Engineering a chimeric Agrin LG3 dimer to probe MuSK-dependent clustering of acetylcholine receptors at the neuromuscular junction

The neuromuscular junction (NMJ) relies on the agrin-LRP4-MuSK signalling complex to assemble and maintain a highly specialized postsynaptic apparatus. However, the precise mechanism by which, agrin binding to LRP4 leads to efficient MuSK activation is not fully understood. Recent structural modelling has revealed a 2:2 agrin-LRP4 assembly, suggesting that ligand valency may amplify MuSK signaling beyond simple receptor dimerization. Understanding the role of ligand valency in this pathway could help to bridge the gap between structural models of the agrin-LRP4 complex and its functional output at the neuromuscular junction. This thesis focuses on the generation of an agrin-derived construct to address this question by engineering a chimeric construct (LGX), designed to contain two copies of the NMJ-active LG3z8 domain covalently linked by a short Gly/Ser-rich peptide. The length and flexibility of this 13-residue linker were inspired by structural models of the agrin-LRP4 complex. The aim was to enforce LG3 bivalency while preserving ample orientational freedom between the two domains, and to avoid unwanted interactions. In this way, LGX was designed to be a constitutive, matrix-independent, LG3 dimer that could isolate the specific contribution of ligand valency within the agrin-LRP4-MuSK axis. The LGX cDNA was generated using a PCR-based insertion strategy coupled to homologous recombination to insert the short peptide and join the DNA fragments together within the backbone of a chosen expression vector. This recombinant construct was then expressed in a bacterial system using the autoinduction method. Subsequently, we obtained a monodisperse LGX preparation by optimizing a multistep purification pipeline, comprising immobilized metal affinity (IMAC) , anion-exchange chromatography (IEC), TEV-mediated tag removal with reverse IMAC, and a final size-exclusion chromatography (SEC). The final SEC confirmed the absence of aggregation, and indicated a behavior consistent with the intended flexible, non-aggregating, designed linker. Subsequent thermal denaturation analysis using NanoDSF showed that LGX is correctly folded, slightly more stable than the monomeric LG3z8, and is capable of interacting with Ca2+ ions like the wt monomer. Functionally, the capacity of LGX to induce the clustering of acetylcholine receptors was tested on differentiated C2C12 myotubes and visualized by α-bungarotoxin staining. Cluster patterns were found to be comparable to those elicited by equimolar monomeric LG3z8. Overall these data indicate that the “chimerization” procedure did not seem to affect the fold of the LG3z8 module, and that an isolated (matrix-free) LG3 dimer retained the capability to promote the formation of agrin-induced AChR clusters in vitro. However, enforcing LG3 bivalency alone was not found to increase cluster density or maturity in the assayed conditions. This suggests that additional features, such as extracellular-matrix tethering, precise LRP4-MuSK geometry, or higher-order membrane organization are likely to significantly contribute to the signal amplification necessary to shape the neuromuscular junction. Taken together, these results evidence that the LGX is likely as biologically active as a single LG3z8 module, but fails to elicit a comparatively stronger response from myotube cultures. Given that this construct was designed on the basis of the proposed 2:2 agrin:LRP4 complex, our data also suggest that the biologically active assembly might present a different conformation, one accommodating and stabilized by components missing in more controlled in-vitro setups. Thus LGX has acted as a molecular probe to interrogate published partial assemblies of the agrin-LRP4-MuSK signaling complex, and provides a starting point for the potential development of future avidity-engineered, and matrix-anchored variants.