I don't think it's that far-fetched. If you accept the proposition that it's possible for a microbe to secrete a protein or something into the bloodstream that

I don't think it's that far-fetched. If you accept the proposition that it's possible for a microbe to secrete a protein or something into the bloodstream that causes you to prefer a certain type of food (and keep in mind that we already have evidence of biologically-driven food cravings, e.g. in pregnant women), then there's a good argument that evolutionary pressure would eventually cause gut flora that could help deliver its own preferred food source to predominate in human digestive systems.