High mobility group protein-1 (HMG-1) is a unique activator of p53

Abstract

The binding of p53 protein to DNA is stimulated by its interaction with covalent as well as noncovalent modifiers. We report the identification of a factor from HeLa nuclear extracts that activates p53 DNA binding. This factor was purified to homogeneity and identified as the high mobility group protein, HMG-1. HMG-1 belongs to a family of highly conserved chromatin-associated nucleoproteins that bend DNA and facilitate the binding of various transcription factors to their cognate DNA sequences. We demonstrate that recombinant His-tagged HMG-1 enhances p53 DNA binding in vitro and also that HMG-1 and p53 can interact directly in vitro. Unexpectedly, HMG-1 also stimulates DNA binding by p53Delta30, a carboxy-terminally deleted form of the protein that is considered to be constitutively active, suggesting that HMG-1 stimulates p53 by a mechanism that is distinct from other known activators of p53. Finally, using transient transfection assays we show that HMG-1 can increase p53 and p53Delta30-mediated transactivation in vivo. HMG-1 promotes the assembly of higher order p53 nucleoprotein structures, and these data, along with the fact that HMG-1 is capable of bending DNA, suggest that HMG-1 may activate p53 DNA binding by a novel mechanism involving a structural change in the target DNA.

Figure 1

(A) HeLa nuclear fraction P-11.85 stimulates p53 DNA binding. p53 protein (90 ng) was bound to ³²P-labeled GADD45 oligonucleotides either alone (lane 1) or in the presence of the P-11.85 fraction and p53 monoclonal antibodies pAb 421 and pAb 1801 or competitor DNA as indicated and analyzed by EMSA and autoradiography. Lanes 3 and 5 contain the P-11.85 fraction heat-treated for 5 min at 70°C. Reaction mixtures in lanes 4 and 5 contain no p53 protein. Wild-type (wt) or mutant (mt) consensus site-containing oligonucleotides were added in 10-fold (lanes 6,8) or 20-fold (lanes 7,9) molar excess over probe. The arrow indicates the p53–DNA complex; the bracket indicates p53–DNA complexes supershifted by antibodies. Not shown is free unbound probe at bottom of gel. (B) Purification of the p53 activator protein from HeLa nuclear fraction P-11.85. (Top) Protocol for purification of p53 DNA-binding stimulatory activity from HeLa nuclear extract. (Bottom) Aliquots of peak fraction from each step in A were analyzed by SDS-PAGE and silver staining. The arrow indicates purified protein.

Figure 1

(A) HeLa nuclear fraction P-11.85 stimulates p53 DNA binding. p53 protein (90 ng) was bound to ³²P-labeled GADD45 oligonucleotides either alone (lane 1) or in the presence of the P-11.85 fraction and p53 monoclonal antibodies pAb 421 and pAb 1801 or competitor DNA as indicated and analyzed by EMSA and autoradiography. Lanes 3 and 5 contain the P-11.85 fraction heat-treated for 5 min at 70°C. Reaction mixtures in lanes 4 and 5 contain no p53 protein. Wild-type (wt) or mutant (mt) consensus site-containing oligonucleotides were added in 10-fold (lanes 6,8) or 20-fold (lanes 7,9) molar excess over probe. The arrow indicates the p53–DNA complex; the bracket indicates p53–DNA complexes supershifted by antibodies. Not shown is free unbound probe at bottom of gel. (B) Purification of the p53 activator protein from HeLa nuclear fraction P-11.85. (Top) Protocol for purification of p53 DNA-binding stimulatory activity from HeLa nuclear extract. (Bottom) Aliquots of peak fraction from each step in A were analyzed by SDS-PAGE and silver staining. The arrow indicates purified protein.

Figure 2

Identification of HMG-1 as the p53 stimulating protein from HeLa P-11.85 fraction. (A) HMG-1 purified from calf thymus or HeLa P-11.85 Superdex-75 peak fraction or immunopurified p53 protein was subjected to SDS-PAGE and transferred to nitrocellulose, and the resulting blots were probed with affinity-purified polyclonal anti-HMG-1 antibody. (B) DNA binding by p53 (90 ng) in the absence (lane 1) or presence of calf thymus-purified HMG-1 (100 ng) (lanes 2–8) was analyzed by EMSA as in Fig. 1. Antibodies and 20-fold excess competitor DNA were added as indicated. Lane 8 contains mixtures in which antibody and proteins were added simultaneously, whereas in lane 9 HMG-1 was preincubated with anti-HMG-1 antibody for 20 min prior to addition to reaction mixture. (C) Recombinant HMG-1 enhances DNA binding by p53. DNA-binding mixtures contained recombinant His–HMG-1 (50 or 100 ng), either in the absence (lanes 5,6) or presence of p53 (90 ng) (lanes 2–4). Lane 1 contains probe alone. (D) DNA-binding reaction mixtures were as above but in the absence of poly[d(I-C)]. Binding mixtures contained either 8 (lanes 2–4,12,13) or 4 ng (lanes 7–9,10,11) of p53 in the absence (lanes 2,7,10,11) or presence (lanes 3,4, 8,9,12,13) of His–HMG-1 (60 or 120 ng). HMG-1 (60 ng) was present in mixtures shown in lanes 12 and 13. Monoclonal antibody pAb 421 was added as indicated. Only the DNA–protein complexes are shown as in Fig 1A.

Figure 2

Identification of HMG-1 as the p53 stimulating protein from HeLa P-11.85 fraction. (A) HMG-1 purified from calf thymus or HeLa P-11.85 Superdex-75 peak fraction or immunopurified p53 protein was subjected to SDS-PAGE and transferred to nitrocellulose, and the resulting blots were probed with affinity-purified polyclonal anti-HMG-1 antibody. (B) DNA binding by p53 (90 ng) in the absence (lane 1) or presence of calf thymus-purified HMG-1 (100 ng) (lanes 2–8) was analyzed by EMSA as in Fig. 1. Antibodies and 20-fold excess competitor DNA were added as indicated. Lane 8 contains mixtures in which antibody and proteins were added simultaneously, whereas in lane 9 HMG-1 was preincubated with anti-HMG-1 antibody for 20 min prior to addition to reaction mixture. (C) Recombinant HMG-1 enhances DNA binding by p53. DNA-binding mixtures contained recombinant His–HMG-1 (50 or 100 ng), either in the absence (lanes 5,6) or presence of p53 (90 ng) (lanes 2–4). Lane 1 contains probe alone. (D) DNA-binding reaction mixtures were as above but in the absence of poly[d(I-C)]. Binding mixtures contained either 8 (lanes 2–4,12,13) or 4 ng (lanes 7–9,10,11) of p53 in the absence (lanes 2,7,10,11) or presence (lanes 3,4, 8,9,12,13) of His–HMG-1 (60 or 120 ng). HMG-1 (60 ng) was present in mixtures shown in lanes 12 and 13. Monoclonal antibody pAb 421 was added as indicated. Only the DNA–protein complexes are shown as in Fig 1A.

Figure 2

Identification of HMG-1 as the p53 stimulating protein from HeLa P-11.85 fraction. (A) HMG-1 purified from calf thymus or HeLa P-11.85 Superdex-75 peak fraction or immunopurified p53 protein was subjected to SDS-PAGE and transferred to nitrocellulose, and the resulting blots were probed with affinity-purified polyclonal anti-HMG-1 antibody. (B) DNA binding by p53 (90 ng) in the absence (lane 1) or presence of calf thymus-purified HMG-1 (100 ng) (lanes 2–8) was analyzed by EMSA as in Fig. 1. Antibodies and 20-fold excess competitor DNA were added as indicated. Lane 8 contains mixtures in which antibody and proteins were added simultaneously, whereas in lane 9 HMG-1 was preincubated with anti-HMG-1 antibody for 20 min prior to addition to reaction mixture. (C) Recombinant HMG-1 enhances DNA binding by p53. DNA-binding mixtures contained recombinant His–HMG-1 (50 or 100 ng), either in the absence (lanes 5,6) or presence of p53 (90 ng) (lanes 2–4). Lane 1 contains probe alone. (D) DNA-binding reaction mixtures were as above but in the absence of poly[d(I-C)]. Binding mixtures contained either 8 (lanes 2–4,12,13) or 4 ng (lanes 7–9,10,11) of p53 in the absence (lanes 2,7,10,11) or presence (lanes 3,4, 8,9,12,13) of His–HMG-1 (60 or 120 ng). HMG-1 (60 ng) was present in mixtures shown in lanes 12 and 13. Monoclonal antibody pAb 421 was added as indicated. Only the DNA–protein complexes are shown as in Fig 1A.

Figure 2

Identification of HMG-1 as the p53 stimulating protein from HeLa P-11.85 fraction. (A) HMG-1 purified from calf thymus or HeLa P-11.85 Superdex-75 peak fraction or immunopurified p53 protein was subjected to SDS-PAGE and transferred to nitrocellulose, and the resulting blots were probed with affinity-purified polyclonal anti-HMG-1 antibody. (B) DNA binding by p53 (90 ng) in the absence (lane 1) or presence of calf thymus-purified HMG-1 (100 ng) (lanes 2–8) was analyzed by EMSA as in Fig. 1. Antibodies and 20-fold excess competitor DNA were added as indicated. Lane 8 contains mixtures in which antibody and proteins were added simultaneously, whereas in lane 9 HMG-1 was preincubated with anti-HMG-1 antibody for 20 min prior to addition to reaction mixture. (C) Recombinant HMG-1 enhances DNA binding by p53. DNA-binding mixtures contained recombinant His–HMG-1 (50 or 100 ng), either in the absence (lanes 5,6) or presence of p53 (90 ng) (lanes 2–4). Lane 1 contains probe alone. (D) DNA-binding reaction mixtures were as above but in the absence of poly[d(I-C)]. Binding mixtures contained either 8 (lanes 2–4,12,13) or 4 ng (lanes 7–9,10,11) of p53 in the absence (lanes 2,7,10,11) or presence (lanes 3,4, 8,9,12,13) of His–HMG-1 (60 or 120 ng). HMG-1 (60 ng) was present in mixtures shown in lanes 12 and 13. Monoclonal antibody pAb 421 was added as indicated. Only the DNA–protein complexes are shown as in Fig 1A.

Figure 3

HMG-1 can stimulate DNA binding by p53Δ30 as well as by core p53. (A) Ten nanograms (lanes 1–3) or 20 ng (lanes 4–6) of p53Δ30 was bound to ³²P-labeled GADD45 oligonucleotides and analyzed by EMSA in the absence (lanes 1,4) or presence (lanes 2,3,5,6) of HMG-1 (50 or 100 ng). (B) DNA-binding mixtures contained either core domain p53 (lanes 1–3) or p53Δ96 (lanes 4–6) in the presence (lanes 2,3,5,6) or absence (lanes 1,4) of HMG-1 (50 or 100 ng). Shown are p53–DNA complexes only as in Fig. 1A.

Figure 4

HMG-1 promotes co-operative p53 interactions on DNA. (A, Left) Reaction mixtures contained increasing amounts of HMG-1 (50, 75, 100, or 150 ng) in the absence (lanes 7–9) or presence (lanes 3–6) of full-length p53 (50 ng) using a 120-bp ³²P-labeled DNA fragment containing the GADD45 p53 binding site. Lane 1 contains no protein. Mixtures in lane 2 contained 50 ng of p53 alone. (Right) Reaction mixtures contain increasing amounts (50, 100, 200, or 400 ng) of p53 (lanes 2–5) or increasing amounts of HMG-1 (50 or 125 ng) with 50 ng of p53 (lanes 6,7). Lane 1 contains no protein. (B) Reaction mixtures contained p53Δ30 (10 ng), alone (lane 2) or with 75 (lanes 3,6), 100 (lanes 4,7), or 150 (lanes 5,8) ng of HMG-1 with or without excess anti-HMG-1 antibody as indicated. Mixtures in lanes 9–11 contained HMG-1 protein at the above concentrations without p53.

Figure 5

p53 binds HMG-1 in vitro. Calf-thymus-purified HMG-1 (0.25, 0.5, 1.0, and 2.0 μg), truncated dl166 HDM2 (1 μg) or full-length HDM2 (1 μg) as indicated, and molecular mass marker polypeptides were resolved by SDS-PAGE and transferred to nitrocellulose. (Right) Filter was incubated with human p53 protein (1 μg in 5 ml) followed by detection with pAb 421 as described in Materials and Methods. (Left) Ponceau S staining of filter. The lower amounts of HMG-1 were below the level of detection by Ponceau S. Arrows indicate position of HMG-1 protein.

Figure 6

HMG-1 stimulates transactivation by wild-type p53 as well as p53Δ30 in transfected cells. (A,B) H1299 cells were transiently transfected with wild-type p53 (300 ng) (A) or p53Δ30 (35 ng) (B) or HMG-1 expression plasmid (2.5 μg) or both p53 and HMG-1 expression plasmids as indicated. In all cases, total DNA transfected was normalized with equivalent amounts of control parental vectors. The reporter construct used was cyclin G–luc (1 μg). (Y-axis) Luciferase activity is represented as fold transactivation relative to samples transfected with reporter constructs and parental control plasmids. The fold activation represents an average of triplicate samples.

Figure 6

HMG-1 stimulates transactivation by wild-type p53 as well as p53Δ30 in transfected cells. (A,B) H1299 cells were transiently transfected with wild-type p53 (300 ng) (A) or p53Δ30 (35 ng) (B) or HMG-1 expression plasmid (2.5 μg) or both p53 and HMG-1 expression plasmids as indicated. In all cases, total DNA transfected was normalized with equivalent amounts of control parental vectors. The reporter construct used was cyclin G–luc (1 μg). (Y-axis) Luciferase activity is represented as fold transactivation relative to samples transfected with reporter constructs and parental control plasmids. The fold activation represents an average of triplicate samples.