2165
Development 127, 2165-2176 (2000)
Printed in Great Britain © The Company of Biologists Limited 2000
DEV8700
Regulation of cell proliferation and patterning in Drosophila oogenesis by
Hedgehog signaling
Yan Zhang and Daniel Kalderon*
Department of Biological Sciences, Columbia University, 1212 Amsterdam Ave., New York, NY 10027, USA
*Author for correspondence (e-mail: [email protected])
Accepted 22 February; published on WWW 18 April 2000
SUMMARY
The localized expression of Hedgehog (Hh) at the extreme
anterior of Drosophila ovarioles suggests that it might
provide an asymmetric cue that patterns developing egg
chambers along the anteroposterior axis. Ectopic or
excessive Hh signaling disrupts egg chamber patterning
dramatically through primary effects at two developmental
stages. First, excess Hh signaling in somatic stem cells
stimulates somatic cell over-proliferation. This likely
disrupts the earliest interactions between somatic and
germline cells and may account for the frequent mispositioning of oocytes within egg chambers. Second, the
initiation of the developmental programs of follicle cell
lineages appears to be delayed by ectopic Hh signaling. This
may account for the formation of ectopic polar cells, the
extended proliferation of follicle cells and the defective
differentiation of posterior follicle cells, which, in turn,
disrupts polarity within the oocyte. Somatic cells in the
ovary cannot proliferate normally in the absence of Hh or
Smoothened activity. Loss of protein kinase A activity
restores the proliferation of somatic cells in the absence of
Hh activity and allows the formation of normally patterned
ovarioles. Hence, localized Hh is not essential to direct egg
chamber patterning.
INTRODUCTION
3 (Gonzalez-Reyes and St. Johnston, 1998a). The posterior
location of the oocyte in the nascent egg chamber depends on
a selectively strong interaction with posterior somatic cells that
is mediated by DE-cadherin and is key to future development
of oocyte and embryonic polarity (Godt and Tepass, 1998;
Gonzalez-Reyes and St. Johnston, 1998a).
A distinct lineage of somatic cells is determined soon after
cyst envelopment to give rise to a quiescent pair of ‘polar cells’
at each pole of the egg chamber and the 5-7 ‘stalk cells’ that
separate consecutive egg chambers as they bud from region 3
of the germarium (Tworoger et al., 1999). The remaining
proliferating somatic cells show no obvious lineage restrictions
and form a continuous epithelium around the growing cyst of
fifteen nurse cells and one oocyte (Margolis and Spradling,
1995). These follicle cells proliferate until stage 6, midway
through oogenesis, after which they undergo three rounds of
endo-replication and then synchronously cease general DNA
replication in favor of amplification of a small number of
genes, including specific chorion genes (Royzman and OrrWeaver, 1998).
Several different follicle cell types can be distinguished in
developing egg chambers. ‘Main body follicle cells’ separate
two large fields of ‘terminal cells’. This division of fates
depends on Notch function (Gonzalez-Reyes and St. Johnston,
1998b; Larkin et al., 1999). Within the anterior terminal field
‘border cells’ form adjacent to the polar cells, followed
successively to the posterior by ‘stretched cells’ and
Drosophila oogenesis (reviewed by de Cuevas et al., 1997; Lin,
1997; Spradling, 1993; van Eeden and St. Johnston, 1999)
offers the opportunity to study how cell signaling regulates
proliferation and patterning of both germline and somatic cell
lineages. Each of the two ovaries in an adult fly contains 1518 ovarioles. Egg chambers form in the germarium, at the
anterior of each ovariole, and mature through successive stages
as they move posteriorly to form a polarized oocyte,
surrounded by a patterned epithelium of somatic follicle cells.
At the anterior tip of the germarium are non-proliferating,
somatic ‘terminal filament cells’ and ‘cap cells’ (see Fig. 6A)
(Forbes et al., 1996a; Godt and Laski, 1995). Two or three
germline stem cells maintain contact with the base of the
terminal filament, releasing daughter cystoblasts to the
posterior (Deng and Lin, 1997). Cystoblasts undergo four
mitotic divisions with incomplete cytokineses to produce a cyst
of sixteen interconnected cells that is enveloped by somatic
cells as cysts move through region 2 of the germarium. The
enveloping somatic cells derive from stem cells located at the
border between regions 2a and 2b of the germarium (Margolis
and Spradling, 1995). The germline cyst adopts a flattened disc
shape spanning the width of the germarium before entering
region 3. At this time, one of the centrally located germline
cells, which has been determined to be the oocyte, adopts the
most posterior position as the cyst becomes spheroid in region
Key words: Hedgehog, Oogenesis, Polarity, Drosophila
melanogaster
2166 Y. Zhang and D. Kalderon
‘centripetal cells’ (Gonzalez-Reyes and St. Johnston, 1998b).
At stage 9, border cells and the anterior polar cells delaminate
and migrate as a group to the border between the nurse cells
and oocyte, while the rest of the follicular epithelium moves in
a posterior direction, leaving only thin, elongated stretched
cells over the nurse cells. Posterior terminal cells behave
differently from anterior cells because they receive a signal
from the oocyte, encoded by the TGFα homolog gurken (grk)
(Lee and Montell, 1997; van Eeden and St. Johnston, 1999).
Correctly specified posterior follicle cells signal back to the
oocyte, leading to a redistribution of oocyte microtubules at
stage 6 (van Eeden and St. Johnston, 1999). This allows correct
localization of mRNAs such as bicoid and oskar within the
oocyte, specifying future anterior/posterior polarity in the
embryo. Microtubules and correctly specified posterior follicle
cells are also required for the oocyte nucleus to move to the
anterior cortex of stage 7-8 oocytes, where it produces a locally
high concentration of grk RNA that specifies dorsal follicle
cells and ultimately the future dorsal side of the embryo.
Within this framework there are several important
unanswered questions. How do somatic cells at the posterior of
a germline cyst in region 2b develop higher affinity than more
anterior somatic cells for the oocyte? How are polar cell, stalk
cell, and other follicle cell fates progressively specified and how
is their proliferation regulated? Which posterior follicle cells
produce the hypothetical signal that re-organizes the oocyte’s
microtubules to set up correct polarity within the oocyte and
what is the nature of this signal? Previous studies have
suggested that Hedgehog signaling may be involved in some of
these processes (Forbes et al., 1996a,b; Tworoger et al., 1999).
Hedgehog (Hh) signaling contributes to pattern formation
and regulates proliferation in a number of settings in Drosophila
and other animals (Ingham, 1998). In adult Drosophila ovaries,
Hh is expressed in terminal filament and cap cells (Fig. 6A) and
is required for normal somatic cell proliferation in the
germarium and for budding of egg chambers (Forbes et al.,
1996a). Hh acts by binding to the transmembrane protein,
Patched (Ptc), which normally restricts the activity of the 7transmembrane domain protein, Smoothened (Smo), ultimately
leading to the activation of the transcription factor, Cubitus
interruptus (Ci) (Ingham, 1998). Loss of ptc activity stimulates
intracellular Hh signal transduction, even in the absence of Hh
(Ingham, 1998), and in the ovary produces phenotypes similar
to the consequences of heat-shock induced hh expression
(Forbes et al., 1996b). We therefore refer to ptc mutant
phenotypes as being due to ectopic Hh signaling. These
phenotypes include the production of excessive somatic cells,
mis-positioning of oocytes within egg chambers and the
generation of ectopic polar cells. It has therefore been
postulated that Hh signaling regulates somatic cell proliferation
at an early stage in the germarium and may also regulate oocyte
positioning, perhaps via effects on somatic cell-type
determination.
Here we show that loss of ptc activity can also induce follicle
cell proliferation beyond stage 6, disrupt polarity within the
oocyte, delay polar cell differentiation and induce ectopic
border cells. We investigate the origins of phenotypes due to
ectopic Hh signaling and suggest (i) that polar cells can induce
a border cell fate in adjacent cells, (ii) that excessive cell
proliferation and not induction of ectopic polar cells causes
oocyte mis-positioning, (iii) that a delay in the initiation of
follicle cell developmental programs may underlie production
of ectopic polar cells, future defects in follicle cell proliferation
and aberrant development of posterior follicle cells, and (iv)
that very few correctly specified posterior follicle cells suffice
to direct normal polarity in the oocyte. We also use mutations
in several Hh signaling components to corroborate recent
models suggesting that Hh signals through at least two distinct
intracellular pathways that combine to regulate the magnitude
or quality of a cell’s response. Finally, we investigate whether
Hh normally regulates ovarian somatic cell proliferation and
those patterning events that can be disrupted by excessive or
ectopic Hh signaling.
MATERIALS AND METHODS
Fly strains and clonal analysis
Mutant clones were generated by mitotic recombination using the
FLP/FRT system (Xu and Rubin, 1993), an X-chromosomal hs-flp
(Ohlmeyer and Kalderon, 1998), and the following alleles: ptcS2, PKA
catalytic subunit null allele DC0H2, Su(fu)LP, fumH63, P[fu+], and smo2
(formerly designated smo11X43) (Ohlmeyer and Kalderon, 1998), smo3
and smoD16 (Chen and Struhl, 1998), hhts2 (Ma et al., 1993), and
cos2W1 (Sisson et al., 1997). Clones were marked by using
αtub84BlacZ (‘tublacZ’) transgenes on 2L and 2R (Harrison and
Perrimon, 1993), or FRT42Dπ-myc (Xu and Rubin, 1993).
Recombination was induced by heat shocking 3rd instar larvae at
38°C for 1 hour. The relevant genotypes (parentheses) of animals for
inducing clones lacking specific activities (bold) are listed below.
ptc (FRT42D ptcS2/FRT42D tublacZ); pka (DCOH2
FRT40A/tublacZ FRT40A); smo (smo2,3 or D16 FRT40A/tublacZ
FRT40A); fu (fumH63; P[fu+] FRT40A / tublacZ FRT40A or fumH63;
FRT42D P[fu+] / FRT42D tublacZ); pka fu (fumH63; P[fu+]
FRT40A/DC0H2 FRT40A); ptc fu (fumH63; FRT42D P[fu+]/FRT42D
ptcS2); pka Su(fu) (DC0H2 FRT40A/FRT40A; Su(fu)LP); pka hh
(DC0H2 FRT40A/tublacZ FRT40A; hhts2); ptc hh (FRT42D
ptcS2/FRT42D tublacZ; hhts2). Flies that carry hhts2 were raised at
18°C, then switched to 29°C after eclosion. The following follicle cell
markers were incorporated into some of the genotypes above: neulacZ (A101; Clark et al., 1994); 5A7, BB127, L53b, and 998/12
(Gonzalez-Reyes and St. Johnston, 1998b); pnt-lacZ (PZ07825; Lee
and Montell, 1997), and kin-lacZ (KZ503; Clark et al., 1994).
Staining ovaries
β-galactosidase activity staining was performed as described
previously (Cooley et al., 1992). For antibody staining, ovaries were
dissected into phosphate-buffered saline (PBS), washed 10 minutes in
PBST (0.1% Tween-20 in PBS), and fixed in 1:1 heptane: 8%
formaldehyde in PBST for 10 minutes. The ovaries were
permeabilized with 1% Triton X-100 in PBS for 1 hour, blocked in
5% BSA for 30 minutes, then incubated with antibody for 2 hours at
room temperature or at 4°C overnight. The antibodies used in this
study were anti-β-galactosidase (1:500, Cappel), anti-Fasciclin III
(1:3, gift from N. Patel), anti-phospho-histone H3 (1:500, Upstate
Biotech.), anti-Delta (1:1, gift from M. Muskavitch), anti-myc (1:20,
Oncogene), anti-Bicaudal D (1:20, Suter and Steward, 1991), anti-Ci
m2A1 (1:1, Slusarski et al., 1995). The secondary antibodies used
were fluorescein anti-rabbit IgG, Texas red anti-rabbit IgG and Texas
red anti-mouse IgG (1:500, Molecular Probes). Sometimes,
biotinylated secondary antibodies (1:500, Vector) were used to
amplify the primary antibodies, followed by a 2-hour ABC reagent
(Vector ABC Elite Kit) incubation. Microscopy was performed on a
Nikon Optiphot-2 microscope with Nomarski differential interference
contrast optics. Fluorescent staining was examined using a BioRad
MRC-600 confocal microscope system.
Hedgehog in oogenesis 2167
RESULTS
Ectopic Hh signaling disrupts many somatic cell
behaviors, oocyte position and oocyte polarity
We characterized the consequences of ectopic Hh signaling by
examining adult ovaries 7-9 days after inducing mitotic
recombination (Xu and Rubin, 1993) to produce ptc mutant
germline and somatic cell clones. We confirmed previous
observations (Forbes et al., 1996a,b) that ovarioles containing
ptc mutant stem cell derivatives accumulated an excess of
follicle cells, especially between egg chambers (Fig. 1A,B),
included egg chambers in which the oocyte was not at the
posterior (32%; 42/133) and many egg chambers with ectopic
polar cells, revealed by neu-lacZ expression (Fig. 1C,D). We
also found some novel phenotypes, which indicated more
extensive alteration of follicle cell behavior than previously
shown and the potential for ectopic Hh signaling to disrupt
polarity within the oocyte.
First, expression of neu-lacZ in the two pairs of polar cells
at either end of the egg chamber normally appears by stage
2 (Fig. 1C) but was delayed in most (80%; 28/35) ptc mutant
ovarioles until stage 4 (Fig. 1D). Similarly, refinement of Fas
III expression from its initial pattern in all follicle cells to
being exclusively in polar cells normally takes place by stage
2 (Fig. 1A), but was delayed to beyond stage 4 in most
ovarioles containing ptc mutant clones (Fig. 1B). Hence,
differentiation of polar cells and perhaps also the
developmental program of other follicle cells were delayed in
most ptc mutant ovarioles.
Second, follicle cells proliferated beyond stage 6 in ptc
mutant ovarioles. Antibody to a phospho-epitope of histone H3
(Wei et al., 1999) showed cells undergoing mitoses no later
than stage 6 in wild-type ovarioles (Fig. 2A) but up to stage 10
in ptc mutant ovarioles (Fig. 2B). BrdU labeling showed that
all follicle cells switched uniformly from genome-wide DNA
replication to chorion gene amplification at stage 10 in ptc
mutant ovarioles (data not shown) as in wild-type (Royzman
and Orr-Weaver, 1998). Many nuclei in stage 10 egg chambers
from ptc mutant ovarioles were smaller than those of other
follicle cells (Fig. 2D), consistent with the idea that one or
more rounds of endo-replication had been replaced by mitotic
cell cycles.
Third, one or two ectopic groups of border cells were often
found in stage 9-10 egg chambers of ptc mutant ovarioles (Fig.
3A,B). These ectopic border cells were often found overlying
nurse cells in stage 10 egg chambers, indicating defective
migration.
Fourth, defects in anterior-posterior polarity within the
oocyte were observed in stage 8-10 egg chambers even when
the oocyte occupied its normal posterior position. In 25%
(12/49) of stage 8-10 egg chambers from ovarioles with ptc
mutant clones, the oocyte nucleus had failed to migrate from
the posterior to the anterior cortex (Fig. 3C,D). Also, in 41%
(22/54) of stage 9 egg chambers, a kinesin-β-galactosidase
fusion protein, which normally localizes at this stage to the
extreme posterior (Clark et al., 1994) (Fig. 3E), was found in
the middle of the oocyte (Fig. 3F). Similar oocyte polarity
phenotypes have also been observed as a result of ectopic
expression of Hh (Tworoger et al., 1999).
We explored the origin of these diverse phenotypes by using
additional follicle cell markers, by determining their cellular
and temporal origins, and by assaying the effects of weaker
ectopic Hh signaling to dissociate phenotypes.
Defective posterior follicle cell development causes
oocyte polarity defects
Normal oocyte polarity depends on a re-organization of
microtubules that is triggered at stage 6 by a signal from
posterior follicle cells (van Eeden and St. Johnston, 1999). This
signal has not been identified but can only be delivered by
correctly specified posterior follicle cells. Several aspects of
posterior cell development appeared to be normal in ptc mutant
ovarioles. These cells never expressed markers characteristic
of terminal anterior follicle cells (5A7, border cells; BB127,
centripetal cells; or L53b, border cells and anterior stretched
cells) in egg chambers with the oocyte at the posterior (for
example see Fig. 3B and Fig. 4A). Also, Delta expression was
selectively reduced in posterior follicle cells at stage 6 (Fig.
4B), as in wild-type ovarioles (Larkin et al., 1999). However,
expression of the posterior follicle cell markers, 998/12 (Fig.
4C-E) (Gonzalez-Reyes and St. Johnston, 1998b) and pnt-lacZ
(data not shown) was sometimes partly or completely absent.
In all egg chambers where the oocyte nucleus remained
aberrantly at the posterior of the oocyte beyond stage 7, pntlacZ and 998/12 expression were completely missing from the
posterior follicle cells (Fig. 4E). Thus, a failure of posterior
follicle cells to differentiate normally is likely to underlie the
observed defects in oocyte polarity.
Cell autonomous effects on follicle cells underlie all
ptc mutant phenotypes
Marked clones of ptc mutant cells were generated 8 days prior
to ovary dissection to examine the cell autonomy of mutant
phenotypes. Both ectopic ptc-lacZ (data not shown) and FasIII
expression (Fig. 5A) were found to be fully penetrant cell
autonomous responses to ptc inactivation in proliferating
somatic cells. The majority of ptc mutant cells accumulated in
stalks between egg chambers. Ectopic induction of the polar
cell marker, neu-lacZ, was limited to a subset of the FasIIIexpressing ptc mutant cells that were in contact with germline
cells (Fig. 5B). Thus, ptc mutant cells tended to aggregate,
exhibited abnormally low affinity for germline cells and
appeared to require contact with the germline to become polar
cells.
Ectopic expression of the border cell marker, 5A7, was not
cell autonomous. Instead, ectopic border cells were found
surrounding ectopic polar cells, implying that polar cells can
induce a border cell fate over a short distance (Fig. 5C).
Only ptc mutant follicle cells were found to proliferate
beyond stage 6 (Fig. 2C). Furthermore, in stage 10 egg
chambers all ptc mutant cell nuclei were smaller than those of
surrounding follicle cells (Fig. 2D), suggesting that the
substitution of mitoses in place of endo-replication cycles is
fully penetrant. Absent or reduced expression of the posterior
follicle cell marker 998/12 was also autonomous to ptc mutant
clones (Fig. 4G-I), but was not fully penetrant. Thus, all
changes in follicle cell behavior other than border cell
induction were autonomous to ptc mutant cells.
Oocyte positioning was examined at early stages by using
antibody to Bicaudal-D (BicD) (Suter and Steward, 1991) (Fig.
6B). Mis-positioned oocytes were seen as early as region 3 of
the germarium (Fig. 6C), implying disruption of initial
2168 Y. Zhang and D. Kalderon
Fig. 1. Polar cell specification in ptc mutant ovarioles. (A) Fas III
antibody staining of wild-type ovarioles and (B) ptc mutant
ovarioles. (C) neu-lacZ expression pattern of wild-type ovarioles and
(D) ptc mutant ovarioles. In all panels anterior is to the left. Numbers
indicate the approximate stage of egg chambers. Initial neu-lacZ
expression and refinement of Fas III expression to polar cells are
evident by stage 2 in wild-type ovarioles, but are delayed until at
least stage 4 in most ptc mutant ovarioles. ptc mutant ovarioles also
show ectopic neu-lacZ and Fas III expression.
positioning rather than a failure to maintain a posterior position
in the egg chamber. Most of the egg chambers with altered
oocyte position retained ptc activity in the germline (Fig. 6D);
moreover, several egg chambers with a normally positioned
oocyte contained ptc mutant germline cells. Hence, correct
oocyte position does not depend on germline ptc activity. Misplaced oocytes were found only within ovarioles that contained
some ptc mutant follicle cells, implying that ptc activity in
somatic cells is critical. There was, however, no correlation
between oocyte position and the location of ptc mutant cells.
For example, the proportion of egg chambers with mis-placed
oocytes was similar whether ptc mutant clones occupied the
posterior (52/154=34%) (Fig. 6F), the anterior (24/118=20%)
(Fig. 6D), or the whole of the egg chamber (102/394=26%).
Also, the mis-placed oocyte was sometimes in contact with ptc
mutant cells (41/103=40%) and sometimes with cells that
retained ptc activity (62/103=60%) (Fig. 6D,E). Indeed, in a
few egg chambers (13/428=3%) there was a mis-placed oocyte
but no ptc mutant cells in the surrounding epithelium. Thus,
ptc mutant somatic cells can affect oocyte position without
stably contacting the germline cyst.
Fig. 2. Follicle cell proliferation in ptc mutant ovarioles. Antibody to
phosphohistone H3 (green) labels cells in mitosis. (A) In wild-type
ovarioles follicle cells stop proliferation after stage 6, but (B) in ptc
mutant ovarioles proliferation continues until stage 10. (C) Stage 9
egg chamber double stained for tub-lacZ (red) and phosphohistone
H3 (green). Only ptc mutant follicle cells (no tub-lacZ staining)
proliferate in egg chambers beyond stage 6. (D) Stage 10 ptc mutant
egg chamber double stained for tub-lacZ (red) and hs-nlsGFP (green,
to label all nuclei). Nuclei of ptc mutant follicle cells (green) are
smaller than wild-type follicle cell nuclei (yellow).
Abnormal oocyte polarity, assessed by the position of the
oocyte nucleus, did not correlate with the germline ptc
genotype (Fig. 6H, for example). Most of the egg chambers
with altered oocyte polarity were composed entirely of ptc
mutant somatic cells. However, even in these egg chambers the
phenotype was partially penetrant (15/62=24%) (Fig. 6G).
Mosaic egg chambers only had polarity defects if the ptc
mutant clone covered the entire posterior (4/17=23%) (Fig.
6H). This is consistent with the idea that aberrant posterior
follicle cell differentiation is responsible for defective oocyte
polarity.
Two critical developmental periods for generating
ptc mutant phenotypes
To determine which of the effects of ectopic Hh signaling
might be causally related we determined the critical period for
inducing each phenotype by examining ovaries at various times
after inducing ptc mutant clones by mitotic recombination. The
non-stem cell daughter of a somatic stem cell normally divides
roughly every 10 hours for about 1.5 days in region 2b of the
germarium (3-4 divisions), less than a day in germarial region
3 (~2 divisions) and more than one day in the vitellarium (~3
Hedgehog in oogenesis 2169
Fig. 3. Ectopic border cells and oocyte polarity
defects in ptc mutant egg chambers. (A,C,E) Wild
type egg chambers. (B,D,F) ptc mutant egg
chambers. (A,B) Border cell marker 5A7 in stage 10
egg chambers shows (A) one group of cells in wildtype and (B) two groups of border cells in ptc
mutant egg chamber. (C,D) The oocyte nucleus
(arrow) has always migrated to the anterior of the
oocyte in (C) wild-type stage 9 egg chambers but
not in (D) ptc mutant ovarioles. C is also stained for
5A7 and D for neu-lacZ. (E,F) The product of the
kin-lacZ transgene normally localizes to the extreme
posterior in (E) wild-type stage 9 egg chambers, but
(F) often fails to do so in ptc mutant ovarioles.
divisions) before reaching mitotic quiescence
in a stage 7 egg chamber of 500-1000 follicle
cells (Margolis and Spradling, 1995; Spradling,
1993). Stages 7, 8 and 9 each last for
approximately 6 hours (Spradling, 1993).
Ectopic ptc-lacZ could be observed readily in
egg chambers up to and including stage 7 as
early as 2 days after ptc mutant clone induction
(Table 1). Induction of the ptc promoter is a
direct and seemingly ubiquitous response to Hh signaling in
several tissues (Ingham, 1998) and therefore is unlikely to be
developmentally restricted in the ovary. In this case, the 2-day
delay that we observed provides a rough estimate of the time
between induction of a ptc mutant clone and loss of ptc activity
by dilution and degradation of inherited gene products.
Ectopic FasIII and neu-lacZ expression, loss of 998/12
expression and follicle cell proliferation beyond stage 6 were
all first seen 4 days after clone induction and were restricted
at this time to stage 7 or younger egg chambers (Table 1). At
5 days after clone induction each of these phenotypes was seen
more frequently and in egg chambers up to stage 9. We
therefore estimate that somatic cells must lose ptc activity at
least 2 days before forming part of a stage 7 egg chamber in
order to induce this set of follicle cell phenotypes. This
corresponds to a developmental stage just prior to germarial
region 3.
Fig. 4. Anterior and posterior follicle cell determination in ptc and
PKA mutants. (A) Expression of anterior follicle cell marker L53b is
restricted to anterior cells in ptc mutant egg chambers. (B) Delta
expression (red) is reduced, as in wild-type, in posterior follicle cells
after stage 6 in ptc mutant egg chambers. In this egg chamber all
follicle cells are ptc mutant, marked by lack of tub-lacZ expression
(green). (C) In wild-type ovarioles the posterior follicle cell marker
998/12 is first visible at stage 7 and becomes strongest at stage 10
(shown here). (D) A stage 10 ptc mutant egg camber in which some
posterior cells lack 998/12 expression, and some cells ectopically
express 998/12. (E) A stage 9 ptc mutant egg chamber completely
lacking 998/12 expression, as observed whenever the oocyte nucleus
(arrow) remained aberrantly at the posterior of the oocyte. Some
normally polarized oocytes also lacked 998/12 expression.
(F) Reduced 998/12 expression in egg chambers from PKA mutant
ovarioles. (G-I) Double staining of 998/12 (red, G) and π-myc (green,
H). ptc mutant follicle cells are marked by lack of π-myc expression.
(I) Merged picture of G and H. In most ptc mutant follicle cells
998/12 expression is lost but in some weak expression is seen
(arrow).
Egg chambers with mis-placed oocytes were observed only
5 days or more after induction of ptc mutant clones (Table 1).
Stage 8 and older egg chambers were only affected if ptc
mutant clones had been induced 6 days earlier. Similarly,
defects in oocyte polarity, assessed by the position of the
oocyte nucleus at stage 8-10 were observed only if ptc mutant
clones were generated 6 or more days previously. We therefore
estimate that egg chamber polarity and oocyte polarity can only
be disrupted if somatic cells lose ptc activity at least four days
2170 Y. Zhang and D. Kalderon
Table 1. Latest developmental stage at which phenotypes
can be observed following timed induction of ptc mutant
clones
Days after heat-shock
Ectopic ptclacZ
Ectopic FasIII
Ectopic neulacZ
Ectopic mitoses
Loss of 988/12
Oocyte not at P
Mislocalization of
kin-lacZ product
Oocyte nuclei at P
1
2
3
4
5
6
-
stage 7
-
stage 7
-
stage 9
stage 7
stage 7
stage 7
stage 7
-
stage 9
stage 9
stage 9
stage 9
stage 9
stage 5
-
stage 10
stage 10
stage 10
stage 10
stage 10
stage 8
stage 9
-
-
-
-
-
stage 8
prior to forming part of a stage 8 egg chamber. This suggests
a critical action in somatic stem cells.
We infer that there are at least two developmental periods at
which inappropriate Hh signaling can disrupt oogenesis and
that phenotypes induced at the same stage may share a
common primary cause.
Two branches of hedgehog signal transduction in
ovaries
In imaginal discs and embryos, a response to Hh activity or to
mutation of ptc requires smo function and leads to changes in
the activity of the transcription factor Ci in a number of ways
(Ingham, 1998). In these tissues, Hh both inhibits partial
proteolysis of full-length Ci (Ci-155) to a repressor form, Ci75, and stimulates conversion of Ci-155 to a nuclear
transcriptional activator. Conversion of Ci-155 to Ci-75
requires protein kinase A (PKA) and probably Costal-2 (Cos2)
activity, whereas full activation of Ci-155 by Hh requires Fused
(Fu) kinase activity and is opposed by Suppressor of fused
(Su(fu)) (Ohlmeyer and Kalderon, 1998). We investigated
whether Hh signaling in oogenesis is regulated in a similar
way.
Follicle cells mutant for ptc or PKA expressed very high
levels of Ci-155 detected in situ by an antibody specific for this
form of Ci (Fig. 7B,D), indicating blocked proteolysis to Ci75 (Aza-Blanc et al., 1997; Slusarski et al., 1995). In contrast
to other tissues, ptc (but not PKA) mutant clones also induced
ectopic expression of ci RNA (data not shown) and expression
of a transcriptional reporter fusion, ci-lacZ (Schwartz et al.,
1995) (Fig. 7G).
PKA mutant somatic stem cell clones induced ectopic ptclacZ (Fig. 8A-C), polar cells (Fig. 8D) and border cells (data
not shown) but caused only mild over-proliferation of follicle
cells (Fig. 8E) and did not affect oocyte positioning or oocyte
polarity, as indicated by normal migration of the oocyte
nucleus to the dorsal anterior corner. Expression of 998/12 was
never completely lost from the posterior of PKA mutant egg
chambers but was frequently found to be mosaic or reduced in
intensity (Fig. 4F). PKA mutant ovarioles also included follicle
cells that proliferated beyond stage 6, although rarely beyond
stage 8 (data not shown). Thus, PKA mutations elicited only
the subset of ptc mutant phenotypes that were inferred to be
induced in late region 2b of the germarium.
The follicle cell over-proliferation induced by cos2 mutant
somatic stem cell clones was less than seen for ptc mutations
but greater than observed for PKA mutations (data not shown).
In a few egg chambers the oocyte was not at the posterior
(<1%) but the migration of the oocyte nucleus, indicative of
oocyte polarity, was always normal (data not shown). Ectopic
polar cells, border cells (see Liu and Montell, 1999) and Fas
III expression were all induced by cos2 mutations (data not
shown), as for ptc and PKA mutations. The expression of neulacZ in normally positioned polar cells was also delayed in
cos2 mutant ovarioles, but at lower penetrance (13/40=30%)
than in ptc mutant ovarioles (28/35=80%).
Somatic stem cell clones lacking both ptc and Fu kinase
activity (fumH63 has a point mutation in the kinase domain)
exhibited only very mild over-proliferation and only very
occasional ectopic expression of FasIII (data not shown).
Somatic cell clones deficient for both PKA and Fu kinase
activity produced similar phenotypes to PKA mutant clones;
they induced ectopic ptc-lacZ, FasIII and neu-lacZ, but
exhibited slightly greater follicle cell over-proliferation than
PKA mutant ovarioles, characteristically yielding egg
chambers with substantial multi-layering at the posterior poles
(Fig. 8H).
Loss of Su(fu) enhanced the PKA mutant phenotypes
dramatically. Somatic stem cells lacking PKA and Su(fu)
activity induced extensive follicle cell over-proliferation (Fig.
8F), mis-placed oocytes (15/52=28%) (Fig. 8G) and aberrant
oocyte polarity (4/14=28%).
The severity of over-proliferation induced by the various
clones examined can be summarized as: ptc > PKA Su(fu) >
cos2 >> PKA fu > PKA >> ptc fu. These results are consistent
with observations in other tissues that loss of PKA or cos2
cause weaker intracellular signaling in the Hh transduction
pathway than loss of ptc because only the latter mutation
overcomes the inhibitory action of Su(fu) through a mechanism
that requires Fu kinase activity (Ohlmeyer and Kalderon,
1998). Furthermore, only strong ectopic signaling in the Hh
pathway (due to ptc or PKA + Su(fu) mutations) resulted in
penetrant oocyte position or polarity defects. Weak ectopic
signaling (due to cos2, PKA fu or PKA mutations) sufficed to
alter many aspects of follicle cell behavior, although the
severity and penetrance of ptc mutant phenotypes were
generally greater.
Is Hh signaling required for patterning egg
chambers?
The cessation of somatic cell proliferation and egg chamber
budding in the absence of Hh activity has limited direct
investigation of the normal role of Hh in egg chamber
patterning. When hhts2 animals were shifted to the restrictive
temperature, budding of egg chambers diminished after about
3 days and virtually ceased by 7 days. During this period some
egg chambers budded with multiple complements of germline
cells. In these compound egg chambers an oocyte was always
seen at the posterior and FasIII was expressed in a pair of cells
at each pole (Fig. 9A). In some cases, one or two additional
pairs of FasIII-expressing cells were seen in central regions of
the compound egg chamber. When hhts2 animals were held at
the restrictive temperature for 7 days and then returned to the
permissive temperature egg chamber budding resumed within
3 days, showing no significant evidence of defects in oocyte
positioning or polar cell differentiation in the most recently
budded egg chambers (data not shown).
To try to examine the effects of blocking Hh signaling
Hedgehog in oogenesis 2171
Fig. 5. Cell autonomy of ectopic polar cell and border
cell induction. (A) ptc mutant ovariole double stained for
tub-lacZ (green) and Fas III (red). Only ptc mutant
follicle cells express ectopic Fas III. (B) ptc mutant egg
chamber double stained for neu-lacZ (green) and Fas III
(red). Ectopic neu-lacZ (solid arrowhead) is only
expressed in the nuclei of ptc mutant follicle cells
(outlined by membrane FasIII staining) that contact
germline cells, but not in all such cells (open
arrowhead). ptc mutant follicle cells that do not contact
the germline never express ectopic neu-lacZ (arrow).
Unstained regions of the egg chamber are outlined by
white dots. (C) Ectopic border cells, stained by 5A7
expression (green), closely surround an ectopic polar cell
(arrowhead), labeled by neu-lacZ (red), in a stage 10 ptc
mutant egg chamber. Wild type polar cells (arrow) and
border cells are at the border of the nurse cells and the
oocyte in the same egg chamber. The extreme anterior (left) and posterior (right) ends of this egg chamber are beyond the limits of the field shown.
without arresting egg chamber production we induced marked
smo mutant clones at various times prior to ovary dissection.
Three days after induction, smo mutant clones were found to
have proliferated like similarly marked wild-type cells and did
not induce any aberrant phenotypes. However, 8 days after
clone induction, smo mutant clones were much smaller and
were recovered at much lower frequency than control clones.
Also, roughly 10% of ovarioles included an enlarged germarial
region 3, indicative of arrested budding,
compound egg chambers (Fig. 9C), or closely
apposed egg chambers with no stalk between
them (Fig. 9B,C). The affected egg chambers
were often composed largely of marked smo
mutant cells (Fig. 9B). One possible
interpretation is that Hh signaling is required
in some somatic cells for germline cyst
encapsulation or for the process of egg
chamber budding. However, in some
compound, or partially fused, egg chambers no
Fig. 6. Oocyte positioning in ptc mutant ovarioles.
(A) Germarium structure with germline cysts
shaded. (B) The oocyte, preferentially labeled by
antibody to BicD (white), is at the posterior of the
nascent egg chamber by region 3 (indicated by
arrows) in wild-type germaria but (C) is often
mislocalized in ptc mutant germaria. (D) A mislocalized oocyte (arrow) in a stage 5 egg chamber
with an anterior ptc mutant follicle cell clone. The
germline still retains ptc activity, shown by tublacZ staining (green). (E) The oocyte (red BicD
staining) is occasionally mis-localized in egg
chambers with no ptc mutant follicle cells (green
tub-lacZ in all follicle cells). (F) Oocytes (red
BicD staining) are properly localized in some egg
chambers when adjacent to ptc mutant follicle cells
(lacking green tub-lacZ staining). (G) The oocyte
nucleus of a stage 9 egg chamber occasionally
migrates to the anterior properly, even though it is
completely surrounded by ptc mutant follicle cells
(absence of green tub-lacZ). (H) The oocyte
nucleus remains at the posterior only when a large
ptc mutant clone (no green tub-lacZ) surrounds the
oocyte.
smo mutant cells were seen (Fig. 9C). This suggests an
alternative, or additional possibility, wherein defective
proliferation of smo mutant cells results in insufficient somatic
cells to envelop and separate germline cysts, leading to delayed
budding of compound egg chambers with an abnormally high
ratio of germline to somatic cells.
In many morphologically normal ovarioles smo mutant cells
were found to contribute substantially to individual egg
2172 Y. Zhang and D. Kalderon
Fig. 7. Ci protein and ci RNA expression in ptc and PKA mutant
ovarioles. (A) In wild-type ovarioles, antibody 2A1, which
recognizes full-length Ci (Ci-155), stains strongly in the germarium
(red), but staining fades in increasingly mature egg chambers.
(B,C) In ptc mutant ovarioles and (D,E) PKA mutant ovarioles
double-stained for (B,D) Ci-155 and (C,E) ptc-lacZ, ectopic Ci-155
staining (red) is seen in the same follicle cells that express ectopic
ptc-lacZ (green). (F) ci-lacZ expression (blue) in wild-type ovarioles
is strong in the germarium but reduced and preferentially retained at
the posterior as egg chambers mature (Forbes et al., 1996b).
(G) Ectopic ci-lacZ (blue) in ptc mutant ovarioles is most obvious in
anterior regions of mature egg chambers.
chambers of normal polarity. In such egg chambers smo mutant
cells were found in a variety of positions, and sometimes
included FasIII-expressing cells with typical polar cell
morphology at the poles of the egg chamber (Fig. 9D). Thus,
Hh signaling is not required cell autonomously for positionally
appropriate specification of polar cells. We did not, however,
recover any normal egg chambers enveloped entirely by smo
mutant clones and cannot therefore exclude a role for Hh
signaling in patterning or budding of egg chambers.
Localized Hedgehog is not essential for oogenesis
To test whether Hh has any positionally instructive role in
oogenesis we induced PKA mutant stem cell clones in third
instar hhts2 larvae, which were then shifted to the restrictive
temperature for 7 days as adults. PKA mutant clones rescued
the proliferation defects of the hhts2 animals so that eight fully
developed ovarioles were recovered from about 600 defective
Fig. 8. PKA, PKA Su(fu) and PKA fu mutant ovarioles. (A) ptc-lacZ
(blue) is expressed only in the germarial region of wild-type
ovarioles. (B) Ectopic ptc-lacZ expression in PKA mutant ovarioles.
(C) Ectopic ptc-lacZ (green) is expressed ectopically only in PKA
mutant follicle cells (shown by lack of red π-myc staining).
(D) Ectopic neu-lacZ (blue) and (E) Fas III expression (red) in PKA
mutant ovarioles. (F) PKA Su(fu) double mutants induce excess
follicle cells, ectopic Fas III (red) and (G) oocyte (brown ooplasm)
mis-positioning. (H) PKA fu double mutants induce excess follicle
cells and ectopic ptc-lacZ (blue).
ovarioles dissected from 20 flies. In every case, all of the
proliferating somatic cells were mutant for PKA (Fig. 9E),
indicating that all stem cells must lose PKA activity to restore
normal somatic cell proliferation. Egg chambers in these
ovarioles showed normal Fas III staining confined to
appropriately positioned polar cells (20/20), normal oocyte
positioning (16/16) and normal oocyte nucleus migration to the
anterior cortex in mid-oogenesis (5/5). Thus, in a situation
where all proliferating somatic cells experience the same level
of Hh signal transduction (due to PKA mutation in the absence
of a source of functional Hh) all aspects of polarity that we
scored were normal. We therefore conclude that Hh has no
essential positionally instructive role in assigning polar cell
fates or directing the oocyte to the posterior of the egg chamber.
We also recovered ovarioles containing ptc mutant clones in
hhts2 animals at the restrictive temperature (Fig. 9F,G). Ectopic
polar cells, ectopic Fas III expression, mild over-proliferation
and mis-positioning of oocytes were all consistently observed
in these ovarioles. Some ovarioles included a few somatic cells
Hedgehog in oogenesis 2173
Fig. 9. Localized Hh signaling is not required for polar cell
specification and oocyte positioning. (A) A compound egg
chamber from a hhts2 fly held at the restrictive temperature for
7 days. Fas III (red) stains a pair of polar cells at each pole.
(B-D) Ovarioles from flies in which smo mutant clones
(absence of green tub-lacZ staining) were induced 8 days
earlier. (B) Several egg chambers are incompletely separated
and contain many smo mutant somatic cells. (C) A swollen
germarial region 3 (indicated by bracket), followed by a
compound egg chamber (arrow), and two incompletely
separated egg chambers that contain no smo mutant follicle
cells. (D) Polar cells are normally specified, as indicated by
Fas III antibody staining (red), even when they are part of a
smo mutant clone (arrow). (E) Egg chamber budding in hhts2
animals held for 7 days at 29°C is restored when ovarioles are
entirely composed of PKA mutant follicle cells (lacking green
tub-lacZ staining). Polar cell specification (red FasIII staining)
and oocyte positioning (small bright green nucleus) are
normal in these ovarioles. (F,G) ptc mutant follicle cells
(lacking green tub-lacZ staining) not only restore egg
chamber budding to hhts2 ovarioles at 29°C, but also produce
excess follicle cells and ectopic polar cells stained with FasIII
(red). Most of the proliferating cells in these ovarioles are
mutant for ptc but a few cells retain ptc activity (green,
marked by arrows).
that retained ptc function. Thus, ptc mutations need not
affect all somatic stem cells to rescue oogenesis and
induce ectopic Hh signaling phenotypes even in the
absence of normal Hh function.
DISCUSSION
Earlier studies of hh loss of function in oogenesis suggested an
essential role in somatic cell proliferation early in the
germarium (Forbes et al., 1996a) that is confirmed in this study
by the phenotype of smo mutant somatic cells. The induction
of ectopic polar cells and mis-positioned oocytes by ectopic
Hh signaling, had also suggested that Hh may normally
determine follicle cell fates and egg chamber polarity (Forbes
et al., 1996a). We have explored the origin of these and other
ectopic Hh signaling phenotypes and suggest that they result
from at least two primary effects. First, excessive Hh signaling
in stem cells leads to over-proliferation of somatic cells and
partially penetrant mis-positioning of the oocyte in the egg
chamber. We suggest that oocyte mis-positioning may be due
to disruption of early interactions between somatic cells and
germline cysts in the over-populated germarium. Second,
ectopic Hh signaling acts on somatic cells of cysts just prior to
region 3 of the germarium to affect multiple aspects of their
future behavior. The primary effect of ectopic Hh signaling
may be to delay the differentiation program of somatic cells in
favor of continued proliferation of a precursor state. This could
explain the tardy appearance of increased numbers of polar
cells, the protracted proliferation of other follicle cells, and the
abnormal development of posterior follicle cells, which can
lead to disruption of oocyte polarity.
Essential roles for Hh signaling in egg chamber patterning
were neither proven nor refuted by hh and smo mutant
phenotypes because of accompanying effects on cell
proliferation. We could, however, show that spatially localized
Hh signaling was not required to pattern egg chambers. We
therefore suggest that Hh signaling in the germarium is
necessary to control somatic cell proliferation and to allow
other instructive signals to position the oocyte and determine
positionally appropriate somatic cell fates.
Our studies also substantiate the model that Hh signals
through two complementary intracellular pathways and
provide insight into several regulatory mechanisms in
oogenesis. We found that anterior polar cells can induce a
border cell fate in adjacent cells, that the program of follicle
cell mitotic cycles may be determined intrinsically at the time
lineage commitments are made in the germarium, and that very
few properly differentiated posterior follicle cells can suffice
to induce re-organization of microtubule polarity in the oocyte
in mid-oogenesis.
Excessive Hh signaling affects ovarian cells at two
developmental stages
We observed three different minimal periods following ptc
mutant clone induction before specific mutant phenotypes could
be observed in stage 7-8 egg chambers. These time periods were
2 days (ectopic ptc-lacZ), 4 days (ectopic polar cells,
proliferation beyond stage 6 and loss of 998/12 expression in
posterior follicle cells) and 6 days (altered egg chamber polarity
and oocyte polarity). From these data, we argued that ptc activity
must be lost in somatic stem cells to affect egg chamber and
oocyte polarity. We also estimated that the critical time for ptc
dysfunction to induce the several follicle cell phenotypes is when
germline cysts are surrounded by 16-32 somatic cells, just prior
to germarial region 3. Such cysts take approximately 2 days to
2174 Y. Zhang and D. Kalderon
develop from a somatic stem cell, and mature into stage 7-8 egg
chambers in a further 2 days.
PKA mutations affected follicle cell proliferation beyond
stage 6, FasIII, neu-lacZ and 998/12 expression, but did not
affect egg chamber or oocyte polarity and did not induce gross
over-proliferation of somatic cells. We therefore propose that
abnormally high activity in the Hh signal transduction pathway
has two primary consequences, each leading to a subset of
phenotypes. Strong ectopic Hh signaling (induced by ptc or
PKA + Su(fu) mutations) acts in somatic stem cells to induce
over-proliferation in the germarium, together with defects in
egg chamber and oocyte polarity, whereas even weak ectopic
Hh signaling (due to PKA or cos2 mutations) alters the
behavior of somatic cells just prior to region 3 of the
germarium to disrupt future follicle cell proliferation and
differentiation.
Excess somatic cells may disrupt oocyte
positioning
The posterior position of the oocyte in the nascent egg chamber
depends on a selectively strong homophilic DE-cadherin
interaction between the oocyte and posterior somatic cells
around the germline cyst (Godt and Tepass, 1998; GonzalezReyes and St. Johnston, 1998a). It is not known how the more
posterior somatic cells of a late region 2b cyst acquire
selectively high levels of DE-cadherin, but one hypothesis is
that an anteriorly localized extracellular signal, such as Hh,
could serve to repress DE-cadherin expression selectively in
more anterior cells. However, our mosaic analyses did not
support this model. There was no correlation between oocyte
mis-positioning and the presence of ptc mutant cells at the
posterior (or any other region) of the egg chamber, mispositioned oocytes did not always contact wild-type (or ptc
mutant) cells and mis-positioned oocytes were occasionally
seen in egg chambers with no ptc mutant follicle cells in
contact with the germline cells. Furthermore, the distribution
of ptc mutant cells was not biassed toward the anterior or
posterior in egg chambers of normal polarity, implying that
ectopic Hh signaling does not affect selective affinity for the
oocyte relative to nurse cells.
Several mutations disrupt egg chamber polarity by delaying
differentiation of the oocyte (Ghabrial et al., 1998; GonzalezReyes et al., 1997), demonstrating that interactions with
follicle cells must take place in a very restricted time frame. In
the over-populated germaria of ptc mutant ovarioles it seems
quite likely that the time at which somatic cells initially make
contact with germline cysts is altered. Also, since most ptc
mutant cells eventually populate the regions between egg
chambers it is possible that many ptc mutant cells first contact
the germline cyst and are then replaced by other cells, thereby
delaying stable association between germline and somatic
cells. Thus, we suggest that oocyte mis-positioning in ptc
mutant ovarioles depends on altered timing, stability or
geometry of the earliest interactions between somatic and
germline cells. Over-proliferation of ptc mutant somatic cells
may be directly causative or may exacerbate the effects of
changes in the adhesive qualities of ptc mutant cells.
Ectopic Hedgehog signaling delays developmental
programs of follicle cell lineages
Is there a causal connection between the various follicle cell
phenotypes that can be induced by ptc or PKA mutations with
a minimal latency of about 4 days? It has previously been
argued that precursors for stalk cells and polar cells separate
from the general follicle cell lineage prior to germarial region
3 and that excessive Hh signaling can maintain cells of the
polar cell/stalk cell lineage in a precursor state for an
abnormally long period of time (Margolis and Spradling, 1995;
Tworoger et al., 1999). The delayed appearance of neu-lacZ in
polar cells in ptc mutant ovarioles supports this argument.
Delayed specification of polar cells would permit more
proliferation than usual in this lineage, and might explain the
large clusters of polar cells that we frequently observed around
either pole of egg chambers. However, polar cells were also
seen singly or in small groups at ectopic positions, suggesting
that ectopic Hh signaling can also interfere with the normal
assignment of cells to the polar cell/stalk cell lineage
(Tworoger et al., 1999).
Several other ptc mutant phenotypes could be explained if
we propose that excessive Hh signaling delays the
developmental programs of all somatic cells. Continued FasIII
expression in most follicle cells beyond stage 1 (up to stage 4)
in ptc mutant ovarioles provides some evidence for this
proposal. If we suppose that follicle cells normally undergo an
intrinsically fixed number of divisions following allocation to
a non-polar cell/ stalk cell lineage, delayed initiation of this
developmental program would result in follicle cell divisions
beyond stage 6, as observed. This model can account for the
otherwise puzzling observation that proliferation of ptc mutant
cells in stage 7-9 egg chambers depends on loss of ptc activity
much earlier in the germarium. The time at which follicle cells
cease general DNA replication is not altered by ptc mutations
and therefore does not appear to be regulated in the same way
as withdrawal from mitotic cell cycles.
The inability of posterior ptc mutant follicle cells to
differentiate normally in mid-oogenesis also resulted from ptc
inactivation in the germarium. Posterior follicle cells normally
express 998/12 from stage 7 onward in response to prior
activation of the EGF receptor by Grk protein concentrated at
the posterior of the oocyte (van Eeden and St. Johnston, 1999).
Posterior follicle cells do not express posterior markers like
998/12 if components of Grk-EGFR signaling are disrupted or
if the follicle cells are not competent to respond because of
prior Notch or Delta inactivation (Gonzalez-Reyes and St.
Johnston, 1998b; Larkin et al., 1999). In the former case,
posterior cells express anterior cell markers (van Eeden and St.
Johnston, 1999). In the latter case neither type of terminal cell
marker is expressed and posterior cells also fail to repress Delta
expression at stage 6 and beyond (Gonzalez-Reyes and St.
Johnston, 1998b; Larkin et al., 1999). Although posterior ptc
mutant cells did not express posterior markers normally,
anterior markers were not ecopically activated and Delta was
repressed normally. This unprecedented defect suggests that
ptc mutant cells behave like terminal cells and can respond to
grk, but in a way that is compromised by their developmental
competence. This may result from the postulated
developmental immaturity of ptc mutant cells.
Disruption of oocyte polarity requires extensive misspecification of posterior follicle cells
Oocyte polarity defects were observed only in egg chambers
where 998/12 expression was completely absent from the
Hedgehog in oogenesis 2175
posterior, and only in egg chambers where all, or the vast
majority of posterior cells lacked ptc activity. This suggests
that only a small number of posterior cells are required to
deliver the signal to re-polarize the oocyte cytoskeleton and
that these cells are not constrained to occupy a specific position
within the posterior terminus. Only ptc mutant clones induced
at least 6 days previously produced over-proliferation of ptc
mutant cells sufficient for them to occupy the entire posterior
of the follicular epithelium. Thus, both cell autonomous early
over-proliferation and aberrant differentiation of posterior
follicle cells were required for ptc mutant cells to disrupt
oocyte polarity. PKA mutant somatic stem cells did not cause
massive somatic cell over-proliferation, never eliminated
998/12 expression from all posterior cells and never disrupted
anterior migration of the oocyte nucleus.
Hedgehog signaling pathway in oogenesis
Hh signaling in Drosophila generally regulates the abundance
and activity of Ci proteins without altering ci mRNA levels
(Ingham, 1998). By contrast, vertebrate Hh homologs
frequently regulate transcription of the Ci-related GLI family
of transcriptional effectors (Ruiz i Altaba, 1999). The induction
of ci RNA in ptc mutant follicle cells provides the first evidence
that this circuitry can also be found in Drosophila.
Other consequences of altering the activity of Hh signaling
components in ovarian somatic cells substantiate the
hypothesis that Hh signaling activates at least two distinct
intracellular pathways (Methot and Basler, 1999; Ohlmeyer
and Kalderon, 1998). One pathway, involving protection of Ci155 from proteolysis and perhaps also release from
cytoplasmic anchoring, is phenocopied by PKA and cos2
mutations. In the ovary, cos2 mutations elicited stronger
phenotypes than PKA mutations, perhaps because cos2
mutations preferentially disrupt cytoplasmic anchoring of Ci155. The second pathway increases the specific activity of Ci155 in opposition to the inhibitory effects of Su(fu) (Ohlmeyer
and Kalderon, 1998). This pathway is elicited by ptc, but not
by PKA mutations and requires Fu kinase activity. In
accordance with this model, PKA Su(fu) double mutant cells
produced phenotypes almost as strong as for ptc mutants in
ovaries, whereas ptc fu double mutant cells exhibited minimal
phenotypes and PKA mutant phenotypes were not greatly
altered by additional loss of Fu kinase activity.
In imaginal discs high level Hh signaling to nearby cells is
phenocopied by ptc mutations and requires Fu kinase activity,
whereas only low level Hh signaling to more distant cells can
be phenocopied by PKA mutations and does not require Fu
kinase activity. We found that PKA mutations in somatic
ovarian cells can effectively substitute for Hh activity, that Fu
kinase activity is not essential for somatic cell proliferation (Y.
Z. unpublished data) and that ptc mutations engender excessive
Hh signaling phenotypes even in the absence of Hh activity.
Hence, we surmise that ovarian somatic cells normally undergo
only low levels of Hh signaling, in keeping with the
observation that the source of Hh in the germarium is separated
from its target cells by several cell diameters (Forbes et al.,
1996a).
Requirements for Hh signaling in oogenesis
The rescue of apparently normal oogenesis in hhts animals at
the restrictive temperature by PKA mutations in somatic stem
cells implies that there is no essential role for spatially graded
Hh levels in the germarium. However, the level of Hh signaling
must fall within certain bounds for oogenesis to proceed
normally. Normal rates of somatic cell proliferation require
some Hh signaling but also require that Ptc limits Hh signaling.
We have argued that Ptc must also restrain Hh signaling in
order to allow somatic cells to enter the developmental
program appropriate to their lineage in a timely fashion.
It is not clear at this stage whether Hh signaling has any
essential function in oogenesis other than stimulating cell
proliferation. On one hand, normal egg chambers can include
smo mutant cells in a variety of positions. In particular, polar
cells can form in normal numbers and at the correct position
from within a group of smo mutant cells, which are presumed
to be unable to transduce any Hh signal (Ingham, 1998). On
the other hand, in smo mutant ovarioles, egg chamber budding
is sometimes arrested or defective, and we have never observed
normal egg chambers completely enveloped by smo mutant
follicle cells. These phenotypes might derive solely from an
insufficient supply of somatic cells, resulting directly from
impaired proliferation of smo mutant cells. However, we
cannot dismiss the possibility that Hh signaling has a more
direct role in germline cyst encapsulation, promoting egg
chamber budding, or delaying somatic cell lineage decisions
until the appropriate developmental stage.
We thank Y. Chen, R. Holmgren, T. Kornberg, R. Lehmann, M.
Muskavitch, N. Patel, T. Schupbach, D. St. Johnston, and the
Drosophila Stock Center for generous gifts of reagents, J. Erickson,
J. Mohler, M.A. Price and K. Chung for comments on the manuscript
and K. Chung for excellent technical support. This work was
supported by NIH grant GM41815 to D. K.
REFERENCES
Aza-Blanc, P., Ramirez-Weber, F.-A., Laget, M.-P., Schwartz, C. and
Kornberg, T. B. (1997). Proteolysis that can be inhibited by Hedgehog
directs Cubitus interruptus protein to the nucleus and changes its activity as
a transcriptional regulator. Cell 89, 1043-1053.
Chen, Y. and Struhl, G. (1998). In vivo evidence that Patched and
Smoothened constitute distinct binding and transducing components of a
Hedgehog receptor complex. Development 125, 4943-4948.
Clark, I., Giniger, E., Ruohola-Baker, H., Jan, L. Y. and Jan, Y. N.
(1994). Transient posterior localization of a kinesin fusion protein
reflects anteroposterior polarity of the Drosophila oocyte. Curr. Biol. 4,
289-300.
Cooley, L., Verheyen, E. and Ayers, K. (1992). chickadee encodes a profilin
required for intercellular cytoplasm transport during Drosophila oogenesis.
Cell 69, 173-184.
de Cuevas, M., Lilly, M. A. and Spradling, A. C. (1997). Germline cyst
formation in Drosophila. Annu. Rev. Genet. 31, 405-428.
Deng, W. and Lin, H. (1997). Spectrosomes and fusomes anchor mitotic
spindles during asymmetric germ cell divisions and facilitate the formation
of a polarized microtubule array for oocyte specification in Drosophila. Dev.
Biol. 189, 79-94.
Forbes, A. J., Lin, H., Ingham, P. W. and Spradling, A. C. (1996a).
Hedgehog is required for the proliferation and specification of ovarian
somatic cells prior to egg chamber formation in Drosophila. Development
122, 1125-1135.
Forbes, A. J., Spradling, A. C., Ingham, P. W. and Lin, H. (1996b). The
role of segment polarity genes during early oogenesis in Drosophila.
Development 122, 3283-3294.
Ghabrial, A., Ray, R. P. and Schupbach, T. (1998). okra and spindle-B
encode components of the RAD52 DNA repair pathway and affect meiosis
and patterning in Drosophila oogenesis. Genes Dev. 12, 2711-2723.
Godt, D. and Laski, F. A. (1995). Mechanisms of cell rearrangement and cell
2176 Y. Zhang and D. Kalderon
recruitment in Drosophila ovary morphogenesis and the requirement of bric
a brac. Development 121, 173-187.
Godt, D. and Tepass, U. (1998). Drosophila oocyte localization is mediated
by differential cadherin-based adhesion. Nature 395, 387-391.
Gonzalez-Reyes, A., Elliott, H. and St. Johnston, D. (1997). Oocyte
determination and the origin of polarity in Drosophila: the role of the spindle
genes. Development 124, 4927-4937.
Gonzalez-Reyes, A. and St. Johnston, D. (1998a). The Drosophila AP axis
is polarized by the cadherin-mediated positioning of the oocyte.
Development 125, 3635-3644.
Gonzalez-Reyes, A. and St. Johnston, D. (1998b). Patterning of the follicle
cell epithelium along the anterior-posterior axis during Drosophila
oogenesis. Development 125, 2837-2846.
Harrison, D. and Perrimon, N. (1993). A simple and efficient method for
generation of marked clones in Drosophila. Curr. Biol. 3, 424-433.
Ingham, P. W. (1998). Transducing Hedgehog: the story so far. EMBO J. 17,
3505-2511.
Larkin, M. K., Deng, W.-M., Holder, K., Tworoger, M., Clegg, N. and
Ruohola-Baker, H. (1999). Role of Notch pathway in terminal follicle cell
differentiation during Drosophila oogenesis. Dev. Genes Evol. 209, 301-311.
Lee, T. and Montell, D. J. (1997). Multiple Ras signals pattern the Drosophila
ovarian follicle cells. Dev. Biol. 185, 25-33.
Lin, H. (1997). The tao of stem cells in the germline. Annu. Rev. Genet. 31,
455-491.
Liu, Y. and Montell, D. J. (1999). Identification of mutations that cause cell
migration defects in mosaic clones. Development 126, 1869-1878.
Ma, C. Y., Zhou, Y., Beachy, P. A. and Moses, K. (1993). The segment
polarity gene hedgehog is required for progression of the morphogenetic
furrow in the developing Drosophila eye. Cell 75, 927-938.
Margolis, J. and Spradling, A. C. (1995). Identification and behavior of
epithelial stem cells in the Drosophila ovary. Development 121, 3797-3807.
Methot, N. and Basler, K. (1999). Hedgehog controls limb development by
regulating the activities of distinct transcriptional activator and repressor
forms of Cubitus interruptus. Cell 96, 819-831.
Ohlmeyer, J. T. and Kalderon, D. (1998). Hedgehog stimulates maturation
of Cubitus interruptus into a labile transcriptional activator. Nature 396, 749753.
Royzman, I. and Orr-Weaver, T. L. (1998). S phase and differential DNA
replication during Drosophila oogenesis. Genes to Cells 3, 767-776.
Ruiz i Altaba, A. (1999). Gli proteins and Hedgehog signaling. Trends Genet.
15, 418-425.
Schwartz, C., Locke, J., Nishida, C. and Kornberg, T. B. (1995). Analysis
of cubitus interruptus regulation in Drosophila embryos and imaginal discs.
Development 121, 1625-1635.
Sisson, J. C., Ho, K. S., Suyama, K. and Scott, M. P. (1997). Costal2, a
novel kinesin-related protein in the hedgehog signaling pathway. Cell 90,
235-245.
Slusarski, D. C., Motzny, C. K. and Holmgren, R. (1995). Mutations that
alter the timing and pattern of cubitus interruptus gene expression in
Drosophila melanogaster. Genetics 139, 229-240.
Spradling, A. C. (1993). Developmental genetics of oogenesis. In The
Development of Drosophila melanogaster. (ed. M. Bate, A. Martinez Arias)
pp1-70. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory
Press.
Suter, B. and Steward, R. (1991). Requirement for phosphorylation and
localization of the Bicaudal-D protein in Drosophila oocyte differentiation.
Cell 67, 917-926.
Tworoger, M., Larkin, M. K., Bryant, Z. and Ruohola-Baker, H. (1999).
Mosaic analysis in the Drosophila ovary reveals a common Hedgehoginducible precursor stage for stalk and polar cells. Genetics 151, 739-748.
van Eeden, F. and St. Johnston, D. (1999). The polarization of the anteriorposterior and dorsal-ventral axes during Drosophila oogenesis. Curr. Op.
Genet. Dev. 9, 396-404.
Wei, Y., Yu, L., Bowen, J., Gorovsky, M. A. and Allis, C. D. (1999).
Phosphorylation of histone H3 is required for proper chromosome
condensation and segregation. Cell 97, 99-109.
Xu, T. and Rubin, G. M. (1993). Analysis of genetic mosaics in developing
and adult Drosophila tissues. Development 117, 1223-1237.