- Email: [email protected]
Activity and enhancer binding factors for BK virus regulatory elements in differentiating embryonal carcinoma cells
VIROLOGY
183, 374-380
(1991)
SHORT COMMUNICATIONS Activity and Enhancer Binding Factors for BK Virus Regulatory Elements in Differentiating Embryonal Carcinoma Cells HARIKRISHNA
NAKSHATRI,’
MARY
Division of Basic Medical Sciences, Faculty of Medicine,
Memorial
M. PATER, AND ALAN PATER* University
Received January 28, 199 1; accepted
of Newfoundland,
March
St. John’s, NF A 1B 3V6 Canada
19, 199 1
We have studied cell type specificity of expression of the human papovavirus BK regulatory elements in undifferentiated and differentiated embryonal carcinoma (EC) cells as a model system. While the activity of the regulatory elements of this virus was marginal in undifferentiated cells, differentiation by retinoic acid and DMSO resulted in a dramatic increase in the activity. To correlate in vivo activity of the regulatory elements with interaction with cellular transcription factors, we performed DNase I footprinting experiments. A GC-rich region was protected in both undifferentiated and differentiated cells. An additional four protected sites were detected in retinoic acid-differentiated cells and at least one of these additional sites was weakly protected in DMSO-differentiated cells. The sequences of the differentiated cell type-specific protected regions showed homology to a nuclear factor 1 (NF-1) binding motif and to a muscle creatine kinase gene enhancer motif. The intensity, competition, and pattern of protection of these sites were different in the two differentiated cell types, suggesting the involvement of different transcription factors regulating the _ -0 1991 Academic Press, Inc. activity of BKV regulatory elements in the two cell types.
difference in homology may contribute to the differences in host range, pathogenicity, and cell type specificity of these viruses. We have used the embryonal carcinoma (EC) cell line, P19, which can be induced to differentiate into a mixture of glial cells, neurons, and astrocytes upon retinoic acid (RA) treatment and into cardiac and skeletal muscle cells upon dimethyl sulfoxide (DMSO) treatment (14). Previous studies have shown that the intact BKV regulatory region containing three 68-bp enhancer-containing repeats is expressed ubiquitously while individual repeats are expressed in a cell-specific manner (15- 17). Our present studies provide evidence for the induction of two different sets of cellular factors important in the regulation of the control region of this virus following differentiation of P19 cells by RA and DMSO. The P19 EC cell line was maintained in a-MEM medium containing 1OYo fetal calf serum, and cells were induced to differentiate as described (14). DNA transfections were with the calcium phosphate precipitation procedure (18). Plasmid pBKcat (77), containing the bacterial chloramphenicol acetyl transferase (CAT) reporter gene (18) and BKV regulatory elements, was transfected into the undifferentiated and the retinoic acid- and DMSO-differentiated EC cell line, P19. The SV40 enhancer pSV2cat and the enhancerless pSVOcat were used as positive and negative controls,
One of the central challenges in eukaryotic molecular biology has been to determine the mechanisms of tissue-specific gene expression and the molecular events that alter gene expression during the differentiation process. By molecular genetic and biochemical approaches, it has been possible to identify c&-acting promoter and enhancer sequences of selected genes and transacting protein factors interacting with these elements and to assess their role in tissue-specific gene expression (1-3). The regulatory regions of viruses have served as excellent model systems for the study of tissue-specific and inducible gene expression (4, 5). Simian virus 40 (SV40) enhancer has been studied extensively and has been described as a “prototypic enhancer.” A systematic mutational analysis of this enhancer has permitted the identification of a number of domains, several of which are expressed in a cell-specific manner (6- 12). The human papovavirus BKV is a closely related virus which has 70-80% DNA sequence homology to that of SV40 ( 13); also, BKV and SV40 large tumor antigens cross-react immunologically. However, the regulatory elements of these viruses are only approximately 40% homologous. This ’ Present address: Inserm-U. 184 lnstitut de Chimie Biologique, Faculte de Medecine 11, Rue Humann. 67085 Strasbourg Cedex, France. ‘To whom request for reprints should be addressed. 0042-6822/91
$3.00
CopyrIght 0 1991 by Academic Press, Inc. All rtghts of reproduction in any form reserved.
374
SHORT COMMUNICATIONS
A
UD
RA
375 UD
DM F
B
UD
RA
B
1 2 3 4
5 6 7
8
DM
FIG. 1. Activity of BK regulatory elements in the undifferentiated (UD) and the retinoic acid (RA)- and DMSO (DM)-differentiated EC cell line, P19. Cells were transfected with 10 pg of the indicated plasmids and incubated for 48 hr. (A) Autoradiogram of the CAT assays. The percentage of conversion of [‘4C]chloramphenicol to acetylated chloramphenicol is given at the bottom and numbers greater than 100% were from assays for which diluted extracts had been used. (B) Spot-blot hybridization of episomal BKcat DNA from the Hirt supernatant with CAT coding sequence DNA as a probe.
respectively. CAT enzyme assays were used to measure quantitatively transcriptional activity of the BKV regulatory elements and the results are illustrated and presented numerically in Fig. 1A. In undifferentiated P19 cells, BKV regulatory elements were expressed only at low levels. However, differentiation resulted in approximately a 200-fold increase in retinoic acid-differentiated P19 cells and a greater than lOO-fold increase in DMSO-differentiated P19 cells. To evaluate whether the increased activity was due to differences in plasmid uptake, transfected DNA was isolated by the Hirt procedure and spot blotted. Autoradiography and liquid scintillation counting demonstrated that both DMSO-differentiated and undifferentiated cells contained the same amount of transfected DNA while retinoic acid-differentiated cells contained only 1.7-fold higher levels of transfected DNA (Fig. 1 B). These results introduced the suggestion that differentiation of EC cells was correlated with quantitative
FIG. 2. DNase I footprinting analysis of BK regulatory elements with undifferentiated P19 nuclear extracts. Chemical cleavage of purines (28) is in lane A/G. Probe treated with DNase I in the absence and presence of 25 pg of protein from nuclear extracts (29) are in lanes F and B, respectively. Lanes 1 to 8 represent competition experiments in which the same amount of extracts and probe were incubated with unlabeled competitor DNA before DNase I digestion. Competitor enhancer DNA fragments are: SV40 nucleotides 107240 (6) in lanes 1 and 2, JC nucleotides 5112-270 (30) in lanes 3 and 4, and the unlabeled probe BK in lanes 5 and 6. A control, nonspecific competitor pUCl9 DNA fragment, nucleotide 306-638, is in lanes 7 and 8. Amounts of competitors were 4 ng (lo-fold excess) and 20 ng in odd and even numbered lanes, respectively. The protected region is designated by the square bracket and numbered Ill. Arrows indicate hypersensitive sites.
SHORT COMMUNICATIONS
3i ‘6
A
RA
A’G F B 1 2 3
4
5 6
7 8 F
FIG. 3. DNase I footprinting with retinoic acid-differentiated nuclear extracts. (A) Competition is by papovaviral enhancer DNA fragments. Procedures and labeling are as in Fig. 2 except that five regions are designated. (B) Oligonucleotide competition experiments: Competition was with 100 ng (3200- to 1O,OOO-fold excess) of the indicated double-stranded oligonucleotides. Oligonucleotides corresponding to BK region It (lane 1; 5’ TGGGCAGCCAGCCA 3’; Fig. 6) and region Ill (lane 2; 5’ GAAACCCCGCCCC 3’; Frg. 6). and control oligonucleotide (lane J; BAAGGGAAGGGATGG 3’; a JC virus enhancer-promoter region sequence not affecting DNase protection; 31) were incubated with nuclear extract prior to addition of probe and DNase I digestion. Other procedures and labeling are as in Fig. 3A. Densitometric scans were performed to quantify competition by BK region II oligonucleotide. Scanned bands are indicated by the lines between lanes B and 1 and were selected since they were
and/or qualitative changes in cellular transcription factors. Therefore, DNase I footprinting experiments were undertaken to compare the binding of transcription factors to the regulatory region of this virus in various P19 cell types. The probe was a sense strand 32P-, 3’ end-labeled BKV DNA fragment, nucleotides 3 186 to 3479 (13). Figure 2 illustrates footprints of the BKV regulatory region in undifferentiated P19 cell extracts. Only one protected region was detected. This region was partially competed by BKV and SV40 DNA but not by JCV and pUC 19 DNA. This protected region spans a GC-rich sequence and has homology to the transcription factor Spl binding motif and to a transcription repressor protein binding site (19). In addition to the protected region, two hypersensitive sites were also obtained. The footprint pattern was completely different for retinoic acid-differentiated cell extracts (Fig. 3A), in which BKV regulatory elements function with greater efficiency (Fig. 1). A total of five protected regions were detected and four among them, designated I, II, IV, and V, have similarity to the binding site of transcription factor CTF/NF-I . Protected regions I and IV contained the sequence TGGN,CCA, which has been shown to be the most favorable NF-1 binding site, while regions II and V contained the sequence TGGN,CCA. Significantly, JCV regulatory elements, which also contain NF-1 binding sequences, competed, although more weakly than those of BKV, for these four protected regions. SV40 regulatory elements, which lack NF-1 binding sequences, competed, although poorly, only to region III. Region Ill is the GC-rich region which was also protected in undifferentiated cells. In addition to protected regions I to V, another region (Fig. 3A) which was not protected for UD (Fig. 2) was observed below region V. However, competition for protection was unconvincing for all four DNA competitors. To further examine the nature of the DNA-protein interactions, we performed oligonucleotide competition experiments of DNase I protection by the RA-differentiated cell extracts. Oligonucleotides corresponding to regions II and III were tested. Oligonucleotide for region II competed, albeit weakly, for regions I, II, IV, the most sensitive to competition in Fig. 3A. Relative absorbances of protected bands were calculated by comparison to the most proximal interregion bands, which are similarly indicated and are located between the regions I and II and IV and V. These bands were selected since they were proximal to protected regions and sufficiently distant from hypersensitive sites surrounding the protected regions. The percentage increase in relative absorbance in lane 1 compared to that of lane B was calculated. For the various regions competition was: I, 540%; II, 490%; IV, 380%; V. 300%. For region IV, “competition” values in control lanes were: lane 2, 50%; lane J, 99%.
SHORT COMMUNICATIONS
FIG. 4. DNase I footprinting by DMSO-differentiated nuclear extracts. Procedures and labeling are as in Figs. 2 and 3.
and V (Fig. 3B, lane 1). This competition, while visibly less than that for total enhancer fragments (Fig. 3A), was significant, as verified bydensitometric scan analysis (Fig. 3B legend). The relatively weak nature of the competition observed with this oligonucleotide could be due to the absence of flanking sequences which could confer positive cooperativity among adjacent protected regions in the enhancer fragments. Cooperativity could be involved in the competition for the four regions with homology to NF-1 binding sites. Region III oligonucleotide competition produced a pattern of pro-
377
tection distinct from that of region Il. Region III oligonucleotide competed only for the homologous region, indicating that the factor(s) binding to this region is unique and suggesting that binding is not positively cooperative with factors binding with neighboring regions. Conversely, competition with this oligonucleotide increased the degree of protection in region IV, especially toward the region Ill proximal side (Fig. 38; lane 2). Even the protection of the band in the center of region IV was 100% increased, as measured by densitometric scans (see the legend to Fig. 3B). This suggests that the interaction of a protein(s) with region III has a negative influence on protein binding to region IV, possibly through steric hinderance. For the DMSO-differentiated cells, the cell system in which BKV enhancer was active but to a lesser extent than in retinoic acid-differentiated cells, only sequences in region III were clearly protected (Fig. 4). BKV and SV40 regulatory elements, at high concentration, competed for protection of this site. The absence of convincing protection of sites in addition to those observed for undifferentiated cells is unexpected, since the activity of BKV in DMSO-differentiated cells was significantly higher than in undifferentiated cells. It is possible that differentiation of P19 cells by DMSO increases transcription factors which are unique from those protected in RA-induced cells and are not as readily detectable by DNase protection assays. To test this possibility, we performed footprints with double the amount of protein in extracts from undifferentiated and DMSO-treated cells. This double amount of protein was selected since the activity of the enhancer in RA-treated cells had also been double that for DMSOtreated cells (Fig. 1A). With undifferentiated cell extracts a partial protection of sequences corresponding to region IV was obtained (Fig. 5). The same region was reproducibly more completely protected when the increased level of DMSO-differentiated cell extracts was used (see Fig. 5 and the legend to Fig. 5 for densitometric scanning results). Furthermore, there were three additional weakly protected regions, corresponding to regions I, II, and V of retinoic acid-differentiated cells (for densitometric scanning results see Fig. 5 legend). These results support the unique natures of NF-1 site binding factors in the two differentiated cell types, since there is a visible difference in the degree of protection between that shown in Fig. 3 and that in Fig. 5. In addition, the protein concentration-dependent variation in footprints for DMSO-differentiated cell extracts suggests that some of the quantitative differences in in viva expression in the retinoic acid- and DMSO-differentiated cells (Fig. 1) could be due to quantitative differences in levels of NF-1 site binding factors.
SHORT COMMUNICATIONS
378
UD kGFB12345676F
kG F 6 1 2 3 4
5 6
7 8
F
FIG. 5. DNase I footprinting with increased levels of nuclear extracts. Procedures and labeling are as in Frgs. 2 to 4 except that 50 ag of protein in extracts was used. Brackets with solid lines and broken lines represent complete and partial protection, respectively. Densitometric scans of selected protected regions were performed to quantify protection and competition. The relative absorbances for the protected regions were determined as described in the legend to Fig. 3. The percentage of protection was determined from the extinction of the relative absorbance of protected bands in lane B compared to that rn lane F. For competition values, the percentages are the level to which the competitor had restored the relative absorbance from the level in lane B (0%) to that in lane F (100%). Protection in the various regions was as follows: UD. IV, 37%; DM, I, 90%; DM, II, 82%; DM, IV, 86%; DM, V, 66%. Competition for the various DM, region IV lanes was as follows: 2 (SV40), 88%; 4 (JC), 73%; 6 (BK), 90%; 8 (pUC), 30%.
The protected regions were differentially competed by all tested DNAs, with competition by DNA of BKV greater than that of SV40, greater than that of JCV, much greater than that of pUC (for densitometric scanning results for region IV competition see Fig. 5 leg-
end). Competition we have observed protected regions tif with homology which differs from
by SV40 DNA is not surprising since significant homology between BKVand SV40 enhancer motifs. One mois the sequence 5’ TGGAAIG 3’, the SV40 enhancer core region only
SHORT COMMUNICATIONS
NF-1 -
MCK
MCK -
I
FIG. 6. DNase l-protected nucleotide sequences of BK regulatory elements with all three cell types. Protected sequences are overlined for each of the P19 cell types indicated on the left. The numerals above the lines correspond to those in Fig. 3. Sequences homologous to consensus sequences within the protected regions, for known transcription factors, are underlined and the abbreviations for the transcription factors are indicated. MCK, the muscle creatine kinase gene K enhancer domain; SV-C, the SV40 core enhancer motif.
by the single underlined T (Fig. 6). The second is the sequence 5’ GCAlTCCAT in regions I and IV, which differs from the SV40 GT-IIC motif only by the two underlined nucleotides. It is interesting to note that in an analogous series of experiments examining protection/competition of labeled SV40 enhancer DNA, we observed protection of the Sphllll motif of the SV40 enhancer with DMSO-differentiated cell extracts and this protection was competed by BKV DNA (our unpublished observations). This suggests that the BKV enhancer’s protected regions I and IV interact with the transcription factor TEF-1 in DMSO-differentiated cells since the three SV40 motifs, Sphllll and GT-IIC, are all binding sites for the same TEF-1 factor (20). The results presented above demonstrate that there are dif-
379
ferences in both footprint and competition patterns in DMSO-differentiated cell extracts compared to those of the RA system. This leads us to suggest that BKV enhancer interacts with distinct factors in the two differentiated cell types. These factors would most likely be NF-1 or NF-l-like factors in RA-differentiated cells and TEF-1 or TEF-1 -like factors in DMSO-differentiated cells. The NF-1 and the related CCAAT motifs have been shown to bind a family of transcription factors, some of which are tissue specific (21-24). The evidence presented above for distinct factors binding to individual BKV NF-1 sites concurs with the results of Grinnell et a/. (16). They have shown with footprints of the BKV-P2 strain regulatory elements that regions II and V were protected only by HeLa cell extracts but not by 293 or MK2 cell extracts, whereas protection for regions I and IV was common to all three cell types. Furthermore, on the basis of results of in viva competition studies, they suggested that factors binding to regions II and V are distinct from those for regions I and IV and act as negative regulators of transcription. However, different conclusions were drawn from studies by Markowitz and Dynan (25) who were able to affinity purify protein NFBK, using oligonucleotides corresponding to region II, and to show that the same factor(s) binds to all NF-1 motifs of BK (Dunlop), BK (WW), and BK (MM) variants. The discrepancies between the results could be due to (1) BKV variant-specific differences in the sequences adjoining NF-1 motifs, (2) cell type-specific differences in NF-1 factors, and (3) inability to copurify, by affinity chromatography, noncovalently associated proteins. In addition to the NF-1 motifs and SV40 enhancer homologous sequences, the protected regions contained sequences that may act as binding sites for other known transcription factors. These include the sequence 5’ GCAGCCA 3’ (Fig. 6; labeled MCK) of regions I, II, IV, and V which were all protected in muscle cell type DMSO-treated cells. This sequence is also present in the K motif of muscle creatine kinase gene enhancer (26, 27). Another sequence of interest is TATGC, present in regions I and IV, which is identical to the EF-II sequence motif except for one nucleotide (I 1; Fig. 6). These motifs may provide diversity in BKV enhancer activity in various cell types. Thus, although the BKV enhancer and the SV40 prototypic enhancer have only 409/o sequence homology, both enhancers may be organized in an analogous fashion. This hypothesis can be tested by examining the effects of mutations of SV40 homologous and nonhomologous motifs on cell-specific expression by BKV enhancer. The advantages of the P19 EC cell system are that populations of cells in a variety of differentiated states
380
SHORT COMMUNICATIONS
can inexpensively, quickly, effectively, and directly be derived from undifferentiated cells with a single genotype. Thus, EC cells have been shown to represent a good system to study cell type-specific expression of an enhancer and to correlate this with molecular events. An unlikely disadvantage of this system concerns the presence of some heterogeneity within the population of cells. Therefore, one subpopulation of differentiated cells might allow expression while another subpopulation might contain the transcription factors that bind DNA. This potential but highly unlikely complication could be evaluated by in situ detection of both viral enhancer driven gene expression and biochemical markers for different cell types. ACKNOWLEDGMENTS We thank H. Hamada for providing the PI 9 cell line and K. Ross, S. Atkins, and W. Tucker for typing the manuscript. Harikrishna Nakshatri has been a predoctoral fellow of the Cancer Research Society Inc., Montreal, Quebec, Canada and more recently of the National Cancer Institute of Canada. This work was supported in part by the Medical Research Council and the National Cancer Institute of Canada.
REFERENCES 1. JOHNSON, P. F. and MCKNIGHT, S. L., Annu. Rev. Biochem. 58, 799-839 (1989). 2. MANIATIS, T., GOODBOURN, S.. and FISCHER,1. A., Science 236, 1231-1245 (1987). 3. MITCHELL, P. J., and TJIAN, R., Science 245, 371-378 (1989). 4. JONES, N. C., RIGBY, P. W. J., and ZIFF, E. B., Genes Dev. 2, 267-281 (1988). 5. MCKNIGHT, S., and TJIAN, R., Cell 46, 795-805 (1986). 6. FROMENTAL, C., KANNO, M., NOMIYAMA, H., and CHAMBON, P., Cell 54, 943-953 (1988). 7. KANNO, M., FROMENTAL,C., STAUB, A., RUFFENACH,F., DAVIDSON, I., and CHAMBON, P., EMBOJ. 8,4205-4214 (1989). 8. MACCHI, M., BORNERT,J. M.. DAVIDSON, I., KANNO, M.. ROSALES, R., VIGNERON,M., XIAO, J-H., FROMENTALC.. and CHAMBON, P., EMBO J. 8,42 15-4227 (1989).
9. ONDEK, B., GLOSS, L., and HERR, W., Nature 333, 40-45 (1988). 10. ONDEK, B., SHEPARD, S., and HERR, W., EMBOJ. 6, 1017-1025 (1987). 11. SCHIRM, S., JIRICNY,J., and SCHAFFNER,W., GenesDev. 1, 65-74 (1987). 12. ZENKE, M., GRUNDSTROM, T.. MATHES, H., WINTZERITH, M., SCHATZ, E., WILDEMAN, A., and CHAMBON, P., EMBOJ. 5,387397 (1986). 13. YANG, R. C. A., and Wu, R., Science 206, 456-462 (1979). 14. RUDNICKI, M. A., and MCBURNEY, M. W., ln “Teratocarcinoma and Embryonic Stem Cells: A Practical Approach” (E. J. Robertson, Ed.), pp. 19-49. IRL Press, Oxford, 1987. 15. DEYERLE. K. L., and SUBRAMANI, S. J., Virology 62, 3378-3387 (1988). 16. GRINNELL, B. W., BERG, D. T., and WALLS, J. D., Mol. Cell. Biol. 8, 3448-3457 (1988). 17. PATER,A., and PATER, M. M., Virology 163, 625-628 (1988). 18. GORMAN, C. M., MOFFAT, L. F., and HOWARD, B. H., Mol. Cell. Biol. 2, 1044-1051 (1982). 19. KAGEYAMA, R., and PASTAN, I., Cell 59, 815-825 (1989). 20. DAVIDSON, I., XIAO, J. H., ROSALES.R., STAUB, A., and CHAMBON. P., Cell54, 931-942 (1988). 21. BIRKENMEIER.E. H.. GWYNN, B., HOWARD, S., JERRY,J., GORDON, J. I., LANDSCHULTZ,W. H., and MCKNIGHT, S. L., GenesDev. 3, 1146-1156 (1989). 22. CHODOSH, L. A., BALDWIN, A. S., CARTHEW, R. W., and SHARP, P. A., Cell 53, 1 l-24. 23. COHEN, R. B., SHIFFERY, M., and KIM, C. G., Mol. Cell. Biol. 6, 821-832(1986). 24. JOHNSON, P. F., LANDSCHU~Z, W. H., GRAVES, B. J., and MCKNIGHT, S. L., Genes Dev. 1, 133-146 (1987). 25. MARKOWITZ, R. B., and DYNAN, W. S., /. l&o/. 62, 3388-3398 (1988). 26. LASSAR.A. B., BUSKIN,J. N., LOCKSHON. D., DAVID, R. L., APONE, S., HAUSCHKA, S. D., and WEINTRAUB, H. Ce// 58, 823-831 (1989). 27. MULLER, P. R., and WOLD, W., Science 246, 780-786 (1989). 28. MAXAM, A., and GILBERT, W., Merh. Enzymol. 65, 499-560 (1980). 29. HENNIGHAUSEN,L., and LUBON. H., In “Methods in Enzymology” (S. L. Berger and A. R. Kimmel, Eds.), Vol. 152, pp. 721-735. Academic Press, San Diego, 1987. 30. FRISQUE, R. J., BREAM, G. L., and CANNELLA, G. L., J. Viol. 51, 458-469 (1984). 31. NAKSHATRI,H., PATER,A., and PATER, M. M., Virology 177,784789 (1990).
183, 374-380
(1991)
SHORT COMMUNICATIONS Activity and Enhancer Binding Factors for BK Virus Regulatory Elements in Differentiating Embryonal Carcinoma Cells HARIKRISHNA
NAKSHATRI,’
MARY
Division of Basic Medical Sciences, Faculty of Medicine,
Memorial
M. PATER, AND ALAN PATER* University
Received January 28, 199 1; accepted
of Newfoundland,
March
St. John’s, NF A 1B 3V6 Canada
19, 199 1
We have studied cell type specificity of expression of the human papovavirus BK regulatory elements in undifferentiated and differentiated embryonal carcinoma (EC) cells as a model system. While the activity of the regulatory elements of this virus was marginal in undifferentiated cells, differentiation by retinoic acid and DMSO resulted in a dramatic increase in the activity. To correlate in vivo activity of the regulatory elements with interaction with cellular transcription factors, we performed DNase I footprinting experiments. A GC-rich region was protected in both undifferentiated and differentiated cells. An additional four protected sites were detected in retinoic acid-differentiated cells and at least one of these additional sites was weakly protected in DMSO-differentiated cells. The sequences of the differentiated cell type-specific protected regions showed homology to a nuclear factor 1 (NF-1) binding motif and to a muscle creatine kinase gene enhancer motif. The intensity, competition, and pattern of protection of these sites were different in the two differentiated cell types, suggesting the involvement of different transcription factors regulating the _ -0 1991 Academic Press, Inc. activity of BKV regulatory elements in the two cell types.
difference in homology may contribute to the differences in host range, pathogenicity, and cell type specificity of these viruses. We have used the embryonal carcinoma (EC) cell line, P19, which can be induced to differentiate into a mixture of glial cells, neurons, and astrocytes upon retinoic acid (RA) treatment and into cardiac and skeletal muscle cells upon dimethyl sulfoxide (DMSO) treatment (14). Previous studies have shown that the intact BKV regulatory region containing three 68-bp enhancer-containing repeats is expressed ubiquitously while individual repeats are expressed in a cell-specific manner (15- 17). Our present studies provide evidence for the induction of two different sets of cellular factors important in the regulation of the control region of this virus following differentiation of P19 cells by RA and DMSO. The P19 EC cell line was maintained in a-MEM medium containing 1OYo fetal calf serum, and cells were induced to differentiate as described (14). DNA transfections were with the calcium phosphate precipitation procedure (18). Plasmid pBKcat (77), containing the bacterial chloramphenicol acetyl transferase (CAT) reporter gene (18) and BKV regulatory elements, was transfected into the undifferentiated and the retinoic acid- and DMSO-differentiated EC cell line, P19. The SV40 enhancer pSV2cat and the enhancerless pSVOcat were used as positive and negative controls,
One of the central challenges in eukaryotic molecular biology has been to determine the mechanisms of tissue-specific gene expression and the molecular events that alter gene expression during the differentiation process. By molecular genetic and biochemical approaches, it has been possible to identify c&-acting promoter and enhancer sequences of selected genes and transacting protein factors interacting with these elements and to assess their role in tissue-specific gene expression (1-3). The regulatory regions of viruses have served as excellent model systems for the study of tissue-specific and inducible gene expression (4, 5). Simian virus 40 (SV40) enhancer has been studied extensively and has been described as a “prototypic enhancer.” A systematic mutational analysis of this enhancer has permitted the identification of a number of domains, several of which are expressed in a cell-specific manner (6- 12). The human papovavirus BKV is a closely related virus which has 70-80% DNA sequence homology to that of SV40 ( 13); also, BKV and SV40 large tumor antigens cross-react immunologically. However, the regulatory elements of these viruses are only approximately 40% homologous. This ’ Present address: Inserm-U. 184 lnstitut de Chimie Biologique, Faculte de Medecine 11, Rue Humann. 67085 Strasbourg Cedex, France. ‘To whom request for reprints should be addressed. 0042-6822/91
$3.00
CopyrIght 0 1991 by Academic Press, Inc. All rtghts of reproduction in any form reserved.
374
SHORT COMMUNICATIONS
A
UD
RA
375 UD
DM F
B
UD
RA
B
1 2 3 4
5 6 7
8
DM
FIG. 1. Activity of BK regulatory elements in the undifferentiated (UD) and the retinoic acid (RA)- and DMSO (DM)-differentiated EC cell line, P19. Cells were transfected with 10 pg of the indicated plasmids and incubated for 48 hr. (A) Autoradiogram of the CAT assays. The percentage of conversion of [‘4C]chloramphenicol to acetylated chloramphenicol is given at the bottom and numbers greater than 100% were from assays for which diluted extracts had been used. (B) Spot-blot hybridization of episomal BKcat DNA from the Hirt supernatant with CAT coding sequence DNA as a probe.
respectively. CAT enzyme assays were used to measure quantitatively transcriptional activity of the BKV regulatory elements and the results are illustrated and presented numerically in Fig. 1A. In undifferentiated P19 cells, BKV regulatory elements were expressed only at low levels. However, differentiation resulted in approximately a 200-fold increase in retinoic acid-differentiated P19 cells and a greater than lOO-fold increase in DMSO-differentiated P19 cells. To evaluate whether the increased activity was due to differences in plasmid uptake, transfected DNA was isolated by the Hirt procedure and spot blotted. Autoradiography and liquid scintillation counting demonstrated that both DMSO-differentiated and undifferentiated cells contained the same amount of transfected DNA while retinoic acid-differentiated cells contained only 1.7-fold higher levels of transfected DNA (Fig. 1 B). These results introduced the suggestion that differentiation of EC cells was correlated with quantitative
FIG. 2. DNase I footprinting analysis of BK regulatory elements with undifferentiated P19 nuclear extracts. Chemical cleavage of purines (28) is in lane A/G. Probe treated with DNase I in the absence and presence of 25 pg of protein from nuclear extracts (29) are in lanes F and B, respectively. Lanes 1 to 8 represent competition experiments in which the same amount of extracts and probe were incubated with unlabeled competitor DNA before DNase I digestion. Competitor enhancer DNA fragments are: SV40 nucleotides 107240 (6) in lanes 1 and 2, JC nucleotides 5112-270 (30) in lanes 3 and 4, and the unlabeled probe BK in lanes 5 and 6. A control, nonspecific competitor pUCl9 DNA fragment, nucleotide 306-638, is in lanes 7 and 8. Amounts of competitors were 4 ng (lo-fold excess) and 20 ng in odd and even numbered lanes, respectively. The protected region is designated by the square bracket and numbered Ill. Arrows indicate hypersensitive sites.
SHORT COMMUNICATIONS
3i ‘6
A
RA
A’G F B 1 2 3
4
5 6
7 8 F
FIG. 3. DNase I footprinting with retinoic acid-differentiated nuclear extracts. (A) Competition is by papovaviral enhancer DNA fragments. Procedures and labeling are as in Fig. 2 except that five regions are designated. (B) Oligonucleotide competition experiments: Competition was with 100 ng (3200- to 1O,OOO-fold excess) of the indicated double-stranded oligonucleotides. Oligonucleotides corresponding to BK region It (lane 1; 5’ TGGGCAGCCAGCCA 3’; Fig. 6) and region Ill (lane 2; 5’ GAAACCCCGCCCC 3’; Frg. 6). and control oligonucleotide (lane J; BAAGGGAAGGGATGG 3’; a JC virus enhancer-promoter region sequence not affecting DNase protection; 31) were incubated with nuclear extract prior to addition of probe and DNase I digestion. Other procedures and labeling are as in Fig. 3A. Densitometric scans were performed to quantify competition by BK region II oligonucleotide. Scanned bands are indicated by the lines between lanes B and 1 and were selected since they were
and/or qualitative changes in cellular transcription factors. Therefore, DNase I footprinting experiments were undertaken to compare the binding of transcription factors to the regulatory region of this virus in various P19 cell types. The probe was a sense strand 32P-, 3’ end-labeled BKV DNA fragment, nucleotides 3 186 to 3479 (13). Figure 2 illustrates footprints of the BKV regulatory region in undifferentiated P19 cell extracts. Only one protected region was detected. This region was partially competed by BKV and SV40 DNA but not by JCV and pUC 19 DNA. This protected region spans a GC-rich sequence and has homology to the transcription factor Spl binding motif and to a transcription repressor protein binding site (19). In addition to the protected region, two hypersensitive sites were also obtained. The footprint pattern was completely different for retinoic acid-differentiated cell extracts (Fig. 3A), in which BKV regulatory elements function with greater efficiency (Fig. 1). A total of five protected regions were detected and four among them, designated I, II, IV, and V, have similarity to the binding site of transcription factor CTF/NF-I . Protected regions I and IV contained the sequence TGGN,CCA, which has been shown to be the most favorable NF-1 binding site, while regions II and V contained the sequence TGGN,CCA. Significantly, JCV regulatory elements, which also contain NF-1 binding sequences, competed, although more weakly than those of BKV, for these four protected regions. SV40 regulatory elements, which lack NF-1 binding sequences, competed, although poorly, only to region III. Region Ill is the GC-rich region which was also protected in undifferentiated cells. In addition to protected regions I to V, another region (Fig. 3A) which was not protected for UD (Fig. 2) was observed below region V. However, competition for protection was unconvincing for all four DNA competitors. To further examine the nature of the DNA-protein interactions, we performed oligonucleotide competition experiments of DNase I protection by the RA-differentiated cell extracts. Oligonucleotides corresponding to regions II and III were tested. Oligonucleotide for region II competed, albeit weakly, for regions I, II, IV, the most sensitive to competition in Fig. 3A. Relative absorbances of protected bands were calculated by comparison to the most proximal interregion bands, which are similarly indicated and are located between the regions I and II and IV and V. These bands were selected since they were proximal to protected regions and sufficiently distant from hypersensitive sites surrounding the protected regions. The percentage increase in relative absorbance in lane 1 compared to that of lane B was calculated. For the various regions competition was: I, 540%; II, 490%; IV, 380%; V. 300%. For region IV, “competition” values in control lanes were: lane 2, 50%; lane J, 99%.
SHORT COMMUNICATIONS
FIG. 4. DNase I footprinting by DMSO-differentiated nuclear extracts. Procedures and labeling are as in Figs. 2 and 3.
and V (Fig. 3B, lane 1). This competition, while visibly less than that for total enhancer fragments (Fig. 3A), was significant, as verified bydensitometric scan analysis (Fig. 3B legend). The relatively weak nature of the competition observed with this oligonucleotide could be due to the absence of flanking sequences which could confer positive cooperativity among adjacent protected regions in the enhancer fragments. Cooperativity could be involved in the competition for the four regions with homology to NF-1 binding sites. Region III oligonucleotide competition produced a pattern of pro-
377
tection distinct from that of region Il. Region III oligonucleotide competed only for the homologous region, indicating that the factor(s) binding to this region is unique and suggesting that binding is not positively cooperative with factors binding with neighboring regions. Conversely, competition with this oligonucleotide increased the degree of protection in region IV, especially toward the region Ill proximal side (Fig. 38; lane 2). Even the protection of the band in the center of region IV was 100% increased, as measured by densitometric scans (see the legend to Fig. 3B). This suggests that the interaction of a protein(s) with region III has a negative influence on protein binding to region IV, possibly through steric hinderance. For the DMSO-differentiated cells, the cell system in which BKV enhancer was active but to a lesser extent than in retinoic acid-differentiated cells, only sequences in region III were clearly protected (Fig. 4). BKV and SV40 regulatory elements, at high concentration, competed for protection of this site. The absence of convincing protection of sites in addition to those observed for undifferentiated cells is unexpected, since the activity of BKV in DMSO-differentiated cells was significantly higher than in undifferentiated cells. It is possible that differentiation of P19 cells by DMSO increases transcription factors which are unique from those protected in RA-induced cells and are not as readily detectable by DNase protection assays. To test this possibility, we performed footprints with double the amount of protein in extracts from undifferentiated and DMSO-treated cells. This double amount of protein was selected since the activity of the enhancer in RA-treated cells had also been double that for DMSOtreated cells (Fig. 1A). With undifferentiated cell extracts a partial protection of sequences corresponding to region IV was obtained (Fig. 5). The same region was reproducibly more completely protected when the increased level of DMSO-differentiated cell extracts was used (see Fig. 5 and the legend to Fig. 5 for densitometric scanning results). Furthermore, there were three additional weakly protected regions, corresponding to regions I, II, and V of retinoic acid-differentiated cells (for densitometric scanning results see Fig. 5 legend). These results support the unique natures of NF-1 site binding factors in the two differentiated cell types, since there is a visible difference in the degree of protection between that shown in Fig. 3 and that in Fig. 5. In addition, the protein concentration-dependent variation in footprints for DMSO-differentiated cell extracts suggests that some of the quantitative differences in in viva expression in the retinoic acid- and DMSO-differentiated cells (Fig. 1) could be due to quantitative differences in levels of NF-1 site binding factors.
SHORT COMMUNICATIONS
378
UD kGFB12345676F
kG F 6 1 2 3 4
5 6
7 8
F
FIG. 5. DNase I footprinting with increased levels of nuclear extracts. Procedures and labeling are as in Frgs. 2 to 4 except that 50 ag of protein in extracts was used. Brackets with solid lines and broken lines represent complete and partial protection, respectively. Densitometric scans of selected protected regions were performed to quantify protection and competition. The relative absorbances for the protected regions were determined as described in the legend to Fig. 3. The percentage of protection was determined from the extinction of the relative absorbance of protected bands in lane B compared to that rn lane F. For competition values, the percentages are the level to which the competitor had restored the relative absorbance from the level in lane B (0%) to that in lane F (100%). Protection in the various regions was as follows: UD. IV, 37%; DM, I, 90%; DM, II, 82%; DM, IV, 86%; DM, V, 66%. Competition for the various DM, region IV lanes was as follows: 2 (SV40), 88%; 4 (JC), 73%; 6 (BK), 90%; 8 (pUC), 30%.
The protected regions were differentially competed by all tested DNAs, with competition by DNA of BKV greater than that of SV40, greater than that of JCV, much greater than that of pUC (for densitometric scanning results for region IV competition see Fig. 5 leg-
end). Competition we have observed protected regions tif with homology which differs from
by SV40 DNA is not surprising since significant homology between BKVand SV40 enhancer motifs. One mois the sequence 5’ TGGAAIG 3’, the SV40 enhancer core region only
SHORT COMMUNICATIONS
NF-1 -
MCK
MCK -
I
FIG. 6. DNase l-protected nucleotide sequences of BK regulatory elements with all three cell types. Protected sequences are overlined for each of the P19 cell types indicated on the left. The numerals above the lines correspond to those in Fig. 3. Sequences homologous to consensus sequences within the protected regions, for known transcription factors, are underlined and the abbreviations for the transcription factors are indicated. MCK, the muscle creatine kinase gene K enhancer domain; SV-C, the SV40 core enhancer motif.
by the single underlined T (Fig. 6). The second is the sequence 5’ GCAlTCCAT in regions I and IV, which differs from the SV40 GT-IIC motif only by the two underlined nucleotides. It is interesting to note that in an analogous series of experiments examining protection/competition of labeled SV40 enhancer DNA, we observed protection of the Sphllll motif of the SV40 enhancer with DMSO-differentiated cell extracts and this protection was competed by BKV DNA (our unpublished observations). This suggests that the BKV enhancer’s protected regions I and IV interact with the transcription factor TEF-1 in DMSO-differentiated cells since the three SV40 motifs, Sphllll and GT-IIC, are all binding sites for the same TEF-1 factor (20). The results presented above demonstrate that there are dif-
379
ferences in both footprint and competition patterns in DMSO-differentiated cell extracts compared to those of the RA system. This leads us to suggest that BKV enhancer interacts with distinct factors in the two differentiated cell types. These factors would most likely be NF-1 or NF-l-like factors in RA-differentiated cells and TEF-1 or TEF-1 -like factors in DMSO-differentiated cells. The NF-1 and the related CCAAT motifs have been shown to bind a family of transcription factors, some of which are tissue specific (21-24). The evidence presented above for distinct factors binding to individual BKV NF-1 sites concurs with the results of Grinnell et a/. (16). They have shown with footprints of the BKV-P2 strain regulatory elements that regions II and V were protected only by HeLa cell extracts but not by 293 or MK2 cell extracts, whereas protection for regions I and IV was common to all three cell types. Furthermore, on the basis of results of in viva competition studies, they suggested that factors binding to regions II and V are distinct from those for regions I and IV and act as negative regulators of transcription. However, different conclusions were drawn from studies by Markowitz and Dynan (25) who were able to affinity purify protein NFBK, using oligonucleotides corresponding to region II, and to show that the same factor(s) binds to all NF-1 motifs of BK (Dunlop), BK (WW), and BK (MM) variants. The discrepancies between the results could be due to (1) BKV variant-specific differences in the sequences adjoining NF-1 motifs, (2) cell type-specific differences in NF-1 factors, and (3) inability to copurify, by affinity chromatography, noncovalently associated proteins. In addition to the NF-1 motifs and SV40 enhancer homologous sequences, the protected regions contained sequences that may act as binding sites for other known transcription factors. These include the sequence 5’ GCAGCCA 3’ (Fig. 6; labeled MCK) of regions I, II, IV, and V which were all protected in muscle cell type DMSO-treated cells. This sequence is also present in the K motif of muscle creatine kinase gene enhancer (26, 27). Another sequence of interest is TATGC, present in regions I and IV, which is identical to the EF-II sequence motif except for one nucleotide (I 1; Fig. 6). These motifs may provide diversity in BKV enhancer activity in various cell types. Thus, although the BKV enhancer and the SV40 prototypic enhancer have only 409/o sequence homology, both enhancers may be organized in an analogous fashion. This hypothesis can be tested by examining the effects of mutations of SV40 homologous and nonhomologous motifs on cell-specific expression by BKV enhancer. The advantages of the P19 EC cell system are that populations of cells in a variety of differentiated states
380
SHORT COMMUNICATIONS
can inexpensively, quickly, effectively, and directly be derived from undifferentiated cells with a single genotype. Thus, EC cells have been shown to represent a good system to study cell type-specific expression of an enhancer and to correlate this with molecular events. An unlikely disadvantage of this system concerns the presence of some heterogeneity within the population of cells. Therefore, one subpopulation of differentiated cells might allow expression while another subpopulation might contain the transcription factors that bind DNA. This potential but highly unlikely complication could be evaluated by in situ detection of both viral enhancer driven gene expression and biochemical markers for different cell types. ACKNOWLEDGMENTS We thank H. Hamada for providing the PI 9 cell line and K. Ross, S. Atkins, and W. Tucker for typing the manuscript. Harikrishna Nakshatri has been a predoctoral fellow of the Cancer Research Society Inc., Montreal, Quebec, Canada and more recently of the National Cancer Institute of Canada. This work was supported in part by the Medical Research Council and the National Cancer Institute of Canada.
REFERENCES 1. JOHNSON, P. F. and MCKNIGHT, S. L., Annu. Rev. Biochem. 58, 799-839 (1989). 2. MANIATIS, T., GOODBOURN, S.. and FISCHER,1. A., Science 236, 1231-1245 (1987). 3. MITCHELL, P. J., and TJIAN, R., Science 245, 371-378 (1989). 4. JONES, N. C., RIGBY, P. W. J., and ZIFF, E. B., Genes Dev. 2, 267-281 (1988). 5. MCKNIGHT, S., and TJIAN, R., Cell 46, 795-805 (1986). 6. FROMENTAL, C., KANNO, M., NOMIYAMA, H., and CHAMBON, P., Cell 54, 943-953 (1988). 7. KANNO, M., FROMENTAL,C., STAUB, A., RUFFENACH,F., DAVIDSON, I., and CHAMBON, P., EMBOJ. 8,4205-4214 (1989). 8. MACCHI, M., BORNERT,J. M.. DAVIDSON, I., KANNO, M.. ROSALES, R., VIGNERON,M., XIAO, J-H., FROMENTALC.. and CHAMBON, P., EMBO J. 8,42 15-4227 (1989).
9. ONDEK, B., GLOSS, L., and HERR, W., Nature 333, 40-45 (1988). 10. ONDEK, B., SHEPARD, S., and HERR, W., EMBOJ. 6, 1017-1025 (1987). 11. SCHIRM, S., JIRICNY,J., and SCHAFFNER,W., GenesDev. 1, 65-74 (1987). 12. ZENKE, M., GRUNDSTROM, T.. MATHES, H., WINTZERITH, M., SCHATZ, E., WILDEMAN, A., and CHAMBON, P., EMBOJ. 5,387397 (1986). 13. YANG, R. C. A., and Wu, R., Science 206, 456-462 (1979). 14. RUDNICKI, M. A., and MCBURNEY, M. W., ln “Teratocarcinoma and Embryonic Stem Cells: A Practical Approach” (E. J. Robertson, Ed.), pp. 19-49. IRL Press, Oxford, 1987. 15. DEYERLE. K. L., and SUBRAMANI, S. J., Virology 62, 3378-3387 (1988). 16. GRINNELL, B. W., BERG, D. T., and WALLS, J. D., Mol. Cell. Biol. 8, 3448-3457 (1988). 17. PATER,A., and PATER, M. M., Virology 163, 625-628 (1988). 18. GORMAN, C. M., MOFFAT, L. F., and HOWARD, B. H., Mol. Cell. Biol. 2, 1044-1051 (1982). 19. KAGEYAMA, R., and PASTAN, I., Cell 59, 815-825 (1989). 20. DAVIDSON, I., XIAO, J. H., ROSALES.R., STAUB, A., and CHAMBON. P., Cell54, 931-942 (1988). 21. BIRKENMEIER.E. H.. GWYNN, B., HOWARD, S., JERRY,J., GORDON, J. I., LANDSCHULTZ,W. H., and MCKNIGHT, S. L., GenesDev. 3, 1146-1156 (1989). 22. CHODOSH, L. A., BALDWIN, A. S., CARTHEW, R. W., and SHARP, P. A., Cell 53, 1 l-24. 23. COHEN, R. B., SHIFFERY, M., and KIM, C. G., Mol. Cell. Biol. 6, 821-832(1986). 24. JOHNSON, P. F., LANDSCHU~Z, W. H., GRAVES, B. J., and MCKNIGHT, S. L., Genes Dev. 1, 133-146 (1987). 25. MARKOWITZ, R. B., and DYNAN, W. S., /. l&o/. 62, 3388-3398 (1988). 26. LASSAR.A. B., BUSKIN,J. N., LOCKSHON. D., DAVID, R. L., APONE, S., HAUSCHKA, S. D., and WEINTRAUB, H. Ce// 58, 823-831 (1989). 27. MULLER, P. R., and WOLD, W., Science 246, 780-786 (1989). 28. MAXAM, A., and GILBERT, W., Merh. Enzymol. 65, 499-560 (1980). 29. HENNIGHAUSEN,L., and LUBON. H., In “Methods in Enzymology” (S. L. Berger and A. R. Kimmel, Eds.), Vol. 152, pp. 721-735. Academic Press, San Diego, 1987. 30. FRISQUE, R. J., BREAM, G. L., and CANNELLA, G. L., J. Viol. 51, 458-469 (1984). 31. NAKSHATRI,H., PATER,A., and PATER, M. M., Virology 177,784789 (1990).