Mechanistic model of oxytocin/vasopressin-stimulated GPCR.
Nicolas Azzopardi1, Caroline Gora1, Lucile Drobecq1, Romain Yvinec1,2, Lucie Pellissier1.
(1) INRAE UMR85, CNRS UMR7247, University of Tours, Tours, France. (2) Musca Inria, Inrae Paris-Saclay, Palaiseau, France.
Objectives: Around 30% of commercialised drugs interacts at cell membrane with a G protein-coupled receptor (GPCR). Oxytocin (OT) and vasopressin (AVP) are hypothalamic peptides that bind to four GPCR: oxytocin receptor OTR and vasopressin receptors V1A, V1B and V2. These binding lead to different interactions with downstream signalling transducers. Known for their prosocial effect, these pathways are an active area of research in autism spectrum disorders. Understanding downstream intracellular signal transduction dynamics is mandatory to ensure drugs safety and efficacy.
Bioluminescence resonance energy transfer (BRET) assays the proximity of a bioluminescent donor and a fluorescent acceptor. Coupling donor and acceptor with two proteins of interest gives a unique insight in these protein-protein interactions in real time. Thus, it is a pertinent tool for quantitative monitoring of these actors and their dynamics in signal transduction. Data extracted from BRET are mainly analysed as the area under curve (AUC) of signal intensity ratio of the acceptor's fluorescence and the donor's bioluminescence. However, this approach presents some loss of information e.g., the inter-assay intensity variability, the signals time-dependent relationship modulation or the influence of the limits of quantification of the signals.
Therefore, this work investigates the pertinence of a novel approach of the BRET data analyses based on mechanistic population compartmental modelling and proposes new quantifications of the interaction of oxytocin or vasopressin-stimulated GPCR with downstream G proteins and β-arrestins.
Methods: A total of 480 BRET assays were performed to quantify the interaction of murine OTR, V1A V1B, and V2 or human receptors tagged with the Renilla luciferase (Rluc8) with downstream β-arrestin 1, β-arrestin 2, Gs, Gi or Gq tagged with enhanced yellow fluorescent protein (Venus), after stimulation with OT or AVP. OT and AVP doses were ranging from 0 to 10⁻4 M. Upon substrate coelenterazine addition Rluc8 emits a 480nm signal (S1). If Venus is present within a 100Å area, it is excited and emits a 530nm signal (S2). The intensities of S1 and S2 are monitored and used for analysis. BRET assays were performed in murine neuroblastoma cell (mNeuro2a).
The structure of the compartmental model is identical for all the assays. The delay of the coelenterazine to reach Rluc8 is described with a five transit-compartments model with estimated mean absorption time MAT. The degradation of the substrate by Rluc8 is quantified by the estimated first order rate kdeg. This elimination is directly equal to the emission fluorescence signal S1 and proportional with the bioluminescence signal S2 with a factor ρ. To work around the high inter-assay variability in signal intensity, total S1 dose is set to 1. A scale correction is assured by the estimation of a global gain G defined as OBS_S1=S1·10^(G/10) and OBS_S2=S2·10^(G/10). Parameters were estimated using non-linear mixed effect model (Monolixsuite 2023R1, lixoft, Antony, Frannce).
Results: BRET data were correctly described. All the parameters values and their variabilities can be estimates with r.s.e. < 10%. MAT was estimated for all assays to 4.66 minutes. Substrate consumption rate kdeg was 0.038 min-1 (r.s.e = 1%) with ωkdeg = 0.1. The global gain G was estimated as a normal distribution with high variability ωG = 2.2 (r.s.e. = 3.4%) and with mean 77.8 dB for V1A, 75 dB for V1B and 71.2 dB for V2. The signal proportionality ρ was estimated to 0.77 and was positively influenced by ligand dose β_dose = 2043 (r.s.e = 8,3 %). No significative correlation between estimated parameters were observed.
Conclusions: This modelling approach gives robust description of BRET data. Estimation of a global gain informs about cell state and level of receptor expression variabilities. Mixed effect modelling allows to quantify the potential influences of the different conditions assays on signals.
Despite OT and AVP share similar molecular characteristics they display important differences in signalling transduction at the four receptors. The same modelling approach can be used for flowing downstream protein-protein interactions of the signal transduction cascade.
