Mapping of the Transcriptional Repression Domain of the Lymphoid-specific Transcription Factor Oct-2A*

The lymphoid-specific transcription factor Oct-2a is implicated in B cell-specific transcriptional activity via the octamer motif. Structure/function analysis of various Oct-2a effector regions in the context of the GAL4 DNA-binding domain revealed that Oct-2a contains two functionally different activation domains at the N and the C termini. The transcriptional activity of both domains is strongly potentiated by interactions with distinct B cell-specific coactivators. Recently, we have identified a repression domain located within the N terminus of Oct-2a (amino acids 2–99). When this domain was transferred to a potent activator, transcription was strongly inhibited. In this study we present a deletion analysis of the N-terminal region of Oct-2a to determine the minimal repression domain. We identified a stretch of 23 amino acids, rich in serine and threonine residues, which was responsible for most of the repression activity. We show that repression is strongly dependent on the type of enhancer present in the reporter plasmid as well as on the cell line tested. The possibility that Oct-2a can act as an activator and/or a repressor may have important consequences for the function of Oct-2a in B cell differentiation and other developmental processes. The B cell-restricted 1 genes is promoter and confer B cell-specificity. The its transcription disruption of in a critical the final stage of B cell differentiation and more recently a new role for Oct-2, as a nuclear mediator of B cell proliferation, has been suggested (8). chain reaction fragments N3, N4, and N5 into the respective sites of pGAL4-C-C. The N3 fragment encodes the amino acids 2–64 of Oct-2a and was amplified with the primers 5 (cid:57) -TACTC-AGGATCCGTTCACTCCAGCATGGG-3 (cid:57) and 5 (cid:57) -GTCCTTGGTACCCT- TGATCTTTGTACTG-3 (cid:57) . The N4 region, encoding residues 42–99 of Oct-2a, was generated with the oligonucleotide primers 5 (cid:57) -TAGAG- GATCCCATCAGAACCCCCAA-3 (cid:57) and 5 (cid:57) -CATCTTGGTACCTAGCTG-GCTGCCCGTC-3 (cid:57) . The N5 fragment, encoding Oct-2a residues 65–99,

The lymphoid-specific transcription factor Oct-2a is implicated in B cell-specific transcriptional activity via the octamer motif. Structure/function analysis of various Oct-2a effector regions in the context of the GAL4 DNA-binding domain revealed that Oct-2a contains two functionally different activation domains at the N and the C termini. The transcriptional activity of both domains is strongly potentiated by interactions with distinct B cell-specific coactivators. Recently, we have identified a repression domain located within the N terminus of Oct-2a (amino acids 2-99). When this domain was transferred to a potent activator, transcription was strongly inhibited. In this study we present a deletion analysis of the N-terminal region of Oct-2a to determine the minimal repression domain. We identified a stretch of 23 amino acids, rich in serine and threonine residues, which was responsible for most of the repression activity. We show that repression is strongly dependent on the type of enhancer present in the reporter plasmid as well as on the cell line tested. The possibility that Oct-2a can act as an activator and/or a repressor may have important consequences for the function of Oct-2a in B cell differentiation and other developmental processes.
The B cell-restricted expression of immunoglobulin (Ig) 1 genes is dependent on promoter and enhancer elements that confer B cell-specificity. The octamer motif ATGCAAAT or its inverse complement is a pivotal regulatory element for transcription of all Ig genes. It is conserved in all Ig promoters and most of the Ig enhancers (1).
Recently, a variety of transcription factors binding to the octamer site have been cloned (2), including the ubiquitously expressed Oct-1 protein (3) and the lymphoid-specific factor Oct-2 (4). These octamer site-binding factors are implicated in developmentally regulated gene expression (2,5,6). Targeted disruption of the Oct-2 gene in mice has shown that Oct-2 plays a critical role in the final stage of B cell differentiation (7), and more recently a new role for Oct-2, as a nuclear mediator of B cell proliferation, has been suggested (8).
By transient transfections and in vitro transcription reactions, it has been demonstrated that Oct-1 and Oct-2, together with the B cell coactivator Bob1/OBF-1/OCA-B, efficiently stimulate Ig promoter activity (9 -11). This B cell coactivator interacts with the DNA-binding POU domains of Oct-1 and Oct-2 to potentiate transcription.
Furthermore, we have presented evidence for additional distinct B cell-specific activities that interact with the functionally different activation domains of Oct-2a to enhance transcriptional stimulation (12). In the same study, we identified a repression domain within the N-terminal 99 residues of Oct-2a.
Recently, different negative regulatory regions have been defined within other isoforms of Oct-2 (13,14).
For a long time research has focused mainly on activator and coactivator proteins (15)(16)(17), but recently it became clear that repressors and corepressors also play an important role in regulation of gene expression (18 -20). Repressors can act by passive repression strategies such as competition for DNAbinding sites, quenching, squelching, or the establishment of a repressive chromatin structure around the target promoter. Active repression mechanisms are also involved and imply a direct or indirect (using a corepressor) interaction with the transcription machinery, thus interfering with preinitiation complex assembly (18). Interestingly, several of the transcription factors investigated function as both activators and repressors depending on the target promoter and the cellular context (21). This could also be the case for Oct-2a, which belongs to a group of transcription factors with a complex modular structure consisting of separable domains for activation and repression (12), similar to Krü ppel (22), Even-skipped (22), Egr-1 (23), and YY1 (24).
In this study we describe a deletion analysis of the N-terminal domain of Oct-2a (N1 region) in order to define the region involved in transcriptional repression. Different segments of the N1 region were fused to the potent activator protein GAL4-C-C and tested in transient transfections of B cells and nonlymphoid cells together with luciferase-based reporter plasmids containing different enhancers. We show that the major repressor activity mapped to a region of 23 amino acids with a high abundance of serine and threonine residues (30.4%). Furthermore, we demonstrate that transcriptional repression displayed cell type-specific characteristics, which were determined by the enhancer present in the reporter plasmid.
The chimeric constructs pN3-GAL4-C-C, pN4-GAL4-C-C, and pN5-GAL4-C-C were generated by introduction of the Asp718-BamHIcleaved polymerase chain reaction fragments N3, N4, and N5 into the respective sites of pGAL4-C-C. The N3 fragment encodes the amino acids 2-64 of Oct-2a and was amplified with the primers 5Ј-TACTC-AGGATCCGTTCACTCCAGCATGGG-3Ј and 5Ј-GTCCTTGGTACCCT-TGATCTTTGTACTG-3Ј. The N4 region, encoding residues 42-99 of Oct-2a, was generated with the oligonucleotide primers 5Ј-TAGAG-GATCCCATCAGAACCCCCAA-3Ј and 5Ј-CATCTTGGTACCTAGCTG-GCTGCCCGTC-3Ј. The N5 fragment, encoding Oct-2a residues 65-99, * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. was amplified using the oligonucleotide primers 5Ј-GGTAGGATCCG-CTGAAGACCCCAGTG-3Ј and 5Ј-CATCTTGGTACCTAGCTGGCTGC-CCGTC-3Ј. The constructs were verified by restriction analysis and DNA sequencing. The luciferase reporter plasmids p2GlucE and p2GlucBG5 have been described previously (12). The target vector p2GlucE contains two Gal4-binding sites in the promoter region 13 base pairs upstream of the TATA box in the context of a minimal rabbit ␤-globin promoter and a simian virus 40 (SV40) enhancer 2.85 kilobase pairs downstream of the transcription start site. The luciferase reporter plasmid p2GlucBG5 has the same structure as p2GlucE but contains, instead of the SV40 enhancer, a synthetic enhancer consisting of five Gal4-binding sites.
Cell Culture and Transient Transfections-Namalwa cells were kept in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS) and 50 M ␤-mercaptoethanol. HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with 3% FCS and 3% newborn calf serum. COS-7 cells were grown in Dulbecco's modified Eagle's medium with 10% FCS. For the transient activation assays, HeLa and Namalwa cells were cotransfected with 6 g of reporter and 4 g of effector plasmids. For protein expression, COS-7 cells were transfected with 3 g of the expression vector. Transfections contained a total of 20 g of DNA for HeLa cells or 10 g for Namalwa cells and 5 g for COS-7 cells. Sonicated salmon sperm DNA (Pharmacia) was used as carrier. Transfections were performed by the calcium phosphate coprecipitation method for HeLa cells (25) and by the DEAE-dextran procedure (26) for Namalwa and COS-7 cells as described previously (12).

Preparation of Nuclear Extracts and Electrophoretic Mobility Shift
Assay (EMSA)-Nuclear extracts from transfected COS-7 cells were prepared by a small scale procedure (27). EMSA reactions (28,29) were set up in 20 l with 2.5 g of nuclear extract, 40 fmol of an end-labeled probe containing a single Gal4-binding site, 0.5 g of denatured sonicated salmon sperm DNA (Pharmacia), and 1 g of poly(dI-dC) (Pharmacia) in a buffer containing 20 mM HEPES, pH 7.9, 50 mM KCl, 5 mM MgCl 2 , 10 M ZnCl 2 , 6% glycerol, and 200 g/ml bovine serum albumin. Binding reactions were incubated for 15 min at room temperature and resolved at 15 V/cm for 2 h on a native pre-run 4% polyacrylamide gel (80:1) containing 0.5 ϫ TBE and 1% glycerol. The gel was fixed for 30 min in 20% (v/v) methanol and 10% (v/v) acetic acid, dried, and autoradiographed.
Luciferase Assays-Luciferase assays were performed essentially as described previously (30, 31) using a Berthold MicroLumat LB96P machine. Whole cell extracts were prepared by three cycles of freezethaw lysis in 0.1 M potassium phosphate, pH 7.8, and 1 mM dithiothreitol. The luciferase activity was measured in a total volume of 200 l containing 100 g of the protein extract, 11.25 mM MgSO 4 , 18.75 mM glycine, 2.0 mM ATP (Boehringer Mannheim), and 0.15 mM of the substrate D-luciferin (Sigma), pH 7.8. Light output was measured for 15 s.

RESULTS AND DISCUSSION
In our previous experiments, we have shown that Oct-2a is a complex transcription factor consisting of separable modules for activation and repression. The repression domain was localized to the N-terminal 99 amino acids of Oct-2a, the N1 region (12).
To further delineate the domain important for repression, we have performed a deletion analysis of the N1 region. Various expression vectors were generated containing different regions of the N1 segment, comprising the repression domain, fused to a potent activator protein. The regions N1 (amino acids 2-99), N2 (amino acids 163-207), N3 (amino acids 2-64), N4 (residues 42-99), and N5 (residues 65-99) of Oct-2a were linked to GAL4-C-C (Fig. 1, A and B). As judged by EMSA with nuclear extracts of transfected COS-7 cells, all of these fusion proteins were expressed at comparable levels (Fig. 2).
Transient transfections were performed in HeLa and Namalwa cells with the luciferase reporter plasmids p2GlucE and p2GlucBG5 (Fig. 3). These vectors differ with respect to their enhancers; p2GlucE contains an SV40 enhancer, and p2GlucBG5 contains a synthetic enhancer consisting of five Gal4-binding sites (see also "Materials and Methods").
In Namalwa cells, a Burkitt's lymphoma B cell line, cotrans- fection of the expression vectors together with the reporter plasmid p2GlucE (Fig. 3A) led to transcriptional repression with the constructs N1-GAL4-C-C, N3-GAL4-C-C, and N4-GAL4-C-C, but no repression was seen either with N5-GAL4-C-C or with the control N2-GAL4-C-C. This indicates that repression is dependent on residues common to N1, N3, and N4, which comprise a region at amino acids 42-64 of Oct-2a. The localization of the repression domain to this short region of 23 amino acids is supported by the fact that N4 shows the same degree of repression as the N1 region, containing the complete repression domain, whereas region N5 does not show any inhibitory activity.
None of the constructs tested in HeLa cells, an epitheloid cancer cell line, could repress transcription, and all displayed the same activity as GAL4-C-C (Fig. 3A). This is in agreement with our previous results (12) and shows that repression is impaired by a HeLa cell-specific factor(s).
In Fig. 3B the same series of expression vectors was tested together with the reporter plasmid p2GlucBG5. Repression was obtained with the constructs N1-GAL4-C-C, N3-GAL4-C-C, and N4-GAL4-C-C, and no repression was seen with N5-GAL4-C-C and the control N2-GAL4-C-C in agreement with the results presented in Fig. 3A. Interestingly, transcriptional repression with the reporter plasmid p2GlucBG5 was observed in all of the cell lines tested. Again N4-GAL4-C-C repressed transcription to the same extent as N1-GAL4-C-C, whereas N5-GAL4-C-C did not inhibit transcription in HeLa cells and had only a minor effect in Namalwa cells.
As shown in Fig. 3, the repression domain of Oct-2a can be mapped to residues 42-64. The relative position of the various domains tested (N1, N2, N3, N4, and N5) within the Oct-2a N terminus and the minimal repression domain (dark box) are outlined in Fig. 4.
As mentioned above, region N4 represses transcription to the same extent as the extended N1 region. However, the N3 domain is slightly less efficient than N4 in inhibiting transcription despite containing the minimal repression domain. This may be due to a slight destabilization of the protein structure of the N3 domain, since a high probability for a ␤ sheet was predicted for exactly the C-terminal end of N3 (program Pre-dictProtein), containing the inhibitory domain. This region could be stabilized by the additional C-terminal residues present in the N4 domain. These residues could favor the correct folding of the inhibitory domain giving rise to maximal transcriptional repression.
Repression of the reporter plasmid p2GlucBG5 by N4-GAL4-C-C was observed in both HeLa and Namalwa cells. In contrast, repression of the target vector p2GlucE was only seen in Namalwa cells and not in HeLa cells. This is consistent with our previous data (12) and underlines that repression by Oct-2a follows cell type-specific characteristics that are strongly determined by the enhancer present on the reporter plasmid. Therefore, we conclude that repression by N4-GAL4-C-C is impaired by a HeLa cell-specific factor(s) interacting with the SV40 enhancer. Such a factor(s) could interfere with repression by masking the inhibitory domain of Oct-2a. This working model is in agreement with the finding that the N terminus of Oct-2a can exert its negative effect when linked to other Oct-2a activation domains only if HeLa cell-specific factors binding to the SV40 enhancer are not present (Namalwa cells) or are not able to bind (synthetic Gal4 enhancer) (12).
Region N4, containing the minimal repression domain between residues 42-64, is a transferable inhibitory domain that works independently of its cognate DNA-binding POU domain. N4 linked to the activator GAL4-C-C efficiently inhibits transcription and can therefore be classified as an active transcriptional repressor (19,20,32). This class of repression domains is implicated in contacting components of the basal transcription apparatus, thus interfering with the formation of a transcription-competent initiation complex. This was recently demonstrated for Dr1 (33) and the Drosophila proteins Krü ppel (34) and Even-skipped (35). In the context of the native Oct-2a protein, repression could follow a combination of different mechanisms including competition for DNA-binding sites and quenching, both requiring the presence of the homologous DNA-binding POU domain, or interaction with basal transcription factors. This is not unexpected since different mechanisms of repression, depending on the promoter architecture and the type and amount of distinct factors present in the cellular context, are used by repressor proteins such as Krü ppel (22,34,36) and Even-skipped (35,37).
The repression domain of Oct-2a, which mapped to amino acids 42-64, is rich in serine and threonine residues (30.4%). Protein data base searches with this region did not reveal a significant homology to other known transcriptional repressors, and no particular structural motifs were found. Therefore, this region could represent a novel type of repression domain. Experiments to identify possible targets of the Oct-2a inhibitory domain using a yeast two-hybrid system are currently in progress in our laboratory.
Recently, negative regulatory domains have also been identified within other isoforms of Oct-2 (13,14). These domains are neither related to each other nor to the repression domain mapped in this study. One of these regions was assigned to an N-terminal domain common to both the neuronally expressed mouse isoforms Oct 2.4 and Oct 2.5 (13). This region specifically inhibited activation of a herpes simplex virus immediate-early promoter containing the octamer/TAATGARAT motif and was implicated in the production of latent infection in neuronal cells upon exposure to herpes simplex virus (13). The other repression domain identified was localized to a region that is unique to the mouse isoform Oct 2.3 (14), but is not present in any of the other Oct-2 isoforms (38).
The identification of non-related repression domains within splice variants of Oct-2 indicates that transcriptional repression by the various Oct-2 isoforms might follow distinct regulatory mechanisms, which could be dependent on the cellular context and the structure of the target promoter.
The complex modular structure of Oct-2a, which includes distinct domains for activation and repression, could have important implications for the expression of B cell-specific genes involved in the control of B cell differentiation. However, very little is known about Oct-2-regulated target genes. The promoters of several genes, including the Ig heavy and light chain genes, and of genes coding for B29 (Ig␤), CD21 (Cr2), and CD20 have been shown to be octamer site-dependent and B cellspecific, but not strictly dependent on Oct-2 expression (8,39). Only the expression of the murine CD36 gene has been demonstrated to be critically dependent on Oct-2 in B cells and macrophages so far, although the precise function of this gene remains to be determined (40).
In addition to its role in the terminal phase of B cell development, Oct-2 may also be critical for developmental processes outside of the immune system, because Oct-2-deficient mice die within hours of birth (7). This strongly indicates that developmentally relevant genes exist, whose activation or repression may be required for the postnatal survival of the newborn mouse (7). Since Oct-2 is also expressed in the developing nervous system as well as in the adult brain of mice (41,42), it is tempting to speculate that activation and repression by Oct-2 may have an impact on the fine-tuning of gene regulation in the central nervous system.
The identification of Oct-2a target genes will contribute significantly to our understanding of the complex regulatory functions of Oct-2a and its biological role in B cell differentiation and general development.