Identification of TSPAN4 as Novel Histamine H4 Receptor Interactor

The histamine H4 receptor (H4R) is a G protein-coupled receptor that is predominantly expressed on immune cells and considered to be an important drug target for various inflammatory disorders. Like most GPCRs, the H4R activates G proteins and recruits β-arrestins upon phosphorylation by GPCR kinases to induce cellular signaling in response to agonist stimulation. However, in the last decade, novel GPCR-interacting proteins have been identified that may regulate GPCR functioning. In this study, a split-ubiquitin membrane yeast two-hybrid assay was used to identify H4R interactors in a Jurkat T cell line cDNA library. Forty-three novel H4R interactors were identified, of which 17 have also been previously observed in MYTH screens to interact with other GPCR subtypes. The interaction of H4R with the tetraspanin TSPAN4 was confirmed in transfected cells using bioluminescence resonance energy transfer, bimolecular fluorescence complementation, and co-immunoprecipitation. Histamine stimulation reduced the interaction between H4R and TSPAN4, but TSPAN4 did not affect H4R-mediated G protein signaling. Nonetheless, the identification of novel GPCR interactors by MYTH is a starting point to further investigate the regulation of GPCR signaling.


Introduction
Histamine is a key mediator of allergic inflammation and is released from mast cells and basophils upon allergen binding. The histamine H 4 receptor (H 4 R) is predominantly expressed on immune cells and mediates histamine-induced chemotaxis and production of inflammatory cytokines [1,2]. Importantly, H 4 R-deficient mice revealed a role of this receptor in pruritus, dermatitis, asthma, and arthritis disease models [2]. Consequently, the H 4 R has been recognized as potential anti-inflammatory drug target, and selective antagonists are currently in clinical trials to counteract atopic dermatitis, psoriasis, allergic rhinitis, bronchial allergen challenge, asthma, and rheumatoid arthritis [2].
In this study, the MYTH approach was used to identify GIPs of the human H 4 R from an unstimulated Jurkat T cell cDNA library [32]. In short, the H 4 R was fused at the N-terminus to the yeast STE2 leader sequence to target the fusion protein to the membrane of yeast, whereas the C-terminal half of ubiquitin (Cub) and the LexA-VP16 transcriptional activator were fused in frame to the intracellular C-terminal tail of H 4 R. The prey proteins encoded by the naïve Jurkat T cell cDNA library were N-terminally tagged with the N-terminal half of ubiquitin that contains an isoleucine 13 to glycine mutation (NubG) to prevent spontaneous ubiquitin reconstitution (Dualsystems Biotech, Switserland) ( Figure 1A). Interaction of unliganded H 4 R-Cub-LexA-VP16 (bait) with NubGtagged prey proteins in yeast allows reconstitution of a pseudoubiquitin moiety, which is subsequently recognized by endogenous ubiquitin-specific proteases resulting in the cleavage of the LexA-VP16 transcriptional regulator. The released transcriptional regulator translocates into the nucleus and activates transcription of LexA-promoter fused reporter genes for selection (LacZ, HIS3, ADE2) ( Figure 1B). yeast co-expressing the membrane targeted H 4 R-Cub-LexA-VP16 bait and cytosolic (depicted) or membrane-associated Nterminally tagged NubG prey proteins was generated by mating of haploid yeasts expressing the individual bait and prey constructs. (B) Interaction of the H 4 R-Cub-LexA-VP16 'bait' with NubG-protein 'prey' results in functional reconstitution of a pseudo-ubiquitin, which is subsequently recognized by cytosolic ubiquitin-specific proteases leading to the cleavage of the LexA-VP16 transcriptional activator and the expression of (HIS3, ADE2, LacZ) reporter genes. (C) Subcellular localization and biological function summary of the 43 hits were retrieved from the Uniprot database (https://www.uniprot.org; accessed on 24 June 2021). Gene names are indicated and full description of proteins with Uniprot codes are presented in Supplementary Table S1. Hits in bold have been found to interact with other GPCRs in MYTH screens (see Supplementary  Table S1), whereas frequent MYTH screen hits are indicated in italics. ER = endoplasmic reticulum; PM = post-translational modification (D) STRING analysis (v.11.0; https://string-db.org; accessed on 25 June 2021) of hits from the MYTH screen of H 4 R (bait) on an unstimulated Jurkat T cell DUALmembrane cDNA library. Known interactions between proteins in the STRING database are indicated by connecting lines. The thickness of the line represents the degree of confidence prediction of the interaction. This MYTH approach identified 43 novel GIPs for the H 4 R. The interaction of one potential GIP (i.e., human tetraspanin 4; TSPAN4) was further validated in transfected HEK293T cells using bioluminescence resonance energy transfer (BRET), bimolecular fluorescence complementation (BiFC), and co-immunoprecipitation. In addition, the consequence of this H 4 R-TSPAN4 interaction was evaluated in ligand binding and H 4 R-mediated G protein activation assays.

MYTH Constructs and Screen
The MYTH bait vector pBT3-STE-H 4 R was generated by PCR. A SfiI fragment containing the full open reading frame of H 4 R was subcloned in frame downstream of the STE2 leader sequence and upstream of the sequence encoding for Cub (i.e., amino acids 34-76 of ubiquitin) a GG-containing linker and the LexA-VP16 transcriptional regulator in the pBT3-STE bait vector that contains a LEU2 marker (Dualsystems, Switzerland). The DUALmembrane prey library from Dualsystems (catalog # P02205) consists of cDNA sequences from unstimulated Jurkat T cell containing~9 × 10 6 independent clones, with an average size of 1.5 kb, that were N-terminally fused to NubG in the TRP1 markercontaining pDSL-Nx prey vector (Dualsystems, Switzerland). Hits from the H 4 R MYTH screen were analyzed using Uniprot database (https://www.uniprot.org; accessed 24 June 2021) [33] and the protein-protein network tool STRING (v11.0; https://string-db.org; accessed on 25 June 2021) [34].

Cell Culture and Transfection
HEK293T cells (ATCC; Manassas, VA, USA) were cultured in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% FBS and 1% penicillin/streptomycin (50 µg/mL) at 37 • C, 5% CO 2 . Cells (2 × 10 6 ) were seeded in a 10 cm dish and transiently transfected the next day with indicated amounts of DNA plasmids using 20 µg 25 kDa linear polyethylenimine (PEI), as previously described. Empty pcDEF3 plasmid was used to keep total DNA amounts at 5 µg for each transfection. For saturation BRET experiments by gene-dosing, 2 × 10 4 cells/well were seeded in 0.1% poly-L-lysine-coated white bottom 96well plates and transiently transfected the next day using 25 kDa linear PEI, as previously described [36].

Mammalian Expression Constructs
The human H 4 R-Nluc fusion (H 4 R; NM_021624.3) construct was generated by subcloning H 4 R-Rluc8 into H 1 R-Nluc/pcDEF3 using flanking KpnI and SpeI restriction sites, as previously described [14,37]. Human TSPAN4 (TSPAN4; NM_003271.4) was genetically fused to mVenus at the intracellular N-or C-terminus by substituting start or stop codon, respectively, with SpeI-NotI restriction sites (coding for TSAAA linker) using PCR as previously described [14]. The H 4 R fusion to the N-terminal split-fragment of mVenus (V 1 : amino acids 1-155) in pcDEF3 was previously reported [38], whereas the C-terminal splitfragment of mVenus (V 2 : amino acids 156-240) was genetically fused to the N-terminus of TSPAN4 via the aforementioned TSAAA-linker sequence using PCR and subsequent subcloned into pcDNA3.1 or pcDEF3 using the introduced restriction enzymes. HA-H 4 R in pcDEF3 was previously described [39]. Gα i2 protein biosensor plasmid was kindly provided by Dr. Schihada (Karolinska Institutet, Department of Physiology and Pharmacology, Stockholm, Sweden) [40]. All constructs were verified by DNA sequencing.

Bioluminescence Resonance Energy Transfer (BRET)-Based Close Proximity Detection
HEK293T cells were transiently transfected with 12.5 ng H 4 R-Nluc/pcDEF3 in combination with 0 to 500 ng mVenus-TSPAN4/pcDNA3.1 or TSPAN4-mVenus/pcDEF3 plasmids per 10 6 cells in white-bottom 96-well plates. Forty-eight hours after transfection, medium was aspirated from the cells and replaced by assay buffer Hank's Balanced Salt Solution (HBSS). Luminescence (lum) was measured in time upon stimulation with vehicle or 10 µM histamine in the presence of Nanoglo (3.2 µL/mL) at 37 • C using the Mithras LB940 multimode microplate reader (Berthold, Germany) at 540-40 nm and 480-20 nm. The expression of mVenus-TSPAN4 or TSPAN4-mVenus was measured as fluorescence (fluo) at 540 nm emission upon excitation at 485 nm in the Mithras LB940 plate reader. The BRET ratio signal was calculated as the 540 lum/480 lum emission ratio and presented as function of the mVenus/Nluc expression levels as calculated by 540 fluo/480 lum ratio [41].
Saturation BRET curves were fitted using the nonlinear One site-specific binding model in GraphPad Prism 8.0. For histamine concentration-response curves, cells were co-transfected in 10 cm dishes with 100 ng H 4 R-Nluc/pcDEF3 and 2 µg mVenus-TSPAN4/pcDNA3.1 or TSPAN4-mVenus/pcDEF3 plasmids and transferred the next day into white-bottom 96-well plates (5 × 10 4 cells/well). Two days after transfection, BRET signal was measured in the presence of increasing concentration histamine at 37 • C.

Biomolecular Fluorescence Complementation (BiFC)-Based Close Proximity Detection
HEK293T cells were co-transfected with 0.5 µg H 4 R-V1/pcDEF3 and 0.5 µg V2-TSPAN4/pcDNA3.1 per dish. The next day, 8 × 10 5 cells/well were transferred on 0.1% poly-L-lysine-coated cover slides in 6-well plates. Forty eight hours post-transfection, the cells were stimulated with vehicle or 10 µM histamine for 30 min at room temperature. After fixation by 4% paraformaldehyde (PFA), the cells were stained with DAPI for nuclear staining and the reconstituted green fluorescence were visualized with an Olympus FSX-100 microscope at 475/30 nm excitation and 535/30 nm emission.

[ 3 H]Histamine Binding Assay
HEK293T cells were transiently transfected with 100 ng H 4 R-Nluc and/or 2 µg mVenus-TSPAN4 or TSPAN4-mVenus plasmids and collected after two days, as previously described [7]. The cells were homogenized in binding buffer (50 mM Tris-HCl, pH 7.4) and incubated with increasing concentrations of [ 3 H]histamine (0-40 nM) in duplicate for 2 h at 25 • C in the presence and absence of 50 µM JNJ7777120 to detect total and nonspecific binding, respectively. Incubation was stopped by rapid filtration through 96-well GF/C plates that were pre-soaked with 0.5% (v/v) PEI, using a 96-well Filtermate harvester (PerkinElmer, Groningen, The Netherlands). Next, the GF/C filter plates were rapidly washed four times with ice-cold wash buffer (50 mM Tris-HCl, pH 7.4, 4 • C) and dried at 52 • C for at least 30 min before addition of 25 µL/well Microscint-O to quantify radioactivity using the Microbeta Wallac Trilux scintillation counter (Perkin-Elmer). Protein concentration of the cell homogenates was measured by a BCA protein assay, according to the manufacturer's recommendation. Binding affinity and B max values were determined using the One site-total and nonspecific binding model in GraphPad Prism 8.0.

Data Analysis
Regression and statistical analyses of experimental data was performed by GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, USA). ImageJ software (National Institutes of Health, Bethesda, MD, USA) was used for Western blot quantification.

Identification of H 4 R Interactors by MYTH Screen of Jurkat T Cell cDNA Library
To identify H 4 R interactors, a MYTH screen was performed by Dualsystems Biotech AG (Schlieren, Switzerland) using their unstimulated Jurkat T cell DUALmembrane cDNA library. Diploid yeast clones that grew under selection were picked from primary screening plates and transferred to liquid screening medium. Clones were passaged for five rounds to eliminate non-specific interactors and subsequently assayed for the lacZ reporter gene using a β-galactosidase assay. Screening of 2.5 × 10 6 transformants yielded 43 potential interactors of H 4 R ( Figure 1C; Supplementary Table S1). Seven (16%) of these interactors (i.e., MT-ATP6, MT-CO2, MT-CO3, COX8A, SSR3, OST4, and SERP1) were recognized as highly connected proteins that are hits in the majority of MYTH screens by Dualsystems. The 43 hits were analyzed using the STRING protein-protein association network analysis tool but did not reveal known interconnections between bait (i.e., HRH4) and identified prey proteins ( Figure 1D). The relative abundance of membrane (-associated) proteins (84%) amongst the prey proteins seems evident and the presence of two members of the membrane-associated tetraspanin family (CD63, TSPAN4) was noted ( Figure 1C; Supplementary Table S1).

MYTH between H 4 R and Tetraspanins TSPAN4 and CD63 Is Decreased by Histamine Stimulation
We have selected CD63 and TSPAN4 for in-house MYTH validation as tetraspanins have been reported in immune cells to act as scaffolds for signaling cascade components, mediate cell-cell communication, and are involved in cellular migration [42,43].
To this end, we used the STE2-H 4 R-Cub-LexA-VP16 construct as bait and validated its expression and potential to interact with Nub-tagged baits. The bait construct was validated, generating diploid strains carrying either wild type Nub expressing vector (pNub wt) or empty vector (pXN21) as prey plasmids. Besides growing on the -leu-trp medium, selecting for diploids, the positive control displayed growth on media without histidine, indicative of interaction between bait and prey (Supplementary Figure S1).
Validation of the prey plasmids was made by generating diploids either with an empty bait plasmid (pMETYc) or the H 4 R-Cub-LexA-VP16 bait plasmid. The diploids with the pMETYc negative control as bait only grew on media with histidine, whereas the H 4 R-TSPAN/CD63 combinations displayed growth on selective media without histidine (Supplementary Figure S1).
Two-hybrid interactions between H 4 R-Cub-LexA-VP16 and NubG-TSPAN4 ( Figure 2A) and NubG-CD63 ( Figure 2C) were confirmed as growth in liquid media. Again, the diploid strains expressing bait and prey are able to grow in media without histidine, indicative of the interaction between H 4 R and TSPAN4 and CD63, respectively. Next, we examined the possibility to modulate these interactions by addition of 10 µM histamine. Growth of the diploids expressing the H 4 R-Cub-LexA-VP16 as bait in combination with NubG-TSPAN4 ( Figure 2B) or NubG-CD63 ( Figure 2D) was diminished by histamine, suggesting reduced interactions.

Saturation BRET Confirms Interaction between H 4 R and TSPAN4 in HEK293T Cells and Can Be Reduced by Histamine Stimulation
To monitor the interaction between H 4 R and TSPAN4 in mammalian cells by bioluminescence resonance energy transfer (BRET), the bioluminescent enzyme NanoLuc (Nluc) was fused in-frame to the intracellular C-terminal tail of H 4 R, whereas the fluorescent protein mVenus was fused to either the intracellular N-or C-terminus of TSPAN4 ( Figure 3A,B, respectively).
Co-transfection of HEK293T cells with a constant amount of H 4 R-Nluc (i.e., BRETdonor) in combination with increasing amounts of either mVenus-TSPAN4 or TSPAN4-mVenus (i.e., BRET-acceptors) plasmids resulted in hyperbolic BRET signals ( Figure 3C,D, respectively). These saturable BRET signals suggest that Nluc and mVenus are brought in close proximity (<10 nm) as a consequence of specific interactions between H 4 R and TSPAN4 in transfected HEK293T cells rather than random collisions [41].

BiFC Microscopy Reveals H 4 R-TSPAN4 Complexes in HEK293T Cells
To confirm the close proximity between H 4 R and TSPAN4 in HEK293T cells, we fused the N-terminal fragment of mVenus in frame to the intracellular C-terminal tail of H 4 R (i.e., H 4 R-V1) and the C-terminal fragment of mVenus to the intracellular N-terminal tail of TSPAN4 (i.e., V2-TSPAN4) to monitor bimolecular fluorescence complementation (BiFC) of these mVenus fragments that is driven by the interaction by the proteins to which they are fused ( Figure 4A,B), as previously described for complexes between GPCRs [38,44]. Indeed, co-expression of H 4 R-V1 and V2-TSPAN4 in HEK293T resulted in BiFC ( Figure 4C), which matches the localization of mVenus-TSPAN4 ( Figure 4D). BiFC detection of proteinprotein interactions by mVenus reconstitution is known to be virtually irreversible and as anticipated no decrease in fluorescence was observed upon histamine stimulation ( Figure 4E,F) [45,46]. In addition, histamine seemed not to affect the localization of the H 4 R-TSPAN4 complexes as compared to vehicle-stimulated cells.

TSPAN4 Co-Immunoprecipitates with H 4 R from Co-Expressing HEK293T Cells
To confirm the physical interaction between H 4 R and TSPAN4, HEK293T cells were transfected with HA-H 4 R and mVenus-TSPAN4, and cell lysates were subjected to immunoprecipitation using an anti-HA agarose beads. Both lysates and immunoprecipitates were resolved by SDS-PAGE and subsequently immunoblotted using rat anti-HA and goat anti-mVenus antibodies. To verify that co-immunoprecipitated H 4 R-TSPAN4 complexes are not the consequence of non-specific aggregation during the cell solubilization procedure, cells transfected with either HA-H 4 R or mVenus-TSPAN4 were mixed prior to the solubilization step. HA-H 4 R and mVenus-TSPAN4 were detected in immunoblots of both co-expressing (Co-) and mixed (Mix) cells ( Figure 5A left panel), whereas mVenus-TSPAN4 was only detected in co-expressing cells (Co-) upon immunoprecipitation of HA-H 4 R confirming their physical interaction ( Figure 5A right panel). Stimulation of these co-expressing cells (Co+) with 10 µM histamine did not seem to affect the co-immunoprecipitation of TSPAN4 with H 4 R ( Figure 5A,B), which is in line with the only small decrease of basal BRET and MYTH. . Statistical difference versus Cowas analyzed using one-way ANOVA followed by Dunnett's multiple comparison test and indicated by an asterisk p < 0.05. NS = no significant difference.

Overexpression of TSPAN4 Does Not Affect Histamine Binding to H 4 R or Signaling
To evaluate whether the basal interaction of the TSPAN4 affect H 4 R functioning, histamine binding and signaling was evaluated. Co-expression of mVenus-TSPAN4 or TSPAN4-mVenus did not significantly affect the binding affinity of radiolabeled [ 3 H]histamine for H 4 R as compared to cell homogenates expressing only H 4 R (Table 1; Supplementary  Figure S2A-C). The 5.5-fold reduced total number of H 4 R (B max ) in cells co-transfected with TSPAN4 as compared to cells expressing only H 4 R is most likely the consequence of overexpression of TSPAN4 resulting in an inefficient transcription and translation of H 4 R ( Table 1).
Co-expression of TSPAN4 did also not affect H 4 R-mediated heterotrimeric Gα i2 protein activation in response to histamine (pEC 50 = 7.8 ± 0.13) as compared to cells expressing only H 4 R (pEC 50 = 7.8 ± 0.03) as detected by BRET-based G protein activation sensor ( Figure 6A,B).

Discussion
In the last decade, studies have shown that GPCRs can interact with numerous membrane-associated and cytosolic protein in addition to their well-known coupling to heterotrimeric G proteins, GRKs, and β-arrestins [21,[25][26][27][28][29][30][31]. Indeed, GPCR signaling is primarily mediated via the heterotrimeric G proteins, whereas GRKs and β-arrestins play key functions in regulating the duration of G protein signaling. These so-called GPCRinteracting proteins (GIPs) have been found to modulate GPCR activity by for example anchoring GPCRs with their signaling partners in subcellular membrane compartments (e.g., lipid rafts) for directional signaling (e.g., cell migration), or forming signalosome complexes to signal in G protein-independent manner [16][17][18]. In contrast to most G proteins, GRK, and β-arrestin, GIPs are often less widely expressed allowing cell-type specific fine-tuning of GPCR activity [16].
The MYTH approach has been successfully used to screen a human fetal brain DUALmembrane NubG-X cDNA library using~50 other GPCRs as bait and yielding 700 potential GIPs [26,27,29], whereas validated GIPs have also been identified in MYTH screen using human liver and pancreatic islets cDNA libraries for the GLP1-R [30,31]. As the H 4 R is predominantly expressed in the immune cells, we have used the MYTH assay to screen a Jurkat T cell Dualmembrane cDNA library for H 4 R interactors, which yielded 43 potential GIPs. Remarkably, known H 4 R interactors such as Gα i/o proteins, GRKs, and β-arrestins were not retrieved in this MYTH screen. This might be related to the fact that an inactive, unliganded H 4 R conformation was used as bait in yeast cells, whereas an active (agonist-bound) H 4 R conformation is required to interact with these proteins, and additional phosphorylation of the H 4 R C-terminal tail in the case of β-arrestins1/2 [14]. In-deed, GRKs and β-arrestins have not been reported MYTH screens with~50 GPCRs as bait, whereas only few GPCRs showed interaction with Gα s (i.e., cysteinyl leukotriene receptor 2 (CYSLTR2), muscarinic acetylcholine receptor M3 (M3), Relaxin-3 receptor 1 (SALPR), and glucagon-like peptide 1 receptor (GLP1R)) and Gα 12 (i.e., type-1 angiotensin II receptor (AGTR1)) but not the other Gα subtypes in these screens [27,[29][30][31]. The unliganded H 4 R bait identified 11 putative GIPs from the Jurkat T cell DUALmembrane cDNA library that were previously observed, but not further validated, in human fetal brain DUALmembrane cDNA library for other GPCRs ( Figure 1C and Supplementary Table S1) [29]. Two of these H 4 R putative hits ATP6AP2 and CD63 were also found to interact with GLP1R in human pancreatic and liver islets cDNA libraries MYTH screens, respectively, whereas six other interactors were only shared with GLP1R but none of the other tested GPCRs ( Figure 1C and Supplementary Table S1) [30,31]. The interaction of SERP1 and ATP6AP2 with GLP1R were confirmed by co-immunoprecipitation experiments, and found to be important for Nlinked glycosylation in HEK293 cells and GLP1-induced Ca 2+ influx and insulin secretion in INS-1 beta cell lines, respectively [30,31]. Interestingly, CD63 has been earlier identified in a lentiviral cDNA library screen that made T cells resistant to human immunodeficiency virus (HIV)-induced cell death [47]. To this end, CD63 prevented viral entry by targeting the HIV co-receptor CXCR4 from the Golgi apparatus to late endosomes instead of cell surface [48]. In addition, CD63-mediated downregulation of CXCR4 in activated B cells was found to be essential for their migration from the dark into the light zone of germinal centers in secondary lymphoid organs formation [49]. Indeed, interaction of CD63 and CXCR4 was confirmed by co-immunoprecipitation and BiFC in transfected cells [48,50]. CD63 belongs to the structurally conserved tetraspanin family of 33 membrane-associated proteins in humans consisting of four transmembrane helices with an intracellular N-and C-termini, one intracellular loop, and two extracellular loops. In this study, H 4 R was also found to interact with TSPAN4, another tetraspanin family member, belonging to a different subfamily (i.e., CD versus CD63 subfamilies, respectively) based on sequence conservation [43]. TSPAN4 was also found to interact with the leukotriene B4 receptor 2 (LTB4R2) in a MYTH screen of human fetal brain DUALmembrane cDNA library, but was not further validated [29]. In the same study, five other tetraspanin family members (i.e., TSPAN3, TSPAN27, TSPAN33, and in particular TSPAN7 and TSPAN28) were also found to interact with a number of GPCRs (i.e., TSPAN3 with three GPCRs, TSPAN7 with 12 GPCRs, TSPAN27 with two GPCRs, TSPAN28 with 12 GPCRs, and TSPAN33 with one GPCR), but again no validation in mammalian cell lines was provided [29]. Considering that tetraspanins have been recognized as regulators of cellular signaling with a potential for therapeutical targeting [42,43], we decided to focus our validation on TSPAN4. Interestingly, the interaction between H 4 R and TSPAN4 was slightly reduced (<5-10%) in both MYTH yeast growth and BRET assays in HEK293T cells by histamine stimulation. In contrast, no effect was observed of TSPAN4 overexpression on histamine binding or H 4 R-mediated G i protein activation in response to histamine. Although we cannot exclude that a fraction of H 4 R might not be in complex with TSPAN4 in these cells, the ratios between H 4 R-and TSPAN4-expressing plasmids that were used in the binding and G protein activation corresponds to approximately 80-90% and 60-70% of the saturation BRET maximum plateaus ( Figure 3C,D), suggesting that the majority of H 4 R was indeed in complex with TSPAN4. Instead of overexpression studies, future research might use CRISPR/Cas9 genomic editing to inhibit TSPAN4 (and/or CD63) expression in immune cell lines to shed more light on the potential role of tetraspanins in the regulation of H 4 R functioning, or vice versa. Interestingly, TSPAN4 was recently reported to play a role in the formation of migrasomes on retracting fibers at the rear end of migrating cells [51,52]. Migrasomes are extracellular vesicles that contain signaling molecules and can be taken up by other cells to mediate cell-to-cell communication. Importantly, migrasome formation requires cell migration and the H 4 R has been reported to mediate chemotaxis of various immune cells towards histamine [9,11,[53][54][55][56][57][58][59]. Hence, it is tempting to speculate whether histamine might stimulate migrasome generation by reducing the interaction between H 4 R and TSPAN4, so that the latter can be enriched in the migrasomes.
The H 4 R is also expressed in the brain and a bioinformatics approach was previously used to predict brain-specific H 4 R interacting proteins [60]. The integration of Knowledgegram and Predictogram computational analyses revealed 15 potential brain-specific GIPs for the H 4 R, which did not overlap with the potential GIPs in the here reported MYTH screen on Jurkat T cell cDNA. Experimental confirmation of these predicted brain-specific H 4 R GIPs by performing for example MYTH on brain cDNA might reveal cell type-specific regulation of H 4 R functioning.
In conclusion, our MYTH screen using H 4 R as bait on Jurkat T cell cDNA library identified 43 novel interactors, of which 17 have been previously reported for other GPCRs in MYTH screens on different cDNA libraries. These MYTH datasets might collectively provide valuable clues in the role of GIPs in regulating GPCR activities, but do require experimental validation for example by analyzing the consequences of GIP knockdown using CRISPR/Cas9 or using this genomic editing approach in combination with sensitive biosensors (e.g., NanoBiT) to measure the dynamics of these interactions at physiologically relevant expression levels [61].
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/biom11081127/s1, Table S1: Proteins identified by MYTH screen on Jurkat T cell DUALmembrane cDNA library using unliganded hH 4 R as bait. Figure S1: Validation of MYTH constructs.