He's wrong on a technicality, actually.
In reality, you don't divide by 2 for a protein. See, proteins have an inherent order to their sequence due to the fact that on one end of the protein, you have a carboxylic acid, and on the other terminal you have an amine. Chemically, these are distinct - as a result, the two proteins ABB and BBA ARE NOT equivalent. They're just ~mostly~ equivalent - in reality, they aren't one and the same. If you ignore the fact that proteins can't be read the same backwards as forwards, though, the answer is (20^1000-20^500)/2.