Mechanism and energetics of membrane transporters and channels

Living cells have evolved specialized transport proteins called membrane transporters and channels that catalyze exchange of materials across the cell membrane. Membrane trans- porters couple the active transport of their specific substrates against their electrochemical gradient. On the other hand, membrane channels facilitate passive diffusion of polar or charged molecules down their electrochemical gradient. We present here molecular dy- namics (MD) investigation of a membrane transporter, glycerol-3 phosphate transporter (GlpT) and two membrane channels, urea transporter (UT) and aquaporin (Aqp1). Each simulations presented here provided a dynamical and atomistic picture of the protein of interest in a collaborative effort with an experimental lab. Membrane transporters use various source of cellular energy, e.g., ATP binding and hydrolysis in primary active transporters, and pre-established electrochemical gradient of molecular species other than their substrate in the case of secondary active trans- porters. All membrane transporters use the widely-accepted "alternating-access mecha- nism", which ensures that the substrate is only accessible from one side of the membrane at a given time, and relies on complex protein conformational changes between outward- facing (OF) and inward-facing (IF) states, going through several intermediate states. The first system that we investigated is the glycerol-3-phosphate transporter (GlpT), an antiporter member of the MFS. GlpT transports glycerol-3-phosphate (G3P) into the cell in exchange for inorganic phosphate (Pi ). Major facilitator superfamily (MFS) is the largest superfamily of secondary active transporters and catalyze the transport of an enormous variety of small solute molecules across biological membranes. Individual MFS members, despite their architectural similarities, exhibit strict specificity toward the substrates that they transport. The structural basis of this specificity, however, is poorly understood. Our collaborators, Da-Neng Wang Lab (New York University, NY) performed mutagenesis studies and transport assays, while we performed equilibrium sim- ulations of wild-type GlpT and several of its mutant forms in membrane in the presence of all physiologically relevant substrates (Pi­ , Pi2 ­ , G3P ­ , and G3P 2 ­ ) to characterize the determinants of substrate selectivity and conformational response of the protein to substrate binding. The positive electrostatic potential of the lumen of GlpT recruits substrate and drives binding. Only a few amino acid residues that line the transporter lumen act as specificity determinants. The phosphate moiety of Pi and G3P bind to a common binding site and residues involved solely in recognition of the glycerol moiety of G3P confers it a higher binding affinity. Furthermore, the simulations characterized the process and mechanism of substrate binding, and the protein's initial conformational re- sponse. All substrate-bound systems resulted in partial closing of the cytoplasmic half of GlpT. Extended simulations of substrate-bound systems also captured a water-conducting "channel-like" state. These states were also observed in several other transporters, sug- gesting that alternating-access mechanism tolerates transient states that are partially open to both sides of the membrane. We, later, obtained a model of the outward-facing (OF) state of GlpT using nonequilibrium molecular dynamics and calculated free energies to investigate the energetics associated with the transport cycle of GlpT. The second system we report here is a membrane channel that facilitates passive diffu- sion of urea across the membrane, namely the urea transporter (UT). Urea is ubiquitously used as a nitrogen source by bacteria and a safe end product of protein catabolism. Due to its highly polar nature, urea relies on the UTs to permeate through the cell membrane. UTs are most frequently found in kidneys of mammals and allow rapid equilibration of urea between the urinary space and the hyperosmotic tissue fluid to prevent osmotic diuresis. Our collaborators, Ming Zhou lab (Columbia University, NY), crystallized struc- ture of a mammalian UT (UT-B). UT-B is a homotrimer and each monomer contains a urea conduction pore with a narrow selectivity filter. We performed an extensive set of molecular dynamics simulations combined with free energy calculations to elucidate the structural determinants of the selectivity in UT-B and the associated energetics. The orientation of the urea as its goes through the channel, as well as specific water-urea and protein-urea interactions determine the specificity of the channel. The free energy barrier at the selectivity filter appears to be approximately 5.0 kcal/mol. We, then, investigated the gas permeability of UT in collaboration with Walter F. Boron lab (Case Western Uni- versity, Cleveland). Our free energy calculations along with the physiological experiments indicate that UTs can function as gas channels and identified the monomeric pores as the main conduction pathway for both water and NH3 . Our work characterized UTs as the third family of gas channels along with aquaporins (water channels), and Rh-associated glycoprotein (RhAG) (ammonia channels). The other membrane channel system that we investigated is an aquaporin. Aqua- porins are ubiquitous integral membrane channels that maintain water homeostasis of the cell by facilitating selective diffusion of water across the membrane while preventing proton diffusion. Two conserved regions located along the pore are responsible for the selectivity: the dual asparagine, proline, alanine (NPA) aquaporin signature motif, and the aromatic/arginine selectivity filter (SF). Recently, our collaborators, Richard Neutze lab (University of Gotenburg, Sweden), have crystallized a yeast aquaporin at 0.88 ° res- A olution, the highest resolution achieved to date for a membrane protein. The structure reveals a great deal of novel information on the structure of hydrogen-bonded network of water and protein side chains. To determine the dynamics and energetics of water diffu- sion along the channel, we performed molecular dynamics simulations of this impressively high quality crystal structure. The results show disruption of the water chain in both NPA and SF regions in this aquaporin, due to characteristic hydrogen-bonding patterns that dictate specific orientations to water molecules. The motion of water molecules is highly correlated on either side of the NPA region. The correlation, however, is lower at the NPA region, attesting yet another possible mechanism for this region to contribute to a barrier against proton transport. Besides, the NPA region appears as a barrier region with low occupancy for water, a feature not seen in other aquaporins. The correlated motion of adjacent water molecules along with their binary co-occupancies in the SF show that water molecules move in pairs in this region. Specific hydrogen-bonding patterns in the SF region may also play a role in exclusion of hydronium (H3 O+ ) and/or hydroxide ions (OH ­ ). These simulations have helped elucidate the dynamical basis of many intricate features revealed by this new structure