ELF-1 interacts with and transactivates the IgH enhancer pi site.

We previously identified a B-cell-specific regulatory element in the immunoglobulin heavy chain (IgH) enhancer, pi, with striking similarity to binding sites for ets-related transcription factors. Whereas the ability of ets-related factors to bind to and transactivate the pi site has been substantiated, the identification of the particular member of the ets family responsible for B-cell-specific regulation of the pi site has remained controversial. We have used antibodies specific for individual members of the ets family to evaluate which ets-related factor in B-cell nuclear extracts interacts with the IgH pi site. We present strong evidence that ELF-1 is highly expressed in B-cells and is one of two major factors specifically interacting with the murine IgH enhancer pi site in B-cell nuclear extracts. Binding of ELF-1 correlates with activity of the pi site, since mutations abolishing function of pi also inhibit binding of ELF-1. Furthermore, we demonstrate that ELF-1 can transactivate the IgH enhancer in HeLa cells, suggesting a role for ELF-1 in B-cell-specific IgH gene expression.

In our approach to understand the molecular mechanisms underlying B lymphocyte development, we have focused on the role of specific transcription factors in the regulation of IgH 1 gene expression (1)(2)(3)(4). Several distinct regulatory regions have been characterized in the IgH gene that are dispersed over the whole IgH gene cluster (1). These regulatory regions confer both cell type and developmental stage specificity to IgH transcription and include an upstream promoter as well as enhancer regions in introns and 3Ј of the IgH gene (1,5). We have focused our attention on the 700-base pair intronic enhancer located between the last joining region exon and the first coding region exon (1,6,7). This intronic IgH enhancer functions as a B-cell-specific enhancer that is already active at very early stages of B-cell differentiation prior to IgH gene rearrangement. B-cell specificity is defined by both positively acting B-cell-specific enhancer elements and negatively acting non-Bcell silencer elements cooperating with ubiquitously active regulatory elements (1).
We and others have previously identified two novel IgH enhancer elements, and B, the activations of which appear to be central for B-cell-specific expression of the IgH gene (1-3, 8 -10). Whereas the B site is active throughout B-cell development, the site is primarily active at early stages of B-cell development (2,3,11). Since the and B enhancer elements show striking similarity to binding sites for transcription factors of the ets gene family (12,13), we have attempted to determine the nature of the factors interacting with these sites. The ets gene family shares a highly conserved DNA binding domain and comprises a group of now more than 20 different transcription factor genes, the aberrant expression of which has been directly linked to tumorigenesis in humans (12)(13)(14)(15). ets factors have been implicated in the transcriptional regulation of a whole variety of genes, in particular genes involved in differentiation, proliferation, and cell type specificity (12,13). Within the immune system, ets factors appear to be involved in the regulation of many B-and T-cell-specific as well as monocyte/macrophage-specific genes (3, 4, 9, 16 -33). Most ets factors do not manifest absolute tissue specificity, but expression as well as activation can be limited to only a few cell types and/or highly regulated upon differentiation or activation. Thus, Pu.1, for example, is a member of the ets family that is primarily expressed in the B-cell and myeloid cell lineage (26,34,35). Different members of the ets family can act either as enhancers or as repressors of transcription, the activity of which is regulated many times by an interplay of different signal transduction pathways leading to phosphorylation and dephosphorylation of the particular ets factor (12,13). Since each cell type normally expresses a variety of ets-related genes, the presence of an ets binding site in a gene cannot immediately be correlated with a specific member of the ets family.
At least 10 different members of the ets family are expressed in the B-cell lineage, and the assignment of one specific ets factor for the activity of the IgH enhancer site has remained controversial. Nelsen et al. (9) have reported that ets-1 can bind to the site and, when overexpressed, can activate the site. Rivera et al. (10), however, showed that fli-1 or erg-3 when overexpressed activates the site. We have recently cloned a new member of the ets family, ERP, which is within the B-cell lineage highly expressed at the preB-cell stage (4). We demonstrated that ERP is able to bind to the site as well (4), thus leaving open the question of which of these ets factors, if any, regulates the site in B-cells.
Due to high conservation of the DNA binding domain among all members of the ets family, DNA motifs recognized by different Ets family members are very similar (12,13). Thus, the DNA sequence of a particular ets binding site by itself will not immediately reveal which Ets-related factor is the functionally relevant protein. We have, therefore, used antibodies specific for individual members of the Ets family to evaluate in electrophoretic mobility shift assays (EMSA) which Ets-related factor in B-cell nuclear extracts interacts with the IgH enhancer site. We demonstrate here that the ets factor ELF-1 is one of two major protein⅐DNA complexes formed by B-cell nuclear extracts when incubated with the IgH site, suggesting a role of ELF-1 in the B-cell-specific function of the IgH site.
Nuclear Extracts-Nuclear extracts were prepared according to the method of Dignam et al. (36). All buffers included leupeptin at 0.3 g/ml, 5 mM phenylmethylsulfonyl fluoride, antipain at 0.3 g/ml, and aprotinin at 2 g/ml.
Electrophoretic Mobility Shift Assay-DNA binding reactions and EMSAs were performed as described (3,4). Samples of 20 l containing 5 g of nuclear extract were incubated with 0.1-0.5 ng of 32 P-labeled IgH wild-type site DNA fragment (5,000 -25,000 cpm), 10 mM Tris-Cl, pH 7.5, 50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 5% glycerol, 2 g of bovine serum albumin (Boehringer Mannheim), and 0.1 g of poly[d(I⅐C)] (Pharmacia Biotech Inc.). For supershift assays, samples were preincubated for 10 min in the presence of 0.5 l of antiserum or preimmune serum prior to the addition of the labeled probe for an additional 15 min. Polyclonal rabbit antibodies against Ets-1, Ets-2, PEA3, Pu.1, ERG-1/2, ERG-1, ELK-1, SAP-1A, and Fli-1 were purchased from Santa Cruz Biotechnology. The generation and characterization of polyclonal rabbit peptide antibodies against ERP will be described elsewhere. 2 Polyclonal rabbit antibodies against ELF-1 were as described (30) and kindly provided by Craig Thompson. Samples were incubated in the presence or absence of increasing amounts of competitor oligonucleotides (0.1, 1, and 10 ng) for 15 min at room temperature and run on 4% polyacrylamide gels, containing as buffer 25 mM Tris-HCl, pH 8.5, 190 mM glycine, and 1 mM EDTA. Oligonucleotides used as probes and for competition studies are as described in the legend to Fig. 3B.
Expression Vector and Luciferase Reporter Gene Constructs-A blunted XbaI-SalI fragment of the ⌬56-c-fos-CAT plasmid as described previously by Gilman et al. (38) containing the c-fos minimal promoter region from Ϫ56 to ϩ107 was inserted into the blunted HindIII site upstream of the luciferase gene in the pGL3 vector (Promega). Synthetic murine IgH wild-type and mutant M2 site oligonucleotides containing SalI and XhoI ends were inserted as trimers into the SalI site of the ⌬56-c-fos-pGL3 plasmid.
DNA Transfection Assays-Cotransfections of 3 ϫ 10 5 HeLa cells were carried out with 3.5 g of reporter gene construct DNA and 2 g of expression vector DNA using 12.5 l of Lipofectamine (Life Technologies, Inc.). Cells were incubated with the liposomes and DNA for 4 h at 37°C, harvested 16 h after transfection, and assayed for luciferase activity as described (39). Transfections for every construct were performed independently in triplicates and repeated 2-4 times with at least two different plasmid preparations with similar results. Cotransfection of a second plasmid for determination of transfection efficiency was omitted because potential artifacts with this technique have been reported (40) and because many commonly used viral promoters contain potential binding sites for ets factors. The protein concentration was measured with a kit from Bio-Rad and normalized for all samples in each individual experiment.
Transfections of PD31 preB-cells were carried out with 10 g of DNA using the DEAE-dextran method (37,41). The cells were harvested 48 h after transfection and assayed for CAT activity as described (37,42) in a 2-h incubation at 37°C. Transfections for every construct were performed independently in duplicates and repeated two to four times. Samples were analyzed using thin layer chromatography (42).

Two Major Nuclear Factors Bind to the Murine IgH Site in
B Cells-We recently identified and characterized a novel regulatory element in the murine IgH enhancer, (Fig. 1) (3). Having realized that the IgH enhancer site exhibits similarities to ets binding sites, we set out to characterize which members of the ets family might be involved in transcriptional regulation of this element in B-cells. EMSA analysis of the murine IgH site with nuclear extracts from the murine late preB-cell line NFS5.3 reveals the formation of two predominant protein⅐DNA complexes and several additional weaker complexes ( Fig. 2A).

ELF-1 Is One of Two Predominant Proteins Specifically
Binding the IgH Site in B-Cells-To determine whether any of the protein⅐DNA complexes formed by the IgH site with nuclear factors from B-cells is due to the interaction of an ets-related factor, we used a panel of antibodies against different members of the ets family in an EMSA supershift assay. We compared the ability of these antibodies to either inhibit binding of a protein⅐DNA complex to the IgH site or to form a slower migrating antibody⅐protein⅐DNA complex with the IgH site ( Fig. 2A). The majority of antibodies did not affect the mobility of any protein⅐DNA complex formed by murine NFS 5.3 late preB-cell nuclear extracts. Antibodies against ELF-1, however, completely shift one of the two predominant protein⅐DNA complexes, as indicated by the arrows in Fig. 2 suggesting that ELF-1 or a highly related factor binds with high affinity to the IgH site in preB-cell nuclear extracts. The ELF-1 antibody did not cross-react with any other ets family member. 3 To determine whether nonhematopoietic cells also form complexes with the IgH site that contain ELF-1, we performed EMSA with nuclear extracts from HeLa cervical carcinoma cells (Fig. 2B). HeLa cell nuclear extracts formed several of the minor complexes visible with B-cell extracts, but none of the complexes comigrated with the ELF-1-specific complex seen in 38B9 preB-cells or U-937 monocytic cells. Indeed, none of the antibodies reacted with complexes formed by HeLa cell nuclear extracts, indicating that ELF-1 is not highly expressed in HeLa cells.
Mutation of the ets Consensus Sequence Abolishes ELF-1 Binding to the IgH Site-To analyze the DNA sequence requirements for the binding of ELF-1 and the other B-cell nuclear factors to the IgH site, we designed mutant IgH site oligonucleotides containing two-nucleotide changes in different regions of the IgH site (Fig. 3B). EMSA analysis was performed using the wild-type IgH site oligonucleotide as probe and nuclear extract from NFS 5.3 B-cells. Competition with increasing amounts of unlabeled wild-type oligonucleotide demonstrates that the majority of the protein⅐DNA complexes (A to D) are specific (Fig. 3). Competition analysis with increasing amounts of mutant oligonucleotides reveals that mutations affecting the core "GGAA" recognition motif for ets-related factors, namely mutations M2, M3, and M6, abolish binding of ELF-1 (complex C) and the second predominant factor (complex B) without diminishing binding of the other factors (complexes A and D) (Fig. 3). Mutations in other regions of the IgH site have either no effect or weaker effects on binding of these two predominant factors. Similarly, an unrelated oligonucleotide of the same length does not compete with any factor binding to the IgH site (Fig. 3). Interestingly, mutant M5 that introduces changes at the 3Ј end of the IgH site abrogates binding of a slow migrating complex A, suggesting that this factor most likely interacts with the 3Ј part of the IgH site, whereas mutant M1, which introduces changes at the 5Ј end of the IgH site, abolishes binding of a faster migrating complex D, indicating binding of this factor to the 5Ј end of the IgH site (Fig. 3). These results support the notion that the two major B-cell protein⅐DNA complexes B and C are formed by ets-related factors and specifically interact with the IgH site. None of the mutations was able to distinguish between the two predominant factors, indicating that both have identical sequence requirements.
Mutations That Inhibit Binding of ELF-1 Knock Out the Activity of the IgH Site-To examine whether mutations that inhibit binding of ELF-1 to the IgH site affect the enhancer function of the IgH site, we tested trimers of either the wild-type or the mutant M6 and M7 IgH site oligonucleotides that had been placed upstream of a minimal c-fos promoter driving the CAT gene in the ⌬56 plasmid in transient transfection assays. Upon transfection of these constructs into the preB-cell line, PD31, the wild-type IgH site expresses strong enhancer activity. Mutation M6, which replaces the two-core GG with TT, abrogates the activity of the IgH site, whereas mutation M7 has only partial effects (Fig. 4). These data demonstrate that a mutation that inhibits interaction of ELF-1 3 Y. Akbarali, P. Oettgen, J. Boltax, and T. A. Libermann, unpublished results. with the IgH site also drastically diminishes the enhancer activity of the IgH site. Mutations affecting binding of ELF-1 also inhibit binding of the second major factor, but we are thus far unable to determine which of these two factors might be the functionally more important.
ELF-1 Transactivates the IgH Enhancer Site-To determine whether ELF-1 has the capacity to transactivate the IgH enhancer site, full-length ELF-1 was inserted into the eukaryotic expression vector pCI downstream of the cytomegalovirus promoter and then cotransfected into HeLa cells, together with reporter gene constructs containing trimers of either the wild-type or the mutant M2 IgH site oligonucleotides, which had been placed upstream of a minimal c-fos promoter driving the luciferase gene in the pGL3 plasmid (38). ⌬56-pGL3 containing only the minimal c-fos promoter expressed very little luciferase activity above the background of the parental promoterless pGL3 vector (Fig. 5A). Three copies of the wild-type IgH enhancer site exhibited significant activity, whereas a construct containing the same copy number of the mutant IgH enhancer site was transcriptionally inactive. Cotransfection with an ELF-1 expression vector resulted in a 3-fold transcriptional stimulation of the wild-type IgH enhancer site (Fig.  5A). Mutation of the IgH enhancer site abolished transactivation by ELF-1, confirming the specificity of transactivation.
To further define the ability of ELF-1 to transactivate the single site in the context of the IgH enhancer, we inserted a fragment of the IgH enhancer containing the region from the E2 site to the site into ⌬56-pGL3 (E2-/⌬56-pGL3). Cotransfection experiments into HeLa cells show that E2-/⌬56-pGL3 expresses very little enhancer activity. However, ELF-1 activates transcription of E2-/⌬56-pGL3 by 4 -5-fold (Fig.  5B). Similarly, the 140-base pair PstI-DdeI fragment of the IgH enhancer (140) (Fig. 1) containing a minimal B-cell-specific enhancer region, which includes E2, , E3, and B, was inserted into ⌬56-pGL3 (140/⌬56-pGL3). 140/⌬56-pGL3 expresses only marginal activity in HeLa cells. However, ELF-1 stimulated transcription of 140/⌬56-pGL3 by 6 -8-fold (Fig.  5B). We conclude from these experiments that ELF-1 can efficiently stimulate transcription of the IgH enhancer via the site, supporting the notion that ELF-1 might be a critical factor in IgH enhancer regulation. DISCUSSION We recently described the discovery of a novel enhancer element in the murine IgH enhancer, , which is more active in preB-cells than mature B or plasma cells and shows homology to binding sites for ets-related transcription factors (1,3). To determine which member of the ets family regulates the IgH site in B-cells, we have used antibodies against different ets factors in an EMSA-supershift assay. We were able to identify one of two predominant factors binding in B-cell nuclear extracts to the IgH enhancer site as ELF-1. The nature of the second major protein⅐DNA complex formed by the site remains to be determined. Binding of ELF-1 correlates with the ability to transactivate the IgH enhancer as well as an isolated IgH enhancer site, suggesting a biological role for ELF-1 in IgH gene regulation.
Previously, ELF-1 has been implicated primarily in the regulation of T-cell-specific genes including interleukin 2, interleukin 2 receptor ␣, granulocyte-macrophage colony-stimulating factor, interleukin 3, and CD4 (22,25,30,32,43). Only recently it was shown that ELF-1 is involved in the regulation of the IgH 3Ј enhancer in response to antigen receptor crosslinking in mature B-cells (44). Our results demonstrate that ELF-1 is expressed in B-cells, even at the preB-cell stage, and binds to the intronic IgH enhancer, suggesting a broader biological function of ELF-1. Indeed, we have evidence that ELF-1 is the major nuclear factor binding to ets binding sites in several other B-cell-specific genes. 4  SpiB, and ERP (4,10,26,(45)(46)(47). That ELF-1 is also highly expressed in B-cells, therefore, does not come as a surprise. Ets-related transcription factors appear to play a very central role in the regulation of B-cell-specific gene expression. Many B-cell-specific genes contain functionally important ets-related binding sites in their regulatory regions, including among others the IgH promoter, IgH intronic enhancer, IgH 3Ј enhancer, Ig, Ig, J chain, mb-1, TdT, B29, and lyn genes (3, 9, 10, 20, 21, 23, 28, 29, 44, 48 -50). 4 The nature of the particular member of the ets family regulating a specific ets site has been revealed in only a few cases. In the case of the IgH site, three different members of the ets family, fli-1, ets-1, and erg-3, have been suggested to be involved in activation of the IgH site (9, 10). All three ets-related factors are able to transactivate the IgH site when overexpressed in non-B-cells. We, furthermore, recently cloned a novel member of the ets family, ERP, which is highly expressed in preB-cells and down-regulated upon maturation of B-cells (4). ERP binds to the IgH site, and its expression in the B-cell lineage correlates with the activity of the IgH site (4). To our surprise, neither ets-1, fli-1, erg-3, nor ERP are the proteins in B-cell nuclear extracts that interact with the IgH site with high affinity. The IgH site forms two major specific protein⅐DNA complexes that correlate with the function of the IgH site as well as with the consensus binding motif for ets-related factors. One of these two complexes is formed by the interaction of ELF-1 with the IgH site. The second complex does not react with any of the antibodies tested, indicating that a factor against which we do not have antibodies is involved in the formation of this complex.
Since the DNA binding domain among all members of the ets family is highly conserved, DNA motifs recognized by different Ets family members are very similar. How does a particular ets binding site in a specific regulatory region select a distinctive member of the ets family? One explanation could be differential expression of different members of the ets family in different cell types. Some members of the ets family are indeed expressed only in certain tissues such as Pu.1 (26,51). Most members of the family, nevertheless, are less restricted in expression. In addition, as mentioned above, B-cells express a whole variety of different ets factors, suggesting that specificity of a particular ets site has to be determined by additional criteria. One clue might come from the distinctive structures of different ets family members. Whereas the DNA binding domain is the common link of all members of the ets family, only limited homology exists in other regions of these transcription factors (13). Therefore, most if not all ets family members contain domains involved in protein-protein interactions and differ in their ability to form heterodimers with factors from various different transcription factor families, suggesting that the selection of a unique member of the ets family might be at least partially due to the regulatory elements surrounding a particular ets binding site. A third criterion for specificity might be related to posttranslational modifications due to phosphorylation, which may have an impact on DNA binding, cellular localization, protein stability, protein-protein interaction, and/or transactivation capacity of the individual ets factor. Indeed, many ets factors contain negative regulatory domains that inhibit DNA binding (4,50,52).
Why do we see only ELF-1 and the additional factor binding to the IgH site in B-cell nuclear extracts, even though recombinant ets-1, fli-1, erg-3, and ERP can bind as well and are expressed in B-cells? The combination of different mechanisms described above could explain this. (a) ELF-1 might be more abundant in B-cells than the other ets factors. This is unlikely, since Northern blot analysis shows relatively similar levels of expression for many ets factors, and expression differs for the different ets factors in different B-cell lines. (b) The binding affinity of the individual ets factors toward the IgH site is distinct, such that ELF-1 binds with stronger avidity to the IgH site than ets-1, fli-1, or erg-3. We believe that differences in the DNA binding domains of different ets factors contribute to binding affinity, and there is evidence that nucleotides flanking the core ets binding motif "(A/G)GA(A/T)" influence binding affinity. Hence, ELF-1 might be at least partially selected out from the other ets factors in B-cell nuclear extracts due to stronger affinity toward the IgH site. Interaction of ELF-1 with another protein might increase its affinity above the affinity of the other ets factors. We have no evidence yet that the protein⅐DNA complexes contain additional proteins. Phosphorylation or other posttranslational modifications of ELF-1 in B-cells might, via conformational changes, increase the DNA binding affinity above the affinity of other ets factors. Most likely a combination of several of these mechanisms is involved in the process of selectivity.
In conclusion, we show strong evidence that ELF-1 is involved in regulation of IgH enhancer function. The fact that ELF-1 appears to play an important role in T-cell gene regulation warrants further studies as to its relevance in B-cell gene regulation and differentiation.