Thanks, that's a good one one. I was wondering where exactly you disagreed with the previous definition, and that's reasonable.
Not trying to gotcha or anything, I just think it's fun, but niacin is an example of a substance that is a nutrient but can also produce a drug effect ("niacin flush") when taken in large (non-nutritive) quantities.
So for me personally, whether something is a drug or not depends less on the identity of the substance, and more on the specific interaction with the organism, and what effect is most salient to observers. I don't have a pithy definition handy, though.

Jism