Synergy of aromatic residues and phosphoserines within the intrinsically disordered DNA-binding inhibitory elements of the Ets-1 transcription factor

Abstract

The E26 transformation-specific (Ets-1) transcription factor is autoinhibited by a conformationally disordered serine-rich region (SRR) that transiently interacts with its DNA-binding ETS domain. In response to calcium signaling, autoinhibition is reinforced by calmodulin-dependent kinase II phosphorylation of serines within the SRR. Using mutagenesis and quantitative DNA-binding measurements, we demonstrate that phosphorylation-enhanced autoinhibition requires the presence of phenylalanine or tyrosine (ϕ) residues adjacent to the SRR phosphoacceptor serines. The introduction of additional phosphorylated Ser-ϕ-Asp, but not Ser-Ala-Asp, repeats within the SRR dramatically reinforces autoinhibition. NMR spectroscopic studies of phosphorylated and mutated SRR variants, both within their native context and as separate trans-acting peptides, confirmed that the aromatic residues and phosphoserines contribute to the formation of a dynamic complex with the ETS domain. Complementary NMR studies also identified the SRR-interacting surface of the ETS domain, which encompasses its positively charged DNA-recognition interface and an adjacent region of neutral polar and nonpolar residues. Collectively, these studies highlight the role of aromatic residues and their synergy with phosphoserines in an intrinsically disordered regulatory sequence that integrates cellular signaling and gene expression.

Keywords: fuzz complex; intrinsically disordered region; protein dynamics; transcription factor regulation.

Fig. 1.

Intramolecular interactions of the SRR with the ETS domain are dependent on aromatic residues and phosphoserines. (A) Ets-1 schematic: PNT domain (cyan), SRR (yellow), IM (purple), ETS domain (red). ΔN279 with SRR residues, CaMKII phosphoacceptors S282 and S285 (*), and four mutated aromatic residues (4ϕA; red, orange). (B) CSPs for main-chain amides and side-chain amides/indoles (Inset) in ΔN279 (orange bars) and ΔN279^2P (black lines) on 4ϕA substitutions. See Fig. S2A for other comparisons. (C) CSPs for residues 301–440 of ΔN279^2P on 4ϕA mutation mapped on ΔN301 [Protein Data Bank (PDB) ID 1R36] with perturbed side chains in stick format (red: Δδ > 0.125 ppm, yellow: 0.125 ppm > Δδ > 0.08 ppm, gray: Δδ < 0.08 ppm or prolines). (D) Surface representation of ΔN301 showing the positively charged (blue; Arg, Lys), negatively charged (red; Asp, Glu), hydrophobic (green), and neutral polar (white) residues.

Fig. 2.

Intermolecular interactions of the trans-SRR peptides with the ETS domain are dependent on adjacent aromatic residues and phosphoserines. (A) SRR peptides (residues 279–300) and ΔN301. (B) SRR peptide affinities were obtained by fitting the NMR chemical shift changes (data points) of perturbed ΔN301 nuclei. See Table 3 for K_D values and Fig. S5A for full spectra.

Fig. 3.

Characterizing the trans-SRR peptide structure, dynamics and ΔN301-binding interface. (A) Overlaid ¹⁵N-HSQC and ¹³C-HSQC spectra of ¹³C/¹⁵N-labeled SRR^2P on progressive addition of unlabeled ΔN301 in the indicated protein:peptide molar ratios. pS285* is from a population of peptide with a cis Val280-Pro281 bond. See Fig. S6 for full spectra. (B) Amide CSPs observed upon addition of ΔN301 to the labeled SRR^0P (∼25% saturation; blue bars) and SRR^2P (∼85% saturation; red line). Peptide interface (gray), aromatic residues (ϕ), phosphoacceptors (*). (C) Propensities of the SRR^0P (green), SRR^2P (black), and ΔN301-bound SRR^2P (∼85% saturation; red) to adopt random coil conformations, as calculated with δ2D (13) (trans X-Pro conformers shown, cis X-Pro yielded similar results; Fig. S8A). (D) ¹H{¹⁵N}-NOE ratios for SRR^0P (green), SRR^2P (black), and SRR-4ϕA^0P (orange). Decreasing values indicate increasing flexibility on the subnanosecond timescale. Missing data are prolines or residues with overlapped signals.

Fig. 4.

Trans-SRR peptides interact with the ETS domain DNA-recognition interface of ΔN301. Small regions of the superimposed ¹⁵N-HSQC spectra of ¹⁵N-labeled ΔN301 (∼250 μM) in the absence (cyan contours) or presence of SRR-4ϕA^0P (1.1-mM peptide yielding <37% saturation based on the K_D values of Table 3; orange), SRR-4ϕA^2P (1.1 mM, 78%; purple), SRR^0P (1.0 mM, 37%; blue), or SRR^2P (640 μM, 84%; red). Residues with large (>0.1 ppm; red) and medium (0.07–0.1 ppm; yellow) amide or indole CSPs on titration with SRR^2P are mapped on ΔN301 (PDB ID 1R36). Perturbed side chains in stick format. Other peptides were similar (Fig. S5 B and C). Residues (Left) exhibited a pattern of increasing shift perturbations in the order SRR-4ϕA^0P < SRR-4ϕA^2P < SRR^0P < SRR^2P, suggestive of similar structural perturbations. Residues (Right) showed peptide-specific changes.