Common coupling map advances GPCR-G protein selectivity

Abstract

Two-thirds of human hormones and one-third of clinical drugs act on membrane receptors that couple to G proteins to achieve appropriate functional responses. While G protein transducers from literature are annotated in the Guide to Pharmacology database, two recent large-scale datasets now expand the receptor-G protein 'couplome'. However, these three datasets differ in scope and reported G protein couplings giving different coverage and conclusions on G protein-coupled receptor (GPCR)-G protein signaling. Here, we report a common coupling map uncovering novel couplings supported by both large-scale studies, the selectivity/promiscuity of GPCRs and G proteins, and how the co-coupling and co-expression of G proteins compare to the families from phylogenetic relationships. The coupling map and insights on GPCR-G protein selectivity will catalyze advances in receptor research and cellular signaling toward the exploitation of G protein signaling bias in design of safer drugs.

Keywords: G protein; GPCR; computational biology; human; pharmacology; signal transduction; systems biology.

Figure 1.. Coverage and agreement of the Bouvier lab, Inoue group, and Guide to Pharmacology (GtP) datasets.

(A) Biosensor principles used by the Bouvier and Inoue groups, and literature annotation stored in the GtP database. (B) Intersection of the G protein-coupled receptors (GPCRs) included in the Bouvier (n = 100), Inoue (n = 150), and GtP (n = 254) datasets. (C) Relative family distributions of G protein couplings across datasets. (D) Comparison of the G protein family coupling profiles of 70 GPCRs present in all of the Bouvier, Inoue, and GtP datasets. More detailed analysis of common G protein couplings for just the Bouvier and Inoue datasets is given in Figure 1—figure supplement 1. (A–D) Note: All analyses herein cover the 12 G proteins: G_s, G_i1, G_i2, G_oA, G_oB, G_z, G_q, G₁₁, G₁₄, G₁₅, G₁₂, and G₁₃. G_olf and G_i3 could not be analyzed, as they had not been tested by the Bouvier group. The Inoue data for the pairs G_i1-G_i2, G_oA-G_oB, and G_q-G₁₁ were generated with identical chimera inserting the Gα C-terminal hexamer into a G_q backbone (Inoue et al., 2019) meaning that identical datapoints were used to assess their coverage and agreement with other datasets (Table 1).

Figure 1—figure supplement 1.. Common G protein-coupled receptor (GPCR-G) protein couplings in the Bouvier and Inoue datasets.

Black bars represent couplings not in Guide to Pharmacology (GtP) and therefore considered novel (novel couplings are listed in Figure 3). Frequencies are percent couplings among all receptors common to Bouvier and Inoue (the 100% includes noncoupling receptors). G_i3 and G_olf (chimeras) are not included herein as they were only tested by the Inoue group. Inoue used the same chimeras to represent the pairs G_oA-G_oB, G_i1-G_i2, and G_q-G₁₁ (Table 1).

Figure 2.. Map of G protein-coupled receptor (GPCR-G) protein couplings supported by at least two studies.

(A) Normalization approach. E_max minimum-maximum normalization, EC₅₀ log transformation and use of log(E_max/EC₅0) as a combined measure with member-to-family aggregation by maximum G protein value. (B) Heatmap representation of log(E_max/EC₅₀) values for 166 GPCRs tested by the Bouvier and/or Inoue labs (for couplings with dual data sources a mean is used). For Guide to Pharmacology (GtP), a G protein subtype is considered supported if the family has a known coupling. G_i2 and G_i3 are represented by the same/identical chimeric G protein in Inoue’s dataset (Table 1). (C) Heatmap representation of primary and secondary transducers for 90 GPCRs which couplings are only covered by the GtP database. (B–C) Empty cells (white) indicate no coupling. All source values are available in tab ‘Fig_4’ in Source data 5. Note: Researchers wishing to use this coupling map, optionally after applying own reliability criteria or cut-offs, can do so for any set of couplings in GproteinDb (Pándy-Szekeres et al., 2022).

Figure 3.. Novel G protein couplings identified by both the Bouvier and Inoue groups.

Novel couplings identified by the Bouvier and Inoue groups but not in Guide to Pharmacology. G protein families are here considered shared if at least one specific subtype is found to couple in both dataset, which is not the case for the bottommost four receptors.

Figure 4.. Unifying the three datasets reveals a large diversity in G protein-coupled receptor (GPCR) coupling selectivity.

(A) GPCR-G protein selectivity on the G protein family (top) and subtype levels: Venn diagrams showing the numbers of shared and unique receptors. The 0 values (no receptors) are omitted for clarity. (B) Receptor coupling promiscuity: number of receptors that couple to 1–4 G protein families (top) and 1–5 subtypes. (C) G protein coupling promiscuity: number of receptors that couple to each G protein family (top) or subtype. (A–C) Panels B–C are based on the couplings from the Bouvier and Inoue groups that are also supported by a second dataset and panel A additionally includes GtP couplings. This analysis of the G_s family leaves out 11 receptors tested for coupling to G_s but not to G_olf, and G_olf couplings are only counted if there is a supported G_s coupling. All source data are available in tab ‘Fig_4’ in Source data 5.

Figure 5.. Correlated coupling of G proteins based on their receptor profiles.

(A) Overall correlated coupling of G proteins quantified by the Pearson standard correlation coefficient, which gives a measure of the strength of the linear relationship between two G proteins the log(E_max/EC₅₀) value of noncoupling datapoints was set to 0. Statistically significant pairwise correlations are indicated in cells by *p≤0.05; **p≤0.005, and ***p≤0.0005. (B) Overall coupling correlation of G proteins shown as a tree. (C) Correlated coupling explained by shared coupling/noncoupling quantified as Jaccard indices (% of couplings to the same G protein-coupled receptor [GPCRs]). (D) Correlated coupling explained by activation level quantified as the differences in average log(E_max/EC₅₀) when a G protein pair couples to the same receptor. The values are averages of the log(E_max/EC₅₀) averages of Bouvier and Inoue values, except where data is only available in one dataset, that is, Bouvier only: G_i1-G_i2, G_oA-G_oB, G_q-G₁₁, and Inoue only: G_olf-all G proteins. (A, C–D) All G protein couplings are for class A GPCRs and supported by two datasets (Bouvier or Inoue group or Guide to Pharmacology), and their source values are available in tab ‘Fig_5’ in Source data 5. (A–D) For G proteins (all but G_olf) and receptors tested by both the Bouvier and Inoue groups, an average of averages from the two groups is used. The G_oA and G_oB couplings from Bouvier were compared to the G_o values from Inoue which does not distinguish isoforms.

Figure 6.. G protein tissue and organ expression profiles.

(A) Expression heatmap of all 16 human G proteins across 50 tissues and 16 organs (here grouped in 8 categories) extracted from the Human Protein Atlas (Uhlén et al., 2015), which also includes data from the Genotype-Tissue Expression project (Lonsdale et al., 2013). The coloring denotes the normalized transcripts per million (nTPM) for each G protein and tissue capped at the median of all maximum G protein expression values (G_t3 at 80 nTPM in retina). (B) Number of tissues for which the given G protein has an expression ≥ the first quartile expression threshold (28.4 nTPM) across all 16 G proteins and 50 tissues.

Figure 6—figure supplement 1.. Correlated expression of G proteins.

(A) Expression heatmap like in Figure 6 but with a z-transformation across tissues, thereby visualizing their relative expression for each G protein. (B) Correlated G protein expression across tissues as a function of normalized transcripts per million (nTPM), which is a consensus normalized transcript expression value across the Genotype-Tissue Expression and Human Protein Atlas datasets (see Materials and methods). Pairwise correlation of nTPM expression values by Pearson standard correlation coefficient gives positive and negative correlations between G protein pairs. Expression clusters are distinct from the G protein families, which are denoted with color-coded labels.

Figure 7.. Unique and missing couplings in the Bouvier and Inoue datasets.

Unique and missing couplings (defined in Table 2) among all 70 receptors common to Bouvier, Inoue, and Guide to Pharmacology (GtP) (the 100% includes noncoupling receptors). Missing couplings are shown as negative values. For GtP, we consider coupling to a G protein subtype possible if a coupling has been observed for the respective family. Unique couplings are hidden by default in the online G protein couplings browser in GproteinDb, as they await the independent support by a second group (Pándy-Szekeres et al., 2022).