Electron microscopy

Download Report

Transcript Electron microscopy

Electron microscopy
electron diffraction and
X-ray crystallography
Sven Hovmöller
Structural Chemistry, Stockholm University
SE-106 91 Stockholm, Sweden
[email protected]
The resolution of the light microscope
is limited by the wavelength of light
400 – 700 nm
Electrons in an electron microscope have a
wavelength under 0.1 Ångström
Electron gun
Lenses
Sample holder
Electron
microscope
Bacteria are so thick that they
have to be sliced for the EM
200 nm
Electron microscopy started cell biology
Electron microscopy
(EM) image of human
muscle fibres.
Viruses and bacteriophages were
first seen by EM
Bacteriophages
attacking lactobillus
Image by Nurmiaho-Lassila,
Helsinki.
Organelles can be seen by EM
1000Å
Mitochondrion with parallel inner membranes
ify!
magn
400 Å
Higher magnification
without increased resolution is pointless
400 Å
Spherical viruses determined by cryo transmission
electron microscopy at Purdue
Bettina Böttcher
SEM: Scanning Electron Microscopy
gives 3D images but only of the surface and with low resolution
Transmission Electron Microscopy (TEM) gives higher resolution


Figure 2. a) Electron microscopy image of the membrane protein
complex cytochrome reductase. The individual protein complexes have
a molecular weight of 550 kDa and are seen as small rings. From
Hovmöller et al. 1983.
b) The calculated Fourier transform of the crystal in a) shows clear
diffraction spots, extending to 5 orders in one direction and 7 in the
other. The crystallographic resolution is thus 137Å/5 = 27 Å along h
and 174Å/7 = 25 Å along k. The diffraction spots are indexed along h
and k as shown by the two reflections 3 5 and 2 -6.
Resolution.
k


h

The
resolution
of
this
diffractogram is 25 Å, since the
outermost reflections 5 0 and
0 7 (at the outermost ring) have
d-spacings of 25 Å.
The innermost ring corresponds
to a resolution of 80 Å. The two
middle rings are at 45 and 35 Å
resolution, respectively. The
kind of images we can see using
these different data are shown
in Figure 3.
Atomic resolution is reached at
about 2 Å resolution, i.e. with a
radius 12 times larger than the
largest circle shown here.
The effect of increased resolution
80 Å
45 Å




35 Å
25 Å
Left: the resolution is 80 Å, i.e. only the data inside
the innermost ring has been used. b) 45 Å, c) 35 Å
and d) 25 Å resolution.
At 80 Å resolution we see only a blob.
At 45 Å the molecule looks (incorrectly) as a
tetramer.
At 35 and 25 Å resolution we see that the molecule
is an ellipsoidal dimer. We can also see that the
molecules in the corner and middle of the unit cells
are differently oriented (one is flipped over relative
to the other).
Atomic resolution in the EM
20Å
Electron microscopy image of a metal oxide showing an interesting
intermediate stage of ordering between crystalline and disordered.
Each small black dot is an NbO6 octahedron.
This near atomic resolution image was taken by Gunnar Svensson.
The first known membrane protein
structure was bacteriorhodopsin
– by low dose EM
2D projection perpendicular to the membrane
3D reconstruction from several tilted views
Unwin & Henderson, MRC Cambridge 1975
The hydrophobicity plot is based on this work
The transmembrane protein bacteriorhodopsin visualised
in five different ways.
a) wire frame
b) ball-and-stick
c) space-filling
d) secondary structure and
e
e) original balsa-wood model from EM
a to d are made with RASMOL on data from PDB.
EM 25 Å resolution
Weiss,H., Hovmöller,S. and Leonard,K. Preparation of
Membrane Crystals of Ubiquinol-Cytochrome c Reductase
from Neurospora Mitochondria and Structure Analysis by
Electron Microscopy.
Methods in Enzymology (1986) 126, 191-201
Chicken cytochrome bc1 complex
X-ray 2.8 Å resolution
The structure is built up by successively adding
in the reflections (=Fourier coefficients




a) The reflection 1 1 cuts the a edge of
the unit cell once and the b edge also
once per unit cell. This cosine wave
has its maximum (0o) at the origin of
the unit cell, so its phase is 0.
b) The reflection 1 –1 also has the
phase 0.
c) The summation of the 3 reflections
1 1, 1 –1 and 0 2 results in a blob at
the origin of the unit cell and another
blob in its center.
d) The reflection 0 2 cuts the b edge
twice but never the a edge. The phase
is 0.
Solving the phase problem in X-ray
crystallography for proteins using heavy
atom derivatives.
a
b
c
d
Native protein Heavy atom Protein + heavy metal
e
Protein crystal structures can be
refined to about 0.1 Å accuracy
The disagreement between observed and calculated
amplitudes (or intensities I(hkl) = F(hkl)2 ) are
calculated as the crystallographic R-value
EM on whole cells - tomography
Visualization of actin network,
membranes, and cytoplasmic
macromolecular complexes.
A volume of 815 nm by
870 nm by 97 nm
Baumeister et al. MPI Martinsreid
Fig. A and B: 26S proteasome as visualized in its cytoplasmic context.
D
D. X-ray at 2.4 Å
resolution
Fig. C: Purified and negatively stained 26S proteasome
Baumeister et al. MPI Martinsreid
Ribosomes from cryo-EM and X-rays
Marin van Heel London
Joachim Franck Albany
The Protein Data Base PDB is on the Internet

The wealth of accurate structural data,
mainly on proteins but also on DNA, RNA
and carbohydrates is freely available on the
Internet at http://www.rcsb.org/pdb/
Conclusions
X-ray diffraction and electron microscopy are the
main sources of structural information in
biological systems. Both methods can reach
atomic resolution.
 X-ray crystallography is the method of choice if
3D-crystals of at least 20x20x20μm3 (containing
> 108 ordered identical molecules) can be made.
 For single particles or larger systems that are not
ordered, such as organelles or whole cells, electron
microscopy is the main technique.
The Ramachandran plot

The torsion angles in
the polypeptide
backbone
The Ramachandran plot
Alanine/non-Glycine
Glycine
Smoothing the plots:
-sheets
Random coil
Random coil

Ramachandran plots are very informative
Ramachandran was right in all essentials, but the details were
wrong
All distributions are along the diagonal, i.e. the  and  angles
are coupled
The -helices are very rigid
The -sheet region is divided into two parts, but parallel and
anti-parallel -sheets are nearly identical
The Gly plot is very far from the predicted

Hovmöller, Zhou & Ohlson Acta Cryst D58 (2002) 768-776





Shape strings
Amino acids: comformational preferences
Amino acid shape-propensities from nrPDB, H-sorted
1
0.9
Fraction of the seven shapes
0.8
0.7
G
0.6
U
V
0.5
T
R
S
0.4
H
0.3
0.2
0.1
0
E
A
Q
K
L
M
R
D
W
N
H
S
F
Y
T
I
C
P
V
G
Shapes are conservative
Shape String SSSSHHSSSS


107 cases of SSSSHHSSSS
– 103 are well clustered,
all with 3 or 4 H-bonds
(yellow and green) in antiparallel β-sheets. Only 4
cases differ in shape –
these are not connected by
H-bonds.
carbon
oxygen
nitrogen
One sequence can have many shapes
sequence
3-frag  hits N shapes
shape1:hits shape2:hits ... shapeN:hits
ALV
256
19
HHH:145 HHR:3 HHS:2 HRS:3 HSH:1 HSS:3 HUS:1 HVR:1 RHH:2 RRS:2 RSH:1
RSR:4 RSS:10 SHS:1 SRR:1 SRS:5 SSH:1 SSR:5 SSS:65
GGG
237
101
GGG:2 GGH:1 GGR:1 GGS:5 GHG:3 GHH:6 GHR:6 GHS:1 GHT:2 GHV:1 GRG:1
GRH:2 GRR:1 GRS:2 GRT:2 GSH:1 GSR:2 GSS:1 GTG:1 GTT:6 GVR:3 HGG:2
HGH:1 HGR:3 HGS:3 HGT:3 HHG:1 HHH:10 HHR:1 HHS:2 HHT:5 HHV:1 HRS:2
HSH:1 HSR:1 HSS:1 HTG:1 HTH:2 HTR:4 HTS:1 HTT:3 RGR:1 RGT:1 RHG:1
RHH:4 RHR:1 RHS:2 RRR:1 RRS:1 RRT:2 RSG:1 RSH:1 RSR:2 RSS:3 RTG:2
RTR:4 RTT:8 SGG:1 SGR:1 SGS:3 SGT:4 SHG:1 SHH:1 SHR:1 SHS:2 SHT:1
SRH:3 SRR:1 SRS:2 SRT:6 SSG:3 SSH:1 SSR:2 SSS:7 STG:1 STS:1 STT:2 TGG:2
TGH:1 TGS:1 TGT:1 THG:8 THR:2 THS:1 THV:1 TRG:5 TRH:5 TRR:1 TRS:7 TSG:3
TSR:2 TSS:1 TTG:1 TTH:5 TTR:4 TTS:1 TTT:2 VGG:1 VGH:2 VHR:1 VHT:2
VVI
160
17
HHH:15 HRH:1 RHS:3 RRR:1 RSR:4 RSS:12 RSU:1 SRH:2 SRR:2 SRS:6 SRT:1
SRV:1 SSH:2 SSR:26 SSS:81 SSU:1 SUR:1
VLI
141
17
HHH:36 HRR:1 RHU:1 RSR:2 RSS:4 RVH:2 SRH:4 SRR:1 SRS:7 SRU:1 SSR:27
SSS:50 STR:1 STS:1 SUR:1 SUS:1 SVR:1
LPP
142
14
GRH:1 HHH:5 HRR:1 RHH:3 RRH:37 RRR:35 RRV:1 SHH:1 SRH:18 SRR:32 THH:1
TRR:1 URR:5 URV:1
PPL
84
12
HHH:12 HHS:1 RHG:2 RHH:8 RHR:3 RHS:4 RRH:19 RRR:15 RRS:17 RRT:1 RRV:1
RVS:1
PPP
65
5
HHH:1 RHH:2 RHR:2 RRH:12 RRR:48
Predicting Protein Structure
Helix 35%
Sheet 20%
Random coil 45%
Gripping the N-terminal end
a) All proteins have both ends
(N:blue, C: red) on the surface of
the molecule. Here is a typical
example.
b) The first and last amino acids in proteins are
equally often found on the surface (red) and
rarely in the centre of proteins (black curve).
Hovmöller & Zhou, PROTEINS (2003), in press.
Dr. Tuping Zhou
programmer
Jonas Almqvist
Ph.D. student
Tomas Ohlson
Ph.D. student
Dr. Roger Ison
Author of Frags