On Dioxygen and Substrate Access to Soluble Methane Monooxygenases: An all‐Atom Molecular Dynamics Investigation in Water Solution

In a preliminary exploration of the dummy model for diiron proteins, random‐acceleration molecular dynamics (RAMD) revealed that a pure four‐helix bundle structure, like hemerythrin, constitutes an efficient cage for dioxygen (O₂), which can only leave from defined, albeit very broad, gates. However, this well ordered structure does not constitute an archetype on which to compare O₂ permeation of other diiron proteins, like the complex of soluble methane monooxygenase hydroxylase with the regulatory protein (sMMOH‐MMOB). The reason is that with this complex, unlike hemerythrin, the four helices of the four‐helix bundle are heavily bent, and RAMD showed that most traps for O₂ lie outside them. It was also observed that, in spite of a nearly identical van der Waals radius for O₂ and the natural substrate CH₄, the latter behaves under RAMD as a bulkier molecule than O₂, requiring a higher external force to be brought out of sMMOH‐MMOB along trajectories of viable length. All that determined with sMMOH‐MMOB multiple gates and multiple pathways to each of them through several binding pockets, for both O₂ and CH₄. Of the two equally preferred pathways for O₂, at right angle with one another, one proved to be in accordance with the Xe‐atom mapping for sMMOH. In contrast, none of the pathways identified for CH₄ proved to be in accordance with such mapping, CH₄ looking for more open avenues instead.     

On Dioxygen Permeation through a Dehydrogenase‐Pyrroloquinoline Quinone Complex. A Molecular‐Dynamics Investigation

In this work, an all atom model of the quinoprotein dehydrogenase PqqC in complex with the PQQ (=4,5‐dihydro‐4,5‐dioxo‐1H‐pyrrolo[2,3‐f]quinoline‐2,7,9‐tricarboxylic acid) cofactor and dioxygen (O₂), solvated with TIP3 water in periodic boxes, was subjected to random‐acceleration molecular dynamics (RAMD). It was found that O₂ leaves the active binding pocket, in front of PQQ, to get to the solvent, as easily as with a variety of other O₂‐activating enzymes, O₂ carriers, and gas‐sensing proteins. The shortest pathway, orthogonal to the center of the mean plane of PQQ, was largely preferred by O₂ over pathways slightly deviating from this line. These observations challenge the interpretation of an impermeable active binding pocket of PqqC‐PQQ, as drawn from both X‐ray diffraction data of the crystal at low temperature and physiological experimentation.     

From Dioxygen Storing to Dioxygen Sensing with Neuroglobins: An Insight from Molecular Mechanics

This work deals with two neuroglobins from phylogenetically distant organisms. Deriving from the acoelomorph Symsagittifera roscoffensis, SrNgb is functionally pentacoordinated, and is assumed to function as a reserve of dioxygen (O₂). Obtained from mice, mNgb is functionally hexacoordinated, and presumably triggers signals from sensing O₂. Here, it is investigated how these two globins are permeated by diatomic gases, SrNgb by O₂ and mNgb by CO. With protein atomic coordinates available from high‐resolution X‐ray diffraction analysis, O₂ and CO pathways were traced from molecular‐dynamics simulations in H₂O solution, which makes no difference between the two gases, accelerated by applying an external randomly‐oriented minimal force to the center of mass of the diatomic gas molecule. This allowed us to explore a statistically significant large number of trajectories. It emerged that CO leaves mNgb from preferentially peripheral gates located on the side of the heme propionate chains, whereas O₂ leaves SrNgb from the opposite side. This shows no analogy with either the functionally pentacoordinated, O₂‐transporting, myoglobin (Mgb), or the hexacoordinated, O₂‐sensing, cytoglobin, despite the same three‐over‐three typical α‐helical globin folding. The sole analogy that could be observed was a preference for the shortest diatomic gas pathways with both SrNgb and Mgb. It is tempting to speculate that this fulfills the need of being quick in delivering O₂ to depleted organs.     

On the Quest of Dioxygen by Monomeric Sarcosine Oxidase. A Molecular Dynamics Investigation

It is reported here on random acceleration molecular dynamics (RAMD) simulations with the 2GF3 bacterial monomeric sarcosine oxidase (MSOX), O₂, and furoic acid in place of sarcosine, solvated by TIP3 H₂O in a periodic box. An external tiny force, acting randomly on O₂, accelerated its relocation, from the center of activation between residue K265 and the si face of the flavin ring of the flavin adenine dinucleotide cofactor, to the surrounding solvent. Only three of the four O₂ gates previously described for this system along a composite method technique were identified, while two more major O₂ gates were found. The RAMD simulations also revealed that the same gate can be reached by O₂ along different pathways, often involving traps for O₂. Both the residence time of O₂ in the traps, and the total trajectory time for O₂ getting to the solvent, could be evaluated. The new quick pathways discovered here suggest that O₂ exploits all nearby interstices created by the thermal fluctuations of the protein, not having necessarily to look for the permanent large channel used for uptake of the FADH cofactor. To this regard, MSOX resembles closely KijD3 N‐oxygenase. These observations solicit experimental substantiation, in a long term aim at discovering whether gates and pathways for the small gaseous ligands inside the proteins are under Darwinian functional evolution or merely stochastic control operates.     

Binding Pockets and Permeation Channels for Dioxygen through Cofactorless 3‐Hydroxy‐2‐methylquinolin‐4‐one 2,4‐Dioxygenase in Association with Its Natural Substrate, 3‐Hydroxy‐2‐methylquinolin‐4(1H)‐one. A Perspective from Molecular Dynamics Simulations

This work describes an investigation of pathways and binging pockets (BPs) for dioxygen (O₂) through the cofactorless oxygenase 3‐hydroxy‐2‐methylquinolin‐4‐one 2,4‐dioxygenase in complex with its natural substrate, 3‐hydroxy‐2‐methylquinolin‐4(1H)‐one, in aqueous solution. The investigation tool was random‐acceleration molecular dynamics (RAMD), whereby a tiny, randomly oriented external force is applied to O₂ in order to accelerate its movements. In doing that, care was taken that the external force only continues, if O₂ moves along a direction for a given period of time, otherwise the force changed direction randomly. Gates for expulsion of O₂ from the protein, which can also be taken as gates for O₂ uptake, were found throughout almost the whole external surface of the protein, alongside a variety of BPs for O₂. The most exploited gates and BPs were not found to correspond to the single gate and BP proposed previously from the examination of the static model from X‐ray diffraction analysis of this system. Therefore, experimental investigations of this system that go beyond the static model are urgently needed.     

Unveiling the Pathways of Dioxygen Through the C2 Component of the Environmentally Relevant Monooxygenase p‐Hydroxyphenylacetate Hydroxylase from Acinetobacter baumannii: A Molecular Dynamics Investigation

In this work, models of the homotetrameric C2 component of the monooxygenase p‐hydroxyphenylacetate hydroxylase from Acinetobacter baumannii, in complex with dioxygen (O₂) and, or not, the substrate p‐hydroxyphenylacetate (HPA) were built. Both models proved to be amenable to random‐acceleration molecular dynamics (RAMD) simulations, whereby a tiny randomly oriented external force, acting on O₂ at the active site in front of flavin mononucleotide (FMNH^?), accelerated displacement of O₂ toward the bulk solvent. This allowed us to carry out a sufficiently large number of RAMD simulations to be of statistical significance. The two systems behaved very similarly under RAMD, except for O₂ leaving the active site more easily in the absence of HPA, but then finding similar obstacles in getting to the gate as when the active site was sheltered by HPA. This challenges previous conclusions that HPA can only reach the active center after that the C4aOOH derivative of FMNH^? is formed, requiring uptake of O₂ at the active site before HPA. According to these RAMD simulations, O₂ could well get to FMNH^? also in the presence of the substrate at the active site.     

Understanding How H‐NOX (Heme Nitric Oxide/Oxygen) Domain Works Needs First Clarifying How Diatomic Gases Are Relocated Inside This Sensing Protein. A Molecular‐Mechanics Approach

H‐NOX (Heme Nitric Oxide/Oxygen) domain has widespread occurrence, either standalone or associated with functional proteins, sending signals for functions that span from modulating vasodilation and neurotransmission with humans to competition and symbiosis with bacteria. Understanding how H‐NOX works, and possibly intervening on degeneration for health purposes, needs first clarifying how diatomic gases are relocated through this protein in relation to the deeply buried heme. To this end, a biased form of molecular dynamics, i.e., Random Accelaration Molecular Dynamics (RAMD), is used by applying a randomly oriented tiny force to heme‐dissociated CO of Nostoc sp. H‐NOX, while changing randomly the direction of the force, if CO travels less than specified for the evaluated block. The result is that a large area of the protein, comprising amino acids from serine 44 to leucine 67 along two adjacent helices, offers a broad portal to CO from the surrounding medium to the deeply buried heme. Most traffic is concentrated through a channel lined by tyrosine 49, valine 52, and leucine 67. This modifies the picture drawn from mapping Xe cavities on pressurizing Nostoc sp. H‐NOX with Xe gas. What is the main pathway with Xe‐cavity mapping becomes a minor pathway with RAMD, and vice versa. The reason is that the fluctuating protein under MD creates clefts for CO slipping through, as it is expected to occur in nature.     

Mechanisms of neuroprotection by hemopexin: modeling the control of heme and iron homeostasis in brain neurons in inflammatory states

Hemopexin provides neuroprotection in mouse models of stroke and intracerebral hemorrhage and protects neurons in vitro against heme or reactive oxygen species (ROS) toxicity via heme oxygenase‐1 (HO1) activity. To model human brain neurons experiencing hemorrhages and inflammation, we used human neuroblastoma cells, heme–hemopexin complexes, and physiologically relevant ROS, for example, H₂O₂ and HOCl, to provide novel insights into the underlying mechanism whereby hemopexin safely maintains heme and iron homeostasis. Human amyloid precursor protein (hAPP), needed for iron export from neurons, is induced ~twofold after heme–hemopexin endocytosis by iron from heme catabolism via the iron‐regulatory element of hAPP mRNA. Heme–hemopexin is relatively resistant to damage by ROS and retains its ability to induce the cytoprotective HO1 after exposure to tert‐butylhydroperoxide, although induction is impaired, but not eliminated, by exposure to high concentrations of H₂O₂ in vitro. Apo‐hemopexin, which predominates in non‐hemolytic states, resists damage by H₂O₂ and HOCl, except for the highest concentrations likely in vivo. Heme–albumin and albumin are preferential targets for ROS; thus, albumin protects hemopexin in biological fluids like CSF and plasma where it is abundant. These observations provide strong evidence that hemopexin will be neuroprotective after traumatic brain injury, with heme release in the CNS, and during the ensuing inflammation. Hemopexin sequesters heme, thus preventing unregulated heme uptake that leads to toxicity; it safely delivers heme to neuronal cells; and it activates the induction of proteins including HO1 and hAPP that keep heme and iron at safe levels in neurons.     

Structural studies of constitutive nitric oxide synthases with diatomic ligands bound

Crystal structures are reported for the endothelial nitric oxide synthase (eNOS)–arginine–CO ternary complex as well as the neuronal nitric oxide synthase (nNOS) heme domain complexed with l-arginine and diatomic ligands, CO or NO, in the presence of the native cofactor, tetrahydrobiopterin, or its oxidized analogs, dihydrobiopterin and 4-aminobiopterin. The nature of the biopterin has no influence on the diatomic ligand binding. The binding geometries of diatomic ligands to nitric oxide synthase (NOS) follow the {MXY} ⁿ formalism developed from the inorganic diatomic–metal complexes. The structures reveal some subtle structural differences between eNOS and nNOS when CO is bound to the heme which correlate well with the differences in CO stretching frequencies observed by resonance Raman techniques. The detailed hydrogen-bonding geometries depicted in the active site of nNOS structures indicate that it is the ordered active-site water molecule rather than the substrate itself that would most likely serve as a direct proton donor to the diatomic ligands (CO, NO, as well as O₂) bound to the heme. This has important implications for the oxygen activation mechanism critical to NOS catalysis.     

On the Dynamical Behavior of the Cysteine Dioxygenase‐l‐Cysteine Complex in the Presence of Free Dioxygen and l‐Cysteine

In this work, viable models of cysteine dioxygenase (CDO) and its complex with l ‐cysteine dianion were built for the first time, under strict adherence to the crystal structure from X‐ray diffraction studies, for all atom molecular dynamics (MD). Based on the CHARMM36 FF, the active site, featuring an octahedral dummy Fe(II) model, allowed us observing water exchange, which would have escaped attention with the more popular bonded models. Free dioxygen (O₂) and l ‐cysteine, added at the active site, could be observed being expelled toward the solvating medium under Random Accelerated Molecular Dynamics (RAMD) along major and minor pathways. Correspondingly, free dioxygen (O₂), added to the solvating medium, could be observed to follow the same above pathways in getting to the active site under unbiased MD. For the bulky l ‐cysteine, 600 ns of trajectory were insufficient for protein penetration, and the molecule was stuck at the protein borders. These models pave the way to free energy studies of ligand associations, devised to better clarify how this cardinal enzyme behaves in human metabolism.     

Gates and Binding Pockets for Nitric Oxide with Cytochrome c′, According to Molecular Dynamics

Random‐acceleration molecular‐dynamics (RAMD) simulations with models of homodimeric 6‐ligated distal‐NO and 5‐ligated proximal‐NO cytochrome c′ complexes, in TIP3 H₂O, showed two distinct, non‐intercommunicating worlds. In the framework of a long cavity formed by four protein helices with heme at one extremity, NO was observed to follow different pathways with the two complexes to reach the solvent. With the 6‐ligated complex, NO was observed to progress by exploiting protein internal channels created by thermal fluctuations, and be temporarily trapped into binding pockets before reaching the preferred gate at the heme end of the cavity. In contrast, with the 5‐ligated complex, NO was observed to surface the solvent‐exposed helix 7, up to a gate at the other extremity of the protein, only occasionally finding an earlier, direct way out toward the solvent. That only bulk NO gets involved in forming the 5‐ligated proximal‐NO complex is in agreement with previous experimental observations, while the occurrence of binding pockets suggests that also reservoir NO might play a role with the distal‐NO complex.     

Upregulation of heme oxygenase‐1 inhibits the maturation and mineralization of osteoblasts

Heme‐oxygenase‐1 (HO‐1), an important enzyme involved in vascular disease, transplantation, and inflammation, catalyzes the degradation of heme into carbon monoxide and biliverdin. It has been reported that overexpression of HO‐1 inhibits osteoclastogenesis. However, the effect of HO‐1 on osteoblast differentiation is still not clear. We here used adenoviral vector expressing recombinant human HO‐1 and HO‐1 inducer hemin to study the effects of HO‐1 in primary cultured osteoblasts. The results showed that induction of HO‐1 inhibited the maturation of osteoblasts including mineralized bone nodule formation, alkaline phosphatase activity and decreased mRNA expression of several differentiation markers such as alkaline phosphatase, osteocalcin, and RUNX2. Furthermore, downstream products of HO‐1, bilirubin, carbon monoxide, and iron, are involved in the inhibitory action of HO‐1. HO‐1 can be induced by H₂O₂, lipopolysaccharide and inflammatory cytokines such as TNF‐α and IL‐1β in osteoblasts and also in STZ‐induced diabetic mice. In addition, endogenous PPARγ ligand, 15‐deoxy‐Δ^12,14‐prostaglandin‐J2 (15d‐PGJ2) markedly increased both mRNA and protein levels of HO‐1 in osteoblasts via PI3K‐Akt and MAPK pathways. Blockade of HO activity by ZnPP IX antagonized the inhibitory action on osteocalcin expression by hemin and 15d‐PGJ2. Our results indicate that upregulation of HO‐1 inhibits the maturation of osteoblasts and HO‐1 may be involved in oxidative‐ or inflammation‐induced bone loss. J. Cell. Physiol. 222: 757–768, 2010. © 2009 Wiley‐Liss, Inc.     

Influence of Distal Residue B10 on CO Dynamics in Myoglobin and Neuroglobin

For many years, myoglobin has served as a paradigm for structure–function studies in proteins. Ligand binding and migration within myoglobin has been studied in great detail by crystallography and spectroscopy, showing that gaseous ligands such as O₂, CO, and NO not only bind to the heme iron but may also reside transiently in three internal ligand docking sites, the primary docking site B and secondary sites C and D. These sites affect ligand association and dissociation in specific ways. Neuroglobin is another vertebrate heme protein that also binds small ligands. Ligand migration pathways in neuroglobin have not yet been elucidated. Here, we have used Fourier transform infrared temperature derivative spectroscopy at cryogenic temperatures to compare the influence of the side chain volume of amino acid residue B10 on ligand migration to and rebinding from docking sites in myoglobin and neuroglobin.     

Understanding How H-NOX (Heme Nitric Oxide/Oxygen) domain works needs first clarifying how diatomic gases are relocated inside this sensing protein. A molecular-mechanics approach

H-NOX (Heme Nitric Oxide/Oxygen) domain has widespread occurrence, either standalone or associated with functional proteins, sending signals for functions that span from modulating vasodilation and neurotransmission with humans to competition and symbiosis with bacteria. Understanding how H-NOX works, and possibly intervening on degeneration for health purposes, needs first clarifying how diatomic gases are relocated through this protein in relation to the deeply buried heme. To this end, a biased form of molecular dynamics, i.e., Random Accelaration Molecular Dynamics (RAMD), is used by applying a randomly oriented tiny force to heme-dissociated CO of Nostoc sp. H-NOX, while changing randomly the direction of the force, if CO travels less than specified for the evaluated block. The result is that a large area of the protein, comprising amino acids from serine 44 to leucine 67 along two adjacent helices, offers a broad portal to CO from the surrounding medium to the deeply buried heme. Most traffic is concentrated through a channel lined by tyrosine 49, valine 52, and leucine 67. This modifies the picture drawn from mapping Xe cavities on pressurizing Nostoc sp. H-NOX with Xe gas. What is the main pathway with Xe-cavity mapping becomes a minor pathway with RAMD, and vice versa. The reason is that the fluctuating protein under MD creates clefts for CO slipping through, as it is expected to occur in nature.     

On the Permeation by Dioxygen of Urate Oxidase from Aspergillus flavus in Complex with Xanthine Anion: Dioxygen Pathways and a Portrait of the Enzyme Cavities from Molecular Dynamics Simulations in Water Solution

This work describes molecular dynamics (MD) simulations in aqueous media for the complex of the homotetrameric urate oxidase (UOX) from Aspergillus flavus with xanthine anion ( 5 ) in the presence of dioxygen (O₂). After 196.6 ns of trajectory from unrestrained MD, a O₂ molecule was observed leaving the bulk solvent to penetrate the enzyme between two subunits, A/C. From here, the same O₂ molecule was observed migrating, across subunit C, to the hydrophobic cavity that shares residue V227 with the active site. The latter was finally attained, after 378.3 ns of trajectory, with O₂ at a bonding distance from 5 . The reverse same O₂ pathway, from 5 to the bulk solvent, was observed as preferred pathway under random acceleration MD (RAMD), where an external, randomly oriented force was acting on O₂. Both MD and RAMD simulations revealed several cavities populated by O₂ during its migration from the bulk solvent to the active site or backwards. Paying attention to the last hydrophobic cavity that apparently serves as O₂ reservoir for the active site, it was noticed that its volume undergoes ample fluctuations during the MD simulation, as expected from the thermal motion of a flexible protein, independently from the particular subunit and no matter whether the cavity is filled or not by O₂.     

A permanent ion binding site located between two gates of the Shaker K+ channel.

K+ channels can be occupied by multiple permeant ions that appear to bind at discrete locations in the conduction pathway. Neither the molecular nature of the binding sites nor their relation to the activation or inactivation gates that control ion flow are well understood. We used the permeant ion Ba2+ as a K+ analog to probe for K+ ion binding sites and their relationship to the activation and inactivation gates. Our data are consistent with the existence of three single-file permeant-ion binding sites: one deep site, which binds Ba2+ with high affinity, and two more external sites whose occupancy influences Ba2+ movement to and from the deep site. All three sites are accessible to the external solution in channels with a closed activation gate, and the deep site lies between the activation gate and the C-type inactivation gate. We identify mutations in the P-region that disrupt two of the binding sites, as well as an energy barrier between the sites that may be part of the selectivity filter.     

Ligand Entry and Exit Pathways in the β2-Adrenergic Receptor

The recently determined crystal structure of the human β₂-adrenergic (β₂AR) G-protein-coupled receptor provides an excellent structural basis for exploring β₂AR-ligand binding and dissociation process. Based on this crystal structure, we simulated ligand exit from the β₂AR receptor by applying the random acceleration molecular dynamics (RAMD) simulation method. The simulation results showed that the extracellular opening on the receptor surface was the most frequently observed egress point (referred to as pathway A), and a few other pathways through interhelical clefts were also observed with significantly lower frequencies. In the egress trajectories along pathway A, the D192-K305 salt bridge between the extracellular loop 2 (ECL2) and the apex of the transmembrane helix 7 (TM7) was exclusively broken. The spatial occupancy maps of the ligand computed from the 100 RAMD simulation trajectories indicated that the receptor-ligand interactions that restrained the ligand in the binding pocket were the major resistance encountered by the ligand during exit and no second barrier was notable. We next performed RAMD simulations by using a putative ligand-free conformation of the receptor as input structure. This conformation was obtained in a standard molecular dynamics simulation in the absence of the ligand and it differed from the ligand-bound conformation in a hydrophobic patch bridging ECL2 and TM7 due to the rotation of F193 of ECL2. Results from the RAMD simulations with this putative ligand-free conformation suggest that the cleft formed by the hydrophobic bridge, TM2, TM3, and TM7 on the extracellular surface likely serves as a more specific ligand-entry site and the ECL2-TM7 hydrophobic junction can be partially interrupted upon the entry of ligand that pushes F193 to rotate, resulting in a conformation as observed in the ligand-bound crystal structure. These results may help in the design of β₂AR-targeting drugs with improved efficacy, as well as in understanding the receptor subtype selectivity of ligand binding in the β family of the adrenergic receptors that share almost identical ligand-binding pockets, but show notable amino acid sequence divergence in the putative ligand-entry site, including ECL2 and the extracellular end of TM7.     

Nuclear magnetic resonance‐based determination of dioxygen binding sites in protein cavities

The location and ligand accessibility of internal cavities in cysteine‐free wild‐type T4 lysozyme was investigated using O₂ gas‐pressure NMR spectroscopy and molecular dynamics (MD) simulation. Upon increasing the concentration of dissolved O₂ in solvent to 8.9 mM, O₂‐induced paramagnetic relaxation enhancements (PREs) to the backbone amide and side chain methyl protons were observed, specifically around two cavities in the C‐terminal domain. To determine the number of O₂ binding sites and their atomic coordinates from the 1/r⁶ distance dependence of the PREs, we established an analytical procedure using Akaike's Information Criterion, in combination with a grid‐search. Two O₂‐accessible sites were identified in internal cavities: One site was consistent with the xenon‐binding site in the protein in crystal, and the other site was established to be a novel ligand‐binding site. MD simulations performed at 10 and 100 mM O₂ revealed dioxygen ingress and egress as well as rotational and translational motions of O₂ in the cavities. It is therefore suggested that conformational fluctuations within the ground‐state ensemble transiently develop channels for O₂ association with the internal protein cavities.