Intracellular Bioorthogonal Labeling of Glucagon Receptor via Tetrazine Ligation

Abstract

The third intracellular loop (ICL3) in the cytosolic face of glucagon receptor (GCGR) experiences significant conformational transition during receptor activation. It thus offers an attractive site for the introduction of organic fluorophores in our efforts to construct fluorescence-based GPCR biosensors. Herein, we report our confocal microscopic study of intracellular fluorescent labeling of ICL3 using a bioorthogonal chemistry strategy. Our approach involves the site-specific introduction of a strained alkene amino acid into the ICL3 through genetic code expansion, followed by a highly specific inverse electron-demand Diels-Alder reaction with the fluorescent tetrazine probes. Among the three strained alkene amino acids examined, both SphK and 2’-aTCOK offered successful fluorescent labeling of GCGR ICL3 with the appropriate tetrazine probes. At the same time, 4’-TCOK gave high background fluorescence due to its intracellular retention. The fluorescent tetrazine probes were designed following a computational model for background-free intracellular fluorescent labeling; however, their performance varied significantly in live-cell imaging as the strong non-specific signals interfered with the specific ones. Among all GCGR ICL3 mutants bearing a strained alkene, the H339SphK/2’-aTCOK mutants provided the best reaction partners for the BODIPY-Tz1/4 reagents in the bioorthogonal labeling reactions. The results from this study highlight the challenges in identifying bioorthogonal reactant pairs suitable for intracellular labeling of low-abundance receptors in live-cell imaging studies.

1. Introduction

The class B G protein-coupled receptors (GPCRs) comprise 15 receptors in humans in response to stimulation by secretin, glucagon, glucagon-like peptide 1 (GLP-1), corticotropin-releasing factor, parathyroid hormone, calcitonin gene-related peptide, and other peptide hormones. These receptors are validated drug targets for treating many diseases, including diabetes, depression, and osteoporosis.¹ A recent surge of structural data on several full-length class B GPCRs has provided insights into the structural basis of peptide ligand binding and receptor activation.^2–7 However, our understanding of the class B GPCR conformational diversity and dynamics during receptor activation and signaling to G proteins and β-arrestins in their native environment remains limited.^8–11 A critical barrier is the lack of an efficient method to construct functional GPCR biosensors that report receptor conformational transitions in real-time in live cells.¹² To this end, the successful biosensor design involves genetic fusion of a pair of fluorescent proteins (FPs), usually CFP and YFP, to generate Förster resonance energy transfer (FRET) biosensors for single-cell fluorescence-based analysis of GPCR signaling.^13–17 However, the FP fusion is restricted to the N- or C-termini of GPCRs and not suitable for the intracellular loops where the most profound conformational changes occur because of their large size.

We have recently begun developing a chemical biology strategy to construct fluorescence-based GPCR biosensors to address this problem.¹⁸ Our approach took advantage of two significant advances in chemical biology, namely, genetic code expansion¹⁹ and bioorthogonal reactions,^20,21 allowing us to introduce an organic dye into any position of GPCRs site-specifically in live cells.^22,23 Indeed, a growing repertoire of bioorthogonal reactions, together with a diverse range of genetically encoded unnatural amino acids (UAAs) carrying unique functionalities, have opened new venues for protein functional studies in living cells.²⁴

We have recently reported fluorescent labeling of the extracellular loops of the class B GPCRs using tetrazine ligation²² or tetrazole-based photoclick chemistry.²³ In this work, we turn our attention to the intracellular loops because they transduce the signals directly from the activated receptor to downstream G proteins and β-arrestins.²⁵ From a chemistry standpoint, the intracellular bioorthogonal labeling presents a significant challenge because the excess reagents need to be washed away from the cytosol after the labeling reaction. Using glucagon receptor (GCGR) as a model class B GPCR, we systematically investigated the various factors that affect the overall bioorthogonal labeling efficiency. These include selecting a suitable genetically encoded alkene reporter, identifying appropriate incorporation sites within the intracellular loop, designing fluorescent bioorthogonal reagents with low cytosolic retention, and examing labeling reaction specificity in the cytosol.

2. Results and discussion

Overall design strategy.

In a previous FRET-based biosensor design,²⁶ a fluorescein arsenical hairpin binder (FlAsH) reacted selectively to a genetically encoded tetracysteine tag encoded in the third intracellular loop of the human adenosine A_2A receptor-CFP fusion protein.²⁷ In functional assays, this CFP/FlAsH-tetracysteine pair gave fivefold greater agonist-induced FRET signals, along with similar receptor activation kinetics and normal downstream signaling.²⁶ Nonetheless, this approach suffers two shortcomings: 1) FlAsH reagent also binds weakly to other cysteine-containing proteins, giving rise to high fluorescence background; 2) the insertion of the 16-residue tetracysteine tag, AEAAARECCPGCCARA, into the intracellular loops of the class B GPCRs may affect its coupling with downstream G proteins. We envision introducing an organic fluorophore at the ICL3 of class B GPCRs via bioorthogonal labeling to minimize structural perturbation. In brief, we will incorporate a strained alkene amino acid at the ICL3 of GCGR-mCherry via genetic code expansion followed by a fast ligation reaction with a fluorescent tetrazine reagent (Fig. 1A). Among many reported genetically encoded alkene amino acids such as cyclopropene,²⁸ norbornene,^29–31 trans-cyclooctene (TCO),^32,33 and spirohexene,³⁴ we selected SphK,²² 2’-aTCOK,³⁵ and 4’-TCOK³² because of their fast kinetics in the tetrazine ligation (Fig. 1B). Separately, we designed a series of BODIPY-tetrazine probes for bioorthogonal labeling (Fig. 1C) because BODIPY shares similar fluorescent properties with EGFP and could serve as a substitute in the EGFP/mCherry FRET pair.³⁶ Importantly, the C-terminal mCherry allows us to monitor the incorporation efficiency of the strained alkene amino acids by their corresponding aminoacyl-tRNA synthetases.

Fig. 1.

(A) Scheme for bioorthogonal labeling of GCGR ICL3 in HEK293T cells. (B) Structures of strained alkene amino acids used in this study. (C) Two classes of designed BODIPY-tetrazine reagents for intracellular bioorthogonal labeling.

Selection of appropriate labeling sites.

To select appropriate locations on GCGR for the introduction of a BODIPY dye, we superimposed the full-length, activated GCGR structure²⁵ with the inactive one⁵ (Fig. 2A). We note that while the extracellular domains diverge significantly, the transmembrane regions are highly superimposable except transmembrane helices 5 and 6 (TM5/6) and the intervening intracellular loop 3 (ICL3). It was suggested that the activation of class B GPCR causes a significant kink in the middle of TM6 helix,²⁵ resulting in a drastic outward movement of ICL3 (Fig. 2B). Indeed, a recent GCGR−G_i1 complex structure showed that the H339 residue in ICL3 forms a salt bridge with D315 of G_i1, leading to G_i1 activation.³⁷ Thus, we generated a series of GCGR-mCherry mutants in which the codons for 11 residues in ICL3 (R336−R346; Fig. 2C) were switched to amber codons for charging the strained alkene amino acids by the appropriate PylRS variants as reported in the literature.

Fig. 2.

Selection of suitable sites for incorporation of chemical reporters. (A) Superimposition of the full-length inactive (PDB code: 5XEZ) and active (PDB code: 6WPW) conformations of GCGR. (B) A close-up view of the ICL3 regions showing the significant outward movement of ICL3 after receptor activation. The distance was measured between the relevant α-carbons. (C) The eleven ICL3 residues selected for unnatural amino acid mutagenesis, R336-R346, are rendered in a tube model on the active receptor.

Design and synthesis of BODIPY-tetrazine probes.

In choosing bioorthogonal chemical probes, the fitness factors comprised of reactivity, selectivity, physicochemical properties, and biological context need to be carefully considered.³⁸ For labeling GPCRs, which are often expressed at low levels in intact cells, we decided to employ the tetrazine/strained alkene reactant pairs known to have some of the fastest reaction kinetics in bioorthogonal chemistry. For example, a 2’-aTCOK-encoded EGFP gave a second-order rate constant of 3.59 × 10⁴ M⁻¹ s⁻¹ in the ligation reaction with the H-Tet-Cy5 probe.³⁹ On selectivity, we considered monoaryl tetrazine that is relatively stable but less reactive, as well as the dipyridyl-tetrazine that is more reactive but less stable.³² Regarding biological context, the tetrazine probes need to not only permeate into the cytosol to label the alkene-encoded ICL3 but also be washed away readily after the reaction. To this end, we adopted a computational model that has been used successfully to guide the development of cell-permeable background-free fluorescent probes.⁴⁰ This model identifies three molecular descriptors: SlogP for lipophilicity, LogS for water solubility, and Q_VSA_FNEG for negatively charged van der Waals surface area as the determinants. The ideal background-free fluorescent probes need to satisfy the following three criteria: adequate lipophilicity with SlogP in the range of 1 ~ 4, high water-solubility with LogS in the range −2 ~ −6, and moderate negative surface charge with Q_VSA_FNEG in the range of 0.15 ~ 0.35. Using this model, we designed four BODIPY-tetrazine probes (BODIPY-Tz1, −2, −3, and −4), and calculated their molecular descriptor values as listed in Chart 1. For the most part, these values fall into the desired ranges.

Chart 1.

Calculated molecular descriptor values of the designed BODIPY-tetrazine probes.

The syntheses of BODIPY-tetrazine probes are shown in Scheme 1. BODIPY-Tz1 was obtained in 22% yield by coupling meso-carboxypropyl-BODIPY (1) with mono-substituted tetrazine amine 2. BODIPY-Tz2 was derived from the coupling of BODIPY-FL carboxylic acid with tetrazine amine 2 in 63% yield. For BODIPY-Tz3, BODIPY 1 was first coupled with methyl 3-aminopropionate 3, followed by hydrolysis and subsequent coupling with tetrazine 2 to give the final product in an overall 11% yield. Lastly, BODIPY-Tz4 was obtained by coupling BODIPY 1 with the dipyridyl-tetrazine glycine amine 4 in an overall 47% yield.

Scheme 1.

Synthesis of the BODIPY-tetrazine probes

Assessing ICL3 labeling efficiency by tetrazine probes.

With BODIPY-tetrazine probes in hand, we performed bioorthogonal labeling of the ICL3 mutants bearing the various strained alkenes. Because Y343 is solvent-exposed and undergoes long-distance transition upon receptor activation (19.5 Å, Fig. 2B), we decided to use the GCGR-Y343TAG mutants to assess bioorthogonal labeling efficiency by BODIPY-Tz1–4 probes in both live and fixed cells (Fig. 3 and Fig. S1). The incorporation of three alkene amino acids into position-343 was confirmed based on the fluorescent signals in the mCherry channel (Fig. 3). On the other hand, the BODIPY channel showed varying fluorescent intensity and specificity after treatment with BODIPY-Tz1–4 probes. In fixed cells, specific ICL3 labeling was observed for the reactant pairs: SphK/BODIPY-Tz1, 2’-aTCOK/BODIPY-Tz3, 2’-aTCOK/BODIPY-Tz4. Notably, 4’-TCOK consistently showed a high level of intracellular retention and poor reaction selectivity. In live cells, strong UAA-independent fluorescence in the BODIPY channel was detected for BODIPY-Tz1/2, but not BODIPY-Tz3/4, suggesting that BODIPY-Tz1/2 can readily permeate into cells but are hard to remove. On the other hand, despite its high reactivity, BODIPY-Tz4 may have limited cell permeability compared to the charge-neutral BODIPY-Tz1–3 probes.

Fig. 3.

Confocal micrographs of HEK293T cells after bioorthogonal labeling of the GCGR-mCherry-Y343 mutants encoding the strained alkene. Cells were treated with 10 μM BODIPY-Tz1–4 for 1 hour before image acquisition.

Labeling studies of ICL3 position mutants.

Since the 2’-aTCOK/BODIPY-Tz4 pair gave the lowest intracellular retention and highest labeling specificity, we employed this pair to further assess their bioorthogonal labeling efficiency with the remaining ICL3 position mutants (R336−R346). We observed uniform and strong fluorescence in the mCherry channel throughout, indicating excellent incorporation of 2’-aTCOK at all positions (Fig. 4 and Fig. S2). Furthermore, the mutants encoding 2’-aTCOK at position-337, 339, 342, and 343 showed bright fluorescence in the BODIPY channel overlaying well with fluorescence in the mCherry channel, indicating that bioorthogonal labeling is specific (Fig. 4). The high labeling efficiency observed at these four positions is consistent with the loop structure as these residues are located at the tip of ICL3 and likely more solvent-accessible during the labeling reaction (Fig. 2C).

Fig. 4.

Confocal micrographs of HEK293T cells expressing GCGR-mCherry mutants carrying 2’-aTCOK at positions 336−346. Cells were treated with 10 μM BODIPY-Tz4 for 1 hour and then fixed with ice-cold methanol before image acquisition.

Since GCGR-mCherry-339(2’-aTCOK) exhibited the highest labeling efficiency, we expanded the labeling studies to include SphK and other BODIPY-Tz reagents. To our satisfaction, we found that BODIPY-Tz1/3 efficiently labeled the GCGR-mCherry-339SphK mutant, whereas BODIPY-Tz1/4 efficiently labeled the GCGR-mCherry-339(2’-aTCOK) mutant (Fig. 5 and Fig. S3). However, the live-cell imaging of the fluorescent labeling of GCGR-mCherry-339SphK/2’-aTCOK by BODIPY-Tz reagents was unsuccessful, which we attributed to the high intracellular retention of the fluorescent probes. By examining background fluorescence of BODIPY-3/4 in live cells (Fig. S4), we found that BODIPY-Tz3/4 showed high intracellular fluorescence regardless whether UAA was present, suggesting that the relatively high intrinsic fluorescence of these BODIPY-tetrazine reagents may have prevented the live-cell imaging.

Fig. 5.

Confocal micrographs of HEK293T cells expressing GCGR-mCherry-339SphK/2’-aTCOK. Cells were treated with 10 μM BODIPY-Tz4 for 1 hour and then fixed with ice-cold methanol before image acquisition.

3. Conclusions

In summary, we have performed a confocal fluorescent microscopic study of the intracellular bioorthogonal labeling of the third intracellular loop (ICL3) of glucagon receptor. Our labeling strategy involves the site-specific introduction of a strained alkene amino acid into the ICL3 through genetic code expansion, followed by a highly specific inverse electron-demand Diels-Alder reaction with the fluorescent tetrazine probes. Among the three strained alkene amino acids examined, both SphK and 2’-aTCOK offered successful fluorescent labeling of GCGR ICL3 with the appropriate tetrazine probes. At the same time, 4’-TCOK gave high background fluorescence due to its intracellular retention. The fluorescent tetrazine probes, BODIPY-Tz1–4, were designed following a computational model for background-free intracellular fluorescent labeling; however, their performance varied significantly in live-cell imaging as the strong non-specific signals interfered with the specific ones. Among all GCGR ICL3 mutants bearing a strained alkene, the H339SphK/2’-aTCOK mutants provided the best reaction partners for BODIPY-Tz1/4 in the bioorthogonal labeling reactions based on the fixed cell imaging results.

Owing to extensive retention of the fluorescent tetrazine probes in the cytosol, we could not obtain sharp live-cell fluorescent images nor conduct FRET-based receptor dynamics studies, even after multiple rounds of extensive washing. While BODIPY-tetrazine reagents have shown strong turn-on fluorescence in the literature,⁴¹ we found the turn-on effect was modest in vitro and insufficient for live-cell imaging studies. In meeting this challenge, one way is to use no-wash fluorogenic bioorthogonal probes that are nonfluorescent in solution but highly fluorescent when bound to the receptor surface.^42,43 Alternatively, a fluorophore could be generated in situ at the alkene-encoded ICL3 site through the tetrazole-based photoclick chemistry.²⁸ Efforts along these lines are currently underway and will be reported in due course.

4. Experimental

Solvents and chemicals were purchased from commercial sources and used directly without further purification. Flash chromatography was performed with Yamazen AKROS flash system equipped with SiliaSep HP pre-packed columns. ¹H and ¹³C NMR spectra were recorded with Varian Mercury-300 or −500 MHz spectrometer. Electrospray mass analysis was performed using an Agilent 6530 QTOF-LC/MS instrument. Plasmid pMyc-GCGR-mCherry-N1 was a generous gift provided by Prof. Graham Ladds at University of Cambridge. Compound 2’-aTCOK and the plasmid pCMV-FLAG-Mm-2’-TCOKRS-AF were generous gifts provided by Prof. Howard Hang at Rockefeller University.

4.1. Chemistry

4.1.1. N-(4-(1,2,4,5-Tetrazin-3-yl)benzyl)-3-(5,5-difluoro-1,3,7,9-tetramethyl-5H- 4λ⁴,5λ⁴-dipyrrolo[1,2-c:2’,1’-f][1,3,2]diazaborinin-10-yl)propenamide (BODIPY-Tz1)

To a solution of 1 (12 mg, 0.037 mmol) and 2 (11 mg, 0.045 mmol) in DMF (2 mL) was added HATU (17 mg, 0.045 mmol) and DIPEA (24 mg, 0.19 mmol). The mixture was stirred at room temperature for 2 h, then diluted with ethyl acetate (3 mL), washed with H₂O (5 mL), brine (5 mL), and concentrated. The residue was purified by silica gel flash column chromatography (hexanes/ethyl acetate = 1:1) to afford the titled compound as a red solid (4 mg, 22% yield): ¹H NMR (500 MHz, DMSO-d₆) δ 10.58 (s, 1H), 8.69 (s, 1H), 8.45 (d, J = 8.0 Hz, 2H), 7.52 (d, J = 8.0 Hz, 2H), 6.24 (s, 2H), 4.44 (d, J = 5.8 Hz, 2H), 3.23 (s, 2H), 2.54 (s, 6H), 2.40 (s, 6H); HR-QTOF-MS m/z for C₂₅H₂₆BFN₇O [M – F⁻], calcd 470.2270, found: 470.2248.

4.1.2. N-(4-(1,2,4,5-Tetrazin-3-yl)benzyl)-3-(5,5-difluoro-7,9-dimethyl-5H-4λ⁴,5λ⁴- dipyrrolo[1,2-c:2’,1’-f][1,3,2]diazaborinin-3-yl)propenamide (BODIPY-Tz2)

To a solution of BODIPY-FL (7 mg, 0.024 mmol) and 2 (7 mg, 0.029 mmol) in DMF (2 mL) was added HATU (11 mg, 0.029 mmol) and DIPEA (16 mg, 0.12 mmol). The mixture was stirred at room temperature for 2 h, then diluted with ethyl acetate (3 mL), washed with H₂O (5 mL), brine (5 mL), and concentrated. The residue was purified by silica gel flash column chromatography (hexanes/ethyl acetate = 1:1) to afford the titled compound as a red solid (7 mg, 63% yield): ¹H NMR (500 MHz, CDCl₃) δ 10.21 (s, 1H), 8.49 (d, J = 8.0 Hz, 2H), 7.34 (d, J = 8.0 Hz, 2H), 7.10 (s, 1H), 6.90 (d, J = 3.9 Hz, 1H), 6.32 (d, J = 3.9 Hz, 1H), 6.24 (s, 1H), 6.08 (s, 1H), 4.51 (d, J = 5.9 Hz, 2H), 3.40 – 3.27 (m, 2H), 2.79 (d, J = 7.3 Hz, 2H), 2.53 (s, 3H), 2.24 (s, 3H); HR-QTOF-MS m/z for C₂₃H₂₂BFN₇O [M – F⁻], calcd 442.1963, found 442.1936.

4.1.3. N-(4-(1,2,4,5-Tetrazin-3-yl)benzyl)-3-(3-(5,5-difluoro-1,3,7,9-tetramethyl-5H- 4λ⁴,5λ⁴-dipyrrolo[1,2-c:2’,1’-f][1,3,2]diazaborinin-10-yl)propanamido)propenamide (BODIPY-Tz3)

To a solution of 1 (30 mg, 0.094 mmol) and methyl 3-aminopropionate hydrochloride (16 mg, 0.11 mmol) in DMF (2 mL) was added HATU (42 mg, 0.11 mmol) and DIPEA (61 mg, 0.47 mmol). The mixture was stirred at room temperature for 2 h, then diluted with ethyl acetate (3 mL), washed with H₂O (5 mL), brine (5 mL), and concentrated. The residue was purified by silica gel flash column chromatography (hexanes/ethyl acetate = 1:2) to afford compound 3 as a red solid (21 mg, 55% yield): ¹H NMR (500 MHz, CD₃Cl) δ 6.10 (s, 1H), 6.06 (s, 2H), 3.70 (d, J = 2.4 Hz, 3H), 3.53 (qd, J = 6.1, 2.3 Hz, 2H), 3.36 –3.28 (m, 2H), 2.59 –2.48 (m, 8H), 2.46 –2.42 (m, 8H). To a solution of 3 (21 mg, 0.052 mmol) in THF/H₂O (3 mL, 2:1, v/v) was added LiOH (6 mg, 0.26 mmol) at 0 °C. The mixture was stirred at room temperature for 4 h, then 0.2 N HCl was added to adjust pH (<7). The mixture was extracted with ethyl acetate (2 mL×3), and the organic layer was washed with H₂O (5 mL), brine (5 mL), then concentrated. The crude carboxyl acid was used directly for next step without further purification. To the solution of crude carboxylic acid (18 mg, 0.046 mmol) and 1 (14 mg, 0.055 mmol) in DMF (2 mL) was added HATU (21 mg, 0.055 mmol) and DIPEA (30 mg, 0.23 mmol). The mixture was stirred at room temperature for 2 h, then diluted with ethyl acetate (3 mL), washed with H₂O (5 mL), brine (5 mL), concentrated, and purified by silica gel flash column chromatography (dichloromethane/ethyl acetate = 1:2) to afford the titled compound as an orange solid (5 mg, 20% yield): ¹H NMR (500 MHz, DMSO-d₆) δ 10.57 (s, 1H), 8.53 (t, J = 6.1 Hz, 1H), 8.43 (d, J = 8.1 Hz, 2H), 8.15 (s, 1H), 7.54 (d, J = 8.0 Hz, 2H), 6.20 (s, 2H), 4.40 (d, J = 5.6 Hz, 2H), 3.33 (m, 2H), 3.16 (t, J = 8.9 Hz, 2H), 2.41 (s, 6H), 2.38 (m, 10H); HR-QTOF-MS m/z for C₂₈H₃₁BFN₈O₂ [M – F⁻], calcd 541.2642, found 541.2618.

4.1.4. 3-(5,5-Difluoro-1,3,7,9-tetramethyl-5H-4λ⁴,5λ⁴-dipyrrolo[1,2-c:2’,1’-f][1,3,2] diazaborinin-10-yl)-N-(2-oxo-2-((6-(6-(pyridin-2-yl)-1,2,4,5-tetrazin-3-yl)pyridin-3-yl)amino)ethyl)propenamide (BODIPY-Tz4)

To a solution of 4 (6.3 mg, 0.015 mmol) in dichloromethane (2 mL) was added 4 N HCl in dioxane (1 mL) at 0 °C. The mixture was stirred at room temperature for 1 h, then concentrated. The crude de-Boc product was used directly for next step without further purification. To a solution of 1 (4.6 mg, 0.014 mmol) and crude de-Boc product (5.2 mg, 0.015 mmol) in DMF (2 mL) was added HATU (5.9 mg, 0.015 mmol) and DIPEA (13 mg, 0.098 mmol). The mixture was stirred at room temperature for 2 h, then diluted with ethyl acetate (3 mL), washed with H₂O (5 mL), brine (5 mL), concentrated, and purified by silica gel flash column chromatography (dichloromethane/methanol = 13:1) to afford the titled compound as a red solid (4 mg, 47% yield): ¹H NMR (500 MHz, DMSO-d₆) δ 10.71 (s, 1H), 9.06 (d, J = 2.5 Hz, 1H), 8.93 (d, J = 4.7 Hz, 1H), 8.64 (d, J = 8.7 Hz, 1H), 8.59 (d, J = 7.9 Hz, 1H), 8.52 (s, 1H), 8.42 (d, J = 8.6 Hz, 1H), 8.15 (t, J = 7.9 Hz, 1H), 7.77 – 7.64 (m, 1H), 6.25 (s, 2H), 4.05 (d, J = 5.5 Hz, 2H), 3.23 (t, J = 6.6 Hz, 3H), 2.62 (s, 1H), 2.50 (s, 6H), 2.41 (s, 6H); HR-QTOF-MS m/z for C₃₀H₃₀BF₂N₁₀O₂ [M + H⁺], calcd 611.2614, found 611.2583.

4.2. Calculation of molecular descriptors

Molecular operating environment (MOE) 2019 software package was used in calculating molecular descriptors of BODIPY-tetrazine probes. Hydrogens and lone-pair electrons were adjusted as required. Partial charge was set, and the protonation state was corrected for each structure. The active 3D conformation of the probes was obtained through energy minimization using Hamiltonian AM1 method. Three molecular descriptors (SLogP, LogS, and Q_VSA_FNEG) were generated.

4.3. Generation of amber mutants of GCGR-mCherry

Amber codon was introduced into the selected position in GCGR using site-directed mutagenesis with Platinum Pfx DNA polymerase (Thermo Fisher Scientific) or Phusion high-fidelity DNA polymerase (New England Biolabs) following the manufacturer’ instructions using pMyc-GCGR-mCherry-N1 as the template and the following primer pairs:

GCGR-R336TAG

F: 5’- GCTGCGGGCAtaGCAGATGCACC −3’

R: 5’- TTGGCCACGAGCAGCTGA −3’

GCGR-Q337TAG

F: 5’- GCGGGCACGGtAGATGCACCA −3’

R: 5’- AGCTTGGCCACGAGCAGC −3’

GCGR-M338TAG

F: 5’- GGCACGGCAGtaGCACCACACAGAC −3’

R: 5’- CGCAGCTTGGCCACGAGC −3’

GCGR-H339TAG

F: 5’- ACGGCAGATGtagCACACAGACTACAAGTTCC −3’

R: 5’- GCCCGCAGCTTGGCCACG −3’

GCGR-H340TAG

F: 5’- GCAGATGCACtagACAGACTACAAGTTCCGGC −3’

R: 5’- CGTGCCCGCAGCTTGGCC −3’

GCGR-T341TAG

F: 5’- GATGCACCACtagGACTACAAGTTCCGGCTGGCC −3’

R: 5’- TGCCGTGCCCGCAGCTTG −3’

GCGR-D342TAG

F: 5’- GCACCACACAtagTACAAGTTCCGGCTGG −3’

R: 5’- ATCTGCCGTGCCCGCAGC −3’

GCGR-Y343TAG

F: 5’- ACACAGACTAgAAGTTCCGGCTGG −3’

R: 5’- GGTGCATCTGCCGTGCCC −3’

GCGR-K344TAG

F: 5’- CACAGACTACtAGTTCCGGCTGGC −3’

R: 5’- TGGTGCATCTGCCGTGCC −3’

GCGR-F345TAG

F: 5’- AGACTACAAGtagCGGCTGGCCAAG −3’

R: 5’- GTGTGGTGCATCTGCCGT −3’

GCGR-R346TAG

F: 5’- CTACAAGTTCtaGCTGGCCAAGTCC −3’

R: 5’- TCTGTGTGGTGCATCTGC −3’

4.4. HEK293T cell culture and transfection

HEK293T cells were maintained in Dulbecco’s modified eagle medium (DMEM, Life Technologies) supplemented with 10% (v/v) fetal bovine serum (FBS, Life Technologies) and 10 μg/mL gentamicin (Life Technologies). Transfection was performed at 70−80% confluency using a 3:1 transfection reagent/DNA ratio of polyethylenimine (PEI, Polysciences Inc.) with 3 μg of total DNA per 35 mm dish. For imaging experiments, cells were maintained in 35 mm glass-bottom culture dishes using FluoroBrite DMEM medium (Life Technologies) supplemented with 10% FBS, 4 mM L-glutamine, and 25 mM HEPES.

4.5. Bioorthogonal labeling and confocal microscopy

HEK293T cells were co-transfected with a 4:1 ratio of pCMV-MmPylRS-U6-tRNA and pCMV-GCGR-TAG-mCherry plasmids. A solution of 10 mM UAA in DMSO was diluted in growth medium to obtain a final concentration of 50 μM UAA, and the medium was added to the cells, followed by the transfection mixture. The cells were incubated for an additional 24−48 h before washing the cells with a growth medium. A fresh solution of BODIPY-Tz1–4 was prepared from 10 mM stock solution in DMSO to a final concentration of 10 μM in 1 mL of growth medium. For live-cell imaging, HEK293T cells expressing GCGR-TAG-mCherry mutants were treated with 10 μM BODIPY-Tz1–4 in 1 mL of DMEM medium containing 10% FBS, and the culture was incubated at 37 °C and 5% CO₂ for 60 min. The tetrazine-containing medium was removed, and the cells were washed with growth medium before incubation at 37 °C and 5% CO₂ for an additional 30 min. The medium was then switched to FluoroBrite DMEM before laser scanning confocal microscopy. The confocal images were acquired using a Zeiss LSM 710 equipped with Plan-Apochromat 20×/0.8 M27 or 40×/1.3 Oil DIC M27 objective with ex 488/em 493−598 nm for the BODIPY channel and ex 580/em 596–696 nm for the mCherry channel. For fixed-cell imaging, following tetrazine incubation in culture media, cells were fixed in ice-cold MeOH for 15 min, washed with 3 × 1 mL PBS, and incubated in 1 mL PBS at 4 °C overnight. PBS was removed and FluoroBrite DMEM was added to cells before the image acquisition.