enhancer-binding protein (NF-KB) that activates transcription from

Proc. Natl. Acad. Sci. USA Vol. 85, pp. 4700-4704, July 1988 Biochemistry

Identification and purification of a human immunoglobulinenhancer-binding protein (NF-KB) that activates transcription from a human immunodeficiency virus type 1 promoter in vitro (transcription factor/lymphoid-specific protein/UV crosslinking/DNase I "footprinting"/methylation interference)

KIYOSHI KAWAKAMI, CLAUS SCHEIDEREIT, AND ROBERT G. ROEDER Laboratory of Biochemistry and Molecular Biology, The Rockefeller University, New York, NY 10021

Communicated by James E. Darnell, Jr., March 21, 1988

position 4163 (16, 17) was excised from plasmid pLX40P.1 (18), subcloned into plasmid pUC13, and used as one probe (kindly donated by R. Currie, Rockefeller University, New York). An Xho I-EcoRI fragment containing HIV-1 sequence from position - 167 to position + 80 fused to the bacterial chloramphenicol acetyltransferase gene fragment (HindIIIEcoRI) of p167 (13) was subcloned into pUC13 (pXRIII, kindly provided by M. Sawadogo, Rockefeller University) and used as a second probe. pU3R/ - 121 and pU3R/ -57 (13) were a kind gift from C. Rosen (Roche Institute) and were used as templates for in vitro transcription. Oligonucleotides and Affinity Column. A specific oligonucleotide that forms the recognition site for NF-KB and contains a 6-nucleotide 5'-end extension was prepared by annealing the following chemically synthesized oligonucleotides: A (5'-GGATCCTCAACAGAGGGGACTTTCCGAGGCCA-3') and B (5'-GGATCCTGGCCTCGGAAAGTCCCCTCTGTTGA-3'). A mutant oligonucleotide that contains a 1-bp change (bases in boldface letters) in the recognition site for NF-KB was prepared from synthesized oligonucleotides: C (5'-GGATCCTCAACAGAGGCGACTTl-CCGAGGCCA3') and D (5'-GGATCCTGGCCTCGGAAAGTCGCCTCTGTTGA-3'). CNBr-activated Sepharose CL-4B (Pharmacia) was coupled with the double-stranded oligonucleotides as described by the manufacturer. Approximately 0.25 mg of DNA was coupled per ml of matrix. Purification of NF-KB. The preparation of Namalwa-cell nuclear extracts and the chromatographic buffers BC and C, with dithiothreitol and phenylmethylsulfonyl fluoride, were as described (19). NF-KB activity was monitored with the gel retardation assay by using an end-labeled Dde I-Sau3AI fragment (positions 3886-3973) from the K enhancer (16; 17). The first two steps of the purification were as described for octamer-binding transcription factor 2, except that a typical purification involved 150 ml of Namalwa-cell nuclear extract (900 mg of protein, from 6 x 1010 nuclei) and slightly larger DEAE-cellulose (50 ml) and Bio-Rex 70 (80 ml, equilibrated with buffer BC/100 mM KCI) columns. Active fractions (170 mM-300 mM KCI) from the Bio-Rex 70 column were pooled, dialyzed to 100 mM KCI, and loaded onto a 10-ml singlestranded DNA-agarose column (Pharmacia) equilibrated in buffer BC/100 mM KCI/0.03% Nonidet P-40 (NP-40), and NF-KB was eluted with a 5-column-volume gradient (100-800 mM KCl) in the same buffer. Active fractions (150 mM-350 mM KCl) were pooled, dialyzed against buffer BC containing 0.05% NP-40 to 100 mM KCl, and loaded onto a 0.6-ml specific-oligonucleotide-Sepharose column equilibrated with the same buffer. After washing with buffer BC/190 mM KCl/0.05% NP-40, bound materials were eluted with buffer BC/350 mM KCl/0.05% NP-40. Activity-containing fractions were combined, diluted with buffer BC/0.05% NP-40 to

The enhancer-binding factor NF-cB, which is ABSTRACT found only in cells that transcribe immunoglobulin light chain genes, has been purified from nuclear extracts of Namalwa cells (human Burkitt Iymphoma cells) by sequence-specific DNA affinity chromatography. The purified NF-cB has been identified as a 51-kDa polypeptide by UV-crosslinking analysis. "Footprint" and methylation-interference analyses have shown that purified NF-mcB has a binding activity specific for the K light chain enhancer sequence. The purified factor activated in vitro transcription of the human immmwdeficiency virus type I promoter by binding to an upstream NF-icBbinding site.

The immunoglobulin heavy and light chain genes are transcribed only in cells of the B lineage. This results from tissue-specific gene rearrangements (1) and from the presence of tissue-specific promoters and tissue-specific enhancers (reviewed in refs. 2-4). As found for many other promoters and enhancers (5, 6), those present in immunoglobulin genes each'contain multiple elements that bind either common or tissue-specific factors (reviewed in refs. 2-4). One of these sites in the K enhancer binds a factor called NF-KB (4, 7) and is the single most critical element for lymphoid cell expression (8). The additional observation that an NF-KB binding activity is present in mature B cells and plasma cells but not in pre-B or T cells (4) suggested that NF-KB does play a regulatory role in B-cell-specific K transcription. Consistent with this hypothesis NF-KB is induced in pre-B cells by bacterial lipopolysaccharide or phorbol ester in concert with the induction of K light chain gene transcription (3, 7). Similarly, in HeLa or T-cell lymphoma cells phorbol ester simultaneously induces NF-KB binding and human immunodeficiency virus type 1 (HIV-1) transcription (9, 10). An NF-KB-binding site exists not only in the K enhancer but also in the enhancers of simian virus 40 (SV40) (11), cytomegalovirus (12), HIV-1 (13), and 132-microglobulin (14). Mutation of the NF-KB motif in the SV40 enhancer decreased the enhancer activity in B cells (15). These observations suggest that NF-KB is involved in regulating the transcription of many genes in addition to the K light chain genes. To establish its role as a transcriptional activator and, ultimately, to analyze structure-function relationships, we have purified NF-KB from a mature B-cell line and tested its functional (transcriptional) activity on the HIV-1 promoter in vitro.

MATERIALS AND METHODS Recombinant Plasmids. A 475-base-pair (bp) EcoRI fragment containing K-enhancer sequences from position 3689 to The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: HIV, human immunodeficiency virus; SV40, simian virus 40; NP-40, Nonidet P-40; BrdUrd, bromodeoxyuridine.

4700

Biochemistry: Kawakami et aL a final concentration of 100 mM KCl, and loaded onto a 0.6-ml mutant-oligonucleotide-Sepharose column equilibrated with buffer BC/100 mM KCI/0.05% NP-40. The column was washed with 10 column volumes of the same buffer, and NF-KB was eluted with buffer BC/190 mM KCI/0.05% NP-40. Activity-containing fractions were combined and diluted with buffer BC/0.05% NP-40 to a final concentration of 100 mM KCl. After addition of herring testis DNA to 200 ,4g/ml, the solution was incubated at 00C for 1 hr and then loaded onto a 0.6-ml specific-oligonucleotideSepharose column equilibrated with buffer BC/100 mM KCI/0.5% NP-40/herring DNA (200 gg/ml). The column was washed with buffer BC/190 mM KCI/0.05% NP-40, and NF-KB was eluted with buffer BC/350 mM KCI/0.05% NP-40. The eluted material was frozen in liquid nitrogen and stored at - 70°C. UV Crosslinking. All the procedures were essentially as described by Wu et al. (20). The bromodeoxyuridine (BrdUrd)-substituted oligonucleotide probe was prepared by annealing oligonucleotide A (for the specific probe) or C (for the mutant probe) with the complementary primer 5'TGGCCTCGGA-3' and filling in with the Klenow fragment of DNA polymerase I, [32P]dATP, [32P]dCTP, dGTP, and a 1:1 ratio of dTTP/BrdUrd. One microliter of purified NF-KB fraction was incubated with BrdUrd-substituted probe in the presence of various competitors (see Fig. 2) in 10 ,u of gel-retardation-assay buffer (19) at 30°C for 15 min. The samples were placed 5 cm from a UV lamp (1600 IW/cm2, 260 nm) and irradiated at 4°C for 30 min. One microliter of a solution of 120 mM CaCl2 and 24 mM MgCI2 and 1 ul of a mixture of DNase I and micrococcal nuclease [2.5 units/,tu and 0.2 unit/pul, respectively, in 5 mM TrisHCl (pH 7.5) containing 50 mM NaCl, 35% (vol/vol) glycerol, bovine serum albumin (50 ng/,ul), and 10 mM CaCl2J was added, and the complex was digested for 25 min at 30°C. The reaction was terminated with 1 ,lI of 0.25 M EGTA. After addition of 3 Al of Sx NaDodSO4 sample buffer, the sample was boiled for 2 min, electrophoresed on a 10% polyacrylamide/ NaDodSO4 gel (21), and subjected to autoradiography. In Vitro Transcription. Transcription reaction mixtures (30 /l) contained 20 mM Hepes (pH 8.4), 70 mM KCl, 3 mM MgCl2, 0.6 mM ATP, 0.6 mM CTP, and 0.6 mM UTP, 0.03 mM [32P]GTP (0.1 mCi/ml; 1 Ci = 37 GBq), template DNA (50 ng), and 8.4 ul of HeLa-cell nuclear extract (=60 ,ug of protein). After incubating 1 hr at 30°C, run-off transcripts were analyzed on a denaturing 4.5% polyacrylamide gel.

Proc. Nail. Acad. Sci. USA 85 (1988)

45 kDa (Fig. 1A). Table 1 summarizes the purification of NF-KB from Namalwa-cell nuclear extract. The overall purification and the yield of the specific DNA binding activity were =4000-fold and 0.8%, respectively. Specificity of the Purified NF-KB Binding Complexes. In a gel retardation assay with the K light chain enhancer probe (an 88-bp end-labeled Dde I-Sau3AI fragment), purified NF-KB formed four distinct complexes. By using other probes such as the Dde I-Sau96I fragment (positions 3886-4012) from the K enhancer or Xba I-Pvu II (position - 19) fragment of the HIV-1 promoter, similar complexes were formed (data not shown). The formation of these complexes designated C1, C2, and C3 (a doublet), was inhibited by a specific oligonucleotide (Fig. 1B, lanes 2 and 3) but not by an octamercontaining oligonucleotide (lanes 4 and 5), which readily inhibited the binding of a purified octamer-binding transcription factor to the octamer motif (19). A further analysis of the sites of interaction in complexes C1, C2, and C3 is presented below (see Fig. 3). Identification of the NF-KB-Binding Protein by UV Crosslinking. The co-elution of the 51-kDa protein band with the specific NF-KB-DNA binding activity (above) suggested that it was responsible for the site-specific (NF-KB) activity of the purified fraction. To confirm this, we used the UVcrosslinking procedure that was employed by Wu et al. (20) to identify a purified Drosophila heat shock activator protein. The photoaffinity labeling probe was prepared from specific oligonucleotide A or mutant oligonucleotide C. BrdUrd substitution did not change the specific binding activity as measured in the gel retardation assay (data not shown). As shown in Fig. 2, after binding, UV crosslinking, nuclease digestion, and NaDodSO4/polyacrylamide gel electrophoresis, the 51-kDa protein was labeled by 32p transfer from the [32P]BrdUrd-labeled specific oligonucleotide (lane 2). Controls with the [32P]BrdUrd-labeled mutant-oligonucleotide probe showed no labeled band (lanes 8-10). In competition

A

:3

5-

Purification of Enhancer-Binding Protein NF-#cB. Our purification scheme involved six chromatographic steps. These included passage through DE 52 at a high salt concentration (to remove RNA and nucleoprotein complexes), chromatography on Bio-Rex 70 and on single-stranded DNA-agarose, and chromatography on specific-oligonucleotide columns (20) containing the K-enhancer NF-KB site [in the absence or presence of nonspecific DNA (22)] or a mutated NF-KB site (which reduced NF-KB affinity by a factor of >100). Omission of any one of the three oligonucleotide column steps resulted in several contaminating protein bands (data not shown). Fractions from the final column were assayed in the gel retardation assay and by NaDodSO4/polyacrylamide gel electrophoresis. In those fractions containing the DNA binding activity, two major bands of 68 kDa and 51 kDa were detected on polyacrylamide gels by Coomassie stain. Within the sensitivity limits of this analysis, only the 51-kDa protein correlated in elution position with the DNA binding activity (data not shown). Silver staining of gels containing the peak fraction showed the same (major) 51-kDa and 68-kDa polypeptides and only one additional band of any significance at

2 3 4 5

B

7 CL H1

_

RESULTS

4701

-84 -58 48.5

a ;

36.5

Nl-Ci C3 i }C32 ;4

-26.6

FIG. 1. (A) NaDodSO4/polyacrylamide gel electrophoresis of affinity-purified NF-KB fractions. Forty microliters of the input to the final affinity column and 80 Al of fraction 17 were precipitated with 10%1 (wt/vol) trichloroacetic acid, subjected to electrophoresis on a NaDodSO4 gel, and stained with silver. Positions of molecular size markers are shown. (B) Gel retardation/oligonucleotide competition analysis of purified NF-KB binding to the K enhancer. Gel retardation assays were performed as described (19) with 10 fmol of end-labeled probe (Dde I-Sau3AI fragment) in a 20-Al reaction mixture. The purified NF-KB fraction (1 .l) was incubated with probe in the absence of competitor (lane 1) or in the presence of an NF-KBsequence-containing oligonucleotide (lanes 2 and 3) or an octamersequence-containing oligonucleotide (lanes 4 and 5). Oligonucleotides were added at a 50-fold (lanes 2 and 4) or 20-fold (lanes 3 and 5) molar excess relative to the probe. C1, C2, and C3 indicate the positions of the complexes observed.

4702

Proc. Natl. Acad. Sci. USA 85

Biochemistry: Kawakami et al.

(1988)

Table 1. Purification of the enhancer-binding factor NF-KB Total Total Specific protein, activity, activity, Fold Yield, units Fraction mg units/mg purification t 130 900 Nuclear extract 120,000 100 160 850 1.2 DE 52 137,000 114 123 252 1.6 Bio-Rex 70 31,000 26 24 625 2.5 Single-stranded DNA-agarose 15,000 12.5 0.37 17.3 Specific oligo-Sepharose (first) 4,000 10,800 3.0 2.8 0.092 Mutant oligo-Sepharose 2,800 2.3 30,400 0.002 16.5 Specific oligo-Sepharose (second) 1,000 500,000 0.8 For specific activity, one unit of binding activity is the amount of NF-KB that, under standard gel-retardation-assay conditions in the presence of 1.5 ug of poly[d(I-C)-d(I-C)] and 20 fmol of probe, retards 1 fmol of end-labeled probe. Oligo, oligonucleotide.

experiments an unlabeled specific oligonucleotide (lane 3) and a K-enhancer fragment containing the NF-KB-binding sequence (lane 5) abolished the signal, whereas neither the mutant oligonucleotide (lane 4) nor pUC19 DNA (lane 6) did so. Omission of purified NF-KB protein from the reaction mixture also failed to produce any signal (lane 1). Omission of the DNase treatment resulted in a labeled band remaining at the position of a 60-kDa protein (lane 7), suggesting that the undigested 32-bp probe remained covalently attached to the 51-kDa protein. Interactions of Purified NF-KB with the K Enhancer. DNase I footprinting and methylation-interference experiments were performed on the C1, C2, and C3 complexes. As shown in the DNase-footprint analysis of Fig. 3A, 9 nucleotides of the upper strand and 13 nucleotides of the lower strand were protected. To identify the guanine contacts in this protected region the methylation-interference analysis was performed. Methylation of six of the guanines (marked with arrows) in the NF-KB-binding sequence interfered with the complex formation (Fig. 3B, lanes C1 and C2). Corresponding cleavage enhancement at the positions of these six guanines was observed in the free probe (lane F). These observed enhancements (in the free probe pool) with probes methylated at these positions were expected since only -40% of the probe was uncomplexed in this experiment. The methylation interference and DNase protection results are summarized in Fig. 3C. Transcriptional Activity of the Purified NF-ucB. Although NF-KB was purified with the K-enhancer fragment as probe, the lack of a K enhancer-dependent in vitro transcription 2

3

4

5 6

7

8 9 10 6 84

- 11

51-*

5*

......7 .M

_58 -48.5

--36 5 26.6

FIG. 2. Identification of NF-KB protein by UV crosslinking. The purified NF-KB fraction (1 ul) was incubated with BrdUrdsubstituted specific-oligonucleotide probe (lanes 2-7) or with BrdUrd-substituted mutant-oligonucleotide probe (lanes 8-10), UV crosslinked, digested with the nuclease mixture, and analyzed on an NaDodSO4/polyacrylamide gel. Lanes: 1, no NF-KB fraction added; 3 and 9, 5 ng (100-fold molar excess) of specific oligonucleotide; 4 and 10, 5 ng of mutant oligonucleotide; 5, 45 ng of 213-bp Dde I fragment of the K enhancer (positions 3886-4098); 6, 1 ug of pUC19 DNA; 7, without DNase treatment. Molecular size markers in kDa are shown on the right. The arrow indicates the position of the 51-kDa band.

system has precluded functional analysis of this enhancer. We, therefore, used the HIV-1 promoter, which has two putative NF-KB-binding sites in the enhancer (23), as the A

G F

B

Cl C2C3 G F Cl C2 C3

1 Cl C2 F

I CC2F

wIMM sS48 ;.*a

U N,-

a.

OAS

] 3939 w_

::ea 4*44

3946

_

:..-

Z--m-

¢S

M

_x .

.i.

-m

*0'sR

*4*, Lowe --ran

Upper Strand

C

Lower

Strand

3940

Upper Strand

3950

AACAGAGGGGACTTTCCGAGGCCATCT

TTGTCTCCCCTGAAAGGCTCICGGTAGA A

AA

FIG. 3. DNase I footprinting and methylation-interference analyses of purified NF-KB binding to the K enhancer. (A) DNase I footprints were generated with a 5'-labeled (lower strand) or a 3'-labeled (upper strand) Dde I-Sau96I fragment (positions 38864012) of the K enhancer. The 80-,ul binding reaction mixtures (19) contained 40 fmol of probe and 5 pAl of purified NF-KB, but no poly[d(I-C) d(I-C)]. After 30 min at 30°C, 16 IL of DNase I (1 ug/ml) in 18 mM MgCl2 and 1 mM CaC12 was added and incubated 40 sec at 25°C. Reactions were stopped with 1 ,l1 of 0.5 M EDTA and the samples were analyzed on nondenaturing gels (19). Bands corresponding to the unretarded probe (F) and to the C1, C2, and C3 complexes (Fig. 1B) were excised, and the corresponding DNAs were isolated and analyzed on an 8% (wt/vol) urea sequencing gel as described (19). Maxam-Gilbert guanine reaction mixtures were in adjacent lanes. The regions of protection are indicated. (B) Partially methylated 5'-labeled (lower strand) or 3'-labeled Dde I-Sau96I fragments (positions 3886-4012) were prepared and used for gel retardation experiments as described (19). Bands corresponding to unretarded probe (F), C1, and C2 were excised from the gel after autoradiography. Probe DNA (lane I) and DNAs recovered from the gel slices were cleaved at modified guanines and resolved on an 8% (wt/vol) urea sequencing gel. Guanine positions that interfered with binding are indicated by arrows. (C) Summary of the binding data. Bars above and below the K-enhancer sequence indicate the regions protected against DNase I digestion. The triangles denote guanines whose methylation interferes with binding of NF-KB.

Proc. Natl. Acad. Sci. USA 85 (1988)

Biochemistry: Kawakami et al. template for in vitro transcription. These NF-KB-binding sites are responsible for inducibility of transcription in T cells treated with phorbol esters (9). We have tested two kinds of deletion mutants. pU3R/- 121 contains HIV-1 sequence extending from position -121 to position + 80 and includes the entire promoter/enhancer region. pU3R/ -57 contains sequence from position -57 to position + 80 and lacks the two NF-KB-binding sites (from position -105 to position - 82) (see Fig. SC and ref. 13). When transcribed in HeLa extracts both templates generated 332-nucleotide run-off transcripts corresponding to specific initiation at the in vivo cap site (Fig. 4, lanes 1 and 4). This was expected since these templates contain a "TATA box" and functional SP1 site (10, 24). Addition of affinity-purified (step 6) NF-KB stimulated transcription from the - 121 template =3-fold (quantitated by densitometry of the x-ray film) (lanes 1-3) but had no effect on transcription from the -57 template (lanes 4-6). These results indicate that purified NF-KB activated transcription from the HIV-1 promoter through the interaction with the NF-KB-binding sites rather than in a nonspecific manner. To confirm the idea, we analyzed the binding of the purified NF-KB to the HIV-1 promoter. Interaction of Purified NF-#cB with the HIV-1 Enhancer Sites. Gel retardation assays with an HIV-1 enhancer probe and purified NF-KB revealed complexes (Cl, C2, and C3) analogous to those formed with the K enhancer (data not shown). Of the two putative NF-KB-binding sites in HIV-1 (9), only the distal binding site (positions - 103 to -96) was protected from DNase I treatment in these complexes (Fig. 5A). Decreased signals indicative of methylation interference were not clearly observed in complexed probes (Fig. SB, lanes C1 and C2), but enhanced methylation was seen in the free probe (lane F, see above), suggesting at least a partial interference of binding when one of these six guanine residues was methylated ["-25% of the probe was in the free (uncomplexed) form in this experiment] (Fig. SC).

DISCUSSION To understand the developmental signals that lead to the commitment and differentiation of lymphoid cells, we have undertaken the identification and characterization of celltype-specific factors that regulate expression of the wellcharacterized immunoglobulin heavy and light chain genes. These genes contain tissue-specific promoter and enhancer elements, and factors interacting with these elements have -121 -

-57

++ -++

NF-KB

2 3 45 6 FIG. 4. Transcriptional activity of pure NF-KB in vitro. Template plasmids were digested with EcoRI. pU3R/ -121 (lanes 1-3) or pU3R/ -57 (lanes 4-6) was added as template to the standard reaction mixtures with either 0.8 Al (lanes 2 and 5) or 1.6 ,ul (lanes 3 and 6) of purified NF-KB (second-affinity-column fraction). The arrow indicates the position of the run-off transcript whose size (estimated from markers, data not shown) is that expected for initiation at the cap site and extension to the EcoRI end at position + 332 relative to the cap site.

A

B

GFCIC2C3

G FCIC2C3

4703

ICIC2F

I Cl C2 F

-

,.. .WS

.ell-:

*

-96

-1031

C

S

_

1i96 _ idPV a

_1

Strand

*-40

-

IS

L owe

Upper Strand

C

-110

Upper

Lower Strand

Strand

-100

-90

Lower Strand

-80

GCTACAAGGGACTTTCCGCTGGGGACTTTCCAGGG

CGATGTTCCCTGAAAGGCGACCCCTGAAAGGTCCC A

Ai

FIG. 5. DNase I footprinting and methylation-interference analyses of purified NF-KB with an HIV-1 promoter fragment probe. (A) Footprinting was performed with 5'-labeled (upper strand) or 3'labeled (lower strand) Xba I-Pvu II fragment (from the Xba I site in the vector sequence to position -19) of pXRIII as described in Fig. 3. (B) Partially methylated labeled probes described above were used for gel retardation experiments. Procedures were as described in Fig. 3. Positions of enhanced guanines in the unretarded probe are marked with arrows. (C) Summary of the binding data. Bars above and below the HIV-1 promoter sequence indicate the regions protected against DNase I digestion. The open triangles denote guanines whose methylation was overrepresented in unretarded probe.

been identified by DNA-binding assays (reviewed in ref. 19). We have purified and characterized a B-cell-specific factor (octamer-binding transcription factor 2) that interacts with a sequence (octamer) present in heavy and light chain promoters and in the heavy chain enhancer and is presumed to play a major role in the tissue-specific function of these elements (reviewed in ref. 19). Such a role was demonstrated for the K promoter, which was activated by the purified protein in vitro (19). For the K enhancer, the main determinant for tissuespecific function is an element for which a tissue-specific binding activity (designated NF-KB) has been reported (4, 8). Here we describe the purification of NF-KB by affinity chromatography with an immobilized oligonucleotide containing the K-enhancer NF-KB recognition sequence. The final preparation contained a predominant 51-kDa polypeptide(s) that was shown by UV crosslinking to contain the site-specific binding activity and that formed two major complexes in gel retardation assays. These probably resulted from heterogeneity in the purified protein, since a doublet was observed in NaDodSO4 gels by direct analysis and after UV crosslinking. The molecular basis and possible biological significance of this heterogeneity remains to be demonstrated, but similar observations have been made for other transcription factors such as the CCAAT-binding transcription factor (25) and the B-cell-specific octamer-binding transcription factor 2 (19). Purified NF-KB also formed two minor, more rapidly migrating complexes in the gel retarda-

4704

Biochemistry: Kawakami et al.

tion assay, which may contain degradation products. All of these complexes showed the same binding specificity. NF-KB binding to the K enhancer was analyzed in more detail by DNase-protection and methylation-interference assays, which revealed binding over an -13-bp region containing the sequence GGGACTTTCC (Fig. 3C). Of special note is the fact that the methylation-interference pattern was identical to that reported for the major NF-KB binding activity in crude nuclear extracts (or fractions) (4). This strongly indicates that the protein purified was the B-cellspecific factor responsible for this activity and not a ubiquitous factor (H2TF1 or KBF1) observed by others to interact (through contacts with all guanine residues) with a related sequence (GGGGATTCCCC) in the H-2Kb enhancer (26, 27). This conclusion is supported by our failure to detect the NF-KB binding activity in HeLa cells (cf. ref. 4) and by binding studies (below) on the HIV-1 enhancer. The ubiquitous factor recognizing H-2Kb promoter elements has been purified and identified as a 48-kDa protein (28). Interactions of purified NF-KB with the HIV promoter were restricted to one of the two potential NF-KB-binding sites (Fig. SC). It was shown (9) that one (or both) of these sites mediated the phorbol ester-induced transcription of the HIV promoter in a T-cell line (Jurkat) and that this was accompanied by the induction of an activity (presumably NF-KB) that bound to one (or both) of these sites and to the enhancer. Selective binding of purified NF-KB to the distal site is surprising in view of the identity of the two sites with each other and with the K enhancer site but may reflect a steric problem resulting from the close proximity of the sites. Using crude extracts from HeLa cells, others (29) have observed interactions of a cellular factor(s) with the proximal (but not the distal) element, whereas extracts from lymphoid or phorbol ester-treated cells showed additional interactions with the distal element (ref. 30, see also ref. 10). Thus it appears that the distal element can be occupied only by NF-KB, whereas the proximal site could, in principle, be occupied either by a common cellular factor (possibly the H2TF1 factor discussed above) or by NF-KB. Although our observations with purified NF-KB favor the hypothesis that the proximal site can be occupied only by a factor other than NF-KB, we cannot exclude the possibility that one of the other HIV-1 promoter recognition factors (e.g., SP1; ref. 24) might stabilize interactions of NF-KB at this site (cf. refs. 29 and 31). Binding and functional studies (see below) with purified factors, including NF-KB, H2TF1, SP1, and the TATA factor transcription factor IID (31), will be required to resolve the mechanisms involved in HIV-1 promoter activaK

tion.

The correlations mentioned above between levels of NF-KB binding activity and levels of K or HIV-1 transcription strongly suggest that NF-KB serves as a transcription factor for these genes. We have provided direct proof for this supposition by showing that the purified NF-KB stimulates transcription from the HIV-1 promoter in HeLa extracts and that this stimulation is specific for templates containing NF-KB-binding sites. The observed level of transcriptional stimulation (3-fold) was not as great as that observed for the induction of HIV-1 in T cells by a combination of phorbol ester and phytohemagglutinin (9). However, even without optimization of the in vitro conditions, the level of stimulation approximated the level of induction (3- to 4-fold) of HIV-1 transcription by phorbol ester alone in intact HeLa cells (30) or in derived extracts (10). Thus, apart from implicating NF-KB as a true transcription factor, the present data indicate that the 51-kDa form of NF-KB is functional. These studies represent an important step in studying the structure, function,

and

regulation of

a

factor important not only for

Proc. Natl. Acad. Sci. USA 85

(1988)

immunoglobulin regulation in B cells but also for HIV-1 promoter regulation in T cells. Note Added in Proof: In direct footprint analysis, both of the NF-KB binding sites in HIV promoter were protected from DNase I treatment by the purified NF-KB.

We thank Dr. Masami Horikoshi for valuable discussions throughout this project; Virginia Kozler and Carmen-Gloria Balmaceda for technical assistance; and Drs. Richard Currie, Michele Sawadogo, and Craig Rosen for providing plasmid DNAs. This work was supported by Grant CA42567 from the National Cancer Institute to R.G.R. and by general support from the Pew Trusts to The Rockefeller University. C.S. was supported by a grant from the Deutsche

Forschungsgemeinschaft. 1. Tonegawa, S. (1983) Nature (London) 302, 575-581. 2. Mizushima-Sugano, J. & Roeder, R. G. (1986) Proc. Natl. Acad. Sci. USA 83, 8511-8515. 3. Atchison, M. L. & Perry, R. P. (1987) Cell 48, 121-128. 4. Sen, R. & Baltimore, D. (1986) Cell 46, 705-716. 5. McKnight, S. L. & Tjian, R. (1986) Cell 46, 795-805. 6. Hatzopoulos, A. K., Schlokat, U. & Gruss, P. (1988) in Eucaryotic RNA Synthesis and Processing, Horizons in Molecular Biology Series, eds. Hames, B. D. & Glover, D. M. (IRL, Oxford, Washington, DC), in press. 7. Sen, R. & Baltimore, D. (1986) Cell 47, 921-928. 8. Lenardo, M., Pierce, J. W. & Baltimore, D. (1987) Science 236, 1573-1577. 9. Nabel, G. & Baltimore, D. (1987) Nature (London) 326, 711713. 10. Dinter, H., Chiu, R., Imagawa, M., Karin, M. & Jones, K. A. (1987) EMBO J. 6, 4067-4071. 11. Sassone-Corsi, P., Wildeman, A. & Chambon, P. (1985) Nature (London) 313, 458-463. 12. Boshart, M., Weber, F., Jahn, G., Dorsch-Hasler, K., Fleckenstein, B. & Schaffner, W. (1985) Cell 41, 521-530. 13. Rosen, C. A., Sodroski, J. G. & Haseltine, W. A. (1985) Cell 41, 813-832. 14. Kimura, A., Israel, A., Le Bail, 0. & Kourilsky, P. (1986) Cell 44, 261-272. 15. Nomiyama, H., Fromental, C., Xiao, J. H. & Chambon, P. (1987) Proc. Natl. Acad. Sci. USA 84, 7881-7885. 16. Max, E. E., Maizel, J. V., Jr., & Leder, P. (1981) J. Biol. Chem. 256, 5116-5120. 17. Parslow, T. G. & Granner, D. K. (1983) Nucleic Acids Res. 11, 4775-4792. 18. Foster, J., Stafford, J. & Queen, C. (1985) Nature (London) 315, 423-425. 19. Scheidereit, C., Heguy, A. & Roeder, R. G. (1987) Cell 5l, 783793. 20. Wu, C., Wilson, S., Walker, B., Dawid, I., Paisley, T., Zimarino, V. & Ueda, H. (1987) Science 238, 1247-1253. 21. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 22. Kadonaga, J. T. & Tjian, R. (1986) Proc. Natl. Acad. Sci. USA 83, 5889-5893. 23. Muesing, M. A., Smith, D. H. & Capon, D. J. (1987) Cell 48, 691-701. 24. Jones, K. A., Kadonaga, J. T., Luciw, P. A. & Tjian, R. (1986) Science 232, 755-759. 25. Jones, K. A., Kadonaga, J. T., Rosenfeld, P. J., Kelly, T. J. & Tjian, R. (1987) Cell 48, 79-89. 26. Israel, A., Kimura, A., Kieran, M., Yano, O., Kanellopoulos, J., Le Bail, 0. & Kourilsky, P. (1987) Proc. Natl. Acad. Sci. USA 84, 2653-2657. 27. Baldwin, A. S. & Sharp, P. A. (1987) Mol. Cell. Biol. 7, 305313. 28. Yano, O., Kanellopoulos, J., Kieran, M., Le Bail, O., Israel, A. & Kourilsky, P. (1987) EMBO J. 6, 3317-3324. 29. Garcia, J. A., Wu, F. K., Mitsuyasu, R. & Gaynor, R. B. (1987) EMBO J. 6, 3761-3770. 30. Wu, F., Garcia, J., Mitsuyasu, R. & Gaynor, R. (1988) J. Virol. 62, 218-225.

31. Sawadogo, M. & Roeder, R. G. (1985) Cell 43, 165-175.