INTRODUCTION
The generation of order and asymmetry at the cytoplasmic
level is fundamental to many basic cellular functions
including mitosis, directed cell movement, and developmental
processes such as axis specification. All of these
events are largely dependent on the cytoskeleton, which is
responsible for organizing the cytoplasm of eukaryotic cells.
The problem of cytoplasmic organization is particularly
acute during early embryogenesis in Drosophila: the initial
divisions are syncytial and many of the nuclei undergo a
precise program of movements leading to the formation of
a highly regular blastoderm embryo.
The initial nuclear divisions of early Drosophila embryogenesis
have been thoroughly studied (Rabinowitz, 1941;
Sonnenblick, 1950; Turner and Mahowald, 1977; Zalokar
and Erk, 1976; Foe and Alberts, 1983; Stafstrom and
Staehelin, 1984; Minden et al., 1989). The Drosophila
embryo undergoes 13 rapid (8-21 minutes) synchronous
nuclear divisions without accompanying cytokinesis. The
first eight nuclear divisions occur in the interior of the
embryo. After these divisions, the nuclei embark along one
of three different division programs. The majority of the
nuclei migrate to the cortex during nuclear cycles 8, 9, and
10. By the end of nuclear cycle 10, these nuclei are positioned
at the surface of the embryo just beneath the plasma
membrane, where they undergo four synchronous divisions
as a regularly spaced cortical monolayer. Cellularization
occurs during interphase of nuclear cycle 14. A second
nuclear division pathway involves a few nuclei that migrate
ahead of the main body of nuclei and reach the posterior
cortex during nuclear cycle 9. During cycle 10, these nuclei
are surrounded by plasma membranes to form the pole cells,
the precursors of the germ line. A third nuclear division
program is exhibited by a group of nuclei that fail to migrate
outward toward the cortex during nuclear cycle 8. These
nuclei, known as the yolk nuclei, remain in the interior of
the embryo and become polyploid.
The cortical cytoskeleton of the Drosophila embryo has
been characterized through immunofluorescence analysis of
fixed material (Warn et al., 1984; Warn et al., 1985; Karr
and Alberts, 1986; Warn and Warn, 1986; Warn et al.,
1987). In addition, the dynamics of the cortical cytoskeleton
has been followed in living embryos by microinjecting
fluorescently labeled actin and tubulin (Kellogg et al., 1988).
The injection of fluorescently labeled histones provides a
means of following chromosome behavior (Minden et al.,
1989). During interphase, the cortical actin becomes
organized in a cap between the plasma membrane and the
nucleus. At this stage, microtubule networks, which
originate from a centrosome pair located between the
nucleus and the actin cap, radiate inwards to encompass each
1245Development 118, 1245-1254 (1993)
Printed in Great Britain © The Company of Biologists Limited 1993
Cytoplasmic organization, nuclear migration, and
nuclear division in the early syncytial Drosophila embryo
are all modulated by the cytoskeleton. In an attempt to
identify genes involved in cytoskeletal functions, we have
examined a collection of maternal-effect lethal mutations
induced by single P-element transposition for those that
cause defects in nuclear movement, organization, or
morphology during the syncytial embryonic divisions.
We describe three mutations, grapes, scrambled, and
nuclear-fallout, which define three previously uncharacterized
genes. Females homozygous for these mutations
produce embryos that exhibit extensive mitotic division
errors only after the nuclei migrate to the surface.
Analysis of the microfilament and microtubule organization
in embryos derived from these newly identified
mutations reveal disruptions in the cortical cytoskeleton.
Each of the three mutations disrupts the actin-based
pseudocleavage furrows and the cellularization furrows
in a distinct fashion. In addition to identifying new genes
involved in cytoskeletal organization, these mutations
provide insights into cytoskeletal function during early
Drosophila embryogenesis.
Key words: grapes, scrambled, nuclear-fallout, Drosophila,
maternal-effect lethal mutation, microtubule organization, actin
SUMMARY
Mutations affecting the cytoskeletal organization of syncytial Drosophila
embryos
William Sullivan1,2,*, Patrick Fogarty1 and William Theurkauf2,†
1Department of Biology, Sinsheimer Laboratories, University of California, Santa Cruz, CA 95064, USA
2Department of Biochemistry and Biophysics, School of Medicine, University of California, San Francisco, CA 94143, USA
*Author for correspondence at address1
†Present address: Department of Biochemistry and Cell Biology, State University of New York at Stony Brook, Stony Brook, NY 11794-5215, USA
1246
nucleus. As the nuclei progress into mitosis, the centrosomes
separate further and their microtubules form a bipolar
mitotic spindle parallel to the cortex. In addition, the actin
is redistributed to become associated with transient invaginations
of the plasma membrane, called pseudocleavage
furrows. These furrows surround each of the tightly packed
mitotic spindles (Warn et al., 1984; Karr and Alberts, 1986).
Recently, cytological, biochemical and genetic approaches
have identified a number of proteins responsible for organizing
the cortical cytoskeleton of the Drosophila embryo
(for review see Fryberg and Goldstein, 1990; Schweisguth
et al., 1991).
To identify genes involved in the formation and dynamics
of embryonic cytoskeletal structures, we have screened for
mutations that specifically disrupt the embryonic syncytial
divisions. Our strategy was based on inhibitor studies that
demonstrate that the organized divisions, migrations, and
spacing of the nuclei in the syncytial embryo are highly
dependent on maternally supplied cytoskeletal components
(Zalokar and Erk, 1976; Foe and Alberts, 1983; Warn et al.,
1987; Planques et al., 1991; Hatanaka and Okada, 1991). For
instance, disrupting the cortical actin filaments with
cytochalasin, results in nuclear fusions specifically during
the cortical nuclear divisions (Zalokar and Erk, 1976;
Theurkauf, unpublished observations). Irregularities in the
organization and behavior of the early embryonic nuclei,
therefore, are likely to serve as sensitive indicators of defects
in these cytoskeletal components. These cytoskeletal
inhibitor studies motivated a search for maternal-effect
mutations with similar effects.
A number of genetic screens has generated a large collection
of maternal-effect mutations (Bakken, 1973; Rice
and Garen, 1975; Gans et al., 1975; Zalokar et al., 1975;
Mohler, 1977; Schupbach and Weischaus, 1986; Perrimon
et al., 1986; Erdelyi and Szabad, 1989; Schupbach and
Weischaus, 1989; Perrimon et al., 1989; Szabad et al.,
1989). With only a few exceptions (Zalokar et al., 1975), the
nuclear phenotype of syncytial embryos derived from
females homozgous for these maternal-effect mutations has
been described. We have examined the syncytial nuclear
phenotype in 44 recently isolated P-element induced
maternal-effect lethal mutations. Our studies have identified
a set of mutations that result in the production of embryos
that exhibit abnormal nuclear behavior during the cortical
divisions. However, prior to their migration to the cortex,
the nuclei behave normally in these embryos. During the
cortical divisions, the mutant embryos also display specific
disruptions of their cortical microtubules and actin
networks. We suggest that at least some of the members of
this group of maternal-effect mutations disrupt genes that
are specifically involved in the organization of the cortical
cytoskeleton.
MATERIALS AND METHODS
Stocks
The laboratories of Y. N. Jan (U.C. San Francisco)and A. Spradling
(Carnegie Institution of Washington) generously provided us with
the female-sterile mutations described in this screen (Bier et al.,
1989; Karpen and Spradling, submitted). Oregon-R was used as the
wild-type stock. Other mutations and chromosomes are described
by Lindsley and Zimm (1992). The daughterless-abo-like (dal)
mutation has been previously described (Sandler, 1977; Sullivan,
1987; Sullivan et al., 1990). The stocks were maintained on a
standard corn meal/molasses medium at 25°C.
Genetic analysis
Complementation analysis was performed by obtaining all combinations
of transheterozygote females bearing the maternal-effect
mutations of interest. 10 of these transheterozygote females were
mated to 10 wild-type males for 5 days. After 5 days, the adult flies
were removed. The number of adults eclosing prior to day
seventeen was recorded for each of the matings.
Deficiency analysis was performed by crossing heterozygous
mutant females to heterozygous deficiency males. The hatch rate
of embryos derived from females bearing a copy of the mutation
over a deficiency was determined. A set of deficiencies that
resulted in no hatching when in combination with one of the
maternal-effect mutations enabled us to map the mutations. The
following deficiencies were used to map grapes: Df(2L)osp29,
Df(2L)r10, Df(2L)H20, Df(2L)TW137, Df(2L)M-HS5, and
Df(2L)TW50. The deficiencies Df(3L)vin5, Df(3L)vin7,
Df(3L)Ly, Df(3L)fzGf3b, Df(3L)fzM21, and Df(3L)fzD21 were
used to map nuclear-fallout. The deficiencies Df(2R)st1,
Df(2R)PK78K, and Df(2R)42 were used to map scrambled.
Revertant analysis of our P-element induced maternal-effect
mutations was performed according to the procedure described by
Cooley et al. (1988).
Mapping of the P-element by in situ polytene hybridization was
performed as described by Engels et al. (1986), using a probe from
the P-lac Z vector (Karpen and Spradling, 1992).
Cellular analysis
The embryos used in our analysis were derived from either normal
females, females homozygous for the maternal-effect mutation, or
females bearing the maternal-effect mutation and a deficiency for
the mutation. The embryos were fixed through a modification of
the Mitchison and Sedat (1983) protocol (Theurkauf, 1992). Fixed
embryos were double stained for actin and microtubules, using fluorescein-conjugated
phalloidin and rhodamine-conjugated mouse
anti-tubulin antibodies, as described elsewhere for ovaries
(Theurkauf, 1992). Immunoflourescence with an anti-Drosophila
sperm-tail antibody was used to assay for fertilization in earlyarrested
embryos (Karr, 1991).
Conventional fluorescence microscopy was performed using a
Nikon Microphot FXA photoscope. Laser scanning confocal
microscopy was performed using a Biorad MRC600 confocal
imaging system installed on an Olympus IMT2 inverted microscope
and a Nikon Optiphot photoscope. The lenses used included
the Nikon 60× Planapo (1.4 numerical aperture (na)), the Nikon
Plan 20 DIC (0.5 na), the Olympus S Plan Apo 60×, Oil (1.4 na),
and the Olympus D Plan Apo 20×, UV. Oil (0.8 na).
RESULTS
Nuclear phenotypes of embryos derived from 76
female-sterile mutations
Of the 76 P-element induced female-sterile mutants that we
examined, 44 laid morphologically normal eggs. To define
the time in development that the nuclei arrested or when the
first abnormalities in nuclear behavior occurred, we
examined the morphology and distribution of the nuclei in
0- to 4-hour embryos derived from homozyous mutant
mothers. 12 of the 44 maternal-effect mutations resulted in
homozygous females producing embryos that developed
normally through cellularization at nuclear cycle 14. These
W. Sullivan, P. Fogarty and W. Theurkauf
1247Cytoskeleton in the syncytial Drosophila embryo
lines were not further characterized. Of the remaining 32
maternal-effect mutations, 23 caused severe disruptions or
arrest of the nuclear divisions prior to the completion of
nuclear migration at nuclear cycle 10 (Fig. 1). 2 of the 32
maternal-effect mutations, resulted in a failure of the
embryos to form pole cells. These 2 mutations exhibited no
other discernible nuclear defects (Fig. 1). Females homozygous
for any of the remaining 7 maternal-effect mutations
(AS6034, AS5812, AS2257, AS742, and fs37 on chromosome
two and AS928 and AS1479 on chromosome three) produced
embryos that successfully completed nuclear migration, but
underwent abnormal cortical nuclear divisions prior to cellularization
(Fig. 1). Since this phenotype is similar to that
observed in embryos treated with the specific actin-filament
drug cytochalasin, these 7 mutations were therefore of particular
interest and were subjected to the genetic and cell
biological studies described below.
The maternal-effect mutations that preferentially
disrupt the cortical blastoderm divisions
represent at least four distinct loci
Complementation analysis of the two, third chromosome
mutations, AS928 and AS1479, demonstrates that they fall
into a single complementation group. Deficiency analysis
maps both mutations to region 70D2-70D6. In situ hybridization
maps the P-element in AS928 to region 70 and
revertant analysis demonstrates the maternal effect is caused
by the presence of a P-element. The complementation group
defined by AS928 and AS1479 has been renamed nuclearfallout
(nuf). We designate these alleles nuf1 and nuf2
respectively.
Complementation analysis of the second chromosome
mutations yields three complementation groups: AS6034,
AS5812, and AS2257 fall into a single complementation
group, while fs37 and AS742 define the other two complementation
groups. Deficiency analysis maps AS6034,
AS5812, and AS2257 to region 36A8-36D1. This genetic
mapping is supported by in situ polytene hybridizations,
which localizes the P-element to region 36A-36B in the
AS6034 stock. Revertant analysis confirmed that the
maternal-effect in AS6034 was caused by the presence of a
P-element. The complementation group defined by AS6034,
AS5812, and AS2257 has been renamed grapes (grp). We
designate these alleles grp1, grp2 and grp3 respectively.
Deficiency analysis maps fs37 to region 43B1-43E and in
situ hybridization localizes the P-element to region 43.
Revertant analysis of fs37 demonstrates that the maternal
effect is caused by the presence of a P-element. fs37 has
been renamed scrambled (sced).
AS742 was previously named fsry9 and is a P-element
induced allele of staufen (cytological position 55A; St.
Johnson et al., 1991). Although the division defects
observed in AS742-derived embryos are consistent and
specific to the cortical divisions, the defects are not as
extensive as observed in the other mutations in this set (Fig.
2). In addition, we do not know whether these defects are a
direct consequence of the P-induced mutation. In this paper,
we do not further analyze this mutation.
grp, sced, and nuf are therefore maternal-effect genes that
preferentially disrupt the cortical nuclear divisions. For
sced, grp1, nuf1, we observed the same phenotype whether
the embryos were derived from females homozygous or
hemizygous for these mutations. Although further analysis
is required, this suggests that these mutations represent
amorphic alleles.
As this screen was restricted to a relatively small number
Fig. 1. Examples of typical
nuclear phenotypes found in
embryos derived from the
maternal-effect mutations
examined. (A) Nuclear-staining
of an embryo derived from
females homozygous for the
maternal-effect mutation fs25.
Females homozygous for this
mutation consistently produce
embryos that arrest during the
first nuclear division. Doublestaining
with an anti-sperm tail
antibody reveals that this
embryo had been fertilized.
The maternal-effect mutant fs6
produces embryos that arrest at
variable times during the
premigration nuclear cycles
(B). C depicts the maternaleffect
mutation fs4 in which the
embryos consistently fail to produce pole cells. (D) Embryos derived from homozygous fs37 females. The embryos undergo abnormal
postmigration nuclear divisions and consequently possess unevenly sized and spaced cortical nuclei. Prior to nuclear migration,
development is relatively normal in these embryos.
1248
of available P-element-induced maternal-effect lethals, and
did not include X chromosome mutations, it is clear that we
have not saturated the genome for this class of maternaleffect
mutations. For example, a previously characterized
second chromosome maternal-effect mutation, daughter less-abo-like
(cytological region 32A-F), produces a preferential
disruption of the cortical nuclear divisions, but was
not recovered in this screen (Sandler, 1977; Sullivan et al.,
1990).
Embryos derived from females homozygous for
grapes, scrambled, or nuclear-fallout exhibit
extensive nuclear division abnormalities only
after the nuclei have migrated to the cortex
Initial observations indicated that grp1, nuf1 and sced
develop normally prior to the cortical divisions. To quantify
these observations, we measured both the spacing and
nuclear division synchrony in premigration embryos
(nuclear cycles 5 through 8) derived from mothers homozyW.
Sullivan, P. Fogarty and W. Theurkauf
Fig. 2. 0- to 4-hour embryos derived from wild-type, AS742 (designated stauAS742 see text), sced, nuf1, and grp1 females were fixed with
formaldehyde and stained with an anti-histone antibody. Each row, depicting embryos derived from one of these maternal-effect
mutations, presents a surface view of a late syncytial blastoderm stage embryo (nuclear cycle 12 or 13), and a surface and cross sectional
view of a cellularized embryo.
1249Cytoskeleton in the syncytial Drosophila embryo
gous for each of these maternal-effect mutations. If greater
than 2% of the nuclei of an embryo exhibited abnormal
spacing or were greater than a half-nuclear cycle out of
division synchrony with respect to their neighboring nuclei,
the embryo was scored as abnormal. By this test, 5% of
wild-type embryos between nuclear cycles 5 and 8 were
scored as abnormal (Table 1). Embryos derived from the
three maternal-effect mutations exhibit values indistinguishable
from the wild-type values. Thus, embryos derived
from females homozygous for these maternal-effect
mutations appear to develop normally prior to nuclear
migration.
This assay was also used to examine nuclear cycles 9 and
10. Approximately 3% of the embryos derived from wildtype
females were abnormal during these nuclear cycles.
With the exception of sced-derived embryos, all of the
mutants produced less than 10% abnormal embryos during
nuclear cycles 9 and 10 (Table 1). In about 50% of the sced
-derived embryos, uneven nuclear migration to the cortex
was also observed.
In embryos derived from females homozygous for the
maternal-effect mutations, extensive division irregularities
occur during the postmigration cortical divisions (nuclear
cycles 11-14). In these mutants, the cortical nuclei share the
following abnormal features.
(1) Nuclei vary in size and shape, suggesting nuclear
fusion events, division failures and other division abnor-
malities.
(2) Patches of nuclei differ in density and/or stage in the
nuclear cycle from that of their neighboring nuclei.
(3) Regions of the cortex are free of nuclei; these regions
often contain high concentrations of centrosomes.
(4) Nuclei are present in the cortical cytoplasm beneath
the normal monolayer of cortical nuclei; often there is a
visible increase in the number of nuclei in the interior of the
embryo.
(5) Cellularization occurs at a reduced nuclear density,
and the size of the nuclei and cells at cellularization are proportionally
increased as the cellular density is decreased.
To quantify these observations, we estimated the percentage
of nuclei in each postmigration embryo that were
abnormally shaped, sized, or had moved into the interior of
the embryo. If this was greater than 2%, the embryo was
scored as abnormal. The results of this analysis are summarized
in Table 1. For the embryos derived from the maternaleffect
mutants, the percentage of embryos scored as
abnormal during nuclear cycles 11-14 ranged from 83% to
98% compared with 2% for normal embryos.
The postmigration nuclear phenotypes typical of embryos
derived from wild-type females and females homozygous
for the maternal-effect mutations are documented in Fig. 2.
The first column depicts a late syncytial blastoderm embryo
derived from each of these lines. In comparison with the
wild-type embryo (Fig. 2A), the embryos derived from the
maternal-effect mutations exhibit irregular nuclear distributions,
clumping of nuclei, and large gaps in the surface
monolayer of nuclei (Fig. 2D,G,J,M). Previous studies
suggest that these phenotypes are the consequences of
abnormal cortical nuclear divisions (Sullivan et al., 1990).
The second and third columns of Fig. 2 show surface and
cross sectional views, respectively, of the nuclear phenotype
at cellularization (or the equivalent time in development) of
embryos derived from these mutant females. Fig. 2B,C
shows a normal embryo cellularized at nuclear cycle 14. The
mutant embryos are ordered by an increasing severity of cellularization
defects. Embryos derived from sced females cellularize
with fewer cortical nuclei and an excess of interior
nuclei (Fig. 2H,I). This is probably a consequence of cortical
division errors and the subsequent loss of abnormal nuclei
from the surface monolayer (Minden et al., 1989; Sullivan
et al., 1990) In addition, the distribution of cortical nuclei is
irregular in sced-derived embryos (Fig. 2H). Embryos
derived from nuf1 females also cellularize with fewer surface
nuclei and a corresponding increase of interior nuclei (Fig.
2K,L). In nuf1-derived embryos, the distribution of surface
nuclei during cellularization is extremely irregular. The
nuclei often form clusters of varying size and shape (Fig.
2K). Finally, embryos derived from grp1 females do not cellularize
(Fig. 2N,O). Nomarski images reveal that they fail
even to initiate cellularization. As the cortical divisions
progress, the nuclei clump into clusters of approximately 5-
25 nuclei (Fig. 2M,N). grp1 differs from the other mutations
in that a much smaller proportion of the abnormal nuclei
move into the interior of the embryo.
sced, nuf, and dal mutants develop distinct
defects in the cortical cytoskeleton
This set of maternal-effect mutations all produce extensive
mitotic division errors only after the nuclei have migrated
to the cortex. Although they share this phenotype, analysis
of the cytoskeleton indicates that at least three of these
mutations produce distinct disruptions in embryonic actin
and tubulin organization during the cortical divisions. Fig.
3 depicts the actin and microtubule organization of syncytial
and cellularized embryos derived from wild-type, dal, nuf1
Table 1. Nuclear divisions in embryos from mothers
homozygous for each of the maternal-effect mutations
Nuclear cycle
No. of
5-8 9-10 11-14 embryos
wild-type Normal 39 34 49
Abnormal 2(5%) 1(3%) 1(2%) 126
sced Normal 18 18 6
Abnormal 1(5%) 7(28%) 211(97%) 261
grp1 Normal 49 45 8
Abnormal 1(2%) 2(4%) 60(88%) 165
nuf1 Normal 48 33 16
Abnormal 3(6%) 0(0%) 102(86%) 202
Embryos derived from females homozygous for each of the maternaleffect
mutations listed above were fixed with formaldehyde and their
nuclei DAPI-stained. The nuclear phenotype of these embryos were scored
with respect to their nuclear cycle. For nuclear cycles 5-8, an embryo was
scored as abnormal if more than 2% of the nuclei were spaced less than 3
µm from their nearest neighbor or were greater than a half-nuclear cycle
out of division synchrony with respect to their neighboring nuclei. For
nuclear cycles 9-10, the same scoring scheme was employed. For nuclear
cycles 11-13, an embryo was scored as abnormal if greater than 2% of its
nuclei exhibited abnormal size or shape or had moved into the interior of
the embryo. Estimates of the number of nuclei that had been removed
from the cortical nuclear monolayer could be obtained by scoring the
nuclei in the cortex of the embryo (within 10 µm of the plasma membrane)
but beneath the surface monolayer of nuclei.
1250 W. Sullivan, P. Fogarty and W. Theurkauf
Fig. 3. Confocal images of formaldehyde-fixed embryos derived from wild-type, dal, nuf1 and sced females that have been stained for
both actin (yellow/green) and tubulin (red). The top row depicts a surface view of a normal embryo in metaphase during the cortical
divisions (A), a surface view (B) and a cross sectional view (C) of a cellularized wild-type embryo. The remaining rows depict equivalent
views of embryos derived from the maternal-effect mutations. As described more fully in the text, this analysis indicates that each of the
mutations disrupts the cortical cytoskeleton in a distinct fashion.
1251Cytoskeleton in the syncytial Drosophila embryo
and sced mutant females. grp-derived embryos have not
been included in this analysis because they arrest prior to
cellularization. The first column in Fig. 3 presents a surface
view of metaphase embryos during the cortical divisions.
The second and third columns depict surface and crosssectional
views, respectively, of the embryos during cellularization
at nuclear cycle 14.
dal, nuf, and sced all produce defects in the pseudocleavage
furrows. In wild-type embryos, the actin-based pseudocleavage
furrows completely encircle each of the metaphase
spindle complexes during the syncytial cortical divisions
(Fig. 3). Similar views of dal-derived embryos reveal fused
and unevenly spaced spindles (Fig. 3D). Previous analysis
indicates that these spindle defects are a consequence of
incomplete centrosome separation during the previous interphase
(Sullivan et al., 1990). In dal-derived embryos,
abnormal pseudocleavage furrows are associated with
defective spindles, while normal spindles are surrounded by
normal furrows (Fig. 3D). The defects in actin organization
in dal embryos therefore appear to be a consequence of the
centrosome migration defect and subsequent abnormalities
in spindle distribution. The formation of actin caps during
interphase appears normal in dal-derived embryos.
Fig. 3G depicts a similar view of a nuf1-derived embryo.
The first obvious defect occurs during nuclear cycle 12 with
the appearance of large gaps in the actin-based furrows surrounding
each spindle (see arrow, Fig. 3G). Interestingly,
the gaps usually appear in regions of the furrow that are
closest to the metaphase plate. In spite of the defective
furrows, the spacing and structure of the spindles appear
normal. Thus, unlike dal, the first observable defect occurs
in the actin-based pseudocleavage furrows rather than in the
microtubules or microtubule-associated structures such as
the centrosomes. In nuf1-derived embryos, the formation of
the actin-caps during interphase occurs normally. During the
later cycles, fused and unequally spaced spindles become
apparent in these mutant embryos.
In sced-derived embryos, there is an almost complete
failure in pseudocleavage furrow formation (Fig. 3J).
During nuclear cycle 10, in sced-derived embryos, the
mitotic actin structures are severely disrupted, although
observable defects in the mitotic spindles are infrequent.
Later, defective spindles are seen. Because the actin defects
precede spindle defects in sced-derived embryos, the spindle
abnormalities in these embryos are likely to be a consequence
of failed furrow formation, rather than a cause. As
with dal and nuf, the sced mutation does not block formation
of the interphase actin caps (Theurkauf, Sullivan, and B.
Alberts, unpublished data).
The above analysis demonstrates that sced, nuf1 and dal
embryos can be distinguished by the nature of their
cytoskeletal defects during the syncytial blastoderm
divisions. Analysis of actin and microtubule organization
demonstrates that these mutations also produce distinct
defects during cellularization. An optical section of a cellularized
wild-type embryo at nuclear cycle 14 reveals evenly
spaced nuclei surrounded by microtubules which in turn are
surrounded by a ring of actin (Fig. 3B). The actin filaments
are just beneath the plasma membrane and the microtubules
are tightly associated with each nucleus. During cellularization,
the nuclei elongate perpendicularly to the plasma
membrane. In the cross-sectional views, the tightly associated
microtubules outline the elongated nuclei (Fig. 3C).
In dal-derived embryos, microtubule and actin organization
during nuclear elongation and membrane invagination
appear normal (Fig. 3E,F), although these embryos cellularize
with enlarged nuclei at one-half of the normal density
(Sullivan et al., 1987). Therefore the dal-induced defect is
likely to be specific to the syncytial divisions and does not
affect the process of cellularization.
While cellularization occurs in sced-derived embryos,
neither the honeycomb actin pattern nor the ring of microtubules
outlining each of the cellularized nuclei is as distinct
as observed in wild-type embryos (Fig. 3K,L). The cross
sectional view reveals that nuclear elongation is disrupted
and that the microtubules are not closely associated with the
nuclei in cellularized sced-derived embryos (Fig. 3L). As
with dal, the nuclei and cells are larger and at a lower density
relative to wild-type nuclei. This is likely to be the result of
the products of the abnormal cortical divisions moving into
the interior of the embryo, which leaves fewer, larger
surface nuclei.
The cellularization phenotype of nuf1-derived embryos is
dramatically different from either dal or sced. As panels H
and I illustrate, distinct cellularization furrows occur in nuf1
embryos, but the pattern is irregular with groups of nuclei
being encompassed by a single cellularization furrow.
Unlike sced, distinct rings of microtubules do form around
each nucleus during cellularization, and this occurs whether
or not membrane invagination has occurred. The microtubules
remain tightly associated with each nucleus and
nuclear elongation appears to occur normally.
DISCUSSION
In an attempt to identify genes involved in the organization
of the cytoskeleton of syncytial Drosophila embryos, we
have initiated a cytological screen of embryos derived from
females homozygous for maternal-effect lethal mutations.
This analysis has led to the identification of a class of
mutations in which the females produce embryos that
develop normally until the nuclei migrate to the cortex.
However, once the nuclei reach the cortex, these embryos
acquire abnormal mitotic figures, irregularly shaped nuclei,
and uneven nuclear spacing.
There are a number of possible explanations for the stagespecific
phenotypes found in this class. These mutations
could be hypomorphic alleles of genes required during both
the premigration and postmigration divisions, but the level
or activity of the mutant gene product may be such that it is
used up at the time the nuclei reach the cortex. We do not
believe that this is the case, however, because the transition
from normal to abnormal divisions is rather abrupt,
occurring only after the nuclei reach the cortex. With hypomorphic
alleles one might expect more variation in the
timing of the abnormal divisions. In addition, for sced, nuf1
and grp1, the same phenotype is observed whether the
embryos are derived from females homozygous or hemizygous
for the mutation.
We believe that these mutations are likely to disrupt genes
whose products are specifically involved in the organization
1252
of the cortical cytoskeleton. As described, the dividing
nuclei induce the formation of actin-based pseudocleavage
furrows which are unique to the syncytial cortical divisions.
Unlike normal cleavage furrows, these structures encompass
each metaphase spindle. Previous studies suggest that they
function as barriers to prevent improper connections
between spindles and to prevent the nuclei from colliding
with one another (Sullivan et al., 1990; Postner et al., 1992).
It is likely that pseudocleavage furrow formation requires
specialized proteins. The predicted effect on the nuclear
behavior of a mutation that is disrupted in one of these
proteins would be similar to that observed in the mutations
described above. Studies of permeabilized embryos treated
with the actin disrupting drug cytochalasin support this
notion. This treatment results in a phenocopy of the
mutations described above; embryos exhibit abnormal
nuclear divisions only after the nuclei have migrated to the
surface (Zalokar and Erk, 1976; Theurkauf, unpublished
observation). Indeed, the cytological analysis of actin and
microtubule organization in sced and nuf1 embryos demonstrates
that both of these mutations produce a specific defect
in the actin-containing pseudocleavage furrows (Fig. 4).
The mutations described here support studies demonstrating
that the cytoskeletal rearrangements of the cortical
syncytial divisions and of cellularization involve both
distinct and common elements. The previously described
maternal-effect mutation sponge results in failed actin-cap
and pseudocleavage furrow formation during the syncytial
divisions (Postner et al., 1992). Surprisingly, during nuclear
cycle 14, sponge-derived embryos form regular cellularization
actin networks. Similarly, in dal-derived embryos, cellularization
occurs normally even though the pseudocleavage
furrows are disrupted as a result of failed centrosome
separation during the syncytial divisions (Sullivan et al.,
1990). The newly identified sced mutation dramatically
disrupts the formation of the pseudocleavage furrows but
has surprisingly little effect on the formation of the actin
networks at cellularization. These three maternal-effect
mutations define components or processes that preferentially
disrupt actin dynamics during the syncytial divisions and do
not appear to be significantly involved in the formation of
the actin-based cellularization furrows.
Conversely, exhaustive deficiency screens have identified
a small set of zygotic genes specifically involved in cellularization
(Wieschaus and Sweeton, 1988; Merrill et al.,
1988). Deficiencies for these genes do not disrupt the
cytoskeletal or nuclear dynamics of the syncytial divisions.
serendipity alpha and nullo, two of the genes specifically
involved in cellularization, have been particularly well
studied. Deficiencies for serendipity alpha result in gross
disruptions of the actin/myosin network during cellularization.
serendipity alpha codes for a 58×103 Mr protein that
colocalizes with F-actin and is associated with the leading
front of the invaginating plasma membrane during cellularization
at nuclear cycle 14 (Schweisguth et al., 1990). By
germband extension the protein is no longer present indicating
that it is specifically involved in cellularization.
Careful analysis of the nullo phenotype, reveals that nullo is
involved in stabilizing the actin contractile network during
cellularization, but it is not involved in the initial formation
of this network (Simpson and Wieschaus, 1990). As
expected, the nullo gene is also transiently expressed just
prior to cellularization (Rose and Wieschaus, 1992).
In contrast to the genes described above, the phenotype
of nuf- derived embryos suggests that this gene plays a
similar role both in pseudocleavage furrow formation and
W. Sullivan, P. Fogarty and W. Theurkauf
Fig. 4. This figure summarizes our analysis of the cortical cytoskeleton in dal, sced, and nuf derived embryos. A ‘+’ indicates that the
event occurs normally while a ‘−’ indicates that the event does not occur or occurs abnormally. The two tables indicate that although each
of these mutations exhibit similar nuclear defects, they exhibit distinct cellular defects. Each mutation appears to be disrupting a different
step in the process of pseudocleavage furrow formation and cellularization.
1253Cytoskeleton in the syncytial Drosophila embryo
cellularization. In both cases, the actin-based furrows are
capable of forming properly, but there are extensive gaps in
the overall furrow network. The distribution of furrows
during the syncytial divisions and the distribution of
membrane invaginations during cellularization therefore
both rely on normal nuf activity. These studies reveal that
some gene products are required during the cortical
syncytial divisions and cellularization while others are
specifically required for one of these processes.
The similarity of the nuf phenotype to the zygotic
mutations nullo and serendipity is striking (Simpson and
Wieschaus, 1990; Schweisguth et al., 1990). In all three
mutations, the cellularization furrows fail to encompass
every nucleus creating an embryo with multinucleate cells.
From the analysis of these zygotic cellularization mutations
and extensive deficiency screening, it has been suggested
that the syncytial divisions and even the initial formation of
the actin/myosin contractile network at cellularization are
maternally controlled while the zygotic genes function to
stabilize the cellularization furrows (Simpson and
Wieschaus, 1990). Possibly nuf provides a maternally
supplied component required for pseudocleavage furrow
formation during the syncytial divisions and also the
initiation of furrow formation at cellularization. This
predicts that, unlike nullo, the initial formation of the cellularization
furrows would be perturbed in nuf-derived
embryos. Although preliminary analysis supports this interpretation,
a conclusion must await more extensive phentoypic
analysis. Although there are a number of possibilities
concerning the relationship of nuf, serendipity and nullo, an
attractive working model is that the maternally supplied nuf
acts upstream and in concert with the zygotic genes
serendipity and nullo to form and stabilize the actin
networks during cellularization.
In the zygotic mutation pebbles, pseudocleavage furrow
formation and cellularization are not disrupted, but cytokinesis
fails during the division following cellularization
(Hime and Saint, 1992). Therefore, the three similar actinmediated
events of Drosophila embryogenesis, pseudocleavage
furrow formation, cellularization, and cytokinesis
each require unique gene products.
The sced and nuf mutations also define dependency relationships
during formation of the pseudocleavage furrows
and during cellularization. It is evident from the sced
phenotype that the formation of the spindles is independent
of pseudocleavage formation. In addition, the sced cellularization
phenotype demonstrates that cellularization is independent
of nuclear shaping and the proper close association
of microtubules with each nucleus. In nuf-derived embryos,
microtubules properly encompass each nucleus even in the
absence of the actin-based plasma membrane at cellularization.
This cellularization phenotype demonstrates that
microtubule alignment around the nucleus during cellularization
is independent of the cellularization furrows.
Although each of the maternal effect mutations described
in this paper produces a distinct disruption in the cortical
cytoskeleton, they cause a number of common nuclear phenotypes,
including a deficiency of cortical nuclei and excess
interior nuclei at the time of cellularization. The excess
interior nuclei are explained by previous studies demonstrating
that the products of abnormal cortical divisions
move into the interior of the embryo, leaving fewer cortical
nuclei (Minden et al., 1989, Sullivan et al., 1990). The newly
identified mutations demonstrate that the removal of
abnormal nuclear products into the interior of the embryo is
a general response of the Drosophila embryo. In addition,
previous studies have demonstrated that when the abnormal
nuclei move into the interior they leave their centrosomes
on the surface. We have found that this is also true for the
mutations described here (data not shown).
Recently, a large number of actin and tubulin-associated
proteins have been isolated from the early Drosophila
embryo (Miller et al., 1989; Kellogg et al., 1989). Immunofluorescence
has revealed that some of these proteins are
specifically located in the cortex of the embryo. We believe
the mutational analysis described in our studies provides a
complimentary functional approach towards identifying
cortical cytoskeletal-associated proteins of the Drosophila
embryo. In addition to a molecular characterization of these
mutations, we are also initiating studies of the microtubule
and microfilament dynamics in embryos derived from
females homozygous for these mutations.
We would like to thank Bruce Alberts, Douglas Kellogg, Jordan
Raff, and John Tamkun for their critical reading of this manuscript.
We would especially like to thank Dr. Bruce Alberts in whose laboratory
a significant portion of this work was performed. We also
thank Tim Karr for kindly supplying us with the Drosophila antisperm
tail antibody. Finally we express our appreciation to the
Y.N. Jan and the A. Spradling laboratories for generously
providing us with the maternal-effect mutations.
This work was supported by grants to W. Sullivan from the
National Institutes of Health (R29 GM46409-01) and the American
Cancer Society (JFRA-366) and to W. E. Theurkauf from the
Damon Runyon-Walter Winchell Cancer Research Fund (DRG-
945) and to Bruce M. Alberts from the National Institutes of Health
(R01 GM23928 and P01 GM31286).
REFERENCES
Bakken, A. H. (1973). A cytological and genetic study of oogenesis in
Drosophila melanogaster. Dev. Biol.33, 100-122.
Bier, E., Vaessin, H., Shepherd, S., Lee, K., McCall, S., Barbel, S.,
Ackerman, L., Carretto, R., Uemura, T., Grell, E., Jan, L. Y. and Jan
Y. N. (1989). Searching for pattern and mutation in the Drosophila
genome with a P-lacZ vector. Genes Dev. 3, 1273-1287.
Cooley, L., Kelley, R. and Spradling, A. (1988). Insertional mutagenesis
of the Drosophila genome with single P elements. Science 239, 1121-
1128.
Engels, W. R., Preston, C. R., Thompson, P. and Eggleston, W. B.
(1986). In situ hybridization to Drosophila salivary chromosomes with
biotinylated probes and alkaline phosphatase. Focus 8, 6-8.
Erdelyi, M. and Szabad, J. (1989). Isolation and characterization of
dominant female sterile mutations of Drosophila melanogaster. I.
Mutations on the third chromosome. Genetics 122, 111-127.
Foe, V. E. and Alberts, B. M. (1983). Studies of nuclear and cytoplasmic
behavior during the five mitotic cycles that precede gastrulation in
Drosophila embryogenesis. J. Cell Sci.61, 31-70.
Fryberg, E. A. and Goldstein, L. S. B. (1990). The Drosophila
cytoskeleton. Ann. Rev. Cell Biol. 6, 559-596.
Gans, M., Audit, C. and Masson, M. (1975). Isolation and characterization
of sex-linked female sterile mutants in Drosophila melanogaster.
Genetics 81, 683-704.
Hatanaka, K. and Okada, M. (1991). Retarded nuclear migration in
Drosophila embryos with aberrant F-actin reorganization caused by
maternal mutations and by cytochalasin treatment. Development 111,
909-920.
1254
Hime, G. and Saint, R. (1992). Zygotic expression of the pebble locus is
required for cytokinesis during the postblastoderm mitoses of
Drosophila. Development 114, 165-171
Karpen, G. H. and Spradling, A. C. (1992). Analysis of subtelomeric
heterochromatin in the Drosophila minichromosome Dp1187 by single
P-element insertional mutagenesis.Genetics 132, 737-757
Karr, T. L. and Alberts, B. M. (1986). Organization of the cytoskeleton in
early Drosophila embryos. J. Cell Biol.102, 1494-1509.
Karr, T. L. (1991). Intracellular sperm/egg interactions in Drosophila: A
three-dimensional structural analysis of a paternal product in the
developing egg. Mech. Dev. 34, 101-111.
Kellogg, D. R., Mitchison, T. J. and Alberts, B. M. (1988). Behavior of
microtubules and actin filaments in living Drosophila embryos.
Development 103, 675-686.
Kellogg, D. R., Field, C. M. and Alberts, B. M. (1989). Identification
of microtubule-associated proteins in the centrosome, spindle, and
kinetochore of the early Drosophila embryo. J. Cell Biol. 109, 2977-
2991.
Lindsley, D. L. and Zimm, G. G. (1992). The genome of Drosophila
melanogaster. London, New York: Academic Press, Inc.
Merrill, P. T., Sweeton, D. and Wieschaus, E. (1988). Requirements for
autosomal gene activity during precellular stages of Drosophila
melanogaster. Development 104, 495-509.
Miller, K., Field, C. M. and Alberts, B. M. (1989). Actin binding proteins
from Drosophila embryos: a complex network of interacting proteins
detected by F-actin affinity chromatography. J. Cell Biol. 109, 2963-2975
Minden, J. S., Agard, D. A., Sedat, J. W. and Alberts, B. M. (1989).
Direct cell lineage analysis in Drosophila melanogaster by time lapse
three dimensional optical microscopy of living embryos.J. Cell Biol. 109,
505-516.
Mitchison, T. J. and Sedat, J. W. (1983). Localization of antigenic
determinants in wholeDrosophila embryos. Dev. Biol. 99, 261-264.
Mohler, J. D. (1977). Developmental genetics of the Drosophila egg I.
Identification of 59 sex-linked cistrons with maternal-effects on
embryonic development. Genetics 259-272.
Perrimon, N., Mohler, D., Engstrom, L. and Mahowald, A. P. (1986). Xlinked
female-sterile loci in Drosophila melanogaster. Genetics 113,
695-712.
Perrimon, N., Engstrom, L. and Mahowald, A. P. (1989). Zygotic lethals
with specific maternal-effect phenotypes in Drosophila melanogaster. I.
Loci on the X chromosome. Genetics 121, 333-352.
Planques, V., Warn, A. and Warn, R. M. (1991). The effects of
microinjection of rhodamine-phallodin on mitosis and cytokinesis in
early stageDrosophila embryos. Exp. Cell Res.192, 557-566.
Postner, M. A., Miller, K. G. and Weischaus, E. (1992). Maternal-effect
mutations of the sponge locus affect cytoskeletal rearrangements in
Drosophila melanogaster embryos. J. Cell Biol. 119, 11205-1218
Rabinowitz, M. (1941). Studies on the cytology and early embryology of
the egg ofDrosophila melanogaster. J. Morphol.69, 1-49.
Rice, T. B. and Garen, A. (1975). Localized defects of blastoderm
formation in maternal-effect mutants of Drosophila. Dev. Biol. 43, 277-
286.
Rose, L. S. and Wieschaus, E. (1992). The Drosophila cellularization gene
nullo produces a blastoderm-specific transcript whose levels respond to
the nucleocytoplasmic ratio.Genes Dev. 6, 1255-1268.
Sandler, L. (1977). Evidence for a set of closely linked autosomal genes
that interact with sex-chromosome heterochromatin in Drosophila
melanogaster. Genetics 86, 567-582.
Schupbach, T. and Wieschaus, E. (1986). Maternal-effect mutations
altering the anterior-posterior pattern of the Drosophila embryo. Roux’s
Arch. Dev. Biol.195, 302-317.
Schupbach, T. and Wieschaus, E. (1989). Female sterile mutations on the
second chromosome of Drosophila melanogaster. I. Maternal-effect
mutations. Genetics 121, 101-117.
Schweisguth, F., Lepesant, J. A. and Vincent, A.(1990). The serendipity
alpha gene encodes a membrane associated protein required for the
cellularization of theDrosophila embryo. Genes Dev. 4, 922-931.
Schweisguth, F., Vincent, A. and Lepesant, J. A. (1991). Genetic analysis
of the cellularization of the Drosophila embryo. Biology of the Cell 72,
15-23.
Simpson, L. and Wieschaus, E.(1990). Zygotic activity of the nullo-locus
is required to stabilize the actin-myosin network during cellularization in
Drosophila. Development 110, 851-863.
Sonnenblick, B. P. (1950). The early embryology of Drosophila
melanogaster. In Biology of Drosophila (ed. M. Demerec), pp. 62-167.
New York: John Wiley and Sons. Reprinted 1965, New York and
London: Hafner.
St. Johnston, D., Beuchle, D. and Nusslein-Volhard, C. (1991). staufen, a
gene required to localize maternal RNAs in the Drosophila egg. Cell 66,
51-63.
Stafstrom, J. P. and Staehelin, L. A. (1984). Dynamics of the nuclear
envelope and nuclear pore complexes during mitosis in the Drosophila
embryo. European Journal of Cell Biology 34, 179-189.
Sullivan, W. (1987). Independence of fushi tarazu expression with respect
to cellular density inDrosophila embryos. Nature 327, 164-167.
Sullivan, W., Minden, J. M. and Alberts, B. M. (1990). daughterless-abolike,
a Drosophila maternal-effect mutation that exhibits abnormal
centrosome separation during the late blastoderm divisions. Development
110, 311-323.
Szabad, J., Erdelyi, M., Hoffmann, G., Szidonya, J. and Wright, T. R. F.
(1989). Isolation and characterization of dominant female sterile
mutations of Drosophila melanogaster. II. Mutations on the second
chromosome. Genetics 122, 823-835.
Theurkauf, W. (1992). Behavior of structurally divergent alpha-tubulin
isotypes during Drosophila embryogenesis: evidence for posttranslational
regulation of isotype abundance. Dev. Biol. 154, 204-217.
Turner, R. R. and Mahowald, A. P. (1977). Scanning electron microscopy
of Drosophila melanogaster embryogenesis. I. The structure of the egg
envelope and the formation of the cellular blastoderm. Dev. Biol. 50, 95-
108.
Warn, R. M., Magrath, R. and Webb, S. (1984). Distribution of F-actin
during cleavage of the Drosophila syncytial blastoderm. J. Cell Biol. 98,
156-162.
Warn, R. M., Smith, L. and Warn, A.(1985). Three distinct distributions
of F-actin occur during the divisions of polar surface caps to produce pole
cells in Drosophila embryos. J. Cell Biol. 100, 1010-1015.
Warn, R. M. and Warn, A. (1986). Microtubule arrays present during the
syncytial and cellular blastoderm stages of the early Drosophila embryo.
Exp. Cell Res.163, 201-210.
Warn, R. N., Flegg, L. and Warn, A. (1987). An investigation of
microtubule organization and functions in living Drosophila embryos by
injection of a fluorescently labeled antibody against tyrosinated tubulin. J.
Cell Biol. 105, 1721-1730.
Wieschaus, E. and Sweeton, D.(1988). Requirements of X-linked zygotic
gene activity during cellularization of early Drosophila embryos.
Development 104, 483-493.
Zalokar, M., Audit, C. and Erk, I. (1975). Developmental defects of
female-sterile mutants of Drosophila melanogaster. Dev. Biol. 47, 419-
432.
Zalokar, M. and Erk, I. (1976). Division and migration of nuclei during
early embryogenesis of Drosophila melanogaster. J. Microbiology Cell
25, 97-106.
(Accepted 4 May 1993)
W. Sullivan, P. Fogarty and W. Theurkauf