The Drosophila Tankyrase regulates Wg Signaling depending on the concentration of Daxin

Abstract

The canonical Wnt signaling pathway plays critical roles during development and homeostasis. Dysregulation of this pathway can lead to many human diseases, including cancers. A key process in this pathway consists of regulation of β-catenin concentration through an Axin-recruited destruction complex. Previous studies have demonstrated a role for Tankyrase (TNKS), a protein with poly (ADP-ribose) polymerase, in the regulation of Axin levels in human cancers. However, the role of TNKS in development is still unclear. Here, we have generated a Drosophila tankyrase (DTNKS) mutant and provided compelling evidence that DTNKS is involved in the degradation of Drosophila Axin (Daxin). We show that Daxin physically interacts with DTNKS, and its protein levels are elevated in the absence of DTNKS in eye discs. In S2 cells, DTNKS suppressed the levels of Daxin. Surprisingly, we found that Daxin in turn down-regulated DTNKS protein level. In vivo study showed that DTNKS regulated Wg signaling and wing patterning at a high Daxin protein level, but not at a normal level. Taken together, our findings identified a conserved role of DTNKS in regulating Daxin levels, and thereby Wg/Wnt signaling during development.

1 Introduction

Wingless(Wg)/Wnt signaling is an evolutionarily conserved signaling pathway from invertebrates to vertebrates which plays critical rolesduring embryonic development, stem cell self-renewal, tissue homeostasis, adipogenesis, and neuronal maturation [1–3]. The key process in the canonical Wg/Wnt signaling pathway is the regulation of the concentration of β-catenin/Armadillo by a destruction complex composed of Axin, GSK3/Shaggy, APC and CKI [4–8]. In the absence of the Wg/Wnt ligand, the destruction complex associates with β-catenin/Armadillo and induces its degradation through the ubiquitin-proteasome mechanism [9–12]. When the ligand is present, β-catenin/Armadillo dissociates from the destruction complex and translocates to the nucleus. Within this destruction complex, Axin serves as a central scaffold protein and binds the components of the destruction complex through different domains [5–8, 13]. Previous study showed that Axin was present at a lower concentration than other components of the complex, and overexpression of Axin in cultured cells promoted degradation of β-catenin [14]. Thus, Axin is considered to be a limiting factor during the destruction complex formation, and its concentration is tightly regulated [14].

Tankyrases (TNKSs) are proteins with poly(ADP-ribose) polymerase activity and are evolutionarily conserved in human, mouse, rat, chicken, C.elegans, and Drosophila [15]. Two human TNKSs, hTNKS1 and hTNKS2, share an 85% amino acid identity [16–18]. Both hTNKSs contain an ankyrin repeat domain, a SAM domain, and a PARP domain [16–18]. hTNKSs have been shown to be essential in regulating telomere length. In human cells, hTNKS1 binds to the telomeric-repeat-binding-factor1 (TRF1), a negative regulator of telomere length maintenance [19], removes TRF1 from telomeres and further induces its ubiquitination and degradation [16, 20–23]. In addition, hTNKSs are also involved in GSV trafficking [24–26], spindle structure regulation [27], resolution of sister telomere association [28], and centrosome regulation and maturation [29, 30]. In human cells, hTNKSs mediate PARsylation of Axin1 and Axin2 to regulate the Axin protein at appropriate levels. Knocking-down TNKSs significantly increased the Axin protein levels, and thereby suppressed Wnt signaling [31]. A recent study showed that both human and Drosophila TNKS modulated the activity of the proteasome regulator PI31, and were involved in proteasome assembly [32].

It is less clear about the role of TNKS in development. Homozygous TNKS1 mice and TNKS2 mice are viable, but double mutant mice are embryonic lethal, suggesting that mouse TNKS1 and TNKS2 are functionally redundant [33]. Interestingly, TNKS1^−/− mice fail to show any effect on telomere maintenance, and appear to develop normally [33]. Similar results were obtained from TNKS2^−/− mice regarding telomere regulation [22, 34]. Thus, the role of TNKS during development remains largely unknown and needs to be further examined.

To elucidate the function of TNKS in development, we have generated a Drosophila tankyrase (DTNKS) mutant and examined its role in Wg signaling. We provide compelling evidence that DTNKS is involved in the degradation of Daxin. We show that Daxin interacts with DTNKS and suppreses its protein levels in S2 cells. Daxin levels are increased in the absence of TNKs in the eye disc. Importantly, we found that Daxin in turn down-regulated protein levels of DTNKS. In vivo study showed DTNKS regulated Wg signaling and wing patterning at a high Daxin protein level, but not at normal level. Taken together, our findings identified a conserved role of DTNKS in regulating axin levels, and thereby Wg/Wnt signaling during development.

2. Material and methods

2.1 Drosophila strains

All Drosophila stocks were maintained and crossed at 25°C according to standard procedures. The En-gal4,UAS-GFP, ywhsflp;FRT^82B-ubiGFP, ywhsflp;FRT^82B-hsCD8GFP and SalE-gal4 lines were obtained from the Bloomington stock center. The pW20-DTNKS-RNAi1, pW20-DTNKS-RNAi2 and UAS-V5-DTNKS transgenic fly lines were generated using the PhiC31 integrase-mediated site-specific transgenesis system. The DTNKS^m250 and DTNKS^m23 were generated from fly strain P{EPgHP37069 (BL#22129). The SalE-gal4,UAS-Axin/TM6B line was a gift from Dr H. Song’s lab.

2.2 Generation of transgenic constructs

To generate N-terminal V5-tagged full-length and PARP-domain truncated DTNKS constructs, we amplified the DTNKS cDNA (DGRC #LD22548) by PCR and sub-cloned it into the UAST-attB-V5 vector with XhoI and XbaI sites. The primers were as follows:

• Tank forward: 5′-CCGCTCGAGATGGCCAACAGCAGCCGAAG-3′

• Tank reverse: 5′-GCTCTAGATCATCTTGTATCCTCCGTTCC-3′

• TankΔPARP forward: 5′-CCGCTCGAGATGGCCAACAGCAGCCGAAG-3′

• TankΔPARP reverse: 5′-GCTCTAGATCAATTCACGTTGTTACCAATGC-3′

To obtain the V5 tagged, ankyrin-domain truncated DTNKS construct, we amplified the cDNA fragments from the full-length cDNA by bridge PCR and sub-cloned it into the UAS-attB-V5 vector with XhoI and XbaI sites. The primers were as follows:

• TankΔANK forward-1: 5′-CCGCTCGAGATGGCCAACAGCAGCCGAAG-3′

• TankΔANK reverse-1: 5′-TCCCGCCGTATCCCTGGCGTTC-3′

• TankΔANK forward-2: 5′-GATACGGCGGGA GAGGGGCAGA-3′

• TankΔANK reverse-2: 5′-GCTCTAGATCATCTTGTATCCTCCGTTCC-3′

The PARP-domain truncated construct has a deletion of 961–1181aa, and the ankyrin-domain truncated construct has a deletion of 56–770aa.

A similar strategy was used to generate the UAS-Flag-Daxin and UAS-Flag-Daxin(Δ19–27aa) constructs, which were sub-cloned into the UAS-Flag vector with BglII and XbaI sites. The primers were as follows:

• Daxin forward: 5′-GAAGATCTGATGAGTGGCCATCCATCGGGAATC-3′

• Daxin reverse: 5′-GCTCTAGATTA ATCGGATGGCTTGACAAGACC-3′

• Daxin(Δ19–27aa) forward: 5′-GAAGATCTGATGAGTGGCCATCCATCGGGAATCCGGAAACATGATGATAATGAGTGT GTTAAAAAGATGACCGAAGG-3′

• Daxin(Δ19–27aa) reverse: 5′-GCTCTAGATTA ATCGGATGGCTTGACAAGACC-3′

To generate DTNKS shRNA constructs, the following primers were annealed at 95°C for 5 min in annealing buffer (10mM Tris-HCl,pH7.5,100mM NaCl,1mM EDTA), and slowly cooled to room temperature. The oligos were sub-cloned into the pWALIUM20 vector with NheI and EcoRI sites. The primers were as follows:

• tank-RNAi-1 forward: 5′-CTAGCAGTCGTGCTGTGTCGAACCAAAGA TAGTTATATTCAAGCATATCTTTGGTTCGACACAGCACGGCG-3′

• tank-RNAi-1 reverse: 5′-AATTCGCCGTGCTGTGTCGAACCAAAGA TATGCTTGAATATAACTA TCTTTGGTTCGACACAGCACG ACTG-3′

• tank-RNAi-2 forward: 5′-CTAGCAGTCGGAGTACTTGATAACCTACC TAGTTATATTCAAGCATA GGTAGGTTATCAAGTACTCCG GCG-3′

• tank-RNAi-2 reverse: 5′-AATTCGCCGGAGTACTTGATAACCTACC TATGCTTGAATATAACTA GGTAGGTTATCAAGTACTCCG ACTG-3′

2.3 Generation of DTNKS mutant clones

The DTNKS mutant clones were generated by the FLP-FRT method. The flies were heat shocked at 37°C for 1 hr at 1^st and 2^nd instar larval stages to induce mitotic clones.

2.4 Drosophia imaginal discs preparation, Immunostaining and microscopy

Drosophila imaginal discs were dissected from 3^rd instar larvae in cold PBS solution, fixed in 4% formaldehyde for 15 min at room temperature and rinsed in PBS containing 0.1% Triton X-100 (PBST). For immunostaining experiments, imaginal discs were blocked in blocking buffer (PBST with 5% serum) for 15min and incubated with indicated antibodies. The imaginal discs were photographed using the Zeiss LSM710 Laser Scan Confocal Microscope. Antibodies used in this study were as follows: rabbit anti-Vg (1:25), mouse anti-Dll(1:500), guinea pig anti-Sens(1:200), goat anti-Daxin(dT-20, Santcruz, 1:10), mouse anti-Eya(DSHB, 1:20), and guinea pig anti-Daxin(1:500). Primary antibodies were detected by fluorescent-conjugated secondary antibodies(Invitrogen). The adult wing images were obtained using OLYMPUS BX41 microscope.

2.5 The total RNA isolation and Real-time quantitative PCR

Adult wild-type and DTNKS^m250 flies were used to isolate RNA with RNAprep pure tissue kit (Tiangen). cDNAs were produced with the RT-PCR kit (K1005s, Promega) in a 10μL reaction volume. Quantitative RT-PCR was performed using Applied Biosystems 7500 Real-Time PCR System. Three pairs of primers were used to detect DTNKS expression. The transcription levels of DTNKS were normalized to Ribosomal protein L32 (rp49). The primers were as follows:

• tank-1 forward:5′-GTGTAGGACGGGCAGAGCAACT-3′

• tank-1 reverse: 5′-CATGACCGCATCGAGATTAACG-3′

• tank-2 forward: 5′-ATGGGCACTATGAGGTAACCGAACT-3′

• tank-2 reverse: 5′-TGCAACATCGTGATCAGATTCCTTA-3′

• tank-3 foward: 5′-CTATGCACATTGTTGGTGGACTT-3′

• tank-3 reverse: 5′-GGACTACCGTGGAAGAGCATAC-3′

• Rp49 forward: 5′-ATGCTAAGCTGTCGCACAAA-3′

• Rp49 reverse: 5′-GTTCGATCCGTAACCGATGT-3′.

2.6 Luciferase assay

Drosophila Schneider 2 (S2) cells were seeded in 6-well plates, cotransfected with plasmids expressing DTNKS-shRNA, 12×dTCF Firefly luciferase, Dfz2, and Renilla luciferase by Effectene (Qiagen) or treated with XAV939 (Sigma Aldrich). After 36 hours transfection, medium containing secreted Wingless was added to the cultured cells, and luciferase assays were carried out 4 hours later. The activity of Wg signaling was determined by the ratio of Firefly/Renilla luciferase activity.

2.7 Cell culture and transfection

S2 cells were cultured in Schneider’s medium supplemented with 1U/ml Penicillin and 1μg/ml Streptomycin at 25°C. Plasmid transfection was performed using Effectene (Qiagen) according to the manufactures’ instructions.

2.8 Immunoblotting and immunoprecipitation

Cell lysates were prepared with RIPA buffer (Beyotime). Total proteins were resolved by SDS-PAGE, transferred onto PVDF membrane (BioRad), and probed with indicated antibodies. For co-immunoprecipitation experiments, cells were lysed in lysis buffer (20mM Tris-HCl, pH7.6, 150mM NaCl, 1mM EDTA, 0.1% NP-40, with 1mM protein inhibitor). Precleared cell lysates were incubated with the indicated antibodies and Protein G-sepharose beads (GE healthcare) for 4 hours at 4°C. Beads were washed with lysis buffer 3 times. The bound proteins were eluted by SDS buffer, and monitored by western immunoblotting. The results of western blots were analyzed by LICOR Odyssey imager. Antibodies used in this study were as follows: rabbit-anti-V5 (Sigma, 1:1000), mouse-anti-V5 (Invitrogen, 1:2000) mouse-anti-Flag (Sigma, 1:1000), mouse-anti-GFP (Invitrogen, 1:1000), and rabbit-anti-DTNKS (1:300).

3 Result

3.1 Drosophila tankyrase (DTNKS) is not required for viability

Drosophila CG4719 encodes the Drosophila tankyrase (DTNKS) (Fig. 1). The DTNKS gene is localized on the 3^rd chromosome and encodes two polypeptides, one contains 1181 amino acids and the other contains 1520 amino acids. Both polypeptides contain three protein domains: an ankyrin repeat domain, a PARP domain, and a SAM domain. DTNKS shares high identity with hTNKS within these domains. The ankyrin repeat domain of DTNKS shares 71.6% identity with hTNKS1 (Fig. 1A). The PARP domain and SAM domain share 75% and 51.5% identity, respectively, with those of hTNKS (Fig 1A), indicating that DTNKS is an evolutionarily conserved protein.

Figure 1. Generation of DTNKS mutant.

(A) DTNKS is highly conserved with human TNKS1. It shares 71.6%, 51.5% and 75% identity with human TNKS1 at the ankyrin repeat domain, the SAM domain, and the PARP domain, respectively. (B) Schematic diagram of the genomic region of DTNKS locus. Red boxes represent coding sequences and gray boxes indicate untranslated exon sequences. P{HP37069} indicates the insertion site of the P-element stock. Straight lines show intron sequences and broken lines indicate deleted sequences covering the first exon and ATG of DTNKS. (C) Quantitative RT-PCR using total RNAs from adult DTNKS^m250 homozygotes or wild type flies. The expression level of DTNKS of each line was normalized to the control rp49. All three pairs of primers (tank-1, tank-2 and tank-3) show low expression of DTNKS in DTNKS mutant, as compared with wild type fly. Error bars represents standard error of the mean.

To examine the functions of DTNKS in development, we generated DTNKS mutants through P element-mediated knockout technique. The P element, P{EPg}HP37069 (BL#22129), was inserted in the 5′-UTR of DTNKS gene. The imprecise excision of this P element generated two alleles of DTNKS: DTNKS^m23 and DTNKS^m250. Segments of 1635 bp and 1924 bp of genomic DNA were deleted in DTNKS^m23 and DTNKS^m250, respectively. Both deletions covered the first exon of DTNKS. Moreover, DTNKS^m250 covered part of the regulatory region in front of the transcription starting site (Fig. 1B). Both alleles were homozygous viable. To determine whether our generated mutants are null, we performed real-time PCR to detect DTNKS mRNA transcription level in the DTNKS^m250 mutant with three pairs of primer (tank-1, tank-2 and tank-3). The region covered by primer pair tank-1 was located at the deleted region, which served as a negative control. The regions covered by tank-2 and tank-3 were located at the non-deleted region, detecting the transcription level of DTNKS. The results from RT-PCR revealed that DTNKS mRNA level in DTNKS^m250 mutant were as low as the negative control (Fig. 1C), suggesting that DTNKS^m250 was a null allele. The DTNKS^m250 homozygous mutant was fertile and had no obvious developmental defects, indicating that DTNKS was not required for viability.

3.2 DTNKS interacts with Daxin, and their protein levels are mutually interdependent

A previous study showed that hTNKS interacted with Axin and induced the degradation of Axin [31]. To explore whether DTNKS regulates Daxin in Drosophila, we examined the interaction between Daxin and DTNKS in Drosophila S2 cells. Using co-immunoprecipitation experiments, we found that DTNKS Co-IPed with Daxin, and vice versa (Fig. 2A). Axin contains an N-terminal tankyrase binding domain (TBD), which is conserved in different species and has been shown to mediate the interaction between hTNKS and Axin [31]. Deletion of this domain in Daxin (mapped to 17^th–29^th amino acids) also disrupted the interaction between DTNKS and Daxin (Fig. 2A, lane 2 and 5), confirming that Daxin binds to DTNKS via its N-terminal TBD. DTNKS contains an N-terminal ankyrin repeat domain, a middle SAM domain, and a C-terminal PARP domain. We next examined which domain in DTNKS mediated the interaction between DTNKS and Daxin. We generated two truncated DTNKS constructs, one deleting the PARP domain (961^st–1181^st amino acids) and the other deleting the ankyrin repeat domain (56^th–770^th amino acids) (Fig. 2E). Using co-immunoprecipitation experiment, we found that the ankyrin repeat domain, but not the PARP domain, was required for DTNKS binding to Daxin (Fig. 2B). These data indicated that DTNKS physically interacted with Daxin, and this interaction was mediated via the TBD domain of Daxin, and the ankyrin repeat domain of DTNKS.

Figure 2. DTNKS interacts with Daxin and modulates protein levels of each other in S2 cells.

(A) DTNKS interacts with Daxin in S2 cells. V5-tagged DTNKS and Flag-tagged Daxin expression vectors were transfected into S2 cells (lane 1, 3, 4). V5-DTNKS and Flag-Daxin are co-precipitated with each other (lane 1, 3, 4). TBD of Daxin (mapped to 17^th–29^th amino acids) is required for Daxin to bind to DTNKS (lane 2, 5). A construct lacking this domain does not interact with DTNKS. (B) The ankyrin repeat domain but not the PARP domain of DTNKS is required for DTNKS binding to Daxin. Flag-Daxin does not bind to DTNKS lacking the ankyrin repeat domain. (C) Exogenous expression of DTNKS causes reduction of Daxin protein levels. The protein level of Flag-Daxin was significantly reduced with increased amount of V5-DTNKS. (D) Exogenous expression of Daxin results in down-regulation of the DTNKS protein level. The protein level of V5-DTNKS was significantly reduced with increased amounts of Flag-Daxin. (E) The diagram of the DTNKS mutant constructs. The PARP-domain truncated construct DTNKSΔPARP has a deletion of 961–1181aa. The ankyrin repeat domain truncated construct DTNKSΔANK has a deletion of 56–770aa.

We found that the protein level of Daxin was much lower than that of Daxin (Δ17–29aa) when co-expressed with DTNKS (Fig. 2A, lane 1 and 4 compared to lane 2 and 5), suggesting that the binding of DTNKS might affect Daxin protein levels. We then examined the Daxin protein level in the presence of increasing amounts of DTNKS. Co-transfection of DTNKS dramatically decreased protein levels of Daxin in a dosage-dependent manner (Fig. 2C), suggesting a conserved role of TNKS in the regulation of Axin protein levels.

Surprisingly, we found that the protein level of DTNKS was also reduced in the presence of co-expressed full-length Daxin, but not in the presence of Daxin (Δ19–27aa) (Fig 2A, lane 1 compared with lane 2 and 3). Indeed, co-expression of increasing amounts of Daxin down-regulated DTNKS protein levels in a dosage-dependent manner (Fig. 2D). These data suggest that DTNKS and Daxin can regulate the protein levels of each other.

3.3 DTNKS regulates wing patterning depending on the concentration of Daxin

Daxin acts as a scaffold protein in the destruction complex to regulate the stability of Armadillo (β-catenin), and works as a negative regulator in Wg/Wnt signaling [5–8, 13, 35, 36]. The aforementioned experiments showed that DTNKS was involved in the degradation of Daxin in S2 cells. To determine whether DTNKS is required for Wg signaling, we generated two independent shRNAs targeting DTNKS to examine DTNKS’ function in the Wg signaling pathway. We first examined the Wg-stimulated luciferase activity in S2 cells. The result showed that depletion of DTNKS using DTNKS-shRNA1 or DTNKS-shRNAi2 significantly reduced the Wg reporter activity (p<0.001) (Fig. 3A). Similar result was also obtained after treated S2 cells with XAV939, a small molecule of hTNKS (p<0.001) (Fig. 3B). These results indicate that DTNKS regulates Wg signaling in S2 cells. To confirm this finding in vivo, we generated transgenic flies expressing DTNKS-shRNAs and used genetic approaches to test the function of DTNKS during wing patterning. Wing margin bristles are a particularly sensitive indicator for proper levels of Wg signaling. Excessive Wg signaling causes ectopically expressed bristles, whereas low levels of Wg signaling leads to loss of bristles [37, 38]. However, we found that knocking-down DTNKS by DTNKS-shRNA using SalE-gal4 or En-gal4 (data not shown) had no obvious defects in wing patterning (Fig. 3C). Axin works as a negative regulator in Wg signaling, and a high concentration of Axin protein inhibits Wg signaling. Thus, we speculated that the elevated Daxin protein level in the absence of DTNKS was not high enough to influence Wg signaling. We tested this possibility by knocking-down DTNKS in a genetic background with high levels of Daxin. During wing patterning, ectopic expression of Daxin in the wing pouch by SalE-gal4 generated wing-margin and bristles defects (Fig. 3D). This phenotype was enhanced when DTNKS was knocked-down simultaneously (Fig. 3E). We also found that ectopic expression of DTNKS in the wing significantly rescued the defect caused by overexpression of Daxin (Fig. 3F). Taken together, these results indicate that DTNKS regulates Daxin protein levels, and affects wing patterning under high concentrations of Daxin.

Figure 3. DTNKS regulates Wg signaling and wing patterning.

(A) Paracrine activation luciferase assay. The S2 cells were transfected with W20-DTNKS-shRNAs (or pWALIUM20 vector), 12xdTCF firefly luciferase, Dfz2, and renilla luciferase. The medium containing secreted wingless was added to cultured cells. Knock-down of DTNKS by two independent shRNAs inhibits Wg reporter in S2 cells (*P<0.001, n=5). (B) Paracrine activation luciferase assay as in figure A. Suppression of DTNKS activity using XAV939 inhibits Wg reporter in S2 cells (*P < 0.001, n=6). (C) Knock-down of DTNKS by shRNA driven by SalE-gal4 has no effect on wing patterning. (D) Overexpression of Daxin driven by SalE-gal4 led to wing margin defects. (E) The defect in (D) was enhanced when DTNKS shRNA was coexpressed using SalE-gal4. (F) The wing defect in (D) was rescued by ectopic expression of DTNKS.

3.4 DTNKS is required for Daxin degradation, but has no obvious effect on Wg signaling

To understand further the function of DTNKS in Wg signaling in vivo, we examined Daxin protein levels and expression levels of Wg target genes through RNAi-mediated DTNKS knockdown and mosaic clonal analysis on DTNKS mutant in Drosophila imaginal discs. Consistent with our results in S2 cells, we found that Daxin protein levels were increased in DTNKS mutant cells (Fig. 4A–C). We then monitored Wg target genes in wing imaginal discs, including senseless(Sens), distal-less(Dll), and vestigial(Vg). Endogenous Sens protein is expressed in cells receiving high levels of Wg signaling (Fig. 5C). Knocking-down DTNKS in the posterior compartment of wing imaginal discs by en-gal4 did not affect sens expression, when compared to control discs (Fig. 5C and 5D). Dll and vg are expressed in cells receiving both high and low levels of Wg signaling. Knocking down DTNKS also had no effect on their expression. (Fig. 5A and 5B).

Figure 4. Daxin is increased in DTNKS mutant clones.

Protein levels of Daxin were examined in DTNKS mutant cells. (A) The absence of GFP signals labeled the DTNKS mutant cells. (B) Endogenous Daxin was increased in DTNKS mutant cells in eye discs.

Figure 5. Expression of Wg target gene is not affected in DTNKS mutant.

All wing imaginal discs are oriented dorsal top-left, posterior top-right.

(A–D) Wing imaginal discs expressed UAS-GFP alone or together with W20-DTNKS shRNA by en-gal4. The target gene expression region was labeled by GFP signals. (A–A‴) The expression pattern of Vg (A′) and Dll (A″) in wing imaginal discs. (B–B‴) The expression of Vg(B′) and Dll(B″) was not affected when DTNKS was knocked-down by shRNA in the posterior compartment. (C–C′) Expression pattern of Sens in wild-type wing imaginal discs. (D–D′) Expression level of Sens was not affected when DTNKS was knocked-down by shRNA in the posterior compartment. (E–E‴) The wing imaginal disc containing DTNKS mutant cells was stained by anti-Vg (E′) and anti-Dll (E″) antibodies. The absence of GFP signals labeled the mutant cells. Expression level of Vg and Dll was not affected in mutant cells. (F–F′) The wing imaginal disc containing DTNKS mutant cells was stained by anti-Sens antibody. Expression level of Sens was not affected in mutant cells. (G–G′) The eye imaginal disc containing DTNKS mutant cells was stained by anti-Eya antibody. Expression of Eya was not affected in DTNKS mutant cells.

Similar to the observations by shRNA-mediated DTNKS knockdown, expression of endogenous sens, Dll, and vg were not affected in DTNKS mutant clones cells (Fig. 5E and 5F). We also examined the role of DTNKS in the eye disc, another organ system in Drosophila where Wg signaling plays pivotal roles. In eye imaginal discs, Wg signaling functions to restrict the area of the eye field, which is dependent upon the activities of various transcription factors, including eyeless (eye), eyes absent(eya), sine oculis(so), and dachshund (dac) [39–44]. However, the expression of eya, an eye specification gene regulated by Wg signaling [44], was not affected in the absence of DTNKS (Fig. 5G). Taken together, these results suggest that DTNKS regulates Daxin protein levels, but has no obvious function in Drosophila wing and eye development under normal conditions.

3.5 DTNKS regulates expression of Wg target genes under high levels of Daxin

As DTNKS regulated Daxin protein levels and affected wing patterning at a high level of Daxin, we speculated that the regulation of Wg target genes by DTNKS also depended on the concentration of Daxin. To test this hypothesis, we examined expression of Dll and vg in wing discs in which Daxin was ectopically expressed. As expected, we found that expression of Dll and vg were decreased when Daxin was expressed by Sal-gal4 in the wing imaginal discs (Fig. 6A). This reduction was dramatically enhanced when DTNKS was simultaneously knocked-down in wing imaginal discs (Fig. 6B). Taken together, these data suggest that DTNKS regulates Wg signaling, depending on the concentration of Daxin.

Figure 6. Loss of DTNKS suppresses Wg signaling at a high level of Daxin.

(A–B″) Wing imaginal discs expressed UAS-Daxin alone (A) or together with w20-DTNKS shRNA (B) by SalE-gal4. The target gene expression region was labeled by immunostaining of Daxin (green). (A–A″) Ectopic expression of Daxin by SalE-gal4 caused reduction of Dll (A′) and Vg (A″) in wing imaginal discs. (B–B″) Wing imaginal discs expressing UAS-Daxin with DTNKS by shRNA display further reduction of Dll (B′) and Vg (B″).

4 Discussion

4.1 The conserved role of DTNKS in the regulation of Daxin degradation

Wg/Wnt signaling is a conserved signaling pathway and plays important roles during many developmental processes and homeostasis. A tight regulation on the concentration of Axin is critical for Wg signaling transduction [14]. Although previous studies have demonstrated a role for TNKS in the regulation of Axin levels and Wnt signaling in human cancers, its functions during development remain largely unknown. By generating Drosophila tankyrase mutants, we demonstrated a role for DTNKS in the regulation of Daxin levels during imaginal disc development (Fig. 4A–C). In addition, we also show that DTNKS physically interacts with Daxin and regulates protein levels of Daxin in S2 cells (Fig. 2A, 2C). These results indicate that the function of TNKS in regulating Axin protein levels is conserved among different species [31].

Surprisingly, we found that Daxin was also involved in the regulation of DTNKS protein levels (Fig. 2D). Coexpression of Daxin can dramatically reduce protein levels of DTNKS (Fig. 2D). With increasing amounts of Daxin, this effect on DTNKS was enhanced. Daxin is a scaffold protein without any known catalytic activity. Thus, we speculate that Daxin mediates degradation of DTNKS by recruiting other components and forming a destruction complex for DTNKS. A preview study suggests RNF146, one of the E3 ligases, forms a complex with Axin and TNKS and regulates the degradation of both Axin and TNKS in human cells [45, 46]. Therefore, we propose that degradation of DTNKS by Daxin may depend on the Drosophila RNF146-mediated ubiquitination pathway. Since endogenous levels of Daxin are low, the increased levels of Daxin may recruit more RNF146 and promote DTNKS degradation in S2 cells.

4.2 The role of DTNKS in Wg signaling

Our genetic studies showed that depletion of DTNKS has no obvious effect on wing patterning (Fig. 3C) or expression of Wg target genes (Fig. 5A–G′), suggesting the limited function of DTNKS for Wg signaling under normal developmental conditions. However, the luciferase assay showed that knock-down of DTNKS by shRNA, or suppression of DTNKS activity by XAV939 significantly reduces Wg reporter activity in S2 cells (Fig. 3A–B). Consistent with this finding, Knock-down of DTNKS caused dramatic defects in wing patterning (Fig. 3E) and reduction of Wg target gene expression under a genetic background with a high Daxin protein level (Fig. 6B–B″), indicating that DTNKS can regulate Wg signaling under certain conditions. A previous study showed that tubulin promoter-driven Daxin protein level was expressed ~4.3 fold higher than that of endogenous Daxin, and caused no obvious defect during wing patterning [47]. According to this result, a potential explanation for our findings is that the increase of Daxin caused by depletion of DTNKS is within the physiological range required for normal development, and is not enough to disrupt Wg signaling. Thus, we speculate that DTNKS plays important roles in Wg signaling in certain situations where Daxin protein levels reach a certain threshold.

4.3 The role of DTNKS in other developmental and cellular processes

A critical function of human Tankyrase (hTNKS) is the maintenance of telomere length. In human telomerase-positive cells, overexpression of hTNKS protects the telomere from degradation. hTNKS removes TRF1, a negative regulator of telomere length maintenance, from the telomere and subsequently triggers its ubiquitination and degradation [16, 20, 23, 48, 49]. However, knocking out mouse TNKS1 or TNKS2 showed no effect on telomere length maintenance [22, 33, 34]. Moreover, mouse TRF1 lacks a tankyrase binding consensus motif, RXXG/PDG, that is shared by all TNKS partners, and does not interact with mouse TNKS [15, 50–52]. As a result, TNKS does not remove mouse TRF1 from the telomere, or maintain the telomere length in mouse. Therefore, the function of TNKS on telomere length maintenance is not conserved between human and mouse. We propose that DTNKS is unlikely to be involved in regulation of Drosophila telomere length. This hypothesis is supported by two pieces of evidence. First, TRF1, the negative regulator of telomere length maintenance, is not a conserved protein between human and Drosophila. We failed to identify a TRF1 homologue in Drosphila. Secondly, the structure of telomere and the maintenance of telomere length are different between human and Drosophila. In humans, telomeres are composed of TTAGGG tandemrepeat sequences. The maintenance of telomere length requires telomerase. Drosophila lacks telomerase and the maintenance of telomere length depends on transposition of three specialized retro-transposons TART, HeT-A, and TAHRE, rather than telomerase activity [53]. Previous study showed that the regulation of telomeres by TNKS only occurred in telomerase-positive cells [49], indicating that the regulation of telomere by TNKS requires telomerase activity. Therefore, we speculate that the function of TNKS in telomere length maintenance is not conserved in Drosophila.

Highlights.

• DTNKS interacts with Daxin and modulates its protein levels in vivo and in S2 cells.

• Daxin in turn down-regulates DTNKS protein levels.

• DTNKS regulates Wg target genes expression and wing patterning at high Daxin levels.