Production of a Human Histamine Receptor for NMR Spectroscopy in Aqueous Solutions

G protein-coupled receptors (GPCRs) bind a broad array of extracellular molecules and transmit intracellular signals that initiate physiological responses. The signal transduction functions of GPCRs are inherently related to their structural plasticity, which can be experimentally observed by spectroscopic techniques. Nuclear magnetic resonance (NMR) spectroscopy in particular is an especially advantageous method to study the dynamic behavior of GPCRs. The success of NMR studies critically relies on the production of functional GPCRs containing stable-isotope labeled probes, which remains a challenging endeavor for most human GPCRs. We report a protocol for the production of the human histamine H1 receptor (H1R) in the methylotrophic yeast Pichia pastoris for NMR experiments. Systematic evaluation of multiple expression parameters resulted in a ten-fold increase in the yield of expressed H1R over initial efforts in defined media. The expressed receptor could be purified to homogeneity and was found to respond to the addition of known H1R ligands. Two-dimensional transverse relaxation-optimized spectroscopy (TROSY) NMR spectra of stable-isotope labeled H1R show well-dispersed and resolved signals consistent with a properly folded protein, and 19F-NMR data register a response of the protein to differences in efficacies of bound ligands.


Introduction
G protein-coupled receptors (GPCRs) initiate physiological processes by recognizing and binding a wide range of extracellular stimuli, including hormones, neurotransmitters and numerous other small molecules. Over one third of FDA-approved drugs target human GPCRs [1], and GPCRs remain top targets for drug development in both academic and industrial labs [1]. Histamine receptors are class A GPCRs and one of the earliest GPCR targets of rational drug development to treat illness. In humans, four histamine receptor subtypes regulate vital physiological processes including gastric acid secretion, smooth muscle relaxation, the release of neurotransmitters and neutrophil distribution [2]. The human H 1 receptor (H 1 R) subtype is expressed in several different tissue types including in the brain, vascular smooth muscle, airway, liver and in lymphocytes [3]. Activation of H 1 R elicits physiological processes that include modulation of circadian rhythms [4], responses to allergens, and proinflammatory cytokine production [5]. H 1 R stimulation has also been implicated in inflammation-associated tumorigenesis [6]. H 1 R activation by endogenous histamine is blocked by widely recognized drugs known as antihistamines to prevent or reduce allergic responses. The strong interest in identifying novel H 1 R drugs has led to the development of over 40 H 1 R antagonists approved for use as therapeutics [7,8] An H 1 R complex with the antagonist doxepin, a first-generation antihistamine drug, was among the earliest reported crystal structures of human GPCRs [9]. Molecular docking of antihistamines to the H 1 R structure provided a structural basis for the improved specificity of second-generation antihistamines, which formed interactions with amino acids

Transformation and Colony Screening
The H 1 R plasmid was linearized with PME1 and transformed into electrocompetent Bg12 Pichia pastoris cells. Immediately after electroporation, 1 mL of cold 1M sorbitol was added and the cells were incubated for 1 h at 30 • C and 250 rpm. Following incubation, an aliquot of the culture was plated on YPD agar plates containing 2% dextrose (w/v), 1.7% (w/v) yeast nitrogen base without amino acids, 0.5% (w/v) ammonium sulfate and 0.02% (w/v) biotin and incubated for 48 h at 30 • C.
Fifteen colonies were selected to screen for protein expression. An expression test was carried out by inoculating each colony in 4 mL BMGY media for 3 days and then transferring the cells to fresh BMMY media containing 0.5% methanol (w/v) with additional aliquots of methanol added every 12 h for 48 h total expression time. Individual transformants were then tested for protein expression using a Western blot protocol as described previously [22]. Glycerol stocks of transformants that showed higher levels of expression were prepared and stored at −80 • C for subsequent experiments.

Optimization of H 1 R Expression Conditions
To increase expression yields, induction conditions were varied and tested for relative amounts of expressed H 1 R by a Western blot assay. Small-scale cultures were grown in 4 mL of buffered minimal glycerol yeast (BMGY) media in culture tubes at 30 • C. After 48 h, the cultures were centrifuged and resuspended in 4 mL of buffered minimal methanol yeast (BMMY) media. Induction temperature, duration of protein induction, addition of a stabilizing ligand, and addition of DMSO prior to inducing expression were varied and Western blot analysis was done to determine highest yielding induction conditions. The most intense band was the expression conditions used for all future expression of H 1 R.

H 1 R Production and Purification for NMR Experiments
Several 15 mL culture tubes containing 4 mL buffered minimal glycerol media (BMGY) were inoculated with a glycerol stock of a selected transformant that was determined to have higher levels of protein expression and incubated at 30 • C and 250 rpm until reaching an optical density at 600 nm of 7 to 10. Then, 250 mL baffled flasks containing 50 mL BMGY were inoculated with the 4 mL cultures and grown at 30 • C for 2 days to an optical density of 15-20. Subsequently, 2.8 L baffled flasks containing 500 mL of BMGY were inoculated with the 50 mL cell cultures and grown at 30 • C for 3 days until reaching an optical density of 15-20. The cultures were then centrifuged at 3000× g for 15 min and resuspended in buffered minimal methanol media (BMMY) without methanol and returned to the incubator shaker. The temperature was reduced to 28 • C and after 4 h, doxepin was added to each culture to a final concentration of 20 µM. One hour later, methanol was added to the cultures to a final concentration of 0.5% w/v, and approximately every 12 h after additional aliquots of methanol were added until reaching a total induction time of 36 h. Twelve hours after the final methanol addition, the cells were harvested by centrifuging at 3000× g for 15 min and stored at −80 • C until future use. Uniformly 15 N-labeled H 1 R was expressed by adding 15 N ammonium sulfate to the BMGY and BMMY media.
Cell pellets containing H 1 R were resuspended in lysis buffer (50 mM sodium phosphate pH 7.0, 100 mM NaCl, 5% glycerol (w/v), and in-house-made protease inhibitor cocktail solution) and lysed with a cell disruptor (Pressure Biosciences) operating at 40,000 psi. Membranes were separated and collected by ultracentrifugation at 44,000 rpm (200,000× g), frozen in liquid nitrogen and stored at −80 • C for future use.

Fluorescence Thermal Shift Assays
Fluorescence thermal shift assays were carried out as described previously [26]. H 1 R was extracted from P. pastoris membranes and purified in aqueous buffer containing LMNG/CHS mixed micelles without addition of any ligands. For each thermal shift assay sample, 5 µg of H 1 R was added to the buffer (50mM HEPES pH 7.0, 150mM NaCl, 0.05% DDM/CHS). The fluorescent CPM dye was added to each sample at a final concentration of 10 µM. Ligands were dissolved in a minimum amount of DMSO and added to each sample, which were then incubated in the dark on ice for 30 min and then transferred to quartz cuvettes. Experiments were carried out with a Cary Eclipse spectrofluorometer using quartz cuvettes (Starna Cells, Inc., Atascadero, CA, USA) over a linear temperature range from 20 • C to 90 • C heated at a constant rate of 2 • C/min. The excitation wavelength was 387 nm, and the emission wavelength was 463 nm.
Thermal shift data were analyzed using the program Origin (OriginLab Corporation, Northampton, MA, USA). Raw data were fitted to a Boltzmann sigmoidal curve to determine the melting temperatures of each H 1 R sample.

NMR Spectroscopy
Two-dimensional [ 1 H, 15 N] correlation spectra were recorded with a solution contain-ing~300 µM u-15 N human H 1 R in 20 mM HEPES pH 7.0, 75 mM NaCl, 0.05% LMNG, 0.0025% CHS and 25µM doxepin. Two-dimensional [ 15 N, 1 H]-transverse relaxation-optimized spectroscopy (TROSY) [27] correlation spectra were measured at 800 MHz 1 H nutation frequency on a Bruker Avance III spectrometer running Topspin 3.6.2 and equipped with a 5-mm TXI cryoprobe. Experiments were measured at 42 • C. The sample temperature was calibrated using a standard sample (4% methanol in methanol-d4). The TROSY spectrum was recorded with acquisition periods of 80 ms in 1 H and 16 ms in 15 N, with a 1 s recycle delay for a total experimental time of about 22 h. NMR data were processed and analyzed in Topspin 3.5pl2 (Bruker Biospin, Billerica, MA, USA). Prior to Fourier transformation, the data matrices were zero filled to 1024 (t1) × 4096 (t2) complex points and multiplied by a Gaussian window function in the acquisition dimension and a 75 • -shifted sine bell window function in the indirect dimension. 19 F-NMR data were recorded with a solution containing 200 µM H 1 R in 20 mM HEPES pH 7.0, 75 mM NaCl, 0.05% LMNG, 0.0025% CHS and either 25 µM doxepin or 50 µM histamine. 19 F-NMR experiments were measured at 600 MHz 1 H nutation frequency on a Bruker Avance III HD spectrometer running Topspin 3.6.2 and equipped with a Bruker 5-mm BBFO probe. Experiments were measured at 7 • C, and the temperature was calibrated using a standard sample. 19 F data were acquired with an acquisition period of Biomolecules 2021, 11, 632 5 of 14 360 ms, 16k scans, and 0.3 s recycle delay for a total experimental time of about 3 h per experiment. NMR data were processed and analyzed in Topspin 3.5pl2 (Bruker Biospin, Billerica, MA, USA). Prior to Fourier transformation, the data were zero filled to 64k points and multiplied by an exponential window function with 30 Hz line broadening. NMR signals were referenced to an internal standard of trifluoro acetic acid (TFA), which has a chemical shift of −76.55 ppm relative to trichloro-fluoro-methane.

Design of an H 1 R Expression Vector for NMR Studies
Construction of a human H 1 R expression vector for NMR studies was adapted from the design of the vector used to determine the crystal structure of the H 1 R complex with doxepin [9,28]. This H 1 R vector includes several modifications from the native receptor sequence: truncation of the first 19 amino acids, including two glycosylation sites, and replacement of the third intracellular loop (ICL3) with T4-lysozyme (T4L) to facilitate crystallization. This H 1 R vector expressed in Pichia pastoris was demonstrated to bind antagonists and agonists, including the endogenous ligand histamine, with affinities that were highly similar to native H 1 R expressed in insect cells and mammalian cells [9,28]. To preserve the ligand binding function, we retained all H 1 R endogenous amino acids present in the vector used for crystallization. For NMR studies, we removed the non-endogenous T4L, as the presence of T4L in ICL3 has been demonstrated to alter that native receptor function-related dynamics [29]. In our new vector, residues numbered 223 5.71 to 229 and residues numbered 399 to 404 6.24 , comprising the intracellular ends of helices V and VI, respectively, were reinserted into the vector (superscripts indicate amino acid positions using the Ballosteros-Weinstein nomenclature [30]). The resulting amino acid sequence is shown in Figure 1.
bell window function in the indirect dimension. 19 F-NMR data were recorded with a solution containing 200 µM H1R in 20 mM HEPES pH 7.0, 75 mM NaCl, 0.05% LMNG, 0.0025% CHS and either 25 µM doxepin or 50 µM histamine. 19 F-NMR experiments were measured at 600 MHz 1 H nutation frequency on a Bruker Avance III HD spectrometer running Topspin 3.6.2 and equipped with a Bruker 5-mm BBFO probe. Experiments were measured at 7 °C, and the temperature was calibrated using a standard sample. 19 F data were acquired with an acquisition period of 360 ms, 16k scans, and 0.3 s recycle delay for a total experimental time of about 3 h per experiment. NMR data were processed and analyzed in Topspin 3.5pl2 (Bruker Biospin, Billerica MA). Prior to Fourier transformation, the data were zero filled to 64k points and multiplied by an exponential window function with 30 Hz line broadening. NMR signals were referenced to an internal standard of trifluoro acetic acid (TFA), which has a chemical shift of −76.55 ppm relative to trichloro-fluoro-methane.

Design of an H1R Expression Vector for NMR Studies
Construction of a human H1R expression vector for NMR studies was adapted from the design of the vector used to determine the crystal structure of the H1R complex with doxepin [9,28]. This H1R vector includes several modifications from the native receptor sequence: truncation of the first 19 amino acids, including two glycosylation sites, and replacement of the third intracellular loop (ICL3) with T4-lysozyme (T4L) to facilitate crystallization. This H1R vector expressed in Pichia pastoris was demonstrated to bind antagonists and agonists, including the endogenous ligand histamine, with affinities that were highly similar to native H1R expressed in insect cells and mammalian cells [9,28]. To preserve the ligand binding function, we retained all H1R endogenous amino acids present in the vector used for crystallization. For NMR studies, we removed the non-endogenous T4L, as the presence of T4L in ICL3 has been demonstrated to alter that native receptor function-related dynamics [29]. In our new vector, residues numbered 223 5.71 to 229 and residues numbered 399 to 404 6.24 , comprising the intracellular ends of helices V and VI, respectively, were reinserted into the vector (superscripts indicate amino acid positions using the Ballosteros-Weinstein nomenclature [30]). The resulting amino acid sequence is shown in Figure 1.  (a) Snake plot of the H 1 R receptor employed in the current study, as generated from the GPCRdb [31]. The red dashed line represents residues from the third intracellular loop that have been truncated from the native sequence. The "tag" labels at the N-terminus and C-terminus represent a FLAG tag and polyhistidine tag, respectively; (b) amino acid sequence of the human H 1 R employed in this study.

Optimization of H 1 R Expression and Purification Conditions
The H 1 R gene was placed into the open reading frame of a pPIC9K vector and transformed into electrocompetent P. pastoris Bg12 cells. Transformants containing a higher copy number of the H 1 R vector were screened in 4 mL cultures, and transformants showing higher expression of H 1 R relative to other transformants in a Western blot assay were selected for initial expression efforts. Expression of one of these transformants in standard buffered minimal glycerol (BMGY) and buffered minimal methanol (BMMY) media resulted in a yield of less than 0.25 mg of purified protein pure liter cell culture, as estimated from a Bradford assay.
To increase the yield of expressed H 1 R, multiple expression parameters were explored in 4 mL cell cultures, including induction temperature, duration of protein induction, addition of H 1 R ligands to the cell culture, and addition of DMSO to the cell culture. Three different temperatures of induction in BMMY media were explored: 22, 28, and 30 • C. Methanol feeding was varied from 24 up to 48 h. The addition of 1% (w/v) and 2% (w/v) of DMSO was also studied. The relative amounts of expressed proteins were analyzed via Western blot ( Figure 2 and Table 1).
loop that have been truncated from the native sequence. The "tag" labels at the N-terminus and C-terminus represent a FLAG tag and polyhistidine tag, respectively; (b) amino acid sequence of the human H1R employed in this study.

Optimization of H1R Expression and Purification Conditions
The H1R gene was placed into the open reading frame of a pPIC9K vector and transformed into electrocompetent P. pastoris Bg12 cells. Transformants containing a higher copy number of the H1R vector were screened in 4 mL cultures, and transformants showing higher expression of H1R relative to other transformants in a Western blot assay were selected for initial expression efforts. Expression of one of these transformants in standard buffered minimal glycerol (BMGY) and buffered minimal methanol (BMMY) media resulted in a yield of less than 0.25 mg of purified protein pure liter cell culture, as estimated from a Bradford assay.
To increase the yield of expressed H1R, multiple expression parameters were explored in 4 mL cell cultures, including induction temperature, duration of protein induction, addition of H1R ligands to the cell culture, and addition of DMSO to the cell culture. Three different temperatures of induction in BMMY media were explored: 22, 28, and 30 °C. Methanol feeding was varied from 24 up to 48 h. The addition of 1% (w/v) and 2% (w/v) of DMSO was also studied. The relative amounts of expressed proteins were analyzed via Western blot ( Figure 2 and Table 1).

Figure 2.
Optimization of H1R protein production from a Western blot analysis of protein expression conditions. Each lane represents a distinct set of experimental conditions used to express H1R in 4 mL cell culture (see Table 1). The intensity of a protein band appearing just above the 30 kDa molecular weight marker was used to identify optimal expression conditions. Cell culture conditions tested in lanes 6 and 7 were found to be optimal, and conditions in lane 6 were used for subsequent large-scale protein production.   Table 1). The intensity of a protein band appearing just above the 30 kDa molecular weight marker was used to identify optimal expression conditions. Cell culture conditions tested in lanes 6 and 7 were found to be optimal, and conditions in lane 6 were used for subsequent large-scale protein production. Most of the tested parameters resulted in no to modest increases in the amount of expressed H 1 R. The addition of ligand showed the most pronounced effects. The presence of either 20 or 100 µM doxepin in the cell culture resulted in an approximate 10-fold increase in the yield of expressed H 1 R (Figure 2 and Table 1). The addition of 20 µM doxepin to Biomolecules 2021, 11, 632 7 of 14 BMMY media was thus implemented for all subsequent H 1 R production efforts. Production of H 1 R with the optimized expression conditions resulted in a yield of approximately 2 to 2.5 mg purified H 1 R per liter cell culture.
H 1 R was isolated from P. pastoris membranes following a protocol adapted from the production and purification of human A 2A AR from P. pastoris [20,22]. Purification of H 1 R in aqueous solutions containing doxepin and mixed micelles of either n-dodecylb-D-maltoside (DDM) and cholesterol hemisuccinate (CHS) or lauryl maltose neopentyl glycol (LMNG) and CHS resulted in a homogenous preparation. Analytical size exclusion chromatography of the purified H 1 R product showed a monodispersed peak with an estimated molecular weight above 50 kDa and below 150 kDa (Figure 3), which was similar to the observed elution times of samples of A 2A AR prepared in LMNG/CHS mixed micelles. 9 28 °C 36 h 0 2% Most of the tested parameters resulted in no to modest increases in the amount of expressed H1R. The addition of ligand showed the most pronounced effects. The presence of either 20 or 100 µM doxepin in the cell culture resulted in an approximate 10-fold increase in the yield of expressed H1R (Figure 2 and Table 1). The addition of 20 µM doxepin to BMMY media was thus implemented for all subsequent H1R production efforts. Production of H1R with the optimized expression conditions resulted in a yield of approximately 2 to 2.5 mg purified H1R per liter cell culture.
H1R was isolated from P. pastoris membranes following a protocol adapted from the production and purification of human A2AAR from P. pastoris [20,22]. Purification of H1R in aqueous solutions containing doxepin and mixed micelles of either n-dodecyl-b-Dmaltoside (DDM) and cholesterol hemisuccinate (CHS) or lauryl maltose neopentyl glycol (LMNG) and CHS resulted in a homogenous preparation. Analytical size exclusion chromatography of the purified H1R product showed a monodispersed peak with an estimated molecular weight above 50 kDa and below 150 kDa (Figure 3), which was similar to the observed elution times of samples of A2AAR prepared in LMNG/CHS mixed micelles.

H1R Response to Ligand Binding
To characterize the ability of purified H1R to bind ligands, we employed a fluorescence thermal shift assay using the thiol-specific fluorophore N-[4-(7-diethylamino-4-methyl-3-coumarinyl)phenyl]maleimide (CPM). This fluorophore interacts with cysteines present in the protein hydrophobic core as they become exposed to aqueous solvent due to thermal unfolding of the protein and has been used to characterize GPCR-ligand interactions [26]. To record the fluorescence thermal shift experiments, H1R was purified in an aqueous solution containing LMNG/CHS and without any ligand added. Several ligands were then individually mixed with aliquots of purified H1R and incubated for half an hour

H 1 R Response to Ligand Binding
To characterize the ability of purified H 1 R to bind ligands, we employed a fluorescence thermal shift assay using the thiol-specific fluorophore N-[4-(7-diethylamino-4-methyl-3coumarinyl)phenyl]maleimide (CPM). This fluorophore interacts with cysteines present in the protein hydrophobic core as they become exposed to aqueous solvent due to thermal unfolding of the protein and has been used to characterize GPCR-ligand interactions [26]. To record the fluorescence thermal shift experiments, H 1 R was purified in an aqueous solution containing LMNG/CHS and without any ligand added. Several ligands were then individually mixed with aliquots of purified H 1 R and incubated for half an hour on ice prior to recording the temperature ramp experiments. To determine the melting temperature (T M ) of each sample, the fluorescence data were fitted to a Boltzmann sigmoidal curve and the T m was determined at the inflection point of the curve. All samples with ligands added showed moderate (3-4 • C) to more significant (8 • C) increases in the calculated T M values over the sample with no ligand added (Figure 4), indicating that the purified protein recognized and could bind ligands. The T M for H 1 R with no ligand added was determined to be 71.0 • C. The H 1 R sample with histamine added showed the highest determined T m at 77.2 • C, and the curve showed a clear sigmoidal shape, consistent with cooperative unfolding. on ice prior to recording the temperature ramp experiments. To determine the melting temperature (TM) of each sample, the fluorescence data were fitted to a Boltzmann sigmoidal curve and the Tm was determined at the inflection point of the curve. All samples with ligands added showed moderate (3-4 °C) to more significant (8 °C) increases in the calculated TM values over the sample with no ligand added (Figure 4), indicating that the purified protein recognized and could bind ligands. The TM for H1R with no ligand added was determined to be 71.0 °C. The H1R sample with histamine added showed the highest determined Tm at 77.2 °C, and the curve showed a clear sigmoidal shape, consistent with cooperative unfolding.

Spectroscopic Characterization of Solutions Containing Purified H1R by 2-Dimensional TROSY NMR
Production of H1R for NMR studies followed the optimized expression and purification conditions described above. To assess the foldedness of purified H1R in solutions containing LMNG/CHS mixed micelles, we recorded [ 15 N, 1 H]-TROSY two-dimensional spectra of u-15 N H1R in complex with the ligands doxepin and histamine (Figure 5a,b). A TROSY spectrum of u-15 N H1R in complex with doxepin recorded at 42 °C shows approximately 100 to 130 signals that are well dispersed from 6.5 to 10.5 ppm along the 1 H chemical shift axis. About 20 to 25 signals of greater intensity are narrowly dispersed between 7.5 and 8.5 ppm along the 1 H axis, which likely arise from more flexible loop regions of the protein. The approximately 25-30 well-resolved signals observed above 8.5 ppm are likely from amide groups in regular secondary structure and indicate that the protein is folded in conditions used for our NMR studies. A TROSY spectrum of u-15 N H1R in complex with the agonist histamine shows fewer resolved signals than the complex with doxepin, reminiscent of earlier comparisons of TROSY spectra for antagonist and agonist complexes of human A2AAR [22]; however, a number signals are observed above 8.5 ppm and below 7.5 ppm, consistent with the thermal shift assays that indicate that the complex with histamine is folded.

Spectroscopic Characterization of Solutions Containing Purified H 1 R by 2-Dimensional TROSY NMR
Production of H 1 R for NMR studies followed the optimized expression and purification conditions described above. To assess the foldedness of purified H 1 R in solutions containing LMNG/CHS mixed micelles, we recorded [ 15 N, 1 H]-TROSY two-dimensional spectra of u-15 N H 1 R in complex with the ligands doxepin and histamine (Figure 5a,b). A TROSY spectrum of u-15 N H 1 R in complex with doxepin recorded at 42 • C shows approximately 100 to 130 signals that are well dispersed from 6.5 to 10.5 ppm along the 1 H chemical shift axis. About 20 to 25 signals of greater intensity are narrowly dispersed between 7.5 and 8.5 ppm along the 1 H axis, which likely arise from more flexible loop regions of the protein. The approximately 25-30 well-resolved signals observed above 8.5 ppm are likely from amide groups in regular secondary structure and indicate that the protein is folded in conditions used for our NMR studies. A TROSY spectrum of u-15 N H 1 R in complex with the agonist histamine shows fewer resolved signals than the complex with doxepin, reminiscent of earlier comparisons of TROSY spectra for antagonist and agonist complexes of human A 2A AR [22]; however, a number signals are observed above 8.5 ppm and below 7.5 ppm, consistent with the thermal shift assays that indicate that the complex with histamine is folded.
To provide further interpretation of the NMR data, we compared the TROSY spectrum with the crystal structure of the H 1 R complex with doxepin and more closely examined signals from the tryptophan indole 15 N-1 H region (Figures 5 and 6a). There are four resolved signals of greater intensity in the tryptophan indole 15 N-1 H region of the TROSY spectrum and approximately five to six additional signals that are less intense but resolved. The crystal structure of the H 1 R complex with doxepin shows 9 tryptophans located in the transmembrane helices and extracellular loops (Figure 6b). Thus the number of observed signals in the Trp indole 15 N-1 H region of the TROSY spectrum is similar to the number of expected signals from all tryptophans. Comparing the spectra of the agonist and an-tagonist complexes shows some initial differences, for example in the region of the NMR spectra where glycine and threonine signals are typically observed (Figure 7). These initial spectra are highly encouraging considering that the protein is not deuterated, and thus we anticipate additional improvement in resolution upon incorporation of deuterium in the protein.  Figure 6. Both spectra were recorded with ~300 µM H1R at a 1 H NMR frequency of 800 MHz and experimental temperature of 42 °C for a total acquisition time of ~22 h per experiment. To provide further interpretation of the NMR data, we compared the TROSY spectrum with the crystal structure of the H1R complex with doxepin and more closely examined signals from the tryptophan indole 15 N-1 H region (Figures 5 and 6a). There are four resolved signals of greater intensity in the tryptophan indole 15 N-1 H region of the TROSY spectrum and approximately five to six additional signals that are less intense but resolved. The crystal structure of the H1R complex with doxepin shows 9 tryptophans located in the transmembrane helices and extracellular loops (Figure 6b). Thus the number of observed signals in the Trp indole 15 N-1 H region of the TROSY spectrum is similar to the number of expected signals from all tryptophans. Comparing the spectra of the agonist and antagonist complexes shows some initial differences, for example in the region of the NMR spectra where glycine and threonine signals are typically observed (Figure 7). These initial spectra are highly encouraging considering that the protein is not deuterated, and thus we anticipate additional improvement in resolution upon incorporation of deuterium in the protein.    Figure 5a. (b) The crystal structure of the H1R complex with doxepin is shown in ribbon representation (PDB ID 3RZE [9]), and the 9 tryptophans in H1R are shown in red in stick representation. The blue dashed line represents missing electron density in the crystal structure for amino acids 168 to 174, and the black dotted line represents the location of T4L in the crystal structure. The bound doxepin ligand is shown in brown in stick representation.
To provide further interpretation of the NMR data, we compared the TROSY spectrum with the crystal structure of the H1R complex with doxepin and more closely examined signals from the tryptophan indole 15 N-1 H region (Figures 5 and 6a). There are four resolved signals of greater intensity in the tryptophan indole 15 N-1 H region of the TROSY spectrum and approximately five to six additional signals that are less intense but resolved. The crystal structure of the H1R complex with doxepin shows 9 tryptophans located in the transmembrane helices and extracellular loops (Figure 6b). Thus the number of observed signals in the Trp indole 15 N-1 H region of the TROSY spectrum is similar to the number of expected signals from all tryptophans. Comparing the spectra of the agonist and antagonist complexes shows some initial differences, for example in the region of the NMR spectra where glycine and threonine signals are typically observed (Figure 7). These initial spectra are highly encouraging considering that the protein is not deuterated, and thus we anticipate additional improvement in resolution upon incorporation of deuterium in the protein.

Response to Changes in Efficacy of Bound Drugs Monitored by 19 F-NMR Spectroscopy
We assessed the potential for labeling with 19 F-trifluoroethanethiol (TET) and whether we could detect a response to changes in the efficacy of bound ligands by preparing samples of H 1 R complexes with the antagonist doxepin and the endogenous agonist histamine. Based on analysis of the crystal structure of the H 1 R complex with doxepin, three endogenous cysteines are predicted to be more solvent accessible and thus available for 19 F labeling via chemical modification with trifluoroethanethiol (TET). These cysteines are C221 5.69 , located at the intracellular end of helix V, cysteine 441 6.61 , located at the extracellular end of helix VI, and C471 7.56 , located at the intracellular end of helix VII (Figure 8a). While the TROSY experiments were recorded at 42 • C, the 19 F NMR data were measured at 7 • C so that we could assess the data in the larger context of many previously reported 19 F NMR GPCR studies [21,25,29,32,33]. Both 19 F spectra of the complexes with doxepin and histamine show complicated line shapes containing more than three individual components (Figure 8b), suggesting the presence of multiple, simultaneously observed conformers for one or more of the labeled cysteines. Comparison of the 19 F data between the complex with doxepin and the complex with histamine shows that individual components appear to have similar chemical shifts, for example the components near 9.0 and 9.8 ppm, with changes observed in the relative intensities of the individual components. This suggests that changes in relative populations of different conformers are potentially observed for complexes with different efficacies of bound ligands and that further 19 F-NMR experiments with H 1 R variants containing single labeled cysteines may be used to characterize these different populations.
for 19 F labeling via chemical modification with trifluoroethanethiol (TET). These cysteines are C221 5.69 , located at the intracellular end of helix V, cysteine 441 6.61 , located at the extracellular end of helix VI, and C471 7.56 , located at the intracellular end of helix VII ( Figure  8a). While the TROSY experiments were recorded at 42 °C, the 19 F NMR data were measured at 7 °C so that we could assess the data in the larger context of many previously reported 19 F NMR GPCR studies [21,25,29,32,33]. Both 19 F spectra of the complexes with doxepin and histamine show complicated line shapes containing more than three individual components (Figure 8b), suggesting the presence of multiple, simultaneously observed conformers for one or more of the labeled cysteines. Comparison of the 19 F data between the complex with doxepin and the complex with histamine shows that individual components appear to have similar chemical shifts, for example the components near 9.0 and 9.8 ppm, with changes observed in the relative intensities of the individual components. This suggests that changes in relative populations of different conformers are potentially observed for complexes with different efficacies of bound ligands and that further 19 F-NMR experiments with H1R variants containing single labeled cysteines may be used to characterize these different populations.

Production of Stable-Isotope Labeled Human GPCRs in Pichia Pastoris
Production of human GPCRs for structure determination by crystallography or cryo-EM has been achieved mostly using insect cell expression [34], with some full-length class B and class F receptors expressed in mammalian cells. Though P. pastoris has been used to express a much smaller fraction of GPCRs for structure determination so far, the benefits of working with P. pastoris have motivated several studies of the potential to express functional GPCRs in this system. Surveys of the expression of about 30 different human GPCRs

Production of Stable-Isotope Labeled Human GPCRs in Pichia Pastoris
Production of human GPCRs for structure determination by crystallography or cryo-EM has been achieved mostly using insect cell expression [34], with some full-length class B and class F receptors expressed in mammalian cells. Though P. pastoris has been used to express a much smaller fraction of GPCRs for structure determination so far, the benefits of working with P. pastoris have motivated several studies of the potential to express functional GPCRs in this system. Surveys of the expression of about 30 different human GPCRs demonstrated that a sizeable fraction of functional receptors could be expressed at milligram to multi-milligram scales [13,35]. Systematic evaluation of expression conditions also demonstrated that yields of more poorly expressed receptors could be significantly improved with optimized protocols [11].
P. pastoris offers multiple advantages for the expression of human membrane proteins for NMR studies, which include the ability to produce functional proteins with high levels of deuterium incorporation [36]. A growing number of membrane proteins have been expressed in P. pastoris for NMR experiments in both aqueous solutions and solid state samples, including human aquaporin 1 [37], human aquaporin 2 [38], and fungal rhodopsin from Leptosphaeria maculans [39]. Initial NMR data with the human GPCRs CB1 and OX2 also appear to be promising [24]. The NMR data presented here demonstrate that functional human H 1 R can be produced at sufficient quantities for NMR studies and establish an important step in the application of NMR to study the energy landscape of additional human GPCRs.

Tryptophans as NMR Probes of GPCR Structure-Function Relationships
Both endogenous and extrinsically introduced tryptohans are sensitive reporters of function-related changes in the GPCR structure, particularly when nearby other aromatic amino acids that create significant ring current shifts [40]. Further, tryptophan indole 15 N-1 H signals are typically well-separated from the majority of amide signals in [ 15 N, 1 H] heteronuclear two-dimensional experiments, which facilitates their assignment by amino acid replacement. We explored the potential use of tryptophans in future NMR studies of H 1 R structure-function relationships by calculating the ring current shifts for H 1 R tryptophan indole 15 N-1 H signals from the crystal structure of the H 1 R complex with doxepin using the program Molmol [41]. The calculated ring current shifts show a relatively narrow range spanning approximately 0.3 ppm, with the notable exception of Trp 428 6.48 (Table 2; superscripted text denotes the Ballosteros-Weinstein nomenclature used to report GPCR amino acid positions relative to a consensus sequence [30]). The calculated ring current shift for Trp 428 6.48 shows a pronounced shift of −0.56 and −0.78 ppm for the Hε1 and Nε1 chemical shifts, respectively. Comparably large ring current shifts were also observed for the corresponding Trp246 6.48 indole 15 N-1 H lines in TROSY NMR spectra of human A 2A AR [19,22]. The single largest contribution to these ring current shifts was the nearby Phe242 6.44 of the highly conserved P-I-F activation motif [19,22], and the chemical shifts of Trp246 6.48 were found to be highly sensitive to the efficacy of bound drugs. Thus, we anticipate that Trp 428 6.48 in H 1 R will be a sensitive probe of structure-function relationships.

Conclusions
We present a protocol for the expression and purification of the human H 1 histamine receptor in Pichia pastoris, which can be applied to economically produce stable-isotope labeled GPCR samples for NMR experiments. Evaluation of protein production parameters revealed that addition of ligand during protein expression resulted in a ten-fold increase in the yield of expressed receptor. Purification of expressed H 1 R in buffer containing either DDM/CHS mixed micelles or LMNG/CHS mixed micelles yielded a monodispersed and homogeneous preparation. Purified H 1 R responded to the addition of known ligands, as monitored by fluorescence thermal shift assays, with the addition of the ligand histamine resulting in the largest increase in the protein melting temperature. Two-dimensional [ 15 N, 1 H]-TROSY NMR spectra of u-15 N H 1 R in buffer containing LMNG/CHS mixed micelles show well-dispersed signals consistent with a folded protein and comparison of TROSY spectra of H 1 R complexes with an antagonist and an agonist show changes in the distribution of signals. Evaluation of the tryptophan indole 15 N-1 H region of the TROSY correlation spectrum of the antagonist complex shows the presence of a similar number of signals as expected from the crystal structure of H 1 R. 19 F-NMR spectra of H 1 R show differences related to changes in the efficacy of bound drugs. Expression of stable-isotope labeled human receptors is a bottleneck in the application of NMR spectroscopy to map the energy landscapes for most human GPCRs. While P. pastoris is a promising host for GPCR expression, only a few receptors have been expressed successfully in P. pastoris for NMR studies. This work demonstrates an additional example of a human GPCR that can be prepared from P. pastoris for NMR experiments.