Recovery of the histamine H3 receptor activity lost in yeast cells through error-prone PCR and in vivo selection

G protein-coupled receptors (GPCRs) are the largest protein family in humans and are important drug targets. Yeast, especially Saccharomyces cerevisiae, is a useful host for modifying the function and stability of GPCRs through protein engineering, which is advantageous for mammalian cells. When GPCRs are expressed in yeast, their function is often impaired. In this study, we performed random mutagenesis using error-prone PCR and then an in vivo screening to obtain mutants that recovered the activity of the human histamine H3 receptor (H3R), which loses its signaling function when expressed in yeast. Four mutations with recovered activity were identified after screening. Three of the mutations were identified near the DRY and NPxxY motifs of H3R, which are important for activation and are commonly found in class A GPCRs. The mutants responded exclusively to the yeast YB1 strain harboring Gi-chimera proteins, showing retention of G protein specificity. Analysis of one of the mutants with recovered activity, C415R, revealed that it maintained its ligand-binding characteristics. The strategy used in this study may enable the recovery of the activity of other GPCRs that do not function in S. cerevisiae and may be useful in creating GPCRs mutants stabilized in their active conformations.


Yeast strains and growth media
The yeast strains, YB1 (also known as MPY578t5), YB11 (MPY578q5) and YB14 (MPY578s5) 27 , were kindly provided by Professor Bryan Roth's laboratory at the University of North Carolina, Chapel Hill (North Carolina, USA).The -Ura and -Ura-His dropout supplements were prepared according to a published formulation 28 .YPAD medium (20 g/L peptone, 10 g/L yeast extract, 2% glucose, 20 mg/L adenine sulfate) was used for non-selective applications, while synthetic complete (SC) medium without uracil (-Ura/SC; 1.7 g/L yeast nitrogen base without amino acids, 5 g/L ammonium sulfate, 1.92 g/L -Ura dropout supplement, and 2% glucose) or SC medium without uracil and histidine (-Ura-His/SC; 1.7 g/L yeast nitrogen base without amino acids, 5 g/L ammonium sulfate, 1.92 g/L -Ura-His dropout supplement, and 2% glucose) was used for selective applications.The Escherichia coli strain, DH5α, was grown in LB medium containing 100 μg/mL ampicillin for plasmid preparation.

Plasmids
Plasmid p416GPD was kindly provided by Professor Bryan Roth.A codon-optimized cDNA of the full-length histamine H 3 receptor (H 3 R_FL), which has an N11Q mutation that prevents glycosylation, was kindly provided by Professor Iwata's laboratory at Kyoto University (Kyoto, Japan).A series of genes, including the TEV protease cleavage site, enhanced green fluorescent protein (GFP), and histidine tag, were integrated downstream of the GPD promoter in the p416GPD plasmid, resulting in plasmid p416GPD_GFP.The H 3 R_FL or H 3 R variants with the third intracellular loop removed from 237 to 340 (H 3 R_i3d) were integrated immediately before the TEV site of the p416GPD_GFP plasmid, resulting in an expression plasmid in which GFP was fused to the C-terminus of the receptor (p416GPD_H 3 R_FL-GFP or p416GPD_H 3 R_i3d-GFP).The H 3 R_i3d-GFP gene was integrated into the pPIC9K plasmid (pPIC9K_H 3 R_i3d-GFP), a Pichia pastoris plasmid, to prevent template-derived colonies when transforming DNA fragments of error-prone PCR into S. cerevisiae.

Random mutagenesis and generation of the yeast library
Random mutations were introduced via error-prone PCR using a GeneMorph II Random Mutagenesis Kit (Agilent Technologies).The reactions were performed according to the manufacturer's instructions.Approximately 4.8 μg of template pPIC9K_H 3 R_i3d-GFP (~ 11.3 kb) was used per reaction volume (50 μL).This amount corresponds to approximately 420 ng of target DNA (H 3 R_i3d; ~ 1 kb) per reaction.Five reaction solutions (250 μL in total) were prepared for each transformation.To amplify the H 3 R_i3d gene for random mutation, a forward primer 81_GPD_F (5′-AGT TTC GAC GGA TTC TAG AAC TAG GGA TCC ATG GAA AGA GCT CCA-CCA GAT GG-3′) and a reverse primer 81_YHR_R (5′-AAA TTG ACC TTG AAAA-TAT AAA TTT TCC CCC TTC CAA CAA TGT TCC AAA GAA GAATG-3′) were used.The primers contained an overlapping region in the p416GPD_GFP vector for gap repair recombination cloning.The reaction was carried out in 30 cycles at an annealing temperature of 50 °C.The desired fragments were confirmed using electrophoresis.
After random mutagenesis, both the SmaI-digested p416GDP_GFP and DNA fragment of H 3 R_i3d were cotransformed into the yeast YB1 strain.Briefly, an overnight culture of YB1 in YPAD medium was subcultured in 500 mL of fresh YPAD medium to an OD 600 of 0.6 at 30 °C.The cells were recovered from the culture medium via centrifugation at 3000 rpm for 5 min using a swinging-bucket rotor, washed with sterile water, and then suspended in 4 mL of 0.1 M LiAc/TE.Single-strand carrier DNA (10 mg), 20 μg SmaI-digested p416GPD_GFP, 250 μL of random mutagenesis reaction mix, and 30 mL of 40% PEG3350/0.1 M LiAc/TE were added to the cell suspension, which was then stirred well using a vortex mixer and then incubated for 30 min with shaking at 30 °C.DMSO (3.5 mL) was added to the cell suspension, and heat shock was performed at 42 °C for 25 min.The cells www.nature.com/scientificreports/were then harvested and washed once with water.The cell pellet was resuspended in 3 mL of water and plated on 15 × 15 cm plates of -Ura-His/SC medium containing 10 mM 3-AT and 100 μM histamine.The plates were incubated at 30 °C for 4 days.A portion of the cell suspension was plated on a -Ura/SC medium and incubated at 30 °C for 3 days to determine the colony-forming units per amount of plasmid (cfu/μg).

Recovering plasmid DNA from yeast
Yeast colonies were inoculated in 5 mL of -Ura/SC medium, and left to grow for 18-20 h at 30 °C with shaking.The cells were collected via centrifugation and the supernatant was discarded.The cell pellet was resuspended in 500 μL of 1 M sorbitol solution, and 20 μL of 12.5 mg/mL Zymolyase 20 T (Nacalai Tesque, Japan) was added.The cell suspension was incubated at 37 °C for 40 min and then centrifuged at 12,000g for 1 min at 22-25 °C.After removal of the supernatant, the cell pellet was resuspended in 200 μL of mP1 solution from the FastGene Plasmid Mini Kit (NIPPON Genetics, Japan).A 150-μL volume of acid-washed glass beads (Sigma) was added, and the cells were disrupted by vortexing at 4 °C for 2 min.The disrupted cell suspension was transferred to a clean tube and 200 μL of mP2 solution was added.Subsequent operations were performed according to the manufacturer's instructions for plasmid extraction from E. coli.

Ligand-concentration dependent growth assay
Yeast colonies were inoculated in 5 mL of -Ura/SC medium in test tubes, and grown at 30 °C for 20-22 h at 160 rpm in a shaking incubator.The cells were collected via centrifugation at 8000g for 3 min, and the supernatant was discarded.The cells were then resuspended in sterile water and collected via centrifugation at 8000g for 3 min.After the supernatant was discarded, the cells were then resuspended in -Ura-His/SC medium at an OD 600 of 1.Thereafter, the cells were diluted to an OD 600 of 0.02 in 500 μL of the -Ura-His/SC medium with 10-40 mM 3-AT and a fivefold dilution series of ligands in a 96-well deep well plate, and cultured at 25 °C for 70 h at 1450 rpm in an M/BR-022UP shaker (TAITEC, Japan).The culture (100 μL) was transferred to a 96-well microplate, and the OD 595 was measured using an iMark micro plate reader (Bio-Rad, USA).Curve fitting was performed using GraphPad Prism 6.0 (GraphPad Software, San Diego, CA, USA).

Measuring whole-cell GFP fluorescence
Yeast cells cultured in 500 μL -Ura/SC medium in 96-well deep well plates were collected at 3000g for 30 min at 4 °C using a microplate swinging-bucket rotor.The supernatant was then removed via suction using an aspirator.The cells were resuspended in 160 μL/well of the resuspension buffer (50 mM Tris-HCl pH 7.5, 5 mM EDTA, 10% glycerol, 0.12 M sorbitol) with the complete protease inhibitor EDTA-free cocktail (Merck).The fluorescence of a 20-μL cell suspension was measured in a small volume 384-well microplate (Greiner Bio-One, Frickenhausen, Germany) at an excitation wavelength of 490 nm and a detection wavelength of 525 nm using a 515 nm cutoff filter and a Spectra Max Gemini XPS fluorescence plate reader (Molecular Devices, USA).

Yeast membrane preparation
Colonies were inoculated in 10 mL of -Ura/SC medium in test tubes and grown at 30 °C for 20-22 h at 160 rpm in a shaking incubator.The cells were diluted to an OD 600 of 0.12 in a fresh 1-L medium of -Ura/SC in a Tunair flask, and grown at 30 °C for 20 h at 160 rpm.The cells were collected via centrifugation at 6000g for 10 min at 4 °C, and then resuspended in 12 mL of the resuspension buffer (50 mM Tris-HCl pH 7.5, 5 mM EDTA, 10% glycerol, 0.12 M sorbitol) with 1 × concentration of the complete protease inhibitor EDTA-free cocktail (Merck).For cell disruption, the cells were mixed with 15 mL of 0.5 mm glass beads and then disrupted using a vortex mixer that operated for 12 cycles (1 min of disruption and 8 min of incubation on ice).Undisrupted cells and cell debris were pelleted via centrifugation at 3000g for 30 min at 4 °C, and the supernatant, including the yeast membrane, was collected in microtubes.The membrane was pelleted via centrifugation at 100,000g for 30 min at 4 °C using a Himac CS100FNX Micro Ultracentrifuge (Eppendorf-Himac, Japan).The membrane pellet was resuspended in membrane buffer (50 mM Tris-HCl pH 7.5, 120 mM NaCl, 20% glycerol) with 1 × concentration of the complete protease inhibitor EDTA-free cocktail.The protein concentration of the membrane suspension was determined using the TaKaRa BCA Protein Assay Kit (Takara Bio, Japan).

In-gel fluorescence
The membrane suspensions of the wild type and each mutant were adjusted to a protein concentration of 5 mg/ mL.Fifteen microliters of the membrane suspension (5 mg/mL) was mixed with an equal volume of 2 × sample buffer (50 mM Tris-HCl pH 7.5, 5 mM EDTA, 5% β-mercaptoethanol, 5% glycerol, 4% SDS, 0.02% bromophenol blue, 2 × concentration of complete EDTA-free protease inhibitor cocktail), and 20 μL of the mixture was applied to the SDS-PAGE gel of the Tris-glycine buffer system without boiling the sample.One microliter of the Benchmark Fluorescent Protein standard (Thermo Fisher Scientific, USA) was applied to the gel as a molecular size marker.Electrophoresis was performed at 100 V and 4 °C.Fluorescence images were obtained using Typhoon FLA 7000 (GE Healthcare, USA).

Radioisotope (RI)-labeled ligand binding assay
The membrane suspension, 3 H-histamine (PerkinElmer, USA), and cold ligand (histamine) were diluted with the assay buffer (20 mM HEPES pH7.5, 150 mM NaCl).Briefly, 100 μL of the reaction mixture was dispensed in triplicate in a 96-well microplate.Membrane proteins (300 μg) were added to each well.The final concentrations of 3 H-histamine were 200, 100, 50, 25, 12.5, 6.25, and 3.13 nM.To achieve non-specific binding, the experiment was performed in the presence of a cold ligand at a 1000-fold concentration of the hot ligand.The reaction www.nature.com/scientificreports/mixture was incubated at room temperature (22-24 °C) for 1 h.The membrane was harvested on a glass fiber Filtermat B (PerkinElmer, USA) filter paper presoaked in 0.3% polyethylene imine using a FilterMate cell harvester (PerkinElmer, USA).The filter paper was washed with distilled water and dried.The solid scintillator, MeltiLex B/HS (PerkinElmer, USA), was melted on the filter paper.Radioactivity was detected using a Microbeta 2 system (PerkinElmer, USA).Specific binding was determined by subtracting nonspecific binding from total binding.

Results H 3 R expressed in S. cerevisiae does not possess signaling activity
In this study, we used the engineered yeast strain, YB1, which possesses a chimeric yeast Gα protein, in which five C-terminal residues of Gpa1 are replaced with human Gα i .If a heterologously expressed GPCR couples with a chimeric protein, the yeast pheromone signaling pathway is activated.Of note, YB1 lacked the HIS3 gene; therefore, it could not grow in histidine-depleted medium.When the expressed GPCR is activated by an agonist, the yeast pheromone signaling pathway is activated, which finally activates the FUS1-HIS3 gene, resulting in cell growth in a histidine-depleted medium.The degree of cell proliferation depends on the strength of the signal (i.e., the agonist concentration).
The full-length H 3 R (H 3 R_FL) used in this experiment had an N11Q mutation introduced to avoid glycosylation via post-translational modifications.Glycosylation is important for receptor activity in certain GPCRs.However, Peng et al. confirmed the signal transduction activity of the H 3 R variant by deleting residues 1-26 at the N-terminus 18 .Therefore, the N11Q mutation did not significantly affect activity.The YB1 strain was transformed with the expression vector, H 3 R_FL containing GFP fused to the C-terminus of the receptor.After the transformants were cultured in -Ura/SC medium, an increase in GFP fluorescence was observed (Fig. 1a).Ingel fluorescence 29 of the yeast membrane showed that bands were observed at a position approximately 32 kDa although the expected molecular mass of H 3 R_FL-GFP was approximately 78 kDa, which was considered to be due to receptor degradation (Fig. 1b).The band observed between 32 and 21 kDa, which was observed in all mutants, may be derived from a GFP released from the C-terminal of H 3 R by digestion.A ligand concentrationdependent growth assay using the natural agonist, histamine, was performed with the H 3 R_FL transformant; however, no growth was observed (data not shown).
As reported in a study that investigated muscarine M 1 , M 3 , and M 5 receptors, deletion of a long intracellular third loop (I3L) can improve expression and coupling efficiency in S. cerevisiae 30 .Therefore, we constructed an H 3 R_i3d variant, in which the long I3L (from residues 237 to 340) was deleted.The GFP fluorescence of the transformant of H 3 R_i3d was similar to that of H 3 R_FL (Fig. 1a).In the in-gel fluorescence, a band that appeared to be derived from H 3 R_i3d-GFP (molecular mass of 67 kDa) was observed between 63 and 40 kDa (Fig. 1b, indicated by an arrow).In many cases, when electrophoresis is performed without boiling, the band representing membrane proteins is approximately 20% smaller than that representing their actual molecular mass 31 .However, ligand concentration-dependent growth was not observed (WT; Fig. 3a).

Random mutagenesis and selection on the histamine plate
The selection strategy for the H 3 R mutants is shown in Fig. 2. The YB1 strain was transformed with the SmaIdigested p416GPD_GFP plasmid and H 3 R-i3d via random mutagenesis using error-prone PCR.Based on the number of colonies on the -Ura/SC plate medium, the transformation efficiency was approximately 5200 cfu/ μg per plasmid.After selection on -Ura-His/SC plate medium containing 10 mM 3-AT and 100 μM histamine, www.nature.com/scientificreports/24 colonies grew from approximately 1.1 × 10 5 transformants.After these colonies were cultured in -Ura/SC medium overnight, the culture was transferred to -Ura-His/SC medium containing 10 mM 3-AT and 100 μM histamine or the medium without histamine in 96-well deep well plates, and then cultured at 25 °C for 48 h.Seven independent clones (r1, r2, r3, r5, r6, r8, and r10) were finally identified, which grew in medium with histamine but not in medium without histamine.All the clones showed histamine concentration-dependent growth (Fig. 3a).The EC 50 values for histamine were 10-39 μM (Table 1).These clones contained one to four base substitutions in H 3 R_i3d (Table 2).The integrity and activity of the receptor protein expressed in YB1 were analyzed in activity-recovered clones.The YB1 strain transformed with H 3 R_i3d (wild-type) or the activity-recovered clones was cultured in -Ura/SC (containing histidine) medium at 25 °C for 18 h, and the fluorescence was measured.No significant difference was found between wild-type and each clone in terms of GFP fluorescence intensity (Fig. 3b).In-gel fluorescence  analysis of the cell membrane after SDS-PAGE revealed a band between 40 and 63 kDa for all mutants, similar to the wild-type H 3 R_i3d (Fig. 3c).A ligand-binding assay using 3 H-histamine was performed using a membrane fraction prepared from the wild-type and clone r8 (C415R mutant).The dissociation constants (K d s) could not be determined due to lower binding to the amount of membrane used; however, similar specific binding was confirmed for the wild-type and clone r8 (Fig. 3d).

Determination of mutations for the recovery of signaling activity
Of the obtained, r3 (L73M 2.43 ), r6 (F193S ECL2 ), and r8 (C415R 7.56 ) were single mutants that were important for the recovery of signal transduction activity.Furthermore, C415R 7.56 was confirmed in two clones, r2 and r8.We opened to focus on clones r1, r2, r5 and r10, in which multiple amino acid substitutions were introduced, and searched for mutations important for activity recovery (Table 2).The location of these mutations was confirmed using an H 3 R homology model registered in the GPCR database (GPCRdb) 32,33 (Fig. 4a).Ala39 1.37     www.nature.com/scientificreports/and Val407 7.48 pointed to the exterior of the receptor, and the intramolecular interactions seemed to be small.His187 ECL2 , present in the second extracellular loop, located far from a potential ligand-binding site.Thus, mutations at these positions did not seem to significantly alter the receptor function.Therefore, the following six mutations, A61T 1.59 (clone r5), V83I 2.53 (r10), L117S 3.35 (r5), M156I 4.46 (r10), S359Y 6.36 (r1), and F423L 8.54 (r10), were introduced into H3R_i3d alone, and a growth assay was performed.Sufficient growth was only observed with S359Y 6.36 , with low growth activity in the other mutations (Fig. 5).Therefore, S359Y 6.36 alone could recover signal activity, and a combination of mutations was needed for A61T 1.59 , L117S 3.35 , V83I 2.53 and F423L 8.54 to recover the signal activity of clones r5 and r10.Finally, we identified four mutations, L73M 2.43 , F193S ECL2 , S359Y 6.36 , and C415R 7.56 that markedly recovered signal transduction activity.Based on the recently elucidated crystal structure of H 3 R, three mutations, namely L73M 2.43 , S359Y 6.36 , and C415R 7.56 , were present at positions likely to be involved in the movement of TM6 during activation; these mutations were replaced with bulkier (M, Y and R) and/or positively charged (R) side chains (Fig. 4b).F193S ECL2 located in the second extracellular loop and faced to a ligand-binding pocket.

G protein-coupling specificity of mutants with recovered activity
H 3 R specifically couples to the G i heterotrimeric G protein 12 .To confirm the specificity of the activity-recovered mutants for the G protein, four single mutants (L73M 2.43 , F193S ECL2 , S359Y 6.36 , and C415R 7.56 ) were transformed into the YB11 strain, in which the five C-terminal residues of Gpa1 were replaced with those of human G q , and then transformed into the YB14 strain with human G s .Thereafter, the growth assay was performed.As positive controls, YB11 transformed with the histamine H 1 receptor and YB14 transformed with the adenosine A 2A receptor grew in an agonist concentration-dependent manner (Fig. 6a-d).YB11 and YB14 transformed with H 3 R mutants did not exhibit histamine-dependent growth (Fig. 6a-d).The expression of H 3 R mutants in YB11 and YB14 was confirmed by in-gel fluorescence after SDS-PAGE (Fig. 6e).

Evaluation of mutants with recovered activity using agonists and antagonists
We evaluated the function of the mutants with recovered activity expressed in YB1 using agonists and antagonists for H 3 R. Synthetic agonist (imetit) and antagonists/inverse agonists (A331440, BF2649, clobenpropit, iodophenpropit, and JNJ-5207852) were used.Imetit is a more potent agonist than histamine 34 .All clones showed agonist concentration-dependent growth of imetit, although wild-type H 3 R_i3d did not show an activity response (Fig. 7a).The maximum growth (OD 595 ) was approximately 0.2-0.7,which was generally lower than that of histamine (~ 0.7-1).In particular, the growth of clones r5, r6, and r10 was lower than that of the other clones.This result indicates that these mutations did not completely transmit the structural changes of the receptor to the G protein following imetit binding.The EC 50 values of the clones with recovered activity were approximately 1/15 those of histamine (Table 3).
Competitive inhibition studies using antagonists A331440, clobenpropit, iodophenpropit, and JNJ-5207852 were performed in the presence of 100 μM histamine (Fig. 7b).Clobenpropit and iodophenpropit are antagonists with imidazole groups, whereas A-331440 and JNJ-5207852 are antagonists without imidazole groups.Each ligand inhibited YB1 cell growth in a concentration-dependent manner by inhibiting histamine binding.The IC 50 values were markedly higher than the K i values listed in the International Union of Basic and Clinical Pharmacology (IUPHAR) database; however, the magnitude of the IC 50 was in the order of the K i values (JNJ5207852 < clobenpropit < iodephenpropit, A331440, and BF2649) (Table 4).The inverse agonist activity was examined under conditions of increased constitutive activity by lowering the 3-AT concentration.The inverse agonist (BF2649) and four antagonists (A331440, clobenpropit, iodophenpropit, JNJ-5207852) inhibited cell Figure 5. Agonist (histamine) concentration-dependent growth curve of the A61T 1.59 , V83I 2.53 , L117S 3.35 , M156I 4.46 , S359Y 6.36 , and F423L 8.54  In this study, we investigated the reactions of mutants with restored activity when expressed in yeast in the presence of various H 3 R ligands.In a ligand-dependent growth assay, the EC 50 of the synthetic agonist, imetit, was approximately 11-28 times lower than that of histamine.In addition, the maximal activation of imetit decreased to less than 70% of that of histamine.A calcium mobilization assay using HEK-293 cells also revealed a lower  4.For all experiments, at least three independent experiments were performed, and the representative graph is shown.Each experiment was performed in triplicate.The growth assay data were fitted to a Log (agonist) vs. response model.The inhibition assay data were fitted to a Log (inhibitor) vs. response model.www.nature.com/scientificreports/maximal activity of imetit than that of histamine 35 .Maximum imetit activity differed markedly depending on the clone used.Further detailed experiments are required to determine the cause of this phenomenon.According to experiments conducted using the C415R mutant as a representative, all the tested antagonists exhibited inverse agonist activity, as described in a previous review 36 .As described above, the EC 50 and IC 50 values were higher than those in animal cells; however, the order of relative H 3 R activity to various ligands was similar to that reported previously.Thus, mutations with recovered activity may retain the characteristics of H 3 R against its ligands.
Receptor stabilization is necessary for the large-scale preparation of proteins for the structural and physicochemical functional analyses of GPCRs.However, stabilizing GPCRs, especially in their active state, remains a challenge.A screening system using S. cerevisiae signal transduction would enable efficient production of thermostable mutants while retaining receptor activity.Rag23, a thermally stabilized A 2A R mutant, retains agonistinduced receptor activity in S. cerevisiae, and is thus an example of successful stabilization 7,37 .Recently, we succeeded in stabilizing A 2A R by combining statistical thermodynamics and molecular evolution 38 .A screening system that uses auxotrophy for selection, as used in this study, is effective for selecting receptor mutants from a vast library of mutants.When examining receptor activity, the time can be shortened by using an assay system with reporter genes, such as GFP or β-galactosidase.In this study, by using yeast, the EC 50 of histamine binding to the receptor with recovered activity (e.g., for clone r1) was approximately two orders of magnitude higher than that of binding to the wild-type H 3 R analyzed using animal cells 39 .The EC 50 may be further lowered by increasing sensitivity using tuned yeast strains 26 .Alternatively, sensitivity can be improved by increasing the expression of active receptors by adding a signal sequence or by stabilizing the receptors 40,41 .
In conclusion, random mutagenesis followed by in vivo selection in a histidine-deficient plate medium containing an agonist (histamine) enabled us to obtain mutants of H 3 R with recovered activity, in which the wildtype receptor had no signal transduction activity when expressed in yeast.These mutants retained H 3 R activity in terms of G protein specificity and ligand binding.The strategy for obtaining mutants with recovered activity may enable the recovery of the activity of other GPCRs that do not function in S. cerevisiae and may be useful in creating GPCRs mutants stabilized in their active conformations.However, further experiments are needed to elucidate why these mutations recovered the activity.Table 4. Inhibition profile of the H 3 R mutant, C415R, with antagonists/inverse agonists in the presence (100 μM) or absence of histamine.IC 50 values were determined from the inhibition assays shown in Fig. 7.The values are expressed as mean ± SD of three independent experiments.The K i values for human H 3 R were converted from the pK i in the IUPHAR database, (https:// www.guide topha rmaco logy.org).

Figure 1 .
Figure 1.Expression of full-length H 3 R (FL) and the I3L-deleted H 3 R (i3d) in the YB1 strain of S. cerevisiae.NC is the YB1 strain transformed with empty p416GPD plasmid as a negative control.(a) Whole-cell GFP fluorescence intensity.Average ± SEM of eight colonies.(b) In-gel fluorescence after SDS-PAGE.Arrow indicates the band derived from H 3 R_i3d-GFP observed between 63 and 40 kDa.M molecular size marker.

Figure 2 .
Figure 2. Schematic diagram of the screening experiment to obtain mutants with recovered activity.

Figure 3 .
Figure 3. Characterization of the H 3 R mutants with recovered activity.(a) Agonist (histamine) concentrationdependent growth assay of the H 3 R clones (r1-r10) with recovered activity.Data shown here are representative of three independent experiments.Each experiment was performed in triplicate.The growth assay data were fitted to a Log (agonist) vs. response model.The mean EC 50 ± SEM is shown in Table 1.(b) Whole-cell GFP fluorescence of the clones.Average ± SEM of eight colonies.(c) In-gel fluorescence after SDS-PAGE.Arrow indicates the band derived from H 3 R_i3d-GFP observed between 63 and 40 kDa.M; molecular size marker.(d) Ligand ( 3 H-histamine) binding assay using the membrane of the wild-type and C415R.Specific binding was plotted.Each experiment was performed in triplicate.

Figure 4 .
Figure 4. (a) Homology model of the H 3 R active-state built based on the H 1 R-G q complex structure (PDBID: 7DFL) with the GPCRdb homology modeling pipeline, version 2019-03-14.The model was obtained from GPCRdb.(b) Crystal structure of H 3 R (PDBID: 7F61) viewed from the intracellular side.The mutations found in the activity-recovered clones are shown in magenta.
Figure 5. Agonist (histamine) concentration-dependent growth curve of the A61T 1.59 , V83I 2.53 , L117S 3.35 , M156I 4.46 , S359Y 6.36 , and F423L 8.54 H 3 R mutants.All data shown here are representative of three independent experiments.Each experiment was performed in triplicate.The growth assay data were fitted to a Log (agonist) vs. response model.The average EC 50 of S359Y was 8.8 ± 2.3 [μM].

Figure 7 .
Figure 7. (a) Agonist (imetit) concentration-dependent growth assay of wild-type and H 3 R clones with recovered activity.The mean EC 50 ± SEM is shown in Table 3.(b) Antagonist/inverse agonist-concentration dependent inhibition curve of the C415R mutant in the presence of 100 μM histamine.The experiments related to the data in (a) and (b) were performed in the presence of 20 mM 3-AT.(c) Inverse agonist-concentration dependent inhibition curve of the C415R mutant in the absence of histamine.The experiments related to the data in (c) were performed in the presence of 10 mM 3-AT.The mean IC 50 ± SEM values of (b) and (c) are shown in Table4.For all experiments, at least three independent experiments were performed, and the representative graph is shown.Each experiment was performed in triplicate.The growth assay data were fitted to a Log (agonist) vs. response model.The inhibition assay data were fitted to a Log (inhibitor) vs. response model.

Table 1 .
Activation profile of the clones with recovered activity by the agonist histamine.EC 50 values were obtained from the growth assays.The values are expressed as mean ± SD of three independent experiments.

Table 2 .
Mutations found in the clones with recovered activity.The position (Ballesteros-Weinstein numbering) of the residue is shown as a superscript.

Table 3 .
Activation profile of clones with recovered activity by the agonist imetit.EC 50 values were obtained from the growth assays.The values are expressed as mean ± SD of three independent experiments.