From elementary logic it is obvious that protein ligand receptors do not bind to any part of the ligand molecule, but only to a certain primary structure (amino acid sequence), moreover, organized into a very specific secondary structure. OTHERWISE, ANY PROTEIN COULD WORK AS ANY HORMONE, so that there would be absolute chaos in each cell and in the body as a whole. There are many works in the world literature on the binding of proteins and peptides, you just have to look for it.
A binding site on the receptor, and on the ligand if it is big, e.g. a protein, is rather a specific piece of the surface, not necessarily a classical deep pocket. So to say, when macromolecules bind each other, they stick together at complementary surfaces. There are numerous examples that a receptor, having a classical binding pocket for a small ligand, still able to bind allosteric molecules, including proteins, at a different part of the receptor surface.