Transcriptional regulation of the gene encoding the human C-type lectin leukocyte receptor AIM/CD69 and functional characterization of its tumor necrosis factor-alpha-responsive elements.

The human activation antigen CD69 is a member of the C-type animal lectin superfamily that functions as a signal-transmitting receptor. Although the expression of CD69 can be induced in vitro on cells of most hematopoietic lineages with a wide variety of stimuli, in vivo it is mainly expressed by T-lymphocytes located in the inflammatory infiltrates of several human diseases. To elucidate the mechanisms that regulate the constitutive and inducible expression of CD69 by leukocytes, we isolated the promoter region of the CD69 gene and carried out its functional characterization. Sequence analysis of the 5′-flanking region of the CD69 gene revealed the presence of a potential TATA element 30 base pairs upstream of the major transcription initiation site and several putative binding sequences for inducible transcription factors (NF-κB, Egr-1, AP-1), which might mediate the inducible expression of this gene. Transient expression of CD69 promoter-based reporter gene constructs in K562 cells indicated that the proximal promoter region spanning positions −78 to +16 contained the cis-acting sequences necessary for basal and phorbol 12-myristate 13-acetate-inducible transcription of the CD69 gene. Removal of the upstream sequences located between positions −78 and −38 resulted in decreased promoter strength and abolished the response to phorbol 12-myristate 13-acetate. We also found that tumor necrosis factor-α (TNF-α) is capable of inducing the surface expression of the CD69 molecule as well as the promoter activity of fusion plasmids that contain 5′-flanking sequences of the CD69 gene, suggesting that this cytokine may regulate in vivo the expression of CD69. In addition, cotransfection experiments demonstrated that the CD69 gene promoter can be activated by the NF-κB/Rel family members c-Rel and RelA. The deletion of the sequence spanning positions −255 to −170 abolished both the response to TNF-α and the transactivation by NF-κB. These results indicate that the NF-κB-binding site located at position −223 is necessary for the TNF-α-induced expression of the CD69 gene. Mobility shift assays showed that the two NF-κB motifs located in the proximal promoter region (positions −223 and −160) bind various NF-κB-related complexes, including the heterodimers p50/RelA and p50/c-Rel and homodimers of p50 (KBF-1) and RelA. Our findings help to explain the regulated synthesis of CD69 in vivo and suggest that TNF-α has a key role in the expression of this molecule at sites of chronic inflammation.

ity shift assays showed that the two NF-B motifs located in the proximal promoter region (positions ؊223 and ؊160) bind various NF-B-related complexes, including the heterodimers p50/RelA and p50/c-Rel and homodimers of p50 (KBF-1) and RelA. Our findings help to explain the regulated synthesis of CD69 in vivo and suggest that TNF-␣ has a key role in the expression of this molecule at sites of chronic inflammation.
The human activation inducer molecule (AIM/CD69) is a phosphorylated disulfide-linked 27/33-kDa transmembrane homodimeric glycoprotein (Sá nchez-Mateos and Sá nchez- . The CD69 molecule is rapidly expressed on the surface of T-lymphocytes upon in vitro activation with a wide variety of agents, including anti-CD3/T cell receptor and anti-CD2 mAbs, 1 activators of protein kinase C, and phytohemagglutinin (Cebrián et al., 1988;Hara et al., 1986). Similarly, the expression of CD69 is inducible on the surface of NK cells, B-lymphocytes, and eosinophils (Hartnell et al., 1993;Lanier et al., 1988). In contrast, CD69 is constitutively expressed in vivo on platelets and on a small percentage of resident T-and B-cells of different lymphoid tissues (Sá nchez-Mateos et al., 1989;Testi et al., 1990). Although a physiologic ligand for CD69 has not been identified so far, experiments with specific mAbs indicate that this antigen functions as a signal-transmitting receptor. Signals triggered by CD69 mAbs in T-lymphocytes include increase in intracellular calcium concentration and result in the synthesis of different cytokines and their receptors, enhancement of the expression of c-myc and c-fos protooncogenes, and cell proliferation (Cebrián et al., 1988;Nakamura et al., 1989;Santis et al., 1992;Testi et al., 1989;Tugores et al., 1992). In NK cells and platelets, CD69 also acts as a triggering molecule, being involved in the redirected target cell lysis by interleukin-2-activated NK cells (Moretta et al., 1991) and in the induction of platelet aggregation, Ca 2ϩ influx, and hydrolysis of arachidonic acid (Testi et al., 1990).
The molecular cloning of cDNAs encoding human and mouse CD69 revealed that this antigen is a member of the Ca 2ϩdependent (C-type) animal lectin superfamily of type II transmembrane receptors (Hamann et al., 1993;Ziegler et al., 1993). This superfamily includes the human NKG2, rat and mouse NKR-P1, and mouse Ly-49 families of NK cell-specific antigens as well as the low avidity IgE receptor (CD23), the Kupffer cell receptor, and the hepatic asialoglycoprotein receptor (Drickamer, 1993;Yokoyama, 1993). Characterization of the structure of genes encoding human and mouse CD69 further evidenced that CD69 is evolutionarily related to these C-type lectin receptors (Santís et al., 1994;Ziegler et al., 1994). In addition, the CD69 gene is clustered with other C-type lectin-encoding genes within a genetic region named the NK cell complex (Drickamer, 1993;López-Cabrera et al., 1993;Yokoyama, 1993;Ziegler et al., 1994).
The CD69 antigen is undetectable on peripheral blood lymphocytes; however, it is expressed at high levels by the majority of T-cells in the inflammatory cell infiltrates of several human diseases such as rheumatoid arthritis and chronic viral hepatitis (García-Monzón et al., 1990;Laffón et al., 1991), suggesting that inflammatory cytokines may be involved in CD69 expression. In this context, it has been reported that a large amount of TNF-␣ is produced by human hepatocytes in chronic viral hepatitis (Gonzá lez-Amaro et al., 1994). Thus, TNF-␣ and other cytokines may play a key regulatory role in the inducible expression of the CD69 antigen in vivo.
Northern blot analysis demonstrated that the constitutive and inducible expression of the CD69 molecule is regulated at the transcriptional level (Hamann et al., 1993Ziegler et al., 1993Ziegler et al., , 1994. To determine the molecular basis for the pattern of CD69 expression, we have isolated the 5Ј-region of the CD69 gene and analyzed its inducible promoter activity. We describe herein that the upstream sequence of the CD69 gene functions as a PMA-inducible promoter element. In addition, we report that TNF-␣ is capable of inducing both the surface expression of the CD69 antigen and the promoter activity of the 5Ј-region of the CD69 gene. The presence of NF-B motifs within the proximal promoter region of the CD69 gene may account for the TNF-␣-inducible promoter activity.

Isolation of Genomic Clones Containing the 5Ј-Region of the CD69
Gene-A total of 5 ϫ 10 4 clones from an EcoRI-digested human chromosome 12-specific library in Charon 21A (ATCC 57727) were screened as described (Santís et al., 1994) using a 456-bp 32 P-labeled probe from the 5Ј-end of the CD69 cDNA (nucleotides 12-467) . After four rounds of screening, three positive clones were isolated. These clones were then hybridized with 32 P-labeled oligonucleotides derived from the 5Ј-untranslated (nucleotides 12-31) and transmembrane-encoding (nucleotides 209 -228) sequences of the CD69 cDNA . Since these clones hybridized only with the oligonucleotide derived from the 5Ј-untranslated sequence, they contained DNA inserts corresponding to the 5Ј-portion of the CD69 gene.
Restriction analysis of the clones indicated that they possessed the same 3-kbp EcoRI insert, which was subcloned into the pBluescript plasmid (Stratagene, La Jolla, CA). DNA sequencing was performed by the dideoxy termination method either by subcloning restriction fragments into pBluescript vectors or by direct oligonucleotide-primed DNA sequencing with internal primers.
Plasmid Constructions, Transfections, and Luciferase Assays-Restriction endonuclease site mapping and partial DNA sequencing of the cloned 3-kbp insert revealed that it contained the first exon, a part of the first intron, and a 2.1-kbp fragment containing the putative upstream regulatory region of the CD69 gene. Five restriction fragments of the latter region (EcoRI-SacI, PvuII-SacI, HindIII-SacI, BglII-SacI, and XbaI-SacI of 2.1, 1.4, 0.64, 0.49, and 0.27 kbp, respectively) were inserted upstream of the luciferase gene in the pXP2 plasmid (Nordeen, 1988). Three additional deletion fragments of 186, 94, and 54 bp were generated by polymerase chain reaction and cloned into the same reporter plasmid.
K562 erythroleukemic cells were transiently transfected with 4 g of each recombinant plasmid using 10 g of Lipofectin reagent (Life Technologies, Inc.) according to the manufacturer's recommendations. To normalize transfection efficiency, 1 g of pCMV␤-gal (CLONTECH, Palo Alto, CA), which contains the cytomegalovirus promoter ligated to the ␤-galactosidase gene, was included in each transfection. After 48 h of transfection, luciferase activity was determined in cell extracts containing identical ␤-galactosidase activity according to the instructions of the luciferase assay kit (Promega). Light emission was measured in a Lumat LB9501 luminometer (Berthold, Wildbad, Germany), and the results are expressed as relative light units. To analyze the effect of PMA and TNF-␣ on the CD69 promoter activity, half of the transfected cells were treated for 12-16 h with either 20 ng/ml PMA or 50 ng/ml human recombinant TNF-␣ (3.2 ϫ 10 7 units/mg; Wichem, Vienna). As positive controls for PMA and TNF-␣ stimulation, we used the plasmids pGL2P (Promega), in which the luciferase gene is under the control of the SV40 early promoter, and pKBF-Luc (a gift from Dr. A. Israël, Institut Pasteur, Paris), which contains a trimer of the NF-B motif of the H-2K b gene (Yano et al., 1987) upstream of the herpes simplex virus thymidine kinase gene promoter.
To obtain K562 stable transfectants, cells were cotransfected by electroporation with 50 g of pAIM1.4-Luc and 10 g of pSV2-neo as described previously (Nueda et al., 1993), selected in the presence of G418 (1 mg/ml), and cloned by limiting dilution. To analyze the transactivation of the CD69 promoter by NF-B/Rel family members, K562 cells were cotransfected, in the presence of 20 g of Lipofectin, with 5 g of CD69 promoter-derived plasmids and 5-10 g of the expression vector pRc/CMV-p50, pRc/CMV-c-Rel, or pRc/CMV-RelA (kindly provided by Dr. A. Israël), which contained the full-length cDNA encoding each protein, into the pRc/CMV plasmid (Invitrogen, San Diego, CA) (Le Bail et al., 1993).
Nuclear Extracts and Electrophoretic Mobility Shift Assays-COS-7 cells were cotransfected by the calcium phosphate precipitation procedure with the expression vectors pRc/CMV-p50 (0.5 g/ml) and pRc/ CMV-RelA (1.0 g/ml). Small-scale nuclear extracts from these cells and untransfected K562 cells were prepared according to a procedure described elsewhere (Schreiber et al., 1989). Nuclear extracts from stimulated K562 cells were obtained after a 6-h treatment with PMA (20 ng/ml) or human recombinant TNF-␣ (50 ng/ml).

RESULTS
Structure of the 5Ј-Region of the CD69 Gene-The gene encoding the human CD69 antigen has previously been mapped to chromosome 12p13-p12 . To isolate the 5Ј-regulatory region of the CD69 gene, an EcoRIdigested chromosome 12-specific library was screened with a 456-bp specific probe that contained the 5Ј-end of CD69-encoding cDNA. Three positive clones containing the same 3-kbp insert were obtained. Restriction mapping analysis and hybrid-ization experiments with oligonucleotides derived from the 5Јuntranslated and transmembrane-encoding sequences of the CD69 cDNA indicated that the cloned DNA fragment contained the 5Ј-portion of CD69 gene (Fig. 1A). Partial sequencing of the genomic DNA insert revealed the presence of an exon that includes the whole reported 5Ј-untranslated region and that codes for the first 21 amino acid residues of the N-terminal region of the CD69 protein (Fig. 1B). The coding sequence of the DNA cloned was interrupted by the first intron of the gene (Fig.  1B). Upstream of the first exon, a 2.1-kbp fragment containing the putative CD69 gene promoter region was identified (Fig.  1A).
The major transcription initiation site of the CD69 gene has been located 81 nucleotides upstream from the translation start codon (Santís et al., 1994). Nucleotide sequence analysis 1050 bp upstream from the CAP site revealed the presence of a canonical TATA box at position Ϫ30 and a GC-rich sequence at position Ϫ52. In addition, three potential NF-B/Rel-binding sequences were identified at positions Ϫ160, Ϫ223, and Ϫ373 (Fig. 1B). The two proximal NF-B-binding sites (kB-1 and kB-2 in Fig. 1B) were identical to those found in the gene promoters of c-myc and interleukin-6, respectively (Baeuerle, 1991), whereas the most upstream NF-B motif (kB-3 in Fig.  1B) was similar to that found in the major histocompatibility complex class I (H-2K b ) gene promoter (Yano et al., 1987).
Functional Analysis of the CD69 Gene Promoter-To functionally characterize the 5Ј-flanking region of the CD69 gene, eight genomic fragments containing the transcription initiation site were ligated to the luciferase gene (Fig. 2), and their basal and PMA-inducible promoter activities were assessed in K562 cells, which express CD69 in an inducible manner. Comparison of the relative promoter activities of the different constructs indicated that the progressive removal of 5Ј-sequences up to position Ϫ78 did not affect significantly the uninduced promoter activity, suggesting that the 94-bp fragment, spanning positions Ϫ78 to ϩ16, contained the cis-acting elements necessary for basal promoter activity. PMA treatment of cells transfected with constructs up to position Ϫ78 resulted in augmented promoter activities, with inductions that ranged from 15 to 125-fold (Fig. 2). Further deletion of upstream sequences up to position Ϫ38 resulted in decreased promoter strength and abolishment of the response to PMA (Fig. 2). These results demonstrate that PMA up-regulates the promoter activity of the CD69 gene in hemopoietic cells, a phenomenon that is in agreement with the described pattern of expression of the CD69 molecule. Moreover, the sequence located between positions Ϫ78 and Ϫ38 contained, at least in part, the cis-acting elements involved in the PMA-inducible promoter activity. Interestingly, this 41-bp fragment possesses a GC-rich sequence First exon nucleotides are underlined, and the major transcription initiation site is denoted by ϩ1. The first intron sequence appears in lower-case letters. Shaded areas correspond to consensus binding sequences for transcription factors. The TATA box is indicated by double underlining. Prediction of putative transcription factor-binding sites was carried out by the C-coded program SITIOS (Dr. M. A. Vega, Instituto López Neyra, Consejo Superior de Investigaciones Científicas), which includes Release 5.0 of D. Ghosh's transcription factor data base (Ghosh, 1991).
FIG. 2. Functional analysis of the CD69 promoter. The schematic representation of the CD69 promoter-based reporter gene constructs is shown on the left, and the upstream region of the CD69 gene is represented at the top. The positions of the first exon and the first intron are indicated by filled and hatched boxes, respectively. The nomenclature of the deletion plasmids is based on the most 5Јnucleotide of the CD69 gene sequence present, and its position is denoted relative to the transcription initiation site (position ϩ1). The basal and PMA-induced promoter activity of the 5Ј-region of the CD69 gene was determined by transient expression of luciferase gene-based constructs in K562 cells. Each transfection was carried out at least four times, and the data from a representative experiment are shown on the right. INDUCT., induction.
( Fig. 1B) that could be recognized by the immediate-early growth-response transcription factor Egr-1/Krox-24, which has been described to be inducible by PMA (Krä mer et al., 1994).
TNF-␣ Treatment Induces the Expression of the CD69 Antigen and the Activity of the CD69 Gene Promoter-The restricted expression of the CD69 antigen in vivo, at places where inflammation occurs (García-Monzón et al., 1990;Laffón et al., 1991), and the presence of three putative NF-B-binding sites within the CD69 promoter region led us to study the effect of TNF-␣, a cytokine that promotes inflammation and induces NF-B (Baeuerle, 1991;Baeuerle and Henkel, 1994), on the expression of the CD69 gene. First, we analyzed by flow cytometry the surface expression of the CD69 antigen on K562 cells treated with human recombinant TNF-␣. As shown in Fig. 3A, TNF-␣ treatment resulted in an increase in CD69 expression, although to a lesser extent compared with PMA-treated cells. To determine whether the increase in CD69 expression in response to TNF-␣ was mediated by changes in the CD69 promoter activity, K562 cells were stably transfected with the CD69 promoter-derived construct pAIM1.4-Luc (Fig. 2), and the TNF-␣-induced promoter activity was analyzed in three independent clones. Comparison of the luciferase activity produced by this plasmid in unstimulated and TNF-␣-stimulated transfectants showed that the promoter activity was augmented 4 -5 times upon cytokine treatment (Fig. 3B). This result demonstrates that the CD69 promoter contains TNF-␣-responsive elements that are accounting, at least in part, for the TNF-␣-mediated induction of CD69 expression.
It is known that the NF-B/Rel family of transcription factors plays an important role in the cytokine induction of many cellular and viral genes (Baeuerle, 1991). To characterize the functional role of NF-B-related proteins in the activation of the CD69 gene promoter, K562 cells were cotransfected with the construct pAIM1.4-Luc and expression vectors encoding the p50, RelA, and c-Rel members of the NF-B/Rel family. As shown in Fig. 3C, p50 was unable to transactivate the promoter, whereas c-Rel and RelA efficiently induced the promoter activity (4 -5-and 12-18-fold, respectively). Interestingly, cotransfection with a combination of p50-and RelA-encoding plasmids did not result in activation of the CD69 promoter (data not shown).
The B-2 Site Is Responsible for TNF-␣ Inducibility of the CD69 Gene Promoter-To identify the cis-acting sequences of the CD69 gene promoter involved in the response to TNF-␣, K562 cells were transiently transfected with the different CD69 promoter-derived constructs, which contained three, two, one, or none NF-B motifs of the promoter (Figs. 1B and 2). Comparison of the luciferase activity produced by the different plasmids in unstimulated and TNF-␣-stimulated cells showed that deletion of the sequences located between positions Ϫ480 and Ϫ255, which eliminated the B-3 site, did not significantly affect the response to TNF-␣, whereas further removal of the sequences from positions Ϫ255 to Ϫ170, which deleted the B-2 site, abolished the induction by TNF-␣ (Fig. 4A). It is interesting to note that this unresponsive construct (pAIM170-Luc) still conserved the B-1 site. These results indicate that the response to TNF-␣ is mediated, at least in part, by the binding of NF-B/Rel-related proteins to the B-2 site.
To confirm this point, K562 cells were cotransfected with the CD69 promoter constructs and the expression vector pRc/CMV-RelA. As shown in Fig. 4B, the pattern of transactivation of the different CD69 promoter fragments by RelA correlated with the response of these constructs to TNF-␣ (Fig. 4A). Therefore, deletion of the sequences containing the B-2 site greatly diminishes the transactivation of the CD69 promoter by RelA.
Members of the NF-B/Rel Family Bind to the B-2 and B-1 Motifs of the CD69 Gene-The data presented above indicate that the TNF-␣-induced expression of the CD69 gene is mainly mediated by the B-2 site. However, since the shortest promoter fragment responsive to TNF-␣ (pAIM255-Luc) contained the B-2 and B-1 sites, the contribution of the latter NF-B motif to the overall TNF-␣ response cannot be ruled out. To analyze whether these two putative NF-B-binding sites of the CD69 promoter were capable of binding NF-B/Rel family members, gel retardation assays were performed using as probes two double-stranded oligonucleotides (CD69-B-1 and CD69-B-2) containing these motifs. First, these probes were incubated with nuclear extracts from COS-7 cells cotransfected with the expression vectors pRc/CMV-p50 and pRc/CMV-RelA. Two major DNA-protein complexes were observed with both oligonucleotides when they were incubated with nuclear extracts from transfected COS-7 cells (Fig. 5A). These bands corresponded to the homodimer of p50 (KBF1) and the heterodimer p50/RelA (NF-B) and were not observed using extracts from untransfected cells (Fig. 5A, lane 1). Competition assays with an excess of oligonucleotide KBF, which contains the NF-B motif of the H-2K b gene, completely prevented the formation of the two DNA-protein complexes. In contrast, the addition of an excess of the unrelated oligonucleotides AP-1 and OCT-1 did not compete the binding to both probes (Fig.  5A).
To characterize the NF-B/Rel-related proteins that bind to the NF-B sites of the CD69 promoter in CD69-expressing cells, the two oligonucleotide probes were incubated with nuclear extracts prepared from untreated K562 cells and from K562 cells treated with either PMA or TNF-␣. Four inducible DNA-protein complexes (a-d) were observed with both oligonucleotide probes when they were incubated with extracts from PMA-and TNF-␣-treated K562 cells (Fig. 5B). The specific DNA-protein complexes observed with both probes displayed identical electrophoretic mobility, suggesting that they bind the same nuclear factors. It is interesting to note that oligonucleotide CD69-B-1, which bound a lesser amount of nuclear proteins, contained one mismatch with respect to the consensus NF-B-binding site (5Ј-GGGRNTYYC-3Ј), whereas oligonucleotide CD69-B-2 perfectly matched the consensus sequence (Baeuerle, 1991). In both cases, the addition of an excess of unlabeled specific oligonucleotide to the binding reaction completely abolished the formation of the inducible DNA-protein complexes (Fig. 5B). Similarly, an equal amount of the heterologous oligonucleotide VP9, which contained two NF-B-binding sites from the vascular cell adhesion molecule-1 gene promoter, efficiently competed the specific complexes (Fig. 5B). In contrast, the formation of these DNA-protein complexes was not blocked by the addition of a heterologous competitor that lacked NF-B binding sequences (Fig. 5B).
To identify the nature of the NF-B/Rel family members that bind to the NF-B motifs of the CD69 gene promoter, the binding reactions were preincubated with antiserum specific to p50, RelA, c-Rel, or p52 (Fig. 5C). The anti-p50 antiserum induced the disappearance or reduced the intensity of complexes b-d. The anti-RelA antiserum blocked the formation of complexes a and b, whereas the anti-c-Rel antiserum inhibited only complex c. In contrast, the anti-p52 antiserum did not interfere with the formation of any of these complexes. These results indicate that the slower migrating complex (complex a) corresponds to a homodimer of RelA, which is more easily detected when using oligonucleotide CD69-B-2. Then, from top to bottom, the complexes correspond to p50/RelA (complex b) and p50/c-Rel (complex c) heterodimers and to a p50 homodimer (complex d). Taken together, these results confirm that the NF-B motifs of the proximal CD69 promoter are capable of binding to inducible NF-B/Rel factors, and therefore, they may be involved in the TNF-␣-mediated induction of CD69 gene expression.

DISCUSSION
The transcription of the CD69 gene appears to be tightly regulated in vivo, as it is almost exclusively expressed at sites were inflammation takes place, suggesting that inflammatory cytokines may participate in the control of the expression of this gene. In this report, we describe the isolation and functional characterization of the human CD69 gene promoter region. We have focused mainly on the identification of the cisacting sequences and the nuclear factors involved in the inducible expression of the CD69-encoding gene.
Deletion analysis of the 5Ј-flanking region of the CD69 gene has allowed the identification of a proximal fragment of 94 bp (nucleotides Ϫ78 to ϩ16) that is sufficient to govern basal and PMA-induced promoter activity. This proximal promoter domain contains a canonical TATA box and a GC-rich sequence that could be a target for Sp1, a ubiquitous and constitutive transcription factor that binds to the promoter of many genes, and Egr-1/Krox-24, a zinc-finger transcription factor whose expression can be induced by various agents, including phorbol esters, and that has been implicated in the activation of T-and B-lymphocytes (McMahon and Monroe, 1995;Krä mer et al., 1994;Pérez-Castillo et al., 1993). Since the synthesis of CD69encoding mRNA is rapidly induced upon PMA stimulation of FIG. 4. Activation of the CD69 promoter by TNF-␣ is achieved through the sequence spanning positions ؊255 to ؊170, which contains the NF-B-2 motif. A, CD69 promoter-based luciferase plasmids were transfected into K562 cells, and half of the transfected cells were treated with TNF-␣ for 16 h. Each transfection was carried out five times, and a representative experiment is shown. B, 5 g of each of the CD69 promoter-derived constructs were cotransfected into K562 cells with 10 g of expression vector pRc/CMV-RelA. Numbers above the bars indicate -fold transactivation over the values obtained by cotransfection with the empty vector pRc/CMV. the cells , additional pre-existing transcription factors should mediate the PMA-induced expression of the CD69 gene to ensure a rapid transcriptional response. In this context, computer-aided analysis of the 5Ј-flanking region of the CD69 gene revealed the presence of three NF-B motifs (positions Ϫ160, Ϫ223, and Ϫ374) and at least one putative AP-1-binding site (position Ϫ956), which may cooperate with the proximal cis-acting elements in the response to phorbol esters.
We demonstrate herewith that TNF-␣ is capable of inducing the expression of the CD69 antigen and that this induction is mediated by an increase in the CD69 promoter activity. Cotransfection experiments with CD69 promoter-derived reporter constructs and NF-B-encoding vectors show that this promoter is transactivated by members of the NF-B/Rel family, specially by RelA (formerly p65). Thus, the TNF-␣-induced expression of the CD69 gene may be mediated by the binding of NF-B/Rel proteins to one or more NF-B motifs of the CD69 promoter. This point is further supported by the fact that deletion of the sequence that contains the B-2 motif (nucleotides Ϫ255 to Ϫ170) completely abolishes the response to TNF-␣ and markedly reduces RelA-mediated transactivation. Since no additional TNF-␣-responsive elements are present in this sequence, our results strongly indicate that the B-2 motif plays a key role in the TNF-␣-induced promoter activity. However, to define unambiguously the contribution of B-2 and the other NF-B motifs to the response to TNF-␣, we are currently performing site-directed mutagenesis of these motifs in the context of the intact promoter.
Our results also demonstrate that the two most proximal NF-B motifs of the CD69 promoter (B-2 and B-1) bind four PMA-and TNF-␣-inducible NF-B/Rel-related complexes. The higher affinity of these proteins for the B-2 sequence further emphasizes the functional role of this motif and may explain why constructs containing only the B-1 site are unresponsive to TNF-␣. Antisera directed against the different NF-B/Rel proteins were used to identify the family members present in the DNA-protein complexes detected with the NF-B motifderived probes. These experiments revealed that the DNAbinding activities consisted of a RelA homodimer, p50/RelA and p50/c-Rel heterodimers, and a p50 homodimer. The most prominent complexes were composed of p50/RelA and p50/c-Rel heterodimers. The detection of RelA homodimer binding to the B-2 motif, a complex that is not easily detected with the NF-B sites of other promoters (Ganchi et al., 1993), is consistent with the observation that the sequence of this motif matches perfectly with the consensus binding sequence for RelA, 5Ј-GGGRNTTTCC-3Ј (Kunsch et al., 1992).
We have recently demonstrated that the expression of the CD69 gene is regulated at the post-transcriptional level by a rapid degradation pathway associated with AU-rich sequence motifs (Santís et al., 1995). This mechanism of regulation has FIG. 5. The B-2 and B-1 motifs of the proximal CD69 promoter region are recognized by NF-B/Rel family members. A, mobility band shift assays were performed with oligonucleotide probes CD69-B-1 (lanes 1-5) and CD69-B-2 (lanes 6 -9), which contained the putative NF-B motifs located at positions Ϫ160 and Ϫ223, respectively. Nuclear extracts from COS-7 cells (lane 1) and from COS-7 cells transfected with pRc/CMV-p50 and pRc/CMV-RelA (lanes 2-9) were incubated with 0.5 ng of each double-stranded oligonucleotide probe. Competitor oligonucleotides were added at 50-fold molar excess and included oligonucleotide KBF, which contains the NF-B motif of the promoter of the H-2K b gene (lanes 3 and 7) and the unrelated oligonucleotides OCT-1 (lanes 4 and 8) and AP-1 (lanes 5 and 9). B, oligonucleotide probes CD69-B-2 (lanes 1-6) and CD69-B-1 (lanes 7-12) were incubated with nuclear extracts from uninduced K562 cells (lanes 1 and   7) and from K562 cells stimulated with PMA (P; lanes 2 and 8) or TNF-␣ (lanes 3-6 and 9 -12). Competitor oligonucleotides were added at 50fold molar excess and included the specific competitors (lanes 5 and 11), oligonucleotide VP9 (which contains two NF-B-binding sites from the vascular cell adhesion molecule-1 gene promoter) (lanes 6 and 12), and the unrelated oligonucleotide OCT-1 (lanes 4 and 10). Specific DNAprotein complexes (a-d) are indicated. C, nuclear extracts from uninduced K562 cells (lanes 1 and 7) or from K562 cells stimulated with TNF-␣ (lanes 2-6 and 8 -12) were preincubated either with preimmune serum (PRE; lanes 1, 2, 7, and 8) or with antiserum specific to p50 (lanes 3 and 9), RelA (lanes 4 and 10), c-Rel (lanes 5 and 11), or p52 (lanes 6 and 12) prior to the addition of probe CD69-B-2 (lanes 1-6) or CD69-B-1 (lanes 7-12). The DNA-protein complexes obtained with probe CD69-B-1 were exposed three times longer than those obtained with CD69-B-2 to get a similar intensity of the bands. been found in many genes involved in inflammatory and activation responses. Most of these genes code for cytokines and oncoproteins, which are implicated in the initial events leading to activation and proliferation of the cells. In addition, the transcription of these genes appears to be highly regulated by the NF-B motifs of their promoters (Baeuerle, 1991;Baeuerle and Henkel, 1994). Therefore, our findings, which highlight the functional relevance of the NF-B motifs of the CD69 gene, further support the idea regarding a general mechanism involved in the control of expression of activation-associated genes during the early phase of the immune response.
The expression of CD69-encoding mRNA can be easily induced in most leukocytes by treatment "in vitro" with a wide range of stimuli such as PMA or anti-CD3/T cell receptor mAb Ziegler et al., 1994). In addition, the expression of CD69 on activated T-lymphocytes from cell infiltrates of various chronic inflammatory diseases such as rheumatoid arthritis and chronic viral hepatitis has been documented (García-Monzón et al., 1990;Laffón et al., 1991). It is also known that TNF-␣ is an important inflammatory mediator that is actively secreted by several cell types at inflammatory sites (Gonzá lez-Amaro et al., 1994;Vassalli, 1992). Our findings strongly suggest that the pro-inflammatory cytokine TNF-␣ may play a key role "in vivo" in the expression of CD69 by inflammatory cells. Interestingly enough, we have previously described that CD69 antigen is in turn capable of generating signals that induce the synthesis of TNF-␣ by lymphoid cells (Santís et al., 1992). Thus, it is feasible that, at sites of inflammation, a positive feed-back loop is established between the expression of CD69 and the production of TNF-␣. The putative ligand of CD69 should be clearly involved in this condition. The above phenomenon could have an important role in the perpetuation of several inflammatory diseases and could be related to the resistance of some patients with these conditions toward the current anti-inflammatory therapy.