VIROLOGY

182, 239-249

(1991)

Multiple Sets of Adjacent pE1 and act-1 Binding Sites Upstream of the Pseudorabies Virus immediate-Early Gene Promoter ZBYNEK Institute

of Molecular

Genetics

KOZMiK,

and *institute Flemingovo Received

LUBOS ARNOLD,*

AND

VACLAV

PACES’

of Organic Chemistry and Biochemistry, Czechoslovak n8m. 2, CS- 16637 Prague 6, Czechoslovakia December

6, 1990;

accepted

January

Academy

of Sciences,

23, 199 1

Binding of cellular proteins to specific motifs in the promoter region of the immediate-early gene of pseudorabies virus was studied. The region was dissected into several portions that were used with HeLa cell nuclear proteins in mobility shift assays, DNase I footprinting, and a methylation interference assay. Close to the transcription start site (nucleotide +l) are a TATA-box (-26 to -29), an Spl-binding motif (GGGGCGGGC) (-45 to -54), a CCAAT motif (-66 to -7O), and, further upstream, an NF-lE1 -binding site (AAGATGGC) (-161 to -168). Binding of a protein to the Spl site was demonstrated. Competition experiments show that the CCAAT motif might bind the NF-rE1 factor, rather than any of the known CCAAT-specific factors. Four domains were identified further upstream (nucleotides -200 to -500) from this promoter, each of which contained closely associated motifs where cellular transcription factors NF-pE1 and act-1 could bind. These domains comprise what we call the upstream element. The orientation of the four NF-pE1 motifs in the upstream element is opposite to the orientation of the two NF-pE1 motifs located closer to the transcription start site. Transient expression of reporter genes was used to study the activity of the upstream element after transfection into 3T3 and HeLa cells. The upstream element was necessary for efficient expression of the pseudorabies virus immediate-early gene and increased somewhat the efficiency of the herpes simplex virus thymidine kinase promoter. 0 1991 Academic Press, Inc.

INTRODUCTION

distantly related human cytomegalovirus (CMV), where the so-called 2 l-, 19-, 18- and 16-bp repeats form one of the most efficient enhancers found so far (Boshat-t et a/., 1985). Some of these repeats contain binding sites for cellular transcription factors, such as the CAMP responsive element binding factor (CREB or ATF) within the 19-bp repeat and nuclear factor kappa-B (NFkappa-B) and activator protein 1 (AP-1) within the 18-bp repeat (Akrigg et al., 1985; Henninghausen and Fleckenstein, 1986; Montminy et al., 1986; Sen and Baltimore 1986a,b; Lee et al., 1987; Hai et al., 1988). Sambucetti et al. (1989) have recently shown that the inducible transcription factor NF-kappa-B, which binds to the 18-bp repeat, has a key function in enhancer activation in infected human fibroblasts. The CMV promoter/enhancer is also constitutively strong in uninfected human fibroblasts and is subject to both positive and negative regulation during virus infection. The most important enhancer element for strong constitutive expression is the 19-bp repeat (Boshart eta/., 1985; Stinski and Roehr, 1985; Jeang eta/,, 1987). Campbell and Preston (1987) noted that the 18bp repeat of CMV shows homology with sequences in the PRV IE gene promoter region. Here we report results showing that cellular proteins specifically bind to both unique and repeated elements within the PRV IE gene promoter region. The region consists of the promoter, immediately upstream of the

Many animal viruses use cellular transcription factors to control expression of their genes. For instance, expression of the herpes simplex virus (HSV) immediate-early (IE) gene requires formation of a ternary complex between the viral transactivator Vmw65 (also known as TIF), the cellular octamer-binding protein act-1 , and the target TAATGARAT sequence (review: Goding and O’Hare, 1989). Pseudorabies virus (PRV), the herpesvirus causing Aujezky’s disease in swine, lacks a potent viral transactivator analogous to the HSV TIF (Campbell and Preston, 1987) and expression of its IE gene (Cheung, 1989; VlCek et a/., 1989; Cheung eta/., 1990) seems to be exclusively under the control of cellular factors. The PRV IE gene promoter contains sequence elements that have been implicated in eukaryotic transcriptional control: TATA, Spl, and CCAAT motifs (Campbell and Preston, 1987; VlCek et al., 1990). The upstream promoter region consists of nine imperfect direct repeats (approximately 80 bp each), each of them containing an ordered array of potential binding sites for known transcription factors (VICek et a/., 1990). In this respect, it is similar to the well-characterized upstream region of the major IE gene (IEl) of the ’ To whom

requests

for reprints

should

be addressed. 239

0042.6822/91 Copyright All rights

$3.00

0 199 1 by Academic Press, Inc. of reproducuon in any form reserved.

240

KOZMiK,

ARNOLD, TABLE

SYNTHETIC

OLIGODEOX~RIBONUCLEOTIDES

1 USED IN THIS STUDY

CGAGTGGGCAAGATGGCCGCCGCGGGGG CGAGTGGGCAAGAGTTACGCCGCGGGGG CGGGGGCCGGGCATGCAAATGGTCCTAGC CGAGTGGGCAAGATGGCCGCCGCGGGGGCCGGGCATGCAAATGGTCCTAGC CGAGTGGGCAAGAG-iTACGCCGCGGGGGCCGGGCATGCAAATGGTCCTAGC CGAGTGGGCAAGATGGCCGCCGCGGGGGCCGGGCATGCCCCTGGTCCTAGC AATTGCTTACCTGGGGGCGGGCTCTCCTCCCGG AATTGCGAAAATCGGCCATrGGTCCGCll-CCGG

1

SPl CCAAT

transcription start site, and the upstream element (UE), characterized by the array of repeated binding motifs for cellular transcription factors NF-pE1 and act-1. By using transient transfection, we show that the UE (nucleotides -178 to -615) strongly stimulates transcription from this promoter in cultured cells and that it also stimulates a heterologous promoter. MATERIALS Plasmids

PACES

Nucleotide sequence (in 5’to 3’ orientation)

Designation El Elmut OCTA E 1 IOCTA E 1 mut/OCTA OCTAmut/E

AND

AND

METHODS

and oligonucleotides

Plasmid pPRVP was prepared from pPRV28 (VICek et with HindIll, removal of the insert, and religation. Plasmid ~180 was derived from pPRVP by digestion with Xmal, removal of the released fragment, and religation. Plasmid pXX was created by subcloning the released Xmal fragment (position -425 to the Xmal site in the polylinker region next to the BamHl site at position +48 of pPRVP) into pUC19. Plasmid pPRVPdel was constructed by digestion of pPRVP with lVael and religation. Plasmid p-107CAT was prepared by ligation of the BamHl (position +48) to Nael (position -107) fragment with BarnHI-Xbalcleaved pBLCAT3 (Luckow and Schutz, 1987), filling, and religation. Plasmid p-l 07/UECAT was prepared by inserting the /-/indIll (position -615) to BarnHI (position +48) fragment of pPRVPdel into the BarnHI-HindIllcleaved pBLCAT3. Plasmid pBLCAT2/UE was prepared by cloning the filled HindIll (position -615) to Nael (position -178) fragment into the filled BamHl site of pBLCAT2 (Luckow and Schutz, 1987). In pBLCAT2/ UE, the UE is cloned in the orientation opposite to that in the IE gene promoter region. pBLCAT2IUEd was prepared by inserting the same fragment into the Smal site of pBLCAT2. Oligonucleotides were synthesized on the automatic synthesizer Syngen-2 (Czechoslovak Academy of Sciences) by the H-phosphonate method. Oligonucleotides used in this study are in Table 1. Several oligonucleotide constructs are described in the figure legends.

a/., 1990) by digestion

Cell lines and transfections NIH 3T3 cells and HeLa cells were grown in Dulbecco’s modified Eagle’s medium, supplemented with 10% fetal calf serum and antibiotics. Transfections were carried out using calcium phosphate coprecipitation essentially as described (Graham and Van der Eb, 1973). Cells were transfected with 10 pg of the tested plasmid, harvested after 48 hr, and assayed for chloramphenicol acetyltransferase (Gorman et al., 1982). In addition, 2.5 pg of a plasmid (pCH1 10) containing the p-galactosidase reporter gene (Herbomel et al., 1984) was transfected to normalize transfection efficiencies. Nuclear

extracts

Nuclear extracts from HeLa cells were prepared according to Dignam et a/. (1983). Nuclear extracts from sp2/0, COS, NIH 3T3, and MDBK cells, as well as from mouse kidney, spleen, and liver, were prepared by a small scale procedure (Schreiber et al., 1989). Protein concentration was determined according to Bradford (1976). Mobility

shift assay

Mobility shift assays were performed essentially as described (Fried and Crothers, 1981; Garner and Revzin, 1981) with minor modifications. Restriction fragments or synthetic oligonucleotides used as probes were end-labeled with 32P using Klenow enzyme (Boehringer). Every binding reaction mixture contained 0.5 to 1 pg poly(dl-dC) to disrupt weak and unspecific interactions, and 1 to 6 pg of crude nuclear extract in 20 ~1. The reaction mixtures were incubated 5 min on ice in binding buffer containing 25 mM HEPES, pH 7.5, 5 mn/l MgCI,, 1 mM DTT, 1 mM EDTA, and 10% glycerol and then incubated on ice with the labeled fragment for another 30 min. For competition, unlabeled competitor DNA was included in the preincubation mixture. In some mobility shift assays, glycerol was

PSEUDORABIES

replaced by 4% Ficoll. Reaction mixtures onto a nondenaturing 5% polyacrylamide TBE buffer and electrophoresed at 4”. DNAse

I footprinting

VIRUS

were loaded gel in 0.5X

assay

Binding conditions were as described for the mobility shift assay, except that higher amounts of nuclear proteins (usually 12 and 30 pg) were used along with 1 pg of poly(dl-dC). After 30 min of incubation on ice, DNase I was added and incubation on ice was extended for 1 min. The reactions were terminated as described (Kozmik and Paces, 1990), and mixtures were run on a 6% polyacrylamide sequencing gel, along with a Maxam-Gilbert G + A reaction (Maxam and Gilbert, 1977). Methylation

interference

assay

For the methylation interference assay, DNA was partially methylated with dimethyl sulfate (DMS) (Siebenlist and Gilbert, 1980) and used in a mobility shift assay experiment under the conditions described above, except scaled up lo-fold. Free and protein bound DNAs were located by wet gel autoradiography, isolated by overnight extraction (37”) with elution buffer (0.5 M ammonium acetate, 2 mM EDTA, 10% methanol), and ethanol precipitated. DNAs were cleaved at modified G residues with 10% piperidine and analyzed on a 6% polyacrylamide sequencing gel. RESULTS Similar proteins bind to multiple the PRV IE gene

sites

upstream

of

The PRV IE gene promoter region contains nine imperfect repeats (VICek et al., 1990). Mobility shift assays and competition with three overlapping probes encompassing nucleotides +48 to -615 were performed to determine whether identical or similar protein factors bind in this region (Fig. 1). After incubation of HeLa cell nuclear extract with probe A (nucleotides -615 to -425), two major complexes (Al and A2) were observed (Fig. la, lane 2). Both bands were competed out by the unlabeled fragment A (Fig. 1a, lane 3), but not by an unrelated control oligonucleotide (Fig. 1 a, lane 6) or by the fragment containing four Spl -binding sites (VIEek et al., 1990) (Fig. 1 a, lane 5). Fragment B (-425 to -174) competed with the probe for both bands, suggesting additional binding site(s) for the same protein(s) within the fragment B. Competition with the fragments A and B when the labeled fragment B was used as a probe (Fig. 1 b) provided further evidence for binding of the same or similar protein(s) to multiple sites within the PRV IE promoter region. A mobility shift assay with the cap site

IE GENE

PROMOTER

241

containing probe D (-153 to +48) (Fig. lc) suggested the presence of an Spl-binding site in this DNA fragment, because one of the complexes (Dl) was eliminated by the Spl-specific probe (VICek et al., 1990) (Fig. 1c, lane 6). Moreover, o-phenanthroline, shown to prevent binding of Spl to target DNA, interfered with Dl complex formation (Fig. lc, lane 8). DNAse region

I footprints

in the PRV IE gene promoter

Various DNA fragments derived from the PRV IE gene promoter region (nucleotides -615 to +48) were end-labeled and used as probes in DNase I footprinting analyses. Altogether, seven protected sites were found (Fig. 2). Six of them are upstream of the transcription initiation site (position +l), and one is downstream of it. The downstream site (nucleotides +20 to +38 in Fig. 2a) contains, with two mismatches, a consensus motif for nuclear factor 1 (TcGtN,GCCAA) (Wingeneder, 1988). Upstream from the TATA-box, the first protected site (nucleotides -40 to -74 in Fig. 2a, nucleotides -38 to -72 in Fig. 2b) contains a GC-box (Spl binding site) and a CCAAT-box. Next comes a rather short domain (nucleotides -153 to -174 in Fig. 2b) containing the El motif AAGATGGC, which was shown previously to bind nuclear factor NF-pE1 both in vitro (Weinberger et al., 1986) and in viva (Ephrussi et a/., 1985). The orientation of this El motif is opposite to that of the other El motifs found in the other four upstream protected domains (Figs. 2b,2c). These upstream domains are similar to each other in that each of them contains the El motif A/GAGATGGC closely associated with the octamer ATGCAAAT/G. This arrangement explains why the different restriction fragments derived from this region competed for the same proteins in the mobility shift experiments (Fig. 1). All of these binding sites are part of the upstream element (nucleotides -178 to -615). The overall arrangement of the protein binding sites and the nucleotide sequence of the promoter/UE region are shown in Fig. 3. Simultaneous binding of proteins (NF-pE1 and act-1) to the PRV IE gene upstream element To provide additional evidence that the binding domains in Fig. 2 are protected from DNAse I digestion by transcription factors NF-pE1 and act-1 , the mobility shift assay experiment was performed using a -2 10 to -257 domain made from synthetic oligonucleotides (Fig. 4). Four complexes were observed with this El/ OCTA oligonucleotide (Fig. 4, lane 1). Competition with the El oligonucleotide (Fig. 4, lanes 2 to 4) or the El/ OCTAmut oligonucleotide (mutated at three positions in the octamer motif) (Fig. 4, lanes 1 1 to 13) prevented

KOZMiK, ARNOLD, AND PACES

242

a COMPETITOR: b COMPETITOR:

2 E ap” F c (o”aal

$ z’“OU

Al-

Al-

F12345678 I -425

c -615 -

12345678

12345678 -174 -153

ly t4b

A

-174

,

-615

-425

-153

r;* t46

BtC

B FIG. 1. Mobility shift assays with restriction fragments derived from the PRV IE gene promoter/upstream element region. DNA restriction fragments HindIll-Xmal (nucleotides -615 to -425) (a), Xmal-NotI (nucleotides -425 to -175) (b), and Sphl-BarnHI (nucleotides -153 to +46) (c) were labeled with 32P and used for interaction with HeLa cell nuclear extract. Competing fragments indicated by A, B, B + C, C, and D at the top and bottom of the figures were added in 50.fold excess over the labeled probes. The Spl probe was described earlier (VIEek ef al., 1990); Random 12mer is a mixture of various synthetic oligonucleotides 12 nucleotides long; o-phenanthroline was present at a final concentration of 500 Mmol/liter. F. free DNA; Al, A2. Bl , B2, Dl , D2. and D3 are retarded complexes.

formation of the two fastest complexes (El and El’) and of the slowest one (O/El). These competitors did not interfere with the complex 0 formation. Oligonucleotide OCTA (Fig. 4, lanes 5 to 7) and oligonucleotide El mut/OCTA mutated at four positions in the El motif (Fig. 4, lanes 8 to 10) interfered with formation of the two slowest complexes (O/El and O), but not with formation of complexes El and El’. A GC-rich Spl oligonucleotide derived from promoter region -38 to -62 served as an unrelated competitor DNA (Fig. 4, lanes 14 to 16) and did not compete for any of the specific complexes. These competition studies show that the El and El’ complexes in Fig. 4 may represent binding of NF-pE1 protein, that 0 may be the act-1 -specific complex, and that the O/El complex is probably formed by simultaneous binding of NF-pE1 and act-1 proteins to the same molecule of the probe. Specific identification of the binding proteins by purification or immunocharac-

terization must be performed, hypothesis.

however,

Characterization of El-binding methylation interference assay

motif

to validate this by the

The NF-pE1 -binding sites were further characterized by the methylation interference assay. DNAfragment A (nucleotides -615 to -425) was end-labeled either at position -613 (lower strand) or -424 (upper strand). This DNA was partially methylated with DMS and used as a probe in the mobility shift assay with HeLa cell nuclear extract. The results of the methylation interference assay obtained for complexes Al and A2 (see Fig. 1a) indicate that these complexes are NF-pE1 -specific (Fig. 5). The guanines in positions -488, -487, and -485 of the lower strand and in positions -486, -489, -490, and -493 of the upper strand are in close contact with the binding protein (Table 2). Similar contact points were detected on this probe using mouse

PSEUDORABIES

b*

6

0

2

5

JII N.E.

VIRUS

I

2

IE GENE

5

PROMOTER

243

YI N.E.

do

2

52

ul N.E.

a4 &

0

2

5

JJI N.E.

-497

1 -484

-474

1 462

12

34

LOWER

1

2

3

4 12

UPPER

FIG. 2. DNase I footpnnts detected on both fragments HindIll-Xmal (nucleotides -615 to were singly end-labeled and used for binding Frg. 3. Lane 1, Maxam-Gilbert G + A reaction 5 PI of nuclear extract. Nucleotrde sequences

LOWER

LOWER

34

5

UPPER

strands of DNAfragments derived from the PRV IE promoter/upstream element region. Restriction -425) (a), Xmal-BarnHI (nucleotides -425 to +46)(b). S’phl-BarnHI (nucleotides -153 to +46)(c) of HeLa nuclear proteins. Lower and upper strand refer to the nucleotide sequence presented in (Maxam and Gilbert, 1980); lane 2, no nuclear extract; lanes 3 and 5, 2 ~1 of nuclear extract, lane 4, of the protected regions shown in this figure are in Fig. 3. l DNase I hypersensitive sites.

kidney nuclear extract (data not shown). These results are in agreement with the analysis of the NF-pE1 -binding sites in the immunoglobulin heavy chain enhancer both in viva (Ephrussi et al., 1985) and in vitro (Weinberger et al., 1986). We found NF-pE1 binding activity in all cell types tested (see Materials and Methods; data not shown). This is in agreement with the previously noted ubiquitous distribution of the NF-pE1 protein (Weinberger et al., 1986). The CCAAT box at position -66 likely binds the NF-pE1 protein and not a CCAAT-binding protein Protein binding to the CCAAT motif at position -66 was studied by mobility shift assays using the end-la-

beled synthetic oligonucleotide representing nucleotides -59 to -83. For competition, strong binding sites for well-characterized CCAAT-binding proteins were used. While a homologous oligonucleotide competed for all three observed complexes (Fig. 6b, lanes 7 to 1 l), CPl and UEBP oligonucleotides (Superti-Furga et a/., 1988) competed for only the slowest complex (data not shown); a CTF/NFl -specific oligonucleotide (Superti-Furga et al., 1988) did not compete at all (data not shown). Surprisingly, the El oligonucleotide, used in these competition experiments as an unrelated control DNA, appeared to be a specific competitor for the two faster complexes (Fig. 6b, lanes 2 to 6) while the same oligonucleotide mutated in four positions in the El mo-

244

KOZMiK,

ARNOLD,

AND

PACES

Hind III

a.

AAGCTTCCCCGAAAATCATCTGATTGGCTCGCTAGCACCA ~GGGGCTTTTAGTAGRCTAACCGAGCGATCGTFGT

-615

-4b7

Xma I

-4ilo

C~GGCATGCAAATChGAGGCGCGCGGGAGACGCCTCCGCGCGCCCATTGGCCCGG~CGAGCCGAGATGGCCGCCGCGGGGGCCGGACATGCAAAGTAGAC AGTCTCCGCGCGCCCTCTGCGGAGGCGCGCGGGTAACCGGGCCCGCT

-425

-465

-4i5

GCGAGAGGAAGTAGGGAGAGAAATCCCATTGGCC CGJZTC;:CTTCATCCCTCTCTTTAGGGTAACCGG

GAGAGGAAGTGGGCGA CTCTCCTTCACCCGCT l *

-3i4

40

-2b5

GAGAAATCCCATTGGCCdPCGAGTGGGCAAGATGGCCGCCGCGGGGGCCGGGCATGCAAAT~TCCTCGCGAGGAAGTTCCTCGCGAAATCCCATTGGCC CTCTTTAGGGTAACCGG~GCTCACCCGTTCTACCGGCGGCGCCCCCGGCCCGTACGTTTACCAGG~GCGCTC~~TCAAGGAGCGCTTTAGGGTAACCGG .

Not I

-257

-2io

Sph I

bGCGGCCG