Flexible Fitting of Small Molecules into Electron Microscopy Maps Using Molecular Dynamics Simulations with Neural Network Potentials

Despite significant advances in resolution, the potential for cryo-electron microscopy (EM) to be used in determining the structures of protein-drug complexes remains unrealized. Determination of accurate structures and coordination of bound ligands necessitates simultaneous fitting of the models into the density envelopes, exhaustive sampling of the ligand geometries, and, most importantly, concomitant rearrangements in the side chains to optimize the binding energy changes. In this article, we present a flexible-fitting pipeline where molecular dynamics flexible fitting (MDFF) is used to refine structures of protein-ligand complexes from 3 to 5 Å electron density data. Enhanced sampling is employed to explore the binding pocket rearrangements. To provide a model that can accurately describe the conformational dynamics of the chemically diverse set of small-molecule drugs inside MDFF, we use QM/MM and neural-network potential (NNP)/MM models of protein-ligand complexes, where the ligand is represented using the QM or NNP model, and the protein is represented using established molecular mechanical force fields (e.g., CHARMM). This pipeline offers structures commensurate to or better than recently submitted high-resolution cryo-EM or X-ray models, even when given medium to low-resolution data as input. The use of the NNPs makes the algorithm more robust to the choice of search models, offering a radius of convergence of 6.5 Å for ligand structure determination. The quality of the predicted structures was also judged by density functional theory calculations of ligand strain energy. This strain potential energy is found to systematically decrease with better fitting to density and improved ligand coordination, indicating correct binding interactions. A computationally inexpensive protocol for computing strain energy is reported as part of the model analysis protocol that monitors both the ligand fit as well as model quality.

Immune protection is dependent on the gut microbiome in a lethal mouse gammaherpesviral infection

Immunopathogenesis in systemic viral infections can induce a septic state with leaky capillary syndrome, disseminated coagulopathy, and high mortality with limited treatment options. Murine gammaherpesvirus-68 (MHV-68) intraperitoneal infection is a gammaherpesvirus model for producing severe vasculitis, colitis and lethal hemorrhagic pneumonia in interferon gamma receptor-deficient (IFNγR-/-) mice. In prior work, treatment with myxomavirus-derived Serp-1 or a derivative peptide S-7 (G305TTASSDTAITLIPR319) induced immune protection, reduced disease severity and improved survival after MHV-68 infection. Here, we investigate the gut bacterial microbiome in MHV-68 infection. Antibiotic suppression markedly accelerated MHV-68 pathology causing pulmonary consolidation and hemorrhage, increased mortality and specific modification of gut microbiota. Serp-1 and S-7 reduced pulmonary pathology and detectable MHV-68 with increased CD3 and CD8 cells. Treatment efficacy was lost after antibiotic treatments with associated specific changes in the gut bacterial microbiota. In summary, transkingdom host-virus-microbiome interactions in gammaherpesvirus infection influences gammaherpesviral infection severity and reduces immune modulating therapeutic efficacy.

Serp-2, a virus-derived apoptosis and inflammasome inhibitor, attenuates liver ischemia-reperfusion injury in mice

BACKGROUND

Ischemia-reperfusion injury (IRI) is an antigen-independent, innate immune response to arterial occlusion and ischemia with subsequent paradoxical exacerbation after reperfusion. IRI remains a critical problem after vessel occlusion and infarction or during harvest and surgery in transplants. After transplant, liver IRI (LIRI) contributes to increased acute and chronic rejection and graft loss. Tissue loss during LIRI has been attributed to local macrophage activation and invasion with excessive inflammation together with hepatocyte apoptosis and necrosis. Inflammatory and apoptotic signaling are key targets for reducing post-ischemic liver injury.Myxomavirus is a rabbit-specific leporipoxvirus that encodes a suite of immune suppressing proteins, often with extensive function in other mammalian species. Serp-2 is a cross-class serine protease inhibitor (serpin) which inhibits the inflammasome effector protease caspase-1 as well as the apoptotic proteases granzyme B and caspases 8 and 10. In prior work, Serp-2 reduced inflammatory cell invasion after angioplasty injury and after aortic transplantation in rodents. In this report, we explore the potential for therapeutic treatment with Serp-2 in a mouse model of LIRI.

METHODS

Wildtype (C57BL/6 J) mice were subjected to warm, partial (70%) hepatic ischemia for 90 min followed by treatment with saline or Serp-2 or M-T7, 100 ng/g/day given by intraperitoneal injection on alternate days for 5 days. M-T7 is a Myxomavirus-derived inhibitor of chemokine-GAG interactions and was used in this study for comparative analysis of an unrelated viral protein with an alternative immunomodulating mechanism of action. Survival, serum ALT levels and histopathology were assessed 24 h and 10 days post-LIRI.

RESULTS

Serp-2 treatment significantly improved survival to 85.7% percent versus saline-treated wildtype mice (p = 0.0135), while M-T7 treatment did not significantly improve survival (p = 0.2584). Liver viability was preserved by Serp-2 treatment with a significant reduction in serum ALT levels (p = 0.0343) and infarct scar thickness (p = 0.0016), but with no significant improvement with M-T7 treatment. Suzuki scoring by pathologists blinded with respect to treatment group indicated that Serp-2 significantly reduced hepatocyte necrosis (p = 0.0057) and improved overall pathology score (p = 0.0046) compared to saline. Immunohistochemistry revealed that Serp-2 treatment reduced macrophage infiltration into the infarcted liver tissue (p = 0.0197).

CONCLUSIONS

Treatment with Serp-2, a virus-derived inflammasome and apoptotic pathway inhibitor, improves survival after liver ischemia-reperfusion injury in mouse models. Treatment with a cross-class immune modulator provides a promising new approach designed to reduce ischemia-reperfusion injury, improving survival and reducing chronic transplant damage.

Computational Study of Oxidation of Guanine by Singlet Oxygen (1 Δg ) and Formation of Guanine:Lysine Cross-Links

Oxidation of guanine in the presence of lysine can lead to guanine-lysine cross-links. The ratio of the C4, C5 and C8 crosslinks depends on the manner of oxidation. Type II photosensitizers such as Rose Bengal and methylene blue can generate singlet oxygen, which leads to a different ratio of products than oxidation by type I photosensitizers or by one electron oxidants. Modeling reactions of singlet oxygen can be quite challenging. Reactions have been explored using CASSCF, NEVPT2, DFT, CCSD(T), and BD(T) calculations with SMD implicit solvation. The spin contamination in open-shell calculations were corrected by Yamaguchi's approximate spin projection method. The addition of singlet oxygen to guanine to form guanine endo- peroxide proceeds step-wise via a zwitterionic peroxyl intermediate. The subsequent barrier for ring closure is smaller than the initial barrier for singlet oxygen addition. Ring opening of the endoperoxide by protonation at C4-O is followed by loss of a proton from C8 and dehydration to produce 8-oxoG^ox . The addition of lysine (modelled by methylamine) or water across the C5=N7 double bond of 8-oxoG^ox is followed by acyl migration to form the final spiro products. The barrier for methylamine addition is significantly lower than for water addition and should be the dominant reaction channel. These results are in good agreement with the experimental results for the formation of guanine-lysine cross-links by oxidation by type II photosensitizers.

Theoretical Determination of One-Electron Oxidation Potentials for Nucleic Acid Bases

The oxidation potentials for N-methyl substituted nucleic acid bases guanine, adenine, cytosine, thymine, uracil, xanthine, and 8-oxoguanine were computed using B3LYP and CBS-QB3 with the SMD solvation model. Acid-base and tautomeric equilibria present in aqueous solution were accounted for by combining standard redox potentials with calculated pKa and tautomerization energies to produce an ensemble averaged pH dependent potential. Gas phase free energies were computed using B3LYP/aug-cc-pVTZ//B3LYP/6-31+G(d,p) and CBS-QB3. Solvation free energies were computed at the SMD/B3LYP/6-31+G(d,p) level of theory. Compared to experimental results, calculations with the CBS-QB3 level of theory have a mean absolute error (MAE) of ca. 1 kcal/mol for the gas phase proton affinity/gas phase basicity and an MAE of ca. 0.04 eV for the adiabatic/vertical ionization potentials. The B3LYP calculations have a MAE of ∼2 kcal/mol for the proton affinity/gas phase basicity data but systematically underestimated ionization potentials by 0.14-0.21 eV. Solvent cavities for charged solute species were rescaled uniformly by fitting computed pKa data to experimentally measured pKa values. After solvent cavity scaling, the MAEs for computed pKa's compared to experimental results are 0.7 for B3LYP and 0.9 for CBS-QB3. In acetonitrile, the computed E°(XH(+•)/XH) redox potentials are systematically lower than experimentally measured potentials by 0.21 V for CBS-QB3 and 0.33 V for B3LYP. However, the redox potentials relative to adenine are in very good agreement with experimental results, with MAEs of 0.10 V for CBS-QB3 and 0.07 V for B3LYP. In aqueous solution, B3LYP and CBS-QB3 have MAEs of 0.21 and 0.19 V for E7(X(•),H(+)/XH). Replacing the methyl substituent with ribose changes the calculated E7 potentials by 0.1-0.2 V. The calculated difference between the guanine and adenine oxidation potentials is too large compared to experimental results, but the calculated difference between guanine and 8-oxoguanine is in good agreement with the measured values.

Mechanistic aspects of the formation of guanidinohydantoin from spiroiminodihydantoin under acidic conditions

Experimentally, it was observed that the oxidized guanine lesion spiroiminodihydantoin (Sp) contained in highly purified oligodeoxynucleotides slowly converts to guanidinohydantoin (Gh). The reaction is accelerated in the presence of acid. The possible mechanisms of this transformation have been analyzed computationally. Specifically, the potential energy surface for formation of Gh from Sp has been mapped using B3LYP density functional theory, the aug-cc-pVTZ and 6-31+G(d,p) basis sets, and the integral equation formalism for the polarizable continuum model (IEF-PCM) solvation model. The results favor a mechanism in which proton-assisted hydration of the C6 carbonyl group forming a gem-diol leads to ring opening of the iminohydantoin ring. The resulting species resembles a beta-ketoacid in its ability to decarboxylate; tautomerization of the resulting enol forms Gh. The results of these studies indicate that incubation of nucleosides or oligonucleotides containing Sp should be avoided in acidic media when high purity or an accurate assessment of the amounts of hydantoin lesions is desired.

An exploration of mechanisms for the transformation of 8-oxoguanine to guanidinohydantoin and spiroiminodihydantoin by density functional theory

The potential energy surface for formation of 2-amino-5-hydroxy-7,9-dihydropurine-6,8-dione (5-OH-OG), guanidinohydantoin (Gh) and spiroiminodihydantoin (Sp) from 8-oxoguanine (8-oxoG) has been mapped out using B3LYP density functional theory, the aug-cc-pVTZ and 6-31+G(d,p) basis sets and the IEF-polarizable continuum model (PCM) solvation model. Three pathways for formation of 5-OH-OG from 8-oxoG were evaluated: (A) stepwise loss of two electrons and two protons to form the quinonoid intermediate 2-amino-7,9-dihydro-purine-6,8-dione (8-oxoG(ox)) followed by hydration; (B) stepwise loss of two electrons and one proton and net addition of hydroxide, in which the key step is nucleophilic addition to the 8-oxoG radical cation; and (C) stepwise loss of one electron and one proton and addition of hydroxyl radical to the 8-oxoG radical cation. The data suggest that all three pathways are energetically feasible mechanisms for the formation of 5-OH-OG, however, Pathway A may be kinetically favored over Pathway B. Although lower in energy, Pathway C may be of limited biological significance since it depends on the local concentration of hydroxyl radical. Pathways for hydrolysis and decarboxylation of 5-OH-OG to form Gh via either a carboxylic acid or substituted carbamic acid intermediate have been evaluated with the result that cleavage of the N1-C6 bond is clearly favored over that of the C5-C6 bond. Formation of Sp from 5-OH-OG via stepwise proton transfer and acyl migration or ring opening followed by proton transfer and ring closure have also been explored and suggest that deprotonation of the hydroxyl group facilitates a 1,2 acyl shift. Results of the calculations are consistent with experimental studies showing dependence of the Gh/Sp product ratio on pH. Under neutral and basic conditions, the data predict that formation of Sp is kinetically favored over the pathways for formation of Gh. Under acidic conditions, Gh is predicted to be the kinetically favored product.

Exploration of Mechanisms for the Transformation of 8-Hydroxy Guanine Radical to FAPyG by Density Functional Theory

The potential energy surface for the transformation of 8-hydroxy guanine radical to formamidopyrimidine adducts via four pathways has been mapped out using B3LYP density functional theory and the IEF-polarizable continuum model (PCM) solvation model. Results of the calculations are consistent with experimental studies indicating that numerous compounds may be formed during the oxidation and subsequent reduction of guanine, some of which can react over time to form the observed 2,6-diamino-4-hydroxy-5-formamidopyrimidine (FAPyG) adduct. All four pathways begin with the 8-hydroxyguanine radical (8-OHGrad) species. Pathway 1 proceeds with reduction of the 8-OHGrad to a hemiaminal species, which undergoes ring opening to either FAPyG or 2,5-diamino-4-hydroxy-6-formamidopyrimidine (2,5FAPyG). Pathway 2 begins with a water-assisted proton transfer from the hydroxyl group of 8-OHGrad to form an 8-oxyguanine radical. This radical species can undergo ring opening and subsequent reduction to form either FAPyG or 2,5FAPyG. Pathways 3 and 4 lead to formation of only the FAPyG ring-opened adduct. Both begin with ring opening of 8-OHGrad to yield a formimidic acid radical, which can either be reduced to the formimidic acid and then undergo tautomerization to FAPyG (pathway 3) or initially tautomerize to form one of two FAPyG radicals before being reduced to FAPyG (pathway 4). Of the four possible reaction pathways explored, pathway 2 appears to be slightly lower in energy than pathway 4, which in turn is lower in energy than pathways 1 and 3. The calculations indicate that reactions proceeding via pathway 2 may yield a 2,5FAPyG adduct, which is thermodynamically less stable than the FAPyG adduct but may be formed at least initially. Interconversion of the two isomers is possible via a hemiaminal adduct. In the presence of water, it is energetically possible to form the FAPyG adduct from the formimidic acid, the hemiaminal, and the 2,5FAPyG adducts. Calculations at the B3LYP/6-31+G(d) level of theory suggest that it will be possible to differentiate between the different intermediate adducts using IR and NMR spectroscopy.