Protein-protein and protein-nucleic acid interactions are often considered difficult drug targets because the surfaces involved lack obvious druggable pockets. Cryptic pockets could present opportunities for targeting these interactions, but identifying and exploiting these pockets remains challenging. Here, we apply a general pipeline for identifying cryptic pockets to the interferon inhibitory domain (IID) of Ebola virus viral protein 35 (VP35). VP35 plays multiple essential roles in Ebola’s replication cycle but lacks pockets that present obvious utility for drug design. Using adaptive sampling simulations and machine learning algorithms, we predict VP35 harbors a cryptic pocket that is allosterically coupled to a key dsRNA-binding interface. Thiol labeling experiments corroborate the predicted pocket and mutating the predicted allosteric network supports our model of allostery. Finally, covalent modifications that mimic drug binding allosterically disrupt dsRNA binding that is essential for immune evasion. Based on these results, we expect this pipeline will be applicable to other proteins.

Significance A protein is a shape-shifter, but it is currently unclear which of the many structures a protein can adopt are relevant for its function. Here, we examine conformations that contain a “cryptic” pocket (i.e., a pocket absent in ligand-free structures). Cryptic pockets have potential utility in drug discovery efforts because they provide a means to target “undruggable” proteins (i.e., proteins lacking known pockets) or enhance rather than inhibit protein function. In this study, we use a combination of thiol-labeling and kinetic assays, NMR, and molecular dynamic simulations to identify the function of the Ω-loop cryptic pocket in β-lactamase enzymes. We find that an open pocket population is beneficial for hydrolysis of the substrate benzylpenicillin.

The retinoids all-trans retinoic acid (ATRA) and bexarotene are active in acute myeloid leukemia (AML), but responses beyond acute promyelocytic leukemia (APL) have been more modest than APL responses. To determine whether chemical modification of bexarotene might augment retinoid responses in AML, we screened a series of 38 bexarotene derivatives for activity in a mouse MLL-AF9 leukemia cell line, which exhibits strong synergistic sensitivity to the combination of ATRA and bexarotene. We found that RXRA potency correlated with anti-leukemic activity and that only one compound (103-4) with dual RARA/RXRA activity was capable of ATRA-independent anti-leukemic activity. We evaluated bioisostere and cyclohexane modifications for potential resistance to P450 metabolism and found that bioisosteres reduced potency and that bezopyran, cyclopentane, and cyclohexene modifications only modestly reduced susceptibility to metabolism. Collectively, these studies provide a map of the structure-activity relationships of bexarotene with outcomes related to RXRA and RARA activity, corepressor binding, compound stability, and anti-leukemic potential.

Many proteins are classified as ‘undruggable,’ especially those that engage in protein-protein and protein-nucleic acid interactions. Discovering ‘cryptic’ pockets that are absent in available structures but open due to protein dynamics could provide new druggable sites. Here, we integrate atomically-detailed simulations and biophysical experiments to search for cryptic pockets in viral protein 35 (VP35) from the highly lethal Ebola virus. VP35 plays multiple essential roles in Ebola’s replication cycle, including binding the viral RNA genome to block a host’s innate immunity. However, VP35 has so far proved undruggable. Using adaptive sampling simulations and allosteric network detection algorithms, we uncover a cryptic pocket that is allosterically coupled to VP35’s key RNA-binding interface. Experimental tests corroborate the predicted pocket and confirm that stabilizing the open form allosterically disrupts RNA binding. These results demonstrate simulations’ power to characterize hidden conformations and dynamics, uncovering cryptic pockets and allostery that present new therapeutic opportunities.

The functional mechanisms of multidomain proteins often exploit interdomain interactions, or "cross-talk." An example is human Pin1, an essential mitotic regulator consisting of a Trp-Trp (WW) domain flexibly tethered to a peptidyl-prolyl isomerase (PPIase) domain, resulting in interdomain interactions important for Pin1 function. Substrate binding to the WW domain alters its transient contacts with the PPIase domain via means that are only partially understood. Accordingly, we have investigated Pin1 interdomain interactions using NMR paramagnetic relaxation enhancement (PRE) and molecular dynamics (MD) simulations. The PREs show that apo-Pin1 samples interdomain contacts beyond the range suggested by previous structural studies. They further show that substrate binding to the WW domain simultaneously alters interdomain separation and the internal conformation of the WW domain. A 4.5-μs all-atom MD simulation of apo-Pin1 suggests that the fluctuations of interdomain distances are correlated with fluctuations of WW domain interresidue contacts involved in substrate binding. Thus, the interdomain/WW domain conformations sampled by apo-Pin1 may already include a range of conformations appropriate for binding Pin1's numerous substrates. The proposed coupling between intra-/interdomain conformational fluctuations is a consequence of the dynamic modular architecture of Pin1. Such modular architecture is common among cell-cycle proteins; thus, the WW-PPIase domain cross-talk mechanisms of Pin1 may be relevant for their mechanisms as well.

Severe fever with thrombocytopenia syndrome virus (SFTSV) is a tick-borne virus with 12%-30% case mortality rates and is related to the Heartland virus (HRTV) identified in the United States. Together, SFTSV and HRTV are emerging segmented, negative-sense RNA viral (sNSV) pathogens with potential global health impact. Here, we characterize the amino-terminal cap-snatching endonuclease domain of SFTSV polymerase (L) and solve a 2.4-angstrom X-ray crystal structure. While the overall structure is similar to those of other cap-snatching sNSV endonucleases, differences near the C terminus of the SFTSV endonuclease suggest divergence in regulation. Influenza virus endonuclease inhibitors, including the US Food and Drug Administration (FDA) approved Baloxavir (BXA), inhibit the endonuclease activity in in vitro enzymatic assays and in cell-based studies. BXA displays potent activity with a half maximal inhibitory concentration (IC50) of similar to 100 nM in enzyme inhibition and an EC50 value of similar to 250 nM against SFTSV and HRTV in plaque assays. Together, our data support sNSV endonucleases as an antiviral target.

Ebola virus nucleoprotein (eNP) assembles into higher-ordered structures that form the viral nucleocapsid (NC) and serve as the scaffold for viral RNA synthesis. However, molecular insights into the NC assembly process are lacking. Using a hybrid approach, we characterized the NC-like assembly of eNP, identified novel regulatory elements, and described how these elements impact function. We generated a three-dimensional structure of the eNP NC-like assembly at 5.8 Å using electron cryo-microscopy and identified a new regulatory role for eNP helices α22-α23. Biochemical, biophysical, and mutational analyses revealed that inter-eNP contacts within α22-α23 are critical for viral NC assembly and regulate viral RNA synthesis. These observations suggest that the N terminus and α22-α23 of eNP function as context-dependent regulatory modules (CDRMs). Our current study provides a framework for a structural mechanism for NC-like assembly and a new therapeutic target.

Increasing evidence shows that active sites of proteins have non-trivial conformational dynamics. These dynamics include active site residues sampling different local conformations that allow for multiple, and possibly novel, inhibitor binding poses. Yet, active site dynamics garner only marginal attention in most inhibitor design efforts and exert little influence on synthesis strategies. This is partly because synthesis requires a level of atomic structural detail that is frequently missing in current characterizations of conformational dynamics. In particular, while the identity of the mobile protein residues may be clear, the specific conformations they sample remain obscure. Here, we show how an appropriate choice of ligand can significantly sharpen our abilities to describe the interconverting binding poses (conformations) of protein active sites. Specifically, we show how 2-(2'-carboxyphenyl)-benzoyl-6-aminopenicillanic acid (CBAP) exposes otherwise hidden dynamics of a protein active site that binds β-lactam antibiotics. When CBAP acylates (binds) the active site serine of the β-lactam sensor domain of BlaR1 (BlaRS), it shifts the time scale of the active site dynamics to the slow exchange regime. Slow exchange enables direct characterization of inter-converting protein and bound ligand conformations using NMR methods. These methods include chemical shift analysis, 2-d exchange spectroscopy, off-resonance ROESY of the bound ligand, and reduced spectral density mapping. The active site architecture of BlaRS is shared by many β-lactamases of therapeutic interest, suggesting CBAP could expose functional motions in other β-lactam binding proteins. More broadly, CBAP highlights the utility of identifying chemical probes common to structurally homologous proteins to better expose functional motions of active sites.

Computational approaches can facilitate the design of clinically relevant small molecule inhibitors. Classical drug design focuses primarily on targeting a protein's active site based on its x-ray crystal structure. One drawback to this approach is crystal structures do not report on protein dynamics and do not capitalize on the presence of hidden cryptic pockets. Cryptic pockets are transient allosteric sites invisible to conventional structural biology techniques and present new opportunities for drug design. Here we introduce Boltzmann docking as a method to identify small molecules that bind to cryptic pockets and allosterically modulate enzymatic activity. We leverage knowledge of protein dynamics from molecular dynamics simulations and Markov State Models (MSMs) to identify cryptic pockets. MSMs are a network representation of a protein's free energy landscape that provides representative states and their populations. Boltzmann docking screens a library of small molecules against multiple MSM states containing a target cryptic pocket and generates a new Boltzmann score that is a population weighted average of the individual states’ docking scores. We successfully applied this approach to identify multiple allosteric modulators of TEM β-lactamase activity, and are currently looking to identify novel inhibitors of the Ebola viral protein VP35. Boltzmann docking should provide a general framework to identify allosteric modulators of protein activity, especially for targets widely considered “undruggable” owing to the lack of an obvious binding site in their crystal structures.

Protein stabilization is fundamental to enzyme function and evolution, yet understanding the determinants of a protein's stability remains a challenge. This is largely due to a shortage of atomically-detailed models for the ensemble of relevant protein conformations and their relative populations. For example, the M182T substitution in TEM β-lactamase, an enzyme that confers antibiotic resistance to bacteria, is stabilizing but the precise mechanism remains unclear. Here, we employ Markov state models (MSMs) to uncover how M182T shifts the distribution of different structures that TEM adopts. We find that M182T stabilizes a helix that is a key component of a domain interface. We then predict the effects of other mutations, including a novel stabilizing mutation, and experimentally test our predictions using a combination of stability measurements, crystallography, NMR, and in vivo measurements of bacterial fitness. We expect our insights and methodology to provide a valuable foundation for protein design.

Allosteric drugs, which bind to proteins in regions other than their main ligand-binding or active sites, make it possible to target proteins considered “undruggable” and to develop new therapies that circumvent existing resistance. Despite growing interest in allosteric drug discovery, rational design is limited by a lack of sufficient structural information about alternative binding sites in proteins. Previously, we used Markov State Models (MSMs) to identify such “cryptic pockets,” and here we describe a method for identifying compounds that bind in these cryptic pockets and modulate enzyme activity. Experimental tests validate our approach by revealing both an inhibitor and two activators of TEM β-lactamase (TEM). To identify hits, a library of compounds is first virtually screened against either the crystal structure of a known cryptic pocket or an ensemble of structures containing the same cryptic pocket that is extracted from an MSM. Hit compounds are then screened experimentally and characterized kinetically in individual assays. We identify three hits, one inhibitor and two activators, demonstrating that screening for binding to allosteric sites can result in both positive and negative modulation. The hit compounds have modest effects on TEM activity, but all have higher affinities than previously identified inhibitors, which bind the same cryptic pocket but were found, by chance, via a computational screen targeting the active site. Site-directed mutagenesis of key contact residues predicted by the docking models is used to confirm that the compounds bind in the cryptic pocket as intended. Because hit compounds are identified from docking against both the crystal structure and structures from the MSM, this platform should prove suitable for many proteins, particularly targets whose crystal structures lack obvious druggable pockets, and for identifying both inhibitory and activating small-molecule modulators.

The retinoid X receptor α (RXRA) has been implicated in diverse hematological processes. To identify natural ligands of RXRA that are present in hematopoietic cells, we adapted an upstream activation sequence-green fluorescent protein (UAS-GFP) reporter mouse to detect natural RXRA ligands in vivo. We observed reporter activity in diverse types of hematopoietic cells in vivo. Reporter activity increased during granulocyte colony-stimulating factor (G-CSF)-induced granulopoiesis and after phenylhydrazine (PHZ)-induced anemia, suggesting the presence of dynamically regulated natural RXRA ligands in hematopoietic cells. Mouse plasma activated Gal4-UAS reporter cells in vitro, and plasma from mice treated with G-CSF or PHZ recapitulated the patterns of reporter activation that we observed in vivo. Plasma from mice with dietary vitamin A deficiency only mildly reduced RXRA reporter activity, whereas plasma from mice on a fatty acid restriction diet reduced reporter activity, implicating fatty acids as plasma RXRA ligands. Through differential extraction coupled with mass spectrometry, we identified the long-chain fatty acid C24:5 as a natural RXRA ligand that was greatly increased in abundance in response to hematopoietic stress. Together, these data suggest that natural RXRA ligands are present and dynamically increased in abundance in mouse hematopoietic cells in vivo.

To address the growing need for new antimicrobial agents, we explored whether inhibition of bacterial signaling machinery could inhibit bacterial growth. Because bacteria rely on two-component signaling systems to respond to environmental changes, and because these systems are both highly conserved and mediated by histidine kinases, inhibiting histidine kinases may provide broad spectrum antimicrobial activity. The histidine kinase ATP binding domain is conserved with the ATPase domain of eukaryotic Hsp90 molecular chaperones. To find a chemical scaffold for compounds that target histidine kinases, we leveraged this conservation. We screened ATP competitive Hsp90 inhibitors against CckA, an essential histidine kinase in Caulobacter crescentus that controls cell growth, and showed that the diaryl pyrazole is a promising scaffold for histidine kinase inhibition. We synthesized a panel of derivatives and found that they inhibit the histidine kinases C. crescentus CckA and Salmonella PhoQ but not C. crescentus DivJ; and they inhibit bacterial growth in both Gram-negative and Gram-positive bacterial strains.

The transmembrane antibiotic sensor/signal transducer protein BlaR1 is part of a cohort of proteins that confer β-lactam antibiotic resistance in methicillin-resistant Staphylococcus aureus (MRSA) [Fisher, J. F., Meroueh, S. O., and Mobashery, S. (2005) Chem. Rev. 105, 395-424; Llarrull, L. I., Fisher, J. F., and Mobashery, S. (2009) Antimicrob. Agents Chemother. 53, 4051-4063; Llarrull, L. I., Toth, M., Champion, M. M., and Mobashery, S. (2011) J. Biol. Chem. 286, 38148-38158]. Specifically, BlaR1 regulates the inducible expression of β-lactamases that hydrolytically destroy β-lactam antibiotics. The resistance phenotype starts with β-lactam antibiotic acylation of the BlaR1 extracellular domain (BlaRS). The acylation activates the cytoplasmic protease domain through an obscure signal transduction mechanism. Here, we compare protein dynamics of apo versus antibiotic-acylated BlaRS using nuclear magnetic resonance. Our analyses reveal inter-residue interactions that relay acylation-induced perturbations within the antibiotic-binding site to the transmembrane helix regions near the membrane surface. These are the first insights into the process of signal transduction by BlaR1.

In methicillin-resistant Staphylococcus aureus, β-lactam antibiotic resistance is mediated by the transmembrane protein BlaR1. The antibiotic sensor domain BlaR(S) and the L2 loop of BlaR1 are on the membrane surface. We used NMR to investigate interactions between BlaR(S) and a water-soluble peptide from L2. This peptide binds BlaR(S) proximal to the antibiotic acylation site as an amphipathic helix. Acylation of BlaR(S) by penicillin G does not disrupt binding. These results suggest a signal transduction mechanism whereby the L2 helix, partially embedded in the membrane, propagates conformational changes caused by BlaR(S) acylation through the membrane via transmembrane segments, leading to antibiotic resistance.

Bis(monoacylglycero)phosphate (BMP) is an endosomal lipid with a unique structure that is implicated in the formation of intraendosomal vesicular bodies. Here we have characterized the effects of dioleoyl-BMP (BMP18:1) at concentrations of 5, 10, 15 and 20mol% on the thermotropic behavior of dipalmitoyl phosphatidylcholine (DPPC) vesicles, and compared them to those of equimolar concentrations of dioleoyl phosphatidylglycerol (DOPG), a structural isoform of BMP18:1. Because BMP is found in the acidic environments of the late endosome and intralysosomal vesicles, samples were prepared at pH 4.2 to mimic the pH of the lysosome. Both 2H NMR of perdeuterated DPPC and spin-labeled EPR with 16-doxyl phosphatidylcholine were utilized in these investigations. NMR and EPR results show that BMP18:1 induces a lowering in the main phase transition temperature of DPPC similar to that of DOPG. The EPR studies reveal that BMP18:1 induced more disorder in the Lβ phase when compared to equimolar concentrations of DOPG. Analysis from dePaked 2H NMR spectra in the Lα phase reveals that BMP18:1 induces less disorder than equal concentrations of DOPG. Additionally, the results demonstrate that BMP mixes with other phospholipids as a phospholipid and not as a detergent molecule as once speculated.

Bis(monoacylglycero)phosphate (BMP18:1) is an uniquely structured anionic phospholipid found in enriched concentrations in the internal membranes of the lysosome and late endosome.1 BMP18:1, differs from typical phospholipids, possessing two glycerol moieties each with a single oleoyl acyl-chain as well as an atypical sn-1:sn-1′ stereoconfiguration. It has been suggested that the unusual structure of BMP18:1 plays a role in lipid and protein trafficking, as well as glycosphingolipid degradation.2,3 Dimyristoyl-BMP (BMP14:0) and dipalmitoylphosphatidylcholine (DPPC) undergo a phase transition from the gel to liquid crystalline phase at 42 °C, whereas the phosphatidylcholine with myristoyl acyl-chains, dimyristoylphosphatidylcholine (DMPC), undergoes a gel to liquid crystalline phase transition at 23 °C. It is therefore of interest to investigate the thermotropic phase behavior of saturated BMP analogues. Temperature dependent 2H-NMR was used to characterize analogues of BMP with lauroyl, myristoyl, palmitoyl, or stearoyl acyl-chains. First moment (M1) analyses of 2H-NMR spectra can be used to monitor the gel to liquid crystalline phase transition. The phase transition temperature for each BMP analogue was found to be higher than the phosphatidylcholine consisting of the corresponding acyl-chain. BMP18:0 was studied under three different buffer conditions to determine the effects of pH and ionic strength on the phase transition. 1. Kobayshi et al., J. Biol. Chem., 2002, 277, 32157. 2. Gruenberg, Curr. Opin. Cell Biol., 2003, 15, 382. 3. Schulze, Kolter, and Sandhoff, BBA, 2009, 1793, 674.

The morphology and size of hydrated lipid dispersions of bis(monoacylglycero)phosphate (BMP) mixed with varying mole percentages of the ganglioside GM1 were investigated by dynamic light scattering (DLS) and transmission electron microscopy (TEM). Electron paramagnetic resonance (EPR) spectroscopy of these same mixtures, doped at 0.5mol% with doxyl labeled lipids, was used to investigate acyl-chain packing. Results show that for 20–30% GM1, hydrated BMP:GM1 mixtures spontaneously form small spherical vesicles with diameters ∼100nm and a narrow size distribution profile. For other concentrations of GM1, hydrated dispersions with BMP have non-spherical shapes and heterogeneous size profiles, with average vesicle diameters>400nm. All samples were prepared at pH 5.5 to mimic the lumen acidity of the late endosome where BMP is an essential component of intraendosomal vesicle budding, lipid sorting and trafficking. These findings indicate that GM1 and BMP under a limited concentration range spontaneously form small vesicles of homogeneous size in an energy independent manner without the need of protein templating. Because BMP is essential for intraendosomal vesicle formation, these results imply that lipid–lipid interactions may play a critical role in the endosomal process of lipid sorting and trafficking.

Bis(monoacylglycero)phosphate (BMP) is an unusually shaped lipid found in relatively high percentage in the late endosome. Here, we report the characterization of the morphology and molecular organization of dioleoyl-BMP (DOBMP) with dynamic light scattering, transmission electron microscopy, nuclear magnetic resonance (NMR) spectroscopy, and electron paramagnetic resonance spectroscopy. The morphology of hydrated DOBMP dispersions varies with pH and ionic strength, and DOBMP vesicles are significantly smaller in diameter than phosphatidylcholine dispersions. At neutral pH, DOBMP forms highly structured, clustered dispersions 500 nm in size. On the other hand, at acidic pH, spherically shaped vesicles are formed. NMR and spin-labeled electron paramagnetic resonance demonstrate that DOBMP forms a lamellar mesophase with acyl-chain packing similar to that of other unsaturated phospholipids. 31P NMR reveals an orientation of the phosphate group in DOBMP that differs significantly from that of other phospholipids. These macroscopic and microscopic structural characterizations suggest that the biosynthesis of BMP on the inner luminal membrane of maturing endosomes may possibly produce budded vesicles high in BMP content, which form small vesicular structures stabilized by the physical properties of the BMP lipid.