The structural basis of lipid interactions in lipovitellin, a soluble lipoprotein.

BACKGROUND: The conformation and assembly of lipoproteins, protein containing large amounts of noncovalently bound lipid, is poorly understood. Lipoproteins present an unusual challenge as they often contain varying loads of lipid and are not readily crystallized. Lipovitellin is a large crystallizable oocyte protein of approximately 1300 residues that contains about 16% w/w lipid. Lipovitellin contains two large domains that appear to be conserved in both microsomal triglyceride transfer protein and apolipoprotein B-100. To gain insight into the conformation of a lipoprotein and the potential modes of binding of both neutral and phospholipid, the crystal structure of lamprey lipovitellin has been determined. RESULTS: We report here the refined crystal structure of lipovitellin at 2.8 A resolution. The structure contains 1129 amino acid residues located on five peptide chains, one 40-atom phosphatidylcholine, and one 13-atom hydrocarbon chain. The protein contains a funnel-shaped cavity formed primarily by two beta sheets and lined predominantly by hydrophobic residues. CONCLUSIONS: Using the crystal structure as a template, a model for the bound lipid is proposed. The lipid-binding cavity is formed primarily by a single-thickness beta-sheet structure which is stabilized by bound lipid. This cavity appears to be flexible, allowing lipid to be loaded or unloaded.

Small angle x-ray scattering from the inner and outer membranes from Escherichia coli.

Small angle X-ray data from purified forms of inner or cytoplasmic and outer membranes from Escherichia coli have been obtained and appear to be qualitatively similar. Transitory changes are apparent in the circularly averaged X-ray profiles from inner membranes. Such results could be due to the loss or denaturation of peripheral membrane proteins. Some partially dried forms of outer membrane are partly ordered and produce diffraction patterns which support an underlying bilayer structure. An extra light membrane fraction which results from membrane preparations utilizing a French pressure cell for spheroplast disruption has been characterized and shown to be similar to inner membrane. The purified membranes produce small angle X-ray diffraction patterns which are much different from those of lipid dispersions and the differences are attributable to the high protein content of the intact membranes. While the small angle X-ray region may be useful for characterizing the membrane preparations, the paucity of detail in the diffraction pattern suggest that it will be of little value in describing the complex underlying membrane structure.

Mercury complexes of thionicotinamide adenine dinucleotide.

One-to-one mercury complexes of thionicotinamide adenine dinucleotide (TNAD+) were prepared by using HgSO4 and Hg(CH3 COO-)2. Optical absorption spectroscopy indicated that the mercury probably binds to the TNAD+ through the thio-keto group on the pyridine ring. X-ray diffraction patterns of crystals of mitochondrial malate dehydrogenase soaked in solution containing TNAD+ . mercury complex indicated binding and the X-ray intensity differences are different from mercurials alone.

Crystal structure of a ternary complex of Escherichia coli malate dehydrogenase citrate and NAD at 1.9 A resolution.

The structure of malate dehydrogenase from Escherichia coli complexed with the substrate analog, citrate and the cofactor NAD, has been determined by X-ray crystallography. A monoclinic crystal of the malate dehydrogenase, grown in citrate buffer, was soaked in 10 mM NAD solution and found to be isomorphous with the apo-form. The X-ray data extended to 1.9 A, nearly the same resolution limit as the apo-enzyme crystals. The ternary complex of malate dehydrogenase has very few conformational differences from that of the pseudo binary complex of enzyme with bound citrate. In addition, the NAD molecule has a very similar conformation to the NAD as found in the crystal structure of the cytosolic eukaryotic malate dehydrogenase. Similar hydrogen bond interactions are made by both enzymes from polar groups belonging to the NAD. Such interactions include hydrogen bonds from the ribose oxygens and the phosphate oxygens, to backbone amide and carbonyl atoms of the protein and to side-chains of a select few conserved hydrophilic residues. The only notable difference occurs in the active site region where the nicotinamide moiety is obstructed from further entering the active site by the C-6 carbonyl atoms of citrate. In this position there are no direct polar interactions between the protein and the nicotinamide moiety. Energy minimization of the structure with malate substituted for citrate in the active site shows that the nicotinamide moiety assumes the same position in the active site as the NAD in cytosolic malate dehydrogenase. The carboxamide atoms of the energy minimized model make significant hydrogen bond interactions with the catalytic residue, H177, and with the main-chain atoms of I117 and V146 in the vicinity of the active site, while the position of the rest of the cofactor remains unchanged.

Crystal structures of holo and apo-cellular retinol-binding protein II.

Apo and holo-cellular retinol-binding protein II have been crystallized, and their crystal structures have been determined to 2.1 A and 1.9 A respectively. The apo and holo-crystals have different but related triclinic space groups. The X-ray phases for both structures were determined using the molecular replacement method. The crystal co-ordinates were refined to an R-factor of 0.200 for apo, and 0.173 for holo-cellular retinol-binding protein II. The holo and apo-models have nearly the same tertiary structures. Cellular retinol-binding protein II consists of a ten-stranded anti-parallel beta-barrel with the ligand binding cavity within the barrel. Two alpha-helices cover the open end of the beta-barrel making it almost solvent inaccessible. A single portal large enough to admit a water molecule was observed opening into the binding cavity. Exogenously added retinol was found within the cavity of each holo-cellular retinol-binding protein II molecule. Each retinol was surrounded by both polar and non-polar residues. The hydroxyl group of the bound retinol hydrogen bonds to the amide group of glutamine 108. The overall conformation of the bound retinol was derived from the four different molecules of holo-cellular retinol-binding protein II present in the triclinic form. The four copies of bound retinol had essentially the same conformation as found in crystalline retinaldehyde.

Purification and crystallization of fumarase C from Escherichia coli.

Fumarase C purified from Escherichia coli has been crystallized in the presence of polyethylene glycol in both a citrate buffer at pH 5.3 and a 3-(4-morpholino)-propanesulfonic acid buffer at pH 7.5 yielding two crystal forms. An orthorhombic C222(1) form was obtained in citrate at pH 5.3 and an orthorhombic I222 form was obtained in 3-(4-morpholino)-propanesulfonic acid (pH 7.5). Complete native data sets have been collected on both crystal forms: the C222(1) form is complete to 2.10 A and the I222 form is complete to 2.20 A.

Crystal structure of cellular retinoic acid binding protein I shows increased access to the binding cavity due to formation of an intermolecular beta-sheet.

A recombinant form of murine apo-cellular retinoic acid binding protein I (apo-CRABPI) has been purified and crystallized at pH 5.0, and the crystal structure has been refined to an R-factor of 19.6% at a resolution of 2.7 A. CRABPI binds all-trans retinoic acid and some retinoic acid metabolites with nanomolar affinities. Coordinates of the holo form of CRABP were not available during the early stages of the study, and in spite of numerous homologs of known structure, phases were not obtainable through molecular replacement. Instead, an interpretable electron density map was obtained by multiple isomorphous replacement methods after improvement of the heavy-atom parameters with density modified trial phases. Two molecules of apo-CRABPI occupy the P3121 asymmetric unit and are related by pseudo 2-fold rotational symmetry. Unique conformational differences are apparent between the two molecules. In all of the family members studied to date, there is a lack of hydrogen bonds between two of the component beta-strands resulting in a gap in the interstand hydrogen bonding pattern. In the crystallographic dimer described here, a continuous intermolecular beta-sheet is formed by using this gap region. This is possible because of an 8 A outward maximum displacement of the tight turn between the third and fourth beta-strands on one of the molecules. The result is a double beta-barrel containing two apo-CRABPI molecules with a more open, ligand-accessible binding cavity, which has not been observed in other structures of a family of proteins that bind hydrophobic ligands.

Crystal structure of rat intestinal fatty-acid-binding protein. Refinement and analysis of the Escherichia coli-derived protein with bound palmitate.

Rat intestinal fatty-acid-binding protein (I-FABP) is a small (15,124 Mr) cytoplasmic polypeptide that binds long-chain fatty acids in a non-covalent fashion. I-FABP is a member of a family of intracellular binding proteins that are thought to participate in the uptake, transport and/or metabolic targeting of hydrophobic ligands. The crystal structure of Escherichia coli-derived rat I-FABP with a single molecule of bound palmitate has been refined to 2 A resolution using a combination of least-squares methods, energy refinement and molecular dynamics. The combined methods resulted in a model with a crystallographic R-factor of 17.8% (7775 reflections, sigma greater than 2.0), root-mean-square bond length deviation of 0.009 A and root-mean-square bond angle deviation of 2.85 degrees. I-FABP contains ten antiparallel beta-strands organized into two approximately orthogonal, beta-sheets. The hydrocarbon tail of its single C16:0 ligand is present in a well-ordered, distinctively bent conformation. The carboxylate group of the fatty acid is located in the interior of I-FABP and forms a unique "quintet" of electrostatic interactions involving Arg106; Gln 115, and two solvent molecules. The hydrocarbon tail is bent with a slight left-handed helical twist from the carboxylate group to C-16. The bent methylene chain resides in a "cradle" formed by the side-chains of hydrophobic, mainly aromatic, amino acid residues. The refined molecular model of holo-I-FABP suggests several potential locations for entry and exiting of the fatty acid.

Crystal structure of Escherichia coli malate dehydrogenase. A complex of the apoenzyme and citrate at 1.87 A resolution.

The crystal structure of malate dehydrogenase from Escherichia coli has been determined with a resulting R-factor of 0.187 for X-ray data from 8.0 to 1.87 A. Molecular replacement, using the partially refined structure of porcine mitochondrial malate dehydrogenase as a probe, provided initial phases. The structure of this prokaryotic enzyme is closely homologous with the mitochondrial enzyme but somewhat less similar to cytosolic malate dehydrogenase from eukaryotes. However, all three enzymes are dimeric and form the subunit-subunit interface through similar surface regions. A citrate ion, found in the active site, helps define the residues involved in substrate binding and catalysis. Two arginine residues, R81 and R153, interacting with the citrate are believed to confer substrate specificity. The hydroxyl of the citrate is hydrogen-bonded to a histidine, H177, and similar interactions could be assigned to a bound malate or oxaloacetate. Histidine 177 is also hydrogen-bonded to an aspartate, D150, to form a classic His.Asp pair. Studies of the active site cavity indicate that the bound citrate would occupy part of the site needed for the coenzyme. In a model building study, the cofactor, NAD, was placed into the coenzyme site which exists when the citrate was converted to malate and crystallographic water molecules removed. This hypothetical model of a ternary complex was energy minimized for comparison with the structure of the binary complex of porcine cytosolic malate dehydrogenase. Many residues involved in cofactor binding in the minimized E. coli malate dehydrogenase structure are homologous to coenzyme binding residues in cytosolic malate dehydrogenase. In the energy minimized structure of the ternary complex, the C-4 atom of NAD is in van der Waals' contact with the C-3 atom of the malate. A catalytic cycle involves hydride transfer between these two atoms.

Preparation of single crystals of a yolk lipoprotein.

Single crystals of a yolk lipoprotein from two species of lamprey oocytes have been prepared and shown to belong to the monoclinic space group C2. The unit cell dimensions combined with molecular weight determinations support the idea that the lipoprotein is a symmetrical dimer centered on a crystallographic 2-fold symmetry axis. The lipid content is estimated to be about 15% (w/w) and the crystals are suitable for structural studies by single crystal X-ray analysis. The X-ray results are correlated with those obtained by electron diffraction.

Purification and crystallization of recombinant Escherichia coli malate dehydrogenase.

Malate dehydrogenase from Escherichia coli has been crystallized with polyethylene glycol and citrate buffer at pH 5.7. The enzyme was obtained from an E. coli strain in which the chromosomal malate dehydrogenase gene was contained on a pBR322 vector. Two types of crystals have been observed; a monoclinic C2 form and an orthorhombic C222(1) form, which is found infrequently. Monoclinic crystals were used as seeds in several rounds of crystallization until large crystals suitable for diffraction analysis were available. A complete X-ray data set to 2.0 A has been collected.

The structure of vitellogenin provides a molecular model for the assembly and secretion of atherogenic lipoproteins.

The assembly of atherogenic lipoproteins requires the formation in the endoplasmic reticulum of a complex between apolipoprotein (apo)B, a microsomal triglyceride transfer protein (MTP) and protein disulphide isomerase (PDI). Here we show by molecular modelling and mutagenesis that the globular amino-terminal regions of apoB and MTP are closely related in structure to the ancient egg yolk storage protein, vitellogenin (VTG). In the MTP complex, conserved structural motifs that form the reciprocal homodimerization interfaces in VTG are re-utilized by MTP to form a stable heterodimer with PDI, which anchors MTP at the site of apoB translocation, and to associate with apoB and initiate lipid transfer. The structural and functional evolution of the VTGs provides a unifying scheme for the invertebrate origins of the major vertebrate lipid transport system.

Engineering the quaternary structure of an enzyme: construction and analysis of a monomeric form of malate dehydrogenase from Escherichia coli.

The citric acid cycle enzyme, malate dehydrogenase (MDH), is a dimer of identical subunits. In the crystal structures of 2 prokaryotic and 2 eukaryotic forms, the subunit interface is conformationally homologous. To determine whether or not the quaternary structure of MDH is linked to the catalytic activity, mutant forms of the enzyme from Escherichia coli have been constructed. Utilizing the high-resolution structure of E. coli MDH, the dimer interface was analyzed critically for side chains that were spatially constricted and needed for electrostatic interactions. Two such residues were found, D45 and S226. At their nearest point in the homodimer, they are in different subunits, hydrogen bond across the interface, and do not interact with any catalytic residues. Each residue was mutated to a tyrosine, which should disrupt the interface because of its large size. All mutants were cloned and purified to homogeneity from an mdh- E. coli strain (BHB111). Gel filtration of the mutants show that D45Y and D45Y/S226Y are both monomers, whereas the S226Y mutant remains a dimer. The monomeric D45Y and D45Y/S226Y mutants have 14,000- and 17,500-fold less specific activity, respectively, than the native enzyme. The dimeric S226Y has only 1.4-fold less specific activity. All forms crystallized, indicating they were not random coils. Data have been collected to 2.8 A resolution for the D45Y mutant. The mutant is not isomorphous with the native protein and work is underway to solve the structure by molecular replacement.

A model for the regulation of D-3-phosphoglycerate dehydrogenase, a Vmax-type allosteric enzyme.

Escherichia coli D-3-phosphoglycerate dehydrogenase (ePGDH) is a tetramer of identical subunits that is allosterically inhibited by L-serine, the end product of its metabolic pathway. Because serine binding affects the velocity of the reaction and not the binding of substrate or cofactor, the enzyme is classified as of the Vmax type. Inhibition by a variety of amino acids and analogues of L-serine indicate that all three functional groups of serine are required for optimal interaction. Removing or altering any one functional group results in an increase in inhibitory concentration from micromolar to millimolar, and removal or alteration of any two functional groups removes all inhibitory ability. Kinetic studies indicate at least two serine-binding sites, but the crystal structure solved in the presence of bound serine and direct serine-binding studies show that there are a total of four serine-binding sites on the enzyme. However, approximately 85% inhibition is attained when only two sites are occupied. The three-dimensional structure of ePGDH shows that the serine-binding sites reside at the interface between regulatory domains of adjacent subunits. Two serine molecules bind at each of the two regulatory domain interfaces in the enzyme. When all four of the serines are bound, 100% inhibition of activity is seen. However, because the domain contacts are symmetrical, the binding of only one serine at each interface is sufficient to produce approximately 85% inhibition. The tethering of the regulatory domains to each other through multiple hydrogen bonds from serine to each subunit apparently prevents the body of these domains from undergoing the reorientation that must accompany a catalytic cycle. It is suggested that part of the conformational change may involve a hinge formed in the vicinity of the union of two antiparallel beta-sheets in the regulatory domains. The tethering function of serine, in turn, appears to prevent the substrate-binding domain from closing the cleft between it and the nucleotide-binding domain, which may be necessary to form a productive hydrophobic environment for hydride transfer. Thus, the structure provides a plausible model that is consistent with the binding and inhibition data and that suggests that catalysis and inhibition in this rare Vmax-type allosteric enzyme is based on the movement of rigid domains about flexible hinges.

Pig heart short chain L-3-hydroxyacyl-CoA dehydrogenase revisited: sequence analysis and crystal structure determination.

Short chain L-3-hydroxyacyl CoA dehydrogenase (SCHAD) is a soluble dimeric enzyme critical for oxidative metabolism of fatty acids. Its primary sequence has been reported to be conserved across numerous tissues and species with the notable exception of the pig heart homologue. Preliminary efforts to solve the crystal structure of the dimeric pig heart SCHAD suggested the unprecedented occurrence of three enzyme subunits within the asymmetric unit, a phenomenon that was thought to have hampered refinement of the initial chain tracing. The recently solved crystal coordinates of human heart SCHAD facilitated a molecular replacement solution to the pig heart SCHAD data. Refinement of the model, in conjunction with the nucleotide sequence for pig heart SCHAD determined in this paper, has demonstrated that the previously published pig heart SCHAD sequence was incorrect. Presented here are the corrected amino acid sequence and the high resolution crystal structure determined for pig heart SCHAD complexed with its NAD+ cofactor (2.8 A; R(cryst) = 22.4%, R(free) = 28.8%). In addition, the peculiar phenomenon of a dimeric enzyme crystallizing with three subunits contained in the asymmetric unit is described.