UbcD1 regulates Hedgehog signaling by directly modulating Ci ubiquitination and processing

Abstract

The Hh pathway controls many morphogenetic processes in metazoans and plays important roles in numerous pathologies and in cancer. Hh signaling is mediated by the activity of the Gli/Ci family of transcription factors. Several studies in Drosophila have shown that ubiquitination by the ubiquitin E3 ligases Slimb and Rdx(Hib) plays a crucial role in controlling Ci stability dependent on the levels of Hh signals. If Hh levels are low, Slimb adds K11‐ and K48‐linked poly‐ubiquitin chains on Ci resulting in partial degradation. Ubiquitin E2 enzymes are pivotal in determining the topologies of ubiquitin chains. However, which E2 enzymes participate in the selective ubiquitination‐degradation of Ci remains elusive. Here, we find that the E2 enzyme UbcD1 negatively regulates Hh signaling activity in Drosophila wing disks. Genetic and biochemical analyses in wing disks and in cultured cells reveal that UbcD1 directly controls Ci stability. Interestingly, UbcD1 is found to be selectively involved in Slimb‐mediated Ci degradation. Finally, we show that the homologs of UbcD1 play a conserved role in modulating Hh signaling in vertebrates.

Introduction

Hedgehog (Hh) pathway has a multitude of functions in tissue‐patterning and organogenesis during embryonic development 1, 2, 3, 4 and also participates in adult tissue maintenance and repair 5, 6 in metazoans. Perturbation of Hh signaling is associated with numerous human disorders including congenital malformations 7 and cancers 8, 9, 10, 11.

The key determinant of Hh signaling output is the activity of Gli/Ci transcription factors, primarily regulated by ubiquitination‐degradation machines. Drosophila is a good model for studying Hh pathway. In the absence of Hh signals, Ci is sequentially phosphorylated by protein kinase A (PKA), glycogen synthase kinase 3 (GSK3) and casein kinase I (CKI), which triggers phosphorylation‐dependent ubiquitination of Ci by supernumerary limbs (Slimb)–cullin 1 (Cul1) E3 ligase complex. This causes partial degradation (or processing) of full‐length Ci (Ci^FL) by proteasomes and leaves its N terminus as a transcriptional repressor (Ci^R) to inhibit the expression of a subset of Hh target genes including dpp 12, 13, 14, 15, 16, 17, 18. In the presence of high‐level Hh signals, phosphorylation and Slimb‐dependent processing of Ci is blocked and Ci is further activated to turn on the expression of other Hh target genes, including ptc, en, and Col (also known as kn). Meanwhile, roadkill (Rdx, also known as Hib)‐cullin 3 (Cul3) E3 ligase ubiquitinates Ci, leading to a total degradation of Ci 19, 20, 21, 22, 23.

Interestingly, our laboratory has reported that not only the two different E3 ligases, but also the distinct ubiquitin (Ub) chains are implied in the selective degradation of Ci. Slimb‐Cul1 E3 ligase adds both K48‐ and K11‐linked Ub chains while Rdx‐Cul3 E3 ligase adds mainly K48‐linked Ub chains on Ci, and this difference of Ub chains is important for distinguishing specifically ubiquitinated Ci and further determining the fate of Ci, either partial or total degradation 24. The family of E2 enzymes, in the center of ubiquitination enzymatic cascade, is now considered to be the main determinant for the types of Ub modification 25, 26, 27, 28. However, which E2 enzymes are specifically responsible for the selective ubiquitination of Ci is still unknown.

Here, we report that ubiquitin E2 enzyme UbcD1 controls Ci stability through directly and particularly regulating Slimb‐based ubiquitination of Ci. Moreover, the homologs of UbcD1 play a conserved role in regulating Hh signaling in vertebrates.

Results and Discussion

Loss of UbcD1 function promotes Ci accumulation and increases Hh signaling activity

Drosophila wing imaginal disk is a classic model to study basic mechanisms of Hh signal transduction. In order to identify new components that regulate Hh pathway, we conducted a genetic screening by using MS1096‐Gal4; UAS‐dicer2 to induce RNAi of individual genes in the wing pouch. The wing disks of the third‐instar larva (Fig 1A) were dissected and immunostained with anti‐Ci (2A1) antibodies 29, which recognize only full‐length Ci (Ci^FL). In this screening, UbcD1, also known as effete/eff, was identified as a negative regulator of Ci. Compared with control, Ci^FL proteins accumulate, especially in the dorsal part of UbcD1 RNAi wing disks (Fig 1B and C), which is consistent with previous studies indicating that MS1096‐Gal4 has a higher transcriptional activity in the dorsal compartment compared with the ventral compartment 24, 30, 31, 32, 33, 34. In response to Ci^FL accumulation, the expression of dpp, one of the Hh target genes, was elevated as indicated by both real‐time qPCR analyses (Fig 1D) and dpp‐LacZ 20, 31, 33, 34 staining (Fig 1E–F′), suggesting the up‐regulation of Hh signaling activity.

Figure 1. Loss of UbcD1 promotes Ci accumulation and increases Hh signaling activity.

• A
  A diagram of wing disk. The frame indicates the view in which Ci^FL staining is shown. A, anterior; P, posterior; V, ventral; D, dorsal.
• B, C
  Ci^FL staining in wing disks of indicated genotypes, by using Ci(2A1) antibody. Scale bars, 30 μm.
• D
  Relative mRNA levels of UbcD1 and dpp in extracts of wing disks with or without UbcD1 RNAi driven by MS1096‐Gal4;UAS‐dicer2. Data are presented as mean ± SD, n = 3. Two‐tailed unpaired Student's t‐test: **P < 0.01; ***P < 0.001.
• E, F
  Ci^FL (red) and dpp‐LacZ (green) staining (single slice) in wing disks of indicated genotypes. The arrows indicate cells distant from anterior/posterior compartment (A/P) boundary. Scale bars, 30 μm.
• G
  Ci^FL (red) staining in wing disk with UbcD1 RNAi clones (GFP⁺). The arrows indicate a clone distant from A/P boundary. Scale bars, 30 μm.
• H
  Ci^FL (red) staining in wing disk with UbcD1 ^mer1 mutant clones (GFP⁺). The arrows indicate clones. Scale bars, 30 μm.
• I, J
  Ci^FL (red) staining in wing disks with over‐expression of UbcD1^WT (I) or UbcD1^C85A (J) in UbcD1 ^mer1 mutant clones (GFP⁺). Scale bars, 30 μm.
• K, L
  Enlarged views of framed regions in (I) and (J). Scale bars, 20 μm.

To ascertain the phenotypes, we used another driver AG4;UAS‐GFP;UAS‐dicer2, to induce UbcD1 knockdown in randomly formed GFP‐positive (GFP⁺) clones. Through in situ hybridization (ISH) assays, we found that UbcD1 is ubiquitously expressed throughout the wild‐type wing disk (Fig EV1A and B) and UbcD1 mRNA levels were decreased in its RNAi clones (Fig EV1C–C′′). Combined with results using MS1096‐Gal4;UAS‐dicer2, these observations indicate an elevation of Ci^FL protein levels in the UbcD1 RNAi clones (Fig 1G and G′).

Figure EV1. Repression of UbcD1 elevates Ci^FL .

• A, B
  In situ hybridization of UbcD1 in wild‐type wing disks, with DIG‐labeled mRNA probe against UbcD1 (B). The sense probe acts as negative control (A). The dashed line outlines the disk in (A). Scale bar, 75 μm.
• C
  In situ hybridization of UbcD1 (red) in wing disk with UbcD1 RNAi clones (GFP⁺). Scale bar, 30 μm.
• D
  Ci^FL (red) staining in wing disk with UbcD1 ^s1782 mutant clones (GFP⁺), using Ci(2A1) antibody. Scale bar, 30 μm.
• E
  Enlarged views of framed regions in (D). The arrows indicate clones. Scale bar, 15 μm.
• F
  Ci^FL (red) staining in wing disk with UbcD1 over‐expression clones (GFP⁺). Scale bars, 30 μm.
• G
  The structural model of UbcD1, performed in SWISS_MODEL (https://swissmodel.expasy.org/). Human UbcH5b has a 95% identity to Drosophila UbcD1, and thus, its structure (PDB code 5DDG) was used to generate the structural model of UbcD1. The image was generated by PyMOL (http://pymol.org). The structural model of UbcD1 is shown as cartoon and colored green. The unconserved residues between UbcD1 and UbcH5b are colored orange. The conserved catalytic site of UbcD1 (85^th Cys, C85) is highlighted in ball representation.

In order to confirm the negative role of UbcD1 on Ci protein level, we utilized MARCM system 35 to generate GFP⁺ clones with homozygous UbcD1 ^mer1 36, 37 mutant. Consistent with the results of RNAi assays, Ci^FL protein levels were increased in UbcD1 ^mer1 mutant clones (Fig 1H and H′). In addition, we also observed increased Ci^FL protein levels in UbcD1 ^s1782 36, 37, 38 mutant clones (Fig EV1D–E′). The above results indicate that UbcD1 is required for repressing Hh signaling activity and Ci protein level. Surprisingly, over‐expression of UbcD1 could not decrease Ci^FL protein levels (Fig EV1F and F′). We suppose that UbcD1 and its homologs are highly expressed in vivo, supported by our ISH results (Fig EV1A and B), and thus, the effects of additional E2 proteins might be overwhelmed or covered up by endogenous ones which are sufficient for regulating Hh signaling.

UbcD1 is a characteristic Ub E2 enzyme. Given that UbcD1 has a 95% identity to UbcH5b (the homolog in human; Fig EV5A), the structure of UbcD1 (Fig EV1G) is very similar to that of UbcH5b 39. Besides, both UbcD1 and UbcH5b contain a conserved cysteine (Cys, C) residue at position 85 (C85; Figs EV1G and EV5A), which is reported to be the active site of UbcH5b for thioester formation with Ub 39, 40. In our following in vitro ubiquitination assays, C85 is also verified to be the active site of UbcD1. We then individually over‐expressed UbcD1^WT or UbcD1^C85A [with the C85 mutated to alanine (Ala, A)] in UbcD1 ^mer1 mutant clones using MARCM system. The accumulation of Ci^FL in UbcD1 ^mer1 mutant clones was diminished by over‐expression of UbcD1^WT, but not UbcD1^C85A (Fig 1I–L′), suggesting that the catalytic activity of UbcD1 is required for its function in modulating Ci protein levels.

Figure EV5. UbcD1/Ube2d2 plays a conserved role in inhibition of Hh signaling in zebrafish.

• A
  Alignment of amino acid sequences of UbcD1 and its homologs in indicated species. Asterisk indicates the conserved cysteine residue at position 85.
• B
  Relative mRNA levels of Hh target genes in extracts of 24 hpf zebrafish embryos, which were injected with 0.15 mM indicated MOs. Data are presented as mean ± SD, n = 4.
• C–J
  In situ hybridization of hhip (C–F), and fkd4 (G–J) in 24 hpf zebrafish embryos, which were injected with 0.15 mM indicated MOs. The proportion of injected zebrafish exhibiting the staining pattern depicted in the images is indicated. Scale bars, 200 μm.
• K
  Relative mRNA levels of gli1 and ptch2 in extracts of 24 hpf zebrafish embryos, which were injected with 50 ng/μl GFP (Control), Ube2d2 ^WT or Ube2d2 ^C85A mRNA. Data are presented as mean ± SD, n = 4. Two‐tailed unpaired Student's t‐test: n.s., not significant; *P < 0.05; ***P < 0.001.
• L–P
  In situ hybridization of ptch2 in 10 hpf zebrafish embryos, which were injected with indicated MOs (0.15 mM) and mRNAs (200 ng/μl). Scale bars, 100 μm.
• Q
  Western blot analyses of lysates from NIH‐3T3 cells treated with or without SAG, or transfected with indicated siRNAs (three independent replicates). The relative mRNA levels of Ube2d2a were analyzed. Data are presented as mean ± SD, n = 3. Two‐tailed unpaired Student's t‐test: ***P < 0.001.

UbcD1 directly regulates Ci stability

Given that UbcD1 is an Ub E2 enzyme and can negatively regulate Ci^FL protein level, it is conceivable that UbcD1 might directly affect the stability of Ci. To test this projection, we first detected the endogenous Ci^FL protein level in wing disk using Western blot assays. Consistent with immunostaining results displayed in Fig 1, endogenous Ci^FL proteins accumulated in UbcD1 RNAi disks (Fig 2A). Besides, exogenous HA‐Ci^FL was also increased upon UbcD1 knockdown (Fig 2B), suggesting that UbcD1 directly represses Ci protein levels in vivo. To further confirm the role of UbcD1 in control of Ci stability, we employed cultured S2 cells. Western blot results and statistical densitometry analyses showed that Ci^FL protein levels were increased upon UbcD1 knockdown (Fig 2C), further confirming that UbcD1 negatively regulates Ci stability.

Figure 2. UbcD1 is presumably related to Slimb‐mediated Ci partial degradation.

• A
  Western blot analyses of lysates from wing disks with or without UbcD1 RNAi driven by MS1096‐Gal4;UAS‐dicer2. Endogenous Ci^FL was detected by anti‐CiN antibody. Data in the statistical analysis are presented as mean ± SEM, n = 3. Two‐tailed unpaired Student's t‐test: *P < 0.05.
• B
  Western blot analyses of lysates from wing disks with or without indicated transgenes driven by ci‐Gal4. The arrow indicates HA‐Ci^FL. The asterisk indicates nonspecific band. Data in the statistical analysis are presented as mean ± SEM, n = 3. Two‐tailed unpaired Student's t‐test: *P < 0.05.
• C
  Western blot analyses of lysates from S2 cells co‐transfected with Myc‐Ci and Myc‐CFP evenly, as well as indicated dsRNA targeting LacZ (Control) or UbcD1. Data in the statistical analysis are presented as mean ± SEM, n = 3. Two‐tailed unpaired Student's t‐test: *P < 0.05.
• D, E
  Ci^FL (red) and HA (blue) staining in wing disks with over‐expression of HA‐Rdx in wild‐type (D) or UbcD1 ^mer1 mutant (E) clones (GFP⁺). Scale bars, 30 μm.
• F, G
  Enlarged views of framed regions in (D) and (E). The arrows indicate clones near A/P boundary. Scale bars, 10 μm.
• H, I
  Ci^FL (red) staining in eye disks with wild‐type (H) or UbcD1 RNAi (I) clones (GFP⁺). Scale bars, 50 μm. A, anterior; P, posterior. The degradation of Ci is mediated by Slimb in A compartment while mediated by Rdx in P compartment.
• J, K
  Enlarged views of framed regions in (I and I′), showing clones in A compartment (J) and P compartment (K). The dashed lines outline clones. Scale bars, 15 μm.
• L
  Ci^FL (red) staining in eye disk with UbcD1 ^mer1 mutant clones (GFP⁺). The arrow indicates a clone in A compartment, and the arrowhead indicates a clone in P compartment. Scale bars, 50 μm.

Source data are available online for this figure.

Notably, the effects of UbcD1 RNAi on Ci^FL accumulation were eliminated by over‐expression of UbcD1^WT, but not the “catalytic dead” forms UbcD1^C85A or UbcD1^C85V (with the C85 mutated to Ala or Valine; Fig EV2A), indicating that the catalytic activity is required for UbcD1 to destabilize Ci. Collectively, these above biochemical results strongly support that UbcD1 destabilizes Ci and this function is dependent on its catalytic activity.

Figure EV2. UbcD1 is presumably related to Slimb‐mediated Ci partial degradation.

• A
  Western blot analyses of lysates from S2 cells co‐transfected with Myc‐Ci and Myc‐CFP evenly, as well as indicated constructs and dsRNAs. Data in the statistical analysis are presented as mean ± SEM, n = 3. Two‐tailed unpaired Student's t‐test: n.s., not significant; *P < 0.05; **P < 0.01. For lanes 2–5, UbcD1 was knocked down by dsRNA targeting 5′UTR of UbcD1. The arrows indicate endogenous and exogenous UbcD1 proteins.
• B, C
  Ci^FL (red) and dpp‐LacZ (green) staining in wing disks of indicated genotypes. The arrows indicate cells distant from the anterior/posterior compartment (A/P) boundary. Scale bars, 30 μm.
• D–F
  Ci^FL (red) and Flag (blue) staining in wing disks with or without over‐expression of Flag‐Slimb in wild‐type or UbcD1 ^mer1 mutant clones (GFP⁺). Scale bars, 30 μm.
• G–I
  Enlarged views of framed regions in (D–F). The dashed lines outline clones. Scale bars, 10 μm.
• J, K
  Ci^FL (red) staining in eye disks with Slimb RNAi (J) or Rdx RNAi (K) clones (GFP⁺). Scale bars, 50 μm.
• L
  Ci^FL (red) staining in eye disks with UbcD1 ^s1782 mutant clones (GFP⁺). The arrows indicate clones in A compartment. Scale bars, 50 μm.

UbcD1 is presumably related to Slimb‐mediated Ci partial degradation

The degradation of Ci is, respectively, mediated by two different Ub ligases in a manner dependent on the levels of Hh signals. We then started to explore which degradation processes UbcD1 is involved in. In wing disk, Ci is wholly expressed in anterior (A) compartment. Hh proteins produced by posterior (P) compartment cells are diffused from P compartment into A compartment 41. Normally, in cells distant from the A/P boundary, Ci is cleaved to generate Ci^R dependent on Slimb‐based E3 ligase and Ci^R is required for repression of dpp 16, 18. In cells near the A/P boundary, Ci processing is blocked in response to high‐level Hh signals. Ci is activated as a transcriptional activator to turn on the expression of other Hh target genes including ptc, en, and Col, and the degradation of Ci is mediated by Rdx 19, 20, 21, 22, 23. We noticed that knockdown of UbcD1 could up‐regulate Ci^FL protein levels as well as the expression of dpp in cells distant from the A/P boundary (Fig 1F and G, arrows), similar to the phenotypes occurred upon Slimb knockdown (Fig EV2B and B′, arrows), but not Rdx knockdown (Fig EV2C and C′). Moreover, the levels of ptc, en, and Col were not affected by Slimb knockdown or UbcD1 knockdown, confirmed by generating UbcD1 ^mer1 mutant clones (Fig EV3). Taken together, both Ci^FL protein levels and Hh target genes are regulated by UbcD1 and Slimb in a same manner, suggesting that UbcD1 is specifically involved in Slimb‐mediated Ci partial degradation.

Figure EV3. Repression of UbcD1 releases the expression of dpp in the absence of Hh signals, but does not influence Hh target genes responsive to high‐level Hh signals.

• A–C
  Ci^FL (red) and ptc‐LacZ (green) staining in wing disks of indicated genotypes. Scale bars, 30 μm.
• D
  Ci^FL (red) and Ptc (blue) staining in wing disks with UbcD1 ^mer1 mutant clones (GFP⁺). Scale bars, 30 μm.
• E
  Enlarged views of framed regions in (D). The arrows indicate a clone. Scale bars, 15 μm.
• F–H
  Ci^FL (red) and En (green) staining in wing disks of indicated genotypes. Scale bars, 30 μm.
• I
  Ci^FL (red) and En (blue) staining in wing disks with UbcD1 ^mer1 mutant clones (GFP⁺). Scale bars, 30 μm.
• J
  Enlarged views of framed regions in (I). The arrows indicate a clone. Scale bars, 15 μm.
• K–M
  Ci^FL (red) and Col (green) staining in wing disks of indicated genotypes. Scale bars, 20 μm.
• N
  Ci^FL (red) and Col (blue) staining in wing disks with UbcD1 ^mer1 mutant clones (GFP⁺). Scale bars, 20 μm.
• O
  Enlarged views of framed regions in (N). The arrows indicate a clone. Scale bars, 10 μm.

In order to verify the above notion, we explored the genetic interaction between UbcD1 and Slimb/Rdx using MARCM system. Over‐expression of Rdx decreased Ci^FL protein levels in both WT clones and UbcD1 ^mer1 mutant clones near the A/P boundary (Fig 2D–G′′), indicating that Rdx‐mediated degradation of Ci^FL could not be blocked by loss of UbcD1 function. In contrast, over‐expression of Slimb decreased Ci^FL protein levels in WT clones (Fig EV2D, G and G′), but not in UbcD1 ^mer1 mutant clones (Fig EV2E, H and H′), suggesting the requirement of UbcD1 for Slimb‐mediated partial degradation of Ci. Interestingly, Slimb‐mediated degradation of Ci was not completely blocked in some UbcD1 ^mer1 mutant clones (comparing Fig EV2H′ with EV2I′), suggesting the existence of other E2 enzymes functioning redundantly with UbcD1 in Slimb‐mediated Ci degradation.

To further determine in which Ci degradation process UbcD1 is involved, we employed Drosophila eye disk. Slimb‐Cul1 E3 ligase controls partial degradation of Ci exclusively in cells anterior to the morphogenetic furrow (MF), whereas Rdx‐Cul3 E3 ligase dominates complete degradation of Ci in cells posterior to the MF 20, 21, 42 (Fig EV2J–K′). Interestingly, UbcD1 knockdown specifically caused the accumulation of Ci^FL proteins in clones (GFP⁺) anterior to the MF but not posterior to the MF (Fig 2H–K′). These phenotypes were confirmed in MARCM system by generating UbcD1 ^mer1 or UbcD1 ^s1782 homozygous mutant clones (Figs 2L and L′, and EV2L and L′). The above observations clearly indicate that UbcD1 regulates Ci stability in a Slimb‐dependent but not Rdx‐dependent manner.

UbcD1 is essential for ubiquitination and partial degradation of Ci mediated by Slimb‐Cul1 E3 ligase

The above findings promoted us to verify the role of UbcD1 in Slimb‐mediated Ci partial degradation. First, we analyzed the level of Ci processing in wing disks through Western blot assays. Knockdown of UbcD1 resulted in a decrease of Ci^R protein levels and increased the ratios of Ci^FL/Ci^R (Fig 3A), suggesting the involvement of UbcD1 in Ci processing. Consistently, UbcD1 and Slimb knockdown disturbed the partial degradation of over‐expressed Ci in wing disks (Fig 3B). Next, we examined the effects of UbcD1 loss in cultured S2 cells. Over‐expression of PKA, Slimb, and Cul1 strongly promoted wild‐type Ci processing, while did not affect the processing of Ci^−PKA with substitution mutations at all PKA phosphorylation sites 23 (Fig EV4A). In this background, knockdown of UbcD1 stabilized Ci^FL and decreased the generation of Ci^R (Figs 3C and EV4B, comparing lane 3 with lane 2). More importantly, knockdown of UbcD1 also disturbed the processing of endogenous Ci in Drosophila Clone‐8 (Cl8) cells (Fig 3D). These data support that UbcD1 plays an important role in Slimb‐mediated partial degradation of Ci.

Figure 3. UbcD1 is essential for ubiquitination and partial degradation of Ci mediated by Slimb‐Cul1 E3 ligase.

• A
  Western blot analyses of lysates from wing disks with or without UbcD1 RNAi driven by MS1096‐Gal4;UAS‐dicer2. The arrows indicate endogenous Ci^FL and Ci^R, detected by anti‐CiN antibodies. Data in the statistical analysis are presented as mean ± SEM, n = 3. Two‐tailed unpaired Student's t‐test: *P < 0.05.
• B
  Western blot analyses of lysates from wing disks with over‐expression of HA‐Ci, as well as indicated transgenes driven by ci‐Gal4. The arrows indicate HA‐Ci^FL and HA‐Ci^R. The asterisk indicates nonspecific band. Data in the statistical analysis are presented as mean ± SEM, n = 3. Two‐tailed unpaired Student's t‐test: **P < 0.01.
• C
  Western blot analyses of lysates from S2 cells co‐transfected with Myc‐Ci and Myc‐CFP evenly, as well as indicated constructs and dsRNAs. The arrows indicate Myc‐Ci^FL and Myc‐Ci^R. The asterisk indicates nonspecific band. Data in the statistical analysis are presented as mean ± SEM, n = 3. Two‐tailed unpaired Student's t‐test: ***P < 0.001.
• D
  Western blot analyses of lysates from Cl8 cells transfected with indicated siRNAs. Data in the statistical analysis are presented as mean ± SEM, n = 3. Two‐tailed unpaired Student's t‐test: *P < 0.05.
• E
  Western blot analyses of extracts pulled down with Ni‐NTA agarose in denaturing conditions and lysates from S2 cells transfected with indicated constructs and dsRNAs (three independent replicates). Pan Ub signals on Ci‐His or Ci^−PKA‐His were detected by anti‐Ub antibody.
• F
  Western blot analyses of Ub signals on GST‐Ci (phosphorylated) after in vitro ubiquitination assays with or without indicated E2 enzymes and Slimb‐Cul1 E3 ligases (three independent replicates). Pan Ub signals on GST‐Ci (phosphorylated) were detected by anti‐Ub antibody.
• G, H
  Western blot analyses of K48‐ and K11‐lined Ub chains on GST‐Ci (phosphorylated) after in vitro ubiquitination assays with or without indicated E2 enzymes and Slimb‐Cul1 E3 ligases (three independent replicates). K48/K11‐linked Ub chains were detected by specific antibodies.

Source data are available online for this figure.

Figure EV4. UbcD1 is a unique Ub‐E2 negatively regulating Ci stability.

1. Western blot analyses of immunoprecipitates from S2 cells co‐transfected with Myc‐Ci or Myc‐Ci ^−PKA, as well as indicated constructs. The arrows indicate Myc‐Ci^FL and Myc‐Ci^R. The asterisks indicate nonspecific bands.
2. Western blot analyses of immunoprecipitates from S2 cells co‐transfected with indicated constructs and dsRNAs. The arrows indicate Myc‐Ci^FL and Myc‐Ci^R. The asterisk indicates nonspecific bands. Data in the statistical analysis are presented as mean ± SEM, n = 3. Two‐tailed unpaired Student's t‐test: *P < 0.05.
3. Western blot analyses of extracts pulled down with Ni‐NTA agarose in denaturing conditions and lysates from S2 cells transfected with indicated constructs and dsRNAs (three independent replicates). Pan Ub signals on Ci‐His or Ci^−PKA‐His were detected by anti‐Ub antibody.

Source data are available online for this figure.

Given that UbcD1 can directly interact with Slimb‐Cul1 E3 ligase 43 and is important for Slimb‐mediated Ci degradation, we speculated that UbcD1 may directly serves as the E2 enzyme responsible for Slimb‐mediated Ci ubiquitination. Over‐expressed Ci proteins were pulled down by Ni‐NTA agaroses in denaturing conditions, and the ubiquitination levels on Ci were detected (Fig 3E and EV4C). In line with the effects on Ci degradation, over‐expression of PKA, Slimb, and Cul1 in S2 cells promoted ubiquitination of wild‐type Ci and this increased ubiquitination was greatly suppressed by knockdown of UbcD1. As the negative control, the ubiquitination of Ci^−PKA mutants was not affected neither by over‐expression of PKA, Slimb, and Cul1 nor by knockdown of UbcD1 (Figs 3E and EV4C). These findings suggest that UbcD1 is essential for Slimb‐Cul1 E3 ligase's function in ubiquitination of Ci.

In order to clearly demonstrate whether UbcD1 has the ability to ubiquitinate Ci, we performed in vitro ubiquitination assays. As shown in Fig 3F, in the absence of either Slimb‐Cul1 ligase (lane 1) or an E2 enzyme (lane 2), Ci could not be modified with Ub chains in vitro. Here, we employed UbcH5c, which is a general E2 enzyme widely used in in vitro ubiquitination assays, as the positive control (lane 5). UbcD1 added Ub proteins with higher efficiency than UbcH5c (lane 3), suggesting the capacity of UbcD1 to ubiquitinate Ci mediated by Slimb‐Cul1 E3 ligase. As the negative control, UbcD1^C85A failed to ubiquitinate Ci (lane 6). Importantly, when we replaced UbcD1 with UbcD2 (another ubiquitin E2 enzyme in Drosophila), there were negligible Ub signals on Ci (lane 4). The results from in vitro ubiquitination assays further underline the critical role of UbcD1 and its catalytic activity in Slimb‐mediated ubiquitination of Ci.

We previously reported that Slimb‐Cul1 E3 ligase added both K48‐linked and K11‐linked Ub chains on Ci 24. We then tried to figure out whether UbcD1 can modify Ci with either one or both types of Ub chains. Antibodies recognizing K48‐linked and K11‐linked Ub chains were used to detect corresponding Ub signals on Ci in in vitro ubiquitination assays. We found that wild‐type UbcD1, but not UbcD1^C85A can efficiently add both K48‐ and K11‐linked Ub chains on Ci, with stronger capacity than UbcH5c (Fig 3G and H), suggesting that UbcD1 is involved in Slimb‐mediated both K11‐ and K48‐linked Ub chain modifications of Ci.

UbcD1 plays a conserved role in regulating Hh signaling across metazoans

Most of the core components of the Hh signaling are conserved between vertebrates and Drosophila, and also the mechanism of Hh signal transduction shares some degree of similarity 44. Given that the amino acid sequence of fly UbcD1 has a 95% identity to its homologs UbcH5b (in human) and Ube2d2 (in zebrafish; Fig EV5A), we then tested whether UbcD1 exerts a conserved function in regulating Hh signaling in vertebrates. Antisense morpholinos (MOs) were designed to, respectively, block splicing (SB‐MOs) or translation (TB‐MOs) of Ube2d2. In response to the loss of Ube2d2 (Fig 4A), the expression levels of Hh‐responding genes, such as gli1 and ptch2 45, 46, 47, 48, 49, 50, were increased, and such expression elevation could be blocked by injection with Ube2d2 mRNA (Fig 4B). Of note, the effects of Ube2d2 knockdown on Hh target genes were not as strong as those upon knockdown of both ptch1 and ptch2 (ptch1;2; Fig EV5B), well‐known repressors of Hh signaling 34. There are at least three Gli transcription factors (Gli1, Gli2, and Gli3) involved in vertebrate Hh signal transduction, but only Gli2 and Gli3 are proteolytically processed into the repressor forms 27. According to the regulation mechanism of UbcD1 we revealed in the fly system, loss of Ube2d2 might only up‐regulate target genes repressed by Gli2/3^R in the absence of Hh signals. Therefore, the function of Ube2d2 on repressing Hh signaling could be limited in zebrafish.

Figure 4. UbcD1/Ube2d2 plays a conserved role in regulating Hh signaling in zebrafish.

• A
  Relative mRNA levels of Ube2d2 in extracts of 24 hpf zebrafish embryos, which were injected with gradient concentrations of Ube2d2 SB‐MOs designed to block splicing. Data are presented as mean ± SD, n = 3. Two‐tailed unpaired Student's t‐test: *P < 0.05; **P < 0.01; ***P < 0.001.
• B
  Relative mRNA levels of Hh target genes in extracts of 24 hpf zebrafish embryos, which were injected with Control MOs or Ube2d2 SB‐MOs (0.15 mM), and GFP mRNAs (Control) or Ube2d2 mRNAs (400 ng/μl). Data are presented as mean ± SD, n = 3. Two‐tailed unpaired Student's t‐test: *P < 0.05; **P < 0.01.
• C–N
  In situ hybridization of gli (C–F), ptch2 (G–J), and nkx2.2b (K–N) in 24 hpf zebrafish embryos, which were injected with 0.15 mM indicated MOs. The proportion of injected zebrafish exhibiting the staining pattern depicted in the images is indicated. Scale bars, 200 μm.
• O–R
  Prox1 (green) and Eng (red) staining in 24 hpf zebrafish embryos, which were injected with indicated MOs (0.15 mM) and mRNAs (400 ng/μl). Scale bars, 30 μm. The arrows indicate ectopic MPs (Prox1⁺ Eng⁺). The arrowheads indicate ectopic MFFs (Eng⁺). The hollow arrowhead indicates ectopic SSFs (Prox1⁺).
• S
  The relative number of Eng⁺ cells in each somite of 24 hpf zebrafish embryos, which were injected with indicated MOs (0.15 mM) and mRNAs (400 ng/μl). The horizontal lines inside the boxes represent the medians of each group; the lower and upper limits of each box represent the lower and upper quartiles of each group respectively; the lower and upper whiskers represent the minimum and maximum value of each group respectively; the circles represent the outliers of each group. Data are presented as mean ± SD, n > 20. Two‐tailed unpaired Student's t‐test: n.s., not significant; ***P < 0.001.
• T
  Relative mRNA levels of Ube2d2, UbcD1, and Hh target genes in extracts of 24 hpf zebrafish embryos, which were injected with indicated MOs (0.15 mM) and mRNAs (200 ng/μl). Data are presented as mean ± SD, n = 3. Two‐tailed unpaired Student's t‐test: n.s., not significant; **P < 0.01; ***P < 0.001.
• U–Y
  In situ hybridization of ptch2 in 10 hpf zebrafish embryos, which were injected with indicated MOs (0.15 mM) and mRNAs (200 ng/μl). The proportion of injected zebrafish exhibiting the staining pattern depicted in the images is indicated. Scale bars, 100 μm. Additional embryos are shown in Fig EV5L–P.

To support the real‐time qPCR results, we further detected the mRNA levels of downstream target genes using ISH assays. Similar with the phenotypes observed in ptch2 morphants, the expression levels of Hh target genes were elevated in Ube2d2 morphants injected with SB‐MOs or TB‐MOs (Figs 4C–N and EV5C–J). These results all together suggest that Ube2d2 is required for inhibition of Hh signaling.

Previous studies showed that Hh signaling specifies cell fate in the zebrafish myotome 51, 52. In order to confirm the function of Ube2d2 in regulating Hh signaling activity, we explored the muscle cell type identity in Ube2d2 morphants. In wild‐type embryos by 24 hpf, all slow muscle fibers including muscle pioneers (MPs) and superficial slow fibers (SSFs) express Prox‐1 while only MPs in addition express Eng. Eng proteins are also expressed in medial fast fibers (MFFs) at relatively low levels 51, 52. We found that embryos injected with Ube2d2 MOs showed an increase in the number of Eng‐expressing cells and MFFs within the myotome, which resembled, although not as strong as, the effects of activating Hh signaling by ptch2 knockdown (Fig 4O–Q′ and S). Importantly, these phenotypes could be rescued by simultaneous injection with Ube2d2 mRNA (Fig 4R–S). In conclusion, Ube2d2 functions as a repressor of Hh signaling in zebrafish.

Moreover, similar to over‐expression of UbcD1 in fly wing disks, injection with Ube2d2 mRNA in zebrafish embryos showed marginal effects (Fig EV5K), probably due to abundant endogenous Ube2d2, which are sufficient for regulating Hh signaling. In contrast, embryos injected with Ube2d2 ^C85A mRNA displayed up‐regulation of Hh target genes (Fig EV5K), suggesting the dominant negative function of this mutant. Another important thing is that injection of UbcD1 mRNA could eliminate the up‐regulation of Hh target genes occurred in Ube2d2 morphants (Fig 4T). Similarly, another report demonstrated that UbcD1 can functionally substitute for yeast UBC4 53, suggesting that UbcD1 and its homologs share a high level of evolutionary conservation across species. Thus, through assays in both fly system and zebrafish system, we conclude that UbcD1/Ube2d2 exerts a conserved biological function in repressing Hh signaling, dependent on its catalytic activity.

The above notion is further confirmed by examination of ptch2 expression in earlier development stage—bud stage (10 hpf)—embryos (Fig 4U–Y and EV5L–P). The expression of ptch2 was elevated in Ube2d2 morphants, which could be rescued by co‐injection with Ube2d2 mRNA, as well as UbcD1 mRNA, consistent with the effects observed in 24 hpf embryos. Moreover, over‐expression of UbcD1 ^C85A mutant could not eliminate the increase of ptch2 expression in Ube2d2 morphants, supporting that the catalytic activity is required for the function of UbcD1.

In order to confirm the conservation of this E2 enzyme directly, we analyzed its function on the processing of vertebrate Gli proteins. Several studies have provided evidence to support the involvement of β‐TRCP (the homolog of Slimb) in Gli3 processing 15, 54, 55, and thus, we examined mouse Gli3 processing in cultured NIH‐3T3 cells. The results showed that knockdown of Ube2d2a (the homolog of UbcD1 in mouse) impeded Gli3 processing as indicated by the increased ratio of Gli3^FL/Gli3^R (Fig EV5Q), suggesting the involvement of Ube2d2a in β‐TRCP‐mediated Gli3 degradation. Taken together, UbcD1 plays a conserved role in regulating Hh signaling across metazoans.

However, we noticed that the phenotypes induced by loss of UbcD1/UbcH5b function seem modest. The ectopic activation of dpp‐lacZ upon UbcD1 knockdown was weak compared to that upon Slimb knockdown (Figs 1F and F′, and EV2B and B′). Besides, although knockdown of UbcD1 resulted in the accumulation of Ci^FL and attenuated Ci processing, the effects are not remarkable (Fig 3A–D). Similarly, knockdown of mouse Ube2d2a also showed modest effects on Gli3 processing (Fig EV5Q). The weak effects might be due to low efficiency of UbcD1 RNAi. However, given that over‐expression of Slimb still partially worked in the absence of UbcD1 function (Fig EV2D–I′), there seem to be other E2 enzyme(s) function redundantly with this E2 in regulation of Ci/Gli3 processing.

In conclusion, we for the first time demonstrated that an Ub E2 enzyme UbcD1 selectively and directly cooperates with Slimb‐Cul1 but not Rdx‐Cul3 E3 ligase in modulation of Ci ubiquitination‐degradation. Remarkably, Wang's laboratory recently found that UbcD1 exerts similar function to Slimb‐Cul1 E3 ligase in regulating self‐renewal of neuroblast 56. Combined with our findings, UbcD1 is very likely responsible for transferring Ub modifier on multiple Slimb‐targeted substrates. Given that β‐TRCP targets many specific substrates related to a variety of human cancers 57, 58, it will be interesting and significant to reveal the role of human UbcH5b in those diseases in the future.

Materials and Methods

RNA isolation, reverse transcription, and real‐time qPCR

Wing disks, cells, or zebrafish embryos were lysed in TRIzol (Invitrogen) for total RNA isolation following standard protocol. For real‐time qPCR, 0.5 μg RNA was used for reverse transcription with ReverTra Ace qPCR RT Master Mix with gDNA Remover (TOYOBO, FSQ‐301). Real‐time qPCR was performed on CFX96 Touch Real‐Time PCR Detection System (Bio‐Rad) with SYBR^® Green Realtime PCR Master Mix (TOYOBO, QPK‐201). The $2^{-\mathrm{\Delta }\mathrm{\Delta }{C}_{\mathrm{t}}}$ method was used for quantification. The primer pairs used are as follows:

(1) Drosophila melanogaster (fly)
UbcD1 (forward)	5′‐TCAATAAGGAACTGCAAGATCTGG‐3′
UbcD1 (reverse)	5′‐GCCCATTATTGTAGCTTGCCA‐3′
dpp (forward)	5′‐AGCATTTAATCCATTCAGCGAG‐3′
dpp (reverse)	5′‐TCTAATTGGTCTGTTCTTGGCT‐3′
Rpl32 (forward)	5′‐CTAAGCTGTCGCACAAATGG‐3′
Rpl32 (reverse)	5′‐AGGAACTTCTTGAATCCGGTG‐3′
(2) Danio rerio (zebrafish)
Ube2d2 (forward)	5′‐CCAAATTCAGACCTCCACTC‐3′
Ube2d2 (reverse)	5′‐GCCATAAACACTTTCTACAGCA‐3′
gli1 (forward)	5′‐TTCTTGGTTTACTTGAAGGCAGAG‐3′
gli1 (reverse)	5′‐GCTCATTATTGATGTGATGCACC‐3′
ptch2 (forward)	5′‐TCCTCCTTATGAGTCCCAAACAG‐3′
ptch2 (reverse)	5′‐CATGAACAACCTCAACAAACTTCC‐3′
fkd4 (forward)	5′‐GCTTCACTGAACCATTTCGCA‐3′
fkd4 (reverse)	5′‐CTGAGCCATAATACATCTCGCTG‐3′
hhip (forward)	5′‐CTTACGAGCCAAGTGTGAACTG‐3′
hhip (reverse)	5′‐TGCTGTCTTTCTCACCGTCC‐3′
nkx2.2b (forward)	5′‐CAAATATCCAGTGCCGTCAGC‐3′
nkx2.2b (reverse)	5′‐CGCTCTAACTCAAAGGTTTGAGTC‐3′
gapdh (forward)	5′‐CATCACAGCAACACAGAAGACC‐3′
gapdh (reverse)	5′‐ACCAGTAAGCTTGCCATTGAG‐3′
(3) Mus musculus (mouse)
Ube2d2a (forward)	5′‐TCCATCCAATCTCAAACAGTC‐3′
Ube2d2a (reverse)	5′‐AACCTTAGACCTTAAAGCACAG‐3′
gapdh (forward)	5′‐CGTCCCGTAGACAAAATGGT‐3′
gapdh (reverse)	5′‐GAATTTGCCGTGAGTGGAGT‐3′

The real‐time qPCR data presented in figures are all relative mRNA levels normalized to the levels of Rpl32 (Drosophila samples) or gapdh (Danio rerio and Mus musculus samples).

DNA constructs

Full‐length UbcD1/UbcD2/Ci cDNA was cloned from Drosophila melanogaster wing disks RNA using RT–PCR. PCR‐based site‐directed mutagenesis was used to mutate the 85^th Cys of UbcD1 and all PKA binding sites of Ci. UbcD1 ^WT /UbcD1 ^C85A/UbcD1 ^C85V cDNA was inserted into pUAST‐attb, pUAST‐V5 vector or pET28a‐His vector to construct pUAST‐attb‐UbcD1 ^WT /UbcD1 ^C85A, pUAST‐UbcD1 ^WT ‐V5/UbcD1 ^C85A ‐V5/UbcD1 ^C85V ‐V5 or pET28a‐His‐UbcD1 ^WT /UbcD1 ^C85A. UbcD2 cDNA was inserted into pET28a‐His vector to construct pET28a‐His‐UbcD2. Ci/Ci ^−PKA cDNA was inserted into pUAST‐His vector to construct pUAST‐Ci‐His/Ci ^−PKA ‐His. pUAST‐Flag‐Slimb, pUAST‐Flag‐Cul1, pUAST‐Myc‐Ci, and pGEX‐4T‐1‐GST‐Ci were previously reported 24.

Fly stocks

Flies were raised on standard yeast/molasses medium at 25°C unless otherwise stated. MS1096‐Gal4 drives transgene expression in the wing pouch, with a higher level in the dorsal half compared with the ventral half 34. act>CD2>Gal4;UAS‐GFP (AG4; UAS‐GFP) induces random Gal4 expression by heat shock‐induced “jump out” of CD2, so that Gal4‐expressing clones were marked as CD2 minus and GFP‐positive region (Flybase). UAS‐dicer2 was used to increase RNAi efficiency (Flybase). apterous (AP)‐Gal4, ci‐Gal4, dpp‐LacZ, ptc‐LacZ, UAS‐HA‐Ci, UAS‐HA‐Rdx, UAS‐Flag‐Slimb, and Slimb RNAi have been described previously 17, 20, 24, 59, 60, 61, 62, 63. UbcD1 RNAi (NIG 7425R‐1, VDRC 26011), Rdx RNAi (NIG 9924R‐1), UbcD1 ^mer1 (BS 4656), and FRT82‐UbcD1 ^s1782 (Kyoto DGGR 111415) were obtained from the National Institute of Genetics, Vienna Drosophila RNAi Center, Bloomington Drosophila Stock Center, or Kyoto Stock Center. FRT82‐UbcD1 ^mer1 was generated via a recombination of FRT82 and UbcD1 ^mer1. pUAST‐attb‐UbcD1 ^WT /UbcD1 ^C85A was used to generate attp‐UAS‐UbcD1 ^WT/UbcD1 ^C85A transgenic flies, through a P element‐mediated insertion of cDNAs at the 25C site of the second chromosome, and their expression is induced by Gal4 binding with upstream activating sequence UAS.

Genotypes for MARCM are as follows:

• yw hsflp/Y; tubGal4 UAS‐GFP; FRT82 UbcD1 ^mer1 /FRT82 tubGal80,

• yw hsflp/Y; tubGal4 UAS‐GFP; FRT82 UbcD1 ^s1782 /FRT82 tubGal80,

• yw hsflp/Y; tubGal4 UAS‐GFP; FRT82/FRT82 tubGal80,

• yw hsflp/Y; tubGal4 UAS‐GFP/HA‐Rdx; FRT82 UbcD1 ^mer1 /FRT82 tubGal80,

• yw hsflp/Y; tubGal4 UAS‐GFP/HA‐Rdx; FRT82/FRT82 tubGal80,

• yw hsflp/Y; tubGal4 UAS‐GFP/Flag‐Slimb; FRT82 UbcD1 ^mer1 /FRT82 tubGal80,

• yw hsflp/Y; tubGal4 UAS‐GFP/Flag‐Slimb; FRT82/FRT82 tubGal80.

Immunostaining and in situ hybridization (ISH) of imaginal disks

Third‐instar larvae were cut in half and fixed in freshly made 4% formaldehyde in PBS buffer at room temperature for 25 min, and then washed four times with buffer PBT (PBS, 0.1% Triton X‐100). Larvae were incubated with primary antibody diluted in PBT for overnight at 4°C, then washed three times with PBT, and incubated with secondary antibody diluted in PBT for 2 h at room temperature. After wash, wing disks and/or eye disks were dissected and mounted in 40% glycerol. Leica LAS SP8 confocal microscope was employed to take immunostaining images. Primary antibodies used in this study are as follows: mouse anti‐β‐galactosidase (1:500, Cappel), anti‐En (DSHB), rabbit anti‐HA (1:1,000, Sigma), anti‐Flag (1:1,000, Sigma), rat anti‐Ci (2A1; 1:100, Developmental Studies Hybridoma Bank). Rabbit anti‐Col antibody is a gift from Prof. Jin Jiang 31. Secondary antibodies used in this study were bought from Millipore and were diluted at 1:1,000. Wing disks in situ hybridization was performed following standard protocols.

Cell culture, RNAi, and transfection

S2 cells were cultured at 25°C in Schneider's Drosophila Medium (Invitrogen) with 10% fetal bovine serum, 100 U/ml of penicillin, and 100 μg/ml of streptomycin. Cl8 cells were maintained in M3 medium. To perform knockdown experiment in S2 cells, double‐strand RNA (dsRNA) was synthesized using the in vitro Transcription T7 Kit from Takara. S2 cells were diluted into 1 × 10⁶ cells/ml and mixed with 15 μg dsRNA per 1 × 10⁶ cells for 1‐h starvation in serum‐free medium. The dsRNA targeting LacZ coding sequence was used as control. Transfection was carried out by using an X‐tremeGENE™ HP DNA Transfection Reagent (Roche) according to manufacturer's instructions. NIH‐3T3 cells were cultured at 37°C in Dulbecco's modified Eagle's medium (DMEM; Thermo)/10% fetal bovine serum/antibiotics, in 5% CO₂. Cl8 cells derived from wing imaginal disks were cultured using Shields and Sang M3 insect medium (Sigma, S‐3652) containing 2.5% fetal calf serum, and 2.5% fly extract and 10 μg/ml of insulin, and PS antibiotics (100 units/ml penicillin G, 100 μg/ml streptomycin). To perform knockdown experiment in NIH‐3T3 cells, siRNA was transfected by using RNAiMAX reagent (Invitrogen) according to manufacturer's instructions. Forty‐eight hours after transfection, cells were harvested with indicated buffers for immunoprecipitation, pull‐down, or direct Western blot analyses. The siRNA sequences are as follows:

Negative Control (N.C.)	5′‐UUCUCCGAACGUGUCACGUTT‐3′
Ube2d2a #1	5′‐CCAAUCCAGAUGAUCCUUUTT‐3′
Ube2d2a #2	5′‐GGAAUGGACUCAGAAGUAUTT‐3′
UbcD1 #1	5′‐UCCAACAGACUAUCCCUUUTT‐3′
UbcD1 #2	5′‐GGACUAGAAAGUAUGCUAUTT‐3′

Immunoprecipitation and Western blot analysis

For regular immunoprecipitation, S2 cells were lysed using NP‐40 buffer (50 mM Tris–HCl pH 8.0, 0.1 M NaCl, 10 mM NaF, 1 mM Na₃VO₄, 1% NP‐40, 10% glycerol, 1.5 mM EDTA, protease inhibitor cocktail). For in vivo ubiquitination assays with Myc tagged Ci, cells were first lysed in denaturing buffer (1% SDS, 50 mM Tris–HCl pH 7.5, 0.5 mM EDTA, 1 mM DTT). After incubation for 5 min at 100°C, the lysates were diluted tenfold with NP‐40 buffer. After centrifugation, lysates were incubated with 2 μg indicated antibodies for 2 h and then 20 μl protein A/G PLUS agarose (Santa Cruz) 1 h on a rotator at 4°C. For in vivo ubiquitination assay with His tagged Ci, cells were lysed with 1% SDS and boiled and diluted with denaturing buffer (0.1 M Na₂HPO₄/NaH₂PO₄, 8 M urea, 10 mM imidazole). After centrifugation, lysates were incubated with Ni‐NTA agarose for 1 h on a rotator at 4°C. Beads were washed three times with 1 ml NP‐40 buffer and then boiled in 40 μl SDS loading buffer. In a typical Western blot, samples from immunoprecipitation or cell/disk lysates were resolved by SDS–PAGE electrophoresis on 8, 10, or 12% gels, transferred to PVDF membranes (Millipore), blotted with the primary antibody for 2 h and the secondary antibody for 1 h, and then visualized by chemiluminescent substrate (Thermo). Band intensity was measured by ImageJ 1.44p. The following antibodies were used for immunoprecipitation and Western blot: mouse anti‐HA, anti‐Myc, anti‐His, anti‐GST, anti‐tubulin (Sigma), anti‐V5 (Invitrogen), anti‐Ub (P4D1; Santa Cruz), rabbit anti‐Flag (Sigma), anti‐K11‐linkage Ub, anti‐K48‐linkage Ub (Millipore), and goat anti‐Gli3 (R&D). Rabbit anti‐CiN antibody used in Fig 2A was generated in rabbit with 1–440aa of Ci as antigen. Rabbit anti‐UbcD1 antibody was generated in rabbit with 128–140aa of UbcD1.

In vitro ubiquitination assay

E3 complex was immunoprecipitated from S2 cells expressing 3× Flag‐tagged or Flag‐tagged Cul1‐Slimb E3 components and eluted by 3× Flag peptides (Sigma). GST‐Ci, His‐UbcD1^WT, His‐UbcD1^C85A, or His‐UbcD2 was purified with glutathione agarose beads (GE Healthcare) or Ni2⁺ NTA column (QIAGEN), from Escherichia coli (E. coli) BL21 cells carrying specific vector. To create recognition sites for Cul1‐Slimb‐based E3 ligase, GST‐Ci‐loaded beads were incubated with 500 U CKI, 1,250 U PKA, and 250 U GSK3 (NEB) in GSK3 reaction buffer (20 mM Tris–Cl pH 7.5, 10 mM magnesium chloride, 5 mM DTT, 1 mM ATP) at 30°C for 30 min, before in vitro ubiquitination assay. For in vitro ubiquitination assay, GST‐Ci‐loaded beads were incubated with 100 ng E1 (Rabbit), 200 ng E2 (indicated), 5 μg ubiquitin (Boston Biochem), 80 mM ATP and E3 complex in 20 μl ubiquitination buffer (50 mM Tris‐Cl pH 7.5, 5 mM magnesium chloride, 2 mM sodium fluoride, 0.6 mM DTT, 10 nM okadaic acid) at 30°C for 2 h.

MO and mRNA injection

Antisense MOs (Gene Tools) and mRNA were microinjected into one‐cell‐stage embryos according to standard protocols. A 4‐nl volume of MOs or mRNA was injected at indicated concentration. MO sequences used are as follows:

• Ube2d2 SB‐MO, 5′‐CCGTGACTTCCTTCTCTTACCTTGT‐3′;

• Ube2d2TB‐MO, 5′‐TGTGGATTCTTTTCAGAGCCATTGT‐3′;

• ptch2 MO, 5′‐CATAGTCCAAACGGGAGGCAGAAGA‐3′;

• ptch1 MO, 5′‐AGGAGACATTAACAGCCGAGGCCAT‐3′; and standard control‐MO (Gene Tools).

In situ hybridization of zebrafish embryos

Zebrafish (Danio rerio) embryos were obtained by natural spawning of adult AB strain zebrafish. Embryos were maintained at 28.5°C on a 14‐h light/10‐h dark cycle. MOs or mRNA were injected into 1‐cell‐stage zebrafish embryos and collected at 24 h postfertilization (hpf) or 10 hpf (bud stage). Whole mount in situ hybridization of zebrafish embryos was performed according to standard protocols 64. To avoid variability of in situ hybridization to the utmost extent, we injected the MOs, collected the fish, performed the in situ assays, and took the picture at the same time under the same conditions between a batch of samples. Probes for hhip, fkd4, nkx2.2b, gli1, and ptch2 (previously named as ptc1) were labeled with digoxigenin and have been described previously 33, 46.

Structure modeling

The structural modeling for UbcD1 is performed in SWISS_MODEL (https://swissmodel.expasy.org/). Human UbcH5b has a 95% identity to Drosophila UbcD1, and thus, its structure (PDB code 5DDG) was used to generate the structural model of UbcD1. The image was generated by PyMOL (http://pymol.org).

Statistical analyses

For analysis of experiments, statistically significant estimates were obtained from at least three independent experiments. The statistical significance between two groups was evaluated with a two‐tailed unpaired Student's t‐test. Significant differences were accepted at P < 0.05. For immunostaining and in situ hybridization (ISH) of imaginal disks, images from at least 10–20 disks were randomly taken to get conclusive results. For in situ hybridization of zebrafish embryos, images from at least 30 zebrafish embryos were randomly taken to get conclusive results.

Author contributions

The authors have made the following declarations about their contributions. CP conceived this study, designed, and performed experiments, analyzed data, and prepared manuscript; YX performed experiments and analyzed data; XL and HW analyzed data and edited manuscript; YX, SZ, JF, HC, WW, and FL contributed reagents, materials, and analysis tools; ZZ and LZ analyzed the data; YZ supervised and conceived this study, analyzed data, and edited manuscript.

Conflict of interest

The authors declare that they have no conflict of interest.

Supporting information

Expanded View Figures PDF

Source Data for Expanded View

Review Process File

Source Data for Figure 2

Source Data for Figure 3

EMBO Reports (2017) 18: 1922–1934