Wolynes argued that the track of a protein’s folding was directed by the tendency of lowering its energy, and thus when a local minimum of its energy was reached, a relatively stable conformation was formed. However not all of the local minimums will lead the protein to a biologically useful conformation, for those otherwise are called energy traps. Wolynes energy landscape theory and natural selection have well explained the high efficiency of protein folding in vivo, instead of being stuck in energy traps. As to whether a protein can assume different conformations with the same bioactivity, there is no clear answer yet. In this paper, two conformational states of a pHLA-A*2402 are discovered after refolding, and by studying their interactions with TCR and CD8αα, two conformations of pHLA-A*2402 are confirmed of having escaped from natural selection.