J. Cell Sci. 23, 43-55 (1977)
Printed in Great Britain
43
MITOTIC SPINDLES OF DROSOPHILA
MELANOGASTER: A PHASE-CONTRAST AND
SCANNING ELECTRON-MICROSCOPE STUDY
AMY MILSTED* AND WILLIAM D. COHEN
Department of Biological Sciences, Hunter College of The City University of Neiv
York, New York, New York 10021, U.S.A.
AND NINA LAMPEN
Memorial Sloan-Kettering Cancer Center, 410 East 6Sth Street, New York,
Netv York 10021, U.S.A.
SUMMARY
Mitotic spindles have been isolated from the blastema stage of Drosophila melanogaster
embryos using modified tubulin-polymerizing medium. 'Clean' spindles, relatively free of
contaminating cytoplasmic material, are obtained. Under phase contrast, mitotic stages appear
remarkably similar to those seen in situ, as reported in early literature. This preservation of
morphological integrity, coupled with relative structural simplicity due to low chromosome
number (zn = 8), makes these spindles ideal subjects for study. Use of the scanning electron
microscope provides excellent visulization of their general structural organization, changes in
whole spindle structure during the course of mitosis, and higher resolution viewing of surface
detail than is permitted with light microscopy.
INTRODUCTION
The structure of isolated mitotic and meiotic spindles has been examined previously
by a variety of techniques, including thin sectioning (Kane, 1962; Goldman &
Rebhun, 1969 ; Cohen & Gottlieb, 1971) and negative staining (Kiefer, Sakai, Solari &
Mazia, 1966) for transmission electron microscopy (TEM), high voltage TEM
(Mclntosh, Cande, Snyder & Vanderslice, 1975), phase-contrast, polarization and
interference light microscopy (Kane & Forer, 1965 ; Rebhun & Sander, 1967 ; Cohen,
1968 ; Forer & Goldman, 1972 ; Inoue, Borisy & Kiehart, 1974).
We describe here a study of isolated mitotic spindles of Drosophila melanogaster
utilizing the scanning electron microscope (SEM). With the great depth of field and
3-dimensional imaging of objects which it can give, the SEM is superior to other types
of microscope for visualization of general structural organization. In usual applications, with specimens such as whole cells or pieces of tissue, internal structure cannot
be seen. Recently, however, it has become apparent that the SEM is applicable to the
study of organelles, with greatest value in the case of structures too large or thick for
TEM whole mounts (Kersey & Wessells, 1976 ; Kirschner, Rusli & Martin, 1975).
The isolated mitotic spindle falls in this category; with appropriate preparation
• Present address: Carnegie-Mellon University, Department of Biological Sciences,
Pittsburgh, Pennsylvania, 15213, U.S.A.
44
A. Milsted, W. D. Cohen and N. Lampen
methods there will be no surrounding membrane or other material obscuring the view
of its fibrous structure.
Drosophila melanogaster mitotic spindles isolated at the blastema stages appear under
phase contrast to have relatively simple structure (Milsted & Cohen, 1973). With their
low chromosome number (zn = 8) and obvious fibres they seemed a promising
subject for study with the SEM. In the course of this work we found that clean
spindles, free of most adhering cytoplasmic material, could be prepared by using a
modified tubulin polymerization medium (Rebhun, Rosenbaum, Lefebvre & Smith,
1974). In such specimens spindle and astral fibres, chromosomes, and a complex
structure in the polar region can be seen. Phase-contrast light microscopy and scanning
electron microscopy together allow characterization of major changes in spindle
organization during the mitotic cycle in these embryos.
MATERIALS AND METHODS
Wild-type Drosophila melanogaster were obtained from the Carolina Biological Supply Co.
(Burlington, North Carolina, U.S.A.) and were cultured in 'Instant Drosophila Medium'
(Carolina). Embryos were collected and dechorionated as described previously (Milsted &
Cohen, 1973) and allowed to develop to the blastema stage during which hundreds of nuclei
undergo synchronous syncytial mitosis (Sonnenblick, 1950). Isolated spindles can be prepared
at this stage using hexylene glycol medium (Milsted & Cohen, 1973). However, better preparations for microscopy and for physiological experiments can be obtained using modified microtubule polymerization medium (TPM) based on that of Rebhun et al. (1974). The T P M
contained o-i M PIPES (piperazine-iV-iV'-bis-[2-ethane sulphonic acid]), 1 mM MgCl,, 5 mM
EGTA (ethyleneglycol-bis-[-aminoethyl ether] JV,iV'-tetraacetic acid), 10 mM TAME (p-tosyl
arginine methylester HC1), 1 mM G T P (guanosine s'-triphosphate) when used, and 0 4 %
Triton X-100, brought to pH 6-8 with KOH.
Preparations intended only for phase-contrast microscopy were made as follows : one or more
dechorionated embryos of the appropriate stage were transferred from the culture dish to a
slide, in a drop of water. The water was totally removed (embryos do not dehydrate easily) and
replaced by a small drop of T P M containing G T P . A clean No. 1 coverslip was added, and its
weight was generally sufficient to rupture the vitelline membrane of the embryo, allowing the
spindles to spill out and spread between coverslip and slide. In some cases, lysis was induced by
gentle tapping of the coverslip. The term 'isolated' is thus applied to the spindles in such
preparations in its broad sense, indicating physical separation of spindles from cytoplasm and
substitution of an experimental medium for the in vivo environment.
Observations were made using a Zeiss phase-contrast microscope (oil immersion lens N.A.
1 3) and spindles were photographed on Kodak Plus-X 35-mm film. Glass coverslips (as opposed
to plastic ones used for SEM) permitted the clearest viewing and photography.
For routine preparations, the embryos were cultured at room temperature and lysed in T P M
at room temperature (21-22 °C). Additional experiments were performed in which the dechorionated embryos were cooled prior to lysis, in attempts to obtain spindle polar regions in
which the microtubule organizing centre ( ' M T O C : Pickett-Heaps, 1969) might be seen more
easily. Some success was achieved by chilling for approximately 25 min at 6 °C, followed by
lysis in T P M without G T P at room temperature.
For SEM, plastic coverslips were employed (A. H. Thomas Co., Philadelphia, Pa., U.S.A.).
These were pre-cleaned for a minimum of 1 h in dilute 'Micro' cleaning solution (Int. Prod.
Corp., Trenton, N.J.), rinsed with water, and incubated in a solution of polylysine in water at
1 mg/ml (Mazia, Schatten & Sale, 1975) for up to 48 h, under refrigeration to prevent bacterial
contamination. Subsequently the coverslips were rinsed thoroughly in water and allowed to dry.
Glass slides were silicone coated (Siliclad, Clay Adams, Inc., N.Y., N.Y.) in order to favour
adhesion of the spindles to the polylysine-coated coverslips. Embryos in the 9th to 12th divisions
post-fertilization were lysed in T P M without GTP, and the spindles perfused with additional
Drosophila spindles in phase contrast and SEM
45
T P M to remove much of the debris. Within 5 min of lysis, T P M containing 2 0 or 2 5 % glutaraldehyde at pH 6-8 was perfused under the coverslip, and the preparation, still on the slide,
was placed in a moist chamber for 1-1-5 h. The chamber consisted of a Petri dish with moist
filter paper in the bottom. Glutaraldehyde was removed by perfusion with several changes of
o-i M PIPES buffer, pH 6-8. Subsequently 1 % OsO 4 ino-i M PIPES, pH 6 8 , was perfused
under the coverslip and allowed to remain there for 30—60 min, with the preparation again held
in the moist chamber. Osmium was then removed by perfusion with several changes of buffer.
Preparations were examined under phase contrast for possible candidate spindles suitable for
further SEM processing. In general, the best specimens were those lying alone, free of other
spindles and adhering debris. That there were relatively few such spindles remaining is probably
due to perfusion, during which spindles occasionally are seen to detach from the coverslips.
The pre-selected coverslip preparations were carefully removed from their slides and placed
in small Petri dishes of 0 1 M PIPES, pH 6-8. They were then dehydrated in an ethanol series
and passed through a graded series of ethanol-Freon 113 mixtures into 100% Freon 113
(CCljF-CCljF). Critical point drying (Anderson, 1951) was carried out in Freon 13 (CC1F3)
(Cohen, Marlow & Garner, 1968).
Critical-point-dried specimens were examined, while dry, under phase contrast. Spindles
were still recognizable, and they were photographed for later reference and their location
marked by circling or scratching the region with a Scheaff micro-object marker (A. H. Thomas
Co., Philadelphia, Pa., U.S.A.) mounted on the microscope. Appropriate areas of the plastic
coverslips were cut out with scissors and mounted on SEM stubs with double-sticky tape.
Plastic coverslips were used for all SEM preparations because they could be marked and
trimmed easily.
Silver conductive paint was applied to the coverslip edges on the stub, and the preparation
was coated with approximately 20 run of gold or gold-palladium in a Denton rotating, tilting
vacuum evaporator. Specimens were examined and photographed in a Cambridge Stereoscan S4 SEM at 20 kV, using zero tilt for easiest recognition of preselected individual spindles
when compared with light micrographs (phase contrast).
RESULTS
Phase-contrast observations
When viewed under phase contrast, blastema stage spindles prepared in TPM are
easy to see and usually free of adhering material. Occasionally they are clumped
together in masses, but in most preparations there are many individual spindles which
can be examined. Spindle fibres, chromosomes, midbodies, and rather large structures
at the expected location of centrioles are obvious. In any one preparation, all of the
spindles are in approximately the same mitotic stage, as expected. By slight alterations
in the timing of embryo lysis, most of the major mitotic stages can be obtained in
different preparations. These are illustrated in Figs. 1 to 5, with mitotic stages
designated according to Huettner (1933).
In Fig. 1, the triangular morphology typical of the Drosophila melanogaster blastema
stage prophase spindle is apparent. Black spots with fibrous material radiating from
them appear at the 2 presumptive polar positions (c and c'). The third corner of the
triangle lacks this feature and is more rounded, so that the entire structure is sac-like.
The chromosomes (ch) are grouped in the midregion and seem to be organized into
2 subgroups, each of which has a distinct focal point (arrows) located midway along
the line between c and c .
The polar structures are about 0-7 /mi in diameter as measured in Fig. 1 and in all
other spindle preparations. In phase contrast they are always seen in prophase
spindles, usually in metaphase, and less frequently in later stages. Since their size is
4
CEL
23
A. Milsted, W. D. Cohen and N. Lampen
,<<*
•-c'
Drosophila spindles in phase contrast and SEM
47
greater than that expected for a single centriole or centriole pair, we have tentatively
assigned to them the term 'centriole complex'.
At metaphase (Fig. 2) the spindle poles are well denned by centriole complexes. A
small number of major fibres, generally three or four, are evident between chromosomes and poles. The fibres seem to converge at a point a short distance from the
centriole complex, where they coalesce into a constricted neck region (x, also visible in
Huettner's (1933) spindles in situ) which apparently connects directly to the centriole
complex. Chromosomes (ch) are seen positioned in the spindle mid-plane.
Figs. 3 and 4 show typical late anaphase spindles. There are a small number of
major interzonal fibres, probably four, each of which is thickened in the midregion
forming an apparent midbody. In Fig. 3 a mass of spindles is seen; even in such
masses structural details such as midbodies (m) and apparent centriole complexes can
often be detected with careful focusing. As this figure shows, synchrony in any one
preparation is not absolutely perfect: while most spindles were in anaphase (a), a few
had entered telophase (f). In the spindle shown in Fig. 4, four midbodies are visible
(m). Here the chromosomes are massed at the poles except for a trailing arm in each
chromosome set (x). The arms are identifiable as such on the basis of careful focusing
through several planes, which shows them to be continuous with the polar chromatin
and considerably thicker than the interzonal fibres. Such trailing arms were seen by
Rabinowitz (1941) in situ, and we have confirmed their presence by thin-sectioning
(Cohen & Milsted, unpublished observations). As seen in phase contrast, these arms
Fig. 1. A group of prophase spindles, in isolation medium, c, c, ' centriole complexes' ;
ch, chromosomes ; arrows indicate chromosome sub-group focal points. Phase contrast,
x 1740.
Fig. 2. A metaphase spindle, in isolation medium ; c, c', centriole complexes, ch,
chromosomes ; x, constricted region at fibre convergence point, apparently connecting
directly to centriole complex. Phase contrast, x 2060.
Fig. 3. Spindles trapped in a mass. Most of them were in late anaphase, such as spindle
(a); one is in telophase, spindle (t); c, presumed centriole complexes ; m, midbody.
Phase contrast, x 2060.
Fig. 4. A late anaphase spindle, in isolation medium ; ch, chromosomes massed near
poles ; if, interzonal fibre ; m, one of 4 adjacent midbodies ; x, trailing chromosome
arm (one in each half-spindle). Phase contrast, x 2060.
Fig. 5. A telophase spindle, typically bent at this stage, in isolation medium, m, single
midbody. Phase contrast, x 940.
Fig. 6. Spindle polar region obtained by lysis of precooled embryo in T P M ; central
centriole complex is present. Phase contrast, x 1740.
Fig. 7. Three early prophase spindles, isolated in T P M and fixed, dehydrated, and
critical-point dried on polylysine-coated coverslip, as described in Materials and
methods ; c, c , apparent focal points for fibrils, although polar structure is not visible ;
X, crossed scratch lines on coverslip ; Y, another scratch line ; 1, 2, 3, particles on
coverslip (scratches and particles used as landmarks for comparison with Fig. 8).
Scanning EM, x 1400.
Fig. 8. The same 3 early prophase spindles shown in Fig. 7, as seen under phase
contrast after critical-point drying, prior to gold coating. Note landmarks corresponding to those in Fig. 7 : X, crossed scratch lines on coverslip ; Y, another scratch line ;
1, 2, 3, particles on coverslip ; c, c\ apparent focal points. Phase contrast, x 940.
4-2
A. Milsted, W. D. Cohen and N. Lampen
Drosophila spindles in phase contrast and SEM
extend toward the same midbody in each half-spindle, as if associated with the same
interzonal fibre.
Even telophase spindles can be isolated (Fig. 3, spindle (t); Fig. 5). This stage is
distinguished from late anaphase by the fact that there is but one relatively thick
interzonal fibre with a single large midbody (m), and by the reconstituting daughter
nuclei at the polar extremities. The entire structure is usually curved or bent.
If embryos are cooled prior to lysis, it is possible to obtain preparations with
incomplete spindles and small, sometimes separate polar regions, in which the centriole
complex is seen to advantage (Fig. 6). A number of 'points' radiate from the black
central structure, the diameter of which is in the range 0-7-0-8 /tm.
Scanning EM observations
In the description which follows, the term 'fibril' is used to indicate the relatively
thin structures seen in the SEM, and the term 'fibre' is applied to obvious bundles of
fibrils.
Early prophase spindles, as seen in the SEM (Fig. 7), all present the appearance of
masses of tangled fibrils. In one of the spindles shown, 2 major focal points for the
fibrils are apparent (c and c'). As seen at higher magnification in Fig. 9, no major
fibres are evident, and chromosomes are not visible at this stage using the SEM. For
comparison, Fig. 8 shows the same 3 spindles as Fig. 7, after critical point drying and
prior to coating, under phase contrast. They are recognizable as such both by their
relative positions and by the particles and scratches which correspond (for example,
particles numbered /, 2, 3, crossed scratch lines labelled X, a scratch labelled Y, in
both Figs. 7 and 8). Since the spindles are viewed dry at this point, not much detail is
seen in phase contrast, but the light micrograph is valuable in locating good specimens
for subsequent scanning.
In late prophase spindles such as that in Fig. 10, relatively thick, elongate smoothFig. 9. Higher magnification view of the prophase spindle in upper right corner of
Fig. 7 ; /, fine fibrils, in approximately parallel array. Chromosomes are not visible at
this stage in the SEM. Scanning EM, x 5830.
Fig. 10. A late prophase spindle, somewhat flattened on coverslip ; ch, chromosomes ;
crossed fibrils at arrow. Scanning EM, x 5200.
Fig. 11. A late anaphase spindle, showing fibres composed of thinner fibrils, af,
astral fibrils ;/, central fibre. Scanning EM, x 4170. Inset: The same spindle, as seen in
phase contrast after critical-point drying, prior to gold coating. Highly refractile areas
correspond to regions of greatest density. Phase contrast, X 785.
Fig. 12. Higher magnification of the lower polar region of the spindle shown in
Fig. 11 : cc, centriole complex; ch, chromosome arms ; /, a thin fibril on coverslip
surface ; p, a point at periphery of centriole complex, to which 3 very fine fibrils are
connected. Scanning EM, x 8330.
Fig. 13. A telophase spindle, with curved structure ; interzonal fibril bundle seems
twisted, as does path of fibril (/) ; n, daughter nuclei at spindle extremities. Scanning
EM, x 5200.
Fig. 14. Higher magnification of one of the daughter nuclei in the telophase spindle
shown in Fig. 13 ; r, ring-like surface features. Scanning EM, x 9830.
49
SO
A . Milsted,
W. D. Cohen and N.
Lampen
surfaced structures are found in the position expected for chromosomes on the basis of
phase-contrast observations of the same preparations. Based upon their thickness range
(0-3-0-4 /<m) and location, we believe these to be the chromosomes and have designated them as such (ch). At this stage the number of individual fibrils seems relatively
small, and they are not yet organized into a recognizable spindle with major fibres
and distinct poles. Apparent multiple focal points for the fibrils give the spindle
end-region a criss-crossed appearance (arrow).
A late anaphase spindle is shown in Fig. 11. Here the poles are well defined by
convergence of fibrils. In the midregion the spindle is composed of a relatively small
number of interzonal fibres, each consisting of smaller fibrils. The central fibre (/) is
thicker than the others, measuring about 0-7 /tm in diameter, with component fibrils
in the range of 0-1-0-2 /tm. Apparent astral fibrils converge to a series of 'points' at
the upper pole (a/), while additional astral material seems skewed in the opposite
direction at the lower pole.
The lower polar region is shown in greater detail in Fig. 12. Chromosome arms
(ch), again identified by surface texture, thickness, and position, can be seen protruding
from the fibrous mass. A roughly polygonal structure (cc) is present at the position
expected for the centriole, adjacent to the fibril convergence point. It appears to
have a central body surrounded by a ring of material which itself has regular substructure, with several very thin fibrils leaving the outer edge of the ring at point p.
The diameter of the entire structure, including surrounding ring, is approximately
0-7 /tm, and its attachment to the rest of the spindle appears rather tenuous. With the
SEM such structures have not been readily identifiable at earlier mitotic stages in
which centriole complexes are most frequently seen in phase contrast. We think it
most likely that they are obscured from view in the SEM by overlying material at
these stages; however, our identification of this polygonal body (Fig. 12, cc) as a
'centriole complex' should be regarded as tentative pending further investigation.
For comparison, the same spindle is shown as it appeared under phase contrast
after critical-point drying, before coating (Fig. 11, inset). The 2 highly refractile areas
(specimen viewed in air) correspond to the 2 polar regions of greatest density in Fig. 11
(SEM).
At telophase (Fig. 13) reconstituting daughter nuclei (n) are seen at the spindleends. The spindle now apparently consists entirely of non-chromosomal fibrils,
massed together into a compact interzonal bundle measuring about 0-9 /tm in width
at its midpoint, and approximately 13-5 /mi in length. The entire structure seems
twisted, as if one end of the spindle had been rotated with respect to the other. An
individual fibril (/) can be seen following a route corresponding to such a twist. On
the surface of the daughter nucleus at the left in Fig. 13, and seen at higher magnification in Fig. 14, there are a number of ring-like features (r) visible which measure
about 0-25-0-3 /tm in diameter. Structures which might correspond to centriole
complexes have not been seen at this stage.
Drosophila spindles in phase contrast and SEM
51
DISCUSSION
Properties of D. melanogaster spindles isolated in TPM
Use of an isolation medium based upon tubulin-polymerizing conditions and
containing protease inhibitor (TAME) and detergent (Triton X-ioo) as described by
Rebhun et al. (1974) makes possible the preparation of D. melanogaster mitotic
spindles suitable for examination with the SEM. Under phase contrast such spindles
are clean; that is, their fibrous structure is readily apparent, as are chromosomes,
midbodies, and centriolar complexes. This lack of contamination with other cytoplasmic material is most likely due not only to the selective stabilization of microtubulecontaining structures, but also to the dispersive action of the detergent on membranous
components.
The morphology of these spindles after isolation is remarkably faithful to that
reported for blastema-stage spindles in situ (paraffin sections) in the older literature
(Huettner, 1933 ; Rabinowitz, 1941). This includes details such as midbodies and
trailing chromosome arms in anaphase. The latter were identified by Rabinowitz
(1941) as the long arm of the X chromosome (Fig. 4). Bending of the telophase
spindle as seen in isolates can also be found in the older in situ micrographs. One
difference is noted however: polar structures are lacking in our telophase preparations, as judged by phase contrast. According to Huettner (1933), a pair of centrioles is
present at each pole in situ at this stage, a relatively long distance away from the
reconstituting nuclei. This suggests that they are tenuously connected to the telophase
spindle, and probably fall off during isolation.
The morphology of spindles at various stages seemed the same under phase contrast whether GTP was or was not included in the isolation medium. However, GTP
did seem to stimulate formation of background fibrils in the preparations. The latter
were especially prominent in material from non-mitotic embryos containing interphase nuclei, and presumably reflected the availability of a polymerizable tubulin pool
during interphase.
In order to study spindle structure with the SEM, spindles must be isolated free
of debris, and must retain the in situ morphology which characterizes different mitotic
stages. These criteria are met by the TPM preparations.
Phase-contrast versus scanning electron microscopy
In general, there is good agreement between phase-contrast and SEM observations
in terms of spindle size and shape, indicating that spindle morphology is little altered
during the SEM preparative steps. In phase contrast, of course, the material is
effectively transparent so that structures such as chromosomes and centriole complexes
are readily visible. Comparison of Figs. 1 and 7 shows that this is not the case for the
SEM, as expected; in prophase spindles the chromosomes are hidden beneath the
fibrous surface. At later mitotic stages there appear rather smooth-surfaced oblong
structures, thicker than surrounding fibrils, which we believe to be chromosomes. In
the late anaphase spindle (Fig. 11) in which some of these are visible, it is likely that
others are concealed within the fibrous polar regions.
52
A. Milsted, W. D. Cohen and N. Lampen
The major fibres seen in metaphase and anaphase spindles under phase contrast are
found to be bundles of smaller fibrils in the SEM. Many of these fibrils are in
the thickness range o-i-o-2/<m, which could accommodate a maximum of about
6—25 microtubules. (This estimate excludes the thickness of the metal coating,
and ignores possible clear zones or inter-microtubule material.) In some instances,
fibrils observed in the SEM are thin enough to be accounted for by single metalcoated microtubules (for example, fibril / and the fibrils radiating from point p in
Fig. 12). The apparently single interzonal fibre observed in telophase spindles
under phase contrast (Fig. 5) is similarly found to be constructed of smaller fibrils
(Fig- 13)The structures seen at spindle poles under phase contrast and referred to as
centriole complexes have a maximum dimension of approximately 0-7 /tm, and the
central black body of polar regions isolated from cooled embryos is about the same
size. This value seems too great to be accounted for by a simple centriole or centriole
pair, even if measurement error is considered. Mahowald (1963) reported that the
centrioles of Drosophila melanogaster in the blastula stages were o-i6/tm in diameter
and 0-15-0-175 /tm long, and were never found in fully formed pairs. Fullilove &
Jacobson (1971) and Fritz-Niggli & Suda (1972) show centrioles surrounded by
satellites in the forming blastoderm of Drosophila montana and in spermatocytes of
Drosophila melanogaster, respectively. The diameter of these complexes is in the
range 0-6—0-8 /.cm, which correlates well with the size of the structures seen at poles
under phase contrast.
In the polar position of some spindles the SEM reveals a structure of roughly
polygonal shape, about 0-7 /un in diameter, with a central body approximately 0-2 /tm
in diameter (Fig. 12). This object could be a complex of centriole plus accessory
material, possibly satellites, which would serve as the polar microtubule organizing
centre for these spindles (Pickett-Heaps, 1969). While satellites have rarely been
reported surrounding mitotic centrioles (de Harven, 1968), their presence has been
shown around meiotic centrioles of D. melanogaster (Fritz-Niggli & Suda, 1972) and
jellyfish (Szollosi, 1964).
One structure seen easily in phase contrast but not yet identified in the SEM is the
midbody. While its absence could indicate loss of some midbody material during
SEM preparative steps, a different explanation may hold : the midbody, as seen in
phase contrast, could simply be a region of increased density due to overlap of interzonal microtubules of opposite polarity and/or inter-microtubule material (Paweletz,
1967 ; Mclntosh et al. 1975). Such differences in internal density would not be
expected to appear under the SEM.
An interesting feature of the telophase spindles is the substructure at the surface
of reconstituting nuclei seen in the SEM (Fig. 14). There appear to be coils or rings
of material, about 0-25-0-3 /tm in diameter. In a study of isolated mouse liver nuclei
with the SEM, Kirschner et al. (1975) found that nuclear pore complexes are resistant
to treatment with Triton X-100. However, the features described here are several
times larger than typical pore complexes. It is possible that they are exposed chroma-
Drosophila spindles in phase contrast and SEM
53
tin coils, but further work is obviously needed, using a higher resolution SEM on
isolated telophase and interphase nuclei.
Events of the mitotic cycle in D. melanogaster embryos
The SEM reveals the progressively more ordered fibrous structure of spindles at
different stages of the mitotic cycle. The relatively thin and somewhat tangled fibrils
of prophase assume a more linear arrangement in late prophase and are organized into
bundles at anaphase. By telophase, one thick interzonal bundle remains. In metaphase
and anaphase, as seen in phase contrast, there is a fairly small number of major fibres
visible, perhaps 3 or 4. Four midbodies, but not more, have been seen in a number of
anaphase spindles, again suggesting that there may be 4 major fibres involved. Since
Drosophila and other Diptera exhibit the phenomenon of somatic pairing of homologues during the mitotic cycle (Metz, 1916), there will be 4 pairs of chromosomes per
half-spindle in D. melanogaster (zn = 8). The number of major fibres may therefore be
significant, suggesting a relationship between each chromosome pair and a major nonchromosomal fibre. It is noteworthy that the trailing chromosome arms of anaphase
spindles often seem associated with one of the interzonal fibres (for example, Fig. 4);
this could be fortuitous, however.
In the transition from anaphase to telophase separate interzonal fibres disappear,
giving the impression that they associate laterally to form the telophase interzonal
fibril bundle. The one large midbody at telophase would then be interpreted as the
fusion product of separate smaller anaphase midbodies. The telophase fibril bundle
seen in the SEM is about 0-9 /un thick, and this is sufficient to accommodate all of the
anaphase interzonal fibrils if they were to associate laterally. Since there is no cytokinesis at the blastema stage of development, the presence of midbodies indicates
their direct involvement in mitosis, possibly as regions of overlap between nonchromosomal microtubules of opposite polarity which have been implicated in a
sliding spindle elongation mechanism (Pickett-Heaps, McDonald & Tippitt, 1975).
The observed bending of telophase spindles might be significant in relation to
spindle mechanics. The SEM reveals an apparent twist of the interzonal fibre bundle,
suggesting that the 2 half-spindles rotate with respect to each other.
Use of scanning electron microscopy in analysis of spindle structure
The SEM improves upon both resolution and depth of field obtainable with light
microscopy, providing a 3-dimensional view of the entire spindle. Comparable spindle
reconstructions from serial thin sections and TEM are difficult to achieve and require
excessive amounts of labour by comparison. For examination of general structural
organization of isolates, the SEM might serve as an alternative to high voltage transmission electron microscopy as well (Mclntosh et al. 1975).
With respect to small sample preparation methods for the SEM, use of polylysinecoated coverslips (Mazia et al. 1975) permits the following of individual spindles
under phase contrast through all of the processing steps. They can be photographed
after critical point drying for subsequent identification in the SEM using spindle
orientation and adjacent landmarks. Polylysine-coated plastic coverslips have proved
54
A. Milsted, W. D. Cohen and N. Lampen
more convenient for our purposes than either polylysine- or Formvar-coated glass
(Lung, 1974). While adequate for the work thus far, significant losses of material have
been encountered ; perhaps better retention of material might be achieved by covalent
linking to the substratum (Vial & Porter, 1975).
To our knowledge, this is the first study of spindle structure in which the SEM is
utilized. Thus we have not had the advantage of comparison with the data of others.
However, based on our results, we believe that the SEM will become an increasingly
powerful tool for such work as improvements are made in the areas of instrument
resolving power, specimen retention, and reduction or elimination of the metal
specimen coating.
We wish to express our appreciation to Dr Etienne de Harven for use of the Cambridge
Scanning Electron Microscope and preparative facilities, for his interest, and for critical reading
of the manuscript. This work was supported in part by Faculty Research Award no. 1220 of the
Research Foundation of The City University of New York, and by grant no. GB-25578 of the
National Science Foundation to Dr W. D. Cohen. Submitted in partial fulfilment of the
requirements for the Doctor of Philosophy degree at The City University of New York (A. M.).
REFERENCES
ANDERSON, T. F. (195I). Techniques for the preservation of three-dimensional structure in
preparing specimens for the electron microscope. Trans. N.Y. Acad. Sci. 13, 130—135.
COHEN, A. L., MARLOW, D. P. & GARNER, G. E. (1968). A rapid critical point method using
fluorocarbons ('Freons') as intermediate and transitional fluids. J. Microscopie 7, 331-342.
COHEN, W. D. (1968). Polyelectrolyte properties of the isolated mitotic apparatus. ExplCell
Res. 51, 221-236.
COHEN, W. D. & GOTTLIEB, T. (1971). C-microtubules in isolated mitotic spindles. J . Cell Sci.
9, 603-619.
DE HARVEN, E. (1968). The centriole and the mitotic spindle. In Infrastructure of Biological
Systems, vol. 3 (ed. A. J. Dalton & F. Hoguenau), pp. 197-227. New York : Academic Press.
FORER, A. & GOLDMAN, R. D. (1972). The concentration of dry matter in mitotic apparatuses in
vivo and after isolation from sea-urchin zygotes. J. Cell Sci. 10, 387-418.
FRITZ-NIGGLI, H. & SUDA, T. (1972). Bildung und Bedeutung der Zentriolen : Eine Studie und
Neuinterpretation der Meiose von Drosophila. Cytobiologie 5, 12-41.
FULLILOVE, S. L. & JACOBSON, A. G. (1971). Nuclear elongation and cytokinesis in Drosophila
montana. Devi Biol. 26, 560-577.
GOLDMAN, R. D. & REBHUN, L. I. (1969). The structure and some properties of the isolated
mitotic apparatus. J. Cell Sci. 4, 179-209.
HUETTNER, A. F. (1933). Continuity of the centrioles in Drosophila melanogaster. Z. Zellforsch.
mikrosk. Anat. 19, 119-134.
INOUE, S., BORISY, G. G. & KIEHART, D . P. (1974). Growth and lability of Chaetoptertis oocyte
mitotic spindles isolated in the presence of porcine brain tubulin. J. Cell Biol. 62, 175-184.
KANE, R. E. (1962). The mitotic apparatus : fine structure of the isolated unit. J. Cell Biol. 15,
279-287.
KANE, R. E. & FORER, A. (1965). The mitotic apparatus. Structural changes after isolation.
J. Cell Biol. 25 (No. 3, Part 2), 31-39KERSEY, Y. M. & WESSELLS, N. K. (1976). Localization of actin filaments in internodal cells of
Characean algae. A scanning and transmission electron microscope study. J. Cell Biol. 68,
264-275.
KIEFER, B., SAKAI, H., SOLARI, A. J. & MAZIA, D. (1966). The molecular unit of the micro-
tubules of the mitotic apparatus. J. molec. Biol. 20, 75-79.
KIRSCHNER, R. H., RUSLI, M. & MARTIN, T. E. (1975). Scanning electron microscopy of
isolated mouse liver nuclei. J. Cell Biol. 67, 213a.
Drosophila spindles in phase contrast and SEM
55
LUNG, B. (1974). The preparation of small, particulate specimens by critical point drying:
application for scanning electron microscopy. J. Microscopy 101, 77-80.
MAHOWALD, A. P. (1963). Ultrastructural differentiations during formation of the blastoderm
in the Drosophila melanogaster embryo. Devi Biol. 8, 186-204.
MAZIA, D., SCHATTEN, G. & SALE, W. (1975). Adhesion of cells to surfaces coated with polylysine. Applications to electron microscopy. J. Cell Biol. 66, 198-200.
MCINTOSH, J. R., CANDE, Z., SNYDER, J. & VANDERSLICE, K. (1975). Studies on the mechanism
of mitosis. Ann. N. Y. Acad. Set. 253, 407-427.
METZ, C. W. (1916). Chromosome studies in the Diptera. II. The paired association of chromosomes in the Diptera, and its significance. J. exp. Zool. 21, 213-279.
MILSTED, A. & COHEN, W. D. (1973). Mitotic spindles from Drosophila melanogaster embryos.
Expl Cell Res. 78, 243-246.
PAWELETZ, N. (1967). Zur Funktion des ' Fleming-korpus' bei der Teilung tierisches Zellen.
Naturwissenschaften 54, 533-535PICKETT-HKAPS, J. D. (1969). The evolution of the mitotic apparatus : an attempt at comparative ultrastructural cytology in dividing plant cells. Cytobios 1, 257-280.
PICKETT-HEAPS, J. D., MCDONALD, K. L. & TRIPPITT, D. H. (1975). Spindle structure and
function in diatoms. J. Cell Biol. 67, 336a.
RABINOWITZ, M. (1941). Studies on the cytology and early embryology of the egg of Drosopliila
melanogaster. J. Morph. 69, 1-49.
REBHUN, L. I., ROSENBAUM, J., LEFEBVRE, P. & SMITH, G. (1974). Reversible restoration of the
birefringence of cold-treated isolated mitotic apparatus of surf clam eggs with chick brain
tubulin. Nature, Lond. 249, 113-115.
REBHUN, L. I. & SANDER, G. (1967). Ultrastructure and birefringence of the isolated mitotic apparatus of marine eggs. J. Cell Biol. 34, 859-883.
SONNENBLICK, B. P. (1950). The early embryology of Drosopliila melanogaster. In Biology of
Drosophila (ed. M. Demerec), pp. 62-167. New York and London: Hafner.
SZOLLOSI, D. (1964). The structure and function of centrioles and their satellites in the jellyfish
PhiaUdium gregarium. J. Cell Biol. 21, 465-479.
VIAL, J. & PORTER, K. R. (1975). Scanning electron microscopy of dissociated tissue cells. J.
Cell Biol. 67, 345-360.
{Received 27 January—Revised 3 June 1976)