Cells were grown on glass coverslips and transfected using Fugene 6 (Roche). Cells were fixed after 24 h with 4% paraformaldehyde for 15 min at room temperature (all of the following steps were done at room temperature) and permeabilized in 0.2% Triton X-100-phosphate-buffered saline for 5 min. Cells were washed with phosphate-buffered saline and blocked with 5% goat serum for 1 h, incubated with primary antibodies for 1 h, washed, and incubated with secondary antibodies for 30 min. Following extensive washing, the cells were mounted and immunostaining was visualized with a Leica DM5500 B microscope.

A schematic representation of all of the constructs used in this study can be found in Fig. S3 in the supplemental material and also in the relevant figures. Mammalian cell lines were cultured in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum and antibiotics. Drosophila Schneider cells were cultured at 25°C in Schneider cell medium (Sigma) supplemented with 10% fetal bovine serum. Mammalian cells were transfected with Fugene 6 (Roche) and insect cells with Effectene reagent (Qiagen), according to the manufacturers' protocol. In general, 1 μg of each construct was transfected into cells seeded at 40 to 60% confluence in 6-cm tissue culture dishes, and cells were lysed 24 to 48 h posttransfection. TelHA,TelFlag, YanHA, and YanFlag constructs were fused in frame with either a hemagglutinin (HA) or a Flag epitope tag and cloned into either the pCS2 expression vector for expression in mammalian cells or the pMT vector (Invitrogen) for expression in insect cells. Glutathione S-transferase (GST)-Tel was cloned in frame with GST in the pGEX-2TK vector. Mutants were generated by PCR. For immunoprecipitations, cells were lysed in 1 ml of either radioimmunoprecipitation assay-sodium dodecyl sulfate buffer (1 mM EDTA, 50 mM Tris [pH 8.0], 150 mM NaCl, 10% glycerol, 1% Triton X-100) or protein lysis buffer (50 mM Tris [pH 7.5], 1% NP-40, 0.1% sodium dodecyl sulfate, 0.5% Na deoxycholate, 150 mM NaCl) with protease inhibitors (phenylmethylsulfonyl fluoride, trypsin, pepstatin A, leucine, aprotinin) and NaF. Cell lysates were syringed and centrifuged. Immunoprecipitations were carried out with 0.75 μl of either anti-HA polyclonal antibody (Abcam) or anti-V5 monoclonal antibody (Invitrogen). Lysates were preincubated with the antibody for 1 h, after which suitable beads, protein A-Sepharose 4 fast flow (Amersham Pharmacia Biotech) for HA immunoprecipitations or protein G-Sepharose (Sigma-Aldrich) for V5 immunoprecipitations, were added for a further 2 h of incubation. Associated proteins were analyzed by Western blotting. For pulldown assays, Fbl6 proteins were labeled with [³⁵S]methionine by using the TNT coupled reticulocyte in vitro translation system (Promega) and then incubated with GST-Tel fusions that were immobilized on glutathione-Sepharose beads as previously described (2).

Cells were washed free of medium and seeded into 6-cm tissue culture dishes (Gibco) in methionine- and serum-free Dulbecco modified Eagle medium (Gibco). Cells were routinely incubated for 3 h, and then the medium was supplemented with 50 μCi ³⁵S-labeled methionine for an additional 3 h of labeling. Labeled HA epitope-tagged Tel proteins were immunoprecipitated from the cell lysates as described above.

All methods were performed essentially as described previously (37). The following primer sets were deployed for real-time PCR analysis: Telq15 (CCCTGCGCCACTACTACAA) and Telq13 (TGATTTCATCTGGGGTTTTCA), Telq25 (CTTTCGCTATCGATCTCCTCA) and Telq23 (AGGGTGGAAGAATGGTGAAA), hFbl615 (ATGCCCAATCGGTTTTCA) and hFbl613 (AGGACAGCACTCACCTACCAG), hFbl625 (GGATCTTCGTGGCTGTGC) and hFbl623 (CATACAGGCCCAGATGAAGC), hSKP1A15 (GGAGATTCGCAAGACCTTCA) and hSKP1A13 (CACTGGTTCTCTTTGCGTACC), and Gapdh5 (TGCCATGTAGACCCCTTGAAG) and Gapdh3 (ATGGTACATGACAAGGTGCGG).

P{GawB}NP0326, Df(2R)stan1, and Df(2R)Exel6060 fly stocks were obtained from the Bloomington Stock Center. Stocks were maintained using balancers expressing the green fluorescent protein. Reverted flies were generated by transposase-mediated precise excision of the P element in P{GawB}NP0326 flies. Precise excision was confirmed by the sequencing of fly genomic DNA. Stocks were maintained under standard conditions, and crosses were performed by following standard procedures. For expression analyses, embryos were carefully staged, and total RNA was prepared using the SV total RNA isolation kit in accordance with the manufacturer's protocol (Promega). Following cDNA synthesis using a TaqMan kit (Roche), expression levels of relevant transcripts were determined by PCR using the following primer sets: the fbl6 sense primer (AGGTCCTGCGTGTGACAAACTCTC) and the fbl6 antisense primer (GAGGCAATGAGTTGGATGTGAAC); the argos sense primer (GGATTTTCTGTTTGATCAGTTC) and the argos antisense primer (ATTATTGGATATTTCATTCACT); the CG18335 sense primer, (ATAACGCCGGAACCGCACTTGGTG) and the CG18335 antisense primer (GTAAACGGAGTCCACCAAACCCTC); the CG13220 sense primer (ACCTGACGAACGTGTTCCTCATCTC) and the CG13220 antisense primer (GGTTCGAATACCTCAGTCCTATATC); the tubulin sense primer (CGTTCACATCCAAGCTGGTCAG) and the tubulin antisense primer (GGGTGCGGAAGCAGATATCG).

Full-length yan or tel cDNA was cloned into pAS2-1 (Clontech) in frame with the GAL4 DNA binding domain. Independent transformants (approximately 5 × 10⁵) from a GAL4 activation domain Drosophila embryo library (for Yan) or a human bone marrow library (for Tel) were screened in Saccharomyces cerevisiae Hf7c cells according to the manufacturer's protocol (Clontech). We confirmed the specificities of interactions between Yan and Drosophila Fbl6 and between Tel and human Fbl6 by transforming yeast with the cDNA of fbl6, in either the presence or the absence of the yan bait (for Drosophila fbl6), the tel bait (for human fbl6), or an unrelated bait. Full-length cDNAs of the human fbl6 gene and of the Drosophila fbl6 gene were isolated by reverse transcriptase (RT)-PCR using total RNA derived from either insect Schneider cells (for Drosophila fbl6) or U2OS human osteosarcoma cells (for human fbl6).

We engineered a V5 epitope-tagged version of wild-type human Fbl6 and confirmed the biochemical interaction between Tel and Fbl6 in tissue culture cells (Fig. 1B and C). Furthermore, we found that this interaction requires the SAM domain of Tel (Fig. (Fig.1B)1B) and the putative LRDs (LRD1 and LRD2) of Fbl6 (Fig. (Fig.1C).1C). Interestingly, disruption of the F-box domain of Fbl6 also impairs binding to Tel (see Fig. Fig.1C).1C). In cells, both Tel and Fbl6 colocalize in the nucleoplasm, and Fbl6 is also found in the nucleolus (Fig. (Fig.1D1D and see Fig. Fig.2C).2C). Recent reports have highlighted differential regulation of protein function by different isoforms of the same F-box protein that are localized in either the nucleolus or the nucleus (31). It would be interesting as a future study to determine whether nuclear and nucleolar Fbl6 proteins play different roles in protein regulation.

It is firmly established that F-box-containing proteins couple their substrates to the conserved ubiquitinating machinery that targets the protein for proteasome-dependent degradation (14, 18, 48). Thus, we investigated whether Tel is ubiquitinated and whether the ubiquitination of Tel might be regulated by Fbl6. Figure Figure2A2A shows that Tel is ubiquitinated in tissue culture cells and that ubiquitinated Tel is targeted for degradation by the proteasome because the incubation of cells with inhibitors of the proteasome stabilizes the fraction of ubiquitin-conjugated Tel. In support of these data, Fig. Fig.2A2A also shows that Tel that is covalently conjugated by a version of ubiquitin that lacks all lysines (and therefore fails to form chains of ubiquitin that are recognized by the proteasome) is not stabilized following inhibition of the proteasome with MG132. Figure Figure2B2B shows that both the SAM domain and the Ets DNA binding domain (EDBD) are indispensable for Tel ubiquitination. Tel with EDBD deleted (TelΔEDBD) can interact with Fbl6 both in cells (Fig. (Fig.1B)1B) and in vitro (see Fig. Fig.3A);3A); however, we have shown previously (37) (Fig. (Fig.2C)2C) that the deletion of the EDBD leads to the mislocalization of Tel from the nucleus to the cytoplasm. Therefore, to determine that the failure of TelΔEDBD to be ubiquitinated in cells does not result simply from subcellular mislocalization, we engineered a Tel mutant, TelRR-AA. TelRR-AA contains two single point mutations of arginine residues in the EDBD that are known to make contact with DNA (25) and are sufficient to promote Tel mislocalization from the nucleus to the cytoplasm (Fig. (Fig.2C).2C). However, unlike TelΔEDBD, TelRR-AA is efficiently ubiquitinated (Fig. (Fig.2B),2B), suggesting that the failure of TelΔEDBD to be ubiquitinated does not result simply from its mislocalization from the nucleus. Some possibilities are that lysine residues in the EDBD are targeted for ubiquitin conjugation and/or that the EDBD mediates a protein interaction that is essential for Tel ubiquitination. Indeed, the EDBDs of all known Ets proteins harbor four highly conserved lysine residues (three of which make contact with DNA) (25) that may be substrates for ubiquitin conjugation. Precise definition, by mutagenesis, of the exact site of ubiquitination is hampered by redundancy (M. Alloul-Ramdhani, M. Op Den Brouw, and D. A. Baker, unpublished data); however, improved antibodies and mass spectrometry techniques should, in time, resolve this issue. Collectively, these results show that Tel is ubiquitinated and that ubiquitination requires both the Tel SAM domain and the EDBD.

We next explored the role of Fbl6 in Tel ubiquitination. Several lines of evidence support a role for Fbl6 in promoting Tel downregulation and ubiquitination. First, deletion of the SAM domain, which is essential for Fbl6 binding to Tel (Fig. (Fig.1B),1B), abolished Tel ubiquitination (Fig. (Fig.2B).2B). Second, ectopic expression of wild-type Fbl6 stimulated Tel ubiquitination, whereas Fbl6 harboring mutations that disrupt the interaction with Tel (Fig. (Fig.1C)1C) failed to augment Tel ubiquitination (Fig. (Fig.2D).2D). Furthermore, in contrast to wild-type Tel, Fbl6 failed to stimulate ubiquitination of either Tel with the SAM domain deleted (TelΔSAM), which is unable to bind Fbl6, or TelΔEDBD. Third, shRNA-mediated knockdown of Fbl6 both stabilized steady-state endogenous Tel protein levels (Fig. (Fig.2E)2E) and inhibited ubiquitin conjugation of Tel (Fig. (Fig.2F).2F). If ubiquitination of Tel is indeed F-box dependent, one should expect Tel to associate with SKP1, which is a core component of SCF ubiquitin ligases, and one should also expect SKP1 to play a role in regulating steady-state Tel protein levels. Figure Figure2E2E shows that endogenous Tel and endogenous SKP1 were copurified as a complex from tissue culture cells, and it also shows that, similarly to lowering fbl6 levels, the shRNA-mediated knockdown of SKP1 stabilized steady-state endogenous Tel protein levels. In cells, antibodies specific for endogenous Tel detect two differently migrating protein species (29, 43), and the higher-molecular-weight form is likely to represent phosphorylated Tel (29). Ectopic expression of Tel also produces two differently migrating proteins, and the higher-molecular-weight Tel form is likely to be phosphorylated Tel because incubation with phosphatase leads to a loss of this band and a concomitant increase in the amount of the faster-migrating, unmodified Tel species (Fig. (Fig.2G,2G, upper panel). Figure Figure2G2G (lower panel) shows that wild-type Fbl6, but not Fbl6 harboring mutations in either the F-box domain or the leucine-rich regions, strongly promotes the degradation of phosphorylated Tel. This suggests that in cells, phosphorylation of Tel might further sensitize it to Fbl6-mediated downregulation, perhaps by enhancing the affinity of Fbl6 for binding to Tel. Together, these data show that Tel is ubiquitinated and that Fbl6 stimulates Tel downregulation and ubiquitination.

Numerous studies have established that the SAM domain is essential for the monomers of Tel to form homotypic oligomers (17, 33, 37). Moreover, we showed earlier that the SAM domain is also indispensable both for the binding of Fbl6 and for Tel ubiquitination (Fig. (Fig.1B1B and and2B).2B). In light of these findings, we compared the ubiquitination of Tel oligomers and that of Tel monomers. First, we established that Fbl6 can associate with both Tel oligomers and Tel monomers. Figure Figure3A3A shows that both wild-type Tel and a Tel mutant that is unable to self-associate (TelA*) (37) efficiently interacted with Fbl6 in vitro, whereas TelΔSAM, which is also monomeric, failed to bind Fbl6. In contrast, in cells, wild-type Tel efficiently coimmunoprecipitated with Fbl6, but TelA*, like TelΔSAM, did not (Fig. (Fig.3B).3B). However, the interaction between TelA* and Fbl6, but not that between TelΔSAM and Fbl6, was stabilized in the presence of proteasome inhibitors (Fig. (Fig.3B),3B), and furthermore, TelA* was far more readily ubiquitinated than the wild-type Tel (Fig. (Fig.3C),3C), suggesting that Tel monomers are more labile than Tel oligomers. We recently reported a similar phenomenon with regard to the sumoylation of Tel (37). In cells, Tel is efficiently sumoylated on lysine 11, whereas TelA* is detectably sumoylated on lysine 11 in cells only in the presence of the proteasome inhibitor MG132. Figure Figure3C3C shows that the ubiquitination of TelA*, like wild-type Tel ubiquitination, was also stimulated by Fbl6, in contrast to the ubiquitination of TelΔSAM, which is unable to bind Fbl6. Collectively, these findings revealed that Tel monomers can associate with Fbl6 and are efficiently ubiquitinated.

The SAM domain defines a subfamily of Ets transcription factors, including Tel, Tel2, Ets1, and Ets2 from vertebrates as well as Yan and Pointed P2 from Drosophila (23, 34). Since Fbl6 associates with Tel via its SAM domain, we tested whether Fbl6 also interacts with other SAM domain-containing Ets family members. Figure Figure3F3F shows that Fbl6 efficiently associated with both Tel and Tel2 but did not copurify detectably with Ets1 or Ets2, both of which express SAM domains, or PU.1, which does not contain a SAM domain. Tel2 strongly resembles Tel, both structurally and functionally (12, 30, 32), and may have resulted from an earlier gene duplication of the ancestral Tel gene. We found that, like Tel, Tel2 is ubiquitinated and that Fbl6 augments Tel2 ubiquitination (Fig. (Fig.3F),3F), suggesting that Tel and Tel2 may share a common F-box-dependent degradation pathway. In the future, it will be of interest to determine whether Fbl6 plays a role in regulating the activities of other factors in vivo.

Having established a biochemical and functional interaction between human Tel and Fbl6, we next explored whether this mechanism might be evolutionarily conserved. To that end, we isolated a full-length Drosophila fbl6 (also referred to as CG13213) cDNA from insect Schneider cells. Figure 4A and B show that, like the binding of human Fbl6 to Tel (Fig. 1B and C), Drosophila Fbl6 efficiently associated with Yan, and this interaction depended on the Yan SAM domain (Fig. (Fig.4A)4A) and the LRD2 domain of Fbl6 (Fig. (Fig.4B).4B). In contrast to Tel-Fbl6 binding, Yan-Fbl6 binding appeared to not require the LRD1 sequence (Fig. (Fig.4B4B).

Next, we determined whether Fbl6 influences Yan protein stability. To that end, we established a number of stable Schneider S2 cell lines and monitored their endogenous Yan protein levels. We and others have established that in Drosophila, a protein named Mae (modulator of the activity of Ets) orchestrates Yan derepression by binding to Yan, thereby disrupting Yan self-association and binding to DNA and sensitizing it to downregulation that is dependent on mitogen-activated protein kinase (2, 33, 40, 41). Figure Figure4C4C shows that ectopic expression of a constitutively active form of Ras alone in Drosophila S2 cells had little detectable effect on steady-state endogenous Yan protein levels. In contrast, ectopic Mae expression alone strongly promoted endogenous Yan downregulation, and this effect was enhanced by activated Ras. Under these conditions, incubation of the cells with MG132 failed to fully rescue Yan protein levels (Roukens and Baker, unpublished), suggesting that Ras/Mae might destabilize Yan by proteasome-independent and proteasome-dependent pathways. However, we found that double-stranded-RNA-mediated knockdown of Fbl6 and SKPA (suppressor of kinetochore protein A), which is a core component of the ubiquitinating machinery (1, 26, 39, 47, 48), inhibited Mae/Ras-dependent Yan downregulation (Fig. (Fig.4C).4C). These results suggest that in cells, SKP and Fbl6 play a role in facilitating Yan degradation. To ascertain whether Yan can be ubiquitinated and whether Fbl6 augments Yan ubiquitination, we employed the baculovirus system for Yan expression in insect Sf9 cells (6). Figure Figure4D4D shows that ubiquitin was efficiently and covalently conjugated to Yan and that the coexpression of Fbl6 strongly stimulated this ubiquitination. Together, these data suggest that Fbl6 promotes Yan downregulation and stimulates Yan ubiquitination.

We found that the SAM domain of both Tel (Fig. (Fig.1B)1B) and Yan (Fig. (Fig.4A)4A) is required for the binding of Fbl6. The SAM domain defines a subfamily of Ets transcription factors, including Tel, Tel2, Ets1, and Ets2 from vertebrates as well as Yan and Pointed P2 from Drosophila (23, 33), raising the possibility that at least some of these factors are also targeted by Fbl6 or a related F-box protein. Indeed, we found that human Fbl6 also associates with the Tel-related Tel2 protein and promotes its ubiquitination (Fig. (Fig.3F).3F). Furthermore, we found biochemical interactions between Drosophila Fbl6 and the SAM domain-containing proteins Pointed P2 (3, 27) (but not P1, which lacks the SAM domain) and Mae (2, 33, 41) in tissue culture cells (Roukens and Baker, unpublished). However, we were unable to detect associations between Fbl6 and either Ets1 or Ets2, each of which harbors a SAM domain, suggesting that there is specificity, perhaps structural, that determines these interactions. Many other families of proteins express SAM domains, such as the polycomb group proteins and components of the mitogen-activated protein kinase cascade (34). It should be interesting to investigate whether Fbl6 (or related F-box proteins) plays a role in regulating their activity.

Our data support the idea that Tel/Yan monomers are more susceptible to ubiquitination and degradation than the oligomeric forms of these proteins are. Apparently, and paradoxically, the biochemical interactions between Tel/Yan monomers and Fbl6 were found to be far weaker than the interactions between Tel/Yan oligomers and Fbl6 (Fig. (Fig.3B3B and and4A).4A). However, these interactions were strongly stabilized in the presence of proteasome inhibitors (Fig. (Fig.3B),3B), suggesting that the failure to find strong associations between these proteins resulted from the instability of the Tel/Yan monomer and of the Fbl6 complex. Consistent with this, we found that Tel monomers were more readily ubiquitinated than wild-type Tel (Fig. (Fig.3C).3C). It is worth noting that, in general, wild-type Tel is detected in cells as a phosphorylated or nonphosphorylated protein species and that the phosphorylation of Tel appears to further sensitize it to Fbl6-mediated degradation (Fig. (Fig.2G).2G). In contrast, phosphorylated forms of monomeric Tel are evidently especially unstable and are detected only in the presence of proteasome inhibitors (M. G. Roukens, M. Alloul-Ramdhani, and D. A. Baker, unpublished data). It has previously been described that Tel is negatively regulated by extracellular signal-regulated kinase-induced phosphorylation of serine residues 213 and 257 (24). Phosphorylation of these sites could be a trigger for promoting Tel ubiquitination, although in our ubiquitination assays, we were unable to detect an obvious effect of mutating these residues or an additional putative mitogen-activated protein kinase target, serine residue 22 in the N terminus of Tel (M. Alloul-Ramdhani and D. A. Baker, data not shown). Our previous work also supports the contention that Tel/Yan monomers are relatively labile (37). Monomeric Tel was detectably and very efficiently sumoylated in cells only in the presence of proteasome inhibitors, suggesting that this species of Tel is normally very unstable. The current model of Tel function holds that Tel monomers directly associate via their conserved SAM domains and that the resulting DNA-bound oligomers oppose the transcription-activating apparatus (reviewed in references 34, 42, and 44). Since Fbl6 also interacts with the Tel SAM domain (Fig. (Fig.11 and and4),4), perhaps self-associated Tel oligomers are relatively more resistant to F-box-mediated degradation than are Tel monomers because of reduced accessibility to exposed SAM domains. In the future, it should be of considerable interest to elucidate how Tel sumoylation and ubiquitination are precisely integrated to regulate Tel activity.