Transition rate estimation from along-the-path, unbiased, molecular dynamics simulations of a major facilitator superfamily transporter
by Ghaan, A.; Losey, J.; Goolsby, C.; Tajkhorshid, E.; Moradi, M.
Secondary active membrane transporters undergo large-scale conformational changes to couple the uphill transport of their substrates to the downhill transport of co-transported species. Such conformational changes are challenging to study using conventional molecular dynamics (MD) simulations due to the timescale gap between local and global conformational dynamics. While many enhanced sampling techniques have been developed to address the timescale gap in MD simulations by using external biases, some methods try to achieve this goal without the use of an explicit bias. An effective approach involves the use of Markovian analysis of transition probabilities between discrete states obtained from short but numerous unbiased MD trajectories, commonly known as Markov state modeling (MSM). While MSM analysis allows for several orders of magnitude speedup in sampling as compared to regular MD, exploring high-barrier free energy landscapes such as those associated with membrane transport requires additional means to achieve reliable sampling. One such approach involves combining MSM with path-finding algorithms, where the short, unbiased MD trajectories start along the minimum free energy path. We have applied this methodology to study thermodynamics and kinetics of glycerol-3-phosphate transporter (GlpT), a representative member of the major facilitator superfamily (MFS) of transporters. Similar to other active membrane transporters, MFSs use an alternating-access mechanism to transport their substrates across the membrane by undergoing structural transitions between inward- and outward-facing (IF & OF) states. Our approach, which considers a tube of pathways rather than a single restrained pathway, enables us to estimate transition rates, reconstruct the free energy profile of the IF-OF transition, and identify multiple intermediate state between the IF and OF states. Our results show close agreement with those obtained from biased enhanced sampling techniques as well as with experimental data.