In Search of Small Molecules That Selectively Inhibit MBOAT4

Ghrelin is a 28-residue peptide hormone produced by stomach P/D1 cells located in oxyntic glands of the fundus mucosa. Post-translational octanoylation of its Ser-3 residue, catalyzed by MBOAT4 (aka ghrelin O-acyl transferase (GOAT)), is essential for the binding of the hormone to its receptor in target tissues. Physiological roles of acyl ghrelin include the regulation of food intake, growth hormone secretion from the pituitary, and inhibition of insulin secretion from the pancreas. Here, we describe a medicinal chemistry campaign that led to the identification of small lipopeptidomimetics that inhibit GOAT in vitro. These molecules compete directly for substrate binding. We further describe the synthesis of heterocyclic inhibitors that compete at the acyl coenzyme A binding site.


Introduction
Ghrelin is a 28-residue lipopeptide ( Figure 1A) discovered by Kojima and co-workers as the endogenous ligand for the growth hormone secretagogue receptor (GHS-r) [1]. It was found that the octanoylation of Ser3 was required for ghrelin's endocrine activities. At the time, the enzyme responsible for this modification was unknown. Later, Yang hypothesized this atypical lipidation was likely performed by a member of the membrane-bound O-acyl transferase (MBOAT) family of enzymes [2]. The MBOATs catalyze the lipidation of a variety of substrates including phospholipids, neutral lipids, and proteins using saturated and unsaturated acyl coenzyme-A derivatives as acyl donors [3]. In 2008, Yang et al. utilized a candidate cloning approach to demonstrate that MBOAT 4, now termed ghrelin O-acyl transferase (GOAT), the only MBOAT capable of catalyzing proghrelin octanoylation [2]. Gutierrez used a candidate gene silencing approach to independently reach the same conclusion for human GOAT [4]. Short-interfering RNAs (siRNAs) were produced and the effects of silencing individual MBOAT genes were observed. MS/MS fragmentation analyses of GOAT acylated ghrelin showed that the acylation is on serine-3, identical to acyl ghrelin produced in the stomach.
Active ghrelin (hereafter referred to as ghrelin) was shown to induce adiposity in rodents and food intake in both rodents and humans [5,6]. Ghrelin was also found to play a role in glucose homeostasis by inhibiting insulin secretion from the pancreas [7,8]. Levels of active ghrelin in the blood peak during fasting and decrease after meals [9,10]. Ghrelin is primarily produced by peripheral tissues, rather than the central nervous system [11]. In addition, no other hormone is known to be octanoylated [12]. Given the biological specificity of GOAT and ghrelin, it was thought that, unlike GHS-r ligands, an inhibitor of GOAT would not need to penetrate the brain to cause pharmacological effects. The  [24] using DYNAFIT software [25]. The Top was defined as reaction conditions without inhibitor and the Bottom was defined as reaction conditions with full inhibition by 5 µM concentration of 2. Top and Bottom conditions were included in every experiment.
Peptides chains longer than three amino acids are generally not taken up into cells passively. However, their cyclic counterparts can show an improved stability, pharmacokinetics and, if their structure facilitates internal shielding of the polar surface area, useful levels of cell permeability [26]. We used Schrödinger's Macromodel software to conduct molecular modeling studies, using AMBER in a continuum solvation model (octanol) to inform the design of hypothetical macrocyclic analogs of compound 8, which was our best performing tetrapeptide. During these studies, we were pleased to find that the peptide backbone of the macrocyclic analog 9b overlaid almost perfectly with that of 8 ( Figure   Figure 2. (A) Ghrelin truncation peptide series and respective IC 50 values. Enzyme activity was quantified using the method described by Yang et al. [18] (see Materials and Methods); (B) Macrocycles derived from the parent peptides and respective IC 50 values; (C) Structural overlay of 9b (orange) against 8 (green), octanamide chains are indicated as line bonds for clarity; (D) In vitro inhibition of proghrelin octanoylation mediated by GOAT using the shown compounds. The experimental data were fit to a 4-parameter logistic model, (Bottom − Top)/(1 + (I/IC 50 ) n ) + Top [24] using DYNAFIT software [25]. The Top was defined as reaction conditions without inhibitor and the Bottom was defined as reaction conditions with full inhibition by 5 µM concentration of 2. Top and Bottom conditions were included in every experiment.
Peptides chains longer than three amino acids are generally not taken up into cells passively. However, their cyclic counterparts can show an improved stability, pharmacokinetics and, if their structure facilitates internal shielding of the polar surface area, useful levels of cell permeability [26]. We used Schrödinger's Macromodel software to conduct molecular modeling studies, using AMBER in a continuum solvation model (octanol) to inform the design of hypothetical macrocyclic analogs of compound 8, which was our best performing tetrapeptide. During these studies, we were pleased to find that the peptide backbone of the macrocyclic analog 9b overlaid almost perfectly with that of 8 ( Figure 2C). The root mean square deviation (RMSD) between the two structures was 0.06 Å at the α-carbon atoms for the lowest energy conformations of both molecules. We ignored variations of the octanamide chain in these analyses as this group was expected to be highly flexible. Based on our observations, we synthesized cyclic variants 9 and 17 using catalyzed ring-closing metatheses of bis-allyl ether precursors, (Figure 2), resulting in mixtures of separable alkene geometric isomers. The ability of these compounds to inhibit GOAT activity in vitro was comparable to their linear peptide counterparts. Moreover, enzymatic reaction rates measured while varying substrate concentrations in the presence of a fixed concentration of 17a suggested that the compound inhibited GOAT by competing for the peptide substrate (see Materials and Methods). With both linear and cyclic inhibitors in hand, we began an extensive program to synthesize and assay analogs, wherein the octanamide side chain, the P4 aromatic moiety, N-methylation patterns, ring substituents and connectivity (for 9/17) and stereochemistry were varied in search of molecules that inhibited GOAT potently in vitro, which could block acyl ghrelin secretion from an engineered insulinoma cell line (INS-1, see Materials and Methods). Over time hundreds of compounds were evaluated [27,28]. While several of those components inhibited GOAT activity in membrane fractions in the 50-200 nM range, none were superior to prototypic peptide 8 and none, including 8, showed significant activity in cell culture. Actually, the best performing oligomer in cell culture was lipopentapeptide 13 ( Figure 2). Unfortunately, that molecule showed poor pharmacokinetics and a low oral bioavailability in mice.
During this time, pharmaceutical companies began publishing patent literature on GOAT inhibitors discovered through high-throughput screening. Takeda reported data indicating a series of benzoxazole carboxylates (including 7 in Figure 1D), which inhibited the enzyme by competing at its co-enzyme A binding site [21]. Based on their structures, we believed it likely the aminopyrimidines reported by Eli Lilly [20] and the oxadiazolopyridines described by Boehringer Ingelheim [22] functioned similarly. Because numerous metabolic enzymes utilize co-enzyme A derivatives for catalysis, there was the risk of the unanticipated off-target effects for inhibitors of this kind. Nonetheless, Eli Lilly reported that compound 5 was orally bioavailable and inhibited ghrelin production in vivo at doses that did not cause observable adverse side effects.
We sought a hybrid molecule that would contain an oxadiazolopyridine linked to an octanoyl motif through the piperidinyl ethyl scaffold used in Eli Lilly compound 5 (and also a feature of certain ACAT inhibitors) [29]. We first synthesized compounds 5 and 6 in house and tested them in our own assays. We found the membrane fraction IC 50 = 88 nM and 64 nM, as well as INS-1 cellular IC 50 = 670 nM and 540 nM, respectively. Having roughly confirmed literature activities, we targeted new hybrids 10 and 11 wherein the oxadiazolopyridine would replace the aminopyrimidine in 5 and, in 10, its dipeptidyl segment would be replaced by an octanoyl unit, a feature essential for activity in our previous peptidomimetic efforts.
Amino-iodinated oxadiazolopyridine 19 (Scheme 1) was synthesized from amino cyano oxadiazole 18 and ethyl acetoacetate, as described by Boehringer Ingelheim. Sonogashira coupling of 19 with 4-ethynyl piperidine 20 (prepared via Ohira-Bestmann homologation of the corresponding aldehyde) gave chromophore 21. The attempted saturation of the alkyne in this molecule by hydrogenation over PtO 2 gave mainly cis-alkene 22 alongside a by-product that lacked an oxadiazole ring, presumably formed via N-O bond reduction. The attempted hydrogenation over various other heterogeneous and homogeneous cata-lysts gave similar results. Only when catalytic Pd(OH) 2 /C was used (EtOH, 1 atm H 2 (g)), were small amounts of alkane 23 formed in a mixture that predominately contained 22. Treatment of the crude mixture with TFA followed by acylation with butylated glycolic acid 24 allowed pure target 10 to be isolated (along with alkene congener 25) following column chromatography on silica gel, albeit in low yield. peptidomimetic efforts.
Amino-iodinated oxadiazolopyridine 19 (Scheme 1) was synthesized from amino cyano oxadiazole 18 and ethyl acetoacetate, as described by Boehringer Ingelheim. Sonogashira coupling of 19 with 4-ethynyl piperidine 20 (prepared via Ohira-Bestmann homologation of the corresponding aldehyde) gave chromophore 21. The attempted saturation of the alkyne in this molecule by hydrogenation over PtO2 gave mainly cis-alkene 22 alongside a by-product that lacked an oxadiazole ring, presumably formed via N-O bond reduction. The attempted hydrogenation over various other heterogeneous and homogeneous catalysts gave similar results. Only when catalytic Pd(OH)2/C was used (EtOH, 1 atm H2(g)), were small amounts of alkane 23 formed in a mixture that predominately contained 22. Treatment of the crude mixture with TFA followed by acylation with butylated glycolic acid 24 allowed pure target 10 to be isolated (along with alkene congener 25) following column chromatography on silica gel, albeit in low yield. To avoid the problematic hydrogenation of 21, we developed a different route to access target 11, a path also applicable to 10. In this sequence, the oxadiazolopyridine unit is installed late via annulation onto an intermediate already at the desired oxidation state. We required a protected 5-piperidinyl-2-pentanone for this purpose. Boc derivative 28 (Scheme 2) was known, although it was prepared via a 3-step sequence beginning with a relatively costly starting material [30]. We instead developed a one-step synthesis of 28 from commercial acid 27 using Me 2 CuLi.LiCN, as described by Posner and Genna [31]. Multiple grams of 28 were prepared readily using this method. Degradation of the carbamate in 28 with TFA and acylation of the incipient amino ketone salt with Boc-L-Ala-OSu produced piperidyl amide 29. Further N-terminal extension with N-Me pyrazole derivative 30 afforded compound 31. Structure 35 contains the dipeptide segment of Eli Lilly compound 5 tethered to a methyl ketone handle from which varied heterocycles can derive. To prepare target oxadiazolopyridine 11, we condensed 31 with cyanofurazan 18 [32] using SnCl 4 as promoter. This construction was originally developed by Vasil'ev et al. [33] and, in the current example, affords oxadiazolopyridine 11, directly following aqueous workup. A second regioisomer (35) was also isolated from the reaction (dr = 1.5:1), a result we interpret in terms of the competing formation of regioisomeric enamine intermediates 32/33 that cyclize onto the pendant nitrile. Subsequent tautomerization would afford oxadiazolopyridines.
Synthetic compounds 10, 11, alkene 25 and regiosiomer 35 were tested in membrane fractions containing GOAT and in ghrelin-secreting INS-1 cells. As shown in Figure 3, both 25 and 35 were significantly impaired, likely a result of their altered geometries. However, both 10 and 11 inhibited membrane GOAT as potently as did 5 and 6. Notably, compound 11 blocked ghrelin secretion from INS-1 cells 4-fold more potently than 5. To date, it is the most potent compound we have tested in our cellular assay. Because 10 is much less active in cells, yet it inhibits GOAT in membrane fractions just as well as 11, we speculate that the adenine mimcry provided by the oxadiazolopyridine is key, with a linked functionality contributing mainly to stability and cellular uptake, or lack thereof. tive 30 afforded compound 31. Structure 35 contains the dipeptide segment of Eli Lilly compound 5 tethered to a methyl ketone handle from which varied heterocycles can derive. To prepare target oxadiazolopyridine 11, we condensed 31 with cyanofurazan 18 [32] using SnCl4 as promoter. This construction was originally developed by Vasil'ev et al. [33] and, in the current example, affords oxadiazolopyridine 11, directly following aqueous workup. A second regioisomer (35) was also isolated from the reaction (dr = 1.5:1), a result we interpret in terms of the competing formation of regioisomeric enamine intermediates 32/33 that cyclize onto the pendant nitrile. Subsequent tautomerization would afford oxadiazolopyridines.  Figure 3, both 25 and 35 were significantly impaired, likely a result of their altered geometries. However, both 10 and 11 inhibited membrane GOAT as potently as did 5 and 6. Notably, compound 11 blocked ghrelin secretion from INS-1 cells 4-fold more potently than 5. To date, it is the most potent compound we have tested in our cellular assay. Because 10 is much less active in cells, yet it inhibits GOAT in membrane fractions just as well as 11, we speculate that the adenine mimcry provided by the oxadiazolopyridine is key, with a linked functionality contributing mainly to stability and cellular uptake, or lack thereof.  (bottom) In vitro inhibition of proghrelin octanoylation mediated by GOAT using the shown compounds. The experimental data was fit to a 4-parameter logistic model, (Bottom − Top)/(1 + (I/IC50) n ) + Top [24] using DYNAFIT software [25].
The Top was defined as reaction conditions without inhibitor and the Bottom was defined as reaction conditions with full inhibition by 5 µM concentration of 2. Top and Bottom conditions were included in every experiment.

Materials and Methods
NMR spectra were recorded on Bruker Advance spectrometers (300 MHz, 400 MHz, 500 MHz) and are reported as δ values in ppm relative to CDCl3 (calibrated to 7.26 ppm in 1 H NMR and 77.16 ppm in 13 C NMR, unless otherwise indicated). Splitting patterns are abbreviated as follows: singlet (s), doublet (d), triplet (t), quartet (q), multiplet (m), broad (br), and combinations thereof. Column chromatography was conducted on silica gel 60 (240-400 mesh) purchased from Sillicycle. Thin-layer chromatography (TLC) was performed using pre-coated, glass-backed plates (silica gel 60 PF254, 0.25 mm) and visualized using a combination of UV and potassium permanganate staining. HPLC analyses were (bottom) In vitro inhibition of proghrelin octanoylation mediated by GOAT using the shown compounds. The experimental data was fit to a 4-parameter logistic model, (Bottom − Top)/(1 + (I/IC 50 ) n ) + Top [24] using DYNAFIT software [25]. The Top was defined as reaction conditions without inhibitor and the Bottom was defined as reaction conditions with full inhibition by 5 µM concentration of 2. Top and Bottom conditions were included in every experiment.

Materials and Methods
NMR spectra were recorded on Bruker Advance spectrometers (300 MHz, 400 MHz, 500 MHz) and are reported as δ values in ppm relative to CDCl 3 (calibrated to 7.26 ppm in 1 H NMR and 77.16 ppm in 13 C NMR, unless otherwise indicated). Splitting patterns are abbreviated as follows: singlet (s), doublet (d), triplet (t), quartet (q), multiplet (m), broad (br), and combinations thereof. Column chromatography was conducted on silica gel 60 (240-400 mesh) purchased from Sillicycle. Thin-layer chromatography (TLC) was performed using pre-coated, glass-backed plates (silica gel 60 PF254, 0.25 mm) and visualized using a combination of UV and potassium permanganate staining. HPLC analyses were carried out using an Agilent 1200 HPLC system equipped with an Agilent Quadrupole 6130 ESI-MS detector. Mobile phase was prepared with 0.1% TFA.

(S)-3-(allyloxy)-2-((tert-butoxycarbonyl)amino)-3-methylbutanoic acid (xii):
To a solution of xi (80 mg, 0.27 mmol) in acetone (3 mL), Jones' reagent (2.5 M in H 2 O, 0.16 mL, 0.40 mmol) was added at 0 • C. The resulting mixture was warmed to room temperature, and then stirred at this temperature overnight. To the reaction mixture, celite (100 mg) and isopropanol (0.5 mL) were added, and the resulting precipitate was filtered through off through a plug of celite. The filtrate was adjusted to pH 9 with aqueous NaHCO 3 , and then concentrated under reduced pressure. The aqueous layer was washed with ether (×2) and acidified to pH 3 with citric acid. The resulting solution was extracted with EtOAc (×3) and the combined extracts were washed with brine (×2), dried over MgSO 4 , filtered, and concentrated under reduced pressure to afford xii (50 mg, 68%), which was used without further purification. 1 13 13

In Vitro Assay
pFastBac1-mouseGOAT and pGEX-GST-proGhrelin8His plasmid encoding for mouse proghrelin, fused to GST with TEV cleavage site to release proghrelin moiety, was a kind gift from the Brown and Goldstein laboratory [18]. The BL21Gold E. coli chemically competent cells were transformed with the pGEX-GST-proGhrelin8His plasmid and selected on agar plates with ampicillin. A few colonies were used to inoculate 50 mL LB, and cultures were grown overnight at 37 • C. The next morning, 10 mL of overnight culture was used to inoculate 1L LB. A total of 4 flasks with 1L LB each were inoculated with 10 mL of the overnight culture, and bacterial cultures were grown at 37 • C until they reached OD 600~0 .6. The cultures were chilled on ice for 30 min and then 0.25 mM of IPTG was added to each culture. The cultures were then moved back to a shaker set to 18 • C and incubated overnight. Cells were harvested and resuspended in 200 mL buffer containing 25 mM Tris-HCl pH 8.0, 150 mM NaCl, 1 mM EDTA and 0.5 mg/mL lysozyme. The cell suspension was sonicated, and cell debris was removed by centrifugation. The soluble fraction was passed through glutathione resin by gravity. The resin was washed with 75 mL of lysis buffer without lysozyme. The GST-proGhrelin8His protein was eluted with lysis buffer without lysozyme, supplemented with 15 mM reduced glutathione, GSH. About 50 mg of target protein, as evaluated by SDS-PAGE, was eluted. The eluate was supplemented with 1 mM DTT and about 7.5 mg of recombinant GST-TEV protease (see below) was added. The solution was incubated at 16 • C overnight to allow complete release of proghrelin8His from GST. Prior to Ni-NTA chromatography, 5 mM CaCl 2 was added, and the solution was spun down at 4000× g for 10 min. After elution from Ni-NTA resin, protein solution was dialyzed against buffer containing 10 mM Tris-HCl pH 8.5, 50 mM NaCl, 10% glycerol and 0.01% CHAPS using 3 kDa cutoff membrane with 3 buffer exchanges. The protein was quantified by absorbance at A 280 and qualified by SDS-PAGE.
Generation of GST-TEV protease. The pMHT vector [34] (gift from Dr. Arbing, UCLA) was linearized by PCR reaction with primers outside of MBP gene: forw-5 gcgaccatcctccaaaatcgggagaaagcttgtttaaggggccg 3 and rev-5 caataacctagtataggggacatggttaatttctcctctttaatg 3 . The resulting linear plasmid was used to clone GST gene in-frame at 5 of TEV gene by in vitro Gibson assembly. The GST gene was prepared by PCR amplification with primers: form 5 atgtcccctatactaggttattg and rev 5 cgattttggaggatggtcgc 3 from pGEX-GST-proGhrelin8His plasmid as a template. The final bacterial expression vector, pGST-TEV was used for expression and purification of GST-TEV fusion protein. BL21Gold chemically competent E. coli cells were transformed with pGST-TEV plasmid and entire transformation mixture was used to inoculate 50 mL LB supplemented with kanamycin. Next day, 10 mL of the overnight culture was used to inoculate 1 L of LB. Four flasks with 1 L LB each were inoculated with 10 mL of the overnight culture, and bacterial cultures were grown at 37 • C till OD600~0.8. was reached. After adding 1 mM IPTG, culture was allowed to grow for 4 h, and cells were harvested by centrifugation. The cell pellet was resuspended in 200 mL lysis buffer containing 25 mM Tris-HCl pH 8.0, 150 mM NaCl, 1 mM EDTA, 5% glycerol, 0.5 mg/mL lysozyme and 1 mM DTT. Cell suspension after freeze-thaw cycle was sonicated and cellular debris was removed by centrifugation. The soluble fraction was bound to glutathione resin by gravity. The resin was washed with 50 mL washing buffer (lysis buffer without lysozyme and DTT) and fusion GST-TEV protein was eluted with 25 mL washing buffer supplemented with 15 mM reduced glutathione (GSH). The elution was fractionated into 5 fractions and after SDS-PAGE/Coomassie analysis, fractions containing most of the GST-TEV protein were combined and dialyzed against buffer containing 25 mM Tris-HCl pH 8.0, 150 mM NaCl and 50% glycerol using 3 kDa cutoff membrane with 3 buffer exchanges. The total yield was roughly 8 mg per 1 L bacterial culture.
The baculovirus expression of mouse GOAT. DH10Bac E. coli strain (ThermoFisher, Waltham, MA, USA) was transformed with pFastBac1-mouseGOAT plasmid, and Bacmid DNA from 12 white colonies was analyzed by DNA sequencing and PCR reaction. All clones were transfected to sf9 cells and level of mGOAT expression was compared between clones from cells producing P2. The clone with the highest mGOAT expression level was used to produce P3. For membrane isolation, sf9 cells were seeded at 2-4 × 10 6 /mL in total 1 L volume of sf-900 serum-free medium (ThermoFisher) and infected with P3 virus. At day 2 post-infection, cells were harvested and resuspended in 40 mL of buffer containing 50 mM NaPi pH 7.2, 150 mM NaCl, 1 mM EDTA, 100 µM bis (4-nitrophenyl) phosphate, 2.5 µg/mL aprotinin, 10 µg/mL leupeptin, 10 µg/mL pepstatin A. Cell suspension was briefly sonicated and cell debris was removed by centrifugation at 3000× g for 10 min.
Membrane fraction was collected from supernatant by centrifugation at 100,000× g for 1 h. Membrane pellet was resuspended in storage buffer (50 mM NaPi pH 7.2, 150 mM NaCl and 10% glycerol) and kept at −80 • C. Acyltransferase assay: The assay conditions included per 50 µL reaction: 50 µg of total sf9 membranes, various (for modality experiment) or fixed (for IC 50 ) concentrations of recombinant proghrelin8His peptide and various concentrations of tested compound (for IC 50 ), 100 µM palmitoyl CoA, 50 mM HEPES pH 7.0 and 1 µM [ 3 H] octanoyl CoA (~5.5 dpm/fmol-American Radioactive Chemicals). After incubation of the reaction mixture at 37C for 10 min, tubes were placed on ice and 10 µL of 1M HCl was added to each tube. Following the addition of 740 µL of cold quench buffer (50 mM NaPi pH 7.4, 10 mM Imidazole, 150 mM NaCl, 100 µM bis (4-nitrophenyl) phosphate, 1 mM phenylmethylsulfonyl fluoride and 0.1% Triton), 0.2 mL of 50% Ni-NTA slurry was added to each reaction. Tubes were incubated at 4 • C for 1 h with rotation to capture proGhrl8His. After washing Ni resin with 40 mM Imidazole, all bound proghrelin8His was eluted with 250 mM Imidazole, and an amount of octanylated proghrelin was assessed with scintillation counting.

INS1-Cellular Assay
Generation of Recombinant Retrovirus for mGOAT Expression. The mouse GOAT cDNA with C-terminal HA tag was amplified from pcDNA3.1-mouseGOAT-HA vector (gift from Brown and Goldstein lab [2]) using the following primers: mGOAT_attb1 (forward primer) 5 -ggggacaagtttgtacaaaaaagcaggctaccatggattggctccagctc-3 (attb1 recombination site is in italics, 5 of mGOAT coding sequence is in bold, and Kozak coding sequence is underlined) and mGOAT_attb2 (reverse primer) 5 -ggggaccactttgtacaagaaagctgggtctaagcgtaatctggaacatc -3 (attb2 recombination site is in italics, 3 of mGAOT-HA coding sequence is in bold, and stop codon is underlined). The PCR product was cloned into donor vector pDONR221 (ThermoFisher) with BP clonase according to manufacturer's instructions. Positive clones were verified by DNA sequencing with M13F and M13R primers. The resulting entry plasmid pDONR221-mGOAT-HA was used to transfer mGOAT-HA cDNA to destination vector pBabe-puro (Addgene cat# 51070 [35]) using LR clonase according to manufacturer's instructions. Positive clones were verified by DNA sequencing with pBABE-5 and pBABE-3 primers. The resulting plasmid, pBabe-puro-mGOAT-HA, was used to generate retrovirus. The retrovirus was packed in 293T PhoE (Phoenix-ECO AT0CC ® CRL-3214 ™ ) cells and used to infect INS-1 cells. The stable INS/GOAT cells were selecting on 1 µg/mL puromycin and GOAT expression was confirmed with immunoblot of membrane fraction isolated from antibiotic resistant culture using anti-HA antibody.
Generation of Lentivirus for Expression of Ghrelin. The cDNA for mouse preproghrelin was subcloned from pcDNA3.1-preproghrelin (gift from Brown and Goldstein lab [18]) into pULTRA vector (Addgene cat# 24129 [36]) with XbaI and BamHI restriction enzymes. Using of these restriction sites for cloning will result in the creation of bi-cistronic expression of ghrelin along with EGFP to facilitate identification of positive cells by fluorescence. The above restriction sites were engineered via PCR reaction using the following primers: ghrl_pultra_forw 5 -taccgagctctctagaatgctgtcttcaggc-3 (5 of preproghrelin coding sequence is in bold; XbaI site is in italics) and ghrl_pultra_rev 5 -agcggccgcggatccttacttgtcagctggc -3 (3 of preproghrelin coding sequence is in bold, stop codon is underlined, and BamHI site is italics). The recombinant lentivirus encoding preproghrelin cDNA was packed in Lenti-X 293T cells (Takara cat# 632180) by co-transfection of pUltra-EGFP-mouse-preproghrelin, with packaging encoding plasmid pCMV ∆R8.2 (Addgene cat# 12263) and envelope encoding plasmid pCMV-VSV g (Addgene cat# 8454). The INS/GOAT cells were infected with recombinant lentivirus to generate INS/GOAT/GHRL cell line. After few passages, the population of cells with puromycin resistance and~90% fluorescence was saved for cell-based assay.
Cellular assay. INS/GOAT/GHRL cell line was routinely cultured in RPMI medium supplemented with 10% FBS, 1% Pen/Strep, 10 mM HEPES pH 7.2 and 50 µM 2-mercaptoethanol. Day 0: One 10 cm dish at full confluency was used to seed one 96-well plate. Day 1: Growth media was removed and cells were washed with PBS prior to adding fresh growth media as above but without 2-mercaptoethanol. The serial dilutions of tested compound at 50× concentration were prepared in vehicle containing growth media without 2-mercaptoethnol supplemented with 6% DMSO and were added to cells in duplicates. Day 2: 10 µL of growth media were removed from each well and amount of secreted acyl-ghrelin was measured with ELISA (Cayman cat# 10006307).

Conclusions
Ghrelin signaling continues to be an active area of basic biology research and drug discovery. The discovery of GOAT and its role in maturing ghrelin in the stomach provided an opportunity for pharmacological intervention without accessing the central nervous system. These efforts were largely empirical due to a lack of structural data for the system. However, we have learned a great deal. Peptidomimetic inhibitors that compete for substrate binding can perform well in vitro but, thus far, their utility in cell culture and in animals has been limited. The high-throughput screens developed in the private identified compounds uniformly appear to complete for binding at the co-enzyme site. Several of these molecules advanced through pre-clinical developments and into human trials as therapy for diabetes type II, Prader-Willi syndrome and alcohol use disorder. We developed a six-step synthesis of heterocyclic GOAT inhibitors based on a hybrid design that utilizes a piperidinyl ethyl linker and a late-stage oxadiazolopyridine, forming annulation. The potent activity of novel product 11 both in vitro and in cells indicates this approach has the potential to identify new GOAT inhibitors with increasingly refined properties.