Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://cellmotility.genebee.msu.ru/html/articles/grigoriev97.pdf Äàòà èçìåíåíèÿ: Fri May 17 16:25:14 2002 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 20:11:05 2012 Êîäèðîâêà:

Membr. Cell Biol, 1997, Vol. 11 (2), pp. 195-211 Reprints available directly from the publisher Photocopying permitted by license only

© 1997 OPA (Overseas Publishers Association) Amsterdam B. V. Published in The Netherlands by Harwood Academic Publishers GmbH Printed in India

Quantitative Analysis of the Movements of Cytoplasmic Granules in Polarized Fibroblasts
I. S. Grigoriev, A. A. Chernobelskaya, and I. A. Vorobjev
Laboratory of Cell Motility, Belozersky Institute of Physico-Chemical Biology, Moscow State University, Moscow

Movements of cytoplasmic organelles were analyzed in Vero fibroblasts. In the cells polarized at the edge of an experimental wound, cytoplasmic granules moved randomly (Brownian motions) and by separate jumps (saltatory movements). The displacement of granules by the Brownian motions exceeded by more than an order of magnitude that of the mitochondria similar by weight. Lipid droplets moved predominantly by saltations, whereas mitochondria and lysosomes moved much less often. In a front part of the polarized cells, the main directions of saltatory movements were from the nucleus to the leading edge of a cell and back, whereas the tangential movements (across the long axis of a cell) were less than 1%. 90% of saltatory movements occurred in the area starting 10-12 µm from the nucleus and ending 10-12 µm from the leading edge of a cell. The average rate of saltatory movements of the granules (2.38 µm/s) was identical in both directions. The average length of the track was 7.49 ; the maximum track length reached 30 µm. An increase in the granule diameter from 0.3 to 1.4 µm resulted in a minor (statistically insignificant) decrease in the average rate of the movements. The average rate of saltatory movements of mitochondria was 1.00 µm/s, and the average track length was 6.04 µm. Therefore, mitochondria, in contrast to lipid droplets, are rigidly fixed in the cytoplasm, and the force holding mitochondria is equal to the force produced by the microtubule-associated motors. Taking into account the characteristic of the centrifugal saltations, we suggest that they are mediated by an unusual dynein.
(Received 25 October, 1996; accepted for publication 4 November, 1996)

In phase-contrast microscopic observations of living cells, one can see in the cytoplasm mitochondria and numerous granules which can move at different speeds in different directions. The cytoplasmic granules about 0.5 µm in diameter prove in electron-microscopic studies to be various organelles which, depending on their internal contents, are lipid droplets, lysosomes or endosomes [1]. Most movements of the organelles to distances of about 1 µm and more are of uneven character, i.e., the single movements of several 195


196

I. S. GRIGORIEV et al.

seconds long alternate with pauses between them. Such movements were termed jump-like or saltatory [1-7]. A single saltatory movement was termed a track. In between the saltatory movements, the organelles are in the state of rest or chaotic Brownian motion. The displacements of the organelles as a result of Brownian motion depend on their weight and the viscosity of the cytoplasm [8]. The theoretically calculated and directly measured displacements as a result of Brownian oscillations for the organelles of 0.5 µm in diameter do not exceed 0.5 µm per 10 s [8, 9]. The rates of saltatory movements of the organelles, measured by various authors in tissue culture cells, vary within the wide range: from 0.03 µm/s (endosomes in the primary culture of the ovary granulose cells [5]) to 2.45 µm/s (granules less than 0.5 urn in diameter in fibroblasts and epithelial cells from somites or the ectoderm of Xenopus laevis embryos [6]). In thin cytoplasmic protrusions from Reticulomixa amoebas the granules and mitochondria move at rates of up to 25 µm/s [10, 11, 12]. The saltatory movements are directly related to microtubules (MT). At the depolymerization of the MT network by the cold or under the action of various mitostatics (colchicine, nocodazole) the saltatory movements are suppressed [1, 4, 12, 13]. The movements of the organelles along individual MT were observed using the method of differential interference contrast with electronic amplification (VE-DIC) [14, 15, 16]. Finally, in a cell-free system the moving forces in transition of the organelles along MT were shown to be protein motors dynein and kinesins [16, 17]. In in vitro experiments, the rates of MT movements were 1-2 µm/s along the substrate covered with dynein, and 0.5 µm/s along the substrate covered with kinesin [18, 19]. Investigated in vitro, the mechanism of organelle movements does not fully explain the character of motion in a living cell. A bridge between a cell-free system and living cells can be works in which the parameters of motion were studied by the same methods in both systems. Such works were performed on the giant axon of squid and the axoplasm isolated from it [20-24]. In a cell-free system the average rate of motion of the granules was from 1.8 to 3.56 µm/s [22-24]. In a living axon, the granules moved at speeds of up to 5 µm/s [20]. A direct comparison of the rates of movements of the granules in both systems is given in [22]. In a cell-free system, small and big vesicles and mitochondria moved at a similar average speed of 2.2 µm/s. In an intact axon, the rates of motion of these three types of organelles differed and were 2.2, 1.1 and 0.4 µm/s, respectively. However, the squid axon is not the most typical model to study the movements of granules in a cell, because the unique morphology and high specificity of the axon impose certain restrictions on the organization of granule motion. Thus, on the one hand, in some cells granules move at speeds strongly exceeding those obtained for isolated protein motors in cell-free systems. On the other hand, some authors point to the motion rates significantly lower than


MOTION OF CYTOPLASMIC GRANULES IN

FIBROBLASTS

197

those of in vitro motors. A large spread of the motion rates of the granules can be due to not only the differences in the subjects of research, but to the differences in methods of image analysis. The first works on cultured cells were performed using time-lapse filming [1,5]. The images were recorded frame by frame and then changes in the positions of granules on the adjacent frames were analyzed. The track of a granule was registered as the distance between two points. Subsequently, works appeared in which images were recorded continuously, i.e., the motion of the granules could be observed and analyzed [6, 11]; however, the movements themselves were not considered by the authors in detail. In this work, we made an attempt at following the movements of the granules by using various methods of analysis of one subject - polarized fibroblasts. We used the method of phase contrast which makes it possible to observe the cytoplasmic structures from 0.3 µm in diameter.
EXPERIMENTAL

Cell culture. As a subject of research, we took the cell culture Vero (the fibroblasts from the kidney of a green monkey) from the Cell Culture Collection at the Institute of Cytology, Russian Academy of Sciences. The cells were cultured at 37°C and 5% CO2 in a medium DMEM ± F12 (Sigma, USA) with 3% calf embryo serum. Gentamicin was used as an antibiotic. Vital observations. For experiments, the cells were plated onto round cover glasses. For vital observations, the glasses with cells were mounted into a modified Dvorak-Stottler chamber [25]. Some experiments used the model of an experimental wound [26, 27]. In 2 h after part of the monolayer was removed by a razor blade, the glasses were mounted into the chamber; the vital observations began immediately after and did not exceed 4 h. During the observations, the temperature on the microscopic stage was maintained 37°C by means of a drying fan. For video recording, use was made of a high-sensitivity matrix video camera with the resolution of 500 lines and an AG-6730 time-lapse video tape recorder (Panasonic, Japan). The final magnification on the monitor, using a Planapo 40/1.0 objective lens, was x4000; using a Planapo 63/1.4, . reduce the photodamage of the cells during the recording we used an orange light filter (OS-4). Each cell was recorded continuously for 10 min. Cells at the margin of the wound, which had a pronounced polarity, were chosen for analysis. Analysis of the data. The video images obtained on a phase contrast microscope and recorded were analyzed in two ways: (1) individual granules were continuously followed on the monitor and their movements (tracks) were copied by a felt-tip pen on a transparency put onto the monitor. The movements of all granules were summed up after repeated playbacks of the record; (2) to obtain a time-lapsed (frame-by-frame) record, use was made of a


198

I. S. GR1GORIEV et al

Figure 1. A scheme of a fibroblast polarized at the margin of an experimental "wound". A, the long axis of the leading edge of the cytoplasm. 1, the region adjacent to the lamella; 2, the central region; 3, the nucleus-adjacent region of the leading edge of the cytoplasm.

DT-2851 analog-to-digital converter (Data Translation, USA) and a DT-IRIS program. The interval between the frames was 10 s. The displacements of the granules were registered as described earlier [9]. The following parameters were chosen to characterize the granule movements: the length and direction of a track, the speed of motion, and the share of moving granules. The speed was calculated as the transition from the beginning to the end of a track, divided by the time of movement. The movement time of a granule was determined from the video record using a stopwatch. The share of the moving granules in a lamella at a given time was calculated using the DT-2851 analog-to-digital converter and the DT-IRIS program. The successive frames were recorded with 6-s intervals. The granules displaced relative to their positions on the previous frame to distances greater than their diameter were considered to be motile. To analyze the distribution of the tracks in the leading edge of the polarized cells' cytoplasm, we singled out three zones in it in the following way: a long axis was drawn from the nucleus to the edge of the lamella (Fig. 1), then two perpendiculars were drawn across it at distances of 1/3 and 3/4 of the length of the leading edge from the nucleus. Thus, the leading edge was divided into the zones: adjacent to the nucleus, central and adjacent to the lamella. The number of tracks in these zones was calculated separately. Staining of lipids. The lipid granules were revealed using the reaction of osmium tetroxide (OsO4) reduction [28]. The cells grown on cover glasses were fixed with 2.5% glutaraldehyde for 30 min. After the washout with the physiological phosphate saline (PBS, pH 1.2-1 A), a glass with the cells was incubated with 1% solution of OsO4 on PBS for 24 h. Then the preparation was washed for 30 min in 70% alcohol and put into glycerol.


MOTION OF CYTOPLASMIC GRANULES IN FIBROBLASTS

199

Figure 2. Fibroblasts in a dense monolayer. Phase contrast. Scale, 10 µm.

Staining of lysosomes with acridine orange. The vital staining of the cells with acridine orange was carried out as follows. The cells were incubated for 20 min in a medium containing 1 (µg/ml acridine orange. After the triple washing with the pure culture- medium the cells were observed and photographed in a fluorescent microscope (Opton, Germany) with a set of light filters for tetramethylrhodamine isothiocyanate.
RESULTS

Cell morphology. In the cytoplasm of the cells in the dense monolayer the number of clearly visible granules and mitochondria is not large (Fig. 2). The cells are about 50 µm long and 30 µm wide. If one takes into consideration that the diameter of the nucleus is about 15 µm and that there is no motion of the granules in the subnuclear and cortical layers of the cytoplasm, which are not less than 1.5 µm each, then the cytoplasmic area within which the granules can move is not large. Besides, the large thickness of the cells in the monolayer does not always make it possible to follow a track from the beginning to the end. At a low density the cells spread well on the glass. Part of the cells are polarized and move actively, having a typical locomotor cycle (stretching of the cell, movement of the nucleus forward, detachment of the tail and its invagination, formation of a new leading edge). The complete cycle takes about 40 min, and the phases of the motion alternate rather quickly. Different cells are in different phases, many of them being at the stage of changing the direction of motion, and have more than one leading edge.


200

I. S. GRIGORIEV et al.

Figure 3. Locations of the tracks in a non-polarized fibroblast. Scale, 10 µm. The arrows show the directions of movement of the granules.

Figure 4. Locations of the tracks in a polarized fibroblast. Scale, 10 µm. The arrows show the directions of movement of the granules.

The observations of individual cells in a loose culture showed a strong variability of the cell population by the intensity of the saltatory movements of the granules, which in different cells differed by an order of magnitude. The activity was observed to be lower in the cells having no clear-cut polarization. The movements of the granules in non-polarized cells proved rather chaotic (Fig. 3). In such cells, we observed radial tracks (going from the nucleus to the edge of the lamella) and tangential tracks (when the movement of the granule was tangential to the nucleus). In polarized cells the granules moved mainly in two directions: from the nucleus and towards the nucleus (Fig. 4).


MOTION OF CYTOPLASMIC GRANULES IN FIBROBLASTS

201

Figure 5. A polarized fibroblast. Phase contrast. The arrows indicate the granules. The arrowheads show mitochondria. Scale, 10 µm.

Thus, the polarized stretched fibroblasts with the pronounced leading edge of the cytoplasm proved the most convenient for analysis. In polarized fibroblasts, the cytoplasm can be conventionally divided into two regions: the leading region ahead of the nucleus and the tail region after the nucleus. The movements of the granules in the tail region were less intensive than in the leading region. The leading region was better spread. Therefore, for a detailed investigation of the movements of the granules we would consider only the leading region of the cytoplasm of a polarized fibroblast (Fig. 5). To increase the number of cells of interest for us, we used the model of an experimental wound [26, 27]. Some time after part of the monolayer was removed, the cells from the first row of the remaining part began to
Table 1. The number of granules in the leading edge of the cytoplasm of a fibroblast polarized at the margin of the wound. Average number of granules in the leading edge of the cytoplasm diameter 0.3-0.5 diameter 0.51µm 0.8 µm 27.6 ± 11.8 (53%) 16.6 ± 7.9 (32%) diameter 0.81 - out of focus 1.4 µm 6.5 ±5.8 (12%) 1.5 ± 2.3 (3%) total 52.1 ± 12.6


202

I. S. GRIGORIEV et

al Figure 6. Staining of lipids with osmium tetroxide. Scale, 10 µm. a, Light field; b, phase contrast. The arrows show the lipid granules; the arrowheads, the non-lipid granules.

simultaneously put out the leading edge of the cytoplasm. Herewith, the tails of the cells remained at their previous places for at least 2-2.5 h. As a result, the fibroblasts acquired a clearly polarized shape and spread additionally. In 2-2.5 h after inflicting the wound the length of the leading edge of the cytoplasm of polarized fibroblasts was 40-50 µm (at the width of 20-30 µm), and the total length of the cells reached 80 µm. The major part of the leading edge was sufficiently well spread and contained granules and mitochondria arranged in the thickness of one layer of the edge. Besides, it comprised a thin lamella and a less spread zone adjacent to the nucleus (see Fig. 5).


MOTION OF CYTOPLASMIC GRANULES IN FIBROBLASTS

203 Figure 7. Staining of lysosomes in the living cells with acridine orange. Scale, 10 µm. a, Fluorescence; b, phase contrast.

Cytoplasmic organelles. Phase-contrast observations of the cytoplasm of the fibroblasts reveal spherical granules of different contrast and diameter, and mitochondria of various shapes. The granules and mitochondria moved in the cytoplasm at various speeds and in various directions. The motion was of two types: long directed (saltatory) and short chaotic (Brownian).


204

I. S. GRIGORIEV et al.

The leading edge of the cytoplasm of a polarized fibroblast had about 50 granules on average (see Fig. 5). Depending on the diameter, the granules divided into three categories: small (0.3-0.5 µm), average (0.5-0.8 µm) and large (0.8-1.4 µm). The large granules were 14% of the total number of the granules in the leading edge of the cytoplasm. The small and average granules were the major portion (Table 1). In the spread part of the leading edge of a cell, as a rale, not more than two granules (3-4%) proved out of focus. By the results of staining with osmium tetroxide more than half of the granules located in the leading edge of the cytoplasm were lipid droplets (Fig. 6). The lysosomes in the spread part of the leading edge of the cytoplasm were absent and were present only in the nucleus-adjacent zone (Fig. 7). Besides the various granules in the leading edge of the cytoplasm, mitochondria are also well seen (Fig. 5). The mitochondria can be conventionally divided into two groups: short (0.8-3 µm) and long (3-10 µm). The diameter of both is the same - about 0.5 µm. The number of mitochondria in the leading edge of the cytoplasm strongly varied, but was about 30 on average. Saltatory movements. As shown by preliminary observations, all displacements of the organelles to distances over 1 µm were in jumps, i.e., they can be called saltatory. In the leading edge of the cytoplasm of polarized fibroblasts the share of the saltatorily moving granules was 36%. Among the directions of movements, two predominated: from the nucleus (centrifugal) and towards the nucleus (centripetal). The tracks perpendicular to the long axis of the cell were very rare. The average speed of granule motion was 2.38 µm/s. The maximal speed reached 6.34 µm/s. The average length of the granules' tracks was 7.49 µm, and the maximal length of the tracks approached 30 µm (Table 2). In other words, some granules are capable of overcoming, without stops, distances equal to 2/3 of the length of the leading edge of the cytoplasm. The movement of the granules towards the nucleus did not differ by its parameters from that towards the leading edge of the lamella (Table 3). Among the directions of granule motion, that towards the leading edge of the lamella slightly predominated (56%) (Table 4). Part of tracks of the granules were straight lines. The predominant majority of the tracks were slightly curved lines. When measuring the lengths of such
Table 2. Rates of motion and lengths of tracks of the granules in Vero cells. Direction of motion Speed ± SD, µm/s Speed range, Track Length range, Number µm/s lengths, µm µm of tracks 0.80-6.34 0.86-5.53 0.80-6.34 7.67 ± 2.63 7.30 ± 2.61 7.49 ± 2.62 2.79-20.64 3.36-30.16 2,79-30.16 313 241 565

From the nucleus 2.52 ±0.77 2.25 ± 0.76 To the nucleus Both directions 2.38 ± 0.77


MOTION OF CYTOPLASMIC GRANULES IN HBROBLASTS

205

tracks, we fitted them as linear ones, the error not exceeding 8%. About 2% of the tracks were zig-zag lines. In this case, the granule, without stopping, changed its initial direction to an angle close to 90°. Still another category of tracks included the movements during which the granules, without stops, changed the direction to the opposite. Such tracks were extremely rare.
Table 3. Speeds of the granules and lengths of the tracks. Diameter of granules, µm Speed ± SD, µm/s Length Speed range, Track µm/s lengths, µm range, µm Number of tracks

Towards the leading edge of the lamella 0.3-0.5 0.51-0.8 0.81-1.4 0.3-0.5 0.51-0.8 0.81-1.4 0.3-0.5 0.51-0.8 0.81-1.4 2.62 ± 0.84 2.13 ±0.66 2.00 ± 0.79 2.81 ± 0.86 2.56 ± 0.76 2.00 ± 0.79 2.66 ± 0.83 2.37 ± 0.76 2.28 ± 0.82 1.09-5.53 0.86-4.23 0.88-4.58 1.11-6.34 0.88-5.60 0.73-4.06 1.09-6.34 0.86-5.56 0.73-4.58 8.11 ±3.18 6.96 ± 2.28 6.82 ± 2.38 7.88 ± 3.09 7.87 ± 2.88 7.52 ± 1.93 7.85 ± 3.06 7.38 ± 2.70 6.15 ± 2.36 3.36-30.16 3.97-21.27 3.81-16.67 2.22-20.64 2.54-20.00 3.38-17.14 2.22-30.16 2.54-21.27 3.38-17.14 86 87 68 111 102 100 201 189 175

Towards the nucleus

Depending on their diameter

Table 4. Predominant directions of motion of the granules depending on their diameter. Diameter of granules, µm away from the nucleus 0.3-0.5 0.51-0.8 0.81-1.4 Percentage of the granules towards the nucleus 56 54 59 44 46 41 moving

We measured the parameters of movement of the granules with respect to their diameter (Table 3). As the diameter increased, the average speed decreased from 2.66 µm/s (d < 0.5 µm) to 2.28 µm/s (d > 0.8 µm), i.e., by 15%. However, due to the large spread in the speeds of individual granules this difference is statistically insignificant (p > 0.5). Larger granules also featured tracks of smaller lengths. The bulk of the tracks in the leading edge of the cytoplasm were in its central part, while the regions close to the nucleus and the leading edge of the


206

I. S. GRIGORIEV et al.

lamella were virtually devoid of tracks. The region of the leading edge of the cytoplasm, adjacent to the nucleus (see Fig. 1), had on average 10% of tracks; the region, adjacent to the edge of the lamella, 0.3% of tracks. The other tracks were in the central region of the leading edge of the cytoplasm. The region adjacent to the nucleus and including the cell centre was poorly spread and the granules in it were arranged in more than one layer. In this region, we considered only the granules which were in focus during the recording. They are 46% of the total number of granules in focus in the leading edge of the cytoplasm. The central region of the leading edge of the cytoplasm contained 50% of the granules. The granules in it are arranged in one layer. However, the central region has 90% of the tracks, and the nucleus-adjacent region, only 10%. Even if one takes into consideration that the real number of granules in the nucleusadjacent region is larger and the tracks which are out of focus were not considered, even then tracks in the central region are considerably more in terms of one granule. The region adjacent to the edge of the lamella is very well spread and has 4% of the granules. In this region we considered all granules and in terms of one granules the tracks at the edge of the lamella are about 10 times as less as in the central region. The long mitochondria virtually did not move in the cytoplasm during the time of continuous observations. Their movements were mainly reduced to bending. The short mitochondria sometimes made saltatory movements. In this case, they moved at an average speed of 1.00 ± 0.33 µm/s (n = 37) to distances of 1.6-16 µm (on average, to 6.04 µm). The share of motile mitochondria was less than 1%. The measuring error of the quantitative characteristics of the saltatory movements was 0.16 µm in track length measurements and 0.24 µm/s in movement rate measurements. Brownian motions of the cytoplasmic organdies. Besides the saltatory movements, most granules constantly made Brownian oscillatory motions. The amplitudes of individual oscillations did not exceed the diameter of the granule, and the frequency of oscillations was no less than 1 Hz. As a result of such oscillations, the granules gradually moved to small distances. An average displacement in 10 s was 1.12 ± 0.53 µm (n = 37). The mitochondria were in a state of rest or a weak Brownian motion. The displacements of the long mitochondria are impossible to assess because they bent. The amplitudes of the displacements of the short mitochondria as a result of the Brownian motion was only 0.09 ± 0.19 µm per 10 s (n = 95). Comparison of the methods of continuous and time-lapse (frame-by-frame) observations of the saltatory movements. To compare the methods of continuous and frame-by-frame (time-lapse) observations of the saltatory movements, we sampled three cells, where 90 tracks were analyzed independently by the two methods.


MOTION OF CYTOPLASMIC GRANULES IN FIBROBLASTS

207

On average, a individual saltatory movement of a granule lasted for 2 s. The average speed measured in real time was 2.90 ± 1.33 µm/s, and that measured by means of the frame-by-frame analysis of an image, only 0.53 ± 0.23 µm/s. Considering all errors, the average speed of granule movements using the frameby-frame analysis with the interval of 10 s, was underestimated by 82%. In the frame-by-frame analysis of images, in 18% of the cases several tracks were summed up as one. If the jumps followed one another in one direction within the 10-s interval, then the frame-by-frame analysis overestimated the length of the track. When the varidirectional tracks were summed up, the length of the track was underestimated. In the rare cases when the track of a granule was a strongly curved line, the path passed by the granule exceeded its displacement as a result of an individual jump.
DISCUSSION

Brownian motions and the viscosity of the cytoplasm. For particles in a liquid, the root-mean-square displacement as a result of Brownian oscillations is proportional to the observation time and is inversely proportional to the viscosity of the solution [29, 30]:

where k is the Boltzmann constant, T is absolute temperature, is viscosity in poises, r is the radius of a particle, t is the observation time, s. If the average diameter of the lipid droplets is taken to equal 0.85 µm ([0.3 ± 1.4]/2 µm), then the respective viscosity of the cytoplasm in Vero cells proves to be about 0.1 P. This magnitude is lower than the estimated 2.6 P for HeLa cells in [8] and it shows that in between the saltations the lipid granules in the cytoplasm of the Vero cells are, most probably, not fixed. Using a similar calculation for PKEC (pig kidney embryonal cells) where, as a result of Brownian motions (at the action of nocodazole on the microtubules or the effect of sodium azide), the displacements of the granules in 10 s did not exceed 0.4 µm [9], we obtain the viscosity of the cytoplasm higher than 1 P. Thus, the cytoplasm in Vero fibroblasts has a much lower viscosity than in epithelial cells (HeLa and PKEC). On the other hand, as shown by our observations, the lipid granules and mitochondria behave differently in the cytoplasm of Vero fibroblasts. The amplitude of displacement in Brownian motions of short mitochondria, having the equivalent radius (the radius they would have if they would be a ball of the same volume) similar to that of the lipid granules, is at least 15 times as low. Respectively, the "viscosity" of the cytoplasm for the mitochondria is at least 200 times larger than for the granules. Therefore,


208

I. S. GRIGORJEV et al.

in contrast to the granules, the mitochondria in the cytoplasm of the fibroblasts are constantly fixed. Probably, the granules in the epithelial cells are also fixed, and then the above estimates do not reflect the true viscosity of the cytoplasm in these cells. Quantitative characteristics of the saltatory movements of the granules. In Vero cells, individual cells are capable of moving without stops to distances of up to 30 urn (about 2/3 of the length of the leading edge of the cytoplasm). Thus, the factor limiting the maximal length of the track proves to be the size of the fibroblast itself. One act of movement (jump) can be enough for transferring the granules to the point of "destination". The most motile are lipid droplets. They move in the centrifugal and centripetal directions at approximately equal speeds. Against the background of the intensive motion of the lipid droplets, the mitochondria and lysosomes in Vero cells were virtually immobile. The mitochondria were not subject to Brownian motions, either, though the weight of short mitochondria approximately corresponds to that of the lipid granules. This implies that the mitochondria are rigidly fixed in the cytoplasm, and only their minor portion interacts with the system providing the movement along the microtubules. The fact that the speed of the saltatory motions of the mitochondria is considerably lower than that of the lipid droplets, suggests that the force keeping the mitochondria in the resting state is approximately equal to the force developed by the motors associated with the microtubules. It is known that in a polarized fibroblast the minus ends of the microtubules face the centre of organization of MT - the centrosome, and the plus ends of MT are turned to the periphery of the cell [31]. The centrosome in polarized fibroblasts is located ahead of the nucleus and at a distance from it [32, 33]. Therefore, we suggest that the granules move towards the nucleus in the direction of the minus ends of MT, and in the direction of the leading edge of the lamella, towards the plus ends of MT. Thus, the plus- and minus-directed motors moved the granules similarly. An analogous rapid motion to both sides at equal speeds has not been described for animal cells and has been known only for giant amoebas [10, 11, 12]. The saltatory movements of the granules in opposite directions were investigated in detail in melanophore cells [35] which also have a radial system of MT [36]. Towards the periphery, the granules moved in a series of short jumps, and towards the centre, at a rate of up to 5 µm/s, overcoming distances of many micrometers without stops [35]. Interestingly, particular saltations did not change either at the dispersion of the granules, or at their aggregation under the action of hormones [35]. A later work showed the centrifugal transfers of the granules in melanophores to occur due to the action of kinesin [37]. In axons of spinal ganglia from chicken embryo the granules of less than 0.5 µm in diameter moved retrogradationally at an average rate of 1.05 µm/s;


MOTION OF CYTOPLASMIC GRANULES IN FIBROBLASTS

209

and anteriogradationally, 0.75 µm/s. As the diameter of the granules increased, the average speed decreased, but the differences between the retrogradational and anteriogradational movements were preserved [38]. The mitochondria in these cells moved by about twice slower than the granules, but the rates of their retrogradational and anteriogradational movements correspond in a similar way [39]. Thus, in melanophore cells the movements in the direction of the different ends of MT differ in character; and in axons, in speed. In cell-free systems, the rates of motion using cytoplasmic dynein and kinesin also differ significantly. In the presence of kinesin, the speed of MT interacting with protein motors immobilized on the surface of a glass was 0.5 µm/s; and in the presence of dynein from cilia, 6.7 µm/s [19]. Using cytoplasmic dynein, MT moved at 1-2 µm/s [18]. On the glass covered with a mixture of kinesin and dynein from cilia, MT moved at 0.2-0.3 µm/s with the minus end forward and 3.5-4.0 µm/s with the plus end forward. Thus, the speeds of MT on the glass in opposite directions strongly differ [18, 19]. It should be noted that for all kinesins studied as of today, the rates of motion in cell-free systems do not exceed 1.2 µm/s [40]. Thus, the data on the various speeds of varidirectional motors obtained in vitro are well consistent with the differences in the rates of retrogradational and anteriogradational transport in cells. One more peculiar feature of saltatory movements in Vero cells is that both large (more than 1 µm) and small (0.3-0.5 µm) granules move at almost the same average speed. At the difference of the diameters of the granules, which achieves 1 µm, the difference in their weight proves to be more than 30 times larger. Nevertheless, the decrease in the average rate of motion in our experiments did not exceed 15%. This implies that in Vero fibroblasts the motors moving the lipid droplets operate with a greater stock of power, and the speed with which the granule moves is regulated by factors external with respect to the motor. In other cells, the speeds of the organelles depend on their diameter. In cultures of fibroblasts and epithelial cells [6], as well as in the neurons of the spinal ganglion from chicken embryo [38] the speed of the granules noticeably decreased as their diameter went up. In the squid giant axon, small vesicles moved at an average rate of 2.2; medium vesicles, 1.12; and mitochondria, only 0.4 µm/s [16]. Thus, in Vero fibroblasts the plus-directed motor moves the granules with a larger speed than kinesin in melanophores and kinesins in a cell-free system. Moreover, it has a larger stock of power. These features indicate that in Vero cells an unusual dynein can be involved in the motion towards the plus end of MT. Analysis of the saltatory movements: the results of comparing the methods. We compared the methods of analysis of frame-by-frame images and that of


210

I. S. GRIGORIEV et al

continuous images. At the analysis of the continuous images we determined that the granules moved at an average rate of 2.38 µm/s. The maximal speed was 6.34 µm/s. The rates of saltatory movements, obtained in the analysis of the frame-byframe images with 10-s intervals between frames, proved 5 times lower (the average speed, 0.53 µm/s; the maximal speed, 1.52 µm/s). Most works performed earlier on the movement of organelles in the cell, beginning from [1] and up to the recent time [9], were carried out using the frameby-frame analysis. Those works give different average speeds of the organelles, from 0.03 to 2 µm/s [1, 5, 7, 12]. The appearance of high-sensitivity video cameras with high spatial resolution made possible continuous observations of the cells, without splitting the dynamic process of the motion of the granules into a number of static frames. The published works on granule motion in the cells added diversity to the data of the parameters of granule motion. The speeds of saltatory movements of the organelles in living cells, obtained in the analysis of continuous images also vary within a wide range - from 0.3 to 25 µm/s [11, 6, 39]. A diversity of the data available before these works appeared could be explained by the differences of subjects of study. Now, one should also take into consideration the differences in the methods of analysis of the images. The data we obtained indicate that the analysis of a continuous image characterizes the saltatory movements of the granules in the cell much more exactly. The speeds of the granules in the cells, measured using continuous recording, are commensurable with those of the granules in cell-free systems. Therefore, the data of the works analyzing frame-by-frame images (time-lapse recording) are strongly underestimated as compared with the real speeds. The authors are grateful to M. S. Votchal and R. E. Uzbekov for assistance in video recording. The present work was supported by the Russian Foundation for Basic Research (grants No 993-04-06523 and No 96-04-50935).
REFERENCES
1. J. J. Freed and M. M. Lebowitz, J. Cell Biol. 45: 334-354 (1970). 2. L. I. Rebhun, Biol. Bull. 117:518-545 (1959). 3. L. I. Rebhun, In Primitive Motile Systems in Cell Biology, ed. R. D. Alien and N. Kamaya (New York: Academic Press, 1964): 503 pp. 4. V. E. Foe and B. M. Alberts, J. Cell Sci. 61:31-70 (1983). 5. B. Herman and D. F. Albertini, J. Cell Biol. 98:565-576 (1984). 6. P. C. Bridgeman, B. Kacher, andT. S. Reese, J. Cell Biol. 102:1510-1521 (1986). 7. R. Matteoni andT. E. Kreis, J. Cell Biol. 105:1253-1265 (1987). 8. L. I. Rebhun, Intern. Rev. Cytology 32:93-137 (1972).


MOTION OF CYTOPLASMIC GRANULES IN FIBROBLASTS

211

9. S. P. Speranskaya, M. S. Votchal, and I. A.Vorobyev, Tsitologiya 37:783-790 (1995) (in Russian). 10. J. T. Travis and R. D. J. Alien, Cell Biol. 90:211-221 (1981). 11. M. P. Koonce and M. J. Schliwa, Cell Biol. 100:322-326 (1985). 12. D. Murphy and L. Tilney, J. Cell Biol. 61:757-779 (1974). 13. O. Baumann and D. B. Murphy, Cell Motil. Cytoskeleton 32:305-317 (1995). 14. J. T. Travis, J. F. Kenealy, and R. D. Alien, /. Cell Biol. 97:1668-1676 (1983) 15. J. H. Hayden and R. D. Alien, J. Cell Biol. 99:1785-1793 (1984). 16. R. D. Vale, T. S. Reese, and M. P. Sheetz, Cell 42:39-50 (1985). 17. S. P. Gilbert, R. D. Alien, and R. D. Sloboda, Nature 315:245-248 (1985). 18. J. R. Mclntosh and M. E. Porter, J. Biol. Chem. 264:6001-6004 (1989). 19. R. D. Vale, F. Malik, and D. Brown, J. Cell Biol. 119:1589-1596 (1992). 20. R. D. Alien, J. Metural, I. Tasaki, S. T. Bradi, and S. P. Gilbert, Science 218:1127-1129(1982). 21. S. T. Brady, R. J. Lasek, and R. D. Alien, Science 218:1129-1131 (1982). 22. R. Vale, B. Schnapp, T. Reese, and M. P. Sheetz, Biophys. J. 45:164a (1984). 23. R. D. Alien, D. G. Weiss, J. H. Hayden, D. T. Brown, H. Fujiwake, and M. Simpson,/ Cell Biol. 100:1736-1752 (1985). 24. S. P. Gilbert and R. D. Sloboda, J. Cell Biol. 99:445-452 (1994). 25. R. E. Uzbekov and I. A. Vorobyev, Tsitologiya 33:15-21 (1991) (in Russian). 26. M. Abercrombie, Exp. Cell Research. Suppl. 8:188-198 (1961). 27. Yu. M. Vasiliev, I. M. Gelfand, and L. V. Erofeyeva, Dokl. Akad. Nauk SSSR 171:721-724 (1966) (in Russian). 28. G. I. Roskin, Mikroskopicheskaya Tekhnika (Microscopy Instruments) (Moscc 1946) (in Russian). 29. W. Seifriz, Deformation and Flow Biological System, ed. A. Frey-Wyssling (Amsterdam: North-Holland Publ., 1952): 1-156. 30. K. J. Musels, Introduction to Colloid Chemistry (New York: Wiley-Interscienc 1959). 31. V. I. Gelfand and A. D. Bershadsky, Annu. Rev. Cell Biol. 7:93-116 (1991). 32. A. Kupfer, D. Louvard, and S. J. Singer, Proc. Nat. Acad. Sci. USA 79:2603-2607 (1982). 33. G. O. Gudima, I. A. Vorobyev, and Yu. S. Chentsov, Tsitologiya 25:883-888 (1983). 34. A. Ashkin, K. Schutze, J. M. Dziedzic, U. Euteneur, and M. Schliwa, Nature 348:346348 (1990). 35. L. Green, Proc. Nat. Acad. Sci. USA 59:1179-1186 (1968). 36. D. Bickle, L. G. Tilney, and K. R. Porter, Protoplastna 61:323-345 (1966). 37. V. I. Rodionov, F. K. Gyoeva, and V. I. Gelfand, Proc. Nat. Acad. Sci. USA 88:49564960(1991). 38. P. J. Hollenbeck and D. Bray, J. Cell Sci. 105:2827-2835 (1993). 39. R. Morris and P. Hollenbeck, J. Cell Sci. 104:917-927 (1993). 40. J. D. Moore and S. A. Endow, BioEssays 18:207-219 (1996).