Potassium two-pore domain (K2P) family ion channels are increasingly appreciated as important electrogenic sensors that are used by diverse cell types including neurons, cardiomyocytes, and epithelial cells. The activity of these channels can be activated up to one-hundred fold by diverse physical and chemical stimuli including pH, intracellular calcium, temperature, phospholipids and fatty acids, volatile anesthetics, and membrane tension. In addition to this “stimulus gated” activity, channels in this family also have a basal “leak gated” activity that drives the cellular resting membrane close to the potassium ion reversal potential in the absence of their respective stimuli. While the structural basis of closed states and certain “stimulus gated” open states have been previously determined, the structural basis of a “leak gated” open state in K2P channels is unknown. Here, we show the physical basis for basal “leak gated” activity of the mechanosensitive K2P family ion channel TRAAK, and propose a gating model that includes both “leak gated” and “stimulus gated” openings. Next, we identify the structural basis of intracellular and extracellular alkaline pH “stimulus gated” activation in the acid-sensitive K2P family ion channel TASK2. We observe structural similarities between the “leak gated” structure of TRAAK and other K2P channel structures, including TASK2. This observation supports a general mechanism for “leak gated” activity in K2P family ion channels.

Proteins are biological machines composed of a single or multiple threads of amino acids which fold into three-dimensional shapes. These shapes and surfaces directly participate in many of the molecular interactions essential to life. Some proteins exist in lipid membranes and thus have been termed membrane proteins. Membranes serve as sites of import, export, and signaling and allow for compartmentalization; all critical functions for living organisms. After decades of research throughout the 20th century, many families of membrane proteins were revealed to form transport proteins that facilitate movement of substrates across membranes. The general group of transport proteins can be further divided into two groups: active and passive transporters. Passive transport proteins have been called channels, given their observed activity of opening to allow substrate diffusion down an electrochemical gradient. Channels are diverse both in their architecture and substrates, which include ions, small molecules, and other proteins. Despite the broad importance of channels in a vast array of biology many aspects of how these machines operate at a molecular level are yet to be revealed. Determining the structures of channels, which in many cases was not technically feasible until recently, has provided key insights into the form and function of these intricate machines. Presented here are studies of two different channels: the Translocase of the Outer Mitochondrial Membrane (TOM complex) and the light-gated ion channel, ChRmine.

The first study in this thesis examines the long-awaited architecture and transport mechanisms of the major protein gate of the outer mitochondrial membrane. The high-resolution structure of the TOM complex revealed its overall organization-- including all its core subunits, multiple assemblies of the complex, and a glimpse at the pore which provides insights into mechanisms of protein transport. The subsequent study examines the first ever structures of a light-gated ion channel in a native-like membrane environment, using lipid nanodiscs and cryogenic electron-microscopy(cryo-EM). Determining cryo-EM structures of ChRmine was a technical achievement and revealed its surprising trimeric assembly, ion conduction path, and retinal binding pocket. Through analysis of the structure, we were able to tune the kinetics of ChRmine using a rational engineering approach. Lastly, the latest work on ChRmine paves the path for the determination of structures previously too dynamic to capture. These elusive structures are attractive targets because they could lay the foundation for building the next generation of tools for studying the brain.

The coronavirus genome encodes four structural proteins that are incorporated into virions, spike (S), envelope (E), nucleocapsid (N), and membrane (M). M is the most abundant structural protein and is known to drive viral particle assembly and budding through direct protein-protein interactions. Throughout the COVID-19 pandemic caused by the virus SARS-CoV-2, M has displayed high sequence conservation, making it a compelling target for vaccines and therapeutics. Earlier studies of viral particles from related betacoronaviruses had suggested that M may be able to adopt two distinct conformations – a compact and elongated form – yet prior to 2019, the structure and molecular basis for its functions in infectious particle formation were unknown. The work presented in this dissertation focuses on the structural characterization of the M protein from SARS-CoV-2 to gain insight into how M functions in the viral life cycle and subsequently leverage our findings to develop new avenues to combat disease.

First, we develop routines to express and purify full-length M protein and determine its structure in nanodiscs to describe its architecture in a lipid environment using cryo-electron microscopy (cryo-EM). We show that M is a 50 kDa homodimer that is structurally homologous to an accessory SARS-CoV-2 protein, ORF3a, suggesting that the two share a common ancestor. Subtle structural differences between the two proteins lend insight into how they have evolved to fill drastically different roles in the coronavirus life cycle. One striking feature of M is the overwhelmingly electropositive cytosolic surface that is likely important for protein-protein interactions in viral assembly, as well allowing for close contact with the dense core of viral RNA. We performed molecular dynamics simulations and showed that this conformation is stable in a simple lipid bilayer. Together, these results produced the first atomic model of a coronavirus M protein and provide insight into roles for M in viral assembly and structure.

The next chapter of this thesis describes the structural basis and biophysical characterization of a neutralizing antibody that targets the N-terminus of the SARS-CoV-2 M protein. We used biolayer-interferometry (BLI) to show that the SARS37 neutralizing antibody specifically recognizes the alternate long conformation of M, rather than the short form we had previously characterized in nanodiscs. We used cryo-EM to determine the structure of the SARS37 Fab in complex with long form M and show that the epitope involves residues from the N-termini of both subunits, as well as the outer surface of the transmembrane domain through the insertion of a long CDR3 loop into the lipid bilayer. In agreement with our BLI data, we found that the complex formed between the SARS37 Fab and M in the short conformation was structurally heterogeneous and unstable. Importantly, we find that the residues of M that the Fab recognizes are highly conserved, remaining unchanged as the virus has evolved. Given the strong and stable preference for the long conformation, we infer that this form of M is present in circulating viral particles presented to the immune system and can be targeted to prevent further infection.

The last chapter of this dissertation investigates the role specific M-lipid interactions may play in the process of viral assembly. We show M directly binds a Golgi resident anionic lipid, ceramide-1-phosphate (C1P) and investigate the effect that C1P binding has on M conformational dynamics. We find an overall stabilizing effect of C1P binding to M in the short conformation and describe a specific C1P binding site that involves M residues spanning the transmembrane, hinge and C-terminal domains. Comparing our C1P bound structure with the previously determined apo short conformation M structures, we observe a previously uncharacterized potassium ion binding site and a series of side chain rotamer changes distributed across the C-terminal domain and dimer interface. We design point mutants at these sites and use AlphaFold to show that the predicted models exist in a continuum between the established short and long conformations. The residues involved are highly conserved amongst sarbecoviruses, suggesting a conserved potential principle underlying M protein conformational dynamics.

Two-pore domain potassium channels (K2Ps) are widely expressed ion channels which generate leak-type currents regulated by diverse chemical and physical stimuli including temperature, membrane tension, signaling lipids, and pH, to control the resting membrane potential (RMP) of cells. While these K2Ps are critical to the maintenance of the RMP, the structural mechanisms by which they respond to physiological stimuli remain largely unknown. Additionally, studies of K2Ps in vivo have been hampered by the lack of molecular tools which can be used to probe K2P function and structure.

Here, we describe the structural basis for gating of the K2P TWIK1 and describe a series of molecular tools generated against the K2P TREK1. Utilizing cryo-electron microscopy (cryo-EM), we determine high resolution structure of TWIK1 in lipid nanodiscs at a high and low pH, and in high extracellular K+, and describe the mechanistic basis for pH-gating in this channel. We then describe the generation of a nanobody which specifically binds another K2P, TREK1, with nanomolar affinity. Finally, we describe the computational prediction of a series of novel small-molecule TREK1 agonists. The data and molecular tools presented here further our understanding of the structural and functional properties of the K2P family of ion channels.

Ion channels are a large and diverse family of proteins broadly implicated in human physiology and that of other organisms. The study of the molecular mechanism of ion channels benefits from both “top-down” physiological and “bottom-up” biophysical studies. Here, we use a reductionist biophysical approach, with a focus on the use of cryo-electron microscopy (cryo-EM) to study the structure and function of three putative ion channels. The first, ORF3a is a putative ion channel encoded in the SARS-CoV-2 genome. Here, we determined the first ORF3a structure, and characterized ORF3a activity through electrophysiological recordings and fluorescent imaging of reconstituted ORF3a proteoliposomes. The second, Tweety-homolog family (TTYH) proteins, previously characterized as putative volume-regulated anion channels. Through parallel functional and structural studies, our work determined that TTYHs are not pore-forming subunits of anion channels. Additionally, we discovered a previously unknown “trans-dimeric” conformation of TTYH2. The third putative ion channel is TMEM87A, previously implicated in mechanosensitive conduction of cations. Here we determine the first structure of an TMEM87 protein, and through structural and bioinformatic analysis, identify related proteins forming the basis for a new protein family. Collectively, this work has demonstrated the value of cryo-EM for structural studies of small (<100 kDa) membrane proteins, while laying the foundation for deeper mechanistic understanding of each of these different proteins.

Cation-chloride-cotransporters (CCCs) play critical roles in cellular volume regulation, neural development and function, audition, regulation of blood pressure, and renal function. Disruption of CCCs has been implicated in pathophysiology including epilepsy, hearing loss, and genetic disorders. In this study, we present the structure of a CCC, the Mus musculus K+-Cl- cotransporter (KCC) KCC4, determined by cryo-EM. The structure, captured in an inward-open conformation, reveals the architecture of KCCs and offers mechanistic insights into the function and regulation of this important transporter family. Our findings provide a structural explanation for the varied substrate specificity and ion transport ratio among CCCs, identify binding sites for substrate K+ and Cl- ions, and demonstrate the importance of key coordinating residues for transporter activity. These insights have significant implications for developing novel drugs and treatments targeting CCCs.