|
Preparation of the supercomplex.
PSII supercomplexes were isolated from market spinach leaves (Spinacea
oleracea) according to a method published previously (1).
Electron Microscopy.
A Philips CM200 FEG-equipped electron microscope was used at 200
kV in combination with a Gatan liquid N2 cryostage. To optimise
the imaging conditions in terms of defocus and astigmatism, a Tietz
Video Systems' fast FFT processor was used. The vitrified samples
(24) were imaged at a calibrated
magnification of 48,600x, using different defocus values of between
1.35 µm and 1.9 µm. Second images were taken at an underfocus value
of 7.2 µm, in order to make focal pairs to facilitate the identification
of single particles (Fig. 1b and c).
Image Processing.
A total of 108 micrographs, showing minimal astigmatism and drift,
were digitised with an 'Emil' densitometer (Image Science GmbH,
Berlin) to produce a pixel size corresponding to 5.76 Å per pixel
at the specimen level. All image processing was performed within
the Imagic-5 software environment (25),
running under the UNIX operating system on a COMPAQ XP1000 computer.
A data set of 15,650 single particle images was obtained by picking
all discernible particles from the close to focus micrographs. The
contrast transfer function (CTF) was calculated for each micrograph
and the extracted molecular images were corrected by phase reversal
in the appropriate frequency ranges. An analysis of the data set
was made starting with the reference free alignment-by-classification
procedure (26). This identified
several sub-populations of particles differing in size and shape.
Each sub-population was in turn analysed as a separate data set,
with the reference free alignment giving the initial class averages
necessary for multi-reference alignment (27),
leading to class averages with an enhanced signal to noise ratio.
Relative orientations were determined for the class averages by
angular reconstitution (28),
resulting in initial 3D reconstruction gained from implementation
of the exact back projection technique (27).
Reprojections were taken from each 3D model and used to identify
additional atypical views and further refine the class averages
within each sub-population data set. At each stage, roughly 40%
of the class averages were discarded after assessment using cross
correlation functions. Through this iterative refinement, the data
converged to give the best 3D reconstructions shown. The resolution
was determined by Fourier Shell Correlation between two independent
3D reconstructions (27). Co-ordinate
data sets derived from the 8 Å CP47-RC (15)
and 9 Å PSII core dimer (13)
cryoelectron crystallographic maps were modelled into the supercomplex
and subpopulation reconstructions using the 'O' modelling software
package (29). The AVS software
package (30) was used to visualise
the 3D maps in Figs. 2, 3
and 4.
|
