Oncogene (2013) 32, 4913–4920 & 2013 Macmillan Publishers Limited All rights reserved 0950-9232/13 www.nature.com/onc

ORIGINAL ARTICLE SIAH ligases target the nonreceptor tyrosine kinase ACK1 for ubiquitinylation and proteasomal degradation

M Buchwald1,4, K Pietschmann1,4, P Brand1,AGu¨ nther2, NP Mahajan3, T Heinzel1 and OH Kra¨mer1

Activated Cdc42-associated kinase 1 (ACK1) is a nonreceptor tyrosine kinase linked to cellular transformation. The aberrant regulation of ACK1 promotes tumor progression and metastasis. Therefore, ACK1 is regarded as a valid target in cancer therapy. Seven in absentia homolog (SIAH) ubiquitin ligases facilitate substrate ubiquitinylation that targets proteins to the proteasomal degradation pathway. Here we report that ACK1 and SIAH1 from Homo sapiens interact in a yeast two-hybrid screen. Protein–protein interaction studies and protein degradation analyses using deletion and point mutants of ACK1 verify that SIAH1 and the related SIAH2 interact with ACK1. The association between SIAHs and ACK1 depends on the integrity of a highly conserved SIAH-binding motif located in the far C-terminus of ACK1. Furthermore, we demonstrate that the interaction of ACK1 with SIAH1 and the induction of proteasomal degradation of ACK1 by SIAH1 are independent of ACK1’s kinase activity. Chemical inhibitors blocking proteasomal activity corroborate that SIAH1 and SIAH2 destabilize the ACK1 protein by inducing its proteasomal turnover. This mechanism apparently differs from the lysosomal pathway targeting ACK1 after stimulation with the epidermal growth factor. Our data also show that ACK1, but not ACK1 mutants lacking the SIAH binding motif, has a discernable negative effect on SIAH levels. Additionally, knockdown approaches targeting the SIAH2 mRNA uncover specifically that the induction of SIAH2 expression, by hormonally-induced estrogen receptor (ER) activation, decreases the levels of ACK1 in luminal human breast cancer cells. Collectively, our data provide novel insights into the molecular mechanisms modulating ACK1 and they position SIAH ubiquitin ligases as negative regulators of ACK1 in transformed cells.

Oncogene (2013) 32, 4913–4920; doi:10.1038/onc.2012.515; published online 3 December 2012 Keywords: ACK1; estrogen; ; SIAH1; SIAH2; TNK2

INTRODUCTION SIAH-mediated degradation of targeted proteins is associated with Seven in absentia homolog (SIAH) proteins are homologs of the severe consequences for the tumorigenicity of certain types of 1,2 Drosophila melanogaster Seven In Absentia (SINA) protein. All cancer. The targets of SIAHs are, for example, the deacetylase 6,7 8 7,9 members of the SIAH family are evolutionally conserved across HDAC3, the polycomb protein HPH2, the corepressor N-CoR, 10 species.1,2 In Homo sapiens, SIAH1 and SIAH2 are expressed in a broad the factor T-STAR, the cell cycle regulator range of tissues and share over 80% of amino acid homology and p27,11 the transcription factor b-catenin,12 the leukemia fusion 69.8% sequence identity.1–3 Both of them contain a highly conserved protein AF4-MLL13 and the TNF signaling adapter molecule Really Interesting New (RING) domain, two zinc finger domains TRAF2.14 We recently demonstrated that SIAH1 and SIAH2 and a C-terminal substrate-binding domain. Sequence deviations cooperate with the ubiquitin conjugase UBCH8 to induce mainly occur in the N-termini of SIAH1 and SIAH2 and both proteins proteasomal degradation of leukemia-associated oncoproteins, function as ubiquitin ligases via their RING domains.1,4 such as the mutated constitutively active receptor tyrosine kinase In eukaryotic cells, proteins are tightly regulated by the FLT3-ITD and the chimeric fusion proteins PML-RARa and AML1- ubiquitin–proteasome system, in which polyubiquitinylated pro- ETO. Such processes promote the death of leukemic cells.15–18 teins are cleaved by a multiprotein complex termed the protea- Remarkably, SIAH1 and SIAH2 are targets of the tumor suppressor some.5 Ubiquitinylation is usually carried out by three classes of p53.19,20 Activation of p53 due to genotoxic stress, for example, enzymes. These enzymes act in a concerted manner to recognize induces SIAH1 expression and degradation of the transcriptional and transfer ubiquitin molecules to substrate proteins. The three activator b-catenin, which leads to the suppression of cancer cell classes are a ubiquitin activating enzyme, ubiquitin conjugases and growth.1 Moreover, SIAH1 can restrict EGF-dependent cell growth ubiquitin ligases (E3s).4 The selectivity of ubiquitinylation is by degradation of the phospholipase Ce,21 and SIAH1 suppresses primarily determined by ubiquitin ligases. E3s containing a RING the survival and motility of transformed breast cells by a JUN domain, such as the SIAH proteins, possess no intrinsic catalytic kinase-dependent induction of the proapoptotic protein BIM.22,23 activity. By binding ubiquitin conjugases, they coordinate the The aim of this study was to identify novel SIAH targets that formation of polyubiquitin chains on their substrates.4 might play a role in cancer development. By a systematic search, SIAH ubiquitin ligases facilitate the proteasomal degradation of we identified the (proto) oncogenic nonreceptor tyrosine kinase diverse substrate proteins with multiple functions. Accordingly, activated Cdc42-associated kinase 1 (ACK1) as a novel interaction

1Institute of Biochemistry and Biophysics, Center for Molecular Biomedicine, Friedrich-Schiller University, Jena, Germany; 2Department of Internal Medicine II, University of Giessen Lung Center, Giessen, Germany and 3Drug Discovery Department, Moffitt Cancer Center, Tampa, FL, USA. Correspondence: Dr OH Kra¨mer, Institute of Biochemistry and Biophysics, Center for Molecular Biomedicine, Friedrich-Schiller University, Hans-Kno¨ll-Street 2, 07745 Jena, Germany. E-mail: [email protected] 4These authors contributed equally to this work. Received 26 December 2011; revised 28 September 2012; accepted 2 October 2012; published online 3 December 2012 Estrogen- and SIAH2-dependent proteasomal ACK1 degradation M Buchwald et al 4914 partner of SIAH1. We demonstrate that SIAH1 and SIAH2 facilitate the ubiquitinylation and degradation of ACK1 by the ubiquitin– proteasome system. Furthermore, we show that the association of SIAH1 and SIAH2 with ACK1 relies on a highly conserved binding site within ACK1. In addition, estrogen induces the expression of SIAH2, which we have identified here as a limiting factor for the degradation of ACK1 in transformed breast cancer cells. Our results suggest a new biological role for SIAH proteins as regulators of ACK1 expression in cancer cells.

RESULTS Identification of ACK1 as a binding partner of SIAH1 In order to find novel interaction partners of SIAH1, we performed yeast two-hybrid screening with full-length SIAH1 (as the bait) against human brain, testis and open reading frame cDNA libraries (as the prey). Under stringent screening conditions, 16 positive clones were identified to interact with SIAH1. Out of these clones, five have already been described as SIAH1 interaction partners (Figure 1a), for example, the SIAH1 interacting protein/CacyBP or SIAH1 itself.1 Two positive clones from the brain cDNA library encoded a large C-terminal fragment of human ACK1 (amino acids 819–1038). The association of ACK1 with SIAH1 has so far not been described. We performed protein–protein binding assays to test the initial result of an interaction between SIAH1 and ACK1. Since SIAH1 displays over 80% sequence similarity with the related protein SIAH22,3 and as both share an overlapping spectrum of target proteins, we also tested for a putative interaction between SIAH2 and ACK1. We performed an in vitro GST pulldown assay with bacterially expressed GST-SIAH1 or GST-SIAH2 and full-length ACK1- Myc from HEK293T cells. ACK1 efficiently bound to the GST-SIAH fusion proteins but not to GST (Figure 1b), which corroborated the interaction of ACK1 with SIAH1 and SIAH2. Due to the strong self- ubiquitinylation and degradation of SIAH1, the interaction of SIAH1 its with binding partners is difficult to detect in coimmunopreci- Figure 1. Identification of ACK1 as SIAH1 interacting protein. (a) pitation assays.1,6,16,24 Therefore, we used established RING mutants 2 Yeast two-hybrid screens were performed with full-length SIAH1 (as of SIAH1 and SIAH2. This approach is commonly used to test the bait) against human testis and brain cDNA libraries as well as a interaction of SIAHs with their substrates, since these mutants still library of human full-length open reading frame cDNAs (as prey). bind to their targets but fail to recruit ubiquitin conju- The table displays a summary of the yeast two-hybrid results. Listed gases.3,14,16,17,25 Indeed, we could coprecipitate SIAH1 as well as are prey proteins found in the screens that were already described SIAH2 with endogenous ACK1 from MCF-7 cells (Figure 1c). Taken as SIAH1-interacting proteins and the newly identified SIAH together, these data suggest that SIAH1 and SIAH2 occur in protein interaction partner ACK1. For each prey the numbers of isolated complexes containing ACK1. clones and numbers of different libraries in which this pair has been found are shown. (b) Detection of ACK1-Myc bound to GST-SIAH1 (left panel) or GST-SIAH2 (right panel) or GST (as control) in a GST SIAH proteins mediate the degradation of ACK1 by the ubiquitin– pulldown assay. Bacterially expressed purified GST and GST-SIAH1/2 proteasome system were incubated with total cell lysate protein of ACK1-Myc SIAH proteins mediate the ubiquitinylation of substrate proteins to transfected HEK293T cells (IN, input representing 20% of cell lysate). Bound ACK1-Myc was detected by immunoblotting. Levels of GST promote their proteasomal turnover. However, the interaction fusion proteins were analyzed by membrane staining with Ponceau with SIAH1 and SIAH2 does not necessarily destabilize their 2,3,26 red (left panel) or by immunoblotting (right panel); kDa, relative binding partners. Therefore, we investigated whether SIAH molecular weight. (c) RING mutants of SIAH1 or SIAH2 were proteins mediate the degradation of ACK1. HEK293T cells were overexpressed in MCF-7 cells. SIAH1/SIAH2 was immunoprecipitated transiently transfected with plasmids encoding ACK1-Myc and from cell lysates. The presence of endogenous ACK1 and SIAH1/ SIAH1 or SIAH2. Expression of the SIAHs resulted in a strong SIAH2 was analyzed by immunoblotting (Input, 10% of cell lysate; reduction of ACK1 protein levels and this was dependent on the pre, preimmune serum; IP, immunoprecipitation). concentration of SIAHs (Figure 2a). This inverse correlation between ACK1 and SIAH1 or SIAH2 identifies ACK1 as a new allowed the decay of ACK1 to be judged. The stability of ACK1 was proteasomal target of SIAH1 and SIAH2. The protein levels of drastically decreased when SIAH1 or SIAH2 were present. cotransfected green fluorescent protein as an internal control Expression of the inactive SIAH2 corroborated that SIAH2’s remained stable, which illustrates the specific effect of SIAHs on catalytic activity as is a prerequisite for the ACK1 (Figure 2a). accelerated turnover of ACK1 (Supplementary Figure S1). Since these data suggest that SIAH ubiquitin ligases accelerate Next, we addressed whether endogenously expressed SIAHs the proteasomal degradation of ACK1, we also assessed whether also regulate ACK1 levels. We expressed ACK1-Myc in HEK293T the half-life of ACK1 is altered in the presence of SIAHs. We cells and lowered endogenous SIAH expression by transfecting expressed ACK1 alone or with SIAH1, SIAH2 or an inactive RING these cells with small interfering RNAs (siRNAs) against SIAH1/ mutant of SIAH2 in HEK 293T cells. Treatment with the ribosomal SIAH2 or with small hairpin RNA (shRNAs) specifically targeting inhibitor cycloheximide stopped de novo protein synthesis and SIAH1 or SIAH2.27 The reduction of SIAH protein expression

Oncogene (2013) 4913 – 4920 & 2013 Macmillan Publishers Limited Estrogen- and SIAH2-dependent proteasomal ACK1 degradation M Buchwald et al 4915 ACK1-Myc with SIAH1 in HEK293T cells and added the pro- teasome inhibitor MG132 to block proteasomal degradation of the ubiquitinylated proteins. To detect ubiquitinylated ACK1, ACK1- Myc was immunoprecipitated from cell lysates with anti-Myc antibodies followed by immunoblotting with anti-ubiquitin antibodies. A slower migrating smear was detected in the ACK1 immunoprecipitates from lysates with SIAH1 and ACK1 expression (Figure 2c). The higher molecular weight of these ACK1 species indicates that SIAH1 mediates the ubiquitinylation of ACK1. Polyubiquitinylation of proteins is connected to their proteolytical cleavage by the proteasome. In order to establish whether ACK1 is degraded by the proteasome as a consequence of SIAH- dependent polyubiquitinylation, we tested if ACK1-Myc remains stable in the presence of increased levels of SIAH1 or SIAH2 in HEK293T cells with chemically inactivated . Since ACK1 was found to be subjected to lysosomal degradation under specific cellular conditions28 and MG132 was also described to inhibit certain lysosome-associated proteases, we used the highly proteasome-specific inhibitor lactacystin.29 Coexpression of ACK1- Myc and SIAH1 or SIAH2 led to a robust degradation of the ACK1 protein that could be blocked by treatment with lactacystin (Figures 2d and e). These experiments demonstrate that SIAHs regulate the proteasomal degradation of ACK1. Of note, ACK1-Myc protein was also stabilized by lactacystin in the absence of overexpressed SIAHs (Figure 2d). This finding suggests a basal proteasomal turnover of ACK1-Myc that is in agreement with the notion that endogenous SIAH1 and SIAH2 can control the protein levels of ectopically expressed ACK1 in HEK293T cells (Figure 2b and Supplementary Figure S2). Moreover, SIAH1 and SIAH2 were stabilized in the presence of lactacystin (Figures 2d and e), which is consistent with previous reports demonstrating their high proteasomal turnover.1,16,17,25 To assess whether overexpression of other ubiquitin ligases can Figure 2. SIAH ubiquitin ligases trigger the proteasomal degradation cause degradation of ACK1, we coexpressed ACK1-Myc together of ACK1. (a) Myc-tagged ACK1 (0.5 mg) was transfected alone or with HA-tagged RLIM, Myc-tagged TRIAD1 and MDM2 ubiquitin together with different amounts of SIAH1 or SIAH2 (0.05–0.2 mg) into ligases in HEK293T cells. Coexpression of SIAH1 and SIAH2 HEK293T cells. Green fluorescent protein was cotransfected to resulted in a complete loss of ACK1, whereas ACK1 remained monitor equal transfection efficiency. Protein expression was stable in the presence of coexpressed RLIM, TRIAD1 and MDM2 analyzed by immunoblotting (SIAH1/2: polyclonal antibodies were (Figure 2f). These findings suggest that SIAH proteins function as used for the detection of SIAH1 and SIAH2). Detection of tubulin specific ubiquitin ligases for polyubiquitinylation and subsequent served as loading control. (b) ACK1-Myc (0.2 mg) was expressed in Hestrogen 93T cells and the knockdown of SIAH expression was proteasomal degradation of ACK1. achieved by transfection of shRNA plasmids (1 mg) directed against SIAH mRNA (shSIAH1/2) or unspecific shRNA as control. Immunoblot- ting was performed to detect protein expression. (c) ACK1-Myc and SIAH1 binds and degrades ACK1 independent of its kinase activity SIAH1 were expressed for 24 h in HEK293T cells. Where indicated, cells SIAH1 and SIAH2 have recently been reported to undergo were incubated 6 h before cell lysis with the proteasome inhibitor phosphorylation at multiple tyrosine and serine residues and MG132 (10 mM) to block the degradation of ubiquitinylated proteins. these modifications have been shown to regulate their substrate For detection of ubiquitinylated ACK1, ACK1-Myc was immunopre- affinity and degradation.19–21,26,30 Furthermore, it is known that cipitated with an anti-Myc antibody and precipitates were immuno- the specificity of SIAHs for certain targets can depend on the blotted for the presence of the ubiquitin (Ub) modification (IgG, 15 immunoglobulin heavy chain). (d) ACK1-Myc was coexpressed with substrate phosphorylation status. ACK1 is a tyrosine kinase that SIAH1 in HEK293T cells in the absence or presence of the very catalyzes autophosphorylation and phosphorylation of interacting 31–34 proteasome specific inhibitor lactacystin (Lact., 10 mM), added 8 h prior proteins. To elucidate the relevance of ACK1 kinase activity cell lysis. ACK1 and SIAH1 protein levels were analyzed by for the degradation by SIAH1, we expressed SIAH1 together with immunoblotting. (e) ACK1-Myc was coexpressed with SIAH2-Myc in wild-type ACK1, a kinase dead mutant (kdACK1) or a constitutively HEK293T cells in the absence or presence of lactacystin and analyzed active mutant (caACK1) of ACK1 in HEK293T cells.31 ACK1 was as described in (d). (f) To determine the specificity of the SIAH- immunoprecipitated and the phosphorylation status of ACK1 was mediated ACK1 turnover, ACK1 was coexpressed with SIAH1, SIAH2, verified by immunoblotting with antibodies recognizing RLIM-HA, TRIAD1-Myc and MDM2 ubiquitin ligases. Cell lysates were phosphorylated tyrosine residues (Figure 3a, 2nd panel). Wild- analyzed by immunoblotting. Green fluorescent protein and tubulin served as controls for equal transfection efficiency and equal loading, type ACK1, caACK1 or kdACK1 exhibited no significant differences respectively (n.s., nonspecific band). regarding their SIAH1-mediated degradation (Figure 3a, top panel). To investigate whether the interaction between SIAH1 and ACK1 depends on the kinase activity of ACK1, we performed a increased the levels of ACK1-Myc (Figure 2b and Supplementary GST pulldown assay using GST-SIAH1 and wild-type ACK1, caACK1 Figure S2). These experiments show that endogenous as well as or kdACK1 expressed in HEK293T cells. GST-SIAH1 associated well ectopically expressed SIAH proteins regulate ACK1 protein levels. with all the three ACK1 variants in the pulldown assay (Figure 3b). As SIAH ubiquitin ligases promote polyubiquitinylation target- These data coherently suggest that the interaction of these factors ing proteins for proteasomal degradation,1,4 we analyzed if ACK1 and the SIAH1-induced proteasomal degradation of ACK1 are undergoes SIAH1-mediated ubiquitinylation. We coexpressed independent of its kinase activity.

& 2013 Macmillan Publishers Limited Oncogene (2013) 4913 – 4920 Estrogen- and SIAH2-dependent proteasomal ACK1 degradation M Buchwald et al 4916

Figure 3. SIAH1 facilitates ACK1 degradation independent of ACK1 activity. (a) Cells were transfected with SIAH1 and wild-type (WT), constitutively active (ca) or kinase dead (kd) ACK1-Myc followed by analysis of cell lysates by immunoblotting. Kinase activity of ACK1 was determined by immunoprecipitation (IP) of ACK1-Myc and immunoblotting (IB) with a phospho-tyrosine (pY) specific antibody for detection of phosphorylated ACK1. (b) HEK293T cells were transfected with ACK1-Myc expression vectors as described in (a). Pulldown experiment was performed by incubating HEK293T cell lysates with bacterially expressed GST or GST-SIAH1. Interaction of GST-SIAH1 with ACK1-Myc was assessed by immunoblotting against precipitated ACK1-Myc (IN, input representing 10% of cell lysate). GST fusion proteins were detected by Ponceau-staining of the membrane.

Identification of a SIAH binding site in the far C-terminus of ACK1 Estrogen-induced SIAH2 regulates the levels of ACK1 in MCF-7 The majority of SIAH interacting proteins have a SIAH-binding site breast cancer cells also named degron motif (abbreviation for ‘degradation on’).2,35,36 Since estrogen upregulates SIAH2 in various breast cancer cell The consensus peptide sequence for a degron motif comprises lines,9,38,39 we examined the impact of induced SIAH2 on ACK1 PxAxVxP, with the core sequence VxP (x, random amino acid; V, expression in this system. We incubated MCF-7 breast cancer cells valine; P, proline). Although the degron motif is not a prerequisite with different concentrations of estrogen and monitored SIAH2 for the association with SIAHs, all of the SIAH1-interacting proteins and ACK1 protein expression by immunoblotting. The incubation that we identified in our yeast two-hybrid assay possess conserved of MCF-7 cells with estrogen upregulated SIAH2 expression and degrons (Figure 4a). In silico analysis of the ACK1 peptide seq- this correlated with a reduction of endogenous ACK1 protein uence revealed an amino acid stretch, PTATVRP, which matched levels (Figure 5a). We subsequently analyzed whether estrogen perfectly with the degron consensus sequence (Figure 4a). This might have a differential effect on SIAH1 and SIAH2. Real-time PCR sequence is located in the C-terminal proline rich region of ACK1 analyses and immunoblotting for the SIAHs revealed that estrogen (Figure 4b, 1.). The ACK1 clone that we identified in the yeast two- increases the protein and mRNA expression of SIAH2 but not hybrid assay encodes the C-terminal fragment of ACK1 from of SIAH1 (Figures 5b and c). Experiments with lactacystin verified amino acid 819–1038 (Figure 4b, 3.), which also comprises the that estrogen induces the proteasomal degradation of ACK1 putative SIAH degron. Collectively, these data suggest that SIAH1 (Figure 5d). This compound even reduced ACK1 mRNA levels bound to ACK1 within this region. To further analyze the putative strongly (Figure 5e), but stabilized the ACK1 protein in the role of the SIAH degron in ACK1, we performed a pulldown presence of estrogen (Figure 5d). These data illustrate that ACK1 binding assay with a C-terminally truncated mutant of ACK1 levels are mainly controlled at the protein level. The loss of ACK1 lacking the degron sequence (wACK1, Figure 4b, 2.). GST-SIAH1 mRNA in the presence of a proteasomal inhibitor is likely due to bound to full-length ACK1 but not to C-terminally truncated the frequently observed stabilization of transcriptional corepres- wACK1 (Figure 4c). In agreement with this finding, ACK1 was sors by such agents.7,40 efficiently degraded by SIAH1 whereas wACK1 was unaffected Next, we investigated whether the estrogen-induced down- (Figure 4d). Already in the absence of overexpressed SIAH regulation of ACK1 depends on increased SIAH2 expression in proteins, we detected a much higher expression of wACK1 MCF-7 cells. We observed that endogenous ACK1 levels were compared to the basal expression of full-length ACK1. This is significantly reduced upon estrogen stimulation of cells trans- congruent with previous publications reporting that full-length fected with control siRNA, whereas the reduction in ACK1 was 31,37 ACK1 is expressed poorly compared to wACK1. prevented in cells transfected with siRNA targeting SIAH2 It has been reported that SIAH1 targets T-STAR. The peptide (Figure 5f and Supplementary Figure S3). Similar to endogenous sequence of murine and human T-STAR differs by a single amino ACK1, cotransfection of MCF-7 cells with ACK1-Myc and shRNA acid in the degron VxP core consensus motif (glycine instead of a specifically targeting SIAH2 blocked the estrogen-mediated valine present in human T-STAR). This difference suffices to render downregulation of ACK1 (Figure 5g), indicating that induced murine T-STAR insensitive to SIAH1-mediated degradation.10 To SIAH2 negatively regulates ACK1. The knockdown efficiency of further verify the SIAH1 binding site in ACK1, we mutated the core SIAH2 mRNA by shRNA or siRNA is demonstrated in Supple- sequence of the degron in ACK1 by a single amino acid substitution of valine to glycine, at the position 909 (ACK1V909G). mentary Figures S4 and S5. To further verify the role of SIAH2 expression in reducing the levels of ACK1, we transfected MCF-7 Of note, this V909G point mutation in the ACK1 degron abolished V909G the interaction of ACK1V909G with GST-SIAH1 (Figure 4e). Con- cells with wild-type ACK1 or SIAH-binding deficient ACK1 . The sistently, the SIAH1 binding-deficient mutant ACK1V909G displayed cells were then treated with estrogen and ACK1 protein stability an enhanced protein stability compared with wild-type ACK1 was assessed. Remarkably, wild-type ACK1 was strongly degraded when coexpressed with SIAH1 or SIAH2 in HEK293T cells upon estrogen treatment and the SIAH-binding deficient mutant (Figure 4f). Interestingly, we likewise observed that ACK1V909G ACK1V909G remained stable (Figure 5h). significantly increased SIAH1 and SIAH2 levels (Figure 4f). This was Finally, we tested whether the control of ACK1 by SIAH2 is similarly observed for wACK1 (Figure 4d), which like ACK1V909G limited to estrogen-responsive breast cancer cells. We used MDA- cannot bind to SIAHs (Figures 4c and e). Taken together, these MB-231 basal-like breast cancer cells lacking a functional estrogen data suggest that both SIAH proteins promote the proteasomal receptor (ER) and T47D breast cancer cells. Like MCF-7 cells, T47D degradation of ACK1 dependent on an intact SIAH degron motif in cells are luminal breast cancer cells harboring a functional ER.41–43 ACK1. Furthermore, the inability of SIAHs to accelerate proteaso- We found that estrogen induces SIAH2 and reduces ACK1 in T47D mal degradation of ACK1V909G increases the stability of SIAH cells, but not in MDA-MB-231 cells (Figure 5i). Moreover, MDA- proteins. MB-231 cells reconstituted with ERa have reduced ACK1 levels

Oncogene (2013) 4913 – 4920 & 2013 Macmillan Publishers Limited Estrogen- and SIAH2-dependent proteasomal ACK1 degradation M Buchwald et al 4917

Figure 4. SIAH proteins bind to a distinct C-terminal binding site in ACK1. (a) Alignment of the consensus sequence of the SIAH binding motif (degron) to the peptide sequences of ACK1 and already known SIAH-interacting proteins found in the yeast two-hybrid screens. (b) Schematic representation of ACK1 variants with domain structure: (I) full-length ACK1 (ACK1), (II) C-terminal truncated ACK1 (wACK1); (III) prey fragment of ACK1 identified in the yeast two-hybrid screen (Y2H-ACK1) (kinase domain, SH3 domain, Cdc42/Rac binding domain (C), proline-rich domain (PR), ubiquitin binding domain (U), SIAH degron). (c) Cell extracts of Myc-tagged ACK1 or wACK1 transfected HEK293T cells were used for pulldown experiments with recombinant GST-SIAH1 or GST (IN, Input representing 10% of cell lysate; equal input levels of ACK1 WT and wACK1 were obtained by adjusted lysate volumes). ACK1-Myc protein was detected by immunoblotting. Levels of GST fusion proteins were analyzed by Ponceau-staining of membrane. (d) HEK293T cells were transfected with expression vectors for SIAH1 and full-length wild-type (WT) or C-terminal truncated (wACK1) mutant of ACK1-Myc. Protein expression was analyzed by immunoblotting (asterisk indicates adjusted exposition signals of WT ACK1 and wACK1). (e) The core sequence of the putative SIAH-binding motif in ACK1 was mutated by a single amino acid substitution of valine to glycine at amino acid position 909 (ACK1V909G). Pulldown experiments were performed with lysates from wild- type ACK1 (WT) or ACK1V909G transfected HEK293T cells and GST or GST-SIAH1 as described in (c). The lower panel shows the densitometric analysis of precipitated ACK1 (WT) or ACK1V909G protein levels (means±s.d., n ¼ 3). (f) Myc-tagged wild-type ACK1 (WT) or mutant ACK1V909G were coexpressed with SIAH1 or SIAH2 in HEK293T cells and a following immunoblot analysis was done as stated.

(Supplementary Figure S6). These data suggest that intact ER degradation of the tumor suppressor WWOX.31,32,47 Moreover, ACK1 signaling is necessary to induce SIAH2 and to reduce ACK1. enhances oncogenic epidermal growth factor (EGFR) signaling and In summary, our results suggest that estrogen and SIAH2 act as was shown to increase proliferation and invasiveness of renal and novel regulatory factors for ACK1 in breast cancer cells. The model breast cancer cells.48,49 In a murine breast cancer metastasis model, shown in Figure 6 summarizes the findings of the work we present ACK1 overexpression tied in with increased mortality of the mice.46 here. Due to these oncogenic properties, ACK1 has emerged as a valid target for cancer therapy.34,50 We show here for the first time that ACK1 can be subjected to DISCUSSION SIAH-dependent proteasomal degradation. Earlier, lysosomal degra- ACK1 is a B140 kDa Cdc42-associated nonreceptor tyrosine kinase dation of ACK1 was reported. The HECT E3 ubiquitin ligase Nedd4-1 that is activated by multiple extracellular stimuli.32,33,44,45 It is was shown to mediate an EGFR activation-dependent degradation overexpressed in many cancer types and is involved in tumori- of ACK1 via this pathway.28 Interestingly, the downregulation of genesis.31,33,46 For example, activated ACK1 promotes prostate ACK1 by Nedd4-2 requires ACK1 kinase activity.51 SIAH proteins cancer progression by activation of the androgen receptor and interact with and degrade ACK1 regardless of its kinase activity. The

& 2013 Macmillan Publishers Limited Oncogene (2013) 4913 – 4920 Estrogen- and SIAH2-dependent proteasomal ACK1 degradation M Buchwald et al 4918 proteins.11,12 Experiments addressing this possibility are underway and will decipher the details of the SIAH-ACK1 interaction in more detail. Interestingly, we detected lower SIAH1 and SIAH2 levels when coexpressed with degradable wild-type ACK1, and higher levels of the SIAHs in the presence of nondegradable, binding- deficient wACK1 or ACK1V909G. Apparently, ACK1 has a negative effect on SIAHs. Activation of the ER has been reported to induce the expression of SIAH2 in breast cancer cells.9,38,39 Here, we demonstrate that the upregulation of SIAH2 by estrogen negatively affects the levels of ACK1 in the luminal breast cancer cells MCF-7 and T47D. These findings suggest that induced SIAH2 regulates endogenous ACK1 by accelerating its proteasomal turnover. Since both SIAH1 and SIAH2 can reduce ACK1, modulating their expression with small molecular compounds may attenuate oncogenic ACK1 expression. Furthermore, SIAH1 and SIAH2 are induced by the tumor suppressor p531,19,20 and this may regulate ACK1 and thereby cancer cell growth. To test a putative role for p53 in the estrogen- dependent control of SIAH2, we used T47D cells. Like MCF-7 cells, T47D cells harbor a functional ER41 but contrary to MCF-7 cells, T47D cells lack functional p53.42 We found that T47D cells induce SIAH2 and reduce ACK1 when exposed to estrogen. These data suggest that estrogen does not require intact p53 to induce SIAH2 in luminal breast cancer cells. On the other hand, it needs to be stressed that these data do not exclude that p53 is necessary to activate SIAHs in response to other stimuli. Data collected with ER-negative MDA-MB-231 cells43 propose that estrogen requires functional ER signaling to activate the expression of SIAH2. Consistently, it was found that the ER antagonist ICI182,780 prevents the accumulation of SIAH2 in estrogen-stimulated MCF-7 cells and that SIAH2 is induced by estrogen in several human ER-positive breast cancer cell lines (MCF-7, ZR75-1, BT474).9,38,39 Curiously, the stimulation of estrogen-dependent ZR75-1/HERc cells with EGFR opposed the Figure 5. SIAH2 regulates ACK1 stability in the breast cancer cell line 38 MCF-7. (a) MCF-7 cells were left untreated or incubated with estrogen induction of SIAH2 by estrogen. It is possible that modulators of (E2; 10 nM,20nM) for 48 h and ACK1 protein levels were examined by estrogen signaling disrupt such mechanisms and this might in immunoblotting. For detection of SIAH2 protein, SIAH2 was turn promote ACK1 induction linked to EGFR signaling.38,52 Truly, immunoprecipitated (IP) followed by immunoblotting (IB). (b) Cells other effects such as the suppressive effect of SIAH2 on the tumor were treated with estrogen (20 nM) for 48 h and SIAH1 and SIAH2 suppressor C/EBPd53 may also play an important role in the vast protein levels were detected as described in (a). (c) mRNA levels of complexity of breast tumorigenesis. Further experiments are SIAH1 and SIAH2 were analyzed by real-time qPCR (n ¼ 3; ***Po0,001). necessary to fully elucidate the many roles that SIAH proteins Treatment conditions were as in (b). (d) Cells were incubated with play during cancer development. estrogen (20 nM) for 48 h in presence or absence of lactacystin (10 mM), added 24 h prior cell lysis and ACK1 protein levels were examined by In summary, we reveal a novel molecular mechanism regulating immunoblotting. (e) After treatment of cells as in (d), mRNA levels of the expression of ACK1. We demonstrate that the ubiquitin ligases ACK1 were analyzed by real-time qPCR. (f) MCF-7 cells were SIAH1 and SIAH2 control the protein stability of this kinase. These transfected witch siRNAs directed against SIAH1/SIAH2 mRNA (siSIAH) results might have implications for tumor therapy, especially in or nontargeting control siRNA (siControl). Cells were treated with designing small molecules targeting ACK1 activity or stability. estrogen (20 nM, 48 h) and cell lysates were examined for endogen- ously expressed ACK1 level by immunoblotting. (g) MCF-7 cells were transfected with ACK1-Myc cDNA and SIAH2 specific shRNA plasmids MATERIALS AND METHODS (or non-targeting control shRNA) and further treated and analyzed as Drugs and chemicals described in (f). Ectopical ACK1-Myc levels were detected by immunoblotting. (h) Wild-type ACK1 (WT) or SIAH-binding deficient The proteasome inhibitor Z-Leu-Leu-Leu-al (MG132), N-ethyl-maleimide, ACK1V909G were expressed in MCF-7 cells and cells were left untreated estrogen (17b-estradiol), doxycycline hyclate (doxocycline), cycloheximide or treated for 48 h with estrogen (20 nM). Immunoblotting was and polyethylenimine were purchased from Sigma-Aldrich (Steinheim, performed (asterisk indicates adjusted exposition signals of WT ACK1 Germany); clasto-lactacystin b-lactone (lactacystin) was purchased from and ACK1V909G). (i) ER-negative MDA-MB-231 and ER-positive T47D Axxora Alexis (Lo¨rrach, Germany). cells were incubated with estrogen (20 nM) for 48 h and protein levels of ACK1, SIAH2, and tubulin were analyzed as stated in (a). Cell culture HEK293T cells were maintained in Dulbecco’s modified Eagle’s medium results of our study hence suggest that two distinct biochemical (DMEM) supplemented with 10% fetal calf serum, 1% penicillin/strepto- pathways (that is, via lysosomes or proteasomes) exist for ACK1 mycin and 2% L-glutamine. MCF-7 cells were maintained in Rosewell Park in vivo. We additionally show that the SIAH binding motif located in Memorial Institute (RPMI) medium with the same additives. For experi- the very C-terminus of ACK1 dictates its SIAH-dependent turnover. ments with estrogen, MCF-7, T47D and MDA-MB-231 cells were cultured for a minimum of 6 days as described,9 in phenol-red free RPMI medium This so-called SIAH degron motif is an established interaction surface 1,2,35 containing 1% penicillin/streptomycin, 2% L-glutamine and 5% charcoal for SIAHs and their targets. Nevertheless, we cannot exclude stripped fetal calf serum (PAA Laboratories, Co¨lbe, Germany). All cell lines that additional factors may be required for the interaction of SIAHs were cultured at 37 1C in a 5% CO2 atmosphere. MCF-7 and T47D cells are and ACK1. An example is SIAH1 interacting protein/CacyBP that of the luminal A type (ER þ , prolactin receptor þ /– and EGFR HER2–) and has been shown to mediate interactions of SIAH1 with other MDA-MB-231 cells are of the claudin-low type (ER–, prolactin receptor– and

Oncogene (2013) 4913 – 4920 & 2013 Macmillan Publishers Limited Estrogen- and SIAH2-dependent proteasomal ACK1 degradation M Buchwald et al 4919

Figure 6. Model summarizing our findings for a proteasomal mechanism dictating ACK1 protein stability. In the absence of estrogen breast cancer cells express ACK1 and low levels of SIAH2 (left panel). Induction of ER signaling with estrogen induces SIAH2 to mediate proteasomal degradation of ACK1. SIAH2 interacts with ACK1 and can regulate the ubiquitinylation of ACK1 and its subsequent degradation by proteasomes (right panel).

EGFR HER2–).41,43 MCF-7 cells express wild-type p53 and T47D cells lack Immunoprecipitation volumes were adjusted to 600 ml with NETN buffer. functional p53.42 Precipitation was performed under continuous rotation for 16 h at 4 1C. The beads were washed and bound proteins were eluted with Laemmli buffer Construction of DNA Plasmids and mutagenesis; Yeast two-hybrid and analyzed by immunoblotting. For the detection of ubiquitinylated screening ACK1, all the buffers were supplemented with 10 mM N-ethyl-maleimide to Details on these materials and methods are provided in the Supplementary inhibit deubiquitinylating isopeptidases and 10 mM MG132 to block Methods section accompanying the article. proteasome-mediated degradation. Antibodies were purchased from Santa Cruz Biotechnology (Heidelberg, Germany; ACK1, sc-28336; SIAH1, sc-5506; SIAH2, sc-5507; green fluorescent Transfection assays and RNA interference protein, sc-9996), New England Biolabs (Frankfurt/Main, Germany; Myc-Tag, HEK293T cells were transfected with polyethylenimine as described.16 #2276; Phospho-Tyrosine p-Tyr-100, #9411), Sigma-Aldrich (tubulin, T5168; Unless otherwise stated, transient protein expression in HEK293T cells was Ubiquitin, U5379), Covance (Freiburg im Breisgau, Germany; HA-Tag, MMs- carried out for 48 h using the following amounts of cDNA: Myc-tagged 101P); Calbiochem (Darmstadt, Germany; MDM2, OP115). ACK1, caACK1, kdACK1, wACK1 and ACK1V909G (0.5 mg); SIAH1, SIAH2, MDM2, RLIM-HA and TRIAD-Myc (0.2 mg). A plasmid encoding green fluorescent protein (green fluorescent protein, 0.05 mg) was cotransfected GST pulldown assays and quantitative real-time PCR to monitor transfection efficiency. Empty vector pcDNA3.1 was used to Details on these methods are provided in the Supplementary Methods obtain equal amounts of transfected DNA (total of 1 mg per 5 Â 105 cells). section accompanying the article. The knockdown of SIAH1 and SIAH2 levels in HEK293T cells was performed by transfecting 40 pmol siRNA targeting SIAH1 and SIAH2 or nontargeting control siRNA (siSIAH1/2, sc-44102; siRNA-A, sc-37007; Santa Cruz CONFLICT OF INTEREST Biotechnology, Heidelberg, Germany) with lipofectamine2000 for 48 h (manufacturer’s protocol; Invitrogen, Darmstadt, Germany). Myc-tagged The authors declare no conflict of interest. ACK1, ACK1V909G and the SIAH2 specific shRNA were expressed in MCF-7 cells by transfection of 2–4 mg plasmid-DNA with Attractene (Qiagen, manufacturer’s protocol). siRNA-mediated knockdown in MCF-7 cells was ACKNOWLEDGEMENTS carried out by nucleofection. 40 pmol siRNA or 2 mg plasmid were We thank C. Kosan, M. Korfei and S. Scheiding for their discussions and their excellent transfected with the Amaxa Nucleofector Kit as recommended by the help. We are grateful to M. Ko¨gl, N. Varin-Blank, Z. Ronai, R. Marschalek, H. Bursen, A. manufacturer (Lonza, Ko¨ln, Germany) using solution V and program E-14. Baniahmad, O. Werz and O. Huber for providing material. Grant support: German Cancer Aid (FKZ102362); Wilhelm-Sander Foundation (No. 2010.078.1). Immunoblotting and immunoprecipitation Cell lysis, SDS–PAGE and immunoblotting were described by us pre- viously.15,17 All the lysates were assessed by Bradford assay for protein REFERENCES concentrations. Tubulin served as loading control for all immunoblots. 1 House CM, Mo¨ller A, Bowtell DD. Siah proteins: novel drug targets in the Ras and For immunoprecipitation of SIAH1 or SIAH2, the harvested cells were hypoxia pathways. Cancer Res 2009; 69: 8835–8838. lysed in NaCl-EDTA-Tris-Nonidet (NETN) buffer.15,17 300 mg of total lysate 2Kra¨mer OH, Stauber RH, Bug G, Hartkamp J, Knauer SK. SIAH proteins: Critical roles protein was added to 1 mg of antibody and 10 ml protein-G sepharose in leukemogenesis. Leukemia 2012 (in press). beads (GE Healthcare, Mu¨nchen, Germany). For the detection of 3 Liu M, Hsu J, Chan C, Li Z, Zhou Q. The ubiquitin ligase SIAH1 controls Ell2 stability phosphorylated ACK1, the buffers were supplemented with 5 mM and formation of super elongation complexes to modulate gene transcription. sodium-fluoride (NaF) and 1 mM sodium-orthovanadate (Na3OV4). Mol Cell 2012; 46: 325–334.

& 2013 Macmillan Publishers Limited Oncogene (2013) 4913 – 4920 Estrogen- and SIAH2-dependent proteasomal ACK1 degradation M Buchwald et al 4920 4 Metzger MB, Hristova VA, Weissman AM. HECT and RING finger families of E3 29 Korzeniewski N, Cuevas R, Duensing A, Duensing S. Daughter centriole elongation ubiquitin ligases at a glance. J Cell Sci 2012; 125: 531–537. is controlled by proteolysis. Mol Biol Cell 2010; 21: 3942–3951. 5 Mogk A, Schmidt R, Bukau B. The N-end rule pathway for regulated proteolysis: 30 Khurana A, Nakayama K, Williams S, Davis RJ, Mustelin T, Ronai Z. Regulation prokaryotic and eukaryotic strategies. Trends Cell Biol 2007; 17: 165–172. of the ring finger E3 ligase Siah2 by p38 MAPK. J Biol Chem 2006; 281: 6 Zhao HL, Ueki N, Hayman MJ. The Ski protein negatively regulates Siah2-mediated 35316–35326. HDAC3 degradation. Biochem Biophys Res Commun 2010; 399: 623–628. 31 Mahajan NP, Whang YE, Mohler JL, Earp HS. Activated tyrosine kinase Ack1 pro- 7 Perissi V, Scafoglio C, Zhang J, Ohgi KA, Rose DW, Glass CK et al. TBL1 and TBLR1 motes prostate tumorigenesis: role of Ack1 in polyubiquitination of tumor sup- phosphorylation on regulated gene promoters overcomes dual CtBP and NCoR/ pressor Wwox. Cancer Res 2005; 65: 10514–10523. SMRT transcriptional repression checkpoints. Mol Cell 2008; 29: 755–766. 32 Mahajan NP, Liu Y, Majumder S, Warren MR, Parker CE, Mohler JL et al. Activated 8 Wu H, Lin Y, Shi Y, Qian W, Tian Z, Yu Y et al. SIAH-1 interacts with mammalian Cdc42-associated kinase Ack1 promotes prostate cancer progression via polyhomeotic homologues HPH2 and affects its stability via the ubiquitin-pro- androgen receptor tyrosine phosphorylation. Proc Natl Acad Sci USA 2007; 104: teasome pathway. Biochem Biophys Res Commun 2010; 397: 391–396. 8438–8443. 9 Frasor J, Danes JM, Funk CC, Katzenellenbogen BS. Estrogen down-regulation of 33 Mahajan K, Coppola D, Challa S, Fang B, Chen YA, Zhu W et al. Ack1 mediated the corepressor N-CoR: mechanism and implications for estrogen derepression of AKT/PKB tyrosine 176 phosphorylation regulates its activation. PLoS One 2010; 5: N-CoR-regulated . Proc Natl Acad Sci USA 2005; 102: 13153–13157. e9646. 10 Venables JP, Dalgliesh C, Paronetto MP, Skitt L, Thornton JK, Saunders PT et al. 34 Mahajan K, Mahajan NP. Shepherding AKT and androgen receptor by Ack1 tyr- SIAH1 targets the alternative splicing factor T-STAR for degradation by the pro- osine kinase. J Cell Physiol 2010; 224: 327–333. teasome. Hum Mol Genet 2004; 13: 1525–1534. 35 House CM, Hancock NC, Mo¨ller A, Cromer BA, Fedorov V, Bowtell DD et al. Elu- 11 Nagano Y, Fukushima T, Okemoto K, Tanaka K, Bowtell DD, Ronai Z et al. Siah1/SIP cidation of the substrate binding site of Siah ubiquitin ligase. Structure 2006; 14: regulates p27(kip1) stability and cell migration under metabolic stress. Cell Cycle 695–701. 2011; 10: 2592–2602. 36 Twomey E, Li Y, Lei J, Sodja C, Ribecco-Lutkiewicz M, Smith B et al. Regulation of 12 Fukushima T, Zapata JM, Singha NC, Thomas M, Kress CL, Krajewska M et al. MYPT1 stability by the E3 ubiquitin ligase SIAH2. Exp Cell Res 2010; 316: 68–77. Critical function for SIP, a ubiquitin E3 ligase component of the beta-catenin 37 Yokoyama N, Miller WT. Biochemical properties of the Cdc42-associated tyrosine degradation pathway, for thymocyte development and G1 checkpoint. Immunity kinase ACK1. Substrate specificity, authphosphorylation, and interaction with Hck. 2006; 24: 29–39. J Biol Chem 2003; 278: 47713–47723. 13 Bursen A, Moritz S, Gaussmann A, Dingermann T, Marschalek R. Interaction of AF4 38 Jansen MP, Ruigrok-Ritstier K, Dorssers LC, van Staveren IL, Look MP, Meijer-van wild-type and AF4.MLL fusion protein with SIAH proteins: indication for t(4;11) Gelder ME et al. Downregulation of SIAH2, an ubiquitin E3 ligase, is associated pathobiology? Oncogene 2004; 23: 6237–6249. with resistance to endocrine therapy in breast cancer. Breast Cancer Res Treat 14 Habelhah H, Frew IJ, Laine A, Janes PW, Relaix F, Sassoon D et al. Stress-induced 2009; 116: 263–271. decrease in TRAF2 stability is mediated by Siah2. EMBO J 2002; 21: 5756–5765. 39 Stebbing J, Filipovic A, Lit LC, Blighe K, Grothey A, Xu Y et al. LMTK3 is implicated 15 Buchwald M, Pietschmann K, Mu¨ller JP, Bo¨hmer FD, Heinzel T, Kra¨mer OH. Ubi- in endocrine resistance via multiple signaling pathways. Oncogene 2012 (in press). quitin conjugase UBCH8 targets active FMS-like tyrosine kinase 3 for proteasomal 40 Kra¨mer OH, Zhu P, Ostendorff HP, Golebiewski M, Tiefenbach J, Peters MA et al. degradation. Leukemia 2010; 24: 1412–1421. The histone deacetylase inhibitor valproic acid selectively induces proteasomal 16 Kra¨mer OH, Mu¨ller S, Buchwald M, Reichardt S, Heinzel T. Mechanism for ubi- degradation of HDAC2. Embo J 2003; 22: 3411–3420. quitylation of the leukemia fusion proteins AML1-ETO and PML-RARalpha. Faseb J 41 Lu M, Mira-y-Lopez R, Nakajo S, Nakaya K, Jing Y. Expression of estrogen receptor 2008; 22: 1369–1379. alpha, retinoic acid receptor alpha and cellular retinoic acid binding protein II 17 Pietschmann K, Buchwald M, Mu¨ller S, Knauer SK, Ko¨gl M, Heinzel T et al. Dif- genes is coordinately regulated in human breast cancer cells. Oncogene 2005; 24: ferential regulation of PML-RARalpha stability by the ubiquitin ligases SIAH1/ 4362–4369. SIAH2 and TRIAD1. Int J Biochem Cell Biol 2012; 44: 132–138. 42 Li C, Lin M, Liu J. Identification of PRC1 as the p53 target gene uncovers a novel 18 Pietschmann K, Bolck HA, Buchwald M, Spielberg S, Polzer H, Spiekermann K et al. function of p53 in the regulation of cytokinesis. Oncogene 2004; 23: 9336–9347. Breakdown of the FLT3-ITD/STAT5 axis and synergistic apoptosis induction by the 43 Holliday DL, Speirs V. Choosing the right cell line for breast cancer research. Breast histone deacetylase inhibitor Panobinostat and FLT3-specific inhibitors. Mol Cancer Res 2011; 13: 215. Cancer Ther 2012 (in press). 44 Prieto-Echague V, Miller WT. Regulation of ack-family nonreceptor tyrosine kina- 19 Winter M, Sombroek D, Dauth I, Moehlenbrink J, Scheuermann K, Crone J et al. ses. J Signal Transduct 2011; 2011: 742372. Control of HIPK2 stability by ubiquitin ligase Siah-1 and checkpoint kinases ATM 45 Pao-Chun L, Chan PM, Chan W, Manser E. Cytoplasmic ACK1 interaction with and ATR. Nat Cell Biol 2008; 10: 812–824. multiple receptor tyrosine kinases is mediated by Grb2: an analysis of ACK1 20 Calzado MA, de la Vega L, Mo¨ller A, Bowtell DD, Schmitz ML. An inducible effects on Axl signaling. J Biol Chem 2009; 284: 34954–34963. autoregulatory loop between HIPK2 and Siah2 at the apex of the hypoxic 46 van der Horst EH, Degenhardt YY, Strelow A, Slavin A, Chinn L, Orf J et al. response. Nat Cell Biol 2009; 11: 85–91. Metastatic properties and genomic amplification of the tyrosine kinase gene 21 Yun S, Mo¨ller A, Chae SK, Hong WP, Bae YJ, Bowtell DD et al. Siah proteins induce ACK1. Proc Natl Acad Sci USA 2005; 102: 15901–15906. the epidermal growth factor-dependent degradation of phospholipase Cepsilon. 47 Mahajan K, Challa S, Coppola D, Lawrence H, Luo Y, Gevariya H et al. Effect of Ack1 J Biol Chem 2008; 283: 1034–1042. tyrosine kinase inhibitor on ligand-independent androgen receptor activity. 22 Wen YY, Yang ZQ, Song M, Li BL, Yao XH, Chen XL et al. The expression of SIAH1 is Prostate 2010; 70: 1274–1285. downregulated and associated with Bim and apoptosis in human breast cancer 48 Chua BT, Lim SJ, Tham SC, Poh WJ, Ullrich A. Somatic mutation in the ACK1 tissues and cells. Mol Carcinog 2010; 49: 440–449. ubiquitin association domain enhances oncogenic signaling through EGFR reg- 23 Wen YY, Yang ZQ, Song M, Li BL, Zhu JJ, Wang EH. SIAH1 induced apoptosis by ulation in renal cancer derived cells. Mol Oncol 2010; 4: 323–334. activation of the JNK pathway and inhibited invasion by inactivation of the ERK 49 Howlin J, Rosenkvist J, Andersson T. TNK2 preserves epidermal growth factor pathway in breast cancer cells. Cancer Sci 2010; 101: 73–79. receptor expression on the cell surface and enhances migration and invasion of 24 Nadeau RJ, Toher JL, Yang X, Kovalenko D, Friesel R. Regulation of Sprouty2 human breast cancer cells. Breast Cancer Res 2008; 10: R36. stability by mammalian Seven-in-Absentia homolog 2. J Cell Biochem 2007; 100: 50 Mahajan K, Coppola D, Chen YA, Zhu W, Lawrence HR, Lawrence NJ et al. Ack1 151–160. tyrosine kinase activation correlates with pancreatic cancer progression. Am J 25 Depaux A, Regnier-Ricard F, Germani A, Varin-Blank N. Dimerization of hSiah Pathol 2012; 180: 1386–1393. proteins regulates their stability. Biochem Biophys Res Commun 2006; 348: 51 Chan W, Tian R, Lee YF, Sit ST, Lim L, Manser E. Down-regulation of active ACK1 is 857–863. mediated by association with the E3 ubiquitin ligase Nedd4-2. J Biol Chem 2009; 26 Xu Z, Sproul A, Wang W, Kukekov N, Greene LA. Siah1 interacts with the scaffold 284: 8185–8194. protein POSH to promote JNK activation and apoptosis. J Biol Chem 2006; 281: 52 Osborne CK, Neven P, Dirix LY, Mackey JR, Robert J, Underhill C et al. Gefitinib or 303–312. placebo in combination with tamoxifen in patients with hormone receptor- 27 Ahmed AU, Schmidt RL, Park CH, Reed NR, Hesse SE, Thomas CF et al. Effect of positive metastatic breast cancer: a randomized phase II study. Clin cancer res disrupting seven-in-absentia homolog 2 function on lung cancer cell growth. 2011; 17: 1147–1159. J Natl Cancer Inst 2008; 100: 1606–1629. 53 Sarkar TR, Sharan S, Wang J, Pawar SA, Cantwell CA, Johnson PF et al. Identifi- 28 Lin Q, Wang J, Childress C, Sudol M, Carey DJ, Yang W. HECTE3 ubiquitin ligase cation of a Src tyrosine kinase/SIAH2 E3 ubiquitin ligase pathway that regulates Nedd4-1 ubiquitinates ACK and regulates epidermal growth factor (EGF)-induced C/EBPdelta expression and contributes to transformation of breast tumor cells. degradation of EGF receptor and ACK. Mol Cell Biol 2010; 30: 1541–1554. Mol Cell Biol 2012; 32: 320–332.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene (2013) 4913 – 4920 & 2013 Macmillan Publishers Limited