Hierarchy in regulator interactions with distant transcriptional activation domains empowers rheostatic regulation

Abstract

Transcription factors carry long intrinsically disordered regions often containing multiple activation domains. Despite numerous recent high-throughput identifications and characterizations of activation domains, the interplay between sequence motifs, activation domains, and regulator binding in intrinsically disordered transcription factor regions remains unresolved. Here, we map sequence motifs and activation domains in an Arabidopsis thaliana NAC transcription factor clade, revealing that although sequence motifs and activation domains often coincide, no systematic overlap exists. Biophysical analyses using NMR spectroscopy show that the long intrinsically disordered region of senescence-associated transcription factor ANAC046 is devoid of residual structure. We identify two activation domain/sequence motif regions, one at each end that both bind a panel of six positive and negative regulator domains from biologically relevant regulators promiscuously. Binding affinities measured using isothermal titration calorimetry reveal a hierarchy for regulator binding of the two ANAC046 activation domain/sequence motif regions defining these as regulatory hotspots. Despite extensive dynamic intramolecular contacts along the disordered chain revealed using paramagnetic relaxation enhancement experiments and simulations, the regions remain uncoupled in binding. Together, the results imply rheostatic regulation by ANAC046 through concentration-dependent regulator competition, a mechanism likely mirrored in other transcription factors with distantly located activation domains.

Keywords: activation domains; intrinsically disordered region; isothermal titration calorimetry; nuclear magnetic resonance; protein–protein interactions; short linear motif; transcription factor.

FIGURE 1

Identification of SLiMs and ADs in the ANAC II‐3 subgroup. (a) Thirteen NAC transcription factors from Arabidopsis ordered by the phylogenetic relationship reported in Jensen et al. (2010). For each, the IDR is shown with SLiM mapping; M1 (yellow), M2 (purple), M3 (pink), and M4 (green). Scores from AD screens are shown as a line with dots at the tile center (40 residue length) (Morffy et al., 2024). AD threshold (dashed line) was defined as one standard deviation from the mean AD score of the full library in Morffy et al. (2024). Disorder is represented by normalized pLDDT scores in gray (<0.5 regarded as disordered). Below the names of each transcription factor, the transcriptional activity is indicated as minimally active (+), activator (++), and strong activator (+++) according to Hummel et al. (2023). (b) Logo plots of the four identified SLiMs generated based on the SLiM instances in (a) and colored as in (a). (c) Alignment of the last 21 residues from the transcription factors in (a) based on a single tryptophan (or phenylalanine) in the sequences. Aromatic or leucine residues are gray and acidic residues pink. The isoelectric points of the sequences are shown to the right. The sequence boxed in blue is the known SLiM for RCD1‐RST (RIM, M5) binding shown below in light blue. (d) Tiles used in AD scoring of the ANAC046 IDR (Morffy et al., 2024). AD score threshold shown as dotted line and defined as one standard deviation from the mean AD score of the full library. Tiles I–IV are shown below for identification of the ADs (gray lettering), with the positions of M1–M4 highlighted.

FIGURE 2

The ANAC046 IDR is disordered without residual structure. (a) Domain structure of ANAC046 with fragments used in this work and with an AlphaFold 2 structure in an extended version illustrated above. The fragment noted * is from previous work (O'Shea et al., 2015). (b) C ^α SCS calculated from Newcombe et al. (2021) (top), R ₂ relaxation rates including values for a 167‐residue random coil protein shown as a line (Wirmer et al., 2006) (middle) and hetNOEs (bottom) of ANAC046_172–338. (c) IDDomainspotter (Millard et al., 2020) analysis of ANAC046_172–338 with a sliding window of 15 residues. (d) ParSe (Wilson et al., 2023) prediction of ANAC046_172–338 with Identified regions of 20 or more contiguous residues at least 90% PS prone. For (a)–(d), the coloring of the IDR is defined in Figure 1. (e) Double log plot of the experimental SAXS scattering curve of ANAC046_172–338 (light gray) and calculated SAXS scattering curves from the ensemble generated using CALVADOS 2 (Tesei & Lindorff‐Larsen, 2023) before (χ² = 1.32) (dark purple) and after (χ² = 1.00) (light purple) reweighting. Residuals are shown below. The scaling exponent (ν) is shown in the plot. (f) 50 models sampled randomly from the simulation ensemble with weights from reweighting with one model highlighted.

FIGURE 3

Regulator binding to ANAC046 IDR is promiscuous at specific sites. (a) Overview of proteins used for analyses with colors used in (b) and (c). (b) ¹⁵N‐HSQCs of ¹⁵N‐ANAC046_172–338 with and without (black) added partner. Residues that disappear upon regulator addition are annotated. (c) Zooms on S189 and S261 (left), and M315 (right) in ANAC046_172–338 in the absence (black) and presence of regulators (for colors see A). (d) and (e) NMR peak intensity ratios (I/I ₀) of ¹⁵N‐ANAC046_172–338 determined from titration with RCD1‐RST (h) and HAC1‐TAZ1 (i), respectively. Titration points vary from no addition of regulator to four molar excesses. The SLiMs and ADs identified in the ANAC046 IDR are highlighted. Gaps in the plot show unassigned residues, and residues in the area marked with * were not observed. (f)–(i) AlphaFold 3 models of the two ADs in complex with RCD1‐RST and HAC1‐TAZ1. The highest‐ranking peptide model is shown in purple with other models transparent.

FIGURE 4

Thermodynamics and intramolecular contacts in ANAC046 IDR. (a) Bar plots of the thermodynamic parameters from ITC with different ANAC046 fragments and regulator domains. (b) The three top panels show PRE‐NMR with positions labeled by MTSL shown as orange stars at S217C, S259C, and C323. The two lower panels show PRE‐NMR with ANAC046‐C323 MTSL labeled and added RCD1‐RST(1:1 stoichiometric ratio), and the difference PRE between ANAC046‐C323‐MTSL and ANAC046‐C323‐MTSL:RCD1‐RST (wo‐w(RCD1‐RST)). For each sample, 90 μM ¹⁵N‐ANAC046_172–338 was used and best‐TROSY HSQCs recorded in the paramagnetic and diamagnetic states (+ascorbic acid) of the label. The intensity ratios (I _Para/I _Dia) are plotted for each residue. *unassigned residues. Gaps correspond to ambiguous assignments due to peak overlap. For the top three panels, black dots represent PREs extracted from the simulated ensemble reweighted with SAXS. (c) Normalized pair distance distribution function of the experimental SAXS data of ANAC046‐IDR alone (ANAC046) and at 1:1 molar ratio with RCD1‐RST (ANAC046 + RCD1‐RST), a random coil ensemble of ANAC046‐IDR generated by RANCH (Bernadó et al., 2007) (Random coil) and the simulated ensemble of ANAC046 IDR reweighted with SAXS (Simulation).

FIGURE 5

Rheostatic regulator binding model of ANAC046. (a) Perturbation in ensemble dynamics of the ANAC046 IDR can lead to the formation of large clusters, supramolecular complexes, or phase separation, also induced by regulator interactions. (b) (Left) For four different regulator domain concentrations and an ANAC046 concentration of 50 μM, site saturation was calculated (Wang, 1995) using the K _d values for AD1/M1‐2 (ANAC046_172‐222) and AD2/M5 (ANAC046_300‐338) in the full‐length ANAC046 IDR. (Right) The overall saturation of the two AD/SLiMs at different regulator concentrations (output) depends on differences in affinities with a preference for AD2/M5. Depending on the relative levels of positive and negative regulators, we suggest a model with a graded regulator response depending on concentrations and affinities, illustrated as a dial.