Graph-based algorithms to dissect long-distance water-mediated H-bond networks for conformational couplings in GPCRs.

Changes in structure and dynamics elicited by agonist ligand binding at the extracellular side of G protein coupled receptors (GPCRs) must be relayed to the cytoplasmic G protein binding side of the receptors. To decipher the role of water-mediated hydrogen-bond networks in this relay mechanism, we have developed graph-based algorithms and analysis methodologies applicable to datasets of static structures of distinct GPCRs. For a reference dataset of static structures of bovine rhodopsin solved at the same resolution, we show that graph analyses capture the internal protein-water hydrogen-bond network. The extended analyses of static structures of rhodopsins and opioid receptors suggest a relay mechanism whereby inactive receptors have in place much of the internal core hydrogen-bond network required for long-distance relay of structural change, with extensive local H-bond clusters observed in structures solved at high resolution and with internal water molecules.

A Proteorhodopsin-Related Photosensor Expands the Repertoire of Structural Motifs Employed by Sensory Rhodopsins.

Microbial rhodopsins are light-activated retinal-binding membrane proteins that perform a variety of ion transport and photosensory functions. They display several cases of convergent evolution where the same function is present in unrelated or very distant protein groups. Here we report another possible case of such convergent evolution, describing the biophysical properties of a new group of sensory rhodopsins. The first representative of this group was identified in 2004 but none of the members had been expressed and characterized. The well-studied haloarchaeal sensory rhodopsins interacting with methyl-accepting Htr transducers are close relatives of the halobacterial proton pump bacteriorhodopsin. In contrast, the sensory rhodopsins we describe here are relatives of proteobacterial proton pumps, proteorhodopsins, but appear to interact with Htr-like transducers likewise, even though they do not conserve the residues important for the interaction of haloarchaeal sensory rhodopsins with their transducers. The new sensory rhodopsins display many unusual amino acid residues, including those around the retinal chromophore; most strikingly, a tyrosine in place of a carboxyl counterion of the retinal Schiff base on helix C. To characterize their unique sequence motifs, we augment the spectroscopy and biochemistry data by structural modeling of the wild-type and three mutants. Taken together, the experimental data, bioinformatics sequence analyses, and structural modeling suggest that the tyrosine/aspartate complex counterion contributes to a complex water-mediated hydrogen-bonding network that couples the protonated retinal Schiff base to an extracellular carboxylic dyad.

Protons taken hostage: Dynamic H-bond networks of the pH-sensing GPR68.

Proton-sensing G Protein Coupled Receptors (GPCRs) sense changes in the extracellular pH to effect cell signaling for cellular homeostasis. They tend to be overexpressed in solid tumors associated with acidic extracellular pH, and are of direct interest as drug targets. How proton-sensing GPCRs sense extracellular acidification and activate upon protonation change is important to understand, because it may guide the design of therapeutics. Lack of publicly available experimental structures make it challenging to discriminate between conflicting mechanisms proposed for proton-binding, as main roles have been assigned to either an extracellular histidine cluster or to an internal carboxylic triad. Here we present a protocol to derive and evaluate structural models of the proton-sensing GPR68. This approach integrates state-of-the-art homology modeling with microsecond-timescale atomistic simulations, and with a detailed assessment of the compatibility of the structural models with known structural features of class A GPCRs. To decipher structural elements of potential interest for protonation-coupled conformational changes of GPR68, we used the best-compatible model as a starting point for independent atomistic simulations of GPR68 with different protonation states, and graph computations to characterize the response of GPR68 to changes in protonation. We found that GPR68 hosts an extended hydrogen-bond network that inter-connects the extracellular histidine cluster to the internal carboxylic triad, and which can even reach groups at the cytoplasmic G-protein binding site. Taken together, results suggest that GPR68 relies on dynamic, hydrogen-bond networks to inter-connect extracellular and internal proton-binding sites, and to elicit conformational changes at the cytoplasmic G-protein binding site.

Graph-Based Analyses of Dynamic Water-Mediated Hydrogen-Bond Networks in Phosphatidylserine: Cholesterol Membranes.

Phosphatidylserine lipids are anionic molecules present in eukaryotic plasma membranes, where they have essential physiological roles. The altered distribution of phosphatidylserine in cells such as apoptotic cancer cells, which, unlike healthy cells, expose phosphatidylserine, is of direct interest for the development of biomarkers. We present here applications of a recently implemented Depth-First-Search graph algorithm to dissect the dynamics of transient water-mediated lipid clusters at the interface of a model bilayer composed of 1-palmytoyl-2-oleoyl-sn-glycero-2-phosphatidylserine (POPS) and cholesterol. Relative to a reference POPS bilayer without cholesterol, in the POPS:cholesterol bilayer there is a somewhat less frequent sampling of relatively complex and extended water-mediated hydrogen-bond networks of POPS headgroups. The analysis protocol used here is more generally applicable to other lipid:cholesterol bilayers.

Fentanyl and the Fluorinated Fentanyl Derivative NFEPP Elicit Distinct Hydrogen-Bond Dynamics of the Opioid Receptor.

The development of safe therapeutics to manage pain is of central interest for biomedical applications. The fluorinated fentanyl derivative N-(3-fluoro-1-phenethylpiperidin-4-yl)-N-phenylpropionamide (NFEPP) is potentially a safer alternative to fentanyl because unlike fentanyl─which binds to the µ-opioid receptor (MOR) at both physiological and acidic pH─NFEPP might bind to the MOR only at acidic pH typical of inflamed tissue. Knowledge of the protonation-coupled dynamics of the receptor-drug interactions is thus required to understand the molecular mechanism by which receptor activation initiates cell signaling to silence pain. To this end, here we have carried out extensive atomistic simulations of the MOR in different protonation states, in the absence of opioid drugs, and in the presence of fentanyl vs NFEPP. We used graph-based analyses to characterize internal hydrogen-bond networks that could contribute to the activation of the MOR. We find that fentanyl and NFEPP prefer distinct binding poses and that, in their binding poses, fentanyl and NFEPP partake in distinct internal hydrogen-bond networks, leading to the cytoplasmic G-protein-binding region. Moreover, the protonation state of functionally important aspartic and histidine side chains impacts hydrogen-bond networks that extend throughout the receptor, such that the ligand-bound MOR presents at its cytoplasmic G-protein-binding side, a hydrogen-bonding environment where dynamics depend on whether fentanyl or NFEPP is bound, and on the protonation state of specific MOR groups. The exquisite sensitivity of the internal protein-water hydrogen-bond network to the protonation state and to details of the drug binding could enable the MOR to elicit distinct pH- and opioid-dependent responses at its cytoplasmic G-protein-binding site.

Structures of channelrhodopsin paralogs in peptidiscs explain their contrasting K+ and Na+ selectivities.

Kalium channelrhodopsin 1 from Hyphochytrium catenoides (HcKCR1) is a light-gated channel used for optogenetic silencing of mammalian neurons. It selects K+ over Na+ in the absence of the canonical tetrameric K+ selectivity filter found universally in voltage- and ligand-gated channels. The genome of H. catenoides also encodes a highly homologous cation channelrhodopsin (HcCCR), a Na+ channel with >100-fold larger Na+ to K+ permeability ratio. Here, we use cryo-electron microscopy to determine atomic structures of these two channels embedded in peptidiscs to elucidate structural foundations of their dramatically different cation selectivity. Together with structure-guided mutagenesis, we show that K+ versus Na+ selectivity is determined at two distinct sites on the putative ion conduction pathway: in a patch of critical residues in the intracellular segment (Leu69/Phe69, Ile73/Ser73 and Asp116) and within a cluster of aromatic residues in the extracellular segment (primarily, Trp102 and Tyr222). The two filters are on the opposite sides of the photoactive site involved in channel gating.

Structural Foundations of Potassium Selectivity in Channelrhodopsins.

Potassium-selective channelrhodopsins (KCRs) are light-gated K+ channels recently found in the stramenopile protist Hyphochytrium catenoides. When expressed in neurons, KCRs enable high-precision optical inhibition of spiking (optogenetic silencing). KCRs are capable of discriminating K+ from Na+ without the conventional K+ selectivity filter found in classical K+ channels. The genome of H. catenoides also encodes a third paralog that is more permeable for Na+ than for K+. To identify structural motifs responsible for the unusual K+ selectivity of KCRs, we systematically analyzed a series of chimeras and mutants of this protein. We found that mutations of three critical residues in the paralog convert its Na+-selective channel into a K+-selective one. Our characterization of homologous proteins from other protists (Colponema vietnamica, Cafeteria burkhardae, and Chromera velia) and metagenomic samples confirmed the importance of these residues for K+ selectivity. We also show that Trp102 and Asp116, conserved in all three H. catenoides paralogs, are necessary, although not sufficient, for K+ selectivity. Our results provide the foundation for further engineering of KCRs for optogenetic needs. IMPORTANCE Recently discovered microbial light-gated ion channels (channelrhodopsins) with a higher permeability for K+ than for Na+ (potassium-selective channelrhodopsins [kalium channelrhodopsins, or KCRs]) demonstrate an alternative K+ selectivity mechanism, unrelated to well-characterized "selectivity filters" of voltage- and ligand-gated K+ channels. KCRs can be used for optogenetic inhibition of neuronal firing and potentially for the development of gene therapies to treat neurological and cardiovascular disorders. In this study, we identified structural motifs that determine the K+ selectivity of KCRs that provide the foundation for their further improvement as optogenetic tools.

Interplay between local protein interactions and water bridging of a proton antenna carboxylate cluster.

Proteins that bind protons at cell membrane interfaces often expose to the bulk clusters of carboxylate and histidine sidechains that capture protons transiently and, in proton transporters, deliver protons to an internal site. The protonation-coupled dynamics of bulk-exposed carboxylate clusters, also known as proton antennas, is poorly described. An essential open question is how water-mediated bridges between sidechains of the cluster respond to protonation change and facilitate transient proton storage. To address this question, here I studied the protonation-coupled dynamics at the proton-binding antenna of PsbO, a small extrinsinc subunit of the photosystem II complex, with atomistic molecular dynamics simulations and systematic graph-based analyses of dynamic protein and protein-water hydrogen-bond networks. The protonation of specific carboxylate groups is found to impact the dynamics of their local protein-water hydrogen-bond clusters. Regardless of the protonation state considered for PsbO, carboxylate pairs that can sample direct hydrogen bonding, or bridge via short hydrogen-bonded water chains, anchor to nearby basic or polar protein sidechains. As a result, carboxylic sidechains of the hypothesized antenna cluster are part of dynamic hydrogen bond networks that may rearrange rapidly when the protonation changes.

Graphs of Hydrogen-Bond Networks to Dissect Protein Conformational Dynamics.

Dynamic hydrogen bonds and hydrogen-bond networks are ubiquitous in proteins and protein complexes. Functional roles that have been assigned to hydrogen-bond networks include structural plasticity for protein function, allosteric conformational coupling, long-distance proton transfers, and transient storage of protons. Advances in structural biology provide invaluable insights into architectures of large proteins and protein complexes of direct interest to human physiology and disease, including G Protein Coupled Receptors (GPCRs) and the SARS-Covid-19 spike protein S, and give rise to the challenge of how to identify those interactions that are more likely to govern protein dynamics. This Perspective discusses applications of graph-based algorithms to dissect dynamical hydrogen-bond networks of protein complexes, with illustrations for GPCRs and spike protein S. H-bond graphs provide an overview of sites in GPCR structures where hydrogen-bond dynamics would be required to assemble longer-distance networks between functionally important motifs. In the case of spike protein S, graphs identify regions of the protein where hydrogen bonds rearrange during the reaction cycle and where local hydrogen-bond networks likely change in a virus variant of concern.

Structural basis of adenylyl cyclase 9 activation.

Adenylyl cyclase 9 (AC9) is a membrane-bound enzyme that converts ATP into cAMP. The enzyme is weakly activated by forskolin, fully activated by the G protein Gαs subunit and is autoinhibited by the AC9 C-terminus. Although our recent structural studies of the AC9-Gαs complex provided the framework for understanding AC9 autoinhibition, the conformational changes that AC9 undergoes in response to activator binding remains poorly understood. Here, we present the cryo-EM structures of AC9 in several distinct states: (i) AC9 bound to a nucleotide inhibitor MANT-GTP, (ii) bound to an artificial activator (DARPin C4) and MANT-GTP, (iii) bound to DARPin C4 and a nucleotide analogue ATPαS, (iv) bound to Gαs and MANT-GTP. The artificial activator DARPin C4 partially activates AC9 by binding at a site that overlaps with the Gαs binding site. Together with the previously observed occluded and forskolin-bound conformations, structural comparisons of AC9 in the four conformations described here show that secondary structure rearrangements in the region surrounding the forskolin binding site are essential for AC9 activation.

Preproteins couple the intrinsic dynamics of SecA to its ATPase cycle to translocate via a catch and release mechanism.

Protein machines undergo conformational motions to interact with and manipulate polymeric substrates. The Sec translocase promiscuously recognizes, becomes activated, and secretes >500 non-folded preprotein clients across bacterial cytoplasmic membranes. Here, we reveal that the intrinsic dynamics of the translocase ATPase, SecA, and of preproteins combine to achieve translocation. SecA possesses an intrinsically dynamic preprotein clamp attached to an equally dynamic ATPase motor. Alternating motor conformations are finely controlled by the γ-phosphate of ATP, while ADP causes motor stalling, independently of clamp motions. Functional preproteins physically bridge these independent dynamics. Their signal peptides promote clamp closing; their mature domain overcomes the rate-limiting ADP release. While repeated ATP cycles shift the motor between unique states, multiple conformationally frustrated prongs in the clamp repeatedly "catch and release" trapped preprotein segments until translocation completion. This universal mechanism allows any preprotein to promiscuously recognize the translocase, usurp its intrinsic dynamics, and become secreted.

Conserved hydrogen-bond motifs of membrane transporters and receptors.

Membrane transporters and receptors often rely on conserved hydrogen bonds to assemble transient paths for ion transfer or long-distance conformational couplings. For transporters and receptors that use proton binding and proton transfer for function, inter-helical hydrogen bonds of titratable protein sidechains that could change protonation are of central interest to formulate hypotheses about reaction mechanisms. Knowledge of hydrogen bonds common at sites of potential interest for proton binding could thus inform and guide studies on functional mechanisms of protonation-coupled membrane proteins. Here we apply graph-theory approaches to identify hydrogen-bond motifs of carboxylate and histidine sidechains in a large data set of static membrane protein structures. We find that carboxylate-hydroxyl hydrogen bonds are present in numerous structures of the dataset, and can be part of more extended H-bond clusters that could be relevant to conformational coupling. Carboxylate-carboxyamide and imidazole-imidazole hydrogen bonds are represented in comparably fewer protein structures of the dataset. Atomistic simulations on two membrane transporters in lipid membranes suggest that many of the hydrogen bond motifs present in static protein structures tend to be robust, and can be part of larger hydrogen-bond clusters that recruit additional hydrogen bonds.

Algorithm to catalogue topologies of dynamic lipid hydrogen-bond networks.

Lipid membrane interfaces host reactions essential for the functioning of cells. The hydrogen-bonding environment at the membrane interface is particularly important for binding of proteins, drug molecules, and ions. We present here the implementation and applications of a depth-first search algorithm that analyzes dynamic lipid interaction networks. Lipid hydrogen-bond networks sampled transiently during simulations of lipid bilayers are clustered according to main types of topologies that characterize three-dimensional arrangements of lipids connected to each other via short water bridges. We characterize the dynamics of hydrogen-bonded lipid clusters in simulations of model POPE and POPE:POPG membranes that are often used for bacterial membrane proteins, in a model of the Escherichia coli membrane with six different lipid types, and in POPS membranes. We find that all lipids sample dynamic hydrogen-bonded networks with linear, star, or circular arrangements of the lipid headgroups, and larger networks with combinations of these three types of topologies. Overall, linear lipid-water bridges tend to be short. Water-mediated lipid clusters in all membranes with PE lipids tend to be somewhat small, with about four lipids in all membranes studied here. POPS membranes allow circular arrangements of three POPS lipids to be sampled frequently, and complex arrangements of linear, star, and circular paths may also be sampled. These findings suggest a molecular picture of the membrane interface whereby lipid molecules transiently connect in clusters with somewhat small spatial extension.

Mechanisms of long-distance allosteric couplings in proton-binding membrane transporters.

Membrane transporters that use proton binding and proton transfer for function couple local protonation change with changes in protein conformation and water dynamics. Changes of protein conformation might be required to allow transient formation of hydrogen-bond networks that bridge proton donor and acceptor pairs separated by long distances. Inter-helical hydrogen-bond networks adjust rapidly to protonation change, and ensure rapid response of the protein structure and dynamics. Membrane transporters with known three-dimensional structures and proton-binding groups inform on general principles of protonation-coupled protein conformational dynamics. Inter-helical hydrogen bond motifs between proton-binding carboxylate groups and a polar sidechain are observed in unrelated membrane transporters, suggesting common principles of coupling protonation change with protein conformational dynamics.

Graphs of protein-water hydrogen bond networks to dissect structural movies of ion-transfer microbial rhodopsins.

Microbial rhodopsins are membrane proteins that use the energy absorbed by the covalently bound retinal chromophore to initiate reaction cycles resulting in ion transport or signal transduction. Thousands of distinct microbial rhodopsins are known and, for many rhodopsins, three-dimensional structures have been solved with structural biology, including as entire sets of structures solved with serial femtosecond crystallography. This sets the stage for comprehensive studies of large datasets of static protein structures to dissect structural elements that provide functional specificity to the various microbial rhodopsins. A challenge, however, is how to analyze efficiently intra-molecular interactions based on large datasets of static protein structures. Our perspective discusses the usefulness of graph-based approaches to dissect structural movies of microbial rhodopsins solved with time-resolved crystallography.

Opioid Receptors and Protonation-Coupled Binding of Opioid Drugs.

Opioid receptors are G-protein-coupled receptors (GPCRs) part of cell signaling paths of direct interest to treat pain. Pain may associate with inflamed tissue characterized by acidic pH. The potentially low pH at tissue targeted by opioid drugs in pain management could impact drug binding to the opioid receptor, because opioid drugs typically have a protonated amino group that contributes to receptor binding, and the functioning of GPCRs may involve protonation change. In this review, we discuss the relationship between structure, function, and dynamics of opioid receptors from the perspective of the usefulness of computational studies to evaluate protonation-coupled opioid-receptor interactions.

C-Graphs Tool with Graphical User Interface to Dissect Conserved Hydrogen-Bond Networks: Applications to Visual Rhodopsins.

Dynamic hydrogen-bond networks provide proteins with structural plasticity required to translate signals such as ligand binding into a cellular response or to transport ions and larger solutes across membranes and, thus, are of central interest to understand protein reaction mechanisms. Here, we present C-Graphs, an efficient tool with graphical user interface that analyzes data sets of static protein structures or of independent numerical simulations to identify conserved, vs unique, hydrogen bonds and hydrogen-bond networks. For static structures, which may belong to the same protein or to proteins with different sequences, C-Graphs uses a clustering algorithm to identify sites of the hydrogen-bond network where waters are conserved among the structures. Using C-Graphs, we identify an internal protein-water hydrogen-bond network common to static structures of visual rhodopsins and adenosine A2A G protein-coupled receptors (GPCRs). Molecular dynamics simulations of a visual rhodopsin indicate that the conserved hydrogen-bond network from static structure can recruit dynamic hydrogen bonds and extend throughout most of the receptor. We release with this work the code for C-Graphs and its graphical user interface.

CHARMM Force-Field Parameters for Morphine, Heroin, and Oliceridine, and Conformational Dynamics of Opioid Drugs.

Opioid drug binding to specialized G protein-coupled receptors (GPCRs) can lead to analgesia upon activation via downstream Gi protein signaling and to severe side effects via activation of the ß-arrestin signaling pathway. Knowledge of how different opioid drugs interact with receptors is essential, as it can inform and guide the design of safer therapeutics. We performed quantum and classical mechanical computations to explore the potential energy landscape of four opioid drugs: morphine and its derivatives heroin and fentanyl and for the unrelated oliceridine. From potential energy profiles for bond twists and from interactions between opioids and water, we derived a set of force-field parameters that allow a good description of structural properties and intermolecular interactions of the opioids. Potential of mean force profiles computed from molecular dynamics simulations indicate that fentanyl and oliceridine have complex energy landscapes with relatively small energy penalties, suggesting that interactions with the receptor could select different binding poses of the drugs.

Interactive Interface for Graph-Based Analyses of Dynamic H-Bond Networks: Application to Spike Protein S.

Dynamic hydrogen-bond networks are key determinants of protein conformational dynamics. In the case of macromolecular protein complexes, which can have a large number of hydrogen bonds giving rise to extensive hydrogen-bond networks, efficient algorithms are required to analyze interactions that could be important for the dynamics and biological function of the complex. We present here a highly efficient, standalone interface designed for analyses of dynamical hydrogen-bond networks of biomolecules and macromolecular complexes. To facilitate a comprehensive description of protein dynamics, the interface includes analyses of hydrophobic interactions. We illustrate the usefulness and workflow of the interface by dissecting the dynamics of the ectodomain of SARS-CoV-2 protein S in its closed conformation. We find that protein S contains numerous local clusters of dynamic hydrogen bonds and identify hydrogen bonds that are sampled persistently. The receptor binding domain of the spike protein hosts only a handful of persistent hydrogen-bond clusters, suggesting structural plasticity. Our data analysis interface is released here for open use.