I8-arachnotocin–an arthropod-derived G protein-biased ligand of the human vasopressin V2 receptor

The neuropeptides oxytocin (OT) and vasopressin (VP) and their G protein-coupled receptors OTR, V1aR, V1bR, and V2R form an important and widely-distributed neuroendocrine signaling system. In mammals, this signaling system regulates water homeostasis, blood pressure, reproduction, as well as social behaviors such as pair bonding, trust and aggression. There exists high demand for ligands with differing pharmacological profiles to study the physiological and pathological functions of the individual receptor subtypes. Here, we present the pharmacological characterization of an arthropod (Metaseiulus occidentalis) OT/VP-like nonapeptide across the human OT/VP receptors. I8-arachnotocin is a full agonist with respect to second messenger signaling at human V2R (EC50 34 nM) and V1bR (EC50 1.2 µM), a partial agonist at OTR (EC50 790 nM), and a competitive antagonist at V1aR [pA2 6.25 (558 nM)]. Intriguingly, I8-arachnotocin activated the Gαs pathway of V2R without recruiting either β-arrestin-1 or β-arrestin-2. I8-arachnotocin might thus be a novel pharmacological tool to study the (patho)physiological relevance of β-arrestin-1 or -2 recruitment to the V2R. These findings furthermore highlight arthropods as a novel, vast and untapped source for the discovery of novel pharmacological probes and potential drug leads targeting neurohormone receptors.

Cell culture and transient receptor expression. All cell culture work was performed with human embryonic kidney cells 293 (HEK293) 24 . Unless otherwise stated, cells were incubated at 37 °C and 5% CO 2 and grown in Dulbecco's Modified Eagle's Medium (DMEM; Thermo Fisher Scientific, Australia and Fisher Scientific, Austria) containing 10% fetal bovine serum (GE Life Sciences, Australia and Sigma-Aldrich, Germany), 50 U/ mL penicillin and 50 U/mL streptomycin (Thermo Fisher Scientific, Australia and Sigma-Aldrich, Germany). Transient transfections were performed via Lipofectamine 2000 (Thermo Fisher Scientific, Australia) or jet-PRIME (Polyplus transfection, France) using 2 µg of pEGFP-N1 plasmid DNA coding for EGFP-tagged human OT/VP receptors 23 or a combination of 2 µg each of receptor-encoding plasmids and a plasmid coding for β-arrestin-1or -2-Nluc (subcloned into pcDNA3, with NanoLuc provided under a Limited Use Label License from Promega, Madison, USA).

Second messenger quantification.
Inositol-1-phosphate (IP 1 ) accumulation in response to Gα q coupled activation of human OTR, V 1a R, and V 1b R were measured using the IP-One Gq assay kit (Cisbio, France). Cells were seeded 4 h after transfection onto 384-well plates at a density of 10,000 cells per well and incubated for 2 days. At the time of the assay, all media was removed and the cells equilibrated to the provided stimulation buffer for 15 min at 37 °C. The cells were stimulated with peptide ligands at varying concentration (10 pM -30 µM) for 1 h at 37 °C. Cyclic adenosine monophosphate (cAMP) accumulation induced by Gα s coupling of V 2 R activation was determined using the LANCE Ultra cAMP detection kit (Perkin Elmer, Waltham, USA). Cells were re-passaged 4 h after transfection at a 1:2 ratio and incubated overnight. The next day, all media was removed, and cells were suspended with cAMP stimulation buffer (5 mM HEPES, 0.5 mM 3-isobutyl-1-methylxanthine, 0.1% bovine serum albumin in Hank's balanced salt solution, HBSS, pH 7.4) and the cells transferred onto a 384-well plate at a density of 300 cells per well. Stimulation was carried out for 30 min at 25 °C. Second messenger levels were measured by homogenous time-resolved fluorescence resonance energy transfer via fluorescence measurement on a Flexstation 3 (Molecular Devices, San Jose, USA) using the ratios 665/620 nm (IP 1 ) and 665/615 nm (cAMP) at an excitation wavelength of 340 nm. For antagonist screening, cells were stimulated with the endogenous ligand OT at OTR (50 nM), or VP for V 1a R (1 nM), V 1b R (3 nM) and V 2 R (0.5 pM), in presence (1 or 10 µM) and absence of I8-arachnotocin, as well as with 10 µM of the respective endogenous ligand.
Antagonism of I8-arachnotocin at V 1a R was characterized by Schild regression analysis (as published earlier) 19 . Briefly, several concentration-response curves of the endogenous agonist VP (as described above) were measured Figure 1. Phylogenetic relationship and molecular sequence of OT/VP-like neuropeptides. Taxonomic groups and neuropeptide names (in brackets). Conserved cysteines are highlighted in yellow, and the disulfide bond between Cys 1 and Cys 6 is indicated. *C-terminus amidated; #Renamed to arachnotocin to simplify nomenclature. Analogous to the existing phylum nomenclature, we refer to the mite-derived peptide by I8-arachnotocin, to distinguish it from arachnotocin (also referred to as crustacean OT/VP-like peptide in the past) 20,21 . β-arrestin-1 and -2 recruitment. Recruitment of β-arrestin-1and -2 upon receptor stimulation was measured via real-time measurement of bioluminescence resonance energy transfer (BRET) between β-arrestin-1/2-luciferase and EGFP-tagged receptors. Cells were co-transfected with β-arrestin-1/2-Nluc and OT/ VP receptor encoding plasmids at a ratio of 1:10. At 6 h post-transfection, the cells were transferred onto a white, clear bottom 96-well plate at 50,000 cells/well in phenol-red free DMEM containing 10% fetal bovine serum. The following day, the cells were serum starved for 1 h in phenol-free DMEM. Furimazine (Promega, Madison, USA), diluted 1:50 in HBSS, was added to the cells 5 min prior to monitoring at a 1:1 ratio. Light emissions were measured at 460 nm (Nluc) and 510 nm (EGFP) on a Flexstation 3 (Molecular Devices, San Jose, USA). After establishment of a baseline for 5 min, peptides diluted in HBSS were added and the response measured for 35 min. The ligand-induced BRET signal was calculated as: (emission EGFP ligand /emission Nluc ligand ) − (emission EGFP HBSS / emission Nluc HBSS ). Concentration-response curves were generated from the BRET signal at 5 min after addition of various peptide concentrations (10 pM -30 µM).
Immunoblotting for ERK 1/2. Immunoblotting was performed as described previously 25 . Briefly, following an overnight incubation and 16 h starvation HEK293 cells transiently expressing human V 2 R were treated with 1 µM I8-arachnotocin or 1 µM VP prepared in DMEM. Cells were incubated with peptide ligands at 37 °C for indicated periods and 1 mL of chilled 1x phosphate-buffered saline were used to terminate the incubation. After freeze-thaw cycle in liquid nitrogen cells were solubilized in 100 µl of lysis buffer (50 mM HEPES, 0.5% NP-40 substitute, 50 mM glycerol-2-phosphate, 250 mM NaCl, 5 mM EDTA, 2 mM imidazole, 1 mM Na 3 VO 4 , 1 mM hepta-molybdate, pH adjusted to 7.0 with NaOH), freshly added 1 mM PMSF, one cOmplete Mini EDTA-free tablet (Roche) and one PhosSTOP tablet (Roche) and then centrifuged at 10,000 × g for 10 min. The bicinchoninic acid assay (Micro-BCA kit, Pierce) was used to measure total protein content. Phosphorylated and total ERK 1/2 were detected by immunoblotting on the same membrane using the same exposure method with an anti-phospho-p44/42 MAPK antibody (ERK 1/2) (1:1,000; Cell Signaling Technology) and an anti-p44/42 MAPK (ERK 1/2) (1:1,000; Cell Signaling Technology), respectively. Detection and quantification of resulting bands were executed by secondary antibodies (Donkey anti rabbit 680RD and 880RD) and Odyssey Clx (LiCor Biosciences) infrared fluorescent imaging system, respectively. Data analysis. All data were analyzed using GraphPad Prism (GraphPad Software, San Diego) and all graphs were normalized to the activity of OT/VP above baseline. Concentration response curves were fitted to three-parameter non-linear regression curves with a bottom constrained to zero, a slope of one and sigmoidal shape at logarithmic scale to derive estimates of potency (EC 50 ) and maximum efficacy (E max ). Concentration response curves for Schild regression analysis 26 were additionally constrained to a top value of one hundred. All data were presented as mean ± SEM of at least three independent experiments (unless otherwise stated) conducted in triplicate.

Ethics Statement.
The study presented in this manuscript did not involve human or animal subjects.

I8-arachnotocin is a competitive antagonist at human V 1a R. Based on the lack of V 1a R activation,
we performed an antagonist screen of I8-arachnotocin across all four receptors (Fig. 4a). Receptor stimulation with VP (0.55 nM) in the presence of 1 and 10 µM I8-arachnotocin, yielded lower IP 1 levels than VP alone (Student's t-test, p = 0.0226). The concentration-response curves of VP on V 1a R in the absence and presence of I8-arachnotocin (1 µM, 3 µM, 10 µM) indicated an I8-arachnotocin-mediated dextral shift of the potency proportional to its concentration, without affecting E max , typical for a competitive antagonist (Fig. 4b). The dextral shift was evaluated via Schild regression analysis 26 yielding a linear regression slope of 1.28 ± 0.02 and a pA2 of 6.253 (~558 nM), thus demonstrating that I8-arachnotocin is a competitive antagonist of the V 1a R.
I8-arachnotocin does not induce β-arrestin-1 or -2 recruitment at V 2 R. With ligand bias becoming more relevant in understanding (patho)physiological responses, we also measured I8-arachnotocin-induced β-arrestin-2 recruitment. A concentration of 10 µM of the endogenous ligand induced rapid recruitment of β-arrestin-2 across all human OT/VP receptors, as judged by the increasing BRET signal from basal to maximum in 50-100 s. However, this effect was absent upon stimulation with 10 µM of I8-arachnotocin at OTR-, V 1a R-and V 2 R-expressing cells (Fig. 5).

I8-arachnotocin modulates V 2 R-mediated ERK signaling. Having established that I8-arachnotocin is
a G protein-biased peptide ligand, we examined its ability to modulate V 2 R-mediated ERK 1/2 phosphorylation. Previous studies reported that V 2 R activates ERK by two different pathways hypothesized to be dependent on either Gα s or β-arrestin signaling with distinct temporal patterns; the early phase (<5 min) being mediated by the G protein-dependent pathway, while the later phase (>10 min) being β-arrestin-2 dependent 29 . More recently, a study utilizing a combination of CRISPR/Cas9 genome-editing and pharmaceutical inhibition to generate HEK293 cells with 'zero functional G' indicated that ERK signaling mediated by V 2 R may still be dependent upon G protein even at later time points at which it may also be β-arrestin-dependent 30 . HEK293 cells transiently transfected to express the human V 2 R were stimulated with 1 µM I8-arachnotocin and 1 µM vasopressin for time periods between 1 min and 2 h. The immunoblotting data demonstrate that the late phase of V 2 R-dependent ERK 1/2 activation induced by I8-arachnotocin is significantly different from the late phase of VP-stimulated ERK 1/2 activation (30-120 min; Fig. 7c,d). While both peptides provoked an early and rapid ERK 1/2 phosphorylation peaking at 5 min, I8-arachnotocin-stimulated ERK 1/2 phosphorylation decreased over time in contrast to VP-elicited ERK 1/2 activation that resulted in a more sustained and prolonged ERK 1/2 activation (Fig. 7c,d). These data further support that I8-arachnotocin is a biased peptide ligand that modulates ERK 1/2 activity by preferentially activating the early phase G protein-dependent pathway at the plasma membrane over the late phase pathway that is β-arrestin-dependent.

Discussion
Peptides are gaining momentum in the drug development field, due to their (i) ability to interact with proteins on a large surface and (ii) structural and chiral complexity, which allows for improved discrimination between highly homologous targets as compared to small molecules 14 . This is particularly the case for the OT/VP signaling system, where multiple small molecule drugs failed due to selectivity issues and the majority of approved therapeutics are peptide drugs 15,31 . The study of the complex signaling pathways of the widely-distributed and fundamental OT/ VP signaling system remains however challenging due to the limited availability of pharmacological probes 14,15,18 . Differentiating between signaling events that occur pre or post β-arrestin recruitment has become an important focus [32][33][34] for studies looking at understanding the (patho)physiological roles of β-arrestin-mediated receptor internalization, desensitization and trafficking 35,36 . In addition, β-arrestin-dependent signaling has been linked to chronic stress-evoked melanoma metastasis via OTR 37 , increased neonatal rat cardiac fibroblast proliferation via V 1a R 38 , morphine tolerance via V 1b R 39 and sustained non-canonical signaling after receptor internalization via V 2 R, resulting in strong antidiuretic and anti-natriuretic effects 40 . Biased ligands such as I8-arachnotocin discovered in this work are thus important tools to advance our understanding in these areas of interest.
By utilizing a drug discovery strategy 14,41 on the synthesis of evolutionarily-conserved, yet distinct, peptides, we were able to bypass the time-and resource-consuming fractionation, isolation and identification steps  www.nature.com/scientificreports www.nature.com/scientificreports/ associated with the discovery of plant-or venom-derived compounds such as kalata B7 23 , inotocin 19 and conopressin T 22 . This strategy led us to explore the vast and untapped arthropod kingdom and resulted in the discovery and pharmacological characterization of I8-arachnotocin.
I8-arachnotocin activated the Gα s (cAMP) pathway, inducing ERK 1/2 phosphorylation without detectable recruitment of β-arrestin-1 or -2 at V 2 R, despite the capability of this receptor to form stable and strong interactions with β-arrestins 42 . These findings are consistent with the observation that I8-arachnotocin induced substantially lower levels of ERK 1/2 phosphorylation at later time points compared to VP, leading us to conclude that I8-arachnotocin displays a clear bias away from β-arrestin-dependent signaling at V 2 R. We are not aware of another V 2 R ligand capable of selectively activating the non-β-arrestin-dependent (EC 50 = 50 nM) vs. the β-arrestin-dependent (EC 50 > 100 µM) signaling pathway (>2,000-fold difference). Such biased ligands are highly sought-after for research tools that allow for the discrimination between multiple active conformations of GPCRs 32,33 . Since the pharmacology of the β-arrestin-1 and -2 pathways in respect to the OT and VP receptors is not fully elucidated yet 43,44 , I8-arachnotocin represents a valuable first probe to advance our understanding of this pathway at the human V 2 R.
Our data suggest that I8-arachnotocin can only effectively recruit β-arrestin-2 at the V 1b R (E max = 65%). To try to understand the structural differences resulting in bias between the four receptors, we compared the binding site residues of the V 1b R vs. the OTR, V 1a R and V 2 R; there are two positions that differ in these receptors, i.e. position 7.30 (Thr vs. Glu) and 7.42 (Asn vs. Ser) 19 . Although there are only a limited number of reports dealing with structural changes in GPCRs responsible for arrestin recruitment, it has been suggested for instance that residues in TM6 and TM7 are important for pathway selectivity 45,46 . Hence, the identified residues in positions Figure 7. Characterization of β-arrestin-1 recruitment and ERK 1/2 phosphorylation induced by I8arachnotocin vs. VP at the human V 2 R. (a) Kinetic profile of VP-and I8-arachnotocin-mediated β-arrestin-1 recruitment in HEK293 cells co-expressing EGFP-V 2 R and β-arrestin-1-Nluc. Cells were stimulated by 10 µM of VP or I8-arachnotocin, respectively, 5 min after addition of the luciferase substrate (furimazine). The results are shown as differences in the BRET signals in the presence of ligands and are expressed as the mean value ± SD; n = 2. (b) Concentration-response curves of VP and I8-arachnotocin at V 2 R using HEK 293 cells co-expressing EGFP-tagged V 2 R and β-arrestin-1-Nluc. Cells were pretreated with furimazine and measurements were taken 5 min after addition of ligands. Ligand-induced BRET was calculated as: (emission EGFP ligand /emission NLuc ligand ) − (emission EGFP HBSS /emission NLuc HBSS ). Results were normalized to β-arrestin-1 recruitment in response to VP. Data points were fitted by nonlinear regression curves (sigmoidal, slope = 1); error bars indicate SEM; n = 3. (c) Representative Western blot images of ERK 1/2 phosphorylation by stimulation of V 2 R with I8-arachnotocin vs. VP and (d) quantification of I8-arachnotocin-and VP-induced phosphorylated ERK 1/2 (pERK) relative to total ERK 1/2 (tERK) from four independent experiments (±SEM). Cells were transiently transfected with EGFP-V 2 R encoding plasmid and treated with 1 µM I8-arachnotocin or 1 µM VP at 37 °C for indicated time periods. Immunoblots were prepared from the same membranes using the same exposure method. Regions of interest were cropped from the full image (see Supplementary Information). Statistical significance was determined by Student's t test (*P < 0.05; **P < 0.01; ns non-significant).

Scientific RepoRtS |
(2019) 9:19295 | https://doi.org/10.1038/s41598-019-55675-w www.nature.com/scientificreports www.nature.com/scientificreports/ 7.30 and 7.42 of the ligand binding pocket of OT/VP receptors, could contribute to the observed bias of the bound I8-arachnotocin ligand, by altering the interaction of the receptor C-tail with the N-terminal domain of arrestin 47 . However, this remains speculative until future studies reveal further details.
Overall, GPCRs are prime drug targets 48 and the vast chemical diversity of nature will continue to deliver novel pharmacological and therapeutic leads, particularly with technological advances that accelerate the drug discovery pipeline 49 . This work follows this innovative trend by exploiting the evolutionary conservation and ubiquity of neuropeptides across the animal kingdom 50,51 . In particular, it highlights the abundance of neuropeptides in arthropods: e.g., there are >50-150 neuropeptides reported in the model species Tribolium castaneum 52 , Nasonia vitripennis 53 , Apis mellifera 54 and Drosophila melanogaster 55 . We thus argue that arthropods represent a novel, vast and untapped source for the discovery of pharmacological probes and potential therapeutic leads for a broad range of signaling systems.