Activation of collagen IV gene expression in F9 teratocarcinoma cells ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 267, No. 7, Issue of March 5, pp. 4988-4991, 1992 Printed in U.S A .

Activation of Collagen IV Gene Expression in F9 Teratocarcinoma Cells by 3-Deazaadenosine Analogs INDIRECTINHIBITORS

OF METHYLATION* (Received for publication, June 21, 1991)

Peter K. ChiangS,Peter D. Burbelo#, Sheryl A. Brugh, Richard K. Gopdon,Katsunori FukudaB, and Yoshihiko Yamada# From the Walter Reed Army Institute of Research, Washington, D. C. 20307-5100 and the §National Institute of Dental Research, National Institutes of Health, Bethesda, Maryland 20892

3-Deazaadenosine analogs can function as inhibitors and also as alternative substrates of S-adenosylhomocysteine (AdoHcy) hydrolase. In cells treated with the analogs, AdoHcy invariably accumulates, leading to inhibition of cellular methylation. F9 teratocarcinoma cells, stably transfected with two collagen (IV) promoter-enhancer-CAT constructs and treated with 10 M M 3-deazaadenosine, 3-deaza-(+)-aristeromycinor 3deazaneplanocin, showed a strong induction of CAT activities without affecting differentiation. In comparison, the same 3-deaza analogs did not affect the CAT activity in F9 cells transfected with the &actin promoter-CAT construct. Furthermore, Northern blot analysis of endogenous mRNA from wild-type F9 cells treated with the 3-deaza nucleosides all showed an induction of the collagen al(1V) chain mRNA. Thus, the 3-deaza analogs most likely affect DNA methylation because their results are consistent with the previous observation that the integrated collagen al(1V) promoter-enhancer constructs were activated with 6azacytidine.

and a2(IV) genes is regulated by a bidirectional promoter located between the two genes and by a common enhancer located in the first intron of the al(1V) chain gene (19). Promoter-enhancer-CAT constructs of the al(1V) collagen gene, when introduced stably into undifferentiated F9 teratocarcinoma cells, are not expressed, but are activated in the presence of retinoic acid and dibutyryl CAMP(18).Treatment of the cells stably transfected with the collagen promoterenhancer constructs with 5-azacytidine also activates transcription andacts synergistically with retinoic acid and CAMP.A correlation between the degree of demethylation of the 5”flanking region of al(1V) gene and its increased transcriptional activity hasbeen established (18).We report here that 3-deaza-Ado, 3-deaza-(+)-aristeromycin (3-deaza-Ari), and 3-deaza-neplanocin (3-deaza-Nep) all activated the al(1V)- or a2(IV) collagen promoter-enhancer-CAT constructs in stably transfectedF9 teratocarcinoma cells and also led to a strong expression of the endogenous al(IV) collagen chain mRNA in the F9 cells. EXPERIMENTALPROCEDURES

3-Deazaadenosine (3-deaza-Ado)’analogs have been found to be among the most potent inhibitors of S-adenosylhomocysteine (AdoHcy) hydrolase (Fig. 1).By inhibiting AdoHcy hydrolase, cellular levels of AdoHcy are increased with a concomitant inhibition of methylation reactions (1-6). Most often, S-adenosylmethionine (AdoMet) also accumulates as a reflection of the inhibition of methylation reactions (7-10). 3-Deaza nucleosides have been shown to have a variety of biological and biochemical effects (11-15). However, in spite of many interesting effects observed for the 3-deaza nucleosides, specific activation of transcriptionaleventshasnot been hereto reported, in comparison with the wide use of 5azacytidine to achieve DNA hypomethylation (16-18). Collagen IV is the major component of basement membranes and is composed of a heterotrimer consisting of two al(1V) and one a2(IV) chains. The transcription of the al(1V)

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 4 To whom correspondence should be addressed Applied Biochemistry Branch, Walter Reed Army Inst. of Research, Washington, D. C. 20307-5100. Tel.: 202-576-1361;Fax: 202-576-1304. The abbreviations used are: 3-deaza-Ado, 3-deaza-adenosine; CAT, chloramphenicol acetyltransferase; AdoMet, S-adenosylmethionine; AdoHcy, S-adenosylhomocysteine; 3-deaza-Ari, 3-deaza-(+)aristeromycin; 3-deaza-Nep, 3-deaza-neplanocin; kb, kilobase pair(s).

Cell Culture and DNATransfections-F9 teratocarcinoma cells were maintained in an undifferentiated state as described (18). The promoter-enhancer CAT constructs were established as described previously (18-20). The collagen al(1V) promoter-enhancer-CAT construct, p48, consists of 2.5 kb of the 5”flanking promoter region and 5.0 kb of the first intron (20). The ry2(IV) construct, p53, consists of a 0.7-kb fragment of the 5”flanking region (same as the 0.7-kb fragment of al(1V) of the 5”flanking region but in reverse) and the 5.0-kb fragment of the first intron of the al(1V)gene (19). The constructs were stably cotransfected in undifferentiated F9 cells with 2 pg of the CAT plasmids and 18 pg of RSVneo using the calcium phosphate method and subsequently selected by geneticin (18). For activation experiments, the cells were treated with one of the 3-deaza nucleosides at various concentrations for 48 h. 3-DeazaAdo and 3-deaza-Ari were supplied by Southern Research Institute, Birmingham, AL, and 3-deaza-Nep was kindly provided by Victor Marquez of the National Cancer Institute, Bethesda, MD. CAT Assuy-Treated cells (lo6)were rinsed with phosphate-buffered saline and then were scraped in 1 ml of the same solution into an Eppendorf tube. After centrifugation at 4 “C for 2 min, the cell pellet was frozen on dry ice for 2 min. The cell pellet was.thawed after adding 100 p1 of 0.2 M Tris-HC1, pH 7.9, and thensonicated for 10 s. The mixture was heated for 10 min a t 60 “C and then centrifuged; 75 pl of the supernatant was used for the CAT assay. The assay mixture contained (in a final volume of 90 pl), 0.6 mM acetyl-coA and 0.25 pCi of [14C]chloramphenicol.After incubation for 2 h at 37 “C, 500 p1 of ethyl acetate was added, shaken for 5 min, and then centrifuged. Next, 450 pl of the organic layer was dried and taken up in 30 p1 of ethyl acetate, spotted on silica gel thin-layer plates, and developed with chloroform/methanol(95:5,ascending). Radioactivity was detected by autoradiography. Densitometry of CAT Actiuity-X-ray film of CAT activity was scanned by a Hewlett-Packard scanner connected to a Macintosh

4988

3-Deaza Nucleosides Induce

Collagen I V Expression Gene a

adenosine

3deaza-adenosine 3-deaza-arkteromycin (3-deaza-Ado) (3deaza-Nep) (3deaza-Ari) Ki=lfl Ki= 1 n M

4989

pS3 ( I V ) promoter-enhancer

3-deaza-neplanocin

& = 50pM FIG. 1. 3-Deazaadenosine and its analogs as inhibitors of Sadenosylhomocysteine hydrolase (isolated from hamster liver) with their respective K i values (Refs. 2 and 5).

computer using the Image 1.31software developed by Wayne Rasband of the National Institute of Mental Health, Bethesda, MD. Northern Blot Hybridization of otl(IV) Collagen cDNA and 0-Actin cDNA-A 1.8-kb (BarnHI-PstI) of cDNA flanking the 3' end of the tul(1V) collagen gene was used as a probe in Northern (transfer) blot analysis of mRNA isolated from wild-type F9 cells. 0-Actin cDNA from chick brain(21) was digested with PstI togenerate 1.8-kb fragments for rehybridization to the sameblot.

b

8-Actin-CAT

RESULTS

CAT reporter gene activity of F9 teratocarcinoma cells stably transfected with collagen a2(IV) promoter-enhancer p53 is shown in Fig. 2a. After treatment for 48 h with 10 p~ each of 3-deaza-Ado, 3-deaza-Ari, or 3-deaza-Nep, there was equal activationof the CAT activity inthe F9 cells transfected with the p53 construct. However, no activation of the CAT activity could be observed after treatment for only 24 h with the nucleosides (not shown). In comparison, the @-actin CAT construct, which served as acontrol, showed similar high levels of CAT activity whether the F9 cells were treated with each of the 3-deaza nucleosides or not (Fig. 2b). In F9 cells transfected with another collagen al(1V) promoter-enhancer CAT construct p48, the CAT activity was undetectable until the cells were treated with the 3-deaza nucleosides. A dose-dependent activationof the CAT activity was observed after incubation with 3-deaza-Ado for 48 h (Fig. 3). The CAT activity was almost linear between 5 and 20 p~ 3-deaza-Ado. The inclusion of homocysteine thiolactone up to 250 p~ in the culture medium did not potentiate theCAT activity. At 1 or 10 p~ 3-deaza-Ari, higher CAT activity was observed when compared with 3-deaza-Ado. In contrast,cells treated with 20 p~ 3-deaza-Ari expressed lower CAT activity. Interestingly, 0.3 p~ 3-deaza-Nep achieved a level of CAT activity almost equal to 3.0 or 10.0 pM 3-deaza-Nep. The above results suggested that the treatment of the F9 cells with the 3-deaza nucleosides would lead to anexpression of collagen (IV) genes. To test this hypothesis, 1.8-kb cDNA from the 3' end of the al(1V) collagen gene was used as a probe in the Northernblot analysis of mRNA from wild-type F9 cells treated with each of the 3-deaza nucleosides. Whereas the al(1V) collagen gene was not detectable in control F9 cells, treatment of the cells with each of the 3-deaza nucleosides activated the expression of the gene (Fig. 4). Consistent with the observed dose-dependent effect of 3-deaza-Ado (Fig. 3), there was a stronger expression of the d(1V) collagen gene in F9 cells treated with 20 p~ 3-deaza-Ado than those treated with 10 p ~ Unlike . the induction of al(1V) collagen mRNA by the 3-deaza nucleosides, ,&actin mRNA did not show any significant change regardless of treatment (Fig. 4). The latterobservation was consistent with the lack of effects of the 3-deaza nucleosides on the CAT activity of @-actin

FIG. 2. a, effect of 3-deaza-Ado, 3-deaza-Ari, and 3-deaza-Nep on the stably integrated collagen a2(IV) promoter-enhancer-CAT p53 constructs in F9 teratocarcinoma cells. b, effect of 3-deaza-Ado, 3deaza-Ari, and 3-deaza-Nep on stably integrated @-actin-CAT construct in F9 teratocarcinomacells.

promoter-CAT construct (Fig. 2b). In contrast to the effect of retinoic acid and dibutyryl CAMPon the differentiation of F9 into parietal endoderm-like cells (22, 23), no visual effect on the differentiation or morphology of F9 cells was observed with any of the 3-deaza nucleosides. DISCUSSION

As shown previously, DNA methylation appears to be an important mechanism in the controlling the gene expression of collagen IV genes (18).5-Azacytidine activates collagen IV transcription synergistically with retinoic acid and CAMP, and in thecells treated with the latter two agents, a specific demethylation of the DNA within the 5'-flanking region takes place (18).The results presented here for the activation of the promoter-enhancer-CAT activities and the expression of d ( 1 V ) collagen gene by the 3-deaza nucleosides lent additional evidence to the regulation of collagen IV gene activity by methylation/hypomethylation.It has also been reported that the expression of al(1) collagen gene is associated with DNA hypomethylation and changes in DNase I sensitivity sites (23). To our knowledge, this is the first time that a specific gene activation has been observed with each of the 3-

4990

IV Gene Expression

3-Deaza Nucleosides Induce Collagen

5

10

34eaza-Ado

20

I

1

10 3-deaza-An

20

I

03

30

100

3-dcsza-Ncp

bM)

FIG. 3. Effect of 3-deaza-Ado, 3-deaza-Ari, and 3-deazaNep on the stably integrated collagen (rl(1V) promoter-enhancer-CAT p48 construct in F9 teratocarcinoma cells. Top panel shows computer-scanned promoter-enhancer-CAT activities; bottompanel shows densitometry of the promoter-enhancer-CAT activities, as represented by the front row in the top panel. The first column (unmarked) is the control (untreated).

collagen IV

within the constructshave effects on the differential expression of CAT activities (19, 20). Unless 3-deaza-Ado had secondary biological effects on F9 cells, no additional reasons could be established to explain the difference in the sensitivity of the two collagen CAT constructs to 3-deaza-Ado (Figs. 2 and 3). Nevertheless, the ability of 3-deaza-Nep and 3-deazaAri to activate the collagen al(1V) promoter-enhancer-CAT construct p48 at much lower concentrations than 3-deazaAdo (Fig. 3) seemed to correlate with the Kivalues of these nucleosides for the AdoHcy hydrolase (see Fig. 1). Unlike the previous induction of cellular differentiation of 3T3-Ll fibroblasts by 3-deaza-Ado (15, 28) and HL-60 promyelocytic leukemia cells by 3-deaza-Ari (29), the three 3deaza nucleosides appeared not to affect the differentiation state of the F9 cells. Although the data presented here were suggestive of the hypomethylation of the collagen IV promoter-enhancer, as expected from the overall inhibition of AdoMet-dependentmethylationreactions including DNA methylation (29) by the 3-deaza nucleosides (5, 6, 11, 13), it could not be completely ruled out that another more critical methylation reaction, other thanDNA methylation, might be inhibited instead. It is also possible that the activation of critical transcription factors(s) by the 3-deaza nucleosides may be responsible for the induction of collagen IV expression. Further experiments usinggenomic sequencing may elucidate whether sites within the collagen IV promoter-enhancer have altered CpG methylation when treated with the 3-deaza nucleosides. REFERENCES

1. Liotta, L. A., Mandler, R., Murano, G., Katz, D. A., Gordon, R. K., Chiang, P. K., and Schiffmann, E. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,3302-3306 2. Kim, I.-K., Zhang, C.-Y., Chiang, P. K., andcantoni, G. L. (1983) Arch. Biochem. Biophys. 226,65-72 3. Miura, G.A., Santangelo, J. R., Gordon, R. K., and Chiang, P. K. (1984) Anal. Biochem. 141, 161-167 28 s 4. Montgomery, J. A., Clayton, S. J., Thomas, H. J., Shannon, W. 18 s M., Arnett, G., Bodner, A. J., Kim., I.-K., Cantoni, G. L., and Chiang, P. K. (1982) J. Med. Chem. 25,626-629 5. Glazer, R. I., Hartman, K. D., Knode, M. C., Richard, M. M., Chiang, P. K., Tseng, C. K. H., and Marquez, V. E. (1986) Biochem. Biophys. Res. Commun. 135,688-694 p-Actin 6. Greenberg, M. L., Chaffee, S., and Hershfield, M. S. (1989) J. Biol. Chem. 264,795-803 *um@lGR 7. Shattil, S. J., Montgomery, J. A., and Chiang, P. K. (1982) Blood 59,906-912 FIG. 4. Northern blot analysis of expression of mRNA of 8. Smith, J. D., and Ledoux, D. N. (1990) Biochim. Biophys. Acta al(1V) collagen gene and @-actingene in wild-type F9 cells 1047,290-293 treated with 3-deaza nucleosides. 9. Chiang, P. K., and Miura, G. A. (1986) in Biological Methylation (Borchardt, R. T., Creveling, C. R., and andDrugDesign Ueland, P. M., eds) pp. 239-251, Humana Press, Clifton, NJ deaza nucleosides. Thetranscription activation of certain genes has been correlated with the hypomethylation of spe- 10. Hyman, B. T., Stoll, L. L., and Spector, A. A. (1982) Biochim. Biophys. Acta 713,375-385 cific CpG sites in DNA (24-27). The lack of effect of the 311. Chiang, P. K. (1985) Methods Pharmacol. 6,127-145 deaza nucleosides on P-actin promoter-CATactivity or 12. Tseng, C. K. H., Marquez, V. E., Fuller, R. W., Goldstein, B. M., mRNA was consistent withan earlier observationthat P-actin Haines, D. R., McPherson, H., Parsons, J. L., Shannon, W. M., promoter-CAT construct is not regulated by methylation (17). Arnett, G., Hollingshead, M., and Driscoll, J. S. (1989) J. Med. Chem. 32, 1442-1446 In F9 cells transfectedwith collagen a2(IV) promoterenhancer-CAT construct p53, similar activation of the CAT 13. Razin, A., Levine, A. Kafri,T., Agostini, S., Gomi, T.,and Cantoni, G. L. (1988) Proc. Natl. Acad. Sci. U. S. A. 85,9003activity was observed (Fig. 2a). Incontrast, much higher 9006 concentrations of 3-deaza-Ado were required to activate the 14. Wiesmann, W. P., Johnson, J. P., Miura, G. A,, and Chiang, P. CAT activity of collagen al(1V) promoter-enhancer-CAT K. (1985) Am. J. Physiol. 248, F43-F47 construct p48 (Fig. 3), or theexpression of al(1V) mRNA in 15. Chiang, P. K. (1981) Science 211, 1164-1166 wild-type cells, than those of 3-deaza-Ari or 3-deaza-Nep. The 16. Jones, P.A., and Taylor, S. M. (1980) Cell 20,85-93 latter observation might be explained by the 1000-fold or 17. Lassar, A. B. , Paterson, B. M., and Weintraub, H. (1986) Cell 47,649-656 larger Ki value of 3-deaza-Ado than 3-deaza-Ari or 3-deazaP. D., Horikoshi, S., and Yamada, Y. (1990) J. Biol. Nep for AdoHcy hydrolase (see Fig. 1). On the other hand, 18. Burbelo, Chem. 265,4839-4843 DNA methylase(s) involved might exhibit differential sensi- 19. Burbelo, P. D., Martin, G. R., and Yamada, Y. (1988) Proc. Natl. tivities toAdoHcy. As has been reported, variations in length, Acad. Sci. U. S. A. 85,9679-9682 topological, or directional arrangement of the promoter region 20. Killen, P. D., Burbelo, P. D., Martin, G.R., and Yamada, Y.

3-Deaza Nucleosides Induce Collagen IV Gene Expression

4991

(1988)J. Biol. Chem. 263, 12310-12314 27. Cedar, H. (1988)Cell 53,3-4 21. Cleveland, D. W V LoPat% M.A.9 MacDonald, R. Cowan, N. 28. Chiang, P. K., Brown, N. D., Padilla, F. N., and Gordon, R. K. J., Rutter, W. J., and Kirschner, M. W. (1980)Cell 20,95-105. (1987)in Tumor Cell Differentiation: Biology and Pharmacology 22. Strickland, S., Smith, K. K., and Marotti, K. R. (1980)Cell 24, (Aarbakke, J., Chiang, P. K., and Koeffler, H. P.,eds) pp. 231347-355 240,Humana Press, Clifton, NJ 23. Lakshmi, T.M. (1988)Indian J. Ezp. Biol. 26,922-928 24. Bird, A. P. (1986)Nature 321, 209-21329. Aaarbakke, J., Miura, G. A., Prytz, P. S., Bessesen, A., SlBrdal, 25. Doerfler, W. (1983)Annu. Rev. Biochem. 52.93-124 L., Gordon, R. K., and Chiang, P. K. (1986)Cancer Res. 46, 26. Riggs, A. D., and Jones, P. A. (1983)Adu. Cancer Res. 40, 1-30 5469-5472 J.9