University of Groningen
Cryo-electron crystallography from protein reconstitution to object reconstruction
Koning, Roman Ivan
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to
cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2003
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Koning, R. I. (2003). Cryo-electron crystallography from protein reconstitution to object reconstruction
Groningen: s.n.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
Download date: 17-06-2017
Chapter 3
29
Preparation of Flat Carbon Support Films
Roman I. Koning, Gert T. Oostergetel and Alain Brisson
Biophysical Chemistry, Groningen Biomolecular Sciences and Biotechnology Institute
(GBB), University of Groningen, Nijenborgh 4, 9747 AG, Groningen, The Netherlands.
Published in:
Ultramicroscopy 94 (3-4), 183-191 (2003).
30
Abstract
Wrinkling of carbon support films is known to limit the resolution of electron
microscopy images of protein two-dimensional crystals. The origin of carbon wrinkling
during preparation of the support films was investigated by reflected light microscopy.
We observed that carbon films go through several states during their preparation. While
dried carbon films have a tendency to be wrinkled, a flat state is observed transiently
before complete drying. This state could be stabilized by the addition of sugars or tannic
acid to the embedding medium. An alternative method for preparing flat carbon films
was developed, in which a sandwich is formed by two symmetrical carbon films
positioned on both sides of a grid. The formation of sandwiched carbon films was
facilitated by the use of grids with thin bars. The carbon sandwich films were flat, stable,
and easily and reproducibly prepared.
31
Introduction
The three-dimensional structure determination to near-atomic resolution of protein
molecules by cryo-electron microscopy has been achieved so far in a limited number of
cases (Henderson et al., 1990), (Kuhlbrandt et al., 1994), (Nogales et al., 1998), (Murata et
al., 2000). Instead, two- or three-dimensional protein structures are usually resolved to a
lower resolution (Reviews in (Jap et al., 1992), (Kuehlbrandt, 1992), (Walz and Grigorieff,
1998)). Several factors that limit the resolution in cryo-electron crystallography have been
identified, including the limited crystal quality, radiation damage, the lack of crystal
flatness, specimen movement and charging (Henderson, 1992).
Carbon films are commonly used as support for biological specimens in electron
microscopy for their high transparency for electrons, good conductive properties, and
mechanical stability (Bradley, 1965). On the other hand, carbon films have a high
tendency to wrinkle, which limits the resolution of electron crystallographic data. Carbon
films have been shown to wrinkle when they come into contact with electron microscopy
grids (Schmutz and Brisson, 1996), (Schmutz et al., 1994) and cryo-induced wrinkling has
been observed when supported carbon films are deposited on copper grids and cooled to
liquid nitrogen temperature (Booy and Pawley, 1993). The presence of irregularities in
carbon films has also been shown to limit the quality of electron crystallographic data
(Glaeser, 1992), (Han et al., 1994). The resolution in cryo-electron crystallography is also
limited by beam-induced specimen and image movement, caused by charging of
biological specimens (Henderson and Glaeser, 1985), (Bottcher, 1995), (Brink et al.,
1998b). Several methods have been proposed to reduce the effects of charging, including
the addition of one or two conductive carbon layers on unsupported specimens
(Jakubowski et al., 1989) and the preparation of biological specimens between two
symmetrical carbon support layers (Ren et al., 2000).
To get further insight into the formation of wrinkles in carbon support films, we
followed their formation by reflected light microscopy, a method previously introduced
for investigating the planarity of carbon films (Schmutz and Brisson, 1996). This allowed
us to visualize the formation of wrinkles in single carbon films, to get further
understanding on the effect of sugars, and to develop a new preparation technique in
which the specimen is sandwiched between two flat carbon films.
32
Materials and Methods
Materials.
Mica (item G250-1), carbon rods (item E410-2), copper grids (G2400C, 400 Mesh) and
thin-bar copper grids (G2700C, 200 Mesh) were purchased from Agar Scientific Ltd.
Lactose and glucose were purchased from Merck, sucrose from BDH Lab. Supplies,
trehalose from Sigma Chemical Co., and tannic acid from Polysciences Inc.
Preparation of carbon support films.
Carbon support films were prepared by depositing carbon by resistive heating onto
freshly cleaved mica in a vacuum evaporator (Edwards E380A). The carbon was carefully
stripped off from the mica surface (3 x 3 mm) onto a clean water surface. The carbon
film was picked up from underneath by an electron microscope grid that was held
horizontally and dried by blotting with a filter paper.
Microscopy.
Reflection light microscopy was performed with a Nikon microscope equipped with a
Hitachi KP-160 CCD camera. The behavior of the carbon film during specimen
preparation was recorded using a video-recorder and transferred to an Apple Macintosh
computer.
Electron microscopy was performed on a Philips CM200-FEG microscope equipped
with a 14-bit 1k x 1k Gatan type 794 Multi-Scan CCD, which was controlled by Digital
Micrograph software [20].
Atomic force microscopy (AFM) was performed using a Nanoscope IIIa-MultiMode
AFM (Digital Instruments) operated in constant force mode using oxide-sharpened
silicon nitride tips (spring constant: 0.06 N/m, scanning rate: 3 - 8 Hz, scan angle: 90°).
Results
Formation of wrinkles during the preparation of single carbon support film.
The formation of wrinkles during the deposition of carbon films on electron microscopy
grids was investigated by reflection light microscopy (Schmutz and Brisson, 1996). Five
states could be distinguished between picking up the carbon films with a grid and
complete drying (Fig. 1). These different states differed by the level of flatness/extent of
33
wrinkling of the carbon film. Directly after the carbon film was picked up with the grid
from the water surface it adopted the convex shape of the water droplet (Fig. 1a).
Wrinkles mainly emerged from the edges of cracks that are present in the carbon films
after they have been floated from mica on water. After lowering the water level to the
level of the grid bars carbon films adopted an overall straight shape. Wrinkles were found
to form locally due to the interaction of the carbon with the (uneven) surface of the grid
(Fig. 1b). Subsequently, the carbon film bent inwards into the grid squares, following the
concave-shaped water meniscus. Wrinkles in the carbon film still protruded from the grid
bars, but the wrinkles in the middle of the squares were smoothened (Fig. 1c). Just before
total drying occurred, carbon films were transiently flat, almost devoid of wrinkles (Fig.
1d). A logical interpretation is that the descent of the water level and its accumulation at
the wedge-shaped spaces that were created between the grid bars and the carbon film
resulted in a stretching of the carbon film and the disappearance of the local wrinkles. It
must be noted that the plane of the flat areas was located below the upper surface of the
grid (Fig. 1d). Ultimately, water evaporated or was removed by blotting from the wedgeshaped regions. Hereby the tension on the carbon film was released resulting in a
wrinkled carbon film (Fig. 1e).
The influence of sugar on the formation of flat carbon films.
While the transitions between the concave and convex states (Fig. 1a-c) were slow and
continuous, the transitions between the concave and wrinkled states (Fig. 1c-e) were fast
and sudden. We found that the carbon film could be trapped in any of the different
states by adding appropriate amounts of sugars or tannic acid to water and varying the
extent of blotting. When a 1 to 3 % sugar solution was used and a minimal amount of
blotting was applied, the carbon film in the central grid meshes was predominantly
trapped in the flat state (Fig. 2). At higher sugar percentages, carbon films were
predominantly trapped in a convex, straight or concave state and often sugar deposited
in the middle of the squares.
34
Figure 1.
Reflection light microscopy images of the same 36 squares during the formation of single carbon
support films. During formation five states can be distinguished. Schematic drawings (top) show side
views of a square in the described state. (a) The carbon film has a convex shape and interacts barely
with the grid bars. Wrinkles mainly protrude from cracks in the carbon film or from trapped air
bubbles. (b) Upon blotting the carbon film touches the grid bar and adopts an overall planar form.
Additional wrinkles are induced by the local interaction with the grid bar. (c) The carbon film is pulled
into the grid mesh by the capillary force of the water. Carbon wrinkles in the middle of the squares are
smoothened while irregularities protrude from the grid bars. (d) The carbon in the center of the
squares is flattened due to the capillary force of the water that acts on the rim of the carbon film near
the grid bars. The flat state is short lived and upon further drying suddenly changes. (e) After complete
drying, the carbon film is extensively wrinkled.
35
solution via a groove (Fig. 3c) and positioned beneath the carbon film (Fig. 3d). The
thin-bar grid then lifted the carbon film vertically from the solution (Fig. 3e). A flat
carbon sandwich support film was formed (as shown in Fig. 5 and section 3.4) after
water removal by slight blotting or evaporation (Fig. 3f).
Figure 3.
The preparation of a carbon sandwich support film using a well containing 100 µl of specimen
solution. (a) The Teflon well with internal dimensions of 3.5 x 3.5 x 7.0 mm with a groove of 3.5 x 3.5 x
1.0 mm is filled with specimen solution. (b) A carbon film is stripped off from a piece of mica of 3.0 x
6.0 mm onto the water surface, while the mica sinks to the bottom of the well. (c) A 200 mesh thin-bar
copper grid is inserted in the groove underneath the water surface and (d) is positioned under the
center of the carbon film. (e) The carbon film is picked up from the water surface by lifting the grid
vertically. Excess solution is automatically squeezed out and blotted off on the water surface when the
grid is lifted through the water-air interface. (f) The specimen is trapped in the squares of the grid and
in-between the two carbon layers. After removal of the excess of water by blotting or evaporation the
carbon sandwich is formed.
37
Figure 2.
Reflective light microscopy images of 16 squares recorded in the middle of EM grids with single
carbon support films. The films were picked up from aqueous solutions containing 1-4% of glucose,
lactose, tannic acid, trehalose or sucrose. The volume of liquid transferred was about 3-5 µl. The films
were dried in air after minimal blotting. With solutions containing 2% of sugars or tannic acid the
carbon films were flat. When less than 2% was used, some of the squares showed wrinkles. When more
than 2% was used, sugars or tannic acid formed dried aggregates over the carbon film.
Methods of preparing flat carbon sandwich support films.
Attempts to stabilize flat carbon films by a process similar to that observed in Fig. 1d led
to the development of a reproducible method for preparing flat sandwich carbon films.
The basic principle is that two carbon films are deposited in sandwich on both sides of a
grid, exerting on each other the pulling forces required for maintaining them flat.
A Teflon well was designed to contain a minimal volume of liquid solution (100 µl as
depicted in Fig. 3a or 60 µl when the grid insertion groove was tilted 45°). A carbon film
was deposited from a mica piece (3 x 6 mm) onto the solution filling the well (Fig. 3b). A
200 Mesh thin-bar copper grid, with bars of 10±2 µm thick, was inserted into the
36
A second approach to prepare flat carbon sandwich support films was developed which
required only 1 µl of protein solution (Fig. 4). A carbon film was floated onto a water
surface from a 3 x 3 mm piece of mica and was picked up by a horizontally held thin-bar
200 mesh copper grid (Fig. 4a). The sample solution (1 µl) was ‘back-injected’ into the
remaining solution (Fig. 4b). A second piece of carbon (2 x 2 mm) was deposited on the
water surface and picked up using a thin metal wire loop with a diameter of about 5 mm
(Fig. 4c). This piece of carbon was deposited on the side of the grid opposite to the side
where the first carbon layer was deposited, thus sandwiching the specimen solution (Fig.
4d). Simultaneously, the remaining solution was carefully blotted away by a filter paper
(Fig. 4e). After removal of water a carbon sandwich support film was formed (Fig. 4f).
Figure 4.
Preparation of a carbon sandwich support film that only requires 1 µl of specimen solution. (a) A piece
of carbon film of 3.0 x 3.0 mm is picked up from a water surface with a 200 mesh thin-bar copper grid.
(b) The grid is positioned with the carbon layer down and 1 µl specimen solution is injected into the
water droplet. (c) A second carbon film is picked up with a thin loop with a diameter of about 5 mm
and (d) is placed on the side of the grid opposite to the first carbon layer. (e) While the second carbon
film is held on the grid, the water is carefully blotted away by a filter paper. (f) After further blotting
from the side the carbon sandwich layer is formed.
38
The formation of carbon sandwich support films.
The formation of carbon sandwich support films, as described above, was followed with
reflected light microscopy (Fig. 5). In many aspects it resembled the formation of the
single layer carbon film described in Figure 1. The straight and concave states were
observed first when the carbon film was picked up (Fig. 5a). Further removal of water
was accomplished by touching the side of the grid with a filter paper, since evaporation
was slow. Upon water removal, the two carbon layers on both sides of the grid bent
inwards, adopted a concave shape (Fig. 5b) until they got ‘in touch’ and formed a flat
carbon sandwich film (Fig. 5c). When standard grids with 40 µm bars were used the
sandwiched carbon film was formed in only few of the squares and in a small area in
these squares.
Figure 5.
Analysis by reflective light microscopy of the same 9 squares during preparation of a carbon sandwich
support film, using 200 mesh thin-bar copper grids. (a) Straight and convex carbon film states were
observed directly after the carbon film was picked up from the water. Wrinkles were mainly induced by
the cracks in the carbon. (b) Concave carbon film states were induced by water removal. The capillary
force of the water pulled the carbon films on either side of the grid towards each other. (c) When the
carbon layers touched each other they rapidly formed a flat carbon film.
39
Figure 6.
A representative example of a double carbon film, prepared using a 200 Mesh thin-bar grid and a
sample solution containing IICmtl crystals. On the outside of the mesh two single carbon layers
approach each other from either side of the grid and form a double carbon layer in the middle of the
mesh. In a major part of the mesh, at the inside of the black rim, IICmtl crystals are trapped between the
two layers of the carbon sandwich, which are stabilized by a reproducibly very thin water layer.
Thin-bar grids of 200 Mesh (with 10 µm thick bars) were necessary to allow the
formation of a flat carbon film over a major part of the grid meshes (Fig. 6). The water
layer that was formed between the carbon sandwich was reproducibly thin and crystals
could be trapped in-between the two carbon layers.
40
Flatness of carbon sandwich support films.
Atomic force microscopy was used to investigate the flatness of the carbon sandwich
films. The central areas of carbon sandwich films were shown to be mostly devoid of
wrinkles. The curvature of flat carbon films was less than 0.2°/µm (Fig. 7).
Figure 7.
(a) AFM topography image
of the surface of a carbon
sandwich film. Cracks in
the carbon and white dots,
probably
being
sparked
carbon, are visible. The
height profile ranges form
0 to 250 nm. The black
horizontal
and
white
vertical lines with arrows
depict height sections that
are shown in (b) and (c)
respectively. (b) The height
profile
from
the
black
horizontal line in a. shows
a height difference of 1.96
nm over a distance of 6.88
µm
between
the
arrowheads. The maximal
curvature in this area was
0.14 º/µm. (c) The height
profile
from
the
white
vertical line shows that the
height difference near the
cracks is in the order of 5
to 25 nm over a distance of
a few µm.
41
Discussion
It was shown earlier that wrinkling of carbon films occurred during their deposition on
electron microscopy grids (Schmutz and Brisson, 1996), (Schmutz et al., 1994). We show
here that during the process of depositing carbon support films on a grid, the carbon
film passes through several states, including a transient flat state. The successive states
are mainly controlled by the level or amount of water associated with the grid.
The formation of flat supports attracted our interest, since supports that are not flat are
unfavorable for high-resolution electron crystallography (Glaeser, 1992). The transient
occurrence of flat films during the deposition of carbon films on a grid is obtained by
stretching the wrinkled film upon lowering the water level.
This flat state could be stabilized by addition of sugars -glucose, lactose, sucrose or
trehalose- or tannic acid to the embedding medium. It is possible that upon drying,
sugars accumulate by draining at the level of the grid bars and form a solid material to
which the carbon film remains associated. Several studies have established the beneficial
effect of sugars or tannic acid in high-resolution electron crystallography (Fujiyoshi,
1998), (Nogales et al., 1995), (Wang and Kuehlbrandt, 1991), (Vonck, 2000). The main
effect is likely due to the formation of a protective layer surrounding the protein crystals.
This study reveals a direct effect of sugars on the film flatness, which might be partly
responsible for improved resolution at high tilt angles.
Flat carbon support films can also be created by positioning two carbon films on both
sides of a 200 mesh thin-bar grid, as shown on the scheme of Fig. 4. The curvature of
sandwiched carbon films was shown by AFM to be less than 0.2°/µm, which is
significantly less than was reported earlier for wrinkled carbon grids (Schmutz et al.,
1994), and is theoretically sufficient to collect high-resolution diffraction data from highly
tilted protein crystals (Glaeser, 1992). This preparation method has several advantages: (i)
the method of preparation is simple and reproducible; (ii) there is almost no waste of
specimen solution during preparation; (iii) more than half of the squares of a grid are in a
sandwiched form; these squares are easily distinguished from squares with a single
carbon layer, since the former is marked by a thin dark ring at the transition of
sandwiched to single layered carbon; (iv) the sandwiched film is very stable both in air
and in vacuum, so the grids are easily handled and stored; (v) using the sandwich
technique specimens including protein 2D crystals can routinely be prepared in vitreous
ice.
42
The exact thickness was not measured and we were not able to accurately control the
thickness of the water layer that was present in-between the two carbon layers. However,
the thickness of the water layer that is automatically formed in the sandwich was
reproducibly thin and suitable for high-resolution cryo-electron microscopy (Fig. 6).
Conclusions
(1) The presence of sugars or tannic acid in the embedding medium when specimens are
prepared on single carbon support films using the back-injection technique stabilizes the
carbon support film in a flat state.
(2) A new method was developed for preparing flat carbon sandwich support films for
cryo-electron microscopy.
43
44