I am researching the topic about how signalling molecules find their intra- and extracellular targets amongst multiple potential "distractions" along their way. What you describe as 'lock-and-key' or 'induced-fit' relationship is established only at the direct contact of a ligand with the receptor. While the ligand in vivo if not injected locally first needs to move fast through vast distances (relative to its size) until it reaches the receptor. It seems that it is just accepted for granted, as if the ligand "knew" where the receptor is.
Your link and explanation are very informative, it seems like this intramolecular hydrogen bond completes a "virtual" third heterocycle in the whole molecule. While senolytic quercetin, for example, also has 3 benzene rings.
There's a doctor in my contacts who weighs heavily on heterocycles due to the delocalized electron they contain.