Differential binding of quadruplex structures of muscle-specific genes regulatory sequences by MyoD, MRF4 and myogenin

Abstract

Four myogenic regulatory factors (MRFs); MyoD, Myf-5, MRF4 and Myogenin direct muscle tissue differentiation. Heterodimers of MRFs with E-proteins activate muscle-specific gene expression by binding to E-box motifs d(CANNTG) in their promoters or enhancers. We showed previously that in contrast to the favored binding of E-box by MyoD-E47 heterodimers, homodimeric MyoD associated preferentially with quadruplex structures of regulatory sequences of muscle-specific genes. To inquire whether other MRFs shared the DNA binding preferences of MyoD, the DNA affinities of hetero- and homo-dimeric MyoD, MRF4 and Myogenin were compared. Similarly to MyoD, heterodimers with E47 of MRF4 or Myogenin bound E-box more tightly than quadruplex DNA. However, unlike homodimeric MyoD or MRF4, Myogenin homodimers associated weakly and nonpreferentially with quadruplex DNA. By reciprocally switching basic regions between MyoD and Myogenin we demonstrated dominance of MyoD in determining the quadruplex DNA-binding affinity. Thus, Myogenin with an implanted MyoD basic region bound quadruplex DNA nearly as tightly as MyoD. However, a grafted Myogenin basic region did not diminish the high affinity of homodimeric MyoD for quadruplex DNA. We speculate that the dissimilar interaction of MyoD and Myogenin with tetrahelical domains in muscle gene promoters may differently regulate their myogenic activities.

Figure 1.

Homodimeric Myogenin and MRF4 bind preferentially bimolecular quadruplex structures of integrin DNA. Increasing amounts of homodimers of recombinant Myogenin or MRF4 were incubated under binding conditions with, respectively, 0.18 or 1.0 pmol of 5′-³²P quadruplex integrin DNA. Protein-bound G′2 DNA was resolved from free DNA by nondenaturing 4% polyacrylamide electrophoresis. (A) Binding of Myogenin to G′2 integrin DNA. Shown are the retarded protein–G′2 integrin DNA complex and the G′2 and G′4 forms of free DNA. (B) Binding of MRF4 to G′2 integrin DNA.

Figure 2.

Homodimeric Myogenin raises the heat stability of bound G′2 integrin DNA. (A) Heat denaturation of Myogenin-bound and free G′2 integrin DNA. Reaction mixtures that contained each 0.18 pmol of 5′-³²P labeled G′2 integrin DNA were incubated under binding conditions with or without saturating amounts of recombinant homodimeric Myogenin. Following completion of the binding reaction, the mixtures were heated at the indicated temperatures for 10 min. Heat denaturation was terminated by rapid cooling of the mixtures to 4°C and the addition of 0.5% SDS to strip the protein off the DNA. Residual G′2 integrin DNA was resolved from G′4 DNA by nondenaturing 10% polyacrylamide gel electrophoresis and their relative amounts were quantified by phosphor imaging analysis. Shown is a plot of the remaining amount of G′2 DNA in mixtures that contained or were devoid of Myogenin as a function of the increasing temperature. (B) Kinetics of heat denaturation of Myogenin-bound or free G′2 integrin DNA. Binding conditions, heat denaturation and electrophoretic resolution of remaining G′2 DNA were as described in (A) except that all the mixtures were heated at 57°C for the indicated increasing periods of time. Shown is a plot of the residual amount of G′2 DNA in mixtures that contained or were devoid of Myogenin as a function of increasing time at 57°C.

Figure 3.

Homodimers of MRF4, but not of Myogenin, bind bimolecular quadruplex integrin DNA more tightly than E-box. Constant amounts of homodimers of recombinant MRF4 or Myogenin were incubated under DNA-binding conditions with increasing amounts of either 5′-³²P labeled E-box or G'2 integrin DNA. Formed protein–DNA complexes were resolved from free DNA by nondenaturing polyacrylamide gel electrophoresis and their relative proportions were determined by phosphor imaging analysis (see Materials and methods section). Shown are representative electrophoregrams (insets) and Scatchard plots of the measured ratios of bound to free DNA as a function of the concentration of protein-bound DNA. The indicated K_d values in different plots were derived from the negative reciprocal of their slopes.

Figure 4.

Heterodimers of MRF4 and Myogenin with E47 bind E-box more tightly than G′2 quadruplex integrin DNA. Constant amounts of MRF4-E47 or Myogenin-E47 heterodimers that were prepared as described under Materials and methods section, were incubated under DNA-binding conditions with increasing amounts of either 5′-³²P labeled E-box or G′2 integrin DNA. Electrophoretic resolution and phosphorimage analysis quantification of protein-bound and free DNA were conducted as detailed in the legend to Figure 5. Shown are representative electrophoregrams (insets) and Scatchard plots of the ratios of bound to free DNA as a function of the concentration of protein-bound DNA. The K_d values that were calculated for each plot were derived from the negative reciprocal of their respective slopes.

Figure 5.

Mutated Myogenin with clusters of basic amino acids identical to those of MyoD and MRF4 maintains low relative affinity for G′2 integrin DNA. To render the basic amino acids clusters in the Myogenin basic region identical to those of MyoD and MRF4, two mutations R84K and K91R were introduced into full-length Myogenin cDNA in pGEX-6P vector as detailed under Materials and methods section. G′2 integrin DNA binding by unmodified and mutated Myogenin (mut-Myogenin) recombinant proteins and K_d values of the respective complexes were measured (see Material and methods section). (A) Basic regions of MyoD, MRF4 and Myogenin. The three conserved clusters R₁, R₂ and R₃ of three basic amino acids each are highlighted. The two amino acids, R84 and K91 that distinguish Myogenin from MyoD and MRF4 are circled in red. (B) Binding of quadruplex integrin DNA by Myogenin and mut-Myogenin. The indicated increasing amounts of the respective proteins were incubated under DNA-binding conditions with 0.18 pmol of 5′-³²P labeled G′2 integrin DNA and the formed protein–DNA complexes were resolved by electrophoresis and quantified as detailed under Materials and methods section. Inset: average K_d values ± SD of complexes of G′2 integrin DNA with unmodified or mutated Myogenin that were determined in [N] independent measurements.

Figure 6.

The MyoD basic region and its peripheral domains dominantly dictate high affinity for G′2 integrin DNA. MyoD or Myogenin cDNA in pGEX-6P vectors were modified and restricted to remove their basic regions, which were then replaced by synthetic DNA duplexes that reciprocally encoded the respective Myogenin or MyoD basic regions (see Materials and methods section). Unmodified or proteins with exchanged basic regions were expressed and their capacities to bind 5′-³²P labeled G′2 integrin DNA were determined. (A) Schemes of unmodified and basic region-switched chimeras of MyoD and Myogenin. The tract transplanted into MyoD at positions 102–121 corresponded to the amino acids sequence of the Myogenin basic region (b_mg), whereas the flanking inserted tracts (residues 99–101 and 122–126) maintained the respective MyoD native sequences. The amino acid run transplanted into Myogenin (residues 74–93) represented the MyoD basic region (b_MD), whereas the flanking inserted tracts (residues 71–73 and 94–98) retained the respective Myogenin native sequences. (B) Binding of G′2 quadruplex integrin DNA by MyoD, Myogenin and Myogenin/MyoD_b chimerical protein. Increasing amounts of the respective proteins were incubated under DNA-binding conditions with 0.18 pmol of 5′-³²P labeled G′2 integrin DNA and the protein–DNA complexes that were formed were resolved by electrophoresis and quantified as detailed under Materials and methods section. The presented results are averages ± SD of 3–5 independent measurements. (C) Binding of G′2 quadruplex integrin DNA by MyoD, Myogenin and MyoD/Myogenin_b chimerical protein. Binding of G′2 integrin DNA by the MyoD/Myogenin_b protein was assayed as described in (B) above. The results represent averages of four independent determinations ± SD. The binding curves for unmodified MyoD and Myogenin are the same as in (B).