JOURNAL OF ELECTRON MICROSCOPY TECHNIQUE 18:61-73 (1991)

The Three-Dimensional Architecture of the Mitotic Spindle, Analyzed by Confocal Fluorescence and Electron Microscopy ANDREAS MERDES, ERNST H.K. STELZER, AND J A N DE MEY European Molecular Biology Laboratory, Heidelberg, Germany

Fluorescence microscopy techniques, Poleward chromosome movement, MicroKEY WORDS tubule dynamics Fluorescence microscopy techniques have become important tools in mitosis reABSTRACT search. The well-known disadvantages of fluorescence microscopy, rapid bleaching, phototoxicity and out-of-focus contributions blurring the in-focus image are obstacles which still need to be overcome. Confocal fluorescence microscopy has the potential to improve our capabilities of analyzing cells, because of its excellent depth-discrimination and image processing power. We have been using a confocal fluorescence microscope for the study of the mechanism of poleward chromosome movement, and report here 1) a cell preparation technique, which allows labeling of fixation sensitive spindle antigens with acceptable microtubule preservation; 2) the use of image processing methods to represent the spatial distribution of various labeled elements in pseudocolour; 3) a novel immunoelectron microscopic labeling method for microtubules, which allows the visualization of their distribution in semithin sections at low magnification; and 4)a first attempt to study microtubule dynamics with a confocal fluorescence microscope in living cells, microinjected with rhodamine labeled tubulin. Our experience indicates that confocal fluorescence microscopy provides real advantages for the study of spatial colocalization of antigens in the mitotic spindle. It does not, however, overcome the basic limits of resolution of the light microscope. Therefore, i t has been necessary to use a n electron microscopic method. Our preliminary results with living cells show that it is possible to visualize the entire microtubule network in stereo, but that the sensitivity of the instrument is still too low to perform dynamic time studies. It will be worthwhile to further develop this new type of optical instrumentation and explore its usefulness on both fixed and living cells.

INTRODUCTION Recent mitosis research has focused on the dynamics of microtubule subpopulations, the sites of microtubule assemblyidisassembly in the spindle, and the interactions of microtubules with each other and with the kinetochores (McIntosh and Koonce, 1989; Mitchison, 1988). The use of digital imaging fluorescence microscopy techniques has been particularly valuable. For example, fluorescence recovery after photobleaching (FRAP) of microinjected fluorescent tubulin, analyzed by low light level microscopy, has been used to demonstrate the astonishingly high microtubule dynamics in mitosis (Salmon et al., 1984). In addition, photobleaching was also used to show that kinetochore microtubules during anaphase poleward movement were shortened by loss of subunits a t the kinetochore (Gorbsky et al., 1988). Finally, the dynamic rearrangements of actin filaments and microtubules have been followed second by second in early syncytial Drosophila embryos (Kellogg et al., 1988). Although these studies excel by their elegance, they suffer from the major problem of conventional fluorescence microscopy: the out-of-focus contribution of ex-

0 1991 WILEY-LISS, INC

cited fluorescent molecules blurring the information contained in the plane of focus. This pertains particularly to studies involving the various stages of the mitotic cycle, since the spindle undergoes continuous alterations and reorganizations. The spindle contains subpopulations of microtubules of distinct dynamic behaviour and half-life (Cassimeris et al., 1988; Mitchison, 1988).The microtubules of chromosomal fibers (for terminology see Cassimeris et al., 1988) vary in number during the different phases and form a minority relative to the polar microtubules. For example, using FRAP, no poleward tubulin flux could be detected in metaphase cells (Cassimeris et al., 1988). However, such a flux has convincingly been shown to exist by the use of photoactivation of a carboxyfluorescein derivative of tubulin microinjected into the cell (Mitchison, 1989). This has been possible, because the technique of photoactivation is better suited to follow fluorescent

Received March 5, 1990; accepted in revised form July 26, 1990. Address reprint requests to Dr. J a n De Mey, European Molecular Biology Laboratory, Cell Biology Program, Postfach 10.2209, Meyerhofstr. 1, D-6900Heidelberg, Federal Republic of Germany.

62

A. MERDES ET AL.

tubulin incorporated in stable microtubules. The opposite approach of monitoring the movement of a bleached bar in a fluorescent spindle (Gorbsky et al., 1988) could only be applied to a cell type in which the chromosomal fibers composed the bulk of the spindle microtubules during anaphase, when the fibers only shorten. It could not be applied to the much more labile chromosomal fibers in a n early prometaphase spindle. Immunofluorescence of tubulin labeled spindles is not entirely satisfactory, because it is difficult to detect details of microtubule distribution inside the very highly fluorescent spindle. To overcome this, the population of labile microtubules has been removed by lysis in Ca2 containing buffers (Gorbsky et al., 1987). This solution is neither optimal nor generally applicable. In some cases, non-fluorescent methods such as immunogold staining are superior (De Mey et al., 1982) and have been used in conjunction with high resolution contrast enhancement video microscopy (Inoue e t al., 1985) and Nanovid microscopy (Geuens et al., 1989). It has recently been shown that confocal fluorescence microscopy can produce images of immunolabeled spindles in a single focal plane, containing much sharper structural information (White et al., 1987; Wright et al., 1989). The optical instruments and the accompanying image processing capabilities have since been developed into powerful tools (Pawley, 1989). For example, used on living cells microinjected with fluorescent derivatives of tubulin or microtubule associated proteins (MAPS), confocal microscopy could in principle lead to more accurate measurements and imaging. So far, however, the available instruments lack the necessary sensitivity for that purpose. Confocal microscopy has not previously been used to address a specific problem in spindle biology. We have studied the mechanism of poleward chromosome movement by analyzing the microtubule interaction with moving kinetochores in prometaphase cells fixed immediately after nuclear envelope breakdown. We have worked out a correlative approach on cultured rat kangaroo cells (PtK,) in which chromosome movement in living cells was followed by video light microscopy, and the same cell was analyzed after fixation and staining of microtubules, centromeres, and chromosomes with confocal fluorescence microscopy. These procedures were used to analyze single optical sections and to calculate three-dimensional stereo images of whole cells or parts thereof (Rosa et al., 1989) in which the differently labeled structures were simultaneously visible in pseudocolour. A novel immunoelectron microscopy method has enabled us to make the same analysis a t the ultrastructural level, and to compare the resolution of the light and electron microscope. In this paper, we want to describe the procedures we have developed, and discuss the potential and limitations of confocal microscopy for mitosis and microtubule research in thicker cells in general. +

1%non-essential amino acids, 1%sodium pyruvate, and antibiotics (Gibco, Paisley, Scotland, UK).

Video Microscopy For phase contrast observations, coverslips were mounted in a perfusion chamber of our own design, allowing microscopy with high numerical aperture lenses and rapid fixation of mitotic cells on the microscope stage. Cells were observed with a Leitz 63 x 11.3 NA, Phaco 3 lens on a Leitz Orthoplan microscope. A panchromatic green filter (Leitzj was used to protect cells from deleterious illumination. Observations were made a t room temperature (23-24°C). To record video time lapse series of mitotic events, a video camera (model 81, Dage-MTI) was connected to the microscope via a zoom lens (5-12.5 x ; Leitz) to increase magnification. Images were recorded on video tape with a National VTR NV-8030 time lapse recorder. Micrographs were taken by photographing the images with a prefocused camera from a flat screen monitor (Knott Elektronik, FRG).

Immunofluorescence Cells on coverslips were fixed with a formaldehyde procedure comprising a pH-shift, modified from Berod et al. (1981) and Bacallao and Stelzer (1989). The cells were perfused first with 3 8 formaldehyde (Art. 4001; Merck, Darmstadt, FRG) in PEM (80 mM Pipes, 5 mM EGTA, 2 mM MgCl,, pH 6.5) for 1 minute a t room temperature. Then, the perfusion was continued with 3% formaldehyde in 100 mM Na,B,07, pH 11, for 8 minutes. The cells were then removed from the perfusion chamber and washed three times for 5 minutes with phosphate-buffered saline (PBS), treated twice with freshly prepared sodium borohydride (1mgiml) in PBS, pH 8, for 15 minutes, washed with PBS, and permeabilized twice with 0.2% Triton X-100 in PBS for 5 minutes. They were equilibrated in immunocytochemistry (ICC) buffer (PBS, pH 7.4, containing 0.8% bovine serum albumin (BSA) and 0.1% gelatin) for 5 minutes at room temperature. Microtubules and centromeres were labeled with a mixture of monoclonal mouse antialpha and anti-beta tubulin (Amersham), both diluted 1:500, and a human autoimmune serum from a patient with CREST syndrome, diluted 1:300 (gift from R. Bischoff and H. Ponstingl, German Cancer Research Center, Heidelberg, FRG), for 90 minutes at 37°C. After washing three times for 10 minutes in ICC buffer, cells were incubated with a mixture of fluorescein isothiocyanate (FITC) labeled goat anti-mouse and anti-human IgG antibodies, diluted 1:100 each (Jackson). The primary and secondary antibodies were diluted in ICC buffer containing 1%normal goat serum (Janssen, Olen, Belgium) and 0.1% Triton X-100. The cells were washed three times for 10 minutes in ICC buffer, and rinsed three times in PBS. They were postfixed with 4% paraformaldehyde in PBS for 30 minutes and quenched with 50 mM NH,Cl for 15 minutes. MATERIALS AND METHODS Chromosomes were counterstained with propidiumioCell Culture dide (Sigma) a t 2.5 Fgiml in 10 mM Tris-buffered saRat kangaroo PtK, cells were grown on 20 mm square line (150 mM NaC1) a t pH 7.6 for 15-30 minutes. After coverslips a t 37°C in modified Eagle's medium (MEM) washing in PBS, samples were mounted on glass slides supplemented with 10% fetal calf serum, 1% glutamin, with four spacers, made with acrylic nail polish, in 50%

CONFOCAL MICROSCOPY OF THE MITOTIC SPINDLE

63

Fig. 1. Comparison between A conventional and B: a single confocal optical section. Both images were taken at the same plane of focus. (The conventional image was taken on a Zeiss-Axiophot with a Plan-Neofluor 100 x11.3 lens.) x 1,600.

glycerol in twice concentrated PBS with 100 mg/ml DABCO (1,4 diazabicyclo (2,2,2) octane; Sigma) as a n antifading agent (Bacallao et al., 1989).

Microinjection of Rhodamine Tubulin Rhodamine tubulin was a gift of T. Kreis, European Molecular Biology Laboratory (EMBL), Heidelberg, FRG. Preparation of rhodamine tubulin and microinjection have been performed according to Kreis (1987). Microinjected living cells were observed in a chamber filled with cell culture medium. Confocal Microscopy The samples for light microscopy were investigated on a confocal beam scanning laser microscope (CBSLM), developed and constructed a t the EMBL (Stelzer et al., 1988). An Argon-ion laser (model 2016; Spectra Physics) was used in multi-line mode; distinct wavelengths of 476 nm (excitation of FITC labeled microtubules and kinetochores) and 528.7 nm (excitation of propidiumiodide labeled chromosomes) were selected by filter sets (Ploemopak). Confocal optical sections were recorded a t 0.4 Fm per vertical step, and four times averaging. An average cell required 15 optical sections. Images were composed of 512 x 512 pixels. For double immunofluorescence micrographs, each cell was scanned twice with the same coordinates for the confocal series. Stereo images of each series were obtained by pixel shift as previously described (Rosa et al., 1989). Double colour images were generated with the program “VIEW’ (Lawrence Livermore National Laboratory, University of California, adapted for the confocal microscope by Clemens Storz, EMBL, Heidelberg, FRG) on a VAX 3200 workstation, by superimposing green pseudocoloured FITC fluorescence images on red pseudocoloured propidiumiodide fluorescence images. Images were photographed from the screen of the workstation on Ektachrome 100 Plus film (Kodak).

Electron Microscopy Cells were fixed in 0.5% glutaraldehyde (Polysciences) in PHEM (60 mM Pipes, 25 mM Hepes, 1 mM EGTA, 2 mM Mg acetate, pH 6.9) for 10 minutes. They were subsequently washed and treated with 0.5% Triton X-100 in PHEM. After 15 minutes, cells were washed in PBS, pH 7.4, and treated twice with sodium borohydride (1 mglml) in PBS, pH 8, for 15 minutes. After washing with PBS, pH 7.4, they were blocked for 30 minutes at room temperature with a solution containing 0.8% BSA, 0.1% gelatin (Sigma G7765), and 5% normal goat serum in PBS, pH 7.4. Microtubules were labeled with a mixture of monoclonal mouse anti-alpha and anti-beta tubulin, as described for immunof luorescence. As secondary antibody, AuroProbe One goat anti-mouse, coupled with 1 nm colloidal gold particles (Janssen, Olen, Belgium), diluted 1:20 in PBS, pH 7.4, containing 0.8% BSA, 0.1% gelatin, and 1% normal goat serum, was used. Incubation was performed once with a n aliquot of 40 ~1 for 50 minutes at 37”C, and after adding 40 ~1of new solution overnight at room temperature. After washing in PBS, cells were postfixed with 2% glutaraldehyde in PBS for 5 minutes and with 0.2% tannic acid in PBS for 30 minutes. The cells were rinsed in excess distilled water. Silver amplification was performed according to Danscher (1981) and modified after Namork and Heier (1989) for 9 minutes in a darkroom under red safelight. After thoroughly washing the specimens in distilled water, the cells were treated with 0.5% osmium tetroxide in S@rensenbuffer for 10 minutes on ice. The cells were stained en bloc with 0.5% uranyl acetate and 1% phosphotungstic acid in 70% ethanol. Cells were dehydrated in ethanol and flat embedded in epon. Semithin serial sections (blue interference colour) were cut with a diamond knife, collected on formvadcarbon coated slot grids, stained with lead citrate for 1 minute (Reynolds, 1963) and observed in a Philips EM 301 electron microscope.

64

A. MERDES ET AL.

A

B

Fig. 2. Confocal series of a PtK, cell in metaphase. A. Microtubules. B: Chromosomes. Five representative frames of the complete confocal series (containing 15 frames) are shown. Stereo pairs ( L and R ) of each series were calculated. Stereo pairs: x 850

RESULTS A comparison between the conventional and the confocal fluorescent micrograph of the same cell (Fig. 1) clearly demonstrates the improvement of contrast obtained by the reduction of out-of-focus blur by confocal microscopy. The conventional micrograph (Fig. 1A) loses details in the equatorial plane of the mitotic spindle. The polar regions appear as white areas without further resolution. The confocal image (Fig. 1B) provides increased visibility of fine structures and a higher contrast in the same region. Aster microtubules, however, which extend into the cytoplasm of the cell, appear slightly broader and partially “beaded” in the confocal micrograph. The reason may be that the confocal image is composed of scanning lines and that the resolution therefore is restricted to a limited number of pixels (512 x 512). The principle of optical sectioning and the subsequent calculation of a pseudocoloured stereo image is shown in Figures 2 and 3. A confocal series of FITC labeled microtubules (Fig. 2A) was taken and stored on the computer. The same scanning coordinates were

used to record a second series of the chromosome staining (Fig. 2B) with another filter combination for excitation and emission. Both series have been used to calculate left and right images of a stereo pair, which have subsequently been superimposed. Each structure has been colour-coded, and the resulting pseudocolour images are shown in Figure 3A-H. High mechanical stability was required to ensure exact overlap of the images which were scanned at slightly different times. For the display of black and white images, a n 8 bit gray scale with 256 stepsipixel was used. For the display of double coloured images the values of the scale were reduced to 4 bitipixel. Consequently, a decreased number of gray steps resulted in a loss of details; especially fine structures, which were weakly displayed, appeared to have coarse surfaces. Figures 3A-H show PtKz cells in different steps of mitosis. For our investigations on the interactions between microtubules and kinetochores we were interested in a procedure which allowed simultaneous staining of both structures. Glutaraldehyde fixation was well suited for the preservation of microtubules, but

CONFOCAL MICROSCOPY OF THE MITOTIC SPINDLE

immunostaining of the centromeres was no longer possible, due to the strong cross-linking properties of the glutaraldehyde. It has been shown that the formaldehyde pH-shift method (Bacallao and Stelzer, 1989) allowed good staining of microtubules. Here we show that an antigen, which was sensitive to glutaraldehyde and located at the centromeres, could also be well preserved. We have observed this for other antigens as well (not shown). Figures 3A-C illustrates the transformation of the cytoplasmic microtubule network into the bipolar mitotic spindle. All the features of microtubule and chromosome dynamics which are known from previous studies can be seen with much improved detail. The fact that the whole cell can be seen in three dimensions is a real advantage. In the interphase cell (Fig. 3A), the microtubules form an irregular cytoplasmic network, whereas in early prophase (Fig. 3B), when the centrosomes have already split, a concentration in radial fibers, emanating from the two microtubule organizing centers, is visible. At the same time, the chromatin starts to condense, whereby the centromeric region can be seen as pairs of fluorescent spots. In the stereo view, the cytoplasmic regions of the cells appear flattened on the substratum; in the nuclear region, however, the cell is rounded up and microtubule fibers emanating from the two aster centers are shown to surround the top of the cell. We plan to perform a detailed study of the microtubule distribution in such cells. It is presently not known whether the microtubules of the two asters display interactions which could be of functional importance for the separation of the aster centers. In late prophase (Fig. 3C), the majority of microtubules is concentrated in the aster. The centrosomes have fully split and migrated to opposite poles of the cell. The chromosomes have further condensed; in the stereo view, even the relative spatial orientation of the two spots at the centromere is visible. At the onset of prometaphase, most chromosomes undergo a fast initial poleward movement. In Figure 3D, part of the chromosomes can be seen to be mono-oriented to the left, respectively the right spindle pole, whereas another part remained stationary in the equatorial plane of the spindle. The spindle has already developed, microtubules in the cytoplasm are depolymerized, and aster microtubules have drastically shortened. Later in prometaphase, all chromosomes are directed to the equator, now called the “metaphase plate,” and in Figure 3E the chromosomes are aligned in a bipolar orientation. Their centromeres are connected with the spindle poles via thick microtubule bundles. It can be observed that the proximal ends of the chromosomal fibers do not sharply focus on the centrosomes. This could mean that the minus ends of these microtubules are less well capped. The data on poleward fluxes of tubulin subunits (Mitchison, 1989) are consistent with this observation. These chromosomal fibers shorten during anaphase (Fig. 3F), after the chromatid pairs have split and moved towards the poles. Later, in anaphase b, the distance between the poles itself increases, while interdigitating interzonal microtubules between the poles become clustered in bundles (Fig.

65

3G). A contractile ring of actin is known to be formed around the equatorial region, leading to a constriction between the daughter nuclei. Thereby, the interzonal microtubules become concentrated and form the so called “midbody.”This midbody connects the daughter cells for a longer period. Meanwhile, new cytoplasmic microtubules have been formed (Fig. 3G), and during telophase, an extensive cytoplasmic network has been reconstituted (Fig. 3H). Little is known about the transition between prophase and early prometaphase. One of the questions we are addressing is how the interphase microtubule network depolymerizes when the new mitotic spindle is formed. We were particularly interested in analyzing the microtubule interaction with kinetochores during their first poleward movement, when they become mono-oriented to one pole, in the hope of understanding better the mechanism of poleward chromosome movement in the cell. To this end, a human CREST antiserum was used which recognizes components of the centromere. The exact localization of these components has not yet been determined. In view of recent results (Cooke et al., 1990) they are probably localized very close to the kinetochores and can therefore be used as a marker for their approximate localization at the light microscopic level. We will use the term “centromere” in the context of this paper. A series of phase contrast micrographs shows a PtK, cell in late prophase (Fig. 4A), when the nuclear envelope is still visible as a faint dark line (arrowheads). This dark line disappears a t the onset of prometaphase (Fig. 4B) when the first movement of the chromosomes is detected (compare Figs. 4B and 4C, arrows). The whole fixed and stained cell is shown in stereo in Figure 4D, E. Microtubules in the cytoplasm of the cell have depolymerized. A few microtubules extend from the polar regions of the spindle to the periphery of the cell, most of them as continuous lines. The majority of microtubules form part of the spindle asters, consisting of an extensive array of radially oriented short fibers. Preliminary results of Vandre and Borisy (1986) have indicated that cytoplasmic microtubules are lost abruptly at the time of nuclear envelope breakdown. The interaction between microtubules and centromeres was studied using single optical sections. Figure 5 shows a cell in very early prometaphase where a contact between a centromere and a polar microtubule fiber is visible (arrow). The distal end of the microtubule fiber clearly extends past the centromere, suggesting a lateral contact between the centromere and the fiber. According to the hypothesis of Koshland et al. (1988), chromosome movement should be driven by the depolymerization of microtubule plus-ends. The chromosome fiber should therefore end at the kinetochore. Recently, Rieder and Alexander (1990) have obtained strong evidence for a lateral sliding of the kinetochore along the chromosomal fiber, independently of the depolymerization of its microtubule plus-end (Rieder and Alexander, 1990). The question remains open, however, whether the fiber shown in Figure 5 really consists of a single mi-

A

B

C

Fig. 3. (Figure appears in color in Color Figure Section.) Double pseudocoloured stereo pairs of the field in Figure 2. Microtubules and kinetochores (green) and chromosomes (red) are shown in different stages of mitosis. The kinetochores are visible as small dots. A: interphase; B: early prophase; C: late prophase; D: prometaphase; E: metaphase; F: anaphase a; G: anaphase b; H telophase. A: x 710; B-H: x 850.

CONFOCAL MICROSCOPY OF THE MITOTIC SPINDLE

E

F

G

Fig. 3E-H.

67

68

A. MERDES ET AL.

D

Fig. 4. A-C: Phase contrast video images showing a PtK, cell undergoing transition from prophase (A) to prometaphase (B, C). In A, the nuclear envelope is still visible (arrowheads). In prometaphase, chromosome movement can be observed (B, C: arrows). The cell was fixed immediately after time point C. Confocal stereo micrographs of the fixed cell were calculated, showing D: the microtubule distribution and E: the chromosomes. A-C: x 1,200; D,E: x 1,000.

crotubule or whether it is built of several microtubules with one part ending a t the centromere and another part extending past it. To solve this problem, we developed a different approach using the electron microscope to follow microtubules over long distances a s was possible by immunofluorescence. To this end, microtubules were labeled

with anti-tubulin antibodies and with a secondary antibody coupled with 1 nm colloidal gold. The 1nm gold particles were enhanced with a silver lactate solution to increase the particle size. This small gold conjugate had good penetration properties, and provided a very dense and homogenous staining, Semithin sections of epon flat-embedded cells allowed us to view kineto-

CONFOCAL MICROSCOPY OF THE MITOTIC SPINDLE

69

dense than by immunofluorescence and there was soluble microinjected fluorescent tubulin in the cytoplasm. The fluorescent signal of the structures was relatively low, necessitating high photomultiplier gain, and therefore producing additional electronic noise. Four confocal frames were averaged to reduce this noise. Although single microtubules were visible, we weren’t able to follow microtubule dynamics over a longer time period, because the rhodamine, excited with a wavelength of 528.7 nm, was rapidly bleached. We conclude that experiments on living cells, using confocal microscopy, are limited by the sensitivity of the instrument used. In addition, the available wavelengths of the commercial argon-ion lasers (454 nm, 476 nm, 488 nm, 514 nm, and 528.7 nm) are not within the range which is optimal for rhodamine excitation (546 nm is normally used).

DISCUSSION Our present data confirm that the quality of immunof luorescence microscopy images can be greatly improved by the use of a confocal microscope (White et al., 1987; Wright e t al., 1989). Reduction of out-of-focus fluorescence results in significantly better images than chore microtubules all over their entire length between obtained by conventional microscopy and allows a detailed three-dimensional analysis of the whole cell or spindle pole and chromosome at low magnification. A cell in early prometaphase is shown at low mag- part of it. Confocal microscopy has major advantages nification (Fig. 6). Two sections a t different heights of for studying thick material such as histological secthe cell have been selected. In Figure 6A, the basal tions or cell types that develop into a polarized epitheregion is cut, showing that only a few microtubules lium and reach a cubic form. Madin-Darby canine kidhave penetrated the inner region of the nucleus. These ney cells (MDCK), grown on polycarbonate supports, fibers can clearly be identified as single microtubules. are a widely used example of the latter (Bacallao et al., A few chromosomes have already connected to the spin- 1989; Bomsel et al., 1989). These cannot be investidle poles, are mono-oriented to a single pole, and are gated with a conventional fluorescence microscope, due located near the centrosomal region. Figure 6B shows a to the fact that the filter support causes unacceptable region more on the top of the cell. In the middle of this background fluorescence and because of the enormous section, a chromosome is visible, which is already bi- height of the specimens. Confocal microscopy provides a powerful tool for copolarly and symmetrically oriented, i.e., chromosomal fibers to both poles are formed. These fibers consist of localization studies of antigens. Compared with conthick bundles of microtubules, whereby the single mi- ventional immunofluorescence, exact localization of crotubules in the fiber can still be resolved at the rel- the labeled structures in all three dimensions is possible. The relative spatial orientation of differently laatively low magnification. These data show that the approach using immu- beled antigens can be visualized by double pseudoconogoldkilver staining of semithin sections is well lour imaging as shown in this paper. The cell fixation/ suited for our morphological studies. It provides a good permeabilization procedure adapted here to cells grown overview of large areas in the cell and a n increased on coverslips has been shown to work in many cases for resolution compared with fluorescence light micros- antigens that are otherwise inactivated (Bacallao et copy. Stereo pictures of these semithin sections further al., 1989; Bomsel et al., 1989). It is further essential to increase the vertical resolution and help in making process the samples after immunostaining in a way which does not introduce geometrical distortions. three-dimensional reconstructions (not shown). In our studies on the mechanism of poleward chroIn another experiment we made a first attempt to investigate the dynamics of the microtubules in the mosome movement, the shortcomings of confocal f luoliving cell with the confocal microscope. Rhodamine rescence microscopy as much as its advantages have labeled pure tubulin was microinjected in the cyto- become apparent. Light microscopy is limited by its plasm of PtK, cells. Incorporation of the fluorescent resolution of 0.2 pm in the horizontal plane. Microtutubulin derivatives occurred within a few minutes. The bule fibers in mitotic spindels cannot be further recoverslips with cells were mounted in a chamber, sup- solved in the confocal microscope. There are no rigorplied with culture medium and investigated with the ous criteria for distinguishing single microtubules confocal microscope a t room temperature. Figure 7 from bundles of two or more. In the example shown in shows a stereo view of a PtKz cell in interphase. The Figure 5 , therefore, the question remains open, contrast of the microtubules appears to be lower than whether the observed chromosome fibers in early obtained on fixed immunolabeled specimens, since the prometaphase consist of a single microtubule or labeling of tubulin dimers with fluorophore was less whether they are formed by bundles of several microFig. 5. Single confocal frame of a R K , cell in early prometaphase, stained for microtubules, kinetochores, and chromosomes. The arrow indicates a polar chromosome fiber, extending past the kinetochore. x 1.600.

70

A. MERDES ET AL.

Fig. 6. Immunoelectron micrographs of a PtK, cell in early prometaphase, semithin sectioned. Microtubules were stained with l nm immunogold and silver enhanced. A: The base of the cell was cut. B: The same cell was cut a t the top of the nucleus. x 5,100.

CONFOCAL MICROSCOPY OF THE MITOTIC SPINDLE

71

Fig. 7. Confocal stereo pair of a living PtK, cell in interphase after microinjection of rhodamine tubulin. The stereo view was calculated of a focus series containing eight frames. x 1,000,

tubules. In cases where nanometer resolution is not opens the possibility of following microtubules over required, confocal microscopy represents a n interesting long distances, provided that the microtubules are link between conventional light microscopy and elec- stained with sufficient contrast. This was accomplished by immunostaining them with 1 nm gold particles and tron microscopy. Compared with electron microscopy, the confocal flu- subsequent silver amplification. This results in densely orescence microscope makes it possible to section sam- labeled microtubules, giving the impression of a n “anples optically instead of physically. With the collected alogue of immunofluorescence a t the electron microthree-dimensional data, it is possible to display vertical scopic level.” It is possible to use rigorous criteria to optical sections and stereo views. These developments distinguish single microtubules from bundles of two or represent considerable progress, since any optical sec- more. tion in any orientation through the cell can be calcuOur present experience leads us to conclude that conlated in a short time. In addition, it is possible to cal- focal fluorescence microscopy of fixed specimens is a culate projections that view the samples from different very powerful additional approach, especially when angles and to obtain a n animated view by displaying used in combination with other contemporary microsthem in rapid succession. Our experience indicates that copy approaches such as video microscopy, correlated this represents a novel and very useful way to visualize livingifixed cell analysis, microinjection, and electron the spatial aspects of the specimen by confocal micros- microscopy. The recent work of Rieder and Alexander copy. Unfortunately, it is not possible to document this (1990) has shown that it is possible to determine here. We were made aware of the possibility that the whether single microtubules are associated laterally examples presented here might leave the reader with with the kinetochores of mono-orienting prometaphase the impression that confocal microscopy does not pro- chromosomes using straightforward conventional techvide any additional information beyond an aesthetic niques. This was however only possible because an opthree-dimensional pseudocolor view, and the question timal material was used. Our study (to be published whether it is all worth the added effort and expense. In elsewhere) was conducted on normal mitotic cells, just the case of the examples shown here, this may be true entering prometaphase, and largely confirms their findto some extent although the aesthetic quality is useful ings. It would have been much more difficult to obtain for didactive purposes and the details of the spindle are our results without the use of the methods described shown with unprecedented clarity. It is often more use- here which will be useful for non-mitotic cells as well. ful to calculate three-dimensional data from a few careIt is not possible to observe living material with a n fully sequential optical sections only (not shown). electron microscope. Confocal fluorescence microscopy, The strongest argument in favour of the use of the however, is not limited in this way and holds tremenelectron microscope is its high resolution. However, dous potential for studies of this type. sectioning for electron microscopy is arduous and time Within the last few years, confocal microscopy has consuming. Another disadvantage of conventional thin developed to a routine technique for biologists, since sections is that a t the high magnification used only a the instrument is easy to use and software for image small area and volume of the cell can be viewed. In our processing and storage is available. Microinjection of study, it would have been very time consuming and fluorescent proteins into the living cell leads to the difficult to investigate the spatial orientation of single specific labeling of structures, the dynamics of which microtubules in this way. Since the light microscopic can be studied in the living state. Our first attempts in data were not yielding unambiguous results, a n ultra- the field of microtubule research, however, point to structural analysis was necessary. Semithin sectioning problems of sensitivity and speed of data collection

72

A. MERDES ET AL.

which will have to be overcome. A new “Modular Con- Berod, A., Hartman, B.K., Pujol, J.F. (1981) Importance of fixation in immunohistochemistry: Use of formaldehyde solutions at variable focal Microscope” (MCM) with improved speed and senpH for the localization of tyrosine hydroxylase. J . Histochem. Cysitivity has been constructed at the EMBL to overcome tochem., 29:844-850. these limitations. Bomsel, M., Prydz, K., Parton, R.G., Gruenberg, J., and Simons, K. (1989) Endocytosis in filter-grown Madin-Darby Canine Kidney With this microscope, it is possible to perform phocells. J . Cell Biol., 109:3243-3258. tobleaching or photoactivation experiments: illuminaA,, Jones, S.J., Taylor, M.L., Wolfe, L.A., and Watson, T.F. tion bars or spots can be exactly oriented and spatially Boyde, (1990) Fluorescence in the tandem scanning microscope. J. Milimited by vertically scanning single image lines. Obcrosc., 157:39-49. servations are made with much improved spatial pre- Cassimeris, L., Inoue, S., and Salmon, E.D. (1988) Microtubule dynamics in the chromosomal spindle fiber: Analysis by fluorescence cision, by collecting signals from one or a few succesand high-resolution polarization microscopy. Cell Motil. Cytoskelsive horizontal frames. For example, this use of eton, 10:185-196. confocal laser scanning microscopes has already Cooke, C.A., Bernat, R.A. and Earnshaw, W.C. (1990)CENP-B a mayielded very useful information on membrane transjor human centromere protein located beneath the kinetochore. J. Cell Biol., 110:1475-1488. port within the living Golgi apparatus of astrocytes Cooper, M.S., Cornell-Bell, A.M., Chernjavsky, A., Dain, J.W., and labeled with N-(7-[4-nitrobenzo-2-oxa-1,3-diazolel)-6Smith, S.J. (1990) Tuhulovesicular processes emerge from Transaminocaproyl sphingosine (NBD-ceramide) (Cooper e t Golgi cisternae, extend along microtubules, and interlink adjacent al., 1990). The signal intensity of these lipid derivatrans-Golgi elements into a reticulum. Cell, 61:135-145. tives is very high. Following labeled microinjected pro- Danscher, G. (1981) Histochemical demonstration of heavy metals. A revised version of the sulphide silver method suitable for both light teins will require more sensitive instruments. electron microscopy. Histochemistry, 71:l-16. For live experiments, the confocal scanning micro- Deand Mey, J., Lambert, A.M., Bajer, A.S., Moeremans, M., and De Brascope can also be used in other than the fluorescence bander, M. (1982) Visualization of microtubules in interphase and detection mode: differential interference contrast (DIC) mitotic plant cells ofHuemunthus endosperm with the immuno-gold staining method. Proc. Natl. Acad. Sci. IJSA, 79:1898-1902. with transmitted light, collected by a second detector, can be used to observe the whole cell, while the fluo- Geuens, G., Hill, A.M., Adoutte, A., and De Brabander, M. (1989) Microtubule dynamics investigated by microinjection of Paramerescence mode is used to observe specifically labeled cium axonemal tubulin: Lack of nucleation but proximal assembly structures. of microtubules at the kinetochore during prometaphase. J. Cell Biol., 108:939-953. For future use, further developments on the microscopy as well as the computational sector are expected: Goldstein, S.A., Hubin, T., Rosenthal, S., and Washburn, C. (1990) A confocal video-rate laser-beam scanning reflected-light microscope confocal microscopes operating under greater speed with no moving parts. J . Microsc., 157:29-38. (Boyde et al., 1990; Goldstein et al., 1990) are being Gorbsky, G.J., Sammak, P.J., and Borisy, G.G. (1987) Chromosomes developed. Experiments on living cells, which have move poleward in anaphase along stationary microtubules that coordinately disassemble from their kinetochore ends. J . Cell Biol., to be carried out at near video speed, require large 104:9-18. data storage and processing capacities. Fast computers, Gorbsky, G.J., Sammak, P.J., and Borisy, G.G. (1988) Microtubule for example Transputer systems, may become a predynamics and chromosome motion visualized in living anaphase ferred element in hardware conceptions. Last but not cells. J. Cell Biol., 106:1185-1192. least, improved, more laser-compatible fluorophores Inoue, S., Mole-Bajer, J., and Bajer, AS. (1985) Three-dimensional distribution of microtubules in Hemanthus endosperm cells. In: Miwill hopefully appear soon and help solve problems of crotubules and Microtubule Inhibitors. M. De Brabander and J . De bleaching, cytotoxicity, and quantum efficiency. If Mey, eds. Elsevier, Amsterdam, pp. 269-276. these circumstances can be fulfilled, three-dimen- Kellogg, D.R., Mitchison, T.J., and Alberts, B.M. (1988) Behaviour of microtubules and actin filaments in living Drosophzla embryos. Desional microscopy as a function of time may open new velopment, 103:675-686. horizons for microscopical studies in cell biology reKoshland, D.E., Mitchison, T.J., and Kirschner, M.W. (1988) Polesearch. wards chromosome movement driven by microtubule depolymeriza-

ACKNOWLEDGMENTS The authors thank Clemens Storz, Rainer Pick, Reiner Stricker, and Pekka Hanninen for their excellent technical work on the construction and the software design of the confocal microscope, and Thomas Kreis and Rainer Pepperkok for their help with the microinjection studies. Our special thanks to Dr. J. Turner for revising the English of our manuscript. REFERENCES Bacallao, R.. Antony, C., Dotti, C., Karsenti, E., Stelzer, E.H.K., and Simons, K. (1989) The subcellular organization of Madin-Darby Canine Kidney cells during the formation of a polarized epithelium. J. Cell Biol., 109:2817-2832. Bacallao, R., and Stelzer, E.H.K. (1989) Preservation of biological specimens for observation in a confocal fluorescence microscope and operational principles of confocal fluorescence microscopy. In: Methods in Cell Biology. A.M. Tartakoff, ed. Academic Press, San Diego, pp. 437-452.

tion in vitro. Nature, 331:499-504. Kreis, T. (1987) Microtubules containing detyrosinated tubulin are less dynamic. EMBO J., 6:2597-2606. McIntosh, J.R., Koonce, M.P. (1989) Mitosis. Science, 246:622-628. Mitchison, T.J. (1988) Microtubule dynamics and kinetochore function in mitosis. Ann. Rev. Cell. Biol., 4527-549. Mitchison, T.J. (1989) Polewards microtubule flux in the mitotic spindle: Evidence from photoactivation of fluorescence. J. Cell Biol., 109:637-652. Namork, E., Heier, H.E. (1989) Silver enhancement of gold probes (5-40 nm): Single and double labeling of antigenic sites on cell surfaces imaged with backscattered electrons. J. Electron Microsc. Techn., 11:102-108. Pawley, J. (1989) The Handbook of Biological Confocal Microscopy. IMR Press, San Antonio. Reynolds, E.S. (1963) The use of lead citrate at high pH as a n electron opaque stain in electron microscopy. J. Cell Biol., 17:208-212. Rieder, C.L., and Alexander, S.P. (1990) Kinetochores are transported poleward along a single astral microtubule during chromosome attachment to the spindle in newt lung cells. J. Cell Biol., 110: 81-95. Rosa, P., Weiss, U., Pepperkok, R., Ansorge, W., Niehrs, C., Stelzer, E.H.K., and Huttner, W.B. (1989) An antibody against secretogranin I (chromogranin B) is packaged into secretory granules. J. Cell Biol., 109:17-34.

CONFOCAL MICROSCOPY OF T H E MITOTIC SPINDLE Salmon, E.D., Leslie, R.J., Saxton, W.M., Karow, M.L., and McIntosh, J.R. (1984) Spindle microtubule dynamics in sea urchin embryos: Analysis using a fluorescein-labeled tubulin and measurements of fluorescence redistribution after laser photobleaching. J . Cell Biol., 99:2165-2174. Stelzer, E.H.K., Stricker, R., Pick, R., Storz, C., and Hanninen, P. (1988) Confocal fluorescence microscopes for biological research. SPIE, 1028:146-151. Vandre, D.D., and Borisy, G.G. (1986) The interphase-mitosis trans-

73

formation of the microtubule network in mammalian cells. In: Cell Motility: Mechanism and Regulation. H. Ishikawa, S. Hatano, and H. Sato, eds. Alan R. Liss, Inc., New York, pp. 389-401. White, J.G., Amos, W.B., and Fordham, M. (1987) An evaluation of confocal versus conventional imaging of biological structures by fluorescence light microscopy. J . Cell Biol., 10541-48. Wright, S.J., Walker, J.S., Schatten, H., Simerly, C., McCarthy, J.J., Schatten, G. (1989)Confocal fluorescence microscopy with the tandem scanning light microscope. J. Cell Sci., 94:617-624.