Su(fu) switches Rdx functions to fine-tune hedgehog signaling in the Drosophila wing disk

Authors


  • Communicated by: Tadashi Yamamoto

Correspondence: sishii@rtc.riken.jp

Abstract

Hedgehog (Hh) signaling plays a central role in pattern formation by regulating transcription factor Cubitus interruptus (Ci). Previously, Roadkill (Rdx, also called HIB) was shown to inhibit Ci activity by two distinct mechanisms, depending on the Hh signal strength (Seong et al. 2010, PLoS One 5, e15365). In the anterior region abutting the anterior/posterior (A/P) boundary of the wing disk, where cells receive a strong Hh signal, Rdx blocks the nuclear entry of Ci-155. In contrast, in the region farther from the A/P boundary, where cells receive moderate levels of Hh, Rdx induces Ci-155 degradation in the nucleus. Here, we report that Suppressor of fused, Su(fu), causes the Rdx switch between mechanisms. A strong Hh signal induces rdx expression and suppresses su(fu) expression, whereas moderate levels of Hh induce moderate levels of rdx expression and high levels of su(fu) expression. Rdx blocks entry of Ci-155 into the nucleus in the absence of Su(fu) and Rdx induces the degradation of Ci-155 in the nucleus in the presence of a threshold level of Su(fu). Thus, the Su(fu)-induced switch between the dual actions of Rdx in response to the Hh signal strength plays a role in fine-tuning Hh signaling.

Introduction

The hedgehog (Hh) family of morphogens plays crucial roles in pattern formation and cell growth in both vertebrates and invertebrates (Hooper & Scott 2005). In Drosophila wing imaginal disks, Hh is secreted by posterior (P) compartment cells and acts over a short range on the anterior (A) compartment cells by forming a local concentration gradient (Tabata & Kornberg 1994; Hooper & Scott 2005). The A compartment is divided into three regions, regions 1, 2, and 3, which receive strong, moderate, and weak Hh signals, respectively (see Fig. 1A). These three regions have distinct patterns of Hh target gene expression. In region 1, adjacent to the A/P boundary, the transcription factor Engrailed (En) (Blair 1992) and the COE family transcription factor Knot (Kn) (Vervoort et al. 1999; Dubois & Vincent 2001) are expressed. In region 2, at intermediate distances from the A/P boundary, the bone morphogenetic protein (BMP)-related signaling molecule Decapentaplegic (Dpp) is expressed (Basler & Struhl 1994). In region 3, farthest from the A/P boundary, the expression of Hh target genes is suppressed.

Figure 1.

Strong Hh signals down-regulated Su(fu) and up-regulated Rdx. (A) Schematic showing the expression patterns of the Hh target genes, Dpp, Kn, and En, and the Ci-regulators, Dbr, Rdx, and Su(fu). Regions 1, 2, and 3 in the A compartment are shown. (B) Su(fu) is down-regulated in region 1. Expression of rdx-lacZ (magenta) (a) and Su(fu) immunostaining (green) (b) are superimposed in (c). Anterior is to the left; dorsal is up. Vertical lines indicate the A/P boundary. (C) Su(fu) is down-regulated in the region where Ci-155 is also down-regulated. Ci immunostaining (magenta) (a) and Su(fu) immunostaining (green) (b) are superimposed in (c). The region where both Su(fu) and Ci-155 are down-regulated is indicated by the bar.

The zinc finger transcription factor Cubitus interruptus (Ci) is expressed in the A compartment cells and regulates various Hh target genes depending on the Hh signal strength (Hooper & Scott 2005). In region 3, where cells receive no Hh signal, full-length Ci (Ci-155) is phosphorylated by cAMP-dependent protein kinase, PKA, and other kinases, which results in the generation of the C-terminally truncated form Ci-75 (Aza-Blanc et al. 1997; Jiang & Struhl 1998; Jia et al. 2002, 2005; Price & Kalderon 2002; Smelkinson et al. 2007). Ci-75 is localized in the nucleus and represses the expression of Hh target genes (Domínguez et al. 1996; Méthot & Basler 1999). In regions 1 and 2, Hh signaling prevents the processing of Ci-155 to Ci-75 and stimulates Ci-155 activity (Aza-Blanc et al. 1997; Chen et al. 1999). The relatively high level of Ci-155 in regions 1 and 2 is maintained by expression of Debra (Dbr). In the bands outside regions 1 and 2, poly-ubiquitination of phosphorylated Ci-155 is induced by Dbr, and in cooperation with Slimb, leads to lysosomal degradation of Ci-155 (Dai et al. 2003). In regions 1 and 2, Hh signaling enhances the nuclear import of Ci-155 by releasing Ci-155 from a microtubule-associated complex composed of the kinesin-like protein Costal2 (Cos2), Suppressor of fused [Su(fu)], and the Ser/Thr kinase Fused (Fu) (Robbins et al. 1997; Sisson et al. 1997; Monnier et al. 1998; Amanai et al. 2000; Méthot & Basler 2000; Stegman et al. 2000). Following nuclear entry, Ci-155 interacts with the transcriptional coactivator CBP to induce transcription of target genes (Akimaru et al. 1997).

In regions 1 and 2, which receive strong and moderate Hh signals, respectively, different Hh target genes are selectively expressed. However, the mechanism for this selective gene expression remains obscure. Roadkill (Rdx) (also called HIB), which contains BTB and MATH domains, was first identified as the substrate recognition component of a Cul3 E3 ubiquitin ligase to target Ci for degradation in the nucleus (Kent et al. 2006; Zhang et al. 2006). However, this activity of Rdx is limited to region 2, and Rdx inhibits the nuclear import of Ci-155 in region 1 by releasing Ci from importin-α3, which is a nuclear import receptor for Ci (Seong et al. 2010). Thus, Rdx negatively regulates the activity of Ci-155 by different mechanisms in these two regions. Expression of Rdx is induced by Hh, and the level of Rdx in region 1 is higher than that in region 2. The level of Ci-155 in region 1 is lower than that in region 2, possibly due to the lower stability of cytosolic Ci-155 compared with nuclear Ci-155 (Seong et al. 2010). The amount of nuclear Ci-155 in region 1 is less than that in region 2 due to Rdx blocking the nuclear entry of Ci-155. The relatively small amount of nuclear Ci-155 may contribute to the expression of a subset of Hh target genes, such as kn and en, in region 1. However, the molecular basis for the dual functions of Rdx in these two regions is unknown.

Here, we report that Su(fu) induces a switch in Rdx functions. In region 1, where the Su(fu) level is low, Rdx inhibits the nuclear entry of Ci-155. However, in region 2, where the Su(fu) level is relatively high, Rdx induces the degradation of nuclear Ci-155.

Results

Down-regulation of Su(fu) by the strong Hh signal in region 1

This study focused on Su(fu) as a candidate protein to induce a switch in Rdx function because Su(fu) was previously shown to compete with Rdx for binding to Ci and to inhibit the Rdx-mediated degradation of Ci (Zhang et al. 2006). The expression pattern of Su(fu) in the wing disk was examined. In region 1, where the Hh signal is strongest and Rdx is highly expressed and inhibits the nuclear import of Ci-155 (Seong et al. 2010), the levels of Su(fu) were low (Fig. 1B). The levels of Ci-155 were also low in region 1 (Fig. 1C). Thus, Su(fu) is down-regulated in region 1, while Rdx is up-regulated in this region. Su(fu) was also downregulated in the region directly adjacent to the A/P boundary of the leg disk (Fig. S1 in Supporting Information).

To directly examine whether the strong Hh signal reduces the level of Su(fu) protein in the anterior region of the wing disk, we generated clones overexpressing Hh. In clones overexpressing Hh of the anterior region of wing disk, the levels of Su(fu) protein were reduced, while the levels of Rdx were increased (Fig. 2A). To perform this experiment, we confirmed that anti-Su(fu) antibody 25H3 can be utilized to detect Su(fu) protein (Fig. S2A in Supporting Information). In contrast, the decrease in Su(fu) protein levels and the increase in Rdx protein levels were not observed in Hh-overexpressing clones located in the posterior region of the wing disk (data not shown). In clones of smo3, which carries a mutation of the Hh transducer Smoothened and abolishes Hh signal transduction, the levels of Su(fu) protein at region 1 were up-regulated to which were similar to those at region 2 in wild-type cells (Figs 2B and S2B in Supporting Information). However, in the smo3 clones at region 2, the Su(fu) levels were not affected. In wing disks overexpressing Hh using the MS1096 driver, which is active in the entire wing disk, the level of Su(fu) protein was decreased in all of the anterior region (Fig. 2C-b). Overexpression of Hh did not reduce the levels of Su(fu) protein in the posterior region of the wing disk where endogenous Hh is high. To investigate whether the down-regulation of Su(fu) by Hh occurs at the transcriptional level, we examined the levels of Su(fu) mRNA Real-time RT-PCR analysis using RNA prepared from the wing disks indicated that Hh expression reduced the level of su(fu) mRNA to approximately 45% of the control level (Fig. 2D), suggesting that Hh almost completely suppressed Su(fu) transcription in the anterior region of the wing disk. In mammals, Sonic hedgehog (Shh) signaling promotes the ubiquitin–proteasome-dependent degradation of Su(fu) (Yue et al. 2009). To test this, the clones overexpressing Su(fu) were generated in wing disks, and the level of Su(fu) in the clones, which are located in the region expressing different levels of Ci, were compared. The level of Su(fu) protein was similar in the regions 1, 2, 3, of the anterior compartment and in the posterior region (Fig. S2C in Supporting Information). Furthermore, activation of Hh signaling by either Hh or Ci overexpression in clone-8 cell line, which was derived from developing wing imaginal disks (Peel & Milner 1992) and is responsive to Hh (Denef et al. 2000), did not induce the degradation of exogenously expressed Su(fu) (Fig. 2E). These results suggest that the strong Hh signal in region 1 down-regulates Su(fu) expression at the transcriptional level and up-regulates Rdx expression.

Figure 2.

Overexpression of Hh down-regulated Su(fu) and up-regulated Rdx in wing disks. (A) Clones of cells overexpressing Hh (GFP-positive, green, in (c) are outlined in red. The rdx-lacZ (red) (a) and Su(fu) immunostaining (blue) (b) are merged with Hh in (c). In a and b, the rdx-lacZ and Su(fu) signals are shown in white for clarity. Anterior is to the left; dorsal is up. Vertical lines indicate the A/P boundary. (B) Abolition of Hh signaling did not affect the basal level of Su(fu). Clones of smo3 expressing cells (GFP-negative area) (a) are outlined in red and superimposed with Su(fu) immunostaining (red) (b) in (c). In a and b, the Ci-155 and Su(fu) signals are shown in white for clarity. Anterior is to the left; dorsal is up. Vertical lines indicate the A/P boundary. (C) Hh overexpression in the wing pouch suppressed Su(fu) expression only in the anterior compartment, in which Ci is expressed. Ci-155 immunostaining (blue) (a) and Su(fu) immunostaining (red) (b) in the presence of ectopically expressed Hh (MS1096 driver) are superimposed in (c). In a and b, Ci-155 and Su(fu) signals are shown in white for clarity. (D) Hh down-regulated Su(fu) mRNA levels. A real-time RT-PCR analysis was used to compare Su(fu) mRNA in RNA extracted from the wild-type wing disks and wing disks overexpressing Hh using the MS1096 driver [mean ± SD, ***< 0.001 (n = 5)]. (E) Hh signaling did not induce Su(fu) protein degradation. S2 cells were infected with the Su(fu) expression vector pUAS-Su(fu) (0.5 μg) and the GAL4 expression plasmid pact5c-GAL4 (50 ng) together with increasing amounts of the HA-Ci-155 expression vector (0, 0.5, or 1.5 μg) and the Hh expression vector (0, 1, or 3 μg). The levels of Su(fu) were examined by Western blotting using an anti-Su(fu) antibody.

Su(fu) disrupts the Rdx-dependent inhibition of nuclear import of Ci-155

The expression pattern of Su(fu) in the wing disk described above raised the possibility that a decrease in Su(fu) levels in region 1 is required for Rdx to function as an inhibitor of Ci nuclear import. To examine the effect of Su(fu) on the Rdx function, we have used the clone-8 cell line derived from developing wing imaginal disk (Peel & Milner 1992), because it expresses endogenous Smo and Ptc proteins and is responsive to Hh (Denef et al. 2000). In the absence of overexpression of Hh, Rdx reduced the levels of total Ci-155 and nuclear Ci-155 in clone-8 cells, while this reduction was recovered by treatment with MG132, a proteasome inhibitor (Fig. S3 in Supporting Information). On the other hand, in the presence of overexpressed Hh, Rdx reduced the levels of total Ci-155 and nuclear Ci-155, but this reduction was not abrogated by MG132 treatment. This may be due to the Rdx-dependent nuclear import of Ci-155 and that the turnover of cytosolic Ci is more rapid than nuclear Ci. In fact, we previously reported that Rdx inhibits the nuclear import of Ci-155 fused to GFP (Ci-GFP) in the absence of Su(fu) in clone-8 cells (Seong et al. 2010).

To test the effect of Su(fu) on the Rdx activity to block the nuclear entry of Ci-, fluorescence recovery after photobleaching (FRAP) experiments were performed using Ci-GFP. After bleaching Ci-GFP fluorescence in the nucleus, compartmental equilibration (which, in this case, occurs via nuclear import of Ci-GFP) was measured by time-lapse fluorescence microscopy (Fig. 3A). When Ci-GFP was coexpressed with Su(fu) and Hh in clone-8 cells, the rate of nuclear import of Ci-GFP was 2.3 ± 0.6 per s/100 (Fig. 3B). This rate was similar to the previously reported import rate in the absence of Su(fu), 2.4 ± 0.5 per s/100 (Seong et al. 2010), although Su(fu) apparently decreased the relative nuclear Ci-GFP intensity. In the presence of both Hh and Su(fu), Rdx did not inhibit the nuclear entry of Ci-155 (Fig. 3B,C). Thus, Su(fu) prevents the Rdx-dependent inhibition of C-155 nuclear import.

Figure 3.

Su(fu) abrogated the Rdx-dependent inhibition of Ci-155 nuclear import. (A) Schematic of fluorescence recovery after photobleaching (FRAP) experimental design. Nuclear Ci-GFP was photobleached, and FRAP was monitored. (B) FRAP experiments were performed using the clone-8 cells transfected with the Ci-GFP and Hh expression plasmids in the presence or absence of the Su(fu) and Rdx expression plasmids. The data without Su(fu) expression plasmid was previously published (Seong et al. 2010). The averages of multiple experiments at each time point (Left, −Su(fu): n = 8 for −Rdx and n = 6 for +Rdx; Right, +Su(fu): n = 15 for −Rdx and n = 13 for +Rdx) are shown. (C) Individual images at selected times are shown.

To further examine the role of Su(fu) in the nuclear import of Ci-155, we investigated the effect of Su(fu) on the subcellular localization of Ci-155. When Ci-155 was expressed in clone-8 cells in the presence of leptomycin B (LMB), an inhibitor of nuclear export, Ci-155 was detected primarily in the nucleus in most cells (Fig. 4A,B). Expression of increasing amounts of Su(fu) enhanced the cell population in which Ci-155 was detected in both the nucleus and the cytoplasm (Fig. 4A,C), suggesting that Su(fu) caused Ci-155 to be retained in the cytoplasm. These results are consistent with previous reports indicating that Su(fu) causes retention of Ci-155 in the cytoplasm (Méthot & Basler 2000; Wang et al. 2000). When Rdx was coexpressed with Ci-155, the cell populations in which Ci-155 was localized in both the nucleus and cytoplasm, or primarily in the cytoplasm, increased. These results indicate that Rdx increased cytoplasmic levels of Ci-155 by inhibiting the nuclear import of Ci-155 (Fig. 4B,C), as previously reported (Seong et al. 2010). However, co-expression of Su(fu) with Rdx and Ci-155 increased the cell population in which Ci-155 was localized primarily in the nucleus. These results further support the hypothesis that Su(fu) disrupts the Rdx-dependent inhibition of nuclear import of Ci-155.

Figure 4.

Effect of Su(fu) on the subcellular localization of Ci-155. (A) Su(fu) sequestered Ci-155 in the cytosol in the absence of Rdx. Clone-8 cells were transfected with a mixture of the HA-Ci-155 expression vector (500 ng) and the Hh expression vector (500 ng), together with increasing amounts of the Su(fu) expression vector (0, 100, 200 or 500 ng). Cells were treated with LMB and immunostained with an anti-HA antibody. The cells were scored by counting those in which HA-Ci-155 was detected predominantly in the nucleus (N>C), in both the nucleus and the cytoplasm (N=C), or predominantly in the cytoplasm (N<C). The total number of cells examined was 300–370. (B) Su(fu) enhanced nuclear entry of Ci-155 in the presence of Rdx. Clone-8 cells were transfected with a mixture of the HA-Ci-155 (500 ng), Hh (500 ng) and Rdx (500 ng) expression plasmids, together with increasing amounts of the Su(fu) expression vector (0, 100, 200, 300 or 500 ng), and cells were then treated as described above. The total number of cells examined was 300–330. (C) Typical cells immunostained with the anti-HA antibodies in which the HA-Ci signal was mainly detected in the nucleus (N>C), in both the nucleus and the cytoplasm (N=C), or in the cytoplasm (N<C) are shown. Clone-8 cells were transfected with a mixture of the HA-Ci-155 (500 ng) and Hh (500 ng) expression plasmids, together with Rdx (0 or 500 ng) and Su(fu) (0, 200 or 500 ng) expression plasmids, and cells were then treated as described above.

Ectopic expression of Su(fu) allows Rdx to degrade Ci-155 in region 1

The results described above suggest that decreased levels of Su(fu) in region 1 may be required for the Rdx-dependent inhibition of nuclear import of Ci-155 and that, in the presence of Su(fu), as in region 2, Ci-155 is imported into the nucleus and is degraded by Rdx. To test this hypothesis, we examined whether Rdx could induce Ci-155 degradation in region 1, when Su(fu) was ectopically expressed. Overexpression of Su(fu) led to the up-regulation of Ci-155 (Fig. 5A, a and b), as previously reported (Ohlmeyer & Kalderon 1998). This Su(fu)-induced increase in the Ci-155 levels could be due to an increase in the nuclear Ci-155 levels as previous data suggested that the turnover of cytosolic Ci is more rapid than nuclear Ci (Seong et al. 2010). When both Su(fu) and Rdx were overexpressed in region 1, up-regulation of Ci-155 was not observed (Fig. 5A, c and d), suggesting that nuclear Ci-155 was degraded by Rdx.

Figure 5.

Su(fu) acts as a key regulator of Rdx function. (A) Rdx down-regulated Ci-155 in region 1 when Su(fu) was ectopically expressed. Clones of cells expressing Su(fu) alone (a) or both Su(fu) and Rdx (c) (GFP-positive, green) are superimposed with Ci-155 immunostaining (magenta) in (b) and (d). Anterior is to the left; dorsal is up. Vertical lines indicate the A/P boundary. The level of Ci-155 in the clones relative to the other cells in region 1 (indicated by the white bars at the bottom of the images) was quantified and depicted in the bar graph on the right. ***< 0.001 (n = 4 for −Rdx and n = 6 for +Rdx). (B) Rdx did not down-regulate Ci-155 in region 2, when Su(fu) levels were decreased. Clones of cells expressing Su(fu) ds-RNA alone (a), both Su(fu) ds-RNA and Rdx (b), or both Su(fu) ds-RNA and Rdx ds-RNA (c) (GFP-positive, green) are superimposed with Ci-155 immunostaining (magenta) in (b, d, and f). The level of Ci-155 in the clones relative to the other cells in region 2 (indicated by the blue bars at the bottom of the images) was quantified and depicted in the bar graph on the right. NS, no significant difference (n = 17 for −Rdx, n = 18 for +Rdx, and n = 15 for +Rdx ds-RNA).

Su(fu) is required for the Rdx-dependent degradation of nuclear Ci-155 in region 2

In region 2, where the moderate levels of Su(fu) are expressed, Rdx induces the degradation of nuclear Ci-155. If Su(fu) prevents the Rdx-dependent inhibition of Ci-155 nuclear import, a decrease in the levels of Su(fu) may result in Rdx-dependent inhibition of nuclear import of Ci-155 in this region and block the Rdx-induced degradation of nuclear Ci-155. To test this, clones expressing a Su(fu) double-stranded RNA (ds-RNA) were generated. The levels of Su(fu) protein were decreased in these cells (Fig. S2A in Supporting Information). In the clones generated in region 2, the levels of Ci-155 were decreased (Fig. 5B, a and b). Based on previous observations that prevention of nuclear import of Ci-155 by Rdx causes the down-regulation of Ci-155, possibly due to rapid turnover rate of cytosolic Ci-155 compared with nuclear Ci-155 (Seong et al. 2010), the reduced level of Ci-155 seen here suggests that a decrease in the level of Su(fu) prevented the nuclear import of Ci-155. In clones in which Rdx was coexpressed with Su(fu) ds-RNA, the levels of Ci-155 were very similar to Ci-155 levels in clones expressing Su(fu) ds-RNA alone (Fig. 5B, c and d). Furthermore, in clones in which Su(fu) ds-RNA was coexpressed with Rdx ds-RNA, the levels of Ci-155 were also similar to Ci-155 levels in clones expressing Su(fu) ds-RNA alone (Fig. 5B, e and f). These results suggest that a certain level of Su(fu) is required for Rdx to induce the degradation of nuclear Ci.

Ectopic expression of Su(fu) in region 1 lowered En expression

To further confirm the physiological role of the low level of Su(fu) in region 1, Su(fu) overexpression clones in this region were generated, and the expression of key regulators of pattern formation was examined. In these clones, ptc expression was unchanged, but the level of En was significantly reduced (Figs 6 and S3 in Supporting Information). Expression of En in region 1 is critical to maintain the integrity of the A/P compartment boundary (Blair 1992; Dahmann & Basler 2000). Therefore, these results suggest that keeping the Su(fu) levels low in region 1 is needed for proper pattern formation through the induction of En expression in this region.

Figure 6.

Overexpression of Su(fu) down-regulated En expression in region 1. Clones of cells overexpressing Su(fu) (GFP-positive, green) (A) are superimposed with En (B) or ptc (C) expression, and monitored by anti-En antibody (red) or the ptc-lacZ reporter (blue). Clones displaying Su(Fu) overexpression are outlined in white. Anterior is to the left; dorsal is up. Vertical lines indicate the A/P boundary.

Discussion

This study indicates that Su(fu) induces a switch in Rdx function (Fig. 7). When the Su(fu) level is low, as in region 1, Rdx blocks the nuclear import of Ci-155. The block is caused by releasing Ci from importin-α3, which is a nuclear import receptor for Ci (Seong et al. 2010). Conversely, when the Su(fu) level is relatively high, as in region 2, Rdx enters the nucleus and induces the degradation of nuclear Ci-155. In region 1, the strong Hh signal reduces the Su(fu) level and increases the Rdx level. In region 2, the lower levels of Hh result in a high Su(fu) level and a moderate Rdx level. Previously, Su(fu) was shown to compete with Rdx for binding to Ci-155 (Zhang et al. 2006). When the Su(fu) level is high, Su(fu) may disrupt the interaction between Rdx and Ci-155 in the nuclear pore, and Su(fu) may be released from Ci-155 after entering the nucleus.

Figure 7.

Model for the Su(fu)-induced switch in Rdx function. In region 1, abutting the A/P boundary, the Hh signal is strongest and the Rdx levels are high, while the Su(fu) levels are low. In this region, Rdx inhibits the nuclear import of Ci-155 by competing with importin-α3 for binding to Ci-155 at the nuclear pore complex, resulting in low levels of Ci-155 in the nucleus. Low levels of nuclear Ci-155 selectively activate the transcription of kn. Low levels of nuclear Ci-155 may be central to the selective induction of specific Hh target genes, such as kn and en. In region 2, distant from the A/P boundary, the Hh signal is moderate, the Rdx levels are moderate, and the Su(fu) levels are high. In this region, Su(fu) allows the nuclear import of Ci-155, possibly by blocking the Rdx-Ci-155 interaction at the nuclear pore, resulting in relatively high levels of Ci-155 in the nucleus. The nuclear Ci-155 level is negatively regulated by Rdx via proteasome-dependent degradation. In addition, Su(fu) could work as a corepressor to suppress Ci-dependent transactivation, as shown in vertebrate (Cheng & Bishop 2002).

Su(fu) was previously shown to inhibit the nuclear entry of Ci-155 by sequestering Ci-155 in the cytoplasm (Méthot & Basler 2000). In addition, Su(fu) can enter the nucleus together with Ci-155 in response to Hh signaling in some contexts (Sisson et al. 2006). Su(fu) has no candidate nuclear localization sequence, suggesting that it enters the nucleus when complexed with an interacting partner. Thus, Su(fu) may be carried into the nucleus only with the Ci-155/importin-α3 complex. The Ci-155/importin-α3/Su(fu) complex may enter the nucleus by blocking the Rdx nuclear import inhibitor function at the nuclear pore, allowing Su(fu) to then suppress the proteasome-dependent degradation of Ci by Rdx. In addition, nuclear Su(fu) may act as a corepressor to inhibit Ci-dependent trans-activation of gene expression because mammalian Su(fu) recruits a corepressor to the Gli transcription factor to mediate Gli-dependent transcriptional repression (Cheng & Bishop 2002). Thus, Su(fu) may regulate Ci activity by modulating the interaction between Ci and multiple interacting proteins.

Why does Rdx negatively regulate Ci-155 by two different mechanisms? The regulatory mechanisms may be correlated with the selection of target genes. In region 1, kn and en are expressed in response to the strong Hh signal. Loss of Rdx diminishes kn expression (Seong et al. 2010), and a loss of Rdx function by Su(fu) overexpression disrupts nuclear entry of Ci-155 and diminishes en expression (Fig. 6). To selectively activate those target genes, a small number of active Ci-155 might be needed. In contrast, in region 2, which receives moderate level of Hh signal, different target genes, such as dpp, are induced by Ci-155. In region 2, Su(fu) is also localized to the nucleus. If Drosophila Su(fu) has a corepressor function as in mammals, the amount of active Ci-155 that is not associated with Su(fu) may be small in region 2, and kn and en cannot be induced. In region 1, most of the nuclear Ci-155 may be the active form due to the absence of Su(fu). This speculation is consistent with the hypothesis of Ohlmeyer & Kalderon (1998) that the active form of Ci-155 exists only in region 1.

Experimental procedures

Drosophila strains

The following stocks were used in this study: rdx-lacZ (named P{PZ}mei-P190347 in Kent et al. 2006), ptc-lacZ (Tabata & Kornberg 1994), UAS-rdx (Seong et al. 2010), UAS-hh (Capdevila & Guerrero 1994), smo3 (Chen & Struhl 1996), UAS-Su(fu) (obtained from D. Busson; Dussillol-Godar et al. 2006), UAS-Su(fu)IR (obtained from R. A. Holmgren; Sisson et al. 2006), and MS1096-GAL4 (Capdevila & Guerrero 1994).

The flies used in this study had the following genotypes:

  • rdx-lacZ/+ (Figs 1B, 2A);

  • hs-flp/+; AyGAL4 UAS-GFP/+; UAS-hh/+ (Fig. 2A);

  • hs-flp/+; smo3 FRT40A/ubi-GFP FRT40A (Fig. 2B);

  • MS1096/+; UAS-hh/+ (Fig. 2C,D);

  • hs-flp/ UAS-Su(fu); AyGAL4 UAS-GFP/+ (Figs 5A-a, A-b and S2B in Supporting Information);

  • hs-flp/ UAS-Su(fu); AyGAL4 UAS-GFP/+; UAS-rdx/+ (Fig. 5A-c, A-d);

  • hs-flp/+; AyGAL4 UAS-GFP/ UAS-Su(fu) IR (Figs 5B-a, B-b and S2A in Supporting Information);

  • hs-flp/+; AyGAL4 UAS-GFP/ UAS-Su(fu) IR; UAS-rdx/+ (Fig. 5B-c, B-d);

  • hs-flp/+; AyGAL4 UAS-GFP/ UAS-Su(fu) IR; UAS-rdxIR/+ (Fig. 5B-e, B-f);

  • hs-flp/UAS-Su(fu); AyGAL4 UAS-GFP/+;ptc-lacZ/+ (Figs 6 and S3A in Supporting Information);

  • hs-flp/UAS-Su(fu); AyGAL4 UAS-GFP/+ (Fig. S3B in Supporting Information).

Flies were reared at 25 °C on a standard yeast/cornmeal/glucose/agar medium. Somatic mutant clones and GAL4/UAS-mediated overexpression clones were generated by FLP/FRT-mediated mitotic recombination and the Flp-out technique, respectively (Xu & Rubin 1993; Pignoni & Zipursky 1997). Each clone was generated by heat shock for 60 min at 37 °C, 48–72 h after egg laying.

Immunohistochemistry

Imaginal disks from Drosophila third instar larvae were fixed and stained using standard techniques. The primary antibodies used in this study were as follows: rat anti-Ci (2A1) (diluted 1:50, gift from R. Holmgren), rabbit anti-b-galactosidase (1:500, CAPPEL), mouse anti-Ptc (1:50, gift from I. Guerrero), mouse anti-En (4D9) [1:50, Developmental Studies Hybridoma Bank (DSHB)], and mouse anti-Su(fu) (25H3) (1:100, DSHB). We confirmed that the anti-Su(fu) 25H3 antibody can be utilized to specifically detect Su(fu) using the wing disks that contained the Su(fu) overexpressing clones or Su(fu) mutant clones (Fig. S2 in Supporting Information). Secondary antibodies coupled to Alexa-488 (Molecular Probes), Cy3 and Cy5 (Jackson) were used at 1:250. Immunofluorescence staining was visualized using a Zeiss 510 confocal microscope.

Western blotting

Clone-8 cells (4 × 106 cells/6 cm dish) were transfected with a mixture of the following plasmids: act5C-GAL4 (50 ng) and Su(fu) expression vector (0.5 μg), together with increasing amounts of the HA-Ci-155 expression vector (0, 0.5, or 1.5 μg) and the Hh expression vector (0, 1, or 3 μg). All transfections used the same total amount of DNA by adjusting the amount of empty vector added. We also used pRL, the Renilla luciferase reporter vector (100 ng), as an internal control for transfection efficiency. At 48 h post-transfection, the cells were lysed with lysis buffer [50 mm HEPES, pH 7.4, 250 mm NaCl, 0.5% NP-40, 0.2 mm EDTA, and protease inhibitor cocktail (Roche)]. The lysates were processed for Western blotting using standard protocols and then probed with an anti-Su(fu) antibody (25H3) (1:500, DSHB).

Quantitative RT-PCR

For real-time RT-PCR, total RNA was isolated from 20 wing imaginal disks by homogenization in 500 μL TRIzol reagent (Invitrogen), followed by DNase I treatment. PCR primers for Su(fu) and rp49 were as follows: Su(fu), forward 5′–TTCACCTTCAAGGCTCAGCAT–3′, reverse 5′–CGAGCCGGTGACGGATT–3′; and rp49, forward 5′–GACGCTTCAAGGGACAGTATCTG–3′, reverse 5′–AAACGCGGTTCTGCATGAG–3′. Real-time quantitative RT-PCR was performed using the SuperScript II Platinum SYBR Green One-Step qRT-PCR Kit (Invitrogen) and an Applied Biosystems ABI Prism 7000 Sequence Detection System.

Western blotting analysis of Ci-155

Clone-8 cells (4 × 106 cells/ 6 cm dish) were transfected with the plasmid to express Hh (1.5 μg), 3xFlag-Ci-155 (1.5 μg), HA-rdx (1.5 μg) using Cellfectin. In some cases, cells were treated with 50 μm MG132 or DMSO for 6 h before the preparation. Forty-eight hours after transfection, cells were lysed using the SDS sample buffer, and the lysates were used for Western blotting with anti-Flag (M2) (1:3000, SIGMA). To prepare the nuclear fraction, cells were lysed in hypotonic buffer (20 mm HEPES (pH 7.9), 10 mm KCl and protease inhibitors) and incubated on ice for 15 min. After a centrifugation, nuclei in the pellet fraction were resuspended into washing buffer (20 mm HEPES (pH 7.9), 10 mm KCl, 350 mm sucrose), and homogenized with the B pestle of a Dounce homogenizer. Nuclei were then separated by centrifugation, and resuspended into nuclear extraction buffer (20 mm HEPES (pH 7.9), 10 mm KCl, 420 mm NaCl, and protease inhibitors). After incubation on ice for 15 min, the soluble nuclear fraction was separated after centrifugation. Nuclear extracts were analyzed by immunoblotting with either anti-Flag (M2) (1:3000, SIGMA), anti-α-Tubulin (DM1A) (1:2000, SIGMA), or anti-Lamin (ADL67.10) (1:1000, DSHB).

Fluorescence recovery after photobleaching analysis

Clone-8 cells (2 × 106 cells/2.5 cm dish) were transfected with a mixture of the following plasmids: act5c-Hh (500 ng), UAS-Ci-GFP (Méthot & Basler 2000) (200 ng), act5C-GAL4 (50 ng), act5C-HA-Rdx (0 or 500 ng), and UAS-Su(fu) (400 ng). The total amount of DNA transfected was adjusted to 1.65 μg by adding empty vector. Cells were observed 48 h after transfection. FRAP analyses were performed using a 100× oil immersion objective, a 3X zoom on a Zeiss LSM510 confocal microscope, and the 488 nm line of a Kr/Ar laser operating at 85% laser power and 2% transmission. The stage temperature was maintained at 25 °C. For FRAP experiments, a specific region of the nucleus was selected and bleached using 100 iterations with 100% transmittance. Images were recorded at 1 s intervals. Cells were imaged 50 times prephotobleaching and 50 times postbleach, with maximum scan speed. For data analysis, the LSM software (Zeiss) was used to calculate the mean fluorescence intensity of the nuclear region. A fluorescence recovery curve was graphed using Excel, according to Rabut & Ellenberg (2005).

Subcellular localization of Ci

Using Cellfectin (Invitrogen), clone-8 cells (2 × 106 cells) in 6-well dishes were transfected with a mixture of plasmids to express the following proteins: for Fig. 3A, HA-Ci-155 expression vector (500 ng) and the Hh expression vector (500 ng), together with increasing amounts of the Su(fu) expression vector (0, 100, 200, or 500 ng); for Fig. 3B, HA-Ci-155 (500 ng), Hh (500 ng) and Rdx (500 ng) expression plasmids, together with the Su(fu) expression vector (0 or 500 ng). The total amount of DNA transfected was adjusted to 2 μg by adding empty vector. At 48 h post-transfection, the cells were treated with leptomycin B (LMB) (10 ng/mL) for 4 h, fixed with 4% paraformaldehyde/PBS for 30 min, and washed with PBS containing 0.3% Triton X-100 (PBTx) for 10 min. Samples were incubated with 10% goat serum/PBTx for 1 h and then incubated with an anti-HA antibody (3F10). After washing with PBTx, samples were incubated with Cy3-conjugated rat IgG. Cells were mounted with Gel/Mount (Biomeda) and examined by fluorescence microscopy (Zeiss).

Acknowledgements

We thank K. Basler, D. Busson, R.A. Holmgren, A. Vincent, and the Bloomington stock center for fly stocks, constructs, and antibodies. This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Ancillary