NMR-Guided Studies to Establish the Binding Interaction between a Peptoid and Protein

Abstract

Ligand discovery of nonenzymatic proteins can be accomplished through screening methods utilizing libraries comprising small molecules, peptides, and peptidomimetics. Incorporating peptoids, which are oligomers of N-substituted glycine monomers, into high-throughput screens can produce libraries of large structural diversity. Due to their malleable structures, peptoids can occupy unique protein binding sites, but determination of the peptoid binding pose is challenging. For example, the peptoid KDT-11 is reported to bind with low micromolar binding affinity to the proteasome subunit Rpn-13. Poor solubility of initial compound screening hits, like KDT-11, can greatly hinder progress in drug discovery since it limits in vitro characterization. The work reported here overcomes this hurdle with the addition of a solubility tag to KDT11, enabling elucidation of the biologically relevant surface of the peptoid through a variety of structure-activity relationships and biophysical studies. NMR paramagnetic relaxation data guided a structural modeling protocol using multiple molecular dynamics (MD) trajectories and extensive sampling. The final peptoid-protein structure is conformationally stable in equilibrium MD trajectories for >1 μs time period. KDT-11 binds across the β6/β7/β8 strands and α-helix of Rpn-13, revealing an interface for inhibition that could be targeted in future computational drug discovery efforts to obtain more potent ligands for Rpn-13. It is reasonable that the methodology described here can extend to other flexible peptoid or peptide ligands in complexes with proteins.

Figure 1.

Biophysical structure-activity relationship studies to characterize peptoid-protein binding. (A) KDT-11’s 6-mer peptoid structure and representative dot structures of the sarcosine (Sar) and truncated peptoid series. These molecules were synthesized with or without a fluorescein tag to identify the essential properties of KDT-11. (B - C) Binding affinity analyses of fluorescein-tagged KDT-11 analogs were performed through fluorescence polarization (FP) assay. 10 nM of fluorescein-tagged analog was incubated with increasing concentrations of purified Rpn-13, and the polarization signal was measured by a Biotek Synergy Neo2 Multimode Microplate Reader (excitation wavelength: 485 nm; emission wavelength: 520 nm). Each fluorescein-tagged analog was tested in triplicate. (D) Circular dichroism (CD) to assess secondary structure of non-fluorescent KDT-11 analogs. After synthesizing non-fluorescent versions of the KDT-11 analogs, all analogs were dissolved in methanol and the CD spectrum measured. (E) Structures of KDT-11 analogs designed to examine the role of secondary structure for Rpn-13 binding. The analogs are two peptidomimetics that incorporated a S chiral center in P4 or P5. (F) Secondary structure analysis of KDT-11 analogs 4-L-Cha and 5-L-Homophe by CD, with unmodified KDT-11 as the control. (G) Binding analysis of fluorescein-tagged KDT-11 analogs 4-L-Cha and 5-L-Homophe, with unmodified FL-KDT-11 as the control. Samples were prepared and analyzed similarly as (B - C).

Figure 2.

Biophysical analysis of Rpn-13 PRU with TCM-2, a soluble analog of KDT-11. (A) Binding affinity analysis to Rpn-13 domains with fluorescein-tagged KDT-11 was performed through FP assay. 10 nM of fluorescein-tagged analog was incubated with increasing concentrations of purified PRU, or DEUBAD domain, with full-length Rpn-13 as the control. Polarization signal was measured by a Biotek Synergy Neo2 Multimode Microplate Reader (excitation wavelength: 485 nm; emission wavelength: 520 nm). Each protein construct was tested in triplicate, and the data were analyzed utilizing binding saturation, one site, specific binding with Hill slope. FL-KDT-11 bound to the PRU domain with a K_d of 2.5 ± 0.6 μM. No binding to the DEUBAD domain was observed (B) Structure of tripiperazine-linked (teal) KDT-11, referred to as TCM-2. (C) FP binding assay of fluorescent TCM-2 (green) titrated with Rpn-13, compared to FL-KDT-11 (black). The estimated affinity is 7.4 ± 0.6 μM, which is suitable for NMR structural studies. (D) CD spectrum of TCM-2 in methanol (dotted green line), is similar to that of KDT-11 (dotted black line). In water, the TCM-2 (solid green line). maximum at 208 nm is stronger than the 220 nm absorbance, suggesting dominance of the trans-amide bond conformation. (E) Overlaid ¹H-¹⁵N HSQC spectra of ¹⁵N-labeled PRU (200 μM) in the absence (black) and presence (green) of TCM-2 (300 μM), with zoomed regions showing changes in chemical-shifts. (F) PRU domain residue profile of non-saturating chemical shift perturbations (CSP_NS). Residues that exhibit large CSP_NS are labeled. Upon TCM-2 addition, distinct peaks (purple) arise for amino acids G35, T36, C88, and G91.

Figure 3.

NMR analysis of Rpn-13 PRU with TCM-3, a spin-labeled analog of KDT-11/TCM-2. (A) Chemical structure of TCM-3, the spin-labeled derivative of TCM-2. (B) Peak volume ratio of oxidized and reduced sample at T₀ were plotted against residue number. G91, M109, and Q110 have a relatively small volume ratio (<0.8), indicating they were proximal to the spin-label on TCM-3. (C) Paramagnetic relaxation rate from estimates of transverse relaxation of oxidized and reduced TCM-3 and ¹⁵N-labeled PRU. PRE values exceeding10 s⁻¹ indicate strong paramagnetic effects. Severely attenuated peaks that could not be quantified are indicated as pink bars. (D) Integrated analysis of chemical shift perturbation and PRE data for TCM-2 and TCM-3 reveals a localized binding site in the cleft between the α-helix and β7/β8 strands. PDB ID 2KR0.

Figure 4.

Structure model of the KDT-11/PRU complex. (A) Flowchart of the three MD stages needed to model the PRU/KDT-11 complex. (B) Assessment of the final complex from Stage 3 with PRE-restrained MD, followed by unrestrained, equilibrium MD, showing collective agreement with all PRE distances (left panel) and structural stability over a 1-μs period based on intermolecular energy (middle panel), and the KDT-11 position on the PRU surface (right panel). (C-D) In the final complex, KDT-11 is located on the surface of PRU in the orientation shown with P4 in close contact with the PRU residues exhibiting NMR PREs and CSPs (purple and cyan spheres, respectively), see Supporting Figure S5. Substantial reorientation of KDT-11 sidechains on the PRU surface occurs during the three stages of NMR-guided simulation. (E) Specific polar interactions between the P3 chloro-atom of KDT-11 and PRU-domain sidechains of Lys-83 and Lys-97.