Transcription of the blk gene in human B lymphocytes is controlled by two promoters.

Genomic DNA containing the first exon and 5′-flanking region of the human protein tyrosine kinase, blk, was isolated. Sequence analysis identified a TG repeat element in this region with enhancer activity, but no TATA or CCAAT sequences were found. Two blk transcripts of 2.2 and 2.5 kilobases were identified in various B-cell lines by Northern blot analyses, and primer extension experiments demonstrated two clusters of multiple transcription start sites. Subsequent promoter analyses by transient transfection assays with a reporter gene identified two promoter elements in the human blk gene. Promoter P1 contains sequences that have been shown to regulate the expression of immunoglobulin genes and promoter P2 contains elements that are highly conserved in the promoter of major histocompatibility complex class II genes, as well as a B-cell-specific activator protein- (BSAP) binding site. Electrophoretic mobility shift assays demonstrated that the binding of a protein to the BSAP-binding site was correlated with the presence of the 2.5-kilobase blk transcript. These data suggest that the two human blk RNAs arise from the transcription of the blk gene by two distinct promoters and that these promoters may be subject to regulation by different trans-acting factors.

Genomic DNA containing the first exon and 5-flanking region of the human protein tyrosine kinase, blk, was isolated. Sequence analysis identified a TG repeat element in this region with enhancer activity, but no TATA or CCAAT sequences were found. Two blk transcripts of 2.2 and 2.5 kilobases were identified in various B-cell lines by Northern blot analyses, and primer extension experiments demonstrated two clusters of multiple transcription start sites. Subsequent promoter analyses by transient transfection assays with a reporter gene identified two promoter elements in the human blk gene. Promoter P1 contains sequences that have been shown to regulate the expression of immunoglobulin genes and promoter P2 contains elements that are highly conserved in the promoter of major histocompatibility complex class II genes, as well as a B-cell-specific activator protein-(BSAP) binding site. Electrophoretic mobility shift assays demonstrated that the binding of a protein to the BSAP-binding site was correlated with the presence of the 2.5-kilobase blk transcript. These data suggest that the two human blk RNAs arise from the transcription of the blk gene by two distinct promoters and that these promoters may be subject to regulation by different trans-acting factors.
The blk gene is a member of the src family of protein tyrosine kinases (1,2). The product of the blk gene, as well as other src-family members including fyn, lyn, and lck, has been shown to associate with the immunoglobulin receptor complex (3,4). Since signal transduction via the B-cell antigen receptor is mediated by protein tyrosine phosphorylation (5,6), blk may play a role in B-cell activation and the initial steps of the intracellular signal pathway.
Little is known about the regulatory mechanisms controlling the restricted expression of blk. In the mouse, blk expression is restricted to B-lymphoid cells and is developmentally regulated (7). blk transcripts are first detected in pro-B-cells and persist through differentiation to mature-B-cells, but are absent in plasma cells. This expression pattern is similar to that of two other B-cell-specific genes, mb-1 and CD19 (8,9). The control of CD19 gene expression has been shown to involve a B-cellspecific activating protein (BSAP) 1 (10). BSAP is a member of the paired domain family of transcription factors and is encoded by the paired box gene Pax-5 (11,12). Recently, a BSAPbinding site was identified in the murine blk promoter region (13). The correlation of the expression of BSAP and blk suggested that BSAP may, at least partially, account for the B-cellspecific expression of murine blk.
In contrast to the murine blk gene, the human blk gene, although predominantly expressed in B-cells, is also found in some T-cells (14, and this study). In order to understand the mechanisms regulating human blk expression, we have isolated and characterized the first exon and the 5Ј-flanking region of the human blk gene. Northern blot analyses identified the presence of two blk RNAs in various B-cell lines, and the transcription start sites of these RNAs were mapped to two clusters. Electrophoretic mobility shift assays demonstrated that expression of one of the blk RNAs was correlated with the presence of BSAP. Luciferase reporter gene assays and deletion analyses identified two promoter elements in the 5Ј-flanking region of the blk gene. An enhancer-like element, containing a TG repeat sequence, was also identified upstream of these promoters.
Isolation of the blk Promoter-Genomic DNA clones were isolated from a human lymphocyte genomic library in the bacteriophage Lambda Dash (Stratagene, La Jolla, CA), using a 32 P-labeled blk cDNA as the probe. Six positive hybridizing clones were isolated after screening approximately 1 ϫ 10 6 plaques. Two clones containing blk 5Јflanking sequences were identified by restriction mapping and Southern blot analyses with a 32 P-labeled first exon region probe of the blk cDNA (Ϫ386 to ϩ75). A 6.0-kb EcoRI fragment of the blk29 clone was inserted into the EcoRI site of Bluescript II SK (Stratagene) to yield plasmid pL62. The nucleotide sequence of the blk promoter was determined by the dideoxy-chain termination method following the manufacturer's protocol (Sequenase; United State Biochemical Corp.).
Northern Blotting and Reverse Transcribed-PCR Analysis-Total cellular RNA was prepared according to Gough (15). The RNA samples (20 g/lane) were separated by electrophoresis on a 1% agarose-formaldehyde gel and transferred to a Nytran membrane (Schleicher & Schuell) as described previously (16). cDNA fragments were labeled with [ 32 P]dCTP by random oligonucleotide priming (17). The membranes with transferred RNA were probed with either a 450-bp EcoRI fragment containing the 5Ј-untranslated region of the human blk cDNA or a 900-bp blk cDNA fragment that encodes the amino-terminal region of the blk protein.
To amplify blk RNAs, two sense primers, blkpro26 (5Ј-TGAAAACT-GATTGAGATGAG-3Ј, ϩ51 to ϩ70) and blkpro28 (5Ј-GAAGGGCATT-GTGACCCACG-3Ј, Ϫ192 to Ϫ173), derived from the first exon, were used for the detection of 2.2-and 2.5-kb RNAs, respectively. To exclude the possibility of contamination with genomic DNA template, the antisense primer, race 1 (5Ј-GCGGTGTAGTCATACAGAGCCACCACG-3Ј, ϩ391 to ϩ417), was derived from the third exon of blk. One g of total RNA was reverse transcribed and then followed by 40 cycles of PCR using a GeneAmp PCR system 9600 (Perkin-Elmer). Each cycle consisted of a denaturing temperature of 95°C for 30 s, an annealing temperature of 55°C for 30 s, and an extension temperature of 72°C for 1 min. The PCR products were analyzed by gel electrophoresis and by Southern blot analysis (17).
Primer Extension-Primer extension was performed as described previously (18). Two synthetic oligonucleotide primers were used in the extension reaction. Primer blkEXT (5Ј-CAAGTCCGGCAGAGTTG-CAAACCTCCATGC-3Ј, Ϫ256 to Ϫ287) was used to map upstream start sites, while primer blk-pro29 (5Ј-CTCATCTCAATCAGTTTTCATG-GCTTGT-3Ј, ϩ43 to ϩ70) was used to map the downstream start sites. Each 32 P-end-labeled primer was annealed to 40 g of total RNA in a hybridization buffer (5 l) containing 0.4 M NaCl, 40 mM PIPES (pH 6.4), and 1 mM EDTA at 80°C for 2 min and then at 60°C for 16 h. The annealed products were extended with 200 units of superscript reverse transcriptase (Life Technologies, Inc.) in an extension mixture (50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl 2 , 0.8 mM deoxynucleotide triphosphates, 10 mM dithiothreitol) at 42°C for 1 h. The extension products and a sequencing ladder generated with the same primers were analyzed on a 7 M urea, 6% polyacrylamide sequencing gel.
Plasmid Constructs-To generate luciferase reporter plasmids, the 6.0-kb SacI-XhoI fragment from pL62 was cloned into the SacI and XhoI sites of pGL2-basic (Promega Corp., Madison, WI). The plasmid was digested with HindIII and religated to remove extra multiple cloning sites. The resulting construct, pL74, contained 6.0 kb of the blk 5Ј-flanking region upstream of the luciferase coding sequence and the simian virus 40 splice and polyadenylation sites present in the original pGL2-basic vector.
A series of 5Ј deletions were constructed from pL74. Plasmids pL76, pL78, pL82, and pL90 were generated by cleavage of pL74 with SmaI, KpnI, StuI combined with SmaI and PstI, respectively, and religation to delete the intervening sequences. Plasmids pL86 and pL88 were produced by cleavage with HindIII and either BglII or StuI, respectively, filling in with Klenow and religation. For internal promoter deletion, plasmid pL84 was generated by cleavage of pL76 with PstI and then religation. Plasmid pL27 was constructed in several steps. The 1.8-kb HindIII to BamHI fragment of blk 29 was first cloned into the HindIII and BamHI sites of Bluescript. The resulting plasmid was digested with EcoRI and HindIII, filled in, and religated to create plasmid pL23. Then, the XhoI to BamHI fragment of pL23 was cloned into the BglII site of pGL2-basic vector. Plasmid pL97 was made by polymerase chain reaction (PCR) amplification of a fragment (from Ϫ801 to Ϫ256) using a mutant primer, blkpro5 (5Ј-CAGgaGCTCACTGTGTCTGGCT-3Ј, Ϫ801 to Ϫ780), and the blkEXT primer. Blkpro5 introduced a SacI site at the 5Ј end of the PCR product. The fragment was cleaved with SacI and BglII and inserted into the corresponding sites of the pGL2-basic vector.
The RSSV-CAT plasmid, in which the RSSV promoter was placed upstream of the chloramphenicol acetyltransferase (CAT) gene, was a gift from Dr. R. Weigel (Stanford University, CA).
Transfection, Luciferase, and CAT Assays-Daudi, SUP-B8, SUP-B12, ARH-77, and Jurkat cells were transfected by electroporation as described previously (19). Cells (6 ϫ 10 6 ) were washed once and resuspended in 1 ml of RPMI (without serum or antibiotics) containing 10 g/ml DEAE-dextran, 5 g of the reporter plasmid, and 1 g of RSSV-CAT plasmid and electroporated using 0.4-cm cuvettes in a Bio-Rad gene pulser (Bio-Rad) at 300 V, 960 microfarads for B-cells and 250 V, 960 microfarads for T-cells. After incubation in RPMI complete medium for 24 -36 h, the cells were harvested and lysed in 150 l of 250 mM Tris (pH 7.7) by three freeze-thaw cycles. The cell extracts were then analyzed for luciferase and CAT activities. DU 145, T-47D, and 293 cell lines were transfected using the calcium phosphate precipitation method as described previously (20).
Luciferase activity was measured from 20 l of the cell extract reacting with the luciferase reagents as described by the supplier (Analytical Luminescence Laboratory, San Diego, CA). The light emission was measured with a Monolight 2010 instrument (Analytical Luminescence Laboratory), reading relative light for 10 s. Luciferase activities were normalized for CAT-specific activity. For CAT assays, cell extracts were heated to 65°C for 15 min to inactivate endogenous acetylases. Subsequently, CAT activity was measured by incubating 50 l of the cell extract with [ 14 C]chloramphenicol and butyryl-coenzyme A as described previously (21,22).
The 5ЈBSAP and 3ЈBSAP oligonucleotides correspond to nucleotides Ϫ404 to Ϫ391 of the blk 5Ј-flanking region and span a putative BSAPbinding site. The 5ЈmBSAP and 3ЈmBSAP oligonucleotides correspond to the same region but contain two mutations, C to A and G to A, as indicated by lower case letters. Each pair of oligonucleotides was annealed and fill-in labeled with Klenow DNA Polymerase (Pharmacia Biotech Inc.) and [␣-32 P]dCTP. Ten fmol of each labeled DNA probe was incubated with 10 g of whole cell extract in 1 ϫ binding buffer (40 mM KCl, 20 mM HEPES (pH 7.7), 1 mM MgCl 2 , 0.1 mM EDTA, 0.4 mM dithiothreitol) containing 4% Ficoll and 1 g of salmon sperm DNA in a volume of 25 l at room temperature for 30 min. Protein⅐DNA complexes were analyzed on a native 4% polyacrylamide gel in 0.25 ϫ Tris borate-EDTA. Gels were dried and autoradiographed.

Cloning and Sequence Analysis of Human blk 5Ј-Flanking
Region-To isolate the genetic elements involved in the regulation of blk gene expression, we initially screened a human lymphocyte genomic library with the full-length blk cDNA (24). Five phage clones identified by Southern blot analysis are diagrammed in Fig. 1. Restriction enzyme mapping showed that the human blk gene spans greater than 60 kb of genomic DNA. Like other members of the src gene family, the first exon of human blk contains only 5Ј-untranslated sequence. A 2.3-kb SmaI-EcoRI fragment containing the 5Ј-flanking region and exon 1 of the human blk gene was subcloned and sequenced (Fig. 2). Nucleotide position Ϫ386 corresponds to the 5Ј end of the cDNA (24). Computer analysis demonstrated an absence of TATA or CCAAT motifs located at the optimal distance (25-27) from the Ϫ386 site or the ϩ1 major transcription initiation site (later identified in this study). However, several potential motifs known to be involved in the expression of other B-cell-specific genes were found in two clusters separated by approximately 400 bp. One cluster contains E (28), PU.1 (29,30), X-and Y-like boxes (31), and c-MYB (32) motifs. Another cluster located upstream of the ϩ1 position contains E, PU.1, PEA3 (33,34), and SP1 (35) motifs. In between, a potential BSAP-binding site (10,12) that has been shown to be important for the expression of the B-cell-specific gene, CD19, was identified at position Ϫ404 to Ϫ391. The other notable feature of the nucleotide sequence in the 5Ј-flanking region of the blk gene was the presence of a 36-bp TG repeat (TG element) at position Ϫ2071 to Ϫ2106.
Two Sizes of blk Transcripts Differentially Expressed in Bcell Lines-The presence of a cluster of B-cell-specific motifs downstream of position Ϫ386 suggests that the expression of blk may have two positive regulatory regions. To test this hypothesis, we first investigated the expression pattern of the blk gene by Northern blot analysis. An RNA blot containing total RNA isolated from various lymphoid cell lines was hybridized with a labeled blk cDNA probe (Fig. 3, upper panel). The same blot was reprobed with a labeled human glyceraldehyde-3-phosphate dehydrogenase DNA probe as a control (Fig. 3, lower panel). Two major blk transcripts with approximate sizes of 2.2 and 2.5 kb were detected (Fig. 3). The relative abundance of these two RNAs were different in the tested cell lines. As shown in Table I, four patterns of blk expression were observed. The 2.2-kb message was expressed at a higher level than the 2.5-kb RNA in ARH-77, SU-DHL4, Daudi, Raji, SUP-B8, and Reh. Both transcripts were detected with equal intensity in SUP-B17, OCI-Ly8, Nalm 6, and normal spleen. SUP-B12, RPMI 6666, RPMI 7666, and Jurkat expressed only the 2.2-kb RNA; the 2.5-kb transcript was not detected. Finally, in two B-lymphoblastoid cell lines (PW and UC 729) and several non-B-cell lines (CEM, U-937, LNCaP, T-47D, and 293), no blk transcripts were detected.
Mapping of the blk Transcription Initiation Sites-To determine whether the 2.2-and 2.5-kb transcripts initiate from the same or different start sites, we carried out primer extension experiments using two different DNA primers. The primers blkpro29 (from ϩ43 to ϩ70) and blkEXT (from Ϫ256 to Ϫ287) were designed for the extension of the 2.2-and 2.5-kb RNA species, respectively. Total RNA isolated from Daudi, RPMI 7666, ARH-77, and CEM was hybridized with each 5Ј-endlabeled primer, extended with reverse transcriptase, and analyzed on polyacrylamide gels. As shown in Fig. 4, a pattern of heterogeneous start sites was observed using the blkpro29 primer. Multiple initiation sites are commonly observed with promoters lacking a TATA box (7, 33, 36 -40). A major transcription start site was found at position ϩ1 in Daudi, ARH-77, and RPMI 7666 cell lines but was absent in the control cell line CEM (Fig. 4A). This suggests that the 2.2-kb RNA that is highly expressed in Daudi, ARH-77, and RPMI 7666 cells initiates mainly from this start site.
A pattern of heterogeneous start sites was also found using the blkEXT primer; however, we failed to detect extension products in the RPMI 7666 cell line using this primer (Fig. 4B). To confirm the lack of expression of the 2.5-kb blk message in RPMI 7666 cells and other cell lines, reverse transcribed-PCR was performed. Total RNAs from these cell lines were reverse transcribed and amplified with an upstream primer, either blkpro28 (Ϫ173 to Ϫ192), for the detection of the 2.5-kb RNA, or blkpro26 (ϩ51 to ϩ70), for the detection of the 2.2-kb RNA, and a downstream primer, race 1 (from exon 3). The results demonstrated that SUP-B12 and Jurkat cell-line cDNA can be amplified by blkpro26, but not by blkpro28, whereas RPMI 6666 and RPMI 7666 cell line cDNA can be amplified by both primers (Table I). This suggests that the 2.5-kb transcript, although undetectable by Northern or primer extension analyses in RPMI 6666 and RPMI 7666 cells, can be detected by the more sensitive reverse transcribed-PCR method. In contrast, the SUP-B12 and Jurkat cell lines were found to lack the 2.5-kb transcript by all these assays. The presence of only the 2.2-kb blk RNA in some cell lines and both transcripts in other cell lines suggested that the expression of these two RNAs may be controlled by different regulatory elements.
Functional Analysis of the Human blk Promoters-To determine whether the 5Ј-flanking region contains regulatory elements that control the tissue-specific expression of the blk gene, transient transfection assays were performed in a variety of cell lines. A SmaI-EcoRI (Ϫ2258 to ϩ75) and KpnI-EcoRI a Relative levels of each transcript are indicated by ϩϩ (strong hybridization signal), ϩ (moderate to weak signal), and Ϫ (no signal). b A ϩ indicates the presence of an amplification product, ϩ/Ϫ indicates a weak amplification product, and Ϫ indicates no amplification product.
c The presence (ϩ) or absence (Ϫ) of BSAP was determined by electrophoretic mobility shift assays. d ND, not determined.

FIG. 4. Primer extension mapping of the transcription initiation sites of the 2.2-and 2.5-kb blk transcripts.
Total RNA was prepared from Daudi, ARH-77, RPMI 7666, and CEM cell lines. Oligonucleotides complementary to nucleotides ϩ43 to ϩ70 (A) and Ϫ256 to Ϫ287 (B) of the blk sequence were hybridized to 20 g of RNA. The hybridized primers were extended with reverse transcriptase. Extension products as well as a sequence ladder were separated on a 6% polyacrylamide gel. The arrow indicates the most abundant product.
(about Ϫ3000 to ϩ75) fragment containing 5Ј-flanking sequences and part of the first exon region of the human blk gene were inserted upstream of the luciferase gene in the plasmid pGL2-basic. The resulting plasmids, pL76 and pL78, were transfected into B-cell lines (Daudi and SUP-B8) and non-Bcell lines (DU 145, T-47D, and Jurkat), and luciferase activity was determined. Two plasmids, an SV40 promoter-driven luciferase construct (pGL2-control) and a promoter-less luciferase plasmid (pGL2-basic), were used for controls. In addition, each construct was cotransfected with the RSSV-CAT plasmid to normalize for transfection efficiency. As shown in Table II, both pL76 and pL78 constructs gave much higher levels of luciferase activity in B cell lines (Daudi or SUP-B8) than in non-B-cell lines (DU 145 or T-47D). These constructs were also active in the T-cell line, Jurkat which we have shown expresses blk (Table I).
To identify the regulatory elements responsible for the tissue-specific expression and the promoter activity of the human blk gene, we have constructed 5Ј deletion mutations of the pL76 luciferase reporter gene construct (Fig. 5). The different deletion constructs were transfected into Daudi cells, and luciferase activities were measured. The results show that removal of sequences from Ϫ2258 to Ϫ1628 (pL82, Ϫ1628 to ϩ75) significantly reduced luciferase expression (Fig. 5), indicating the presence of an enhancer element in this region. Further deletion to Ϫ338 (pL90, Ϫ338 to ϩ75) resulted in an increase in luciferase activity, suggesting the presence of a negative regulatory element upstream of Ϫ338 and the presence of an element with promoter activity between Ϫ338 and ϩ75 (designated P1). Interestingly, combining the P1 promoter fragment with the enhancer containing fragment (Ϫ2258 to Ϫ1628), which yields plasmid pL84, dramatically increased activity to nearly the same level as the pL76 construct.
Deletion of the pL76 construct from the 3Ј end (pL86, Ϫ2258 to Ϫ338) resulted in a construct with significant promoter activity, albeit weaker than the P1-containing pL90 construct. To map this element further, it was subdivided into two fragments, one containing the enhancer element (pL88, Ϫ2258 to Ϫ1628) and the other containing a cluster of potential regulatory motifs (pL97, Ϫ801 to Ϫ338). Neither of these constructs had significant activity, suggesting the presence of a weak promoter that requires an enhancer element for activity.
A TG element within the blk 5Ј-Flanking Region Shows Nontissue-specific Enhancer Activity-To test whether the fragment from Ϫ2258 to Ϫ1628 could enhance transcription in a B-cell-specific manner, the fragment was inserted downstream of an SV40 promoter-driven luciferase gene. The resulting construct pL104 was transfected into Daudi, DU145, T-47D, ARH-77, Jurkat, and 293 cells. As shown in Table III, in all of the cell lines tested, the Ϫ2258 to Ϫ1628 fragment in pL104-enhanced SV40 promoter activity by about 3-fold compared to the pGL2promoter plasmid. Thus this region contains an enhancer, but it is not B-cell-specific. Deletion analyses localized this enhancer to a TG repeat element (see Fig. 1). 2 These results are in agreement with other observations that TG elements can act as enhancers (41,42).
The Binding of a Protein at the BSAP Site Is Correlated with the Expression of the 2.5-kb blk RNA-Recent studies (13) have shown that the expression of murine blk is correlated with the presence of BSAP and have suggested that BSAP is a positive regulator of murine blk transcription. To determine whether BSAP may regulate the expression of one of the human blk RNAs, we examined the putative BSAP site in the human blk gene (position Ϫ404 to Ϫ391) by electrophoretic mobility shift assay. A 50-bp probe (from Ϫ426 to Ϫ375) spanning the BSAPbinding site was prepared from the 5Ј-flanking region of blk. This probe was mixed with cell extracts prepared from various lymphocytic cell lines. As shown in Fig. 6A, a DNA⅐protein complex was found in cell extracts of Daudi, ARH-77, SU-DHL4, OCI-Ly8, SUP-B17, Raji, SUP-B8, RPMI 6666, and RPMI 7666 cell lines, whereas it was not present in extract of SUP-B12, Jurkat, CEM, or UC 729 cells (Fig. 6, A and B). As summarized in Table I, the presence of this DNA-bound protein correlated with the presence of the 2.5-kb transcript, indicating that this protein may be a positive regulator of the 2.5-kb blk transcript. Interestingly, the expression of the 2.2-kb blk transcript in SUP-B12 and Jurkat apparently does not require this protein.
To confirm that the bound protein was BSAP, we mutated the 50-bp BSAP probe at positions Ϫ406 (C to A) and Ϫ402 (G to A). Mutations at the corresponding positions in the murine blk promoter have been shown to impair the binding of BSAP (13). Data presented in Fig. 6B show that the mutated probe, m-BSAP, was unable to interact with the protein identified by the wild-type (wt-BSAP) probe. DISCUSSION In previous studies, it has been shown that the expression of the murine blk gene is B-lineage restricted and developmentally regulated (2,7). The murine blk gene is expressed in pre-B through mature B-cell stages of differentiation but not in plasma cells. Our studies and others (14) have demonstrated the expression of the human blk gene in B-cell lines representing all stages of differentiation (pre-B through plasma cells), and in at least one T-cell line (Jurkat), but not in any of the non-lymphoid cell lines examined. The expression of human blk in the early stages of T-cell development has been reported previously (14). The function of blk in these T-cells is not understood. It is possible that human blk may play a role in signal transduction early in T-cell development.
In most of the human B-cell lines examined in this study, we found two blk transcripts (2.2 and 2.5 kb). Interestingly, these two transcripts appeared to be differentially expressed in the various cell lines. Some cell lines expressed higher levels of the 2.2-kb message as compared to the 2.5-kb transcript, others expressed relatively equal levels of both transcripts, while still others expressed only the 2.2-kb message. To begin to understand the mechanism(s) by which these two blk transcripts are differentially regulated, we examined their structural differences using Northern and PCR analyses.
We have previously described the isolation of a human blk cDNA clone (24). This cDNA corresponds in length to the 2.5-kb blk transcript. Northern analysis with a cDNA probe containing protein-coding sequences detected both the 2.2-and the 2.5-kb transcripts. However, a probe containing the first 210 bp (from Ϫ386 to Ϫ177) of the 5Ј-untranslated region of the cDNA clone specifically hybridized to the 2.5-kb RNA, but not to the 2.2-kb RNA. 2 PCR analyses, using an upstream primer that hybridizes within this 210-bp 5Ј-untranslated region and a downstream primer within exon 3, demonstrated amplification products only in cell lines that expressed the 2.5-kb transcript. In addition, primer extension experiments identified one major transcription initiation site, designated ϩ1, as well as two clusters of multiple transcription initiation sites, one located around the major start site at ϩ1 and another located 300 -400 bp upstream of this site. These data suggest that the 2.2-and 2.5-kb blk transcripts initiate at the ϩ1 and Ϫ300 transcription start site clusters, respectively, and that the two transcripts differ in the length of their 5Ј-untranslated regions by approximately 300 bp.
These observations on the structural differences between the two blk transcripts and the differential expression of these transcripts in different cell lines suggest that the human blk gene may contain two promoters that are regulated by different trans-acting factors. To investigate this possibility, the 5Јflanking region of the human blk gene was sequenced and examined for potential transcriptional regulatory elements surrounding the two clusters of transcription start sites. As with other src-family member genes (1), the 5Ј-flanking sequence of the human blk gene lacks TATA box elements near the transcriptional start sites. However, two clusters of possible binding sites for B-cell-specific transcription factors were found. One cluster, located near the upstream group of transcription start sites, contains X-box, CRE, and Y-box motifs. Interestingly, the spacing between these elements is similar to that in the MHC class II promoter (31,43). The molecular mechanisms controlling transcription of MHC class II genes have been studied extensively, and the nuclear factors binding to these sequences have been identified (44 -47). These boxes all contribute to the B-cell-specific expression of MHC class II genes. Another DNA-binding sequence that may contribute to the B-cell-specific expression of blk is the BSAP site at position Ϫ404 to Ϫ391 (see Fig. 2). Previous studies on the CD19 (10) and murine blk promoters (7,13) have shown the presence of a site recognized by the B-lymphoid transcription factor BSAP. This factor, like Sp1 and GCN4 (48), may play a role in activating transcription from TATA-less promoters.
Another cluster of sequences, located just upstream of the major transcription start site at ϩ1, contains E, PU.1, and PEA3 boxes. These sequences are present in the promoter and enhancer regions of immunoglobulin genes and are important for the B-cell-specific expression of these genes (49 -52). A binding site for the general transcription factor Sp1 (35) is also located in this region. Sp1 sites are often found in TATA-less promoters. Whether cooperation of Sp1 with the other binding factors (E, PU.1, and PEA3) is necessary for the activity of the blk promoter requires further investigation.
To test the functional activity of the 5Ј-flanking region of the human blk gene, a series of luciferase reporter gene plasmids were constructed and transfected into various cell lines. Results from these assays demonstrated that a 2.3-kb DNA fragment from the 5Ј-flanking region of the human blk gene contains tissue-specific promoter function (Table II). This fragment drives transcription in Daudi and SUP-B8 B-cell lines and Jurkat T-cells but not in T-47D breast cancer cells or DU 145 prostate cancer cells. Deletion analysis of this fragment in Daudi cells demonstrated three functionally important regions.
One region (Ϫ2258 to Ϫ1628) does not stimulate transcription independently but does act as an enhancer element. The combination of this region with promoter P1 or P2 (described below) increased transcription activity. However, this region also enhanced expression from the SV40 promoter approximately 3-fold in lymphoid and non-lymphoid cell lines, indicating that this is not a tissue-specific enhancer. We have identified a TG repeat element within this region as being FIG. 5. Deletion analysis of the human blk 5 regulatory region. The schematic diagram represents the human blk promoter from Ϫ2261 to ϩ75. The two clusters of transcription initiation sites are indicated by arrows. Putative nuclear factor-binding sites and restriction sites used to construct the deletion plasmids are also shown. The region of the blk promoter contained in each construct is indicated by the bars at the left, and the open box represents the luciferase coding sequence. Daudi cells were cotransfected with the indicated luciferase plasmids and pRSSV-CAT. Activity is presented relative to the pGL2-basic vector after normalization for transfection efficiency. Data represent the mean ϩ S.E. from four independent experiments. responsible for the enhancer function. Poly(dT-dG) is capable of forming right-handed Z-DNA and has been reported to modulate promoter activity from a distance (41,42). Recently, the study of Z-DNA-binding proteins has suggested that the TG element is bound by the high mobility group proteins (53). These proteins contain an high mobility group protein domain that appears to be able to bend or loop DNA to achieve the correct conformation for transcription and various classes of DNA rearrangement (54 -56). The finding of a TG element within the 5Ј-flanking region of the blk gene suggests that the expression of blk may be under such influences. The other two regions with functional activity, Ϫ1628 to Ϫ338 and Ϫ338 to ϩ75, are able to promote transcription independently. Deletion of either region only partially reduces transcription of the reporter gene. Removal of both regions completely abolished the transcriptional activity. These regions each contain one of the clusters of transcription initiation sites as well as a cluster of potential transcription factor-binding sites as described above. These results are consistent with the hypothesis that the blk gene contains two promoters. We have designated the putative promoter between Ϫ338 and ϩ75 as "P1" and the promoter between Ϫ1628 to Ϫ338 as "P2". Two promoters controlling single gene expression are also present in another member of the src family, lck (39). Similarly, the mouse Thy-1.2 glycoprotein gene (57), ␣-amylase gene (58) rat ␤ B (59) gene, and B-50/GAP-43 gene (60) are also transcribed from two promoters.
The presence of two promoters in the blk gene was further supported by the data from electrophoretic mobility shift assays. The P2 promoter, which drives the expression of the 2.5-kb blk transcript, contains a possible BSAP-binding site. We have demonstrated that the expression of the 2.5-kb transcript is correlated with the presence of a protein that binds to a probe containing this BSAP site. In contrast, cell lines that express the 2.2-kb blk transcript but lack the 2.5-kb transcript, lack this binding protein. These results suggest that the expression of the 2.5-kb transcript is regulated by BSAP or a related protein whereas the expression of the 2.2-kb transcript is not under regulation by this factor.
In summary, we have reported studies to support the conclusion that the human blk gene is expressed from two distinct promoters. The presence of different clusters of known B-cellspecific regulatory motifs within these two promoter regions and the different expression patterns of the two blk transcripts in various B-cell lines suggest that these promoters may be subject to regulation by different trans-acting factors. Consistent with this hypothesis is our identification of a protein that binds to the BSAP site in promoter P2 and the demonstration that the presence of this protein is correlated with the expression of the 2.5-but not the 2.2-kb blk transcript. The human blk gene is expressed throughout B-cell development, from very immature B-cells (as well as immature T-cells) to plasma cells. Two distinct blk promoters may be required to allow efficient transcription of the blk gene throughout B-cell development, since different trans-acting factors may be present at different stages of differentiation. Investigations are underway to identify these factors and to further elucidate the regulatory mechanisms controlling blk expression in human B lymphocytes.