Proc. Natl. Acad. Sci. USA Vol. 89, pp. 1875-1879, March 1992 Biochemistry The 65-kDa subunit of human NF-KB functions as a potent transcriptional activator and a target for v-Rel-mediated repression DEAN W. BALLARD*t, ERIC P. DIXON, NANCY J. PEFFER, HAL BOGERD, STEFAN DOERRE, BERND STEIN, AND WARNER C. GREENE* Howard Hughes Medical Institute, Department of Medicine and Department of Microbiology and Immunology, Duke University Medical Center, Durham, NC 27710 Communicated by Harold E. Varmus, November 6, 1991

ABSTRACT Molecular cloning of the polypeptide compo- (7, 18, 23). The p55 and p50 adducts contain NF-KB p50 and nent of the Rel-related human p75 nucleoprotein complex has a closely related , respectively (18), while the p85 revealed its identity with the 65-kDa (p65) subunit of NF-KB. adduct is composed of the c- protooncogene product Functional analyses of chimeric composed of NF-KB (c-Rel) (18), which has been shown to contain a functional p65 C-terminal sequences linked to the DNA-binding domain of transactivation domain (20, 24, 25). the yeast GAL4 polypeptide have indicated that the rinal 101 The ability of p75 to dimerize with NF-KB p50 suggested a amino acids of NF-#cB p65 comprise a potent transcriptional potential relationship to NF-KB p65 (18); however, the lack activation domain. Transient transfection of human T cells ofintrinsic DNA-binding activity reported for NF-KB p65 (21, with an expression vector encoding NF-KB p65, but not NF-tcB 22) implied that these two proteins were distinct. In the p50, produced marked transcriptional activation of a basal present study, we have cloned p75 cDNAs based on peptide promoter containing duplicated KB enhancer motifs from the sequence information derived from the affinity-purified pro- long terminal repeat of type 1 human immunodeficiency virus. tein and have demonstrated the complete identity of p75 and These stimulatory effects of NF-acB p65 were synergistically NF-KB p65 at the protein level. Transient transfection studies enhanced by coexpression of NF-ucB p50 but were completely show that this enhancer-binding protein functions as a pow- inhibited by coexpression of the v-rel oncogene product. To- erful transcriptional transactivator and a target for v-Rel- gether, these functional studies demonstrate that NF-KB p65 is mediated repression. a transactivating subunit of the heterodimeric NF-mcB complex and serves as one cellular target for v-Rel-mediated transcrip- MATERIALS AND METHODS tional repression. Isolation of p75 cDNA. Human p75 was partially purified The NF-KB complex plays a central role from HeLa cytosolic extracts (18) and further fractionated by in the activation of type 1 human immunodeficiency virus preparative SDS/PAGE (26) prior to digestion with cyanogen (HIV-1) and many immunologically relevant cellular genes bromide (18 hr at 25°C) (27). Cleavage products were frac- (1). Nuclear NF-KB DNA-binding activity is constitutively tionated by reverse-phase HPLC and analyzed on an Applied expressed in mature B cells (2) and induced in T cells by Biosystems gas-phase sequenator. A degenerate antisense phorbol esters (3, 4), tumor necrosis factor a (5, 6), and the oligonucleotide probe (5'-GTCATCIGTRTCIGGIAGRTAC- Tax protein of the type I human T-cell leukemia virus TGGAA-3') was designed (28) based on peptide B (see text) (HTLV-I) (7-9). Nuclear NF-KB expression is regulated in and used to screen (29) a human HL-60 cDNA library. part at a posttranslational level (3) involving its dissociation Nucleotide sequences of cross-hybridizing clones were de- from the cytoplasmic inhibitor IKB (10, 11). This dissociation termined by the dideoxy chain-termination method (30). reaction is apparently catalyzed by the phosphorylation of DNA Binding and Peptide Mapping Studies. UV crosslink- IKB, which in turn allows the rapid translocation ofNF-KB to ing analysis and peptide mapping were performed (18, 31, 32) the nucleus (10-12). using a palindromic variant of the KB enhancer of the Sequence analyses of human and murine cDNAs encoding interleukin 2 a-chain gene (5'-CAACGGCAGGG- the DNA-binding subunit (NF-KB p50) (13, 14) and the GAATTCCCCTCTCCTT-3'); this variant serves as a partic- IKB-binding subunit (NF-KB p65) (15, 16) of NF-KB have ularly potent binding site (18). Endoproteases used for pep- revealed extensive homology between the N-terminal 300 tide mapping (23) were added at an enzyme/protein ratio of amino acids of these two polypeptides, the v-Rel oncopro- 1:25. tein, and the dorsal gene product, a ventral morphogen in Expression Vectors. To produce chimeric GAL4 expression Drosophila (17). v-Rel specifically engages the KB enhancer vectors, C-terminal coding sequences of various Rel family and acts as a dominant negative repressor ofNF-KB in mature members were ligated downstream of the first 147 codons of T cells (18-20). However, the molecular basis for the tran- GAL4 in the pSG424 expression vector (33). The GAL4/ scriptional activation mediated by the NF-KB p5O/p65 com- VP16 expression vector has been described (33). The human plex or transcriptional repression mediated by v-Rel has c-rel cDNA clone (34) was generated by RNA amplification remained unknown. (18). All Rel expression vectors, including those encoding Although gel retardation analyses with purified NF-KB NF-KB p50 (amino acids 1-462; ref. 14), NF-KB p65 (16), and originally suggested that only NF-KB p50 had intrinsic DNA- v-Rel (35), were constructed in pCMV4 (36). The chimeric binding activity (21, 22), recent DNA/protein crosslinking gene encoding the NF-KB p50 Rel domain fused to the NF-KB studies with human T cell extracts have demonstrated four p65 C terminus was constructed by amplification with over- Rel-related KB-specific proteins that are expressed in a temporally biphasic manner (pS5/p75, early; pSO/p85, late) Abbreviations: CAT, chloramphenicol acetyltransferase; HIV, hu- man immunodeficiency virus; HTLV, human T-cell leukemia virus; LTR, long terminal repeat. The publication costs of this article were defrayed in part by page charge *Present address: Gladstone Institute of Virology and Immunology, payment. This article must therefore be hereby marked "advertisement" P.O. Box 419100, San Francisco, CA 94141-9100. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed at *.

1875 Downloaded by guest on September 30, 2021 1876 Biochemistry: Ballard et al. Proc. Natl. Acad. Sci. USA 89 (1992) lapping primers (37). A blunt-ended Esp I-EcoRI fragment A B encoding the C-terminal 214 amino acids of human NF-KB D r e Anti .4 p65 was fused to the HinclI site in the v-Rel coding sequence irnmune 275 ".. -D nvo: Ar.4 .... i..;--N - to create the v-Rel/NF-KB p65 expression vector. An ex- 2 29i4 5. pression vector encoding the C-terminal 226 amino acids of : 4 NF-KB p65 alone was constructed by amplification of the Ce corresponding 3' cDNA sequences with oligonucleotide primers that introduced optimal translation initiation se- quences (38). 9 7 4 Cell Transfections and Chloramphenicol Acetyltransferase S. (CAT) Assays. COS cells or human Jurkat T cells were transfected with plasmid DNA by the DEAE-dextran method 68` P (39) and were assayed for CAT activity as described (40, 41). Reporter plasmids containing either the full-length HIV-1 long terminal repeat (LTR) (42) or five GAL4 binding sites tR5 (pG5BCAT; ref. 33) have been described. The pBLCAT- 4 43 C' 46 4w based (43) constructs contained either two HIV-1 KB en- 4 .W-- hancer repeats (-105 to -79; KB-TATA-CAT) or five tan- 11 dem phorbol ester-responsive elements (-72 to -65; AP1- j.;,- I:,. 1-;.. <.- t, .:.

,pL -I' TATA-CAT) cloned immediately upstream of a basal S .9 l-'. albumin promoter (TATA-CAT) (44). I 0? 104 RESULTS Molecular Cloning of p75 and Its Relationship to NF-KB p65. FIG. 2. Human NF-KB p65 corresponds to a KB enhancer-binding To determine the primary structure of p75, microsequencing protein identical to p75. (A) Partially purified HeLa p75 (lanes 4 and 7) or extracts from COS cells transfected with a pCMV4 expression was performed on three peptides derived from the affinity- vector lacking (lanes 2 and 5) or containing (lanes 3 and 6) an NF-KB purified protein. Peptides A and B corresponded precisely to p65 cDNA were mixed with a 32P-radiolabeled KB enhancer probe and sequences present in human NF-KB p65 (16) (Fig. 1A). In UV-irradiated. Radiolabeled crosslinked adducts were immunopre- contrast, peptide C was lacking in the original reported cipitated with either preimmune serum (lanes 2-4) or anti-peptide A sequence of NF-KB p65. However, sequencing of an inde- antiserum (lanes 5-7) and analyzed by SDS/PAGE. Molecular weight pendently isolated p75 cDNA revealed the presence of three standards are shown in lane 1 (Mr X 10-3 at left). (B) Radiola- undetected nucleotides that alter the amino acid sequence beled adducts corresponding to HeLa p75 (lanes 2, 4, 6, and 8) and between residues 372 and 380 to conform precisely with that COS NF-KB p65 (lanes 1, 3, 5, and 7) were subjected to partial of peptide C (Fig. 1B). This correction has been confirmed proteolysis with the indicated endoproteases. Cleavage products retaining covalently bound 32P-labeled DNA were separated by SDS/ (C. Rosen, personal communication) and leads to 551 amino PAGE and detected by autoradiography. Control samples run in lanes acids rather than 550 in the full-length NF-KB p65 polypep- 1 and 2 correspond to undigested adducts. tide. These primary structural correlates suggested that p75 was either identical to NF-KB p65, a posttranslational vari- Based on renaturation studies, the factors giving rise to these ant, or the product of an alternatively spliced transcript. two adducts were each capable ofindependently engaging the To further explore the relationship between p75 and NF-KB KB enhancer and comigrated during SDS/PAGE (ref. 18 and p65, DNA/protein crosslinking studies were performed with data not shown). Furthermore, the partial cleavage patterns partially purified p75 (18) and extracts from COS cells for these two nucleoprotein adducts produced by digestion transfected with an NF-KB p65 cDNA expression vector. with Arg-C, Asp-N, or Lys-C endoprotease proved identical First, radiolabeled DNA/protein adducts were subjected to (Fig. 2B). These findings indicate that NF-KB p65 mediates immunoprecipitation with antiserum specific for peptide A. the formation of the p75 crosslinked adduct and, together Radiolabeled adducts containing NF-KB p65 (Fig. 2A, lane 6) with NF-KB p50, corresponds to the two KB enhancer- and p75 (lane 7) were specifically immunoprecipitated with binding proteins that are rapidly expressed as a heterodimeric this antiserum, but not by preimmune serum (lanes 3 and 4). complex in the nucleus following human T-cell activation (18, Notably, the immunoreactive adducts (70-75 kDa) formed 23). with p75 and NF-KB p65 displayed indistinguishable electro- NF-KB p65 Contains a Potent Transcriptional Activation phoretic mobilities in the presence of the bound DNA probe. Domain. To determine whether NF-KB p65 contains a func- A 100 200 300 400 500

NH2 REL HOMOLOGY I COH _-~i ELFPLIFPAEPAQASGP FQYLPDTDD FPSGQISQA (peptide A) (peptide B) (peptide C) B TTT CCT TCT GGG CAG ATC AGC CAG GCC TCG GCC TTG GCC CCG F P S G Q S Q A S A L A P 368 381 FIG. 1. Comparison of the primary structures of p75 and NF-KB p65. (A) Sequence analysis was performed on three HPLC-purified p75 peptides (peptides A, B, and C). Relative positions of the resultant amino acid sequences in the predicted primary structure of human NF-KB p65 (16) are indicated. Numbers above the NF-KB p65 block diagram refer to amino acid position. (B) DNA sequence analysis of a partial p75 cDNA. Previously undetected nucleotides that alter the original published amino acid sequence for NF-KB p65 (underlined residues), but that are consistent with the amino acid sequences of peptide C, are shown in bold type. Downloaded by guest on September 30, 2021 Biochemistry: Ballard et al. Proc. Natl. Acad. Sci. USA 89 (1992) 1877 tional transactivation domain, chimeric proteins containing loss of function. In contrast, further deletion of amino acids the DNA-binding domain (amino acids 1-147) (45) of the 451-501 markedly reduced (=90%) the capacity of the re- yeast transcription factor GAL4 were coexpressed in COS sulting hybrid protein to activate GAL4-directed transcrip- cells together with a CAT reporter construct containing five tion. In reciprocal 3' deletion studies, removal ofamino acids GAL4 binding sites. Fusion of the GAL4 DNA-binding 502 to 551 from the C terminus of NF-KB p65 diminished the domain with the acidic activation domain of the herpes transactivation capacity conferred to GAL4 by =50%. Fur- simplex virus activator VP16 (46) produced a predictably ther truncation to amino acid 451 virtually abolished this strong chimeric activator (33) that stimulated transcription residual activity. These studies suggest that the C-terminal 40- to 50-fold above basal levels (Fig. 3A). However, replace- 101 amino acids of NF-KB p65 are required for maximal ment of the acidic activation domain of VP16 with the C activation. However, C-terminal deletion analysis of the terminus of NF-KB p65 (amino acids 313-551) yielded a full-length NF-KB p65 protein indicated that additional se- quences between amino acids 414 and 450 also are important chimeric protein that activated GAL4-directed transcription for the activation of KB-directed transcription (data not even more potently than GAL4/VP16. Consistent with pre- shown). vious findings (20, 24, 25), hybrid GAL4 proteins containing Fusion with NF-#cB p65 Converts v-Rel or NF-#cB p50 into the C-terminal halfofhuman c-Rel (amino acids 297-587) also Strong kB-Specific Transactivators. To extend this functional activated transcription, albeit at 6- to 10-fold lower levels analysis, expression vectors encoding various Rel-related than that observed with GAL4/NF-KB p65. In contrast, polypeptides (Fig. 4A) were cotransfected in Jurkat human T GAL4 fusion constructs containing the C-terminal portion of cells with a CAT reporter plasmid containing two HIV-1 KB v-Rel (amino acids 305-503) failed to significantly activate enhancers (KB-TATA-CAT). Cotransfection of expression transcription from this promoter (20). vectors encoding full-length or truncated forms of the v-Rel To localize the NF-KB p65 transactivation domain, we next oncoprotein failed to stimulate transcription from this pro- constructed a series of GAL4/NF-KB p65 hybrids with moter (Fig. 4B). However, replacement of v-Rel sequences deletions between amino acids 313 and 551 (Fig. 3B). Pro- downstream of the Rel homology domain with NF-KB p65 gressive deletion from amino acid 313 to 450 resulted in no C-terminal sequences resulted in a chimeric protein that induced KB-specific promoter activity by >50-fold. While A Relative CAT Activity NF-KB p50 alone produced <5-fold stimulation, fusion ofthe (Fold Induction) C terminus of NF-KB p65 to the Rel homology domain of 0 50 100 150 200 NF-KB p50 produced 55-fold stimulation. Neither of these *47 chimeric Rel polypeptides activated CAT reporter constructs containing five AP-1 binding sites (AP1-TATA-CAT) or the 413 490 enhancer-deleted basal promoter (TATA-CAT), suggesting 313 551 A B 297 587 RELATIVE CAT ACTIVITY fl (FOLD INDUCTION) 10 20 30 40 50 60 306 503 u--I~ kB-TATA-CAT VECTOR ALONE API-TATA-CAT TATA-CAT B 313 551 1 503 FGAL4GA3 \/N-kV~ p65 NH21 v-Rel COOH 414 551 1 331 NH2h v-Relj COOH 451 551 1 331 338 551 501 551 I 313 501 1 463 CO Vp5 NH2 NF-NFkB_p~p50 COOH 33 41 1 367 306 551 NH2 NF-kB p0 pOCOOH FIG. 3. Transcriptional activation by GAL4/NF-KB p65 chime- ras. (A) COS cells were cotransfected with a CAT reporter plasmid 286 551 containing five GAL4 binding sites (pG5BCAT, 2 /Ag) and expression NH2 NF-kB p6est COOH vectors (4 ,ug) encoding various chimeras (schematically represented on left) containing the DNA-binding domain of GAL4 (amino acids 1 551 1-147) fused to either the transactivation domain of VP16 or the C NH2i NF-kB P185 F~COOH termini of NF-KB p65, c-Rel, or v-Rel. Cells were assayed for CAT after 48 hr and the data are presented (histogram on right) as 4 activity NH2v-ReIC fold induction over basal levels obtained with the truncated GAL4 FIG. 4. Functional analysis of chimeric and native Rel-related DNA-binding domain. Results were normalized for protein recovery polypeptides. Jurkat T cells were cotransfected with CAT reporter and expressed as the mean fold increase in CAT activity ± SEM. (B) plasmids (5 ,g) containing either two tandem KB enhancer elements COS cells were cotransfected with pG5BCAT (2 jg) and GAL4 from the HIV-1 LTR (KB-TATA-CAT), five tandem AP1 binding expression vectors (4 tkg) encoding truncated forms ofthe NF-KB p65 sites (AP1-TATA-CAT), or the enhancer-deleted basal promoter C terminus (amino acids 313-551) fused to the DNA-binding domain (TATA-CAT) in the presence or absence of pCMV4 expression of GAL4 (amino acids 1-147). The N- and C-terminal endpoints of vectors encoding various Rel-related polypeptides. (A) Fusion and these deletion mutants are schematically depicted on the left. CAT deletion endpoints ofeach chimeric or truncated protein, which were activities are expressed as the mean fold increase in CAT activity + expressed at comparable levels when introduced into COS cells (data SEM relative to basal levels measured with GAL4 alone. Compa- not shown). (B) Relative CAT activities measured after 48 hr of rable levels of protein expression were confirmed by immunopre- culture and expressed as the mean fold increase in CAT activity cipitation with anti-GAL4 (data not shown). SEM. Downloaded by guest on September 30, 2021 1878 Biochemistry: Ballard et al. Proc. Natl. Acad. Sci. USA 89 (1992) that these transactivation events are contingent upon KB LTR-CAT transcription units were induced approximately enhancer binding by the N-terminal Rel homology domain. 24- and 3-fold, respectively, by the limited amounts ofNF-KB Consistent with this interpretation, transient expression of p65. When transfected alone, NF-KB p50 produced only the NF-KB p65 C terminus (residues 286-551), which lacks a weak activation of the KB-TATA-CAT and essentially no DNA-binding domain, failed to stimulate KB-directed tran- stimulation of the HIV-1 LTR, compared with NF-KB p65. scription. However, fusion of this disarmed C-terminal seg- However, combinations of NF-KB p50 and NF-KB p65 pro- ment to the Rel homology domain of NF-KB p65 reconsti- duced synergistic activation of both reporter units in a tuted a potent native transactivator protein. dose-dependent fashion (Fig. 5 A and B). At high concentra- with NF-KB p65, tions of p5O, moderate inhibition of this response was ob- NF-KB p50 Functionally Synergizes tained, presumably due to the formation of transcriptionally Whereas v-Rel Represses NF-KB p65-Mediated Activation. inactive p50 homodimers. The predominant inducible nucleoprotein complex detected v-Rel is a dominant negative repressor of NF-KB- with crude nuclear extracts corresponds to an NF-KB p50/ dependent transcription induced by cellular activation with p65 heterodimer (18, 21, 23). These findings prompted us to either phorbol ester or the Tax protein of HTLV-I (18, 19). examine whether coexpression of these two NF-KB subunits We therefore assessed the ability of v-Rel to specifically might synergistically activate KB-dependent transcription. repress NF-KB p65-activated transcription in the absence of To test this hypothesis, Jurkat T cells were transfected with these cellular activation agents. In contrast to the costimu- either KB-TATA-CAT or the HIV-1 LTR-CAT reporter latory effects of NF-KB p5O and NF-KB p65, cotransfection plasmid and graded doses ofan NF-KB p50 expression vector of NF-KB p65 with a v-rel expression vector encoding the in the presence or absence of a fixed suboptimal amount of full-length oncoprotein completely inhibited NF-KB p65- the NF-KB p65 expression vector (Fig. 5 A and B). In the mediated activation of these same KB-directed transcription absence of NF-KB p50, the KB-TATA-CAT and HIV-1 units in a dose-dependent manner (Fig. 5 C and D). These observed antagonistic effects suggest that the 65-kDa subunit B HIV-I LTR-CAT of NF-KB represents one cellular target for v-Rel-mediated 8 transcriptional repression.

_- 6- DISCUSSION <*0 Human NF-ecB p65 Is a Potent Transcriptional Activator. 0- +NF-kB p65 C' 4 ii Transient cDNA expression studies have revealed that C) *° NF-KB p65, but not NF-KB p50, functions as a potent co l.L activator of KB enhancer-directed transcription in human 2 -NF-kB p65 Jurkat T cells. The domain required for this transactivation i function has been localized to the C-terminal 101 amino acids of NF-KB p65. Consistent with the modular organization of 1 2 eukaryotic transcription factors (47, 48), this transactivation < -U NF-kB p50 (/xg) function is fully preserved when the C-terminal halfofNF-KB p65 is fused to a heterologous DNA-binding domain (GAL4) or to the Rel homology domains ofeither v-Rel or NF-KB p50. However, the Rel homology domain could indirectly modu- late this transactivation function through its capacity to mediate dimerization with other Rel-related polypeptides I._ C) C (17). In this regard, we found that coexpression ofNF-KB p65 < ._ and NF-KB p50 led to synergistic functional activation of < KB-dependent transcription units, relative to the expression 'O functional effects a), o of either protein alone. These synergistic U. C-_ most likely reflect the ability of NF-KB p65 and NE-KB p50 cr to form a heterodimeric complex that more avidly engages the KB enhancer than does the NF-KB p65 protein alone (18, 21). Human NF-ecB p65 Is a Cellular Target for v-Rel-Mediated 0 1 2 3 0 2 4 Transcriptional Repression. We have previously shown that v-Rel (tLg) v-Rel (/.g) v-Rel, but not nontransforming v-Rel mutants, binds to the KB enhancer and inhibits NF-KB-activated transcription from FIG. 5. Divergent functional effects of NF-KB p5O and the v-Rel the interleukin 2 receptor a-chain promoter and the HIV-1 oncoprotein on NF-KB p65-activated transcription. (A and B) Syn- LTR in human T cells (18). These findings suggested a of by NF-KB p5O and ergistic activation KB-dependent transcription potential functional link between this NF-KB inhibitory ac- NF-KB p65. Jurkat T cells were cotransfected with either the v-Rel. We KB-TATA-CAT or the HIV-1 LTR-CAT reporter plasmid and tivity and cellular transformation mediated by graded amounts of an NF-KB p5O cDNA expression vector (pCMV4- show here that fusion of the NF-KB p65 transactivation p5O) in the presence (e) or absence (o) of a suboptimal amount (0.5 domain to the N-terminal 331 amino acids of v-Rel converts ,g) ofthe NF-KB p65 cDNA expression vector. (C and D) Repression this protein into a potent transactivator of KB-directed tran- of NF-KB p65-activated transcription by v-Rel. Jurkat T cells were scription. Furthermore, v-Rel completely inhibits NF-KB cotransfected with either the KB-TATA-CAT or the HIV-1 LTR- p65-activated transcription when provided in trans, thus CAT reporter plasmid and graded amounts of a v-Rel cDNA expres- establishing NF-KB p65 as at least one cellular target for sion vector (pCMV4-v-Rel) in the presence (0) or absence (o) of a v-Rel-mediated transcriptional repression. Given their re- fixed amount of the NF-KB cDNA vector (0.5-2 ,ug). p65 expression lated DNA-binding activities (18), v-Rel may act as a com- The amount of DNA for all transfections was held constant by the addition of unmodified pCMV4 expression vector. Results from petitive inhibitor of NF-KB p65 binding to the KB enhancer. is duplicate assays are expressed as the mean fold increase in CAT This potential mechanism for v-Rel-mediated repression activity relative to that measured with CAT reporter constructs supported by the finding that mutations which disrupt its alone. DNA-binding function but not heterodimerization with Downloaded by guest on September 30, 2021 Biochemistry: Ballard et al. Proc. Natl. Acad. Sci. USA 89 (1992) 1879 NF-KB also abolish its repressor activity (ref. 18 and unpub- 13. Ghosh, S., Gifford, A. M., Riviere, L. R., Tempst, P., Nolan, lished data). G. P. & Baltimore, D. (1990) Cell 62, 1019-1029. Regulation ofNF-#cB Function. 14. Kieran, M., Blank, V., Logeat, F., Vandekerckhove, J., Potential Implications for the Lottspeich, F., Le Bail, O., Urban, M. B., Kourilsky, P., The data presented here and in prior reports (13-16, 18) Baeuerle, P. A. & Israel, A. (1990) Cell 62, 1007-1018. establish that the predominant KB enhancer-binding proteins 15. Nolan, G., Ghosh, S., Liou, H.-C., Tempst, P. & Baltimore, D. detected in DNA/protein crosslinking assays ofhuman T-cell (1991) Cell 64, 961-969. extracts correspond to at least three unique gene products, 16. Ruben, S., Dillon, P., Schreck, R., Henkel, T., Chen, C.-H., including NF-KB p50, NF-KB p65, and c-Rel. Our results Maher, M., Baeuerle, P. & Rosen, C. A. (1991) Science 251, demonstrate that NF-KB p65, but not NF-KB p50, functions 1490-1493. as a potent transcriptional activator of KB enhancer- 17. Gilmore, T. D. (1990) Cell 62, 841-843. a trans- 18. Ballard, D. W., Walker, W. H., Doerre, S., Sista, P., Molitor, containing promoters. Of note, c-Rel also contains J., Dixon, E., Peffer, N., Hannink, M. & Greene, W. C. (1990) activation domain in its C terminus (20, 24, 25) but appears Cell 63, 803-814. able to only weakly activate KB-specific transcription (19). 19. Inoue, J.-I., Kerr, L., Ransone, L., Bengal, E., Hunter, T. & Furthermore, in contrast to the rapid nuclear induction of Verma, I. (1991) Proc. Natl. Acad. Sci. USA 88, 3715-3719. NF-KB p65 occurring during T-cell activation, the nuclear 20. Richardson, P. & Gilmore, T. (1991) J. Virol. 65, 3122-3130. expression of c-Rel is induced with delayed kinetics and, in 21. Baeuerle, P. & Baltimore, D. (1989) Genes Dev. 3, 1689-1698. part, is dependent upon de novo protein synthesis (23). 22. Urban, M. & Baeuerle, P. (1990) Genes Dev. 4, 1975-1984. These findings suggest an attractive mechanism to regulate 23. Molitor, J., Walker, W., Doerre, S., Ballard, D. W. & Greene, the functional activities of NF-KB, which involves the ability W. C. (1990) Proc. Natl. Acad. Sci. USA 87, 10028-10032. 24. Bull, P., Morley, K., Hoekstra, M., Hunter, T. & Verma, I. of NF-KB p50 to dimerize with either NF-KB p65 or c-Rel. (1990) Mol. Cell. Biol. 10, 5473-5485. Specifically, following T-cell stimulation, NF-KB p50/p65 25. Kamens, J., Richardson, P., Mosialos, G., Brent, R. & Gil- heterodimers are rapidly mobilized to the nucleus to partic- more, T. (1990) Mol. Cell. Biol. 10, 2840-2847. ipate in transcriptional activation ofearly genes under NF-KB 26. Hager, D. & Burgess, R. (1980) Anal. Biochem. 109, 76-86. control (23). The extent of this induced expression, and 27. Gross, E. (1967) Methods Enzymol. 11, 238-255. perhaps the transcriptional activation of late genes under 28. Lathe, R. (1985) J. Mol. Biol. 183, 1-12. NF-KB control, may be temporally governed by the delayed 29. Ullrich, A., Berman, C., Dull, T., Gray, A. & Lee, J. (1984) nuclear import and action of distinct NF-KB p50/c-Rel het- EMBO J. 3, 361-364. erodimers (23). Further studies on the biological relationship 30. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Nail. Acad. Sci. USA 74, 5463-5467. between these differentially induced heterodimers should 31. Wu, C., Wilson, S., Walker, B., David, L., Paisley, T., provide important insights concerning their regulatory roles Zimarino, V. & Ueda, H. (1987) Science 238, 1247-1253. in the genetic program controlling T-cell activation and 32. Ballard, D. W., Bohnlein, E., Hoffman, J., Bogerd, H., Dixon, growth. E., Franza, B. R. & Greene, W. C. (1989) New Biol. 1, 83-92. 33. Sadowski, I., Ma, J., Triezenberg, S. & Ptashne, M. (1988) We thank J. Kadonaga, M. Hnatowich, and J. Ostrowski for Nature (London) 335, 563-564. critical discussions regarding this work; S. Ballard and J. Nevins for 34. Brownell, E., Mittereder, N. & Rice, N. (1989) Oncogene 4, HeLa cytoplasmic extracts; M. Ptashne for GAL4 plasmids; S. 935-942. Ruben and C. Rosen for the NF-KB p65 cDNA clone; R. Randall for 35. Wilhemsen, K., Eggleton, K. & Temin, H. M. (1984) J. Virol. amino acid sequence analysis; J. Didsbury for the HL-60 cDNA 52, 172-182. library; and B. Kissell for manuscript preparation. 36. Andersson, S., Davis, D., Dahlback, H., Jornvall, H. & Rus- sell, D. W. (1989) J. Biol. Chem. 264, 822-829. 1. Lenardo, M. J. & Baltimore, D. (1989) Cell 58, 227-229. 37. Horton, R., Cai, Z., Ho, S. & Pease, L. (1990) Biotechniques 2. Sen, R. & Baltimore, D. (1986) Cell 46, 705-716. 8, 528-535. 3. Sen, R. & Baltimore, D. (1986) Cell 47, 921-928. 38. Kozak, M. (1987) J. Mol. Biol. 196, 947-950. 4. Bohnlein, E., Lowenthal, J. W., Siekevitz, M., Ballard, D. W., 39. Holbrook, N., Gulino, A. & Ruscetti, F. (1987) Virology 157, Franza, B. R. & Greene, W. C. (1988) Cell 53, 827-836. 211-219. 5. Lowenthal, J. W., Ballard, D. W., Bogerd, H., Bohnlein, E. & 40. Neumann, J. F., Morency, C. & Russian, K. (1987) Biotech- Greene, W. C. (1989) J. Immunol. 142, 3121-3128. niques 5, 444-447. 6. Osborn, L., Kunkel, S. & Nabel, G. (1989) Proc. Nail. Acad. 41. Bradford, M. (1976) Anal. Biochem. 72, 248-254. Sci. USA 86, 2336-2340. 42. Sodroski, J., Rosen, C., Goh, W. C. & Haseltine, W. (1985) 7. Ballard, D. W., Bohnlein, E., Lowenthal, J. W., Wano, Y., Science 228, 1430-1434. Franza, B. R. & Greene, W. C. (1988) Science 241, 1652-1655. 43. Luckow, B. & Schutz, G. (1987) Nucleic Acids Res. 15, 5490. 8. Leung, K. & Nabel, G. (1988) Nature (London) 333, 776-778. 44. Stein, B., Rahmsdorf, H., Steffen, A., Litfin, M. & Herrlich, P. 9. Ruben, S., Poteat, H., Tan, T.-H., Kawakami, K., Roeder, R., (1989) Mol. Cell. Biol. 9, 5169-5181. Haseltine, W. & Rosen, C. A. (1988) Science 241, 89-91. 45. Ma, J. & Ptashne, M. (1987) Cell 48, 847-853. 10. Baeuerle, P. & Baltimore, D. (1988) Cell 53, 211-217. 46. Triezenberg, S., Kingsburg, R. & McKnight, S. (1988) Genes 11. Baeuerle, P. & Baltimore, D. (1988) Science 242, 540-546. Dev. 2, 718-729. 12. Ghosh, S. & Baltimore, D. (1990) Nature (London) 344, 678- 47. Brent, R. & Ptashne, M. (1985) Cell 43, 729-736. 682. 48. Frankel, A. & Kim, P. (1991) Cell 65, 717-719. Downloaded by guest on September 30, 2021