenosine receptor family subtypes (Stoddart et al., 2012). Since the review by Kozma and colleagues, there has been one additional account comparing the pharmacology and imaging properties of three new agonist-based fluorescent adenosine A3-receptor probes (Kozma et al., 2013a) to five alternatives that had been previously reported (Tosh et al., 2009). The new compounds included an IR dye 700 DX conjugate (1) linked through the C2 position of the adenine nucleoside ring, and two N6-linked Alexa Fluor 488 probes (2) and (3) synthesized by click-coupling between an azide and alkyne. The three novel conjugates unfortunately displayed a weaker adenosine A3-receptor potency when compared with the originally reported fluorescent ligands, and therefore, the authors proceeded with imaging studies using the previously reported Cy5-containing MRS5218 (Tosh et al., 2009) as their firstchoice fluorescent probe to visualize and study both the human and the mouse A3-receptor. Martikkala et al. (2009) constructed three europium(III)labelled probes for the β2-adrenoceptor by coupling amino pindolol derivatives containing different linker lengths to isothiocyanate-activated europium chelates [the chemical moiety(s) of these chelates were not disclosed]. The compounds with the shortest (4) and longest (5) linker-length were employed in a competitive time-resolved fluorescence emission-binding assay using the beta-blocker propranolol as the model drug. IC50 values of 60 and 37 nM for the human β2-adrenoceptor were obtained for propranolol using (4) and (5), respectively, compared to a value of 33 nM calculated from a [ 3 H]dihydroalprenolol radioligand displacement assay. It was interesting to note that an intermediate linker-length europium conjugate (5 minus one of the heptanamide units) could not displace propranolol from the β2-adrenoceptor. This is intriguing, as one might expect the structure-activity relationship trend of linker length to be consistent, up to a point, in any one direction. This demonstrates that a fluorescent conjugate can possess a complex, unique and often unpredictable pharmacological profile compared with the parent pharmacophore.

Adrenoceptor
In a comprehensive study from our laboratories, a series of red-fluorescent β-adrenoceptor ligands were synthesized based on three different orthosteric β-antagonist head groups; namely propranolol, alprenolol and pindolol (Baker et al., 2011). Using alkyl-or polyether-based linker extensions, the resultant propranolol (6) and alprenolol-based (7) fluorescent β-blockers displayed high affinity for the human β2-adrenoceptor. This study provided a further example of how subtle changes in the structural nature of the linker can exert a significant impact on the final conjugate's pharmacology. The 8-carbon linker analogue of 6, where the 'PEGlike' linker was replaced with a hydrocarbon chain, showed a 10-fold lower affinity for the β2-adrenoceptor compared to 6. In contrast, when the linker of 6 was replaced with a shorter 4-carbon linker, the conjugate's affinity for the β2-adrenoceptor was comparable to 6. Conjugate 6 was used to visualize ligand-receptor binding in CHO-β2 cells expressing the human form of the β2-adrenoceptor using confocal microscopy, and displayed clear labelling of the membranebound receptors at 3 nM. This specific binding could be attenuated by incubation with various concentrations (1-100 nM) of the β2-selective antagonist ICI 118551. In this study, it was also of interest to note that the pindolol-based fluorescent conjugates showed significant loss of affinity when compared with the native drug molecule. Even with three orthosteric ligands acting upon the equivalent receptorbinding pocket, one cannot therefore assume that installation of a fluorophore onto the analogous position of a congener will afford similar pharmacological outcomes with regard to the final conjugate. Morishima et al. (2010) developed a high-affinity fluorescent probe (8) selective for the α1A-adrenoceptor and α1Ladrenoceptor (thought to be an α1A-adrenoceptor phenotype) over the α1B-adrenoceptor and α1D-adrenoceptor subtypes. The α1A-adrenoceptor subtype selective antagonist silodosin, which is used to treat bladder outlet obstruction, was labelled with an Alexa Fluor 488 fluorophore. The authors did not disclose which isomer of the fluorophore was used, and therefore the Alexa Fluor 488 mixture of 5′ and 6′ isomers has been depicted in Table 1. While fluorescent probe 8 displayed a 10-fold reduction in binding affinity across the human adrenoceptor receptor subtypes as compared with silodosin, it retained an α1A-adrenoceptor selectivity profile (100-and 15-fold selective over α1B-adrenoceptor and α1D-adrenoceptor respectively). Fluorescent confocal microscopy demonstrated that 8 localized to the membrane of CHO cells overexpressing the α1A-adrenoceptor, and this binding could be significantly reduced using the high-affinity selective antagonist prazosin. Building on this promising result, 8 was used to visualize the α1L-adrenoceptor, an α1A-adrenoceptor phenotype, localized to the muscle layer of the human prostate.

Angiotensin receptor
A fluorescent angiotensin II AT1 receptor (AT1R) ligand has been reported, derived from a sartan-based pharmacophore (Giarrusso et al., 2012). In place of the thiophene carboxylate moiety of the antagonist milfasartan, Giarrusso et al. (2012) instead installed various polyaromatic hydrocarbons or a coumarin fluorophore (e.g. 9) by reacting an alkyl halide with the pyrimidinone pharmacophore core. The authors remark that polyaromatic conjugates, such as those containing a naphthalene moiety, were clearly unsuitable for in vivo use; therefore, coumarin-conjugate 9 was further evaluated as a potential visual AT 1R probe. Functional analysis using CHO cells expressing the rat AT1aR revealed that 9 was an antagonist of the AT1R, with an estimated pKb value similar to the native drug. Although the authors state that 9 was a selective ligand for the AT1R, no pharmacological data were provided for other angiotensin receptors to confirm this statement. The ability of 9 to label the AT1R was evaluated, but unfortunately, significant accumulation of fluorescence in the cell cytoplasm was observed even with non-AT1R transfected CHO cells. Development of less lipophilic fluorescent ligands, to eliminate this intracellular localization, is an ongoing work in the authors' laboratory.

Cannabinoid receptor
In a recent report by Sexton et al. (2011), two newly designed CB2 cannabinoid receptor (CB2R) fluorescent probes (10) and (11) were compared with the previously reported fluorescent antagonist NIRmbc94 (12) (Bai et al., 2008). The purpose of these new conjugates was to examine the influence of linker location around the core of the antagonist. The parent pharmacophore, CB2R selective antagonist SR144528, lacked an intrinsic biological handle such as an amine or carboxylic acid. Therefore, a 6-(aminohexyl)aminomethyl tether was incorporated in additional positions to that previously reported for 12, and then coupled to the near-infrared IRDye 800CW-NHS ester. The two new compounds did not demonstrate measurable binding to mouse delayed brain tumour cells that heterologously express the mouse CB2R; consequently, the fluorescent conjugate of choice remained the previously reported NIRmbc94 (12). This study reinforces the importance of identifying a tolerant location on the pharmacophore for linker attachment and, as anticipated, demonstrates that different linker positions can have dramatic effects on final conjugate pharmacology. NIRmbc94 (12) was subsequently used as the competing probe in a competitionbinding assay (Sexton et al., 2011), and this methodology was further elaborated by screening a small compound library to reliably identify known CB2R binders. The authors then went on to demonstrate that NIRmcb94 can identify endogenously expressed CB2R in a mouse microglia cell line, BV-2.
Instead of the more common approach of conjugating a known, discrete orthosteric ligand via a linker to a fluorophore, the fluorescent moiety can instead be designed as part of the primary 'pharmacophore' scaffold with rational receptor-ligand interactions in mind. This approach was employed by Petrov et al. (2011), who constructed isatin acylhydrazone-based antagonist 13 that demonstrated selectivity for the human CB2R (over human CB1R). A methoxyisatin derivative was linked to a 7-nitro-2,1,3-benzoxadiazole (NBD) fluorophore to afford 13, which, although displaying slightly reduced affinity to the comparable non-fluorescent compound fragment, retained the desired CB2R selectivity profile. Using fluorescent confocal microscopy, the association of 13 with T-cells could be visualized, and this interaction could be blocked using a non-fluorescent selective CB2R

Fluorescent ligands for GPCRs
antagonist. Using flow cytometric analysis, the authors also revealed that 13 is associated with the CB2R in B lymphocytes.

Histamine receptor
A high-affinity fluorescent antagonist (14) for the human histamine H1 receptor (H1R) has been reported by Rose et al.
(2012) consisting of the high-affinity and H1R-selective antagonist mepyramine linked to the BODIPY 630/650 fluorophore. Although fluorescent ligand 14 was purchased from a commercial supplier and not synthesized by the authors in this publication, it represents the first disclosure of the chemical structure of 14. Conjugate 14 displayed comparable affinity for the H1R compared with the parent pharmacoph- ore mepyramine. Confocal microscopy revealed specific and displaceable binding of 14 to the H1R localized to the cell membrane, and despite significant non-specific intracellular uptake, this probe proved very useful for studying the receptor in single living cells. The diffusion coefficient of 14-H1R complexes was quantified using FCS, and these values were different for CHO-K1 cell lines transiently expressing the H1R with and without the yellow fluorescent protein receptor tag. The authors then developed this further, by showing that FCS experiments using 14 can detect endogenously expressed H1Rs in HeLa cells. Kuder et al. (2009) have reported 15 as a selective H3R fluorescent antagonist, which consists of a piperidine-containing pharmacophore (related to the known H3R selective antagonist pitolisant) linked to a nitrobenzoxadiazole-based fluorophore. The only difference between 15 and a fluorescent probe previously developed by Amon et al. (2007) is the presence of a 3-methyl group on the piperidine ring of 15. Conjugate 15 was twofold less potent for the human H3R than the previously reported nonmethylated derivative (Amon et al., 2007), and the authors did not evaluate the selectivity or imaging properties of 15.
In another report of an H3R fluorescent ligand, but with a goal of making red-shifted probes, Tomasch et al., (2012c) have tethered selective H3R antagonists, again based on a piperidine moiety, to substituted chalcones. A series of pharmacophore-fluorophore combinations containing different linker positions and lengths were synthesized, and all exhibited nanomolar affinity for the human H3R and selectivity over the H1R (one log unit) and H4R (two log units). The authors then examined the ability of three conjugates (16-18) to visualize the H3R in hH3-HEK-293 cells. Confocal microscopy revealed enrichment of the fluorescent signal to the cell membrane, and the authors conclude that this was specific binding to the H3R as when HEK-293 cells that do not transiently express the H3R were treated with the fluorescent ligands, no enriched membrane fluorescence was observed.
From the same laboratory, Tomasch et al. (2012b) have used the same piperidine-based pharmacophore but now with the boron-dipyrromethene scaffold as the fluorophore (19). Synthesis of the fluorophore moiety was completed by reaction with boron trifluoroetherate as the final reaction in a stepwise synthesis, rather than the more common convergent approach of coupling a pre-activated (and often commercially available as the NHS ester) fluorophore to a complementary pharmacophore/linker. Conjugate 19, named Bodilisant by the authors, displayed a low nanomolar affinity for the human H3R that was 10 times more potent that the previous generation (16-18) of conjugates (Tomasch et al., 2012c). Again, H3R subtype selectivity was maintained. Fluorescence microscopy was used to visualize the human H3R in H3-HEK-293 cells and showed that 19 predominately localized to the cell membrane. In these experiments, the authors concluded that fluorescent probe 19 was not internalized, as it did not overlap with the nuclear stain DAPI. Lacivita et al. (2010) designed and synthesized a fluorescent probe (20) for the serotonin1A receptor (5-HT1AR) in an effort to improve on previous ligands from the same laboratory that showed high levels of non-specific binding. A chromenone-containing fluorophore was synthesized in-house and coupled to a 1-arylpiperizine-based antagonist, affording a conjugate with nanomolar affinity for the human 5-HT 1AR; approximately 10-fold less potent than the parent piperizine pharmacophore. The authors then evaluated the ability of 20 to visualize the 5-HT1AR, and using a high concentration of 20, showed fluorescent labelling of CHO-5-HT1A cells that was reduced by application of serotonin. Along with the lead ligand (20), a conjugate containing a near-infrared fluorophore was also synthesized, and despite a similar affinity for the 5-HT1AR compared to 20, the authors commented that due to the loss of fluorescent properties as measured in aqueous buffer, it was not useful as an imaging probe. However, given the location of the 5-HT1AR, as with all Class A GPCRs, in the cell membrane, it would be interesting to study the properties of this ligand when bound to the receptor. It can be advantageous to have a fluorescent probe that is quenched in an aqueous environment (Baker et al., 2010) but fluoresces when associated with the receptor in a lipophilic membrane environment. For example, in competitionbinding assays, this property can eliminate the need for thorough washing steps prior to analysing membrane-localized fluorescence.

5-Hydroxytryptamine (serotonin) receptor
Fluorescent 5-HT1AR probes have also been developed by Alonso et al. (2010) based on an arylpiperazine agonist previously reported from the same research group. A series of compounds were synthesized using the 7a-position of the bicyclohydantoin moiety to tether a dansyl fluorophore. Several conjugates showed an affinity for the human 5-HT1AR that were comparable to the starting arylpiperazine scaffold. Conjugate 21 was identified as the lead fluorescent ligand due to a high fluorescent intensity emission value. Fixed CHO-5-HT1AR cells could be labelled with 21 and this could be blocked using a reference non-fluorescent ligand. Despite being based on a known agonist, only radioligand competition binding assays were carried out to determine the affinity of 21 for the 5-HT1AR -no information was provided regarding how linkage to the dansyl fluorophore influenced ligand efficacy. The authors did not comment on localized membrane fluorescence or the potential of the fluorescent conjugate (if indeed an agonist) to internalize with the receptor. However, the timescale for the visualization experiments was only 10 min pre-incubation followed by a wash, fix and mount process, which may therefore have precluded this from occurring.

Muscarinic receptors
A fluorescent antagonist (22) of the muscarinic M3 receptor (M3R) has been reported by Jones et al. (2008), by linking the non-subtype selective M3R antagonist tolterodine to the commercially available fluorophore BODIPY 630/650-NHS ester. Conjugate 22 displayed a threefold loss in affinity for the human M3R compared with tolterodine, and also approximately the same fold-loss in affinity across the other human muscarinic receptor subtypes. The authors did not examine the use of 22 as an imaging tool, but indicated that this was an ongoing work in their laboratory. Hern et al. (2010) have reported 23 as an M1R probe, synthesized by the reaction of a telenzepine amino congener to an Alexa Fluor 488 fluorophore. Conjugate 23 had a nanomolar affinity for the human M1R and very slow dissociation kinetics, making it an ideal tool to visualize and monitor receptor-ligand complexes in living cells. Alongside the previously reported high-affinity M1R ligand 24 (Harris et al., 2003), total reflection fluorescence microscopy was used to track the position of fluorescent ligand-receptor complexes in live CHO cells expressing the human M1R. Information about M1R mobility, clustering and, in particular, dimerization could be obtained by simultaneously using probes 23 and 24 that have different fluorescence emission wavelengths. Daval et al. (2012) synthesized fluorescent M1R ligand 25 by coupling an agonist (based on AC-42) with high functional selectivity for the M1R to the fluorophore lissamine rhodamine B sulfonyl chloride. The emphasis in this study was on investigating how the fluorescent conjugate binds to the M1R, and teasing out possible 'non-orthosteric' binding mode(s) and receptor-ligand interactions. The parent agonist pharmacophore and conjugate 25 displayed a similar affinity towards the human M1R receptor, but interestingly, 25 could no longer elicit a typical agonist-induced calcium response in a functional assay. Instead, 25 fully reduced the functional response to a known agonist, thereby classifying 25 as an antagonist. In a series of very comprehensive and interesting experiments involving assays with reference allosteric ligands, receptor truncation, molecular modelling and even application of 25 itself as a FRET tracer, the authors concluded that binding of 25 to the M1R showed a bitopic  nature. This work further demonstrates the importance of treating the ligand-linker-fluorophore conjugate as a new chemical entity, which can have subtle or quite profound pharmacological differences compared with the starting drug molecule. These differences are often not captured in a single competition-based affinity assay, and there is a fascinating road ahead in terms of rationalizing ligand-receptor interactions of fluorescent conjugates beyond the confines of the orthosteric ligand-binding pocket.

Prostanoid receptor
From the same research laboratory as fluorescent H3R ligands 16-18 (Tomasch et al., 2012b;2012c) was the report of a fluorescent prostanoid EP3 receptor (EP3R) antagonist (26) (Tomasch et al., 2012a). Based on an ortho-substituted cinnamic acid antagonist, a series of fluorescent conjugates were synthesized containing different fluorophores, with pyrylium-containing 26 showing the most promise as an imaging tool. Although with approximately threefold reduced affinity for the human EP3R compared with the parent drug molecule, 26 maintained selectivity over the EP1R, EP2, and EP4R subtypes. The authors then proceeded to demonstrate that the EP3R receptor could be visualized using 26 in murine kidney and human brain tissue. Loison et al. (2012) reported the first examples of selective, fluorescent, non-peptidic ligands for the vasopressin V2 receptor (V2R) based on a tetrahydro-1H-benzo[b]azepine antagonist. A series of conjugates containing different linker lengths and fluorophores were synthesized, and from these, three lead compounds (27-29) were identified. Interestingly, although compounds 27-29 displayed a slight decrease in affinity for the human V 2R compared with the parent ligand, these fluorescent conjugates were comparatively more selective for the V2R over the V1AR and oxytocin receptor. Cyanine probe 28 and terbium-containing 29 (the nature of the linker housed within the fluorophore is not represented, it has been presumed this is the commercially available Lumi4-Tb-NHS ester as stated in the publication) were then used to develop an acceptor/donor V2R TR-FRET-based assay and utilized as a tool to study V2R-V1AR dimerization in association with a previously reported V1AR probe (Albizu et al., 2010).

Other fluorescent Class A GPCR ligands
In addition to fluorescent probes specifically designed to target Class A GPCRs there have been recent reports of fluorescent ligands for alternative receptors, but which one would predict might also bind to Class A GPCRs. There are two recent reports (Céspedes-Guirao et al., 2011;Jose et al., 2011) outlining the synthesis and application of fluorescent oestradiol-based probes (30) and (31) to target the nuclear hormone oestrogen receptor (ER). Although the ER is not classified as a GPCR, the other type of oestrogen receptor, GPER (or GPR30), is classified as a Class A GPCR. Oestradiol is a high-affinity ligand for both the ER and the GPER; therefore, oestradiol-based conjugates developed for the ER may well be useful tools for studying GPER. In addition to the GPCR families covered in this review, there continues to be an interest in developing small-molecule fluorescent probes for the dopamine receptor, another Class A GPCR, although there have been no new reports published of novel ligands since the preceding review articles (Kuder and Kieć-Kononowicz, 2008;Böhme and Beck-Sickinger, 2009). Another Class A GPCR for which developing fluorescent ligands is an exciting prospect is the opioid receptor, but since the last GPCR fluorescent ligand reviews, there have only been reports of peptide, or peptide-based ligands conjugated to a fluorophore (e.g. Josan et al., 2009), and is therefore outside the scope of this review article. In addition to the novel fluorescent conjugates reported via peer-reviewed publications that have been discussed in this review, there are a growing number of fluorescent conjugates available commercially, with and without the chemical structure disclosed. In an excellent example that showcased the power of using several commercially available fluorescent probes that emit at different wavelengths, Daly et al. (2010) investigated the distribution of adrenoceptors and 'cannabinoid-like' receptors in different cell types.

Conclusion
While fluorescent ligands boast over a 30-year pedigree in their application for the study of cellular receptors, it remains clear from all reported research activity within the past decade (including the more recent developments highlighted in this review) that their usage is indeed escalating. This has been accompanied by advances in microscopy and live cell imaging, which have made it possible to visualize and measure the receptor life cycle. One such receptor family class that has benefitted immensely from fluorescent ligand development is GPCRs. The fluorescent conjugates described within this review have helped to crystallize many of the key factors for consideration when appending a fluorophore to a relatively small orthosteric drug molecule. The linker that connects the fluorescent moiety to the drug of interest must be attached to a position that is relatively insensitive to structural modification and that can potentially tolerate bulky substituents. In order to attenuate non-specific binding of the fluorescent-drug conjugate, one must also be mindful of the final physicochemical properties of the molecule. Finally, when using the generated probes for cellular-based imaging studies or certain types of assay, it has become clearer that judicious choice of the correct fluorophore is also of paramount importance. The work summarized in this and previous reviews unequivocally substantiates that if all these considerations are taken into account, the resulting fluorescent conjugates are extremely powerful pharmacological tools that can be utilized in numerous cutting-edge assay formats. An interesting road lies ahead for all researchers in this area in attempting to rationalize the ligand-receptor interactions of fluorescent conjugates beyond the confines of the orthosteric ligand-binding pocket. In so-doing though, this will undoubtedly help cement the use of these fluorescent reagents as versatile and extremely useful tools for modern-day molecular pharmacology and drug discovery.