Zovko, S.
Citation
Zovko, S. (2010, June 22). Dynamics and structural features of the microtubule plus-ends in interphase mouse fibroblasts. Retrieved from https://hdl.handle.net/1887/15711
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/15711
Note: To cite this publication please use the final published version (if
applicable).
CHAPTER II
Electron Microscopy and Electron Tomography as a tool to study the microtubules and microtubule plus-end
conformations
34
35
Electron Microscopy and Electron Tomography as a tool to study the microtubules and microtubule plus-end conformations
ynamics of the microtubule (MT) network of interphase mammalian cells has been extensively studied by live-cell imaging using light microscopy (LM) techniques. In order to visualize MTs, the LM-approach requires fluorescently labeled molecules either involved in the construction of a MT or physically associated with it. MTs are typically visualized by injection or overexpression of fluorescently tagged tubulin in the cells (Komarova et al., 2002b; Ferenz and Wadsworth, 2007). When it comes to investigation of the dynamic instability of the plus-ends of the cellular MTs, an additional strategy has been proven useful, namely visualization of the growing plus-ends using fluorescently tagged proteins which bind specifically to the growing plus ends, the so called plus-end tracking proteins or +TIPs (Mimori-Kiyosue et al., 2000; Komarova et al., 2002a; Krylyshkina et al., 2003; Stepanova et al., 2003)
MT plus-end is a highly dynamic structure, exhibiting phases of slow growth and rapid shrinkage. Moreover, the plus-ends of in vitro assembled MTs have specific morphological features which are related to their actual dynamic state. One of the major goals of our study was to identify these structural features and to investigate their relationship to the dynamic state of the plus-end in situ.
In order to investigate the morphological features of the plus-end, which dimensions are in the range of a few tens of nanometers, an adequate microscope resolution is required.
Although the resolution of LM, which is approximately 200 nm (between 100-50nm when
D
36
Total Internal Fluorescence Resonance, TIRF, microscopy is applied) is good enough to study behavior of individual MTs in the cell periphery, it is a limiting factor when the visualization of the fine scale structure of the MT plus-end is considered. However, transmission electron microscopy (EM) provides the resolution required for the visualization of the structure of the MT plus-end EM has the resolving power in the nanometer range and the possibility to visualize an object with a 50-100 times higher magnification than LM. Combined with electron tomography (ET) which provides information on the structure of the object in 3D, EM is the ideal technique to study fine scale structures like MT plus-ends in their original cellular context (Chretien et al., 1995;
O'Toole et al., 2003; VandenBeldt et al., 2006).
In the next few paragraphs a number of important issues regarding visualization and investigation of MT plus-end features by means of EM/ET will be addressed.
First, techniques of transmission electron microscopy will be introduced, followed by a more detailed description of the main specimen preparation methods used in this thesis for MT visualization. Here both chemical versus cryo-fixation and sectioning versus whole-cell preparations will be discussed. All the data are taken from my study of the MTs and MT plus-ends present in the area near the cell membrane (periphery) of the interphase mouse fibroblasts.
Introduction to Electron Microscopy and Electron Tomography
In order to comprehend the electron microscopical approach of visualization of the cellular ultrastructure it is important to be familiar with the principles of electron microscopy. These are further discussed in brief below.
Unlike a light microscope which uses light to illuminate the specimen, an electron microscope (EM) uses a beam of highly energetic electrons. The superior resolution of the EM is due to the wavelength of an electron, which is much smaller than that of a light photon.
Electrons, which are produced by the electron-gun, are accelerated by a voltage of more than 100.000 volts. In vacuum, accelerated to close to the speed of light, the path of the (negatively charged) electrons is controlled by the electro-magnetic lenses of the EM. As the electrons are not blocked by the presence of air, they will move freely and pass through the specimen. A projection-image of the electrons that transmit through the specimen is recorded by a digital camera or on a photographic plate. Note that the vacuum in the electron microscope is essential; in air the electrons would only travel a few micrometers and in vacuum more than a meter. Consequently, vacuum-resistant specimens are required for observation by EM. During the last 50 years several procedures where developed to prepare biological specimen that are vacuum resistant.
When, the so called, conventional EM is considered, this usually implies the use of chemically fixed, stained and dried (dehydrated) specimens. Depending the specific protocols applied, the (heavy-metal) stain will be distributed at different surfaces in the specimen, e.g. along membranes. In addition, the last decade novel approaches were
37
developed based on rapid freezing (vitrification) of the sample and study of the vitrified samples at cryogenic (liquid nitrogen, liquid helium) temperatures. This so-called cryo- electron microscopical (cryo-EM) approach will be discussed in more detail in next paragraphs and Chapter IV of this thesis.
When the beam of electrons encounters stained thin specimen, some of the electrons are deflected. This deflection (scattering) depends on the mass density of the specimen:
the greater the atomic-mass of the atoms present in the specimen, the stronger the scattering. Depending on the scattering angle, the electrons that are scattered are blocked by an aperture underneath the sample. The unscattered electrons are not blocked by the aperture. Hence, those electrons in the electron beam that traverse the specimen and are not blocked by the aperture, hit the screen coated with fluorescent material, forming an image of the specimen, a 2D projection. Areas and features in the image that correspond to areas in the specimen with high mass density appear darker in the image.
Although EM has the advantage to “see details through” the specimen (projection), its drawback is that the structural details from different depths within the specimen are superimposed upon one another to form a 2D projection. This superposition of structural detail can make interpretation sometimes difficult, especially when a fully packed system, like a cell, is studied.
In order to overcome the problem of superimposed structures, one can visualize the structure of interest in three dimensions using electron tomography (ET). ET is a method for the reconstruction of three dimensional (3D) structures using a series of projections from the multiple tilt views of the structure acquired by means of an electron microscope (Koster et al., 1997; McEwen and Marko, 2001; McIntosh, 2001; McIntosh et al., 2005).
Basically, the specimen is automatically tilted in the electron beam of an EM and at every tilt-angle an image (projection) of the structure is recorded (Figure 1A). The acquired tilt series is then digitally processed to generate a 3D volume (often referred to as a tomogram). Several software packages for electron tomographic reconstructions are available, e.g. the IMOD package (Kremer et al., 1996). In short, the procedure is as follows: First, the collected 2D images are aligned to each other using cross-correlation.
Next, the alignment of the whole tilt series is refined, by means of fiducial gold markers, which were deposited onto the specimen during the specimen preparation process.
Fiducial markers function as reference points on each 2D image of the tilt series. After alignment, the 2D image sequence is then further digitally processed by projecting for each 2D image with known tilt angle the image intensities into a reconstruction volume (Figure 1B). In the 3D volume, image intensities that correspond to the same 3D mass density in the specimen overlap and reinforce one another. The outcome is a 3D volume (tomogram), which can be visualized as a set of 2D images showing parallel slices through the 3D volume.
Assessment of resolution in electron tomography is not a straightforward issue as it involves several aspects. Firstly, the resolution in tomograms is anisotropic (i.e. varies with viewing direction). Through the middle of the specimen, parallel to the tilt axis, and
38
Figure 1. Principles of electron tomography. (A) A series of 2D projections is recorded while tilting an object (in this illustration a knot) incrementally about an axis perpendicular to the electron beam (single-tilt t); (B) The aligned 2D image sequence is then further digitally processed by projecting the image intensities into a reconstruction volume for every 2D image using weighted back- projection algorithms. (C) The four images illustrate the effects of tilt range and tilt increments on the quality of a tomographic reconstruction. The first image (left) represents a reconstruction obtained by tilting the object from +90˚ to –90˚ with 2˚ increment, or +60˚ to –60˚ with 2˚ increment (second image from the left). Due to the missing wedge, the details (second image from the left) which are perpendicular to the virtual z-direction are smeared out. The third and fourth images from the left are obtained by tilting the object with an 5˚ increment from +90˚ to –90˚ and from +60˚ to –60 respectively. The third and fourth images have a lower resolution, when compared to the first and the second image. Adapted image from Baumeister et al., Cell Biology, Vol. 9, (1999).
in the plane of the specimen, the resolution is determined by the resolution of the images acquired, assuming perfect image alignment of the tilt-series. However, in the direction perpendicular to the specimen plane, in the z-direction (depth), the resolution is dependent on the number of projections and the specimen thickness (sampling angle).
For a cylindrical object it can be approached by the formula d=πD/N, in which d is resolution, D is diameter of the cylinder and N is the number of projections acquired (McEwen and Marko, 2001). Unfortunately, this formula is only an approximation as it is based on the assumption that the images are taken over a tilt range of +90˚ to -90˚. In practice, due to the specimen holder constraint, the tilt range is limited to 70˚. Since it is not possible to collect 2D images at tilt angles higher than 70˚ information is missing to generate an unambiguous 3D reconstruction from the tilt series („missing wedge‟
phenomenon). Additionally, at the high tilt angles the effective thickness of the sample increases twice at 60˚, and almost three times at 70˚. Both „missing wedge‟ and increasing effective thickness of the specimen cause a degradation of resolution in tomograms. All these factors affect the resolution of tomograms in negative manner (Figure 1C).
39
Loss of information due to „missing wedge‟ phenomenon can be minimized by recording a “double-tilt” series: after first tilt-series (“single-tilt”) the grid is rotated 90˚ and an additional series of the same spot on the specimen is recorded. After reconstruction of this second tilt series, these two tomograms are then combined into one tomogram. This approach results in reduction of the amount of missing information in tomograms and leads to better resolution in different directions (Mastronarde, 1997). It is of importance to note that the „missing wedge‟ phenomenon in particular has large consequences on the visualization of linear structures like MTs that fade when oriented perpendicularly to the tilt axis (Mastronarde, 1997; McEwen and Marko, 2001).
Beside the sample thickness and number of the tilt views, the type specimen and the stain distribution are of high importance for the quality of a tomographic reconstruction.
Visualization of the objects reconstructed with electron tomography is usually achieved by using graphic models, possibly based on surface- or volume rendering. For objects with very high signal to noise, and without the presence of nose, algorithms are implemented in commercially available software packages to automatically extract and segment interesting structures on the basis of the choice of the appropriate density threshold. Note that this is possible when the signal-to-noise ratio in the images is relatively high, which is not the case for transmission electron microscopy of stained sections.
Another way is to manually trace the features of interest marking their contours on each of the successive optical slices of the volume (in XY, YZ and XZ direction). All the contours together are then stacked to produce a 3D model. Consequently, this part of the data processing is highly sensitive to data interpretation. Visualization, segmentation and volume rendering is performed using software like AMIRA® 3D visualization software (TGS).
Specimen preparation for visualization of MTs and MT plus-ends by EM/ET
In order to optimize the sample preparation for visualization of the cellular ultrastructure, two aspects of particular importance for MT plus-end visualization will be discussed: a) choice of fixative and b) cell sectioning versus whole-cell (whole-mount sample).
Chemical versus cryo-fixation
Proper preservation of the cellular ultra structure depends largely on the fixation procedure. A sample for EM can generally be fixed by 1) chemical fixation or by 2) cryo- fixation (cryo-immobilization). Chemical fixation is the most widely used method for preservation of biological specimens. It is performed at 4˚C, room temperature (RT) or at 37˚C and involves the use of chemical compounds like glutaraldehyde and paraformaldehyde, which cross-link the proteins, and osmium tetraoxide, which fixes and stains the lipids. Subsequent dehydration of the sample is carried out at RT. Generally, the penetration/fixation rate of the above mentioned compounds is in the range of seconds to minutes. Due to the relatively slow diffusion rate of the chemical compounds,
40
combined with the vulnerability of the specimen to various factors, i.e. temperature, pH, osmolarity, buffer type, it is thought that this way of sample preparation results in the fixation of the response of the cells to the treatment rather than the fixation of the native state (Hayat et al, 1981).
The more sophisticated specimen preservation techniques include: a) cryo-fixation, resulting in a frozen-hydrated or vitrified1 specimen and b) cryo-fixation followed by freeze-substitution (FS). By ultra rapid freezing the cells/tissues are assumed to correspond most to the original state. The cryo-fixed samples can directly be imaged by cryo-EM while the cryo-fixed/freeze-substituted specimens have to be processed further before becoming suitable for investigation with EM.
Cryo-preservation of the sample can be achieved by ultra rapid freezing of the specimen at -180 ˚C or lower, and is usually performed using a cryo-plunging device or by applying the high pressure freezing (HPF) method. In both cases, fixation is attained within milliseconds, which results in the near-instantaneous arrest of the physiological processes within the specimen. This is essential for capturing dynamic processes within the cell, e.g. rapid shrinkage of MTs. Plunge-freezing of the specimen (e.g. cells grown on a coverslip or EM-grid) involves rapid dipping of the specimen in a liquid cryogen (ethane, propane, nitrogen), resulting in an almost instant cryo-immobilization and embedding of the specimen in amorphous ice. In theory, the vitrification depth for plunge-freezing is in the range of 0.5 µm. However, in practice vitrification throughout the whole sample can seldom be accomplished, due to the ice formation on the surface of the sample. This ice forms an isolating layer thereby hampering the proper freezing of deeper layers of the specimen. Poor temperature conductance of water and isolation by the ice lead to the formation of hexagonal and cubic ice crystals, which expand and disrupt the ultrastructure, thus causing „freeze-damage.‟ For this reason even in relatively thin samples some freeze-damage can be observed. On the other hand, in our laboratory, cells with well preserved nuclei (>0.5 µm thickness) were not exceptional, meaning that other (unknown) factors might play a role in the freezing process (e.g.
composition of water within the various cellular compartments).
The plunge-freezing device is typically used for freezing of cell monolayers, whereas HPF is used for cryo-preservation of non-adhering cells, very thick cells (e.g. mitotic, rounded cells) or for tissue samples. In the HPF method the sample is frozen by shooting the sample into a chamber filled with liquid nitrogen under high pressure conditions in order to reduce ice nucleation and thus ice crystal formation and can reach a vitrification depth up to 0.2mm.
After successful cryo-fixation by plunge-freezing, the thin (≤ 0.5 µm) parts of the whole vitrified cells can directly be investigated by cryo-EM using the intrinsic density of the structures within the sample for visualization (see below and Chapter IV).
Alternatively, the water in vitrified cells can slowly be substituted by an organic medium (acetone, methanol, ethanol etc.) at temperatures below the point of water crystallization
1 Vitrification of water = formation of the non-crystalline, vitreous, amorphous ice.
41
(< -85˚C) during the process called freeze-substitution (Giddings TH Jr, 2001). A typical freeze substitution medium often also contains heavy metal-containing (and thus contrast-enhancing) compounds (uranyl acetate, osmium tetraoxide, tannic acid etc.).
Freeze-substitution is followed by flat-embedding of the specimen in a resin (epon, lowycryl, etc.) and sectioning parallel to the attachment plane of the cells in order to visualize the major parts of the MT network and to easily trace the membrane oriented plus-ends (see also Chapter III).
In order to optimize the preservation, and thus visualization, of the MTs in our model cells, i.e. mammalian interphase fibroblasts, we compared chemical fixation at 37˚C followed by dehydration at RT, with cryo-fixation with or without subsequent freeze- substitution (see protocols in appendix of this thesis). Figure 2 gives an overview of work flow of each preparation method described here.
Figure 2. Work flow of specimen preparation procedure for visualization of cellular ultrastructure by EM/ET.
Three specimen preparation paths used in our experiments are depicted (a-c). Cells cultured on a substrate (Thermanox coverslip or an EM-grid) are chemically fixed at RT or cryo- immobilized by plunge-freezing. Upon cryo-fixation, frozen-hydrated cells can either be directly visualized (path (B), cryo-EM/ET) or freeze-substituted (path C). For cryo-ET, fiducial markers are added onto the substrate before growing the cells on it. Following chemical fixation or freeze substitution, cells are infiltrated with resin, flat- embedded and sectioned.
Subsequently, plastic sections are post-stained in order to enhance the contrast. For ET, fiducial gold markers are applied on top of the plastic section. Sections are then investigated by EM/ET. ET includes a number of steps, that is: 1) data acquisition, 2) tomographic reconstruction (using IMOD software) and 3) analysis of the tomogram (IMOD, AMIRA software).
42
Since a MT has a cylindrical shape in 3D, its 2D projection in EM shows two dark lines created by the superposition of protofilaments in the direction of projection (i.e. the
„sides‟ of the MT), having a higher signal than the „bottom‟ and „top‟ of the MT (Wade and Chrétien, 1993). Interestingly, the structural appearance of a MT wall appears to be significantly dependent on the fixation and dehydration method employed. Figure 3 shows the effect of chemical fixation at 37˚C/dehydration at RT versus cryo-fixation (with or without freeze-substitution) on the appearance, and thus preservation, of MTs in situ and in vitro. MTs in chemically fixed cells at 37˚C (3A and 3C) appear to have irregular MT wall when compared to the MTs in cryo-fixed/freeze-substituted/resin embedded cells (3B and 3D, respectively). The latter ones show similar features as those from the periphery of a vitrified whole fibroblast directly studied by cryo-EM (3E). This indicates that freeze-substitution does not cause (major) perturbations of the MTs and that the sample preservation is very close to its native state. This is supported by the cryo-EM of in vitro assembled MTs shown in Figure 3F (see also (Mandelkow and Mandelkow, 1985; Wade and Chrétien, 1993). Furthermore, when the same cryo-immobilized in vitro assembled MTs are additionally subjected to the sample preparation like in (A) and (B), the irregular appearance of MTs is observed again in the case of chemical fixation but not when freeze-substitution is applied (G and H, respectively). These results indicate that for the visualization and study of the MTs and MT plus-end structures it is of great importance that the samples are prepared by cryo-fixation (either alone or followed by freeze-substitution) instead of chemical fixation at 37˚C and dehydration at RT.
Cell sections versus whole-cell sample
Interphase mammalian cells like fibroblasts adhere onto a substrate and stretch out.
Their thickest part is the nucleus (up to 20µm), while the cell periphery is usually relatively thin (typically ranging from 2µm to a few hundred nm). In order to visualize the features inside the thicker part of the cell by EM, the cell has to be sliced into thin (100 nm) sections. Thin sectioning of the sample is important not only to allow good transmission of the electrons through the sample, but also to avoid beam-associated heat formation and subsequent damage of the sample. On the other hand, for ET it is essential that the structure of interest is present within the volume of the section as complete as possible. Here thicker (up to 300m) sections are more desirable.
The MTs of an interphase fibroblast form a radial array. They originate from the MT organizing center (MTOC; located near the nucleus), in which the MT minus-ends are embedded, and grow towards the cell periphery where their plus-ends exhibit dynamic instability (see Chapter I). To facilitate the visualization of the major parts of the MT network and therefore reliably locate the plus-ends near the cell edge, the cells are flat- embedded and sectioned (100nm sections) in the plane of cell attachment. Unfortunately by doing so the MTs which are not positioned precisely within the plane of section are cut. Recent studies have shown that MT plus-ends target focal adhesions, the structures responsible for attachment of the cell onto the cell substrate, approaching the cell surface in an angle of 5-10 degrees (Krylyshkina et al., 2003). Consequently, these MT plus-ends are often cut off. Therefore, sectioning often results in the creation of “artificial”
ends.
43
Electron Microscopy and Electron Tomography as a tool to study the microtubules and microtubule plus-end conformations
44
Figure 3 (previous page). Effect of specimen preparation on structural preservation of MTs and their appearance in EM. (A-B) Electron micrograph of a 100nm thin section of a flat-embedded interphase mouse fibroblast in the cell periphery. Cells were fixed by (A) chemical fixation at room temperature using 1.5%GA in PHEM buffer /1% osmium tetraoxide (OsO4) in PHEM buffer or (B) by plunge-freezing in liquid ethane followed by FS in acetone containing 0.01% OsO4 and 0.25%
UrAc. Bar, 2.5µm. Tomographic slice showing a single microtubule in the periphery of a sectioned interphase mouse fibroblast prepared by (C) chemical fixation as in (A), (D) cryo-fixation/freeze substitution as in (B). (E) ~25nm thick tomographic slice showing a microtubule in the periphery of a whole (not sectioned) mouse interphase fibroblast which was plunge-frozen and directly studied by cryo-EM. (F-H) ~25nm thick tomographic slice of an in vitro assembled microtubule (F) cryo- immobilized by plunge-freezing and directly studied by cryo-EM . Magnification is the same in (C-F).
Bar, 50µm. (G) Micrograph of in vitro assembled MT which were cryo-immobilized, chemically fixed, plastic-embedded and sectioned, as in (A,C), or treated like in (B, D) Note the irregularities in the MT wall structure in samples prepared by chemical fixation at room temperature (A, C, G) compared to (B, D, E, F, H). Arrows point to the MTs. With courtesy to dr. Roman Koning for images E and F.
Some cell types have an extremely large and thin periphery with a thickness in the range of a few hundreds of nanometers and less. A major advantage of these cells is that they can be used for investigation and visualization of their ultrastructure by means of cryo- EM without the need for sectioning (Garvalov et al., 2006; Bouchet-marquis et al., 2007);
see also Chapter IV from this thesis). Thin cells are easily vitrified by plunge-freezing, resulting in excellent preservation of the ultrastructure comparable to its native living state (no dehydration, no chemical fixation and plastic infiltration). In the vitrified samples no resolution limitations are set by the stain distribution. However, a number of aspects regarding cryo-EM of the whole vitrified cell have to be taken into account:
1. Cryo-EM is limited to the thinner parts of the vitrified cell (a few hundred nanometers thick or less). A thinner sample will result in improved resolution and a reduction of noise in the tomogram. In addition, to achieve these benefits the ice layer in which the cell is embedded also needs to be thin, which can be achieved by optimizing the procedure for blotting the grids prior to the vitrification.
2. The sample (ice) is sensitive to electrons, thus working with low-dose settings is necessary. For the same reason, when performing ET on these samples the total electron-dose needs to be divided over the number of projections that needs to be acquired, which often results in a lower tilt angle range and/or number of tilts /larger increments. In turn this leads to resolution degradation.
3. Vitrified whole-cell samples are not stained, so only intrinsically electron-dense structures can be observed. However, the intrinsic density (contrast) of many cell structures is quite low. The consequence is that the investigation is performed more or less blind. The structures are better visible when a larger defocus is applied.
4. Inelastic scattering of the electrons results in high noise levels in the cryo-EM specimens. Zero-loss energy filters can be applied to eliminate the inelastic scattered electrons. Also, in many cases post-acquisition processing (de-noising) can be performed.
45
MT plus-end structure visualization by EM/ET
A detailed overview of the previous studies performed on the subject of MT plus-end structures can be found in this thesis, Chapter I. Briefly, the first studies on the structure of the MT ends were performed on in vitro assembled or extracted MTs, by using the negative stain and cryo-EM approach. These studies reported three different plus-end structures: sheet-like, blunt and frayed morphology. In situ research mainly focused on the kinetochore-associated MT plus-ends (kMTs) investigating the dynamics of MT ends during various phases of mitosis. These in situ studies introduced a number of new morphologically distinct ends (Austin et al., 2005; VandenBeldt et al., 2006; Höög et al., 2007). Recently, a whole-cell reconstruction study published by Höög and colleagues confirmed the existence of a number of these plus-end conformations at the tips of peripheral MTs in yeast cell (Höög et al., 2007).
In our study we used two different approaches to visualize MT plus-ends in cells:
1. EM/ET using plastic sections of cryo-fixed/freeze-substituted/flat-embedded 3T3 and MEF cells
2. Cryo-EM/ET using vitrified whole MEF cells.
In both approaches the MT plus-ends were scored in the peripheral area of the cells, followed by characterization in 2D. Some of the various MT plus-ends found in the periphery of cryo-fixed/freeze-substituted/resin embedded and sectioned cells are shown in Figure 4. Based on the morphological features characterizing the plus-ends, the scored ends were divided in groups. Doing this, we were able to identify nine distinct plus-end conformations in sectioned 3T3 samples (Zovko et al., 2008).
While attempting to visualize MT plus-ends in the cell periphery of interphase fibroblasts, a number of technical aspects needed special attention:
1. Cytoplasm of the interphase mammalian cells is filled with electron dense structures (free ribosome, cortical actin fibers, intermediate fibers etc.) which hinder the visualization of the MT plus-ends in 2D.
In order to circumvent the obstructing effects of surrounding objects on the MT plus-end visualization in the cryo-fixed/freeze-substituted samples we attempted to reduce the visibility of the objects surrounding MT plus-end by adapting the staining protocol. For this purpose we used relatively low percentages of osmium tetraoxide (0.01%-0.1%) and uranyl acetate (0.25%-0.5%) in the freeze-substitution medium which resulted in a fairly light cytoplasmic staining, while the MTs were still properly visible. Some staining protocols include tannic acid (VandenBeldt et al., 2006) which specifically stains cytoskeletal structures. However, in our case it was important to reduce the staining of the abundant cortical actin filaments, with no need for tannic acid.
46
Figure 4. Microtubule plus- ends gallery. Here the 2D images of various MT plus- ends are presented. These were found in the periphery of cryo-fixed/ freeze-substituted, flat-embedded interphase mouse fibroblasts were visualized with conventional (RT) electron microscopy.
Each micrograph (on the left) is accompanied by the same micrograph (on the right) that has been image-processed to make it easier to interpret (inverted + Gaussian blur).
The plastic embedded cells were cut parallel to the plane of the attachment to the substratum into approximately 100nm thin slices. The electron microscope used for visualization was a CM-10 (Phillips) operating at 80kV.
To confirm the plus-end conformations we identified in 2D, and to improve the visualization of the plus-end structures, we performed ET. An example of an area in the periphery of a mouse fibroblast is shown in Figure 5. In order to dissect the volume of, here, 100nm, optical slices of for example 1 nm in thickness are “sectioned”. This allows observing the features situated within a volume in more detail (Figure 5A-C”).
By means of ET we were able to gain information on the 3D arrangement of the structures around the plus-end, adding to a better identification of the plus-end features themselves. Figure 6A shows a plus-end and its features observed in chronological slices when scrolling through the tomographic volume. Manual tracing of the plus-end features in each sequential optical section gives rise to a 3D model depicted in Figure 6B and 6C.
2. The MT plus-ends are often cut off when sectioning due to the fact that 1) the MTs are not entirely oriented parallel to the plane of section and often contact with their plus-ends the structures at the ventral side of the cell; 2) it is extremely difficult to section the cells absolutely parallel to their substrate.
Consequently, it is extremely laborious to find a plus-end in 2D. To compensate for this effect, many cells and sections have to be scanned. A solution to this problem is to perform tomography of thicker sections or whole-cells.
47
Figure 5. Searching for a microtubule and its plus-end in 2D and 3D. An example of an area in the periphery of an interphase mouse fibroblast demonstrating the poor visibility of the microtubules and microtubule plus-ends in a 100nm plastic section due to the dense cytoplasm around them. The arrowheads point towards the plus-ends. (A) Zero-angle projection of a 100nm thick section (B) a 5nm thick optical slice from the tomographic reconstruction of the same spot as in (A) with MTs clearly visible (C) a 1nm thick optical slice from the tomographic reconstruction of the same spot, (C‟ and C”) higher magnification of the microtubule plus-ends from A, B and C. The visibility is highly improved through tomographic reconstruction allowing detailed analysis of the microtubule plus-end features. (Koning et al., 2008)
3. The plus-ends in the periphery of an interphase mammalian cell are scattered over a relatively large area which needs to be searched systematically in order to find a plus-end. This process of finding plus-ends is for this reason particularly laborious. Moreover, in order to properly visualize the plus-end morphology by EM or ET magnifications of 20.000x-30.000x need to be used. Consequently, when recording a micrograph, a relatively small area of the cell is covered, which often yields only one MT plus-end in the micrograph/tomogram.
These issues possibly form the major reasons why the previous MT plus-end research has focused mainly on either kMTs (VandenBeldt et al., 2006) or the peripheral MTs of an extremely small cell like budding yeast (Höög et al., 2007).
48
4. For ET, the orientation of a MT within the specimen plays an important role. In order to gain adequate resolution and contrast of the MT walls with ET, the MTs oriented with their long axis more or less parallel to the tilt axis need to be avoided due to the missing wedge phenomenon (McEwen and Marko, 2001).
5. Locating a particular plus-end after rotating the grid 90 degrees, in order to perform dual-axis ET, is in most of the cases impossible due to the size and thus visibility of a MT plus-end when using low-magnifications. This problem could, in principle, be solved by using finder-grids with slots which do not interfere with tilting the grid to higher tilt-angles.
Despite the challenges mentioned above, we were able to 1) visualize and investigate the features characterizing the plus-ends of peripheral MTs in 2D (Figure 4 and Figure 2) confirm their existence using 3D electron tomography approach in both sectioned interphase 3T3 fibroblasts and mouse embryonic fibroblasts (MEFs) (Figure 6 and Figure 7). For a detailed study of MT plus-ends in cryo-preserved/freeze substituted/ sectioned 3T3 cells see Chapter III (Zovko et al., 2008) and for the cryo-EM/ET whole cell study in MEFs see Chapter IV (Koning et al., 2008) of this thesis.
One of the promising recent developments on the subject of visualization of plus-end structures is the implantation of the automatic recognition and extraction of the MT plus- ends within a tomographic volume which would enable a quicker and more objective analysis of the plus-ends inside a tomographic volume (Jiang M, 2006).
49
Figure 6. Visualization of a microtubule plus-end using electron tomography. (A) Nine sequential
~1nm thick optical sections (slices) extracted from a tomographic volume depicting a microtubule plus-end. In the upper part of the first slice three fiducial gold particles are visible used for alignment of the images. The sequential slices are to be read from left to right, continuing in the next row. This particular plus-end is characterized by curving protofilaments (arrowheads) situated at different plus-end heights. (B and C) A 3D representation derived from the same tomogram as in (A) showing the plus-end (anterior and posterior view, respectively). The x,y, and z axis show the position of the plus-end within the examined plastic-section. The image segmentation and surface rendering were performed using AMIRA® 3D visualization software (TGS). The plus-end features were manually traced marking their contours on each of the successive optical slices of the volume (in XY, YZ and XZ direction). All the contours together are then stacked giving a rise to a 3D model.
50
Figure 7. Visualization of MT plus-ends by means of electron tomography in situ. A ~5nm thick optical tomographic slice (on the left) is accompanied by the same slice (on the right) that has been image-processed (Photoshop, inverted + Gaussian blur) to make it easier to interpret the features.
Briefly, mouse embryonic fibroblasts were grown on coverslip, cryo-fixed/ freeze-substituted, flat- embedded and sectioned parallel to the plane of attachment to the substratum. Next, the ~100nm thick plastic sections were investigated with an EM. Single axis tilt-series of various MT plus-ends were recorded tilting the specimen holder from -60˚ to +60˚, with increments of 1 in a Tecnai-12 (FEI) equipped with a LaB6 filament operating at 120 kV. 2kx2k binned images were recorded in focus using a CCD camera (4k Eagle) at a magnification of 30 000x. Tomographic tilt series were analyzed and processed using IMOD (Colorado University, Boulder).
51
AppendixThe detailed protocols used in the above mentioned experiments are described below.
Specimen preparation by chemical fixation
The mouse fibroblasts (3T3s) were grown in a Petri dish for 2 days in a DMEM supplemented with 10% fetal bovine serum, 2mM L-Glutamine and antibiotics, in a humidified atmosphere with 5% CO2 at 37°C. Pre-warmed double concentrated solution of glutaraldehyde in 0.2M PHEM buffer was added to the cells with medium reaching the end concentration of glutaraldehyde of 1.5%. The cells were incubated at 37˚C in culture chamber for one hour. Next the cells were washed twice with 0.1M PHEM buffer and subsequently incubated with 1% OsO4/0.1M PHEM buffer at 4˚C for one hour. The cells were then washed twice with 0.1M PHEM buffer, followed by wash steps with 70%
ethanol. The cells were further dehydrated using ethanol-series (80-100% ethanol). The cells were subsequently infiltrated with ethanol and plastic (Epon) mix 1:1 for one hour, followed by infiltration with pure Epon for one hour. The cells were then flat embedded by putting an Epon-filled capsule onto the cells. The Epon was allowed to polymerize for two days at 60˚C. Next the capsules were broken off from the plate leaving the cells on top of the flat surface of the capsule. Sections of 100 were made parallel to the cell attachment plane. The samples were post-stained with uranyl acetate and lead-salt mixture for 15 and 5 minutes respectively.
In vitro grown MTs in glycerol (kindly provided by Dr. Roman Koning) were pelleted for 5 minutes at 1000 rpm. The pellet was then fixed as described above for the cells, except that the fixation with glutaraldehyde was at RT for 30 minutes and with OsO4 for 30 minutes. Dehydration, resin-infiltration, sectioning and post-staining were performed as described above for the cells.
Specimen preparation by cryo-fixation and freeze substitution
Fibroblasts were grown on 13mm diameter Thermanox (NUNC) coverslips (3T3s) or on carbon-coated formvar-supported golden EM-grids (MEFs) for two days in a DMEM supplemented with 10% fetal bovine serum and antibiotics (3T3 cells) or in a 1:1 mixture of Dulbecco's modified Eagle's medium (DMEM) (GIBCO, Invitrogen) and Ham‟s F10 medium (Cambrex), supplemented with 2mM L-Glutamine, 10% fetal calf serum and antibiotics (MEF cells), in a humidified atmosphere with 5% CO2 at 37°C. The cells were cryo-fixed by plunge-freezing the coverslip/EM-grid in liquid ethane in an automated system with a temperature and humidity controlled chamber.
Cryo-immobilized samples on coverslips were further subjected to freeze-substitution in a mixture of OsO4 (0.01% or 0.1%) and uranyl acetate (0.25%) in absolute acetone (Figure 3, Figure 4 and Figure 6; Figure 7 with the exception of images 1,2,3,8) or in acetone containing OsO4(1%), uranyl acetate (0.5%) and 1% H2O (Figure 5 and Figure 7, images 1,2,3,8). Freeze-substitution was performed in Automatic Freeze Substitution apparatus (AFS, Leica) for 72 hours at -90°C. The temperature was then gradually raised (10°/h) to -20°C and maintained at this temperature for 12 hours. Afterwards, the temperature was raised to 0°C (10°/h). Subsequently, cells were washed twice for 15 minutes with acetone. The specimens were then infiltrated with epoxy resin and flat-
52
embedded as described above. Sections of 100 were made parallel to the cell attachment plane. The samples were post-stained with uranyl acetate and lead-salt mixture for 15 and 5 minutes respectively.
In vitro grown MTs in glycerol (kindly provided by Dr. Roman Koning) were pelleted for 5 minutes at 1000 rpm. The pellet was then smeared out very fast onto a coverslip which was immediately plunge-frozen and further treated as described above for the cells on coverslips.
Electron microscopy and electron tomography
Thin sections of chemically fixed and cryo-fixed/freeze-substituted specimens were analyzed using a CM-10 Transmission Electron Microscope with a LaB6 filament operating at 80kV.
The MT plus-ends in the area of maximally 5 µm and ~10µm from the plasma membrane towards the cell interior were scored in 3T3 cells MEFs, respectively.
Using Inspect 3D (FEI Company) software, tomographic series of MT plus-ends were collected performing single-axis tilt series from –60 to +60 degrees with increments of 1
in a Tecnai-12 (FEI) equipped with a LaB6 filament operating at 120 kV. 2kx2k binned images were recorded in focus using a CCD camera (4k Eagle) at a magnification of 30 000x. Tomographic tilt series were analyzed and processed using IMOD (Colorado University, Boulder). Projection images were preprocessed by hot pixel removal and roughly aligned by cross-correlation. Final alignment was performed using 10nm colloidal gold as fiducial markers. Tomograms were obtained using weighted back-projection.
The plus-end features were segmented manually by contour tracing in each successive optical section of a tomogram in xy,yz and xz directions using AMIRA® 3D visualization software (Mercury Computer Systems, Merignac, France). With SurfaceView module the contours are stacked into a 3D object, as represented for a plus-end in Figure 6B and 6C.
53
ReferencesAustin, J.R., II, Segui-Simarro, J.M., and Staehelin, L.A. (2005). Quantitative analysis of changes in spatial distribution and plus- end geometry of microtubules involved in plant-cell cytokinesis. J Cell Sci 118, 3895- 3903.
Bouchet-marquis, C.d., Zuber, B.t., Glynn, A.-M., Eltsov, M., Grabenbauer, M., Goldie, K.N., Thomas, D., Frangakis, A.S., Dubochet, J., and Chrétien, D. (2007).
Visualization of cell microtubules in their native state. Biology of the Cell 099, 45-53.
Chretien, D., Fuller, S.D., and Karsenti, E.
(1995). Structure of growing microtubule ends: two-dimensional sheets close into tubes at variable rates. J. Cell Biol. 129, 1311-1328.
Ferenz, N.P., and Wadsworth, P. (2007).
Prophase Microtubule Arrays Undergo Flux- like Behavior in Mammalian Cells. Mol. Biol.
Cell 18, 3993-4002.
Garvalov, B.K., Zuber, B., Bouchet-Marquis, C., Kudryashev, M., Gruska, M., Beck, M., Leis, A., Frischknecht, F., Bradke, F., Baumeister, W., Dubochet, J., and Cyrklaff, M. (2006). Luminal particles within cellular microtubules. J. Cell Biol. 174, 759-765.
Giddings TH Jr, O.T.E., Morphew M, Mastronarde DN, McIntosh JR, Winey M.
(2001). Using rapid freeze and freeze- substitution for the preparation of yeast cells for electron microscopy and three- dimensional analysis. Methods Cell Biol. 67, 17.
Höög, J.L., Schwartz, C., Noon, A.T., O'Toole, E.T., Mastronarde, D.N., McIntosh, J.R., and Antony, C. (2007). Organization of Interphase Microtubules in Fission Yeast Analyzed by Electron Tomography.
Developmental Cell 12, 349-361.
Jiang M, J.Q., McEwen BF. (2006).
Automated extraction of fine features of kinetochore microtubules and plus-ends from electron tomography volume. IEEE Trans Image Process. 15, 13.
Komarova, Y.A., Akhmanova, A.S., Kojima, S.-i., Galjart, N., and Borisy, G.G. (2002a).
Cytoplasmic linker proteins promote microtubule rescue in vivo. J. Cell Biol. 159, 589-599.
Komarova, Y.A., Vorobjev, I.A., and Borisy, G.G. (2002b). Life cycle of MTs: persistent growth in the cell interior, asymmetric transition frequencies and effects of the cell boundary. J Cell Sci 115, 3527-3539.
Koning, R.I., Zovko, S., Bárcena, M., Oostergetel, G.T., Koerten, H.K., Galjart, N., Koster, A.J., and Mieke Mommaas, A.
(2008). Cryo electron tomography of vitrified fibroblasts: Microtubule plus ends in situ.
Journal of Structural Biology 161, 459-468.
Koster, A.J., Grimm, R., Typke, D., Hegerl, R., Stoschek, A., Walz, J., and Baumeister, W. (1997). Perspectives of Molecular and Cellular Electron Tomography. Journal of Structural Biology 120, 276-308.
Kremer, J.R., Mastronarde, D.N., and McIntosh, J.R. (1996). Computer Visualization of Three-Dimensional Image Data Using IMOD. Journal of Structural Biology 116, 71-76.
Krylyshkina, O., Anderson, K.I., Kaverina, I., Upmann, I., Manstein, D.J., Small, J.V., and Toomre, D.K. (2003). Nanometer targeting of microtubules to focal adhesions. J. Cell Biol. 161, 853-859.
Mandelkow, E.-M., and Mandelkow, E.
(1985). Unstained microtubules studied by cryo-electron microscopy : Substructure, supertwist and disassembly. Journal of Molecular Biology 181, 123-135.
Mastronarde, D.N. (1997). Dual-Axis Tomography: An Approach with Alignment Methods That Preserve Resolution. Journal of Structural Biology 120, 343-352.
McEwen, B.F., and Marko, M. (2001). The Emergence of Electron Tomography as an Important Tool for Investigating Cellular Ultrastructure. J. Histochem. Cytochem. 49, 553-564.
McIntosh, J.R. (2001). Electron Microscopy of Cells: A New Beginning for a New Century. J. Cell Biol. 153, 25F-32.
54
McIntosh, R., Nicastro, D., and Mastronarde, D. (2005). New views of cells in 3D: an introduction to electron tomography. Trends in Cell Biology 15, 43-51.
Mimori-Kiyosue, Y., Shiina, N., and Tsukita, S. (2000). The dynamic behavior of the APC-binding protein EB1 on the distal ends of microtubules. Current Biology 10, 865- 868.
O'Toole, E.T., McDonald, K.L., Mantler, J., McIntosh, J.R., Hyman, A.A., and Muller- Reichert, T. (2003). Morphologically distinct microtubule ends in the mitotic centrosome of Caenorhabditis elegans. J. Cell Biol. 163, 451-456.
Stepanova, T., Slemmer, J., Hoogenraad, C.C., Lansbergen, G., Dortland, B., De Zeeuw, C.I., Grosveld, F., van Cappellen, G., Akhmanova, A., and Galjart, N. (2003).
Visualization of Microtubule Growth in Cultured Neurons via the Use of EB3-GFP (End-Binding Protein 3-Green Fluorescent Protein). J. Neurosci. 23, 2655-2664.
VandenBeldt, K.J., Barnard, R.M., Hergert, P.J., Meng, X., Maiato, H., and McEwen, B.F. (2006). Kinetochores Use a Novel Mechanism for Coordinating the Dynamics of Individual Microtubules. Current Biology 16, 1217-1223.
Wade, R.H., and Chrétien, D. (1993).
Cryoelectron Microscopy of Microtubules.
Journal of Structural Biology 110, 1-27.
Zovko, S., Abrahams, J.P., Koster, A.J., Galjart, N., and Mommaas, A.M. (2008).
Microtubule Plus-End Conformations and Dynamics in the Periphery of Interphase Mouse Fibroblasts. Mol. Biol. Cell 19, 3138- 3146.
