Cellular Uptake of a Fluorescent Ligand Reveals Ghrelin O-Acyltransferase Interacts with Extracellular Peptides and Exhibits Unexpected Localization for a Secretory Pathway Enzyme

Ghrelin O-acyltransferase (GOAT) plays a central role in the maturation and activation of the peptide hormone ghrelin, which performs a wide range of endocrinological signaling roles. Using a tight-binding fluorescent ghrelin-derived peptide designed for high selectivity for GOAT over the ghrelin receptor GHSR, we demonstrate that GOAT interacts with extracellular ghrelin and facilitates ligand cell internalization in both transfected cells and prostate cancer cells endogenously expressing GOAT. Coupled with enzyme mutagenesis, ligand uptake studies support the interaction of the putative histidine general base within GOAT with the ghrelin peptide acylation site. Our work provides a new understanding of GOAT’s catalytic mechanism, establishes that GOAT can interact with ghrelin and other peptides located outside the cell, and raises the possibility that other peptide hormones may exhibit similar complexity in their intercellular and organismal-level signaling pathways.

G hrelin is a unique peptide hormone implicated in regulation of physiological pathways impacting appetite, energy storage and metabolism, neurological responses to stress, and reward processing associated with addictive behavior. 1−7 Ghrelin exists in two distinct chemical forms in the bloodstream, acylated ghrelin with an eight-carbon fatty acid covalently linked to a serine side chain at position 3 and desacyl-ghrelin with a free serine hydroxyl at this position. 8 Ghrelin is the only known peptide predicted to undergo this serine octanoylation modification, which is catalyzed by ghrelin O-acyltransferase (GOAT). 9 Ghrelin octanoylation is required for binding to its cognate receptor, the growth hormone secretagogue receptor (GHSR), a member of the G-protein coupled receptor family. 7,10 Beyond their direct involvement in ghrelin maturation and signaling, the ghrelin binding proteins GHSR and GOAT are potential disease biomarkers. The GHSR receptor is being explored for cancer detection and imaging due to its altered expression in neoplasms including prostate, testicular, ovarian, breast, and neuroendocrine tumors. 11−13 GOAT overexpression is also observed in multiple cancers including breast, endocrine tissue, and prostate. 14 −18 In a study accounting for metabolic alterations in prostate cancer (PCa) patients, GOAT was shown to be differentially overexpressed in these patients compared to noncancer controls. 14 GOAT has also been detected by an ELISA-based assay in the urine and blood plasma of PCa patients, with GOAT plasma concentrations acting as a more consistent and sensitive detector of aggressive PCa than the traditional PSA diagnostic biomarker. 14,15 To fully explore the potential for GOAT to serve as a novel disease biomarker, significant advancements are needed in the design of efficient probes for detecting GOAT and our understanding of the trafficking of this enzyme within both the cell and the body.
Originally annotated and observed in the endoplasmic reticulum membrane, 19 GOAT has been recently suggested to also be distributed to the plasma membrane. 20,21 Extracellular exposure on the plasma membrane would position GOAT to serve as a cancer cell biomarker through GOAT-specific ligands coupled to the appropriate imaging groups. The high selectivity of GOAT for ghrelin, the only predicted GOAT substrate in the human proteome, 9 supports the ability to design the potent GOAT-targeted ligands required to exploit this novel diagnostic and potential therapeutic target. In this work, we employed parallel structure−activity analyses to develop a synthetic ghrelin analogue with nanomolar binding to GOAT without any measurable binding to GHSR. This ligand was further functionalized with a sulfo-Cy5 fluorophore to afford imaging of ligand binding to human cell lines expressing GOAT. The ligand probes designed in this study and our analysis of GOAT localization and ligand binding offer mechanistic insight into GOAT binding and catalytic strategies, demonstrate that GOAT interacts with extracellular peptides at the cell surface, and support further exploration of GOAT as an imaging target for disease diagnostics.
■ RESULTS AND DISCUSSION Development of a Specific High-Affinity Peptide Ligand for GOAT. The high affinity and specificity of ghrelin binding to GHSR has enabled development of imaging agents targeting this receptor such as a fluoronaphthyl acylated ghrelin (1−8) analogue for use in PET imaging. 22 While GOAT can also bind ghrelin mimics acylated with fatty acids, 9,23 the affinity of these molecules for both GOAT and GHSR renders these molecules unsuitable for specifically detecting and imaging GOAT in a cellular or organismal context. The GO-CoA-Tat bisubstrate GOAT inhibitor developed by Barnett and co-workers binds GOAT without exhibiting antagonism of GHSR, 24 demonstrating selectivity between these two ghrelin binding proteins. In GO-CoA-Tat, selectivity for GOAT is presumably due to the inclusion of coenzyme A attached to the acyl side chain. To provide an easily functionalizable and synthetically accessible scaffold for ligand development, we explored a new class of GOAT ligands inspired by a class of substrate-mimetic GOAT inhibitors incorporating a free amino group in place of the serine hydroxyl at the acylation site. 25 With these ligands lacking a hydrophobic moiety at the acylation site, we predicted they would exhibit significantly reduced binding affinity for GHSR. 10 Compared to peptides acylated with either an octanoyl group (ligand 2) or a 6-fluoro-2-naphthoyl group (ligand 3), the peptide with a free Dap amino group (ligand 1) exhibits tight binding to GOAT without detectible binding to the ghrelin receptor (Table 1 and Supporting Figures S1 and  S2).
Intriguingly, while the octanoylated ligand 2 binds tightly to both GOAT and GHSR, incorporation of the fluoronaphthoyl group at the acylation site in ligand 3 enhances receptor binding while blocking interaction with GOAT. This suggested the potential to tune ligand selectivity for both GOAT and GHSR to generate orthogonal ligands for each ghrelin-binding protein. Noting the strict selectivity for recognition of the Nterminal glycine residue (G1) by GOAT, which requires both the N-terminal amino group and lack of side chain/steric bulk at the G1 position, 9,25 we replaced G1 with two unnatural amino acids�aminoisobutyric acid (Aib, ligand 4) and isonipecotic acid (Inp, ligand 5)�and determined binding affinities of these ligands for both GOAT and GHSR. Incorporation of either Aib or Inp at the peptide N-terminus had a pronounced negative impact on binding to GOAT while GHSR readily tolerated replacement of G1 in the context of acylated ligands 6 and 7. 22 Taken together, these studies show that modifications at only two sites within the ghrelin peptide sequence, the G1 position and the acylation site, are sufficient to achieve ligand specificity toward either GOAT or GHSR (Table 1).
Having achieved ligand selectivity for GOAT over GHSR through modifications at the acylation site, we sought to optimize ligand binding affinity for GOAT. We explored substitutions at glutamate 8 (E8) and phenylalanine 4 (F4) within the ghrelin-derived ligand based on previous studies demonstrating the involvement of these amino acids in ghrelin recognition by GOAT and GHSR. 9,22,25,26 Incorporation of a threonine residue at E8 strengthens binding of the ligand to GHSR, and we found this substitution similarly increases binding of the ligand to GOAT (ligand 11) much more than binding of either tyrosine or asparagine at this position. Both GOAT and GHSR recognize the F4 residue, 22,27 with GOAT exhibiting a preference for large hydrophobic/aromatic amino acids at this position. 9,25 Exploiting this preference using unnatural amino acids, we found that incorporation of 1naphthylalanine (Nal-1, compound 12) or 2-naphthylalanine (Nal-2, compound 13) at the F4 position substantially increased ligand binding to GOAT in the context of unacylated ligands, which exhibited no detectible binding to GHSR (Table 1).
Correlating binding affinities for GOAT and GHSR exhibited by ligands in this study highlights three classes of compounds, nonselective ligands and those with >100-fold binding preference for either GOAT or GHSR ( Figure 1). Combining the most successful substitutions in these studies has generated a highly selective ligand 12 for GOAT with nanomolar affinity, with ligand 14 previously reported as a highly potent ligand for GHSR which combined elements of ligands 3 and 4 ( Figure 2). 22 Each of these ligands exhibits nanomolar (or better) potency for its intended target without any detectable interaction with the other ghrelin-binding protein up to a 100 μM ligand concentration. To equip the GOAT selective ligand for use in cell imaging, we synthesized  ligand 15 containing a lysine residue at its C-terminus and a sulfo-Cy5 fluorophore attached to the lysine side chain ( Figure  2). Earlier studies of labeled Cy5-ghrelin (1−19) demonstrated strong binding to GHSR, with less susceptibility to photobleaching and high detection of GHS-R expression in live differentiating cardiomyocytes. 28 Ligand 15 exhibits potent binding to GOAT with a nanomolar IC 50 value when assayed as an inhibitor, supporting our ability to functionalize GOAT peptide ligands with imaging groups without compromising the GOAT binding ability. Ligand Binding and Uptake by Human Cells Transfected with hGOAT. To demonstrate the interaction between ligand 15 and hGOAT in a biologically relevant cellular setting, we transiently transfected HEK 293 cells with either a FLAG-tagged hGOAT construct or an empty vector and imaged cells to determine hGOAT expression and ligand 15 binding (Figures 3 and S3). Following live cell incubation with the fluorescent ligand 15, cells were washed, fixed, and labeled with antibodies against FLAG. Using spinning-disk confocal microscopy, FLAG-hGOAT expressing cells were positive for ligand 15 Cy5 fluorescence with ligand fluorescence distributed throughout the cell, whereas empty vector control expressing cells were not, consistent with GOAT expression being required for ligand binding and cellular uptake ( Figure 3 and Supporting Table S2). In contrast, cells transfected with an empty vector excluded ligand 15, demonstrating that this peptide is not intrinsically cell permeable or nonspecifically associated with cellular membranes.
The uptake of ligand 15 by hGOAT-expressing cells provides the opportunity to define the molecular interactions between the ligand and hGOAT responsible for ligand binding affinity. For example, the enhancement in binding to hGOAT upon amine substitution at the acylation site could arise from formation of a ground state hydrogen bond to an enzyme side chain, which normally serves as a general base for serine acylation (Figure 3b). To explore ligand−enzyme interactions required for binding and cellular uptake, we introduced three alanine mutations to functionally essential amino acids ( Figure  3c). These three mutations all result in a complete loss of hGOAT enzyme activity, but this loss of activity reflects interference in contributions by these residues at different steps of the hGOAT catalytic cycle. 29 Histidine 338 is an absolutely conserved histidine in MBOAT family members and is proposed to serve as a general base interacting with the serine hydroxyl group. 26,29 We propose that an interaction between the Dap side chain amine and H338 is partially responsible for the tight binding observed for Dap-containing peptide ligands to hGOAT. In contrast, both arginine 304 and asparagine 307 form interactions within the octanoyl-CoA acyl donor-binding site, 29 which would not directly impact ghrelin-mimetic peptide binding to hGOAT ( Figure 3 and Supporting Table  S2).
The H338A hGOAT variant does not maintain cell uptake of ligand 15 when expressed in HEK293 cells, supporting the idea that an interaction between the Dap amino group and H338 is required for tight ligand binding. Furthermore, the loss of ligand uptake in the presence of H338A hGOAT expression argues against a general loss of cell membrane integrity as the mechanism for ligand internalization. Given the integral membrane nature of GOAT, 19,29 we considered it possible that hGOAT expression may destabilize or disrupt membrane integrity, which could allow ligand cell penetration without requiring direct binding to hGOAT. In contrast to the H338A variant, both R304A and N307A variants supported ligand 15 internalization as expected for mutations predicted to compromise the octanoyl-CoA binding pocket but not the ghrelin binding site within the hGOAT catalytic channel. The ability of the noncatalytically competent R304A and N307A variants to support ligand uptake also provides support; the ligand 15 uptake occurs through binding to GOAT rather than an alternate model wherein the ligand is acylated in situ and then undergoes uptake by binding to GHSR. This is also consistent with our previous studies of Dap-containing ghrelin mimetic peptides that demonstrated these peptides to be very inefficient substrates for GOAT-catalyzed acylation even under forcing conditions. 25 Endogenous GOAT Expression in Prostate Cancer Cells Supports Ligand Uptake. While unlikely, it is possible that the hGOAT interaction with extracellular peptides in transfected HEK 293 cells reflects aberrant trafficking to the plasma membrane resulting from enzyme overexpression or effects from the transient transfection itself. To complement ligand uptake studies using hGOAT overexpression in transfected HEK293 cells, we examined human cell lines with elevated levels of endogenous GOAT expression for similar ligand binding and uptake. These experiments maintain a high level of GOAT expression in a more biologically relevant context to probe GOAT exposure to extracellular peptides at the plasma membrane. For these studies, we utilized LNCaP and 22Rv1 prostate cancer lines which have been reported to overexpress GOAT when compared to normal prostate cells. 14,15,30 Immunofluorescence imaging using an anti-MBOAT4 antibody revealed robust hGOAT expression in both prostate cancer lines (Figure 4a,b and Supporting Figures S4−S6).
We next examined the binding and cellular uptake of ligand 15 in prostate cancer cells expressing hGOAT to compare with  Table 1. our earlier studies of hGOAT-transfected cells. Confocal imaging confirmed ligand binding and uptake in both 22Rv1 and LNCaP cells, with 100% of the prostate cancer cells exhibiting ligand binding (Figure 5a−c). Ligand binding and uptake became more pronounced at higher ligand concentrations ( Figure 5d). Labeling of 22Rv1 PCa cells with ligand 15 was significantly less efficient at 4°C than at 37°C, consistent with ligand binding and uptake requiring unimpeded plasma membrane trafficking for ligand internalization (Figure 5e). Selective uptake of ligand 15 by hGOAT was probed by treating cells with an unlabeled competitor ligand to saturate cell surface-exposed hGOAT to block ligand 15 binding. Co-incubation with a large excess of unlabeled competitor ligand 12 similarly led to a significant reduction of ligand 15 internalization consistent with specific ligand recognition and uptake mediated through binding of the ligand to cell surface exposed hGOAT (Figure 5f,g).
Originally assigned as an ER-resident enzyme responsible for acylating ghrelin during hormone maturation prior to secretion, 7 our work provides the first direct evidence for GOAT exposure to the extracellular space and interaction with soluble peptides. These studies were enabled by the creation of a specific ghrelin-mimetic ligand for GOAT, which allows for direct detection and investigation of this ghrelin-binding protein without interference from GHSR. Our development of ghrelin-based ligands opens the door for creating noninvasive imaging agents targeting GOAT, while also providing the first functional connection between the active site of GOAT and ghrelin through the Dap amine−H338 interaction. Most unexpectedly, our studies indicate that GOAT can bind extracellular peptides and facilitate cellular uptake.
The absolutely conserved histidine residue that serves as one of the defining characteristics of MBOAT family members (H338 in GOAT) has been suggested to act as a general base in the acylation reactions catalyzed by these enzymes. 19,23,26,29,31−42 While this catalytic role has not been conclusively demonstrated in any MBOAT, the dependence of ligand 15 uptake on the presence of H338 in GOAT supports a direct interaction between the acylation site serine in ghrelin and this conserved histidine (Figure 3b). A role for H338 in both ghrelin binding and catalysis is supported by previous studies of ghrelin substrate analogs by Taylor and co-workers, who similarly hypothesized that H338 could also play a role in binding and orienting the peptide substrate for the acylation reaction using hydrogen bonding. 26 The enhanced binding affinity of ghrelin ligands with amine modifications at the serine acylation site likely arises from reapportionment of the transition-state stabilization energy from the serine−histidine hydrogen bond/general base interaction during acyl transfer to ground-state binding enhancement from the amine/ammonium−histidine interaction in ligand 15. This interaction provides the first functional connection between a residue within ghrelin and the GOAT active site, a valuable constraint for ongoing studies to model ghrelin binding within GOAT as no structural or computational models for this complex are currently available. Defining the substrate bindings sites in GOAT is essential for further modeling of this enzyme and its catalytic architecture. 29 Looking beyond GOAT, it will be interesting to examine similar hydroxyl to amine acylation site substitutions in other MBOAT substrates to determine if this simple atomic substitution provides a facile family wide strategy for generating potent MBOAT inhibitors and identifying catalytic interactions in these enzymes. 43  Cellular uptake of ligand 15 requires a subpopulation of GOAT to be exposed on the plasma membrane where it can bind ghrelin and desacyl-ghrelin from outside the cell, consistent with detection of GOAT in intracellular and plasma membranes of lipid trafficking vesicles in blood marrow adipocytes using immunogold staining. 20 Our unambiguous demonstration of ligand binding and cellular uptake supports an expanded view of GOAT's involvement within the ghrelin trafficking and signaling pathway in cells expressing GOAT ( Figure 6). Our finding provides support for a new branch of the ghrelin signaling pathway involving local reacylation of desacyl-ghrelin by plasma membrane exposed GOAT, which could provide a potential explanation for the biological impact of treatment with desacyl ghrelin. 7,20,21,44−47 The question of desacyl-ghrelin biological activity has remained an unresolved point of contention in the ghrelin field, with differing claims of reproducibility between laboratories. The lack of a known receptor for desacyl-ghrelin, despite significant efforts to identify a candidate receptor for this role, also raises questions about the existence of desacyl-ghrelin signaling and its biological mechanism.
Ghrelin reacylation would allow cells and tissues presenting both GOAT and GHSR on their surfaces to detect the total ghrelin level in circulation rather than only the acylated portion of the ghrelin pool in the bloodstream. Cellular integration of the "total ghrelin" signal could provide a readout for chronic organismal stress, whether from metabolic factors or other environmental stressors, rather than the acute signal provided by the rise and fall of the hydrolytically susceptible pool of ghrelin secreted from the gastrointestinal tract. 48−51 This twofactor signaling system may provide mechanisms to explain the complex and sometimes paradoxical biological signaling behavior observed for ghrelin and desacyl ghrelin, essential features of this endocrine system that remain to be fully understood. We note that the existence of this proposed local ghrelin reacylation pathway remains to be conclusively demonstrated (e.g., definitive bioanalytical demonstration of desacyl ghrelin reacylation by cell surface-exposed GOAT), with such a demonstration essential to show that plasma membrane-localized GOAT plays an important role in ghrelindependent physiology.
Given GOAT overexpression reported in prostate and breast cancer, 14−18,30 the role of surface-exposed GOAT in ghrelin signaling in cancer cells and validation of GOAT as a cancer biomarker represent compelling areas for further investigation. Validating GOAT as a cancer biomarker will require GOAT expression analysis and ligand binding/uptake studies in related noncancerous cells to demonstrate that elevated GOAT expression correlates with the presence of cancer and serves as a sign of oncogenic cell behavior. 14,30 In cancer cell lines and tissues found to overexpress GOAT, defining the impact of this overexpression on cell signaling is the next step in understanding how these transformed cells are affected by organismal metabolic and endocrine states as reflected by ghrelin and des-acyl ghrelin concentrations. We anticipate applying selective GOAT ligands, GOAT localization analysis, and ghrelin-dependent cell signaling studies in future work to explore this compelling intersection of ghrelin signaling and cancer biology. 13,52 Ligand internalization can be facilitated by GOAT through at least two distinct mechanisms: (1) GOAT−ligand complex transportation into the cell through membrane trafficking or (2) ligand transit into the cell by traversing GOAT as if the enzyme were a pore. In mechanism 1, the GOAT−ligand complex would be internalized by endocytosis similarly to receptor-mediated ligand uptake. In contrast, mechanism 2 allows GOAT to act as a ligand transporter at the plasma membrane without enzyme internalization. Computational modeling of GOAT supports the presence of an internal transmembrane channel that could act as a pore, 29 and similar channels are observed in the crystal structure of the bacterial MBOAT DltB, the cryo-EM structures of Hhat, and structural models of PORCN. 32,53−56 We note that our imaging studies do not support colocalization of ligand 15 with GOAT in either transfected HEK 293 cells or prostate cancer cells, which could potentially support ligand transport (mechanism 2) rather than enzyme−ligand internalization (mechanism 1). However, our studies were performed with fixed cells, and the short length and small number of amine groups within the peptide ligand may lead to inefficient ligand cross-linking during fixation, allowing intracellular dispersion during sample preparation. Further studies of GOAT and GOAT−ligand cellular trafficking will determine which of the proposed mechanisms is responsible for ligand internalization.
Looking to the future, we will investigate the mechanism and biological impact of peptide internalization by GOAT using the ligands reported in this work and expand our understanding of GOAT trafficking and localization within mammalian cells. Whether through transport of the GOAT− ligand complex by membrane trafficking or GOAT serving as a pore/transporter using its internal channel, this process presents a new and unanticipated function for integral membrane acyltransferases and may provide a novel avenue for intracellular drug/cargo delivery targeting cells expressing GOAT. We also suggest that this work argues for examining the potential involvement of other integral membrane enzymes currently proposed to perform exclusively intracellular roles in extracellular interactions, chemistry, and signaling.   Table S1. GHSR Receptor Binding Assays. Peptide binding affinity for the ghrelin receptor was determined using a competitive radioliganddisplacement binding assay, 22 with details provided in the Supporting Information.
hGOAT Inhibition Assays. Assays were performed using previously reported protocols, 57,58 as described in the Supporting Information.
Construction of hGOAT WT and Mutants. Site-directed mutagenesis was performed on pcDNA 3.1 (+) mammalian expression vector containing an hGOAT insert cloned from our previously reported pFastBacDual vector (Invitrogen) using the EcoRI and XbaI restriction sites, resulting in the pcDNA3.1_Mb4.WT (hGOAT) construct. 57 This construct contains a C-terminal FLAG epitope tag, a polyhistidine (His6) tag, and three human influenza hemagglutinin (HA) tags appended downstream of a TEV protease site. 58 Plasmids and primers are provided in Tables S3 and S4 in the Supporting Information.
hGOAT Transfection in HEK 293 Cells. Mammalian cell line HEK 293 (ATCC) was maintained in 75 mL vented tissue culture flasks (Celltreat) and kept to 70% confluency before splitting. All cells were cultivated in complete DMEM (Dulbecco's Modified Eagle's Medium) supplemented with 10% fetal bovine serum (FBS) and 1% (v/v) penicillin-streptomycin (MediaTech) in a humidified atmosphere with 5% CO 2 at 37°C. For transfection of WT hGOAT, mutant hGOAT and empty vector (EV) cells were plated at a density of 1 × 10 6 per well in 2 mL of complete DMEM in a six-well plate per well (Corning) with sterile 12 mm 1.5 glass coverslips in each well (Warne Instruments). The cells were incubated for 16 h prior to transfection. The DNA−transfection reagent complex was prepared by combining 4 μg of pcDNA3.1_Mb4.WT (hGOAT) or mutant plasmid and 9 μL of Lipofectamine 2000 transfection reagent (Invitrogen) in a total volume of 500 μL of supplement free DMEM followed by incubation for 30 min at RT. The cells were then transfected with the DNA−transfection reagent complex by dropwise addition into the plate wells.
GOAT Ligand Labeling and Immunofluorescence Imaging in HEK293 Cells. Following transfection for 40 h, coverslips with attached cells were removed from the wells, washed with 1× phosphate-buffered saline (PBS; Cellgro), and incubated with 10 μM ligand 15 (500 μL/well) for 30 min at 37°C. Following washing with 1× PBS, cells were fixed with 4% paraformaldehyde for 20 min at RT, washed with 1× PBS, quenched in 50 mM NH 4 Cl for 10 min at RT, and washed with 1× PBS. For antibody staining, all steps were performed in a Parafilm dark chamber at RT (unless otherwise specified) with a humid atmosphere. Cells were blocked with PBSAT buffer (PBS + 1% bovine serum albumin, 0.1% triton) for 30 min in the dark chamber, followed by aspiration of the buffer without allowing the coverslip to dry. Cells were then incubated with 1:250 diluted primary antibody Rabbit anti FLAG (DYKDDDDK) antibody (Sigma, F7425) diluted in 1× PBS overnight in a dark chamber at 4°C . The following day, cells were washed with 1× PBS and incubated with 1:1000 diluted secondary antibody Alexa Fluor 488-conjugated goat antirabbit (Jackson Immuno Research, 709−545−149) for 1 h. Following antibody incubations, cells were washed three times, mounted on slides with DAPI, and analyzed by confocal microscopy. Images were taken on a Leica DMi8 STP800 (Leica, Bannockburn, IL) equipped with an 89 North−LDI laser with a Photometrics Prime-95B camera taken with a Crest Optics X-light V2 Confocal Unit spinning disk. Optics used were HC PL APO 63×/1.40 NA oil CS2 Apo oil emersion objective.
Immunofluorescence Staining of hGOAT in PCa Cell Lines. Confluent PCa cells grown on coverslips were fixed with 4% paraformaldehyde for 20 min at RT and quenched in 50 mM NH 4 Cl for 10 min at RT. For antibody staining, all steps were performed in a Parafilm dark chamber at RT with a humid atmosphere. Cells were blocked under permeabilizing conditions with PBSAT buffer for 30 min in the dark chamber. Cells were then incubated with 1:40 or 1:80 diluted rabbit anti MBOAT4 polyclonal antibody (Cayman, #18614) for 1 h in the dark chamber at 4°C, and then cells were incubated with 1:1000 diluted secondary antibody Alexa Fluor 488-conjugated goat antirabbit (Thermo, A21206) for 1 h at RT. Following antibody incubations, cells were extensively washed, mounted on slides with Prolong containing DAPI (Thermo, P36971), and analyzed by confocal microscopy.
GOAT Ligand Labeling and Imaging in PCa Cells. Upon reaching confluency, coverslips with attached PCa cells were treated with ligands at the indicated concentration for 30 min at RT (unless otherwise stated). Cells were washed 3× with 1× phosphate-buffered saline (PBS) (Cellgro) and then fixed for immunofluorescence staining as described above. For ligand competition experiments, cells were incubated with 5 μM ligand 15 alone, 5 μM ligand 15 with 40 μM ligand 12, or PBS alone for 30 min at RT. For variable temperature experiments, PCa cells were preincubated in a tissue culture refrigerator (4°C), on the benchtop (21°C), or in an incubator (37°C) for 30 min prior to addition of ligand 15. The cells were further incubated with ligand 15 at those temperatures for 30 min and then processed for imaging as described above.
Image Analysis. The entire cell was imaged at 0.2-μm stepintervals and displayed as maximum projections (ImageJ). The fluorescence range of intensity was adjusted identically for each image series. Graphs and statistical analyses were completed using Graphpad Prism software, with specific tests and p values provided in the figure captions. All images were set to a resolution of 300 DPI or greater after image analysis from the raw data. Total cell fluorescence was determined by calculating the integrated density of mean gray value in a cell area compared to background as shown in eqs 4 and 5.