Fan Hai-Fu

Image resolution enhancement by combining information from electron diffraction pattern and micrograph

Abstract An electron micrograph of chlorinated copper phthalocyanine at 2 A resolution taken on the Kyoto 500 kV electron microscope has been enhanced to 1 A resolution by incorporating the information from the corresponding electron diffraction pattern. Structure-factor amplitudes up to 1 A resolution were obtained from the electron diffraction pattern. Phases of the structure factors within 2 A resolution were derived from the Fourier transform of the electron micrograph. A phase-extension technique introduced from X-ray crystallography was then used to derive the phases between 2 and 1 A resolution. The final image was obtained by the inverse Fourier transform using the structure factor amplitudes from the electron diffraction pattern and the phases from the electron micrograph and from the phase-extension procedure.

Solving a superstructure from two-wavelength x-ray powder diffraction data - a simulation

Two different kinds of phase ambiguities are intrinsic in two-wavelength x-ray powder diffraction from acentric crystal structures having pseudo-translation symmetry. In a test calculation we have solved the problem for the first time by two different phasing procedures developed originally in single-crystal structure analysis. They are the direct method of breaking enantiomorphous phase ambiguity in protein crystallography and that of breaking translational phase ambiguity for superstructures. An artificial structure was used in the test, which is based on atomic coordinates of the known structure, SHAS (C5H6O5N3K), with the atom K replaced by Rb. The arrangement of Rb atoms possesses a subperiodicity of t = (a + b + c)/2. Two-wavelength synchrotron x-ray powder diffraction data were simulated with λ1=0.0816nm and λ2=0.1319nm. Overlapped reflections were uniformly decomposed at the beginning and redecomposed afterward when the partial-structure information became available. The enantiomorphous phase ambiguity was resolved only for reflections with h + k + l even. Phases of reflections with h + k + l odd were derived by the direct method of solving superstructures. A fragment was then obtained, which led to the complete structure in five cycles of Fourier iteration.

Direct phasing of one-wavelength anomalous-scattering data of the protein core streptavidin

The direct method [Fan, Hao, Gu, Qian, Zheng & Ke (1990). Acta Cryst. A46, 935-939] was used to break the phase ambiguity intrinsic to one-wavelength anomalous scattering data from a known protein of moderate size, core streptavidin, which was solved originally with three-wavelength anomalous diffraction data [Hendrickson, Pähler, Smith, Satow, Merritt & Phizackerley (1989). Proc. Natl Acad. Sci. USA, 86, 2190-2194]. Unlike that in the previous test with a small protein, the Fourier map calculated with the direct-method phases could not clearly reveal the moderate-sized protein structure. However, the phases can be improved step by step using Wangs solvent-flattening method, non-crystallographic symmetry averaging and the skeletonization method. The final electron-density map clearly shows most Calpha positions and some side chains and it is traceable without prior knowledge of the structure. It is concluded that the direct method is capable of breaking the OAS phase ambiguity of a moderate-sized protein at moderate resolution such as 3 A, while the combination of direct methods with macromolecular techniques may produce phases good enough for unknown protein structure to be traced.

New Techniques of Applying Multi-Wavelength Anomalous Scattering Data

Several different methods of using multi-wavelength anomalous scattering data are described and illustrated by application to the solution of the known protein structure, core streptavidin, for which data at three wavelengths were available. Three of the methods depend on the calculation of Patterson-like functions for which the Fourier coefficients involve combinations of the anomalous structure amplitudes from either two or three wavelengths. Each of these maps should show either vectors between anomalous scatterers or between anomalous scatterers and non-anomalous scatterers. While they do so when ideal data are used, with real data they give little information; it is concluded that these methods are far too sensitive to errors in the data and to the scaling of the data-sets to each other. Another Patterson-type function, the Ps function, which uses only single-wavelength data can be made more effective by combining the information from several wavelengths. Two analytical methods are described, called AGREE and ROTATE, both of which were very successfully applied to the core streptavidin data. They are both made more effective by preprocessing the data with a procedure called REVISE which brings a measure of mutual consistency to the data from different wavelengths. The best phases obtained from AGREE lead to a map with a conventional correlation coefficient of 0.549 and this should readily be interpreted in terms of a structural model.

SAD phasing by OASIS at different resolutions down to 0.30 nm and below

Single-wavelength anomalous diffraction (SAD) phasing is increasingly important in solving de novo protein structures. Direct methods have been proved very efficient in SAD phasing. This paper aims at probing the low-resolution limit of direct-method SAD phasing. Two known proteins TT0570 and Tom70p were used as test samples. Sulfur-SAD data of the protein TT0570 were collected with conventional Cu-Kα source at 0.18 nm resolution. Its truncated subsets respectively at 0.21, 0.30, 0.35 and 0.40 nm resolutions were used in the test. TT0570 Cu-Kα sulfur-SAD data have an expected Bijvoet ratio / ~ 0.55%. In the 0.21 nm case, a single run of OASIS-DM-ARP/wARP led automatically to a model containing 1178 of the total 1206 residues all docked into the sequence. In 0.30 and 0.35 nm cases, SAD phasing by OASIS-DM led to traceable electron density maps. In the 0.40 nm case, SAD phasing by OASIS-DM resulted in a degraded electron density map, which may be difficult to trace but still contains useful secondary-structure information. Test on real 0.33 nm selenium-SAD data of the protein Tom70p showed that even automatic model building was not successful, the combination of manual tracing and direct-method fragment extension was capable of significantly improving the electron-density map. This provides the possibility of effectively improving the manually built model before structure refinement is performed.

Combining SAD/SIR iteration and MR iteration in partial-model extension of proteins

There are two kinds of dual-space partial-model extensions which involve the direct-method program OASIS. The first kind, named SAD/SIR iteration, uses SAD/SIR information, while the second kind, named molecular replacement (MR) iteration, does not use that information. In general, the SAD/SIR iteration is more powerful since more experimental information is used. However, in most cases when protein structures are solved with the molecular replacement method, SAD/SIR information is not available. Thus the MR iteration is particularly useful for the completion of models from molecular replacement. The SAD/SIR iteration will be automatically used in OASIS for data sets containing SAD/SIR signals, while the MR iteration will be dedicated to data sets without SAD/SIR signals. The present paper shows that for data containing SAD/SIR signals, a combination of SAD/SIR iteration and MR iteration could lead to significantly better results than that obtained from the SAD/SIR iteration alone.

Is single-wavelength anomalous scattering sufficient for solving phases? A comparison of different methods for a 2.1 Å structure solution

The structure of rusticyanin is the largest unknown structure (M(r) = 16.8 kDa) which has been recently solved by the direct-methods approach using only single-wavelength anomalous scattering (SAS) data from the native protein [Harvey et al. (1998). Acta Cryst. D54, 629-635]. Here, the results of the Sim distribution approach [Hendrickson & Teeter (1981). Nature (London), 290, 107-113] and of the CCP4 procedure MLPHARE [Collaborative Computational Project, Number 4 (1994). Acta Cryst. D50, 760-763] are compared with those from direct methods. Analysis against the final refined model shows that direct methods produced significantly better phases (average phase error 56 degrees ) and therefore significantly better electron-density maps than the Sim distribution and MLPHARE approaches (average phase error was around 63 degrees in both cases).

OASIS4.0-a new version of the program OASIS for phasing protein diffraction data

The program OASIS4.0 has been released. Apart from the improved single-wavelength anomalous diffraction (SAD) phasing algorithm described in a separate paper, an important new feature in this version is the automation of the iterative phasing and model-building process in solving protein structures. A new graphical users interface (GUI) is provided for controlling and real-time monitoring the dual-space iterative process. The GUI is discussed in detail in the present paper.

SIR phasing by combination of SOLVE/RESOLVE and dual-space fragment extension involving OASIS

A new phasing procedure has been proposed for dealing with single isomorphous replacement (SIR) x-ray diffraction data. The procedure combines SOLVE/RESOLVE with the dual-space fragment extension involving OASIS. Two sets of SIR data at 0.28 nm resolution taken from the protein (R)-phycoerythrin (PDB code: 1LIA) were used in the test. For one of the two SIR data sets, a default run of SOLVE/RESOLVE based on the heavy-atom substructure found by SHLEXD led automatically to an interpretable electron density map. OASIS could not effectively improve the result. For the other set of SIR data, SOLVE/RESOLVE resulted in a fragmented model consisting of 454 of the total 668 residues, in which only 29 residues were docked into the sequence. Based on this model, 7 iteration cycles of OASIS-DM-RESOLVE (build only) yielded automatically a model of 547 residues with 133 residues docked into the sequence. The overall-averaged phase error decreased considerably and the quality of electron density map was improved significantly. Two more cycles of iterative OASIS-DM-RESOLVE were carried out, in which the output phases and figures of merit from DM were merged with that from the original run of SOLVE/RESOLVE before they were passed onto RESOLVE (build only). This led automatically to a model containing 452 residues with 173 docked into the sequence. The resultant electron density map is manually traceable. It is concluded that when results of SOLVE/RESOLVE are not sufficiently satisfactory, the combination of SOLVE/RESOLVE and OASIS-DM-RESOLVE (build only) may significantly improve them.

Ab-initio determination of the incommensurate modulated structure of Bi-2212 from x-ray powder diffraction data—a simulation

A set of x-ray powder diffraction data of the high-Tc superconductor Bi2Sr2Ca1Cu2Oy (Bi-2212) was simulated based on the experimental single-crystal diffraction data by merging together reflections with diffraction angles (2θ) closer to each other than 0.04 degrees. There are three types of overlapping in the powder diffraction data, i.e. (i) overlapping of main reflections; (ii) overlapping of satellite reflections and (iii) overlapping of main and satellite reflections. The third type of overlapping was first separated into main and satellite components according to the ratio between the average intensity of that of types (i) and (ii). Then the overlapped reflections of main reflections and those of the satellites were uniformly partitioned. Heavy-atom sites in the basic/average structure were found using the uniformly decomposed main reflections by the conventional direct method. Phases of the satellites were derived by the multidimensional direct method. The resultant four-dimensional Fourier maps revealed correctly the essential feature of the modulation. No assumption on either the basic structure or the modulation is needed.