Nov 10, 1986 - nonfusion proteins, Germino and Bastia (13) have developed an expression vector, pJG200. We attempted to produce the. E2 protein in E. coli ...
Proc. Natl. Acad. Sci. USA Vol. 84, pp. 1215-1218, March 1987
Biochemistry
The E2 "gene" of bovine papillomavirus encodes an enhancer-binding protein (transcriptional control/recombinant DNA/DNA-protein interaction)
CHRISTOPHER MOSKALUK AND DEEPAK I3ASTIA* Department of Microbiology and Immunology, Duke University Medical Center, Durham, NC 27710
Communicated by Wolfgang K. Joklik, November 10, 1986
ABSTRACT The E2 early open reading frame (presumably gene) of bovine papillomavirus-1 was fused in frame with the collagen-f3-galactosidase-encoding region of the vector pJG200 and was expressed in and partially purified from Escherichia coli. The hybrid protein specifically bound to the enhancer region of bovine papillomavirus at several sites. DNase I-cleavage protection analysis of one such site revealed the protected sequence. A comparison of the protected sequence with the remainder of the DNA sequences that also have affinity for the protein revealed a consensus sequence having the motif TTGGCGGNNC, in which N is any nucleotide. The protected region also includes a sequence with 2-fold rotational symmetry-ATCGGTGICACCGAT.
Papillomaviruses in recent years have become attractive model systems for the study of DNA replication and gene expression. Although there are no tissue culture systems for the propagation of the virus in the laboratory, the pioneering work of Howley and his colleagues (1, 2) demonstrated that the cloned bovine papillomavirus type 1 (BPV-1) DNA can be transfected into mouse C127 cells. The viral DNA not only transforms the mouse cells but also replicates in the mouse cell nuclei as a plasmid with a definite copy number of -100 per cell (3). The complete nucleotide sequence of the viral DNA is known (4), and it reveals several open reading frames (ORFs), presumably corresponding to viral genes (Fig. 1). Of the various early ORFs, El is known to be involved in the replication of the viral DNA (6). The E6 and E7 ORFs are involved in control of copy number (7). The E2 ORF has been identified as the encoder of a transacting protein necessary for transactivation of the viral promoters (5). The expression of the viral promoters is influenced by a viral enhancer element (5, 8, and 9). We wished to find out if the E2 protein transactivates transcription by direct interaction with the viral enhancer or else does the protein control viral transcription without direct DNA-protein interaction (e.g., like adenovirus ElA protein, see refs. 10-12). We have expressed the E2 ORF as a fusion protein in Escherichia coli by fusing the entire ORF, in phase, with the collagen-,3-galactosidase DNA sequence of the vector pJG200 (13). The partially purified hybrid protein was found to interact specifically with the DNA sequences in the region of the enhancer of BPV-1. Immunoprecipitation of the E2 hybrid protein-DNA complexes with anti-p3-galactosidase antibody narrowed down one of the binding sites to a -250-base-pair (bp) DNA sequence. DNase-cleavage protection experiments ("footprinting") enabled us to identify one of the binding sites and provided a clue as to the nature of the "core" binding sequence. 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.
MATERIALS AND METHODS Bacterial and Plasmid Strains. The plasmids pJG200 and pdBPV-1 have been described (2, 13). The E. coli strain TG-1 also has been described (14). Purification of the E2 Hybrid Protein. The E2 hybrid protein was expressed and purified as described (13). The hybrid protein was solubilized by adding 1% Zwittergent and Nonidet P-40 into the extraction buffer. Immunoprecipitation of DNA-Protein Complex. Immunoprecipitation of the E2 hybrid protein-DNA complex with anti-f3-galactosidase antibody has been described by Germino and Bastia (15). DNase-Cleavage Protection Assay. DNase I footprints were prepared as described by Galas and Schmitz (16). Molecular weight ladders for the footprints were derived by performing G > A cleavage reaction described by Maxam and Gilbert (17). RESULTS Expression and Purification of the E2 Protein of BPV-1. To express foreign proteins in E. coli and rapidly purify the proteins either as fusion proteins with ,B-galactosidase or as nonfusion proteins, Germino and Bastia (13) have developed an expression vector, pJG200. We attempted to produce the E2 protein in E. coli by engineering the E2 ORF with synthetic oligonucleotide adaptors in such a way that the DNA corresponding to the 3' end of the E2 ORF (last codon) was fused in the correct translational phase with the collagen linker-,B-galactosidase sequence of pJG200. The first codon of the ORF was fused to a DNA sequenqe that provided an initiator triplet (ATG)-the ribosome binding site of cro gene and the phage X PR promoter. Thus, the entire E2 ORF was fused in phase with the sequence encoding f3-galactosidase. The resultant plasmid called pCM007 (Fig. 2), upon induction, produced a hybrid protein of the correct size, and the hybrid protein could be specifically cut with purified collagenase. The hybrid protein constituted 1% of the total cell protein in the induced cells. We partially purified the hybrid protein by using an affinity column as described (13). The details of the E2 protein purification will be published elsewhere. The partially purified protein contained =30% hybrid protein. Sequence-Specific DNA-Binding by the E2 Hybrid Protein. The partially purified E2 hybrid protein was examined for sequence-specific DNA-binding activity by using a modified immunoprecipitation assay as described (15). The plasmid pdBPV-1 (2) contains the viral enhancer in the Cla I-Hpa I fragment (coordinates 7476-7945; see Fig. 3). The plasmid DNA was digested to completion with Cla I and EcoRI, and the fragments were 3'-end-labeled. The mixture of labeled Abbreviations: BPV-1, bovine papillomavirus type 1; ORFs, open reading frames. *To whom reprint requests should be addressed.
1216
Proc. Natl. Acad. Sci. USA 84
Biochemistry: Moskaluk and Bastia CY
ra) e)
0
iCu Al
0^: 0
base pairs 4000
3000
2000
1000
0
E2
WN
(1987)
20 m
0
20 map units 30
40
n E7 o
50
f.-
40 E4t
$@
E3
El
E5
I_ ro
0 CY
fragments was incubated with various amounts of E2-pgalactosidase hybrid protein and immunoprecipitated with anti-,B-galactosidase antibody. The DNA fragments in the precipitate were recovered, concentrated, and examined by electrophoresis in 1% agarose gels. A representative autoradiogram of the gels is shown in Fig. 4. The results show that the hybrid E2 protein specifically and strongly immunoprecipitated the 2582-bp Cla I-EcoRI fragment of pdBPV-1 (Fig. 3 and Fig. 4, lane B). A larger 4964-bp EcoRI fragment that contains almost all of the bacterial plasmid sequence and 2337 bp of BPV-1 sequence was also less strongly immunoprecipitated. No other fragments of pdBPV-1 showed significant immunoprecipitation. Control experiments were performed with protein fractions from E. coli that contained only the pJG200 vector. No immunoprecipitation of BPV-1 DNA was seen in these experiments (Fig. 4, lane C). We attempted to narrow down the binding site in the following manner. The Cla I-Hpa I enhancer fragment (Fig. 3) contains two Fok I sites (5 GCGAN9 3' in which N = any nucleotide; see Fig. 5 Right and Fig. 7A). We cloned the Cla I-Hpa I fragment between the HindIII and EcoRI sites of pUC19 and cut the fragment out with HindIII and EcoRI. We partially digested the subcloned Cla I-Hpa I fragment with Fok I so that a population of DNA fragments were cut at only one of the two Fok I sites (i.e., cut at coordinate 7670 but not at 7610). A portion of the DNA was allowed to be cut at both Fok I sites. The identity of the resulting fragments of the input DNA (Fig. 5 Left, lane 1) is as follows: 276-bp Fok I-Hpa I fragment (coordinates 7670-7946), 225-bp Fok I partial fragment (coordinates 7476-7670 plus a piece of vector DNA), 165-bp fragment [coordinates 7676 (Cla I)-7610 (Fok I)], and a 60-bp Fok I fragment (7610-7670). The immunoprecipitation assay shows that the 276- and 225-bp fragments were immunoprecipitated, whereas the 165- and 60-bp fragments
,BoMH lambda
EcaRl
Pr
B
orn
HI
Collagen linker
E2
pCM007
N
cm
Cn
_ =mr
= r enqo -
It
FIG. 1. The "early" open reading frames (ORFs), presumably genes, of BPV-1. The coordinates are based on the sequence data of Chen et al. (4). The E2 ORF transactivates simian virus 40 promoters (5). The reading frames are all on one strand of BPV-1 DNA.
apparently did not interact with the hybrid protein. The results lead to two conclusions. First, there are at least two binding sites in the Cla I-Hpa I enhancer fragment, one between coordinates 7476 and 7670 and the other between 7670 and 7946. Second, the Fok I cleavage at 7610 destroys at least one of the E2 protein-binding sites. Identification of the DNA Sequence Recognized by E2 Protein. We 5'-end-labeled the enhancer fragment either at the Hpa I end or at the Cla I end and performed DNase I footprinting with the E2 hybrid protein. We also subcloned the Hae III-HinPI (coordinates 7587-7699 in Fig. 7B) fragment in the pUC19 vector and end-labeled the cloned fragment either at the HindIII end or the EcoRI end of the vector. The labeled DNA was incubated with various amounts of hybrid protein, partially digested with DNase I, and analyzed by electrophoresis in DNA sequence gels. The G > A cleavage reaction described by Maxam and Gilbert (16) was used as a molecular weight ladder. A typical DNase footprint of the top strand is shown in Fig. 6. The E2 protein protected a stretch of DNA between the coordinates 7616 and 7639 of the top strand (Fig. 7). The Fok I cleavage site that destroyed the E2-binding sequence is, as expected, located within the E2-binding sequence (see the arrows in Fig. 7). The bottom-strand footprint was also obtained after labeling the EcoRI end of the enhancer fragment (Fig. 7B) in the usual way. The bottom strand was protected from DNase cleavage between the coordinates 7613 and 7637. Thus, the footprints of the top and bottom strands, as expected, overlap each other. Examination of the autoradiogram shown in Fig. 6 revealed enhancements and attenuation of bands at regions on either side of the main protected sequence. If enhancements are due to distortion of the DNA double helix by the protein, then the Hind
iIl
6958 C
Hpal
Clal
7476 1/7945
i
CIc
54 Bam HI 4450
E co RI FIG. 2. Structure of E2-expression plasmid pCM007. Upon induction by raising the temperature of E. coli TG-1 cells containing the plasmid to 42TC, a tripartite hybrid protein was synthesized. The protein can be used as a hybrid, or the E2 protein can be dissociated from p-galactosidase by cleavage with collagenase (15). AmpR, gene for ampicillin resistance.
pM4L2d
-~
262 7bp
FIG. 3. The structure of pdBPV-1 plasmid. The viral enhancer is located in the Cla I-Hpa I fragment (ref. 5).
Biochemistry: Moskaluk and Bastia A
S
C
Proc. Natl. Acad. Sci. USA 84 (1987)
_
-
%M
b
a
D
d
c
f
e
9416 6557 4361
1217
,... _
_
r
mo
2322
_
r_ .q..,.. -. op
FIG. 4. Autoradiogram of a 1% agarose gel showing the results of immunoprecipitated DNA-binding assay. Lanes: A, Cla IlEcoRIdigested pdBPV-1 input DNA; B, immunoprecipitated DNA fragments that bind to the E2-galactosidase hybrid protein; C, control containing E. coli cell extract without the expressed E2 protein but containing the vector pJG200; D, molecular weight markers.
1
w
_o
E2 protein might be affecting DNA structure in the vicinity of its primary binding site. The protected sequences include the motif ATGGCGGTAC that is also present, with minor variations, at three other places in the enhancer fragment and, thus, might represent other binding sites for E2 protein. The protected sequence includes an inverted repeat sequence ATCGGTGICACCGAT, which is not present elsewhere in the enhancer sequence. However, the sequence YYGGTG (in which Y is an unspecified pyrimidine nucleoside) that forms one arm of the dyad symmetry is formed at three other places in the Hpa I-Cla I enhancer fragment. We also performed exonuclease III-cleavage protection (
a
a-,
4
-
.
Ir.
*
CIa I, 7476
2
01"
Fok l, 7610
501 _ 276
'
225 165w
Fok I, 7670 Hin Pi, 7699
I
60
am
L
Hpa 1, 7946
FIG. 5. Finer localization of the E2-binding site. The Cla I-Hpa I "enhancer fragment" was cloned between the EcoRI and HindIII sites of pUC19, cut out with HindIII and EcoRI, recut with Fok I, and labeled by nick-translation with [32P]dNTPs. (Right) Restriction sites (Fok I and HinPI) in the enhancer fragment. (Left) Autoradiogram of a 5% acrylamide gel. Lanes: 1, Fok I partial digest of the enhancer fragment; 2, specific immunoprecipitation of the 276- and 225-bp fragments. Fok I cleavage at coordinate 7610 destroys an E2-binding site.
FIG. 6. Autoradiogram of a DNA sequence gel showing the DNase 1-cleavage protection pattern of the E2-binding site of the top strand (see Fig. 7) that is interrupted by the Fok I site at coordinate 7610 (see Fig. 5). The curved arrows demarcate the E2 protein-protected area in lanes c and d. Lanes: a and f, Maxam-Gilbert G > A reaction; b and e, DNase I digest with no protein added; c and d, E2 protein added. The autoradiogram was checked by densitometer scanning to confirm the extent of protection by E2 protein. The straight arrows indicate several bands enhanced by the addition of E2 protein. The blank area at the top of lanes c and d was not reproducibly observed and probably is an
artifact.
1218
A,
Biochemistry: Moskaluk and Bastia
(0 FokI ACGTATCGCCTiAGACATGGCGGTAGCCACGTGQCTAT
'IFok / ram G TGCATAGCGGATCTGTACCGCCATCGGTGCACCE ATA C
E2 protein binding sequence )
=
r-
)
B,
5I
Hind ll
E r-
l7
i)
EcoR
FIG. 7. The DNA sequence protected by E2 protein from DNase I digestion. (A) Summary of the DNase I protection experiments. The heavy overlines and underlines show the protected sequences of top and bottom strands. The arrows show actual cleavages by Fok I. (B) DNase fragment used in the footprinting experiment was a HaeIII-HinPI subfragment (coordinates 7587-7699; see ref. 4) that was cloned at the HincIH site of pUC19. The EcoRI-HindIII fragment was isolated, 5'-end-labeled, and cut at either the Sma I site or the Pst site to generate substrates for DNase "footprinting" experiments.
experiments with E2 protein bound to the Cla I-Hpa I enhancer fragment. Two of the exonuclease III stop sites were found at coordinates 7633 and 7640 (data not shown). Thus, these two stop sites closely correspond to the regions protected from DNase cleavage shown in Fig. 7; hence, the identification of E2 binding by Fok I digestion, DNase I footprinting, and exonuclease III cleavage protection are all internally consistent.
DISCUSSION Many mammalian DNA viruses encode gene products necessary for activation of the viral promoters-a phenomenon called transactivation (see ref. 18). There may be, at least, two types of transactivation. First, the transactivating protein might turn on the viral promoter by a largely unknown but indirect mechanism (without direct interaction with enhancer or promoter sequences) such as that in which adenovirus ElA protein apparently participates (10-12). Second, the transacting protein might directly interact with a specific DNA sequence, as is the case with the a gene of herpesvirus (19), Epstein-Barr virus (20) or the GAL4 protein of yeast (21), and activate a distantly located promoter. In the case of BPV-1, the E2 gene product was identified as the viral transactivator (5). The viral enhancer was localized in the Cla I-Hpa I fragment of BPV-1 (5, 9) and subsequently was narrowed down to a DNA region between the coordinates 7611 and 7805 (Barbara Spalholz and Peter Howley, personal communications). In this report, we present evidence that shows that the protein encoded in the E2 ORF of BPV-1 expressed in E. coli specifically binds to the region of the viral enhancer sequence. The protected sequence includes a region with 2-fold rotational symmetry. The binding site also includes the sequence ATGGCGGATG, which is present with minor variations at several other places in the region of the enhancer and, therefore, may be a key recognition signal for E2 protein. The sequence YYGGTG that forms part of the inverted repeat is also formed at three other places in the enhancer fragment.
Proc. Natl. Acad. Sci. USA 84
(1987)
It should be interesting to determine whether the DNA binding per se at the enhancer is the critical event in enhancement of transcription or if the DNA binding domain is replaceable and merely serves to anchor the protein to the enhancer. The domain(s) of E2 involved in protein-protein interaction might be the critical elements in transcription enhancement (22). The conservation of E2 ORF between BPV-1 and human papillomaviruses strongly suggests that the bovine transactivator might be able to transactivate human papillomavirus genes (23). The bacterially expressed E2 protein should be a useful reagent in dissecting the mechanism oftransactivation. The DNA-protein interactions reported in this paper made use of P-galactosidase-tagged E2 protein. Does the f8-galactosidase marker alter or interfere with DNA-protein interaction of E2 protein? This question can be settled definitively by cleaving the hybrid with collagenase. However, in our experience, p-galactosidase has not interfered or altered the pattern of interaction with DNA in two other proteins that we have studied (15). We thank Miss Hilda Smith for the preparation of the manuscript. This work was supported by grants from the National Institutes of Health and the National Cancer Institute. D.B. is an established investigator of the American Heart Association. 1. Law, M.-F., Lowy, D. R., Dvoretzky, I. & Howley, P. M. (1981) Proc. Natl. Acad. Sci. USA 78, 2727-2731. 2. Lowy, D. R., Dvoretzky, I., Shober, R., Law, M.-F., Engel, L. & Howley, P. M. (1980) Nature (London) 287, 72-74. 3. Lancaster, W. F. (1981) Virology 108, 251-255. 4. Chan, E. Y., Howley, P. M., Levenson, A. D. & Seeburg, P. H. (1982) Nature (London) 299, 529-534. 5. Spalholz, B. A., Kang, Y.-C. & Howley, P. M. (1985) Cell 42, 183-191. 6. Lusky, M. & Botchan, M. R. (1985) J. Virol. 53, 955-965. 7. Berg, L. J., Singh, K. & Botchan, M. (1986) Mol. Cell. Biol. 6, 859-896. 8. Lusky, M., Berg, L., Weihar, H. & Botchan, M. (1983) Mol. Cell. Biol. 3, 1108-1122. 9. Howley, P. M., Schwenborn, E. T., Lund, E., Byrne, J. C. & Dahlberg, J. (1985) Mol. Cell. Biol. 11, 3310-3315. 10. Berk, A. J., Lee, F., Harrison, T., Williams, J. & Sharp, P. A. (1979) Cell 17, 935-944. 11. Jones, N. & Shenk, T. (1979) Proc. Natl. Acad. Sci. USA 76, 3665-3669. 12. Ko, J.-L., Dalie, B. L., Goldman, E. & Harter, M. L. (1986) EMBO J. 5, 1645-1651. 13. Germino, J. & Bastia, D. (1984) Proc. Natl. Acad. Sci. USA 81, 4692-4696. 14. Carter, P., Bedouelle, H. & Winter, G. (1985) Nucleic Acids Res. 13, 4431-4443. 15. Germino, J. & Bastia, D. (1983) Cell 32, 131-140. 16. Galas, D. & Schmitz, A. (1978) Nucleic Acids Res. 5, 3157-3170. 17. Maxam, A. & Gilbert, W. (1977) Proc. Natl. Acad. Sci. USA 74, 560-564. 18. Kingston, R. E., Baldwin, A. S. & Sharp, P. A. (1985) Cell 41, 3-5. 19. Kristie, T. M. & Roizman, B. (1986) Proc. Natl. Acad. Sci. USA 83, 4700-4704. 20. Rawlins, D. R., Milman, G., Hayward, S. D. & Hayward, G. S. (1985) Cell 42, 859-868. 21. Giniger, E., Varnum, S. M. & Ptashne, M. (1985) Cell 40, 767-774. 22. Ptashne, M. (1986) Nature (London) 322, 697-701. 23. Danos, O., Engel, L. W., Chen, E. Y, Yaniv, M. & Howley, P. M. (1983) J. Virol. 46, 557-566.
