Many recombinant proteins have optimized codons that give great expression in Cos or HEK cells, and others including baculovirus and other non-mammalian cellular systems such as yeast and insect cells. Certain mammalian proteins do not express well or at all as native sequences in non-native cellular systems, but can be optimized with regard to codon usage. I have substantial experience at this, as this is commonly used as a work around strategy for gene patents in the past.
Vinmantoo mentioned in his original post though that he had known some examples of altered protein activity with the apparent same protein sequence. I am eager to read up on such examples.