Quantification of reaction cycle parameters for an essential molecular switch in an auxin-responsive transcription circuit in rice

Abstract

Protein-based molecular switches play critical roles in biological processes. The importance of the prolyl cis-trans switch is underscored by the ubiquitous presence of peptidyl prolyl isomerases such as cyclophilins that accelerate the intrinsically slow isomerization rate. In rice, a tryptophan-proline (W-P) cis-trans switch in transcription repressor protein OsIAA11 along with its associated cyclophilin LRT2 are essential components in a negative feedback gene regulation circuit that controls lateral root initiation in response to the plant hormone auxin. Importantly, no quantitative characterizations of the individual (microscopic) thermodynamic and kinetic parameters for any cyclophilin-catalyzed W-P isomerization have been reported. Here we present NMR studies that determine and independently validate these parameters for LRT2 catalysis of the W-P motif in OsIAA11, providing predictive power for understanding the role of this switch in the auxin-responsive circuit and the resulting lateral rootless phenotype in rice. We show that the observed isomerization rate is linearly dependent on LRT2 concentration but is independent of OsIAA11 concentration over a wide range, and LRT2 is optimally tuned to maintain OsIAA11 at its cis-trans equilibrium to supply the slower downstream cis-specific proteasomal degradation with maximal OsIAA11 substrate. This indicates that accelerating the LRT2-catalyzed isomerization would not accelerate OsIAA degradation, whereas decreasing this rate via targeted mutation could reveal relationships between circuit dynamics and lateral root development. Moreover, we show that sequences flanking the highly conserved Aux/IAA W-P motif do not impact LRT2 catalysis, suggesting that the parameters determined here are broadly applicable across highly conserved cyclophilins and their Aux/IAA targets.

Keywords: LRT2; OsIAA11; W-P prolyl isomerization; auxin circuit; degron motif.

Fig. 1.

Determination of four-state model parameters by lineshape analysis. (A) The four-state model for the LRT2/OsIAA11 reaction cycle and its associated microscopic parameters. (B) Expanded spectral regions of overlaid ¹⁵N-¹H-HSQC spectra of 0.26 mM ¹⁵N-labeled OsIAA11^72–125 with varying amounts of unlabeled LRT2 [0 mM (red) to 800 µM (pink)] showing resolved G103HN and W104-H_N^ε peaks. (C) The ¹H peak simulations using the four-state model fitted microscopic parameters (SI Appendix, Table S1) and TITAN (33).

Fig. 2.

Independent validation of model and fitted parameters. (A) Simplified schematic of LRT2/OsIAA11 reaction showing the observed exchange rate, ${\mathit{k}}_{\mathit{ex}}^{\mathit{obs}}$, measured using Nzz or ROESY experiments. (B) Region of ¹⁵N-OsIAA11^72–125 Nzz spectrum showing W104-NHε cis and trans peaks at mixing time t_mix = 0.55 s, without LRT2 (Left) and with 16 µM LRT2 (Right). (C and D) Intensities of W104-NHε Nzz (C) autopeaks and (D) cross-peaks for 0.8 mM ¹⁵N-OsIAA11^72–125 in the presence of 16 µM LRT2 as function of t_mix, fitted to obtain ${\mathit{k}}_{\mathit{ex}}^{\mathit{obs}}$. (E) Residues selected to determine ${\mathit{K}}_{\mathbf{d}}^{\mathit{App}}$ (magenta) mapped onto the LRT2 homology model. (F) Binding curve showing the average of the normalized chemical shift change $\left(\overline{\mathbf{\Delta }\mathit{\omega }}\right)$ induced by titration with unlabeled OsIAA11^72–125 over the selected 20 residues (dots) and the corresponding fitted curve (dashed line, ${\mathit{K}}_{\mathbf{d}}^{\mathit{App}}$ = 1.54 mM). Error bars denote the SD of Δω over the 20 residues at each titration point.

Fig. 3.

Flanking X-P peptide bonds do not affect the isomerization of LRT2. (A) Sequences of the core OsIAA11 peptide, of OsIAA11 WT, and of OsIAA11 mutants designed to determine whether flanking X-P motifs (blue boxes) affect isomerization of the ¹⁰⁴W-P¹⁰⁵ peptide bond (magenta box). KR (green box) is important for degradation rates of Aux/IAA proteins. (B) Expanded region of Nzz spectra of ¹⁵N-OsIAA11^72–125 mutants in the presence of 16 µM LRT2 showing the W104-NHε peaks at mixing time of 0.55 s, demonstrating LRT2 catalysis of all four mutants.

Fig. 4.

Binding and isomerization of OsIAA11^78–109 by LRT2. (A) Binding curve showing $\overline{\mathbf{\Delta }\mathit{\omega }}$ induced by titration with unlabeled OsIAA11^78–109 synthetic peptide for seven selected residues in ¹⁵N-(6His)-LRT2 (dots) and the corresponding fitted curve (dashed line, ${\mathit{K}}_{\mathbf{d}}^{\mathit{App}}$ = 1.3 ± 0.4 mM). Error bars denote the SD in Δω over the seven residues at each titration point. The pH of this sample was 6.5. (B and C) Intensities of W104-NHε ROESY (B) autopeaks and (C) cross-peaks for 1.84 mM OsIAA11^78–109 in the presence of 120µM LRT2 as function of t_mix, fitted to obtain ${\mathit{k}}_{\mathit{ex}}^{\mathit{obs}}$ for these sample conditions.

Fig. 5.

Exchange rate as function of free OsIAA11_cis and LRT2. The microscopic thermodynamic and kinetic parameters determined by NMR lineshape analysis predict an exchange rate linearly dependent on the LRT2 concentration, and invariant to OsIAA11 concentrations across a broad range (<1 mM).