The 9 A projection structure of cytochrome b6f complex determined by electron crystallography

Thin three-dimensional crystals of the cytochrome b6 f complex from the unicellular algae Chlamydomonas reinhardtii have been grown by BioBeads-mediated detergent removal from a mixture of protein and lipid solubilized in Hecameg. Frozen-hydrated crystals, exhibiting p22121 plane group symmetry, were studied by electron crystallography and a projection map at 9 A resolution was calculated. The crystals (unit cell dimensions of a=173.5 A, b=70.0 A and gamma=90.0 degrees) showed the presence of dimers, and within each monomer 14 domains of electron density were observed. The combination of the projection map obtained from ice-embedded crystals of cytochrome b6 f with a previous map obtained from negatively stained samples brings new insight in the organization of the complex. For example, it distinguishes some peaks and/or domains that are only extramembrane or transmembrane, and reveals the possible localization of single-stranded transmembrane alpha-helices (Pet subunits). Furthermore, the cross-correlation of our projection map from frozen hydrated samples with the atomic model of the transmembrane part of the cytochrome bc1 complex has allowed us to localize the cytochrome b6 at the dimer interface and to reveal structural differences between the two complexes.     

Three-dimensional structure of the surface layer protein of Aquaspirillum serpens VHA determined by electron crystallography.

The three-dimensional structure of the protein which forms the S layer of Aquaspirillum serpens strain VHA has been determined by electron microscopy. Structures have been reconstructed to a resolution of about 1.6 nm for single-layered specimens and about 4 nm for two-layered specimens. The structure, which has hexagonal symmetry, consists of a core in the shape of a cup, with six projections arising from the rim of the cup to join adjacent subunits at the threefold symmetry axes. The model is consistent with edge views of the S layer which have been obtained in this and other work. It is now clear from this work and from three-dimensional reconstructions of other bacterial S layers that a wide diversity exists in the morphology of surface layers.     

Three-dimensional structural analysis of tetanus toxin by electron crystallography

Two-dimensional crystalline arrays of native tetanus toxin have been formed at the interface between a solution of the toxin and a phospholipid monolayer containing a ganglioside. Electron crystallographic analysis has been used to study these periodic arrays. The arrays obey the symmetry of plane group p12(1), with a = 126 A and b = 84 A, and a thickness of 90 A (1 A = 0.1 nm). The three-dimensional structure of tetanus toxin in negative stain is reconstructed to a nominal resolution of 14 A from multiple tilt images. The molecule presents an asymmetric three-lobed structure and could interact with the monolayer in two possible orientations.     

Three-dimensional fold of the human AQP1 water channel determined at 4 A resolution by electron crystallography of two-dimensional crystals embedded in ice

Here, we present a three-dimensional (3D) density map of deglycosylated, human erythrocyte aquaporin 1 (AQP1) determined at 4 A resolution in plane and approximately 7 A resolution perpendicular to the bilayer. The map was calculated by analyzing images and electron diffraction patterns recorded from tilted (up to 60 degrees ), ice-embedded, frozen-hydrated 2D crystals of AQP1 in lipid bilayer membranes. This map significantly extends the findings related to the folding of the AQP1 polypeptide chain determined by us at a lower, 7 A by approximately 20 A, resolution. The solvent-accessible volume within a monomer has a vestibular architecture, with a narrow, approximately 6.5 A diameter constriction near the center of the bilayer, where the location of the water-selective channel is postulated to exist. The clearly resolved densities for the transmembrane helices display the protrusions expected for bulky side-chains. The density in the interior of the helix barrel (putative NPA box region) is better resolved compared to our previous map, suggesting clearer linkage to some of the helices, and it may harbor short stretches of alpha-helix. At the bilayer extremities, densities for some of the inter-helix hydrophilic loops are visible. Consistent with these observed inter-helix connections, possible models for the threading of the AQP1 polypeptide chain are presented. A preferred model is deduced that agrees with the putative locations of a group of aromatic residues in the amino acid sequence and in the 3D density map.     

Membrane protein structure determination by electron crystallography

During the past year, electron crystallography of membrane proteins has provided structural insights into the mechanism of several different transporters and into their interactions with lipid molecules within the bilayer. From a technical perspective there have been important advances in high-throughput screening of crystallization trials and in automated imaging of membrane crystals with the electron microscope. There have also been key developments in software, and in molecular replacement and phase extension methods designed to facilitate the process of structure determination.     

The structure of membrane associated proteins in eicosanoid and glutathione metabolism as determined by electron crystallography

Membrane associated proteins in eicosanoid and glutathione metabolism (MAPEG) are involved in biosynthesis of arachidonic-derived mediators of pain, fever, and inflammation as well as in biotransformation and detoxification of electrophilic substances. Structure determination of microsomal glutathione transferase 1 using electron crystallography has provided the first atomic model of an MAPEG member. The homotrimer consists of three repeats of a four-helix transmembrane bundle with the largest extramembranous domain connecting the first and second helix and with a short proline rich loop on the same side between helices three and four. Residues of importance for intramolecular or intermolecular contacts as well as for stabilizing the active site have been identified and the results can be applied for interpreting structure-function relationship for similar MAPEG members.     

Primary structure of cholera toxin B-subunit

The primary structure of cholera toxin B-subunit, responsible for the binding of the toxin to cell surfaces, has been elucidated. The polypeptide contains 103 amino acid residues and one intra-chain disulfide bridge between Cys 9 and Cys 86. The molecular weight is calculated to be 11,637, 15–20% higher than the values estimated by physicochemical methods. This value is consistent with a structure containing five moles of B-subunits per mole of cholera toxin.     

Crystal structure of cholera toxin B-pentamer bound to receptor GM1 pentasaccharide.

Cholera toxin (CT) is an AB5 hexameric protein responsible for the symptoms produced by Vibrio cholerae infection. In the first step of cell intoxication, the B-pentamer of the toxin binds specifically to the branched pentasaccharide moiety of ganglioside GM1 on the surface of target human intestinal epithelial cells. We present here the crystal structure of the cholera toxin B-pentamer complexed with the GM1 pentasaccharide. Each receptor binding site on the toxin is found to lie primarily within a single B-subunit, with a single solvent-mediated hydrogen bond from residue Gly 33 of an adjacent subunit. The large majority of interactions between the receptor and the toxin involve the 2 terminal sugars of GM1, galactose and sialic acid, with a smaller contribution from the N-acetyl galactosamine residue. The binding of GM1 to cholera toxin thus resembles a 2-fingered grip: the Gal(beta 1-3)GalNAc moiety representing the "forefinger" and the sialic acid representing the "thumb." The residues forming the binding site are conserved between cholera toxin and the homologous heat-labile enterotoxin from Escherichia coli, with the sole exception of His 13. Some reported differences in the binding affinity of the 2 toxins for gangliosides other than GM1 may be rationalized by sequence differences at this residue. The CTB5:GM1 pentasaccharide complex described here provides a detailed view of a protein:ganglioside specific binding interaction, and as such is of interest not only for understanding cholera pathogenesis and for the design of drugs and development of vaccines but also for modeling other protein:ganglioside interactions such as those involved in GM1-mediated signal transduction.     

The 3-D structure of microsomal glutathione transferase 1 at 6 A resolution as determined by electron crystallography of p22(1)2(1) crystals

Pure solubilised microsomal glutathione transferase 1 (MGST1) forms well-ordered two-dimensional (2-D) crystals of two different symmetries, one orthorhombic (p22(1)2(1)) and one hexagonal (p6), both diffracting electrons to a resolution beyond 3 A. A three-dimensional (3-D) map has previously been calculated to 6 A resolution from the hexagonal crystal form. From orthorhombic crystals we have now calculated a 6 A 3-D reconstruction displaying three repeats of four rod-like densities. These are inclined relative to the normal of the membrane plane and consistent with arising from a left-handed four-helix bundle fold. The rendered volume clearly displays the same structural features as the map previously calculated from the p6 crystal type including similar lengths and substructure of the helices, but several distinguishing features do exist. The helices are more tilted in the map calculated from the orthorhombic crystals indicating conformational flexibility. Density present on the cytosolic side is consistent with the location of the active site. In addition, the current map displays the noted similarity to subunit I of cytochrome c oxidase.     

Solution structure of alpha-conotoxin ImI determined by two-dimensional NMR spectroscopy.

The three-dimensional structure of alpha-conotoxin ImI, a potent antagonist targeting the neuronal alpha7 subtype of nicotinic acetylcholine receptor (nAChR), has been investigated by NMR spectroscopy. On the basis of 181 experimental constraints, a total of 25 converged structures were obtained. The average pairwise atomic root mean square difference is 0.40+/-0.11 A for the backbone atoms. The resulting structure indicates the presence of two successive type I beta-turns and a 310 helix for residues Cys2-Cys8 and Ala9-Arg11, respectively, and shows a significant structural similarity to that of alpha-conotoxin PnIA, which is also selective for the neuronal nAChR.     

Comparison of the structures of three carboxypeptidase A-phosphonate complexes determined by X-ray crystallography

The structures of the complexes of carboxypeptidase A (CPA) with two tight-binding phosphonate inhibitors have been determined by X-ray crystallography. The inhibitors, Cbz-Phe-ValP-(O)-Phe[ZFVP(O)F] and Cbz-Ala-GlyP-(O)-Phe[ZAGP(O)F], bind noncovalently to CPA with dissociation constants (Ki's) of 11 fM and 710 pM, respectively. The CPA-ZFVP(O)F complex crystallizes in the space group P2(1)2(1)2(1) with unit cell parameters a = 65.3 A, b = 63.4 A, and c = 76.0 A, and the CPA-ZAGP(O)F complex crystallizes in the space group P2(1)2(1)2(1) with unit cell parameters a = 63.4 A, b = 65.9 A, and c = 74.4 A. Both structures were determined by molecular replacement to a resolution of 2.0 A. The final crystallographic residuals are 0.189 for the CPA-ZFVP(O)F complex and 0.191 for the CPA-ZAGP(O)F complex. The CPA-ZFVP(O)F complex exhibits the lowest Ki yet determined for an enzyme-inhibitor interaction. Comparison of the CPA-ZFVP(O)F structure with that of the CPA-ZAAP(O)F complex [Kim, H., & Lipscomb, W.N. (1990) Biochemistry 29, 5546-5555] indicates the likely important contributions of hydrophobic and weakly polar enzyme-inhibitor interactions to the exceptional stability of the CPA-ZFVP(O)F complex. Among these interactions is a network of four aromatic rings of CPA and ZFVP(O)F in a configuration that allows stabilizing aromatic-aromatic edge-to-face interactions from one ring to the next. A comparison of the structures of the CPA-ZFVP(O)F, CPA-ZAAP(O)F and CPA-ZAGP(O)F complexes shows that all three phosphonates assume a similar binding mode in the active-site binding groove of CPA. For ZAGP(O)F, the glycyl P1 residue does not lead to an anomalous or a partially disordered binding mode as seen in some previous complexes of CPA involving dipeptide analogue inhibitors with glycyl P1 residues. The additional enzyme-inhibitor interactions for these tripeptide phosphonates secure a binding mode in which a Pi portion of the inhibitor is clearly bound by the corresponding Si binding subsite. These three phosphonates have been implicated as transition-state analogues of the CPA-catalyzed reaction. The phosphinyl groups of these phosphonates coordinate to the active-site zinc in a manner that has been proposed as a characteristic feature of the general-base (Zn-hydroxyl or Zn-water) mechanism for the CPA-catalyzed reaction. Further mechanistic proposals are made for Arg-127, whose probable role in binding substrates is apparent in these CPA-phosphonate complexes.     

The structure of Photosystem I from the thermophilic cyanobacterium Synechococcus sp. determined by electron microscopy of two-dimensional crystals.

The structure of the Photosystem I (PS I) complex from the thermophilic cyanobacterium Synechococcus sp. has been investigated by electron microscopy and image analysis of two-dimensional crystals. Crystals were obtained from isolated PS I by removal of detergents with Bio-Beads. After negative staining, either single layers or two superimposed layers with a rotational different orientation were observed. The layers have a rectangular unit cell of 16.0 x 15.0 nm, which contains two PS I monomers. The monomers are arranged alternating up and down in each layer. For double-layer crystals, the images of the two layers could be separately processed by a combination of Fourier-peak-filtering and correlation averaging. Features in the two-dimensional plane can be seen with a resolution up to 1.5-1.8 nm. A model for the PS I structure was obtained by combining three-dimensional reconstructions from three tilt-series. The model shows an asymmetric PS I complex. On one side (presumably the stromal side) there is a 3 nm high ridge. This is most likely comprised of the psaC, psaD and psaE subunits. The other side (presumably the lumenal side) is rather flat, but in the center there is a 3 nm deep indentation, which possibly separates partly the two large subunits psaA and psaB.     

Phosphorylation of the active, A1 component of cholera toxin by protein kinase.

Cholera toxin, an activator of adenylate cyclase in a wide variety of cells, is a substrate for the phosphotransferase reaction catalyzed by purified cyclic adenosine 5'-monophosphate dependent bovine cardiac muscle protein kinase and the protein associated with human erythrocyte membranes. Phosphorylation occurs when the toxin is dissociated with 5-20 mM dithiothreitol and is restricted to the A1 or "adenylate cyclase activating" subunit of the toxin.     

Electron crystallography of membrane proteins: two-dimensional crystallization and screening by electron microscopy

Structural and functional information of membrane proteins at ever-increasing resolution is being obtained by electron crystallography. While a large amount of work on the development of methods for electron microscopy and image processing has resulted in tremendous advances in terms of speed of data collection and resolution, general guidelines for crystallization are first starting to emerge. Yet two-dimensional crystallization itself will always remain the limiting factor of this powerful approach in structural biology. Two-dimensional crystallization through detergent removal by dialysis is the most widely used technique. Four main factors need to be considered for the dialysis method: the protein preparation, the detergent, the lipid added as well as any constituent lipid, and the buffer conditions. Equally important is proper and careful screening to identify two-dimensional crystals.     

Tertiary structure of mouse epidermal growth factor determined by two-dimensional 1H NMR

The tertiary structure of mouse epidermal growth factor (EGF) in solution (28 degrees C, pH 2.0) was studied by two-dimensional NMR spectroscopy. Proton-proton distance constraints derived from NOESY spectra were used to construct a mechanical molecular model of mouse EGF, which was subsequently checked by means of a preliminary distance geometry calculation. The chain-folds in the two structural domains of mouse EGF were very similar to those previously reported (Montelione et al. (1987) Proc. Natl. Acad. Sci. U.S. 84, 5226-5230). However, the relative orientations of the two domains were different. Because we could assign much more inter-domain NOEs, the relative orientations of the two domains were well determined in our model. The hollow between the two domains may function as a binding site for the EGF receptor.     

Binding of NAD+ by cholera toxin.

1. The Km for NAD+ of cholera toxin working as an NAD+ glycohydrolase is 4 mM, and this is increased to about 50 mM in the presence of low-Mr ADP-ribose acceptors. Only molecules having both the adenine and nicotinamide moieties of NAD+ with minor alterations in the nicotinamide ring can be competitive inhibitors of this reaction. 2. This high Km for NAD+ is also reflected in the dissociation constant, Kd, which was determined by a variety of methods. 3. Results from equilibrium dialysis were subject to high error, but showed one binding site and a Kd of about 3 mM. 4. The A1 peptide of the toxin is digested by trypsin, and this digestion is completely prevented by concentrations of NAD+ above 50 mM. Measurement (by densitometric scanning of polyacrylamide-gel electrophoretograms) of the rate of tryptic digestion at different concentrations of NAD+ allowed a more accurate determination of Kd = 4.0 +/- 0.4 mM. Some analogues of NAD+ that are competitive inhibitors of the glycohydrolase reaction also prevented digestion.     

Conformational instability of the cholera toxin A1 polypeptide

Cholera toxin (CT) moves from the cell surface to the endoplasmic reticulum (ER) by vesicular transport. In the ER, the catalytic CTA1 subunit dissociates from the holotoxin and enters the cytosol by exploiting the quality control system of ER-associated degradation (ERAD). It is hypothesized that CTA1 triggers its ERAD-mediated translocation into the cytosol by masquerading as a misfolded protein, but the process by which CTA1 activates the ERAD system remains unknown. Here, we directly assess the thermal stability of the isolated CTA1 polypeptide by biophysical and biochemical methods and correlate its temperature-dependent conformational state with susceptibility to degradation by the 20S proteasome. Measurements with circular dichroism and fluorescence spectroscopy demonstrated that CTA1 is a thermally unstable protein with a disordered tertiary structure and a disturbed secondary structure at 37 °C. A protease sensitivity assay likewise detected the temperature-induced loss of native CTA1 structure. This protease-sensitive conformation was not apparent when CTA1 remained covalently associated with the CTA2 subunit. Thermal instability in the dissociated CTA1 polypeptide could thus allow it to appear as a misfolded protein for ERAD-mediated export to the cytosol. In vitro, the disturbed conformation of CTA1 at 37 °C rendered it susceptible to ubiquitin-independent degradation by the core 20S proteasome. In vivo, CTA1 was also susceptible to degradation by a ubiquitin-independent proteasomal mechanism. ADP-ribosylation factor 6, a cytosolic eukaryotic protein that enhances the enzymatic activity of CTA1, stabilized the heat-labile conformation of CTA1 and protected it from in vitro degradation by the 20S proteasome. Thermal instability in the reduced CTA1 polypeptide has not been reported before, yet both the translocation and degradation of CTA1 may depend upon this physical property.     

ADP-ribosylation by cholera toxin: functional analysis of a cellular system that stimulates the enzymic activity of cholera toxin fragment A1

We have clarified relationships between cholera toxin, cholera toxin substrates, a membrane protein S that is required for toxin activity, and a soluble protein CF that is needed for the function of S. The toxin has little intrinsic ability to catalyze ADP-ribosylations unless it encounters the active form of the S protein, which is S liganded to GTP or to a GTP analogue. In the presence of CF, S.GTP forms readily, though reversibly, but a more permanent active species, S-guanosine 5'-O-(3-thiotriphosphate) (S.GTP gamma S), forms over a period of 10-15 min at 37 degrees C. Both guanosine 5'-O-(2-thiodiphosphate) and GTP block this quasi-permanent activation. Some S.GTP gamma S forms in membranes that are exposed to CF alone and then to GTP gamma S, with a wash in between, and it is possible that CF facilitates a G nucleotide exchange. S.GTP gamma S dissolved by nonionic detergents persists in solution and can be used to support the ADP-ribosylation of nucleotide-free substrates. In this circumstance, added guanyl nucleotides have no further effect. This active form of S is unstable, especially when heated, but the thermal inactivation above 45 degrees C is decreased by GTP gamma S. Active S is required equally for the ADP-ribosylation of all of cholera toxin's protein substrates, regardless of whether they bind GTP or not. We suggest that active S interacts directly with the enzymic A1 fragment of cholera toxin and not with any toxin substrate.(ABSTRACT TRUNCATED AT 250 WORDS)     

Conditions of two-dimensional crystallization of cholera toxin B-subunit on lipid films containing ganglioside GM1

A systematic study of the lipid-layer two-dimensional crystallization technique has been carried out on the system composed of cholera toxin B-subunit and monosialoganglioside GM1, by electron microscopy, image analysis, and lipid film surface pressure measurements. Concentrations of protein and lipid components required for two-dimensional crystallization of toxin-GM1 complexes have been determined. Crystals were only obtained in the presence of mixed lipid films, composed of GM1 and of unsaturated lipids, such as dioleoylphosphatidylcholine or dioleoylphosphatidylethanolamine, in agreement with a previous report [D. S. Ludwig et al., (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 8585–8588]. Crystals were obtained with cholera toxin B-subunit concentration as low as 5 μg/ml, as well as in the presence of protein contaminants. They were obtained over a wide range of concentrations of both GM1 and unsaturated lipids. The minimal lipid amount needed for crystallization corresponded to a lipid monolayer at, or near, the maximal spreading pressure (50 mN/m). The use of an excess of lipid resulted in a stabilization of lipid monolayers and in a higher reproducibility or crystallization experiments.