Domain architecture, characterization of DRB7.2, and its interaction with DRB4.

(A) Various DRB7.2 constructs used in the study are represented as full-length, dsRBD containing region annotated as DRB7.2M (orange), and the unstructured N-terminus. The nuclear localization sequence (NLS) composed of residues Q20-L37 is marked with a yellow box. (B) Various DRB4 constructs are shown as DRB4, DRB4D1D2, DRB4C and DRB4D3. Structured domains are color coded as dsRBD1 (blue), dsRBD2 (green), and D3 (cyan). (C) 1H-15N TROSY-HSQC spectrum of perdeuterated, U-[15N] DRB7.2. For a ∼ 22 kDa complex, the spectral crowding between 7.5-8.5 ppm in the 1H dimension and overall broad resonances indicate that DRB7.2 is undergoing heterogeneous conformational exchange. (D) The 1H-15N TROSY-HSQC of DRB7.2 in the presence of equimolar DRB4. The formation of the complex shows significant improvement in the spectrum (representative red rectangles in (C and D)). (E) The isothermal calorimetric titration studies between DRB7.2 and DRB4 reveal that these two proteins form a high affinity complex. The top panel depicts the thermogram for the titration of 2 μL consecutive injections of 200 μM DRB4 into 20 μM DRB7.2. The corresponding bottom panel is the normalized binding isotherm, showing integrated changes in enthalpy (ΔH) against the molar ratio. The titration of DRB7.2 with DRB4 yielded Kd = 3.1±1.5 nM with N = 0.80.

Thermodynamic parameters derived from ITC studies.

Interaction studies of DRB7.2M and DRB4D3.

A 1H-15N TROSY-HSQC of (A) DRB7.2M, (B) DRB4D3, (C) equimolar complex of 15N DRB7.2M with unlabelled DRB4D3, and (D) equimolar complex of 15N DRB4D3 with unlabelled DRB7.2M. (E) The binding interaction between DRB7.2M and DRB4D3 titration is probed by ITC using 2 μL consecutive injections of 200 μM DRB4D3 into 20 μM DRB7.2M. The normalized binding isotherm fitted with Kd = 2.84±1.15 nM with N = 0.95 for integrated changes in enthalpy (ΔH) against the molar ratio.

Solution structure of DRB7.2M and DRB4D3.

(A) A superposition of the 10 lowest energy ensemble structures of DRB7.2M (average backbone RMSD = 0.82 Å) shows a highly conserved secondary structure in the dsRBD region. The N-terminal region (T71-S83) adopts random orientation to the DRB7.2 dsRBD. Color representation: carbon: orange; oxygen: red; nitrogen: blue. Residues A85, L89, V120, V122, L133, A150, A151, A154, and L155 form the core of DRB7.2 dsRBD. (B) The structural overlay of DRB7.2M with HsTRBPD1 (PDB ID: 5N8M) yields a backbone RMSD of 1.16 Å. While most of the dsRBD region of DRB7.2 aligns well with the TRBP dsRBD1, the orientation of the β1-β2 loop is different, and DRB7.2M has a longer β2-β3 loop. (C) Overlay of DRB7.2M with DCL3 dsRBD (1.69 Å, PDB ID: 7VG2) and DCL1 dsRBD2 (1.17Å, PDB ID: 2LRS). (D) Overlay of DRB7.2M with other structured dsRBDs involved in the plant RNAi pathway, such as DRB4D1 (0.9 Å, PDB ID: 2N3G), DRB4D2 (1.9 Å, PDB ID: 2N3H), DRB1D1 (1.0 Å, PDB ID: 2L2N) and DRB1D2 (1.7 Å, PDB ID: 2L2M). The alignment shows the overall conservation of the dsRBD fold in DRB7.2 except for the extended DRB7.2 specific β2-β3 loop. (E) Ensemble of 10 lowest energy structures of DRB4D3 calculated using Rosetta with backbone RMSD 0.25 Å. β1: V323-P327, β2: A338-R343, and β3: F347-L352, whereas, V324, Y350, I348, R326 A349, F342, E339, R351, A338, P335 form hydrogen bonds, and F347, P327, W328, P330, V324, and L352 make hydrophobic contacts. Although the solution structures of DRB7.2M and DRB4D3 represented above appear as solo domains, the data was collected on the complex with an unlabeled corresponding partner, which is invisible.

Structural statistics for the solution structure of DRB7.2M and DRB4D3.

Crystallographic data collection and refinement statistics (8IGD)

Crystal structure of DRB7.2M:DRB4D3 complex at 2.9 Å resolution.

(A) The cartoon representation of the heterodimeric complex shows the helical face of DRB7.2M. In the complex, DRB7.2M (orange) adopts a canonical α1-β1-β2-β3-α2 fold, and DRB4D3 (cyan) assumes a β1-β3-β2 topology. The interaction is primarily mediated by the β3 (DRB7.2M) and β1 (DRB4D3) parallel strands. The crystal structure is solved for the regions S83-G162 of DRB7.2M and K306-P355 of DRB4D3, which showed unambiguous electron density, whereas, residues T71-R82 of DRB7.2M and E294-K305 of DRB4D3 were not visible. (B) Residues, represented as sticks and annotated, are involved in the hydrogen bond network formation between β3 (DRB7.2M) and β1 (DRB4D3). The β-sheet formation across two domains stabilizes the complex. (C) The representation of additional stabilization interactions at the binding interface between DRB7.2M and DRB4D3, such as intermolecular/intramolecular salt bridges and cation-π interactions, (D) Weak π-π interaction, and (E) Hydrophobic interaction. Sidechains (as a stick model) of the residues involved in the specific interactions are annotated along with the corresponding bond distances (yellow dashed lines). Atoms are represented using the color coding as oxygen (red); nitrogen (blue); and sulphur (yellow).

NMR-based binding studies of DRB7.2M:DRB4D3 complex with 13 bp dsRNA.

(A-B) Excerpts of the chemical shift perturbation plots for the 150 μM DRB7.2M:DRB4D3 complex upon titration with various molar equivalence concentrations of 13 bp dsRNA are represented. Arrows indicate the direction of changes in the annotated residues. (C) The normalized chemical shift perturbation plot as a function of residue number indicates changes in the chemical shifts at 1:2 molar equivalence of DRB7.2M:DRB4D3 (150 μM) and 13 bp dsRNA (300 μM). The secondary structure is marked with arrows for β-sheet and rectangle for α-helical regions. The dotted lines represent statistical significance at 1α and 2α. (D) Electrostatic potential map of the surface-filled complex structure of DRB7.2M:DRB4D3 shows a contiguous positively charged patch on DRB7.2M on the α-helical face, whereas DRB4D3 shows random charge distribution. The surface charge map at −15 kT/e (red) to +15 kT/e (blue) was calculated using the Adaptive Poisson-Boltzmann Solver (APBS) Tool 2.1 plugin (Jurrus et al, 2018) in Pymol. (E) The surface plot of DRB7.2M:DRB4D3 complex annotated with the normalized chemical shift perturbation derived in panel C indicates that the 13 bp dsRNA induces changes in the chemical shift changes in the canonical residues arising from α1 helix, β1-β2 loop, and N-terminus of α2 helix.

Interaction studies of DRB7.2:DRB4 complex with DCL3 dsRNA substrate.

(A) Gel shift assay of 20 μM DRB7.2, DRB7.2M, DRB7.2N, and DRB4D3 with 750 nM 40 bp RNA indicating that the DRB7.2M is the sole dsRNA recognizer in DRB7.2. (B-D) Gel shift assay of 750 nM 40 bp RNA with increasing concentration of DRB7.2, DRB4, and DRB7.2:DRB4 complex. (E) Non-linear curve fit analysis to estimate the apparent Kd between 40 bp RNA with DRB7.2, DRB4, and their complex. DRB7.2:DRB4 shows improved affinity to the 40 bp DCL3 dsRNA substrate against the affinities exhibited by individual DRB7.2 and DRB4. (F) Gel shift assay of a preformed complex of DRB7.2M:40 bp RNA with increasing concentrations of DRB4D3 implying that the DRB7.2M:dsRNA complex is unperturbed by the DRB7.2M:DRB4D3 binding. The gels in (A-D and F) were visualized using SYBR-Gold nucleic acid gel stain.

Apparent binding affinities (Kd) and parameters derived from gel shift assays.

Proposed model for the dsRNA sequestration mechanism by DRB7.2 and DRB4 complex.

The functional domains, DRB7.2M and DRB4D3 of DRB7.2 and DRB4, respectively, assemble a stable and very high-affinity heterodimeric complex such that DRB7.2M is accessible to the substrate RNA without steric hindrance. The ternary complex composed of DRB7.2, DRB4, and endo-IR precursor RNA is formed by the interaction of N-terminal dsRBDs of DRB4 and the sole dsRBD of DRB7.2 that can recognize the dsRNA fold. The formation of the DRB7.2:DRB4 complex increases the availability of the dsRBD pool that is necessary to prevent the access of DCL3 to the endo-IR precursor RNA, which leads to stalling of the production of 24/22 bp siRNA duplexes. The formation of the DRB4:DRB7.2 complex is a critical step in regulating the endo-IR precursor RNA processing. The structural divergence in DRB4D3 from its homologs in humans and flies (TRBPD3/R2D2D3) is a key event that allows plants to tailor the components of the RNAi machinery for diverse gene regulatory pathways.

Diffusion tensor analysis for DRB7.2M:DRB4D3 complex based on the 15N backbone relaxation data.

Interaction of DRB7.2 with DRB4D1D2, DRB4C, and DRB4D3.

(A) ITC plot of DRB7.2 titrated with DRB4D1D2 shows no interaction, dictating the role of two N-terminal dsRBDs of DRB4 in dsRNA-specific interaction. (B) The binding isotherm of DRB7.2 with DRB4C shows an affinity of 2.8±1.6 nM, suggesting that both N-terminal DRB4 dsRBDs do not take part in the binding with DRB7.2. (C) The ITC plot of DRB7.2 with DRB4D3 showing an affinity of 2.2±1.2 nM suggests that DRB4D3 is necessary and sufficient to interact with DRB7.2.

15N backbone relaxation studies and analysis on DRB7.2M.

15N R1, R2, R2/R1, and R2*R1 were estimated unambiguously after transferring backbone chemical shift of 42 residues. The average R2/R1 corresponds to 14.56, whereas R2*R1 appears to be 23.84. The estimated τc from the data corresponds to 11.89 ns, whereas the theoretical τc for a monomeric 10 kDa globular protein should be ∼ 5-6 ns. The enhanced τc in the case of the DRB7.2M implies that DRB7.2M is probably experiencing slow to intermediate exchange between monomer:dimer and higher-order oligomeric states.

Characterization and primary sequence analysis of DRB7.2M and DRB4D3.

(A) Size exclusion chromatography (SEC) profile of DRB7.2M corresponding to the molecular weight of 10440 Da, (B) MALDI-TOF MS of DRB7.2M resulted in 12009 Da as observed MW. A small peak appearing at 24184 Da is most likely a dimeric fraction, implying DRB7.2M may remain in dynamic equilibrium of monomer:dimer in the free state. (C) Sequence alignment of DRB7.2M with homologous dsRNA binding domains. The dsRNA binding residues involved in making a tripartite contact with dsRNA are highlighted with blue dashed boxes. The extended β2-β3 loop specific to DRB7.2M is marked with red dashed boxes. (D) Size exclusion chromatography profile yields 13319 Da as MW of DRB4D3, whereas (E) corresponds to the MALDI-TOF MS profile of DRB4D3 in the free state. Multiple fragments at higher order oligomeric state in DRB4D3 suggest its dynamic equilibrium of monomer to higher order oligomers (up to tetramer or so) in the free form. (F) The sequence comparison of DRB4D3 with HsTRBPD3, DmR2D2D3, and DmLoqsPBD3 implies that the C-terminal region of DRB4 is divergently evolved from its non-plant higher eukaryotes homologs. (G) Size exclusion chromatography profile of DRB7.2M:DRB4D3 complex gives an MW of 27658 Da. (H) The corresponding MALDI-TOF MS shows the major monomeric peaks corresponding to the complex but also small populations of oligomeric states of individual proteins. The analytical SEC data was collected on 1 mg/mL protein concentration on a G75 Superdex analytical column, and the molecular weights were calculated using standards (Cytiva). MALDI-TOF analysis was performed with 0.5 mg/mL, 2 mg/mL, and 5 mg/mL concentrations. SDS PAGE for samples post-SEC is given as an inset in the corresponding panels. Secondary structural elements, α-helices and β-strands derived from the PDB structures of HsTRBPD1 (PDB ID: 5N8M) and HsTRBPD3 (PDB ID: 4WYQ) in panels C and F, respectively, are represented at the top of the alignment. Highlighted in yellow are the invariant residues, and in red are conserved by >70%.

Rosetta convergence plot and topology map for DRB4D3.

(A) Cα RMSD vs. all-atom energy in DRB4D3 (convergence at 5000 structures). The RMSD was calculated against the lowest energy structure. (B) The topology of the novel fold adopted by DRB4D3 upon binding with DRB7.2M is derived from the PDBsum1 server (Laskowski et al. 2022).

Salt bridge interaction in DRB7.2M:DRB4D3 complex.

1H-15N HSQC-TROSY spectrum of DRB4D3 showing the region of Arginine guanidino group (Nε-Hε) amide protons of the sidechain involved in salt-bridge formation through either intra or intermolecular interaction. For DRB7.2M, the low 15N frequencies (82-86 ppm) peaks in the 1H-15N HSQC-TROSY were not observed, suggesting that the salt-bridge interactions involve the amide sidechains in DRB4D3.

15N backbone relaxation studies of DRB7.2M and DRB4D3 complex.

15N-{1H} hetero-nuclear NOE, R1, R2, and R2/R1 obtained at 600 MHz are represented as a function of residue number. Secondary structural elements, α-helices (circles) and β-strands (arrows), are represented on the top. The data obtained for the N-terminal region of DRB7.2M and DRB4D3 was not used further for the model-free analysis as these regions were missing in the crystal structure.

Backbone dynamics studies for the heterodimer complex formed by DRB7.2M:DRB4D3.

The B-factors derived from the electron density map for backbone nitrogens obtained from the structure of the complex, the square of the order parameter (S2), the fast-timescale internal motion time (τe), and the dynamic model selection using Model-free formalism as described by Mandel et al. (Mandel et al., 1995) as a function of residue number. Secondary structures are annotated as filled circles and filled arrows for α-helices and β-strands, respectively. The dynamic information derived from the NMR relaxation and x-ray electron density map suggests that both interacting domains are overall rigid in the complex and do not exhibit significant internal motions at the interaction interface. However, the selection of Model2/Model5 for E108, G109, M113, H112, and K142 coupled with the inherent resonance broadening in K143 and K144 imply that the dsRNA binding site experiences dynamics at a fast to intermediate timescale (ns-ms).

Comparison of the solution and crystal structures of DRB7.2M and DRB4D3.

(A) Overlay of the lowest energy structure obtained using Xplor-NIH and crystal structure of DRB7.2M with backbone RMSD of 0.87 Å shows a similar arrangement of the dsRBD fold apart from a minute deviation in the length of β2 and β3 strands. The density for residues T71-R82 is absent in the crystal structure mainly due to the inherent flexibility of the region, as mentioned above. (B) Overlay of lowest energy structure obtained using Rosetta and X-ray structure of DRB4D3 with backbone RMSD of 0.59 Å with a similar arrangement.

A full 2D 15N-1H HSQC of DRB7.2M:DRB4D3 complex with different dsRNA.

(A) The overlay of the 1H-15N HSQC of 150 μM 15N DRB7.2M precomplexed with 165 μM unlabeled DRB4D3 (red) and upon titration with 300 μM 20 bp dsRNA (green). (B) The overlay of the 1H-15N HSQC of 150 μM 15N DRB7.2M precomplexed with 165 μM unlabeled DRB4D3 (red) with 300 μM 20bp siRNA duplex (magenta). (C) The overlay of the 1H-15N HSQC of 150 μM 15N DRB7.2M precomplexed with 165 μM unlabeled DRB4D3 (red) with 75 μM 40bp dsRNA (blue).

Gel shift assays for DRB4D3.

EMSA of DRB4D3 with (A) 80 bp RNA and (B) 40 bp RNA, where samples were separated on a 7% native PAGE. Subsequently, electrophoresis was performed at 80 V for 3 h at 4 °C, and the gel was visualized with SYBR Gold staining. The data shows that DRB4D3 does not bind to dsRNA.

Functional characterization of DRB7.2 with 80 bp blunt end dsRNA.

Gel shift assay of (A) DRB7.2, (B) DRB7.2M, and (C) DRB7.2M:DRB4D3 complex with internally labeled 80 bp dsRNA. Eighty nucleotide long sense and antisense RNA (sense: 5’ GGGUGCUGUUUCUCGUGUUCGUGUUCGUUUCUCUUCUCUUGUCCUUGUUCUGUUCUCCUUUGUUCGUUCCUGUUCCCCUU 3’, and antisense: 5’ AAGGGGAACAGGAAC GAACAAAGGAGAACAGAACAAGGACAAGAGAAGAGAAACGAACACGAACACGAGAAACAGCACCC 3’) were internally labeled with [α-32P] CTP using in vitro transcription and purified. The titration in panel A-C with increasing concentration of dsRNA shows saturation of protein at around ∼ 2 μM concentration. For the gel shift assays, recombinant proteins (∼0.1-10 μM) were incubated with [α-32P] labeled 1 nM 80 bp dsRNA for 1 h at 4 °C in buffer containing 50 mM potassium phosphate (pH 7.0), 50 mM NaCl, 50 mM Na2SO4, 2 mM DTT. Samples were separated on a 7% native PAGE. Subsequently, electrophoresis was performed at 80 V for 3 h at 4 °C, and the gel was analyzed using autoradiography. (D) Gel shift assay of a preformed complex of DRB7.2M:80 bp dsRNA with increasing concentrations of DRB4D3. (E) ITC-derived titration study of DRB7.2M:80 bp dsRNA complex with DRB4D3. The Kd corresponding to 9.43±2.04 nM with N=0.81 was obtained by fitting the isotherm to one site binding mode. The data implies that the interactions between DRB7.2M, dsRNA, and DRB4D3 occur independently, and the ternary complex formation by the three components transpires through exclusive binding events.