Genetics: Published Articles Ahead of Print, published on May 17, 2012 as 10.1534/genetics.112.141093
New negative feedback regulators of Egfr signaling in Drosophila
Jonathan P. Butchar, Donna Cain, Sathiya N. Manivannan, Andrea D. McCue, Liana Bonanno,
Sarah Halula, Sharon Truesdell, Christina L. Austin, Thomas L. Jacobsen, and Amanda Simcox
Department of Molecular Genetics, 484 W 12th Ave, Ohio State University, Columbus, OH
43210
Microarray data have been submitted to GEO (GSE34872)
1
Copyright 2012.
Running Title: Feedback regulation of Drosophila Egfr pathway
Key words: Egfr, microarray, Drosophila, wing disc, Sulf1, CG4096, feedback regulation
Corresponding author: Amanda Simcox, Department of Molecular Genetics, 484 W. 12th Ave,
Columbus, OH 43210
Phone: 614-292-8857
Fax: 614-292-8866
E-mail: [email protected]
2
ABSTRACT
The highly conserved Epidermal Growth Factor-receptor (Egfr) pathway is required in all
animals for normal development and homeostasis; consequently, aberrant Egfr signaling is
implicated in a number of diseases. Genetic analysis of Drosophila melanogaster Egfr has
contributed significantly to understanding this conserved pathway and led to the discovery of
new components and targets. Here we used microarray analysis of third instar wing discs, in
which Egfr signaling was perturbed, to identify new Egfr-responsive genes. Upregulated
transcripts included five known targets suggesting the approach was valid. We investigated the
function of 29 previously uncharacterized genes, which had pronounced responses. The Egfr
pathway is important for wing-vein patterning and using reverse genetic analysis we identified
five genes that showed venation defects. Three of these genes are expressed in vein primordia
and all showed transcriptional changes in response to altered Egfr activity consistent with being
targets of the pathway. Genetic interactions with Egfr further linked two of the genes, Sulfated
(Sulf1), an endosulfatase gene, and CG4096, an ADAMTS (A Disintegrin And Metalloproteinase
with ThromboSpondin motifs) gene, to the pathway. Sulf1 showed a strong genetic interaction
with the neuregulin-like ligand vein (vn) and may influence binding of Vn to heparan-sulfated
proteoglycans (HSPGs). How Drosophila Egfr activity is modulated by CG4096 is unknown, but
interestingly vertebrate EGF ligands are regulated by a related ADAMTS protein. We suggest
Sulf1 and CG4096 are negative feedback regulators of Egfr signaling that function in the
extracellular space to influence ligand activity.
3
INTRODUCTION
The EGF receptor (Egfr) pathway is required for cell proliferation, differentiation,
migration, and survival during development (reviewed for Drosophila in (SHILO 2005)).
Drosophila Egfr is activated by four ligands; three in the TGF-α family—spitz (spi), gurken
(grk) and keren (krn) and a neuregulin called vein (vn). The pathway is controlled by multiple
regulatory mechanisms that can either dampen or amplify the signal. Components of these
regulatory mechanisms include transcriptional targets of the signaling pathway and thus serve as
negative and positive feedback loops (reviewed in (AVRAHAM and YARDEN 2011)). Negative
feedback regulators include argos (aos) (GOLEMBO et al. 1996; KLEIN et al. 2004; SCHWEITZER
et al. 1995), sprouty (sty) (CASCI et al. 1999), kekkon-1 (GHIGLIONE et al. 1999), MAPK
Phosphatase 3 (Mkp3) (GOMEZ et al. 2005; KIM et al. 2003), mae (VIVEKANAND et al. 2004) and
d-Cbl (PAI et al. 2000). Positive feedback regulators include the two Egfr activating ligands, vn
(GOLEMBO et al. 1999; WANG et al. 2000; WASSERMAN and FREEMAN 1998; WESSELLS et al.
1999) and spi (WASSERMAN and FREEMAN 1998), pointed (pnt) (GABAY et al. 1996) and a
miRNA, miR7, which positively regulates the pathway by targeting the transcriptional repressor
yan (LI and CARTHEW 2005). Here we provide genetic evidence for two new feedback controls,
which both function as negative regulators of Egfr signaling in the wing imaginal disc.
The Drosophila wing has proven to be a good model system to study Egfr signaling
because Egfr is required for specifying the stereotypical pattern of veins separated by interveins
in this tissue. A prepattern of the veins is apparent in the mature third instar imaginal disc and
can be visualized, for example, by rhomboid (rho) expression (STURTEVANT et al. 1993). Rho is
required to process the TGF-α ligands to an active form and flies mutant for both rho and the
neuregulin-like ligand vn lack all veins (STURTEVANT and BIER 1995; URBAN et al. 2001;
4
URBAN et al. 2002). In contrast to the loss of vein phenotypes seen when Egfr signaling is
reduced, excessive Egfr signaling leads to extra-vein phenotypes. In the third instar wing disc,
vn is expressed along the anterior-posterior boundary in the central intervein territory, where it is
required for specifying the flanking longitudinal veins (3 and 4), especially vein 4 (SIMCOX et al.
1996). vn expression is induced by Hedgehog signaling (WESSELLS et al. 1999), and indeed in
addition to the Egfr pathway, the Hh, Dpp, Wingless, and Notch signaling pathways are required
for positioning veins and determining their thickness (reviewed in (BLAIR 2007)).
There have been multiple genetic screens for venation mutants leading to the discovery of
new components in these signaling pathways. Screening is facilitated because the wing, like the
eye, is dispensable for viability and has a stereotypical pattern that can be easily scored for
changes. Here, rather than conducting another genetic screen, we employed a microarray-based
approach to first identify Egfr-responsive genes. We then tested the function of candidate target
genes using reverse genetics. With this approach we hoped to find novel genes that were targets
of the pathway but that would not necessarily be discovered in genetic screens because they
either had pleiotropic roles causing early death or only small phenotypic effects. We discovered
five genes with venation defects and further genetic tests suggested that two of these, Sulfated
(Sulf1) and CG4096, act as negative regulators of the Egfr pathway. The results provide more
evidence for the elaborate controls that ensure precise regulation of signaling pathways. They
also provide an example of genes that contribute a relatively fine control that when disrupted
have only subtle effects. The work exemplifies the use of transcriptional profiling as a first line
of screening, which can then be followed by reverse genetics to discover new genes in a given
pathway.
5
MATERIALS AND METHODS
Drosophila stocks: The following gene alleles and transgenes were used: vnL6, vnddd3,
spi1, Egfr3F18, krn27, pnt
Δ88
, Ras85De1B, UAS-vn1.1, UAS-Dcr-2, EgfrElp, 71B-GAL4, Act5C-GAL4,
en-GAL4, sd-GAL4, tsh-GAL4, vn-GAL4 (CLA unpublished data). Most RNAi transgenes were
from the KK or GD collections at the Vienna Drosophila RNAi Center (VDRC
http://stockcenter.vdrc.at/control/main). CG31048RNAi was from the TRIP collection available at
the Bloomington Drosophila Stock Center. Additional transgenes generated here are described
below. GAL4-UAS crosses were carried out at 17°, 25° and 29° to provide a range of GAL4
activity levels. Trangenes for Sulf1 were generated as part of this study and also obtained from
Hiroshi Nakato; UAS-Sulf1, UAS-Sulf1-Golgi, UAS-Sulf1-ER (KLEINSCHMIT et al. 2010). The
source of a Sulf1 transgene used in a given experiment is indicated in the text.
Microarray Processing and Analysis: RNA was extracted from third instar wing discs
dissected from 71B-GAL4; UAS EgfrACT and 71B-Gal4; UAS EgfrDN larvae, with about 200
larvae for each genotype. RNA quality was preserved by placing dissected wing discs in groups
of about 10 directly into RLT buffer (Qiagen) on ice for subsequent RNA extractions. RNA
samples were processed and hybridized to Drosophila Genome 1 Arrays (three arrays per
genotype) using standard Affymetrix protocols at the Microarray Shared Resource of the
Comprehensive Cancer Center at The Ohio State University. The R environment (http://www.rproject.org) and BioConductor suite ((GENTLEMAN et al. 2004); http://www.bioconductor.org)
were used for all data analysis. Scanned image files were processed using the Robust Multiarray
Average (RMA) method (BOLSTAD et al. 2003) to normalize across datasets and to calculate
expression values. Genes showing log2 expression values of 8.2 or higher in either or both of the
Egfr groups were retained for statistical analysis, as previous work (BUTLER et al. 2003) had
6
found the most reliable in situ hybridization results with genes expressed at these levels.
Following this filtering, two-tailed t-tests were used to identify genes differentially expressed
between the EgfrACT versus EgfrDN groups. Differentially expressed transcripts were ranked by
fold-change to generate a working list for subsequent biological analyses.
In situ hybridization: RNA probes were generated by transcription of antisense RNA
with T7 RNA polymerase (Roche) from plasmid templates with cDNA inserts from the
Drosophila Genomics Resource Center (DGRC; https://dgrc.cgb.indiana.edu/) or from genespecific PCR-generated templates. For PCR templates the 3' primer for each gene had a T7
recognition sequence: GAATTTAATACGACTCACTATAGG. All hybridizations were carried
out using a protocol involving proteinase K treatment of the tissues (BUTLER et al. 2003) or
using a method with a higher hybridization temperature that does not include this step (FIRTH
and BAKER 2007).
Immunohistochemistry: Late third instar wing discs were stained for dpERK with rabbit
anti-dpERK antibody (Cell Signaling; (GABAY et al. 1997)) using methods as described in
Coppey et al. (COPPEY et al. 2008). Primary antibody was used at 1:250 dilution and Cy3
conjugated goat anti-rabbit antibody (Jackson Immunoresearch) was used at 1:1000 dilution for
detection. Samples were mounted in VECTASHIELD (Vector Laboratories) and imaged using a
Nikon C90i confocal microscope.
Transgenes: We generated transgenes for ectopic expression of Sulf1 and CG6234. All
transgenes were cloned into pUAST and transgenic flies were generated using standard Pelement transformation. Three or more lines were generated and examined for each gene. Sulf1:
the DGC cDNA clone SD04414 that was used to make this construct lacks the first 315 bp of
coding sequence. Genomic PCR was used to generate the missing sequence, which was
7
combined by overlapping PCR with the cDNA clone to give the full coding sequence. CG6234:
an Xho I fragment comprising the entire cDNA of the gene was excised from the DGC cDNA
LP04345 clone.
RT-PCR: S2 and S2-DER cells (SCHWEITZER et al. 1995) were treated for 7h with
CuSO4 at a final concentration of 700 µM to induce Egfr expression in the S2-DER cells.
Controls were mock treated with saline. Cells were harvested and total RNA was isolated using
an RNeasy kit (Qiagen) with an in-column DNase I treatment. RNA (2 µg) from each sample
was subjected to reverse transcription (RT) using an Omniscript RT kit (Qiagen). A fraction
(1/10) of the RT reaction was used as template for PCR reactions with gene-specific primers (25
cycles). Ornithine decarboxylase antizyme (Oda), which is expressed at similar levels in all cells,
was used as a standard for comparison. Primers for each gene were as follows:
Sulf1: 5’ GTCCAGGGATTCAGGCATCTAAGG 3’; 5’ ATCCTCCGGCTGCTGAGC 3’
CG4096: 5'- CCAGGGTGCCACCTACAAG -3'; 5' GAGAATGTGCCGCGC 3'
Oda: 5’ GTCCTTCGGTAGAGCGACAT 3’; 5’ GCACCATCTCGACTTCGTCT 3’
Heparin binding assay: Assays for ligand binding to heparin were performed as
previously described (KLEIN et al. 2004). S2-Vn (SCHNEPP et al. 1998), S2-sSpi::GFP and S2sSpi::His cells (TLJ unpublished) were treated for 7h with CuSO4 at a final concentration of 700
µM to induce gene expression. The conditioned medium (7 mL) from each culture was
concentrated (7 fold) using a 10 kDa cut-off spin filter (Amicon). 1 mL of this concentrate was
incubated with 100 µL of agarose beads covalently linked to heparin (pre-equilibrated with
dilution buffer (20 mM Tris, pH 7.2)) and incubated overnight at 4°. The beads were washed five
times using low salt buffer (50 mM NaCl in dilution buffer). Bound proteins were eluted with 50
8
µL of high salt buffer (900 mM NaCl in dilution buffer) and samples were analyzed using
Western blotting and probing with affinity purified α-Vn antibody (SCHNEPP et al. 1996) or αSpi (Developmental Studies Hybridoma Bank; (SCHWEITZER et al. 1995)) followed by HRP
conjugated
secondary
antibodies
(goat
anti-Rabbit
or
goat
anti-Mouse,
Jackson
Immunoresearch). Each ‘input’ lane contained 1/100 of the total volume that was added to the
beads for binding. Each ‘bound’ lane contained 1/5 of the total eluate from the high salt wash.
9
RESULTS
Transcriptome analysis of wing discs identified known and candidate Egfr-responsive
genes: The 71B-GAL4 driver (BRAND and PERRIMON 1993) was used to induce expression of
transgenes encoding either constitutively active (UAS-EgfrACT) (QUEENAN et al. 1997), or
dominant negative (UAS-EgfrDN) (BUFF et al. 1998) forms of Egfr in the wing disc (Figure 1 AE). 71B-GAL4 is expressed throughout the wing pouch, with slight down regulation at the AP
and DV boundaries (Figure 1B). This is broader than the normal domain of Egfr signaling in the
wing pouch, which is limited to a series of stripes corresponding to the future veins and wing
margin, as visualized by the pattern of dpErk expression (doubly-phosphorylated Erk/MAPK;
Figure 1 A; (GABAY et al. 1997)) or the vein cell markers aos and rho (Figure 1 F and I). In
response to altered signaling, the adult flies had vein-patterning defects; 71B-GAL4; UAS
EgfrACT flies had extra veins and 71B-Gal4; UAS EgfrDN flies had no veins (Figure 1 C-E).
In addition to broad expression, we also chose 71B-GAL4 for this analysis because 71BGal4 expression is initiated at the early-late third instar. This ensures that the time between
induction of the transgenes and harvesting at late-late third instar is only about 48 hours
((WESSELLS et al. 1999); Figure S1 A-C). The short time frame increases the likelihood that
direct Egfr targets will be identified over genes that function further downstream in the gene
regulatory network. These latter genes would include those that are induced or repressed once
cell type reprogramming has occurred. Indeed, we found that aos, a known transcriptional target
was strongly and uniformly induced (Figure 1 F-H), whereas, rho, a gene expressed in all cells of
vein histotype, showed little induction by the time of harvesting (Figure 1 I-K).
Analysis of the microarray data identified a total of 162 transcripts that were expressed at
10
significantly different levels (p≤0.05) between the EgfrACT and EgfrDN samples. Table 1 shows
the top 25 transcripts upregulated by EgfrACT, and Table 2 shows the top 25 transcripts
upregulated by EgfrDN (the full set is shown in Table S1 and the primary data have been
submitted to GEO (GSE34872)). Eight of these genes were also identified in other studies that
conducted whole-genome transcriptional analysis of the Egfr pathway or the Ras gene, which
functions downstream of Egfr (Tables 1 and 2; (ASHA et al. 2003; FIRTH and BAKER 2007;
JORDAN et al. 2005)). Five known target genes were also represented within the top group of 25
genes upregulated by Egfr: ventral veinless (vvl) (DE CELIS et al. 1995; LLIMARGAS and
CASANOVA 1997), sty (CASCI et al. 1999; HACOHEN et al. 1998), pnt (GABAY et al. 1996;
O'NEILL et al. 1994), MASK (SMITH et al. 2002) and Mkp3 (GOMEZ et al. 2005; KIM et al. 2003;
RINTELEN et al. 2003). Finding differential expression of these known targets suggested that the
experimental approach itself was sound and had the potential to identify novel targets.
Functional analysis of the most differentially expressed genes using reverse genetics
identifies genes with roles in vein patterning: We used RNAi to test the function of 29 of the
50 most differentially expressed genes that were either poorly characterized or had unknown
functions (Tables 1 and 2). When ubiquitously expressed using Act5C-GAL4 (Figure S1D),
RNAi transgenes targeting 14/29 genes caused phenotypes—lethality (9), lower viability (2), and
visible phenotypes (3) (Tables 1, 2 and S2). Knockdown of CG4096 and CG4382 caused an
extra-vein phenotype (Figure 2, B and C), which was of particular interest because abnormal
vein phenotypes can indicate a function in the Egfr pathway.
To further test the 29 genes for vein phenotypes, including those for which ubiquitous
RNAi was lethal, we used a number of tissue-specific GAL4 drivers to inhibit gene expression in
a more restricted pattern (Table S2). Knockdown of CG4096 and CG4382 showed consistent
11
extra-vein phenotypes and CG34398 showed a consistent loss of vein phenotype with a variety
of GAL4 lines (Tables 3 and S2). An additional seven genes showed a loss of the anterior
crossvein (ACV) when RNAi expression was driven with vn-GAL4 (Table S2). The ACV is very
sensitive to Vn/Egfr signaling (CLIFFORD and SCHUPBACH 1989; SCHNEPP et al. 1996) and as vnGAL4 is a loss-of-function allele in the vn ligand gene (CLA, unpublished observation) it may
provide a sensitized genetic background for revealing genes functioning in the Egfr pathway.
These seven genes, therefore, warrant further analysis, but here we focused on the genes showing
vein phenotypes that were apparent with drivers in addition to vn-GAL4.
We also tested for phenotypic effects of overexpression of CG6234 and Sulf1, using cDNA
transgenes. We analyzed these genes further because in situ analysis (see below) demonstrated
that both genes were expressed in vein primordia, a major target tissue for Egfr signaling in the
wing disc. Ectopic expression of CG6234 with multiple drivers resulted in an extra crossvein
phenotype (Figure 2D; Tables 3 and S2). Sulf1 causes a vein loss phenotype when strongly
overexpressed (KLEINSCHMIT et al. 2010). Using en-GAL4 (AZA-BLANC et al. 1997) to
misexpress the wild-type Sulf1 transgenes we generated (Figure 5A), or those of Kleinschmit et
al., and growing the flies at 25° (or 29°), caused no vein-loss phenotype ((KLEINSCHMIT et al.
2010); Figure S2A). However, an altered form of Sulf1 that is confined to the Golgi, did cause a
strong vein-loss phenotype consistent with a role for Sulf1 in vein patterning ((KLEINSCHMIT et
al. 2010); Figure S2C). The Golgi-tethered form also had a stronger effect on Wingless signaling
(KLEINSCHMIT et al. 2010), which suggests that retaining the enzyme in the Golgi may enhance
its function. We also found loss of the ACV when Sulf1 was misexpressed with the vn-GAL4
driver, which likely provides a sensitized background because vn-GAL4 is a mutant allele (Table
3; Figure 2E). Together these data are consistent with Sulf1 acting as a negative regulator of vein
12
development.
In situ hybridization confirms the response to Egfr signaling and identifies three
genes expressed in vein primordia, an Egfr-signaling target tissue: The expression patterns
of the three genes that showed venation defects with RNAi and two genes that showed venation
defects following overexpression (Figure 2; Table 3) were determined in wild-type wing discs
and wing discs corresponding to the genotypes examined in the microarray analysis in which
Egfr signaling was altered (Table 3; Figure 3). The genes were expressed in a variety of patterns,
and all showed changes in expression level or distribution, verifying the microarray data. The
expression pattern and how it relates to the observed phenotypes for each of these five genes is
discussed in the next sections.
Sulf1: Sulf1 was expressed in all provein regions and was regulated by Egfr activity
(Figure 3, A-A”). Ectopic Egfr activity promoted expression throughout the wing pouch (Figure
3A’), whereas dominant negative Egfr greatly reduced expression in the central domain where
71B-GAL4 was expressed (Figure 3A”). Sulf1 was also expressed robustly in S2 tissue-culture
cells stably transfected with Egfr (S2-DER cells), which were induced to activate the Egfr
pathway ((SCHWEITZER et al. 1995; ZAK and SHILO 1990); Figure S3). Sequence analysis also
shows the presence of ETS binding sites in the Sulf1 gene that could mediate signaling through
Egfr signaling (Figure S4). ETS family transcription factors such as Pointed mediate
Egfr/Ras/MAPK signaling (O'NEILL et al. 1994). Induction by Egfr signaling has recently been
shown to be important for the role of Sulf1 in the Hh signaling pathway (WOJCINSKI et al. 2011).
RNAi of Sulf1 had no effect and amorphic mutants also have only subtle phenotypes
(Table 2; (KLEINSCHMIT et al. 2010; YOU et al. 2011)). Sulf1 overexpression, however, causes
vein loss (Table 3; Figure 2E; (KLEINSCHMIT et al. 2010)). In further genetic tests, we showed
13
that Sulf1 behaves functionally as a negative regulator of Egfr signaling, most likely by
modulating activity of the ligand Vn (see below).
CG4096: CG4096 encodes an ADAMTS (A Disintegrin And Metalloproteinase with
ThromboSpondin motifs), a family of secreted enzymes with conserved domains conferring a
variety of different functions (PORTER et al. 2005). CG4096 is expressed in presumptive veins
L3 and L4 (Figure 3B). Expression of CG4096 was more extensive in discs expressing EgfrACT
(Figure 3B’); however, ectopic expression is not seen in the entire 71B-GAL4 expression
domain. The specific expression of CG4096 in a subset of veins in wild type and its limited
response to increased Egfr activity suggests that other factors may also be required for its
expression. Also consistent with being a target of the pathway, CG4096 was expressed robustly
in S2 tissue-culture cells stably transfected with Egfr (S2-DER cells), which were induced to
activate the Egfr pathway ((SCHWEITZER et al. 1995; ZAK and SHILO 1990); Figure S3).
Sequence analysis also shows the presence of ETS binding sites in the CG4096 gene that could
mediate signaling through Egfr signaling (Figure S4).
RNAi silencing of CG4096 gave a consistent extra-vein phenotype with an additional vein
fragment above longitudinal vein 2 and some vein deltas (Figure 2C). These phenotypes are
typical of elevated Egfr signaling and seen in flies with the hypermorphic EgfrElp allele (Figure
4B). The phenotype was not enhanced in flies co-expressing Dcr-2, which increases RNAi
effects (CG4096dsRNA/UAS-Dcr-2; Act5C-GAL4, not shown) (DIETZL et al. 2007). Generating a
null mutant, however, will be required to determine the full loss-of-function phenotype. In
further genetic tests, we showed CG4096 behaves functionally as a negative regulator of Egfr
signaling most likely by modulating ligand activity (see below).
CG6234: CG6234 is expressed in all provein regions (Figure 3C). Ectopic activity of Egfr
14
elicits expression in the 71B-GAL4 pattern (Figure 3C’), and inhibition represses expression
(Figure 3C”). CG6234 is also known to be regulated by Wg signaling (BHAMBHANI et al. 2011;
FANG et al. 2006). CG6234 encodes a protein with a predicted signal peptide, but the function of
the CG6234 protein has not been elucidated (Table S3). Ubiquitous downregulation of CG6234
by RNAi caused lethality (Table 1), but no venation defects were seen using more specific
drivers including vn-GAL4, which is likely to provide a sensitized background for observing vein
defects (Tables 3 and S2). Overexpression of CG6234, however, caused an ectopic crossvein
phenotype inducing additional ACVs and/or posterior crossveins (PCVs) (Figure 2D; Tables 3
and S2). This is distinct from that seen following widespread activation of the Egfr pathway, in
which there is a more global extra-vein phenotype (Figure 1D). The crossveins form during
pupal development and, while multiple pathways are involved, there is a major role for signaling
by the Bone Morphogenetic Pathway (BMP) (O'CONNOR et al. 2006). The ectopic crossvein
phenotype seen after overexpression of CG6234 suggests that although the gene is strongly
induced by Egfr signaling, it may function in the BMP pathway. Further molecular analysis is
required to test this hypothesis.
CG34398: CG34398 is an uncharacterized gene containing no known domains apart from
predicted coiled-coil regions. There is an Anopheles gambiae ortholog for CG34398, but no clear
vertebrate ortholog (Table S3). The gene is expressed broadly and exhibits increased expression
in 71B-GAL4; UAS EgfrACT discs and reduced expression in 71B-GAL4; UAS EgfrDN discs
(Figure 3, D-D”). Downregulation of CG34398 via RNAi resulted in the loss of the ACV and/or
PCV with a number of drivers including vn-GAL4 (Tables 3 and S2; Figure 2F). Together these
data suggest that CG34398 is a transcriptional target of the Egfr pathway and plays a positive
role in vein patterning.
15
CG4382: CG4382 encodes a glutathione S-transferase (ALIAS and CLARK 2007).
CG4382 showed robust expression in the peripodial membrane, which was strongly repressed by
Egfr signaling in 71B-GAL4; UAS-EgfrACT individuals (Figure 3, E-E”). The peripodial
membrane is a second cell layer that overlays the disc proper (columnar epithelium). 71B-GAL is
expressed in the peripodial membrane (Figure S1C) and, therefore, there may be direct inhibition
of CG4382 expression in these cells. On the other hand, activation may occur via signaling from
the columnar epithelium. RNAi against CG4382 using expression with multiple different GAL4
drivers resulted in an extra-vein phenotype (Figure 2B; Tables 2 and S2). Together the data
support the idea that CG4382 functions as a negative regulator of vein development, which is
repressed by high Egfr signaling. The expression of CG4382 in the peripodial membrane
overlying the disc proper suggests that its function in venation appears to involve communication
between two cell layers, as had been noted for patterning of the major regions of the wing disc
(BAENA-LOPEZ et al. 2003; PALLAVI and SHASHIDHARA 2003). Our results suggest the peripodial
membrane may also have a more specific role in vein patterning.
Genetic interactions with EgfrElp link CG4096 and Sulf1 to the Egfr pathway: In
order to test for a functional link between the candidate genes and the Egfr pathway, we looked
for modifcation of the phenotype caused by a hypermorphic Egfr allele, EgfrElp. Candidate genes
(CGX) were either silenced by RNAi (EgfrElp; Act5C-GAL4; UAS-CGXRNAi) or overexpressed
(EgfrElp; Act5C-GAL4; UAS-CGX) in EgfrElp heterozygotes. EgfrElp flies have small eyes and an
extra vein phenotype due to constitutive signaling (Figure 4B). All 29 genes were tested although
not all crosses yielded viable adults. Of the nine RNAi lines that were lethal with Act5C-GAL4
(Tables 1 and 2), we were able to examine all but two as pharate adults for gross defects and eye
phenotypes (CG5800 and CG10200 died prior to this stage). From the 27 genes examined, two
16
showed changes in the EgfrElp phenotype: downregulation of CG4096 by RNAi enhanced the
extra-vein phenotype (Figure 4, A-D) and overexpression of Sulf1 caused suppression of the
extra-vein phenotype (Figure 4, E and F). We did not detect any changes in eye phenotype.
Based on the genetic interaction with Egfr we decided to examine Sulf1 and CG4096 in more
detail.
Genetic interaction suggests Sulf1 modulates Vn activity: vn encodes a neuregulin-like ligand
that activates Egfr in the wing (SCHNEPP et al. 1996). Phenotypes caused by missexpression of
vn can be suppressed by co-expression of the inhibitor aos (WESSELLS et al. 1999).
Misexpression of a vn transgene with en-GAL4 (en-GAL4; UAS-vn1.1) caused lethality mainly at
the pupal stage with only 6% of animals surviving long enough to die as pharate adults with
severe thoracic defects (Figure 5B). Remarkably, co-expression of Sulf1 rescued these
individuals such that some survived to adulthood. We tested three independent lines of UASSulf1 and found 14%, 20% or 23% of the expected number of en-GAL4; UAS-vn1.1; UAS-Sulf1
individuals survived to become adults. (The expected number was based on the survival of sibs
without the UAS-vn1.1 gene.) Survivors had extra wing veins and blisters in the posterior region,
indicative of vn overactivity (Figure 5C). The rescue suggests that overexpression of Sulf1,
which likely encodes a pathway inhibitor, compensated for overproduction of the ligand Vn. In
order to eliminate the possibility that the rescue was non-specific and due to a dilution of GAL4
levels caused by competition for binding when an additional UAS-transgene was present, we
also examined a control genotype (en-GAL4; UAS-vn1.1; UAS-GFP); these animals did not
survive. Sulf1 overexpression was unable to rescue the lethality resulting from similar
overexpression of the TGF-α ligand sspi (secreted spi). sSpi is a more potent ligand than Vn
(SCHNEPP et al. 1998) and en-GAL4; UAS-sspi animals died as embryos. Due to the early
17
lethality, this test may be too stringent to rule out an effect of Sulf1 on sSpi function.
At the biochemical level, the effects of Sulf1 on Vn may be an indirect consequence of
the endosulfatase activity of Sulf1 (DHOOT et al. 2001), which modifies heparan sulfate (HS)
chains by removing 6-O-sulfate groups. If the function of Vn is enhanced by binding to HS
chains, specifically at 6-O-sulfate moieties, then the level of Sulf1 could modulate Vn activity by
removing these binding sites. In keeping with this idea we found that Vn bound heparin (a highly
sulfated glycosaminoglycan) in an in vitro assay (Figure 5D). The interaction may be limited to
the ligand Vn, as we did not observe binding of sSpi to heparin (Figure 5D; (KLEIN et al. 2004)).
Genetic analysis suggests CG4096 modulates Egfr signaling at the level of ligand action:
The extra-vein phenotype resulting from RNAi of CG4096 is consistent with the idea that it
functions as a negative regulator of Egfr signaling. In order to test for genetic interactions
between CG4096 and genes in the pathway, we determined whether reducing the dose of these
genes suppressed the extra-vein phenotype (Act5C-GAL4; UAS-4096RNAi; geneX
-/+
). Reducing
the dose of Egfr, pnt (a transcription factor that mediates Egfr signaling), or any single
zygotically-active ligand (spi, vn or krn) had no significant effect on the extra-vein phenotype
(Table 4). Reducing the dose of Ras85D suppressed the phenotype (Table 4). Ras85D is a dose
sensitive component in the Egfr pathway (SIMON et al. 1991). We also tested all possible double
combinations of the ligands. Two, spi; krn and krn, vn, showed a significant suppression of the
extra-vein phenotype (Table 4). The suppression, however, was pronounced when the dose of
three ligand genes was reduced simultaneously (Table 4). Both the predicted secreted nature of
CG4096 and genetic evidence showing suppression at the level of the ligands, support the
hypothesis that CG4096 negatively regulates Egfr signaling by masking ligand activity either
directly or indirectly (Figure 7).
18
DISCUSSION
Transcriptional profiling of Egfr signaling identified 162 genes that represent a
typical genomic cross section of molecular functions: We used a whole-genome microarray
assay to detect transcripts that responded to changed activity of the Egfr pathway in the
Drosophila wing disc. The GO classifications of the 130 genes for which there was information
in the GO database showed a distribution of molecular functions that closely mirrors the
frequencies of each category in the whole the genome (Table S1; Figure S4). This is in contrast
to a large-scale gain-of-function genetic screen for genes involved in vein patterning (MOLNAR et
al. 2006), where the distribution of gene functions was biased towards recovering those encoding
transcription factors and cell signaling molecules (MOLNAR et al. 2006). The genes Molnar et al.
recovered, which included 60% of known genes in the Egfr and other major signaling, were
identified based on their ability to cause a vein phenotype and this increased the recovery of
control genes. Using reverse genetic analysis only 17% (5/29) of the genes we characterized as
having robust transcriptional changes caused vein patterning defects (Tables 1-3). We did,
however, discover two new feedback regulators as described in a following section.
Among the remaining 32 genes with non-annotated molecular functions, 21 (65%) are
predicted to encode either secreted or transmembrane proteins (Table S3). This exceeds the
frequency of such proteins in the whole genome where genes encoding secreted proteins
comprise 19% and genes encoding transmembrane proteins comprise 3.4% of the genome
(http://www.pantherdb.org). BLASTP analysis indicates that 24 of the 32 genes are conserved in
only in insects (17) or unique to Drosophilidae (9) (Table S3). We tested the function of eight of
these genes using RNAi and five showed wing defects (Tables S1 and S3). As a group these 24
genes should be interesting to analyze because over half of those tested showed wing
19
phenotypes, a large proportion are secreted, and they are unique to insects.
Genetic analysis of top candidates led to the discovery of two new negative feedback
regulators of Egfr signaling: We analyzed 29 Egfr-responsive genes using RNAi. This allowed
us to screen rapidly through the candidates. Additional characterization identified two genes,
Sulf1 and CG4096, which behave genetically as negative regulators of Egfr signaling. The
potential role of these genes in the Egfr pathway is discussed next.
Sulf1 is an enzyme that modifies heparan sulfate (HS). The binding of growth factors to
HS chains attached to a protein backbone (HSPG) is important for regulating growth factor
distribution, activity and interaction with other molecules, including receptors (SARRAZIN et al.
2011). HS chains are synthesized in the Golgi and then remodeled by enzymes including
sulfatases that remove specific sulfate groups. These enzymes play important roles because the
final sulfation pattern is a determinant of ligand binding. Products of the Sulfated genes (Sulfs)
are endosulfatases that remove 6-O-sulfate groups from trisulfated glucosamine units (AI et al.
2006; AI et al. 2003; DHOOT et al. 2001; MORIMOTO-TOMITA et al. 2002; SHILATIFARD and
CUMMINGS 1994). There are two vertebrate genes: Sulf1 and Sulf2, and a single Drosophila gene,
Sulf1. The genetic characterization of the mouse and Drosophila genes has confirmed the ability
of the Sulf genes to act as endosulfatases by showing that Sulf loss-of-function mutants
accumulate tri-sulfated disaccharides (KLEINSCHMIT et al. 2010; LAMANNA et al. 2006).
Genetic analysis of Sulfs in vertebrates revealed that the genes have redundant biochemical
functions and are not required for development, though double mutant (Sulf1-/-; Sulf2-/-) mice
died soon after birth with low body weight caused by a defect in innervation of the esophagus
(AI et al. 2007; HOLST et al. 2007; LAMANNA et al. 2006; LUM et al. 2007). In vertebrates, the
Sulfs have been linked to multiple signaling pathways, including FGF, VEGF, WNT, BMP, HH,
20
and EGF and have also been found to be important in cancer (LAI et al. 2008; ROSEN and
LEMJABBAR-ALAOUI 2010). In Drosophila, Sulf1 mutants are adult viable and fertile, but have
subtle phenotypes that demonstrate a role for the genes in Wg and Hh signaling (KLEINSCHMIT et
al. 2010; WOJCINSKI et al. 2011; YOU et al. 2011). Overexpression of Drosophila Sulf1 also
suggests it has a role in FGF signaling (KAMIMURA et al. 2006). Here we provided the first
evidence of a role for Sulf1 and HSPGs in EGF signaling in Drosophila.
We found Sulf1 is a target of Egfr signaling in the Drosophila wing. An observation also
reported by Wojcinski et al. in their analysis of Sulf1 in Hh signaling (WOJCINSKI et al. 2011).
We discovered that Sulf1 is not only induced by Egfr but also functions directly in the pathway
as a negative feedback regulator to repress signaling. Overexpression of Sulf1 suppressed the
EgfrElp phenotype (Figure 4, E and F) and rescued lethality caused by overexpression of vn
(Figure 5A-C). Ectopic expression of Sulf1 caused vein loss, which is characteristic of reduced
Egfr activity and co-expression of vn restored veins (Figures 2E and S2, C and D; (KLEINSCHMIT
et al. 2010)). This genetic evidence is consistent with Sulf1 modulating Vn activity and is
supported by the observation that Vn binds heparin (Figure 5D).
Based on these results, we hypothesize that Sulf1 reduces Vn binding to HSPGs by
removal of 6-O sulfate moieties and hence effects it localization in the ECM (Figure 7). Some
vertebrate EGF ligands are known to bind HSPGs including some neuregulins (NRGs) and
heparin-binding EGF (IWAMOTO et al. 2010; MAHTOUK et al. 2006; PANKONIN et al. 2005). The
NRGs share structural similarity with Vn because both types of growth factors have an Ig
domain in addition to the EGF domain (SCHNEPP et al. 1996). The Ig domain in NRG is required
for binding to heparin, and sulfate groups including 6-O sulfate groups play a role in the
interaction (LI and LOEB 2001; PANKONIN et al. 2005). The addition of a Drosophila EGF ligand
21
to the collection of known ligands regulated by Sulf1 highlights the broad role HSPGs play in
signaling pathways. It will be important to determine more about the Vn-HSPG interaction
including discovering which proteoglycan is involved.
CG4096 contains predicted protein domains characteristic of an ADAMTS family
member, including a zinc-dependent protease and three thrombospondin-like repeats (Figure 6).
It is one of three genes in Drosophila belonging to the ADAMTS family (Figure 6; (NICHOLSON
et al. 2005)). Only one of these, stall, has been analyzed genetically and found to be involved in
ovary development (OZDOWSKI et al. 2009). In mammals there are 19 ADAMTS genes with
diverse biological roles in the extracellular matrix (ECM) (reviewed in (A PTE 2009; PORTER et al.
2005; STANTON et al. 2011)). The genes have also been implicated in diseases including
atherosclerosis, arthritis, and cancer (reviewed in (LE G OFF and CORMIER-D AIRE 2011; LIN and
LIU 2010; SALTER et al. 2010; WAGSTAFF et al. 2011)).
Sequence analysis places CG4096 closest to human ADAMTS7 and ADAMTS12
(NICHOLSON et al. 2005). But CG4096 does not possess a protease and lacunin (PLAC) domain
or a mucin domain, which are seen in mammalian genes (Figure 6; (A PTE 2009; PORTER et al.
2005). Based on domain architecture (predicted using Prodom), CG4096 most closely resembles
ADAMTS1 and its sub-family of proteins ADAMTS4/5/8/15. Of these ADAMTS15 is most
closely related to CG4096 by sequence (Figure 6; NICHOLSON et al. 2005).
Metalloproteases in general are well characterized for their positive roles in cancer
progression through the ability to degrade the ECM and facilitate metastasis. Evidence, however,
is emerging that ADAMTS proteins can also function as tumor suppressors. ADAMTS1,
ADAMTS12 and ADAMTS15 act as tumor suppressors in prostate, colon and breast cancer (EL
22
H OUR et al. 2010; M OLOKWU et al. 2010; M ONCADA-P AZOS et al. 2009; PORTER et al. 2006;
VILORIA et al. 2009). Our genetic evidence shows CG4096 has an inhibitory effect on Egfr
activity. By extension to the role of Egfr/Ras in tumors this would be considered a tumor
suppressor function. Interestingly, there is also a connection to the Egfr pathway in mammals
where ADAMTS1 acts as an activator by promoting the shedding of heparin binding EGFligands (LIU et al. 2006; LU et al. 2009; RICCIARDELLI et al. 2011). In contrast, and in keeping
with the inhibitory function of CG4096, it has also been suggested that a self-cleaved product of
ADAMTS1 could act as a repressor by sequestering ligands (LIU et al. 2006).
It is intriguing to note both the tumor suppressor function and the connection to Egfr
signaling of some mammalian ADAMTS genes. With this in mind, CG4096 could function like
the cleaved form of ADAMTS-1 to sequester the ligands and prevent them from binding the
receptor (Figure 7). It is also possible the effect on ligands is indirect as ADAMTS proteins have
many different ECM substrates (Figure 7). Further molecular and genetic analysis of CG4096
will be important to decipher its role in Drosophila. Any discoveries made in Drosophila are also
likely to further the understanding of the ADAMTS family in other animals including humans.
New feedback controls by genes that play small roles in the Egfr signaling pathway:
The microarray screen described here allowed us to identify two secreted factors that are
negative feedback regulators of the Egfr pathway. We propose that both Sulf1 and CG4096 fine
tune Egfr signaling in the extracellular phase of the signaling pathway by negatively regulating
the interaction between ligands and the receptor (Figure 7). The identification of two new
negative regulators of Egfr signaling highlights the importance of mechanisms that dampen
signaling. Sulf1 mutants are viable with mild morphological changes (KLEINSCHMIT et al. 2010;
23
YOU et al. 2011), and reducing CG4096 function with RNAi has only subtle effects. Yet these
small effects on a vital appendage like a wing could have profound consequences for flies in the
wild. The ability to turn off a pathway is clearly critical for development and homeostasis and as
a result multiple negative regulators exist. This raises the question of how many such controls are
in place and what approaches can be used to find them. Given the genome-wide resources
available for Drosophila, combining transcriptional profiling with reverse genetics is a highly
tractable option that in our experience appears well suited to the discovery of genes with small
effects.
24
FIGURE LEGENDS
FIGURE 1.—Genotypes used in the microarray experiment. (A) Antibody (∝-dpErk) stain of
third instar wing disc to show pattern of Erk/MAPK activation in longitudinal veins (2-5) and the
wing margin (arrowhead). (B) 71B-GAL4; UAS-GFP third instar wing disc showing expression
of GFP in the 71B pattern throughout most of the wing pouch and hinge regions of the disc. The
expression is more extensive than that of dp-Erk (A). (C) Wild-type wing with stereotypical
pattern of veins separated by intervein regions. (D) 71B-GAL4; UAS-EgfrACT wing displaying an
extensive extra-vein pattern. (E) 71B-Gal4; UAS-EgfrDN wing displaying complete vein loss. (FK) in situ hybridization with probes for indicated genes. (F-H) aos in wild type (F) showing
expression in longitudinal veins (3-5) and the wing margin (arrowhead), 71B-GAL4; UASEgfrACT (G) showing expression throughout the wing pouch, and 71B-GAL4; UAS-EgfrDN (H)
showing loss of expression in the wing pouch. (I-K) rho in wild type (I) showing expression in
longitudinal veins (2-5) and the wing margin (arrowhead), 71B-GAL4; UAS-EgfrACT (J) showing
slight expansion of vein width (double headed arrow), and 71B-GAL4; UAS-EgfrDN (K) showing
loss of expression in the wing pouch.
FIGURE 2.—Vein phenotypes displayed by Egfr-responsive genes. (A) Wild-type wing showing
normal wing patterning with longitudinal veins (1-5) and crossveins (ACV and PCV, open
arrows) and indicating other regions for comparison with experimental wings (filled arrows). (B)
Act5C-GAL4; UAS-CG4382RNAi wing, a region of extra-vein material is present above L2 (filled
arrow). (C) Act5C-GAL4; UAS-CG4096RNAi wing, a region of extra vein is present above L2
(filled arrow) and a delta of extra vein is present at the tip of L4 (filled arrow). (D) Act5C-GAL4;
UAS-CG6234 wing, two additional cross veins are present (open arrows). (E) vn-GAL4; UAS25
Sulf1 wing, the ACV is missing (open arrow). (F) vn-GAL4; UAS-CG34398RNAi wing, the ACV
is missing (open arrow). All wings shown are female. In addition to vein defects, wings C-F are
smaller than wild type.
FIGURE 3.—Expression patterns displayed by Egfr-responsive genes. In situ hybridization to
transcripts upregulated (A-D”) or downregulated (E-E”) by Egfr signaling. (A-A”) expression of
Sulf1 in wild type (A) showing expression in longitudinal veins (2-5) and the wing margin
(arrowhead), 71B-GAL4; UAS EgfrACT (A’) showing expression throughout the wing pouch, and
71B-GAL4; UAS EgfrDN (A”) showing loss of expression in the wing pouch. (B-B”) expression
of CG4096 in wild type (B) showing expression in longitudinal veins (3 and 4), 71B-GAL4; UAS
EgfrACT (B’) showing elevated expression in the pouch, pleura and hinge, and 71B-GAL4; UAS
EgfrDN (B”) showing some loss of expression in the pleura and hinge. (C-C”) expression of
CG6234 in wild type (C) showing expression in longitudinal veins (2-5) and the wing margin
(arrowhead), 71B-GAL4; UAS EgfrACT (C’) showing expression throughout the wing pouch, and
71B-GAL4; UAS EgfrDN (C”) showing loss of expression in the wing pouch. (D-D”) expression
of CG34398 in wild type (D) showing expression in broad domains in the disc, 71B-GAL4; UAS
EgfrACT (D’) showing elevated expression, and 71B-GAL4; UAS EgfrDN (D”) showing loss of
expression in the central region of the disc. (E-E”) expression of CG4382 in wild type (E)
showing expression in the peripodial membrane, 71B-GAL4; UAS EgfrACT (E’) showing reduced
expression in the peripodial membrane, and 71B-GAL4; UAS EgfrDN (E”) showing expression in
the peripodial membrane.
FIGURE 4.—CG4096 and Sulf1 interact genetically with the hypermorphic EgfrElp allele. (A)
Act5C-GAL4/+ showing the normal appearance of the distal-anterior portion of the wing blade
(arrows indicate regions for comparison with other genotypes). (B) EgfrElp; Act5C-GAL4 wing
26
with a region of extra-vein material above vein 2 (open arrow). (C) Act5C-GAL4; UASCG4096RNAi wing, a region of extra-vein material is present above vein 2 (open arrow). (D)
EgfrElp; Act5C-GAL4; UAS-CG4096RNAi wing, there is an extensive region of extra-vein material
not present in (B) or (C) (filled arrow). (E) Act5C-GAL4; UAS-Sulf1 wing, exhibiting a normal
pattern. (F) EgfrElp; Act5C-GAL4; UAS-Sulf1 wing showing suppression of the extra-vein
material above vein 2 compare with (B) (open arrow).
FIGURE 5.—Sulf1 interacts genetically with the Egfr ligand vn, which binds heparin. (A) enGAL4; UAS-Sulf1 adult wing showing normal wing pattern. (B) en-GAL4; UAS-vn1.1 rare pharate
adult with thoracic defect (arrow indicates cleft). (C) en-GAL4; UAS-vn1.1; UAS-Sulf1 adult wing
with large posterior blister. (D) Western analysis of Vn, sSpi::GFP, or sSpi::HIS samples
incubated heparin-coated beads. ‘Input’ lanes show starting material. ‘Bound’ lanes show
material eluted from beads in high salt. Only Vn bound to the beads (band in Vn-‘bound’ lane).
Vn lanes probed with α-Vn antibody and Spi::GFP and Spi::His lanes probed with α-Spi
antibody.
FIGURE 6.—Domain architecture of CG4096 and other ADAMTS proteins. The three
Drosophila (Dm) ADAMTS proteins are shown at the top. Seven of the 19 human (Hs)
ADAMTS proteins that are most related by sequence or domain structure to the Drosophila
proteins are shown below. CG4096 is phylogenetically closest to ADAMTS7/12, but lacks the
mucin and PLAC domains. CG4096 has a domain structure most similar to ADAMTS1 and its
subfamily of related proteins ADAMTS4/5/8/15. ADAMTS1, ADAMTS4 and ADAMTS15 are
shown. CG14869 is similar to ADAMTS9 because both have a GON-1 domain. Stall is closest to
ADAMTS16. Conserved functional domains are color coded. The longest isoform of each
protein is shown. Domain analysis was performed using Pfam and the cartoons are drawn to
27
scale.
FIGURE 7.—Working model for the function of Sulf1 and CG4096 in the Egfr pathway. A
schematic of the Egfr pathway is shown with Vn binding to the receptor (the other ligands are
also likely to be involved), which leads to activation of the pathway and expression of the Sulf1
and CG4096 genes. Sulf1 blocks binding of Vn to HSPG by removing 6-O sulfates. We
hypothesize this is inhibitory because binding of Vn to HSPG facilitates interaction with the
receptor. CG4096 negatively regulates pathway by masking the EGF ligands by an unknown
mechanism that may be direct or involve an additional factor (X). X could also be a ligand
activator that CG4096 inhibits.
28
TABLE 1: Upregulated in Egfr
Gene
ImpE1
vvl
ACT
Fold
Molecular function
RNAi
RNAi phenotype
Vein
Hit in Egfr/Ras
change
(Egfr pathway?)
analysis
Act5C-GAL4
phenotype
microarray
unknown
yes
lethal
yes
none
vein loss
Firth and
(RNAi)
Baker 2007
vein loss
Firth and
(OE)
Baker 2007
2.17
2
POU domain TF
(yes)
sty
1.87
unknown (yes)
pnt
1.82
ETS TF (yes)
CG34398
1.66
unknown
nyobe
1.57
PAN / Endoglin
yes
wings held out
yes
none
domains
MASK
1.55
RTK mediator (yes)
Sulf1
1.53
endosulfatase
tou
1.41
chromatin structure
corto
1.39
Polycomb
interactions
qkr58E-1
1.38
RNA binding
bel
1.37
DEAD box helicase
Firth and
Baker 2007
Ect3
1.36
beta-galactosidase
CG5326
1.34
fatty acid elongation
salm
1.33
zinc finger TF
CG4096
1.32
ADAMTS
pk
1.32
yes
none
yes
extra vein
PET LIM domain
29
extra vein
Firth and
(RNAi)
Baker 2007
Cypl
1.31
isomerase
yes
young adult
lethal
CG5800
1.3
DEAD box helicase
egg
1.29
methyltransferase
CG6234
1.29
unknown, pred.
yes
lethal
yes
lethal
transmembrane
CG3857
1.28
unknown
eIF5B
1.27
translation initiation
(OE)
yes
lethal
yes
none
factor
CG31048
1.27
GEF
Mkp3
1.26
MAPK phosphatase
extra vein
(yes)
30
DN
TABLE 2: Upregulated in Egfr
Gene
Fold
Molecular
change
RNAi
RNAi phenotype
Vein
Hit in Egfr/Ras
function
analysis
Act5C-GAL4
phenotype
microarray
CG13053
2.75
unknown
yes
none
CG4382
2.47
carboxylesterase
yes
extra veins
extra vein
(RNAi)
CG31075
2.11
aldehyde
dehydrogenase
Fbp2
2.1
short-chain
dehydrogenase
CG4778
1.95
chitin-binding
yes
none
CG11899
1.89
aminotransferase
yes
none
CG9027
1.86
superoxide
yes
none
dismutase
BM-40-
1.83
calcium binding
CG6448
1.72
dehydrogenase
Ance
1.68
angiotensin
Jordan et al.
2005
SPARC
yes
none
converting
enzyme
GstE1
1.66
glutathione Stransferase
Mgstl
1.66
glutathione Stransferase
CG10359
1.62
unknown
yes
none
CG10200
1.61
unknown
yes
lethal
31
Asha et al. 2003
AnnX
1.56
phospholipid
binding
CG7860
1.54
peptidase T2
yes
lethal
CG17746
1.52
phosphatase
yes
lethal
Idgf4
1.52
growth factor
yes
none
peptidase
yes
low viability
SP1029
1.5
Firth and Baker
2007
Cyp4e2
1.49
cytochrome p450
yes
lethal
Firth and Baker
2007
CG9914
1.48
enoyl-CoA
yes
none
hydratase
CG4210
1.47
acetyltransferase
yes
none
CG9691
1.45
unknown
yes
none
CG12643
1.45
unknown
yes
none
regucalcin
1.43
calcium-
yes
lethal
mediated
signaling
32
TABLE 3: Genes showing vein phenotypes
Gene
RNAi-induced vein
OE-induced vein
Interaction with
phenotype
phenotype
Egfr
CG34398
loss
ND
none
ND
Sulf1
none
loss
Su (OE)
vn
CG4096
extra
ND
E (RNAi)
spi, krn, vn
CG6234
none
extra
none
ND
CG4382
extra
ND
none
ND
Elp
OE, overexpression; ND, not determined; Su, Suppressor; E, Enhancer
33
Interaction with
EGF ligand
TABLE 4: Suppression of Act5C-GAL4; UAS-CG4096
RNAi
extra-vein phenotype with lowered dose
of EGF ligands
Act5C-GAL4; UAS-CG4096
RNAi
Heterozygous mutant allele(s)
Fraction of wings with extra vein (n)
none (control)
0.91 (32)
Egfr
3F18
0.98 (50)
Ras85D
e1B
0.17 (66)**
∆88
0.95 (42)
spi
1
1.00 (26)
3
0.96 (28)
pnt
vn
krn
27
1
spi ; vn
0.97 (62)
L6
0.91 (44)
1
27
0.59 (32)**
L6
27
0.82 (38)*
spi ; krn
vn , krn
1
L6
spi ; vn , krn
27
0.17 (52)**
* significantly different than control P = 0.005
** significantly different than control P < 0.0001
34
ACKNOWLEDGEMENTS
We thank Hiroshi Nakato for Sulf1 transgenic flies, the Bloomington Stock Center for multiple
fly strains, Chiu-Wen Lin for the heparin binding assays and Nicole Werner for conducting some
fly crosses. The anti-sSpi antibody developed by B-Z Shilo was obtained from the
Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and
maintained by The University of Iowa, Department of Biology, Iowa City, IA 52242. Our work
was supported by an award from the National Science Foundation to AS (IBN0920231) and a
Pelotonia fellowship to CLA.
35
LITERATURE CITED
AI, X., A. T. DO, M. KUSCHE-GULLBERG, U. LINDAHL, K. LU et al., 2006 Substrate specificity
and domain functions of extracellular heparan sulfate 6-O-endosulfatases, QSulf1 and
QSulf2. J Biol Chem 281: 4969-4976.
AI, X., A. T. DO, O. LOZYNSKA, M. KUSCHE-GULLBERG, U. LINDAHL et al., 2003 QSulf1
remodels the 6-O sulfation states of cell surface heparan sulfate proteoglycans to promote
Wnt signaling. J Cell Biol 162: 341-351.
AI, X., T. KITAZAWA, A. T. DO, M. KUSCHE-GULLBERG, P. A. LABOSKY et al., 2007 SULF1 and
SULF2 regulate heparan sulfate-mediated GDNF signaling for esophageal innervation.
Development 134: 3327-3338.
ALIAS, Z., and A. G. CLARK, 2007 Studies on the glutathione S-transferase proteome of adult
Drosophila melanogaster: responsiveness to chemical challenge. Proteomics 7: 36183628.
APTE, S. S., 2009 A disintegrin-like and metalloprotease (reprolysin-type) with thrombospondin
type 1 motif (ADAMTS) superfamily: functions and mechanisms. J Biol Chem 284:
31493-31497.
ASHA, H., I. NAGY, G. KOVACS, D. STETSON, I. ANDO et al., 2003 Analysis of Ras-induced
overproliferation in Drosophila hemocytes. Genetics 163: 203-215.
AVRAHAM, R., and Y. YARDEN, 2011 Feedback regulation of EGFR signalling: decision making
by early and delayed loops. Nat Rev Mol Cell Biol 12: 104-117.
AZA-BLANC, P., F. A. RAMIREZ-WEBER, M. P. LAGET, C. SCHWARTZ and T. B. KORNBERG,
1997 Proteolysis that is inhibited by hedgehog targets Cubitus interruptus protein to the
nucleus and converts it to a repressor. Cell 89: 1043-1053.
BAENA-LOPEZ, L. A., J. C. PASTOR-PAREJA and J. RESINO, 2003 Wg and Egfr signalling
antagonise the development of the peripodial epithelium in Drosophila wing discs.
Development 130: 6497-6506.
BHAMBHANI, C., J. L. CHANG, D. L. AKEY and K. M. CADIGAN, 2011 The oligomeric state of
CtBP determines its role as a transcriptional co-activator and co-repressor of Wingless
targets. EMBO J 30: 2031-2043.
BLAIR, S. S., 2007 Wing vein patterning in Drosophila and the analysis of intercellular signaling.
Annu Rev Cell Dev Biol 23: 293-319.
BOLSTAD, B. M., R. A. IRIZARRY, M. ASTRAND and T. P. SPEED, 2003 A comparison of
normalization methods for high density oligonucleotide array data based on variance and
bias. Bioinformatics 19: 185-193.
BRAND, A., and N. PERRIMON, 1993 Targeted gene expression as a means of altering cell fates
and generating dominant phenotypes. Development 118: 401-415.
BUFF, E., A. CARMENA, S. GISSELBRECHT, F. JIMENEZ and A. M. MICHELSON, 1998 Signalling
by the Drosophila epidermal growth factor receptor is required for the specification and
diversification of embryonic muscle progenitors. Development 125: 2075-2086.
BUTLER, M. J., T. L. JACOBSEN, D. M. CAIN, M. G. JARMAN, M. HUBANK et al., 2003 Discovery
of genes with highly restricted expression patterns in the Drosophila wing disc using
DNA oligonucleotide microarrays. Development 130: 659-670.
CASCI, T., J. VINOS and M. FREEMAN, 1999 Sprouty, an intracellular inhibitor of Ras signaling.
Cell 96: 655-665.
36
CLIFFORD, R. J., and T. SCHUPBACH, 1989 Coordinately and differentially mutable activities of
torpedo, the Drosophila melanogaster homolog of the vertebrate EGF receptor gene.
Genetics 123: 771-787.
COPPEY, M., A. N. BOETTIGER, A. M. BEREZHKOVSKII and S. Y. SHVARTSMAN, 2008 Nuclear
trapping shapes the terminal gradient in the Drosophila embryo. Curr Biol 18: 915-919.
DE CELIS, J. F., M. LLIMARGAS and J. CASANOVA, 1995 Ventral veinless, the gene encoding the
Cf1a transcription factor, links positional information and cell differentiation during
embryonic and imaginal development in Drosophila melanogaster. Development 121:
3405-3416.
DHOOT, G. K., M. K. GUSTAFSSON, X. AI, W. SUN, D. M. STANDIFORD et al., 2001 Regulation
of Wnt signaling and embryo patterning by an extracellular sulfatase. Science 293: 16631666.
DIETZL, G., D. CHEN, F. SCHNORRER, K. C. SU, Y. BARINOVA et al., 2007 A genome-wide
transgenic RNAi library for conditional gene inactivation in Drosophila. Nature 448:
151-156.
EL HOUR, M., A. MONCADA-PAZOS, S. BLACHER, A. MASSET, S. CAL et al., 2010 Higher
sensitivity of Adamts12-deficient mice to tumor growth and angiogenesis. Oncogene 29:
3025-3032.
FANG, M., J. LI, T. BLAUWKAMP, C. BHAMBHANI, N. CAMPBELL et al., 2006 C-terminal-binding
protein directly activates and represses Wnt transcriptional targets in Drosophila. EMBO
J 25: 2735-2745.
FIRTH, L. C., and N. E. BAKER, 2007 Spitz from the retina regulates genes transcribed in the
second mitotic wave, peripodial epithelium, glia and plasmatocytes of the Drosophila eye
imaginal disc. Dev Biol 307: 521-538.
GABAY, L., H. SCHOLZ, M. GOLEMBO, A. KLAES, B.-Z. SHILO et al., 1996 EGF receptor
signaling induces pointed P1 transcription and inactivates Yan protein in the Drosophila
embryonic ventral ectoderm. Development 122: 3355-3362.
GABAY, L., R. SEGER and B.-Z. SHILO, 1997 In situ activation pattern of Drosophila EGF
receptor pathway during development. Science 277: 1103-1106.
GENTLEMAN, R. C., V. J. CAREY, D. M. BATES, B. BOLSTAD, M. DETTLING et al., 2004
Bioconductor: open software development for computational biology and bioinformatics.
Genome Biol 5: R80.
GHIGLIONE, C., K. L. R. CARRAWAY, L. T. AMUNDADOTTIR, R. E. BOSWELL, N. PERRIMON et al.,
1999 The transmembrane molecule kekkon 1 acts in a feedback loop to negatively
regulate the activity of the Drosophila EGF receptor during oogenesis. Cell 96: 847-856.
GOLEMBO, M., R. SCHWEITZER, M. FREEMAN and B.-Z. SHILO, 1996 argos transcription is
induced by the Drosophila EGF receptor pathway to form an inhibitory feedback loop.
Development 122: 223-230.
GOLEMBO, M., T. YARNITZKY, T. VOLK and B. Z. SHILO, 1999 Vein expression is induced by the
EGF receptor pathway to provide a positive feedback loop in patterning the Drosophila
embryonic ventral ectoderm. Genes Dev 13: 158-162.
GOMEZ, A. R., A. LOPEZ-VAREA, C. MOLNAR, E. DE LA CALLE-MUSTIENES, M. RUIZ-GOMEZ et
al., 2005 Conserved cross-interactions in Drosophila and Xenopus between Ras/MAPK
signaling and the dual-specificity phosphatase MKP3. Dev Dyn 232: 695-708.
HACOHEN, N., S. KRAMER, D. SUTHERLAND, Y. HIROMI and M. A. KRASNOW, 1998 sprouty
encodes a novel antagonist of FGF signaling that patterns apical branching of the
Drosophila airways. Cell 92: 253-263.
37
HOLST, C. R., H. BOU-RESLAN, B. B. GORE, K. WONG, D. GRANT et al., 2007 Secreted
sulfatases Sulf1 and Sulf2 have overlapping yet essential roles in mouse neonatal
survival. PLoS One 2: e575.
IWAMOTO, R., N. MINE, T. KAWAGUCHI, S. MINAMI, K. SAEKI et al., 2010 HB-EGF function in
cardiac valve development requires interaction with heparan sulfate proteoglycans.
Development 137: 2205-2214.
JORDAN, K. C., S. D. HATFIELD, M. TWOROGER, E. J. WARD, K. A. FISCHER et al., 2005 Genome
wide analysis of transcript levels after perturbation of the EGFR pathway in the
Drosophila ovary. Dev Dyn 232: 709-724.
KAMIMURA, K., T. KOYAMA, H. HABUCHI, R. UEDA, M. MASU et al., 2006 Specific and flexible
roles of heparan sulfate modifications in Drosophila FGF signaling. J Cell Biol 174: 773778.
KIM, S. E., S. H. KIM and K. Y. CHOI, 2003 Regulation of Drosophila MKP-3 by Drosophila
ERK. Ann N Y Acad Sci 1010: 51-61.
KLEIN, D. E., V. M. NAPPI, G. T. REEVES, S. Y. SHVARTSMAN and M. A. LEMMON, 2004 Argos
inhibits epidermal growth factor receptor signalling by ligand sequestration. Nature 430:
1040-1044.
KLEINSCHMIT, A., T. KOYAMA, K. DEJIMA, Y. HAYASHI, K. KAMIMURA et al., 2010 Drosophila
heparan sulfate 6-O endosulfatase regulates Wingless morphogen gradient formation.
Dev Biol 345: 204-214.
LAI, J. P., D. S. SANDHU, A. M. SHIRE and L. R. ROBERTS, 2008 The tumor suppressor function
of human sulfatase 1 (SULF1) in carcinogenesis. J Gastrointest Cancer 39: 149-158.
LAMANNA, W. C., R. J. BALDWIN, M. PADVA, I. KALUS, G. TEN DAM et al., 2006 Heparan
sulfate 6-O-endosulfatases: discrete in vivo activities and functional co-operativity.
Biochem J 400: 63-73.
LE GOFF, C., and V. CORMIER-DAIRE, 2011 The ADAMTS(L) family and human genetic
disorders. Hum Mol Genet 20: R163-167.
LI, Q., and J. A. LOEB, 2001 Neuregulin-heparan-sulfate proteoglycan interactions produce
sustained erbB receptor activation required for the induction of acetylcholine receptors in
muscle. J Biol Chem 276: 38068-38075.
LI, X., and R. W. CARTHEW, 2005 A microRNA mediates EGF receptor signaling and promotes
photoreceptor differentiation in the Drosophila eye. Cell 123: 1267-1277.
LIN, E. A., and C. J. LIU, 2010 The role of ADAMTSs in arthritis. Protein Cell 1: 33-47.
LIU, Y. J., Y. XU and Q. YU, 2006 Full-length ADAMTS-1 and the ADAMTS-1 fragments
display pro- and antimetastatic activity, respectively. Oncogene 25: 2452-2467.
LLIMARGAS, M., and J. CASANOVA, 1997 ventral veinless, a POU domain transcription factor,
regulates different transduction pathways required for tracheal branching in Drosophila.
Development 124: 3273-3281.
LU, X., Q. WANG, G. HU, C. VAN POZNAK, M. FLEISHER et al., 2009 ADAMTS1 and MMP1
proteolytically engage EGF-like ligands in an osteolytic signaling cascade for bone
metastasis. Genes Dev 23: 1882-1894.
LUM, D. H., J. TAN, S. D. ROSEN and Z. WERB, 2007 Gene trap disruption of the mouse heparan
sulfate 6-O-endosulfatase gene, Sulf2. Mol Cell Biol 27: 678-688.
MAHTOUK, K., F. W. CREMER, T. REME, M. JOURDAN, M. BAUDARD et al., 2006 Heparan
sulphate proteoglycans are essential for the myeloma cell growth activity of EGF-family
ligands in multiple myeloma. Oncogene 25: 7180-7191.
MOLNAR, C., A. LOPEZ-VAREA, R. HERNANDEZ and J. F. DE CELIS, 2006 A gain-of-function
38
screen identifying genes required for vein formation in the Drosophila melanogaster
wing. Genetics 174: 1635-1659.
MOLOKWU, C. N., O. O. ADENIJI, S. CHANDRASEKHARAN, F. C. HAMDY and D. J. BUTTLE, 2010
Androgen regulates ADAMTS15 gene expression in prostate cancer cells. Cancer Invest
28: 698-710.
MONCADA-PAZOS, A., A. J. OBAYA, M. F. FRAGA, C. G. VILORIA, G. CAPELLA et al., 2009 The
ADAMTS12 metalloprotease gene is epigenetically silenced in tumor cells and
transcriptionally activated in the stroma during progression of colon cancer. J Cell Sci
122: 2906-2913.
MORIMOTO-TOMITA, M., K. UCHIMURA, Z. WERB, S. HEMMERICH and S. D. ROSEN, 2002
Cloning and characterization of two extracellular heparin-degrading endosulfatases in
mice and humans. J Biol Chem 277: 49175-49185.
NICHOLSON, A. C., S. B. MALIK, J. M. LOGSDON, JR. and E. G. VAN MEIR, 2005 Functional
evolution of ADAMTS genes: evidence from analyses of phylogeny and gene
organization. BMC Evol Biol 5: 11.
O'CONNOR, M. B., D. UMULIS, H. G. OTHMER and S. S. BLAIR, 2006 Shaping BMP morphogen
gradients in the Drosophila embryo and pupal wing. Development 133: 183-193.
O'NEILL, E. M., I. REBAY, R. TIJAN and G. M. RUBIN, 1994 The activities of two Ets-related
transcription factors required for Drosophila eye development are modulated by the
Ras/MAPK pathway. Cell 78: 137-147.
OZDOWSKI, E. F., Y. M. MOWERY and C. CRONMILLER, 2009 Stall encodes an ADAMTS
metalloprotease and interacts genetically with Delta in Drosophila ovarian follicle
formation. Genetics 183: 1027-1040.
PAI, L. M., G. BARCELO and T. SCHUPBACH, 2000 D-cbl, a negative regulator of the Egfr
pathway, is required for dorsoventral patterning in Drosophila oogenesis. Cell 103: 5161.
PALLAVI, S. K., and L. S. SHASHIDHARA, 2003 Egfr/Ras pathway mediates interactions between
peripodial and disc proper cells in Drosophila wing discs. Development 130: 4931-4941.
PANKONIN, M. S., J. T. GALLAGHER and J. A. LOEB, 2005 Specific structural features of heparan
sulfate proteoglycans potentiate neuregulin-1 signaling. J Biol Chem 280: 383-388.
PORTER, S., I. M. CLARK, L. KEVORKIAN and D. R. EDWARDS, 2005 The ADAMTS
metalloproteinases. Biochem J 386: 15-27.
PORTER, S., P. N. SPAN, F. C. SWEEP, V. C. TJAN-HEIJNEN, C. J. PENNINGTON et al., 2006
ADAMTS8 and ADAMTS15 expression predicts survival in human breast carcinoma. Int
J Cancer 118: 1241-1247.
QUEENAN, A. M., A. GHABRIAL and T. SCHUPBACH, 1997 Ectopic activation of torpedo/Egfr, a
Drosophila receptor tyrosine kinase, dorsalizes both the eggshell and the embryo.
Development 124: 3871-3880.
RICCIARDELLI, C., K. M. FREWIN, A. TAN IDE, E. D. WILLIAMS, K. OPESKIN et al., 2011 The
ADAMTS1 Protease Gene Is Required for Mammary Tumor Growth and Metastasis. Am
J Pathol 179: 3075-3085.
RINTELEN, F., E. HAFEN and K. NAIRZ, 2003 The Drosophila dual-specificity ERK phosphatase
DMKP3 cooperates with the ERK tyrosine phosphatase PTP-ER. Development 130:
3479-3490.
ROSEN, S. D., and H. LEMJABBAR-ALAOUI, 2010 Sulf-2: an extracellular modulator of cell
signaling and a cancer target candidate. Expert Opin Ther Targets 14: 935-949.
SALTER, R. C., T. G. ASHLIN, A. P. KWAN and D. P. RAMJI, 2010 ADAMTS proteases: key roles
39
in atherosclerosis? J Mol Med (Berl) 88: 1203-1211.
SARRAZIN, S., W. C. LAMANNA and J. D. ESKO, 2011 Heparan sulfate proteoglycans. Cold
Spring Harb Perspect Biol 3.
SCHNEPP, B., T. DONALDSON, G. GRUMBLING, S. OSTROWSKI, R. SCHWEITZER et al., 1998 EGF
domain swap converts a Drosophila EGF-receptor activator into an inhibitor. Genes &
Dev. 12: 908-913.
SCHNEPP, B., G. GRUMBLING, T. DONALDSON and A. SIMCOX, 1996 Vein is a novel component
in the Drosophila epidermal growth factor receptor pathway with similarity to the
neuregulins. Genes & Dev. 10: 2302-2313.
SCHWEITZER, R., R. HOWES, R. SMITH, B.-Z. SHILO and M. FREEMAN, 1995 Inhibition of
Drosophila EGF receptor activation by the secreted protein Argos. Nature 376: 699-702.
SHILATIFARD, A., and R. D. CUMMINGS, 1994 Purification and characterization of Nacetylglucosamine-6-sulfate sulfatase from bovine kidney: evidence for the presence of a
novel endosulfatase activity. Biochemistry 33: 4273-4282.
SHILO, B. Z., 2005 Regulating the dynamics of EGF receptor signaling in space and time.
Development 132: 4017-4027.
SIMCOX, A. A., G. GRUMBLING, B. SCHNEPP, C. BENNINGTON-MATHIAS, E. HERSPERGER et al.,
1996 Molecular, phenotypic, and expression analysis of vein, a gene required for growth
of the Drosophila wing disc. Dev Biol 177: 475-489.
SIMON, M. A., D. D. BOWTELL, G. S. DODSON, T. R. LAVERTY and G. M. RUBIN, 1991 Ras1 and
a putative guanine nucleotide exchange factor perform crucial steps in signaling by the
sevenless protein tyrosine kinase. Cell 67: 701-716.
SMITH, R. K., P. M. CARROLL, J. D. ALLARD and M. A. SIMON, 2002 MASK, a large ankyrin
repeat and KH domain-containing protein involved in Drosophila receptor tyrosine kinase
signaling. Development 129: 71-82.
STANTON, H., J. MELROSE, C. B. LITTLE and A. J. FOSANG, 2011 Proteoglycan degradation by
the ADAMTS family of proteinases. Biochim Biophys Acta 1812: 1616-1629.
STURTEVANT, M. A., and E. BIER, 1995 Analysis of the genetic hierarchy guiding wing vein
development in Drosophila. Development 121: 785-801.
STURTEVANT, M. A., M. ROARK and E. BIER, 1993 The Drosophila rhomboid gene mediates the
localized formation of wing veins and interacts genetically with components of the EGFR signaling pathway. Genes & Dev. 7: 961-973.
URBAN, S., J. R. LEE and M. FREEMAN, 2001 Drosophila rhomboid-1 defines a family of
putative intramembrane serine proteases. Cell 107: 173-182.
URBAN, S., J. R. LEE and M. FREEMAN, 2002 A family of Rhomboid intramembrane proteases
activates all Drosophila membrane-tethered EGF ligands. Embo J 21: 4277-4286.
VILORIA, C. G., A. J. OBAYA, A. MONCADA-PAZOS, M. LLAMAZARES, A. ASTUDILLO et al., 2009
Genetic inactivation of ADAMTS15 metalloprotease in human colorectal cancer. Cancer
Res 69: 4926-4934.
VIVEKANAND, P., T. L. TOOTLE and I. REBAY, 2004 MAE, a dual regulator of the EGFR
signaling pathway, is a target of the Ets transcription factors PNT and YAN. Mech Dev
121: 1469-1479.
WAGSTAFF, L., R. KELWICK, J. DECOCK and D. R. EDWARDS, 2011 The roles of ADAMTS
metalloproteinases in tumorigenesis and metastasis. Front Biosci 16: 1861-1872.
WANG, S. H., A. SIMCOX and G. CAMPBELL, 2000 Dual role for Drosophila epidermal growth
factor receptor signaling in early wing disc development. Genes Dev 14: 2271-2276.
WASSERMAN, J. D., and M. FREEMAN, 1998 An autoregulatory cascade of EGF receptor
40
signaling patterns the Drosophila egg [In Process Citation]. Cell 95: 355-364.
WESSELLS, R. J., G. GRUMBLING, T. DONALDSON, S. H. WANG and A. SIMCOX, 1999 Tissuespecific regulation of vein/EGF receptor signaling in Drosophila. Dev Biol 216: 243-259.
WOJCINSKI, A., H. NAKATO, C. SOULA and B. GLISE, 2011 DSulfatase-1 fine-tunes Hedgehog
patterning activity through a novel regulatory feedback loop. Dev Biol 358: 168-180.
YOU, J., T. BELENKAYA and X. LIN, 2011 Sulfated is a negative feedback regulator of wingless
in Drosophila. Dev Dyn 240: 640-648.
ZAK, N. B., and B. Z. SHILO, 1990 Biochemical properties of the Drosophila EGF receptor
homolog (DER) protein. Oncogene 5: 1589-1593.
41
Figure 1
42
Figure 2
43
Figure 3
44
Figure 4
45
Figure 5
46
Figure 6
47
Figure 7
48