Cell lines<\/p>\n
HEK293T\/17 (CRL-11268, RRID:\u00a0CVCL_1926<\/a>) cells were obtained from the American Type Culture Collection (ATCC). G-protein-deficient HEK293 cells (\u0394GNAS, \u0394GNAL, \u0394GNAQ, \u0394GNA11, \u0394GNA12 and \u0394GNA13, HEK293 clone 38 (ref. 54<\/a>) and \u03b2-arrestin\u00a01\/2-deficient HEK293 cells (\u0394ARRB1 and \u0394ARRB2; also known as arrestin 2 and arrestin 3, respectively, HEK293 clone 4\u00a0(ref.\u00a027<\/a>)) have been previously described. All cells were cultured in Dulbecco\u2019s Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS) (Invitrogen, CX30346) and 1\u00d7 antibiotic antimycotic solution (100 units ml\u22121 penicillin, 100\u2009\u00b5g\u2009ml\u22121 streptomycin and 250\u2009ng\u2009ml\u22121 amphotericin B; Thermo Fisher Scientific, 15240062). Cells were grown exponentially in an incubator at 37\u2009\u00b0C under 5% CO2 and subcultured at ratios of 1:2\u20131:10 every two to four days using 0.05% trypsin-EDTA (Thermo Fisher Scientific, 25300120).<\/p>\n Chemicals<\/p>\n All chemicals were obtained from MilliporeSigma unless otherwise noted. SBI-0654553 HCl (abbreviated as SBI-553) was synthesized by the Conrad Prebys Center for Chemical Genomics at the Sanford Burnham Prebys Medical Discovery Institute. Coelenterazine h and coelenterazine 400a were obtained from Cayman Chemical. For receptor signalling assessments, NT (Sigma, N6383) and PD149163 (Sigma, PZ0175) were maintained as 2\u2009mM stock in 80% glycerol. SBI-553 and SR142948A (Sigma, SML0015) were maintained as 50\u2009mM stocks in dimethyl sulfoxide (DMSO). SBI-0654553 analogues were synthesized using previously published methods3<\/a>. Derivates were maintained as 50, 25 or 12.5\u2009mM stocks in DMSO, as solubility permitted. For in vivo studies, SBI-553 and SBI-593 were freshly prepared from powdered stocks. When applicable, doses and concentrations were calculated from the formula weight of the compound salts and adjusted for fractional anhydrous base weight.<\/p>\n Recombinant DNA plasmids<\/p>\n The 3\u00d7HA-NTSR1 plasmid consists of N-terminal 3\u00d7HA-tagged WT human NTSR1 cloned into the pcDNA3.1(+) vector (Invitrogen, Thermo Fisher Scientific) at KpnI (5\u2032) and XbaI (3\u2032). This construct was purchased from the University of Missouri (cDNA Bank, NTSR10TN00). TRUPATH was a gift from B. Roth (Addgene kit 1000000163). The pcDNA3.1Zeo(\u2212) vector was acquired from Invitrogen. The mVenus-\u03b2-arrestin\u00a01 (human) and mVenus-\u03b2-arrestin\u00a02 (human) plasmids consist of human \u03b2-arrestin\u00a01 (cDNA Bank, ARRB100002) or human \u03b2-arrestin\u00a02 (cDNA Bank, ARRB200001) cloned in frame, through restriction cloning, into a pcDNA3.1 plasmid that contained the mVenus sequence in the N-terminal position. The Gq\u2013Rluc8 single-point mutants (corresponding Gq residues: V359Y, N357G, Y356C and L351N) were obtained by single-nucleotide mutagenesis using the QuikChange II XL kit (Agilent Technologies), following the manufacturer\u2019s specifications. The mVenus-tagged mini-Go was generated by G-block synthesis of the reported mini-Go sequence55<\/a> and in-frame restriction cloning into the same backbone as the rest of the mini-G constructs. The 5 and 13 C-terminal substitutions in the TRUPATH G\u03b1 subunits were performed by oligonucleotide synthesis and annealing, followed by in-frame restriction cloning. Cloning of \u03b2-arrestin and G protein sensors was performed by the University of Minnesota Viral Vector and Cloning Core and validated through Sanger sequencing.<\/p>\n BRET2 G protein activation assays<\/p>\n In the BRET2-based TRUPATH platform22<\/a>, G protein activation results in a decrease in BRET between an Rluc8-tagged G\u03b1 and a GFP2-tagged G\u03b3 protein. For visualization purposes, we plotted transformed (\u2212\u0394 net BRET) data, such that G protein activation produces upward sloping curves. Curve height is a function of both the number of G\u03b1\u2013G\u03b2\u03b3 complexes dissociating and the relative proximity of the Rluc8 and GFP2 tags. Because tag location differs among the G\u03b1 proteins and distinct \u03b3 subfamily members are used for each G\u03b1 protein, maximal changes in BRET are expected to differ among the G\u03b1 sensors. We retain the original BRET values on these curves to preserve information on sensor dynamic range. On day 1, HEK293T cells were plated in 6-well plates (750,000 per well) in DMEM containing 10% FBS and 1% 1\u00d7 antibiotic antimycotic solution. On day 2, cells were transiently transfected with NTSR1 (200\u2009ng per well), G\u03b21 or G\u03b23 (100\u2009ng per well), GFP2-tagged G\u03b39, G\u03b38, G\u03b313 or G\u03b31 (100\u2009ng per well) and the G\u03b1 of interest using a standard calcium phosphate transfection protocol, and the pairings of G\u03b1, G\u03b2 and G\u03b3 subunits as described previously22<\/a> For mutant G proteins, the preferred \u03b2\u03b3 combination for the G protein base was maintained. A summary of \u03b2\u03b3 selections is provided in Supplementary Table 13<\/a>. This produces a receptor:G\u03b1:G\u03b2:G\u03b3 ratio of 2:1:1:1. On day 3, cells were plated (25,000 cells per well) onto poly-d-lysine (PDK)-coated (100\u2009ng\u2009ml\u22121), clear-bottom, white-walled 96-well plates in Opti-MEM containing 2% FBS and 1% 1\u00d7 antibiotic antimycotic solution. On day 4, cells were incubated in 70 or 80\u2009\u00b5l per well Hanks\u2019 balanced salt solution (HBSS) containing calcium and magnesium and 20\u2009mM HEPES for five to six hours before treatment. Cells receiving an SBI-553 or SR142948A pretreatment were incubated in 70\u2009\u00b5l per well of HBSS containing 20\u2009mM HEPES. Cells not receiving any pretreatment were incubated in 80\u2009\u00b5l per well of HBSS plus 20\u2009mM HEPES. SR142948A was freshly prepared in HBSS from a 50\u2009mM DMSO stock. The 10\u00d7 NT and PD149163 were freshly prepared in HBSS from 2\u2009mM 80% glycerol (NTS, PD149163) stocks, and 10\u00d7 SBI-553 was freshly prepared in HBSS with 5% 2-hydroxylpropyl-\u03b2-cyclodextrin (HP-\u03b2-CD, Tokyo Chemical Industry) from a 50\u2009mM DMSO stock. For SAR studies, 10\u00d7 SBI-553 and all derivatives were freshly prepared in 30%\u00a0HP-\u03b2-CD from 50, 25 or 12.5\u2009mM DMSO stocks. A white vinyl sticker was placed on the bottom of the plate. For room-temperature studies, plates were allowed to cool for 10\u2009min until they reached approximately 25\u2009\u00b0C. Ten microlitres of 10\u00d7 SBI-553, SR142948A or SBI-553 derivative pretreatments were added during the 10-min cooling period, a total of 20\u2009min before reading. For studies at 35\u2009\u00b0C, treatments were incubated at temperature. Ten microlitres of 10\u00d7 NTS, PD149163, SBI-553 or SR142948A treatments were added 10\u2009min before reading. Ten microlitres per well of a 10\u00d7 concentration of coelenterazine 400a (final concentration around 7.5\u2009\u00b5M, Cayman Chemical) was added 5\u2009min before reading. After treatment with coelenterazine 400a, plates were protected from light. Plates were read with a Tecan SparkCyto Microplate reader at room temperature (25\u2009\u00b0C) or 35\u2009\u00b0C in ambient air every 5\u2009min for 20\u2009min. BRET2 ratios were computed as the ratio of GFP2 emission to Rluc8 emission. The \u0394 net BRET ratio was calculated by subtracting the stimulated GFP2\/Rluc8 ratios from control GFP2\/Rluc8 ratios for each read. The minimum \u0394 net BRET ratio over time was averaged within treatments and combined between experiments. Data are presented as negative mean \u0394 net BRET ratio\u2009\u00b1\u2009s.e.m. from at least three independent experiments.<\/p>\n BRET1 \u03b2-arrestin recruitment assays<\/p>\n Recruitment of mVenus-tagged human \u03b2-arrestin\u00a01 and human \u03b2-arrestin 2 to Renilla luciferase (Rluc8)-tagged NTSR1 was assessed in HEK293T cells using a BRET assay, as described4<\/a>. Previous studies evaluating NTSR1 \u03b2-arrestin agonism have used bovine or rodent \u03b2-arrestin constructs4<\/a>,5<\/a>. To assess recruitment of the human \u03b2-arrestins, the following procedure was used. On day 1, HEK293T cells were plated in 6-well plates (750,000 per well) in growth medium. On day 2, cells were transiently transfected with Rluc8-tagged NTSR1 (100\u2009ng per well), Venus-tagged \u03b2-arrestin\u00a02 or Venus-tagged \u03b2-arrestin\u00a01 (1.4\u2009\u00b5g per well), and pcDNA3.1 (1.5\u2009\u00b5g per well) using a standard calcium phosphate transfection protocol. To maximize assay sensitivity, with the exception of the NT combination studies in Fig. 2<\/a>, transfections also included 500\u2009ng GRK2. On day 3, cells were plated onto PDK-coated (100\u2009ng\u2009ml\u22121), clear-bottom, white-walled 96-well plates (40,000 cells per well) in Opti-MEM containing 2% FBS and 1\u00d7 antibiotic antimycotic solution. On day 4, cells were incubated in 70 or 80\u2009\u00b5l per well HBSS containing calcium and magnesium and 20\u2009mM HEPES for five to six hours before treatment. Cells receiving an SBI-553 or SR142948A pretreatment were incubated in 70\u2009\u00b5l per well of HBSS containing 20\u2009mM HEPES. Cells not receiving a pretreatment were incubated in 80\u2009\u00b5l per well of HBSS plus 20\u2009mM HEPES. The 10\u00d7 NT and PD149163 were freshly prepared in HBSS from 2\u2009mM 80% glycerol (NT, PD149163) stocks, and 10\u00d7 SBI-553 was freshly prepared in HBSS with 5% HP-\u03b2-CD from a 50\u2009mM DMSO stock. For SAR studies, 10\u00d7 SBI-553 and all derivatives were freshly prepared in 30% HP-\u03b2-CD from 50, 25 or 12.5\u2009mM DMSO stocks. For compound combination studies, cells were pretreated with 10\u2009\u00b5l of 10\u00d7 SBI-553, SR142948A or vehicle pretreatments 15\u2009min before reading. Ten microlitres per well of a 10\u00d7 concentration of coelenterazine h (final concentration 4.7\u2009\u00b5M; Cayman Chemical) was added 10\u2009min before reading, followed by treatment with 10\u2009\u00b5l of 10\u00d7 NT 5\u2009min before reading. Cells receiving no pretreatment received 10\u2009\u00b5l of 10\u00d7 NT, PD149163, SBI-553 and SR142948A treatments 10\u2009min before reading and 10\u00d7 coelenterazine h 5\u2009min before reading. For experiments in Figs. 1<\/a> and 2<\/a>, plates were maintained and read at room temperature. For the temperature studies, time-course studies and analogue screening and characterization in Extended Data Figs. 2<\/a>, 3<\/a> and 6<\/a>\u20138<\/a>, plates were maintained at 35\u2009\u00b0C during treatment and reading. A white vinyl sticker was placed on the bottom of each plate. Plates were read with a CLARIOstar Plus microplate reader (BMG Labtech) set at 25\u2009\u00b0C or 35\u2009\u00b0C at 10, 15, 20, 25, 30 and 35\u2009min after treatment. BRET1 ratios were computed as the ratio of Venus emission to Rluc8 emission. The \u0394 net BRET ratio was calculated by subtracting the stimulated Venus\/Rluc8 ratios from control Venus\/Rluc8 ratios for each read. The maximum \u0394 net BRET ratio over time was averaged within treatments and combined between experiments. Data are presented as mean \u0394 net BRET ratio\u2009\u00b1\u2009s.e.m. from at least three independent experiments.<\/p>\n TGF\u03b1 shedding assays of G protein activation<\/p>\n TGF\u03b1 shedding assays of G protein activation were performed as originally presented23<\/a>, with previously described modifications4<\/a>,56<\/a>. These modifications included using G-protein-deficient (\u0394GNAS, \u0394GNAL, \u0394GNAQ, \u0394GNA11, \u0394GNA12 and \u0394GNA13) HEK293 cells54<\/a> and the fluorescent substrate 4-methylumbelliferyl phosphate (4-MUP), at a working concentration of 1\u2009mM per well, instead of p-nitrophenyl phosphate (p-NPP). On day 1, G-protein-deficient HEK293 cells were plated in 6-well plates (750,000 per well) in growth medium. On day 2, expression vectors were transiently transfected in HEK293 cells using Lipofectamine 2000 (Invitrogen; 8\u2009\u00b5l per well in a six-well plate). Expression vectors included 1.875\u2009\u00b5g AP-TGF\u03b1, 750\u2009ng HA-tagged NTSR1 and 350\u2009ng of G\u03b1 protein. On day 3, transfected cells were detached with a brief rinse (1\u2009ml, around 30\u2009s) of phosphate-buffered saline (PBS) followed by 0.5\u2009ml per well of 0.05% trypsin-EDTA (Gibco). The cell suspension was pelleted by centrifugation (200g, 5\u2009min), followed by a resuspension in 3\u2009ml HBSS containing 5\u2009mM HEPES (pH 7.4) and a 10-min incubation at room temperature. Cells were again centrifuged (200g, 5\u2009min) and resuspended in 4\u2009ml HBSS containing 5\u2009mM HEPES (pH 7.4). The resuspended cells were plated in 80\u2009\u00b5l per well in a 96-well plate and placed in an incubator at 37\u2009\u00b0C with 5% CO2. After a 30-min incubation, cells were treated with 10\u2009\u00b5l vehicle (HBSS), 10\u00d7 final concentrations of SR142948A or SBI compounds and incubated at 37\u2009\u00b0C for 20\u2009min. For preparation of 10\u00d7 stocks, SBI compounds and SR142938A were diluted directly from 1\u201310\u2009mM DMSO stocks into HBSS. The final concentration of DMSO did not exceed 0.4%. After a 20-min incubation with SBI\u00a0compounds, cells were treated with 10\u2009\u00b5l of 10\u00d7 concentration of NT and incubated at 37\u2009\u00b0C for one hour. For inhibition studies, 10\u2009nM NT was used for the Gq, Gi1\/2 and G12 constructs, and 100\u2009nM was used for the Go construct; these concentrations were selected to elicit maximal NT-induced shedding. Plates were centrifuged (190g, 2\u2009min) and conditioned medium (80\u2009\u00b5l) was transferred into a new 96-well plate. After a 20-min incubation at room temperature, 2\u2009mM 4-MUP-containing solution was added (80\u2009\u00b5l per well) to both the conditioned medium and the cell plate. Alkaline phosphatase activity was measured using a CLARIOstar Plus microplate reader set to 25\u2009\u00b0C. Data were collected before and after a one-hour incubation at 37\u2009\u00b0C. Excitation was set at 360\u2009nm (\u00b110\u2009nm) and emission at 450\u2009nm (\u00b115\u2009nm). TGF\u03b1 shedding activity was calculated by dividing the amount of phosphatase activity present in the conditioned medium by the amount present on the cells plus the conditioned medium. All values were standardized to background shedding activity.<\/p>\n Mini-G protein recruitment assays<\/p>\n Mini-G proteins have a truncated N\u2009terminus and contain a mutation that uncouples GPCR binding from nucleotide release, allowing them to form more stable associations with GPCRs than do unmodified G proteins25<\/a>,55<\/a>; these associations are amenable to monitoring by BRET. The recruitment of Venus-tagged mini-G proteins25<\/a> to Rluc8-tagged NTSR1 was assessed by BRET. Note that although mini-Go and mini-G12 were derived from their full-length versions, the mini-Gq and mini-Gi1 constructs were derived from the Gs backbone with substitution of the \u03b15 helix25<\/a>,55<\/a>. To assess recruitment, on day 1, HEK293T cells were plated in 6-well plates (750,000 per well) in growth medium. On day 2, cells were transiently transfected with Rluc8-tagged NTSR1 (100\u2009ng per well), a Venus-tagged mini-G protein and pcDNA3.1 (1.5\u2009\u00b5g per well or 2.65\u2009\u00b5g per well) using a standard calcium phosphate transfection protocol. Cells transfected with a Venus-tagged mini-Gq or mini-Gi1 received 250\u2009ng per well. Cells transfected with a Venus-tagged mini-G12, mini-Gs or mini-Go received 1.5\u2009\u00b5g per well. On day 3, cells were plated onto PDK-coated (100\u2009ng\u2009ml\u22121), clear-bottom, white-walled 96-well plates (40,000 cells per well) in Opti-MEM containing 2% FBS and 1\u00d7 antibiotic antimycotic solution. On day 4, cells were incubated in 70 or 80\u2009\u00b5l per well HBSS containing calcium and magnesium and 20\u2009mM HEPES for three to four hours before treatment. Cells receiving an SBI-553, PD149163 or SR142948A pretreatment were incubated in 70\u2009\u00b5l per well of HBSS containing 20\u2009mM HEPES. Cells not receiving any pretreatment were incubated in 80\u2009\u00b5l per well of HBSS plus 20\u2009mM HEPES. Cells were pretreated with 10\u2009\u00b5l of 10\u00d7 concentrations of SBI compounds, SR142948A, PD149163 or vehicle (5% or 30% HP-\u03b2-CD) 20\u2009min before reading. Cells were then treated with vehicle (HBSS) or 100\u2009nM NT 10\u2009min before reading. Finally, cells were treated with 10\u2009\u00b5l per well of a 10\u00d7 concentration of coelenterazine h (final concentration 4.7\u2009\u00b5M) 5\u2009min before reading. Plates were read on a CLARIOstar Plus microplate reader set at 25\u2009\u00b0C at 5, 10, 15 and 30 min after treatment. Mini-G protein recruitment was calculated by subtracting the stimulated Venus\/Rluc8 ratios from control Venus\/Rluc8 ratios for each read (\u0394 net BRET).<\/p>\n Western blot analysis<\/p>\n \u03b2-Arrestins 1 and 2 were detected by Western blot. On day 1, HEK293T and \u03b2-arrestin\u00a01\/2-null HEK293 cells were plated at 750,000 cells per well onto 6-well plates. On day 2, cells were lysed on ice in 2\u00d7 sample buffer. Whole-cell lysates were sonicated and then analysed for expression of \u03b2-arrestins 1 and 2. Cell protein samples (10\u2009\u03bcl) were resolved on 10% SDS\u2013polyacrylamide gels (NuPAGE, Bis-Tris; Thermo Fisher Scientific) and transferred to nitrocellulose membranes (0.45 \u03bcm pore size; Thermo Fisher Scientific). The membrane was rinsed with Tris-buffered saline (TBS; Li-Cor Biosciences) and then stained with Ponceau S solution (Sigma-Aldrich, 6226-79-5) to visualize protein loading. Membranes were rinsed three times with 50% TBS containing 0.1% Tween-20 (v\/v; TBST) and incubated in Odyssey Blocking Buffer for one hour at room temperature. Blocked membranes were incubated overnight at 4\u2009\u00b0C with rabbit anti-\u03b2-arrestin\u00a01\/2 (Cell Signaling Technology, 4674, RRID: AB_10547883<\/a>) in 50% blocking buffer and 50% TBST at a 1:1,000 dilution. Membranes were washed with TBST before and after incubation with infrared secondary antibodies Alexa Fluor goat anti-rabbit 680 (Invitrogen, A-21109, RRID: AB_2535758<\/a>) at a 1:5,000 dilution for one hour at room temperature. Membranes were imaged on a LI-COR Biosciences Odyssey imaging system.<\/p>\n Optimization of the structure of the NT\u2013NTSR1\u2013SBI-553\u2013Go The cryo-electron microscopy (cryo-EM) structure of NTSR1 bound to SBI-553 and Go (PDB: 8FN0<\/a>) was used at the starting point for modelling. Structure preparation was performed with Maestro (Schr\u00f6dinger). VDW clashes were removed by constrained minimization using the OPLS4 force field. A full conformational analysis was performed with the conformational ensemble generated by MOE (Chemical Computing Group) and minimization using density functional theory calculations (Gaussian, \u03c9b97XD\/6-311+G** functional). The SBI-553 piperidine preferred a chair conformation over the twist boat by more than 3.1\u2009kcal\u2009mol\u22121. As a result, the SBI-553 conformation from PDB: 8JPB<\/a> was substituted into the model by rigid superposition of the two SBI-553 conformations followed by a second constrained minimization of the minor clashes resulting from this replacement. Molecular mechanics calculations (minimization and molecular dynamics) with multiple force fields orient the fluorine on SBI-553\u2019s cyclopropyl group perpendicular to the quinazoline ring. To investigate this orientation, a rotational analysis was performed on SBI-553, SBI-342 and SBI-593 at 10\u00b0 increments about the quinazoline\u2013cyclopropane C\u2013C bond with full minimization using \u03c9b97XD\/6-311+G**. For each molecule, there was a strong preference (2.5 to 3.2\u2009kcal\u2009mol\u22121) for the conformation reported in a previous study5<\/a> in which the substituent on the cyclopropane (F for SBI-553, CH3 for SBI-342 and H for SBI-593) is in the plane of the quinazoline and pointed away from the appended piperidine. Subsequent protein minimizations and molecular dynamics simulations used explicit constraints to maintain desired small-molecule conformations.<\/p>\n G protein homology modelling<\/p>\n Homology models of each G protein were built with MOE (v.2022.02, Chemical Computing Group), with the conformation of each mutated side chain optimized individually using MOE\u2019s Amber 10:EHT force field. Models of Gq conformations required building the NTSR1 H1\/H2 loop absent in PDB: 8FN0<\/a>. This loop is resolved in the PDB: 6OS9<\/a> structure, so the missing residues were concatenated onto PDB: 8FN0<\/a> using the homology model application in MOE. Minimization of the protein complexes containing SBI-342 and SBI-593 used the OPLS4 force field in Maestro (v.14.1.138, Schr\u00f6dinger).<\/p>\n Molecular dynamics simulations<\/p>\n Molecular dynamics simulations with Gq and SBI-593 were done with Desmond using the Schr\u00f6dinger suite (v.14.3.129 and v.13.7.125). Standard protein preparation included building the G87\u2013E92 loop of mini-Gq. SBI-593 was superimposed onto SBI-553, and Gq was mapped onto Go from PDB: 8FN0<\/a>. The conformation of each side chain of Gq-H5 was independently sampled and quaternary (NTSR1, Gq, NT and SBI-553) complex minimized. Where side chains interacted with each other, multiple side-chain combinations were assessed and their energies compared. The side-chain rotamers were well-sampled during the dynamics runs. Dynamics set-up included explicit water, manually placed 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) membrane, 0.15\u2009N NaCl and a constraint on cyclopropyl rotation. Six discrete simulations were performed (three 100\u2009ns and three 300\u2009ns) using the NP\u0263T ensemble, OPLS5 force field and TIP3P water model (300\u2009K). Equilibration used a 100-ps NVT ensemble simulation with Brownian dynamics at 10\u2009K, a 12-ps NVT ensemble simulation with a Langevin thermostat at 10\u2009K, two constrained 12-ps NPT simulations using a Langevin thermostat and barostat at 1\u2009atm at 10\u2009K then 300\u2009K, and a 24-ps NP\u0263T equilibration using a Langevin thermostat and barostat at 300\u2009K and 1\u2009atm. Membrane relaxation used a 1-ps Nose\u2013Hoover chain thermostat and a 2-ps Martyna\u2013Tobias\u2013Klein barostat. Root mean square deviation (RMSD) was measured for the C\u03b1 carbons, and side-chain movement was monitored to ensure adequate sampling of conformational space. Once the complex stabilized to a consistent C\u03b1 RMSD, 20\u201325 frames from each run were selected from periods of the greatest RMSD stability. Each complex was minimized in Maestro, in which conformations, orientations and energies were compared.<\/p>\n Mice<\/p>\n All mouse studies were performed in accordance with the National Institutes of Health Guidelines for Animal Care and Use of Laboratory Animals and with approved animal protocols from the University of Minnesota University Animal Care and Use Committee. The mice studied included: C57BL\/6J mice (Jackson Laboratory, 000664), global Ntsr1\u2212\/\u2212 mice (B6.129P2-Ntsr1tm1Dgen\/J, Deltagen, Jackson Laboratory strain 005826), \u03b2-arrestin\u00a02-knockout (Arrb2\u2212\/\u2212) mice57<\/a> and their respective WT littermates. All mouse lines were backcrossed onto a C57BL\/6J genetic background for at least ten generations before use. At the start of the study, all mice were adults. Mice were 8\u201320 weeks old, weighed 19\u201330\u2009g, and were age-matched across experimental groups. Experiments included both male and female mice, and experimental groups were sex-matched. Ntsr1\u2212\/\u2212 and Arrb2\u2212\/\u2212 mice for experimental use were exclusively generated by Ntsr1+\/\u2212 \u00d7 Ntsr1+\/\u2212 or Arrb2+\/\u2212 \u00d7 Arrb2+\/\u2212 breeding, such that littermate WT mice could serve as controls. Littermates of the same sex and genotype were randomly assigned to experimental groups. Mice for systemic PD149163-induced hypothermia studies were group housed in conventional cages with Teklad irradiated corncob bedding (Inotiv) and Enviro-Dri nesting material (Fibercore) and maintained on a 14-h\u201310-h light\u2013dark cycle. Mice for local NAc injection studies were singly housed after implantation of intracranial cannulas in conventional cages with Teklad irradiated corncob bedding (Inotiv) and Enviro-Dri nesting material (Fibercore), and maintained on a 14-h\u201310-h light\u2013dark cycle. Experiments began at the start of the light cycle (within three hours of the light cycling beginning). Tap water and standard laboratory chow were supplied ad libitum, except during testing.<\/p>\n Assessments of core body temperature<\/p>\n Core body temperatures were measured using a rectal probe thermometer for mice (Thermalert Model TH-8, Physitemp Instruments), as described4<\/a>. Core body temperatures were recorded from age-matched male and female Ntsr1\u2212\/\u2212 and WT mice before treatment (time 0) and 30, 60, 90, 120 and 300-min after treatment. Mice were gently restrained during the procedure and acclimated to this process during baselining. For single-compound dosing studies, baseline temperatures were recorded, and mice received either PD149163 (0.03\u20131\u2009mg\u2009kg\u22121) or PD149163\u2019s vehicle (physiological saline). All treatments were administered i.p. in a volume of 10\u2009ml\u2009kg\u22121. For systemic multiple-compound dosing studies, mice received SBI-553 (12\u2009mg\u2009kg\u22121, i.p.) or its vehicle (5% cyclodextrin) or SBI-593 (12\u2009mg\u2009kg\u22121, i.p.) or its vehicle (20% DMSO, 5% Tween-80) before the start of the study (\u221245\u2009min). After temperature recording at time 0, mice received PD149163 (0.15\u2009mg\u2009kg\u22121) in volumes of 10\u2009\u00b5l per kg i.p.<\/p>\n Intracranial cannulation and local NAc injections<\/p>\n Local NAc microinjections were accomplished after bilateral guide cannula placement in Ntsr1\u2212\/\u2212 mice and their WT littermates, C57BL\/6J mice. Cannulas were acquired from Plastics One (bilateral guide: 2.0\u2009mm spacing, 26G, 4\u2009mm below pedestal; bilateral internal: 2.0\u2009mm spacing, 33G, 0.5\u2009mm projection; bilateral dummy: 2.0\u2009mm spacing, 0.008\u201d\/0.2\u2009mm, 0\u2009mm projection). Bilateral guide cannulas were inserted into the NAc at +1.3\u2009mm AP with 2.0-mm spacing (\u00b11.0\u2009mm\u2009mediolateral (ML)) and \u22124.5\u2009mm dorsoventral (DV) and fixed to the skull with dental cement. Mice were singly housed after surgery and allowed to recover for at least seven days. After recovery, compounds were injected bilaterally using an automated syringe pump (Harvard Apparatus). SBI-553 (molecular weight 450\u2009g\u2009mol\u22121) was dissolved in 1% (v\/v) DMSO and 20% (v\/v) HP-\u03b2-CD at 74.07\u2009mM. One hundred micrograms of SBI-553 in 3\u2009\u00b5l was injected per side at a rate of 0.2\u2009\u00b5l per min. SBI-593 (molecular weight 458\u2009g\u2009mol\u22121) was dissolved in 80% (v\/v) N,N-dimethylacetamide, 10% (v\/v) Tween-80 and 10% (v\/v) UltraPure distilled water at 72.7\u2009mM. One hundred micrograms of SBI-593 in 3\u2009\u00b5l was injected per side at a rate of 0.2\u2009\u00b5l per min. After local injection of SBI compounds, mice were placed in their home cages for one hour before intra-NAc administration of PD149163. PD149163 (molecular weight 943.91\u2009g\u2009mol\u22121) was diluted from a stock concentration of 2\u2009mM in 80% glycerol to 10 or 100\u2009\u00b5M in separate SBI compound or vehicle preparations, and 18.9, 1.89 or 0.189\u2009ng PD149163 in 0.2\u2009\u00b5l was injected per side at a rate of 0.2\u2009\u00b5l per min. The PD149163 dose used was 18.9\u2009ng in the WT versus Ntsr1\u2212\/\u2212 experiments, 1.89\u2009ng in the 100-\u00b5g SBI-553 experiments and 0.189\u2009ng in the 10-\u00b5g SBI pretreatment experiments. The GRK2\/3 inhibitor Compound 101 (Hello Bio) was diluted in vehicle (10% (v\/v) DMSO and 20% (v\/v) hydroxypropyl cyclodextrin) from a stock concentration of 20\u2009mM in DMSO to 1.99\u2009mM and 0.5\u2009\u00b5g in 0.5\u2009\u00b5l was injected per side at a rate of 0.4\u2009\u00b5l per min one hour before the injection of 1.89\u2009ng PD149163. This Compound 101 dose and preparation was previously shown to inhibit GRK2\/3-mediated effects after intracranial delivery in mice57<\/a>. After local injection of PD149163, body temperature was monitored every 30\u201360\u2009min for 8\u2009h. For all experiments, mice were randomly assigned to treatment groups for no more than two experiments, separated by a minimum of seven days. Cannula placements were verified by ink injection. After the experiment, 0.5\u2009\u00b5l India ink (1:20 (v\/v) saline) was injected at a flow rate of 0.5\u2009\u00b5l per min to visualize the observed injection site. Injection sites were identified using anatomical makers and documented on a coronal mouse brain atlas53<\/a>. Mice were anaesthetized with isoflurane and brains were collected one hour after ink microinjection. Brains were sliced in 250-\u00b5m sections using a vibratome (VT1000S, Leica).<\/p>\n Statistical analysis<\/p>\n All data are represented as mean\u2009\u00b1\u2009s.e.m., unless otherwise indicated. Data were analysed and plotted using the software GraphPad Prism v.10.1.2. Information on curve fitting, statistical tests and n numbers are provided in Supplementary Tables 1<\/a>\u20136<\/a>. In Supplementary Tables 1<\/a>\u20136<\/a>, n represents the number of biological replicates. All data represent the average of at least three biological replicates from three independent experiments. Cell-based experiments included two or three technical replicates for every condition. A P\u2009value of less than 0.05 was accepted as statistically significant.<\/p>\n Figure illustrations<\/p>\n Method and concept figure illustrations were created using BioRender. Images of NTSR1 and G protein structures were created in ChimeraX (UCSF, v.1.6). Cannula placements are presented on Paxinos and Franklin\u2019s \u2018The Mouse Brain In Stereotaxic Coordinates\u201953<\/a>.<\/p>\n Reporting summary<\/p>\n
\n A complex for modelling<\/p>\n