Identification of a Novel Repressive Element in the Proximal lck Promoter*

The T-cell-specific protooncogene lck , a src -related ty- rosine kinase, is under the control of two promoters that give rise to transcripts differing only in their 5 (cid:42) -untrans-lated regions. The distal promoter is transcriptionally active in both peripheral and thymic T-cells, whereas expression of the proximal promoter is highest in devel- oping thymocytes. The proximal promoter has also been shown to be selectively activated in a number of colon carcinoma cell lines. Approximately 570 base pairs of proximal promoter sequence is required for expression in both T-cells and colon carcinoma cell lines. Protein binding studies were initiated with an oligonucleotide homologous to a region that, when deleted, causes an increase in promoter activity in transgenic animals. Two proteins with approximate molecular masses of 35 and 75 kDa were found to bind to this region as deter- mined by UV cross-linking studies. Absence of specific protein binding is correlated with a high level of proxi- mal promoter expression. Competitive gel retardation analysis identified a 9-base pair binding site within the proximal lck promoter that is necessary for repression of transcription in cells that contain specific binding activity. Mutants of this binding site do not repress transcription. Repression does not occur in a cell line that expresses lck and lacks this activity. These data support the hypothesis that activation of lck transcription in colon carcinoma is due, at least in part, to the loss of a transcriptional repressor.

The T-cell-specific protooncogene lck, a src-related tyrosine kinase, is under the control of two promoters that give rise to transcripts differing only in their 5-untranslated regions. The distal promoter is transcriptionally active in both peripheral and thymic T-cells, whereas expression of the proximal promoter is highest in developing thymocytes. The proximal promoter has also been shown to be selectively activated in a number of colon carcinoma cell lines. Approximately 570 base pairs of proximal promoter sequence is required for expression in both T-cells and colon carcinoma cell lines. Protein binding studies were initiated with an oligonucleotide homologous to a region that, when deleted, causes an increase in promoter activity in transgenic animals. Two proteins with approximate molecular masses of 35 and 75 kDa were found to bind to this region as determined by UV cross-linking studies. Absence of specific protein binding is correlated with a high level of proximal promoter expression. Competitive gel retardation analysis identified a 9-base pair binding site within the proximal lck promoter that is necessary for repression of transcription in cells that contain specific binding activity. Mutants of this binding site do not repress transcription. Repression does not occur in a cell line that expresses lck and lacks this activity. These data support the hypothesis that activation of lck transcription in colon carcinoma is due, at least in part, to the loss of a transcriptional repressor.
In T-cells, expression of lck, a src-related tyrosine kinase, is driven by two promoters located approximately 35 kilobases apart that give rise to transcripts differing only in their 5Јuntranslated regions (1). Transgenes containing lck proximal promoter sequences are actively transcribed in thymocytes of transgenic animals, whereas in similar studies the distal promoter can support transcription in both thymocytes and peripheral T-cells (2,3). Therefore, these two apparently independently acting promoters may lead to both the developmental and cell type specific regulation of lck.
In T-cells lck is physically associated with the CD4 and CD8 components of the T-cell receptor complex (4,5) and has been reported to be associated with the ␤-chain of the interleukin-2 receptor (6). Although the exact functions of lck in T-cells are not known, it has been shown to be an essential component of the T-cell receptor signaling pathway (7,8) and may also be involved in T-cell maturation in the thymus (9). lck has been classified as a protooncogene because of its homology to src and its activation by promoter insertion in certain retroviral induced murine lymphomas (1,3) and by translocation in human T-cell leukemias (10). Also, transfection of NIH3T3 cells with an lck mutant that encodes an activated protein kinase results in malignant transformation (11). However, the functional significance of overexpression of wild type lck to the transformed phenotype has yet to be determined.
In addition to its expression in T-cells and T-cell lymphomas, lck is expressed in a subset of colon carcinoma cell lines derived from metastatic lesions as well as in some small cell lung carcinomas and B-cell leukemias (12)(13)(14). In solid tumor cell lines, lck expression results from activation of the proximal promoter, and no obvious rearrangements or amplifications of the gene have been detected in these cell lines (12).
The proximal lck promoter, like that of other src family members (15,16), is TATA-less and contains multiple start sites for initiation of transcription. Previously, deletion analysis has determined minimal promoter sequences required for both in vitro (17) and in vivo (2) expression of this gene. These studies identified a potentially important region between positions Ϫ565 and Ϫ433, which, when deleted, caused an increase in expression of both a reporter gene and of lck driven by the proximal promoter in transgenic animals (2). In order to investigate the transcriptional regulatory mechanisms involved in the expression of lck in colon cancer we studied this region of the proximal promoter. We now report that a novel sequence is located at positions Ϫ474 to Ϫ466 and specifically binds two proteins of approximate molecular mass 35 and 75 kDa. This binding activity is present in a number of cell lines that do not express lck as well as in a cell line that expresses only small amounts of lck. Specific binding was not detected using extracts from two cell lines expressing high levels of lck. The 9-bp 1 sequence acts to repress a heterologous minimal promoter in cell lines that contain specific binding activity. Repression does not occur in a colon carcinoma cell line that expresses high levels of lck and lacks this binding activity. Mutants of the sequence that do not bind the proteins do not repress transcription. We propose that activity of the proximal lck promoter is regulated by a repressor that binds to this sequence and that expression of lck in colon cancer cell lines is, in part, due to loss of this repression.

MATERIALS AND METHODS
Nuclear Extracts and Gel Retardation Analysis-Nuclear extracts were prepared essentially as described by Dingam et al. (18). Protein concentrations were determined using a Bradford assay (Pierce), and aliquots were stored in liquid nitrogen.
Gel retardation analyses were performed with a double-stranded oligonucleotide, synthesized by the SKI Microchemistry Core Facility, representing sequences Ϫ560 to Ϫ420 of upstream lck sequence: AA-CAGGCACACATTTATCACTTTACTCCTATGGAGTTCTGCTTGAT-TCATCAGACAAA (19). Other oligonucleotides used in this study were derived from this sequence and are described in detail in Fig. 2A. The AP1 consensus oligonucleotide was purchased from Stratagene, and the nuclear factor of activated T-cells (NFAT) oligonucleotide was a gift from S. Nimer (Memorial Sloan-Kettering Cancer Center). Doublestranded oligonucleotides were prepared by slow annealing and purified on an 18% polyacrylamide gel, and 25 ng was 32 P-end-labeled using T4 DNA kinase (Life Technologies, Inc.) by standard methods (20). Gel shifts were done as described (21). Gel shift buffer contained 20 mM Hepes, pH7.9, 75 mM NaCl, 0.5 mM EDTA, 1 mM dithiothreitol, 10 mM MgCl 2 . Each reaction contained approximately 0.05 ng of oligonucleotide (20,000 cpm), 6 mg of BSA, 4 g of poly(dI-dC), and 60 g of crude extract. These reactions were incubated at 30°C for 30 min, electrophoresed on a nondenaturing 6% acrylamide gel, dried, and subjected to autoradiography.
A search for known transcription factors that bind the mapped sequence was performed using the Transcription Factor Data Base from the National Center For Biotechnology, National Library of Medicine.
UV Cross-linking-Probes were generated by annealing complimentary oligonucleotides (as described above), one 30 bases shorter than the other and labeled by extension with Klenow (Life Technologies, Inc.) using a synthesis mixture containing 0.25 mM dATP, 25 M each dCTP and dGTP, 0.5 M each of [␣-32 P]dCTP and [␣-32 P]dGTP and 0.25 mM bromodeoxyuridine (Sigma). Gel retardation reaction mixtures were generated as described above. For some experiments, the 6% acrylamide gel was wrapped in plastic and cross-linked on a UV transilluminator (Fisher) for 30 min on each side. Each complex was excised from the gel and run on a 10% SDS-polyacrylamide gel, dried, and exposed to x-ray film. For other experiments, the gel shift reaction mixture was cross-linked by a transilluminator at a distance of 10 cm for 30 min, loaded onto a 10% SDS-polyacrylamide gel, dried, and subjected to autoradiography.
Cloning and Site-directed Mutagenesis-All subclones used in this study were derived from a 1.9-kilobase BamHI-EcoRI upstream lck fragment, provided by R. Benarous (22). A 650-bp PstI-BamHI fragment was subcloned into Bluescript KS ϩ by standard methods. Sitedirected mutagenesis was performed as described (23). Both wild type and mutant fragments were subcloned into pCAT Basic (Promega) by standard methods (20).
Concatamerization of both the binding site and a corresponding mutant was performed by cloning double-stranded oligonucleotides with flanking BglII and BamHI sites into a pCAT vector constructed by H.L. Grimes. By utilizing an EcoRI site within the pCAT sequence and sequential BamHI/EcoRI and BglII/EcoRI ligations, the binding site was duplicated and cloned into a TKCAT vector, also provided by H. L. Grimes. All plasmids were sequenced using Sequenase in a procedure described by the manufacturer (U.S. Biochemical Corp.).
Transfections-All cell lines were maintained in RPMI 1640 supplemented with 5% fetal bovine serum with the exception of HeLa, which was maintained in Dulbecco's minimal essential/F12, 5% fetal bovine serum. All cells were transfected using Lipofectin (Life Technologies, Inc.) by procedures described by the manufacturer. Cells were cotransfected with 0.25 g of a Rous sarcoma virus-luciferase vector kindly provided by E. Lai. Each plasmid used for transfection was titrated to insure that expression was a linear function of DNA concentration and that no competition between the promoters occurred. Cells were lysed after 24 h using 75 l of 1 ϫ reporter lysis buffer (Promega) containing 0.2 M phenylmethylsulfonyl fluoride. A portion of the lysate was assayed for luciferase activity using a Bertold Lumat luminometer. CAT assays were performed as described (24) with minor modifications. Briefly, 50 l of lysate was incubated with a solution containing 25 Ci of [ 14 C]acetyl-coenzyme A, 5 mM acetyl-coenzyme A, and 0.8 mM chloramphenicol for 1 h at 37°C. This mixture was extracted with 1 ml of ethyl acetate, dried by lyophilization, resuspended in 30 l of ethyl acetate, spotted onto TLC plates (Whatman), and chromatographed in 95% chloroform, 5% methanol. All CAT assays were quantitated using a Fuji phosphor-imaging system.
Proper initiation of transcription of the TKCAT templates was determined by RNase protection analysis (12) of total RNA isolated from transfected cells. The Riboprobe was comprised of sequences between Ϫ109 and ϩ52 of the thymidine kinase (TK) promoter.

Gel Retardation Analysis of Protein
Binding to Ϫ520 to Ϫ460 of the Proximal lck Promoter-Deletion analysis of the proximal lck promoter has allowed the identification of a region that, when deleted, causes an increase in transcription of both reporter genes and of lck in transient assays and in transgenic animals (2,17). To investigate the potential importance of this region to the expression of lck in colon carcinoma, gel retardation experiments were performed with an end-labeled 60-bp oligonucleotide representing sequences from Ϫ520 to Ϫ460 of the proximal promoter and nuclear extracts isolated from a variety of cell lines. The results of one of these experiments are shown in Fig. 1. In panel A nuclear extracts from HT29, a well differentiated primary colon carcinoma cell line that does not express lck, were analyzed. A number of complexes are observed in the lane marked 0 (no competitor). To assess the specificity of the complexes, competition analysis was performed using the unlabeled Ϫ520 to Ϫ460 oligonucleotide in 10, 100, and 1000-fold molar excess (lanes 3-5) as well as a 1000-FIG. 1. Gel retardation analysis. Panel A shows the result of a gel shift assay using oligonucleotide complementary to Ϫ520 to Ϫ460 of the proximal lck promoter and 60 g of crude nuclear extracts from HT29. The following competitors were used: 0, no competitor; N, a nonspecific oligonucleotide in 1000-fold molar excess and 10, 100, and 1000-fold molar excess of the Ϫ520 to Ϫ460 oligonucleotide. In panel B, the analysis in panel A was repeated using COLO205 nuclear extracts. On the right of panel B a consensus AP1 oligonucleotide was used in a gel retardation using COLO205 extracts and 1000-fold molar excess of a nonspecific oligonucleotide (N) and the AP1 oligonucleotide (1000 ϫ).
fold molar excess of a nonspecific oligonucleotide (lane N). 1000-fold molar excess of the specific oligonucleotide reduces the level of two of the complexes (arrows). The upper complex is occasionally resolved into two closely migrating complexes that show the same competition pattern (see below). In panel B, the experiment was repeated with nuclear extracts isolated from COLO205, an undifferentiated human colon carcinoma cell line that expresses high levels of lck. As shown in the left side of panel B, COLO205 nuclear extracts do not specifically shift this oligonucleotide. COLO205 extracts do specifically shift an oligonucleotide containing a consensus AP1 site as shown in the right portion of panel B. A summary of the results of the gel retardation experiments performed in several cell lines is shown in Table I. Protein binding is tabulated as a function of lck expression, as determined by RNase protection analysis (24,31) and quantitated by phosphor imaging. The two cell lines that express high levels of lck (CEM, a human T-cell leukemia cell line, which expresses high levels of lck from both the proximal and distal promoters (24,31), and COLO205) do not show specific binding to the Ϫ520 to Ϫ460 oligonucleotide. Three cell lines that do not express lck (HT29, HeLa, a well differentiated cervical carcinoma cell line, and T84, another well differentiated colon carcinoma cell line) show specific binding to the Ϫ520 to Ϫ460 oligonucleotide. SW620, a human colon carcinoma cell line derived from a lymph node metastasis expresses small amounts of lck and also specifically binds this oligonucleotide. Thus, in the two cell lines tested that express high levels of lck, specific protein binding to the Ϫ520 to Ϫ460 oligonucleotide is not detected.
Mapping of the Protein Binding Site in HT29 and SW620 -A competitive gel retardation analysis was used to map the binding site of this protein(s). A series of wild type and mutant oligonucleotides representing different regions of the Ϫ520 to Ϫ460 oligonucleotide ( Fig. 2A) were added in 1000-fold molar excess as competitors in the gel retardation assay. The first five lanes of Fig. 2B demonstrate the specific binding to the original Ϫ520 to Ϫ460 oligonucleotide. Neither the most 5Ј 30-bp oligonucleotide, nor the middle 30-bp oligonucleotide competed for binding of the protein. However, the most 3Ј 30-bp oligonucleotide competed with the original oligonucleotide, localizing the binding site to this region. The 3Ј oligonucleotides containing 6 bp changes ( Fig. 2A) were then analyzed similarly. As shown in Fig. 2B, oligonucleotides 2 and 3 did not compete, further narrowing the binding region. Oligonucleotides, 4, 5, 6, and 7, with 3 bp changes were used as competitors in this analysis (Fig. 2, A and C). Of these only oligonucleotides 5 and 6 did not compete, thus mapping the binding site to the sequence TT-TCATCAG, represented in boldface letters within the original Ϫ520 to Ϫ460 oligonucleotide in Fig. 2A. Identical results were obtained using nuclear extracts isolated from both HT29 and SW620. As described above, the upper complex shown in Fig.  2B is resolved into two complexes as seen in Fig. 2C; the significance of this is, at this time, unknown. Therefore, by competitive gel retardation the binding site was mapped to positions Ϫ474 to Ϫ466 of the lck proximal promoter.
UV Cross-linking Defines Two Proteins of Different Molecular Weight-To determine the approximate molecular weight of proteins that bind to the oligonucleotide, a Ϫ520 to Ϫ460 probe was used in UV cross-linking studies. HT29 nuclear extracts were incubated with a labeled probe, exposed to a UV transilluminator, and immediately subjected to SDS-polyacrylamide gel electrophoresis. The results of this analysis are shown in  Colon-lung met Ͻ0.01 ϩ a Each cell type and the site of derivation is shown. b lck expression was determined by RNase protection and is expressed in phosphor imaging units/mm 2 .

FIG. 2. Localization of the binding site by competition in a gel
retardation assay. Panel A shows the Ϫ520 to Ϫ460 oligonucleotide used as a probe in the gel shift assay as well as the competitor oligonucleotides used to map the binding site. Oligonucleotides 1-7 show the mutations, either C to T or G to A conversions used as competitors. The deduced binding site is shown within the Ϫ520 to Ϫ460 region in boldface letters. Panel B is the gel shift analysis using HT29 nuclear extract and no competitor (0) or various competitors at different levels of molar excess, 1000-fold molar excess of a nonspecific competitor (N), the Ϫ520 to Ϫ460 oligonucleotide in 10, 100, and 1000-fold molar excess, and the 5Ј, middle, 3Ј, and mutant 1-3 competitor oligonucleotides also in 1000-fold molar excess. Panel C is identical to panel B except that the last four lanes are competitions with 1000-fold molar excess of mutant oligos 4 -7. the unlabeled Ϫ520 to Ϫ460 oligonucleotide, two proteins were shown to bind specifically and are indicated by the arrows at the right of Fig. 3A. Since covalently bound oligonucleotides have a limited effect on the migration of the cross-linked proteins in SDS-polyacrylamide gel electrophoresis gels (4), the approximate molecular weight of the proteins that bind this oligonucleotide are 35 and 75 kDa. To determine whether similarly sized proteins were present in both the complexes observed in the gel retardation assays, a similar gel was crosslinked in situ on a UV transilluminator, and both the upper and lower complexes described in Figs. 1 and 2 were excised and run on an SDS-polyacrylamide gel electrophoresis gel. The result of this experiment is shown in Fig. 3B. These experiments demonstrated that both proteins detected by in the solution UV cross-linking studies were present in each complex.
The Binding Site Is Required for Transcriptional Repression-To begin to determine the functional significance of the binding site in the lck proximal promoter, a 650-bp PstI-BamHI fragment containing 570 bp of promoter sequence was cloned into a Basic CAT vector and cotransfected with a Rous sarcoma virus-luciferase reporter gene into SW620, HT29, HeLa, and COLO205 cells. Resultant CAT activity for SW620 and HT29 is shown in Fig. 4. These experiments demonstrated that the fragment containing 570 bp of lck promoter sequence was sufficient for expression of CAT activity in SW620 and COLO205 but not in HT29 or HeLa, in which it was only slightly active. Transient assays were performed using the wild type proximal lck promoter and a proximal promoter carrying a mutation in the binding site (GGGCATCAG), which abolished protein binding in gel retardation experiments. The results of a representative experiment are shown in Fig. 4. SW620 expresses high levels of CAT activity from the wild type lck promoter, while in HT29 this promoter is only slightly active. These data suggest that the proximal lck promoter is activated in trans in SW620. The mutation in the lck promoter has no detectable effect on transcription in SW620 and shows a 2-fold increase in transcription over the wild type promoter in HT29 cells.
Since loss of transcriptional repression at this site may not be sufficient to fully activate this promoter in HT29, the effect of the site on a minimal promoter was investigated. An oligonucleotide containing the binding site was cloned in one or two copies just upstream of the herpes simplex virus TK minimal promoter in a CAT-containing expression vector (Fig. 5A) and these constructs were used in transient transfection assays. The results of a representative experiment for HT29 and SW620 are shown in Fig. 5, B and C, respectively. Comparison of CAT activity driven either by the TK promoter alone or with one or two sites just 5Ј to the TK promoter reveals the ability of this element to completely abolish transcription from this minimal promoter element in HT29 (Fig. 5A), and identical results were obtained in HeLa (data not shown). These experiments were repeated in SW620, and although complete repression was not documented, this site caused an 8-fold reduction in promoter activity in this cell line. As a control, a mutant binding site, GGGCATACT, was also cloned upstream of the TK promoter, and the transient assays were repeated. For both HT29 (Fig. 5B) and SW620 (Fig. 5C) a mutant binding site that did not compete for protein binding in gel retardation analysis had no effect on the transcriptional activity of the TK promoter FIG. 3. Molecular characterization of proteins binding the ؊520 to ؊460 oligonucleotide. Panel A shows a 10% SDS-polyacrylamide gel electrophoresis gel indicating the relative migration of the protein-DNA complexes from either HT29 or SW620 nuclear extracts that bind to a doubly labeled probe corresponding to Ϫ520 to Ϫ460 of the proximal promoter after UV cross-linking. The molecular weight marker is listed at the left. The lane marked 0 is no competitor added; 100 ϫ and 1000 ϫ indicate where either 100-or 1000-fold molar excess of specific competitor was used. Panel B shows the migration patterns of both the upper and lower protein-DNA complexes excised from gel retardations using either HT29 or SW620 nuclear extracts. in either of these cell lines. To insure that the CAT activity measured was due to accurate transcription initiation in the TK promoter, RNase protection analysis was performed using a probe containing the TK promoter. Proper initiation of transcription of the TKCAT construct was observed (data not shown).
Since this sequence acts to repress transcription in cells containing specific binding activity, transfection studies were performed in a cell line lacking this activity. As shown in Fig.  6, the sequence does not suppress the TK promoter in COLO205, a cell line that expresses high levels of the proximal lck transcript. Proper transcript initiation was also observed in this cell line (data not shown). Thus, repression of promoter activity seems to require both the intact binding sequence and the binding activity.

DISCUSSION
In this study we have defined a sequence in the human lck proximal promoter located at positions Ϫ474 to Ϫ466, which acts as a strong repressor of transcription. The approximate molecular masses of proteins that specifically bind this sequence are 35 and 75 kDa. The binding activity was detected in four human tumor cell lines; three of these do not express lck (HT29, T84, and HeLa), and the other expresses nominal amounts (SW620). Specific binding was not detected in two cell lines that express high levels of proximal lck promoter transcripts. This sequence acted to repress transcription of a heterologous promoter in cell lines that contained the binding activity but had no effect on the promoter in a cell line lacking this activity. These data suggest that activation of the proximal lck promoter is at least partly due to a loss of transcriptional repression mediated by the loss of specific protein binding to this 9-bp sequence.
These findings are consistent with results from transgenic mice experiments, which demonstrated protein binding to an analogous 50-bp region in the murine proximal promoter with nuclear extracts from spleen, a tissue in which this promoter is inactive, and a lack of binding in cells expressing the transgene (2). Comparisons between mouse and human have demonstrated a high degree of sequence conservation in the proximal promoter element, indicating that this promoter may be similarly regulated in both species. This conservation also extends to the binding site mapped to Ϫ474 to Ϫ466 in this study (7/9-bp). These data indicate that the murine system may have a functionally similar transcriptional repressor.
Our data, which show measurable promoter activity using 570 bp of the human proximal promoter sequence, varies somewhat from previously reported transient assays in which, when using identical sequence, deletion to Ϫ512 was necessary to show modest promoter activity (17). In transgenic animals, FIG. 5. The binding site functions to repress transcription of the TKCAT construct. Graphs of transient expression of TKCAT as compared with expression obtained with either one or two wild type binding sites as well as mutant binding site (mut) cloned 5Ј of the TK promoter. All transfections were done in at least triplicate in three different experiments. Diagrams of the constructs used in this analysis are shown in panel A. The quantitation of one experiment is shown in panel B and panel C using either HT29 or SW620, respectively. CAT assays were quantitated by phosphor imaging and expressed as relative to luciferase activity measured by luminometry (RLU).
FIG. 6. Promoter activity is not affected by the binding site in COLO205. Transfections were performed as in Fig. 5 using COLO205 as the recipient cell line. CAT assays were repeated three times in at least triplicate, quantitated by phosphor imaging, and expressed relative to luciferase activity. A representative experiment is shown. however, 584 bp of the murine proximal promoter sequence was sufficient for promoter activity (2). In this report CAT activity was measured rather than primer extension of transient ␤-globin mRNA, possibly allowing for the detection of lower amounts of promoter activity. The binding site located within the Ϫ520 to Ϫ460 oligonucleotide was mapped to the sequence TTTCATCAG at positions Ϫ474 to Ϫ466 by competitive gel retardation assays. A computer search of the Transcription Factor Data Base revealed no strong homology to known transcription factor binding sites. However, we noted some homology to the 3Ј portion of the composite binding site for NFAT within the interleukin-2 promoter, i.e. TTTCATA-CAG, the portion thought to bind Fos or Jun (25)(26)(27). The proteins detected in this study are unlikely to be identical to those in the T-cell NFAT complex, as neither an oligonucleotide containing an NFAT motif nor one containing an AP-1 motif competes in 1000-fold molar excess in a gel retardation assay (data not shown).
Although the region from Ϫ520 to Ϫ460 was shown to contain a repressive element (2) and gel retardation experiments correlated a lack of binding to this region with a high level of lck expression, site-directed mutagenesis of the binding site located at Ϫ474 to Ϫ466 had little effect on the expression of the intact 570-bp promoter in transient assays in SW620. However, the binding site does repress the transcriptional activity of a heterologous promoter in this cell line. Since the proximal lck promoter is activated in SW620 and deletion analysis to Ϫ433 of the human promoter had little effect on transcription in SW620 (17), it is possible that the promotor is activated at another site. However, a reproducible 2-fold increase in promoter activity was observed in HT29. It is likely that other factors are required for full activation of this promoter.
To determine the effect of the binding site on a minimal promoter the binding site was cloned upstream of a TK promoter containing CAT vector and used in transient transfection assays. The binding site acted to completely abolish transcriptional activity of this minimal promoter in both HeLa and HT29 and reduced transcriptional activity 8-fold in SW620. A mutated site that abolishes protein binding in gel retardation experiments had no effect on promoter activity, indicating that repression was sequence-specific. Since the binding site had no effect of transcription from this promoter in COLO205 it appears that the specific binding proteins, which were not detected by gel retardation, are either absent in this cell line or at least not able to bind to this site. This paper describes a novel binding site in the proximal lck promoter that acts as a strong repressor of transcription in cell lines that either do not express proximal lck transcripts (HT29 and HELA) or only express small amounts (SW620). The correlation of repressor activity, protein binding activity, and lck expression suggests that this repressor does regulate lck transcription. Whether expression of the wild type lck transcript in colon carcinoma regulates aspects of the transformed phenotype remains to be established. However, the loss of a transcriptional repressor may have more global regulatory effects in colon cancer. In addition, the proximal lck promoter is most active in developing thymocytes; the transcriptional repressor complex defined in this report may play an important role in the silencing of this promoter in peripheral T-cells. Purification and cloning of the binding proteins will allow further characterization of the function of this repressor.