Scientist Solutions - Radioligand binding assay

Using cell-line transfected with a neuropeptide GPCR, I've carried out [125-I] radioligand binding assays to determine agonist affinity and estimate receptor binding sites on these cells in competitive binding assays using a fixed [radioligand] and increasing [unlabelled ligand]. I've calculated the Kd and Bmax using the equations without transforming data into Hill plots or anything. I then looked at the effect of a drug which competitively inhibited radioligand binding to the receptor. Using a fixed [drug] + fixed [radioligand] then adding increasing [unlabelled ligand] I found that the maximum binding was reduced although there was no significant change in Kd. Does this mean that the number of agonist binding sites is reduced in the presence of the drug? If so could this be due to receptor internalization (assays were all carried out on whole cells at 37degC but without inhibitors of internalization!). Or else, since I think the drug could be binding to a site distinct from the agonist binding site, could it be that the drug alters the receptor conformation which also changes the configuration of the agonist binding site and so prevents radioligand binding?
If that was the case, would I necessarily see a change in Kd?
And does the radiolabelled agonist specifically only bind receptors which are coupled to G-proteins (i.e. an active conformation) because I think my drug stabilizes an inactive conformation of receptor and this would then reduce the number of receptors which can potentially adopt the active conformation available for radiolabelled agonist binding. Does this make sense or am I way off target? Pharmacology is not my thing at all and I'm stuck in the middle of all this because of these binding assays I did!
Are Hill plots only used to determine co-operativity between binding sites? Does anyone know if the G-protein coupling to a receptor increases the agonist affinity? In which case, if I were to try and determine whether co-operativity is occuring between the agonist binding site and the drug binding site, how does the presence of G-protein affect this? How do I get from my data which is amount of radioligand bound vs. unlabelled [ligand] to a Hill plot to see if there is any co-operativity between the binding sites? I've seen graphs which plot[bound]/[free] vs [bound] but this looked funny when I tried it. Does [free] refer to total concentration of labelled and unlabelled ligand? The other thing is when I fit the original data using the variable slope sigmoidal curve model as Prism suggests to get a hill slope factor, I don't see any obvious difference between hill slope in presence and absence of drug.
With regards to deriving Bmax and Kd, is there any great advantage in using a scatchard plot? Would this be useful for me to show if drug binding alters Kd which I might not be picking up by using equations to calculate these values?

Is it valid for me to say that my drug binds a distinct site which alters the receptor conformation preventing agonist binding to its site just because I see a change in Bmax?

Any help would be greatly appreciated! I know there are a lot of brainy people who use this forum!