To study the potential role of endosomal chemokine receptor signaling we used the model receptor CCR7, which is expressed by several subsets of immune cells and regulates their homing to the lymph nodes²³. This receptor is an ideal model receptor to study endosomal G protein signaling as it is known to be active at endomembranes²⁴ and is stimulated by two endogenous chemokines CCL19 and CCL21 equally effective at 100 nM²⁵. Compared with the full agonist CCL19 that robustly stimulates G_i/o signaling, receptor phosphorylation, βarr recruitment, and CCR7 internalization, the agonist CCL21 promotes full G_i/o protein activation but only partial receptor phosphorylation, βarr recruitment, and internalization of CCR7^25,26. Using the enhanced bystander bioluminescence resonance energy transfer (EbBRET)-based biosensors RlucII-βarr1/2 and cell surface-anchored rGFP-CAAX to monitor βarr1/2 recruitment to plasma membrane upon chemokine-stimulation in CCR7-expressing HEK293 cells (Fig. 1A)²⁷, we confirmed that activating CCR7 by CCL19 leads to robust βarr1/2-recruitment, whereas CCL21 only promotes partial βarr1/2-recruitment responses (Fig. 1B-C). Co-expressing RlucII-βarr1/2 and the endosomally-located rGFP-Rab5 in HEK293-CCR7 cells (Fig. 1A)²⁷, CCL19 stimulation leads to similar strong βarr1/2 trafficking to endosomes whereas CCL21 stimulation of CCR7 only causes a minor translocation of βarr1/2 to the endosomes (Fig. 1B-C). No enhanced EbBRET between RlucII-βarr1/2 and rGFP-CAAX/Rab5 was detected upon chemokine stimulation in HEK293 cells that do not express CCR7 (Fig. S1A-B). These results suggest that CCR7 recruits and is internalized by βarrs robustly when stimulated with CCL19, but only to a modest degree by CCL21. Similar results were found when using β-galactosidase enzyme complementation-based DiscoverX assay to monitor direct βarr2 recruitment to CCR7 (Fig. S1C-D).

Next, to investigate how CCL19/CCL21-stimulated βarrs recruitment affects CCR7 internalization in real time, we applied a split NanoLuc (or NanoBiT) approach in which the small-BiT (SmBiT) is fused the carboxy-terminal of CCR7 and the large-BiT (LgBiT) is fused to the cell surface anchor CAAX. In resting state, the CCR7-SmBiT and LgBiT-CAAX are in subcellular proximity at the plasma membrane and form functional nanoluciferase enzymes, which catalyzes the conversion of coelenterazine h resulting in emission of a bright luminescence signal (Fig. 1D). Upon activation and internalization of CCR7, the CCR7-SmBiT is removed from the cell surface-located LgBiT-CAAX which results in a decrease in the luminescence signal (Fig. 1D). Strong and immediate reduction in this signal was observed when the cells were stimulated with CCL19 (Fig. 1E-F). In contrast, CCL21-stimulated CCR7 only led to a modest decrease in signal suggesting that CCL19-stimulation leads to robust CCR7 internalization whereas most CCR7 stays at the plasma membrane upon CCL21 activation (Fig. 1E-F). In addition to this approach, we also measured chemokine-stimulated trafficking of CCR7-SmBiT to the early endosomal FYVE-LgBiT marker (Fig. 1D). In this case, CCL19 stimulation led to a robust increase in luminescence whereas CCL21 activation only led to a partial increase in the signal (Fig. 1G-H). Interestingly, the kinetic patterns of the signals indicate that the trafficking of CCR7 to endosomes was slower and more transient as compared to its removal from the cell surface, which could be a result of further trafficking of the receptor to late or recycling endosomes as well as the trans-Golgi network (TGN; Fig. 1G-H). Moreover, CCL19/CCL21-stimulation did not promote internalization or trafficking to endosomes of the vasopressin type 2 receptor (V₂R)-SmBiT construct validating that these chemokines act specifically via the CCR7-SmBiT system (Fig. S1E-F).

Finally, to investigate how βarrs recruitment and receptor internalization affect G protein signaling in real-time, we applied the cAMP sensor CAMYEL²⁸ to monitor G_i/o activity in HEK293-CCR7 cells. In contrast to βarrs recruitment and receptor internalization, CCL19 and CCL21 challenge inhibited forskolin-induced cAMP production to a similar extent in HEK293-CCR7 cells (Fig. 1I-J), but had no effect on the cAMP response in mock transfected HEK293 cells (Fig. S1G-H). This supports previous observations, which suggest that both CCL19 and CCL21 act as full agonists in stimulating G_i/o signaling^25,26. Interestingly, this G_i/o signaling was sustained throughout the experiment and not acutely desensitized by βarrs recruitment and/or receptor internalization as other receptor systems (Fig. 1I-J)²⁹.

Surprisingly, neither the degree of βarr recruitment nor receptor internalization appear to affect CCR7-mediated G protein signaling as expected for most GPCRs. As we recently found a correlation between the presence of phosphorylation site clusters in GPCRs and their ability to form GPCR–G protein–βarr megaplexes⁹, here we hypothesized that CCL19-CCR7 stimulates G_i/o while being internalized into endosomes by βarrs (Fig. 2A). As numerous G_i/o-coupled GPCRs recently were demonstrated to co-couple with G protein and βarrs, this hypothesis seems plausible³⁰. In contrast, since CCL21 challenge only leads to minor CCR7 internalization, most of the CCL21-CCR7-mediated G_i/o activation take place at the plasma membrane (Fig. 2A).

Simultaneous GPCR association with G protein and βarr is facilitated by βarrs complexing exclusively through the phosphorylated receptor carboxy-terminal tail. We have previously demonstrated that other GPCRs with carboxy-terminal phosphorylation site clusters bind to βarrs in this tail conformation, and thus, it seemed likely that CCR7 could form them as well. To test this, we used a βarr1 mutant where the region of βarr1 that interacts with GPCR transmembrane cores, the fingerloop (FL), is deleted (βarr1-ΔFL). We showed previously using negative stain electron microscopy and single-particle averaging that βarr1-ΔFL almost entirely forms tail conformation complexes with GPCRs¹⁶. Using the EbBRET pair RlucII-βarr1-ΔFL and rGFP-CAAX (plasma membrane-anchored rGFP) in CCR7-expressing HEK293 cells, we probed the degree of βarr1-ΔFL recruitment to CCR7 in response to CCL19, CCL21, or vehicle. In addition, we tested how well βarr1-ΔFL internalizes with CCR7 into endosomes using the EbBRET pair RlucII-βarr1-ΔFL and rGFP-Rab5 (early endosome-anchored rGFP). Interestingly, both CCL19 and CCL21 stimulation led to recruitment of βarr1-ΔFL to CCR7 at the membrane although CCL19 promoted this recruitment to a significantly higher degree than CCL21 (Fig. 2B). Furthermore, CCL19 and CCL21 stimulation promote βarr1-ΔFL translocation to endosomes (Fig. 2B). No response to chemokine stimulation was observed in mock transfected HEK293 cells that do not express CCR7 (Fig. S2A). These results indicate that CCR7 can form tail conformation complexes with βarr1 in cells and that the stability of this complex is sufficiently strong for βarr1 to internalize the receptor into endosomes.

Since CCR7 forms tail conformation complexes with βarr1 upon agonist stimulation where the G protein binding site is available, we next sought to interrogate whether the receptor binds to βarrs and G_i/o simultaneously (Fig. 2C). To do this, we used a NanoBiT approach where SmBiT is fused to βarr1 and LgBiT is fused to the Gα subunit. If CCR7 recruits G protein and βarr1 simultaneously, the SmBiT-βarr1 and LgBiT-G will come into close proximity resulting in a luminescence signal. Instead of using the full-length Gα subunit for this approach as done previously⁹, we used a truncated surrogate version commonly referred to as miniGi. MiniG proteins are engineered Gα subunits that have been developed for the major Gα subunit families (G_s, G_i/o, G_q/11, and G_12/13). Key modifications of these Gα subunits include: 1) a truncated N-terminal that deletes membrane anchors and Gβγ-binding surface; 2) deletion of the α-helical domain; 3) mutations that improve protein stability in vitro; and 4) a mutation in the C-terminal α5 helix that uncouples GPCR binding from nucleotide release, therefore stabilizing receptor– miniG protein complexes in the presence of guanine nucleotides^31,32. These alterations also enable miniG proteins to report active receptor conformations in living cells by measuring their recruitment from the cytosol to GPCRs at different membrane compartments³¹. As most membranes involved in GPCR trafficking contain endogenous G proteins³³, the presence of active miniG-coupling receptors at these subcellular compartments suggests that G proteins are stimulated from these sites.

The properties of miniG proteins are also beneficial when monitoring simultaneous recruitment of SmBiT-βarr1 and LgBiT-miniG to GPCRs. Since both probes are expressed in the cytosol where they randomly collide, simultaneous translocation to the receptor in cell membranes will lead to a very specific increase in luminescence (Fig. 2C). In contrast, using a wild-type full length Gα subunit and βarr NanoBiT probes, which are located at membranes and cytosol, respectively, recruitment of βarrs to the receptor could lead to potential bystander effects as a result of βarrs-translocation from the cytosol to the plasma membrane.

In HEK293-CCR7 cells expressing SmBiT-βarr1 and LgBiT-miniGi, CCL19-stimulation leads to a robust increase in luminescence (Fig. S2B-C). This signal was specific for miniGi as stimulation of CCR7 in cells expressing SmBiT-βarr1 and LgBiT-miniGs led to a small decrease in luminescence (Fig. S2B-C). This reduction in basal luminescence might be caused by a reduction of random collision between SmBiT-βarr1 and LgBiT-miniGs upon CCR7 stimulation where SmBiT-βarr1, but not LgBiT-miniGs, translocates to the plasma membrane. These results indicate that CCL19-stimulated CCR7 recruits G_i/o and βarrs simultaneously. Next, we compared the ability of CCL19 and CCL21 to provoke co-coupling of G_i/o and βarrs to CCR7. Interestingly, both CCL19 and CCL21 stimulated co-coupling of G_i/o and βarrs to CCR7, although CCL19 promoted formation of these complexes to a significantly larger degree than CCL21 (Fig. 2D-E). The response to chemokine stimulation was not observed in mock transfected HEK293 cells (Fig. S2D).

GPCR–G protein–βarr megaplexes have been described as a mechanism by which GPCRs that bind βarrs exceptionally well via the phosphorylated C-terminal tail can promote endosomal G protein signaling⁹. Therefore, we investigated whether G_i/o and βarrs are recruited simultaneously to internalized CCR7 in endosomes using confocal microscopy. For this purpose, CCR7-expressing HEK293 cells co-expressing βarr2-Strawberry and Halo-miniGi were stimulated with either CCL19 or CCL21 for 30 min whereafter the subcellular location of these two probes was assessed by confocal microscopy. In this setup, CCL19-CCR7 recruited βarr2-Strawberry and Halo-miniGi to endosomal-shaped intracellular locations whereas this co-localization was minor for CCL21-CCR7 (Fig. 2F-G). Together, these results indicate that CCL19 stimulation leads to robust endosomal and CCR7-mediated recruitment of G_i/o and βarr1.

To test whether the internalized CCR7 stimulates G_i/o from endosomal compartments, we used the NanoBiT biosensor pair Rap1GAP-SmBiT and LgBiT-FYVE (Fig. 3A). Rap1GAP is an effector protein that specifically binds to active G_i/o, but not to inactive G_i/o³⁴, and thus, endosomal G_i/o activation can be detected by measuring the recruitment of Rap1GAP-SmBiT to the endosomal marker LgBiT-FYVE³⁵. Using this approach, we observed that both CCL19 and CCL21 stimulate recruitment of Rap1GAP-SmBiT to LgBiT-FYVE in HEK293-CCR7 cells (Fig. 3B-C). Similar to CCR7 internalization, CCL19 facilitated this response to a significant higher degree as compared to CCL21 (Fig. 3B-C). In addition, the recruitment of Rap1GAP-SmBiT to LgBiT-FYVE was blunted by the G_i/o inhibitor pertussis toxin (PTX), and was not observed in mock transfected HEK293 (Fig. S3A-C). These results indicate that internalized CCR7 promotes endosomal G_i/o activation.

To further validate if CCR7 activates G_i/o at endosomes, we measured translocation of RlucII-miniGi protein to the early endosomal marker rGFP-Rab5a by EbBRET in response to CCR7 challenge (Fig. 3D). Using this approach, CCL19 stimulation led to a strong and significant increase in EbBRET signaling between RlucII-miniGi and rGFP-Rab5a indicating that CCL19-bound CCR7 indeed activates G_i/o robustly from endosomes (Fig. 3E). In contrast, CCL21 stimulation provoked a small yet significant increase of EbBRET response, demonstrating that G_i/o protein is activated from endosomes by CCL21-bound CCR7 to a minor extent (Fig. 3E).

Similar trend was also observed by confocal microscopy visualization of CCR7-expressing HEK293 cells co-expressing Halo-miniGi and either the plasma membrane marker RFP-Lck or the early endosomal marker RFP-EEA1. In cells expressing RFP-Lck, CCR7 activation with either CCL19 or CCL21 leads to recruitment of Halo-miniGi to the cell surface (Fig. 3F-G). As expected, no significant difference between CCL19- and CCL21-stimulated translocation of Halo-miniGi to the plasma membrane was detected. However, in cells expressing RFP-EEA1, activation by CCL19, and to a minor degree CCL21, resulted in co-localization of Halo-miniGi and RFP-EEA1 at endosomes (Fig. 3H-I). Together these results suggest that CCL19-CCR7 promotes robust G protein activation from endosomes whereas CCL21-CCR7 predominantly activates G_i/o at the cell surface and to a modest degree from endosomes.

Limited knowledge exists regarding functional outcomes of endosomal G protein signaling by internalized GPCRs, particularly chemokine receptors. Thus, to examine the impact of compartmentalized CCR7 signaling on downstream effectors, we performed unbiased mass spectrometry (MS)-based proteome profiling using APEX2. APEX2 is an engineered ascorbate peroxidase enzyme that we fused to the CCR7 carboxy-terminal and uses H₂O₂ as an oxidant to catalyze a one electron oxidation of various small molecule substrates including biotin-tyramide (Fig. 4A). Oxidation of biotin-tyramide leads to generation of a highly reactive and short-lived (<1 ms) biotin-phenoxyl radical that conjugates to endogenous proteins that are in close proximity to the APEX2 enzyme (∼20 nm, Fig. 4A)³⁶. Thus, the APEX2 application allows for spatiotemporal control of the biotinylation process with high precision. The resulting biotinylated proteins are then enriched with pull-down experiments using neutravidin beads, and their identities analyzed by MS (Fig. 4A).

The CCR7-APEX2 construct expressed well at the cell surface (Fig. S4A) and receptor functionality construct was confirmed by assessing its ability to promote βarr1 translocation to the plasma membrane and early endosomes in response to CCL19 challenge (Fig. S4B). Additionally, CCL19 or CCL21 stimulation inhibited forskolin-induced cAMP production equally as expected (Fig. S4C). In addition, we verified that the APEX2 portion of fusion proteins only biotinylates proteins in the presence of both biotin-tyramide and hydrogen peroxide as expected (Fig. 4B). Using neutravidin beads, we demonstrated that proteins were pulled-down from HEK293-CCR7-APEX2 cell lysates where the biotinylation process had been initiated, whereas only negligible proteins were pulled-down in untreated cells (Fig. 4C). Together these results indicate that both receptor and APEX2’s enzyme functions of the CCR7-APEX2 fusion are intact.

Using our experimental setup, we stimulated HEK293-CCR7-APEX2 cells pretreated with biotin-tyramide with either CCL19 or CCL21 for 0 minute, 2 minutes, 10 minutes or 25 minutes. Hereafter, the biotinylation reaction was initiated by addition of hydrogen peroxide for exactly 1 min, followed by extensive washing, cell lysis, enrichment of biotinylated proteins, and MS profiling.

From the MS profiling we identified a total of 584 proteins, whose enrichment was changed significantly upon stimulation with either CCL19 or CCL21 (p < 0.05 and change in intensity of log₂ > 1) (Fig. S5A-C). Among robustly enriched proteins by CCR7 activation, several are involved in functions related to vesicle trafficking, signal transduction, cytoskeletal dynamics, and cell migration among others (Fig. 5A). Furthermore, as CCR7 is trafficked from the plasma membrane to endosomes within the first minutes of stimulation, we also included functionally validated APEX2-fused spatial references, PM-APEX2, ENDO-APEX2, and CYTO-APEX2 to control for the potential increase in random collisions between CCR7 and endosomal proteins that might occur upon receptor internalization (Figs. 5B-D and S6)³⁷. Proteins enriched by both ENDO-APEX2 (as compared to both CYTO-APEX2 and PM-APEX2) and chemokine-stimulated CCR7-APEX2 include SNX3, RAB9A, ARH, and CCD93 (Fig. 5A and 5D). Thus, these proteins might be enriched by CCR7-APEX2 due to chemokine-stimulated translocation from the plasma membrane to endosomes rather than participating in multiprotein functional complexes or signalosomes forming at CCR7.

As expected, proteins involved in receptor and vesicular trafficking including RIC1, SNX17, VA0D1, ARRB2, and DYN3, among others, were enriched over the entire time course of CCR7 activation (Fig. 5A). Most of these proteins were more enriched by CCL19 stimulation as compared to CCL21 challenge. In addition, we observed enrichment of endosomal proteins such as STX7, STX12, SNX3 and VTI1B for both CCL19 and CCL21 stimulation (Fig. 5A). Interestingly, proteins involved in lysosomal trafficking including LAMP1, VPS16, VAMP7, WDR91, and PP4P1 were enriched by CCL19 and to a lesser degree by CCL21 after 25 minutes of stimulation (Fig. 5A). Proteins involved in receptor trafficking via recycling endosomes or the trans-Golgi network such as SNX6, RAB7L, and GGA3 were also enriched at the later stages of CCR7 activation particularly when stimulated with CCL21 (Fig. 5A). In addition to these, proteins associated with the Golgi network (SNX3, COG5, YIF1A, SC22B, and AP3S1) were enriched by either CCL19 or CCL21 suggesting general CCR7 trafficking to this subcellular compartment as previously reported²⁴. These results indicate that the trafficking pattern of CCR7 upon chemokine stimulation is tightly regulated between lysosomal degradation and recycling.

Among enriched signaling proteins, we observed robust enrichment of proteins regulating inositol 1,4,5-triphosphate signaling (ERLN1/2, IP3KA), diacyl glycerol signaling (DGKQ), calcium mobilization at the endoplasmic reticulum (TMCO1), and activity of the Rho GTPases (Fig. 5A and 5E-G). These are all involved in signaling events downstream of heterotrimeric G protein activation. Another interesting signaling protein enriched by CCR7 activation includes a member of the eyes absent (EYA) subfamily of protein, EYA2. This protein have been shown to interact with and regulate Gα_i/o and Gα_z subunits^38,39, and EYA4, a member of the same family, was recently identified in a similar APEX2-based study of the μ-opioid receptor⁴⁰. We also detected three novel interaction partners GRIP2, MARK4, and EI24 as being among the highest levels regardless of the activation ligand or time (Figs. 5A and S5B).

Rho GTPases such as RhoA, Rac1, Cdc42 regulate important aspects of chemotactic cell migration²². Activity of these effectors are regulated by other proteins such as Rho GTPase-activating proteins (GAPs), Rho guanine nucleotide exchange factors (GEFs), and Rho GDP dissociation inhibitors (GDIs) (Fig. 5E). Our APEX2-based proteomics results demonstrate that a number of these Rho GTPase regulators are enriched by chemokine-stimulation of CCR7, and thus, potentially form part of larger CCR7 signalosomes that control downstream Rho GTPase activity (Fig. 5F-G). Interestingly, some of these regulators, including RHG26, ARG28, RHG29, ACK1, and BORG5, were differentially enriched upon CCL19 or CCL21 stimulation (Fig. 5F). Therefore, it is possible that CCR7 signaling modulates activity of cell migratory Rho GTPases specifically at distinct cellular locations. To test this hypothesis, we focused on RhoA, Rac1, and Cdc42. Upon activation of RhoA, Rac1, and Cdc42, recruit protein kinase N1 (PKN1), p21 activated kinase 1 (PAK1), and Wiskott-Aldrich syndrome protein 1 (WAS1), respectively. We assessed the proximity of these biosensor pairs^41,42 in real-time upon chemokine stimulation by the NanoBiT approach in HEK293-CCR7 cells (Fig. 6A). Using this setup, we found that both CCL19 and CCL21 stimulated RhoA and Cdc42 activation to a similar extent, which indicates that the spatial aspect of G protein activation does not appear to influence RhoA and Cdc42 signaling (Fig. 6B-C and 6F-G). In contrast, CCL19 promoted Rac1 signaling to a significantly greater extent than CCL21 stimulation (Fig. 6D-E). No change in RhoA, Rac1, or Cdc42 activity upon CCL19/CCL21 challenge was detected HEK293 cells that do not express CCR7 (Fig. S7A-C). However, CXCL12 stimulation led to enhanced activity of RhoA, Rac1, or Cdc42 in HEK293 cells, which express CXCR4 (Fig. S7D-F). Pre-incubating the HEK293-CCR7 cells with the G_i/o inhibitor PTX, virtually eliminated the chemokine-induced activation of Rac1, indicating that CCR7-stimulateted Rac1 signaling is mediated via a G_i/o-dependent mechanism (Fig. 6H-I). Interestingly, pre-treatment of the HEK293-CCR7 cells with the endocytosis inhibitor Dyngo-4a⁷ reduced chemokine-stimulated CCR7 internalization and Rac1 activation robustly (Fig. 6J-K and S7G-H). This reduction was significantly smaller as compared to cells pre-incubated with an inactive Dyngo compound⁷ (Figs. 6J-K and S7G-H). Furthermore, overexpression of the hemagglutinin-tagged dominant negative dynamin 1 K44A mutant (HA-Dyn-K44A) diminished chemokine-activated CCR7 internalization and Rac1 activity suggesting that CCR7-mediated Rac1 activation depends on receptor internalization (Figs. 6L-M and S7I-K). Also, pre-incubating the cells with the endocytosis inhibitor PitStop2, which acts on clathrin rather than dynamin, led to a severe reduction in chemokine-stimulated CCR7 internalization and Rac1 activity (Figs. 6N-O and S7L-M). Finally, we mutated all potential serine/threonine phosphorylation sites in the carboxy-terminal tail of CCR7 and found that this CCR7-ΔST construct expresses well at the cell surface, but does not internalize upon CCL19 stimulation (Fig. S7N-P). When expressed in HEK293 cells, the ability of CCR7-ΔST to activate Rac1 in response to chemokine-stimulation is severely reduced as compared to signaling by internalized CCR7 as a mechanism to activate Rac1.

As Rac1 is a key regulator of chemotaxis, we next tested the ability of CCL19 and CCL21 to promote cell migration of endogenously CCR7-expressing Jurkat T cells⁴³ towards either CCL19 or CCL21. As in HEK293-CCR7 cells, CCL19 and CCL21 stimulation led to similar inhibition of forskolin-stimulated cAMP production in Jurkat T cells confirming that CCR7 activation by either chemokine results in comparable stimulation of G_i/o signaling (Fig. 7A). Interestingly, Jurkat T cells migrated more efficiently towards a gradient of CCL19 as compared with CCL21, which was assessed by the transwell chemotaxis assay (Fig. 7B). These findings raised the possibility that endosomal G_i/o signaling plays a major role in chemotaxis. To test this hypothesis, we pre-treated the Jurkat T cells with the G_i/o inhibitor PTX, which almost completely abolished the ability of the cells to migrate towards both CCL19 and CCL21 (Fig. 7C). Interestingly, pre-incubation with the endocytosis inhibitors Dyngo-4a or PitStop2 led to a significant reduction in chemotaxis of both CCL19 and CCL21-stimulated cells as compared to the inactive Dyngo or PitNot compounds, respectively (Fig. 7D-E). Finally, as CCR7-mediated endosomal G_i/o signaling was shown to enhance Rac1 activity, we pre-treated Jurkat T-cells with the Rac1 inhibitor EHT1864. As expected, EHT1864 reduced Jurkat T-cell chemotaxis towards both CCL19 and CCL21 (Fig. 7F). Collectively, these finding suggest that endosomal G_i/o signaling by internalized CCR7 promote chemotaxis of Jurkat T cells.

Hank’s balanced salt solution (HBSS), forskolin, G418, HEPES, NaCl, phosphate saline buffer (PBS) tablets, 3-isobutyl-1-methylxanthine (IBMX), KCl, MgCl₂, NaHCO₃, NaH₂PO₄, glucose, sodium ascorbate, sodium azide, biotin-tyamide, sodium deoxycholate, Tris-HCl, H₂O₂, sodium dodecyl sulfate (SDS), phenylmethylsulfonyl fluoride (PMSF), ethanol, chloroacetamide (CAA), Tween-20, tris(2-carboxyethyl)phosphine (TCEP), ethylenediaminetetraacetic acid (EDTA), CaCl₂, and rabbit anti-β-tubulin antibody (cat. no. T2200), anti-FLAG M2-HRP conjugated antibody (Cat. A8592, Sigma-Aldrich) were all purchased from Sigma-Aldrich. Dulbecco’s Modified Eagle’s Medium (DMEM) high glucose, Opti-MEM^TM reduced serum/no phenol red, Dulbecco’s PBS (DPBS), fetal bovine serum (FBS), penicillin/streptomycin, Lipofectamine 3000, zeocin, puromycin, donkey anti-mouse IgG antibody conjugated to Alexa Fluor™ 488 (cat. no. A21202), SuperSignal™ West Pico PLUS Chemiluminescent Substrate, and silver stain kit were all purchased from Thermo Fisher Scientific. Coelenterazine h, coelenterazine 400a, and Deep Blue C^TM were all purchased from NanoLight Technology. NeutrAvidin™ agarose, paraformaldehyde, and Triton X-100 were purchased from Fisher Scientific. CCL19 was purchased from Chemotactics. CCL21 was purchased from GenScript. Trolox was purchased from Millipore Halo Tag ligand was purchased from Promega. Salmon sperm DNA was purchased from Invitrogen. Linear polyethyleneimine 25K (PEI) was purchased from Polysciences. Mouse anti-HA antibody (cat. no. sc-7392) was purchased from Santa Cruz. Donkey anti-rabbit IgG F(ab′)2 conjugated to horseradish peroxidase (HRP; cat. no. NA9340) was purchased from Cytiva.

Human CCR7 in pcDNA3 was previously described²⁶. Human CCR7 was N-terminally tagged with an influenza hemagglutinin signal sequence followed by a FLAG-tag (MKTIIALSYIFCLVFA + DYKDDDDK) into pTWIST/CMV/Zeo and synthesized by Twist Bioscience. For CCR7 trafficking NanoBiT assays, a SmBiT⁴² was C-terminally tagged to the human CCR7 sequence through a linker (GGSG) and synthesized by Twist Bioscience. The CCR7-ΔST and CCR7-ΔST-SmBiT constructs were synthesized with 11 mutations of serine and threonine residues in the receptor C-terminal tail (S348A, S356A, S357A, S364A, S365A, S367A, T372A, T373A, T374A, T375A, and S377A). For βarr1 trafficking NanoBiT assays, a SmBiT was N-terminally tagged to the human βarr1 sequence through a flexible linker (GGSGGGGSGGSSSGG) and synthesized by Twist Bioscience. In addition, the plasma membrane anchored LgBiT-CAAX and endosomally anchored LgBiT-FYVE have previously been described⁶¹. For proximity biotin labelling, CCR7 was N-terminally tagged with an influenza hemagglutinin signal sequence followed by a FLAG-tag, and with APEX2 added to the C-terminus through a flexible linker (GGSGGGGSGGSSSGG) and synthesized into pTwist/CMV/Puro using the EcoRI and NheI restriction sites of the vector by Twist Bioscience. Lyn11-GFP-APEX2 (PM-APEX), 2×FYVE-GFP-APEX2 (ENDO-APEX), and GFP-APEX2 (CYTO-APEX) were a kind gift from Dr. Mark von Zastrow and were designed as plasma membrane, endosomal, and cytosolic spatial references, respectively³⁷. The CAMYEL biosensor was previously described²⁸. RlucII-βarr1 and RlucII-βarr1-ΔFL were previously described^16,62. The RlucII-βarr2 construct (Renilla Luciferase II in the carboxy-terminal of βarr2) was built by replacing the GFP10-EPAC sequence from the previously published GFP10-EPAC-RlucII^62,63 with the coding sequence of human βarr2. The rGFP-CAAX and rGFP-Rab5a are cloned into pcDNA3.1(+) and were previously described²⁷. MiniGi³² (also referred to as miniGsi) and miniGs³² were N-terminally tagged with a LgBiT⁴² through a linker (GGSG), and the LgBiT was N-terminally tagged with a nuclear export signal (MLQNELALKLAGLDINKT) via a linker (GGSG), and synthesized into pTwist/CMV using the NotI and BamHI restriction sites of the vector by Twist Bioscience. Rap1GAP³⁴ was C-terminally tagged with a SmBiT⁴² through a linker (GSAGTGGRAIDIKLPAT) and synthesized into pTwist/CMV using the NotI and BamHI restriction sites of the vector by Twist Bioscience. Human βarr1 was also N-terminally tagged with SmBiT⁴² through a linker (GGSG) synthesized by Twist Bioscience. The RlucII-miniGsi was synthesized by Twist Bioscience. The Venus tag from the NES-venus-mGsi contruct³¹ was replaced by a Renilla Luciferase II and cloned in pTwistCMV expression vector using HindIII and NheI restriction sites of the vector. The NanoBiT-RhoA sensor comprising the SmBiT-RhoA and the LgBiT-PKN1-GDB constructs was described previously⁴². The NanoBiT-Rac1 sensor comprising the SmBiT-Rac1 and the LgBiT-PAK1-GDB constructs and the NanoBiT-Cdc42 sensor comprising the SmBiT-Cdc42 and the LgBiT-WAS-GDB constructs were generated by replacing firefly luciferase fragments of previously described Rac1 and Cdc42 constructs⁴¹ with the NanoBiT fragments. Specifically, the human Rac1 (residues 2-192) and the GTPase-binding domain (GBD) of the human PAK1 (residues 67-150) were N-terminally fused to SmBiT and LgBiT, respectively, with a 15-amino acid flexible linker (GGSGGGGSGGSSSGG). Similarly, the human Cdc42 (residues 2-191) and the GBD of the human WAS (residues 220-288) were N-terminally fused to SmBiT and LgBiT, respectively, with the flexible linker. Coding sequence for Rac1, PAK1-GDB, Cdc42 and WAS-GBD were human codon-optimized and gene-synthesized by Genscript and inserted into the pCAGGS plasmid using an NEBuilder assembly kit. βarr2-Strawberry⁹ was a gift from Prof. Larry Barak (Duke University, USA). Halo-miniGi (also referred to as miniGsi) was kindly provided by Prof. Nevin A. Lambert (Augusta University, USA). RFP-EEA1 (TagRFP-T-EEA1 cloned into pEGFP-C1 vector) and HA-Dyn-K44A (cloned into pcDNA3.1) constructs were gifts from Silvia Corvera and Sandra Schmid (respectively Addgene plasmids #42635 and #34683). RFP-Lck (C-tRFP-Lck cloned into PCMV6-AC-RFP expression vector) was purchased from Origene (#RC100049).

HEK293 cells (Life Technologies, cat. no. R705-07) were cultured in DMEM high glucose supplemented with 10% (v/v) fetal bovine serum (FBS) and 100 U/ml penicillin/streptomycin (1% (v/v) P/S). Jurkat T cells (ATCC, cat. no. TIB-152, clone E6-1) were cultures in RPMI 1640 media supplemented with 10% FBS and 1% P/S. The cell lines were not authenticated after receiving them from the supplier.

DNA for BRET or confocal imaging experiments to be transfected was combined with salmon sperm DNA to obtain a total of 1 µg DNA/condition. PEI was combined with DNA and incubated 20 minutes before added to the cells. DNA constructs for all other experiments were transfected into the cells using Lipofectamine 3000.

Transfected and construct-expressing HEK293 cells were placed in culture media containing 500 μg/mL G418 (HEK293-CCR7), 2 μg/mL puromycin (HEK293-CCR7-APEX2), 100 μg/ml zeocin (HEK293-CCR7 cells used to generate cell lines stably expressing PM-APEX2, ENDO-APEX2, or CYTO-APEX2), or 100 μg/mL zeocin + 100 μg/mL G418 (HEK293-CCR7-PM-APEX2, HEK293-CCR7-ENDO-APEX2, or HEK293-CCR7-CYTO-APEX2) to initiate selection of stably transfected cells. Individual clones/colonies were expanded and functional expression of constructs was verified by fluorescence imaging (PM-APEX2, ENDO-APEX2, or CYTO-APEX2), or CCL19-medtaed inhibition of forskolin-stimulated cAMP production (CCR7 and CCR7-APEX2).

The CAMYEL biosensor²⁸ (YFP-Epac-Rluc) transfected cells were equilibrated in HBSS supplemented with 10 mM HEPES (pH 7.4) at 37°C for 30 minutes. Coelenterazine-h was added at a final concentration of 5 μM before starting the measurement. After establishing a baseline response for 2 minutes, cells were stimulated with forskolin at a final concentration of 10 μM and the response was measured. Five minutes after the forskolin stimulation, CCL19 or CCL21 was added at a final concentration of 100 nM and the luminescence was measured for further 15 minutes. The signal was detected at 550 nm using a CLARIOstar instrument (BMG LabTech). No blinding from the different conditions was done.

Transfected cells were washed with DPBS and assayed in Tyrode’s buffer (137 mM NaCl, 0.9 mM KCl, 1 mM MgCl₂, 11.9 mM NaHCO₃, 3.6 mM NaH₂PO₄, 25 mM HEPES, 5.5 mM glucose, 1 mM CaCl₂, pH 7.4) at 37°C. 100 nM CCL19, or 100 nM CCL21, or vehicle were added to cells and incubated at 37°C for 30 minutes. 5 minutes before reading, Renilla luciferase II (RlucII) substrate (coelenterazine 400a; Deep Blue C^TM) was added at a final concentration of 2.5 µM. All BRET measurements were performed using a FLUOstar Omega microplate reader (BMG Labtech) with an acceptor filter (515 ± 30 nm) and donor filter (410 ± 80 nm). BRET was calculated by dividing GFP emission by RlucII emission. No blinding from the different conditions was done.

The PathHunter protein complementation assay (DiscoverX) using the PathHunter® HEK 293 CCR7 β-Arrestin cell line was performed according to the manufacturer’s protocol using and read for chemiluminescent signaling on a NovoStar plate reader (BMG Labtech). No blinding from the different conditions was done.

Multiple NanoBiT assays were performed between CCR7-SmBiT/LgBiT-CAAX, CCR7-SmBiT/LgBiT-FYVE, LgBiT-miniG/SmBiT-βarr1, LgBiT-RhoA/SmBiT-PKN1-GDB, LgBiT-PAK1-GBD/SmBiT-Rac1, and LgBiT-WAS1-GDB/SmBiT-Cdc42 to check the interaction between each biosensor pair. NanoBiT biosensor transfected cells were equilibrated in Opti-MEM™ at 37°C for 60 minutes. Coelenterazine-h was added at a final concentration of 10 μM before starting the measurement. After establishing a baseline response for 2 minutes, cells were stimulated with CCL19 or CCL21 added at a final concentration of 100 nM and the luminescence was measured for further 20 minutes. The signal was detected at 550 nm using a PHERAstar FSX instrument (BMG LabTech). No blinding from the different conditions was done.

Confocal microscopy experiments were conducted on CCR7 expressing HEK293 cells co-transfected with Halo-miniGi/βarr2-Strawberry, Halo-miniGi/RFP-Lck, or Halo-miniGi/RFP-EEA1. On the same day as the experiments, Oregon green Halo Tag ligand was added to Halo-miniGi expressing cells at a final concentration of 1 µM in the media and incubated 15 minutes at 37°C. Cells were washed 3 times with media and incubated 30 minutes for the last wash at 37°C. Media was then aspirated, replaced by Tyrode’s buffer and 100 nM CCL19, 100 nM CCL21, or vehicle were added and cells incubated for 30 minutes at 37°C. Cells were fixed with 4% paraformaldehyde in PBS for 10 minutes at room temperature, washed with DPBS, incubated 10 minutes in DPBS and then DPBS replaced by Tyrode’s buffer. Cells were visualized using a Leica SP8 confocal microscope and analyzed using Leica Application Suite X (LASX). The same microscope was used to determine correct subcellular localization of the PM-APEX2, ENDO-APEX2, and CYTO-APEX2 proteins in HEK293 cells. No blinding from the different conditions was done.

Co-localization was performed using Imaris software version 9.9.1 (Bitplane, Oxford instruments, Switzerland). Cells were selected using the “surfaces” module and cells containing regions of interest were selected. Channels for each fluorophore within the selected cells were masked to exclude all other cells or noise within the image. The Coloc module was then used to calculate co-localized voxels between the different channels. Thresholds were selected based on levels of intensity. Data are reported as percentage of red volume (red voxels) above the threshold that is co-localized with green volume (green voxels) above the threshold. No blinding from the different conditions was done.

HEK293-CCR7-APEX2 or Jurkat T cells were harvested and resuspended in assay buffer (500 nM IBMX + 20 mM HEPES in HBSS buffer, pH 7.4). In a white small volume 384-well plate, 5 μL of ligand buffer (assay buffer + 20 μM forskolin) containing 200 nM CCL19, 200 nM CCL21, or vehicle was mixed with 5 μL of cell suspension. The plate was incubated at room temperature for 30 minutes. Next, 10 μL of lysis buffer containing 1.25% Eu³⁺-anti-cAMP antibody and 1.25% cAMP-d2 was added and the plate was incubated for 1 hour at room temperature. The plate was read on an PHERAstar FSX instrument (BMG LabTech) where wells were excited with light at 340 nm and emission light was measured at 615 nm and 665 nm. The TR-FRET 665 nm/615 nm ratio, which is inversely proportional with the cAMP production, was used in combination with a cAMP standard curve to calculate the cAMP production in the cells. No blinding from the different conditions was done.

To confirm correct biotinylation in HEK293-CCR7-APEX2, HEK293-CCR7-PM-APEX2, HEK293-CCR7-ENDO-APEX2, or HEK293-CCR7-CYTO-APEX2 cell lines, western blotting was conducted using streptavidin-Alexa488. All biotinylated proteins were visualized on an Amersham Typhoon instrument (Cytiva).

To confirm expression of HA-tagged Dynamin-K44A mutant (HA-Dyn-K44A) or β-tubulin, western blotting was performed using a mouse anti-HA antibody followed by donkey anti-mouse IgG antibody conjugated to Alexa Fluor™ 488 or rabbit anti-β-tubulin antibody followed by donkey anti-rabbit IgG F(ab′)2 conjugated to HRP. Protein bands representing HA-Dyn-K44A were visualized on an Amersham Typhoon instrument (Cytiva) while bands corresponding to β-tubulin was visualized in ChemiDoc XRS+ (Bio-Rad) using SuperSignal^TM West Pico PLUS Chemiluminescent Substrate. No blinding from the different conditions was done.

Cell surface expression of CCR7-APEX2 was evaluated using ELISA. HEK293 cells or HEK293 cells stably expressing CCR7-APEX2 were seeded in poly-D-lysine coated 96-well white plates. Cells were fixed with 4% paraformaldehyde and blocked with 2% BSA in PBS, then incubated 1 hour at room temperature with Anti-FLAG M2-HRP conjugated antibody diluted 1:10,000 in blocking buffer. Enzymatic chemiluminescence was generated by addition of SuperSignal West Pico PLUS as substrate and measured on a CLARIOstar plate reader (BMG LabTech) with the emission filter 410-80.

Jurkat T cells suspended in culture medium were transferred into Costar® Transwell® polycarbonate membrane inserts with 5 μm pore size (Millipore Sigma). Each insert was fitted into a well in a 24 well plate containing culture media with either vehicle control solution, 100 nM CCL19, or 100 nM CCL21. Cells were allowed to migrate from the top insert to the lower well chamber of the plate for 2 hours at 37 °C at 5% CO₂. At the end of the experiment, all migrated cells from the lower chamber were collected and transferred to a Countess™ Cell Counting Chamber Slide (Thermo Fisher Scientific) and counted using a Countess™ 3 Automated Cell Counter (Thermo Fisher Scientific). Cells used for assessment of G_i/o protein involvement in CCR7-stimulated migration, were pretreated with 100 ng/ml PTX or vehicle control for 16 hours prior to the experiment. Cells used for assessment of the dependency of receptor internalization in CCR7-stimulated chemotaxis, were pretreated with 30 μM Dyngo-4a or inactive Dyngo control compound for 30 minutes prior to the experiment. In additional experiments the cells were pre-incubated with 10 μM PitStop2 or inactive PitNot control compound for 30 minutes prior to the experiment. Cells used for assessment of Rac1’s role in CCR7-stimulated chemotaxis were pretreated with 10 μM EHT1864 or DMSO control for 30 minutes prior to the experiment. All chemotaxic responses were normalized to passive cell migration towards the control as 0% and CCL19 as 100% response. No blinding from the different conditions was done.

All graphs were generated and analyzed using GraphPad Prism 8 (GraphPad Software). Data are presented as mean ± SEM and N referring to the number of independent experiments (or biological replicates) that have been conducted. Differences were assessed using Student’s t test for two comparisons and one- or two-way ANOVA and Tukey’s post hoc test for multiple comparisons. P < 0.05 was considered significant at the 95% confidence level.

CCR7-APEX2 transfected HEK293 cells were preincubated with 500 μM biotin-tyramide for 1 hour at 37 °C. 100 nM CCL19, 100 nM CCL21 or vehicle were added to stimulate cells for 0, 2, 10, or 25 minutes. The proximity labeling was initiated by addition of freshly diluted H₂O₂ from a 30% (v/v) stock solution to a final concentration of 1 mM. The culture media was removed exactly 1 minute later, and the cells were washed three times with ice cold quenching buffer (PBS supplemented with 10 mM sodium ascorbate, 10 mM sodium azide, and 5 mM Trolox). The cells were detached by adding 5 mL of the quenching buffer supplemented with 5 mM EDTA and incubating at room temperature for 10 minutes with agitation. Detached cells were transferred to a 15-mL conical tube and centrifuged at 1,000×g at 4°C for 10 minutes to remove the supernatant. The pellet was stored at −80°C until cell lysis.

The frozen cell pellet was lysed with 1 mL RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, pH 7.4) supplemented with 1 mM PMSF for 30 minutes at room temperature with agitation. The cell lysate was further solubilized by brief sonication and centrifuged at 12,000×g at 4°C for 10 minutes to remove any insoluble cell debris. NeutrAvidin™ agarose beads were pre-washed three times with 20 bed volumes of the RIPA buffer and incubated with the supernatant overnight at 4°C with agitation. After the incubation, the beads were washed by incubating at room temperature for 10 minutes each with 40 bed volumes of buffer A (2% SDS), buffer B (50 mM HEPES, 500 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, pH 7.4), then buffer C (10 mM Tris-HCl, 500 mM NaCl, 1 mM EDTA, 0.1% Tween-20, 0.5% sodium deoxycholate, pH 8.0) with agitation. The beads were further washed twice with 20 bed volumes of 100 mM HEPES (pH 7.4) and once with PBS (pH 7.4) to remove trace amount of detergent. Washed beads were resuspended in 100 μL PBS and subjected to mass spectrometry analysis. No blinding of the different conditions was done.

Proteins on the NeutrAvidin™ agarose beads were eluted by incubating with 50 μL of 50 mM Tris-HCl, 5% SDS, 10 mM TCEP, 20 mM CAA (pH 7.4) at 90°C for 20 minutes. Eluted proteins were purified and digested on magnetic beads following SP3 workflow⁶⁴. Eluates were transferred into clean Eppendorf tubes and mixed with SP3 magnetic beads. Proteins were precipitated by mixing with equal volume of ethanol and SP3 beads were washed three times with 200 μL of 85% ethanol to remove any traces of SDS from the elution buffer. The SP3 beads were suspended in 50 μL of 50 mM Tris-HCl (pH 7.4) and the proteins on the beads were digested with trypsin (4 ng/μL) at 37°C overnight. Trypsin digestion was quenched by adding TFA to a final concentration of 0.5%. Peptide digests were transferred from the magnetic beads into the clean tubes. SP3 beads were washed with 50 uL of 5% ACN 0.2% TFA. The peptides were loaded on Evotips C18 (Evosep) for subsequent LC-MS/MS analysis using Evosep One LC system coupled to Q Exactive™ HF-X Hybrid Quadrupole-Orbitrap™ MS instrument (Thermo Scientific) operating in data-independent acquisition mode (DIA). Peptides were separated online utilizing 15SPD method (88 min LC gradient length). High resolution full MS1 spectra were acquired with a resolution of 120,000, automatic gain control (AGC) target at 3,000,000 and maximum ion injection time at 60 ms, with a scan range of 350–1,650 m/z. Following each full MS1 scan, 22 data-independent high energy collisional dissociation (HCD) MS/MS scans were acquired at the resolution of 30,000, AGC target of 3,000,000 with stepped normalized collision energy (NCE) of 22.5, 25, and 27.5 Collected DIA data were analyzed using Spectronaut software (Biognosis; https://biognosys.com/shop/spectronaut) and searched in directDIA mode against the SwissProt subset of the human UniProt database (http://www.uniprot.org/). Database search was performed in integrated search engine Pulsar (Biognosis). For the database search, the enzyme specificity was set to trypsin with the maximum number of missed cleavages allowed set to two. Oxidation of methionine was searched as variable modification, whereas carbamidomethylation of cysteines was searched as a fixed modification. The false discovery rate (FDR) for peptide, protein, and site identification was set to 1%. Protein quantification was performed on MS2 level using 3 most intense fragment ions per precursor. Proteins identified with a single peptide was removed from further analysis. The mass spectrometry raw files are accessible under MassIVE ID: MSV000090362 at https://massive.ucsd.edu. Subsequent data analysis steps were performed in Perseus and GraphPad Prism.

For spatiotemporal characterization of CCR7 upon agonist stimulation, global median normalization was performed for dataset from each time points (0, 2, 10, and 25 minutes after agonist stimulation). Pairwise differential expression analysis was performed by comparing the stimulated datasets (2, 10, and 25 minutes) to the unstimulated (0 minute) dataset using t-test model, assuming equal variance for each protein being compared. For each stimulated time point, proteins with p value < 0.05 and log₂(fold-change) > 1 were considered significant. Significance score for the proteins were calculated by taking the absolute value of the log₂(fold-change) and adding −log₁₀(p value)⁶⁵.

Proteins with relevance to Rho, Rac, or CDC42 were selected manually from the list and subjected to further analysis for the role of CCR7 in cell motility. These proteins were searched on the STRING database to build a network functional relevance.

For spatial references PM-APEX2, ENDO-APEX2, and CYTO-APEX2, global median normalization was also performed. Pairwise differential expression analysis between each pair (ENDO against PM, ENDO against CYTO, and PM against CYTO) were performed as above. For each comparison pair, proteins with p value < 0.05 and log₂(fold-change) > 1 were considered significant. Then, proteins commonly appear significant for ENDO and PM from the t-tests were chosen as spatial references.