An IgH Enhancer That Drives Transcription through Basic Helix-Loop-Helix and Oct Transcription Factor Binding Motifs FUNCTIONAL ANALYSIS OF THE E m 3 9 ENHANCER OF THE CATFISH*

, The transcriptional enhancer (E m 3 * ) of the IgH locus of the channel catfish, Ictalurus punctatus , shows strong B cell-specific activity and differs from the mammalian E m enhancer in both location and structure. It occurs between the m and d genes and contains numer-ous transcription factor binding sites, predominantly octamer and m E5 motifs of consensus and variant sequences. It lacks the classical m A- m E3(CBF)- m B core ar-ray of binding motifs seen within mammalian IgH E m enhancers. To determine the functionally important motifs, a series of mutant enhancers was created using sequence-targeted polymerase chain reaction. Whereas the mutation of consensus and variant octamer motifs (individually or in multiples) decreased enhancer function, mutation of a single consensus m E5 motif destroyed the function of this enhancer in mammalian plasmacytomas. Mutation of this consensus m E5 site, combined with mutations of certain octamer sites, destroyed function in catfish B cells. Experiments using artificial enhancers containing multimers of motifs or short regions of the native enhancer suggested that the minimal E m 3 * enhancer ( a ) contains a consensus m E5 site and two octamer sites, ( b ) is B cell-specific, and ( c ) is active across species. The dependence of an Ig

The transcriptional enhancer (E3) of the IgH locus of the channel catfish, Ictalurus punctatus, shows strong B cell-specific activity and differs from the mammalian E enhancer in both location and structure. It occurs between the and ␦ genes and contains numerous transcription factor binding sites, predominantly octamer and E5 motifs of consensus and variant sequences. It lacks the classical A-E3(CBF)-B core array of binding motifs seen within mammalian IgH E enhancers. To determine the functionally important motifs, a series of mutant enhancers was created using sequence-targeted polymerase chain reaction. Whereas the mutation of consensus and variant octamer motifs (individually or in multiples) decreased enhancer function, mutation of a single consensus E5 motif destroyed the function of this enhancer in mammalian plasmacytomas. Mutation of this consensus E5 site, combined with mutations of certain octamer sites, destroyed function in catfish B cells. Experiments using artificial enhancers containing multimers of motifs or short regions of the native enhancer suggested that the minimal E3 enhancer (a) contains a consensus E5 site and two octamer sites, (b) is B cell-specific, and (c) is active across species. The dependence of an Ig enhancer on sites that bind basic helix-loop-helix and Oct transcription factors has not previously been observed and confirms large differences in structure and function between fish and mammalian IgH enhancers.
Teleost (bony) fishes represent the evolutionary lineage that is most divergent from mammals yet still shares with them the organization of the IgH 1 locus. In both mammals and bony fish the V H , 1 D, J H , and C gene segments are in tandem arrays, also known as the translocon arrangement (1). Within the IgH locus of mammals, there are six described enhancers, required for the processes of V-D-J recombination, class switching, somatic hypermutation, and expression of the functionally rearranged locus (2)(3)(4). These enhancers are both internal to and 3Ј of the locus and are tightly regulated in function at different stages of B cell development (5,6). The E enhancer, located in the J H -C intron, is arguably the most important; not only is it required for expression of the rearranged IgH locus, but it also plays an important role in the processes of V H -D-J H recombination and somatic hypermutation and is not deleted by any of the recombination processes that occur during B cell development. The IgH locus of the channel catfish, the best-established model for the teleost immune system, is less complex than that of mammals; only two classes of immunoglobulin (IgM and IgD) are produced through alternative processing of the primary transcript from the IgH locus (7,8). Class switching by chromosomal recombination does not occur in teleost fish. The single identified enhancer (E3Ј) driving IgH transcription is located in the -␦ intergenic region (9). The structure and organization of E and E3Ј differ greatly. Mammalian E is ϳ500 bp in length and contains several DNA binding motifs (E1-5, E-box, A, B, CBF, Oct) each present only once (10 -12). The E3Ј enhancer region of the channel catfish is considerably larger, 1.8 kilobases in length, and it contains many repeats of octamer and E motifs of both consensus and variant sequences (9). Despite these differences E and E3Ј are both strong B cell-specific transcriptional activators that function in both mouse and catfish B cells (9,(13)(14)(15). Here we report the results of an analysis of E3Ј, defining the functional motifs essential for its activity and identifying a minimal enhancer.

Preparation of Plasmids for Mutagenesis and Transfection-CAT
reporter constructs to test enhancer function at a distal site were based upon those previously described (9,13). Briefly, mutated enhancers were inserted at the KpnI site of the pFVH/CAT/modified vector that contains the full promoter region of a goldfish V H gene (9). To test DNA binding motifs and enhancer fragments at a proximal (promoter) position, the p⌬56/CAT construct was used (16). This vector contains the minimal c-fos promoter, from Ϫ56 to ϩ109 bp relative to the transcription start site, and includes a minimal TATA box (16). All PCR-generated enhancer or promoter fragments were made using Taq DNA polymerase (Life Technologies, Inc.), with restriction sites incorporated in the primer sequences to facilitate cloning. Mutagenesis of the full enhancer was performed in a modified pUC19 (Life Technologies, Inc.) vector, pUC19⌬HinDIII. The HinDIII site of pUC19 was destroyed by digestion with HinDIII and Klenow fill-in to permit the screening of clones for mutations by the introduction of a HinDIII site. The 1782-bp enhancer fragment that contains full activity (ELF11), hereafter referred to as the full enhancer, was ligated into pUC19⌬HinDIII via the KpnI site of the polylinker and subjected to mutagenesis. After successful mutagenesis was confirmed by sequencing, the mutant enhancer was recloned into the KpnI site of pFVH/CAT/modified (9). Mutagenic primers were designed to replace the DNA binding motif of interest with a sequence of identical length containing a HinDIII site. The only exception was the combined mutation of octamers 10 and 11, in which 13 bp between the two motifs were deleted. To create multiple DNA binding motif mutations in a single construct, PCR mutagenesis was performed on enhancers in which sites had already been mutated. For PCR amplification of longer sequences (including complete vectors) the Expand TM long template PCR system (Roche Molecular Biochemicals) or GeneAmp® XL PCR kit (PerkinElmer Life Sciences) were utilized. The bp number refers to the position of the DNA binding motif in 12C (recombinant catfish genomic -phage clone, EMBL accession number X79482), and the HinDIII sites are underlined (9). Motifs were selected for mutation based on their degree of similarity to the mammalian consensus motifs and, in the case of octamer motifs, their ability to drive transcription (13). The five motifs that have previously been demonstrated to show the strongest transcriptional activity, i.e. octamers 5, 7, 10, 11, and 14 were selected for mutation. The remaining octamer motifs have been shown to be inactive or only weakly active in catfish B cells (13). Mutagenic primers (upper and lower) for use with the Expand TM and GeneAmp® XL PCR kits were as follows: octamers 10 and 11 at 10,695 and 10,716 bp, 5Ј-TTTAAGCTTGTAGCACTTCT-TCCTGTGCAGGTG-3Ј and 5Ј-TTTAAGCTTGCAGTGTTTTGCCTGTT-ATGTAAA-3Ј; octamer 7 at 10,217 bp, 5Ј-TTTAAGCTTAAGTTTGGT-CTGTTCTCTGTTTCTTC-3Ј and 5Ј-TTTAAGCTTTTGTAGTTTCTGT-ACTCATTTCACCT-3Ј; octamer 14 at 11,308 bp, 5Ј-TTTAAGCTTATA-TTCACTCACCAAAAAGGCAGT-3Ј and 5Ј-TTTAAGCTTTTACTGGCC-AATTATGTTTGACC-3Ј; E3 motif at 10,135 bp, 5Ј-ATAAAGCTTAA-ATGGCTACGTTTTTTGATCTA-3Ј and 5Ј-TTTAAGCTTCAGCAATAT-ATTCTCTTTGTTACT-3Ј; E5-1 and octamer 5 at 9,844 and 9,848 bp, 5Ј-TTTAAGCTTAATAATCAACACCTTGAAATTGTT-3Ј and 5Ј-TTTA-AGCTTAAGAAATATGTCTATCATATTCTA-3Ј; E5-2 (mammalian consensus) at 10,740 bp, 5Ј-TTTAAGCTTTCACTCCATCTCAGAGATG-CAG-3Ј and 5Ј-TTTAAGCTTCAGGAAGAAGTGCTACATTTACAT-3Ј; E5-3 at 10,896 bp, 5Ј-TTTAAGCTTAAACACGAAAAACTGT GAAAC-AAA-3Ј and 5Ј-TTTAAGCTTGGCTAGAGGAAGGAGTGCAGGACA-3Ј; E5-4 at 11,378 bp, 5Ј-TTTAAGCTTCAACTTATGACCT GATTGTCA-CAA-3Ј and 5Ј-TTTAAGCTTAGACAAGAGAGAAAATCATACATT-3Ј.
Multimers of motifs were created to test the ability of an individual motif to drive transcription at a proximal position relative to the CAT reporter gene. Oligonucleotides were designed to contain three direct repeats of a DNA binding motif centered 20 bp apart. The oligonucleotides (Oligonucleotide Synthesis Facility, Medical University of South Carolina, Charleston, South Carolina) were synthesized such that, after annealing, HinDIII overhangs would be created to facilitate cloning. Oligonucleotide pairs containing three copies of the binding motifs (bold, underlined) were as follows: E5-1 variant DNA binding motif, 5Ј-AGCTTGATACACATGCAGAACGCGTACTTACACATGCAGACT-CGAGTACT TACACATGCAGAGA-3Ј and 5Ј-AGCTTCTCTGCATGT- After annealing, the double-stranded DNA fragment was purified by electrophoresis on a 12% non-denaturing polyacrylamide gel followed by electroelution. Purified fragments were cloned into the HinDIII site of the p⌬56 c-fos/CAT construct. Clones were sequenced to confirm the presence and copy number of the motif.
Reporter constructs containing the full enhancer (E3Ј) or small regions (ϳ300 bp) of the full enhancer were generated by PCR and inserted into the p⌬56 c-fos/CAT vector at the HinDIII site in the same orientation as they appear in the full enhancer. Primers designed to amplify the full enhancer (E3Ј) with the addition of HinDIII ends were as follows: 5Ј-TTTAAGCTTACATCATCTAGCCTACCATAAGTA-3Ј and 5Ј-TTTAAGCTTCAGAGAAGACGGCACTACTATATG-3Ј. Primers designed to amplify region 1 (see Fig. 1) of the full enhancer, which contains the mammalian consensus octamer 5, variant E5-1 motif, and variant octamer 6, with the addition of HinDIII ends were as follows: 5Ј-TTTAAGCTTCAAAAAGCAAAAAGCACTCTTTAC-3Ј and 5Ј-TTTAAGCTTGCATCTCTGAGATGGAGTGAAAC-3Ј. Primers designed to amplify region 2 (see Fig. 1) of the full enhancer, which contains variant octamer 10, variant octamer 11, and the mammalian consensus E5-2 motif, with the addition of HinDIII ends were as follows: 5Ј-TTTAAGCTTAAGCCTTTTTAGAATATGATAGAC-3Ј and 5Ј-TTTAAGCTTGTATCTCTCTTTTTTATGGTGAAT-3Ј. Primers designed to amplify region 3 (see Fig. 1) of the full enhancer, which contains the octamer variant 13 motif, variant octamer 14, and variant E5-4 motif, with the addition of HinDIII ends were as follows: 5Ј-TTTAAGCTTACAGAAAATTAAGAGAAAGGATG-3Ј and 5Ј-TTTAAGCTTGTGACAATCAGGTCATAAGTTGAAC-3Ј.
PCR-amplified regions from the full enhancer were purified by electrophoresis on 0.8% agarose gels and extracted using the NucleoTrap gel extraction kit (Clontech). Clones were sequenced to confirm enhancer fragment presence, sequence, and copy number. Plasmids for mutagenesis and transfection were prepared by transformation of chemically competent XL1 Blue MRFЈ cells (Stratagene), grown under ampicillin selection. Subsequent purification of DNA was accomplished using QiaFilter Plasmid maxi and mega kits (Qiagen).
Transfection of Cell Lines Using Reporter Constructs-The catfish B lymphoblastoid cell line 1B10 and the catfish T cell line G14D, derived from mitogenic stimulation of peripheral blood leukocytes, have been described previously (17). These cell lines were maintained in AL-5 medium (50% AIM V (Life Technologies, Inc.) and 50% Leibovitz-L15 medium (Life Technologies, Inc.) adjusted to catfish isotonicity by diluting 9:1 with water). AL-5 was supplemented with heat-inactivated catfish serum to a final concentration of 5% (17), 100 g/ml streptomycin (ICN), 100 units/ml penicillin (ICN), 50 M ␤-mercaptoethanol (Sigma), and 0.1% sodium bicarbonate (Sigma). 1B10 and G14D cells were cultured at 27°C in a 5% CO 2 atmosphere. The murine plasmacytoma cell lines J558L and S194 were maintained in RPMI 1640 medium (Life Technologies, Inc.) with a final concentration of 5% fetal calf serum (Life Technologies, Inc.). Mouse cell lines were grown at 37°C in a 5% CO 2 atmosphere. All transfections were done using a BTX600 electroporator and 2-mm gap cuvettes (Biotechnologies & Experimental Research, Inc.). Cells were harvested during logarithmic growth, washed in serum-free RPMI 1640 (at the appropriate tonicity), and resuspended again in serum-free RPMI 1640 for transfection. Directly prior to transfection, 180 l of cells (8 ϫ 10 6 cells of 1B10, 5 ϫ 10 6 cells of G14D, or 4 ϫ 10 6 cells of J558L and S194) were mixed with 20 l of a DNA mixture. DNA solutions contained 2.3-8.9 pmol (7-21 g) of DNA depending on the cell line transfected. Three to nine g of a second DNA construct, pRSV/Luc (18), expressing firefly luciferase driven by the Rous sarcoma virus promoter, were cotransfected to permit assessment of transfection efficiencies. Optimal electroporation values were determined as previously described (9). Briefly, 1B10 cells harvested at a density of 3.5-4.0 ϫ 10 6 cells/ml were electroporated at 210 V, 1100 microfarads, and 48 ohms. G14D cells harvested at a density of 2.6 -3.0 ϫ 10 6 cells/ml were electroporated at 190 V, 1100 microfarads, and 48 ohms. J558L and S194 cells, harvested at a density of 8.0 ϫ 10 5 cells/ml, were electroporated at 130 or 168 V, respectively, 1100 microfarads, and 48 ohms.
Reporter Assays-Expression of reporter constructs was measured 30 -48 h (depending upon the cell line) after electroporation. CAT and luciferase activities were measured using cell extracts prepared with Promega reporter lysis buffer (Promega). Cells were harvested by centrifugation, washed in phosphate-buffered saline, and resuspended in 100 l of 1ϫ reporter lysis buffer (Promega). After a 15-min incubation at room temperature, cell lysate was cleared by centrifugation and aliquoted for immediate use. Luciferase was measured using the Promega luciferase assay system (Promega) and a TD-20/20 luminometer (Turner Instruments) set to standard sensitivity. Values were corrected to exclude background activity. CAT activity was measured using CAT assay grade D-threo-[dichloroacetyl-1,2-14 C]-chloramphenicol (PerkinElmer Life Sciences) and n-butyryl coenzyme A (Sigma) (9). CAT values were corrected for background activity and for transfection efficiency by normalizing with luciferase values.
Sequencing and Analysis-Sequencing of PCR-mutated enhancers, E3Ј fragment multimers, and multimers of DNA binding was accomplished using the Circumvent DNA sequencing kit (New England Biolabs) or the sequencing facility of the Medical University of South Carolina.

Functional Significance of a Consensus E5
Motif-Although it has previously been demonstrated that transcription of the catfish IgH locus is driven by the E3Ј enhancer (9), the motifs and transcription factors through which it functions are unknown. Indeed, the E3Ј enhancer contains significant numbers of motifs that have the potential to contribute to function. As illustrated in Fig. 1, these include 10 consensus and variant octamer motifs, of which eight have been shown to be transcriptionally active (9), a single variant E3 site, and four E5 sites of consensus or variant sequence. To determine the degree to which each of these sites might contribute to enhancer function, targeted mutations were made in the enhancer. In an initial experiment each E5 site and the E3 site were mutated individually, whereas the consensus octamer 5 (ATG-CAAAT) was mutated in conjunction with the E5-1 site that it overlaps, and the closely spaced dimer of octamers (octamers 10 and 11) of identical sequence (ATGtAAAT) were mutated together. The function of the mutated enhancers was tested in mouse plasmacytomas S194 ( Fig. 2A) and J558L (Fig. 2B) and the catfish B cell line 1B10 (Fig. 2C). The effects of mutating specific motifs varied between the mammalian and catfish cell lines. In the murine cell lines S194 and J558L, mutation of the E5-3 and E5-4 sites was either without effect or resulted in a small increase in activity. Mutation of the E3 motif or the E5-1 site (which overlaps the consensus octamer 5) produced a slight decrease in activity in mammalian cells. The mutation of the octamer pair (octamers 10 and 11) reduced enhancer function by 30 -40%, but the most striking effect was seen with mutation of the mammalian consensus E5 site (number 2), which essentially eliminated function in the enhancer (Fig. 2, A  and B). The effects of these mutations were different when tested in catfish B cells (Fig. 2C). All of the mutations reduced activity to some extent, with loss of the E5-4 site or E5-1 site and overlapping consensus octamer 5 having the least effect (ϳ20% reduction in activity) and the consensus E5-2 having the greatest effect, with a 55% reduction in activity. Thus, whereas the E5-2 motif is essential to the function of the enhancer in murine plasmacytomas, in the homologous (catfish) B cells it can be inferred to cooperate with additional transcription factor binding motifs.
Because the effect of mutating the overlapping consensus octamer (ATGCAAAT) and E5-1 (variant) site reduced enhancer function in all cell lines tested (Fig. 2), the role that each of the two overlapping sites played, individually, in the native enhancer was examined. Site-directed mutagenesis was used to destroy one site while leaving the other intact. Mutation of the variant E5-1 site alone increased enhancer activity by 30%, whereas mutation of the consensus octamer motif decreased transcription by 40% (Fig. 2D), showing that the functional site is the consensus octamer and that the presence of the overlapping E5-1 site reduces its activity.
Mutations of the Consensus E5 Site in Combination with Other Motifs-Constructs in which the consensus E5-2 motif was mutated in combination with octamer motifs were tested in catfish B cells (Fig. 3A). A combination of the E5-2 mutation with mutations of any one of five octamer sites further reduced the residual activity of the enhancer. The degree of this reduction depended on the particular octamer(s). Combining the mutation of E5-2 with the octamer dimer 10 and 11 essentially eliminated all function of the enhancer in catfish B cells (Fig. 3A). When the E5-2 motif mutation was combined with that of octamers 7 or 14, the reductions in activity, in catfish B cells, were 86 and 75%, respectively. Mutation of the consensus octamer 5, in combination with the consensus E5-2 motif, also showed a greater than 90% reduction in activity (Fig. 3A). When tested in mammalian J558L cells (Fig. 3B), enhancers with combined mutations of the consensus E5-2 site and the octamer sites showed no activity, as predicted from the effects of deleting the E5-2 site alone (Fig. 2B). From these results (Figs. 2 and 3) it appears that factors bound to the E5-2 motif and as many as five octamer sites act synergistically in the E3Ј enhancer in catfish B cells.
Effect of Mutating Multiple Octamer Motifs-With the consensus E5-2 motif left intact, mutation of three octamer sites (the variant octamer dimer composed of octamer motifs 10 and 11, along with the consensus octamer 5, which also overlaps the E5-1 site) reduced function in catfish B cells by only around 35% (Fig. 3A), a similar effect to that seen with mutation of the octamer pair (octamers 10 and 11) alone. However, mutation of this particular combination of three octamer motifs reduced enhancer function by Ͼ70% in murine J558L plasmacytoma cells (Fig. 3C), indicating further differences in the function of this enhancer in fish and mouse B cells.
Function of Arrays of Individual Motifs-Because mutations of several octamer motifs and the consensus E5-2 motif strongly affect native enhancer function, the ability of these individual motifs to drive transcription of a reporter construct was tested. Trimers of E5 and E3 failed to function as promoters in both catfish and murine cell lines (Fig. 4, A and  B). In contrast, trimers of octamer sequences, either consensus 5 or variant 10 and 11, increased the rate of transcription in both catfish and mouse B cell lines 14 -17-fold. These results indicate that in catfish B cells, the function of E motifs is dependent upon the presence of additional transcription factors bound to other sites, whereas octamer motifs are functional by themselves when associated with a minimal promoter.
A Minimal Enhancer within E3Ј-Inspection of the organization of the E3Ј enhancer (Fig. 1) reveals that it contains a repeating substructure: a E5 motif associated with two octamer motifs. This occurs three times (regions 1-3; Fig. 1). The results of the motif mutation experiments suggest that the second of these regions (containing the E5-2 consensus and octamers 10 and 11) may be a minimal functional unit within the enhancer, and experiments were undertaken to test this and to determine whether the other two regions (1 and 3; Fig.  1) were transcriptionally active. Accordingly, fragments ϳ300 bp in length were PCR-amplified from the three regions of the native enhancer identified in Fig. 1 and ligated into reporter constructs. In catfish B cells the full enhancer increased reporter activity 6-fold, whereas enhancer region 2 (containing the E5-2 and octamer 10 and 11 motifs) increased reporter gene expression almost 5-fold. Other regions failed to activate transcription above control levels (Fig. 5A). To test the contribution of the consensus E5-2 site to the function of this fragment, this motif was mutated. This mutation reduced the activity of the enhancer fragment, as was also observed upon mutation of this motif within the full enhancer (Fig. 2C). When tested in murine J558L cells (Fig. 5B), the full enhancer increased reporter activity almost 15-fold, whereas region 2 alone increased reporter gene expression almost 10-fold. Other regions failed to activate transcription above control levels. Mutation of the mammalian consensus E5-2 site from region 2 reduced activity by more than 60% in J558L cells (Fig. 5B). To determine whether the activity of region 2 was B cell-specific, its ability to activate transcription in catfish T cells was evaluated (Fig. 5C). Both the full enhancer and all three regions of the enhancer failed to drive transcription of the reporter gene in these T cells. Thus, region 2 of the enhancer shows the same B cell specificity as the full enhancer (9, 13-15).  Fig. 1). Normalized values of CAT activity are shown as the mean Ϯ S.E., relative to the transcriptional activity of the full enhancer control (set to 100%). The number of replicate transfections (n) for each construct is shown above the histogram.

FIG. 3. Effects of multiple mutations on the transcriptional activity of the E3 enhancer upon transfection into 1B10 cells (A) and J558L cells (B).
Constructs based on the pFVH/CAT/modified vector (9) contained no enhancer, the full enhancer E3Ј, or the full enhancer with the identified mutations. Expression was assayed after 30 -48 h depending on the cell line tested. M, mock transfection; V, vector alone, and F, full enhancer with no mutations (set to 100%). Mutated motifs were as follows: E5-1, E5-1; E5-2, E5-2; O7, octamer 7; O10, octamer 10; O11, octamer 11; and O14, octamer 14 (see Fig. 1). Normalized values of CAT activity are shown as the mean Ϯ S.E., relative to the transcriptional activity of the full enhancer control (set to 100%). The number of replicate transfections (n) for each construct is shown above the histogram.

DISCUSSION
The results of this study demonstrate that the function of E3Ј, the enhancer driving expression of the IgH locus of the catfish, depends on transcription factors binding to a single consensus E5 motif and to a number of octamer motifs, only one of which is of consensus sequence. This contrasts with the situation in the mammalian IgH locus, where, although E5 and octamer sites play contributory roles in the enhancers, a number of other motifs play major, indeed essential, roles (19,20). This is particularly clear in the case of the E enhancer, found in the mammalian J H -C intron, where the core of this enhancer has been shown to consist of a small region containing three motifs: a A site, a E3 or a CBF site, and a B site (21). The factors binding to these sites (Ets-1 binding to A, PU.1 binding to B, and TFE 3 or core binding factor binding to the E3/CBF site) interact in a manner that requires precise spacing and orientation of the binding sites (22,23). Although the involvement of additional sites in the catfish E3Ј enhancer cannot yet be completely discounted, there is no evidence that it utilizes ETS family transcription factors; nor does it appear to require TFE 3 binding sites (9,13).
Of the multiple E5 sites identified in the catfish enhancer, only the site of exact mammalian consensus sequence was shown to exert a major positive influence on transcription. In fact, the E5-1 site, which differs by one base from consensus and which overlaps the consensus octamer site, appeared to have a negative influence, by depressing the contribution to function made by the overlapping octamer motif (Fig. 2). In contrast, the octamer sites within E3Ј that differed from consensus sequence by single base changes in the 5Јhalf of the site (i.e. ATGC, the binding site for the POU-specific domain of Oct transcription factors) retained function. This is consistent with the previously described ability of these variant octamer sequences both to bind catfish Oct2 and to drive transcription in B cells when placed at the promoter site of a reporter construct (13)(14)(15). One unusual aspect of the catfish E3Ј enhancer  Fig.  1) was cloned into the p⌬56 c-fos/CAT vector. Empty vector, vector containing the full enhancer (E3Ј), or vector containing a single copy of the region was transfected into 1B10 cells (A), J558L cells (B), and G14D cells (C), and expression was assayed 30 -48 h later depending on the cell line tested. M, mock transfection; V, vector alone (set to 1); and F, full enhancer with no mutations. The pCMV/CAT vector, used as a positive control in the transfection of G14D cells, gave enhanced expression 3168 Ϯ 764-fold (seven replicates) above the level seen with the vector alone. 1, region 1; 1-O5, region 1 in which the consensus octamer 5 was mutated; 2, region 2; 2-E5-2, region 2 in which the consensus E5-2 site was mutated; and 3, region 3 (see Fig. 1). Normalized values of CAT activity are shown as the mean Ϯ S.E., relative to the transcriptional activity of the vector (set to 1). The number of replicate transfections (n) for each construct is shown above the histogram. revealed in the present study is the potential of multiple octamer sites to act cooperatively with a single consensus E5 site. Although two of the octamers (octamers 10 and 11) are spaced two full turns of the DNA helix apart, which might facilitate the interaction of bound transcription factors, other octamers that are shown to have function in the enhancer are distributed over 1.4 kilobases. The consensus E5 site is only 24 bp away from the closest functional octamer, whereas the other octamers that function are as far away from the E5 site as 892 bp. This cooperative action of transcription factors bound to widely separated sites is consistent with the involvement of coactivators, the nature of which is not known. A coactivator, OCA-B2, has been postulated to explain the action of Oct2 bound to octamer sites in the mammalian E enhancer (24 -26). Octamer-binding transcription factors (in particular two isoforms of Oct2) have been cloned in the catfish and show clear similarity with their mammalian homologues, including strong sequence conservation of the DNA binding site, the presence of N-and C-terminal domains that are probably transcriptional activators, and the ability to act synergistically with the OCA-B (BOB-1) coactivator (14,15). They have a restricted tissue distribution similar to that described for mammalian Oct2 (14). The factor(s) binding to the catfish E5 site are not yet defined. In mammals the USF proteins, the E2A-encoded basic helixloop-helix motif proteins E12 and E47, and the zinc finger protein ZEB bind to both E5 and E2 sites and are implicated in regulation of the E enhancer in B cells (27)(28)(29). Enhancer function dependent upon Oct transcription factors and factors bound to a E5 site has not been described in mammalian systems. The definition of a B cell-specific functional subunit of the catfish E3Ј enhancer, which contains the consensus E5 motif and two functional octamer motifs, and in which both E5 and octamer sites are functional (Fig. 1), will permit the study of potential interactions between octamer-and E5binding transcription factors in a simpler system than is afforded by the full, native enhancer.
The comparison of enhancers between the IgH loci of catfish and mammals reveals substantial differences that contrast with the overall conservation of exon structure observed in vertebrate Ig genes. The enhancers differ in position, size, composition, and mechanism of action. This probably reflects the complexity of the mammalian locus, which undergoes class switching by chromosomal recombination (which does not occur in fish), and a very complex program of B cell development (6). In light of their substantial differences, it is surprising that the catfish E3Ј and mouse E enhancers both function effectively and B cell-specifically when transfected across species (9,13). This observation would be explicable if B lineage cells of all vertebrates possessed a large complement of transcription factors, of both tissue-specific and ubiquitous expression. Different Ig enhancers would then be able to use different combinations of these transcription factors to drive B cell-specific transcription, with the details of the particular mechanisms reflecting specific needs for fine control of B cell development and differentiation.