It's easy to evaluate if the energy of one structural conformation is lower than another. (You don't need to determine at each step if the submitted conformation approximates the real protein at equilibrium, only that it's a lower-energy conformation than has been seen before.) But I'm not sure that is the only problem.

You want the proof of work to remain difficult for a PoW scheme. We have pretty good assurances that mindless cryptographic hashing is difficult; ASICs accelerate it considerably, but don't lower the Big O complexity of the problem. On the other hand, people keep searching for algorithmically faster ways to simulate protein folding, not just ways to hardware-accelerate existing algorithms. If one group of folders were to discover superior algorithms for searching the potential energy surface, that would be beneficial to science but potentially disastrous for a currency based on protein folding problems. For that matter, the overall incentives might make science worse off if ProteinCoin encouraged superior algorithms to be kept secret, for PoW advantage, instead of publishing advances in the open scientific literature.