Would it be possible for you to upgrade to R2021b? There's a new feature in that release that would make it much easier to do what you want. It's described in the SimBiology release notes as "Specify different parameter values for different groups during fitting". Basically, you can specify a different variant for each time course in your data (similar to how you specify a different dose for each time course in previous releases). You could use this variant to turn certain reactions on or off for specific time courses. You would also need to "merge" the two different models of interest into a single model and add a parameter to the relevant reactions or rules that enable you to turn them on or off for specific data sets.
For example, let's say you want to toggle between using M3_Ctr and Drug_Ctr in your rule for k1. You could update your rule to something like k1 = Emax*EC50/( X*M3_Ctr + (1-X)*Drug_Ctr +EC50). Then, for each time course where you want to use M3_Ctr you use a variant that sets X to 1, and for each time course where you want to use Drug_Ctr you use a variant that sets X to 0. You could make similar change to reaction rates if you need to turn certain reactions on or off for different time courses.