Advances in direct methods for protein crystallography.

Recent advances in ab initio direct methods have enabled the solution of crystal structures of small proteins from native X-ray data alone, that is, without the use of fragments of known structure or the need to prepare heavy-atom or selenomethionine derivatives, provided that the data are available to atomic resolution. These methods are also proving to be useful for locating the selenium atoms or other anomalous scatterers in the multiple wavelength anomalous diffraction phasing of larger proteins at lower resolution.

Crystal structure of vancomycin.

BACKGROUND: Vancomycin and other related glycopeptide antibiotics are clinically very important because they often represent the last line of defence against bacteria that have developed resistance to antibiotics. Vancomycin is believed to act by binding nascent cell wall mucopeptides terminating in the sequence D-Ala-D-Ala, weakening the resulting cell wall. Extensive NMR and other studies have shown that the formation of asymmetric antibiotic dimers is important in peptide binding. Despite intensive efforts the crystal structure of vancomycin has been extremely difficult to obtain, partly because high-resolution data were unavailable, and partly because the structure was too large to be solved by conventional "direct methods'. RESULTS: Using low-temperature synchrotron X-ray data combined with new ab initio techniques for solving the crystallographic phase problem, we have succeeded in determining the crystal structure of vancomycin at atomic resolution. The structure provides much detailed information that should prove invaluable in modelling and mechanistic studies. CONCLUSIONS: Our structure confirms that vancomycin exists as an asymmetric dimer. The dimer conformation allows the docking of two D-Ala-D-Ala peptides in opposite directions; these presumably would be attached to different glycopeptide strands. In the crystal, one of the binding pockets is occupied by an acetate ion that mimics the C terminus of the nascent cell wall peptide; the other is closed by the asparagine sidechain, which occupies the place of a ligand. The occupied binding pocket exhibits high flexibility but the closed binding pocket is relatively rigid. We propose that the asparagine sidechain may hold the binding pocket in a suitable conformation for peptide docking, swinging out of the way when the peptide enters the binding pocket.

Crystal structure determination at 1.4 A resolution of ferredoxin from the green alga Chlorella fusca.

BACKGROUND: [2Fe-2S] ferredoxins, also called plant-type ferredoxins, are low-potential redox proteins that are widely distributed in biological systems. In photosynthesis, the plant-type ferredoxins function as the central molecule for distributing electrons from the photolysis of water to a number of ferredox-independent enzymes, as well as to cyclic photophosphorylation electron transfer. This paper reports only the second structure of a [2Fe-2S] ferredoxin from a eukaryotic organism in its native form. RESULTS: Ferredoxin from the green algae Chlorella fusca has been purified, characterised, crystallised and its structure determined to 1.4 A resolution - the highest resolution structure published to date for a plant-type ferredoxin. The structure has the general features of the plant-type ferredoxins already described, with conformational differences corresponding to regions of higher mobility. Immunological data indicate that a serine residue within the protein is partially phosphorylated. A slightly electropositive shift in the measured redox potential value, -325 mV, is observed in comparison with other ferredoxins. CONCLUSIONS: This high-resolution structure provides a detailed picture of the hydrogen-bonding pattern around the [2Fe-2S] cluster of a plant-type ferredoxin; for the first time, it was possible to obtain reliable error estimates for the geometrical parameters. The presence of phosphoserine in the protein indicates a possible mechanism for the regulation of the distribution of reducing power from the photosynthetic electron-transfer chain.

The 1.2 A crystal structure of hirustasin reveals the intrinsic flexibility of a family of highly disulphide-bridged inhibitors.

BACKGROUND: Leech-derived inhibitors have a prominent role in the development of new antithrombotic drugs, because some of them are able to block the blood coagulation cascade. Hirustasin, a serine protease inhibitor from the leech Hirudo medicinalis, binds specifically to tissue kallikrein and possesses structural similarity with antistasin, a potent factor Xa inhibitor from Haementeria officinalis. Although the 2.4 A structure of the hirustasin-kallikrein complex is known, classical methods such as molecular replacement were not successful in solving the structure of free hirustasin. RESULTS: Ab initio real/reciprocal space iteration has been used to solve the structure of free hirustasin using either 1.4 A room temperature data or 1.2 A low temperature diffraction data. The structure was also solved independently from a single pseudo-symmetric gold derivative using maximum likelihood methods. A comparison of the free and complexed structures reveals that binding to kallikrein causes a hinge-bending motion between the two hirustasin subdomains. This movement is accompanied by the isomerisation of a cis proline to the trans conformation and a movement of the P3, P4 and P5 residues so that they can interact with the cognate protease. CONCLUSIONS: The inhibitors from this protein family are fairly flexible despite being highly cross-linked by disulphide bridges. This intrinsic flexibility is necessary to adopt a conformation that is recognised by the protease and to achieve an optimal fit, such observations illustrate the pitfalls of designing inhibitors based on static lock-and-key models. This work illustrates the potential of new methods of structure solution that require less or even no prior phase information.

Ab initio determination of the crystal structure of cytochrome c6 and comparison with plastocyanin.

BACKGROUND: Electron transfer between cytochrome f and photosystem I (PSI) can be accomplished by the heme-containing protein cytochrome c6 or by the copper-containing protein plastocyanin. Higher plants use plastocyanin as the only electron donor to PSI, whereas most green algae and cyanobacteria can use either, with similar kinetics, depending on the copper concentration in the culture medium. RESULTS: We report here the determination of the structure of cytochrome c6 from the green alga Monoraphidium braunii. Synchrotron X-ray data with an effective resolution of 1.2 A and the presence of one iron and three sulfur atoms enabled, possibly for the first time, the determination of an unknown protein structure by ab initio methods. Anisotropic refinement was accompanied by a decrease in the 'free' R value of over 7% the anisotropic motion is concentrated at the termini and between residues 38 and 53. The heme geometry is in very good agreement with a new set of heme distances derived from the structures of small molecules. This is probably the most precise structure of a heme protein to date. CONCLUSIONS: On the basis of this cytochrome c6 structure, we have calculated potential electron transfer pathways and made comparisons with similar analyses for plastocyanin. Electron transfer between the copper redox center of plastocyanin to PSI and from cytochrome f is believed to involve two sites on the protein. In contrast, cytochrome c6 may well use just one electron transfer site, close to the heme unit, in its corresponding reactions with the same two redox partners.

High-resolution structure of bovine pancreatic trypsin inhibitor with altered binding loop sequence.

A mutant of bovine pancreatic trypsin inhibitor (BPTI) has been constructed and expressed in Escherichia coli in order to probe the kinetic and structural consequences of truncating the binding loop residues to alanine. In addition to two such mutations (Thr11Ala and Pro13Ala), it has a conservative Lys15Arg substitution at position P(1) and an unrelated Met52Leu change. In spite of the binding loop modification, the affinity for trypsin is only 30 times lower than that of the wild-type protein. At pH 7.5 the protein can be crystallized on the time-scale of hours, yielding very stable crystals of a new (tetragonal) form of BPTI. Conventional source X-ray data collected to 1.4 A at room temperature allowed anisotropic structure refinement characterized by R=0.1048. The structure reveals all 58 residues, including the complete C terminus, which is in a salt-bridge contact with the N terminus. The Cys14-Cys38 disulfide bridge is observed in two distinct chiralities. This bridge, together with an internal water molecule, contributes to the stabilization of the binding loop. The Ala mutations have only an insignificant and localized effect on the binding loop, which retains its wild-type conformation (maximum deviation of loop C(alpha) atoms of 0.7 A at Ala13). Four (instead of the typical three) additional water molecules are buried in an internal cleft and connected to the surface via a sulfate anion. Three more SO(4)(2-) anions are seen in the electron density, one of them located on a 2-fold axis. It participates in the formation of a dimeric structure between symmetry-related BPTI molecules, in which electrostatic and hydrogen bonding interactions resulting from the mutated Lys15Arg substitution are of central importance. This dimeric interaction involves direct recognition loop-recognition loop contacts, part of which are hydrophobic interactions of the patches created by the alanine mutations. Another 2-fold symmetric interaction between the BPTI molecules involves the formation of an antiparallel intermolecular beta-sheet that, together with the adjacent intramolecular beta-hairpin loops, creates a four-stranded structure.

Can anomalous signal of sulfur become a tool for solving protein crystal structures?

A general method for solving the phase problem from native crystals of macromolecules has long eluded structural biology. For well diffracting crystals this goal can now be achieved, as is shown here, thanks to modern data collection techniques and new statistical phasing algorithms. Using solely a native crystal of tetragonal hen egg-white lysozyme, a protein of 14 kDa molecular mass, it was possible to detect the positions of the ten sulfur and seven chlorine atoms from their anomalous signal, and proceed from there to obtain an electron-density map of very high quality.

Cyclic diguanylic acid behaves as a host molecule for planar intercalators.

Cyclic ribodiguanylic acid, c-(GpGp), is the endogenous effector regulator of cellulose synthase. Its three-dimensional structure from two different crystal forms (tetragonal and trigonal) has been determined by X-ray diffraction analysis at 1 A resolution. In both crystal forms, two independent c-(GpGp) molecules associate with each other to form a self-intercalated dimer. A hydrated cobalt ion is found to coordinate to two N7 atoms of adjacent guanines, forcing these two guanines to destack with a large dihedral angle (32 degrees), in the dimer of the tetragonal form. This metal coordination mechanism may be relevant to that of the anticancer drug cisplatin. Moreover, c-(GpGp) exhibits unusual spectral properties not seen in any other cyclic dinucleotide. It interacts with planar organic intercalator molecules in ways similar to double helical DNA. We propose a cage-like model consisting of a tetrameric c-(GpGp) aggregate in which a large cavity ('host') is generated to afford a binding site for certain planar intercalators ('guests').

Crystal structures of [Met5] and [(4-bromo)Phe4,Met5]enkephalins: formation of a dimeric antiparallel beta-structure.

The crystal structure of [(4-bromo)Phe4,Met5]enkephalin (Tyr-Gly-Gly-(4-bromo)-Phe-Met) shows two independent molecular conformations. The molecules are arranged in parallel in a head-to-tail fashion and form an antiparallel beta-sheet structure involving intermolecular hydrogen bonds. This dimeric beta-structure is also observed in the [Met5]enkephalin crystal, in spite of their different crystal packing environments, which shows the energetic stability of this molecular conformation. The three-dimensional similarity between the dimeric beta-structure and the beta-turn form is discussed in the relation to the opioid delta and mu receptors.

X-ray structure of beta-cyclodextrin-2,7-dihydroxy-naphthalene.4.6 H(2)O: an unusually distorted macrocycle.

The inclusion complex beta-cyclodextrin.2,7-dihydroxynaphthalene.4.6 H(2)O crystallized in the monoclinic space group P2(1), with a=14.082(3), b=19.079(4), c=12.417(3) A, beta=109.28(3) degrees, V=3149.0(11) A(3), and Z=2. An X-ray study performed at room temperature shows that the crystal packing is of the herringbone type with one 2,7-dihydroxynaphthalene included completely in the beta-CD cavity, its long axis being oriented along the beta-CD molecular axis, and 4.6 water molecules are placed in the interstitial space. The beta-CD macrocycle is elliptically distorted, and the guest molecule is held in the hydrophobic beta-CD cavity by C-H...O and C-H...pi interactions.

X-ray structure of the cyclomaltohexaicosaose triiodide inclusion complex provides a model for amylose-iodine at atomic resolution.

Cyclomaltohexaicosaose (CA26) is folded into two 1(2)/(3) turns long V-helices that are oriented antiparallel. Crystals of complexes of CA26 with NH(4)I(3) and Ba(I(3))(2) are brown and X-ray analyses show that I(3)(-) units are located in the approximately 5 A wide central channels of the V-helices. In the complex with NH(4)I(3), two CA26 molecules are stacked to form 2 x 1(2)/(3) turns long channels harbouring 3 I(3)(-) at 3.66-3.85 A inter I(3)(-) distance (shorter than van der Waals distance, 4.3 A), whereas in the Ba(I(3))(2) complex, CA26 are not stacked and only one I(3)(-) each fills the V-helices. Glucose...I contacts are formed with C5-H, C3-H, C6-H and (at the ends of the V-helices) with O6 in (+) gauche orientation. By contrast, O2, O3, O4 and O6 in the preferred (-) gauche orientation do not interact with I because these distances are >/=4.01 A and exceed the van der Waals I...O sum of radii by about 0.5 A except for one O2...I distance of 3.68 A near the end of one V-helix. Raman spectra indicate that the complexes share the presence of I(3)(-) with blue amylose-iodine.

2-Ethyl-5-(3-indolyl)oxazole from Streptomyces cinnamomeus discovered by chemical screening. Characterization and structure elucidation by X-ray analysis.

In the lipophilic extracts from Streptomyces cinnamomeus 2-ethyl-5-(3-indolyl)oxazole (1a) was detected by chemical screening methods. The structure of the crystalline 1a was determined by spectroscopic and X-ray analysis. The new mono- and dibromo derivatives 1b and 1c are described. 1a is identical with pimprinethine and belongs to a group of microbial indole alkaloids, which can be regarded as masked tryptamine derivatives.

Ab initio structure determination of the lantibiotic mersacidin.

The crystal structure of mersacidin, a potential novel antibiotic against methicillin- and vancomycin-resistant Staphylococcus aureus strains, has been determined by ab initio methods. Despite all crystals being merohedrally twinned, an accurate structural model with an R value of 13.4% has been obtained at atomic resolution. With six molecules in the asymmetric unit and no atom heavier than sulfur, the structure corresponds to a protein of 120 amino acids and is the largest approximately equal-atom unknown structure solved by direct methods. In the crystal, the molecule assumes a compact fold different from that found by NMR in solution. Comparison of the NCS-related molecules reveals regions of variable flexibility. The region highly homologous to the related antibiotic actagardine is very rigid and possibly defines an essential building block of this class of new antibacterial substances.

The application of direct methods and Patterson interpretation to high-resolution native protein data.

Conventional small-molecule methods of solving the phase problem from native data alone, without the use of heavy-atom derivatives, known fragment geometries or anomalous dispersion, have been tested on 0.9 A resolution data for two small proteins: rubredoxin, from Desulfovibrio vulgaris, and crambin. The presence of three disulfide bridges in crambin and an FeS(4) unit in rubredoxin enabled automated Patterson interpretation as well as direct methods to be tried. Although both structures were already well established, the known structures were not used in the phasing attempts, except for identifying successful solutions. Direct methods were not successful for crambin, although the correct phases were stable to phase refinement and gave figures of merit clearly superior to any obtained in the ca 500 000 random starting phase sets that were refined. It appears that the presence of an iron atom in rubredoxin reduces the scale of the search problem by many orders of magnitude, but at the cost of producing 'over-consistent' phase sets that overemphasize the iron atom and involve partial loss of enantiomorph information. However, about 1% of direct-methods trials were successful for rubredoxin, giving mean phase errors of about 56 degrees (for all E > 1.2) that could be reduced to about 20 degrees by standard E-Fourier recycling methods. Limiting the resolution of the data degraded the quality of the solutions and suggested that the limiting resolution for routine direct-methods solution of rubredoxin is about 1.2 A. With the 0.9 A data, automated Patterson interpretation convincingly finds the three disulfide bridges in crambin and the FeS(4) unit in rubredoxin, and in both cases E-Fourier recycling starting from these 'heavier' atoms yields almost the complete structure. Whereas crambin could only be solved in this way at very high resolution, rubredoxin could be solved by Patterson interpretation down to 1.6 A. These results emphasize the benefits of collecting protein data to the highest possible resolution, and indicate that when a few 'heavier' atoms are present, it may prove possible in favorable cases to solve the phase problem from a single native data set collected to 'atomic resolution'.

Structure of vancomycin and its complex with acetyl-D-alanyl-D-alanine.

Vancomycin, a broad-spectrum antibiotic, inhibits the growth of cell walls by complex formation with peptides terminating in D-alanyl-D-alanine. The structure of vancomycin was determined by X-ray analysis of the degradation product CDP-I. A model of the complex is proposed based on this study and spectroscopic data.

Ab initio solution and refinement of two high-potential iron protein structures at atomic resolution.

The crystal structure of the reduced high-potential iron protein (HiPIP) from Chromatium vinosum has been redetermined in a new orthorhombic crystal modification, and the structure of its H42Q mutant has been determined in orthorhombic (H42Q-1) and cubic (H42Q-2) modifications. The first two were solved by ab initio direct methods using data collected to atomic resolution (1.20 and 0. 93 A, respectively). The recombinant wild type (rc-WT) with two HiPIP molecules in the asymmetric unit has 1264 protein atoms and 335 solvent sites, and is the second largest structure reported so far that has been solved by pure direct methods. The solutions were obtained in a fully automated way and included more than 80% of the protein atoms. Restrained anisotropic refinement for rc-WT and H42Q-1 converged to R(1) = summation operator||F(o)| - |F(c)|| / summation operator|F(o)| of 12.0 and 13.6%, respectively [data with I > 2sigma(I)], and 12.8 and 15.5% (all data). H42Q-2 contains two molecules in the asymmetric unit and diffracted only to 2.6 A. In both molecules of rc-WT and in the single unique molecule of H42Q-1 the [Fe(4)S(4)](2+) cluster dimensions are very similar and show a characteristic tetragonal distortion with four short Fe-S bonds along four approximately parallel cube edges, and eight long Fe-S bonds. The unique protein molecules in H42Q-2 and rc-WT are also very similar in other respects, except for the hydrogen bonding around the mutated residue that is at the surface of the protein, supporting the hypothesis that the difference in redox potentials at lower pH values is caused primarily by differences in the charge distribution near the surface of the protein rather than by structural differences in the cluster region.

Structure of balhimycin and its complex with solvent molecules.

Balhimycin is a naturally occurring glycopeptide antibiotic, related to vancomycin which acts by binding nascent bacterial cell-wall peptide ending in the sequence D-Ala-D-Ala. Crystals of balhimycin are monoclinic, space group P21, a = 20.48 (10), b = 43.93 (21), c = 27.76 (14) A, beta = 100.5 (5) degrees with four independent antibiotic molecules, three molecules of 2-methyl-2,4-pentanediol, two citrate ions, three acetate ions and 127.5 water molecules in the asymmetric unit. With an asymmetric unit larger than those of the smallest proteins and a solvent content of about 32%, the crystals have similar diffraction properties to those of small proteins. 27387 unique reflections were collected using synchrotron radiation. The structure was solved by a standard protein technique, the molecular-replacement method, using ureido-balhimycin as search model. The anisotropic refinement against all F2 data between 0.96 and 45 A converged to a conventional R value of 11.27% with R1= Sigma||Fo|-|Fc||/Sigma|Fo| for the 24623 data with I > 2sigma(I) and 12.58% for all 27387 data. The four monomers possess fairly similar conformations (r.m.s. deviation 0.7 A). Two antibiotic molecules form a tight dimer with antiparallel hydrogen bonds between the peptide backbone as well as between the vancosamine residues and the peptide backbone. In each of the two dimers, one binding pocket is occupied by a citrate ion and the other by an acetate ion. The dimer units are linked in the crystal by hydrogen bonds to form infinite chains.

Comparison of different X-ray data-collection systems using the crystal structure of octreotide.

The octapeptide octreotide crystallizes with three peptide molecules and about 20% water in the asymmetric unit, and in many ways possesses diffraction properties similar to those of a 'mini-protein' consisting of 24 amino-acid residues. It diffracts to about 1.0 A but data in the range 1.4-1.0 A are weak. It provides a suitable test of different macromolecular X-ray data-collection techniques, especially of their ability to measure weak reflections accurately. In contrast to typical proteins it is possible to perform a full anisotropic refinement, that we believe provides a more objective test of the quality of the data than the internal consistency of equivalent reflections. We have collected a total of six data sets. The X-ray sources included synchrotron radiation, Cu Kalpha rotating anodes and Mo Kalpha sealed tubes; position-sensitive two-dimensional detectors from four manufacturers and a four-circle diffractometer with scintillation counter were employed. Two of the six data sets were collected at low temperature. Reasonable anisotropic refinement was possible with all area-detector data sets, although significant differences in the precision of the final model were observed. In addition we tested the ability of automated Patterson interpretation to solve the structure using the six independent data sets. The structure solution was only successful using the synchrotron or rotating-anode data sets, i.e. for the more intense sources. It appears that for structure solution the maximum resolution of the data is critical, whereas for refinement the accuracy of the data is more important.