Results and Discussion Identification of Preferred Mad1 DNA-Binding Sites. The DNA-binding specificity of Myc:Max heterodimers was originally defined from their preferential binding to E-box sequences among populations of completely randomized oligonucleotides ( 9, 21, 22). Although Mad:Max binds to E-box sequences in vitro ( 15), the intrinsic binding preferences of Mad:Max dimers have not been reported. It is conceivable that Mad possesses a specificity distinct from that of Myc. Thus, we performed an in vitro binding site selection to determine which DNA sequences are preferentially bound by Mad when presented with oligonucleotides containing all possible sequence combinations ( 22, 23). Selection was performed with a pool of oligonucleotides containing a central region of 20 randomized nucleotides surrounded by two regions of defined sequence 25 nucleotides in length. Flag epitope-tagged Mad1 (flg-Mad1) was transcribed and translated in vitro with untagged Max and mixed with the oligonucleotide pool, and Mad-Max DNA complexes were separated from unbound DNA by anti-Flag immunoprecipitation. To control for nonspecific antibody binding, untagged Max alone was assayed in parallel as a negative control. We monitored the progress of the selection by using PCR of 15, 20, and 25 cycles to verify selective enrichment of flg-Mad1 over the control. PCR products from the earliest PCR cycle that produced a visible band on an agarose gel were used as starting material for the subsequent round. After six rounds of selection, the PCR products were cloned into a plasmid vector and sequenced (see Methods and Materials). Fig. 1A shows the sequence data from the negative control and from the flag-Mad PCRs. None of the negative control sequences contains an E-box. In contrast, all but two of the flag-Mad selected sequences possess E-boxes, and all but one of these (FMad1) matches the canonical Myc E-box CACGTG. In addition to searching for E-boxes, we analyzed the data to identify other Mad-specific sequences but found none. | Fig 1.Preferential binding of Mad1 protein to canonical E-boxes. (A) 20-mer sequences independently derived from the SAAB (selected and amplified binding site) selection. The upper portion contains the first 15 sequences from the negative control (untagged (more ...) |
In addition to providing information about E-box binding, the data from the selection were used to identify flanking sequence preferences for Mad (Fig. 1B). Because the canonical Mad E-box is palindromic, the 5′ and 3′ flanking sequences were combined to provide a consensus half-site for Mad. Taken together with the core hexamer-binding data, the preferred dodecamer-binding site for Mad is predicted to be GACCACGTGGTC, which matches the preferred consensus site for Myc identified by others ( 24–27). No preferences were noted for any additional base pairs farther from the E-box. On the basis of their use of a chimeric Myc protein containing a basic region from Mxi1 in transient transfection assays, DePinho and colleagues ( 20) concluded that Myc and the Mad family protein Mxi1 differ in their preferences for sequences flanking the central E-box. We believe that the discrepancy with our findings is because these authors used only the basic region of Mxi1 linked to the Myc HLHZ domain in their chimeric protein. Studies of the DNA-binding activity of other bHLH proteins have revealed that changes in both helix 1 and the loop region can influence which sequences are bound by the basic region ( 28, 29). Thus, although the basic domain is the only region that makes direct contact with the nitrogenous bases of the E-box, it is unlikely to function independently of the HLH and zipper domains. Mad DNA Binding Is Sensitive to CpG Methylation. Although the binding sites for Myc and Mad seemed to be identical, we hypothesized that DNA binding at target genes might be affected by DNA modification. For example, mammalian genomes have undergone cytosine methylation of CpG dinucleotides within regions of inactive chromatin. Indeed, c-Myc and Max and Myc-Max heterodimers have been described to exhibit methylation-sensitive binding to the consensus sequence CA mCGTG ( 24, 30). To determine whether Mad possesses the ability to distinguish between methylated and unmethylated sequences, we performed electrophoretic gel mobility-shift assays with oligonucleotides chemically synthesized to contain either a methylated or an unmethylated E-box. The results demonstrated that Myc:Max and Mad:Max heterodimers bound to unmethylated CACGTG, but not to methylated CA mCGTG (Fig. 1C). A Myc Protein Containing the Mad1 bHLHZ Domain Activates Transcription of an E-Box Reporter Gene. The experiments described above suggest that the specificities for DNA recognition by Myc and Mad proteins are identical in our in vitro assays. To address the question of whether the bHLHZ region of Mad directs biological functions that are different from Myc, we generated a protein in which the C-terminal bHLHZ region of c-Myc is replaced by the cognate region derived from Mad1 (Fig. 2B). We used this fusion protein, Myc(MadbHZ), to determine whether its biological activities could recapitulate those of wild-type Myc. | Fig 2.Myc and the Myc(MadbHZ) fusion protein (MM). (A) Transcription assays using a synthetic promoter with four E-boxes controlling expression of the luciferase gene. Assays were performed after transient transfection of NIH 3T3 cells with 1 μg of (more ...) |
In transient transcription assays, Myc is capable of transcriptional activation of a synthetic reporter containing promoter-proximal E-box-binding sites ( 7, 31, 32). We therefore compared the transcriptional activity of wild-type c-Myc with that of the Myc(MadbHZ) fusion protein. Fig. 2A shows the transcriptional activities of Myc and Myc(MadbHZ), relative to the reporter alone. Addition of 5 ng of Myc or Myc(MadbHZ) transactivates the reporter roughly 4-fold with only a slight increase upon transfection of higher concentrations of expression vector. The effects observed here are consistent with the weak transcriptional activity of Myc observed by us and others ( 7, 31, 32). We conclude that Myc(MadbHZ) is capable of activating transcription in an E-box-dependent manner to the same extent as Myc. The Chimeric Myc(MadbHZ) Protein Stimulates Growth and Proliferation of myc-Null Cells. To assess the functions of the fusion protein further, we used a Rat-1 cell line in which both c- myc alleles have been deleted by targeted homologous recombination ( 33, 34). The c- myc−/− HO15.19 cells contain no detectable myc family gene products ( 34). These Myc-null cells display a distinct flattened morphology, relative the TGR-1 parental Rat-1 cells, and divide at a slower rate (doubling time approximately 50 h compared with 20 h for TGR-1 cells). Reintroduction of c- myc significantly decreases the doubling time to TGR-1 levels or less ( 34). HO15.19 myc−/− cells were cotransfected with a drug-resistance plasmid and either of the Flag-epitope-tagged constructs F-Myc or F-Myc(MadbHZ). Resistant cells were selected, pooled, labeled with [ 35S]methionine, and subjected to immunoprecipitation with anti-flag. Fig. 3A demonstrates that these pools of cells express roughly equivalent levels of the two tagged proteins. To assess the ability of Myc(MadbHZ) to rescue the proliferation defect of the c- myc-null cells, proliferation rates and cell cycle distributions were determined. Equal numbers of logarithmically growing parental TGR-1, myc−/−, and myc−/− cells expressing F-Myc and F-Myc(MadbHZ) were cultured and counted on successive days. Fig. 3B shows the data plotted on a semilogarithmic scale. The reduced division rate of the myc−/− cells is evident when compared with that of the parental TGR cells. The slope for cells expressing Myc(MadbHZ) from day 2 onward is equal to that for Myc, and both of these values are slightly lower than the parental TGR cells. | Fig 3.Characterization of activity of Myc(MadbHZ) and Myc in stably transfected myc−/− cells. (A) Equivalent expression of F-Myc and F-Myc(MadbHZ) (MM) proteins. Pools of stably transfected cells were labeled with [35S]methionine, (more ...) |
To measure more precisely the rescue of cellular proliferation during logarithmic growth, DNA content was determined using flow cytometry. The tabulated data showing cell cycle distribution are presented in Fig. 3C. Similar to previous reports, the myc−/− cells have a reduced S phase fraction relative to the TGR cells, and overexpression of Myc or Myc(MadbHZ) results in an increased S phase population in the myc null cells. An earlier analysis of the myc−/− Rat-1 cells had noted a decreased rate of growth as determined by measurement of RNA and protein synthetic levels ( 34). Because regulation of cell growth is thought to constitute a major activity of Myc ( 35–38), we used acridine orange to examine whether the F-Myc and F-Myc(MadbHZ) proteins would influence RNA accumulation. Fig. 3D shows the relative G 1 RNA content distributions within our experimental cell populations. Although the myc−/− cells have a markedly reduced amount of cellular RNA relative to the TGR cells, the introduction of either Myc or Myc(MadbHZ) results in an increase in the average RNA level within the cell population. Treatment with RNase A abolishes the signal, thus confirming that the results are specific for determination of RNA levels (data not shown). In these experiments the Myc(MadbHZ) protein reproduced many of the effects of Myc in cells. It transactivated E-box-driven artificial promoter, rescued the proliferation defect of myc null cells, increased the percentage of cells in S phase, and stimulated cell growth. Because the fusion protein contains the Mad1 bHLHZ, these results are consistent with the notion that Mad binds to the Myc target genes responsible for these biological effects. The Chimeric Myc(MadbHZ) Protein Does Not Stimulate Apoptosis. In our characterization of the myc−/− Rat-1 cell line, it was noted that under growth-arrest conditions a significant portion of the cells accumulated in G 2/M (23%) in contrast to the G 1 accumulation (91%) observed for arrested TGR cells ( 34). To determine whether F-Myc and F-Myc(MadbHZ) differed in their ability to affect cell cycle phasing during growth arrest, we analyzed cells arrested by contact inhibition in the presence of normal serum conditions (10% FBS). Cells were plated at about 80% confluence and grown to complete confluence in the course of 1 week, during which time the media were changed every 48 h. Once at confluence, cells were harvested and analyzed for cell cycle distribution. The data in Fig. 4B demonstrate that the myc−/−, TGR, and F-Myc(MadbHZ) cells have all arrested, as seen by their dramatically reduced S phase fraction. In contrast, 20% of the F-Myc cells are in S phase, indicating that these cells did not completely withdraw from the cell cycle despite their growth to high density. | Fig 4.Myc, but not Myc(MadbHZ), induces apoptosis. (A) Photomicrographs of Rat-1 cells following growth-factor withdrawal. Cells were grown to confluence, washed with PBS, and then refed with medium containing 10% or 0.1% serum. (B) DNA content (more ...) |
F-Myc-expressing cells also possess a substantial sub-G 1 population (18.6%), representing cells undergoing apoptosis, not found when F-Myc cells are grown at logarithmic phase (1.1%) (see Fig. 3C). For myc−/−, TGR, and F-Myc(MadbHZ), the rates of apoptosis are very low. Thus, when overexpressed, F-Myc seemed to override the growth-arrest signals of confluence and to drive cells into S phase. However, this continued division also resulted in significant cell death. Thus, although F-Myc(MadbHZ) seemed to rescue the proliferation defect of Myc-null cells during logarithmic growth and reverse the G 2/M block seen during growth arrest, it does not completely mimic the functions of F-Myc. The discrepancy between Myc and Myc(MadbHZ) in apoptotic function prompted a further examination of apoptotic effects of the two Myc constructs. Earlier studies have demonstrated that Myc can accelerate apoptosis induced by radiation damage or by withdrawal of cytokines or serum ( 39–41). We induced apoptosis by growing cells to confluence and then adding medium containing either 10% or 0.1% serum. Cells were analyzed at 24-h intervals. Fig. 4A shows photomicrographs taken at day 3. In 10% serum, the myc−/− cells and the F-Myc(MadbHZ) Rat-1 cells show little evidence of rounded refractile cells typical of those undergoing apoptosis. The F-Myc-rescued population displays a larger number of presumptive apoptotic cells. In 0.1% serum, the myc−/− and Myc(MadbHZ) cells are largely unchanged and show few signs of apoptosis. In contrast, the vast majority of cells expressing F-Myc are detached from the plate, highly refractile, and extremely small, all signs characteristic of apoptotic cells ( 40). Cells undergoing apoptosis are known to degrade genomic DNA into multiples of 200 bp, producing a “DNA ladder” visible on agarose gels ( 42). To confirm that a significant fraction of these cells are actually undergoing apoptosis, we quantitated the degree of genomic DNA degradation as described in Materials and Methods. Fig. 4C is a plot showing the percent of low molecular weight DNA in the different cell lines. The data demonstrate that a significant amount of genomic DNA (about 65%) is degraded in F-Myc-expressing cells when serum is withdrawn. In contrast, the percent of degradation in the Myc(MadbHZ) is equivalent to the myc−/− and the TGR source cells. These data confirm the results obtained with Myc and Myc(MadbHZ) under conditions of high-density growth arrest; overexpression of the Myc(MadbHZ) fusion protein fails to restore the sensitivity to apoptosis associated with overexpression of wild-type Myc. To exclude the possibility that the apoptotic effects were the result of using a particular myc−/− cell line, we used an independent, low-passage Rat-1 cell line to repeat the experiment. Rat-1 cells were stably transfected with a plasmid encoding drug resistance and either a vector control, Myc, or Myc(MadbHZ). The apoptosis experiment was repeated, and identical results were obtained as with the myc−/− cells (data not shown), which demonstrates that the apoptotic properties of Myc and Myc(MadbHZ) are not restricted to the myc−/− cell lines but are rather a reflection of the biological properties of these proteins in fibroblasts. Uncoupling of Proliferation and Apoptosis. The ability of Myc(MadbHZ) to recapitulate the proliferative and growth effects of Myc (Fig. 3) but not its apoptotic effects (Fig. 4) suggests that these functions are separable. Nonetheless, under apoptosis-inducing conditions we note a decreased S phase fraction in the Myc(MadbHZ) cells, and parental TGR cells, relative to cell-expressing Myc (Fig. 4). Thus, the ability to induce S phase under serum-limiting conditions may be related in an as-yet-unknown manner to Myc's ability to induce cell death. The notion that the proliferative and apoptosis functions of Myc can be uncoupled are consistent with two previous reports in which mutations within the N terminus of c-Myc were observed to have differential effects on Myc-induced proliferation, transformation, and apoptosis ( 43, 44). Our data are also consistent with a study showing that, although the Mad1 protein can attenuate apoptosis, it does not seem to do so by blocking the ability of Myc to stimulate apoptosis through cytochrome c release ( 45). Taken together, these data imply that distinct target genes and/or functions are involved in the proliferative and apoptotic responses to Myc. The effects of point mutations in the N-terminal region of c-Myc have been taken as evidence that a specific transcriptional activity, most notably repression, is involved in driving apoptosis ( 44). Thus, one explanation for an uncoupling of proliferation and apoptosis is that the target genes whose expression is modulated by Myc to accelerate apoptosis may be distinct from those involved in the proliferative response. Because the Myc(MadbHLHZ) protein would be expected to retain the transcriptional activities of c-Myc, the inability of the chimeric protein to induce apoptosis may be related to its inability to recognize and modulate expression of apoptosis-specific target genes. Given that Myc and Mad have identical in vitro DNA-binding specificities, as judged by our selection assays, it is likely that in vivo the Myc bHLHZ domain may mediate specific interactions with other proteins, such as Miz-1, that would in turn influence both Myc target gene specificity and transcriptional activity ( 12). |
References 1. Grandori C., Cowley, S. M., James, L. P. & Eisenman, R. N. (2000) Annu. Rev. Cell Dev. Biol. 16:, 653-699. [PubMed]. 2. Eisenman R. N. (2001) Genes Dev. 15:, 2023-2030. [PubMed]. 3. Baudino T. A. & Cleveland, J. L. (2001) Mol. Cell. Biol. 21:, 691-702. [PubMed]. 4. McMahon S. B., Wood, M. A. & Cole, M. D. (2000) Mol. Cell. Biol. 20:, 556-562. [PubMed]. 5. Amati B., Frank, S. R., Donjerkovic, D. & Taubert, S. (2001) Biochim. Biophys. Acta 1471:, M135-M145. [PubMed]. 6. Amati B., Dalton, S., Brooks, M. W., Littlewood, T. D., Evan, G. I. & Land, H. (1992) Nature (London) 359:, 423-426. [PubMed]7. Kretzner L., Blackwood, E. M. & Eisenman, R. N. (1992) Nature (London) 359:, 426-429. [PubMed]8. Stormo G. D. & Hartzell, G. W., III (1989) Proc. Natl. Acad. Sci. USA 86:, 1183-1187. [PubMed]. 9. Blackwell T. K., Huang, J., Ma, A., Kretzner, L., Alt, F. W., Eisenman, R. N. & Weintraub, H. (1993) Mol. Cell. Biol. 13:, 5216-5224. [PubMed]. 10. Li L., Nerlov, K., Prendergast, G., MacGregor, D. & Ziff, E. B. (1994) EMBO J. 13:, 4070-4079. [PubMed]. 11. Lee T. C., Li, L., Philipson, L. & Ziff, E. (1997) Proc. Natl. Acad. Sci. USA 94:, 12886-12891. [PubMed]. 12. Staller P., Peukert, K., Kiermaier, A., Seoane, J., Lukas, J., Karsunky, H., Möröy, T., Bartek, J., Massague, J., Hanel, F. & Eilers, M. (2001) Nat. Cell Biol. 3:, 392-399. [PubMed]. 13. Philipp A., Schneider, A., Västrik, I., Finke, K., Xiong, Y., Beach, D., Alitalo, K. & Eilers, M. (1994) Mol. Cell. Biol. 14:, 4032-4043. [PubMed]. 14. Marhin W. W., Chen, B. J., Facchini, L. M., Fornace, A. J. & Penn, L. Z. (1997) Oncogene 14:, 2825-2834. [PubMed]. 15. Ayer D. E., Kretzner, L. & Eisenman, R. N. (1993) Cell 72:, 211-222. [PubMed]. 16. Schreiber-Agus N., Torres, R., Horner, J., Lau, A., Jamrich, M. & DePinho, R. A. (1993) Mol. Cell. Biol. 13:, 2456-2468. [PubMed]. 17. Brubaker K., Cowley, S. M., Huang, K., Loo, L., Yochum, G. S., Ayer, D. E., Eisenman, R. N. & Radhakrishnan, I. (2000) Cell 103:, 655-665. [PubMed]. 18. Hurlin P. J., Quéva, C., Koskinen, P. J., Steingrímsson, E., Ayer, D. E., Copeland, N. G., Jenkins, N. A. & Eisenman, R. N. (1995) EMBO J. 14:, 5646-5659. [PubMed]. 19. Roussel M. F., Ashmun, R. A., Sherr, C. J., Eisenman, R. N. & Ayer, D. E. (1996) Mol. Cell. Biol. 16:, 2796-2801. [PubMed]. 20. O'Hagan R. C., Schreiber-Agus, N., Chen, K., David, G., Engelman, J. A., Schwab, R., Alland, L., Thomson, C., Ronning, D. R., Sacchettini, J. C., et al. (2000) Nat. Genet. 24:, 113-119. [PubMed]. 21. Blackwell T. K., Kretzner, L., Blackwood, E. M., Eisenman, R. N. & Weintraub, H. (1990) Science 250:, 1149-1151. [PubMed]. 22. Blackwell T. K. (1995) Methods Enzymol. 254:, 604-618. [PubMed]. 23. Blackwell T. K. & Weintraub, H. (1990) Science 250:, 1104-1110. [PubMed]. 24. Prendergast G. C., Lawe, D. & Ziff, E. B. (1991) Cell 65:, 395-407. [PubMed]. 25. Halazonetis T. D. & Kandil, A. N. (1991) Proc. Natl. Acad. Sci. USA 88:, 6162-6166. [PubMed]. 26. Papoulas O., Williams, N. G. & Kingston, R. E. (1992) J. Biol. Chem. 267:, 10470-10480. [PubMed]. 27. Solomon D. L. C., Amati, B. & Land, H. (1993) Nucleic Acids Res. 21:, 5372-5376. [PubMed]. 28. Davis R. L. & Weintraub, H. (1992) Science 256:, 1027-1030. [PubMed]. 29. Winston R. L. & Gottesfeld, J. M. (2000) Chem. Biol. 7:, 245-251. [PubMed]. 30. Prendergast G. C. & Ziff, E. B. (1991) Science 251:, 186-189. [PubMed]. 31. Amin C., Wagner, A. J. & Hay, N. (1993) Mol. Cell. Biol. 13:, 383-390. [PubMed]. 32. Gu W., Cechova, K., Tassi, V. & Dalla-Favera, R. (1993) Proc. Natl. Acad. Sci. USA 90:, 2935-2939. [PubMed]. 33. Hanson K. D., Shichiri, M., Follansbee, M. R. & Sedivy, J. M. (1994) Mol. Cell. Biol. 14:, 5748-5755. [PubMed]. 34. Mateyak M. K., Obaya, A. J., Adachi, S. & Sedivy, J. M. (1997) Cell Growth Differ. 8:, 1039-1048. [PubMed]. 35. Schmidt E. V. (1999) Oncogene 18:, 2988-2996. [PubMed]. 36. Johnston L. A., Prober, D. A., Edgar, B. A., Eisenman, R. N. & Gallant, P. (1999) Cell 98:, 779-790. [PubMed]. 37. Schuhmacher M., Staege, M. S., Pajic, A., Polack, A., Weidle, U. H., Bornkamm, G. W., Eick, S. & Kohlhuber, F. (1999) Curr. Biol. 9:, 1255-1258. [PubMed]. 38. Iritani B. M. & Eisenman, R. N. (1999) Proc. Natl. Acad. Sci. USA 96:, 13180-13185. [PubMed]. 39. Neiman P. E., Thomas, S. J. & Loring, G. (1991) Proc. Natl. Acad. Sci. USA 88:, 5857-5861. [PubMed]. 40. Evan G. I., Wyllie, A. H., Gilbert, C. S., Littlewood, T. D., Land, H., Brooks, M., Waters, C. M., Penn, L. Z. & Hancock, D. C. (1992) Cell 69:, 119-128. [PubMed]. 41. Askew D., Ashmun, R., Simmons, B. & Cleveland, J. (1991) Oncogene 6:, 1915-1922. [PubMed]. 42. Moyse E. & Michel, D. (1997) in Apoptosis Techniques and Protocols 29, ed. Poirer, J. (Humana, Totowa, NJ), pp. 133–159. 43. Chang D. W., Claasen, G. F., Hann, S. R. & Cole, M. D. (2000) Mol. Cell. Biol. 20:, 4309-4319. [PubMed]. 44. Conzen S. D., Gottlob, K., Kandel, E. S., Khanduri, P., Wagner, A. J., O'Leary, M. & Hay, N. (2000) Mol. Cell. Biol. 20:, 6008-6018. [PubMed]. 45. Gehring S., Rottmann, S., Menkel, A. R., Mertsching, J., Krippner-Heidenreich, A. & Luscher, B. (2000) J. Biol. Chem. 275:, 10413-10420. [PubMed]. 46. Geballe A. P. & Mocarski, E. S. (1988) J. Virol. 62:, 3334-3340. [PubMed]. |