In this study we report an atomic force microscopy (AFM) investigation of the actin cortical cytoskeleton of Xenopus laevis oocytes. Samples consisted of inside-out orientated plasma membrane patches of X. laevis oocytes with overhanging cytoplasmic material. They were spread on a freshly cleaved mica surface, subsequently treated with Triton X-100 detergent and chemically fixed. The presence of actin fibres in oocyte patches was proved by fluorescence microscopy imaging. Contact mode AFM imaging was performed in air in constant force conditions. Reproducible high-resolution AFM images of a filamentous structure were obtained. The filamentous structure was identified as an actin cortical cytoskeleton, investigating its disaggregation induced by cytochalasin D treatment. The thinnest fibres showed a height of 7 nm in accordance with the diameter of a single actin microfilament. The results suggest that AFM imaging can be used for the high-resolution study of the actin cortical cytoskeleton of the X. laevis oocyte and its modifications mediated by the action of drugs and toxins.
The oocytes and embryos of the amphibian anurian Xenopus laevis are an interesting model for the study of many developmental mechanisms because of their dimensions and the ease with which they can be manipulated in experiments. In addition, X. laevis oocytes are a widely employed system for the expression and functional study of heterologous proteins (Romero et al., 1998; Schillers et al., 2001; Mari et al., 2006).
The cytoskeleton, an intricate three-dimensional network composed of three filament systems [tubulin microtubules, actin microfilaments and intermediate filaments (Gard, 1999)] is a cell structure that extends throughout the cytoplasmic region. Dynamic and spatially regulated rearrangements in the cytoskeleton network enable the cell to migrate, divide or maintain its shape (De Curtis, 2001; Ridley, 2001).
The X. laevis oocyte cytoskeleton plays a significant role in cell dynamics. It interacts with plasma membrane proteins (Marten & Hoshi, 1997) and is involved in several developmental functions, such as gastrulation and the localization of maternal RNA needed for cell development (Heasman et al., 1992; Klymkowsky et al., 1992; Bashirullah et al., 1998). The cytoskeletal arrangement also plays a significant role in the establishment and maintenance of the polarized X. laevis oocyte (Gard, 1995).
The analysis of the X. laevis oocyte cytoskeleton has been hampered by the difficulty in applying conventional optical microscopy to oocytes because of their large size and opaque cytoplasm, rich in yolk and granules. Electron microscopy techniques have been applied to the study of the cytoskeleton filamentous arrangement and its spatial organization (Moore et al., 1970; Heuser & Kirschner, 1980; Resch et al., 2002). In recent years, in order to investigate the cytoskeleton network organization of oocytes and its structural modifications at a resolution of a few hundred nanometers, confocal laser scanning microscopy has been used (Gard, 1999; Shimizu, 1997). This approach allowed the examination of the distribution and organization of cytoskeletal proteins in X. laevis oocytes by means of immunofluorescence microscopy. More recently, four-dimensional confocal imaging has also been applied to X. laevis oocytes and eggs to follow cytoskeletal dynamics (Bement et al., 2003). It should be considered that electron microscopy requires sample preparation, including detergent extraction, sample coating and staining, that may distort the native array of cytoskeleton filaments whereas confocal laser scanning microscopy does not provide a quantitative, high-resolution analysis of cytoskeleton filaments.
In this work AFM has been used to visualize the actin cortical cytoskeleton of X. laevis oocytes in inside-out orientated plasma membrane patches with overhanging cytoplasmic material. Reproducible high-resolution AFM images of the actin cortical cytoskeleton were obtained in air, operating in contact mode in constant force conditions. To the best of our knowledge, only AFM investigations on the nuclear envelope (Wang & Clapham, 1999; Oberleithner et al., 2000; Shahin et al., 2005), plasma membrane (Schillers et al., 2000, 2001; Lau et al., 2002) and vitelline envelope (Solletti et al., 1994), but not on the cortical cytoskeleton of X. laevis oocytes, have been reported in the literature to date. Although the cytoskeleton structure of X. laevis oocytes has been studied extensively by other microscopy techniques, in our view, the possibility offered by AFM to quantitatively investigate the arrangement of cytoskeleton filaments from the nanometer (single fibres) to micrometer (fibre bundles) scale appears unique and interesting.
Materials and methods
Preparation of oocytes
Oocytes were isolated from mature X. laevis female frogs and manually defolliculated after treatment with 1 mg mL−1 collagenase A (Roche, Mannheim, Germany) in the Ca2+-free ORII buffer (82.5 mm NaCl, 2 mm KCl, 1 mm MgCl2, 5 mm HEPES/Tris, pH 7.5) for 30 min at room temperature (22 °C). Selected V and VI stadium (Dumont, 1972) defolliculated oocytes were maintained at 17 °C in Barth's medium [88 mm NaCl, 1 mm KCl, 0.82 mm MgSO4, 0.41 mm CaCl2, 0.33 mm Ca(NO3)2, 10 mm HEPES/Tris, pH 7.5] supplemented with 0.005% gentamycin sulphate and 2.5 mm pyruvic acid.
For oocyte patch preparation, oocytes were left in Barth's solution supplemented with 250 mm sucrose for approximately 15 min and manually devitellinized as previously described (Sive et al., 2000). Devitellinized oocytes were placed in Petri dishes containing Barth's solution with a 2% agarose gel bottom for approximately 1 h before patch preparation.
Oocyte patch preparation for atomic force microscopy imaging
A freshly cleaved mica leaf was placed in a Petri dish containing the devitellinized oocytes in Barth's solution. Using a pipette, oocytes were gently placed on the mica surface. After approximately 30 min, to allow their adhesion to the mica, oocytes were gently pulled out from the mica surface by blowing on them with a pipette, in order to leave a membrane patch adhering to the mica. Cytoplasmic material that did not adhere and overhanging the oocyte patch was cleaned away by carefully blowing on it with the pipette. Barth's solution was then removed, taking care not to leave the patch dry in order to avoid dehydration effects, and the oocyte patch was treated with a 0.5% solution of Triton X-100 detergent, in order to remove the cytoplasmic membrane components and soluble proteins, as described elsewhere (Berdyyeva et al., 2005). Briefly, the oocyte patch was treated for 10 min with a 0.5% solution of Triton X-100 in buffer consisting of 0.14 m NaCl, 5 mm MgCl2, 4% polyethylene glycol 6000 and 10 mm Tris-HCl, pH 7.6. The treating solution was removed and the oocyte patch was washed twice for 2 min with the buffer and then fixed in a 4% solution of paraformaldehyde in 120 mm Na phosphate buffer, pH 7.4, for 20 min. The paraformaldehyde solution was removed and the oocyte patch was washed twice for 2 min with milliQ ultrapure water and dried at room temperature before AFM measurements.
For drug treatments, devitellinized oocytes placed on a mica support in a Petri dish containing Barth's solution were injected by means of a manual microinjection system (Drummond, Broomall, PA) with 50 nL of 20 µm cytochalasin D (Sigma-Aldrich, St. Louis, MO) (approx. final concentration in the oocyte 2 µm) and 50 nL of 200 µm nocodazole (Sigma-Aldrich) (approx. final concentration in the oocyte 20 µm). After 30 min, the oocytes were pulled out and membrane patches were prepared as described above.
For fluorescent staining experiments, glass cover slips were used as sample supports. Glass cover slips were cleaned overnight with an aqueous solution of 1 n HCl. The treating solution was removed and glass cover slips were washed three times with deionized water for 30 min and dried in air. Cleaned glass cover slips were coated overnight with an aqueous solution of poly-l-lysine (0.01% w/v, Sigma-Aldrich). The treating solution was removed and glass cover slips were rinsed twice with deionized water for 2 min and dried in air. Oocyte patches, prepared as described above, were stained with 1 µg mL−1 fluorescein isothiocyanate-labelled phalloidin (Sigma-Aldrich) in gelatine dilution buffer (GDB) (0.1% gelatine, 0.45 mm NaCl, 20 mm Na phosphate buffer, pH 7.4) for 20 min to detect actin microfilaments. The treating solution was removed and oocyte patches were washed twice with phosphate-buffered saline buffer (150 mm NaCl, 10 mm Na phosphate buffer, pH 7.4) and dried in air. Fluorescence images were obtained using an Eclipse TE 200 fluorescence microscope (Nikon, Kawasaki, Kanakawa, Japan).
Atomic force microscopy
Atomic force microscopy imaging was performed using an AutoProbe CP Research atomic force microscope (ThermoMicroscopes, Sumyvale, CA). The mica supports were attached to stainless steel punches with double-sided adhesive tape (Ted Pella Inc., Redding, CA) and magnetically fixed to the AFM sample holder. Standard rectangular silicon cantilevers (CSG01, NT-MDT, Zelenograd, Moscow, Russia) with a 0.01 N m−1 spring constant and a conical silicon tip with a 10-nm curvature radius and a tip cone angle of 22° were used to scan the sample in air. The tip radius was previously evaluated by means of electron microscopy. Images of 512 × 512 pixels2 were collected at a scan rate of 2 Hz operating in contact mode (constant force conditions). The set point was manually adjusted and kept below 500 pN to obtain the best resolution and the total force applied on the sample during imaging was approximately 5 nN as measured by the force/distance curve.
Topography and error signals were collected simultaneously. The use of the contact scanning mode, justified by the extremely soft measurement conditions, was chosen in order to obtain both higher resolution and error signal, which is useful for improved visualization of small features on the sample surface (Putman et al., 1992). Moreover, this approach was frequently used in the investigation of cytoskeletal elements as reported in the literature (Rotsch & Radmacher, 2000; You et al., 2000; Berdyyeva et al., 2005).
The comparison of the trace and retrace topographies showed that the optimized scan parameters allowed imaging of the cytoskeleton filaments without any significant structural distortions. Moreover, varying the scan angle at such optimized imaging conditions had no noticeable effect on the generated topographies.
The AFM images reported in this work were flattened using the image processing and data analysis software (ThermoMicroscopes) and the experimental data were obtained from three independent experiments.
Results and Discussion
High-resolution imaging of biological structures and their changes induced by different agents such as drugs and toxins is commonly performed by fluorescence and electron microscopy. Recently, AFM has been shown to be a suitable tool for imaging biological structures and their modifications (Engel & Müller, 2000). In comparison with fluorescence microscopy data, AFM experimental data are reported as sets of x-, y- and z-values that can be analysed to give the parameters pertaining to the examined surface in a quantitative form at nanometer resolution. Consequently, AFM allows detailed quantitative morphological information on the biological samples to be obtained.
Although electron microscopy techniques such as scanning electron microscopy and transmission electron microscopy reach nanometer and subnanometer resolution, they need complex and invasive sample treatments that may significantly modify the native structure. On the contrary, AFM can work at close to physiological conditions with the only requirement being that the sample is flattened and adhering to a flat support.
In this work X. laevis oocytes, adhering to a freshly cleaved mica surface, were pulled out in order to obtain patches of inside-out orientated plasma membrane with overhanging cytoplasmic material. This sample preparation procedure has been previously described (Schillers et al., 2001) for the AFM imaging of X. laevis plasma membrane proteins. As the aim of this work was the AFM imaging of the oocyte cortical cytoskeleton, a few parameters of this procedure were modified in order to obtain patches of plasma membrane with overhanging cytoskeleton structure. In particular, the adhesion time of the oocyte to the support, which was 1 min in the sample preparation protocol of Schillers et al. (2001), was increased to 30–40 min in order to let cytoskeletal-bound transmembrane proteins firmly adhere to the support. It is worth noting that in the work of Schillers et al. (2001) the authors did not find cytoskeletal structure attached to the intracellular side of the plasma membrane, possibly because, as reported by Schillers et al. (2001), with an adhesion time of 1 min cytoskeletal-bound transmembrane proteins were pulled off during the process of membrane excision. Oocyte patches were then treated with the Triton X-100 non-ionic detergent to remove the cytoplasmic membrane components and soluble proteins from the sample, in order to better visualize cytoskeletal elements. Samples were subsequently chemically fixed because, in preliminary experiments, we noticed that air exposure could significantly alter the actin cortical cytoskeleton integrity. The sample adhesion to the support is dominated by the electrostatic interactions between the support and the plasma transmembrane proteins that are associated with the actin cytoskeleton. The removal of the unwanted cytoplasmic material from the membrane patch (Triton X-100 detergent step) was necessary to obtain samples with actin cytoskeleton suitable for AFM imaging. This step did not modify the interaction of the support with the sample as transmembrane proteins associated with the actin filaments are not extracted by Triton X-100 detergent (Nathke et al., 1994; Adams et al., 1996). A freshly cleaved mica leaf was chosen as the sample support because of its ultra-flat surface, easily identifiable surface features and simple preparation that did not require any chemical treatment to improve the sample adhesion. In addition, the use of freshly cleaved mica was convenient and gave highly reproducible results. However, analogous AFM images of the cytoskeleton structure have been obtained using glass cover slips, poly-l-lysine-treated glass cover slips, mica and poly-l-lysine-treated mica as supports.
The presence of actin microfilaments in the samples, spread on a glass cover slip, was confirmed by staining patches with fluorescent phalloidin, a molecule that selectively binds to the actin microfilaments. The fluorescence microscopy image in Fig. 1 shows a filamentous structure confirming the presence of actin microfilaments in the prepared oocyte patches.
Subsequent AFM investigation allowed reproducible images (Fig. 2) of a well-defined filamentous structure overhanging the plasma membrane to be obtained. Figure 2(a) shows an intricate filamentous pattern composed of fibres of different lengths and thicknesses covering an area of 30 × 30 µm2. The filaments appear as generally not parallel to each other but orientated at different angles, often in a fan-shaped pattern (see Figs 2 and 3c) according to the typical arrangement of actin microfilaments as observed by electron microscopy (Svitkina et al., 1984) and fluorescence microscopy (Shimizu, 1996) data.
In addition, the AFM image shows spherical-like particles with diameters ranging from 300 to 2500 nm. The comparison of their shape, dimension and localization with data reported in literature (Larabell & Chandler, 1991; Solletti et al., 1994; Abriel & Horisberger, 1999) suggests that these globular structures could be cortical, pigment and yolk granules but the analysis of these structures was beyond the scope of this study.
The comparison between the AFM image (Fig. 2a), fluorescence microscopy image (Fig. 1) and X. laevis oocyte cytoskeleton confocal laser scanning microscopy images reported in the literature (Gard, 1999; Bement et al., 2003) strongly suggests that the filamentous structure visualized in the AFM image is an actual representation of the X. laevis oocyte actin cortical cytoskeleton.
The filamentous structure shown in Fig. 2(a) appears better defined in the error signal image (Fig. 2b) (Putman et al., 1992) because of the high vertical scale of the topography AFM image due to the presence of the globular structures.
Two AFM images collected on a smaller scan area (Fig. 2c and d) give a more detailed picture of the organization, shape and dimensions of single actin fibres. In the high-resolution AFM image (Fig. 2d) the minimum measured diameter of a single filament is 40 nm. Several filaments have a measured diameter in the order of 50–70 nm while, for the majority of the fibres, the diameter ranges between 100 and 200 nm. It must be pointed out that, when structures with lateral dimensions similar to the AFM tip radius are visualized, tip–sample convolution effects occur (Markiewicz & Goh, 1995). On the contrary, the height of such structures is not affected by tip–sample convolution effects. Thus, we used these data to evaluate the single filament dimensions. The minimum height of the thinnest filaments was 7 nm, in accordance with the well-known value of the diameter of a single actin filament (Moore et al., 1970), indicating that individual actin microfilaments could be resolved by our AFM approach. As an example, Fig. 2(f) shows the height profile relative to the black line in Fig. 2(e); the height of the fibre indicated by the arrow is 7.6 nm. In order to quantify the height of actin filaments, cross-sections perpendicular to the filament long axis were taken. The corresponding height profiles, obtained by the analysis of five AFM images collected on different samples, are shown in the histogram of Fig. 4 (black bars). The average filament height was 15.3 ± 8.8 nm (mean ± SD; 252 filaments analysed). The histogram shows that a large number of actin filament bundles along with some single actin fibres have been visualized in AFM images.
To confirm that the actin cortical cytoskeleton has been visualized in AFM images, the effects of cytochalasin D, an actin-disassembling drug, and nocodazole, a microtubule-depolymerizing agent, have also been investigated.
The modifications induced by cytochalasin D on the cytoskeleton structure are shown in Fig. 3(a) where a plasma membrane patch of an X. laevis oocyte injected with cytochalasin D is shown. The AFM image of the cytochalasin D-treated sample (Fig. 3a) is clearly different from that of an untreated sample (Figs 3c and 2c). The filamentous structure is dramatically modified by the drug treatment. In particular, the filamentous pattern is lost and fragmented and disorganized filament bundles are visible. As an independent experiment, the action of cytochalasin D on the cytoskeleton structure has also been investigated by fluorescence microscopy imaging to validate AFM results. The fluorescence microscopy image of the cytochalasin D-treated sample reported in Fig. 3(b) confirms the results shown in the AFM image (Fig. 3a). The filamentous structure observed in the fluorescence microscopy image (Fig. 1) of untreated X. laevis oocytes is dramatically modified in the image in Fig. 3(b), which appears highly comparable with the AFM image of a cytochalasin D-treated sample (Fig. 3a). In particular, the white square in Fig. 3(b), highlighting a region with analogous x/y dimensions to the area visualized in the AFM image reported in Fig. 3(a), allows a better correlation/comparison between AFM and fluorescence microscopy images.
The investigation of the effects of nocodazole on the cytoskeleton structure is shown in Fig. 4 where a plasma membrane patch of an X. laevis oocyte injected with nocodazole is shown. Unlike the cytochalasin D-treated sample, the nocodazole-treated sample does not show alterations of the filamentous structure, suggesting that microtubules were absent in the prepared samples. Exhibiting a vertical resolution of 1 nm or better, topography AFM images provide accurate information about filament height and therefore allow the quantitative analysis of the effect of nocodazole on filament dimensions. The height profiles, derived from cross-sections perpendicular to the filament long axis and obtained by the analysis of five AFM images collected on different samples, are shown in Fig. 4 (grey bars). The filament average height was 15.8 ± 8.8 nm (mean ± SD; 244 filaments analysed).
The good overlapping of the two histograms in Fig. 4 (Student's t test, P = 0.8), showing the values of the filament height of the untreated (black bars) and nocodazole-treated (grey bars) samples, quantitatively confirms the lack of any effect of nocodazole on the filamentous structure visualized in AFM images.
Considered together, these results prove that the filamentous structure visualized by AFM images is the actin cortical cytoskeleton of the X. laevis oocyte.
In this study AFM imaging has been used for the first time to investigate at high resolution the actin cortical cytoskeleton of X. laevis oocytes.
Atomic force microscopy allowed the quantitative investigation of the arrangement and dimensions of actin filaments from nanometer (single fibres) to micrometer (fibre bundles) scale.
In AFM images, a filamentous structure with spherical-like particles is visible. The fluorescence microscopy imaging together with the sample treatment with cytochalasin D, a cytoskeleton-disassembling drug, and nocodazole, a microtubule-depolymerizing agent, allowed identification of the filamentous structure as actin cortical cytoskeleton. Several actin filaments showed a height of 7 nm, indicating that single actin microfilaments have been visualized in AFM images.
Furthermore, the investigation of the effects of the cytoskeleton drug treatment showed that AFM imaging can be a suitable approach to visualize the action of agents able to induce modifications on the cortical cytoskeleton of X. laevis oocytes at single microfilament level.
The authors thank Dr Nicola Panzeri for his technical assistance and support given to this work. This research was funded by grants from the Italian Ministry of Research and University (FIRB programme No. RBNE03B8KK-08) to V.F.S. and G.P. for the project on ‘Investigation of protein structure and function by AFM and physiological studies’.