Computational solvent mapping reveals the importance of local conformational changes for broad substrate specificity in mammalian cytochromes P450

Computational solvent mapping moves small organic molecules as probes around a protein surface, finds favorable binding positions, clusters the conformations, and ranks the clusters on the basis of their average free energy. Prior mapping studies of enzymes, crystallized in either substrate-free or substrate-bound form, have shown that the largest number of solvent probe clusters invariably overlaps in the active site. We have applied this method to five cytochromes P450. As expected, the mapping of two bacterial P450s, P450 cam (CYP101) and P450 BM-3 (CYP102), identified the substrate-binding sites in both ligand-bound and ligand-free P450 structures. However, the mapping finds the active site only in the ligand-bound structures of the three mammalian P450s, 2C5, 2C9, and 2B4. Thus, despite the large cavities seen in the unbound structures of these enzymes, the features required for binding small molecules are formed only in the process of substrate binding. The ability of adjusting their binding sites to substrates that differ in size, shape, and polarity is likely to be responsible for the broad substrate specificity of these mammalian P450s. Similar behavior was seen at "hot spots" of protein-protein interfaces that can also bind small molecules in grooves created by induced fit. In addition, the binding of S-warfarin to P450 2C9 creates a high-affinity site for a second ligand, which may help to explain the prevalence of drug-drug interactions involving this and other mammalian P450s.     

Binding Hot Spots and Amantadine Orientation in the Influenza A Virus M2 Proton Channel

Structures of truncated versions of the influenza A virus M2 proton channel have been determined recently by x-ray crystallography in the open conformation of the channel, and by NMR in the closed state. The structures differ in the position of the bound inhibitors. The x-ray structure shows a single amantadine molecule in the middle of the channel, whereas in the NMR structure four drug molecules bind at the channel's outer surface. To study this controversy we applied computational solvent mapping, a technique developed for the identification of the most druggable binding hot spots of proteins. The method moves molecular probes—small organic molecules containing various functional groups—around the protein surface, finds favorable positions using empirical free energy functions, clusters the conformations, and ranks the clusters on the basis of the average free energy. The results of the mapping show that in both structures the primary hot spot is an internal cavity overlapping the amantadine binding site seen in the x-ray structure. However, both structures also have weaker hot spots at the exterior locations that bind rimantadine in the NMR structure, although these sites are partially due to the favorable interactions with the interfacial region of the lipid bilayer. As confirmed by docking calculations, the open channel binds amantadine at the more favorable internal site, in good agreement with the x-ray structure. In contrast, the NMR structure is based on a peptide/micelle construct that is able to accommodate the small molecular probes used for the mapping, but has a too narrow pore for the rimantadine to access the internal hot spot, and hence the drug can bind only at the exterior sites.     

Exploring the binding sites of the haloalkane dehalogenase DhlA from Xanthobacter autotrophicus GJ10

The catalytic site of haloalkane dehalogenase DhlA is buried more than 10 A from the protein surface. While potential access channels to this site have been reported, the precise mechanism of substrate import and product export is still unconfirmed. We used computational methods to examine surface pockets and their putative roles in ligand access to and from the catalytic site. Computational solvent mapping moves small organic molecule as probes over the protein surface in order to identify energetically favorable sites, that is, regions that tend to bind a variety of molecules. The mapping of three DhlA structures identifies seven such regions, some of which have been previously suggested to be involved in the binding and the import/export of substrates or products. These sites are the active site, the putative entrance of the channel leading to the active site, two pockets that bind Br- ions, a pocket in the slot region, and two additional sites between the main domain and the cap of DhlA. We also performed mapping and free energy analysis of the DhlA structures using the substrate, 1,2-dichloroethane, and halide ions as probes. The findings were compared to crystallographic data and to results obtained by CAVER, a program developed for finding routes from protein clefts and cavities to the surface. Solvent mapping precisely reproduced all three Br- binding sites identified by protein crystallography and the openings to four channels found by CAVER. The analyses suggest that (i) the active site has the highest affinity for the substrate molecule, (ii) the substrate initially binds at the entrance of the main tunnel, (iii) the site Br2, close to the entrance, is likely to serve as an intermediate binding site in product export, (iv) the site Br3, induced in the structure at high concentrations of Br-, could be part of an auxiliary route for product release, and (v) three of the identified sites are likely to be entrances of water-access channels leading to the active site. For comparison, we also mapped haloalkane dehalogenases DhaA and LinB, both of which contain significantly larger and more solvent accessible binding sites than DhlA. The mapping of DhaA and LinB places the majority of probes in the active site, but most of the other six regions consistently identified in DhlA were not observed, suggesting that the more open active site eliminates the need for intermediate binding sites for the collision complex seen in DhlA.     

Algorithms for computational solvent mapping of proteins

Computational mapping methods place molecular probes (small molecules or functional groups) on a protein surface to identify the most favorable binding positions by calculating an interaction potential. We have developed a novel computational mapping program called CS-Map (computational solvent mapping of proteins), which differs from earlier mapping methods in three respects: (i) it initially moves the ligands on the protein surface toward regions with favorable electrostatics and desolvation, (ii) the final scoring potential accounts for desolvation, and (iii) the docked ligand positions are clustered, and the clusters are ranked on the basis of their average free energies. To understand the relative importance of these factors, we developed alternative algorithms that use the DOCK and GRAMM programs for the initial search. Because of the availability of experimental solvent mapping data, lysozyme and thermolysin are considered as test proteins. Both DOCK and GRAMM speed up the initial search, and the combined algorithms yield acceptable mapping results. However, the DOCK-based approaches place the consensus site farther from its experimentally determined position than CS-Map, primarily because of the lack of a solvation term in the initial search. The GRAMM-based program also finds the correct consensus site for thermolysin. We conclude that good sampling is the most important requirement for successful mapping, but accounting for desolvation and clustering of ligand positions also help to reduce the number of false positives.     

FTMove: A Web Server for Detection and Analysis of Cryptic and Allosteric Binding Sites by Mapping Multiple Protein Structures

Protein mapping distributes many copies of different molecular probes on the surface of a target protein in order to determine binding hot spots, regions that are highly preferable for ligand binding. While mapping of X-ray structures by the FTMap server is inherently static, this limitation can be overcome by the simultaneous analysis of multiple structures of the protein. FTMove is an automated web server that implements this approach. From the input of a target protein, by PDB code, the server identifies all structures of the protein available in the PDB, runs mapping on them, and combines the results to form binding hot spots and binding sites. The user may also upload their own protein structures, bypassing the PDB search for similar structures. Output of the server consists of the consensus binding sites and the individual mapping results for each structure - including the number of probes located in each binding site, for each structure. This level of detail allows the users to investigate how the strength of a binding site relates to the protein conformation, other binding sites, and the presence of ligands or mutations. In addition, the structures are clustered on the basis of their binding properties. The use of FTMove is demonstrated by application to 22 proteins with known allosteric binding sites; the orthosteric and allosteric binding sites were identified in all but one case, and the sites were typically ranked among the top five. The FTMove server is publicly available at https://ftmove.bu.edu.     

Domain motion and interdomain hot spots in a multidomain enzyme

The aim of this article is to analyze conformational changes by comparing 10 different structures of Pseudomonas aeruginosa phosphomannomutase/phosphoglucomutase (PMM/PGM), a four‐domain enzyme in which both substrate binding and catalysis require substantial movement of the C‐terminal domain. We focus on changes in interdomain and active site crevices using a method called computational solvent mapping rather than superimposing the structures. The method places molecular probes (i.e., small organic molecules containing various functional groups) around the protein to find hot spots. One of the most important hot spots is in the active site, consistent with the ability of the enzyme to bind both glucose and mannose phosphosugar substrates. The protein has eight additional hot spots at domain‐domain interfaces and hinge regions. The locations and nature of six of these hot spots vary between the open, half‐open, and closed conformers of the enzyme, in good agreement with the ligand‐induced conformational changes. In the closed structures the number of probe clusters at the hinge region significantly depends on the position of the phosphorylated oxygen in the substrate (e.g., glucose 1‐phosphate versus glucose 6‐phosphate), but the protein remains almost unchanged in terms of the overall RMSD, indicating that computational solvent mapping is a more sensitive approach to detect changes in binding sites and interdomain crevices. Focusing on multidomain proteins we show that the subresolution conformational differences revealed by the mapping are in fact significant, and present a general statistical method of analysis to determine the significance of rigid body domain movements in X‐ray structures.     

Mechanism of reactant and product dissociation from the anthrax edema factor: a locally enhanced sampling and steered molecular dynamics study

The anthrax edema factor is a toxin overproducing damaging levels of cyclic adenosine monophosphate (cAMP) and pyrophosphate (PPi) from ATP. Here, mechanisms of dissociation of ATP and products (cAMP, PPi) from the active site are studied using locally enhanced sampling (LES) and steered molecular dynamics simulations. Various substrate conformations and ionic binding modes found in crystallographic structures are considered. LES simulations show that PPi and cAMP dissociate through different solvent accessible channels, while ATP dissociation requires significant active site exposure to solvent. The ionic content of the active site directly affects the dissociation of ATP and products. Only one ion dissociates along with ATP in the two-Mg(2+) binding site, suggesting that the other ion binds EF prior to ATP association. Dissociation of reaction products cAMP and PPi is impaired by direct electrostatic interactions between products and Mg(2+) ions. This provides an explanation for the inhibitory effect of high Mg(2+) concentrations on EF enzymatic activity. Breaking of electrostatic interactions is dependent on a competitive binding of water molecules to the ions, and thus on the solvent accessibility of the active site. Consequently, product dissociation seems to be a two-step process. First, ligands are progressively solvated while preserving the most important electrostatic interactions, in a process that is dependent on the flexibility of the active site. Second, breakage of the electrostatic bonds follows, and ligands diffuse into solvent. In agreement with this mechanism, product protonation facilitates dissociation.     

DRoP: Automated detection of conserved solvent-binding sites on proteins

Water and ligand binding play critical roles in the structure and function of proteins, yet their binding sites and significance are difficult to predict a priori. Multiple solvent crystal structures (MSCS) is a method where several X-ray crystal structures are solved, each in a unique solvent environment, with organic molecules that serve as probes of the protein surface for sites evolved to bind ligands, while the first hydration shell is essentially maintained. When superimposed, these structures contain a vast amount of information regarding hot spots of protein-protein or protein-ligand interactions, as well as conserved water-binding sites retained with the change in solvent properties. Optimized mining of this information requires reliable structural data and a consistent, objective analysis tool. Detection of related solvent positions (DRoP) was developed to automatically organize and rank the water or small organic molecule binding sites within a given set of structures. It is a flexible tool that can also be used in conserved water analysis given multiple structures of any protein independent of the MSCS method. The DRoP output is an HTML format list of the solvent sites ordered by conservation rank in its population within the set of structures, along with renumbered and recolored PDB files for visualization and facile analysis. Here, we present a previously unpublished set of MSCS structures of bovine pancreatic ribonuclease A (RNase A) and use it together with published structures to illustrate the capabilities of DRoP.     

Hydration energy landscape of the active site cavity in cytochrome P450cam

Hydration of protein cavities influences protein stability, dynamics, and function. Protein active sites usually contain water molecules that, upon ligand binding, are either displaced into bulk solvent or retained to mediate protein–ligand interactions. The contribution of water molecules to ligand binding must be accounted for to compute accurate values of binding affinities. This requires estimation of the extent of hydration of the binding site. However, it is often difficult to identify the water molecules involved in the binding process when ligands bind on the surface of a protein. Cytochrome P450cam is, therefore, an ideal model system because its substrate binds in a buried active site, displacing partially disordered solvent, and the protein is well characterized experimentally. We calculated the free energy differences for having five to eight water molecules in the active site cavity of the unliganded enzyme from molecular dynamics simulations by thermodynamic integration employing a three-stage perturbation scheme. The computed free energy differences between the hydration states are small (within 12 kJ mol⁻¹) but distinct. Consistent with the crystallographic determination and studies employing hydrostatic pressure, we calculated that, although ten water molecules could in principle occupy the volume of the active site, occupation by five to six water molecules is thermodynamically most favorable. Proteins 32:381–396, 1998. © 1998 Wiley-Liss, Inc.     

Multiple solvent crystal structures: probing binding sites, plasticity and hydration

Multiple solvent crystal structures (MSCS) of porcine pancreatic elastase were used to map the binding surface the enzyme. Crystal structures of elastase in neat acetonitrile, 95% acetone, 55% dimethylformamide, 80% 5-hexene-1,2-diol, 80% isopropanol, 80% ethanol and 40% trifluoroethanol showed that the organic solvent molecules clustered in the active site, were found mostly unclustered in crystal contacts and in general did not bind elsewhere on the surface of elastase. Mixtures of 40% benzene or 40% cyclohexane in 50% isopropanol and 10% water showed no bound benzene or cyclohexane molecules, but did reveal bound isopropanol. The clusters of organic solvent probe molecules coincide with pockets occupied by known inhibitors. MSCS also reveal the areas of plasticity within the elastase binding site and allow for the visualization of a nearly complete first hydration shell. The pattern of organic solvent clusters determined by MSCS for elastase is consistent with patterns for hot spots in protein-ligand interactions determined from database analysis in general. The MSCS method allows probing of hot spots, plasticity and hydration simultaneously, providing a powerful complementary strategy to guide computational methods currently in development for binding site determination, ligand docking and design.     

Exploring the binding site structure of the PPAR gamma ligand-binding domain by computational solvent mapping

Solvent mapping moves molecular probes, small organic molecules containing various functional groups, around the protein surface, finds favorable positions, clusters the conformations, and ranks the clusters based on the average free energy. Using at least six different solvents as probes, the probes cluster in major pockets of the functional site, providing detailed and reliable information on the amino acid residues that are important for ligand binding. Solvent mapping was applied to 12 structures of the peroxisome proliferator activated receptor gamma (PPARgamma) ligand-binding domain (LBD), including 2 structures without a ligand, 2 structures with a partial agonist, and 8 structures with a PPAR agonist bound. The analysis revealed 10 binding "hot spots", 4 in the ligand-binding pocket, 2 in the coactivator-binding region, 1 in the dimerization domain, 2 around the ligand entrance site, and 1 minor site without a known function. Mapping is a major source of information on the role and cooperativity of these sites. It shows that large portions of the ligand-binding site are already formed in the PPARgamma apostructure, but an important pocket near the AF-2 transactivation domain becomes accessible only in structures that are cocrystallized with strong agonists. Conformational changes were seen in several other sites, including one involved in the stabilization of the LBD and two others at the region of the coactivator binding. The number of probe clusters retained by these sites depends on the properties of the bound agonist, providing information on the origin of correlations between ligand and coactivator binding.     

Experimental and computational mapping of the binding surface of a crystalline protein

Multiple Solvent Crystal Structures (MSCS) is a crystallographic technique to identify energetically favorable positions and orientations of small organic molecules on the surface of proteins. We determined the high-resolution crystal structures of thermolysin (TLN), generated from crystals soaked in 50--70% acetone, 50--80% acetonitrile and 50 mM phenol. The structures of the protein in the aqueous-organic mixtures are essentially the same as the native enzyme and a number of solvent interaction sites were identified. The distribution of probe molecules shows clusters in the main specificity pocket of the active site and a buried subsite. Within the active site, we compared the experimentally determined solvent positions with predictions from two computational functional group mapping techniques, GRID and Multiple Copy Simultaneous Search (MCSS). The experimentally determined small molecule positions are consistent with the structures of known protein--ligand complexes of TLN.     

Substrate-induced conformational change of a coenzyme B12-dependent enzyme: crystal structure of the substrate-free form of diol dehydratase

Substrate binding triggers catalytic radical formation through the cobalt-carbon bond homolysis in coenzyme B12-dependent enzymes. We have determined the crystal structure of the substrate-free form of Klebsiella oxytoca diol dehydratase*cyanocobalamin complex at 1.85 A resolution. The structure contains two units of the heterotrimer consisting of alpha, beta, and gamma subunits. As compared with the structure of its substrate-bound form, the beta subunits are tilted by approximately 3 degrees and cobalamin is also tilted so that pyrrole rings A and D are significantly lifted up toward the substrate-binding site, whereas pyrrole rings B and C are only slightly lifted up. The structure revealed that the potassium ion in the substrate-binding site of the substrate-free enzyme is also heptacoordinated; that is, two oxygen atoms of two water molecules coordinate to it instead of the substrate hydroxyls. A modeling study in which the structures of both the cobalamin moiety and the adenine ring of the coenzyme were superimposed onto those of the enzyme-bound cyanocobalamin and the adenine ring-binding pocket, respectively, demonstrated that the distortions of the Co-C bond in the substrate-free form are already marked but slightly smaller than those in the substrate-bound form. It was thus strongly suggested that the Co-C bond becomes largely activated (labilized) when the coenzyme binds to the apoenzyme even in the absence of substrate and undergoes homolysis through the substrate-induced conformational changes of the enzyme. Kinetic coupling of Co-C bond homolysis with hydrogen abstraction from the substrate shifts the equilibrium to dissociation.     

The 1.8 Å Resolution Structure of Hevamine,a Plant Chitinase/Lysozyme,and Analysis of the Conserved Sequence and Structure Motifs of Glycosyl Hydrolase Family 18

The three-dimensional structure of hevamine, a plant enzyme with chitinase and lysozyme activity, has been refined at 1.8 Å resolution to an R-factor of 14.9% and a freeR-factor of 19.6%. The final model consists of all 273 amino acid residues and 206 ordered water molecules. Two non-prolinecis-peptides were identified, involving Phe32 and Trp255, both of which are implicated in substrate binding.Other glycosyl hydrolase family 18 proteins with known three-dimen sional structure are bacterial chitinase A, endo-β-N-acetylglucosaminidase F₁, endo-β-N-acetylglucosaminidase H, and the two plant proteins concanavalin B and narbonin, which have no known enzymatic activity. All these structures contain a (βα)₈barrel fold, with the two family 18 consensus regions roughly corresponding to the third and fourth barrel strands. This confirms the grouping of these proteins into family 18, which was only based on weak and local sequence similarity. The substrate specificity of the enzymes is determined by the loops following the barrel strands that form the substrate binding site. All enzymes have an aspartic acid and a glutamic acid residue in positions identical with Asp 125 and the catalytic Glu127 of hevamine. The lack of chitinase activity of concanavalin B and narbonin can be explained by the absence of one of these carboxylate groups, and by differences in the loops that form the substrate-binding cleft in hevamine.     

Identification of catalytic amino acids in the human GTP fucose pyrophosphorylase active site

GTP-l-fucose pyrophosphorylase(GFPP) catalyzes the reversible formation of the nucleotide-sugar GDP-beta-l-fucose from guanosine triphosphate and beta-l-fucose-1-phosphate. The enzyme functions primarily in the mammalian liver and kidney to salvage free fucose during the breakdown of glycoproteins and glycolipids. GFPP shares little primary sequence identity with other nucleotide-sugar metabolizing enzymes, and the three-dimensional structure of the protein is unknown. The enzyme does contain several sequences that could be nucleotide binding sites, but none of them are an exact match to consensus sequences. Using a combination of site-directed mutagenesis and UV photoaffinity cross-linking, we have identified five amino acid residues that are critical for catalysis. Some of these amino acids are found within the poorly conserved nucleotide binding consensus structures, while others represent new motifs. Two active site lysines can be cross-linked to photoaffinity probes. The site of cross-linking depends on the probe used. The identification of these critical residues highlights how distinct GFPP is from other nucleotide-sugar pyrophosphorylases.     

Differences in the interactions of liver alcohol dehydrogenases with probes binding into the substrate pocket

The interactions of three groups of probes (berberine alkaloids, tricyclic psychopharmaca and acridine derivatives) with isoenzymes of horse liver alcohol dehydrogenase and with rat liver alcohol dehydrogenase have been examined. These compounds inhibit the activity of the EE isoenzyme of horse liver alcohol dehydrogenase but differ in their behaviour towards the steroid-active enzymes (i.e. the ES isoenzyme of horse liver alcohol dehydrognase and alcohol dehydrogenase from rat liver): psychopharmaca inhibit, acridines activate and berberines do not bind. The ligands differ also in their influence on the modification of the EE isoenzyme by iodoacetate. Polarities (expressed as Kosower's Z values) of the respective binding sites on the EE isoenzyme were estimated from optical properties of bound probes. Berberines bind into a very hydrophobic area of the enzyme molecule, the binding site for psychopharmaca is moderately hydrophobic and that for acridines is rather polar. Steric arrangements of the binding sites are also discussed. The data presented confirm the existence of three distinct binding sites for these ligands in the substrate pocket of liver alcohol dehydrogenase.     

The crystal structure of polyhydroxybutyrate depolymerase from Penicillium funiculosum provides insights into the recognition and degradation of biopolyesters

Polyhydroxybutyrate is a microbial polyester that can be produced from renewable resources, and is degraded by the enzyme polyhydroxybutyrate depolymerase. The crystal structures of polyhydroxybutyrate depolymerase from Penicillium funiculosum and its S39 A mutant complexed with the methyl ester of a trimer substrate of (R)-3-hydroxybutyrate have been determined at resolutions of 1.71 A and 1.66 A, respectively. The enzyme is comprised of a single domain, which represents a circularly permuted variant of the alpha/beta hydrolase fold. The catalytic residues Ser39, Asp121, and His155 are located at topologically conserved positions. The main chain amide groups of Ser40 and Cys250 form an oxyanion hole. A crevice is formed on the surface of the enzyme, to which a single polymer chain can be bound by predominantly hydrophobic interactions with several hydrophobic residues. The structure of the S39A mutant-trimeric substrate complex reveals that Trp307 is responsible for the recognition of the ester group adjacent to the scissile group. It is also revealed that the substrate-binding site includes at least three, and possibly four, subsites for binding monomer units of polyester substrates. Thirteen hydrophobic residues, which are exposed to solvent, are aligned around the mouth of the crevice, forming a putative adsorption site for the polymer surface. These residues may contribute to the sufficient binding affinity of the enzyme for PHB granules without a distinct substrate-binding domain.     

Inactivated enzymes as probes of the structure of arabinoxylans as observed by atomic force microscopy

The complex structures of water-soluble wheat arabinoxylans have been mapped along individual molecules, and within populations, using the visualisation of the binding of inactivated enzymes by atomic force microscopy (AFM). It was demonstrated that site-directed mutagenesis (SDM) can be used to produce inactive enzymes as structural probes. For the SDM mutants AFM has been used to compare the binding of different xylanases to arabinoxylans. Xylanase mutant E386A, derived from the Xyn11A enzyme (Neocallimastrix patriciarium), was shown to bind randomly along arabinoxylan molecules. The xylanase binding was also monitored following Aspergillus niger arabinofuranosidase pre-treatment of samples. It was demonstrated that removal of arabinose side chains significantly altered the binding pattern of the inactivated enzyme. Xylanase mutant E246A, derived from the Xyn10A enzyme (Cellvibrio japonicus), was found to show deviations from random binding to the arabinoxylan chains. It is believed that this is due to the effect of a small residual catalytic activity of the enzyme that alters the binding pattern of the probe. Control procedures were developed and assessed to establish that the interactions between the modified xylanases and the arabinoxylans were specific interactions. The experimental data demonstrates the potential for using inactivated enzymes and AFM to probe the structural heterogeneity of individual polysaccharide molecules.     

Potent family-18 chitinase inhibitors: x-ray structures, affinities, and binding mechanisms

Six novel inhibitors of Vibrio harveyi chitinase A (VhChiA), a family-18 chitinase homolog, were identified by in vitro screening of a library of pharmacologically active compounds. Unlike the previously identified inhibitors that mimicked the reaction intermediates, crystallographic evidence from 14 VhChiA-inhibitor complexes showed that all of the inhibitor molecules occupied the outer part of the substrate-binding cleft at two hydrophobic areas. The interactions at the aglycone location are well defined and tightly associated with Trp-397 and Trp-275, whereas the interactions at the glycone location are patchy, indicating lower affinity and a loose interaction with two consensus residues, Trp-168 and Val-205. When Trp-275 was substituted with glycine (W275G), the binding affinity toward all of the inhibitors dramatically decreased, and in most structures two inhibitor molecules were found to stack against Trp-397 at the aglycone site. Such results indicate that hydrophobic interactions are important for binding of the newly identified inhibitors by the chitinase. X-ray data and isothermal microcalorimetry showed that the inhibitors occupied the active site of VhChiA in three different binding modes, including single-site binding, independent two-site binding, and sequential two-site binding. The inhibitory effect of dequalinium in the low nanomolar range makes this compound an extremely attractive lead compound for plausible development of therapeutics against human diseases involving chitinase-mediated pathologies.