Bacterial chemotaxis: a field in motion.

Many different types of studies are being combined to provide an increasingly detailed picture of the bacterial chemotaxis system. The structures of periplasmic receptors and a cytoplasmic response regulator, along with structures of domains of a membrane receptor, a receptor-modifying enzyme and a cytoplasmic histidine kinase, have been determined. These structures provide a basis for other work which is likely to open up new structural avenues.

Structure of Escherichia coli ribokinase in complex with ribose and dinucleotide determined to 1.8 A resolution: insights into a new family of kinase structures.

BACKGROUND: D-ribose must be phosphorylated at O5' before it can be used in either anabolism or catabolism. This reaction is catalysed by ribokinase and requires the presence of ATP and magnesium. Ribokinase is a member of a family of carbohydrate kinases of previously unknown structure. RESULTS: The crystal structure of ribokinase from Escherichia coli in complex with ribose and dinucleotide was determined at 1.84 A resolution by multiple isomorphous replacement. There is one 33 kDa monomer of ribokinase in the asymmetric unit but the protein forms a dimer around a crystallographic twofold axis. Each subunit consists of a central alpha/beta unit, with a new type of nucleotide-binding fold, and a distinct beta sheet that forms a lid over the ribose-binding site. Contact between subunits involves orthogonal packing of beta sheets, in a novel dimer interaction that we call a beta clasp. CONCLUSIONS: Inspection of the complex indicates that ribokinase utilises both a catalytic base for activation of the ribose in nucleophilic attack and an anion hole that stabilises the transition state during phosphoryl transfer. The structure suggests an ordered reaction mechanism, similar to those proposed for other carbohydrate kinases that probably involves conformational changes. We propose that the beta-clasp structure acts as a lid, closing and opening upon binding and release of ribose. From these observations, an understanding of the structure and catalytic mechanism of related sugar kinases can be obtained.

Structure of Aspergillus niger epoxide hydrolase at 1.8 A resolution: implications for the structure and function of the mammalian microsomal class of epoxide hydrolases.

BACKGROUND: Epoxide hydrolases have important roles in the defense of cells against potentially harmful epoxides. Conversion of epoxides into less toxic and more easily excreted diols is a universally successful strategy. A number of microorganisms employ the same chemistry to process epoxides for use as carbon sources. RESULTS: The X-ray structure of the epoxide hydrolase from Aspergillus niger was determined at 3.5 A resolution using the multiwavelength anomalous dispersion (MAD) method, and then refined at 1.8 A resolution. There is a dimer consisting of two 44 kDa subunits in the asymmetric unit. Each subunit consists of an alpha/beta hydrolase fold, and a primarily helical lid over the active site. The dimer interface includes lid-lid interactions as well as contributions from an N-terminal meander. The active site contains a classical catalytic triad, and two tyrosines and a glutamic acid residue that are likely to assist in catalysis. CONCLUSIONS: The Aspergillus enzyme provides the first structure of an epoxide hydrolase with strong relationships to the most important enzyme of human epoxide metabolism, the microsomal epoxide hydrolase. Differences in active-site residues, especially in components that assist in epoxide ring opening and hydrolysis of the enzyme-substrate intermediate, might explain why the fungal enzyme attains the greater speeds necessary for an effective metabolic enzyme. The N-terminal domain that is characteristic of microsomal epoxide hydrolases corresponds to a meander that is critical for dimer formation in the Aspergillus enzyme.

Regulation of RNA synthesis in Escherichia coli during a shift-up transition.

These experiments investigate two aspects of RNA synthesis in Escherichia coli ML30 during the transition from a relatively slow rate of growth to a more rapid one: (1) the number of growing RNA molecules per cell, and (2) the average time required for addition of a nucleotide onto a growing RNA chain. Cells were grown at 30 degrees C in a glucose-minimal salts medium and shifted-up by the addition of Casamino acids. Measurements were made of the rates of incorporation over short intervals (e.g. 5,8,12, and 16 s) of [3-H]guanine into the internal and 3'-terminal nucleotides of RNA. After correction for the specific activities of the intracellular GTP pools, and for the rate of [3-H]guanine accumulation at the 3'-terminus of non-growing RNA, the rates of chain elongation were calculated. It was found that cells growing at a rate of 0.9 generations/h contain approx. 4800 RNA molecules, growing at a rate of 28 nucleotides/s per chain. Cells growing exponentially at the postshift-up rate (1.2 generations/h) contain 7000 RNA molecules per unit equivalent cell mass, which are growing at a rate of 32 nucleotides/s per molecule. Three min after shift-up, cells contain the same number or slightly fewer (10%) growing RNA molecules than cells prior to shift-up, 4300, and these are being elongated at a rate of about 32 nucleotides per s. The results are consistent with the view that in the range of growth rates studied, the total rate of RNA synthesis is regulated through a limitation in the number of functioning RNA polymerase molecules, each working at a relatively constant, presumably maximal, average rate.

Ribose and glucose-galactose receptors. Competitors in bacterial chemotaxis.

The periplasmic ribose and glucose-galactose receptors (binding proteins) of Gram-negative bacteria compete for a common inner membrane receptor in bacterial chemotaxis, as well as being the essential primary receptors for their respective membrane transport systems. The high-resolution structures of the periplasmic receptors for ribose (from Escherichia coli) and glucose or galactose (from both Salmonella typhimurium and E. coli) are compared here to outline some features that may be important in their dual functions. The overall structure of each protein consists of two similar domains, both of which are made up of two non-contiguous segments of amino acid chain. Each domain is composed of a core of beta-sheet flanked on both sides with alpha-helices. The two domains are related to each other by an almost perfect intramolecular axis of symmetry. The ribose receptor is smaller as a result of a number of deletions in its sequence relative to the glucose-galactose receptor, mostly occurring in the loop regions; as a result, this protein is also more symmetrical. Many structural features, including some hydrophobic core interactions, a buried aspartate residue and several unusual turns, are conserved between the two proteins. The binding sites for ligand are in similar locations, and built along similar principles, although none of the specific interactions with the sugars is conserved. A comparison shows further that slightly different rotations relate the domains to each other in the three proteins, with the ribose receptor being the most closed, and the Salmonella glucose-galactose receptor the most open. The primary axis of relative rotation is almost perpendicular to that which describes the intramolecular symmetry in each case. These relative rotations of the domains are accompanied by the sliding of some helices as the structures adjust themselves to relieve strain. The hinges which are responsible for most of these relative domain rotations are very similar in the three proteins, consisting of a symmetrical arrangement of beta-strands and alpha-helices and two conserved water molecules that are critical to the hydrogen bonding in the important interdomain region. A region of high sequence and structural similarity between the ribose and glucose-galactose receptors is also located around the intramolecular symmetry axis, on the opposite side of the proteins from the hinge region. This region is that which is altered most by the relative rotations, and is the location of most of the known mutations which affect chemotaxis and transport in the ribose receptor.

1.7 A X-ray structure of the periplasmic ribose receptor from Escherichia coli.

The X-ray structure of the periplasmic ribose receptor (binding protein) of Escherichia coli (RBP) was solved at 3 A resolution by the method of multiple isomorphous replacement. Alternating cycles of refitting and refinement have resulted in a model structure with an R-factor of 18.7% for 27,526 reflections from 7.5 to 1.7 A resolution (96% of the data). The model contains 2228 non-hydrogen atoms, including all 271 residues of the amino acid sequence, 220 solvent atoms and beta-D-ribose. The protein consists of two highly similar structural domains, each of which is composed of a core of parallel beta-sheet flanked on both sides by alpha-helices. The two domains are related to each other by an almost perfect 2-fold axis of rotation, with the C termini of the beta-strands of each sheet pointing toward the center of the molecule. Three short stretches of amino acid chain (from symmetrically related portions of the protein) link these two domains, and presumably act as a hinge to allow relative movement of the domains in functionally important conformational changes. Two water molecules are also an intrinsic part of the hinge, allowing crucial flexibility in the structure. The ligand beta-D-ribose (in the pyranose form) is bound between the domains, held by interactions with side-chains of the interior loops. The binding site is precisely tailored, with a combination of hydrogen bonding, hydrophobic and steric effects giving rise to tight binding (0.1 microM for ribose) and high specificity. Four out of seven binding-site residues are charged (2 each of aspartate and arginine) and contribute two hydrogen bonds each. The remaining hydrogen bonds are contributed by asparagine and glutamine residues. Three phenylalanine residues supply the hydrophobic component, packing against both faces of the sugar molecule. The arrangement of these hydrogen bonding and hydrophobic residues results in an enclosed binding site with the exact shape of the allowed sugar molecules; in the process of binding, the ligand loses all of its surface-accessible area. The sites of two mutations that affect the rate of folding of the ribose receptor are shown to be located near small cavities in the wild-type protein. The cavities thus allow the incorporation of the larger residues in the mutant proteins. Since these alterations would seriously affect the ability of the protein to build the first portion of the hydrophobic core in the first domain, it is proposed that this process is the rate-limiting step in folding of the ribose receptor.

Multiple open forms of ribose-binding protein trace the path of its conformational change.

Conformational changes are necessary for the function of bacterial periplasmic receptors in chemotaxis and transport. Such changes allow entry and exit of ligand, and enable the correct interaction of the ligand-bound proteins with the membrane components of each system. Three open, ligand-free forms of the Escherichia coli ribose-binding protein were observed here by X-ray crystallographic studies. They are opened by 43 degrees, 50 degrees and 64 degrees with respect to the ligand-bound protein reported previously. The three open forms are not distinct, but show a clear relationship to each other. All are the product of a similar opening motion, and are stabilized by a new, almost identical packing interface between the domains. The changes are generated by similar bond rotations, although some differences in the three hinge segments are needed to accommodate the various structural scenarios. Some local repacking also occurs as interdomain contacts are lost. The least open (43 degrees) form is probably the dominant one in solution under normal conditions, although a mixture of species seems likely. The open and closed forms have distinct surfaces in the regions known to be important in chemotaxis and transport, which will differentiate their interactions with the membrane components. It seems certain that the conformational path that links the forms described here is that followed during ligand retrieval, and in ligand release into the membrane-bound permease system.

Conformational changes of ribose-binding protein and two related repressors are tailored to fit the functional need.

The structures and conformational changes of the periplasmic ribose-binding protein and two repressors, PurR and LacI, were compared. Although the closed, ligand-bound structures of the three proteins are very similar, they differ greatly in the degree and direction in which they open, as well as in the amount of internal rearrangement within the domains during that process. Water molecules and a relatively symmetrical inter-domain connection region assist in the large opening observed for the binding protein, while the design of the repressors appears to preclude such dramatic movements. The dimeric nature of the latter proteins, an important aspect in their binding of pseudo-symmetrical DNA sequences, also appears to be a determinant in the allowed motion. Slight differences in the structure of PurR and LacI explain how they can converge to a similar DNA-binding state in response to different binding states of their small molecule effector.

Strange bedfellows: interactions between acidic side-chains in proteins.

The oxygen atoms of two acidic side-chains are frequently found within hydrogen-bonding distance of each other in proteins. Two distinct types of cases are common. In metal-binding sites, the oxygen atoms are brought near (average closest approach 3.0 A) by their common role as metal ligands. In a different location, either buried or on the protein surface, the two acidic groups can share a proton. The corresponding O-O distances in the latter case are shorter (usually 2.7 or less), and the geometry is typical of hydrogen-bonding interactions. The glucose/galactose-binding protein of Salmonella typhimurium provides an example of a well-ordered Asp-Glu pair on the surface of a protein with a very short O-O distance, at a pH of 7.0. Other instances have been found at pH values as high as 8.0, suggesting substantial alteration of the pKa involved. These observations have implications for the study of enzymes that use pairs of acidic residues in binding and catalysts.

Planar stacking interactions of arginine and aromatic side-chains in proteins.

A parallel stacking arrangement of the guanidinium groups of arginines directly over the center of the rings of aromatic side-chains is observed much more frequently in proteins than would be expected by chance. This type of interaction, which is often found in locations critical to the function, apparently serves to orient the arginine side-chain without interfering with its ability to form hydrogen bonds elsewhere. It is distinct from the interactions which involve the side-chains of asparagine or glutamine, which do frequently assume a nearly planar relationship to the ring, but at a position at or beyond the ring edge.

Preliminary X-ray data for the periplasmic ribose receptor from Escherichia coli.

Crystals of the periplasmic ribose receptor for chemotaxis and transport in Escherichia coli have been examined by X-ray analysis. The crystals grow as elongated rectangular prisms with the symmetry of the orthorhombic space group P2(1)2(1)2(1). The unit cell dimensions are a = 74.6 A, b = 88.8 A and c = 40.1 A. There is one molecule of molecular weight 28,500 per asymmetric unit.

Crystallization and preliminary X-ray analysis of the periplasmic dipeptide binding protein from Escherichia coli.

The periplasmic dipeptide-binding protein from Escherichia coli has been purified, freed of bound endogenous ligands, and crystallized. Crystals of the protein in complex with added dipeptides have been subjected to X-ray analysis. The crystals grow as hexagonal bipyramids or eye-shaped disks which have the symmetry of space group P6(1). The unit cell dimensions are a = b = 183 A, c = 212 A, and the diffraction pattern extends to 3.2 A resolution with a conventional X-ray source.

The 1.7 A refined X-ray structure of the periplasmic glucose/galactose receptor from Salmonella typhimurium.

The X-ray structure of the periplasmic glucose/galactose receptor (binding protein) of Salmonella typhimurium (GBP-S) has been refined at 1.7 A resolution with an R-factor of 19.0%. The model contains all 309 residues of the amino acid sequence, 153 water molecules, a calcium ion and beta-D-galactose. The protein consists of two very similar structural domains, each of which is composed a core of parallel beta-sheet flanked on both sides by alpha-helices. Three short stretches of amino acid chain (from symmetrically related portions of the structure) link the domains, and presumably act as a hinge to allow their relative movement in functionally important conformational changes. Galactose is bound between the domains, interacting with a number of side-chains from the loops lining the binding cleft. A combination of hydrogen bonding, hydrophobic and steric effects give rise to tight binding (dissociation constant 0.2 microM) and high specificity. Of nine hydrogen bonding groups, three are aspartate, three asparagine, one histidine (unprotonated), one arginine and one water, contributing 13 hydrogen bonds in total. Additional residues pack against (primarily) non-polar faces of the sugar molecule. The precise arrangement of the hydrogen bonding and hydrophobic residues results in an enclosed binding site with a shape that is a composite of those of the allowed sugar molecules. It is presumed that ligands bind to a more open form of the receptor that then closes by rotation in the hinge. Comparison with the GBP-S structure solved earlier in complex with glucose shows no significant changes, even for the aspartate residue most directly involved with the different sugars. Comparison with the galactose/glucose receptor of Escherichia coli indicates that these two proteins are very similar in overall structure, with the main difference being a 2 to 3 degrees rotation in the hinge. This difference appears to be the result of different crystal packing for the two proteins; it is likely that both conformations are normally found in solution.

Conformational changes of three periplasmic receptors for bacterial chemotaxis and transport: the maltose-, glucose/galactose- and ribose-binding proteins.

Small-angle X-ray scattering experiments were carried out for the maltose-, glucose/galactose- and ribose-binding proteins of Gram negative bacteria. All were shown to be monomers that decrease in radius of gyration on ligand binding. The results obtained for the maltose-binding protein agree well with crystal structures of the closed, ligand-bound, and open, ligand-free protein, suggesting that these are indeed the primary forms in solution. The closed form is stabilized by protein-sugar interactions, while the open conformation is stabilized by close contacts between the two domains. Since it is the proper special relationship of the domains in the closed form that is most important for interaction with chemotaxis and transport partners, the stabilization of the open form would help keep ligand-free molecules from interfering in function. The scattering results also provide evidence that a large conformational change takes place in association with ligand binding to the glucose/galactose- and ribose-binding proteins, and that the two changes are similar. Modeling suggests that the open forms resemble those found in the related leucine and leucine/isoleucine/valine-binding proteins, but are different from those observed for the maltose-binding protein and the related purine repressor.

Crystal structures and solution conformations of a dominant-negative mutant of Escherichia coli maltose-binding protein.

A mutant of the periplasmic maltose-binding protein (MBP) with altered transport properties was studied. A change of residue 230 from tryptophan to arginine results in dominant-negative MBP: expression of this protein against a wild-type background causes inhibition of maltose transport. As part of an investigation of the mechanism of such inhibition, we have solved crystal structures of both unliganded and liganded mutant protein. In the closed, liganded conformation, the side-chain of R230 projects into a region of the surface of MBP that has been identified as important for transport while in the open form, the same side-chain takes on a different, and less ordered, conformation. The crystallographic work is supplemented with a small-angle X-ray scattering study that provides evidence that the solution conformation of unliganded mutant is similar to that of wild-type MBP. It is concluded that dominant-negative inhibition of maltose transport must result from the formation of a non-productive complex between liganded-bound mutant MBP and wild-type MalFGK2. A general kinetic framework for transport by either wild-type MalFGK2 or MBP-independent MalFGK2 is used to understand the effects of dominant-negative MBP molecules on both of these systems.

Induced fit on sugar binding activates ribokinase.

The enzyme ribokinase phosphorylates ribose at O5* as the first step in its metabolism. The original X-ray structure of Escherichia coli ribokinase represented the ternary complex including ribose and ADP. Structures are presented here for the apo enzyme, as well as the ribose-bound state and four new ternary complex forms. Combined, the structures suggest that large and small conformational changes play critical roles in the function of this kinase. An initially open apo form can allow entry of the ribose substrate. After ribose binding, the active site lid is observed in a closed conformation, with the sugar trapped underneath. This closure and associated changes in the protein appear to assist ribokinase in recognition of the co-substrate ATP as the next step. Binding of the nucleotide brings about further, less dramatic adjustments in the enzyme structure. Additional small movements are almost certainly required during the phosphoryltransfer reaction. Evidence is presented that some types of movements of the lid are allowed in the ternary complex, which may be critical to the creation and breakdown of the transition state. Similar events are likely to take place during catalysis by other related carbohydrate kinases, including adenosine kinase.

Structure of D-allose binding protein from Escherichia coli bound to D-allose at 1.8 A resolution.

ABC transport systems for import or export of nutrients and other substances across the cell membrane are widely distributed in nature. In most bacterial systems, a periplasmic component is the primary determinant of specificity of the transport complex as a whole. We report here the crystal structure of the periplasmic binding protein for the allose system (ALBP) from Escherichia coli, solved at 1.8 A resolution using the molecular replacement method. As in the other members of the family (especially the ribose binding protein, RBP, with which it shares 35 % sequence homology), this structure consists of two similar domains joined by a three-stranded hinge region. The protein is believed to exist in a dynamic equilibrium of closed and open conformations in solution which is an important part of its function. In the closed ligand-bound form observed here, D-allose is buried at the domain interface. Only the beta-anomer of allopyranose is seen in the crystal structure, although the alpha-anomer can potentially bind with a similar affinity. Details of the ligand-binding cleft reveal the features that determine substrate specificity. Extensive hydrogen bonding as well as hydrophobic interactions are found to be important. Altogether ten residues from both the domains form 14 hydrogen bonds with the sugar. In addition, three aromatic rings, one from each domain with faces parallel to the plane of the sugar ring and a third perpendicular, make up a hydrophobic stacking surface for the ring hydrogen atoms. Our results indicate that the aromatic rings forming the sugar binding cleft can sterically block the binding of any hexose epimer except D-allose, 6-deoxy-allose or 3-deoxy-glucose; the latter two are expected to bind with reduced affinity, due to the loss of some hydrogen bonds. The pyranose form of the pentose, D-ribose, can also fit into the ALBP binding cleft, although with lower binding affinity. Thus, ALBP can function as a low affinity transporter for D-ribose. The significance of these results is discussed in the context of the function of allose and ribose transport systems.

Family 7 cellobiohydrolases from Phanerochaete chrysosporium: crystal structure of the catalytic module of Cel7D (CBH58) at 1.32 A resolution and homology models of the isozymes.

Cellobiohydrolase 58 (Cel7D) is the major cellulase produced by the white-rot fungus Phanerochaete chrysosporium, constituting approximately 10 % of the total secreted protein in liquid culture on cellulose. The enzyme is classified into family 7 of the glycosyl hydrolases, together with cellobiohydrolase I (Cel7A) and endoglucanase I (Cel7B) from Trichoderma reesei. Like those enzymes, it catalyses cellulose hydrolysis with net retention of the anomeric carbon configuration. The structure of the catalytic module (431 residues) of Cel7D was determined at 3.0 A resolution using the structure of Cel7A from T. reesei as a search model in molecular replacement, and ultimately refined at 1.32 A resolution. The core structure is a beta-sandwich composed of two large and mainly antiparallel beta-sheets packed onto each other. A long cellulose-binding groove is formed by loops on one face of the sandwich. The catalytic residues are conserved and the mechanism is expected to be the same as for other family members. The Phanerochaete Cel7D binding site is more open than that of the T. reesei cellobiohydrolase, as a result of deletions and other changes in the loop regions, which may explain observed differences in catalytic properties. The binding site is not, however, as open as the groove of the corresponding endoglucanase. A tyrosine residue at the entrance of the tunnel may be part of an additional subsite not present in the T. reesei cellobiohydrolase. The Cel7D structure was used to model the products of the five other family 7 genes found in P. chrysosporium. The results suggest that at least two of these will have differences in specificity and possibly catalytic mechanism, thus offering some explanation for the presence of Cel7 isozymes in this species, which are differentially expressed in response to various growth conditions.

C alpha-based torsion angles: a simple tool to analyze protein conformational changes.

A simple method is presented for the analysis of protein conformational changes based on the comparison of torsion angles defined by four consecutive C alpha atoms. The technique was applied successfully to proteins that undergo hinge motion and shear motion. In the case of both MBP and LAO, which represent examples of hinge motion, the plot of the differences in C alpha-torsion angles between the open and closed forms of the proteins helped us to formulate a more thorough description of the conformational change: a large displacement of one domain with respect to the other where one of the domains does not behave like a rigid body but exhibits some degree of flexibility. The analysis of citrate synthase, which is an example of shear motion, shows that the largest differences in C alpha-torsion angles between the open and closed conformations are clustered around residues that belong to segments connecting alpha-helices, whereas the helices themselves appear to be rigid; this is in agreement with previous results obtained by detailed least-squares superpositions (Lesk AM, Chothia C, 1984, J Mol Biol 174:175-191).

Simple models for the analysis of binding protein-dependent transport systems.

Mathematical modeling was used to evaluate experimental data for bacterial binding protein-dependent transport systems. Two simple models were considered in which ligand-free periplasmic binding protein interacts with the membrane-bound components of transport. In one, this interaction was viewed as a competition with the ligand-bound binding protein, whereas in the other, it was considered to be a consequence of the complexes formed during the transport process itself. Two sets of kinetic parameters were derived for each model that fit the available experimental results for the maltose system. By contrast, a model that omitted the interaction of ligand-free binding protein did not fit the experimental data. Some applications of the successful models for the interpretation of existing mutant data are illustrated, as well as the possibilities of using mutant data to test the original models and sets of kinetic parameters. Practical suggestions are given for further experimental design.