Affinity-driven aryl diazonium labeling of peptide receptors on living cells

Abstract

Peptide-receptor interactions play critical roles in a wide variety of physiological processes. Methods to link bioactive peptides covalently to unmodified receptors on the surfaces of living cells are valuable for studying receptor signaling, dynamics, and trafficking, and for identifying novel peptide-receptor interactions. Here, we utilize peptide analogues bearing deactivated aryl diazonium groups for the affinity-driven labeling of unmodified receptors. We demonstrate that aryl diazonium-bearing peptide analogues can covalently label receptors on the surface of living cells using both the neurotensin and the glucagon-like peptide 1 receptor systems. Receptor labeling occurs in the complex environment of the cell surface in a sequence specific manner. We further demonstrate the utility of this covalent labeling approach for the visualization of peptide receptors by confocal fluorescence microscopy, and for the enrichment and identification of labeled receptors by mass spectrometry-based proteomics. Aryl diazonium-based affinity-driven receptor labeling is attractive due to the high abundance of tyrosine and histidine residues susceptible to azo coupling in the peptide binding sites of receptors, the ease of incorporation of aryl diazonium groups into peptides, and the relatively small size of the aryl diazonium group. This approach should prove to be a powerful and relatively general method to study peptide-receptor interactions in cellular contexts.

Graphical Abstract

Introduction

Neuropeptides and peptide hormones play critical roles as cell-to-cell messengers within the central nervous system and the endocrine system, where they modulate a large variety of physiological processes.^1–4 Mimicking or antagonizing peptide signaling represents a proven strategy for treating a number of human health concerns, including diabetes, osteoporosis, cancer, gastrointestinal disorders, and more.^5–6 Despite their importance, much about intercellular communication through endogenous peptides remains unknown. Endogenous peptides are one of the largest classes of chemical transmitters, and the molecular details of ligand-receptor interactions can be quite complex.^7–9 For example, many peptides bind to multiple receptors and many receptors can be activated by multiple peptides.^10–12 A number of bioactive peptides do not have well-characterized receptors, and even more identified peptides have completely unknown functions.^7–9,13 Even for peptides with established receptors, the dynamics of ligand-receptor complexes (e.g., receptor internalization, oligomerization, or recycling) often remains understudied.^14–15

Affinity-driven protein labeling facilitates the study of membrane proteins and ligand-receptor interactions.^16–19 Affinity-driven labeling uses reactive functional groups tuned to spontaneously react only when in close proximity to protein binding partners. Affinity-driven reactions capable of labeling native residues on receptors are especially valuable as they allow probing of ligand-receptor interactions without the need for prior genetic or chemical manipulation of the receptor. Covalent labeling of cell surface proteins in their native conformation by fluorescent probes is a powerful approach to study ligand-receptor interactions in a native cellular environment.^20–25 Prior research by the Hamachi group demonstrated the power of affinity-driven labeling for the study of membrane proteins in native contexts.^17,21,23,25 Moreover, work by the Wollscheid group demonstrated how affinity-driven labeling targeting chemically modified sugar molecules on the receptor surface can be used to discover new ligand-receptor interactions.^26–27

Aryl diazonium groups undergo electrophilic aromatic substitution reactions with aromatic groups (“azo coupling”), and electron withdrawing or donating substituents on the aryl diazonium group can tune this reactivity considerably for different applications.^28–38 Azo coupling reactions with deactivated aryl diazonium groups bearing para-alkyl substituents have been demonstrated to be selective for the proximity-induced labeling of tyrosine and histidine residues, and have proven to be useful to generate cyclic peptides via intramolecular cyclization.^28,39–40 Prior work by the Xia group also demonstrated the utility of a peptide bearing a deactivated aryl diazonium group for the labeling of a purified, soluble SH3 protein.²⁸ However, it is not known if affinity-driven azo coupling can be utilized in systems more complex than an isolated soluble protein (e.g., membrane proteins on surfaces of living cells), can be generalized to other protein targets, or can be applied for visualizing/detecting peptide-protein interactions in biological contexts.

We envisioned that deactivated aryl diazonium groups may serve as powerful tools for the affinity-driven labeling of peptide receptors on living cells to study peptide-receptor interactions in cellular contexts (Figure 1). Methods to target residues such as tyrosine via affinity-driven labeling would be especially valuable because these residues are particularly abundant in protein-protein interaction interfaces, relative to more traditionally targeted nucleophiles such as lysine and free cysteine. In fact, the high density of tyrosine and histidine residues found in the binding sites of known peptide receptors (Figure S1)⁴¹ led us to hypothesize that azo coupling may be especially well-suited for targeting these pockets without the need for genetic or chemical manipulation of the target receptor, and thus may be a relatively general method for targeting peptide receptors. Importantly, due to the relatively small size of the aryl diazonium group and its structural similarity to tyrosine and phenylalanine, we expected that incorporation of an aryl diazonium-bearing amino acid residue into peptide ligands may have minimal impact on receptor binding, even when positioned at certain key interaction sites. To explore these possibilities, here we develop and apply aryl diazonium-bearing peptides for affinity-driven protein labeling on the surface of living cells. We demonstrate that aryl diazonium-containing analogues of neuropeptide and peptide hormone ligands can form covalent bonds with unmodified G protein-coupled receptors (GPCRs), and that these bonds are formed through sequence-specific interactions in the peptide binding cavity. We show that our approach can be carried out in the complex environment of the cell surface, can be generalized across different receptors, can be used to visualize receptors on the surface of living cells, and can be applied to enrich and identify labeled receptors via mass spectrometry. Such affinity-driven labeling will serve as a powerful tool to characterize ligand-receptor interactions, including for the development of chemical probes to study peptide signaling, for monitoring receptor dynamics in living systems, and for identifying new peptide-receptor interactions.

Figure 1.

Cartoon representation of aryl diazonium affinity-driven labeling of membrane receptors. Peptides bearing a deactivated aryl diazonium group are incubated with live cells expressing target receptor. Upon receptor binding, proximity-enhanced reactivity facilitates azo coupling between the peptide and a native residue in the receptor (e.g., tyrosine or histidine). Labeled receptors can then be detected or enriched using tags incorporated into the synthetic peptide.

Results and Discussion

Labeling of NTSR1 by an aryl diazonium-containing NTS analogue on living cells.

To develop the affinity-driven aryl diazonium labeling approach for peptide receptors, we chose the interaction of neurotensin (NTS) with neurotensin receptor 1 (NTSR1) as a model system. NTS is a 13-residue peptide which acts as both a neurotransmitter and a peptide hormone. NTS signaling plays roles in a variety of physiological processes, including modulation of dopaminergic pathways, analgesia, feeding, cell survival and proliferation (including in cancers), and more.^42–45 NTS interacts with several receptor proteins, including acting as a high-affinity agonist for NTSR1, a class A GPCR. We chose the NTS-NTSR1 interaction to develop our approach because this system is well studied, with the availability of extensive structure-activity relationship data and high-resolution structures of ligand-receptor complexes.^46–51

To design aryl diazonium-containing NTS analogues predicted to label NTSR1, we began by examining the structures of ligand-bound NTSR1 (Figure S2).^46–47 The structures show that the C-terminus of NTS (Y11, I12, L13) extends deep into the peptide-binding cavity of NTSR1, and is in close proximity to a number of tyrosine and histidine residues in the receptor. Given the structural similarity of the aryl diazonium group and the tyrosine side chain, we predicted that replacing Y11 in the NTS peptide with 4-aminophenylalanine (a precursor to the aryl diazonium) may result in minimal perturbation of the NTS-NTSR1 binding interaction. The resulting peptide analogue (1P, Figure 2) makes this replacement, along with a Y3F substitution (to prevent intramolecular cyclization with the aryl diazonium) and an N-terminal alkyne (for optional modification via azide-alkyne cycloaddition) with a short linker. The ability of NTS analogues to activate NTSR1 was evaluated via a d-myo-inositol 1-monophosphate (IP1) accumulation assay^52–53 using CHO cells expressing NTSR1 (NTSR1-CHO cells) (Table S1 and Figures S3, S4). Glu1NTS is an analogue of NTS bearing a pyroglutamic acid to glutamic acid substitution at position 1 and has an EC₅₀ value of 4 nM in the IP1 accumulation assay. The 4-aminophenylalanine-containing peptide 1P had a virtually identical EC₅₀ value, indicating that the substitution of Y11 for 4-aminophenylalanine along with other additional modifications did not dramatically reduce the ability of this peptide to bind to and activate NTSR1. When treated with a sodium nitrite and trifluoroacetic acid (TFA) solution (pH ~1),²⁸ 1P reacted with a large excess of phenol to form an azo-bonded conjugate (Figure S5), indicating that the 4-aminophenylalanine-containing NTS analogue could be readily converted to an aryl diazonium-containing peptide capable of undergoing an azo coupling reaction. In contrast, Glu1NTS was unreactive to excess phenol under similar conditions (Figure S6). This is consistent with prior experiments showing that aryl amines can be selectively converted to diazonium groups at low pH, even in the presence of other potentially reactive functional groups such as the side chains of lysine, arginine, or a free N-terminus.³⁹

Figure 2.

(A) Primary sequences of peptide analogues used in this study. X = 4-aminophenylalanine, U = propargylglycine, FAM = 6-carboxyfluorescein, Btn = biotin, Ahx = 6-aminohexanoic acid, n = norleucine. Groups in parentheses indicate covalent attachment to the preceding lysine residue through the ε-amino group. (B) 4-aminophenylalanine residues in peptides can be treated with 0.1 M NaNO₂ and 1% TFA to generate the aryl diazonium group for labeling experiments.

We next evaluated the ability of the aryl diazonium-containing NTS analogue to label NTSR1 on living cells. To achieve this, we designed and synthesized a 4-aminophenylalanine-containing NTS analogue tagged with 6-carboxyfluorescein (6-FAM) on its N-terminus (1F) to allow for visualization using western blot and fluorescence microscopy. We treated 1F with sodium nitrite and TFA to form an aryl diazonium-containing peptide, quenched excess nitrite ions with sulfamic acid, and then added the peptide to NTSR1-CHO cells. The aryl diazonium-containing 1F (final peptide concentration of 10 μM) was allowed to react with NTSR1-CHO cells at pH 7.0–7.6 for 30 min at 4 °C in the dark. Lower temperatures were used to minimize ligand-induced receptor internalization, which is known for the NTS-NTSR1 interaction.^54–56 Following incubation with the ligand, the cells were washed to remove unbound peptide, lysed, and analyzed via western blot (Figure 3). These results reveal that aryl diazonium-containing 1F labels NTSR1, as indicated by a major band running between ~40–50 kDa in both anti-FAM and anti-NTSR1 western blots corresponding to the FAM-labeled peptide-NTSR1 conjugate. Conditions using cells lacking NTSR1, lacking sodium nitrite, or lacking peptide do not show the same band in the anti-FAM blot, demonstrating that peptide labeling requires NTSR1 expression and formation of the reactive aryl diazonium-containing peptide. Together, these results demonstrate that aryl diazonium-containing 1F is able to form a covalent bond with NTSR1 on the surface of living cells. Importantly, labeling by aryl diazonium-containing 1F was able to be blocked by the addition of untagged Glu1NTS as a competitor (Figure 4), strongly suggesting that 1F labels NTSR1 at the NTS-binding site. Finally, the 1F-NTSR1 conjugate formation was able to be reversed by treatment with sodium dithionite^{29–31,57–59} (Figure S7), consistent with the hypothesis that the aryl diazonium is forming an azo bond with residues in NTSR1. The ability of the azo bond to be cleaved selectively by certain reducing agents may make this linkage attractive for several downstream applications, including for simplifying MS/MS spectra during proteomics experiments, elution from immobilized resins, or for further derivatization of the resulting 3-aminotyrosine residues.^{29–31,57–59}

Figure 3.

Western blot analysis evaluating the ability of aryl diazonium-containing 1F to covalently label NTSR1 on living cells. A major band running between ~40–50 kDa in both the anti-FAM blot and the anti-NTSR1 blot illustrates covalent modification of NTSR1 with fluorescein-tagged peptide. This band is absent in the anti-FAM blot under negative control conditions using cells lacking NTSR1, using peptide not pretreated with sodium nitrite, or without any peptide. Two individual western blots (one for anti-FAM and one for anti-NTSR1) were performed on the same cell lysates (3.6 μg protein/lane), and each blot was stripped and reprobed for GAPDH as a loading control. Image representative of three independent experiments.

Figure 4.

Western blot analysis evaluating the ability of Glu1NTS to inhibit NTSR1 labeling by aryl diazonium-containing 1F on living cells. Two individual western blots (one for anti-FAM and one for anti-NTSR1) were performed on the same cell lysates (3.6 μg protein/lane), and each blot was stripped and reprobed for GAPDH as a loading control. Image representative of three independent experiments.

Sequence specificity of NTSR1 labeling.

To investigate the role of the aryl diazonium-containing peptide’s primary sequence in labeling NTSR1, we designed scrambled NTS analogues 2P (bearing alkyne) and 2F (bearing 6-FAM) (Figure 2). As with 1P/1F above, alkyne-bearing peptides were used for NTSR1 potency assays, while 6-FAM-bearing peptides were used for visualization of labeling by western blot. As expected, peptide 2P was a very weak agonist of NTSR1, showing some receptor activation only at the highest concentration tested in the IP1 accumulation assay (Table S1 and Figure S4). Similarly, 2F was a weak inhibitor of NTS peptide binding, as judged by a radioligand competition assay (Table S2). Consistent with these observations, aryl diazonium-bearing 2F did not effectively label NTSR1 on the surface of NTSR1-CHO cells (Figure 5A). This result suggests that the labeling of NTSR1 by 1F is a sequence-driven reaction, and the sole presence of an aryl diazonium cation in the peptide sequence is not responsible for the labeling of NTSR1. As a further test of the dependence of peptide sequence on labeling, we also evaluated two additional aryl diazonium-containing peptides unrelated to NTS: analogues of Aplysia allatotropin-related peptide (AT and ATF)⁶⁰ and analogues of pituitary adenylate cyclase-activating polypeptide (PA and PAF)⁶¹ (Figure 2). Both AT and PA showed virtually no activation of NTSR1 in the IP1 accumulation assays (Table S1 and Figure S4), and ATF was unable to inhibit NTS binding to NTSR1 by radioligand competition assay (Table S2). Consistently, neither aryl diazonium-containing ATF nor PAF showed detectable labeling of NTSR1 on live cells (Figure 5B), once again highlighting the importance of peptide sequence in driving the labeling by aryl diazonium-containing 1F. Together, the results that 2F, ATF, and PAF are unable to effectively label NTSR1 are consistent with a mechanism whereby labeling through aryl diazonium-containing 1F occurs primarily through a sequence-dependent and affinity-driven interaction. Further experiments investigating the effect of peptide sequence revealed a reasonable tolerance to some primary sequence modifications for effective labeling by NTS analogues (See note in Supporting Information, and Figures S8–S12).

Figure 5.

Western blot analysis evaluating the ability of (A) aryl diazonium-containing 1F or 2F, or (B) aryl diazonium-containing 1F, ATF, or PAF to covalently label NTSR1 on living cells. For each panel, two individual western blots (one for anti-FAM and one for anti-NTSR1) were performed on the same cell lysates (3.6 μg protein/lane), and each blot was stripped and reprobed for GAPDH as a loading control. Each panel image representative of two independent experiments.

Labeling of GLP1R by an aryl diazonium-containing exendin-4 analogue on living cells.

To evaluate the generalizability of our approach to other peptide-receptor systems, we designed an aryl diazonium-bearing derivative of exendin-4 (EXF, Figure 2). Exendin-4 is a potent agonist of the glucagon-like peptide 1 receptor (GLP1R), a class B GPCR which is an important therapeutic target for type 2 diabetes.^62–63 The C-terminal region of exendin-4 is known to interact with the extracellular domain of GLP1R and contributes the majority of the affinity for the receptor, while the N-terminal region is responsible for receptor activation.^64–67 EXF is an analogue of exendin-4(2–30) that lacks His1 (to prevent intramolecular cyclization with the aryl diazonium) and incorporates 6-FAM linked through the sidechain of Lys12, a position previously demonstrated to be amenable to fluorophore and biotin modification.^68–69 Based on the structure of exendin-4(9–39) bound to the GLP1R extracellular domain (Figure S13)⁶⁷, we chose to replace Phe22 of exendin-4 with 4-aminophenylalanine due to its close proximity to several tyrosine residues in the extracellular domain of GLP1R. Cell-based GLP1R activity experiments (Figure S14) demonstrate that exendin-4(2–30) analogues are somewhat weaker agonists of the receptor, as expected due to the removal of His1 (EC₅₀ ~ 1 nM for analogues lacking His1 versus ~0.01 nM for an exendin-4(1–30) analogue). The fact that exendin-4(2–30) analogues are able to potently activate the receptor indicates that they still bind GLP1R, consistent with the fact that exendin-4 analogues bearing N-terminal truncations retain high affinity for the receptor.^64–67 Importantly, GLP1R activation is not negatively impacted by the replacement of Phe22 with 4-aminophenylalanine (Figure S14), again highlighting the potential advantage of this relatively modest modification for affinity-driven labeling in peptide ligands.

To evaluate the ability of EXF to label GLP1R on living cells, HEK293 cells stably expressing GLP1R were treated with aryl diazonium-containing EXF, washed to remove unbound peptide, lysed, and examined by anti-FAM and anti-GLP1R western blots (Figure 6). We observed covalent labeling of GLP1R by EXF (Figure 6A), evident by a FAM-labeled band running ~3 kDa higher than GLP1R on the blot, as expected for the receptor conjugated to EXF (which has a molecular mass of ~3.6 kDa). This band could be confidently assigned as being labeled GLP1R, as this FAM-labeled band was absent in HEK293 cells that did not express GLP1R. Importantly, EXF did not label NTSR1 in the NTSR1-CHO cell line and the NTS analogue 1F did not label GLP1R in the GLP1R-expressing cell line (Figure 6B), again highlighting the sequence specificity of the affinity-driven azo coupling reactions. Together, these results demonstrate that the ability to apply aryl diazonium-bearing peptides to label membrane-bound receptors on living cells is not limited to the NTS-NTSR1 system but can be readily applied to other unrelated peptide-receptor systems from different GPCR subclasses.

Figure 6.

(A) Western blot analysis evaluating the ability of EXF to label GLP1R covalently on living cells. A major band running between ~40–50 kDa in both the anti-FAM blot and the anti-GLP1R blot illustrates covalent modification of GLP1R with fluorescein-tagged peptide. Labeling reactions took place at 37 °C. Two individual western blots (one for anti-FAM and one for anti-GLP1R) were performed on the same cell lysates (5.0 μg protein/lane), and each blot was stripped and reprobed for GAPDH as a loading control. (B) Western blot analysis evaluating the specificity of EXF or 1F for their respective receptors, GLP1R or NTSR1. Labeling reactions took place at 4 °C. Three individual western blots (one for each probe) were performed on the same cell lysates (5.0 μg protein/lane). Each panel image representative of two independent experiments.

Evaluation of membrane protein labeling by confocal microscopy.

We next evaluated the utility of affinity-driven azo coupling for live-cell receptor imaging. CHO cells (with or without NTSR1 expression) were treated with aryl diazonium-modified 1F for 30 min on ice. To dissociate non-covalently bound peptide, cells were then washed, incubated in fresh media on ice for 15 min, and then washed again prior to imaging via live-cell confocal fluorescence microscopy (Figure 7). The results of these experiments show that fluorescein-tagged peptide signal on the plasma membrane can be clearly visualized under conditions in which aryl diazonium-bearing 1F is exposed to cells expressing NTSR1 (Figure 7B). In contrast, minimal fluorescein signal is seen on cell membranes in the absence of peptide (Figure 7A), when 1F was not converted to an aryl diazonium (Figure 7C), or on cells that do not express NTSR1 (Figure 7D). These results demonstrate that NTSR1 covalently labeled through an aryl diazonium group can be effectively visualized on the cell surface, while non-covalent binding is significantly reduced after washing and a short incubation in fresh media. Similarly, covalent labeling of GLP1R could be visualized on the surface of HEK293 cells using aryl diazonium-bearing EXF (Figure S15), demonstrating the generalizability of this approach. The relative specificity of labeling suggests affinity-driven aryl diazonium labeling may be useful to study protein localization, dynamics, and functions in living cells.

Figure 7.

Fluorescent labeling of cell surface NTSR1 on CHO cells by 5 μM FAM-labeled NTS analogue 1F (green) at pH 7.2–7.4. Cell nuclei were stained blue (Hoechst 33342) and the plasma membrane was stained red. (A) NTSR1-CHO cells in the absence of 1F. (B) NTSR1-CHO cells incubated with aryl diazonium-containing 1F for 30 min on ice. (C) NTSR1-CHO cells incubated with 1F for 30 min on ice without treatment with sodium nitrite (no aryl diazonium formed). (D) CHO cells without NTSR1 expression incubated with aryl diazonium-containing 1F for 30 min on ice. Scale bar, 50 μm. Image representative of three independent experiments.

Enrichment and identification of labeled proteins by mass spectrometry-based proteomics.

We next aimed to investigate the utility of affinity-driven aryl diazonium labeling to facilitate enrichment and identification of labeled membrane proteins. Peptide 1B is a 4-aminophenylalanine-bearing derivative of NTS containing a biotin group at its N-terminus to facilitate enrichment (Figure 2). NTSR1-CHO cells were treated with aryl diazonium-containing 1B, washed to remove unbound peptide, lysed, and cell lysates incubated with streptavidin beads to enrich labeled biotinylated proteins. To evaluate the specificity of labeling, we also included conditions in which cells were incubated either: 1) with no 1B peptide, or 2) with both aryl diazonium-containing 1B and excess Glu1NTS as a competitor. After extensive washing, enriched proteins were eluted from the beads and analyzed by liquid chromatography-mass spectrometry (LC-MS)-based bottom-up proteomics. From this experiment, we confidently detected NTSR1 enriched from cells treated with aryl diazonium-containing 1B, but not from cells that were not treated with 1B nor from cells pretreated with excess Glu1NTS prior to the addition of aryl diazonium-containing 1B (Figure S16, Supporting Document 1, and Discussion in Supporting Information). Importantly, these results confirm NTSR1 as a major protein labeled by 1B (consistent with western blot and confocal microscopy data for 1F) and the only plasma membrane protein specifically labeled by this peptide (competed away by excess Glu1NTS). These findings demonstrate that membrane proteins labeled by affinity-driven azo coupling can be effectively enriched and identified by mass spectrometry. In addition, the low number of proteins enriched by 1B indicates a relatively high level of selectivity for this labeling and enrichment approach. Overall, these results demonstrate that affinity-driven aryl diazonium-based enrichment can be applied to detect peptide-receptor interactions and may be useful in future studies to study interactions for understudied peptide ligands.

Open questions and limitations.

There are several important unanswered questions and limitations worth noting. First, some labeling of proteins other than NTSR1 was observed for peptides via western blot (see full blot images in the Supporting Information). This background was unable to be competed away by unmodified Glu1NTS, indicating this as a level of non-specific labeling. However, the level of labeling for non-specific proteins is relatively low, as minimal labeling was seen in confocal microscopy experiments in cells that were not transfected with NTSR1 (Figure 7) and few proteins were enriched and detected by proteomics (Supporting Document 1). Importantly, cross-reactivity experiments (Figure 6B) show that EXF did not label NTSR1 and 1F did not label GLP1R, highlighting that the labeling of these GPCRs occurs via sequence-specific interactions. Furthermore, because competition experiments with unmodified ligand can be used to identify specific labeling, future experiments applying this approach to other systems should easily be able to distinguish specific versus non-specific labeling. Such an approach has previously been utilized for affinity-driven labeling of membrane-bound glycoproteins to identify novel ligand-receptor interactions.^{26–27,70–76} Second, we noticed that receptor expression levels played an important role in visualizing peptide-receptor conjugate formation when using western blot analysis. For the NTSR1 system, we found the best success using cells selected for NTSR1 expression (NTSR1-CHO cells) via G418-based selection, although labeling could also be seen via western blot in transiently transfected cells (albeit with more variability). The need for overexpression appears to be important for western-blot based analysis, but can likely be overcome using enrichment and mass spectrometry-based proteomic identification. Finally, our current results do not reveal the exact residues in the receptors forming covalent bonds in our peptide-receptor conjugates. Future mass spectrometry-based studies will allow for more precise localization of receptor labeling at the residue level, which may be useful in certain contexts. Nevertheless, this knowledge is not required for studying peptide-receptor interactions via western blot, confocal microscopy, or enrichment and identification via proteomics, as we have done here.

Conclusions

We have shown that peptides modified with a deactivated aryl diazonium group can be harnessed to label membrane-bound GPCRs on living cells. Our results demonstrate that aryl diazonium-containing analogues of NTS are able to covalently label NTSR1 on the cell surface in a sequence-specific and affinity-driven manner. This approach was not limited to NTSR1 (a class A GPCR) and could also be readily extended to label GLP1R (a class B GPCR) using an aryl diazonium-bearing exendin-4 analogue. Confocal fluorescence microscopy experiments show that aryl diazonium-containing peptides could be used to visualize functional NTSR1 and GLP1R receptors on the cell surface, and a biotinylated aryl diazonium-containing ligand could be used to label and enrich the labeled receptor for identification by proteomics. A rigorous evaluation of specificity of labeling towards target proteins, including competition experiments and cross-reactivity experiments provide strong evidence that labeling of GPCRs occurs in an affinity-driven manner. Some advantages of this approach include the selective targeting of tyrosine and histidine residues (which are enriched in peptide binding sites), the ability to label membrane-bound receptors in the complex environment of the living cell surface, the generalizability of the approach for different peptide-receptor systems, and the ability to do so without the incorporation of any non-natural group or tag into the receptor itself. The ability to reverse some azo bonds formed between the peptide and the receptor via sodium dithionite may prove beneficial for future applications,^57–59 as labeled tyrosine residues in the receptor will be reduced to 3-aminotyrosine, which will greatly simplify MS/MS spectra and can be easily identified during proteomics analysis, or can be further functionalized for downstream analysis.^29–31 In addition, the small size of the aryl diazonium derivative utilized here and its structural similarity to tyrosine/phenylalanine are especially attractive features of this approach, as this reactive group can be easily incorporated into peptide ligand sequences with relatively minimal perturbation of peptide structure. This may be an advantage over other modification approaches in which the reactive group is much larger. Indeed, for both NTS and exendin-4, it was relatively straightforward to design analogues bearing the aryl diazonium group at core interaction sites that had minimal impact on binding and were able to label their respective receptors effectively. In all, affinity-driven azo coupling to unmodified receptors should prove useful for monitoring receptors on living cells (e.g., receptor trafficking and recycling), for the development of novel covalent tools as chemical probes (e.g., covalent agonists or antagonists), or to identify new peptide-receptor interactions.

Materials and Methods

General Materials.

Unless otherwise specified, all reagents and solvents were purchased from ThermoFisher Scientific or MilliporeSigma.

Peptide synthesis.

Peptides were synthesized using Fmoc solid phase peptide synthesis (SPPS) either manually or on a Prelude X peptide synthesizer (Gyros Protein Technologies). Fmoc-protected amino acids were purchased from Novabiochem or Anaspec. Fmoc-4-(Boc-amino)-l-phenylalanine (03755), Fmoc-l-Lys(ivDde)-OH (07109), Fmoc-l-propargylglycine (05138), and d-biotin (00033) were obtained from ChemImpex. 6-Carboxyfluorescein (6-FAM) (AS-81004) was obtained from AnaSpec. Fmoc-6-aminohexanoic acid (852053) was obtained from Novabiochem. Peptides with C-terminal amides were synthesized on NovaPEG Rink Amide resin (Novabiochem, 855047). Peptides with C-terminal carboxylic acids were synthesized on corresponding preloaded Wang resin (Novabiochem). Prior to synthesis, all resins were swelled in dichloromethane for a minimum of 60 min. Fmoc deprotection reactions were carried out using 20% piperidine in N,N-dimethylformamide (DMF) (v/v) for 15 min at room temperature with constant stirring. For standard amino acid couplings, reactions were performed using 6 eq. of amino acids activated with 5.4 eq. of O-(1H-6-Chlorobenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HCTU) and 12 eq. of N-methylmorpholine (NMM) in DMF. For uncommon amino acid monomers (Fmoc-4-(Boc-amino)-l-phenylalanine, Fmoc-6-aminohexanoic acid, Fmoc-l-Lys(ivDde)-OH) and enrichment/labeling tags (Fmoc-l-propargylglycine, d-Biotin, 6-FAM) a minimum of 4 eq. of amino acid activated with 3.4 eq. of HCTU and 8 eq. of NMM in DMF were used for coupling reactions. Coupling reactions were performed for 30–40 min at room temperature with constant stirring. For Fmoc-l-propargylglycine, d-biotin, and 6-FAM, coupling reactions were allowed to proceed overnight at room temperature with constant stirring. For peptides containing 6-FAM, precautions were taken to avoid light exposure during synthesis, purification, and storage. For several peptides, biotin or 6-FAM were coupled through the ε-amino group of lysine by incorporating Lys(ivDde) into the linear chain ending with an N-Boc-protected residue at the N-terminus. Deprotection of ivDde was then carried out using freshly prepared 4% hydrazine in DMF (v/v) for a minimum of 4 × 1 h treatments for PA and PAF and a maximum of 4 × 15 min treatments for EXF. Peptides were cleaved from the resin using a cleavage cocktail containing 95% trifluoroacetic acid (TFA), 2.5% H₂O, and 2.5% triisopropylsilane for 3.5 h to 7 h. Cleaved peptides were precipitated with cold diethyl ether, and centrifuged (3000 × g, 10 min, 4 °C). The resulting pellet was dissolved in a 50% acetonitrile (ACN)/H₂O mixture, and filtered through 0.22 μm syringe filter (Fisher Scientific, 13-100-102) before being purified using a Pursuit 5 μm, 200 Å, 250 × 21.2 mm C18 column (Agilent) by preparatory-scale reversed-phase HPLC at a flow rate of 12 mL/min. The purified peptides were lyophilized and stored at −20 °C. After purification, final peptide purity was assessed via analytical-scale reversed-phase HPLC using a Pursuit 3 μm, 200 Å, 150 × 4.6 mm C18 column (Agilent). The polarity of solvent B was increased in a linear gradient from 5 to 55% over the course of 50 minutes with a constant flow rate of 1 mL/min. Solvent A (H₂O + 0.1% TFA) and Solvent B (ACN + 0.1% TFA) were used as the mobile phase for both preparative and analytical scale HPLC. Peptide peaks were monitored at 220 nm (amide) and 280 nm (tyrosine, tryptophan, 4-aminophenylalanine). The identity of each peptide was confirmed via matrix-assisted laser desorption/ionization mass spectrometry (MALDI MS) using either a Bruker SolariX XR Fourier transform ion cyclotron resonance mass spectrometer or a Bruker Autoflex maX time-of-flight mass spectrometer (Figure S17).

Stock solutions for all peptides were prepared in water, except for EXF which was dissolved in dimethyl sulfoxide (DMSO). Purified peptides were quantified using UV-Visible absorbance. For quantification, peptides containing 6-FAM were dissolved in 10 mM Tris (pH 8.0), and quantified using ε_495nm = 68,000 cm⁻¹ M⁻¹.⁷⁷ Peptides without 6-FAM were dissolved in 10 mM Tris (pH 8.0) and quantified using ε_280nm calculated from Gill et al.⁷⁸, with ε_280nm = 1,430 cm⁻¹ M⁻¹ for 4-aminophenylalanine (see Supporting Information).

Cell culture.

CHO cells (CHO-K1, ATCC, CCL-61) were cultured in Ham’s F-12K media (ThermoFisher Scientific, 21127022) supplemented with 10% fetal bovine serum (FBS) (VWR, 97068–085) and 100 U/mL penicillin with 100 μg/mL streptomycin (Cytiva, SV30010) at 37 °C with 5% CO₂. CHO cells stably expressing human NTSR1 (NTSR1-CHO) were generated using a pcDNA3.1(+) plasmid encoding NTSR1 (cDNA Resource Center, #NTSR100000) under selection using 800 μg/mL G418 sulfate (ThermoFisher Scientific, 11811031). HEK293GS cells (HEK293 cells stably expressing the GloSensor plasmid⁷⁹) were cultured in DMEM (high glucose) media (ThermoFisher Scientific, 11965092) supplemented with 10% FBS at 37 °C with 5% CO₂. GLP1R-HEK293GS cells were cultured either in EMEM media (ATCC, 302003) supplemented with 10% FBS and 50 μg/mL zeocin (InvivoGen, ant-zn-1p) (for cAMP assays) or DMEM (high glucose) media supplemented with 10% FBS and 50 μg/mL zeocin (for live-cell labeling) at 37 °C with 5% CO_2.

IP1 accumulation assays.

NTSR1 activation assays were carried out using IP-One Gq kit (Cisbio, 62IPAPEC).^52–53 On the day prior to the experiment, NTSR1-CHO cells were seeded in an opaque bottom, cell culture-treated 96 well half-area plate (Fisher Scientific, 07-200-309) at a density of 5000 cells/well and maintained at 37 °C, 5% CO₂ overnight. The next day, the media from each well was carefully removed and replaced with 28 μL of appropriate peptide dilution in stimulation buffer. The plate was then incubated at 37 °C, 5% CO₂ for 30–60 min. After stimulation, the plate was removed from the incubator and to each well was added 6 μL of 0.5× IP1-d2 followed by 6 μL of 0.5× anti-IP1-cryptate (both provided with the kit). The plate was then incubated in the dark at room temperature on an end-to-end rocker for 60 min. Homogenous time-resolved fluorescence was read on a BioTek Synergy Neo22 plate reader (excitation = 330/80 nm, emission = 665/8 nm and 620/10 nm). The 665 nm/620 nm fluorescence ratio for wells containing IP1 standards was fit to a four-parameter dose-response model in GraphPad Prism 9. The resulting standard curve was then used to determine the concentration of IP1 in each experimental well, and the response for each peptide was fit to a three-parameter dose-response model in GraphPad Prism 9. Agonists that did not reach the top of the curve in a given experiment were recorded as having an EC₅₀ value greater than the maximum concentration tested. Experiments with intra-assay coefficient of variation <25% were used to calculate logEC₅₀ values (mean and standard deviation) among independent experiments.

cAMP accumulation assays.

GLP1R activation assays were carried out using a cAMP Gs HiRange kit (Cisbio, 62AM6PEC). On the day prior to the experiment, GLP1R-HEK293GS cells cultured in EMEM media supplemented with 10% FBS and 50 μg/mL zeocin were seeded per well in an opaque bottom, cell culture-treated 96 well half-area plate (Fisher Scientific, 07-200-309) at a density of 2000 cells/well and maintained at 37 °C, 5% CO₂ overnight. The next day, the media from each well was carefully removed and replaced with 20 μL of appropriate peptide dilution in stimulation buffer supplemented with 500 μM IBMX. The plate was then incubated at 37 °C, 5% CO₂ for 60 min. After stimulation, the plate was removed from the incubator and to each well was added 10 μL of 1× cAMP-d2 followed by 10 μL of 1× anti-cAMP-cryptate (both provided with the kit). The plate was then incubated in the dark at room temperature on an end-to-end rocker for 60 min. Homogenous time-resolved fluorescence was read on a BioTek Synergy Neo22 plate reader (excitation = 330/80 nm, emission = 665/8 nm and 620/10 nm). The 665 nm/620 nm fluorescence ratio for each peptide was fit to a three-parameter dose-response model in GraphPad Prism 9.

Live cell labeling reactions.

Two days prior to an experiment, NTSR1-CHO cells or CHO cells were seeded in 35 mm tissue culture dishes (0.5×10⁶ – 0.7×10⁶ cells/dish) for NTSR1 labeling experiments and maintained at 37 °C, 5% CO₂. For GLP1R labeling experiments, GLP1R-HEK293GS cells or HEK293GS cells were seeded in poly-d-Lysine (ThermoFisher Scientific, A3890401) coated 35 mm tissue culture dishes (0.5×10⁶ cells/dish) and maintained at 37 °C, 5% CO₂. After 48 h recovery, the media in each of the cell dishes was exchanged with 975–980 μL of Opti-MEM (Fisher Scientific, 31-985-070) and the dishes were incubated at 4 °C for NTSR1-CHO cells and CHO cells or at 37 °C, 5% CO₂ for GLP1R-HEK293GS cells and HEK293GS cells for 20 min prior to the addition of the aryl diazonium-containing peptides. For competition experiments, NTSR1-CHO cells were incubated with untagged Glu1NTS for 30 min at 4 °C prior to the addition of aryl diazonium-containing peptides. For specificity experiments shown in Figure 6B, NTSR1-CHO cells and GLP1R-HEK293GS cells were pre-incubated at 4 °C prior to the addition of aryl diazonium-containing peptides.

To generate aryl diazonium-containing peptides, freshly prepared aqueous stock solutions of TFA and sodium nitrite (NaNO₂) were mixed with 4-aminophenylalanine-containing peptide (final concentrations: 0.1 M NaNO₂, 1% TFA, pH ~1) and the reaction was allowed to proceed on ice for 20 min in the dark. Following the diazotization reaction, a freshly prepared aqueous solution of sulfamic acid (NH₂SO₃H) was added to the reaction (final concentrations: 0.1 M NH₂SO₃H) to quench excess nitrite ions. The reaction mixture was quickly vortexed and centrifuged, and 20–25 μL of this solution was added to the dishes of either pre-cooled NTSR1-CHO cells or CHO cells (final peptide concentration: 10 μM peptide, final pH ~7.0–7.6) or pre-warmed GLP1R-HEK293GS cells or HEK293GS cells (final concentration: 1 μM peptide, 0.1% DMSO, pH ~7.0–7.6). The azo coupling reaction was allowed to proceed for 30 min at 4 °C for NTSR1-CHO cells and CHO cells or 37 °C, 5% CO₂ for GLP1R-HEK293GS cells and HEK293GS cells in the dark. The reaction mixture was then aspirated and cells were washed twice with phosphate buffered saline (PBS), pH 7.4. Cells were lysed using 500 μL of cold RIPA lysis and extraction buffer (ThermoFisher Scientific, 89900) containing 1× protease inhibitor cocktail (MilliporeSigma, 5351401SET). Following cell detachment and lysis, cell lysates were incubated on ice for at least 15–20 min, briefly sonicated (2 × 30 sec), and centrifuged (14,000–15,000 × g, 10–15 min, 4 °C). The cell lysate supernatant was collected, and total protein concentration was determined using a BCA assay (ThermoFisher Scientific, 23235). Cell lysates were diluted to 0.4–1.0 μg/μL in RIPA with 1× protease inhibitor cocktail and stored at −20 °C. Prior to analysis by western blot, cell lysates were treated with a minimum of 500 units of PNGase F (New England Bio Lab, P0709L) with the addition of 1% NP-40 (New England Bio Lab, P0709L), and 715 mM 2-mercaptoethanol (ThermoFisher Scientific, 125470100) for 60 min at room temperature.

Western blots.

Cell lysates (3.5–5 μg/lane) were separated by SDS-PAGE using precast 4–12% polyacrylamide gels (ThermoFisher Scientific, NW04120BOX) along with SeeBlue Plus2 Pre-stained Protein ladder (ThermoFisher Scientific, LC5925). The proteins were then transferred to a 0.45 μm nitrocellulose membrane (ThermoFisher Scientific, LC2001) using wet electroblotting (10 V, 1 h). The membranes were blocked using SuperBlock T20 blocking buffer (ThermoFisher Scientific, 37536) either for 1 h at room temperature or overnight at 4 °C. The membrane was then washed with PBS-Tween (PBST) (2 × 5 min), and primary antibody incubation was performed using a 1:1000–1:2000 dilution in PBST with 1% bovine serum albumin (BSA) for 1 h at room temperature or overnight at 4 °C. For anti-FAM blots, rabbit anti-fluorescein antibody (Abcam, ab19491) was used. For anti-NTSR1 blots, rabbit anti-NTSR1 antibody (Abcam, ab183088) was used. For anti-GLP1R blots, rabbit anti-GLP1R antibody (Proteintech, 26196–1-AP) was used. The membranes were then washed with PBST (2 × 5 min), followed by incubation with 1:10,000 dilution of goat anti-rabbit IgG H&L (HRP) antibody (Abcam, ab205718) in PBST with 1 % BSA for 30 min at room temperature. The membranes were washed with PBST (2 × 5 min) and developed using SuperSignal West Pico PLUS chemiluminescent substrate (ThermoFisher Scientific, 34580). Blots were imaged using iBright CL 1500 Imaging system. To perform glyceraldehyde 3-phosphate dehydrogenase (GAPDH) blotting on the same membrane, anti-fluorescein, anti-NTSR1, or anti-GLP1R blotted membranes were washed with PBST (2 × 10 min) and stripped by incubation in Restore PLUS western blot stripping buffer (ThermoFisher Scientific, 46430) for 45–90 min at 37–50 °C followed by an additional 20–30 min at room temperature. The blots were washed with PBST (2 × 10 min) and then blocked, as previously described. Blotting for GAPDH was performed using a 1:1000–1:10,000 dilution of rabbit anti-GAPDH antibody (Abcam, ab181602) in PBST with 1% BSA, and incubating either for 1 h at room temperature or overnight at 4 °C. Membranes were washed with PBST (2 × 5 min) and probed with a 1:10,000 dilution of goat anti-rabbit IgG H&L (HRP) antibody in PBST with 1% BSA for 30 min at room temperature. Membranes were washed, developed, and imaged, as described above. During optimization, we determined that small amounts of carryover/incomplete stripping prevented performing both an anti-FAM blot and anti-receptor blot on the same membrane. To avoid false positive signals, two individual membranes, one for identifying receptor expression and another for identifying FAM-labeled proteins, were prepared for each experiment. Blot images were cropped and then assembled in Adobe Illustrator. Full blot images can be seen in the Supporting Information (Figures S18–S68).

Confocal fluorescence microscopy.

NTSR1-CHO or CHO cells were seeded at a density of 150,000–250,000 cells per 35 mm glass bottom dish (Mattek, P35G-1.5–20-C) and incubated at 37 °C, 5% CO₂ for 48 h. On the day of the experiment, media in cell dishes was exchanged with fresh room temperature Opti-MEM. Cells were incubated for 15 min at room temperature in the dark with Hoechst 33342 (ThermoFisher Scientific, R37605) and an additional 30 min with the plasma membrane stain (1:1000 dilution in Opti-MEM) (ThermoFisher Scientific, C10046) to stain the cell nuclei and plasma membrane, respectively. The staining solution was removed, and the cells were washed once with cold PBS. Fresh 975 μL Opti-MEM was added to the cells and they were incubated on ice for an additional 20 min. Peptide diazotization of NTS analogue 1F and azo coupling reaction on cells was performed as described above (final peptide concentration: 5 μM, final pH 7.0–7.4). Following the azo coupling reaction, cells were washed twice with PBS, pH 7.4 to remove excess reaction mixture and incubated for an additional 15 min in cold Opti-MEM on ice in the dark. Cells were washed twice with PBS, pH 7.4 and immediately imaged in Opti-MEM at room temperature with a 60× microscope objective lens on a Nikon A1R-Ti2 confocal system, using sequential excitation for Hoeschst 33342 (blue), FAM (green), and CellMask (red) channels. All confocal images were saved in the ND2 format and read using ImageJ with the Bio-Formats plugin. All images were converted to RGB type, and the scale bar was added using ImageJ. Images were saved as TIFF files without any additional modifications and assembled in Adobe Illustrator. For details on confocal fluorescence microscopy for the GLP1R system, refer to the Supporting Methods section in the Supporting Information.

Enrichment and LC-MS-based proteomics of labeled proteins.

The live cell labeling reaction and the diazotization of peptide was performed as described above. Briefly, two days prior to an experiment, NTSR1-CHO cells were seeded in 150 mm tissue culture dishes (3 × 10⁶ cells/dish, 5 dishes per experimental condition) and maintained at 37 °C, 5% CO₂. After 48 h, the media in each of the cell dishes was exchanged with 9875 μL of Opti-MEM and the dishes were incubated at 4 °C for 20 min. For the competition experiment (+1B +Glu1NTS), dishes were pre-incubated with 48 μL Glu1NTS (final concentration: 24 μM) at 4 °C for >30 min before the addition of the aryl diazonium-containing peptide. For all other conditions, 48 μL of water was added and the dishes were incubated at 4 °C for the same duration. After incubation with Glu1NTS or water, 77 μL of the diazotized peptide reaction mixture, quenched with sulfamic acid (final concentration: 0.1 M NH₂SO₃H), was added to the pre-cooled NTSR1-CHO cells for dishes with the “+1B” and “+1B +Glu1NTS” conditions. The azo coupling reaction was allowed to proceed for 30 min at 4 °C in the dark (final concentration: 8 μM peptide 1B, final pH ~7.4). For the control with no peptide (−1B), an equal amount of dummy diazotization reaction mixture was added to pre-cooled NTSR1-CHO cells and the dishes were incubated for 30 min at 4 °C in the dark. The reaction mixture for all three conditions was then aspirated, and cells were washed three times with PBS, pH 7.4. Two biological replicates were performed for each condition. Cells were dissociated from the tissue culture dishes using Versene solution (ThermoFisher Scientific, 15040066) and collected using PBS, pH 7.4. Collected cells were centrifuged at 800 × g, 3 min, room temperature. The supernatant was removed, and the cell pellet was resuspended in 1 mL of RIPA supplemented with 1× protease inhibitor cocktail, benzonase nuclease (1:1000 v/v) (EMD Millipore, 71205–25KUN) and 1.5 mM MgCl_2. Cell lysates were transferred to low-protein-binding microcentrifuge tubes_, incubated on ice for 40 min with intermittent vortexing, sonicated (2 × 1 min), and centrifuged (15,000 × g, 30 min, 4 °C). The supernatant was collected, and total protein concentration was determined using a BCA assay. Cell lysates (6000 μg per sample) were then treated with PNGase F (16,350 units) with the addition of 1% NP-40 and 715 mM 2-mercaptoethanol for 90 min at room temperature.

Prior to incubation with cell lysates, a 350 μL slurry of magnetic streptavidin beads (ThermoFisher Scientific, 65605D) per sample was transferred to low-protein-binding microcentrifuge tubes and incubated at room temperature for 20 min on an end-to-end rocker. Excess supernatant from the streptavidin bead suspension was removed and the beads were washed 3 × 500 μL RIPA buffer. Washed streptavidin beads were then incubated with PNGase F treated cell lysates (6000 μg) overnight at 4 °C under rotation. The next day, supernatant from the streptavidin-conjugated beads was removed and beads were resuspended in 1 mL of 1% sodium dodecyl sulfate (SDS) washing buffer (1% SDS, 100 mM Tris, 250 mM NaCl, 5 mM EDTA pH 8.0) and transferred to fresh low-protein-binding microcentrifuge tubes. Bead bound proteins were reduced with 5 mM tris(2-carboxyethyl)phosphine for 30 min at room temperature. Beads were then collected and the supernatant was removed. Next, the beads were again resuspended in 1 mL SDS washing buffer and alkylated with 40 mM iodoacetamide for 30 min at room temperature in dark. Beads were again collected, and the supernatant was removed.

Beads were then resuspended in 1 mL of SDS wash buffer (pH 8.0) and transferred to glass vials. Stringent washing²⁷ of the streptavidin conjugated beads was performed by washes with 10 × 1 mL SDS wash buffer (pH 8.0), 10 × 1 mL 8 M urea in 100 mM Tris (pH 8.0), 5 × 1 mL 5 M NaCl in PBS (pH 7.2–7 .4), 3 × 500 μL 80% isopropanol in water, 5 × 1 mL 100 mM NaHCO₃ (pH 11), 5 × 1 mL 50 mM aqueous ammonium bicarbonate (freshly prepared), 5 × 1 mL 20% acetonitrile, and 5 × 1 mL PBS (pH 7.4). This washing sequence was then repeated an additional time. LC-MS-grade water and acetonitrile were used in all washing solutions. The beads were then resuspended in a minimum amount of PBS (pH 7.4) and frozen at −20 °C until analysis by mass spectrometry.

Proteomics was carried out by the Proteomics & Metabolomics Facility, Nebraska Center for Biotechnology at the University of Nebraska-Lincoln. The streptavidin-conjugated beads were split into two equal aliquots and one aliquot of each sample was heated in 50 μL 3× NuPAGE LDS sample buffer (ThermoFisher Scientific) containing 2 mM biotin and 15 mM DTT for 15 min at 95 °C. The eluted proteins were then run into the top of a 12% Bolt Bis-Tris Plus mini protein gel before being fixed and subsequently stained with colloidal coomassie blue stain. The stained protein gel lanes were excised, cut up and washed 3 times with ammonium bicarbonate/acetonitrile to remove SDS and stain. Trypsin (250 ng, Promega, V5111) was added and digestion was carried out overnight at 37 °C. Peptides were extracted from the gel pieces and dried down frozen in a Speed-Vac. The dried digests were redissolved in 15 μL of 5% acetonitrile, 0.05% TFA, and 5 μL of this solution was injected per run.

Analysis was carried out using a 2 h run on an Acquity UPLC M-Class Peptide CSH C18 column, 1.7μm, 130Å, 0.075 mm × 250 mm (Waters Corp, Milford, MA) feeding into a high sensitivity Orbitrap Eclipse mass spectrometer run in high resolution OT-OT-HCD mode. All MS/MS samples were analyzed using Mascot (Matrix Science, London, UK; version 2.7.0). Mascot was set up to search 3 databases together: the cRAP common contaminants database (dated 20150130; theGPM.org); a Uniprot (www.uniprot.org/) reference proteome for organism number UP000001075, Chinese hamster, Cricetulus griseus, CHO K1 cell line (dated 20231120 and containing 23,886 sequences); and a custom made database containing only the NTR1_HUMAN protein sequence (downloaded 20231120). The digestion enzyme was trypsin. Mascot was searched with a fragment ion mass tolerance of 0.060 Da and a parent ion tolerance of 10.0 ppm. Deamidation of asparagine and glutamine, oxidation of methionine, and carbamidomethyl of cysteine were specified in Mascot as variable modifications.

Scaffold (version 5.3.0, Proteome Software Inc., Portland, OR) was used to validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 80.0% probability by the Peptide Prophet algorithm⁸⁰ with Scaffold delta-mass correction. Protein identifications were accepted if they could be established at greater than 99.9% probability and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm.⁸¹ Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Proteins sharing significant peptide evidence were grouped into clusters.