![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 12, No. 17, pp. 2724-2734, September 1, 1998
1 Department of Molecular Genetics, National Institute of Neuroscience, National Center of Neurology and Psychiatry, Kodaira, Tokyo 187-8502, Japan; 2 Precursory Research for Embryonic Science and Technology (PRESTO) and 3 Core Research for Evolutional Science and Technology (CREST) of Japan Science and Technology Corporation (JST); 4 Institute for Molecular and Cellular Biology, Osaka University, Suita, Osaka 565-0871, Japan; 5 Department of Developmental Neurobiology, Institute of Development, Aging and Cancer, Tohoku University, Aoba-ku, Sendai 980-8575, Japan
 |
Abstract |
|---|
Top
Abstract
Introduction
Results
Discussion
Materials & Methods
References
|
|---|
Morphogen gradients of secreted molecules play critical roles in the establishment of the spatial pattern of gene expression. During midgut development in Drosophila, secreted molecules of Decapentaplegic (Dpp) and Wingless (Wg) establish unique transcriptional regulation within target cells to specify the resultant cell types. Here we report the identification of a novel homeobox gene, defective proventriculus (dve), which is required for the midgut specification under the control of Dpp and Wg. In dve mutants, two distinct parts of the midgut, the proventriculus and middle midgut, are abnormally organized. The Wg signal regulates dve expression during proventriculus development. On the other hand, dve is a downstream target of Dpp in the middle midgut and defines the functional specificity of copper cells along with another Dpp target gene, labial. Thus, the dve gene acts under the two distinct extracellular signals at distant parts of the midgut primordia.
[Key Words: Homeodomain; Dpp; Wg; midgut; functional specificity]
| |
Introduction |
|---|
|
Top
Abstract
Introduction
Results
Discussion
Materials & Methods
References
|
|---|
During animal development, a mass of homogeneous
cells have distinct developmental fates depending on their positions,
and are organized in a stereotyped manner into a variety of functional tissues. Positional information mediated by extracellular signals plays
major roles in giving rise to such a diversity of cell types within the
uniform cell mass. The transforming growth factor-
(TGF-) and
Wnt superfamilies (Drosophila Dpp and Wg, respectively) have
been well studied among secreted molecules that transmit such
extracellular signals, and regulate a wide variety of cellular responses including differentiation, proliferation, adhesion, and
migration (for review, see Nusse and Varmus 1992
; Kingsley 1994).
Signaling by the TGF- superfamily is elicited by way of two types
of receptor serine-threonine kinase, type-I (thick veins and
saxophone in Drosophila) and type-II (punt)
(for review, see Massagué et al. 1994). Intracellular molecules
essential for the signal transduction by the TGF- family include
Smads family molecules [Mothers against dpp] (for review,
see Massagué et al. 1997), and nuclear factors such as
Xenopus FAST1 (Chen et al. 1996) and Drosophila
schnurri (Arora et al. 1995; Grieder et al. 1995). Signaling by the
Wnt superfamily is mediated by the Frizzled receptor family,
Dishevelled, Glycogen synthase kinase-3 (shaggy/zeste-white 3), -Catenin
(armadillo), and nuclear factors such as
TCF/LEF-1 (pangolin) (for review, see Cadigan
and Nusse 1997; Cavallo et al. 1997). Studies of Drosophila
have provided valuable insights into the roles of these highly
conserved signaling pathways in morphogenesis such as wing patterning
(Blair 1995) and midgut specification (Bienz 1994; Graba et al. 1997).
The gut epithelium of Drosophila is derived from the anterior and posterior primordia at both ends of the blastoderm embryo. These primordia are initially nonsegmental and fused into a single continuum. Secreted molecules, such as Dpp and Wg, induce subsequent morphogenetic events that ultimately compartmentalize the primordia into morphologically distinct sectors. During this process, these signals also act for cells to take distinct developmental paths to establish the functional organization of the midgut.
The proventriculus develops at the junction of the foregut and the
midgut, and functions as a valve regulating the passage of food into
the midgut. It is composed of three layers; the outer layer is derived
from the anterior-most region of the midgut, the middle layer is
derived from the foregut of mesoderm-free keyhole structure, and the
inner layer is derived from the esophagus (for review, see Skaer 1993).
The late steps of proventriculus morphogenesis are attributable to
migration of cells, and are controlled by at least Hedgehog (Hh) and
Wg, which are expressed in the mesoderm-free keyhole structure
(Pankratz and Hoch 1995; Fig. 1E). Little is known so
far about downstream targets that respond to Hh/Wg
signals during proventriculus development.
|
The midgut consists of two germ layers, the visceral mesoderm and the
endoderm. The middle midgut cells derived from the endoderm differentiate into four distinct types of cells: copper, interstitial, large flat, and iron cells. These endodermal cell types are specified by Dpp and Wg, which are expressed in the adhering visceral mesoderm of
the parasegments (PS) 7 and 8, respectively. Copper cells exhibit a
unique morphology with banana shapes and exhibit UV light-induced fluorescence after copper feeding. These characteristics are specified by a homeotic gene, labial (lab), which is activated
by the Dpp signal in the midgut. Two different thresholds of Wg define
copper and large flat cells (Hoppler and Bienz 1995). However, it
remains unclear how Lab confers the transcriptional regulation to
specify copper cells.
Here, we have identified a new gene, defective proventriculus (dve), which encodes a homeodomain protein. dve1 homozygous mutants are defective in proventriculus morphogenesis and in the arrangement of middle midgut cells. The dve gene responds differentially to the Wg or Dpp signal in the anterior-most or middle midgut, respectively. In the anterior-most midgut, dve activity is required to maintain the three-layered structure indispensable for a functional proventriculus. In the middle midgut, the dve gene is expressed in all precursors of four distinct cell types, subsequently it is repressed only in copper cells. This repression is mediated by two Dpp target genes, lab and dve itself, and is also essential for the functional specification of copper cells. Thus, dve is involved in different developmental aspects of the midgut under the control of the different extracellular signals. We discuss the roles of dve in the context of the network of inductive signals that organize midgut development.
| |
Results |
|---|
|
Top
Abstract
Introduction
Results
Discussion
Materials & Methods
References
|
|---|
Identification of the dve locus
We have identified the dve locus by two enhancer-trap insertions at 58D1-2 on the second chromosome; one is a viable allele, dveSH255, and the other is a homozygous lethal allele, dve1 (Fig. 2A). Immediately after hatching, dve1 homozygous larvae exhibit normal locomotion behavior; however, develop into small larvae and die within a day. The lethality of the dve1 allele is attributable to the P-element insertion, because the P-element excision recovered homozygous viable adults (15 of 28). Embryos bearing dve1 in trans with Df(2R)X58-3, which uncovers the dve locus, hatch normally into first instar larvae but die, suggesting that animals of this genotype have a lethal stage similar to that of dve1 homozygotes. In addition, dve1 homozygous embryos express no detectable dve transcript (data not shown) or Dve protein (see Fig. 6B, below) until stage 14; faint staining with the anti-Dve serum was detectable at early stage 17 (Fig. 3H). These observations indicate that dve1 is a strong hypomorphic allele. We term this locus defective proventriculus based on the defects in the proventriculus formation as described below.
|
|
The reduced body sizes of dve1 homozygous larvae
suggest that the feeding is affected by the dve1
mutation. The colored yeast fed to heterozygous larvae stained their
guts red throughout their length (see Fig. 1A). On the other hand, it
accumulated in the proventriculus in dve1 mutant
larvae (arrow in Fig. 1B). Consistent with this observation, dve1 larvae fail to form the proventriculus
correctly (Fig. 1, cf. C and D). In the wild type, cell movement leads
to formation of the internal portion of the proventriculus during
embryonic stages 16-17; cells of the foregut epithelium invaginate
into the anterior-most midgut that normally expresses dve
(Fig. 1E). In dve1 embryos, the cell migration was
greatly delayed and the internalization was only temporary (data not
shown). As a result, dve1 larvae cannot form the
three-layered structure of the proventriculus (Fig. 1D) as observed in
hh or wg mutant embryos (Pankratz and Hoch 1995). As
the dve-expressing anterior-most midgut constitutes the outer
layer of the proventriculus (Fig. 1E), these dve1
phenotypes suggest that dve activity is required for the
functional development of outer layer cells to retain the internal
portion of the proventriculus.
Molecular cloning of the dve gene
To isolate the dve gene, we used genomic DNAs flanking the dveSH255 insertion and identified the full-length cDNA of a 4.9-kb transcript. Analysis of the genomic DNA of the dve gene revealed that the transcript consists of six exons and that the P-element of dveSH255 is inserted at 167 bp upstream from the transcription start site. The P-element insertion of the lethal dve1 allele is located within the second intron of the dve gene (Fig. 2A).
The predicted open reading frame of Dve is 1019 amino acids long.
Homology with the homeodomain is found near its carboxy-terminal region
(yellow in Fig. 2B). The Dve homeodomain contains all four invariant
amino acids located within helix 3 (sharps in Fig. 2C), and matches
well with other highly conserved residues (asterisks in Fig. 2C;
Gehring 1992). However, the Dve homeodomain is unusual, as it has a
10-amino-acid insertion between helices 2 and 3 (Bürglin 1997).
The ninth amino acid of helix 3 confers the recognition specificity for
its binding sequences (red arrowhead in Fig. 2C; Hanes and Brent 1989;
Treisman et al. 1989), and the recognition helix of Dve is closest to
the orthodenticle (otd) class homeodomain. In
contrast, helices 1 and 2 of Dve exhibit homology with POU homeodomains
rather than otd class homeodomains (Fig. 2C). Therefore, the
Dve homeodomain seems to be a novel class of homeodomain that is
intermediate between POU and otd class homeodomains. The Dve staining observed in the nucleus is consistent with its homology with
the homeodomain and suggests that Dve is a putative transcription factor.
Expression pattern of dve in the embryonic midgut
The expression of the Dve protein exhibits a pleiotropic pattern
including three separate domains of the midgut: anterior-most, middle,
and posterior-most regions. Because defects in the lethal dve1 allele are evidently observed in the gut, we
focus on dve expression during gut development. At stage 10, dve expression is first observed in the invaginated stomodeum
(Fig. 3A), where hh and wg are expressed under the
control of the winged-helix transcription factor Fork head (Hoch and
Pankratz 1996). We compared anti-Dve staining with wg gene
expression that was monitored with wg-lacZ. The expression of
dve and wg overlaps initially in the stomodeum, and
is then segregated into nonoverlapping but adjacent regions; the
mesoderm-free keyhole structure expresses wg, whereas
dve is solely expessed in the anterior-most region of the
midgut (Fig. 4A,E), which becomes outer layer of the proventriculus
later at stage 16 (Fig. 4H).
|
From stage 13 onward, dve is also expressed in two other parts of the endodermal midgut; one is the middle region where the anterior and posterior midgut primordia are fused, and the other is the posterior-most region (see Fig. 3C). The middle region expressing dve corresponds to both the anterior and posterior sides of the middle constriction (arrow in Fig. 3D). At stage 16, this region becomes the second and third gut lobes (Fig. 3E). Dve-expressing cells in this region include copper cell precursors that coexpress lab (see Fig. 7A,D, below).
wg signal regulates dve expression during proventriculus development
As the dve-expressing domain in the presumptive proventriculus initially overlaps with the wg- and hh-expressing domain (Fig. 4A), we examined the possibility that Wg or Hh regulates the expression of the dve gene. In wgCX4 null embryos, dve expression is abolished in the stomodeum, whereas epidermal expression is normally detected (Fig. 4B). In hhAC mutant embryos, dve is normally expressed (Fig. 4C). Thus, Wg rather than Hh is required to activate the dve gene in the presumptive proventriculus.
The keyhole structure also expresses cubitus interruptus
(ci). In ciD mutants, the expression of
wg and hh is not maintained, leading to the absence
of a keyhole structure. The maintenance of wg expression also
requires hh activity (Pankratz and Hoch 1995). In
ciD mutants, dve expression is never
detected in the presumptive proventriculus until stage 13 (Fig. 4D),
whereas it begins to be detectable at stage 14 despite the absence of a
keyhole structure (Fig. 4F). In addition, dve expression
expands ectopically into the presumptive first gut lobe after stage 15 (Fig. 4G). Similar ectopic expression of dve is observed in
hhIJ35 or wgIL114 mutant
embryos (Fig. 4I; data not shown). The ectopic dve expression in these mutants suggests that the Wg signal is also required to define
the posterior border of late dve expression in the presumptive proventriculus, in addition to the initial activation of the
dve gene.
Dpp signal regulates dve expression in the middle midgut
In the middle midgut, Dpp and Wg signals from the visceral mesoderm
control gene expression on both the anterior and posterior sides of the
middle constriction. Hence we examined whether dve expression
depends on the Dpp or Wg signals in this region. In thick
veins (tkv) mutants that lack the functional type I
receptor for Dpp, dve expression is completely absent in the
middle midgut, whereas the expression in other parts of the midgut is
not affected (Fig. 5, cf. A with C). This indicates
that dve expression in the middle midgut requires the Dpp
signal. On the other hand, the ubiquitous expression of dpp
throughout the visceral mesoderm, using the combination of
UAS-dpp and 24B-Gal4 driver, induced dve expression
throughout the underlying endoderm (Fig. 5, cf. A with E). This
dependence of dve expression on Dpp is quite similar to that
of lab expression (Staehling-Hampton and Hoffmann 1994).
|
We used abdominal-A (abd-A) mutants to examine the
dependence of dve expression on Wg. In abd-A mutants,
the expression of Ultrabithorax (Ubx) and
dpp in the visceral mesoderm expands throughout the posterior
midgut and endogenous wg is not activated (Immerglück et
al. 1990; Panganiban et al. 1990; Reuter et al. 1990). In
abd-A mutants, dve expression also expands
posteriorly (Fig. 5D). These three observations are consistent with the
notion that Dpp is sufficient to induce dve expression without
Wg in the midgut. To confirm this, we examined the effect of the
ubiquitous wg expression in the visceral mesoderm on
dve expression. The ubiquitous activation of the Wg signal
using shaggy null mutations induces the expression of
Ubx and dpp only anteriorly, but not posteriorly, in
the midgut (Yu et al. 1996). The ubiquitous wg expression in
the whole visceral mesoderm induced ectopic dve expression
only anteriorly (Fig. 5F), indicating that this induction appears to be
indirect and mediated by ectopic dpp activation. It is
noteworthy that the ectopic dve induction is especially strong
in the anterior-most region that is normally responsive to Wg (Fig.
5F). Taken together, these results indicate that dve
expression in the middle midgut does not depend on Wg but on Dpp, which
is in contrast to dve expression during proventriculus
development.
Two distinct transcription factors, schnurri (shn)
and Mad, have been reported to mediate the Dpp signal (Arora
et al. 1995; Grieder et al. 1995; Kim et al. 1997). The expression of
Dve protein in the middle midgut is absent in shn mutants
(Fig. 5G; data not shown). In addition, the ubiquitous mesodermal
expression of dpp in the shn mutant background fails
to induce endodermal dve expression (Fig. 5G), indicating that
dve induction requires shn activity in the midgut
endoderm. In contrast, dve expression is not affected in
Mad mutants, whereas lab expression is completely
absent in the same mutant background (Fig. 5H; see Newfeld et al. 1996
and also Discussion).
|
Phenotypes of dve mutants in the larval middle midgut
Although the dve gene is strongly expressed in the
embryonic middle midgut, morphological defects are not evident in this region in dve1 mutants; gut constriction normally
occurs, and the arrangement of the stage-17 midgut appears to be
normal (see Fig. 3E-H). Cells that express the dve gene at
embryonic stages develop into four types of larval midgut cells:
copper, interstitial, large flat, and iron cells. The expression of the
dve gene in these larval cells was monitored as the expression
of the lacZ gene that is located in the
dve1 enhancer-trap insertion
(dve1-lacZ). -Galactosidase activity
attributable to dve1-lacZ is observed in
interstitial, large flat, and iron cells, but not in copper cells in
heterozygous larvae (Fig. 6A,B). In dve mutant larvae, we
found the ectopic dve1-lacZ expression in copper
cells, in addition to highly disorganized arrangement of these cells
(Fig. 6C). These observations suggest that the dve activity is
required to repress its own expression in copper cells. To determine
whether the dve gene could rescue this mutant phenotype of
ectopic dve1-lacZ expression, the dve gene
was induced ubiquitously in the dve1 mutant
background under the control of Gal4-UAS system (Brand and Perrimon
1993). Weak dve induction in stage-17 embryos suppressed the ectopic dve1-lacZ expression in larval copper
cells (Fig. 6C, F,G), indicating that a mutant phenotype is exactly
caused by the lack of the dve gene product.
To monitor cell types in the larval midgut, the C5-2-7 enhancer-trap insertion, which marks specifically interstitial cells, was introduced into the dveE38 background (see Materials and Methods). In dveE38 mutants, C5-2-7 expression remained unaffected although the arrangement of copper cells was highly disorganized (Fig. 6D,E), suggesting that interstitial cells are not strongly affected by the dve mutation.
In the larval midgut, the dve gene is thus expressed in a cell type-specific manner that requires its own activity. As described below, the repression of the dve gene in copper cells depends on the function of both Lab and Dve itself, and is essential for the functional specification of these cells.
Copper cells are specified by the cross-regulation between dve and lab
In the middle midgut, lab is expressed under the control of
Dpp as is dve, whereas lab is regulated negatively by
the Wg signal to generate a sharp posterior border (Hoppler and Bienz
1995). We compared the expression of lab and dve
using lab upstream 6.3 kb-lacZ
(lab-6.3lacZ), which mimics endogenous lab
expression (Tremml and Bienz 1992). The expression of lab is
observed in the endoderm just beneath the dpp-expressing
visceral mesoderm of PS 7, but not in the inner endodermal cells as
described (Reuter et al. 1990; Fig. 7B). In contrast,
dve is expressed more broadly throughout the inner endodermal
layers including presumptive interstitial cell precursors (Fig. 7A-C).
Another difference in lab and dve expression is that
dve expression is repressed subsequently in lab-expressing cells that become copper cells (see Fig. 6B).
We examined the possibility that Lab might be involved in the
repression of dve in copper cells. In lab mutants,
dveE8-lacZ expression is not repressed in
presumptive copper cells as observed in dve mutants (Fig.
7E-G). This pattern of gene expression is similar to that of
neighboring interstitial cells, which express dve continuously
without lab expression in the wild type. Hence the inability
of dve repression in the presumptive copper cells of
lab mutants could be attributable to a change in cell
fate
copper cells to interstitial cells. In this case, the expression
of an interstitial cell marker C5-2-7 should expand in this region. However, it turned out that its expression is abolished in lab mutants (Fig. 7H,I), suggesting that interstitial cells are also affected by the lab mutation. Thus, it is unlikely that the
lab mutation causes the transformation of copper cells into
interstitial cells. Taken together, the repression of dve
requires the activities of both Lab and Dve itself.
|
To determine whether the dve repression in copper cells is
essential for establishment of their correct identity, dve was overexpressed ubiquitously at stage 17. Strong heat shock-induced dve expression resulted in an abnormal morphology of copper
cells. The typical banana shape of copper cells is frequently lost, the cells becoming circular (Fig. 7J), suggesting an abnormal cytoskeletal organization in these cells. To determine the effect of ectopic dve on the copper cell function, ubiquitous dve
expression was induced alternatively by a milder heat shock. Under this
condition, the copper cells appear to retain their normal morphology
(Fig. 7K), however, the typical character of copper cells, UV
light-induced orange fluorescence on copper feeding, is greatly reduced
in these guts (Fig. 7L-N). This mild heat shock does not affect the
posterior fluorescence attributable to iron cells, suggesting that the
function of copper cells is specifically impaired by this treatment.
Taken together with the results for dve mutants described
above, both the loss of function and ectopic expression of the
dve gene affect the morphology of copper cells, and ectopic
dve expression impairs the function of copper cells without
affecting their morphology. Unfortunately, the function of copper cells
in dve mutants could not be examined because of the feeding
defect in the proventriculus; however, the morphological abnormality
observed in dve mutants suggests that dve activity is
also required for the functional development of copper cell precursors.
Our results indicate that temporally restricted dve repression
is essential for this functional specification in addition to the
lab gene, which is indispensable for copper cell development
(Hoppler and Bienz 1994). This repression depends on Lab and Dve
itself. Thus, the cross-regulation of the two Dpp target genes
lab and dve specifies the functional identity of
copper cells.
| |
Discussion |
|---|
|
Top
Abstract
Introduction
Results
Discussion
Materials & Methods
References
|
|---|
Dve expression is regulated differentially by Dpp and Wg
Accumulating evidence has shown that signal inputs from Dpp and Wg
exert cooperative or antagonistic effects to define the border of
target gene expression. The development of the leg imaginal disc has
provided a typical example for the antagonistic and cooperative functions of these two factors. dpp and wg are
expressed adjacently on the anterior side of the anterior-posterior
compartment boundary of the disc; wg is expressed ventrally,
whereas dpp is expressed at high levels dorsally. The
convergence of both signals is required to induce Distal-less
and aristaless at the center of the disc (Campbell et al.
1993; Cohen et al. 1993). On the other hand, the expression of a
ventral marker, H15, is activated by Wg and repressed by Dpp. Thus, the
antagonistic effects of these two factors restrict the H15 expression
ventrally (Brook and Cohen 1996). Another example of the complex gene
regulation by Dpp and Wg has been found in midgut morphogenesis, where
both dpp and wg are expressed in adjacent regions of
the visceral mesoderm of PS 7 and 8, respectively. Dpp and Wg act
independently but synergistically on the visceral mesoderm enhancer of
Ubx; however, Wg-dependent transcriptional stimulation
requires the Dpp signal (Riese et al. 1997). In contrast, the
lab expression in the endoderm is activated by Dpp and
repressed by Wg to confer the discrete boundary between copper and
large flat cells (Hoppler and Bienz 1995).
The roles of Dpp and Wg in dve expression are quite different from those reported previously. Dve-expressing domains are exposed to both Wg and Dpp; however, dve expression responds differentially to either of them in distinct parts of the midgut. In the middle midgut, dve expression depends mainly on Dpp, although a possible contribution of the Wg signal cannot be excluded, as observed for Ubx expression. In contrast, the dve expression in the anterior-most midgut depends on Wg, but not on Dpp. In this region, the posterior border of dve expression appears to be defined by the Wg signal.
How does the dve gene sort the Wg and Dpp signals depending on
the domains of the midgut? Enhancer elements responding to the two
signals in the midgut have been well studied for the Ubx gene.
In the case of the Ubx gene, the two signal inputs are
mediated by the B enhancer, which includes the Dpp response sequence
(DRS) and the Wg response sequence (WRS). These two response sequences are closely located so that signal inputs from Dpp and Wg coordinately induce the expression in a pattern spanning both sides of the middle
constriction in the visceral mesoderm (Riese et al. 1997). This
expression pattern is similar to that of the dve gene,
although the B enhancer of the Ubx gene drives the expression
mainly in the visceral mesoderm. Interestingly, four copies of this
enhancer derivative (L-CRE) drive Wg-dependent expression in the
anterior-most endoderm of the midgut where the dve gene is
activated. The similarity in the spatial pattern between dve
expression and the Ubx enhancer activation raises the
possibility that Dpp and Wg act on similar regulatory elements that are
present in the dve gene to induce the broad expression in the
middle midgut. We do not know anything about the regulatory mechanisms
that activate dve gene depending on Wg but not on Dpp in the
anterior-most midgut. Characterization of the regulatory elements of
the dve gene would provide useful clues for spatial regulation
during Dpp- and Wg-dependent pattern formation.
Early gradient of Dpp defines the cell identity in the midgut
In the middle midgut, both the lab and dve genes
are activated by the Dpp signal, but they do not respond equally to
Dpp. lab expression is restricted strictly to the endodermal
cells that are in contact with the dpp-expressing visceral
mesoderm (Reuter et al. 1990). In contrast, dve is widely
activated even in endodermal cells that are remote from the Dpp source
(Fig. 7A-C). Assuming a concentration gradient of Dpp in the endoderm as well as wing disc (Lecuit et al. 1996; Nellen et al. 1996), these
observations suggest that lab expression requires a higher concentration of Dpp than dve expression. Therefore,
dve can be activated by low levels of Dpp that are
insufficient to induce lab. Furthermore, we found a difference
between lab and dve gene activation in the
requirement for Mad, which transduces the Dpp signal. Mutant
embryos for Mad fail to express lab but show normal activation of dve (Newfeld et al. 1996; Fig. 5H). If the
maternal supply of Mad transcripts is sufficient to induce
dve, but not lab expression, the difference in Mad
dependence might be explained by the different thresholds for the Dpp
signal between lab and dve activation.
The early gradient of Dpp in the endoderm thus gives rise to two
distinct cells in terms of gene expression; one is lab
and
dve+, and the other is lab+ and
dve+ (Fig. 8). If this situation is
disrupted by ubiquitous lab induction during 7-9 hr of
embryonic development, interstitial cells are transformed into copper
cell-like cells that exhibit orange fluorescence on copper feeding
(Hoppler and Bienz 1994). At later stages, these two types of cells
intermingle, and interstitial cell precursors are also exposed to high
Dpp (Fig. 7D; Reuter et al. 1990). Nonetheless, these cells never
express lab at later stages, indicating that the cell identity
once established by the early Dpp gradient seems to be maintained.
Mechanisms involved in this maintenance of cell identity remain
unclear. One of the simplest explanations would be that Dve continues
to suppress lab in interstitial cell precursors. However, this
is unlikely because ectopic expression of lab-6.3lacZ is not
detectable in dve mutants (data not shown). Another
possibility is that the negative autoregulation of lab is
involved in the continuous lab suppression in interstitial
cell precursors. In fact, heat shock induction of the lab
transgene between 9 and 11 hr of development frequently results in
severe repression of copper cell development, suggesting that
heat-induced Lab during a critical period suppresses the endogenous
lab gene (Hoppler and Bienz 1994). The time window for this
negative lab autoregulation coincides well with when
interstitial cell precursors are exposed to high Dpp. It is also an
intriguing possibility that Dpp-induced Smad molecules, which inhibit
the TGF-/Dpp signal, participate in these processes
as a negative feedback loop (Hayashi et al. 1997; Nakao et al. 1997;
Tsuneizumi et al. 1997).
|
Dve function in copper cell specification
Dve is expressed in all precursors of four distinct types of middle midgut cells, and is repressed subsequently only in copper cells. This repression is mediated by Lab and Dve itself, and turns out to be essential for the functional specification of copper cells, as judged by the ability of copper uptake. This does not necessarily mean that dve activity is required for the cell fate decision of copper cells. In fact, the expression of several cell-type markers for middle midgut cells is unaffected in dve mutants (Fig. 6E; data not shown). dve activity seems to be required to refine the functional specificity at later stages of development after the cell fate decision has been completed.
It is very important to understand the mechanism of dve repression that is critical for the acquisition of functional identity in copper cells. There are two possibilities that explain how the repression of the dve gene occurs in larval copper cells (Fig. 8). One is a cell nonautonomous mechanism; interstitial cells might send a signal to adjacent copper cells to repress dve gene expression. For instance, an interstitial cell marker, C5-2-7, is abolished in lab mutants (Fig. 7H,I), suggesting a cell nonautonomous interaction between interstitial cells and copper cells. The other is a cell autonomous mechanism by which the dve gene is autoregulated negatively in copper cells. In copper cell precursors, Lab and Dve are coexpressed during embryogenesis, and are able to act coordinately as transcriptional regulators. Thus, the latter possibility is more plausible, and Dve might function as a transcriptional cofactor for Lab to specify the functional identity of copper cells. Further characterization of the dve gene would clarify the molecular mechanism conferring the functional specificity to each of the four distinct types of middle midgut cells.
| |
Materials and methods |
|---|
|
Top
Abstract
Introduction
Results
Discussion
Materials & Methods
References
|
|---|
Fly strains
The following mutant strains were used: Df(2R)X58-3
(Kerrebrock et al. 1995) uncover the dve locus;
wgCX4, wgIL114,
hhAC, hhIJ35,
ciD, tkv8,
abd-AM1, shn1,
Mad12, and labvd1 (FLYBASE;
http://morgan.harvard.edu/).
dpp or wg was induced ectopically in the embryonic
mesoderm by using 24B-Gal4, UAS-dpp (Staehling-Hampton and
Hoffmann 1994) and UAS-wg (Lawrence et al. 1995). To monitor
the cell type-specific expression in the midgut, copper cell marker
lab-6.3lacZ (Tremml and Bienz 1992), interstitial cell marker
C5-2-7, and large flat cell markers A3-2-66 and B1-2-49 (Hartenstein
and Jan 1992; Hoppler and Bienz 1995) were used. The following
lacZ balancers were used to distinguish homozygous embryos:
en11CyO, CyOftz-lacZ,
TM3ftz-lacZ, and TM3Ubx-lacZ. The
ciD homozygous embryos were identified as to the
expression pattern of wg-lacZ derived from
en11CyO. dveSH255 and
dve1 are enhancer-trap lines of FZ and PZ,
respectively; dve1 is identical to
l(2)0173801738 generated by the Berkeley Drosophila
Genome Project (Karpen and Spradling 1992; Mlodzik and Hiromi 1992). By
the P-element excision of dve1, an imprecise
excision allele dveE38 was recovered that has the
internal deletion of both the rosy marker and the
lacZ gene. dveE38 mutants showed
essentially the same Dve protein expression and phenotype as the
dve1 allele. To examine roles in cell-type
specification, each of the midgut cell type markers was introduced into
the dveE38 background by recombination.
dveE8 is a hypomorphic allele that lacks the
rosy marker but retains the lacZ gene, and this
allele was used to monitor dve expression in the lab
mutant background.
A 4.5-kb EcoRI fragment of dve cDNA was inserted into
the pUAST vector (Brand and Perrimon 1993) to construct
pUAS-dve, and several independent transformants were
obtained. Of these lines, UAS-dve-9B2 (homozygous viable on
the third chromosome) was used in this study. For the rescue
experiments, yw;
dve1/CyO[y+];
hs-GAL4 flies were crossed with y w;
dve1/CyO[y+];
UAS-dve-9B2, and stage-17 embryos were heat shocked at
35°C for 30 min. Homozygous first instar larvae were identified from the absence of the yellow (y) marker. For ectopic
expression, yw;
dve1/CyO[y+];
UAS-dve-9B2 flies were crossed with hs-Gal4, and
yw;
dve1/CyO[y+];
hs-GAL4 flies were crossed with Oregon-R as a control, and stage-17 embryos were heat shocked at 37°C. Subsequently, the gut
morphology and copper fluorescence were examined.
Isolation of genomic and cDNA clones
The dve gene was identified originally from an enhancer trap line that shows an abnormal object fixation response (dveSH255; H. Nakagoshi and F. Matsuzaki, unpubl.). A 4.5-kb SalI genomic fragment flanked to dveSH255 insertion was used for screening Oregon-R adult head cDNA library. A nearly full-length dve cDNA clone (W12) was isolated, and its 5' end was determined by 5' RACE. Genomic clones covering the dve locus were isolated from Oregon-R genomic DNA library, and the exon-intron boundaries were determined. DNA sequencing was carried out using an ABI 373 DNA sequencer (Applied Biosystems).
Antibody staining
A 2.3-kb EcoRI fragment of dve cDNA was inserted
into the pTrcHisC vector (Invitrogen Xpress System). As a result, the
carboxy-terminal 368 amino acid fragment was fused to the
amino-terminal histidine tag. The resultant fusion protein (HE26) was
purified and immunized to a rabbit. The HE26 antiserum was used for
further immunostaining experiments. The stages of embryonic development
are according to Campos-Ortega and Hartenstein (1985). For whole-mount
staining, embryos were dechorionized and then fixed in a 4%
formaldehyde/phosphate-buffered saline (PBS):
heptane = 1:1 solution for 30 min, devitellinized with methanol
and then washed with PBS- 0.2% Tween 20. Embryos were incubated
with HE26 antiserum (1:1000 dilution) in the blocking solution (5%
skim milk/PBS- 0.2% Tween 20), and then washed with PBS-0.2% Tween 20. For the detection, a secondary antibody
(biotin-conjugated anti-rabbit IgG, diluted 1:1000; Vector Labs)
and a horseradish peroxidase (HRP)-conjugated avidin-biotin complex
solution (ABC Elite; Vector Labs) were used. To detect lacZ
expression, an anti--galactosidase monoclonal antibody (diluted
1:1000; Promega) was used.
For confocal microscopic observations with a Bio-Rad MRC-1024, Cy-3 conjugated anti-mouse IgG and Cy-5 conjugated anti-rabbit IgG (Jackson) were used as secondary antibodies.
Feeding assay and X-gal staining of larval guts
For the feeding assay, larvae were allowed to grow on
sucrose-supplied agar plates of 100-mm dish with a red-colored yeast (dyed with carmine; Sigma). Several hours feeding was enough to visualize gut motility. To analyze the copper cell function, larvae were fed with the yeast containing a high concentration (2-3
mg/ml) of CuSO4 for at least 5 hr to visualize
posterior fluorescence attributable to iron cells. The guts of
dissected larvae were immediately mounted in 80% glycerol and excited
with UV light using a DAPI filter. The procedure was basically
according to Hoppler and Bienz (1994).
For X-gal staining, first instar larvae were dissected in a fixing
solution (1% glutaraldehyde/PBS), and subsequently their guts were washed with PBS and stained in 0.1%
5-bromo-4-chloro-3-indolyl--galactopyranoside (X-gal) for
-galactosidase activity. The stained guts were washed with
PBS-0.2% Tween 20 and then mounted in 80% glycerol.
| |
Acknowledgments |
|---|
We are grateful to M. Bienz for lab-6.3lacZ strain and comments on the manuscript; T.L. Orr-Weaver, S. Cohen, C. Nüsslein-Volhard, A. Bejsovec, S. Goto, S. Hayashi, T. Tabata, T. Uemura, Y. Takamatsu, and the stock centers of Bloomington, Bowling Green and Umea for fly strains; R. Murakami for helpful discussion. We are also grateful to A. Fujisawa-Sehara for the initial indication of the embryonic expression of the dve gene. We also thank T. Jinnai and H. Izumi for their technical assistance, C. Hama for the genomic DNA library, and all members of our laboratory for helpful discussion and encouragement. This work was partly supported by grants to H. N. from PRESTO (Precursory Research for Embryonic Science and Technology) of Japan Science and Technology Corporation.
The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked `advertisement' in accordance with 18 USC section 1734 solely to indicate this fact.
| |
Note |
|---|
The nucleotide sequence data reported in this paper will appear in the DDBJ, EMBL, and GenBank nucleotide sequence databases under the following accession number AB010299.
| |
Footnotes |
|---|
Received May 11, 1998; revised version accepted July 17, 1998.
6 Corresponding authors.
E-MAIL nakagosi{at}ncnaxp.ncnp.go.jp; fumio{at}ncnaxp.ncnp.go.jp; FAX (+81)-423-46-1755.
| |
References |
|---|
|
Top
Abstract
Introduction
Results
Discussion
Materials & Methods
References
|
|---|
signaling pathway and encodes a transcription factor homologous to the human MBP family.
Cell
81:
781-790[CrossRef][Medline]. signalling.
Nature
383:
691-696[CrossRef][Medline]. receptor and functions as an antagonist of TGF signaling.
Cell
89:
1165-1173[CrossRef][Medline]. superfamily: New members, new receptors, and new genetic tests of function in different organisms.
Genes Dev.
8:
133-146 family and its composite receptors.
Trends Cell Biol.
4:
172-178.[CrossRef][Medline] signaling through the Smad pathway.
Trends Cell Biol.
7:
187-192.
-helices are indicated below: (#) Invariant amino acids; (*)
highly conserved amino acids in homeodomains. The ninth amino acid of
the recognition helix (helix 3) is indicated by a red arrowhead. The
amino acids (Q, Y, and R) marked by green are highly conserved in
homeodomains except for POU homeodomains. Invariant amino acids in POU
homeodomains (Herr and Cleary 1995
