Movement, learning, and memory are only a few of the functions that depend on our nervous system. Rapid communication between neurons and modification of neuron-neuron connections involve the interconversion of electrical and chemical signals. In particular, at the junction between nerve cells—the synapse—electrical impulses promote the release of the chemical neurotransmitter into the synaptic cleft, where the electricity of the nervous system is transduced into chemistry. The neurotransmitter diffuses across the synapse to the membrane of the neighboring cell and activates ion channels that open transmembrane pores, initiating an electrical signal that will travel along the postsynaptic neuron. To quench the neurotransmitter stimulus, synapses harbor membrane-bound transporters that pump the transmitter out of the synapse and into neighboring cells, allowing the transmitter to be recycled for additional rounds of stimulation. Although neurotransmitter receptors and transporters are linchpins of the nervous system, information on their three-dimensional structures and molecular mechanisms is scarce.
Glutamate (one of the most common neurotransmitters in the mammalian nervous system) and ionotropic glutamate receptors, which are ligand-gated ion channels, are essential to the normal development and function of the nervous system. A major effort in my laboratory is focused on understanding how the atomic structure of glutamate receptors is related to their complex biological functions. To this end, we have been engaged in crystallographic, biophysical, and electrical measurements over the past few years. Our group has determined detailed molecular mechanisms for the actions of full and partial agonists, competitive antagonists, and allosteric modulators on the GluR2-subtype glutamate receptor. Our accomplishments include insights into the subunit stoichiometry and symmetry properties of glutamate receptors and studies of glutamate receptor desensitization, a process for which there had been little understanding at the molecular level.
Desensitization was first characterized in the acetylcholine receptor, a ligand-gated ion channel that is opened by acetylcholine. In the mid-1950s, del Castillo, Katz, and Thesleff found that prolonged application of acetylcholine, as well as other agonists, to the acetylcholine receptor resulted in the diminution of ion flux across the cell membrane. Katz and colleagues postulated that following activation, the receptor entered into an insensitive or desensitized state, one that is refractory to stimulation by acetylcholine or other agonists. Somehow, the agonist-receptor complex had undergone a conformational change that decoupled agonist binding from the opening of the transmembrane ion channel. Many other ligand-gated ion channels have since been shown to undergo agonist-induced desensitization, yet there has been no detailed molecular understanding of desensitization in ligand-gated ion channels.
Ionotropic glutamate receptors also undergo desensitization in the presence of agonist, entering the thermodynamically stable, desensitized state on the millisecond timescale. In the case of glutamate receptors, the rate of entry into the desensitized state can help to shape the electrical response in the postsynaptic cell, thus influencing the molecular "memory" at a particular synapse. We have found that glutamate receptors desensitize by altering specific protein-protein contacts. In the absence of glutamate, the interface between glutamate-binding subunits is stable. As soon as glutamate binds, each glutamate-binding subunit undergoes a conformational change that opens the ion channel. Almost as rapid, however, is a rearrangement at the dimer interface that allows the two glutamate-binding subunits to reorient relative to each other, thereby compensating for the glutamate-induced conformational change that occurs within each subunit. Indeed, the dimer interface is a bit like a clutch in a car: following glutamate binding and opening of the ion channel, the clutch is pressed in, thereby disengaging the conformational changes produced by glutamate from the ion channel gate. As soon as the protein-protein contacts disengage, the ion channel closes and the receptor enters the desensitized state.
Once glutamate activates receptors on the postsynaptic cell, it is cleared from the cleft by high-affinity, sodium-dependent glutamate transporters, integral membrane proteins located on neighboring glial cells and on neurons. These transporters are members of a family of integral membrane transport proteins that includes five eukaryotic glutamate transporters, two eukaryotic neutral amino acid transporters, and a large number of bacterial amino acid and dicarboxylic acid transporters. Eukaryotic members of this transporter family have an essential role in the nervous system and in the heart, kidney, and intestine. In prokaryotes, these transporters carry out the concentrative uptake of metabolites across the membrane by the cotransport of protons and/or sodium ions. Physiological studies have shown that glutamate uptake is coupled to the cotransport of three sodium ions and one proton, and to the countertransport of one potassium ion. Eukaryotic glutamate transporters also possess a thermodynamically uncoupled, glutamate-gated chloride conductance, illuminating their dual roles as secondary transporters and ligand-gated ion channels. Despite the wealth of functional data on glutamate transporters, however, there is little understanding of their three-dimensional architecture or molecular transport mechanism.
We recently solved the crystal structure of a eukaryotic glutamate transporter homolog from Pyrococcus horikoshii (GltPh). The transporter is a bowl-shaped trimer with a solvent-filled extracellular basin that is as large as 50 Å across and 30 Å deep. Perhaps the most striking feature of the structure is that the bottom of the hydrophilic basin lies halfway across the membrane bilayer. At the bottom of the basin are three independent binding sites, each cradled by two helical hairpins (HP1 and HP2) that reach from opposite sides of the membrane. The depth of the extracellular basin and the immediate proximity of the glutamate-binding sites to the bottom of the basin suggest that glutamate can diffuse, from bulk solution, to binding sites that are halfway across the membrane bilayer. The fact that our structure suggests that glutamate can bind at a nearly diffusion-limited rate agrees with the observation that the quenching of synaptic glutamate by transporters is accomplished by a rapid binding event, on the submillisecond timescale, that is followed by a slow transport process, occurring on the millisecond timescale.
The structure of GltPh allowed us to reach a number of important conclusions about this family of transporters. First, we resolved the transmembrane topology of glutamate transporters; showed that the carboxyl terminus—which is critical for substrate transport—is within the "amino-terminal cylinder"; and discovered additional, noncanonical elements of membrane protein structure, including the β-bridge in transmembrane 7 (TM7) and multiple, short helices in TM4. Second, we identified an electron-density feature that likely corresponds to bound glutamate, at the tips of HP1 and HP2 and adjacent to TM7 and TM8. Third, we suggested that HP1 and HP2 are the elements of protein structure that comprise the gates of the transport pathway, and only a modest movement of HP2 is required to render the glutamate-binding site accessible to the extracellular basin. On the basis of our crystal structure, we propose that glutamate transport is achieved by movements of the hairpins, allowing alternating access to either side of the membrane.