Electrostatic and Hydrophobic Interactions Mediate Single-Stranded DNA Recognition and Acta2 Repression by Purine-Rich Element-Binding Protein B

Abstract

Myofibroblast differentiation is characterized by an increased level of expression of cytoskeletal smooth muscle α-actin. In human and murine fibroblasts, the gene encoding smooth muscle α-actin (Acta2) is tightly regulated by a network of transcription factors that either activate or repress the 5' promoter-enhancer in response to environmental cues signaling tissue repair and remodeling. Purine-rich element-binding protein B (Purβ) suppresses the expression of Acta2 by cooperatively interacting with the sense strand of a 5' polypurine sequence containing an inverted MCAT cis element required for gene activation. In this study, we evaluated the chemical basis of nucleoprotein complex formation between the Purβ repressor and the purine-rich strand of the MCAT element in the mouse Acta2 promoter. Quantitative single-stranded DNA (ssDNA) binding assays conducted in the presence of increasing concentrations of monovalent salt or anionic detergent suggested that the assembly of a high-affinity nucleoprotein complex is driven by a combination of electrostatic and hydrophobic interactions. Consistent with the results of pH titration analysis, site-directed mutagenesis revealed several basic amino acid residues in the intermolecular (R267) and intramolecular (K82 and R159) subdomains that are essential for Purβ transcriptional repressor function in Acta2 promoter-reporter assays. In keeping with their diminished Acta2 repressor activity in fibroblasts, purified Purβ variants containing an R267A mutation exhibited reduced binding affinity for purine-rich ssDNA. Moreover, certain double and triple-point mutants were also defective in binding to the Acta2 corepressor protein, Y-box-binding protein 1. Collectively, these findings establish the repertoire of noncovalent interactions that account for the unique structural and functional properties of Purβ.

Figure 1.

Hydrophobic and electrostatic surface maps of the Mm Purβ homodimer generated by computational homology modeling. (A, B) Ribbon model of the Mm Purβ dimer highlighting sequences corresponding to PUR repeats I (violet), II (blue), and III (green). Two intramolecular subdomains are formed by association of PUR repeats I and II. The central dimerization subdomain is formed by intermolecular interaction of PUR repeat III sequences from each monomer. (C, D) Hydrophobic maps of the Mm Purβ dimer show the non-polar core of each subdomain. Yellow spheres represent hydrophobic amino acids while purple spheres represent non-hydrophobic residues. (E, F) Electrostatic surface maps of the Mm Purβ dimer show regions of charged amino acids. Blue areas indicate positively charged residues and red represents negatively charged residues. Images in B, D, and F are rotated 180° around the horizontal axis with respect to the corresponding images in A, C, and E.

Figure 2.

Effect of anionic detergent and salt on the interaction of Purβ with Acta2 ssDNA. (A) The binding of full-length Purβ (Purβ FL), its core region (Purβ I-II-III), and its intermolecular subdomain (Purβ III) to 0.5 nM PE32-bF ssDNA was assessed in assay buffer containing varying concentrations of sodium deoxycholate. Solid-phase Purβ-PE32-bF complexes were detected by ELISA using rabbit anti-mouse Purβ 210–229 as the primary antibody. (B) The same ELISA format was used to evaluate the effect of varying concentrations of salt on the binding of Purβ FL, Purβ I-II-III, and Purβ III to ssDNA. Data points show absorbance values at 405 nm normalized to the maximum absorbance observed at the lowest concentration of deoxycholate tested (A) or normalized to the absorbance obtained in buffer with 0.15 mM NaCl (B) (mean ± SD, n = 4).

Figure 3.

Effect of solution pH on the interaction of Purβ with Acta2 ssDNA. (A) The binding of full-length Purβ (1.0 nM) to PE32-bF ssDNA (0.5 nM) was assessed in assay buffer without MgCl₂ at pH ranging from 7.5 to 12.5. Solid-phase Purβ-PE32-bF complexes were detected by ELISA using rabbit anti-mouse Purβ 210–229 as the primary antibody. Nonspecific background absorbance at 405 nm in control wells without any DNA was subtracted from the signal generated in PE32-bF-coated wells. Background corrected A405 values measured at each pH were normalized by dividing by the mean A405 value determined at pH 7.5 (mean ± SD, n = 4). (B) Effect of solution pH on the thermostability of Purβ. The unfolding of full-length Purβ was evaulated by thermal shift assay at a protein concentration of 2.8 μM in 20 mM HEPES, 150 mM NaCl, 10 mM β-mercaptoethanol adjusted to pH 7.5, 10.5, 11.5, or 12.5.

Figure 4.

Evaluation of the Acta2 repressor activity of Purβ point mutants expressed in fibroblasts. (A) Subconfluent AKR-2B MEFs were transiently co-transfected with mammalian expression plasmids encoding either wild-type (WT) Purβ or the indicated point mutants and an Acta2 promoter-luciferase reporter plasmid, VSMP8. After 48 h, transfected MEFs were harvested and whole cell extracts were assayed for both luciferase enzyme activity and total protein concentration. Relative luciferase reporter expression measured in MEFs co-transfected with an empty pCI vector control was defined as 1. Bars show the fold repression of the VSMP8 reporter by each Purβ construct (mean ± SEM, n = 6–15). ****, p < 0.0001 compared to Purβ WT. (B) Immunoblot analysis was performed with a His tag antibody to confirm the expression of single, double, and triple NHis-Purβ point mutants in transfected cells. The anti-His tag blot was reprobed with a GAPDH antibody as a loading control. (C) A titration assay was performed with plasmids encoding the indicated double and triple Purβ point mutants in comparison to the wild-type protein. Symbols show the relative VSMP8 repressor activity of each Purβ construct (mean± SEM, n = 3). (D) Immunoblot analysis was performed to confirm the dose-dependent expression of NHis-Purβ in transfected cells. The anti-His tag blot was reprobed with a GAPDH antibody as a loading control. (B, D) Numbers on the left indicate molecular mass in kilodaltons.

Figure 5.

Assessment of the relative thermostability of purified Purβ point mutants. (A) The unfolding of wild-type (WT) Purβ in comparison to point mutants R267A (single), R159A/R267A (double), and K82A/R159A/R267A (triple) was evaulated by thermal shift assay at a protein concentration of 2.8 μM in 50 mM sodium phosphate pH 8.0, 300 mM NaCl, 10 mM imidazole, 10 mM β-mercaptoethanol. (B) Bars show the calculated T_m of each protein determined at multiple protein concentrations ranging from 1.0 to 8.5 μM (mean ± SD, n = 6–10).

Figure 6.

Effect of R/K point mutations on the interaction of Purβ with Acta2 ssDNA. (A and B) The binding of wild-type (WT) Purβ and the indicated single (R267A), double (R159A/R267A), or triple (K82A/R159A/R267A) point mutants to 0.5 nM PE32-bF ssDNA (filled symbols) or 0.5 nM mutated PE32–3I5T7 ssDNA (open symbols) was evaluated by ELISA. Solid-phase Purβ-ssDNA complexes were detected using rabbit anti-mouse Purβ 210–229 as the primary antibody. (A) Protein concentration ranges were chosen to achieve saturable binding to the PE32-bF probe. Binding isotherms were generated by fitting data points obtained from multiple, independent titration experiments (n = 3–5) to the equation for a rectangular hyperbola. (B) Replot of the same datasets fit to a log(agonist) vs. response (four parameters) equation. The apparent K_d (A) or EC50 (B) of each Purβ protein tested for the PE32-bF probe was extrapolated accordingly.

Figure 7.

Effect of R/K point mutations on the interaction of Purβ with MSY1. (A) The binding of purified Purβ wild-type (WT) protein or the indicated single (R267A), double (R159A/R267A), or triple (K82A/R159A/R267A) point mutants to immobilized MSY1 (filled symbols, solid lines) or BSA (open symbols, dashed lines) was evaluated by ELISA. Purβ-MSY1 complexes were detected using rabbit anti-Purβ 210–229 as the primary antibody. Binding isotherms were generated by fitting data points obtained from several independent titration experiments (n = 3) to the equation for a rectangular hyperbola. (B) Binding curves generated after subtracting out the nonspecific absorbance measured in BSA only-coated wells from the absorbance obtained in MSY1-coated wells.