Expression of mammalian GPCRs in C. elegans generates novel behavioural responses to human ligands

Abstract

Background: G-protein-coupled receptors (GPCRs) play a crucial role in many biological processes and represent a major class of drug targets. However, purification of GPCRs for biochemical study is difficult and current methods of studying receptor-ligand interactions involve in vitro systems. Caenorhabditis elegans is a soil-dwelling, bacteria-feeding nematode that uses GPCRs expressed in chemosensory neurons to detect bacteria and environmental compounds, making this an ideal system for studying in vivo GPCR-ligand interactions. We sought to test this by functionally expressing two medically important mammalian GPCRs, somatostatin receptor 2 (Sstr2) and chemokine receptor 5 (CCR5) in the gustatory neurons of C. elegans.

Results: Expression of Sstr2 and CCR5 in gustatory neurons allow C. elegans to specifically detect and respond to somatostatin and MIP-1alpha respectively in a robust avoidance assay. We demonstrate that mammalian heterologous GPCRs can signal via different endogenous Galpha subunits in C. elegans, depending on which cells it is expressed in. Furthermore, pre-exposure of GPCR transgenic animals to its ligand leads to receptor desensitisation and behavioural adaptation to subsequent ligand exposure, providing further evidence of integration of the mammalian GPCRs into the C. elegans sensory signalling machinery. In structure-function studies using a panel of somatostatin-14 analogues, we identified key residues involved in the interaction of somatostatin-14 with Sstr2.

Conclusion: Our results illustrate a remarkable evolutionary plasticity in interactions between mammalian GPCRs and C. elegans signalling machinery, spanning 800 million years of evolution. This in vivo system, which imparts novel avoidance behaviour on C. elegans, thus provides a simple means of studying and screening interaction of GPCRs with extracellular agonists, antagonists and intracellular binding partners.

Figure 1

Expression of mouse Sstr2 and human CCR5 in nociceptive neurons in C. elegans generates agonist specific avoidance behaviour. (A) Flow sorting of transgenic nematodes expressing heterologous GPCRs and gut specific GFP marker (elt-2::GFP). Graph shows a plot of forward scatter (size) against GFP fluorescence intensity. GFP-positive animals are sorted into larval stages L1, L2 and L3 according to size with 90% purity and 95% viability (sort gates R1, R2 and R3 respectively). (B) Wild type and Sstr2 transgenic animals were tested with 1–50 μM of SST-28. As controls, the responses to 1% SDS, M9 diluent, the unrelated neuropeptides, neurotensin and MIP-1α were determined. Sstr2 transgenic animals did not avoid 25 μM cyclosomatostatin. Both strains display a normal avoidance response to 1% SDS. Asterisks indicate a statistically significant difference between wild type and Sstr2 transgenic animals. (C) Avoidance responses of wild type and CCR5 transgenic animals to various concentrations of MIP-1α, M9 diluent, 1% SDS and 25 μM SST-28. Error bars denote the standard error of the mean. Asterisks indicate a statistically significant difference between wild type and CCR5 transgenic animals. (D) Animals were either directly tested for their response to M9, SST-28 or 1% SDS or pre-incubated with 25 μM cyclosomatostatin (denoted cyclosst in figure) for 3 minutes before the assays. Pre-treatment with 25 μM cyclosomatostatin abolished the avoidance response of Sstr2 transgenic animals to 25 μM SST-28. Avoidance responses can be recovered by an additional 5 minutes wash to remove cyclosomatostatin. Asterisks indicate a statistically significant difference between cyclosomatostatin treated and untreated transgenic animals. In all panels, each data point represents an average of at least five independent assays for wild-type and each respective strain. Error bars denote the standard error of the mean. (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.005). Avoidance index = number of worms behind barrier/total number of worms

Figure 2

Heterologously expressed GPCRs are integrated into an endogenous signalling pathway in C. elegans. Mutation of the TRPV channel subunit osm-9 and the Gα subunit gpa-3 fully abolished avoidance of (A) 25 μM SST-28 in Sstr2 transgenic animals and (B) 25 μM MIP-1α in CCR5 transgenic animals. Mutation of Gα subunit odr-3 did not affect these responses. (C) In contrast, mutation of either gpa-3 or odr-3 did not have a significant effect on avoidance response in sra6::CCR5 transgenic animals. Each data point represents an average of at least 3 independent assays for wild type and each respective strain. Error bars denote the standard error of the mean. Asterisks indicate a statistically significant difference between osm-9 or gpa-3 mutant animals and wild type animals carrying the Sstr2 or CCR5 transgenes (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.005).

Figure 3

Sstr2 and CCR5 transgenic animals mutants are desensitised by pre-exposure to agonists. (A) Pre-exposure of Sstr2 transgenic animals to 1 μM and 10 μM SST-28 prior to the assay strongly reduced or even abolished avoidance behaviour. (B)Pre-exposure of CCR5 transgenic animals to 1 μM and 10 μM MIP-1α fully abolished avoidance of 1 μM, 10 μM and 25 μM MIP-1α. In both cases, pre-exposure to the agonist did not affect the responses towards other repellents. Animals were washed in buffer as control. For each panel, each data point represents an average of at least five independent assays for wild type and each respective strain. Error bars denote the standard error of the mean. Asterisks indicate a statistically significant difference between transgenic animals pre-exposed to agonist or only to chemotaxis (CTX) buffer (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.005).

Figure 4

Identification of SST-14 residues involved in Sstr2 activation. Sstr-2 transgenic animals were tested for response to SST-14 and five variants with single alanine substitutions at residues 6–10. No avoidance behaviour was observed with 25 μM of Trp8Δ Ala and Lys9Δ Ala SST-14 analogues while residual avoidance behaviour was found with Phe7Δ Ala and Thr10Δ Ala. Sstr2 transgenic animals were repelled by the Phe6Δ Ala analogue, indicating that Phe6 is not essential for receptor:ligand interaction. Each data point represents an average of at least five independent assays for wild type and each respective strain. Error bars denote the standard error of the mean. Asterisks indicate a statistically significant difference between the avoidance index of the transgenic animals for the Trp8Δ Ala and Lys9Δ Ala analogues and native SST-14 agonist (*p ≤ 0.05, **p ≤ 0.01).