Characterization of Human TCR Vb Gene Promoter ROLE OF THE DODECAMER MOTIF IN PROMOTER ACTIVITY*

During T-lymphocyte development, the T-cell antigen receptor (TCR) gene expression is controlled by its promoter and enhancer elements and regulated in tissueand development stage-specific manner. To uncover the promoter function and to define positive and negative regulatory elements in TCR gene promoters, the promoter activities from 13 human TCR Vb genes were determined by the transient transfection system and luciferase reporter assay. Although most of the TCR Vb gene promoters that we tested are inactive by themselves, some promoters were found to be constitutively strong. Among them, Vb6.7 is the strongest. 5*-Deletion and fragmentation experiments have narrowed the full promoter activity of Vb6.7 to a fragment of 147 base pairs immediately 5* to the transcription initiation site. A decanucleotide motif with the consensus sequence AGTGAYRTCA has been found to be conserved in most TCR Vb gene promoters. There are three such decamer motifs in the promoter region of Vb6.7, and the contribution of each such motif to the promoter activity has been examined. Further site-directed mutagenesis analyses showed that: 1) when two Ts in the decamer were mutated, the promoter activity was totally abolished; 2) when two additional nucleotides 3* to the end of decamer were mutated, the promoter activity was decreased to two-thirds of the full level; and 3) when the element with the sequence AGTGATGTCACT was inserted into other promoters, the original weak promoters become very strong. Taken together, our data suggest that the positive regulatory element in Vb6.7 should be considered a dodecamer rather than a decamer and that it confers strong basal transcriptional activity on TCR Vb genes.

During T-lymphocyte development, the T-cell antigen receptor (TCR) gene expression is controlled by its promoter and enhancer elements and regulated in tissueand development stage-specific manner. To uncover the promoter function and to define positive and negative regulatory elements in TCR gene promoters, the promoter activities from 13 human TCR V␤ genes were determined by the transient transfection system and luciferase reporter assay. Although most of the TCR V␤ gene promoters that we tested are inactive by themselves, some promoters were found to be constitutively strong. Among them, V␤6.7 is the strongest. 5-Deletion and fragmentation experiments have narrowed the full promoter activity of V␤6.7 to a fragment of 147 base pairs immediately 5 to the transcription initiation site. A decanucleotide motif with the consensus sequence AGT-GAYRTCA has been found to be conserved in most TCR V␤ gene promoters. There are three such decamer motifs in the promoter region of V␤6.7, and the contribution of each such motif to the promoter activity has been examined. Further site-directed mutagenesis analyses showed that: 1) when two Ts in the decamer were mutated, the promoter activity was totally abolished; 2) when two additional nucleotides 3 to the end of decamer were mutated, the promoter activity was decreased to two-thirds of the full level; and 3) when the element with the sequence AGTGATGTCACT was inserted into other promoters, the original weak promoters become very strong. Taken together, our data suggest that the positive regulatory element in V␤6.7 should be considered a dodecamer rather than a decamer and that it confers strong basal transcriptional activity on TCR V␤ genes.
T-lymphocyte development is similar to that of B-cell maturation in that the germ line variable (V), 1 diversity (D), and joining (J) gene segments of the T-cell antigen receptor (TCR) are somatically recombined to form a V-D-J (for ␤-chain) or V-J (for ␣-chain) exon encoding the variable domain portion of the receptor. Studies in B-cells support the idea that to achieve the recombination, the rearranging Ig gene segments must be transcriptionally activated (1)(2)(3)(4). If this prerequisite also applies to TCR genes, then the promoter elements that control transcrip-tional activity of TCR V genes will play a pivotal role in rearrangement. Specific transcriptional factors may bind to distinct DNA sequence motifs in the promoter region and control transcription. These factors can also affect the accessibility of germ line loci to the recombinational machinery, thus regulating gene expression in a tissue-or developmental stage-specific way.
A 10-base pair decamer sequence motif, AGTGAYRTCA, was found to be conserved in the promoter regions of most murine and human TCR V␤ genes (5)(6)(7). Elimination of this decamer motif in the murine V␤8.3 promoter reduced transcriptional activity (8). Binding of thymic factors to this decamer motif was found to be developmentally regulated, and no decamer binding activity was detected in nuclear extracts prepared from thymuses of severe combined immunodeficiency mice, suggesting that the decamer motif plays an important role in the connection between murine TCR V gene transcription and rearrangement (9).
The most highly conserved portion in the decamer motif is an inverted repeat with the sequence TGA-TCA. This palindromic feature links the decamer to other regulatory elements, such as the c-AMP response element, TGACGTCA, and the AP-1 binding site, TGACTCA (10,11). We have observed some discrete differences in the location and the consensus sequence of the decamer motif between different human TCR V␤ subfamilies (7). Little is known regarding the role of the decamer motif in promoter function or whether differences in the decamer motif may affect promoter activity.
In this study, we have characterized 14 promoters of human TCR V␤ genes, with a specific emphasis on V␤6.7. By using a transient transfection system, we have shown that human TCR V␤ promoters vary in terms of their strength. Whereas most promoters tested are weak, the promoters of the V␤6 subfamily, particularly V␤6.7, are constitutively strong. Sequence analyses have revealed three decamer-like motifs in the promoter region of V␤6.7, and each of them contributes differently to promoter activity. The one located most 5Ј to the transcription initiation site has essentially no effect on promoter activity. In contrast, the one located proximal to the initiation site forms the major contribution to promoter strength. A 5Ј deletion study has shown that a fragment of 147 bp immediately 5Ј upstream of the transcription initiation site can constitutively drive the reporter luciferase gene expression. Further shortening of this fragment results in a decrease in reporter gene expression. Therefore, this fragment represents the minimal promoter element of V␤6.7. Inside this fragment there are a CACCC motif, a TATA box, and two decamer-like motifs. Further site-directed mutagenesis analyses have shown that a 12-nucleotide sequence, AGTGATGTCACT, is responsible for the high promoter activity. Therefore, the regulatory element in the promoter of V␤6.7 should be considered a dodecamer rather than a decamer. The possible contributions to V␤6.7 promoter activity from additional transcription factor binding sites 3Ј behind the dodecamer are also discussed.

MATERIALS AND METHODS
Cell Culture and Total Cellular RNA Isolation-Human peripheral blood T-cells were cultured in RPMI 1640 medium with 10% fetal calf serum. CD8 ϩ T-cells were depleted following treatment with an anti-CD8 monoclonal antibody (OKT 8) and anti-mouse Ig-coated magnetic beads (Dynal, Great Neck, NY). CD4 ϩ T-cells were activated with anti-V␤6.7-specific monoclonal antibody OT145 (12). After a second retriggering, more than 80% of T-cells were V␤6.7-positive. Cells were washed once with phosphate-buffered saline, and the pellet was spun down. Total cellular RNA was isolated using an acidified guanidinium/ phenol/chloroform isolation kit (RNazol, TEL-TEST, Friendswood, TX) and used to identify the transcription initiation sites by RNase protection and primer extension methods.
RNase Protection and Primer Extension-To prepare an antisense RNA probe, a double-stranded DNA fragment spanning the first and the second exons plus the 5Ј-untranslated portion of V␤6.7 was subcloned into a T/A vector that contains a T7 promoter. The template plasmid DNA was digested upstream at an unique XhoI site. The antisense RNA probe was synthesized with a Riboprobe-T7 system (Promega, Madison, WI) in the presence of T7 RNA polymerase and [␣-32 P]CTP. The labeled RNA probe was hybridized with total cellular RNA isolated from V␤6.7 ϩ T-cells. The protected products were purified once with phenol/chloroform extraction, denatured, and loaded on a 6% polyacrylamide, 7 M urea sequencing gel. The dideoxyl-terminate sequencing reaction product from a plasmid DNA with a known sequence was loaded alongside as a size marker. After electrophoresis, the gel was dried and exposed to x-ray film at Ϫ70°C for 12-48 h.
In primer extension experiments, an antisense oligonucleotide primer with the sequence 5Ј-TGTGATCTGCCCCCAGGA-3Ј was synthesized. The primer was end-labeled with [␥-32 P]ATP, and polynucleotide kinase following the protocol in TaqTrack sequencing system (Promega). The end-labeled primer was annealed to 5 g of total cellular RNA isolated from V␤6.7 ϩ cells at 65°C for 10 min and then extended at 56°C for 1 h in the presence of reverse transcriptase, dNTPs, and RNase inhibitor (cDNA Cycle Kit, Invitrogen, San Diego, CA). The cellular RNA isolated from a mouse thymoma cell line BW5147, which contains no human TCR V␤ genes, was used in a parallel extension reaction as negative control. The extension products were denatured and loaded on sequencing gel with a size marker as described in the RNase protection assay.
Preparation of Constructs and Site-directed Mutagenesis-The 5Јflanking sequences were PCR-amplified from either genomic DNA or phage DNA clones. The sequences of 5Ј sense and 3Ј antisense oligonucleotide primers used for 14 human TCR V␤ genes were listed in Table  I. These PCR amplified fragments were ϳ500 bp and were inserted 5Ј to the reporter luciferase gene segment in pGL-2/Basic vector (Promega). The sequences and orientations of these fragments inside pGL-2 were confirmed by sequencing. The plasmid DNAs used for transfection were purified by cesium chloride banding. The site-directed mutagenesis was achieved using the PCR-overlapping method (13).
Transient Transfection-Human Jurkat T-cell line cells were grown to confluence in RPMI 1640 medium plus 10% fetal bovine serum and other supplements. Cells were transfected with plasmid DNA contain-ing promoter-reporter constructs (described above) by a modified DEAE-dextran electroporation method. Briefly, cells were washed twice with serum-free RPMI 1640 medium and resuspended in the same medium at a concentration of 17 ϫ 10 6 /ml. 600 l of cell suspension was placed in a 0.45-cm electroporation cuvette (Gene Pulser, Bio-Rad) followed by adding 180 l of DEAE-dextran (100 g/ml) and 10 g of promoter-reporter plasmid DNA. To evaluate the transfection efficiency, 5 g of plasmid DNA containing a ␤-galactosidase gene segment was co-transfected into the cells. The electroporation was carried out at a capacitance of 960 microfarads and 250 V. After electroporation, the cells remained at room temperature for 10 min and then were resuspended in 10 ml of complete RPMI 1640 medium, placed in wells (3 ml/well) of a 12-well plate, and incubated for 48 -72 h at 37°C in 5% CO 2 .
Luciferase Assay-5ϫ luc stock buffer, which contains 125 mM glycyl glycine, 75 mM MgSO 4 , and 20 mM EGTA, was premade and used later in preparation of lysis solution, assay solution, and luciferin substrate solution. The transfected cells were washed twice with 1ϫ phosphatebuffered saline, pelleted, and lysed with 500 l of lysis solution (1ϫ luc buffer, 1% Triton X-100, 1 mM of dithiothreitol added immediately prior to use). 100 l of cell lysate was added to a test tube that contained 500 l of assay solution (1ϫ luc buffer, 1 mM KHPO 4 , 1 mM dithiothreitol, and 2 mM of ATP). The tube was loaded into the measurement chamber, and 100 l of luciferin substrate solution (1ϫ luc buffer, 0.4 mM Dluciferin-sodium salt, and 2 mM dithiothreitol) was injected automatically. The relative light units (RLUs) were measured on a Monolight 2001 Luminometer (Analytical Luminescence Laboratory, San Diego, CA). The background RLUs from cells transfected with vector pGL-2 only were measured in each experiment. The luciferase assay data were calibrated with ␤-galactosidase activity and are presented as averages in triplicate measurements.
␤-Galactosidase Assay-The ␤-galactosidase activity was measured by using a Galacto-Light Plus Chemiluminescent Reporter assay kit following the manufacturer's instructions (Tropix, Bedford, MA).
Computer Analysis of 5Ј-Flanking Region Sequences-Putative regulatory elements in the 5Ј-flanking region of human TCR V␤ genes were identified by searching the Genetics Computer Group (version 8.1, Madison, WI) transcription factor binding sites data base (TFSITE) through Rockefeller University Computer Services. The decamer-like motifs were screened by the FINDPATTERNS program.

Decamer Motif and TCR V␤ Promoter
Activities-To determine the role of decamer sequence motifs in promoter activity, we have searched for this regulatory element in promoter regions among 54 functional TCR V␤ genes based on their genomic DNA sequences (Hood et al., GenBank accession no. L36092). The consensus sequence for the decamer is AGTGAR-YTCA. However, we observed some discrete differences in different TCR V gene subfamilies, such as AGTGATGTCA in V␤6, AGTGACATCA in V␤5, and TGANNNNTCA in V␤13 subfamilies (7). Therefore, we have screened two general patterns of the decamer motif, NNTGANNTCA and TGANNNNTCA, in the promoter regions ϳ500 bp 5Ј to the translation initiation  ggctcgagtatgtcagctagttcaa ggaagctttgccagatcaggg a The same 5Ј sense and 3Ј antisense primers were used to amplify 5Ј-flanking sequences for V␤ 6.1, 6.3, 6.7a, and 6.7b.
codon ATG, using the FINDPATTERNS computer program. Among 54 functional human TCR V␤ gene families, 42 V␤ promoters contain these general patterns (Table II). 22 V␤s each contain one NNTGANNTCA element. 8 V␤s each contain two such elements, and the remaining 5 V␤s each contain three such elements. Of these 5 V␤s, 4 come from the V␤6 subfamily. They are V␤6.3, 6.7, 6.11, and 6.14. All V␤13 subfamilies, except V␤13.1, contain the TGANNNNTCA pattern. We randomly selected 13 TCR V␤ genes and compared their promoter activity using a transient transfection system. The 5Ј-flanking sequences of human TCR V␤6.3 and 6.7 were PCR amplified from phage genomic DNA clones 4-1 and 5-2 (7). The promoter segments for other V␤ genes were PCR amplified from genomic DNA. The promoter fragments were ligated into a luciferase reporter vector pGL-2/Basic. These constructs were then used to transfect a human Jurkat T-cell line. The promoter activities were measured as luciferase activities in transfectants. To eliminate the influence of transfection efficiency, all data from luciferase assays were calibrated with ␤-galactosidase activities and are presented as averages of triplicate measurements. Fig. 1 shows the reporter luciferase activities, expressed as RLUs, for 13 human TCRV␤ promoters. Human TCR V␤ promoters may be divided into three groups in terms of their constitutive activities. The weak promoter group includes V␤3.1, 5.7, 9.1, 12.2, 14.1, and 21.4. The promoters with moderate constitutive activity are those from V␤5.1, 5.2, 8.1, and 17.1. The strong promoters are those from V␤6 subfamilies, V␤6.1, 6.3, and 6.7. Of them, the promoter of V␤6.7 is the strongest.
Allelic Variations in Promoter Region of V␤6.7-We previously identified two alleles of the human TCR V␤6.7 gene, V␤6.7a and V␤6.7b (14). They differ at two amino acid positions inside the coding region: V␤6.7a encodes Ser 38 and Gly 72 , whereas V␤6.7b encodes Arg 38 and Glu 72 . An allele-specific monoclonal antibody, OT145, can recognize the product of 6.7a but not 6.7b (15). Later, Vissinga et al. (16) found that the peripheral expression of these two alleles in heterozygous individuals was skewed, indicating that an allelic polymorphism in the coding region can have a significant impact on gene expression in the peripheral repertoire. However, other polymorphisms in the TCR ␤ locus, such as those in the promoter region, may also affect gene expression. To address this possibility, we analyzed the 5Ј-flanking sequences of V␤6.7 alleles a and b. First, a 463-bp sequence upstream of V␤6.7a in phage DNA clone 5-2 (7) was obtained. The promoter segment of V␤6.7 allele b was then PCR amplified from the genomic DNA of a b/b homozygous individual. Sequence analyses revealed three point mutations between alleles a and b within the 463-bp 5Ј-flanking region that were confirmed in a total of 24 plasmid clones derived from three V␤6.7 a/a and three V␤6.7 b/b homozygous individuals (Fig. 2). These mutations represent allelic variations in the promoter region of V␤6.7. We expected that no significant difference in promoter activities would be detected between V␤6.7 alleles a and b because no destruction of any major regulatory elements was caused by these mutations. Our expectation was confirmed by luciferase reporter assay (data not shown).
Transcription Initiation Site of V␤6.7-To determine the transcription initiation site of V␤6.7, primer extension analysis was performed. An antisense oligonucleotide primer, 5Ј-TGT-GATCTGCCCCCAGGA-3Ј, was specifically designed. In this primer, 12 nucleotides match the 3Ј-end sequence of the first exon, and the remaining 6 nucleotides match the 5Ј-beginning of the second exon of V␤6.7 (see the coding sequence of V␤6.7 in Fig. 2). Only cDNA, but not the genomic DNA, can anneal to this primer. In Fig. 3, lanes C, T, A, and G were sequencing products of a plasmid DNA with known sequence, shown here as the size marker. Lane 1 was the primer extension product incubated with RNA isolated from V␤6.7 ϩ T-cells. Lane 2 was the primer extension product incubated with RNA isolated from mouse thymoma cell line BW5147, included here as negative control. Two intense bands were observed in lane 1. The higher one is 63 bp, and the lower one 60 bp. We assumed that the 63-bp band represents the full length extension product. This product aligns with an A 26 bp 5Ј to the ATG codon. Given that adenine is a favored base for eukaryotic transcription initiation (17), we defined it as the initiation site for human V␤6.7; it is marked with an asterisk in Fig. 2. This transcription initiation site was further confirmed by RNase protection assay (data not shown) and is consistent with the findings in most murine TCR V␤ genes, in which the cap sites fall into a range of 19 -40 bp 5Ј to the ATG codon (6). The lower band of 60 bp in Fig. 3, lane 1, may represent the partial extension product.
5Ј-Deletion and Fragmentation Analyses of V␤6.7-Three decamer-like motifs were identified in the promoter region of V␤6.7. They were located at positions Ϫ84, Ϫ111, and Ϫ352 and are designated D1, D2, and D3 (Fig. 2). To evaluate the contribution of each decamer or other transcription factor binding elements to the promoter activity, the 5Ј-flanking sequence of V␤6.7 was fragmented. Each fragment was ligated into the luciferase reporter gene construct (Fig. 4, left panel), and then RLUs were measured (Fig. 4, right panel). Fragment I, from Ϫ1 to Ϫ463, representing the entire 5Ј-flanking sequence of V␤6.7, can drive the luciferase activity to ϳ700-fold higher than the background of pGL-2/Basic vector only. Fragment II, from Ϫ463 to Ϫ336, which contains one decamer-like motif D3, had essentially no effect on luciferase activity. The same result was obtained with fragment III (Ϫ336 to Ϫ127). In contrast, when fragment IV (Ϫ147 to Ϫ1) was used, the RLUs were even higher than the level induced by the entire 5Ј-sequence (fragment I). 5Ј-Deletion further created two small fragments: fragment V (Ϫ147 to Ϫ80), which contains D2 only, and fragment VI (Ϫ89 to Ϫ1), which contains D1 only. As shown in Fig. 4, right panel, the RLUs for fragment V were low, and the RLUs for fragment VI were high. However, fragment VI was unable to reach the same level of activity as that induced by fragment IV. Therefore, the 147-bp of fragment IV represents the basic promoter for V␤6.7.
No T-cell Lineage Specificity for Promoter Element in V␤6.7-To determine whether the constitutive activity of the promoter element in V␤6.7 was specific to the T-cell lineage, several cell lines, including human T-cells (Jurkat), Burkitt's lymphoma cells (Ramos), fibroblast cells (cos 7), mouse thymoma cells (BW5147), and human embryonic kidney cells (293) were used. The luciferase reporter constructs containing TCR V␤6.7 promoter segment and a nonspecific Rous sarcoma virus promoter were used to transfect these different cell lines. When compared with the nonspecific Rous sarcoma virus promoter, the RLUs were relatively high for V␤6.7 promoter in lymphocyte cell lines (Jurkat, Ramos, and BW5147) but low in fibro- FIG. 3. The transcription initiation site of V␤6.7 determined by primer extension analysis. Lanes C, T, A, and G represent sequencing products of a plasmid DNA with known sequence as size markers. Lane 1 represents the primer extension products when 32 P-labeled primer was annealed to the total RNA isolated from human V␤6.7 ϩ cells. Lane 2 represents the extension products when 32 P-labeled primer was incubated with cellular RNA isolated from a mouse hybridoma cell line BW5147. Two distinct bands of 63 and 60 bp in lane 1 represent the full and partial extension products. blast cell cos 7 and embryonic kidney cell lines (data not shown). The effect of T-cell activation by stimuli that mimic TCR ligation, such as phorbol 12-myristate 13-acetate and ionomycin, on the promoter activity was further examined. As shown in Fig. 5, there were no significant changes in promoter activity when cells were stimulated by phorbol 12-myristate 13-acetate and ionomycin for 2-12 h, whether the basic 147-bp or the entire 463-bp V␤6.7 promoter fragments were tested. Our data indicate that V␤6.7 promoter element is lymphocytespecific but not T-cell lineage-specific.
Dodecamer, Rather Than Decamer, Is Critical to V␤6.7 Promoter Activity-To further evaluate the role of decamer D1 and its surrounding nucleotides in promoter activity, site-directed mutagenesis was performed. When D1 (AGTGATGTCA) was mutated to D1 m1 (AGAGATGCCA), in which the two Ts that form the core element of TGA-TCA were substituted, one by an A and the other by a C, the luciferase activity was totally abolished (Fig. 6), suggesting that the palindromic TGA-TCA is critical. We further mutated two nucleotides immediately 3Ј of D1, which changed AGTGATGTCACT to AGTGATGTCAAA (Fig. 6, D1 m2). This mutation destroys the reverse complementation between the first two nucleotides, AG, to the last two nucleotides, CT, in the original sequence. As shown in Fig.  6, the luciferase activity was decreased to two-thirds that of the original, nonmutated sequence, indicating that these two nucleotides are also important to promoter activity. Therefore, the regulatory motif in V␤6.7 should be considered a dodecamer (12-bp) rather than a decamer.
There is a dodecamer-like motif, AGTGACATCACA, with relatively the same location as the dodecamer in V␤6.7, in the promoter of V␤5.2. Although the luciferase assay showed that the original weak V␤5.2 promoter became strong when this imperfect dodecamer was corrected (Fig. 7, pVb5.2/*), the transcription activity was still far below that of V␤6.7. Only when a fragment from Ϫ85 to Ϫ1 of the V␤6.7 promoter, which contains the dodecamer, was replaced at its 3Ј-end (Fig. 7, pVb5.2/dd) did the V␤5.2 promoter became very strong and comparable in activity to V␤6.7. This observation led us to speculate that additional regulatory elements may exist 3Ј behind the dodecamer of V␤6.7. To test this hypothesis, constructs containing the fragment from Ϫ85 to Ϫ1 of V␤6.7, and corresponding fragments from V␤6.1 and V␤6.3, were prepared and used in transient transfection. As shown in Fig. 8, although V␤6.1 and 6.3 also contain the same dodecamer motif as V␤6.7, their activities were below that of V␤6.7. Sequence alignment showed that the 3Ј-ends of the promoters among V␤6.7, 6.1, and 6.3 are highly homologous to each other and that they all have the same dodecamer motif. The most distinct differences occur in the area from Ϫ67 to Ϫ43. We have screened for transcription factor sites and found at least three additional sites that are present in V␤6.7 but absent in V␤6.1 and 6.3. They are Sp1-U2snR.2 at Ϫ49, SIF_core_RS at Ϫ53, and BS15 at Ϫ66 (Fig. 9). Future site-directed mutagenesis FIG. 5. Luciferase assay to evaluate the effect of phorbol 12myristate 13-acetate and ionomycin on V␤6.7 promoter activity. 1 ϫ 10 6 cells were transfected with the construct containing the entire 463-bp promoter sequence of V␤6.7. Phorbol 12-myristate 13-acetate (20 ng/ml) and ionomycin (500 ng/ml) were added 12 h after transfection. Cell lysates were obtained at 2, 4, 6, and 12 h after stimulation and used for luciferase assay.
FIG. 6. Site-directed mutagenesis of decamer motif D1 in the promoter of V␤6.7. The left panel shows a schematic diagram of constructs used. The right panel shows the results of a transfection experiment using these constructs. Construct 1, a 147-bp basic promoter element that contains two decamer-like motifs, D1 and D2, and shows the full promoter activity. Construct 2, the same fragment as Construct 1, but inserted into pGL-2 in the opposite orientation. Construct 3, the same fragment as Construct 1, but D1 was mutated to D1 m1. Construct 4, the same fragment as Construct 1, but D1 was mutated to D1 m2. In D1 m1, two Ts that form the core structure of TGA-TCA were replaced, one by A and another by C. In D1 m2, two nucleotides, CT, 3Ј to the end of the decamer motif, were mutated to AA. experiments will be directed toward addressing the roles that these binding sites may play in promoter activity and function for V␤6.7.

DISCUSSION
The process of T-lymphocyte development is similar to that of B-cell maturation. Transcriptional activation of rearranging Ag receptor gene segments has been hypothesized to regulate their accessibility to the recombinational machinery. Several pieces of evidence support this prerequisite in both B-cells and T-cells (1)(2)(3)(4). Therefore, an understanding of transcriptional regulation may be important in elucidating the mechanisms controlling the lineage-specific patterns of rearrangement of the TCR genes during thymocyte differentiation.
Human TCR V␤ gene promoters themselves, which control the transcription, were reported to be essentially inactive. It was the enhancer, located 3.5-5.0 kilobase pairs 3Ј to the C␤2 gene segment, that conferred the transcriptional activity and T-cell lineage specificity to V␤ promoters (18 -20). However, in this study, we found that V␤ promoters varied in terms of their activities. Although most of the V␤ promoters that we tested were very weak, some were constitutively strong, such as those from V␤6 subfamilies. The promoter of V␤6.7 was found to be the strongest when compared with others. The 5Ј-flanking sequence analyses revealed that there are three decamer-like sequence motifs in the promoter region of V␤6.7. There seems to be a correlation between the promoter activity and decamerlike motifs. For example, the strong promoters, such as V␤6.3 and 6.7, all contain three decamer-like sequences. The most highly conserved portion in the decamer is an inverted repeat with the core sequence TGA-TCA. We identified the decamer as AGTGATGTCA in V␤6.7, AGTGACATCA in V␤5.2, and TGANNNNTCA in V␤13 subfamilies (7). The general idea is that the core regulatory element is composed of TGA-TCA. Those nucleotides in front of TGA, after TCA, or between TGA and TCA may be not so important. However, our results show that it is not so simple. Certainly, TGA-TCA is critical. When this element was destroyed by site-directed mutagenesis, the promoter activity was totally abolished (see Fig. 6, Construct 3). However, TGA-TCA alone is not sufficient to confer promoter activity. For instance, the decamer-like motif D3 in the 5Ј-distal segment of the V␤6.7 promoter was found to have essentially no effect on promoter activity (Fig. 4, II). Motif D2 in V␤6.7 also showed a very low contribution to the promoter activity (Fig. 4, V). The surrounding nucleotides 5Ј and 3Ј to TGA-TCA, AG and CT, were also important. When the CT at the 3Ј-end was mutated to AA, the promoter activity dropped to two-thirds of the original activity (see Fig. 6, Construct 4). These data support our hypothesis that the most critical regulatory element is a 12-nucleotide motif with a core sequence AGTGA-TCACT. Therefore, it is a dodecamer, rather than a decamer. Further evidence to support this idea comes from the study of promoter activities for V␤5.1 and 5.2. Although these promoters also contain a dodecamer-like motif at relatively the same location as the dodecamer in the V␤6.7 promoter, their activities were found to be relatively low (Fig. 1). Not surprisingly, the dodecamer-like motif in the promoters of V␤5.1 and 5.2 has the sequence AGTGACATCACA, which is not a perfect dodecamer as in the V␤6 genes. When this imperfect dodecamer in V␤5.2 was corrected by site-directed mutagenesis, the promoter activity was found to be increased. However, only when its 3Ј-end was replaced by a fragment from Ϫ85 to Ϫ1 of V␤6.7 did the V␤5.2 promoter became very strong, with activity comparable to V␤6.7 (see Fig. 7).
The palindromic features lead us to propose that the dodecamer may form a stem and loop structure by reverse complementation. Such a stem-loop structure will exert the regulatory function by binding with homo-or heterodimers of transcription factors, such as members of the cyclic AMP response element binding protein or activating transcription factor families (21)(22)(23). To test this hypothesis, we simply compared the light outputs for constructs in which the 147 bp of fragment IV were inserted in sense or antisense orientations FIG. 7. Augmented strength of the V␤5.2 promoter fragments containing the V␤6.7 dodecamer motif. A luciferase assay was performed using the following constructs: pGL-2/Basic, vector only; pVb5.2, the original V␤5.2 promoter fragment; pVb5.2/*, the V␤5.2 promoter in which the original imperfect dodecamer was corrected; and pVb5.2/dd, the V␤5.2 promoter in which the original imperfect dodecamer and its 3Ј behind sequence was replaced by an 85-bp fragment from Ϫ85 to Ϫ1 of the V␤6.7 promoter, which contains the dodecamer motif AGTGATGTCACT.
FIG. 9. 3-End sequence alignment for promoters of V␤6.7, 6.1, and 6.3. All three promoters contain the same dodecamer motif (shown in boldface type). Three additional transcription factor binding sites, BS15 (underlined), SIF_core_RS (marked with * above the line), and Sp1-U2snR.2 (double-underlined), were identified in V␤6.7 but not in V␤6.1 or 6.3. The consensus and reverse complementary sequences for these sites are listed in the bottom panel. (Fig. 6, Constructs 1 and 2). No significant difference in luciferase activity was observed between the two constructs, suggesting that this stem-loop model may stand. Together, these results indicate that a perfect dodecamer motif, which can form a stem-loop structure, plays an important role in TCR V␤ promoter activity.
In addition to the dodecamer motif, there are other elements, such as the TATA box (at Ϫ123) and CACCC box (at Ϫ140), that also contribute to the promoter activity of V␤6.7. In murine TCR V␤ genes, the decamer motifs were identified in the promoter region 10 -40 bp upstream of the TATA box (6). In human TCR V␤ genes, the locations of the decamer motif are different. The decamer in the V␤8 subfamily, similar to murine TCR V␤s, is located ϳ20 bp upstream of the TATA box. Interestingly, the dodecamer in the promoters of V␤6.1, 6.3 and 6.7 were located 36 bp downstream of the TATA box. It was unclear whether these different locations might affect the promoter activities. The CACCC box, identified in human V␤8.1 and mouse TCR ␣ gene silencer (24), was found to be another transcription factor binding site with T-cell lineage specificity (25,26). We found the CACCC box in the 150-bp basic promoter fragment from V␤6.7. The fragment that contains the dodecamer only, without TATA and CACCC boxes, showed partial promoter activity (Fig. 4, VI), indicating that TATA and CACCC boxes may play a synergistic role with the dodecamer and contribute to the high constitutive promoter activity of V␤6 genes. Other motifs identified 3Ј behind the dodecamer of V␤6.7, such as BS15, SIF_core_RS, and Sp1-U2snR.2, are also important but have not been characterized. Further analyses will help us to evaluate the contributions of these binding sites.