Download Maternal-Effect Genes That Alter the Fate Map of the Drosophila

DEVELOPMENTALBIOLOGY
129,72-83(1988)
Maternal-Effect Genes That Alter the Fate Map of
the Drosophila Blastoderm Embryo
GARY M. WINSLOW,SEAN
B. CARROLL,'
Department of Molecular, Cellular and Devebpmental
ANDMATTHEW
P.
SCOTT~
Biology, Universi@ of Colorado at Boulder, Boulder, Colorado 8o$o$-osh7
Accepted April 26, 1988
The pattern of segmentation in the Drosophila embryo is controlled by at least 25 zygotically active genes and at least
20 maternally active genes. We have examined the pattern of expression of the protein product of the zygotically active
segmentation gene&s& tarazu (ftz) at the cellular blastoderm stage in progeny of mutant females homozygous for each
of six maternal-effect segmentation genes to observe the early effects of the maternal-effect genes on zygotic gene
expression. The genes included ewperantia
(a member of the anterior class of maternal-effect
segmentation genes);
staqen
and vusa (members of the posterior class); and torso, trunk, and fs(l)N
(members of the terminal class).
Mutations in the genes caused a disruption of the normal pattern of ft.z stripes in regions of the embryo where gene
activity is known to be required. The@ stripes provide a marker for segmental determination at the cellular blastoderm stage, making it possible to correlate aberrant patterns of& protein with defects in cuticle morphology at the end
of embryogenesis. J%Zprotein expression in progeny of females mutant for combinations of the above genes was also
examined. The changes in the& pattern in progeny of females doubly mutant for genes of the anterior and terminal
classes or of the posterior and terminal classes can largely be understood as the result of the additive effects of the
single mutations. In contrast, clearly nonadditive effects on thefiz pattern were seen when a mutation in a gene of the
anterior class (ezuperantia) was combined with mutations in posterior class genes. 0 1988 Academic PEW, he.
INTRODUCTION
abnormalities
that affect either contiguous blocks of
segments, alternate segmental units, or a part of each
segment (Niisslein-Volhard
and Wieschaus,
1980).
Among the zygotically
active segmentation
genes is
fushi taraxu (J%z).Embryos homozygous for null alleles
of ftx exhibit a pair-rule phenotype, i.e., alternate segment boundaries and adjacent pattern elements are
missing so that embryos form only about one-half the
normal number of segments (Wakimoto
et al, 1984).
Molecular and immunochemical
analyses of the products of the ftx gene have demonstrated
that at the cellular blastoderm stage& mRNA and protein molecules
are localized in a series of seven transverse stripes
(Hafen et al, 1984; Carroll and Scott, 1985). Each stripe
of ftz protein initially
corresponds to a parasegmental
unit (Lawrence et al, 1987), sofx products are present
in alternate parasegments.
A parasegment
(PS) consists of the posterior region of one segment and the
anterior region of the next most posterior segment
(Martinez-Arias
and Lawrence, 1985). Thus,ftx expression occurs in the primordia
that give rise to the segments that are missing infix- embryos. Cell fates are at
least partially
determined
by the cellular blastoderm
stage (Schubiger and Newman, 1982; Simcox and Sang,
1983), so the ftx protein stripes are convenient markers
for cell determination.
(For detailed fate maps of the
blastoderm embryo see Lohs-Schardin
et al. (1979) and
Hartenstein
et al. (1985).)
Genetic analyses in Drosophila have shown that the
development of the proper spatial organization
of the
embryo is dependent upon the activity of maternally
encoded factors that are present in the egg cytoplasm
(Anderson and Niisslein-Volhard,
1984a; Boswell and
Mahowald, 1985; Mlodzik et al., 1985; Frohnhiifer
and
Niisslein-Volhard,
1986; Lehmann
and Niisslein-Volhard, 1986; Macdonald and Struhl, 1986). Progeny of
females that are homozygous for mutations at any of a
number of maternally
active genetic loci develop distinct pattern of abnormalities
such as loss of normal
anteroposterior
or dorsal-ventral
polarity or aberrant
numbers of segments (Bull, 1966; Niisslein-Volhard,
19’77; Niisslein-Volhard
et ah, 1980; Anderson and Niisslein-Volhard,
1984b; Mohler and Wieschaus,
1985;
Frohnhiifer and Niisslein-Volhard,
1986; Sehiipbach and
Wieschaus, 1986; reviewed in Ntisslein-Volhard
et al.,
1987).
In addition to the maternally
active genes, at least 20
zygotically active genes are required for proper segmentation (reviewed in Akam, 1987; Scott and Carroll,
1987). Homozygous embryos carrying mutations in zygotically active segmentation
genes develop segment
1Present address: Laboratory of Molecular Biology, University of
Wisconsin-Madison,
Madison, WI 53706.
’ To whom correspondence should be addressed.
0012-1606/88
Copyright
All rights
$3.00
0 1988 by Academic Press, Inc.
of reproduction
in any form reserved.
72
WINSLOW,
CARROLL,
AND SCOTT
In order to understand the role of maternally
encoded
information
during pattern formation and segment development, we have examined the spatial pattern of ftx
protein expression in progeny of females homozygous
for mutations in several maternal-effect
genes that affect normal dorsal-ventral
polarity or anterior-posterior segmentation
(Carroll et aZ., 1986, 1987). Mutations
in several of the genes that have been studied cause
distinctive
alterations
in the normal spatial distribution offtz protein and it has been possible to correlate
changes infix protein distribution
with changes in cell
fate. This information
has been used to identify the
parts of the embryonic
primordia
that require the
function of specific maternal genes.
In this study we have extended our earlier analyses of
ftx protein expression to several other maternal-effect
segmentation genes. In addition, the effects of combinations of mutant genes on the spatial expression of ftz
have been studied. Each of the maternal-effect
mutations alters both the normal pattern of ftx protein expression and proper segmental development. No single
mutation
or any combination
of mutations
causes a
complete repression or derepression of ftx protein expression; we have not observed in any case an obvious
change in the timing or level of ftz expression. In the
progeny of females homozygous for combinations
of
mutant genes the alterations in the& pattern suggest
that specific interactions
occur among some maternaleffect gene products and that the interactions
take
place over long distances in the embryo. Related results
have been reported
by Mlodzik
et al. (1987) and
Frohnhijfer
and Niisslein-Volhard
(1987) using RNA
and antibody probes to the ftx gene products.
MATERIALS
AND
METHODS
The patterns of ftx protein expression in embryos
that were progeny of homozygous females of the following genotypes were examined: exuperantiaPJ, staufenHL,
vasaPD, torsowK, trunpA, and fs(l)NasratY
In addition,
progeny of females that were doubly and triply mutant
for some of the above genes were examined. All of the
mutant genotypes, except fs(l)Nasrat,
are from a collection of second chromosome maternal-effect
mutants
that has been described previously
(Schiipbach
and
Wieschaus, 1986). fs(l)Nasrat
(fs(l)N) is an X-linked
EMS-induced
maternal-effect
mutation
(Counce and
Ede, 1957; Degelmann et al, 1987). A detailed analysis of
the segmentation
defects of fs(l)N has been presented
(Degelmann et aZ., 1987). Although the strongest alleles
of each of the genes available at the time were used in
this study, there is at present no proof that the alleles
examined in this study are nulls.
Mutant embryos were obtained from females homo-
jlz
Expression
in Maternal
73
Mutants
zygous for the second chromosome maternal-effect
mutations. The second chromosome carrying each maternal-effect mutation
was marked with cn bw, which
causes the eyes to be white, so white-eyed homozygous
females were collected and placed in bottles with
yeasted egg-lay plates at 25°C. Females homozygous for
the X-linked mutation fs(l)Nasrat were isolated from a
stock balanced with FM3. Two- to 6-hr-old embryos
were collected, fixed, stained with anti-fix antibody, and
photographed as described by Carroll and Scott (1985).
At least 50 embryos were examined from each genotype.
Cuticle preparations were made essentially as described
by Van der Meer (1977). The wild-type ftx stripes are
numbered l-7, from anterior to posterior. For terminology used to describe cuticular markers see Lohs-Schardin et al. (1979).
RESULTS
Closes of Maternal
Phenotypes
Niisslein-Volhard
et al. (1987) have grouped the
known maternal-effect
genes that affect anteroposterior pattern into three phenotypic classes. The anterior
class includes genes such as tic&d @cd), exuperantia
(exu), and swallow (swa); mutations
in this class of
genes cause defects in anterior structures. Genes of the
posterior class are required for proper development of
abdominal
structures
and, in most cases, the pole
plasm. This class includes oskar (osk), vasa (vm), staufen (stau), and tudor (bud). Finally, torso (tor) and trunk
(trk) are characteristic of a class of genes that functions
in development of pattern at both ends of the embryo.
Mutations in these “terminal”
genes eliminate the patterns in terminal regions of the embryo.
The pattern of ftx protein expression and its correspondence to the cuticle morphology of a wild-type embryo are shown in Figs. la and lb and have been described in detail (Carroll and Scott, 1985). The wild-type
and mutant patterns of ftx expression are summarized
in a diagram (Fig. 4).
ftz Protein
Expression
in Mutants
of the Posterior
Class
Theftz pattern in progeny of females homozygous for
mutations in several posterior class genes has been described (Carroll et al, 1986). Theftx pattern and cuticle
phenotype of vasa progeny (Fig. lc) are characteristic of
that class of genes. Alterations
are seen in the regions
that would normally give rise toftx stripes 3 to 6. There
is a corresponding loss of all abdominal denticles (Fig.
Id), indicating that the abdominal defects are heralded
by changes in theftx pattern at the cellular blastoderm
stage.
FIG. 1. Changes in the pattern of ftz protein expression
and larval cuticle caused by maternal-effect
expression
(a, c, e, g, i, k), in progeny
from females homozygous
mutant
for maternal-effect
segmentation
form (b, d, f, h, j, 1). Polyclonal
antibody
specific to thefts
portion
of aftz+galactosidase
fusion protein
embryos.
The antibody
was visualized
using a fluorescein-conjugated
secondary
antibody.
Except where
tion the anterior
end of the embryo
is at the left and the ventral
side is at the bottom.
Cuticular
ventrolateral
view, with anterior
at the left. The identities
and positions
of some of the denticle
belts are
mutations.
The pattern
of ftz protein
genes, and the cuticle structures
that
was affinity
purified
and used to stain
noted, in photographs
of ftz localizaphenotypes
are shown in ventral
or
indicated.
Scale bars are 100 pm. (a) A
Wn~s~ow, CARROLL, AND SCOTT
For the progeny of each mutant, defects in the ftx
pattern have been correlated as precisely as possible
with the cuticular defects visible in embryos at the end
of embryogenesis. Because in some mutants the phenotype varies somewhat, we cannot always correlate the
defects in theftx pattern to the mutant cuticle morphology with absolute precision.
ftz
Expression
in Maternal
ftx Expression
Class
75
Mutants
in Mutants
of the Terminal
Embryos that are progeny of females homozygous
mutant for trk, tor, or fs(l)N exhibit phenotypes in
which pattern elements at the middle of the embryo
appear “expanded” anteriorly and posteriorly with concomitant loss of pattern elements at the posterior- and
anterior-most
regions of the embryo. The eighth dentiexuperantia, A Member of the Anterior Class, Causes
cle band is entirely absent or greatly reduced in embryos from trk or fs(l)N mutant females (Figs. lh and
an Anterior Shift in the ftx Pattern
lj). All telson structures are absent, and the cephaloIn embryos that are the progeny of homozygous mu- pharyngeal skeleton is reduced (Schiipbach and Wieschaus, 1986). Progeny of tar females often exhibit more
tant exu females the normal number of sevenftx stripes
is seen, but the spacing and width of the stripes are extensive pattern deletions, leaving only abdominal
aberrant. The most anterior fin stripe (hereafter re- denticle bands Al-A6 (Fig. 11).
ferred to as stripe 1) in exu progeny is found at about
The pattern of ftx staining in progeny of trk females
80% egg length (EL) (Fig. le, arrowhead). (The anterior
is altered such that only six stripes of protein are visible
(Fig. lg). The first five of these stripes have approxiedge of stripe 1 in wild-type embryos is normally found
at 65% EL, where the posterior end of the embryo is mately normal positioning, width, and spacing, but are
defined as OS..) The first stripe is followed by an ab- followed by an enlarged interstripe
and a sixth stripe
normally wide interstripe
of about 20 nuclei, a second that is abnormally
wide (five nuclei). The pattern of
stripe offtx protein of 12-13 nuclei, an interstripe of 10 staining in fs(l)N embryos is identical to that seen in
nuclei, and a series of five posterior stripes that are the progeny of females homozygous for trk (Fig. li).
slightly
compressed (Fig. le, bracket). The seventh
Mutations
in a zygotically
active gene, tailless (tZZ),
stripe is often only three nuclei in width (not shown),
cause a similar alteration in theftx pattern (Mahoney
instead of the usual five nuclei. The first stripe some- and Lengyel, 1987).
times lacks a well-defined posterior boundary of ftx exIn embryos from homozygous tar females, one sees a
pression. At the end of embryogenesis, when recognizftx pattern
displacement
similar
to that of trk and
able cuticular markers are visible, defects are seen at fs(l)N progeny, but with the difference that the sixth
the anterior end of the embryo. The head skeleton is stripe is often positioned still more posteriorly (Fig. lk,
reduced (Fig. If, arrow) and the median tooth is absent.
arrowhead) or in some cases is absent (not shown). In
The pattern of thoracic and abdominal denticles is in addition, the first five ftx stripes and interstripes
are
most cases normal, with only occasional defects seen in often slightly wider than those seen in wild-type or trk,
posterior structures. This result was unexpected, be- and are shifted posteriorly. In some cases, there is a
cause the region of the embryo in which the j’tx pattern
slight anterior shifting of stripe 1 as well. Thus, the tar
is most affected by loss of exu function gives rise to mutation appears to cause a shift of all the ftx stripes,
thoracic structures, which nonetheless have a normal or while trk and fs(l)N mutations
affect primarily
the
nearly normal cuticle. Frohnhiifer
and Ntisslein-Volmost posterior two stripes. It is possible that the difhard (1987) have reported that the thorax is slightly
ferences seen between the genes of the terminal class
expanded in progeny of females carrying the same exu are due to characteristics
of the particular alleles used
allele. A more dramatic change in the cuticle might be in this study and that tw, trk, fs(l)N, and tll are involved
expected from the changes in ftx expression.
in the same pattern-regulating
processes.
wild-type embryo at the cellular blastoderm stage (3 hr) exhibits seven stripes offtz protein. (b) A wild-type embryo at the end of embryogenesis (22 hr). Some segments are indicated. T, thoracic; A, abdominal; tl, telson. (c) vdpD/vc&? The broad region offtz staining that occurs in
place of stripes 3 to 6 does not form a continuous stripe around the embryo. vf, ventral furrow. There are often isolated nuclei expressing theftz
protein at normal levels. (d) wosrD/waP! All abdominal denticles are absent. sp, spiracles. (e) ezuPJ/ezuP< Note the anterior shift of the first
ftz stripe (arrowhead), the abnormally wide second stripe, and the compression of the posterior stripes (bracket). (f) exupJ/exup< The pattern
of thoracic and abdominal denticles is essentially normal, but note the defects in the head skeleton (arrow). (g) trkRA/trkRA.This embryo
exhibits fiveflz stripes of normal width and spacing followed by an enlarged interstripe and an abnormally wide sixth stripe. In some embryos,
the posterior stripe extends to the very end of the embryo, forming a “cap” of staining at the end of the embryo (not shown). (h) trkR4/trkRA.
Abdominal segment eight and the telson are absent. (i) fs(l)~l’/fs(l)N”‘l
(dorsal view). The pattern of&z expression is identical to that seen in
trk. Only six stripes of protein are visible. (j) fs(l)A@“/fs(l)N?
Only seven complete abdominal denticle bands are visible. An incomplete
eighth denticle band is found at the terminus of the embryo. (k) tar WK/torWKThe ftz stripes are slightly wider than in trk and shifted
posteriorly. The posterior stripe covers the entire end of embryo (arrowhead). (1) torWK/torWK.Only six abdominal segments are visible.
FIG. Z.ftz protein expression
and cuticular
phenotype
of progeny
from females doubly or triply homozygous for mutations in maternal-effect
segmentation
genes. (a, c, e, g, i) fez localization.
(b, d, f, h, j) Cuticle phenotypes.
(a) stm.&~~/~td~~~tor?
The embryo
has just begun
gastrulation.
Note the anterior
shift and abnormal
width of the first stripe and the wide band offtz staining
that covers the entire posterior
of
the embryo.
The cephalic furrow
(cf) is also shifted
anteriorly.
(b) .+taz&or
wK/sta~HLt~“?
There are patches of abdominal
denticles
at the
m/vd%rwK.
The position
of the anterior
border of the first stripe is
posterior
of the embryo.
No telson structures
are visible.
(c) ~u8~tar
The thorax
is expanded
and a patch of abdominal
denticles
is located at the posterior
end of the embryo
knal,
(d) vn~Dtw~/wa$lDtwWK.
where
the telson should be. (e} t&%rrWK/t~kRKtorwK.
Only five stripes
are visible, The two anterior-most
stripes
are wider
than the
corresponding
stripe in trk or tar and are shifted anteriorly.
The first stripe is found at about 76% EL, as compared
to 66% in wild-type
and trk
WINSLOW,
CARROLL,
AND
SCOTT
The Pattern of ftx Expression in Progeny of Females
Doubly Mutant for tar and stau or vas
In order
to learn whether interactions
occur among
maternal-effect
genes from the different phenotypic
classes, we have examined the ftx pattern in some double-mutant
combinations.
Strictly
additive
effects
would not suggest interactions,
while nonadditive
effects on theftx pattern would suggest that the different
classes of maternal-effect
loci affect each other.
In general, combining mutations of the posterior and
terminal
classes produces additive effects. Embryos
that are progeny of mutant females homozygous for
both stau (a posterior class gene) and tar (a terminal
class gene) exhibit two anterior stripes that are followed by one large region of ftx staining that extends
from what would normally be the anterior border of
stripe 3 to the posterior end of the embryo (Fig. 2a).
Stripe 1 is shifted anteriorly and lies just posterior to
the cephalic furrow, which has also been shifted anteriorly (as in stau progeny). Thus, the tur mutation
does
not alter the effect of the stau mutation on the two most
anterior stripes, and the changes caused by the tar mutation are not further altered (or diminished)
by the
loss of stau function. The cuticular phenotype of the
double-mutant
stau tar progeny reveals defects in pattern elements that correspond to the pattern alterations seen with the ftz antibody. Posterior structures
are absent apparently due to the loss of tar function,
and the abdominal segments are fused, as is characteristically seen in stau progeny (Fig. 2b). Therefore, the
changes in the.@ pattern seen in stau tar progeny appear to be due to the independent, additive effects of the
two mutations.
Subtle deviations from a strictly additive effect occur
in progeny of homozygous VCLStar females (vas, like stau,
is a posterior class gene). A single broad region of posterior ftx staining again replaces stripes 3 to 6, but the
apparent effect of tcw is stronger in the vas tor case than
in the stau tar case (Fig. 2~). The tor mutation causes a
slight posterior shift of the.@ stripes (Fig. lk), but the
degree of posterior shifting is often more pronounced in
vas tar double-mutant
progeny. The anterior border of
the third stripe is now found at only 20% EL, as compared to 50% in wild-type or 44% in stau tar progeny
(Fig. 2~). As in stau tar progeny, the third stripe in vas
to-r progeny
now extends as a broad band to cover the
entire posterior of the embryo. The cuticular phenotype
ftz Expression
in Maternal
Mutants
77
of vas tar progeny shows a posterior shift of pattern
elements relative to either of the single mutant progeny
(Fig. 2d). The thoracic segments are wider than in
wild-type and encompass a large portion of the embryo,
while the residual abdominal denticles are found in only
a small patch at the very posterior.
Combinations
of tar, trk, and exu
Progeny of doubly mutant females were examined in
order to determine whether interactions occur between
genes of the terminal class (trk, tw) and the anterior
class (exu). When progeny of females mutant for both
trk and tar are examined, the defects in the@ pattern
are similar to either single mutant, but stronger (Fig.
2e), even though both genes are in the same class. There
is a clear difference in the positioning and width of the
first two anterior stripes and the posterior stripe when
the progeny of the double-mutant
females are compared to those of the single mutants (Fig. 2e, arrowheads). In trk tar progeny, the fifth.& stripe is shifted
posteriorly, and a sixth stripe is never visible although
it is always present in trk progeny and often visible in
tar progeny. The overall effect is that the extent of posterior and anterior shifting of theftx stripes is greater
than that seen in either of the single-mutant
progeny.
One possible explanation
is that the’ tw and trk gene
products
perform a similar function or participate
in
the same process and each mutation only partially eliminates this function.
The cuticle phenotype (Fig. 2f) of progeny of the
doubly mutant females is not significantly
different
from that seen in progeny of females homozygous for
tar or trk alone, although theftx pattern in the trk tar
progeny suggests that additional morphological
defects
might be observed. Thus, the ftx pattern is a more sensitive marker for changes in the blastoclerm fate map
than is the pattern of cuticle markers seen at the end of
embryogenesis, perhaps because later in embryogenesis
the pattern is partially corrected.
A clearly nonadditive effect on thefiz pattern is seen
when tor (terminal class) is combined with exu (anterior class). Only five stripes of staining are visible and
the two anterior stripes are wider and are shifted anteriorly compared to those of wild-type (Fig. 2g). However, the extent of the shifts and the altered size of the
anterior stripe in exu tar progeny a0 not appear as pro-
and about 67% in tar progeny.
(f) trkR%wWK/trkRKtw
‘? The pattern
of adominal
denticles
is similar
to that of tar. (g) exu’%r m/exupJtor wK.
Thefiz
pattern
shows similarity
to trk and tm-.The fluorescence
at the dorsal anterior
surface is due to the contact with another
embryo
and is
notjk
protein.
(h) exupJtorwK/exup%r
? (i) exupJtormtrkRA/exupJtormtr~A.
The pattern
is essentially
identical
to that in (e) and (g). (j)
exupJtorw~t~kRA/e3cupJtorwxt~~
The defects in the head structures
are more extensive
than in single- and double-mutant
progeny.
In some
cases, abdominal
denticle
bands 5 and 6 are fused or band 6 is entirely
absent (not shown).
78
DEVELOPMENTAL BIOLOGY
nounced as in exu alone. The anterior-most
stripe in exu
to 80%
in exu (Fig. Zg, arrowhead), suggesting that normal tcw
function is involved in the anterior shift seen in the
absence of only exu. The extent of posterior shifting of
theft2 stripes in exu tar is greater than that in tar alone,
and is similar to that seen in the trk tar double-mutant
progeny (Fig. 2e). This result suggests that exu function
is required in the posterior of the embryo, at least in
progeny of tcrr females. The pattern of abdominal dentitles in the exu tar progeny is similar to the tw or trk
tor progeny. Denticle bands A7 and A8 and telson
structures are entirely absent. In contrast to tar, where
head structures are partially
deleted, most head and
gnathal structures are deleted in exu tor progeny (Fig.
3h), indicating additive effects of the two mutations on
head structures.
When progeny of females mutant for trk, exu, and tar
are examined, the pattern of ftx expression is similar to
that seen in exu tar progeny. The anterior-most
stripe
in the triple-mutant
progeny appears wider than that in
exu progeny (compare Figs. lc and Zi) and is often
curved anteriorly along the dorsal-ventral
axis. At the
end of embryogenesis the pattern of abdominal denticles is similar to that of the double-mutant
trk tm- or
exu tw progeny (Fig. Zj), although more extensive pattern deletions sometimes occur. In addition, most, if not
all, head and gnathal structures are absent. Thus, in all
combinations
of tar, trk, and exu, some nonadditive effects are seen on theftx pattern (summarized in Fig. 4),
suggesting that some interactions are occurring among
the gene products during early development.
tar progeny is found at 70-75% EL, as compared
ftx Expression in Progeny of Females Homozygous for
exu and stau or vas
A simple model would predict that additive effects on
theftx pattern would be observed when a mutation that
affects anterior development
(exu) is combined with
ones that affect primarily
posterior development (stau
or vas). However, dramatic nonadditive pattern alterations in the pattern offtx protein expression were seen
in embryonic progeny of females homozygous for exu
and either stau or vas. In stau exu progeny the sevenftx
stripes are replaced by a large solid band of protein that
extends from about 20 to 75% EL (Fig. 3a). In some
embryos a partial stripe is observed at the anterior of
the embryo, at about 90% EL (not shown). Considerable
variation in the stripe pattern is seen in the stau exu
double-mutant
progeny, including embryos that have
three to sixftx stripes of aberrant width and spacing, or
mirror-image
pattern duplications
of the fcx pattern.
The variability
in theftz pattern is likely to be due to
the residual function of the stau allele used in this
VOLUME
129.1988
study. Figure 3a shows the most extreme effect on the
ftx pattern of the loss of stau and exu function.
The cuticle phenotype of stau exu double-mutant
progeny exhibits a similar variability.
In what appears
to be the strongest phenotype, the embryonic cuticle
consists of thoracic segments that form mirror-image
duplications (Fig. 3b, arrows), followed immediately
by
a very reduced telson (Schiipbach and Wieschaus, 1986).
No head structures or abdominal denticles are present,
and an inverted telson is often found at the anterior end
of the embryo.
In progeny of homozygous vas exu females, a similar
broad band of ftx protein is visible (Fig. 3~). When ftx
protein first appears, at the cellular blastoderm stage,
it is of uniform intensity along the embryo, but at the
onset of gastrulation
the nuclei at both the anterior and
posterior borders of the broad band have high levels of
fcx protein, while nuclei at the center of the embryo
have lower levels of ftx protein. The intensity of the ftx
staining appears to decrease toward the middle of the
embryo in a graded fashion (Fig. 3e, arrow). No posterior stripe of the sort seen in stau exu progeny has been
seen in vas exu progeny. In contrast to stau exu, in vas
exu progeny, the fez pattern and the cuticle phenotype
show little variability.
No thoracic or abdominal denticles are visible (Fig. 3d), and the only discernible cuticle
structures are of gnathal origin (Schiipbach and Wieschaus, 1986).
DISCUSSION
In order to learn more about the relationship between
the maternally
controlled processes of anteroposterior
pattern specification and the expression of the wellcharacterized zygotic segmentation
genes, we have examined the spatial pattern of ftx protein expression at
the cellular blastoderm stage in embryos obtained from
mutant females. A summary of the patterns of ftx protein expression is presented in Fig. 4. The effects on ftx
expression were compared to the pattern of embryonic
cuticle observed at the end of embryogenesis. Each of
the maternal-effect
mutations examined causes reproducible alterations
of both the ftx protein expression
pattern and the cuticular morphology.
In general, the
ffx stripes are very good markers for cell fate at the
blastoderm
stage. Because proper segmental development requires the combined activities of multiple segmentation
genes, it is likely that the expression patterns of other segmentation
genes are altered as well.
The analysis of double-mutant
progeny suggests that
the maternal genes of the anterior and terminal classes
or the posterior and terminal classes act independently
or show only subtle interactions.
In contrast, genes of
the anterior
and posterior
classes interact
very
WINSLOW,
CARROLL,
AND SCOTT
jIz Expressionin Maternal
Mutants
79
FIG. 3. ftz protein expression and cuticular phenotypes of progeny from females doubly homozygous for emL and stau or vas. (a) stauHL
exuPJ/stHLexuPJ. The broad band of&z staining is followed by a weak posterior stripe (arrowhead). (b) stauHLexuPJ/stHLexupJ. Arrows indicate
polarity
of segments. vm, vitelline membrane (removed from the embryos shown in Figs. l-2, but not from this preparation). (c) vasPDexupJ/
va8DextiP( The&z protein is distributed in a broad band from about 15-75% EL. (d) vasPDexuPJ/va?DexuPJ. Note the absence of all thoracic
expression within the
and abdominal denticles. (e) A higher magnification view of va.8Dex~pJ/v&Dex~ pJ at the beginning of gastrulation.fiz
large stripe occurs as a gradient (arrows), from high levels anteriorly and posteriorly to low levels in the middle of the embryo. (f) exUPJ/exUPJ.
A gradient of ftz expression is visible at the posterior edge of the most anterior stripe (arrow).
strongly. Simultaneous loss of both anterior and posterior genes may cause the entire central region of the
embryo to assume a similar fate.
Changes in the ftx Pattern Can in Most Cases Be
Correlated with Changes in the Cuticle Morphology
Theftx protein is expressed at the early cellular blastoderm stage in even-numbered parasegments (PS)
(Lawrence et al., 1987), which correspond to the primordia of the posterior maxillary (pMx) and anterior
labial (aLb) segment (PS2), posterior prothorax (pTl),
and anterior mesothorax (aT2) segment (PS4), and so
on (Fig. 4). The stripes (and interstripes) of ftz protein
thus serve as molecular markers for cell fates in the
blastoderm embryo. For the most part, in embryos that
are progeny of mutant females, the segmental primordia of the embryo where the ftx pattern is altered and
the regions of the cuticle that are defective are closely
correlated. This indicates that the maternal genes have
most of their effects on embryonic patterning prior to
the cellular blastoderm stage.
80
DEVELOPMENTAL BIOLOGY
VOLUME 129.1988
In stau progeny, the widening and shifting of the
stripes
that mark the thoracic and gnathal primordia
Wt
(PS2-PS4) also correlate with the cuticle phenotype.
2 4 6 8 10 12 14
Cuticular
structures
derived from PS2-4 occupy a
larger part of the embryo than those in wild-type. The
stau
anterior shift offtx stripes 1 and 2 in stau may occur as
a result of the loss of a function that would under norvas
\\J
mal circumstances prevent the anterior stripes and interstripes from expanding to the anterior. Precedent
for such a mechanism is seen in the interactions among
gap genes. The domain of one gap gene expands in the
absence of an adjacent gap gene function (Jackie et al,
1986).
trk
71
In trk and fs(l)N progeny there is a very good correlation between the segmental primordia
that appear
shifted or absent as shown by theftx pattern and the
structures that are defective in the larval cuticle. The
sixth ftz stripe marks PS12 in the blastoderm
anlage
(the primordia for pA6 and aA7) (Fig. 4). The presence
of the sixthftz stripe in trk and fs(l)N progeny suggests
that the A7 denticle band will be present in the cuticle,
as is the case. (The denticle bands originate from the
anterior of each segment primordium
and form in the
anterior region of each segment.) Although the eighth
exu tar b
denticle band is derived from the region of the embryo
marked by the sixth interstripe
(PS13, i.e., pA7 and
aA8)
in
wild-type
embryos,
no
eighth
denticle band is
trk tar 11
seen in trk progeny. This may be because the sixth interstripe seen at the posterior end of trk embryos contains only the primordia
for pA7 structures and lacks
exutrktar(~J
aA primordia. There are no morphological
markers for
pA7; correspondingly,
only naked cuticle is seen at the
posterior end of trk progeny (Fig. lg). In fs(l)N progeny,
however, the sixth interstripe seen in the embryo probably includes primordia
for at least part of aA8, as a
small patch of A8 denticles are present in the cuticle of
fs(l)N progeny (Fig. lj).
FIG. 4. A summary of the patterns of & staining in progeny of
In tor progeny, the A7 denticle band is often missing,
females homozygous for maternal-effect mutations. In the first panel,
segmental (S) and parasegmental (PS) boundaries are indicated, as and A6 is positioned at the posterior of the embryo. In
well as the approximate positions of & protein stripes in wild-type
addition, all abdominal
denticle bands appear to be
embryos. In the following panels, the positions of ftz stripes in the shifted posteriorly
to a greater degree than those in trk
mutant progeny are shown. In each case, the average positions of the
or fs(l)N. This phenotype correlates with the ftx patftz stripes were determined by measurements of at least four emtern.
Stripes 1 to 5 are shifted posteriorly, and the sixth
bryos. Measurements were made using the ends of the embryo as
j%z stripe (PS12, i.e., pA6 and aA7) is either entirely
reference points. M, maxillary segment; L, labial segment.
missing or is positioned at the very posterior of the
embryo (Fig. 4).
In exu tar, trk tar and exu trk tor progeny, only five ftz
In the stau and vas mutant phenotypes, theft2 pattern is altered primarily
in the region of fix stripes 3 to stripes are seen, and what appears to be the fifth ftz
interstripe
(PSll, i.e., pA5 and aA6) is found at the
6 (PS6-PS12), which comprise the segmental primordia
posterior of the embryos. The&
pattern predicts corfor Al through A7 (Fig. 4). It is these pattern elements
that are largely absent in embryos from stau or vas rectly that A6 is the most posterior segment observed in
homozygous females. Thus, both stau and vas are re- the cuticle of progeny of females of these three genoquired to direct cells to form the primordia for the ab- types. The anterior shifting of stripes 1 and 2 that is
observed is consistent with the defects occurring in
dominal segments at the cellular blastoderm stage.
‘thorax’
S
tOrI
vasexu
I
abdomen’
(M~Lt112131112l3~4151617181
WINSLOW,
CARROLL,
AND
SCOTT
head structures, although the thoracic segments appear
normal in these progeny. Paradoxically,
the head defects in the double- and triple-mutant
progeny appear
more severe than those in exu alone, yet the j?ftx stripes
appear to be shifted anteriorly
to a lesser extent. As
molecular markers for anterior cell fates become available, it maybe possible to resolve this paradox.
Additive and Nonadditive
Eflects Are Observed in
Double-Mutant
Progeny
ftz Expression in Maternal Mutants
81
ening of the anterior-most
ftx stripes in a manner not
observed in either single mutant. Considering the similarity in the cuticle phenotypes of tor and trk, one might
have expected that the double-mutant
progeny would
show a pattern offtx stripes similar to that of the single
mutants. This result would have suggested that tar and
trk act in a common pathway of cell determination.
The observed results suggests either that the two genes
act in parallel pathways, both contributing
to anteroposterior patterning, or that the alleles examined only
partially
inactivate
two genes that act in the
same pathway.
The tar mutation causes a posterior shifting of theftx
stripes, and the stau mutation causes a widening of the Interactions between exu and Genes of
two most-anterior
stripes and interstripes
and the forthe Posterior Class
mation of a large band offtx staining in place of stripes
An exact correlation between the extent of pattern
3 to 6 (Fig. 4). In the double-mutant
stau tar progeny,
the broad band of ftx protein in stau progeny appears to disruption of theftx stripes and the cuticle defects is not
seen in the progeny of homozygous exu females. The
have spread to the posterior end of the embryo, presumably due to the absence of tor function. The ftx pattern
two anterior ftx stripes and interstripes,
which normally
correspond
to
the
primordia
of
PS2
through
PS5
can therefore be understood as the result of additive
stau and tor effects, suggesting that tor (a terminal
(i.e., pMx through aA3), are abnormally positioned and
of incorrect width. In the larval cuticle, however, the
gene) and stau (a posterior gene) act largely independently of one another in the oocyte or embryo.
thoracic segments, which include PS3 through PS5, apThe effect of combining vas with tar is very similar to pear normal or are slightly expanded (Frohnhofer and
that seen in stau tor progeny. In this instance, however,
Niisslein-Volhard,
1987). The most obvious defects in
some deviation from strict additivity
of the ftz pattern
exu progeny appear in the head skeleton. The head dechanges is seen. The posterior shift caused by the tar fects are probably due to respecification of the fate of
mutation is often more pronounced in the double-muthe cells that are normally found at positions anterior
tant progeny. The two anterior stripes and interstripes,
to 65% EL and are therefore outside the region where
which mark PS2 through PS5, are wider than in either
ftx is expressed. The anterior expansion of theftz patsingle mutant and are shifted posteriorly. The embrytern suggests that the cells that normally give rise to
onic cuticle phenotype reflects this posterior shift, in the most anterior structures have adopted more postethat structures derived from PS2 through PS5 (pMx
rior fates.
through aT3) occupy nearly the entire embryo. Thus,
Frohnhijfer and Niisslein-Volhard
(1987) have postuwhile the phenotype of vas tor progeny is to a first
lated that the exu phenotype is associated with the alapproximation
the sum of the vu.s and tar phenotypes,
tered activity of bed, a gene which has been shown to be
there are some changes in theftx pattern that would not required for organizing
anterior development
in the
have been predicted from an examination
of either sin- embryo. Cytoplasmic transplantation
experiments have
gle mutant.
demonstrated that wild-type bed activity can cause the
Other deviations from a strictly additive effect on the development of ectopic anterior structures (FrohnhMer
ftz pattern can be seen when the anterior class mutation
and Niisslein-Volhard,
1986). In addition, bed RNA is
exu is combined with mutations of the terminal class. In observed to form a gradient in the embryo, from high
progeny of exu tar females, the ftx pattern exhibits tar- levels anteriorly to low levels posteriorly (Frigerio et ah,
like posterior shifts and exu-like anterior shifts. How- 1986). It was proposed that exu functions to localize bed
ever, the extent of posterior shifting caused by tar muactivity to the anterior pole and that the level of bed
tations is exaggerated in the absence of exu function as gene activity determines the degree of anterior developit is in the absence of vas, and the extent of anterior
ment. The loss of exu causes a more even distribution
of
shifting is not as pronounced as it is in exu progeny
bed gene product, and consequently the loss of anterior
alone. The tor mutation in the double-mutant
exu tar structures and the expansion of thoracic positional
progeny in some manner reduces the anteriorly directed
values (Frohnhofer
and Niisslein-Volhard,
1987; Bershifting of the ftz pattern that is caused by the exu leth et ah, 1988).
mutation.
When exu, which primarily
affects anterior patternProgeny of mutant females homozygous for trk and ing, is combined with vas, which has its greatest effect
tar exhibit an anterior and posterior shifting and widin the abdominal primordia, the embryonic progeny ex-
82
DEVELOPMENTAL BIOLOGY
hibit one very broad band offtx protein (Fig. 4). During
gastrulation,
this band of ftx expression forms a gradient of ftx protein that decreases toward the middle of
the embryo. A similar gradient of fez staining is sometimes visible at the posterior border of stripe one (PS2)
in exu progeny (Fig. 3f). An interpretation
of the ftx
pattern in exu vas progeny is that the single large ftz
stripe is in fact an expansion and mirror-image
of the
first ftz stripe. This interpretation
is supported by the
cuticular phenotype in which the vestigial cuticle is entirely of gnathal origin, which is derived in part from
PS2 (pMx and aLb).
In contrast to the vas exu case, the large stripe of ftx
staining visible in progeny of homozygous stau exu females may correspond to an enlarged PS4ftx stripe (i.e.,
pT1 and aT2). This interpretation
is supported by the
cuticle phenotype, in which T2 is duplicated and flanked
anteriorly by Tl and, in some cases, telson structures,
and posteriorly
by a reduced telson (Schiipbach and
Wieschaus, 1986). The large stripe of fix staining possibly marks the primordia
for the thoracic structures,
and the ftz stripes seen at the anterior and posterior
ends of the embryo are probably a normal and duplicated/inverted
seventh ftx stripe, which marks the primordia for the terminal segmented structures.
Theftx patterns and the cuticle phenotypes described
above for stau exu and vas exu may be a result of the
altered activity of bed, as proposed for exu (Frohnhofer
and Ntisslein-Volhard,
1987; Berleth et ah, 1988). It has
been inferred that the activity of a gene of the posterior
class, oskur, can inhibit bicoid activity. Injection of cytoplasm from oskur+ embryos into the anterior of wildtype embryos leads to a reduction of anterior structures; cytoplasm from oskar- embryos has no effect
(Lehmann and Niisslein-Volhard,
1986). The findings
suggest that oskur, and possibly other genes of the posterior class such as vas and stau, inhibit bed activity at
the posterior pole. In the absence of exu function, which
would normally
localize bed function in the anterior,
and without vas or stau, which may act negatively upon
bed function in the posterior, bed may act uniformly
in
the embryo. This altered bed activity would then cause
all cells to assume anterior developmental
fates.
In the case of vas exu the anterior cells have a gnathal
identity, while in stuu exu they have a thoracic identity.
This suggests that bed activity is roughly uniform, but
at different levels in the two double-mutant
progeny. In
vas exu we propose that bed gene activity is at a level
that would specify gnathal development (PS2), while in
stau exu, bed gene activity is at a level that specifies
thoracic development (PS4). The appearance of terminal structures and the absence of ftx expression at the
poles in the double mutants may be due to the normal
activity of the terminal genes. The differences in the
VOLUME 129,1988
phenotypes seen in vas exu and stau exu may be because
other genes of the posterior class) functions
in the anterior of the embryo as well and thus interacts
differently with bed and/or exu.
stuu (unlike
We thank Drs. Trudi Schiipbach, Adelheid Degelmann, and Anthony
Mahowald for sending stocks and Vern Twombly for care and maintenance of the stocks. We also thank Trudi Schiipbach for her insightful
comments, encouragement, and suggestions; Joan Hooper and David
Gubb for critical comments on the manuscript; and Wolfgang Driever
for a stimulating discussion. We thank Cathy Inouye for skilled preparation of the manuscript. This work was supported by an NIH Postdoctoral Fellowship GM09756 to S.B.C., NIH Grant HD18163, and a
Searle Scholar Award to M.P.S.
REFERENCES
AKAM, M. (1987). The molecular basis for metameric pattern in the
Drosophila embryo. Develomnt lOl, l-22.
ANDERSON, K. V., and N~SSLEIN-VOLHARD, C. (1984a). Information
for the dorsal-ventral pattern of the Drosophila embryo is stored
as maternal mRNA. Nature (Lmdcm)311,223-227.
ANDERSON, K. V., and NOSSLEIN-VOLHARD, C. (1984b). Genetic analysis of dorsal-ventral embryonic pattern in Drosophila In “Pattern
Formation” (G. M. Malacinski and G. Bryant, Eds.), pp. 269-289.
MacMillan, New York.
BERLETH, T., BURRIN, M., THOMA, G., BOPP, D., RICHSTEIN, S., FRIGERIO, G., NOLL, M., and NUSSLEIN-VOLHARD. (1988). The role of
localization of Zricoid RNA in organizing the anterior pattern of the
Drosophila embryo. EMBO J. 7,1749-1756.
BOSWELL, R. E., and MAHOWALD, A. P. (1985). tudor, a gene required
for assembly of the germ plasma in Drosophilamelanogaster.
Cell
43,97-104.
BULL, A. (1966). Bicaudal, a genetic factor which affects the polarity
of the embryo of Drosophila
melanogaster.
J. Exp. Zoo1161,221-242.
CARROLL,
S. B., and SCOTT, M. P. (1985). Localization of the fushi
tarazu protein during Drosophila embryogenesis. Cell 43,47-57.
CARROLL, S. B., and Scorr, M. P. (1986). Zygotically active genes that
affect the spatial expression of thefushi tarazu segmentation gene
during early Drosophila embryogenesis. Cell 45,113-126.
CARROLL, S. B., WINSLOW,
G. M., SCHOPBACH, T., and SCOTT, M. P.
(1986). Maternal control of Drosophila segmentation gene expression. Nature (London)323,278-280.
CARROLL, S. B., WINSLOW,
G. M., TWOMBLY, V. J., and SCOW, M. P.
(1987). Genes that control dorsoventral polarity affect gene expression along the antero-posterior axis of the Drosophila
embryo. Development 99,327-332.
COUNCE, S. J., and EDE, D. A. (1957). The effects in embryogenesis of a
sex-linked female sterility factor in Drosophilamelanogaster.
J.
Embryol.Exp. Morphol.5,404-421.
A., HARDY, P. A., PERRIMON,
N., and MAHOWALD,
A. P.
(1987). Developmental analysis of the torso-like phenotype in Drosophilaproduced by a maternal effect locus. Dev.Biol. 115.479-489.
FRIGERIO,
G., BURRI, M., BOPP, D., BAUMGARTNER, S., and NO& M.
(1986). Structure of the segmentation gene pairedand the Drosophila PRD gene set as part of a gene network. Cell 47,735-746.
FROHNH~FER,
H. G., and N~SSLEIN-VOLHARD, C. (1986). Organization
of anterior pattern in the Drosophilaembryo by the maternal
gene
DEGELMANN,
bicuidNature (Lmdcm)324,120-125.
H. G., and NOSSLEIN-VOLHARD,
C. (1987). Maternal
genes required for the anterior localization of bicoid activity in the
embryo of Drosophila. Genes andDevelopment
1,880-890.
FROHNH~FER,
WINSLOW, CARROLL, AND SCOTT
HAFEN, E., KUROIWA, A., and GEHRING, W. J. (1984). Spatial distribution of transcripts from the segmentation gene fushi tarazu of
Drosophila. Cell 37,825~831.
HARTENSTEIN, V., TECHNAU, G. M., and CAMPOS-ORTEGA, J. A. (1985).
Fate-mapping in wild-type Drosophila
melanogaster.
III. A fate
map of the blastoderm.
Wilhelm
Roux’s
Arch.
Dev. Biol. 194,
213-216.
JACKLE, H., TAUTZ, D., SCHUH, R., SEIFERT, E., and LEHMANN, R.
(1986). Cross-regulatory interactions among the gap genes of Drosophila
Nature
(London)
324,668-670.
LAWRENCE, P. A., JOHNSTON, P., MACDONALD, P., and STRUHL, G.
(1987). Borders of parasegments in Drosophila embryos are delimited by the fwhi tarazu and even-skipped
genes. Nature (London)
328,440-442.
LEHMANN, R., and NUSSLEIN-VOLHARD, C. (1986). Abdominal segmentation, pole cell formation, and embryonic polarity require the
localized activity of oskur, a maternal gene in Drosophila. Cell 47,
141-152.
LOHS~CHARDIN, M., CREMER, C., and NUSSLEIN-VOLHARD, C. (1979).
A fate map for the larval epidermis of Drosophila
melanogaster:
Localized defects following irradiation of the blastoderm with an
ultraviolet laser microbeam. Dew. Biol. 73,239-255.
MACDONALD, P. M. and STRUHL, G. (1986). A molecular gradient in
early Drosophila embryos and its role in specifying the body pattern. Nature (Lo&m)
324,537-545.
MAHONEY, P. A., and LENGYEL, J. A. (1987). The zygotic mutation
tailless alters the blastoderm fate map of the Drosophila
embryo.
Da?
Biol.
122,464-470.
MARTINEZ-ARIAS, A., and LAWRENCE, P. A. (1985). Parasegments and
compartments in the Drosophila embryo. Nature
(London) 313,
639-642.
MLODZIK, M., DEMONTRION, C. M., HIROMI, Y., KRAUSE, H. M., and
GEHRING, W. J. (1987). The influence on the blastoderm fate map of
maternal-effect genes that affect the antero-posterior
pattern in
Drosophila. Genes Dev. 1,603-614.
ftz Expression
in Maternal
83
Mutants
MLODZIK, M., FJOSE, A., and GEHRING, W. J. (1985). Isolation of caudal, a Drowphila
homeoboxcontaining
gene with maternal expression whose transcripts form a concentration gradient at the preblastoderm stage. EMBO J. 4(11), 2961-2969.
MOHLER, J., and WIESCHAUS, E. (1985). Bicaudal
mutations of fi@
sophila
melanogaster: Alteration
of blastoderm cell fate. Cold
Spring Harbor
Symp. Quad. Biol 50,105-112.
NUSSLEIN-VOLHARD, C. (1977). Genetic analysis of pattern formation
in the embryo of Drosophila melanogaster. Wilhelm Roux’s Arch.
Dev. Biol.
183,249-268.
NUSSLEIN-VOLHARD, C., FROHNHUFER,H. G., and LEHMANN, R. (1987).
Determination of anteroposterior polarity in the Drosophila embryo. Science 238,1675-1681.
NUSSLEIN-VOLHARD, C., LOHS-SCHARDIN, M., SANDER, K., and
CREMER, C. (1980). A dorso-ventral shift of embryonic primordia in
a new maternal-effect mutant of Drosophila.
Nature
(London)
283,
474-476.
NUSSLEIN-VOLHARD, C., and WIESCHAUS, E. (1980). Mutations affecting segment number and polarity in Drosophila. Nature (Ibndon)
287,795-801.
SCHUBIGER,G., and NEWMAN, S. M. (1982). Determination in Drosophila embryos. Amer. Zool. 22,47-55.
SCHUPBACH, T., and WIESCHAUS, E. (1986). Germline autonomy of
maternal-effect mutations altering the embryonic body pattern of
Drosophila.
Dev. Biol.
113,443-448.
SCOTT, M. P., and CARROLL, S. B. (1987). The segmentation
meotic gene network in early Drosophila development.
and hoCell
51,
689-698.
SIMCOX, A. A., and SANG, J. M. (1983). When does determination occur
in Drosophila
embryos? Dev. Biol. 97,212-21.
VAN DER MEER, J. (1977). Optical clean and permanent whole mount
preparations for phase contrast microscopy of cuticular structures
of insect larvae. Drosophila
Infwm.
Serv. 52,160.
WAKIMOTO, B. T., TURNER, F. R., and KAUFMAN, T. C. (1984). Defects
in embryogenesis in mutants associated with the Antennapedia
gene complex of Drosophila melanogaster.
Dev. Biol. 102,147-172.