Specificity of MYB interactions relies on motifs in ordered and disordered contexts

Abstract

Physical interactions between members of the MYB and bHLH transcription factor (TF) families regulate many important biological processes in plants. Not all reported MYB-bHLH interactions can be explained by the known binding sites in the R3 repeat of the MYB DNA-binding domain. Noteworthy, most of the sequence diversity of MYB TFs lies in their non-MYB regions, which contain orphan small subgroup-defining motifs not yet linked to molecular functions. Here, we identified the motif mediating interaction between MYB TFs from subgroup 12 and their bHLH partners. Unlike other known MYB-bHLH interactions, the motif locates to the centre of the predicted disordered non-MYB region. We characterised the core motif, which enabled accurate prediction of previously unknown bHLH-interacting MYB TFs in Arabidopsis thaliana, and we confirmed its functional importance in planta. Our results indicate a correlation between the MYB-bHLH interaction affinity and the phenotypic output controlled by the TF complex. The identification of an interaction motif outside R3 indicates that MYB-bHLH interactions must have arisen multiple times, independently and suggests many more motifs of functional relevance to be harvested from subgroup-specific studies.

Figure 1.

Interactions between MYB and bHLH TFs. (A) Typical domain structures of R2R3 MYB and bHLH TFs. The DBD (R2R3 and bHLH) and JAZ interaction domain (JID; not present in all bHLH proteins) are boxed, and larger N- and C-terminal regions of bHLH proteins are indicated to illustrate interaction regions from (B). (B) Biological functions regulated by MYB–bHLH interactions, and the interacting regions, when known. The phylogenetic subgroups (Subg.) are indicated in italics (3,4).

Figure 2.

Identification of a new bHLH interaction interface in plant R2R3 MYB TFs. (A) Split-ubiquitin assays with full-length bHLH proteins (GL3 and MYC4) as baits (underlined) against full-length or truncated MYB proteins (MYB75 and MYB29). Cultures from three independent colonies were spotted on SD-LW (positive control for culture density and viability), SD-AHLW (test for interaction) and SD-AHLW+Rapa (positive control for ability of bait and prey fusion proteins to interact (35)). (B) Schematic diagram indicating truncated versions of MYB75 that interact with GL3, as previously reported (32), and truncated versions of MYB29 that interact with MYC4. The location of the DBD is shown in grey, conserved subgroup 6 and subgroup 12 motifs are shown in green, and the RB motif mediating interaction between MYB and R/B-like bHLH TFs is shown in light grey (32). Constructs marked in black interact, whilst those displayed as a red dotted line do not interact (see also Supplementary Figures S1 and 2). (C and D) Sequence specific disorder predictions of (C) MYB29 and (D) MYB75 using DISOPRED3 (45), PONDR VSL2 (43,44) and IUPred2 (46).

Figure 3.

The MYB subgroup 12 motif is the MYC-interaction motif (MIM) and accurately predicts novel interactions. (A) Alignment of MIM-containing MYB TFs from Arabidopsis thaliana, with numbering according to MYB29. The subgroup 12 motif is enclosed by a box marked with a green line. (B–D) Split-ubiquitin assays with MYC4 or GL3 as baits (underlined) against (B) WT and mutated (mut) versions of MYB29 or MYB75, (C) MYB95 and MYB47, and (D) MYB21 and MYB32. In (B), the mutations to abolish the MIM are L190K, K192N and A194Y. The RB motif of MYB75 was mutated to the corresponding residues in the MYB29 R3 repeat (DLLLRLHRLLGNRWSLIAGR to QILLMLHRSLGNRWSVIAGH; motif residues underlined, as the motif is [DE]Lx₂[RK]x₃Lx₆Lx₃R (32).

Figure 4.

Decomposition of MIM core residues. (A) Interaction between MYC3 or MYC4 as bait against mutated versions of MYB29 (G120-L222). Results from split-ubiquitin assays are displayed as tall and short bars indicating interaction and no interaction, respectively. Results from the split-ubiquitin assays for each mutant (including combinatorial mutants) against MYC3 and MYC4 can be found in Supplementary Figures S5 and 6. (B) Sequence logo of MIM sequences in Arabidopsis thaliana MYB TFs (the eight aligned sequences in Figure 3A), numbered according to MYB29, generated by WebLogo (47).

Figure 5.

Leaf glucosinolate levels in Col-0, myb29-1, or myb29-1 transgenic plants expressing pro35S:MYB29-L190A, pro35S:MYB29-L190V or pro35S:MYB29-WT shown in nmol/mg fresh weight of short chain aliphatic (A and B), long chain aliphatic (C and D) and indole (E and F) glucosinolates. A, C, E: Data are grouped by genotype/construct. Fill colours indicate relative MYB29 transcript levels. Horizontal lines show the mean of each group. Letters above panels A, C and E indicate significant differences (P< 0.05) between group means (Tukey's Honest Significant Differences test). B, D, F: Glucosinolate levels were plotted against relative MYB29 transcript levels. Dotted lines and surrounding shaded areas show linear regressions and the 95% confidence intervals, respectively. Adjusted R² is shown in the same colour as the fit. Statistical differences between slopes were tested with an ANCOVA (Table 1 and Supplementary Table S1). Individual glucosinolate levels reported in Supplementary Table S2.

Figure 6.

The MIM peptide interacts with MYC4Nt in vitro, and the affinity is affected by mutation of L190. N-terminally biotinylated 22-residue peptides comprising the WT (A and B), L190A or L190V (Supplementary Figure S7) MIM sequences of MYB29 were immobilized on streptavidin biosensors and the binding of a dilution series of MYC4Nt (L55-N253 with C-terminal 6×His tag) was detected. (A) Referenced sensorgrams. (B) Fitting of equilibrium responses for WT-MIM to a Hill equation, including the equilibrium dissociation constant (K_D) ± standard error of the fit. (C) K_D values obtained for WT-MIM, L190A-MIM and L190V-MIM peptides. n.d.: Not determined due to lack of convergence. *Mean ± s.e.m. (of n = 2 independent rounds of protein expression). **K_D from one fit ± standard error of the fit.

Figure 7.

Specificity of MYB–bHLH interactions relies on context dependent motifs. The RB motif, which mediates interaction between e.g. MYB TFs from subgroups 4, 5, 6 and 15 and R/B-like bHLH TFs, constitute surface-exposed residues in a globular domain (32), whilst the MIM identified in this work resides within the predicted disordered non-MYB region.