HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Drewett, V. Articles by Shaw, P. E. Search for Related Content PubMed PubMed Citation Articles by Drewett, V. Articles by Shaw, P. E. Social Bookmarking                      What's this?

J Biol Chem, Vol. 275, Issue 3, 1757-1762, January 21, 2000

Dimer Formation by Ternary Complex Factor ELK-1^*

Victoria Drewett, Silke Muller, Jane Goodall, and Peter E. Shaw^§

From the School of Biomedical Sciences and Institute of Cell Signalling, University of Nottingham Medical School, Queen's Medical Centre, Nottingham NG7 2UH, United Kingdom

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Ternary complex factors (TCFs), a subgroup of the ets protein family, bind with a dimer of serum response factor to the c-fos serum response element. Both DNA binding and transcriptional activation by TCFs are regulated by mitogen-activated protein kinases. When activated, mitogen-activated protein kinases form homodimers that translocate to the nucleus, where they interact with TCFs via specific docking sites. Here we show by three different criteria that Elk-1 is capable of forming dimers in eukaryotic cells through two distinct interaction domains. These observations are consistent with a dynamic model of TCF-promoter interactions.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Ternary complex factors (TCFs)¹ were first identified through their interactions with serum response factor (SRF) at the c-fos serum response element (SRE) (1). To date three TCFs, Elk-1, Sap1a, and Net/Sap2, have been identified (2, 3). Together they form a subgroup of the ets protein family, members of which have in common a winged helix-loop-helix (HLH) DNA binding domain (4). In TCFs the ets domain is amino-terminal and linked by a flexible spacer region (5) to the SRF interaction domain (6). Although these structures serve to direct TCFs to specific promoter elements, where they bind as ternary and quaternary complexes with SRF (7, 8), other conserved domains toward the carboxyl terminus mediate the transactivation functions of TCFs (9-11). Exactly how the latter operate has yet to be elucidated.

Both the DNA binding and transactivation functions of TCFs are regulated by mitogen-activated protein kinases (MAPKs) (12, 13). For example, Elk-1 possesses nine consensus sites for extracellular signal-regulated kinases (ERKs), most of which are phosphorylated in vivo when cells are stimulated with mitogens (9, 10, 14, 15). Some of the same sites are also recognized by stress-activated protein kinases/c-Jun-N-terminal kinases (SAPKs/JNKs) and isoforms of p38^mapk (16-19). MAPKs have been reported to interact with TCFs in vitro and in vivo (17, 20, 21), and recently two docking sites for MAPKs, one of them similar to the binding site for SAPK/JNKs in c-Jun ( domain), were identified in TCFs (22, 23). They appear to be essential for activation by ERKs and SAPK/JNKs but not p38^mapk. MAPKs also form cytoplasmic complexes either directly or through scaffold proteins with their activating kinases (24-26). Activation of ERKs by MAP/ERK kinases (MEKs) induces ERK dimerization, which is a prerequisite for translocation of ERKs into the cell nucleus (27). This, in turn, is essential for ERK2-induced neurite outgrowth and cell transformation (28). Thus the active nuclear form of ERKs is a dimer.

Studies of complexes between TCFs, SRF, and the c-fos SRE suggest that TCF interactions are dynamic (7). Consistent with this idea, it has been proposed that TCFs can be activated selectively (29) and that selective activation leads to changes in SRE occupation by TCFs.² This implies that TCFs are not necessarily bound to DNA when targeted by MAPKs. Unbound inactive TCFs may exist as dimers, which would serve as appropriate substrates for active ERK dimers. This idea is further supported by reports that related ets proteins can form homo- and heterodimers mediated by their ets domains (30). Interactions between Elk-1 carboxyl-terminal domains have also been reported (8). Here we have assessed the ability of Elk-1 to form oligomers and demonstrate that a fraction of Elk-1 can exist in cells as homotypic dimers. These results have implications for our understanding of the mode of action of TCFs.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Plasmid Constructs-- The construction of pQE-Elk-1 and pCMV5-Elk-1(his) has been described elsewhere (15). pCMV5L#1, a generic vector for expressing proteins with an amino-terminal hemagglutinin (HA) epitope, was constructed by inserting the following oligonucleotide pair into the EcoRI site of pCMV5 (31):

5'-AATTAAAGAGGAGAAATTACATATGGCTTACCCATACGATGTTCCAGATTACGCT (upper) and 5'-AATTCAGCGTAATCTGGAACATCGTATGGGTAAGCCATATGTAATTTCTCCTCTTT (lower).

pCMV5NL#1, a generic vector for expressing proteins with an amino-terminal nuclear localization signal and HA epitope, was constructed by inserting the EcoRI-BamHI fragment from pRS-POU-Elk³ into the EcoRI and BglII sites of pCMV5.

pCMV5-HA-Elk-1 was constructed by ligation of the BamHI-XbaI fragment from pBS-Elk-1 (15) into the BglII-XbaI sites of pCMV5L#1.

pCMV5-HA-Elk-1ets was constructed by inserting the AccI-XbaI fragment of pBS-Elk-1, encoding Elk-1 residues 83-428, into the HindIII and XbaI sites of pCMV5NL#1.

pCMV5-Elk-1ets(his) was generated by ligating the EcoRI-XbaI fragment from pSG-Gal-Elk (11) into pCMV5 digested with the same enzymes. A nuclear localization signal was inserted by subsequent EcoRI digestion and insertion of the following oligonucleotide pair:

5'-AATTAAAGAGGAGAAATTACATATGCCCAAGAAGAAGCGGAAGGTAG (upper) and 5'-AATTCGACCTTCCGCTTCTTCTTGGGCATATGTAATTTCTCCTCTTT (lower).

pCMV5-HA-ElkApa was derived by deletion of the ApaI fragment from pCMV5-HA-Elk-1. pCMV5-Elk-119(his) was constructed by ligation of the BglII-HindIII fragment from pBO-ElkC19 (8) into pCMV5-Elk-1(his) digested with BglII-HindIII.

pCMV5-Elk-1B(his) was constructed by amplifying sequences encoding Elk-1 amino acids 166-306 by polymerase chain reaction from pCMV5-Elk-1(his) and inserting them as a BsgIBglII fragment into pCMV5-Elk-1(his) cut with BsgI and BglII.

pCMV5-Elk-1D(his) was constructed by amplifying sequences encoding Elk-1 amino acids 322-428 by polymerase chain reaction from pQE-Elk-1 and inserting them as a BamHI-HindIII fragment into pCMV5-Elk-1(his) cut with BglII and HindIII.

The construction of pCMV5-Sap1a(his) has been described elsewhere (29). All recombinant plasmid constructs were confirmed by dideoxy sequencing. Further details are available upon request.

Cell Culture and Extract Preparation-- Human embryonal kidney (HEK 293) and COS cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum. Cells were transfected with 4-6 µg of the various Elk-1 constructs by the standard DNA-calcium phosphate coprecipitation procedure. Eighteen h after transfection, cells were washed and serum-starved (0.2%) for 24 h. Cells were then harvested after prior stimulation where appropriate with EGF (75 nM) for 20 min in 20 mM Tris/HCl, pH 7.8, containing 50 µM NaF, 40 µM Na₄P₂O₇, 5 µM MgCl₂, 10 µM Na₃VO₄, 1% Triton X-100, 0.1% SDS, 40 µg/ml leupeptin, 3 µM phenylmethylsulfonyl fluoride, 5 mM benzamidine, 1 µM aprotinin, and 1 µg/ml pepstatin. After incubation on ice for 20 min, cell debris was pelleted by centrifugation, and the clear lysate was removed.

Cross-linking of Elk-1 in the Cell Lysates-- HEK 293 cells were transfected with expression constructs for His-tagged Elk-1 and HA-Elk-1 deletions. Cell lysates were prepared as described above, except that 20 mM HEPES, pH 7.9, was used instead of 20 mM Tris/HCl and incubated with various concentrations of disuccinimidyl suberate (DSS; Pierce) for 90 min on ice. Samples were then analyzed by Western blotting with the anti-HA antibody (3F10) (Roche Molecular Biochemicals) or an anti-Elk-1 polyclonal antibody.

Gel Filtration-- Bacterially expressed Elk-1 was purified as described (15), except that the elution buffer consisted of 75 mM borate, pH 8.0, containing 25 mM EDTA and 10% glycerol. The purified protein was then concentrated to 0.5-1 mg/ml (Flowgen concentrators) and applied (0.2 ml) at 0.3 ml/min to a Superose-12 HR 10/30 gel filtration column (Amersham Pharmacia Biotech) pre-equilibrated with 75 mM borate buffer, pH 8.0, and 10% glycerol. Fractions were collected and analyzed by Western blotting with an anti-Elk-1 antibody.

COS cells were transfected with expression constructs for His-tagged Elk-1 or Elk-1 deletions, and cell lysates were prepared as described above. Lysates were concentrated to 3-4 mg/ml (Flowgen concentrators) and subjected to gel chromatography as described. The His-tagged proteins were concentrated by incubation of the eluates (0.5 ml) with nickel-agarose beads (Qiagen) and detected by Western blotting with the anti-His antibody (Qiagen).

Calibration of the column was carried out with the following protein standards: YADH (Sigma), bovine serum albumin, ovalbumin, and chymotrypsinogen A (Amersham Pharmacia Biotech). Apparent molecular masses for Elk-1 complexes were estimated by comparing elution peaks from each run with the corresponding calibration curve.

Protein Interaction Assays-- Cell lysates were incubated for 2 h at 4 °C with nickel-agarose beads (Qiagen) pre-equilibrated with lysis buffer containing 10 mM imidazole. The beads were then washed three times with lysis buffer containing 20 mM imidazole, and complexes were eluted by boiling for 5 min in SDS sample buffer before SDS-polyacrylamide gel electrophoresis. Proteins were transferred onto Hybond P nitrocellulose (Amersham Pharmacia Biotech) and probed with the anti-His (Qiagen), anti-HA (Roche Molecular Biochemicals), anti-Elk-1, or anti-phospho-Elk-1 (New England Biolabs) antibodies and visualized by standard ECL procedures (Amersham Pharmacia Biotech) on a luminescence imaging system (Fuji). Reprobing of membranes was carried out in accordance with the manufacturer's protocol.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Elk-1 Behaves as an Oligomer in Solution-- We previously developed a bacterial system to generate purified Elk-1 protein (15). Although yields were not high (250-500 µg/liter of culture), purity and homogeneity were good, with only two major contaminants from the single step purification. Sequencing revealed one of them to be a previously identified bacterial protein with high affinity for nickel-agarose (32).

Bacterial Elk-1 elutes from a Superose-12 column as a single peak with an apparent molecular mass of 162 ± 11 kDa (Fig. 1). As the calculated molecular mass of Elk-1 is 45 kDa and given our previous demonstrations of the functional integrity of bacterial Elk-1 (8, 15, 29), this observation suggests that Elk-1 is unlikely to exist in solution as a monomer. Although the protein preparations were relatively clean, we considered that the behavior of Elk-1 might be due to its interaction with one or more bacterial proteins or some other consequence of expressing Elk-1 in bacteria. To rule out these possibilities, we expressed the same protein in COS cells and applied it to the same gel filtration column. Again Elk-1 eluted with an apparent molecular mass in excess of 160 kDa, indicative of an oligomeric structure (Fig. 2A).

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Gel filtration analysis of Elk-1. Bacterially expressed Elk-1 was applied to a Superose-12 HR 10/30 gel filtration column. Fractions were collected and analyzed by Western blotting with an anti-Elk-1 specific antibody. The column was calibrated using standard molecular mass markers, which are indicated above the blot (see "Experimental Procedures"). Quantification was performed on the digital image. AU refers to arbitrary units. The apparent molecular mass for bacterial Elk-1(his) was 162 ± 11 kDa (n = 2).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Detection of Elk-1 oligomers in eukaryotic cell lysates. COS cells were transfected with expression constructs for his-tagged Elk-1 (A) or the Elk-1 deletion mutant B (B). Cell lysates were applied to a Superose-12 HR 10/30 gel filtration column, and 0.5 ml fractions were collected. Elk-1(his) was concentrated by incubation with nickel-agarose beads and analyzed by Western blotting using the anti-His antibody (Qiagen). Quantification was performed on the digital image. AU refers to arbitrary units. The apparent molecular mass for full-length Elk-1(his) expressed in COS cells was 185 ± 23 kDa (n = 3).

Elk-1 is known to form complexes with other eukaryotic proteins, specifically with SRF and MAPKs. The domains of Elk-1 responsible for these specific interactions, including the B-box and D-box, respectively (see Fig. 4A), have been mapped (6, 22, 23). To rule out the possibility that the high apparent molecular mass of Elk-1 expressed in COS cells was due to interactions with either SRF or MAPKs, Elk-1 proteins lacking each of these short domains (see Fig. 4A) were overexpressed, and their oligomeric status was assessed. Although the peaks shifted marginally, both deleted proteins eluted in the same apparent molecular mass range (Fig. 2B and data not shown). Based on an apparent molecular mass of 62 kDa for Elk-1 expressed in eukaryotic cells, the most likely explanation for these results is that Elk-1 forms dimers in solution. However, as the behavior of a protein under gel filtration is dependent on its Stokes radius and not its molecular mass, we cannot rule out the formation of Elk-1 tetramers.

Elk-1 Forms Dimers in Eukaryotic Cells-- We reasoned that if Elk-1 dimers exist, they could be covalently linked by a chemical cross-linking reagent such as DSS, which would capture and stabilize them. We therefore overexpressed His-tagged Elk-1 in HEK 293 cells, which were serum-starved or starved and stimulated with EGF. Cell lysates were prepared, treated with varying concentrations of DSS, and analyzed by SDS-polyacrylamide gel electrophoresis and Western blotting with anti-HA antibodies. Incubation of lysates from either starved or EGF-stimulated cells with DSS led to the appearance of a high molecular mass Elk-1 species (Fig. 3, upper panel). The apparent molecular mass of this species is ~130 kDa, which implies that Elk-1 is present in a dimeric complex in HEK 293 cells. In some experiments a second complex in excess of 200 kDa was also detected (Fig. 3, upper panel, upper arrow). This may be due to Elk-1-forming tetramers or Elk-1 in complex with other proteins.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Chemical cross-linking of Elk-1 dimers. HEK 293 cells were transfected with his-tagged Elk-1 (upper panel), the deletion mutant HA-Elk-1ets (middle panel), or HA-Elk-1Apa (lower panel) and serum-starved or starved and stimulated with EGF. Cell lysates were prepared and treated with various concentrations of DSS (as indicated) and analyzed by Western blotting using the anti-HA antibody (Roche Molecular Biochemicals) or an anti-Elk-1 polyclonal serum. The cross-linked complex is indicated with arrows, and the Elk-1 monomer is labeled M. Numbers to the left of each panel refer to the sizes (in kDa) of the molecular mass markers.

It has previously been shown that Elk-1 is able to form a quaternary complex with SRF on the c-fos SRE and that the complex is stabilized, in part, by interactions between Elk-1 carboxyl-terminal domains (8). In addition, other ets family proteins have been shown to interact through their ets domains (30). We therefore looked for complexes containing Elk-1 proteins that lack various domains (see Fig. 4A). Elk-1 from which the ets domain had been deleted (HA-Elk-1ets) yielded a similar low mobility species upon DSS cross-linking (Fig. 3, middle panel). Similarly, the removal of a carboxyl-terminal portion of Elk-1 (HA-Elk-1Apa) allowed a single complex to form. We also tested the two deletion mutants Elk-1B and Elk-1D in this assay. Again both gave rise to a comparable high molecular mass complex (not shown). EGF stimulation had no effect on the formation of these complexes, suggesting that phosphorylation has no effect on the oligomeric status of Elk-1. The simplest explanation for these observations is that Elk-1 forms dimers involving two types of homotypic interaction, one between ets domains and the other mediated by a carboxyl-terminal motif. Consequently, deletion of either the ets domain or carboxyl terminus from Elk-1 still allows dimers to form through the remaining domain.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Elk-1 dimers have two distinct interaction interfaces. Panel A, diagrammatic representation of the Elk-1 deletion mutants used in this study. Letters with boxes represent Ets domain (A), SRF interaction domain (B), MAPK interaction domain (C), and transactivation domain (D). HEK 293 cells were transfected with Elk-1(his) and HA-Elk-1 (panel B), Elk-1-ets(his) and HA-Elk-1-ets (panel C), Elk-1-19(his) and HA-Elk-1-Apa (panel D), or Elk-1-ets(his) and HA-Elk-1-Apa (panel E) either alone or in combination. Cells were serum-starved or starved and stimulated with EGF, and interactions (beads) were detected by incubation of the lysates with nickel-agarose beads followed by Western blotting with the anti-HA antibody (Roche Molecular Biochemicals). Prebeads refers to lysates subjected directly to Western blotting.

Two Elk-1 Domains Undergo Homotypic Interactions-- To probe the involvement of Elk-1 domains in directing the formation of Elk-1 dimers further, we co-expressed versions of Elk-1 in HEK 293 cells (Fig. 4A) in pairs, one with a carboxyl-terminal histidine tag (Elk-1(his)) and the other with an amino-terminal hemagglutinin epitope (HA-Elk-1). Elk-1(his) was then isolated from cell lysates on nickel-agarose beads and detected with an antibody against the tag (anti-His). Its association with HA-Elk-1 was assessed by reprobing with an antibody against the epitope tag (anti-HA). As shown in Fig. 4B, a fraction of HA-Elk-1 co-purifies with Elk-1(his) in both serum-starved and EGF-stimulated cells, again suggesting that interactions occur irrespective of the phosphorylation state of Elk-1, which is indicated by the anti-phospho-Elk antibody.

Co-expression of similarly tagged versions of Elk-1 from which the ets domain had been deleted (Elk-1ets(his) and HA-Elk-1ets) also led to HA-Elk-1ets co-purifying with its His-tagged counterpart on nickel-agarose beads (Fig. 4C). This confirms that Elk-1 interactions can occur independently of the ets domain, consistent with our previous results (8). The association between Elk-1 carboxyl-terminal domains was shown previously to require a short, 19-amino acid motif containing hydrophobic residues. As shown in Fig. 4D, a His-tagged version of Elk-1 lacking this motif (Elk-119(his)) was able to direct the co-purification of HA-Elk-1Apa. In this experiment both proteins can be identified by an antibody specific for Elk-1 (anti-982). This confirms our conclusions from the cross-linking experiments that besides the carboxyl terminus, a second domain in Elk-1 is able to direct Elk-1 dimer formation.

Given that other ets proteins have been shown to undergo dimer formation via their ets domains (30) we considered this to be feasible for Elk-1, which would explain the ability of Elk-119(his) and HA-Elk-1Apa to interact. Thus we predicted that a version of Elk-1 without a carboxyl-terminal domain would be unable to interact with Elk-1 lacking an ets domain. When we tested the ability of Elk-1ets(his) to interact with HA-Elk-1Apa under identical conditions, we were unable to detect co-purification of these two forms of Elk-1 (Fig. 4E).

Elk-1 and Sap1a Do Not Interact-- Elk-1 is one of three known TCFs that share structural and functional homology. All three are co-expressed in several cell lines, including mature B lymphocytes (3).² The formation of TCF heteromers would offer the potential for further levels of transcriptional control. We were therefore interested to see if Elk-1 could interact with other TCFs. To this end we overexpressed Sap1a(his) and HA-Elk-1 together and looked for co-purification on nickel-agarose beads. As shown in Fig. 5, Elk-1 could not be detected in complex with Sap1a, suggesting that heteromers between Elk-1 and Sap1a do not form. Furthermore, this result also confirms that nonspecific interactions are not the cause of Elk-1 dimer formation.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   TCF heteromers do not form. HEK 293 cells were transfected with Sap1a(his) and HA-Elk-1 either alone or in combination. Cells were serum-starved or starved and stimulated with EGF, and interactions (beads) were detected by incubation of the lysates with nickel-agarose beads followed by Western blotting with the anti-HA antibody (Roche Molecular Biochemicals). Prebeads refers to lysates subjected directly to Western blotting.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The role of TCFs in regulating gene expression in response to extracellular stimuli is established, but their function at the mechanistic level is only superficially understood. Both DNA binding and transcriptional activation by TCFs is regulated principally by ERKs, which, once phosphorylated by MAP/ERK kinases (MEK), translocate to the nucleus as active dimers. Given that interactions between promoters and TCFs are dynamic and, in the absence of any data suggesting that TCFs and ERKs can form stable, DNA-bound complexes, we hypothesized that TCFs may also have the capacity to form dimers. This contention is supported here by three lines of evidence.

Elk-1 Is Oligomeric in Solution-- Elk-1 derived from either bacterial or eukaryotic cells elutes as a single peak from gel-filtration columns, with a molecular mass above 160 kDa. Although this result may be due to the formation of complexes between Elk-1 and other proteins, this is unlikely given the different sources of Elk-1 and the fact that it was overexpressed in both systems. We also assessed forms of Elk-1 from which the short domains targeted by SRF and MAPKs had been deleted (ElkB and ElkD, respectively) (6, 22). These two deletion mutants behave in the same way as full-length Elk-1, eluting from the column at the same point as Elk-1 or, in the case of ElkB, slightly later, presumably due to its reduced size. These results provide a more specific means of ruling out that the biophysical characteristics of Elk-1 are due to its interactions with other eukaryotic proteins. As the molecular mass of Elk-1 lies somewhere between a calculated value of 45 kDa and an apparent value in denaturing polyacrylamide gels of 62 kDa, the simplest interpretation of these observations is that Elk-1 forms dimers in solution. However, the possibility that Elk-1 forms a tetramer cannot be ruled out.

Elk-1 Dimers in Cell Lysates-- Protein cross-linking experiments yielded a high molecular mass Elk-1 species from both starved and EGF-stimulated cell lysates. The apparent molecular mass of 130 kDa in denaturing SDS gels is consistent with Elk-1 dimerization. That only a fraction of Elk-1 is captured in complexes in these experiments is expected due to the relative inefficiency of the cross-linking agent in whole cell lysates. However, the degree of cross-linking varied among experiments, and in some instances a second, larger complex was observed, which may have formed by Elk-1 complexing with other cellular proteins or as a result of tetramer formation by Elk-1. At present we cannot distinguish between these two possibilities. However, deletion of domains known to confer interaction with SRF and MAPKs did not influence the major cross-linked species (not shown). We therefore infer that this complex consists exclusively of Elk-1. Moreover, deletion of either the amino- or carboxyl-terminal region of Elk-1 had no effect on the formation of this complex. Thus both amino- and carboxyl-terminal domains of Elk-1 appear to direct homotypic interactions in the absence of the remainder of the molecule.

Other ets proteins have been shown to form dimers. Erg proteins, of which there are three isoforms, form homo- and heterodimers with each other and with Fli-1, Ets-2, Er81, and Pu-1 (30). Two domains of Erg are implicated: the ets domain, which directs both homo- and heterodimerization, and the "pointed" domain, which is implicated only in homodimer formation. Although this would appear to be similar to the case of Elk-1 addressed here, the pointed domain is absent from TCFs.

Two Dimerization Interfaces in Elk-1-- Elk-1 interactions could also be detected in eukaryotic cell lysates by co-immunoprecipitation experiments (results not shown) and by an analogous approach involving affinity purification of histidine-tagged molecules. This kind of experiment is widely established as a means of determining whether proteins associate within cells. Essentially it allowed us to confirm and refine the indications of the cross-linking experiments; the amino-terminal ets domain of Elk-1 interacts with forms of Elk-1 containing the ets domain, and equally, the carboxyl-terminal domain interacts with Elk-1 proteins harboring an intact carboxyl-terminal domain. The deletion of 19 amino acids (375-394) from Elk-1 was sufficient to abolish the carboxyl-terminal interaction (HA-Elk-1ets and Elk-119(his); data not shown), in line with previous results obtained in vitro with glutathione S-transferase-Elk fusion proteins (8). Heterotypic interactions between amino and carboxyl-terminal domains of Elk-1 could not be detected. Thus, Elk-1 contains two distinct interaction interfaces that function independently.

We also assessed the ability of Elk-1 to form heteromers with its closest relative Sap1a. Despite levels of Sap1a(his) expression exceeding those of Elk-1ets(his), we were unable to detect any interaction between Elk-1 and Sap1a. In the first instance, this result demonstrates that the interactions we observe in these assays do not result arbitrarily from the overexpression of the proteins in question. Secondly, given the high degree of similarity between these two proteins, it appears that small differences may dictate dimer formation by TCFs, in contrast with at least one other ets protein, Erg, which appears to be somewhat more promiscuous (30).

Elk-1 Phosphorylation and Dimer Formation-- The phosphorylation state of Elk-1 had little if any effect on its behavior under gel filtration (results not shown), as Elk-1 eluted as a single peak at the same position whether the protein was purified from serum-starved or serum-stimulated cells. However, it is possible that the column was unable to resolve unphosphorylated and phosphorylated oligomers clearly. Differences between tetramers and dimers may also not have been apparent, although we are confident that monomers would have been resolved. In addition, there was no difference apparent in Elk-1 cross-linking experiments performed with lysates from starved and EGF-stimulated cells. More explicitly, co-purification of Elk-1 species was observed regardless of whether the experiments were performed with lysates of serum-starved or EGF-stimulated cells. Although not implying that the conformation of Elk-1 remains unchanged subsequent to Elk-1 phosphorylation by MAPKs, these results do indicate that under our experimental conditions the intermolecular interactions governing dimerization in solution are maintained.

What then is the role of dimerization in Elk-1 function? We have found recently that the occupation of a SRE in the egr-1 promoter by TCFs alters when B cells are stimulated by B cell receptor engagement.² This implies either that upon activation TCFs transfer from one binding site to another within the genome or that when inactive, they can exist in the nucleus unbound to DNA. Presumably, the higher the level of TCF expression in cells, the more unbound TCF will be present. In this state they would be targeted by MAPKs, which are active in the nucleus as dimers. Elk-1 dimers would thus provide an appropriate substrate for activated MAPKs. It is known that TCFs are high affinity substrates of MAPKs, particularly ERKs, in the absence of DNA and that as a consequence of phosphorylation, the DNA binding affinity of TCFs is increased (9, 10, 14, 15). One means of achieving the latter would be if phosphorylation-dependent conformational changes in Elk-1 unmasked the DNA binding interface such that, in the presence of appropriate sites, DNA binding would then ensue. Thus DNA may effectively cause dissociation of the ets domains. This model is outlined in Fig. 6. Notably, our experiments were performed in the absence of DNA. Although the solution structures of several ets domains unequivocally show that this motif binds to DNA as a monomer (33, 34), no data on the structure of the ets domain in the absence of DNA is presently available.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Model illustrating the possible role of Elk-1 dimers in gene regulation. Unphosphorylated Elk-1 may exist in cells as dimers unbound to DNA, providing a suitable substrate for activated ERK/MAPK dimers. Phosphorylation of Elk-1 by MAPKs is predicted to induce conformational changes affecting the ets domain and resulting in its increased affinity for the SRE. Binding to the DNA in a ternary complex with an SRF dimer may cause complete dissociation of Elk-1, or retention of carboxyl-terminal interactions may favor formation of quaternary complexes. A, C, and D refer to the ets domain, carboxyl-terminal regulatory domain, and MAPK docking site of Elk-1, respectively. The ERK dimer, SRF dimer, ternary, and quaternary complexes are labeled.

In the case of c-Jun regulation by SAPK/JNKs, a different mechanism appears to prevail. The viral oncogene v-Jun lacks the  domain of c-Jun, which serves as a docking site in c-Jun for SAPK/JNKs (35). This deletion results in a gain of function, consistent with the idea that inactive SAPK/JNKs serve as inhibitors of c-Jun function (36, 37). In this context it was shown recently that SAPK/JNKs can interact with c-Jun via the  domain and a second interface in the DNA binding domain to form a stable, DNA-bound complex (21). Although TCFs also have a MAPK-docking motif similar to the c-Jun  domain, its deletion from Elk-1 was shown to result in a transcription factor that is activated only weakly (22), and no evidence of a stable DNA-bound complex comprising TCFs and MAPKs has been forthcoming (8).

On the basis of our model, at least two predictions can be made. First, if the DNA binding interface of unphosphorylated Elk-1 is masked by intermolecular interactions, stabilization of the inactive Elk-1 dimer would make it refractory to DNA binding even after phosphorylation by MAPKs. Second, dimers formed by Elk-1 mutants lacking the carboxyl-terminal interface may dissociate in the presence of ets DNA binding sites. Experiments to test these predictions are currently under way.

	ACKNOWLEDGEMENTS

We thank Chris Ellson for the plasmid pCMV5NL#1, Francisco Cruzalegui and Tahir Pillay for reading and commenting constructively on the manuscript, and members of the group for help and stimulating discussions.

	FOOTNOTES

^* This work was supported by the Wellcome Trust.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Present address: Marie Curie Research Institute, The Chart, Oxted, Surrey RH8 0TL, UK.

^§ To whom correspondence should be addressed. Tel.: 49 (0)115 970 9362; Fax: 49 (0)115 970 9926; E-mail: peter.shaw@nott.ac.uk.

² E. D. Gallagher and P. E. Shaw, manuscript submitted.

³ V. Drewett, S. Muller, J. Goodall, and P. E. Shaw, unpublished information.

	ABBREVIATIONS

The abbreviations used are: TCF, ternary complex factor; SRF, serum response factor; SRE, serum response element; MAPK, mitogen-activated protein kinases; ERK, extracellular signal-regulated kinase; SAPK/JNK, stress-activated protein kinase/c-Jun-N-terminal kinase; HA, hemagglutinin; EGF, epidermal growth factor; DSS, disuccinimidyl suberate.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				H.-M. Zhang, L. Li, N. Papadopoulou, G. Hodgson, E. Evans, M. Galbraith, M. Dear, S. Vougier, J. Saxton, and P. E. Shaw Mitogen-induced recruitment of ERK and MSK to SRE promoter complexes by ternary complex factor Elk-1 Nucleic Acids Res., May 1, 2008; 36(8): 2594 - 2607. [Abstract] [Full Text] [PDF]

				A. Lipshtat, S. P. Purushothaman, R. Iyengar, and A. Ma'ayan Functions of Bifans in Context of Multiple Regulatory Motifs in Signaling Networks Biophys. J., April 1, 2008; 94(7): 2566 - 2579. [Abstract] [Full Text] [PDF]

				E. D. Gallagher, S. Xu, C. Moomaw, C. A. Slaughter, and M. H. Cobb Binding of JNK/SAPK to MEKK1 Is Regulated by Phosphorylation J. Biol. Chem., November 22, 2002; 277(48): 45785 - 45792. [Abstract] [Full Text] [PDF]

				V. Drewett, H. Molina, A. Millar, S. Muller, F. v. Hesler, and P. E. Shaw DNA-bound transcription factor complexes analysed by mass-spectrometry: binding of novel proteins to the human c-fos SRE and related sequences Nucleic Acids Res., January 15, 2001; 29(2): 479 - 487. [Abstract] [Full Text] [PDF]

				D. A. Fantz, D. Jacobs, D. Glossip, and K. Kornfeld Docking Sites on Substrate Proteins Direct Extracellular Signal-regulated Kinase to Phosphorylate Specific Residues J. Biol. Chem., July 13, 2001; 276(29): 27256 - 27265. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Drewett, V. Articles by Shaw, P. E. Search for Related Content PubMed PubMed Citation Articles by Drewett, V. Articles by Shaw, P. E. Social Bookmarking                      What's this?