Analysis of Human TAAR8 and Murine Taar8b Mediated Signaling Pathways and Expression Profile

The thyroid hormone derivative 3-iodothyronamine (3-T1AM) exerts metabolic effects in vivo that contradict known effects of thyroid hormones. 3-T1AM acts as a trace amine-associated receptor 1 (TAAR1) agonist and activates Gs signaling in vitro. Interestingly, 3-T1AM-meditated in vivo effects persist in Taar1 knockout-mice indicating that further targets of 3-T1AM might exist. Here, we investigated another member of the TAAR family, the only scarcely studied mouse and human trace-amine-associated receptor 8 (Taar8b, TAAR8). By RT-qPCR and locked-nucleic-acid (LNA) in situ hybridization, Taar8b expression in different mouse tissues was analyzed. Functionally, we characterized TAAR8 and Taar8b with regard to cell surface expression and signaling via different G-protein-mediated pathways. Cell surface expression was verified by ELISA, and cAMP accumulation was quantified by AlphaScreen for detection of Gs and/or Gi/o signaling. Activation of G-proteins Gq/11 and G12/13 was analyzed by reporter gene assays. Expression analyses revealed at most marginal Taar8b expression and no gender differences for almost all analyzed tissues. In heart, LNA-in situ hybridization demonstrated the absence of Taar8b expression. We could not identify 3-T1AM as a ligand for TAAR8 and Taar8b, but both receptors were characterized by a basal Gi/o signaling activity, a so far unknown signaling pathway for TAARs.

The thyroid hormone derivative 3-iodothyronamine (3-T1AM) was described in 2004 as a potent agonist of TAAR1, stimulating cAMP production in in vitro experiments [9]. Moreover, in vivo administration of 3-T1AM in mice induced hypothermia and bradycardia [9]. On isolated rat hearts, application of 3-T1AM resulted in reduced heart rates [9,10]. Collectively, these results highlighted possible cardio-protective effects and suggested that 3-T1AM might also protect from major brain damage upon stroke [22,23]. However, Taar1-deficient mice (Taar1 −/− ) showed no significant differences regarding their general health status, and the hypothermic effect of 3-T1AM was even reproduced in Taar1 −/− mice, pointing to a different mediator of 3-T1AM action in vivo [24,25].
Given the reported effects of 3-T1AM in wild type and Taar1 −/− mice for which the mediator remains unknown (see above), we speculated and tested TAAR8 orthologues as putative targets for 3-T1AM action. To date, the known signaling pathway for TAARs is mediated via Gs activation [6,7,26], but the reported effects of 3-T1AM on the heart should be comprehensively explained by Gi/o coupling. We therefore aimed to find a TAAR subtype which potentially activates Gi/o and not Gs. As the in vivo effects of 3-T1AM have primarily been studied in mice [9,10,25], we focused on the poorly characterized human TAAR8 and its murine orthologue Taar8b. We investigated its expression pattern in different tissues and performed a thorough in vitro analysis of the signal transduction properties of Taar8b in comparison with the human TAAR8.

Results
Here we focus on human and mouse trace amine-associated receptor 8 orthologues as possible targets of 3-T1AM action. We analyzed putatively involved G protein-dependent signaling pathways, i.e., Gs, Gi/o, Gq/11, and G12/13, as triggered by signaling via the murine and human orthologues at basal conditions and following stimulation with known agonists of TAAR1. Additionally, we investigated the expression of Taar8b in various mouse tissues.

The Murine Taar8b Is at Most Marginally Expressed in Several Mouse Tissues
In a previous study performed by Liberles and co-workers, marginal levels of Taar8b expression in the olfactory epithelium were detected [7]. We used RT-qPCR to analyze further tissue expression of murine Taar8b. For this purpose, we first verified the absence of contamination of the RNA preparations with residual DNA traces for all samples. While the expression levels of the reference gene Gapdh as well as the positive controls Thra and Thrb were comparable between the individual animals and detected to be rather high for Gapdh (Ct-value average 18.75) and lower for Thra and Thrb (Ct-value averages 23.68 and 26.38), respectively, Taar8b expression was observed to be considerably lower (Ct-values ranging from 41.21 to 49.14, average 42.95). Clear differences were not generally detected between male and female animals. Low expression is commonly known to be reflected in a higher variation of the Ct-values between biological and technical RT-qPCR replicates. This effect was also evident in our sample set ( Figure 1 and Table S1). We measured detectable Ct-values (<50) for one, two, or three of the technical replicates, but also observed no detectable Ct-values in all technical replicates per individual tissue sample in our qPCR setting which was comprised of 50 cycles of amplification (Table S1). Ct-values less than 50 in at least one technical replicate were only observed for 24 of 101 individual tissue samples analyzed (Figure 1), thereby missing our standard criteria for validated expression (at least two comparable technical replicates in at least three biological replicates). The expression level of Taar8b was at most marginal and close to the detection limit. Taar8b transcripts were not detected in thyroid tissue via qPCR (50 cycles).

Figure 1.
Normalized expression levels of the positive control genes Thra and Thrb and the target gene Taar8b. Normalization was performed by ΔCt-value calculation, i.e., by subtracting the Ct-value of the calibrator gene Gapdh from the Ct-value of the positive control or target gene (ΔCt,(Thra) = Ct(Thra) -Ct(Gapdh); ΔCt(Thrb) = Ct(Thrb) -Ct(Gapdh); ΔCt,Taar8b = Ct(Taar8b) -Ct(Gapdh)). The ΔCt-values of samples which yielded "undetermined" Ct-values in PCR runs of 50 cycles for all technical replicates were artificially set to 40 and, thus, are visible above the group of samples with detectable Ct-values (Ct < 50). In all other cases, ΔCt-values were calculated from mean Ct-values of the technical replicates for which a Ct-value was measured. If only one technical replicate yielded a Ct-value smaller than 50, this Ct-value was used for ΔCt-calculation. Positive control genes Thra and Thrb were not included in the analysis of three thyroid samples because of the limited amount of RNA. Graphs are depicted for the positive control genes Thra (A) and Thrb (B) and for the target gene Taar8b (C). Samples from male and female mice for each tissue type are collectively grouped. CB, cerebellum; CX, cortex; HIP, hippocampus; STR, striatum; H, heart; LIV, liver; KID, kidney; THY, thyroid; OB, olfactory bulb; Ct, threshold cycle; Gapdh, glyceraldehyde-3-phosphate dehydrogenase; Thra, thyroid hormone receptor alpha; Thrb, thyroid hormone receptor beta; Taar8b, trace amine-associated receptor 8B.
As it is known that TAARs, in general, are expressed at low levels except for in the olfactory epithelium [7,26], we applied the very sensitive method "locked nucleic acid" (LNA) in situ hybridization (ISH) for the example of heart tissue. In this particular case, we were able to demonstrate the absence of Taar8b mRNA for this tissue (Figure 2).

Figure 2.
Expression studies of Taar8b in murine heart using in situ hybridization. Expression studies were performed by in situ hybridization (ISH) using a LNA (locked nucleic acid) probe. C57BL/6 WT mouse hearts were sectioned and treated with the LNA probe. Signals were visualized with an avidin-biotin complex using DAB (3,3'-diaminobenzidin) (A-C) or DY-light 488 streptavidin (D) staining. (A) The mGata probe used as a positive control caused significant staining of the nucleus and the perinuclear region of cardiomyocytes, as expected; (B) The negative control performed by hybridization with scrambled ISH probe demonstrated homogenous shading; (C) Treatment with Taar8b specific probes did not result in staining of the heart tissue, pointing to a lack of expression of Taar8b in the murine heart; (D) COS-7 cells transiently transfected with Taar8b served as a positive control for the Taar8b probe, exhibiting prominent DyLight-488 positive signal , thereby proving functionality of the probe.

Cell Surface Expression of TAAR8
In order to exclude lack of receptor signaling due to possible misfolding and/or premature retention of the receptor, we performed cell surface expression studies by ELISA with the N-terminally HA-tagged receptor constructs. The receptor orthologues TAAR8 and Taar8b demonstrated low but robust cell surface expression in comparison to the empty vector (mock, negative control) when expressed in COS-7 cells (Figure 3). The human β2-adrenergic receptor (ADRB2) served as a positive control which showed expression levels of 23.72 ± 1.59 (mean OD492/620 nm ± SEM fold over mock). . Cell surface expression of TAAR8. Cell surface expression was determined in COS-7 cells by ELISA as described in the methods section. Data are represented as mean ± SEM fold over mock (0.02 ± 0.001) (black) relative cell surface expression, measured at OD492/620 nm and based on three to six independent experiments with six replicates each. TAAR8 (light grey), Taar8b (dark grey) and human TAAR1 (white) showed comparable surface expression. The human β2-adrenergic receptor (ADRB2) served as a positive control, demonstrating expression levels of 23.72 ± 1.59 (mean OD492/620 nm ± SEM, fold over basal mock).

Identification of Basal Gi/o Mediated Signaling
The accumulation of cAMP was determined under basal conditions for TAAR8 and Taar8b ( Figure 4) in HEK293 cells. Human TAAR1 and the human 5-hydroxytryptamine (serotonin) receptor 1B (5HT1B) served as positive controls (not shown). Cells transiently expressing Taar8b, TAAR8, TAAR1 or 5HT1B showed no cAMP accumulation ( Figure 4A). These levels remained below the endogenous cAMP levels of mock-transfected cells.
Application of 10 µM 3-T1AM or 10 µM β-phenylethylamine (PEA) did not result in an increase in cAMP accumulation for TAAR8 or Taar8b in comparison to cells expressing TAAR1 ( Figure 4A). Furthermore, we prestimulated transfected cells with forskolin (FSK) to induce an overall activation of the adenylyl cyclase. Indeed, cAMP levels of cells expressing the murine and human TAAR8 showed a reduction of 58% (TAAR8) and 78% (Taar8b), compared to FSK-stimulated mock transfected cells ( Figure 4B), indicating strong constitutive Gi/o activity. To confirm Gi/o activation, we pre-incubated the cells with pertussis toxin (PTX), which catalyzes ADP-ribosylation of Gi/o, preventing the Gi/o protein from interacting with GPCR, which led to a significant increase in cAMP accumulation up to the levels of mock-transfected cells under FSK stimulation. These results strengthened the finding of a Gi/o basal activity of TAAR8 and Taar8b orthologue ( Figure 4B). Murine and human TAAR8 showed basal Gi/o coupling activity, indicated by a low basal cAMP accumulation compared to mock transfected cells (dotted line = mock 100% baseline) and by increased cAMP accumulation following pre-incubation with PTX and stimulation with FSK.

G12/13 or Gq/11 Signaling Pathways
Finally, we examined the activation of G protein-coupling pathways G12/13 and Gq/11 under basal and ligand-incubated conditions (Table 1). For activation of the G12/13 pathway, no significant receptor signaling was detected, regardless of whether basal or 3-T1AM or PEA-stimulated conditions were analyzed. Additionally, we found a marginal increase in IP3 formation (related to Gq/11 activation) for TAAR8 under basal conditions, which should also be attributed to the low signaling activity of the β/γ subunits of Gi/o. However, this was not observed for HEK293 cells expressing Taar8b.

Discussion
The objective of the present study was to characterize the signaling properties of TAAR8 and Taar8b, also following challenge with potential ligands. PEA and 3-T1AM were used for stimulation approaches as both are known agonists for TAAR1 [9,25]. Additionally, we analyzed the expression profile of Taar8b in mice.
Low expression of Taar8b has been previously described for the olfactory epithelium [7]. We describe herein that expression of Taar8b in young male and female C57BL6/J mice is at most marginal and close to the detection limit in the analyzed tissues. Detection was impeded by high variation due to the very low transcript levels and was only possible for a fraction of each sample set ( Figure 1 and Table S1). We measured detectable Ct-values in at least one technical replicate for 24 of 101 analyzed samples, independent from the tissue type. In our experiments, Taar8b expression was not detectable in the thyroid tissue of C57BL6/J mice (Figure 1 and Table S1). It has to be concluded that expression levels of Taar8b were at most marginal and very close to the detection limit. Finally, a clear non-ambiguous assessment of the expression levels in the analyzed tissues was not possible based on the RT-qPCR results.
As expression of Taar8b has been described to be undetectable in the heart tissue of C57BL6/J mice, an observation in line with others [27], we additionally applied the very sensitive LNA in situ hybridization for this tissue. We confirmed the published findings ( Figure 2). Hence, from our data and those of others, we conclude that the negative inotropic effects of 3-T1AM administration in WT and Taar1 knockout mice are highly unlikely to be mediated by Taar8b activation in the mouse heart. Further studies are required to elucidate the mediator of 3-T1AM effects. This is particularly important as 3-T1AM has been proposed as a potential neuroprotective compound for stroke patients [23]. If it targets other ubiquitously expressed receptors, this may lead to further, thus far unexpected effects in other tissues.
Recent findings by D'Andrea and colleagues which demonstrate that TAAR8 transcripts in astroglial cells are up-regulated following stimulation with lipo-polysaccharides strengthened the assumption of a distinct functionality in the brain [28]. However, Taar8b transcripts appear to be scarce in brain tissue, if present at all [29]. Likewise, in the RNA-Seq-based Illumina Body Map, TAAR8 expression was not detected in human brain tissue. To set this into context, it has to be emphasized that the corresponding Body Map RNA-Seq analysis was performed with total RNA prepared from whole human brain [30]. Thus, even if expression of TAAR8 would occur in detectable amounts in defined sub-regions of the brain, it is unlikely that transcript levels are abundant enough to be detected if whole-brain RNA instead of RNA prepared from these defined sub-regions is used. However, the functional implications of the regionally restricted expression of Taar8b in murine liver and specific brain regions require further investigation.
We detected basal Gi/o activity for the murine Taar8b and human TAAR8, which per se does not exclude the activation of other pathways by ligand interaction or in different cell-types as used here. A simultaneous activation of two different signaling pathways is a known concept for GPCRs [31]. The α2C-adrenergic receptor, for instance, is capable of coupling to both, Gs and Gi/o, [32]. Likewise, the β2-adrenergic receptor in cardiomyocytes also couples to both, Gs and Gi/o [33]. In this context, it is important to remember that specificity for a certain G protein is determined by several parameters such as (i) agonist concentration or type, (ii) expression levels of the receptor [32,34], (iii) cell type [35,36], and (vi) the receptor amino acid constitution (intermolecular interaction pattern) [35]. Nevertheless, the basal level of Gi/o-mediated signaling found here for Taar8b and TAAR8 point to a permanent inhibitory regulation.
The increase of cAMP accumulation levels following stimulation with 3-T1AM in FSK pre-incubated cells appears inconsistent regarding the fact that 3-T1AM alone does not significantly activate Gs signaling pathway. However, we could not exclude that TAAR8 forms heterodimers with other receptors (reviewed in [37]), which might promote a different signaling profile and response to 3-T1AM. In addition, although we do not yet know the exact endogenous ligand of TAAR8, one should consider that a ligand-independent regulating receptor-function is feasible [38]. These functions may be the capacity to modulate receptor folding, cell surface expression, and signaling of other GPCRs by forming heterodimers. Such a scenario was observed for the non-orphan GABAB1 receptor that requires the orphan GABAB2 receptor for cell surface expression rendering both fully functional [39][40][41]. The level of orphan receptor expression in relation to that of its non-orphan receptor partner might serve as a regulatory mechanism, as was described for heterodimerization of the orphan GPR50 with the non-orphan melatonin receptors 1, which leads to inhibition of signaling function [38].

Expression Analysis of Taar8b in the Heart by in Situ Hybridization
Murine Taar8b was analyzed for potential expression in the mouse heart by in situ hybridization histochemistry using digoxigenin (DIG)-labeled locked nucleic acid (LNA) probes (Exiqon Inc., Woburn, MA, USA). All studies were approved by and performed in accordance to the guidelines of the Institutional Animal Care and Use Committee (IACUC) of the University of Cincinnati (Cincinnati, OH, USA; Protocol number 06-08-07-01; approved 28 January 2010).
Wild type (WT) C57BL/6 mice were sacrificed by an intra-atrial perfusion with saline, followed by a solution of 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) at 4 °C. Hearts were then isolated and post-fixed in 4% paraformaldehyde for 16 h. Following equilibration for 48 h in RNase free 30% sucrose in 0.1 M Tris-buffered saline, tissue was cut into 25 μm sections.
Sections of heart were washed successively with PBS, 0.2 M HCl, and incubated in 0.2% glycine followed by 0.1% Triton X-100. Free floating sections were then pre-hybridized in 1× pre-hybridization buffer and 50% formamide (Sigma-Aldrich, St. Louis, MO, USA) for 1 h at 55 °C on a rocking platform. For hybridization, heart sections were incubated for 8 h in 200 nM LNA probes containing hybridization buffer (Sigma-Aldrich) at 57 °C. After stringent washing steps with decreasing concentrations of saline-sodium citrate, samples were incubated with 1:500 diluted anti-DIG antibody (goat) at 4 °C overnight. In a next step, samples were washed with Tris-Borat-EDTA-buffer (TBE) and incubated with an avidin-biotin-peroxidase complex (ABC) for 1 h at room temperature. For visualization, heart sections were stained with 3,3'-diaminobenzidin (DAB) for 5 min. Sections were mounted on gelatin-coated glass slides, dried, dehydrated through a graded ethanol series, cleared in xylene and cover-slipped for image collection by light microscopy.
A Gata4 probe served as a positive control for in situ hybridization. To test the functionality of the Taar8b probe, COS-7 cells were transiently transfected with Taar8b and hybridized as described for the heart sections. For visualization, COS-7 cells were stained with anti-DIG antibody as described above, followed by a Dy-Light-488 conjugated secondary anti-goat IgG (Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Images were collected by confocal microscopy (Zeiss confocal 710, Carl Zeiss Microscopy GmbH, Jena, Germany).
Heart, liver, kidney, thyroid and brain tissue from different regions were prepared from young male and female mice. The animals were anesthetized with IsoFlo ® (CP Pharma Handelsgesellschaft GmbH, Burgdorf, Germany), followed by i.p. injection of ketamine (Bela-pharm, Vechta, Germany). The abdominal and thoracic cavities were opened and the abdominal aorta was cut. Perfusion was performed through the left ventricle with 200 IU heparin (B. Braun Melsungen AG, Melsungen, Germany) in 0.9% NaCl solution. Tissue specimens of interest were immediately snap-frozen in liquid nitrogen after preparation and subsequently stored at −80 °C.
Tissue samples were first disrupted into smaller pieces under liquid nitrogen using a mortar and pestle. Sample preparation was continued with a tissue aliquot of suitable size. Tissue specimen homogenization combined with simultaneous RNA extraction was performed using a modification of the Trizol reagent-based extraction method described by Chomczynski and Sacchi [42]. First, tissue samples and 500 µL Trizol reagent (Invitrogen/Life Technologies, Karlsruhe, Germany) were transferred to disruption Teflon vessels containing a disruption ball, where the complete system was pre-cooled in liquid N2. Subsequently, the closed vessel was inserted into a bead mill dismembrator (Braun, Melsungen, Germany), and sample disruption was performed at 2600 rpm for 2 min. An additional 500 µL of Trizol reagent was added to the frozen tissue powder. Following thawing, 1 mL total volume of total RNA-containing TRIzol solution was removed and further processed. The final RNA preparations were additionally DNase-treated (RNase-Free DNase Set, Qiagen, Hilden, Germany) and purified using a Norgen RNA Clean-Up and Concentration Micro Kit (Norgen, Biotech Corp., Thorold, ON, Canada). Concentration and purity of RNA preparations were determined using a ND-8000 spectrophotometer (Thermo Fisher Scientific Inc., Wilmington, DE, USA), and RNA integrity was validated by means of lab-on-a-chip capillary electrophoresis technology using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). RNA quality control values are available as Table S2. In order to control for residual contaminations of genomic DNA in the RNA preparations, no-amplification-control-reactions (NACs) were performed for all samples and targeted genes. Here, qPCR reactions were assembled where the usual cDNA was exchanged with the corresponding amount of RNA. Samples yielding positive signals in any NAC-reaction (Ct < 50) were subjected to an additional DNase-treatment using a Turbo DNA-free kit (Ambion/Life Technologies, Carlsbad, CA, USA), and the NAC-test was repeated. All samples were DNA-negative following the second DNase-digestion. Only one single sample was not subjected to the NAC-test repetition because of a very limited RNA amount. Since all other samples were proven to be free of DNA-contamination following the second DNase treatment, we concluded that the second DNase digestion was successful also for this last, limited sample.
Transcript levels of Taar8b (mm03025167_gH), Thra (mm00579691_m1), and Thrb (mm00437044_m1) were determined by RT-qPCR, using specific TaqMan fluorescent reporter probe assays with Gapdh (mm99999915_g1) serving as the calibrator house-keeping gene (all TaqMan Gene Expression Assays: Applied Biosystems/Life Technologies, Carlsbad, CA, USA). Briefly, 2 µg of total RNA were reverse transcribed into cDNA using a High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA), except for three thyroid samples with restricted RNA yield, where only 1 to 1.2 µg of RNA were available for reverse transcription. Thermocycling for quantification of the initial transcript levels was performed in 20 µL reactions using TaqMan Gene Expression Master Mix, TaqMan Gene Expression Assay (both from Applied Biosystems), 100 ng of the corresponding cDNA and a 7900HT Fast Real-Time PCR System (Applied Biosystems). Initial denaturation at 95 °C for 10 min was followed by 50 cycles of denaturation at 95 °C for 15 s, and annealing/elongation at 60 °C for 60 s. All samples were analyzed in triplicates except for one thyroid pool sample with limited RNA yield, which was analyzed in duplicate. For two further thyroid samples, the qPCR assays -although performed in triplicates-were restricted to the reference gene Gapdh and the target gene Taar8b due to limited sample amount. The number of biological samples ranged for analyzed tissues between eight and 13 (liver: n = 13; heart: n = 13; kidney: n = 13; cerebellum: n = 12; cortex: n = 12; hippocampus: n = 10; striatum: n = 12; thyroid: n = 8; olfactory bulb: n = 8). One of the thyroid samples comprised a mixture of equal amounts of RNA/cDNA from two male mice. Each tissue group included at least three male and five female mice. For the graphical display, the results are provided as ΔCt values normalized to Gapdh mRNA levels determined in the same sample. All Ct-values of technical replicates for all analyzed samples are provided as Table S3.

Cloning of TAAR8
All full-length TAAR8 and control constructs were cloned into the eukaryotic expression vector pcDps and N-terminally tagged with a hemagglutinin (5'-YPYDVPDYA-3') epitope (HA) for functional assays and determination of cell surface expression. In order to enhance cell surface expression, TAAR8 was N-terminally fused (between the HA epitope and the receptor) with the first 20 amino acids of the bovine rhodopsin (Rho)-tag as previously described [6,7,43]. Plasmids were sequenced and verified with BigDye-terminator sequencing (PerkinElmer Inc., Waltham, MA, USA) using an automatic sequencer (ABI3710xl; Applied Biosystems).
Cell surface expression studies were conducted in COS-7 cells (derived from an African Green Monkey SV40-transfected kidney fibroblast cell line), which are more robust facing various washing steps and were cultured in Dulbecco's Modified Eagle Medium (DMEM) (Biochrom AG, Berlin, Germany) at 37 °C with 5% CO2 and supplemented with 5% FBS and 100 U/mL penicillin, 100 µg/mL streptomycin (Biochrom AG, Berlin, Germany) and 2 mM L-glutamine (Invitrogen, Carlsbad, CA, USA).
All functional assays were carried out with cells grown in 96 well plates (1.5 × 10 4 cells per well) coated with poly-L-lysin (Biochrom AG, Berlin, Germany). Cell surface expression was determined in 48 well plates (3.75 × 10 4 cells per well). Twenty-eight hours following seeding, receptor DNA was co-transfected with the respective reporter construct (Promega, Fitchburg, WI, USA). Transient transfection was performed in supplement-free MEM or DMEM, respectively, using Metafectene according to the manufacturer's protocol (Biontex, Munich, Germany).

Determination of Cell Surface Expression
Cell surface expression of the different receptor constructs was determined by an enzyme-linked immunosorbent assay (ELISA). Therefore, COS-7 cells were transiently transfected (six replicates in three to six independent experiments) with N-terminally HA-tagged TAAR1, TAAR8, Taar8b, and the human β2-adrenergic receptor (ADRB2) as a positive control and mock (pcDps). Cells were fixed with 4% para-formaldehyde 48 h after transfection and blocked with 10% FCS overnight. Cells were then incubated with biotin-labeled anti-HA antibody (Roche, Basel, Switzerland) (1:200) and detected with horseradish peroxidase-labeled Streptavidin (BioLegend, London, UK) (1:2500).

Determination of Gs and Gi/o Activation by cAMP Accumulation Assay
Gs and Gi/o signaling properties were determined by a competitive cAMP accumulation assay based on AlphaScreen technology (Perkin-Elmer Life Science, Boston, MA, USA) as previously described [44,45]. HEK293 cells transiently transfected with the TAAR8 orthologues were incubated for 40 min with 10 µM ligand (3-T1AM or PEA). For measurement of Gi/o activation, cells were prestimulated with 50 µM forskolin (FSK) (Applichem, Darmstadt, Germany) or co-stimulated with 50 µM FSK and 10 µM of the respective ligand. Based on our own and previously published data, TAAR1 served as an assay control [9,16] for Gs, and the human 5-hydroxytryptamine (serotonin) receptor 1B (5HT1B) as a positive controls for Gi/o activation [46]. Additionally, we tested for pertussis toxin (PTX) sensitive signaling by pre-incubating the cells with 50 µM PTX (Sigma-Aldrich, St. Louis, MO, USA) 20 h prior to ligand stimulation.

Determination of Gq/11 and G12/13 Activation by Reporter Gene Assay
Gq/11 and G12/13 activities were determined by luciferase reporter gene assay (IP3-luc or RhoA-luc), with the human thyrotropin receptor (TSHR) serving as a positive control [47]. HEK293 cells were co-transfected with TAAR8 or Taar8b and a reporter construct containing a responsive element and the firefly luciferase gene under the control of NFAT (nuclear factor of activated T-cells) in the case of Gq/11 determination, or under the control of a serum response factor (SRF-RE) in the case of G12/13 determination. Two days post-transfection, cells were washed with PBS (Biochrom AG, Berlin, Germany) and stimulated for 6 h with 10 µM 3-T1AM or 10 µM PEA in supplement-free MEM at 37 °C with 5% CO2 saturated air. Reactions were terminated by aspirating the medium. Cells were lysed and then analyzed using the automatic luciferase substrate injection in accordance to the manufacturer's protocols (Promega, Fitchburg, WI, USA), followed by quantification using a Berthold Microplate Reader (Berthold Technologies GmbH & Co. KG, Bad Wildbad, Germany). The human thyrotropin receptor (TSHR) served as a positive assay control [47].

Statistical Data Analysis
Results of functional data are indicated either as raw data or as fold over mock transfected. Basal state is shown as fold over basal mock transfection, and ligand stimulated results are referred to fold over ligand-stimulated mock with the respective substances in order to omit endogenous effects. Bar graphs and concentration response curves in which mean ± SEM are provided, and statistical analyses were generated using GraphPad Prism Version 6.0 (GraphPad Software, San Diego, CA, USA).

Conclusions
In summary, we analyzed the expression profile of Taar8b in different tissues by RT-qPCR and demonstrated Taar8b expression at most marginal expression levels, and no gender-specific differences were detected. Complete absence of expression could be validated for heart tissue using LNA in situ hybridization, thereby confirming previously published results. The physiological function of Taar8b in these tissues remains to be elucidated, particularly in combination with the yet unknown ligand. Moreover, for the first time, we confirmed basal Gi/o signaling activity for TAAR8 and Taar8b. However, 3-T1AM was not observed to induce detectable signaling of the TAAR8 subtype analyzed here and therefore, may not account for the ligand induced in vivo effects of 3-T1AM. It will be of future interest to search for other targets of 3-T1AM, including further known Taar8 subtypes or aminergic receptors.