Probing the Evolutionary History of Human Bitter Taste Receptor Pseudogenes by Restoring Their Function

Abstract

Lineage-specific gene losses can be driven by selection or environmental adaptations. However, a lack of studies on the original function of species-specific pseudogenes leaves a gap in our understanding of their role in evolutionary histories. Pseudogenes are of particular relevance for taste perception genes, which encode for receptors that confer the ability to both identify nutritionally valuable substances and avoid potentially harmful substances. To explore the role of bitter taste pseudogenization events in human origins, we restored the open reading frames of the three human-specific pseudogenes and synthesized the reconstructed functional hTAS2R2, hTAS2R62 and hTAS2R64 receptors. We have identified ligands that differentially activate the human and chimpanzee forms of these receptors and several other human functional TAS2Rs. We show that these receptors are narrowly tuned, suggesting that bitter-taste sensitivities evolved independently in different species, and that these pseudogenization events occurred because of functional redundancy. The restoration of function of lineage-specific pseudogenes can aid in the reconstruction of their evolutionary history, and in understanding the forces that led to their pseudogenization.

Introduction

Toxic substances often taste bitter and the perception of this taste is thought to have evolved to protect animals from ingesting potentially harmful compounds (Garcia and Hankins 1975). In addition, variation in bitter taste perception plays an important role in dietary habits (Drewnowski 2001; Li and Zhang 2014). This has led to the hypothesis that genes involved in bitter taste perception may have undergone strong and species-specific evolutionary constraints (Shi et al. 2003; Dong et al. 2009). Bitter taste perception is mediated by a family of G protein-coupled receptors (GPCRs) named TAS2Rs (Adler et al. 2000; Chandrashekar et al. 2000). The genes encoding these proteins have evolutionarily ancient origins, are relatively small (∼1000 base pairs), do not contain introns in their coding regions, and vary greatly in number across different species (Shi and Zhang 2006; Li and Zhang 2014). In humans, about 25 functional TAS2R genes have been identified (Adler et al. 2000). These genes can show evidence of selective pressures, either in ancient or more recent evolution (Wooding et al. 2004; Campbell et al. 2014; Risso et al. 2016), and some display copy number variations (Wooding et al. 2012; Roudnitzky et al. 2016). In addition, 11 TAS2R pseudogenes have been identified in the human genome (Go et al. 2005). Three of these pseudogenes (hTAS2R2, hTAS2R62 and hTAS2R64) are confined to the human lineage, with two fixed nonsense mutations in hTAS2R62 (CAG->TAG and GAA->TAA at codon positions 235 and 292, respectively) and one fixed nonsense mutation in hTAS2R64 (TGG->TGA at codon position 280) (Go et al. 2005). In contrast, hTAS2R2 has been inactivated because of a polymorphic two-base deletion at codon position 160, with an intact form of this gene still found in some human populations (Go et al. 2005), despite its original annotation as a fixed pseudogene (Conte et al. 2002; Shi et al. 2003). Interestingly, it was found that TAS2R64 also became a pseudogene in the orangutan, although because of a different nonsense mutation (TGG->TGA at codon position 49). Whole-gene deletion variations of this gene were observed in eastern chimpanzees, bonobos and gorillas suggesting that these events happened independently in different species as a product of convergent evolution (Hayakawa et al. 2012, 2014).

Several hypotheses have been put forward to explain this and other lineage-specific losses of function, including different responses to environmental conditions such as food preference, in terms of ligand selectivity and different binding modes, or by genetic events driven by natural selection or demography (Go et al. 2005; Bachmanov et al. 2014; Risso et al. 2014). However, a lack of studies on the original function of species-specific pseudogenes leaves a gap in our understanding of the evolutionary history of these genes, with questions such as the reason these genes became pseudogenes and the nature of the cognate ligands of the receptors encoded by these genes remaining unanswered.

To address these questions, we have restored the open reading frames in these three human-specific pseudogenes and synthesized the reconstructed functional hTAS2R2, hTAS2R62 and hTAS2R64 bitter taste receptors. We performed functional expression studies and identified agonists that differentially activate human and chimpanzee TAS2R2 and TAS2R64. We also performed population genetic analyses in ancient and modern hominins to look for evidence of selection that could account for the frequencies of different alleles of these genes in different populations.

Results

Interspecific Variation of Human TAS2R Pseudogenes

The modified human TAS2R pseudogenes and the corresponding functional chimpanzee TAS2R genes carried a high degree of sequence identity [hTAS2R2 (97.69%), hTAS2R62 (97.76%) and hTAS2R64 (98.38%)], with few interspecific amino acid changes. The alignment of -h and -c TAS2R2, 62 and 64 with structurally well-characterized TAS2Rs (hTAS2R10, hTAS2R16, hTAS2R38 and hTAS2R46) and by using the TOPCONS and I-TASSER secondary structure prediction programs revealed that some of these human–chimpanzee amino acid differences are within an extracellular domain or in a transmembrane domain oriented towards the extracellular side and reside in regions expected (Soranzo et al. 2005; Biarnés et al. 2010; Brockhoff et al. 2010; Sakurai et al. 2010; Behrens et al. 2013; Born et al. 2013; Marchiori et al. 2013) to change agonist selectivity and binding (fig. 1 and supplementary table S1, Supplementary Material online).

Fig. 1.

Multiple sequence alignment of human (h) and chimpanzee (c) TAS2R2, TAS2R62 and TAS2R64 with structurally well-known human TAS2R10, TAS2R16, TAS2R38 and TAS2R46. Different residues between human and chimpanzee same TAS2Rs are in red, positions in structurally well-known TAS2Rs affecting agonist selectivity and ligand-binding pockets are in black. NH₂, amino-terminus; TM, transmembrane domain; ic, intra-cellular loop; ec, extra-cellular loop; COOH, carboxyl-terminus.

Population and Evolutionary Genetics

Analysis of sequence data of human-specific TAS2R pseudogenes identified ten single nucleotide polymorphisms with low to moderate minor allele frequencies (MAFs), summarized in table 1. Specifically, the MAFs of the six SNPs found in hTAS2R2 (rs75043117, rs847923, rs6963925, rs12535592, rs10242038 and rs73058722) ranged from 0.01 to 0.41 and displayed low levels of worldwide genetic differentiation, with F_ST values of 0.024, 0.055, 0.061, 0.031, 0.075 and 0.031 respectively. The derived alleles of some of these variations were also present in the Neanderthal (rs12535592 G and rs73058722 G) and Denisovan (rs6963925 T) hominins. According to a previous study (Go et al. 2005), a polymorphic deletion (TA/–, rs10649477) which disrupted the hTAS2R2 ORF, was also observed in some (79%) individuals. This deletion showed a different geographical distribution among worldwide populations, with Africans carrying higher frequencies of the intact form of the gene (0.46) than either Europeans (0.25) or Asians (0.14) (F_ST= 0.09). The Denisovan hominin carried two functional copies (TA/TA), while the Neanderthal was homozygote for the pseudogenized version (–/–). hTAS2R62 showed similar patterns of genetic differentiation, with four SNPs (rs4726624, rs34039200, rs10239143 and rs6944279) showing intermediate MAFs (0.27, 0.21, 0.24 and 0.32 respectively) and low F_ST values (0.036, 0.070, 0.037 and 0.049, respectively). The archaic hominins carried the ancestral alleles at these SNP sites. No SNPs or INDELs were identified in the hTAS2R64 ORF in either the analyzed subjects or in the 1000 Genomes individuals.

Table 1.

Summary Statistics of the Identified Variations in hTAS2R2 and hTAS2R62 Pseudogenes.

Gene	Variant	MAF-Current Study	MAF-All	MAF-Africa	MAF-Europe	MAF-Asia	FST	Tajima’s D-Africans	Tajima’s D-Non-Africans
hTAS2R2	rs75043117	0.01	0.01	0.03	0	0	0.02	−0.38	0
	rs847923	0.41	0.41	0.35	0.3	0.42	0.05	1.91*	1.87*
	rs6963925	0.24	0.28	0.47	0.27	0.16	0.06	2.27*	1.45*
	rs12535592	0.13	0.17	0.15	0.25	0.08	0.03	1.02	0.97
	rs10242038	0.02	0.03	0.11	0	0	0.07	0.38	−0.79
	rs73058722	0.18	0.17	0.15	0.25	0.08	0.03	1.02	1
hTAS2R62	rs4726624	0.27	0.22	0.34	0.13	0.2	0.04	2.02*	1.01
	rs34039200	0.21	0.13	0.07	0.23	0.02	0.07	−0.06	1.26
	rs10239143	0.24	0.22	0.34	0.13	0.2	0.04	2.03*	1.02
	rs6944279	0.32	0.27	0.35	0.35	0.12	0.05	1.89*	1.93*

^*Significance.

Tests for the effects of natural selection revealed significant evidence of deviation from neutrality. Both hTAS2R2 and hTAS2R62 showed positive Tajima’s D values in both African (2.77 and 2.67, respectively) and non-African populations (2.39 and 2.16, respectively; all P’s <0.05). In addition, the HKA test produced a significant P value (<0.01) for human TAS2R2 sequences when compared with the chimpanzee outgroup. The Li’s MFDM test confirmed these findings, producing significant P values (all P’s <0.05) for both genes, suggesting that the observed Tajima’s D values were not impacted by demography. A sliding-window analysis better characterized the observed patterns of deviation from neutrality in the studied single nucleotide polymorphisms: two hTAS2R2 SNPs (rs847923 and rs6963925) showed positive and significant (all P’s <0.05) D_T values, with the remaining SNPs showing nonsignificant deviations from neutrality in both Africans and non-Africans. Similar results were obtained for hTAS2R62, which showed positive D_T values for most of the tested SNPs although only one (rs6944279) reached significance in both African and non-African populations. However, these SNPs were found to be in strong linkage disequilibrium (all D’>0.98) with a coding SNP (rs4595035) in the neighboring functional TAS2R60 gene which showed positive and significant DT values (2.3, P < 0.01) in all populations.

Identification of Agonists for Orphan TAS2R2 and TAS2R64

The extent of receptor expression and cell surface localization of -h (fig. 2) and -c (supplementary fig. S1, Supplementary Material online) TAS2Rs was visualized in the transfected cells by an anti-HSV glycoprotein D antibody (HSV, green), which recognizes the specific HSV glycoprotein D epitope and does not interfere with the signaling function, over the plasma membrane (ConA, red) and cell nuclei (DAPI, blue). All receptors were expressed, although with some variation in expression levels or localization. Both human and chimpanzee TAS2R62 and TAS2R64 showed comparable expression rates (8–20% of cells, fig. 2C;supplementary table S2 and fig. S1C, Supplementary Material online). TAS2R62 appeared to be more expressed and localized at the intra-cellular level, rather than at the plasma membrane (fig. 2B and supplementary fig. S1B, inserts, Supplementary Material online). Although both -h and -c TAS2R2 showed a reduced expression efficiency (3–15% of cells, fig. 2A;supplementary table S2 and fig. S1A, Supplementary Material online), this was comparable to transfection rates of previous reports of other human and chimpanzee TAS2Rs (Bufe et al. 2002; Wooding et al. 2006).

Fig. 2.

Confocal fluorescence images of HEK293T cells transfected with human TAS2R2 (A), TAS2R62 (B) and TAS2R64 (C) receptors and negative control transfected cells (D). Receptor expression is visualized by a mouse anti-HSV glycoprotein D antibody combined with an Alexa488-labeled anti-mouse antibody (HSV-tag, green). Cell surface is visualized by plasma membrane glycoproteins detected with biotin-conjugated concanavalin A and avidin-conjugated Texas Red (ConA, red). Cell nuclei are visualized with 4′,6-diamidino-2-phenylindole (DAPI, blue).

About 4 agonists out of 83 tested (5%) were identified for TAS2R2 (fig. 3). One of these compounds (thiamine) activated only the human form of this receptor, while another one (umbelliferone) was recognized by only the chimpanzee form. Interestingly, umbelliferone did not activate any other human or primate TAS2Rs in previous studies but did activate two mouse bitter taste receptors (Tas2R105 and Tas2R125) (Lossow et al. 2016). TAS2R64 showed a similar activation pattern: 4 functional ligands out of 83 tested (5%) were found for this receptor, with two (amarogentin and colchicine) activating only the human form of this receptor (fig. 3), with the others (andrographolide and picrotoxinin) activating both human and chimpanzee TAS2R64. We failed to identify any agonists for either human or chimpanzee TAS2R62. To better determine the sensitivities of both human and chimpanzee TAS2R2 and TAS2R64, we tested a 3-fold lower concentration of the identified agonists. As shown in figure 4 and supplementary fig. S2, Supplementary Material online, most of the ligands (chlorhexidine, thiamine and umbelliferone for TAS2R2 and amarogentin, andrographolide and picrotoxinin for TAS2R64) activated these receptors with reduced potency, as expected. However, two of them (chloroquine for TAS2R2 and colchicine for TAS2R64) showed similar relative fluorescence changes (ΔF/F) with respect to the mock transfected negative control at both tested dilutions, suggesting a potential saturation of the signal. We therefore obtained dose–response relationships for these compounds (supplementary fig. S3, Supplementary Material online), and found that chloroquine showed similar sensitivity levels between -h and -c TAS2R64, and confirming that colchicine activated only human TAS2R2 but without reaching saturation of the signal.

Fig. 3.

Identification of agonists for human (black solid line) and chimpanzee (grey solid line) TAS2R2 and TAS2R64. Responses of the negative control transfected cells are shown with dotted lines.

Fig. 4.

Functional screening of human (black bars) and chimpanzee (grey bars) TAS2R2 and TAS2R64 bitter taste receptors. The y axis shows the relative fluorescence changes (ΔF/F), and the x axis is labeled with the activating bitter compounds with relative concentrations. Asterisks indicate a significance difference from control.

Discussion

Loss of gene function is a widespread evolutionary phenomenon that is traditionally considered to be disruptive or selectively disadvantageous. However, Olson’s “less-is-more” hypothesis (Olson 1999) suggested that pseudogenization could serve as an engine of evolutionary change. In human evolution, numerous pseudogenization events have led to human-specific phenotypes and have been linked to adaption to different environmental challenges (Chou et al. 1998; Stedman et al. 2004; Go et al. 2005; Dong et al. 2009). Among these, it has been shown that taste perception pseudogenes are overrepresented or unique to humans (Wang et al. 2004, 2006; Go et al. 2005), either with their functional versions still existing (TAS2R2 and TAS2R62), or being polymorphic or pseudogenized because of independent mutations (TAS2R64) in other primates (Conte et al. 2002; Shi et al. 2003; Go et al. 2005; Hayakawa et al. 2012, 2014). Although species-specific gene losses have been related to specific diet and environmental changes (Li et al. 2005; Wang et al. 2006; Zhao et al. 2010; Jiang et al. 2012), underlining the evolutionary importance of these pseudogenes, a lack of functional studies on lineage-specific pseudogenes has left unclear why these genes lost their function, and what the cognate ligands of these receptors were.

To address these questions, we restored the open reading frame of human-specific TAS2R2, TAS2R62 and TAS2R64 pseudogenes and expressed them in vitro, characterizing the function of these restored human receptors. In addition, we compared them to their homologous functional chimpanzee receptors. We were able to identify ligands for both the restored human and existing chimpanzee TAS2R2 and TAS2R64. Interestingly, some of these compounds differentially activated the human and chimpanzee receptors: thiamine (vitamin B1), which only activated hTAS2R2, is not synthesized in animals which need to obtain it from natural dietary sources (Lonsdale 2006). In contrast, umbelliferone only activated cTAS2R2 and is naturally found in many plants of the order Umbelliferae (Venugopala et al. 2013). Similarly, two compounds activated only the human TAS2R64 and are naturally found in gentian, Swertia chirata (amarogentin) or extracted from plants of the genus Colchicum (colchicine) (Ray et al. 1996; Klintschar et al. 1999).

These findings support the hypothesis that bitter-taste sensitivities and TAS2R ligand recognition evolved independently in different species by replacing specific amino acid residues, perhaps allowing the adaptation to different environments and the identification of compounds of species-specific relevance (Imai et al. 2012; Behrens et al. 2014; Li and Zhang 2014). The fact that hTAS2R2 and hTASR64 only recognized a small number (∼5%) of the tested bitter compounds, suggests that these pseudogenes were once narrowly tuned receptors. The same hypothesis could apply to hTAS2R62, for which we were not able to find any ligands. However, hTAS2R62 was barely expressed at the cell surface with most of the expression found at the intracellular level and our failure to identify its cognate bitter compound(s) may be due to its failure to achieve proper localization in transfected cells. This is consistent with previous results, in which the only five TAS2R receptors that were not de-orphanized in a first large attempt (e.g., TAS2R41, TAS2R42, TAS2R45, TAS2R48 and TAS2R60) were barely or not expressed at the plasma membrane (Bufe et al. 2002; Meyerhof et al. 2010). However, a more recent examination of all the naturally occurring major variants of these receptors tested against a large number of bitter compounds led to finding of just one TAS2R41 agonist, leaving TAS2R42, TAS2R45, TAS2R48 and TAS2R60 currently remaining as orphans. This suggests the possibility that high selectivity could explain the failure to find agonists for these receptors, rather than low expression levels in transfected cells (Thalmann et al. 2013). The compounds activating the restored hTAS2R2 and hTAS2R64 pseudogenes also activated at least one existing functional hTAS2R, most of which are broadly tuned: chlorhexidine activates hTAS2R14, chloroquine hTAS2R3, hTAS2R7, hTAS2R10 and hTAS2R39, thiamine hTAS2R1 and hTAS2R39, amarogentin hTAS2R1, hTAS2R4, hTAS2R39, hTAS2R43, hTAS2R46, hTAS2R47 and hTAS2R50, andrographolide hTAS2R46, hTAS2R47 and hTAS2R50, colchicine hTAS2R4, hTAS2R39 and hTAS2R46 and picrotoxinin activates hTAS2R1, hTAS2R10, hTAS2R14, hTAS2R46 and hTAS2R47 (Meyerhof et al. 2010). Therefore, we hypothesize that hTAS2R2 and hTAS2R64 have undergone pseudogenization in humans because of their functional redundancy.

The finding that human TAS2R2 showed positive Tajima’s D values, significant P values of the HKA test and low F_ST values, suggests that this gene has undergone ancient balancing selection before its pseudogenization event occurred, possibly resulting in an enhanced sensitivity to agonists in common with other functional TAS2Rs. In addition, our data have better defined the polymorphic pseudogenization event of this gene, showing that it was still functional in the Denisovan but not in the Neanderthal hominin. This suggests that it was already polymorphic before the emergence of Homo sapiens, at the point of the Neanderthal and Denisovan split of 550–765 ky ago (Prüfer et al. 2014), or that the inactivation of this gene is the product of a convergent evolution in Homo sapiens and Neanderthals.

The positive Tajima’s D values observed at hTAS2R62 are likely due to genetic hitchhiking with a variant in the functional TAS2R60, which showed evidence of balancing selection in the examined populations. Such a phenomenon has been previously documented in other TAS2R pseudogenes (Risso et al. 2014).

We conclude that restoration of the function of TAS2R human-specific pseudogenes can aid in the reconstruction of their evolutionary history, and in understanding the forces that led to the pseudogenization of some paralogs. This method could be applied to other pseudogenes and shed light on the evolutionary mechanisms that have acted specifically on different species.

Materials and Methods

Subjects

About 168 unrelated individuals were enrolled with written informed consent under protocol 01-DC-0230 approved by the National Institutes of Health Combined Neurosciences/Blue Panel Institutional Review Board. Most of the subjects (N = 113) were Caucasians, with the remaining being of Asian (N = 36) or African (N = 19) ancestry. Saliva samples were collected and genomic DNA was extracted as previously described (Risso et al. 2015). Oligonucleotide primers were designed to completely sequence hTAS2R2, hTAS2R62 and hTAS2R64 (supplementary table S3, Supplementary Material online) using dideoxy Sanger sequencing (Sanger et al. 1977). DNA chromatograms were analyzed with the Lasergene suite (DNASTAR, Madison, WI).

Molecular Biology

Human TAS2R2, TAS2R62 and TAS2R64 open reading frames were modified in vitro to restore their functionality using the chimpanzee (c) functional versions of these genes as a template and substituting the human stop codons with the respective chimpanzee amino acids. In particular, “AT” was inserted at hTAS2R2 positions 447-448, “TAG” was replaced with “CAG” and “TAA” with “GAA” at hTAS2R62 amino acid positions 235/292 and “TGA” was modified to “TGG” at hTAS2R64 amino acid position 283. Both the modified human and the chimpanzee TAS2R2, 62 and 64 were then synthetized using the gBlocks^® Gene Fragments technology (IDT, Coralville, IA) and cloned into an expression vector which contained the Kozak sequence (Kozak 1987) and the first 45 amino acids of the rat somatostatin receptor subtype 3 for cell surface localization at the amino-terminus (Ammon et al. 2001) and the herpes simplex virus (HSV) glycoprotein D epitope for immunocytochemical detection at the carboxy-terminus (Roosterman et al. 1997), as previously described (Bufe et al. 2002; Meyerhof et al. 2010). DNA sequences of five clones were compared with identify and exclude the ones containing mutations generated by PCR. The amino acid sequences of the synthetized human and chimpanzee TAS2Rs are shown in supplementary figure S4, Supplementary Material online.

Population and Phylogenetic Analysis

The calculation of minor allele frequencies (MAFs) of modern human TAS2R2, TAS2R62 and TAS2R64 pseudogenes variants was based on sequencing data obtained in this study and compared with the 1000 Genomes Project Phase 3 samples (1000 Genomes Project Consortium et al. 2015), the Altai Neanderthal (Prüfer et al. 2014) and Denisovan (Reich et al. 2010) hominins using VCFTools (Danecek et al. 2011). Linkage disequilibrium (LD) was calculated with PLINK v.1.07 (Purcell et al. 2007) in a 50-kb window surrounding the analyzed pseudogenes of samples of the 1000 Genomes Project. Signatures of natural selection were tested using DNASP v.5 (Librado and Rozas 2009), DNA slider (McDonald JH 1996) and Arlequin v.3.5 (Excoffier and Lischer 2010) in the 1000 Genomes samples using four indicators of selective pressure: Tajima’s D (DT) (Tajima 1989), FST (Wright 1949), the maximum frequency of derived mutations test (MFDM) (Li 2011) and the HKA test (Hudson et al. 1987). Human and chimpanzee TAS2R sequences were retrieved from the Ensembl database version 86 (Aken et al. 2016) and aligned with structurally well-characterized TAS2R functional genes (Soranzo et al. 2005; Biarnés et al. 2010; Brockhoff et al. 2010; Sakurai et al. 2010; Behrens et al. 2013; Born et al. 2013; Marchiori et al. 2013) with MAFFT v.7.304 (Katoh et al. 2002). Additional evidence for the secondary structure of these TAS2Rs was obtained using TOPCONS (http://topcons.cbr.su.se/ (last accessed March 1, 2017), Hennerdal and Elofsson 2011) and I-TASSER (http://zhanglab.ccmb.med.umich.edu/I-TASSER/ (last accessed March 1, 2017), Zhang 2008) servers.

Functional Studies

About 150 ng of TAS2R plasmid DNA and 300 nL Lipofectamine 2000 (Invitrogen) per well were used to transfect HEK293T cells, which stably express the chimeric G protein subunit Gα16gust44 24 h after seeding them in 96-well plates (Ueda et al. 2003). At 24-h post-transfection, cells were loaded with the calcium-sensitive dye Fluo-4-AM (2 μM, Molecular Probes) and probenecid (2.5 mM, Sigma-Aldrich), to minimize the loss of the calcium indicator dye. Cells were then stimulated with a total of 83 different natural and synthetic bitter compounds (Sigma-Aldrich; LGC Promochem, supplementary table S4, Supplementary Material online), chosen to represent diverse chemical classes and predominantly natural substances, dissolved in either buffer C1 (130 mM NaCl, 5 mM KCl, 10 mM N-(2 hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid), 2 mM CaCl₂, 10 mM glucose [pH 7.4]) or in buffer C1 plus 0.1% dimethyl sulfoxide (DMSO). The highest compound concentrations used for the initial screening were the highest concentrations not leading to artificial calcium signals (Meyerhof et al. 2010). In addition, 3-fold lower concentrations were applied. Calcium signals were recorded at 510 nm following excitation at 488 nm by a fluorometric imaging plate reader (FLIPR^Tetra, Molecular Devices). As positive control, TAS2R16-transfected cells were challenged with 30 mM salicin (Sigma-Aldrich), a well-characterized ligand for this receptor (Bufe et al. 2002). Mean calcium release signals from mock-transfected cells were subtracted from the recorded signals. In addition, signals were normalized by subtraction of the background fluorescence for each well and baseline noise (ΔF/F = (F−F₀)/F₀). All data were collected and averaged from at least two independent transfection experiments, each performed in duplicate. Compounds resulting in statistically significant (P < 0.05; t-test) higher changes in fluorescence in receptor transfected cells compared with cells transfected with empty vector (controls) were considered to be agonists. For selected compound–receptor combinations, dose–response relationships were tested as described before (Meyerhof et al. 2010).

Immunocytochemical Experiments

Cells were stained with a primary monoclonal mouse anti-HSV glycoprotein D epitope antibody (1 : 15,000, Novagen) and a secondary Alexa Fluor^® 488-labeled goat anti-mouse immunoglobulin G (1:2,000, Thermo Fisher Scientific) to estimate the level of receptor expression and analyze cell surface location. Cells were also stained with biotinylated concanavalin A (ConA) (0.5 μg/ml; Sigma–Aldrich), Alexa Fluor^® 633 streptavidin (1:1,000, Thermo Fisher Scientific) and 4′,6-diamidino-2-phenylindole (DAPI) (1:500, Thermo Fisher) to visualize cell surface and nuclei, respectively. Cells were examined using a confocal laser-scanning microscope TCS-SP8 (Leica), images were processed with ImageJ (https://imagej.nih.gov/ij/ (last accessed March 1, 2017), Schneider et al. 2012) and the proportion of cells expressing recombinant receptors was calculated as the number of green fluorescent cells divided by the total cell number, averaged over a total of three microscopic fields.

Supplementary Material

Supplementary data are available at Molecular Biology and Evolution online.