Now let's think about two enzymes that are members of the same gene family but have evolved different functions. It's easiest to think of these as two enzymes that are now specific for similar but distinct substrates. Imagine that you were asked to "prove" evolution by changing one of those enzymes into the other? Would you recognize the same fallacy? Would you realize that the most likely evolutionary scenario is that the two different enzyme specificities evolved from an ancestral enzyme that carried out both reactions? [see: The Evolution of Enzymes from Promiscuous Precursors]
Changing one of the modern enzymes into the other would require many changes because in most cases the common ancestor dates back hundreds of millions of years. Many of the changes that have become fixed in the two lineages were not directly involved in selecting one of the substrates over the other (i.e. increasing specificity). They were neutral mutations fixed by random genetic drift after the enzyme became specialized for one or other of the substrates. Many would have to be "reversed" in order to re-create the dual specificity because they would have been detrimental in the ancestral enzyme.
Think about these facts as you watch Douglas Axe explain why his research shows that evolution is impossible [Video: Doug Axe on Protein Evolution's Magic Number (It's Six)].