Determination of how proteins transform into their biologically relevant structure is an extraordinarily complex research area. They fold into their native state as a population of conformations on time scales that experimental methods struggle to resolve. Molecular dynamic simulations can avoid some of these issues, pro- viding theoretical answers to protein dynamics. To that effect, this work presents two enhanced sampling methods that aim to lower the computational cost of the Replica Exchange protocol with fewer replicas while avoiding the exchange bot- tleneck in an approach called Replica-Exchange-with-Tunneling. The method is used to simulate the folding switch of the metamorphic chemokine Lymphotactin. Go-model potentials bias replicas to fold as either the Ltn10 form or Ltn40 form. This study proposes that the conversion between the two forms is assisted by bi- furcated hydrogen bonding. Resolution-Exchange-with-Tunneling (ResET) builds further on these advancements by reducing the number of replicas to the minimum of 2, while still avoiding exchange rejection and enhancing sampling. The folding performance of the method is shown to outperform other simulation approaches for folding the trpcage protein. It was also used to elicit the contribution that amino acid mutations have on the behavior of Alzheimer associated amyloid beta proteins.