STEM CELLS AND REGENERATION
RESEARCH ARTICLE
brinker
levels regulated by a promoter proximal element support
germ cell homeostasis
Leslie Dunipace*, Susan Newcomb* and Angelike Stathopoulos
‡
ABSTRACT
A limited BMP signaling range in the stem cell niche of the ovary
protects against germ cell tumors and promotes germ cell
homeostasis. The canonical repressor of BMP signaling in both the
Drosophila
embryo and wing disc is the transcription factor Brinker
(Brk), yet the expression and potential role of Brk in the germarium
has not previously been described. Here, we find that
brk
expression
requires a promoter-proximal element (PPE) to support long-distance
enhancer action as well as to drive expression in the germarium.
Furthermore, PPE subdomains have different activities; in particular,
the proximal portion acts as a damper to regulate
brk
levels precisely.
Using PPE mutants as well as tissue-specific RNA interference and
overexpression, we show that altering
brk
expression within either the
soma or the germline affects germ cell homeostasis. Remarkably, we
find that Decapentaplegic (Dpp), the main BMP ligand and canonical
antagonist of Brk, is upregulated by Brk in the escort cells of the
germarium, demonstrating that Brk can positively regulate this
pathway.
KEY WORDS:
brinker
, Promoter-proximal element, BMP signaling,
Oogenesis, Niche, Germline stem cells, Undifferentiated germ cells,
Drosophila melanogaster
INTRODUCTION
Maintenance of germline stem cell (GSC) homeostasis is regulated
by numerous pathways that signal between the germline and somatic
cells that comprise the stem cell niche, as well as by other external
and long-range signals (Nelson et al., 2019; Zhang and Cai, 2020).
In the
Drosophila melanogaster
model system, the ovary, in which
the oocyte develops into a mature egg, contains about fifteen
ovarioles composed of germline and somatic cells (Fig. 1B,B
′
). At
the anterior-most tip of each ovariole lies the germarium; a tapered
structure made up of several distinct cell types that support the
differentiation of one GSC daughter into a cystoblast (CB) and
maintenance of the GSC lineage by the other daughter (Fig. 1C,C
′
).
The anterior-most region of the germarium contains the stem cell
niche, which comprises three somatic cell populations
–
the terminal
filament (TF), cap cells (CCs) and an anterior subset of escort cells
[ECs, alternatively inner germarial sheath (IGS) cells]
–
and
supports the maintenance of two or three GSCs throughout
adulthood (Fig. 1C; Liu et al., 2015; Tan et al., 2018; Xie and
Spradling, 2000). ECs located more posteriorly influence the
differentiation of stem cells, forming what is considered the
‘
differentiation niche
’
(Kirilly et al., 2011). GSCs produce
cystoblasts via asymmetric division aligned along the anterior-
posterior axis of the germarium such that daughter cells that move
out of the niche escape the self-renewal signal and begin to
differentiate, whereas those that remain in contact with the CCs are
maintained as GSCs (de Cuevas and Spradling, 1998; Deng and
Lin, 1997). This is a complex but well-studied phenomenon that
requires the precise localization and interaction of a number of cell
signaling pathways (reviewed by Harris and Ashe, 2011; Hayashi
et al., 2020; Hsu et al., 2019).
The most pivotal cue is arguably extracellular Decapentaplegic
(Dpp), the main
Drosophila
BMP ligand, which is expressed at
high levels in and secreted from CCs to promote self-renewal of
GSCs within approximately one cell diameter (Eliazer and
Buszczak, 2011). In many tissues, Dpp functions as a long-range
morphogen to direct cell fate decisions in a concentration-dependent
manner across tissues; however, in its role in GSC maintenance,
the range of Dpp is limited by the expression of receptors and
extracellular matrix components in the niche that serve as a sink
for extracellular ligand (Guo and Wang, 2009; Liu et al., 2015;
Wang et al., 2008; Wilcockson and Ashe, 2019). This limited
signaling range is crucial for proper germline development as
GSC division pushes daughter cells out of the range of Dpp and
permits differentiation factors, such as the gene
bag of marbles
(
bam
), to be expressed (Song et al., 2004). It was also recently
shown that
dpp
is expressed at low levels in ECs to maintain a
population of partially differentiated germline cells that can de-
differentiate to repopulate the germarium (Liu et al., 2015); a result
suggesting that not only the presence of signals, but also their
expression levels, can be interpreted by the germline to affect
cell fate.
The transcription factor Brinker (Brk) encodes a canonical
repressor of Bmp signaling and has been demonstrated to repress
expression of both
dpp
(Hasson et al., 2001; Theisen et al., 2007)
and Dpp-dependent target genes (Rushlow et al., 2001;
Sivasankaran et al., 2000). Inversely, BMP signaling activates a
complex that directly represses expression of
brk
(Marty et al.,
2000). As a result of this mutual repression,
brk
is typically
expressed in an obverse pattern to
dpp
, for example in the embryo
(Jaz
́
win
́
ska et al., 1999a) and wing imaginal disc (Campbell
and Tomlinson, 1999; Minami et al., 1999), which, in addition to
other mechanisms, helps shape the Dpp gradient (Affolter et al.,
2001; Müller et al., 2003; O
’
Connor et al., 2006). Brk acts
similarly in the ovaries starting at stage 8 of oogenesis when it is
important for establishing the anterior-posterior gradient of Dpp
expression, which patterns the eggshell and is essential for dorsal
appendage and operculum formation (Chen and Schüpbach, 2006).
Handling Editor: Cassandra Extavour
Received 10 June 2021; Accepted 22 December 2021
Division of Biology, California Institute of Technology, 1200 East California
Boulevard, MC114-96, Pasadena, CA 91125, USA.
*These authors contributed equally to this work
‡
Author for correspondence (angelike@caltech.edu)
A.S., 0000-0001-6597-2036
This is an Open Access article distributed under the terms of the Creative Commons Attribution
License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use,
distribution and reproduction in any medium provided that the original work is properly attributed.
1
© 2022. Published by The Company of Biologists Ltd
|
Development (2022) 149, dev199890. doi:10.1242/dev.199890
DEVELOPMENT
Despite extensive studies of Dpp expression in the germarium,
however, the role of Brk in this tissue has not been previously
described.
Here, we show that, not only is
brk
expressed in the
germarium, but its expression coincides with and, remarkably,
positively regulates
dpp
expression in somatic cells. We also
demonstrate that a previously described promoter-proximal
element (PPE), first characterized as supporting distal enhancer
action in the early embryo (Dunipace et al., 2013), also has a role in
the ovary.
RESULTS
The
brk
PPE regulates expression in multiple ovarian tissues
and is required for distal enhancer action
To examine the role of the PPE in supporting ovary
brk
expression,
we first used a set of large reporters in which the
brk
coding
sequence is replaced with
gfp
in the context of
∼
30 kb of flanking
sequence (
brkNFgfp
). We examined wild-type reporter expression
as well as that of reporters with deletions of the full-length PPE
(
PPE2kb
) or its distal or proximal subdomains (
PPEdist
and
PPEprox
, respectively) (Fig. 1A; Dunipace et al., 2013). These
Fig. 1.
brk
is expressed in many cell types in the
adult ovary and its expression depends on a
PPE.
(A) Chromosomal locations of the
brk
gene,
PPE (and subunits, red) and
brkB
cis-regulatory
module (CRM) (blue). Green line diagrams
represent gene regions used in large GFP
reporters. (B,B
′
) Schematic of
Drosophila
ovary
and ovariole. (C,C
′
) Schematics of germline stem
cell niche, germarium and posterior half of stage
9/10 follicle. (D-K
′
)
brk
expression as visualized by
brkNFgfp
shows reporter expression in several
somatic cell populations in the germarium
(TF, CCs, ECs and FSCs) as well as in developing
egg chambers (CFCs, BCs and PCs). Dotted line
delineates the outer edge of the ovarian tissue,
arrows in E-K indicate the border and follicle cell
cluster. Scale bars: 20
μ
m. In this and all other
figures, anterior is to the left.
2
STEM CELLS AND REGENERATION
Development (2022) 149, dev199890. doi:10.1242/dev.199890
DEVELOPMENT
reporters were used previously in the embryo to show that the PPE
does not itself drive expression but instead serves to facilitate the
action of other enhancers located at a distance (Dunipace et al.,
2013). To identify the cell types expressing
brk
reporters, ovaries
were co-stained with antibodies for Traffic jam (Tj) to mark all
follicle cells except for the TFs, Lamin C (LamC) to mark CCs and
TF, and/or
α
-Spectrin (Spec), which outlines all later-stage follicle
cells as well as marking both spectrosomes (a specialized rounded
organelle found in GSCs and cystoblasts) and fusomes (found on
differentiating cysts) (de Cuevas et al., 1996; Li et al., 2003; Xie and
Spradling, 2000). In the germarium,
brkNFgfp
was expressed in a
number of cell types: TF, CCs, ECs and follicle stem cells (FSCs)
(Fig. 1D-D
′′
). Expression persisted in follicle cells throughout egg
chamber development (Fig. S1A), becoming restricted after stage 7
to only the columnar follicle cells (CFCs) and the polar and border
follicle cells (PCs, BCs) (Fig. 1C
′′
,E,E
′
). Deletion of
PPE2kb
from
brkNFgfp
(
brkNFgfp-
Δ
PPE2kb
) abolished GFP reporter expression
in the ovary, except for some low level expression in the FSC region
(Fig. 1F-G
′
), indicating that the PPE is required for all
brk
expression in this tissue.
Previously, we found that the PPE could be divided into subunits
that were largely redundant in their ability to support distal enhancer
action in the early embryo (Dunipace et al., 2013). We therefore
examined expression of the
brkNFgfp
reporter in which either the
distal or proximal PPE domain is deleted (
brkNFgfp-
Δ
PPEdist
and
brkNFgfp-
Δ
PPEprox
, respectively). In the germarium,
PPEdist
and
PPEprox
appeared to be generally redundant in their ability to
support
brkNFgfp
reporter expression (Fig. 1H-H
′′
,J-J
′′
). Similarly,
each domain appeared to be sufficient to support CFC reporter
expression (Fig. 1I,K). However, in
brkNFgfp-
Δ
PPEdist
specifically, GFP signal was lost in many follicle cells associated
with mid-stage egg chambers (Fig. S1B), including stage 10 BCs/
PCs (Fig. 1I
′
), indicating a distinct requirement for that element to
support expression in a subset of cells.
In order to explore further the sufficiency of the PPE to drive
expression in the ovary, we created direct-fusion nuclear-localized
RFP transgenic reporters of the PPE and its distal and proximal
subdomains (
PPE2kb>
,
PPEdist>
and
PPEprox>NLS-mCherry
,
respectively; Fig. 1A; Table S1; see Materials and Methods). These
small transgenic reporters revealed that the PPE acts as a traditional
enhancer in the ovary as it is capable of driving expression in
multiple cell types (Fig. 2A-F
′
). The
PPE2kb
reporter is active in
most of the somatic tissues in the niche: the TF, CCs and ECs
(Fig. 2A-A
′′
), but was not detectably expressed in the germline,
either in the germarium or in later-stage nurse cells (Fig. 2A,B). In
stage 10 egg chambers,
PPE2kb
drove expression in the BCs and
PCs but only weakly in the CFCs (Fig. 2B,B
′
).
When we examined direct reporter expression driven by PPE
subdomains, we observed significant differences in levels of
activity. The nuclear localization of these reporters permitted
quantification of differences in levels of expression that were not
readily observable from the cytoplasmic
brkNFgfp
signal,
especially in ECs because these cells form extensive protrusions
into the germline (Kirilly et al., 2011). Expression levels of small
reporters were quantified in accessible cell types, the ECs and CCs
in the germarium, but excluding TFs, which were variable as a result
of mounting (Fig. 2G,H; Table S2; see Materials and Methods). In
the germarium,
PPEdist
drove higher levels of expression in CCs
and ECs than the full
PPE2kb
(Fig. 2C-C
′′
,G,H). Inversely,
PPEprox
drove low level expression in CCs and a subset of ECs
(Fig. 2E-E
′′
,G,H). In late-stage egg chambers,
PPEdist
was
expressed in both BCs and PCs, but drove little to no CFC
expression (Fig. 2D,D
′
). By contrast,
PPEprox
supported weak
expression in BC and CFCs, but did not support expression in PCs
(Fig. 2F,F
′
). Finally,
PPEdist
was active in the late-stage germline
where
brkNFgfp
and other
brk
reporters were not detected
(Fig. 2D,D
′
; Fig. S1E). These results indicate that both halves of
the PPE can drive reporter expression in both CCs and ECs in the
germarium, but at different levels (distal high, proximal low) and
that wild-type expression levels (i.e. those of the full-length
PPE2kb
) require both halves. In late-stage egg chambers, the two
PPE domains act more additively with each supporting subsets of the
full expression pattern, except in the case of the germline expression
of
PPEdist
, which is repressed in the context of the full
PPE2kb
.
To provide insight into PPE function in CFCs, we examined a
previously described distal
brk
cis-regulatory module (CRM),
brkB
(Charbonnier et al., 2015), which is not active in the germarium or
mid-stage egg chambers (Fig. 2I-J
′
) but does drive strong
expression exclusively in the CFCs (Fig. 2K,K
′
). The fact that the
PPE itself is not a strong CFC driver (Fig. 2B), but
brkNFgfp
reporter expression is lost upon PPE deletion (Fig. 1G), indicates
that the PPE is required for
brkB
activity in CFCs. Also, like the
redundancy previously noted in the early embryo, neither distal
nor proximal deletion affects PPE reporter expression in CFCs
(Fig. 1I,K), indicating that either region is sufficient to support the
action of distal enhancers, such as
brkB
. Taken together, this
reporter analysis suggests that in the ovary the PPE has two
functions: to facilitate the action of other enhancers and to serve as a
direct driver of
brk
expression (Fig. 2L).
brk
PPE supports maintenance of germline homeostasis
The fact that we observe
brk
expression in cells that comprise the
germline stem cell niche (i.e. TF, CCs and ECs) suggests that Brk
plays a role in regulating germline homeostasis. To test this, we
generated deletions of the PPE and its distal and proximal
subdomains in the context of the endogenous
brk
locus using
CRISPR-Cas9 genome editing (
Δ
PPE2kb
,
Δ
PPEdist
and
Δ
PPEprox
, respectively; see Materials and Methods, Tables S1
and S3). We also deleted
brkB
(Charbonnier et al., 2015) in the same
manner (
Δ
brkB
)
.
PPE, but not
brkB
, deletions had significant effects
on germarium morphology, including germline differentiation,
spectrosome number and distribution, and the overall organization
of the germline, as well as expression pattern and level of Bam,
which marks differentiating cystoblasts (Fig. 3). Specifically,
whereas wild-type germaria contained two or three GSCs, which
present rounded spectrosomes, contact the CCs, and can be labeled
by phosphorylated Mothers Against Decapentaplegic (pMad)
antibody staining (Fig. 3A,U; Song et al., 2004), germaria from
PPE mutant females consistently contained significantly more
pMad
+
cells (Fig. 3E,I,M,U). This pMad
+
cell population likely
contains true GSCs (in contact with CCs) as well as dysregulated
cystoblasts (located in proximity to, but not directly contacting
the CCs) and will be referred to collectively hereafter as pMad
+
cells. Counterintuitively, this increase in pMad
+
cell number
occurred in all PPE mutants (i.e.
Δ
PPE2kb
,
Δ
PPEdist
and
Δ
PPEprox
), despite the fact that these deletions had varying
effects on
brk
reporter expression levels (Fig. 2G,H). Furthermore,
pMad
+
cell number was unaffected in
Δ
brkB
germaria (Fig. 3Q,U),
indicating that this is a PPE-specific effect. We also observed
that the number of rounded spectrosomes, which mark GSCs
and cystoblasts, increased correspondingly with pMad
+
cell
number, confirming that these changes represent a delay in the
differentiation of pMad
+
cells and cystoblasts into more mature
cysts (Fig. 3B,B
′
,F,F
′
,J,J
′
,N,N
′
,R,R
′
).
3
STEM CELLS AND REGENERATION
Development (2022) 149, dev199890. doi:10.1242/dev.199890
DEVELOPMENT
These observations of effects on germline homeostasis in the
niche are further supported by our findings of corresponding
changes in overall morphology of the germline in mutant germaria.
The germline in wild-type germaria can be divided into
morphological regions whereby region 1 contains the GSCs,
cystoblasts and 2- to 8-cell cysts; region 2a contains the 16-cell
cysts and 2b the same cysts once they have adopted an elongated,
lens-like shape that spans the width of the germarium; and region 3
contains a spherical cluster comprising 15 nurse cells and one
oocyte completely enclosed by follicle cells (see Fig. 1C
′
;
King, 1970; McKearin and Ohlstein, 1995). PPE mutants showed
aberrant germline morphology with tumorous expansion of regions
1 and 2a with region 2b sometimes affected (Fig. 3H,L,P compared
with 3D). Cyst organization appeared to recover by region 3,
which was structured normally in nearly all samples. Germline
organization appeared normal in
Δ
brkB
germaria (Fig. 3T).
Strikingly,
Δ
PPEprox
ovaries also lacked detectable Bam
expression (Fig. 3O) whereas Bam was present in all other
Fig. 2.
brk
PPE drives expression in the germarium and is
required to facilitate expression of distal CRMs in later-
stage egg-chambers
.
(A-F
′
) Transgenic reporters of
brk
PPE
and its distal and proximal subdomains driving nuclear
mCherry show expression in subsets of
brk
-expressing cell
types. White arrows indicate BC/PC cluster, white dotted circle
indicates posterior PC region. Yellow arrowhead indicates
nurse cell nuclei that express the reporter whereas nurse cell
nuclei that do not express reporter are outlined in yellow. Blue
dashed rectangles indicate the area shown in the insets
(B,D,F). (G,H) Quantification of PPE reporter expression levels
in escort cells of the germarium (G) or in cap cells (H).
One-way ANOVA was used for statistical comparison of each
dataset to
Δ
PPE2kb
(see Materials and Methods).
n
=number
of nuclei error bars represent mean±s.d. (see Table S2).
(I-K
′
) Expression of nuclear
brkB-GFP
reporter in the ovariole.
Dotted lines delineate the outer edge of the ovarian tissue.
Scale bars: 20 μm. (L) Schematic illustrating the bimodal role
of the PPE in supporting CRM-driven
brk
expression in the
ovary (not drawn to scale).
4
STEM CELLS AND REGENERATION
Development (2022) 149, dev199890. doi:10.1242/dev.199890
DEVELOPMENT