3015
Journal of Cell Science 107, 3015-3027 (1994)
Printed in Great Britain © The Company of Biologists Limited 1994
Evidence that kinetochore microtubules in crane-fly spermatocytes
disassemble during anaphase primarily at the poleward end
Paula J. Wilson, Arthur Forer and Cindy Leggiadro
Biology Department, York University, North York, Ontario, Canada M3J 1P3
SUMMARY
Anaphase chromosome motion involves the disassembly of
kinetochore microtubules. We wished to determine the site
of kinetochore microtubule disassembly during anaphase
in crane-fly spermatocytes.
In crane-fly spermatocyte spindles, monoclonal antibody
6-11B-1 to acetylated α-tubulin labels kinetochore microtubules almost exclusively, with an area immediately
adjacent to the kinetochore being weakly or not labelled.
This ‘gap’ in acetylation at the kinetochore serves as a
natural marker of kinetochore microtubules in the kinetochore fibre. We measured the length of the gap on kinetochore fibres in metaphase and anaphase in order to deduce
the fate of the gap during anaphase; we used this information to determine where kinetochore microtubules disassemble in anaphase. Gap lengths were measured from
confocal microscope images of fixed spermatocytes dual
labelled with 6-11B-1 to acetylated α-tubulin and YL1/2 to
tyrosinated α-tubulin, the latter being used to determine
the positions of kinetochores.
In metaphase the average gap length was 1.7 µm. In
anaphase, the gap appeared to decrease in length abruptly
by about 0.4 µm, after which it decreased in length by
about 0.2 µm for every 1 µm that the chromosome moved
poleward. PacMan models of chromosome movement
predict that this ‘gap’ in staining should disappear in
anaphase at a rate equal to that of chromosome movement.
Thus, our results do not support theories of chromosome
motion that require disassembly solely at the kinetochore;
rather, in crane-fly spermatocytes kinetochore microtubule
disassembly in anaphase seems to take place primarily at
the poles.
INTRODUCTION
microbeam of ultraviolet light was used to create areas of
reduced birefringence (ARBs) on kinetochore fibres. The
ARBs moved poleward in both metaphase and anaphase, suggesting that kMTs disassemble at the pole in both metaphase
and anaphase. However, electron microscopic studies revealed
that the ARBs are areas in which microtubules are destroyed
(Wilson and Forer, 1988), so ARB motion might be due to
polymerization and depolymerization at the severed ends of the
kMTs rather than at kinetochore and pole (discussed by Wilson
and Forer, 1989b).
In a different approach, Mitchison et al. (1986) allowed
injected biotinylated tubulin to incorporate into metaphase
BSC1 cell spindles, then fixed the cells and determined the site
of incorporation using immunogold electron microscopic techniques. The labelled tubulin was incorporated at the kinetochore in metaphase but was lost at anaphase, suggesting that
the metaphase ‘flux’ ceases in anaphase. However, using a
similar technique but injecting anaphase cells, Wadsworth et
al. (1989) found that biotin-tubulin was incorporated at the
kinetochore during early- and mid-anaphase.
A third method of marking kMTs is to photobleach local
regions of spindles that have incorporated fluorescently labelled
tubulin. In LLC-PK cells, no movement of the photobleached
area was detected in anaphase (Gorbsky and Borisy, 1989;
Gorbsky et al., 1987, 1988), suggesting that disassembly during
Anaphase kinetochore microtubule (kMT) disassembly forms
a cornerstone of many models of chromosome motion. Various
studies have attempted to determine the site of kMT disassembly in anaphase by marking the kMTs and observing how
the mark behaves. Marks that do not move but disappear as the
chromosomes move into and through the marked zone suggest
that disassembly occurs at the kinetochore, as predicted by
various models including the ‘PacMan model’ (e.g. see Cassimeris et al., 1987; Gruzdev, 1972; Pickett-Heaps et al., 1982).
Marks that move poleward with the chromosomes suggest that
disassembly takes place primarily at the pole, as a continuation of the metaphase ‘flux’, where subunits add at the kinetochore and come off at the pole (see Forer, 1965; Margolis et
al., 1978; Margolis and Wilson, 1981; Mitchison, 1989;
Mitchison and Sawin, 1990; Mitchison et al., 1986; Sawin and
Mitchison, 1991). Marks that disappear without poleward
movement suggest that microtubule turnover occurs throughout the fibre, as originally postulated by Inoué and colleagues
(Inoué and Ritter, 1975; Inoué and Sato, 1967). All three
results have been reported and thus experimental data concerning kMT disassembly during anaphase do not present a
consistent view.
In one of the first marking studies (Forer, 1965), a
Key words: mitosis, meiosis, microtubule, anaphase chromosome
motion, spindle fibre disassembly, acetylated tubulin
3016 P. J. Wilson, A. Forer and C. Leggiadro
anaphase occurs primarily at the kinetochore. In contrast, in
sand dollar eggs the photobleached area both moved poleward
and filled in (Hamaguchi et al., 1987), though it is unclear
whether the latter results were due to kMTs or other spindle
microtubules. The interpretations of photobleaching experiments also are complicated by evidence that under certain conditions irradiation of fluorophores can alter cytoskeletal components such as actin and myosin (Knight and Parsons, 1991),
and can affect the dynamic behaviour and integrity of the
labelled microtubules (Leslie et al., 1984; Vigers et al., 1988).
In PC12 neurites different photobleaching protocols can alter
dramatically the behaviour of the photobleached zones (Keith
and Farmer, 1993). Thus different photobleaching experiments
have yielded different results, with the technique itself possibly
contributing to this variation.
In the most recent approach to marking kMTs, tubulin
tagged with a non-fluorescent fluorescein derivative that
becomes fluorescent when photoactivated is allowed to incorporate into spindles; then a small region of the spindle is photoactivated, marking the microtubules. With this technique
there is unambiguous poleward motion of the photoactivated
area in anaphase A in newt lung cells, though at a rate slower
than chromosome motion (Mitchison and Salmon, 1992). The
results suggest that in newt cells the kMTs disassemble at both
the kinetochore and pole ends: early in anaphase the relative
contributions to kinetochore fibre shortening made by the
kinetochore and the pole are 75% and 25%, respectively,
shifting to 63% and 37%, respectively, later in anaphase.
Using a similar technique in LLC-PK cells, Zhai et al. (1993)
estimate that anaphase chromosome motion is 84% due to disassembly at the kinetochore and 16% due to disassembly at
the pole.
It is difficult to extrapolate any of the above results into other
cell systems, since different techniques (and different cell
types) yield different outcomes. An added concern is evidence
that different microtubule structures may contain different
arrays of tubulin isotypes (Fackenthal et al., 1993; Oka et al.,
1990), that isotype specificity may be crucial to microtubule
organization and functioning (Fackenthal et al., 1993), and that
microtubule behaviour can be altered by altering the ratios of
tubulin isotypes (e.g. see Lu and Ludueña, 1993, 1994;
Matthews et al., 1993; reviewed by Ludueña, 1993). These data
confound interpretations of experiments using microinjected
tubulin, since one does not know if subtle effects on spindle
dynamics and behaviour might be caused by the introduction
of various brain tubulin isotypes. We therefore have used a
different approach in studying chromosome motion in cranefly spermatocytes, which, in any event, are not amenable to
microinjection (Czaban et al., 1993).
In this study we have marked kMTs using antibodies to
different forms of α-tubulin. We previously showed that antibodies to acetylated α-tubulin label crane-fly spermatocyte
kMTs to the general exclusion of other spindle microtubules
(Wilson and Forer, 1989a), and that a region close to the kinetochore binds antibody poorly, resulting in a ‘gap’ in fluorescence in front of the kinetochore. This gap is a natural marker
of kMTs, visible without prior microinjection or disruption of
the cell. In anaphase, if disassembly occurs at the kinetochore,
then the gap should disappear at the same rate as chromosome
motion to the pole; if disassembly occurs at the pole, then the
gap should move poleward with the chromosome.
In the experiments described below we measured the ‘gap’
in metaphase and anaphase cells. The results suggest that
during anaphase in crane-fly spermatocytes the kMTs disassemble initially at the kinetochore and then primarily at the
pole.
MATERIALS AND METHODS
Cell preparations
Crane-fly (Nephrotoma suturalis (Loew)) spermatocytes were
prepared as described previously (Czaban and Forer, 1985a; Wilson
and Forer, 1988). Briefly, spermatocytes from fourth-instar larvae
were attached to coverslips with a fibrin clot (thrombin from Sigma
Chemical Co., St Louis, MO; fibrinogen from CalBiochem., La Jolla,
CA) under low-viscosity oil (Series 7 halocarbon oil from Halocarbon Products Corp., Hackensack, NJ). Before lysis and fixation the
oil was replaced with insect Ringer’s solution (0.13 M NaCl, 0.005
M KCl, 0.002 M CaCl2 and 0.01 M Sørensen’s phosphate buffer, final
pH 6.8).
Immunofluorescence procedures
Cells were lysed by replacing the Ringer’s solution with lysis medium
(100 mM PIPES, 10 mM EGTA, 5 mM MgSO4, 5% DMSO, 0.1%
NP40; final pH 6.9). After 10-20 minutes in lysis medium the cells
were fixed for 10 minutes in 0.5% glutaraldehyde in phosphate
buffered saline (PBS: 0.13 M NaCl, 6 mM phosphate buffer, final pH
6.9), followed by 20 minutes in NaBH4 (1 mg/ml in PBS). Coverslips
were rinsed well in PBS and immunostained as described below, or
stored in PBS/glycerol (1:1, v/v) until use.
For immunostaining, stored coverslips were first rinsed in PBS. All
steps were carried out at room temperature and all antibodies were
diluted in PBS or in PBS with 0.1% bovine serum albumin (Sigma)
and 0.02% NaN3. Coverslips were incubated for 30 minutes with
monoclonal antibody 6-11B-1 against acetylated α-tubulin (α-AC)
(Piperno and Fuller, 1985; a generous gift from Dr G. Piperno) diluted
1:10. Following the rinse steps (2× 5 minutes in PBS and 1× 5 minutes
in PBS with 0.1% Triton X-100) they were incubated for 30 minutes
with rat-absorbed FITC-conjugated goat anti-mouse IgG (Caltag Laboratories, San Francisco, CA) diluted 1:10, then rinsed as above,
incubated for 30 minutes with rat monoclonal antibody YL1/2 against
tyrosinated α-tubulin (α-TYR) (Serotec Limited, Oxford, England)
diluted 1:10,000, rinsed, and incubated for 30 minutes with mouseabsorbed Texas Red-conjugated goat anti-rat IgG (Caltag Laboratories) diluted 1:10. After a final rinse they were mounted in Mowiol
(Osborn and Weber, 1982; Rodriguez and Deinhardt, 1960) (Polysciences, Inc., Warrington, PA), which contained 0.2 g/litre of the
anti-fading agent, paraphenylene-diamine (Platt and Michael, 1983;
Storz and Jelke, 1984) (Sigma).
In control experiments we reversed the order of the staining and/or
the chromophores on the secondary antibodies and obtained the same
results in all circumstances. In order to minimize ‘bleed through’ from
the fluorescein to Texas Red channel when using confocal
microscopy, all of our quantitative data were obtained with the
staining protocol described above.
Confocal microscopy
Cells were observed with a Bio-Rad MRC-600 confocal microscope
equipped with a krypton-argon laser, attached to a Nikon Optiphot
microscope, using a ×60 objective lens (NA = 1.4). Simultaneous dual
channel imaging was used to detect α-AC (fluorescein) and α-TYR
(Texas Red) in single optical sections: each channel occupied half the
video screen (split-screen imaging). For each cell we recorded a series
of optical sections spaced 0.5-1.0 µm apart (a ‘Z-series’) and adjusted
the background and sensitivity values so that all grey-scale values
Anaphase microtubule depolymerization 3017
were from 1 to 255 (none saturated the detector). Collected images
were Kalman averaged but otherwise had no image enhancement.
Intensity analysis and gap length determinations
We determined the length of the gap in acetylation by comparing kMT
fluorescence intensity profiles in the two channels. The intensity profile
of α-TYR labelling was used to determine the position of the kinetochore and the intensity profile of α-AC labelling was used to determine
the initial position of maximum acetylation. The gap was the distance
between these two points. This analysis required that we obtain intensity
profiles along lines that were identically placed in both the fluorescein
and Texas Red channels. Software with this capability was developed
by Mr James Kelley. Briefly, images recorded on 3.5 inch optical disks
were analysed on a DOS personal computer that contained an Oculus
300 frame-grabber (Coreco, Montreal, Quebec). One of the two splitscreen images was displayed on the video monitor and a mouse-driven
cursor was used to mark the two ends of a line along a kinetochore fibre.
Then we offset the image by 384 pixels (the width of a single splitscreen image) in the X-direction to put the line in exactly the same
position in the other image. We visually determined whether the two
lines were accurately placed along the kinetochore fibre with the aid of
pre-viewed intensity plots. Once the lines were satisfactorily placed the
computer extracted the pixel intensity at each point along each line. We
stored the data on our hard disk and made working prints of each image
(with superposed lines) using a Mitsubishi video printer (model P4OU).
A separate programme converted inter-pixel distances into µm to yield
arrays of distance versus intensity for each line. Data were graphed
using Lotus 123 and SlideWrite.
Sample intensity profiles for a kinetochore fibre are shown in Fig.
1. Each profile began on the equator side of the kinetochore and ended
past the initial point of maximum intensity in acetylated tubulin (as
judged by eye). As shown in Fig. 1, the intensity of α-TYR labelling
increased rather abruptly at the kinetochore region, reaching
maximum values at distances <1 µm from the low intensity plateaus.
(These scans are similar to intensity profiles across interference
microscope images of molecularly sharp edges of myoglobin crystals
of 0.8 µm thickness in the Z direction (Huxley and Hanson, 1957).)
The intensity of α-AC labelling, on the other hand, increased with a
less steep slope: in Fig. 1 it reaches a maximum intensity about 2 µm
from the low-intensity plateau. We determined the kinetochore
position from the α-TYR profile as the intersection of a line through
the points representing the upwards slope of the curve and a horizontal
line through the points of peak intensity (point A in Fig. 1). We
estimated the position of the peak intensity at the end of the gap in
the same manner from α-AC profiles (point B in Fig. 1). The length
of the gap is the distance between points A and B.
Intensities were extracted from the raw data recorded on the hard
disk (e.g. Figs 1, 2H), or after passing the images through a low-pass
filter (e.g. Fig. 2E). The low-pass filter smoothed out the curves (by
reducing large jumps in intensities) but did not alter the positions at
which we chose the kinetochore or maximum intensities.
We discuss in the Appendix certain difficulties encountered in our
analyses of the data.
Accuracy of gap measurements
Magnification
Images were obtained over a period of 3 years using different cell
preparations and microscopes. At various ‘zoom’ settings for each
microscope we recorded images of a stage micrometer (lines spaced
10 µm apart) to determine the magnifications of each microscope. On
our personal computer we measured each scale image 4-10 times to
determine average number of pixels per µm for each microscope. The
individual values for a microscope never ranged outside 2% from the
average, generally less than 1%. For the microscope at York University, on which most of the data were taken, we recorded images of the
scale lines at five different times; with the exception of one very hot
day, the magnifications varied by less than 1% from the mean, and
Fig. 1. Intensity profiles of a kinetochore fibre dual labelled with αTYR and α-AC were obtained from confocal images of a metaphase
cell. A line was drawn along a kinetochore fibre beginning at the
equator side of the kinetochore (distance = 0). The intensity of
fluorescence for each of the two antibodies at each pixel position
along the line was then extracted and plotted. (j) Intensity of
labelling with α-TYR. (h) Intensity of labelling with α-AC. Point A
represents the kinetochore position and point B represents the
position of maximum acetylation.
the magnification on that unusual day was 8% from the average of the
other four. Thus we are confident that the magnifications we used
were accurate to within 2%, with an upper limit of 8%.
Reproducibility of gap measurements
Many fibres were analysed independently by both P.W. and A.F. (one
example is illustrated in Fig. 2). Each person also analysed certain
fibres up to 3 times. For inter-person comparisons, when A.F.’s values
were subtracted from P.W.’s values, the average difference between
the two gap measurements for 112 fibres was 0.2±0.4 (s.d.) µm, which
is not statistically different from zero using the Student’s t-test
(P>0.5), indicating that neither of us had a consistant bias in measurements. The average absolute difference between our measurements was 0.3±0.3 µm; the data presented herein are averages of the
different analyses and should therefore be accurate to within about
half that, or 0.2 µm.
Our overall accuracy can be estimated another way. Since we
measured the position of the kinetochore independently from the
position of the maximum intensity in α-AC labelling, and since in
determining ‘gap lengths’ we always subtracted the position of the
kinetochore from the position of the maximum intensity, the fact that
averaged negative values for gap lengths were all within 0.1 µm from
zero (Fig. 6) suggests that the gap lengths were accurate to within 0.10.2 µm.
Photography
Colour prints were made from slides photographed from pseudocolour
images on the Bio-Rad monitor (using a Nikon F3 camera and
Fujichrome film, ASA 50). Prints were made with the Ilfochrome
Classic colour print system (Ilford, Mobberley-Cheshire). Black and
white prints are from the Mitsubishi video printer.
RESULTS
Description of cells
Crane-fly spermatocytes in meiosis-I have three autosomal
bivalents and two univalent sex chromosomes. The sex chromosomes, which are amphitelic (i.e. they have kinetochore
3018 P. J. Wilson, A. Forer and C. Leggiadro
fibres to both poles), remain at the equator in anaphase as the
autosomal half-bivalents move poleward. The sex chromosomes enter anaphase once the autosomes have reached the
poles; as they move to opposite poles each maintains kinetochore fibres to each pole. The spindles do not elongate until
the sex chromosomes segregate in anaphase (Forer, 1964,
1980). ‘Precocious’ sperm tail flagella emanate from the
centriole at each pole: two flagella per pole in meiosis-I and
one per pole in meiosis-II.
We analyzed cells in metaphase-I and anaphase-I. At the
time of data acquisition we determined the division stage of
each cell by phase-contrast microscopy. However, since
mistakes are sometimes made, prior to analysis we reviewed
the fluorescence images of each cell to confirm the stage it was
in. We distinguished meiosis-I from meiosis-II based on cell
size, number of spindle fibres and number of flagella at the
poles. For cells in meiosis-I, we distinguished prometaphase
from metaphase by the following criteria: during metaphase
chromosomes are at the equator, the kinetochore fibres are
thicker and more prominent and spindles label well for acetylated tubulin. It was difficult to distinguish between genuine
metaphase and late prometaphase, though we tried to be
ruthless in eliminating all cells that might be in late
prometaphase. We distinguished metaphase from anaphase on
the basis of chromosome position and kinetochore fibre length.
However, since we could not always see the chromosomes
clearly in the fluorescent images, we may have made some
mistakes in distinguishing metaphase from very early
anaphase.
In cells dual labelled with α-TYR and α-AC, labelling with
α-TYR was similar to that described earlier in cells labelled
with polyclonal antibodies to tubulin (Wilson and Forer,
1989a), revealing kMTs, non-kinetochore microtubules and
astral microtubules (Fig. 2). In contrast, α-AC labelling was
present only in the kinetochore fibres and the flagella at the
poles (Fig. 2).
In kinetochore fibres at metaphase, α-TYR labelling was of
more or less constant intensity from the kinetochore to the pole
(Fig. 2). α-AC labelling, however, was less intense closer to
the kinetochore, leaving a gap in labelling (Fig. 2; Wilson and
Forer, 1989a). In the fluorescent images, the autosomal kinetochore fibres were about 0.9-1 µm in diameter; in the α-TYR
channel the kMT bundles were of more or less constant
diameter but in the α-AC channel the portions closest to the
kinetochores sometimes seemed to taper or decrease in
diameter (Fig. 2). The kinetochore fibres associated with the
sex chromosomes appeared to have diameters of about 0.5-0.6
µm, noticeably smaller than those associated with the
autosomes. (Sex chromosome fibres are more difficult to detect
in the photographs presented here because they fluoresce less
intensely and tend to be lost when exposures are set to optimize
autosomal fibres.) The autosomal fibres often seemed to be
composed of two sub-fibres, apparently one from each of the
two chromatids associated with a spindle pole (data not
shown). Intensity scans of α-AC labelling showed that flagella
generally labelled with very similar intensity to kMT bundles
(data not shown).
We measured the gaps in α-AC labelling in metaphase and
anaphase cells and we used these data to deduce the fate of the
gap in anaphase.
At metaphase-I, the average length of the gap in αAC labelling is 1.7 µm and is independent of
kinetochore fibre length
We measured gaps in 74 kinetochore fibres from 16 metaphase
cells, one of which is illustrated in Fig. 2. The average gap
length was 1.7±0.7 µm (s.d.). We also normalized all lengths
to correspond to the average pole-to-pole length of 21.5 µm,
in case the gap lengths were influenced by spindle size. This
analysis did not reveal any new trends but did narrow the
ranges of kinetochore-to-pole distances (kinetochore fibre
lengths), as listed in Table 1. To be more certain of our conclusions, we analyzed both normalized and non-normalized
data.
Gap sizes varied considerably but were independent of the
kinetochore fibre length, judging by the nearly horizontal bestfit line for the data when plotting gap length versus kinetochore
fibre length (Fig. 3A). Data from Fig. 3A also were grouped
according to kinetochore fibre length within 1 µm ranges (e.g.
6.0-6.9 µm, 7.0-7.9 µm) and the average gap length for each
fibre length category was calculated (Fig. 3B). For kinetochore
fibre lengths between 7 and 11 µm, the average gap lengths are
not statistically different using ANOVA analysis (P>0.25),
again suggesting that the gap length is independent of kinetochore fibre length. Similar results were obtained with not-normalized data (data not shown). The slight positive slope in Fig.
3A is due to slightly lower values in the 6-8 µm range, which
may be due to the small sample size in this range, or to our
mistaking early anaphase for metaphase.
In conclusion, the average gap length is 1.7 µm, near that
estimated crudely by Wilson and Forer (1989a), and, with the
possible exception of small kinetochore fibre lengths, the
length of the gap is independent of kinetochore fibre length in
metaphase.
In anaphase-I, the gap length decreases at a rate of
approximately 0.2 µm per 1 µm of kMT shortening
We measured gaps in 116 kinetochore fibres from 23 anaphase
cells at various stages of anaphase, two of which are illustrated
in Figs 4 (mid-anaphase) and 5 (late anaphase). We plotted
anaphase gap lengths versus kinetochore fibre lengths, both as
individual points (Fig. 6A) and grouped according to kinetochore fibre lengths (Fig. 6B). There was a clear relationship
between decreasing gap size and decreasing kinetochore fibre
lengths. The best-fit lines for the scatter diagram (Fig. 6A) and
the grouped data (Fig. 6B) both had slopes of 0.2 µm/µm, indicating that the gap decreases by approximately 0.2 µm for each
Table 1. Averages and range of values for measurements
from metaphase cells
Average±s.d. (µm)
Range (µm)
Measurement
Not norm.
Norm.
Not norm.
Norm.
Gap
Pole-to-pole
KT-to-pole
1.7±0.7
21.5±2.9
9.0±1.6
1.7±0.8
21.5
9.0±0.7
0.2-3.0
16.6-28.0
5.3-12.8
0.2-3.3
−
6.9-10.6
The gap represents the distance between the kinetochore and the initial
point of maximum intensity in labelling with α-AC. Values are for spindle
lengths that have been normalized for the average pole-to-pole distance of
21.5 µm (Norm.) and for lengths that have not been normalized (Not norm.).
Anaphase microtubule depolymerization 3019
Fig. 2. A metaphase spindle dual labelled with α-TYR and α-AC as visualized by dual channel imaging with the confocal microscope.
(A,C,F) Texas Red channel (α-TYR). (B,D,G) Fluorescein channel (α-AC). Only the kinetochore fibres and the flagella label strongly with αAC. (A,B) Ten optical sections taken through the cell were superposed to form an image of the complete spindle (a Z-series). (C,D) Single
optical sections of the cell in A,B showing the line along the kinetochore fibre from which the intensity profile (E) was derived. (E) Intensity of
pixels along the lines in C (m) and D (n), beginning at the end labelled 0. The image was smoothed using a low-pass filter prior to extracting
pixel intensities. (F-H) A second, independent scan of the same fibre, by a different person, without using the low-pass filter. Bars, 5 µm.
3020 P. J. Wilson, A. Forer and C. Leggiadro
A
B
Fig. 3. Gap length vs kinetochore fibre length in metaphase. (A) For
each fibre, the length of the gap in acetylation was plotted against the
kinetochore-to-pole length. Each point represents data from a single
kinetochore fibre. The best-fit line for the data points is shown.
(B) Values from A were grouped according to kinetochore-to-pole
length within 1 µm ranges, and for each group an average gap value
was calculated and plotted. The group labelled 6.5 represents data for
kinetochore-to-pole lengths of 6.0-6.9 µm, etc. Error bars = s.d.
Numbers above the error bars represent the number of points in that
group. All values are normalized to spindle lengths of 21.5 µm.
1 µm that the kMTs shorten. Similar values were obtained with
not-normalized data (data not shown). If depolymerization
were occurring solely at the kinetochore, we would expect the
slope to be unity, i.e. the gap would disappear at the same rate
as kinetochore-to-pole shortening (Fig. 6A). Clearly, then, the
gap decreases in size during anaphase, but not at the rate
expected if disassembly were solely at the kinetochore. To be
sure that this conclusion is valid, we analysed the data in two
other ways.
First, we plotted gap lengths versus interkinetochore
distances (the distances between the two kinetochores of the
separating half-bivalents). In anaphase the gap decreases by
~0.1 µm for every 1 µm increase in interkinetochore distance,
as seen in both the scatter diagram (Fig. 7A) and the grouped
data (Fig. 7B). Since the interkinetochore distance increases 2
µm for every 1 µm that each kinetochore fibre shortens, the
results are exactly the same as from the analysis with respect
to kinetochore fibre length.
Second, we used the data on gap length in metaphase (Fig.
3) to predict the kinetochore fibre length at which each gap
would disappear if the shortening was exclusively at the kinetochore; we then compared the predicted with the actual
values. For example, consider a point in Fig. 3 in which the
gap length is 1.9 µm and the kinetochore fibre length is 8.6
µm. If in anaphase this fibre shortened only from the kinetochore end, the gap would disappear at a fibre length of (8.6
µm-1.9 µm) = 6.7 µm. We calculated such predicted values for
each point in Fig. 3A and plotted the results as a cumulative
percentage of fibre lengths at which the fibres have zero gaps
(Fig. 8). By the time the kinetochore fibres shorten to 6.4 µm,
75% of the anaphase kinetochore fibres should have zero gap
lengths if disassembly is solely at the kinetochore (broken line
in Fig. 8), yet the actual measurements (Fig. 6) show that the
gaps are only beginning to be zero at kinetochore fibre lengths
of 6.4 µm. Thus, this alternative approach also suggests that
most of the disassembly of kMTs during anaphase is not at the
kinetochore end of the fibre.
Evidence for a short burst of disassembly at the
kinetochore at anaphase onset
We extrapolated the best-fit line for the anaphase graph of gap
length versus kinetochore fibre length (Fig. 6A) back to the
average metaphase kinetochore fibre length of 9.0 µm. The
expected metaphase gap length from the extrapolated line is
1.3 µm, which is shorter by 0.4 µm than the measured
metaphase value of 1.7 µm (Table 1). Similarly, when the bestfit line for gap length versus interkinetochore distance in
anaphase is extrapolated to the average metaphase interkinetochore length of 4.2 µm, the gap length is shorter by 0.4 µm
than the measured metaphase gap length of 1.7 µm. These
statements also are true for the grouped data (Figs 6B and 7B),
and for the best-fit lines through the scatter data omitting kinetochore lengths #4.5 (see legend of Fig. 6). Though the difference of 0.4 µm is small, it suggests to us that there may be
a rapid loss of gap length at the start of anaphase.
We conclude, then, that the gap may decrease quickly by
about 0.4 µm at the onset of anaphase, but it then decreases at
a rate of 0.2 µm per 1 µm of kinetochore fibre shortening.
Gap lengths on amphitelic sex-chromosome fibres
During sex-chromosome anaphase, the amphitelic sex chromosomes move poleward at about one-fifth the speed of the
autosomes (Schaap and Forer, 1979). As each moves, one kinetochore fibre shortens and the other kinetochore fibre elongates
(Forer, 1980). In our earlier study (Wilson and Forer, 1989a)
we noted that the gap on the elongating fibre appeared to
increase in length while that on the shortening fibre decreased
in length. Measurements of gaps on sex-chromosome fibres
during anaphase (Fig. 9) confirm these initial qualitative observations. Thus as the sex chromosomes separate in anaphase the
gap length decreases on the shortening fibre and increases on
the elongating fibre.
DISCUSSION
We have used quantitative confocal microscopy of both
metaphase and anaphase cells to measure the length of the
kinetochore fibre in front of the kinetochore that does not label
strongly with α-AC. At the start of anaphase the ‘gap’ seems
to shorten quickly by about 0.4 µm; during the rest of anaphase
the gap decreases in length at a rate of 0.2 µm for every 1 µm
that the kinetochore fibre shortens. Assuming that the gap is a
Anaphase microtubule depolymerization 3021
Fig. 4. The gap in mid-anaphase. (A,B) Confocal Z-series of a cell in mid-anaphase dual-labelled with α-TYR (A) and α-AC (B). (C,D) Single
optical sections from the Z-series shown in A,B, showing the lines from which fluorescence intensity profiles were taken. (E) The fluorescence
intensity profiles from the lines depicted in C,D for labelling with α-TYR (C, m) and α-AC (D, n). In this cell a gap is still present. Bars, 5 µm.
marker of kMTs in anaphase, these results suggest that during
anaphase the kMTs initially disassemble at the kinetochore but
during the remainder of anaphase at least 80% of kMT shortening is due to disassembly at the pole, in agreement with the
hypothesis that during anaphase both the kinetochore and the
pole can act as sites of microtubule disassembly (Forer and
Wilson, 1994).
The above interpretation rests heavily on certain assumptions regarding the nature of the gap; for this reason further
discussion of the gap is warranted.
We explain the absence of acetylated kMTs near the
metaphase kinetochore as follows: kMTs constantly polymerize (add subunits) at the kinetochore, the subunits move (‘flux’)
towards the pole where they are removed, and the ‘gap’ represents the time lag between polymerization and acetylation.
We favour this hypothesis for the following reasons (also see
Wilson and Forer, 1989a).
With respect to incorporation of tubulin into kMTs, there
now is ample evidence that tubulin incorporates at the kinetochore during metaphase. This evidence arises from different
experimental approaches, such as preventing disassembly by
colcemid using UV light (Czaban and Forer, 1985b), electron
microscopy after microinjection of biotin-labelled tubulin
(Mitchison et al., 1986), fluorescence microscopy after
microinjection of DTAF-tubulin (Wise et al., 1991), studying
fluorescence recovery after photobleaching (Gorbsky and
Borisy, 1989), and photoactivation of ‘caged’ tubulin
(Mitchison, 1989; Mitchison and Salmon, 1992).
With respect to acetylation, the evidence from the literature
supports the view that acetylation occurs after polymerization.
In Chlamydomonas reinhardtii acetylation of tubulin is closely
3022 P. J. Wilson, A. Forer and C. Leggiadro
Fig. 5 The gap in late anaphase. (A,B) Confocal Z-series of a cell in late anaphase dual-labelled with α-TYR (A) and α-AC (B). (C,D) Single
optical sections from the Z-series shown in A,B, showing the lines from which fluorescence intensity profiles were taken. (E) The fluorescence
intensity profiles from the lines depicted in C,D for labelling with α-TYR (C, m) and α-AC (D, n). In this cell a gap is absent. Bars, 5 µm.
tied to flagellar growth and hence to microtubule polymerization (Brunke et al., 1982; L’Herneault and Rosenbaum, 1983).
Only assembled tubulin in the flagellum is acetylated (Greer
and Rosenbaum, 1989), and while tubulin dimers can be acetylated, the polymer is a better substrate than the dimer (Maruta
et al., 1986). In tissue culture cells, acetylated tubulin is found
only in the pellet fraction containing microtubules (Piperno et
al. 1987). For newly polymerized microtubules in cells, acetylation lags behind microtubule assembly by 3-15 minutes
(dePennart et al., 1988; Piperno et al., 1987; Wilson and Forer,
1989a). Lastly, cytoplasmic microtubules stabilized by taxol
become increasingly acetylated over time (Piperno et al.,
1987), suggesting that acetylation is a post-polymerization
event.
In summary, previous work supports the view that in
metaphase tubulin adds at the kinetochore and acetylation takes
place some time after polymerization. That the gap gradually
increases in fluorescence as one follows the kMTs from the
kinetochore towards the pole also is consistent with an increase
in acetylation with time as the subunits flux towards the pole.
To extend our explanation to anaphase, we assume that polymerization at the kinetochore ceases in anaphase. Although
some recent evidence suggests that tubulin can add to kMTs
during anaphase (Shelden and Wadsworth, 1992; Wadsworth
et al., 1989), and we have argued that the growth of UVsevered kMTs during anaphase is due to addition of subunits
at the kinetochore (Forer and Wilson, 1994), these examples
appear to be a result of the experimental treatment itself and
not something that occurs under normal conditions (e.g. see
Shelden and Wadsworth, 1992). Given that polymerization at
the kinetochore ceases in anaphase, the fact that the gap
persists once the chromosomes have moved the 1.7 µm length
Anaphase microtubule depolymerization 3023
A
A
B
B
Fig. 6. Gap length vs kinetochore fibre length in anaphase. (A) The
individual gap lengths from kinetochore fibres in anaphase cells were
plotted against the kinetochore-to-pole length. Each point represents
one kinetochore fibre. The continuous line is the best fit line for the
data (slope = 0.2 µm/µm). The dotted line (...) represents the
expected best-fit curve if disassembly were entirely at the
kinetochore (slope = 1 µm/µm). The broken line (---) represents one
possible best-fit curve if disassembly were entirely at the pole (slope
= 0 µm/µm). Note that the x-axis has been inverted in order to depict
better the shortening of the kinetochore-to-pole distance as anaphase
proceeds. From the graph there appears to be a marked drop in gap
lengths at kinetochore-to-pole lengths of # 4.5. Without repeating
the experiment on a second set of cells, we cannot tell whether this is
meaningful or whether it is an artifact of the small number of fibres
studied. In any event, if one omits the values at kinetochore fibre
lengths # 4.5 µm, the slope of the best-fit curve for the remaining
data decreases slightly in the second decimal place but at our level of
significant digits remains 0.2 µm/µm. (B) The data in A were
grouped according to kinetochore-to-pole length and the average gap
value for each length group was calculated. Each group represents
the data for kinetochore-to-pole lengths within a 1 µm range (e.g. 9.5
represents data for kinetochore-to-pole lengths of 9.0-9.9 µm). Error
bars represent standard deviations. Numbers above error bars
represent the number of points in that group. All values are
normalized to a spindle length of 21.5 µm.
of the gap (Fig. 6) strongly suggests that disassembly does not
occur primarily at the kinetochore. A straightforward interpretation of the loss of gap by 0.2 µm for every 1 µm shortening
of kMTs is that 80% of the shortening is due to disassembly
at the poles and 20% to disassembly at the kinetochores.
There is an alternative interpretation for our results,
however: it is possible that the gap in metaphase is due to
deacetylating enzymes emanating from (or activated by) the
Fig. 7. Gap length vs interkinetochore distance. (A) Individual gap
lengths were plotted against interkinetochore distance. The best-fit
line for the data points is shown (slope = −0.1 µm/µm).
(B) Values from A were grouped according to interkinetochore
distances within 1 µm ranges, and for each group an average gap
value was calculated and plotted. (The group labelled 4.5 represents
data for interkinetochore distances of 4.0-4.9 µm, etc.) The number
above each group represents the number of fibres in the group. Error
bars represent standard deviation. All values are normalized to a
spindle length of 21.5 µm.
kinetochore. Similarly, in anaphase the persisting gap could
represent a ‘wave’ of deacetylation in front of the kinetochore.
We outlined previously reasons why we think that this is
unlikely (Wilson and Forer, 1989a). Briefly, we know that
tubulin can add at the kinetochore and that there likely is a lag
between polymerization and acetylation, so our explanation fits
known data. We know of no precedent for expecting newly
polymerized tubulin to be deacetylated, on the other hand. Nor
is there any precedent we know of for microtubules being
deacetylated without first being disassembled. Secondly, we do
not see a reasonable explanation for why the gap gradually
would decrease in length during anaphase when the chromosomes move poleward if the gap is due to a zone of deacetylation. Thirdly, the gap associated with sex-chromosomal
fibres in metaphase is smaller than that associated with
autosomal fibres in metaphase, 1.1 µm±0.4 µm (Fig. 9)
compared with 1.7 µm±0.7 µm (Table 1), a difference that is
statistically significant (using Student’s t-test) at a level of
P<0.001. Whereas this might be understandable in terms of our
assumptions if this is related to the slower speeds with which
the two kinds of chromosomes move (and therefore slower
rates of incorporation), we see no reason why deacetylation
levels should be different in the two cases. Finally, when sex
chromosomes move poleward in anaphase, the elongating fibre
3024 P. J. Wilson, A. Forer and C. Leggiadro
Fig. 8. Predicting the anaphase outcome for metaphase cells if
disassembly of kMTs is entirely at the kinetochore. If disassembly
occurs entirely at the kinetochore, then the predicted fibre length for
which the gap would be 0, once metaphase cells entered anaphase,
would be equal to the metaphase kinetochore-to-pole length minus
the gap length for that fibre. This value was calculated for each of the
fibres represented in Fig. 3 and plotted above against the cumulative
percentage of the total number of fibres in Fig. 3. The vertical broken
line represents the point at which 75% of the fibres should have no
gap, if disassembly occurs solely at the kinetochore.
has an increased gap length while the shortening fibre has a
decreased gap length (Wilson and Forer, 1989a; Fig. 9): this
is understandable from our working hypothesis, since the elongating fibre must have an increased rate of incorporation in
order to elongate, but is not understandable if one assumes that
the gap is due to a region of deacetylation. All these ‘indirect’
arguments are reasons why we think that the deactylation
explanation is unlikely.
We could decide directly if deacetylation is a possible explanation if we knew the characteristics of the deacetylation
enzymes, and whether they can cause deacetylation of polymerized tubulin. Unfortunately there is very little relevant
information available. To our knowledge, the only study that
has approached this topic directly is that of Maruta et al.
(1986), who detected tubulin deacetylase activity in Chlamydomonas reinhardtii, presumably due to the presence of the
deacetylating enzyme. The activity was detected only in the
cell body, whereas disassembly of acetylated microtubules in
flagella appears to take place within the flagellum (Greer and
Rosenbaum, 1989), suggesting that deacetylation occurs after
disassembly. In addition, the acetylation state of tubulin does
not affect assembly/disassembly kinetics, and therefore there
is no need to deacetylate tubulin prior to disassembly (Maruta
et al., 1986). These data are consistent with the suggestion that
deacetylation occurs after disassembly, but they do not provide
unequivocal evidence that this is the only way in which tubulin
can be deacetylated.
In conclusion, we feel that the existing data favour the
hypothesis that in metaphase the gap arises because of a lag
between polymerization and acetylation of microtubules, with
the corollary that in anaphase the gap persists for the same
reason.
Is gap loss due to disassembly or to acetylation?
If one accepts that the gap in metaphase kMTs arises because
of a time lag between polymerization and acetylation, then one
Fig. 9. Gaps on sex chromosome fibres during sex chromosome
anaphase. As sex chromosomes move poleward, they maintain
kinetochore fibres to both poles, so that one fibre shortens and the
other elongates. Gaps in acetylation measured for sex chromosomes
that were still at the equator (d) and for both elongating (n) and
shortening (crosses) sex chromosome fibres in anaphase were plotted
against the distance between the two separating sex chromosomes.
The continuous line represents the best-fit linear curve for the values
from shortening fibres (slope = −0.1 µm/µm) and the broken line
represents the best-fit linear curve for the values from the elongating
fibres (slope = 0.2 µm/µm). The four data points for chromosome
separation of 5.4 µm are all from the same cell and therefore may not
be representative.
must ask how the gap disappears during anaphase. Two
possible mechanisms can account for it: (a) disassembly at the
kinetochore, or (b) continued acetylation of the polymerized
microtubules once polymerization ceases and the chromosome
begins to move poleward. While we cannot distinguish
unequivocally between the two possibilities, acetylation could
account for the entire loss of the gap, in the following way.
The metaphase gap is 1.7 µm long, and the gap shortens at
a rate of 0.2 µm per 1 µm of shortening of kMTs; thus the gap
would disappear on average after the kinetochore fibre shortens
by about 8.5 µm (1.7 µm/0.2 µm). N. suturalis chromosomes
move poleward at about 0.7-1 µm per minute (Marzec, 1993;
Shaap and Forer, 1979); shortening by 8.5 µm corresponds,
then, to 8.5-12 minutes. Experimentally, microtubules that
repolymerize become acetylated within 3-15 minutes after
reappearing (de Pennart et al. 1988; Piperno et al., 1987;
Wilson and Forer, 1989a). Thus, it is conceivable that during
anaphase the loss of the gap occurs by acetylation, and that
there is no disassembly at the kinetochore.
Our overall interpretation, then, is that there is an initial, brief
disassembly of kMTs at the kinetochore at the start of anaphase,
but that the remainder of anaphase occurs by disassembly at the
pole. We might speculate that this represents the general rule,
that earlier anaphase is predominantly due to disassembly at the
kinetochore whereas later anaphase is due to disassembly at the
pole, since this also seems to occur in at least some newt cells:
early in anaphase there was primarily disassembly at the kinetochore but later in anaphase, “.. in some cells the kinetochores
and the marks (on the ktMTs) tended to move together, indicating that disassembly was now mainly at the polewards end
of the fiber” (Mitchison and Salmon, 1992).
Anaphase microtubule depolymerization 3025
Comparing our results with other studies
Like previous studies using caged fluorescence, we find
evidence for disassembly both at the pole and at the kinetochore (Mitchison and Salmon, 1992; Zhai et al., 1993). In
addition, like Mitchison and Salmon (1992), we find a greater
contribution due to disassembly at the kinetochore early in
anaphase.
The key difference between our results and those of other
marking experiments is the relative contribution made by the
pole during disassembly. Our results suggest that, after the start
of anaphase, at least 80% of the kinetochore fibre shortening is
due to disassembly at the poleward end of the kMTs. This
contrasts dramatically with the range of 0 to 37% seen by others
(Gorbsky and Borisy, 1989; Gorbsky et al., 1987; 1988, Salmon
and Mitchison, 1992; Zhai et al., 1993). While this difference
could be due to different experimental approaches, we think that
both disassembly at the kinetochore and disassembly at the pole
take place during anaphase, to different degrees in different cell
types and different circumstances. We have suggested
elsewhere (Forer and Wilson, 1994) that the kMTs may be
pushed poleward by motor molecules in the spindle matrix, and
that, as a consequence, in different cells (or circumstances) one
would expect disassembly predominantly at the kinetochore (as
in newt cell early anaphase), or predominantly at the pole (as
in most of crane-fly spermatocyte anaphase), or mixtures of the
two (as in some newt and tissue culture cells). Thus, the fact
that different cells have varying degrees of one form of
anaphase disassembly or the other does not necessarily mean
that the basic mechanisms for movement are different.
How many of the kMTs are acetylated?
Without analysis by immunoelectron microscopy, one cannot
tell how many of the kMTs in crane-fly spermatocytes are
acetylated. Nonetheless, rough estimates can be made based on
our observation that the kMT bundles in one optical section
have about the same level of acetylation as the flagella at the
spindle poles in the same cells, and that flagellar microtubules
are highly acetylated (Brunke et al., 1982; Glyn and Gull,
1990; L’Herneault and Rosenbaum, 1983; Sasse et al., 1987).
Given the diameter and microtubule density in a flagellum
(Behnke and Forer, 1967), a kinetochore fibre diameter of 0.9
µm (see results), an estimated depth of field of 0.7 µm (Fricker
and White, 1992), and a kMT density of 125 to 160 kMTs per
µm2 (based on data from LaFountain, 1976; Janicke and
LaFountain, 1984; Scarcello et al, 1986), we estimate that
either 70% to 90% of the kMTs in the bundle are acetylated to
the same degree as the flagellar microtubules, or that all of the
kMTs are acetylated but only to 70% to 90% of the level of
acetylation of the flagellar microtubules. We realize that these
estimates are very crude, but they indicate that in studying the
acetylated kMTs in crane-fly spermatocytes we have not
studied a trivially minor portion of the kMT bundle, but rather
a sizeable fraction of the kMTs.
In conclusion, we have marked kMTs using a novel, noninvasive approach; namely, the differential staining pattern
seen on kMTs labelled with antibodies to acetylated α-tubulin.
We have argued that the gap in acetylation represents the time
lag between polymerization of tubulin and acetylation; that
from measurements of gap sizes in anaphase the initial stage
of anaphase is due to disassembly at the kinetochore and the
remainder of anaphase is due to disassembly at the pole; and
that the acetylated kMTs make up a sizable fraction of the kMT
bundle.
We thank Dr B. Czaban and the reviewers for helpful comments
on the manuscript, S. Perreira and Drs J. Shore, M. Sokolowski and
M. Blows for help and advice on statistical analysis and Dr P. Moens
for help in making colour prints. The work was supported financially
by grants from the Natural Sciences and Engineering Research
Council of Canada.
APPENDIX
Problems and controls
Choosing the best optical section
The purpose of using confocal microscopy is to resolve a single
kinetochore fibre without interference from fluorescence above
and below it. We encountered some experimental difficulties,
however. One concerned focus. Unless the focus steps are very
small (and hence the number of sections unreasonably large)
the images recorded may not be in optimal focus for some of
the fibres in the cell. Consequently, some kinetochore fibres
were in focus in more than one optical section. For these fibres,
and for kinetochore fibres that were at an angle to the plane of
focus (e.g. see Figs 2, 5), we chose the section in which the
region near the kinetochore was best visualized. When we
could not do this satisfactorily, the fibres (or cells) were not
used in our analysis.
Focus shifts
Dual images were recorded simultaneously for each single
optical section, but sometimes there were vertical and horizontal phase shifts between the Texas-Red and the FITC
images. By vertical shift we mean that a kinetochore fibre in
one channel would appear best focussed in one optical section,
but in the other channel the same fibre would appear best
focussed in an adjacent optical section. To study how the
vertical shift arose, we labelled interphase cells with one antitubulin antibody and two different secondary antibodies (fluorescein-labelled and Texas Red-labelled). In imaging single
microtubules we did not see shifts in focus. Thus we attribute
the vertical shifts in focus with the kMT bundles to the fact
that the bundles were of the order of 1 µm in diameter and that
because of the way the series of pictures were taken, the level
of focus might have been slightly off from the best (optimal)
focus. As a consequence, the chromatic aberration in the
lenses, estimated theoretically (Martin, 1966) and experimentally (Fricker and White, 1992) to be of the order of half the
depth of field, was enough to cause the observed shift in focus.
In support of this interpretation we note that in general only
some of the fibres in a given cell were shifted in focus. Regardless of the correctness of our interpretation, each individual
fibre was considered separately and the best optical section for
each channel was chosen.
Lateral shifts in focus also occurred, but generally were not
a problem, in that we usually could obtain a line that accurately
reflected both the kinetochore position (α-TYR labelling) and
the intensity of acetylation along the kMTs (α-AC labelling).
On a very few occasions we had to shift one of the two images
3026 P. J. Wilson, A. Forer and C. Leggiadro
by 1 or 2 pixels (corresponding to ~0.1-0.2 µm) in order to
bring the two lines along the kMTs into closer register.
Overlapping fibres
Out of focus fibres sometimes interfered with our analysis.
P.W. and A.F. both separately determined gap lengths for
many spindle fibres, locating the position of the kinetochore
independently of the end of the gap. We then compared curves
and lengths of the gaps. When the gap lengths differed by more
than 0.5 µm P.W. and A.F. each re-analysed the same fibres.
In the re-analysis we eventually realized that the major cause
of disagreements was vertical overlap of kinetochore fibres.
Vertical overlap occurred when fibres that were superposed
in the X and Y direction but separated vertically at the kinetochores (i.e. were in the same position in the X and Y directions
but were in different sections in the Z-series) came together
(‘merged’) closer to the poles. The overlapping fibre contributed to the intensities that we measured along what we
thought were individual kinetochore fibres and consequently
we sometimes saw more than one peak in intensity along a
kinetochore fibre with α-AC labelling, with the second peak
higher in intensity and closer to the pole. (In our original
analysis (Wilson and Forer, 1993) we mistakenly chose the
second peak.) This difficulty did not occur regularly, but was
not infrequent; it arose primarily when the image was not in
the best plane of focus for the fibre in question. We eliminated
this difficulty by analysing the intensities in the fluorescein
channel (α-AC) in the optical section further removed by one
from the offending kinetochore fibre or by choosing the first
intensity peak. If neither were possible, we eliminated the
fibres from the analysis.
Lateral overlap, from kinetochore fibres that were spatially
separated laterally at the kinetochores but that merged as the
fibres came closer to the pole, also occurred but easily was recognized and avoided.
REFERENCES
Behnke, O. and Forer, A. (1967). Evidence for four classes of microtubules in
individual cells. J. Cell Sci. 2, 169-192.
Brunke, K. J., Collis, P. S. and Weeks, D. P. (1982). Post-translational
modification of tubulin dependent on organelle assembly. Nature 297, 516518.
Cassimeris, L. U., Walker, R. A., Pryer, N. K. and Salmon, E. D. (1987).
Dynamic instability of microtubules. BioEssays 7, 149-154.
Czaban, B. B. and Forer, A. (1985a). The kinetic polarities of spindle
microtubules in vivo, in crane-fly spermatocytes. I. Kinetochore
microtubules that re-form after treatments with colcemid. J. Cell Sci. 79, 137.
Czaban, B. B. and Forer, A. (1985b). The kinetic polarities of spindle
microtubules in vivo, in crane-fly spermatocytes. II. Kinetochore
microtubules in non-treated spindles. J. Cell Sci. 79, 39-65.
Czaban, B. B., Forer, A. and Wise, D. A. (1993). Microinjection into crane-fly
spermatocytes. Biochem. Cell Biol. 71, 222-228.
de Pennart, H., Houliston, E. and Maro, B. (1988). Post-translational
modifications of tubulin and the dynamics of microtubules in mouse oocytes
and zygotes. Biol. Cell 64, 375-378.
Fackenthal, J. D., Turner, F. R. and Raff, E. C. (1993). Tissue-specific
microtubule functions in Drosophila spermatogenesis require the ß2-tubulin
isotype-specific carboxy terminus. Dev. Biol. 158, 213-227.
Forer, A. (1964). Evidence for two spindle fiber components: a study of
chromosome movement in living crane fly (Nephrotoma suturalis)
spermatocytes, using polarization microscopy and an ultraviolet microbeam.
Ph.D. thesis, Dartmouth College, Hanover, NH.
Forer, A. (1965). Local reduction of spindle fiber birefringence in living
Nephrotoma suturalis (Loew) spermatocytes induced by ultraviolet
microbeam irradiation. J. Cell Biol. 25, 95-117.
Forer, A. (1980). Chromosome movements in the meiosis of insects, especially
crane-fly spermatocytes. In Insect Cytogenetics (ed. R. L. Blackman, G. M.
Hewitt and M. Ashburner), pp. 85-95. Blackwell Scientific Publications,
Oxford.
Forer, A. and Wilson, P. J. (1994). A model for chromosome movement
during mitosis. Protoplasma 179, 95–105.
Fricker, M. D. and White, N. S. (1992). Wavelength considerations in
confocal microscopy of botanical specimes. J. Microsc. 166, 29-42.
Glyn, M. and Gull, K. (1990). Flagellum retraction and axoneme
depolymerisation during the transformation of flagellates to amoebae in
Physarum. Protoplasma 158, 130-141.
Gorbsky, G. J., Sammak, P. J. and Borisy, G. G. (1987). Chromosomes
move poleward in anaphase along stationary microtubules that coordinately
disassemble from their kinetochore ends. J. Cell Biol. 104, 9-18.
Gorbsky, G. J., Sammak, P. J. and Borisy, G. G. (1988). Microtubule
dynamics and chromosome motion visualized in living anaphase cells. J.
Cell Biol. 106, 1185-1192.
Gorbsky, G. J. and Borisy, G. G. (1989). Microtubules of the kinetochore
fiber turn over in metaphase but not in anaphase. J. Cell Biol. 109, 653-662.
Greer, K. and Rosenbaum, J. L. (1989). Post-translational modifications of
tubulin. In Cell Movement vol. 2: Kinesin, Dynein and Microtubule
Dynamics (ed. F. D. Warner and J. R. McIntosh), pp. 47-66. Alan R. Liss,
New York.
Gruzdev, A. D. (1972). Critical review of some hypotheses of anaphase
chromosome movements. Tsitologiya 14, 141-149.
Hamaguchi, Y., Toriyama, M., Sakai, H. and Hiramoto, Y. (1987).
Redistribution of fluorescently labeled tubulin in the mitotic apparatus of
sand dollar eggs and the effects of taxol. Cell Struct. Funct. 2, 43-52.
Huxley, H. E. and Hanson, J. (1957). Quantitative studies on the structure of
cross-striated myofibrils. I. Investigations by interference microscopy.
Biochim. Biophys. Acta 23, 229-249.
Inoué, S. and Sato, H. (1967). Cell motility by labile association of molecules.
J. Gen. Physiol. 50, 259-292.
Inoué, S. and Ritter, H. Jr (1975). Dynamics of mitotic spindle organization
and function. In Molecules and Cell Movement (ed. S. Inoué and R. E.
Stephens), pp. 3-30. Raven Press, New York.
Janicke, M. A. and LaFountain, J. R. Jr (1984). Malorientation in halfbivalents at anaphase: Analysis of autosomal laggards in untreated, coldtreated, and cold-recovering crane fly spermatocytes. J. Cell Biol. 98, 859-869.
Keith, C. H. and Farmer, M. A. (1993). Microtubule behavior in PC12
neurites: variable results obtained with photobleach technology. Cell Motil.
Cytoskel. 25, 345-357.
Knight, P. and Parsons, N. (1991). Modification of myofibrils by fluorophoreinduced photo-oxidation. J. Muscle Res. Cell Motil. 12, 183-191.
LaFountain, J. R. Jr (1976). Analysis of birefringence and ultrastructure of
spindles in primary spermatocytes of Nephrotoma suturalis during anaphase.
J. Ultrastruc. Res. 54, 333-346.
Leslie, R. J., Saxton, W. M., Mitchison, T. J., Neighbors, B., Salmon, E. D.
and McIntosh, J. R. (1984). Assembly properties of fluorescein-labelled
tubulin in vitro before and after fluorescence bleaching. J. Cell Biol. 99,
2146-2156.
L’Hernault, S. W. and Rosenbaum, J. L. (1983). Chlamydomonas α-tubulin
is posttranslationally modified in the flagella during flagellar assembly. J.
Cell Biol. 97, 258-263.
Ludueña, R. F. (1993). Are tubulin isotypes functionally significant. Mol. Biol.
Cell 4, 445-457.
Lu, Q. and Ludueña, R. F. (1993). Removal of βIII Isotype enhances taxol
induced microtubule assembly. Cell Struct. Funct. 18, 173-182.
Lu, Q. and Ludueña, R. F. (1994). In vitro analysis of microtubule assembly
of isotypically pure tubulin dimers. J. Biol. Chem. 269, 2041-2047.
Margolis, R. L., Wilson, L. and Keifer, B. I. (1978). Mitotic mechanism
based on intrinsic microtubule behaviour. Nature 272, 450-452.
Margolis, R. L. and Wilson, L. (1981). Microtubule treadmills - possible
molecular machinery. Nature 293, 705-711.
Martin, L. M. (1966). The Theory of the Microscope. Blackie and Son, Ltd,
Glasgow.
Maruta, H., Greer, K. and Rosenbaum, J. L. (1986). The acetylation of
alpha-tubulin and its relationship to the assembly and disassembly of
microtubules. J. Cell Biol. 103, 571-579.
Marzec, K. (1993). Movement of chromosome arms in spindles of crane fly
spermatocytes after ultraviolet microbeam irradiation. M.Sc. thesis, York
University, North York, Ontario.
Anaphase microtubule depolymerization 3027
Matthews, K. A., Rees, D. and Kaufman, T. C. (1993). A functionally
specialized α-tubulin is required for oocyte meiosis and cleavage mitoses in
Drosophila. Development 117, 977-991.
Mitchison, T., Evans, L., Schulze, E. and Kirschner, M. (1986). Sites of
microtubule assembly and disassembly in the mitotic spindle. Cell 45, 515527.
Mitchison, T. (1989). Polewards microtubule flux in the mitotic spindle:
evidence from photoactivation of fluorescence. J. Cell Biol. 109, 637-652.
Mitchison, T. J. and Sawin, K. E. (1990). Tubulin flux in the mitotic spindle:
where does it come from, where is it going? Cell Motil. Cytoskel. 16, 93-98.
Mitchison, T. J. and Salmon, E. D. (1992). Poleward kinetochore fiber
movement and anaphase-A in newt lung cell mitosis. J. Cell Biol. 119, 569582.
Oka, M. T., Arai, T. and Hamaguchi, Y. (1990). Heterogeneity of
microtubules in dividing sea urchin eggs revealed by immunofluorescence
microscopy: spindle microtubules are composed of tubulin isotypes different
from those of astral microtubules. Cell Motil. Cytoskel. 16, 239-250.
Osborn, M. and Weber, K. (1982). Immunofluorescence and
immunocytochemical procedures with affinity purified antibodies: tubulin
containing structures. Meth. Cell Biol. 24, 97-132.
Pickett-Heaps, J. D., Tippit, D. H. and Porter, K. R. (1982). Rethinking
mitosis. Cell 29, 729-744.
Piperno, G. and Fuller, M. T. (1985). Monoclonal antibodies specific for an
acetylated form of α-tubulin recognize the antigen in cilia and flagella from a
variety of organisms. J. Cell Biol. 101, 2085-2094.
Piperno, G., LeDizet, M and Chang, X.-J. (1987). Microtubules containing
acetylated α-tubulin in mammalian cells in culture. J. Cell Biol. 104, 289302.
Platt, J. L. and Michael, A. F. (1983). Retardation of fading and enhancement
of intensity of immunofluorescence by p-phenylenediamine. J. Histochem.
Cytochem. 31, 840-842.
Rodriguez, J. and Deinhardt, J. (1960). Preparation of a semi-permanent
mounting medium for fluorescent antibody studies. Virology 12, 316-317.
Sasse, R., Glyn, M. C. P., Girkett, C. R. and Gull, K. (1987). Acetylated αtubulin in Physarum: immunological characterization of the isotype and its
usage in particular microtubular organelles. J. Cell Biol. 104, 41-49.
Sawin, K. E. and Mitchison, T. J. (1991). Poleward microtubule flux in
mitotic spindles assembled in vitro. J. Cell Biol. 112, 941-954.
Scarcello, L. A., Janicke, M. A. and LaFountain, J. R. Jr (1986).
Kinetochore microtubules in crane fly spermatocytes: untreated, 2°C-treated,
and 6°C-grown spindles. Cell Motil. Cytoskel. 6, 428-438.
Schaap, C. J. and Forer, A. (1979). Temperature effects on anaphase
chromosome movement in the spermatocytes of two species of crane flies
(Nephrotoma suturalis Loew and Nephrotoma ferruginea Fabricius). J. Cell
Sci. 39, 29-52.
Shelden, E. and Wadsworth, P. (1992). Microinjection of biotin-tubulin into
anaphase cells induces transient elongation of kinetochore microtubules and
reversal of chromosome-to-pole motion. J. Cell Biol. 116, 1409-1420.
Storz, H. and Jelke, E. (1984). Photomicrography of weakly fluorescent
objects - employment of p-phenylene diamine as a blocker of fading. Acta
Histochem. 75, 133-139.
Vigers, G. P. A., Coue, M. and McIntosh, J. R. (1988). Fluorescent
microtubules break up under illumination. J. Cell Biol. 107, 1101-1024.
Wadsworth, P., Shelden, E., Rupp, G. and Rieder, C. L. (1989). Biotin
tubulin incorporates into kinetochore fiber microtubules during early but not
late anaphase. J. Cell. Biol. 109, 2257-2266.
Wilson, P. J. and Forer, A. (1988). Ultraviolet microbeam irradiation of
chromosomal spindle fibres shears microtubules and permits study of the
new free ends in vivo. J. Cell Sci. 91, 455-468.
Wilson, P. J. and Forer, A. (1989a). Acetylated α-tubulin in spermatogenic
cells of the crane fly Nephrotoma suturalis: kinetochore microtubules are
selectively acetylated. Cell Motil. Cytoskel. 14, 237-250.
Wilson, P. and Forer, A. (1989b). Identifying the site of microtubule
polymerization during regrowth of UV-sheared kinetochore fibres using
antibodies against acetylated alpha-tubulin. Cell Biol. Int. Rep. 13, 823832.
Wilson, P. J. and Forer, A. (1993). Where do kinetochore microtubules
[kMTs] depolymerize during anaphase? Mol. Biol. Cell 4, 245a.
Wise, D., Cassimeris, L., Rieder, C. L., Wadsworth, P. and Salmon, E. D.
(1991). Chromosome fiber dynamics and congression oscillations in
metaphase PtK2 cells at 23°C. Cell Motil. Cytoskel. 18, 131-142.
Zhai, Y., Kronebusch, P. J. and Borisy, G. G. (1993). Microtubule transport
in mitotic cells. Mol. Biol. Cell 4, 51a.
(Received 28 April 1994 - Accepted 5 July 1994)