3.3 MQ1 uses its two large loops to bind V2R
To characterize the importance of the 5 positions described above for hV2R interaction, we synthesized the variants K10E, F17A, N28A, T34A, and R44E. We couldn’t obtain the folded N28A variant. The 5 remaining variants induced a loss of 20 to 1340 fold in binding affinity (Fig. 4 and supplementary Table 2). Interestingly, they all belong to the MQ1 loop 1 and 2 (Fig. 3A). We expanded our analysis to other neighboring positions excluding structural (Cys, Ala, Gly, Pro) and buried positions (Tyr22, Tyr23, Phe33, Tyr35). We found that residues Val9 and Phe18 but not Ser19, Asn41 or Ser46 are in interaction with the V2R. Surprisingly, while the variants K39E and K39W never modified MQ1 affinity on hV2R, the K39A improved it by almost 9 times (Fig. 4, Supplementary table 2). We confirmed here previously demonstrated data showing that the N and C-terminal extremities (Fig. 4) are not involved in the binding process of MQ1 (Ciolek et al., 2017). Finally, the structured part of the toxin defined by residues implied in α-helix or β-sheets, never participate to V2R binding as the variants N6E, F21A, S24G, Q25A, K26A, K26E, K29A and H31F displayed the same affinities for hV2R as MQ1.