Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2025 Jun 23;65(13):6939–6948. doi: 10.1021/acs.jcim.5c00102

Construction of PROTAC-Mediated Ternary Complex Structure Distribution Profiles Using Extensive Conformational Search

Genki Kudo †, Takumi Hirao ‡,§, Ryuhei Harada ∥, Yasuteru Shigeta ∥, Takatsugu Hirokawa §,⊥, Ryunosuke Yoshino §,⊥,*
PMCID: PMC12264966  PMID: 40550492

Abstract

Proteolysis-targeting chimeras (PROTACs) are heterobifunctional small molecules that recruit E3 ubiquitin ligases to a target protein and induce its ubiquitination by forming a ternary complex. However, the structural dynamics underlying the complex formation and degradation efficiency remain unclear. In this study, we attempted an extensive conformational search using the parallel cascade selection molecular dynamics (PaCS-MD) and outlier flooding (OFLOOD) method for PROTACs differing in linker length. Markov state models revealed that while all PROTACs share a common low free-energy state, their structural distribution profiles differ significantly. These results suggest that linker-dependent conformational distribution profiles modulate degradation activity and cooperativity, offering mechanistic insights into rational PROTAC design.

graphic file with name ci5c00102_0010.jpg

graphic file with name ci5c00102_0008.jpg

Introduction

More than 90% of global sales in the current drug market involve conventional small-molecule drugs. Most small-molecule drugs have a high affinity for their targets and function by occupying hydrophobic pockets, limiting the application of conventional modalities for proteins with nonenzymatic functions, targets without hydrophobic pockets, and proteins with high affinity for their substrates. , Proteolysis-targeting chimeras (PROTACs) are a promising new class of molecules that function via “event-driven pharmacology,” unlike the “occupancy-driven pharmacology” of conventional small molecules. − PROTACs, which are heterobifunctional small molecules, recruit E3 ubiquitin ligases to the target protein and induce its ubiquitination by forming a ternary complex. Polyubiquitinated target proteins are degraded by the ubiquitin-proteasome system. Target protein degradation via PROTAC enables the regulation of “undruggable” target protein levels. , Several PROTACs have recently been designed for drug development. −

During the ternary complex formation process, the E3 ligand and warhead in the PROTAC substructures bind to the E3 ligase and target protein, respectively. These substructures can be identified using approaches similar to those used for traditional small-molecule drugs, such as virtual and high-throughput screening. The linker, another substructure in PROTAC, joins the E3 ligand and the warhead. , The length and composition of the linker significantly affect the positive cooperative assembly of the ternary complex, degradation activity, and target selectivity. − Despite the vital role of the linker in PROTAC activity, the design of the linker is challenging because the combination of ligands and the linker leads to optimal degradation. Although optimizing the linker by referring to the experimental complex structures is a rational design approach, as demonstrated by Kofink et al. and Testa et al., , its applicability is limited due to the small number of available experimental structures. Therefore, simulation-based methods are required for the modeling of ternary complex structures.

For this modeling, Drummond et al. proposed four ensemble methods to predict ternary complexes by combining protein–protein docking and linker conformation analysis. Rosetta-based structural predictions have also been reported, successfully reproducing known ternary complex crystal structures. Although other methods for the predictions close to crystal structures have been reported; − Schiemer et al. demonstrated that BTK-PROTAC–CIAP1 ternary complexes in solution can adopt multiple conformations beyond those crystal structures. In addition, Dixon et al. showed that the dynamic nature of the complex via PROTACs plays a crucial role in understanding their degradation activity. Therefore, for rational PROTAC design, predicting the ternary complex structure using only the known crystal structures is insufficient. Consequently, a comprehensive search for stable ternary complex structures and the distribution profiles between these structures is needed to evaluate and understand the basis of the PROTAC activity.

In this study, for three PROTACs with different linker lengths and activities, we investigated possible conformations of the PROTAC-mediated ternary complexes using the hybrid conformational search based on two rare-event sampling methods, the parallel cascade selection molecular dynamics (PaCS-MD) and outlier flooding (OFLOOD) method. , Distribution profiles from the free-energy landscapes (FELs) and stable conformations were obtained by constructing Markov state models (MSMs) in a specified conformational subspace. An extensive structural search of ternary complexes revealed multiple promising stable structures, providing essential insights into the induction of ubiquitination and rational structure–activity relationships for PROTACs.

Materials and Methods

PROTAC Information

This study involved the conformational search of the Von Hippel–Lindau tumor suppressor (VHL)-second bromodomain of bromodomain-containing protein 4 (BRD4BD2)-PROTAC systems. We used MZ1, MZ2, and MZ4, which induce BRD4BD2 degradation mediated by VHL. The E3 ligands and warheads of the three PROTACs share common structures, while the linker length differs for each PROTAC (Figure ). Table shows the linker length and the pharmacological properties of each PROTAC, which was experimentally estimated by Chan et al. In the table, α and pDC50 values represent the cooperativity of ternary complex formation experimentally defined as the ratio of the binary dissociation constant to the ternary dissociation constant and the protein degradation activity, respectively. Chan et al. reported that MZ1 with middle linker length has the best positive cooperativity. In addition, MZ4 has lower degradation activity than MZ1 and MZ2. Among these PROTACs, the 3D structure of the ternary complex formed via MZ1 has only been reported (Protein Data Bank [PDB] ID: 5T35).

1.

1

Open in a new tab

Common structure of MZ4/1/2.

1. Linker Length and Experimental Information of MZ1/2/4 .

PROTAC linker length (n) α pDC50 (BRD4 short isoform) PDB
MZ4 short (n = 2) 5.7 7.0 N.D.
MZ1 middle (n = 3) 7.4 8.1 5T35
MZ2 long (n = 4) 4.7 8.4 N.D.
Open in a new tab
a

The α values represent the cooperativity of ternary complex formation for each system, experimentally defined as the ratio of the binary dissociation constant to the ternary dissociation constant.

System Setup of PROTAC Systems

A setup for molecular dynamics (MD) simulations of the PROTAC systems was developed. First, an initial ternary complex structure with a PROTAC extension was prepared. Based on the crystal structure of MZ1, a model was generated by manually altering the dihedral angle of the linker substructure to 180° using GaussView, as shown in Figure S1 in the Supporting Information. Note that the binding poses between each protein and ligand substructure were retained from the original crystal structure. Next, each initial model was placed in a rectangular box with a margin of 15 Å apart from the proteins, and each box was filled with water molecules. Na+ and Cl– ions were added to neutralize each system. The force fields of the proteins and water molecules were set to the Amber ff14SB and TIP3P models, respectively. , For the PROTACs, the restrained electrostatic potential procedure was employed to fit and convert the partial charges to reproduce the electrostatic potential, which was calculated using Gaussian 16 Rev C.01. The electrostatic potential was calculated at the HF/6–31G level by using an extended PROTAC structure. The force fields of the PROTACs were derived using the general AMBER force field 2 (GAFF2). These systems were performed using the antechamber and leap module of AmberTools20. Each system was minimized by using the steepest descent algorithm. Energy minimization for 1000 steps was performed twice, with and without harmonic positional restraints on the heavy solute atoms (a force constant of 10 kcal/mol/Å2). Each minimized system was gradually heated to 300 K during a 100 ps NVT MD simulation with positional restraints. After the NVT equilibration, the NPT MD simulation was performed at 300 K and 1 bar for 800 ps with a gradual reduction in the positional restraints from 10 to 0 kcal/mol/Å2. To increase the time step of 2 fs in the MD simulations, the LINCS algorithm was set as a constraint algorithm. The velocity-rescaling thermostat and Berendsen barostat were set to control the temperature and pressure of each system, respectively. − The GROMACS version 2021.5 was used for all MD simulations in this study.

Hybrid Conformational Search Based on PaCS-MD and OFLOOD

PaCS-MD is a rare-event sampling method without any extra bias in MD simulations. This method repeats multiple MD simulations from reasonably selected initial structures for efficient conformational sampling. Restarting structures are selected based on specified reaction coordinates (RCs) of MD trajectories. The sampling efficiency of PaCS-MD depends on the specification of the appropriate RCs. In previous studies using PaCS-MD, a wide variety of biomolecule descriptors, such as the distance between two domains and the root-mean-square deviation (RMSD) for a known structure, were selected as RCs. ,− Additionally, the hybrid conformational search based on the combination of PaCS-MD with OFLOOD enables the efficient search of a broad configurational space. , The OFLOOD method detects outliers among MD trajectories based on a defined RC space and resamples these outliers by restarting multiple MD simulations from them. Similar to PaCS-MD, the sampling efficiency of OFLOOD depends on defining an appropriate RC space. ,

This study adopted PaCS-MD and OFLOOD for each PROTAC system to perform an extensive conformational search of the ternary complex structures. Our method was divided into two steps: Cα-RMSD-based PaCS-MD and OFLOOD. First, the PaCS-MD using the Cα-RMSD for the crystal structure with MZ1 (PDB ID: 5T35) was performed starting from each initial structure constructed in the setup. For each cycle in the PaCS-MD, 100 ps MD simulations were run in parallel with 10 conformations with the regenerated velocity. In each MD simulation, 100 snapshots were recorded with an interval of 1 ps. After multiple MD simulations, the Cα-RMSD values of their snapshots were calculated for the crystal structure of VHL-MZ1-BRD4BD2 (PDB ID: 5T35), and 10 snapshots with low Cα-RMSD values were selected as initial structures in the next cycle. In the first step, this process was repeated for 100 cycles. This step contributes to the recording of one of the conformational transitions from the initial to the reference structure. Second, OFLOOD was performed based on the MD trajectories obtained by the PaCS-MD sampling. Two RCs were selected for the OFLOOD sampling: the distance between the centroids of VHL and BRD4BD2 and the dihedral angle forming VHL and BRD4BD2 (Figure ). The dihedral angle was defined using the Cα atom of Trp117 in VHL, the centroid of VHL, the centroid of BRD4BD2, and the Cα atom of Arg423 in BRD4BD2. These two Cα atoms were selected because the root-mean-square fluctuation (RMSF) of these residues was low in the preconventional MD simulation (Figure S2). In each OFLOOD cycle, outliers were detected with FlexDice from distributions in the subspace defined by the two RCs, and 100 ps MD simulations were restarted from the detected outliers in parallel with 10 outliers at different initial velocities. FlexDice is a hierarchical clustering algorithm for a subspace. During the clustering process, the subspace is divided into two types of cells; a set of dense and sparse cells. The dense cells contain highly populated states as clusters, whereas the sparse cells contain rarely occurring states as outliers. After hierarchical clustering, FlexDice can efficiently detect the rarely occurring states as outliers. Note that outliers with a distance over 51 Å were excluded from the initial structures in the next cycle, because they may dissociate proteins and PROTACs. The OFLOOD sampling was repeated for 150 cycles. This step contributes to the exploration of a wide range of conformational space. For each PROTAC system, the total computational time of the hybrid conformational search was 250 ns.

2.

2

Open in a new tab

Reaction coordinates specified in OFLOOD. (A) Distance between the centroids of both proteins. (B) Dihedral angle defined in both proteins. VHL and BRD4BD2 are colored magenta and cyan, respectively.

Free-Energy Calculation Based on MSM

MSMs were constructed to calculate the FELs and quantitatively extract the stable conformations of the PROTAC systems. Here, conformational resampling was performed using snapshots sampled by the hybrid conformational search to construct reliable MSMs. Specifically, multiple conformations were selected from the grids on the distance-dihedral subspace searched by the hybrid conformational search, in which each grid was generated per 1 Å and 15°. Here, snapshots close to each grid were selected as the initial structures of conformational resampling. For the complex systems with MZ4/1/2, 344/397/461 snapshots were selected as the initial structures of the additional MD simulations, respectively. Subsequently, 50 ns conventional MD simulations were performed for each initial structure, and snapshots were recorded every 100 ps. In summary, the total simulation time per system was approximately 17.2 to 23.0 μs. After the conformational resampling with the additional MD simulations, the recorded snapshots were clustered into 100 microstates based on the k-means clustering. The transition matrix (T ij ) between the i and j microstates was estimated by counting the structural transitions between them. The lag time of the transitions among these microstates was set to 5 ns in estimating the transition matrix. Subsequently, a stationary distribution, π = (π1, π2, ···, and π100), was obtained from the transition matrix, and each component can be used as a weight of each microstate, enabling one to calculate the FEL using π. Finally, the FEL of the microstate i was calculated as follows

Fi=−kBT⁡lnπimaxjπj 1

where k B and T are the Boltzmann constant and absolute temperature, respectively. PyEMMA was used for the FEL calculations. In addition, to compare the conformational space areas, we computed the α-shape area of the plotted points for each system using the Alpha Shape Toolbox in Python, and the areas were normalized to the value for the MZ1 system.

Ternary Complex Structure Analysis

The representative states of the ternary complex structures were identified from the microstates generated during the MSM construction. The microstate with the lowest free-energy in each system was defined as the global minimum state. In addition, we also extracted other representative states based on their topological features in the FEL. Specifically, we considered the low free-energy states located in distinct basins that are well separated from neighboring basins by sufficient energy barriers, which is larger than thermal fluctuations (1k B T). Among these, a state was further defined as the system-specific state if its free-energy was significantly lower than the corresponding coordinates in the free-energy landscapes of other systems, indicating a uniquely conformation. For a set of representative states, their Cα-RMSD values were compared with those of the crystal structure with MZ1 (PDB ID: 5T35). As their measures, the radius of gyration (R g), strain energy of PROTAC, and contact map were also calculated for each state. The R g and contact map was computed using MDTraj, and strain energy was computed using the Prime module in Schrödinger Suite 2023–3. The strain energy is defined as the difference in potential energy between the given conformation and the optimized conformation obtained by performing energy minimization, starting from the given structure. The optimized conformation represents a local minimum on the potential energy surface. A total of 100 structures around these minimum states were extracted to determine the binding free-energy using the generalized Born and surface area continuum solvation method (MM/GBSA). We calculated the binding free-energy for the PROTACs in VHL-BRD4BD2 (ΔG bind ) and protein–protein interactions without the PROTACs (ΔG bind ). The gmx_mmpbsa tool was used for the binding free-energy calculations. In addition, to analyze the accessibility to ubiquitin of each state, we superimposed each state onto the ubiquitination complex assembly (PDB ID: 8RX0 ) based on the VHL. And then, the positions of K368, K445, and K456 in the BRD4BD2, identified as the “ubiquitination zone” by Crowe et al., were focused on the aligned model. The angles between these lysine residues, the centroid of BRD4BD2, and the cysteine residue in UBE2R1, the active site residue for ubiquitination, were calculated. For each residue, the Cα atom was used for the angle calculation.

Results

Hybrid Conformational Search Based on PaCS-MD and OFLOOD

Structural transitions during PaCS-MD on Cα-RMSD are shown in Figure . Figure A shows the Cα-RMSD profile for each system versus cycle. Note that the minimum Cα-RMSD value was plotted in every cycle. The complex structures at cycles 0 (initial), 10, 20, and 100 in each system are shown in Figure B–D. All systems showed Cα-RMSD values rapidly decreasing below 2 Å within the 20 cycles. This rapid decrease could be attributed to the structural flexibility of these transition states, which are only connected by PROTACs and do not have protein–protein interactions. These structural transitions converged after the 20th cycle. Finally, the Cα-RMSD value converged to less than 1 Å. The result that the three systems showed no differences in their Cα-RMSD profiles suggests that all systems are reachable at least for the reference pose, the crystal structure.

3.

3

Open in a new tab

Structural transitions in the PaCS-MD simulations for all systems characterized by the Cα-RMSD. (A) Minimum Cα-RMSD values versus PaCS-MD cycle for each PROTAC system. The profiles of MZ4/1/2 are colored red, yellow, and green, respectively. (B–D) Structural transitions of the ternary complexes for MZ4/1/2 during the PaCS-MD cycles. The structures of VHL, BRD4BD2, and PROTAC are colored in magenta, cyan, and green, respectively.

For further conformational search of the ternary complex structures, OFLOOD was subsequentially performed using the PaCS-MD trajectories. Figure S3 shows the extension of the subspace searched by OFLOOD according to the cycle. After 150 OFLOOD cycles, a broad conformational subspace was searched for each system. The MSMs were constructed on each conformational subspace using the trajectories from additional MD simulations. Figure shows the FEL of each PROTAC system based on each MSM. These FELs indicate that the linker length in each PROTAC is proportional to the extent of each conformational subspace. Indeed, we calculated the α-shape area of each conformational space, and the areas were normalized to the value for the MZ1 system. As a result, the scaled values of the MZ4, MZ1, and MZ2 systems are 0.897, 1.000, and 1.187, respectively. This implies that the linker length affects the diversity of the ternary complex structures. Figure A shows a broad FEL of the MZ4 system with low free-energy values less than 2k B T, colored purple and dark blue. Thus, the MZ4 system has several stable conformations as the ternary complex structures. Figure B shows the FEL of the MZ1 system. Compared with the FEL of the MZ4 system, the area with low free-energy values is narrow and limited in the FEL of the MZ1 system. Specifically, the two regions, which are separated by high free-energy barriers, have significantly low free-energy values (less than 1k B T). One is at approximately [40, −100], which coincides with the position of one of the low free-energy regions in the MZ4 system. The other is approximately [45, 0], and this state does not appear in the FELs of the MZ4 and MZ2 systems. This MZ1 system-specific region is widely spread along the Y-axis. Figure C shows the FEL of the MZ2 system, indicating that the MZ2 system has a single low free-energy region, which coincides with the common position of the low free-energy regions of the MZ4 and MZ1 systems. In brief, all systems have the same region with low free-energy values, approximately [40, −100]. Interestingly, the MZ4 system has many local free-energy minima and the MZ1 system has one more region with low free-energy values. In contrast, the MZ2 system has no other regions with low free-energy values.

4.

4

Open in a new tab

FELs of all of the PROTAC systems. FELs of the MZ4 (A), MZ1 (B), and MZ2 (C) systems. The FEL of each PROTAC system was mapped onto each 2D conformational space. X-axis and Y-axis represent the distance between the centroids of VHL and BRD4BD2 and the dihedral angle forming VHL and BRD4BD2, respectively. The triangle marker means the plot of the ternary complex of PDB ID: 5T35. In each FEL, the global minimum state is shown as a square marker. Round markers indicate the plot of the system-specific states.

Structural Evaluation of Stable Conformations

Based on the FELs, ternary complex structures with low free-energy values were extracted as specific states. For the MZ4 and MZ1 systems, the global minimum and these specific states were obtained from the low free-energy regions. The locations of the extracted states in the FELs are shown in Figure as the markers. For the MZ2 system, the global minimum state was only obtained. The complex structures and the corresponding contact maps are shown in Figures and S4, respectively. For each state, structural evaluation of these structures was performed based on their RMSD and binding free-energy values based on the MM/GBSA method, as shown in Table . The RMSD values were calculated using the Cα atoms for the crystal structure of MZ1 (PDB ID: 5T35).

5.

5

Open in a new tab

Ternary complex structures of the global and local minimum states in each system. VHL, BRD4BD2, and PROTACs are colored in magenta, cyan, and green, respectively.

2. Details of Ternary Complex Structures .

PROTAC state Cα-RMSD with 5T35 Rg (Å) PROTAC strain energy [kcal/mol] ΔG bind [kcal/mol] ΔG bind [kcal/mol]
MZ4 global minimum state 1.5 Å 25.5 44.363 –100.52 ± 5.23 –24.33 ± 7.26
specific state 12.7 Å 25.4 22.531 –100.10 ± 5.61 –22.14 ± 12.13
MZ1 global minimum state 2.0 Å 25.5 40.291 –104.44 ± 5.47 –25.52 ± 6.56
specific state 12.8 Å 27.4 16.712 –97.64 ± 5.00 –16.36 ± 4.95
MZ2 global minimum state 1.6 Å 25.3 43.094 –106.23 ± 5.53 –24.76 ± 5.98
Open in a new tab
a

Cα-RMSD values were calculated based on the crystal structure with MZ1 (PDB ID: 5T35). The 100 structures around each minimum state were extracted to determine the binding free energy using MM/GBSA.

The global minimum states of the MZ4/1/2 systems were noted at [39.84, −104.90], [39.58, −109.81], and [39.44, −116.32], respectively. These ternary complex structures are similar to each other and share a low Cα-RMSD value of 1.5–2.0 Å. These structures also have the binding free-energy in proteins (ΔG bind ) of approximately −25 kcal/mol in common. In addition, the binding free-energies of the MZ4/1/2 systems (ΔG PROTAC ) were −100.52 ± 5.23, −104.44 ± 5.47, and −106.23 ± 5.53 kcal/mol, respectively, meaning that there is no substantial difference in ΔG PROTAC among the systems. R g and strain energies of the PROTACs in these states are also comparable to each other. Notably, the strain energies of these states exhibited remarkable values over 40 kcal/mol. Overall, the global minimum states in the three systems are largely consistent across the different linker lengths.

In the MZ4 system, MZ4 system-specific regions with low free-energy values were observed. We picked up the specific state at [39.72, 45.01] in the FEL. This ternary complex structure differs significantly from the global minimum state, with the orientation of proteins rotated by about 145° (left side in Figure ). Notably, the ΔG PROTAC and ΔG bind of the conformation were −100.10 ± 5.61 and −22.14 ± 12.13 kcal/mol, respectively. These values are comparable to those of the global minimum state. These results indicate that the MZ4 system has another stable conformation, which has similar free-energy and binding free-energy to the global minimum state. Additionally, the strain energy of MZ4 in the specific state is lower than that in the global minimum state, while R g is comparable to that of the global minimum state.

The MZ1 system also has a specific region with low free-energy values, found at [44.66, 14.68] in the FEL. These proteins in the ternary complex structure are distant relative to the global minimum state, and the orientation of proteins rotated by approximately 120° (the center in Figure ). The value of R g is slightly higher than that of the global minimum state, and the strain energy of MZ1 in this state is relatively lower than that in the global minimum state, as shown in Table . The binding free-energy analysis shows that the proteins and MZ1 bind stably, for which the ΔG PROTAC value was −97.64 ± 5.00 kcal/mol. By contrast, the affinity between proteins in this conformation was −16.36 ± 4.95 kcal/mol, which is inferior to the global minimum. These results suggest that this MZ1 system-specific state is less stable than the global minimum state, the so-called metastable structure.

For each state of the ternary complex, the alignment model with ubiquitination complex assembly was constructed. Figure A shows the alignment of the global minimum conformation with MZ1 and the ubiquitination complex assembly. The red spheres represent K368, K445, and K456 in BRD4BD2, and the lime cartoon model highlights the ubiquitin in the assembly. The model indicates that the three lysine residues of the BRD4BD2 are positioned toward the ubiquitin side. Specifically, the minimum angle formed by the lysine residue, the centroid of BRD4BD2, and the cysteine residue in the active site for ubiquitination was 20.9°. In the global minimum states of the MZ4 and MZ2 systems, the lysine residues are also positioned toward the ubiquitin side because these states are consistent with the global minimum state in the MZ1 system. In contrast, the alignment of the MZ4-specific state and assembly (Figure B) shows that these lysine residues are in the opposite orientation to the ubiquitin, and the minimum angle value was 95.4°. Additionally, the alignment with the assembly of the MZ1 system-specific state is shown in Figure C, where BRD4BD2 and UBE2R1 (gray cartoon) are in conflict.

6.

6

Open in a new tab

Alignment model with ternary complexes and ubiquitination complex assembly. (A) Alignment model of ubiquitination complex assembly and the global minimum state in the MZ1 system. (B) Alignment model of ubiquitination complex assembly and the MZ4 system-specific state. (C) Alignment model of ubiquitination complex assembly and the MZ1 system-specific state. K368, K445, and K456 of BRD4BD2 are represented as red spheres.

Discussion

The FELs and stable ternary complex structures of the three PROTAC systems were obtained by using our methods (Figures and ). Quantitative evaluation of these stable conformations enabled estimation of their distribution profiles (Figure and Table ). Two dominant states were identified for the ternary complex in the MZ4 system with a short-length linker. One is the global minimum state in the FEL, and the other is the MZ4 system-specific state. The former structure, the global minimum state, is similar to the crystal structure with MZ1 and has the lowest free-energy in the FEL and low binding free-energies. In addition, for the latter structure, the MZ4 system-specific state has a similar low free-energy and binding free-energies compared to the global minimum state. For the ternary complex in the MZ1 system with a middle-length linker, a set of dominant and metastable states was identified. The dominant state has a similar structure to the crystal structure with MZ1. The metastable state has an inferior binding free-energy in the proteins compared to the global minimum state. For the ternary complex in the MZ2 system with a long-length linker, only one dominant state, which is similar to the crystal structure with MZ1, was identified. In summary, the MZ4/1/2 systems commonly have a dominant state similar to the known complex; however, their distribution profiles differ significantly, even though these PROTACs differ only in their linker length. Furthermore, the system-specific states in these profiles may contribute to the variation in cooperativity and further to that in degradation activity of PROTACs, as described below.

The MZ4, a PROTAC with a short linker, has lower degradation activity than MZ1 and MZ2 (“pDC50” column in Table ), and our results showed that the distribution profile of the MZ4-mediated ternary complex had multiple stable conformations. In a previous study, Crowe et al. suggested that recruiting the target protein to the orientation apart from the ubiquitin would cause low efficient ubiquitination and degradation. As shown by our alignment models in Figure , the MZ4-mediated ternary complex has both favorable and unfavorable conformations for ubiquitination. Given that it readily forms an unfavorable conformation, it likely contributes to the low degradation activity of the MZ4. The MZ1, a PROTAC with a middle linker length, has the best cooperativity of ternary complex formation than MZ2 and MZ4 (“α” column in Table ), and our study revealed that the MZ1 system has the distribution profile with the global minimum state and the specific state, so-called metastable state. This specific state has a weak binding energy in the proteins, while the free-energy in the system is low. Focused on the interface in the proteins of this specific state, Arg69 in VHL and Asp389 in BRD4BD2 formed ionic interaction (Figure S5). This interaction contributes to keeping the proteins together, even though the protein–protein interface area in this specific state is limited (the center side in Figure ). Specifically, for Arg69 in VHL, Kumar and Sobhia have previously suggested that this residue plays a key role in the interface of the ternary complexes, particularly in those formed by PROTACs exhibiting high cooperativity. Moreover, for the MZ1 system, the strain energy in this specific state is lower than that in the global minimum state, as shown in Table . This indicates that this MZ1 conformation is native-like and can be often present in the unbound state. These findings suggest that the MZ1 system-specific state can be one of the intermediate states from binary to ternary formation, which prevents them from dissociating before reaching the global minimum state. In contrast, the MZ4 and MZ2 systems do not have the arrangement of the proteins in the MZ1 system-specific state due to their linker length. In the MZ4 system, attempts to adopt the MZ1 system-specific conformation can cause protein collisions due to the short linker in MZ4 (the left side in Figure ). Conversely, the longer linker can prevent ionic interactions between the proteins when trying to achieve this conformation in the MZ2 system (right side in Figure ). In summary, the appropriate linker length in MZ1 enables the high cooperativity of ternary complex formation.

7.

7

Open in a new tab

Schematic view of the suggested model for assembly from the binary to the ternary complex formation. The α values represent the cooperativity of ternary complex formation experimentally defined as the ratio of the binary dissociation constant to the ternary dissociation constant. VHL, BRD4BD2, and PROTACs are shown in magenta, cyan, and green, respectively. “R” and “D” mean Arg69 in VHL and Asp389 in BRD4BD2, respectively.

As above, the distribution profiles in the ternary complex structures allow us to understand the mechanism of PROTACs activities, the degradation activity, and the cooperativity. Given that Schiemer et al. reported that PROTAC-mediated target proteins and E3 bind flexibly to each other and form multiple ensembles, our study is effective for the rational design of PROTACs that induce both the intermediate and ubiquitin-accessible states. Specifically, sampling and profiling of the ternary complex structures via PROTAC with some activities are available to optimize their linker components and length. These structural insights can be leveraged to modify the PROTC’s substructure, promoting favorable orientations for ubiquitination and having additional intermediate states. Although a comprehensive understanding of the PROTAC function is yet to be achieved, our study provides a valuable framework for guiding future investigations.

Conclusions

We attempted an extensive conformational search of PROTAC-mediated ternary complex structures using the PaCS-MD and OFLOOD method for three PROTACs with different linker lengths, experimentally characterized degradation activities, and cooperativity. Our analysis revealed that all PROTACs share a common low free-energy conformation resembling a known crystal structure but differ markedly in their conformational distribution profiles. Notably, the MZ4 system exhibited an alternative stable state that may hinder ubiquitin transfer, potentially explaining its reduced degradation activity. In contrast, the MZ1 system displayed a metastable conformation that may promote ternary complex formation, which is consistent with its high cooperativity. These results highlight the critical role of linker-dependent structural dynamics in determining the PROTAC efficacy. By providing structural distribution profiles linked to biological outcomes, our approach offers a practical framework for the rational design of PROTACs with an optimized degradation performance.

Supplementary Material

ci5c00102_si_001.pdf (902.9KB, pdf)

Acknowledgments

PaCS-MD and OFLOOD used the Cygnus computational resources provided by the Multidisciplinary Cooperative Research Program at the Center for Computational Sciences (Project Code: CADD) at the University of Tsukuba. Additional MD simulation for MSM construction used computational resources of (supercomputer Fugaku provided by the RIKEN Center for Computational Science) (Project ID: hp230216).

Glossary

Abbreviations

BRD4BD2

second bromodomain of bromodomain-containing protein 4

FEL

free-energy landscape

MD

molecular dynamics

PaCS-MD

parallel cascade selection molecular dynamics

MSM

Markov state model

OFLOOD

outlier flooding

PDB

Protein Data Bank

PROTAC

proteolysis-targeting chimeras

RMSD

root-mean-square deviation

Rg

radius of gyration

VHL

Von Hippel-Lindau tumor suppressor

3D structures of the proteins were downloaded from the Protein Data Bank (PDB). We used AmberTools20 and Gaussian 16 Rev C.01 for PROTAC preparation. AmberTools20 was used to prepare the MD system. GROMACS 2021.5 was used as the MD engine. Rms module in GROMACS was used to calculate Cα-RMSD. Alpha Shape Toolbox was used to calculate the α-shape area. MDTraj was used to calculate R g and the contact map. Prime module in Schrödinger Suite 2023–3 was used to calculate strain energy. PyEMMA was used to estimate MSM. Gmx_MMPBSA was used to calculate binding free-energy based on the MM/GBSA method. PyMOL was used for visualization. Scripts of our methodology and these conformational search results are deposited to Zenodo (DOI: 10.5281/zenodo.15307772).

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jcim.5c00102.

  • Initial models of each PROTAC system (Figure S1); RMSF of the Cα atom in VHL and BRD4BD2 (Figure S2); extension of the conformational search during the OFLOOD cycles (Figure S3); contact maps of global minimum states and system-specific states (Figure S4), and binding site in the MZ1-specific state (Figure S5) (PDF)

#.

G.K. and T.H. contributed equally to this work. G.K., T.H., R.H., Y.S., Taka.H., and R.Y. conceived the study. G.K., T.H., R.Y., and Taka.H. selected the targeted PROTAC-mediated ternary complexes. G.K., T.H., and R.H. implemented the PaCS-MD and OFLOOD method. G.K. and T.H. performed the experiments and G.K. analyzed the data. The manuscript was written with contributions from all the authors. All authors approved the final version of the manuscript.

This research was partially supported by the Research Support Project for Life Science and Drug Discovery [Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)] (grant number JP24ama121029j0003) of the Japan Agency for Medical Research and Development (AMED), MEXT as “Program for Promoting Researches on the Supercomputer Fugaku” (Simulation- and AI-driven next-generation medicine and drug discovery based on “Fugaku”, JPMXP1020230120), and KAKENHI (grant numbers JP23K16987, JP23H04879, JP23H02427, and JP23K18033) from the Japan Society for the Promotion of Science (JSPS).

The authors declare no competing financial interest.

References

  1. Makurvet F. D.. Biologics vs Small Molecules: Drug Costs and Patient Access. Med. Drug Discovery. 2021;9:100075. doi: 10.1016/j.medidd.2020.100075. [DOI] [Google Scholar]
  2. Békés M., Langley D. R., Crews C. M.. PROTAC Targeted Protein Degraders: The Past Is Prologue. Nat. Rev. Drug Discovery. 2022;21:181–200. doi: 10.1038/s41573-021-00371-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Kelm J. M., Pandey D. S., Malin E., Kansou H., Arora S., Kumar R., Gavande N. S.. PROTAC’ing Oncoproteins: Targeted Protein Degradation for Cancer Therapy. Mol. Cancer. 2023;22:62. doi: 10.1186/s12943-022-01707-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Churcher I.. Protac-Induced Protein Degradation in Drug Discovery: Breaking the Rules or Just Making New Ones? J. Med. Chem. 2018;61(2):444–452. doi: 10.1021/acs.jmedchem.7b01272. [DOI] [PubMed] [Google Scholar]
  5. Coleman K. G., Crews C. M.. Proteolysis-Targeting Chimeras: Harnessing the Ubiquitin-Proteasome System to Induce Degradation of Specific Target Proteins. Annu. Rev. Cancer Biol. 2018;2:41–58. doi: 10.1146/annurev-cancerbio. [DOI] [Google Scholar]
  6. Pettersson M., Crews C. M.. Proteolysis TArgeting Chimeras (PROTACs)Past, Present and Future. Drug Discovery Today: Technol. 2019;31:15–27. doi: 10.1016/j.ddtec.2019.01.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Troup R. I., Fallan C., Baud M. G. J.. Current Strategies for the Design of PROTAC Linkers: A Critical Review. Explor. Targeted Anti-Tumor Ther. 2020;1:273–312. doi: 10.37349/etat.2020.00018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Sakamoto K. M., Kim K. B., Kumagai A., Mercurio F., Crews C. M., Deshaies R. J.. Protacs: Chimeric Molecules That Target Proteins to the Skp1-Cullin-F Box Complex for Ubiquitination and Degradation. Proc. Natl. Acad. Sci. U.S.A. 2001;98:8554–8559. doi: 10.1073/pnas.141230798. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Powell C. E., Gao Y., Tan L., Donovan K. A., Nowak R. P., Loehr A., Bahcall M., Fischer E. S., Jänne P. A., George R. E., Gray N. S.. Chemically Induced Degradation of Anaplastic Lymphoma Kinase (ALK) J. Med. Chem. 2018;61(9):4249–4255. doi: 10.1021/acs.jmedchem.7b01655. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Burslem G. M., Smith B. E., Lai A. C., Jaime-Figueroa S., McQuaid D. C., Bondeson D. P., Toure M., Dong H., Qian Y., Wang J., Crew A. P., Hines J., Crews C. M.. The Advantages of Targeted Protein Degradation Over Inhibition: An RTK Case Study. Cell Chem. Biol. 2018;25(1):67–77.e3. doi: 10.1016/j.chembiol.2017.09.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. He Y., Khan S., Huo Z., Lv D., Zhang X., Liu X., Yuan Y., Hromas R., Xu M., Zheng G., Zhou D.. Proteolysis Targeting Chimeras (PROTACs) Are Emerging Therapeutics for Hematologic Malignancies. J. Hematol. Oncol. 2020;13:103. doi: 10.1186/s13045-020-00924-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Lai A. C., Toure M., Hellerschmied D., Salami J., Jaime-Figueroa S., Ko E., Hines J., Crews C. M.. Modulares PROTAC-Design Zum Abbau von Onkogenem BCR-ABL. Angew. Chem. 2016;128(2):818–821. doi: 10.1002/ange.201507634. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Schneekloth A. R., Pucheault M., Tae H. S., Crews C. M.. Targeted Intracellular Protein Degradation Induced by a Small Molecule: En Route to Chemical Proteomics. Bioorg. Med. Chem. Lett. 2008;18(22):5904–5908. doi: 10.1016/j.bmcl.2008.07.114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Lee H., Puppala D., Choi E. Y., Swanson H., Kim K. B.. Targeted Degradation of the Aryl Hydrocarbon Receptor by the PROTAC Approach: A Useful Chemical Genetic Tool. ChemBioChem. 2007;8(17):2058–2062. doi: 10.1002/cbic.200700438. [DOI] [PubMed] [Google Scholar]
  15. Zhou C., Fan Z., Gu Y., Ge Z., Tao Z., Cui R., Li Y., Zhou G., Huo R., Gao M., Wang D., He W., Zheng M., Zhang S., Xu T.. Design, Synthesis, and Biological Evaluation of Potent and Selective PROTAC Degraders of Oncogenic KRASG12D. J. Med. Chem. 2024;67:1147–1167. doi: 10.1021/acs.jmedchem.3c01622. [DOI] [PubMed] [Google Scholar]
  16. Tang R., Wang Z., Xiang S., Wang L., Yu Y., Wang Q., Deng Q., Hou T., Sun H.. Uncovering the Kinetic Characteristics and Degradation Preference of PROTAC Systems with Advanced Theoretical Analyses. JACS Au. 2023;3(6):1775–1789. doi: 10.1021/jacsau.3c00195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Burslem G. M., Crews C. M.. Small-Molecule Modulation of Protein Homeostasis. Chem. Rev. 2017;117:11269–11301. doi: 10.1021/acs.chemrev.7b00077. [DOI] [PubMed] [Google Scholar]
  18. Sun X., Gao H., Yang Y., He M., Wu Y., Song Y., Tong Y., Rao Y.. Protacs: Great Opportunities for Academia and Industry. Signal Transduction Targeted Ther. 2019;4:64. doi: 10.1038/s41392-019-0101-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Roy M. J., Winkler S., Hughes S. J., Whitworth C., Galant M., Farnaby W., Rumpel K., Ciulli A.. SPR-Measured Dissociation Kinetics of PROTAC Ternary Complexes Influence Target Degradation Rate. ACS Chem. Biol. 2019;14(3):361–368. doi: 10.1021/acschembio.9b00092. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gadd M. S., Testa A., Lucas X., Chan K. H., Chen W., Lamont D. J., Zengerle M., Ciulli A.. Structural Basis of PROTAC Cooperative Recognition for Selective Protein Degradation. Nat. Chem. Biol. 2017;13(5):514–521. doi: 10.1038/nchembio.2329. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Zorba A., Nguyen C., Xu Y., Starr J., Borzilleri K., Smith J., Zhu H., Farley K. A., Ding W. D., Schiemer J., Feng X., Chang J. S., Uccello D. P., Young J. A., Garcia-Irrizary C. N., Czabaniuk L., Schuff B., Oliver R., Montgomery J., Hayward M. M., Coe J., Chen J., Niosi M., Luthra S., Shah J. C., El-Kattan A., Qiu X., West G. M., Noe M. C., Shanmugasundaram V., Gilbert A. M., Brown M. F., Calabrese M. F.. Delineating the Role of Cooperativity in the Design of Potent PROTACs for BTK. Proc. Natl. Acad. Sci. U.S.A. 2018;115(31):E7285–E7292. doi: 10.1073/pnas.1803662115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kofink C., Trainor N., Mair B., Wöhrle S., Wurm M., Mischerikow N., Roy M. J., Bader G., Greb P., Garavel G., Diers E., McLennan R., Whitworth C., Vetma V., Rumpel K., Scharnweber M., Fuchs J. E., Gerstberger T., Cui Y., Gremel G., Chetta P., Hopf S., Budano N., Rinnenthal J., Gmaschitz G., Mayer M., Koegl M., Ciulli A., Weinstabl H., Farnaby W.. A selective and orally bioavailable VHL-recruiting PROTAC achieves SMARCA2 degradation in vivo. Nat. Commun. 2022;13(1):5969. doi: 10.1038/s41467-022-33430-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Testa A., Hughes S. J., Lucas X., Wright J. E., Ciulli A.. Structure-Based Design of a Macrocyclic PROTAC. Angew. Chem., Int. Ed. 2020;59(4):1727–1734. doi: 10.1002/anie.201914396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Drummond M. L., Williams C. I.. In Silico Modeling of PROTAC-Mediated Ternary Complexes: Validation and Application. J. Chem. Inf. Model. 2019;59(4):1634–1644. doi: 10.1021/acs.jcim.8b00872. [DOI] [PubMed] [Google Scholar]
  25. Bai N., Miller S. A., Andrianov G. V., Yates M., Kirubakaran P., Karanicolas J.. Rationalizing PROTAC-Mediated Ternary Complex Formation Using Rosetta. J. Chem. Inf. Model. 2021;61(3):1368–1382. doi: 10.1021/acs.jcim.0c01451. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Zaidman D., Prilusky J., London N.. ProsetTac: Rosetta Based Modeling of PROTAC Mediated Ternary Complexes. J. Chem. Inf. Model. 2020;60(10):4894–4903. doi: 10.1021/acs.jcim.0c00589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Drummond M. L., Henry A., Li H., Williams C. I.. Improved Accuracy for Modeling ProTAC-Mediated Ternary Complex Formation and Targeted Protein Degradation via New in Silico Methodologies. J. Chem. Inf. Model. 2020;60(10):5234–5254. doi: 10.1021/acs.jcim.0c00897. [DOI] [PubMed] [Google Scholar]
  28. Ignatov M., Jindal A., Kotelnikov S., Beglov D., Posternak G., Tang X., Maisonneuve P., Poda G., Batey R. A., Sicheri F., Whitty A., Tonge P. J., Vajda S., Kozakov D.. High Accuracy Prediction of PROTAC Complex Structures. J. Am. Chem. Soc. 2023;145(13):7123–7135. doi: 10.1021/jacs.2c09387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Schiemer J., Horst R., Meng Y., Montgomery J. I., Xu Y., Feng X., Borzilleri K., Uccello D. P., Leverett C., Brown S., Che Y., Brown M. F., Hayward M. M., Gilbert A. M., Noe M. C., Calabrese M. F.. Snapshots and Ensembles of BTK and CIAP1 Protein Degrader Ternary Complexes. Nat. Chem. Biol. 2021;17(2):152–160. doi: 10.1038/s41589-020-00686-2. [DOI] [PubMed] [Google Scholar]
  30. Dixon T., MacPherson D., Mostofian B., Dauzhenka T., Lotz S., McGee D., Shechter S., Shrestha U. R., Wiewiora R., McDargh Z. A., Pei F., Pal R., Ribeiro J. V., Wilkerson T., Sachdeva V., Gao N., Jain S., Sparks S., Li Y., Vinitsky A., Zhang X., Razavi A. M., Kolossváry I., Imbriglio J., Evdokimov A., Bergeron L., Zhou W., Adhikari J., Ruprecht B., Dickson A., Xu H., Sherman W., Izaguirre J. A.. Predicting the structural basis of targeted protein degradation by integrating molecular dynamics simulations with structural mass spectrometry. Nat. Commun. 2022;13(1):5884. doi: 10.1038/s41467-022-33575-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Harada R., Shigeta Y.. Hybrid Cascade-Type Molecular Dynamics with a Markov State Model for Efficient Free Energy Calculations. J. Chem. Theory Comput. 2019;15(1):680–687. doi: 10.1021/acs.jctc.8b00802. [DOI] [PubMed] [Google Scholar]
  32. Harada R., Kitao A.. Parallel Cascade Selection Molecular Dynamics (PaCS-MD) to Generate Conformational Transition Pathway. J. Chem. Phys. 2013;139(3):035103. doi: 10.1063/1.4813023. [DOI] [PubMed] [Google Scholar]
  33. Harada R., Nakamura T., Shigeta Y.. Sparsity-Weighted Outlier FLOODing (OFLOOD) Method: Efficient Rare Event Sampling Method Using Sparsity of Distribution. J. Comput. Chem. 2016;37(8):724–738. doi: 10.1002/jcc.24255. [DOI] [PubMed] [Google Scholar]
  34. Harada R., Nakamura T., Takano Y., Shigeta Y.. Protein Folding Pathways Extracted by OFLOOD: Outlier Flooding Method. J. Comput. Chem. 2015;36(2):97–102. doi: 10.1002/jcc.23773. [DOI] [PubMed] [Google Scholar]
  35. Chan K. H., Zengerle M., Testa A., Ciulli A.. Impact of Target Warhead and Linkage Vector on Inducing Protein Degradation: Comparison of Bromodomain and Extra-Terminal (BET) Degraders Derived from Triazolodiazepine (JQ1) and Tetrahydroquinoline (I-BET726) BET Inhibitor Scaffolds. J. Med. Chem. 2018;61(2):504–513. doi: 10.1021/acs.jmedchem.6b01912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Dennington, R. ; Keith, T. ; Millam, J. ; Eppinnett, K. ; Hovell, W. L. ; Gilliland, R. . GaussView, Version 3.0; Semichem, Inc.: Shawnee Mission, KS, 2003. [Google Scholar]
  37. Maier J. A., Martinez C., Kasavajhala K., Wickstrom L., Hauser K. E., Simmerling C.. Ff14SB: Improving the Accuracy of Protein Side Chain and Backbone Parameters from Ff99SB. J. Chem. Theory Comput. 2015;11(8):3696–3713. doi: 10.1021/acs.jctc.5b00255. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Jorgensen W. L., Chandrasekhar J., Madura J. D., Impey R. W., Klein M. L.. Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983;79:926–935. doi: 10.1063/1.445869. [DOI] [Google Scholar]
  39. Frisch, M. J. ; Trucks, G. W. ; Schlegel, H. B. ; Scuseria, G. E. ; Robb, M. A. ; Cheeseman, J. R. ; Scalmani, G. ; Barone, V. ; Petersson, G. A. ; Nakatsuji, H. ; Li, X. ; Caricato, M. ; Marenich, A. V. ; Bloino, J. ; Janesko, B. G. ; Gomperts, R. ; Mennucci, B. ; Hratchian, H. P. ; Ortiz, J. V. ; Izmaylov, A. F. ; Sonnenberg, J. L. ; Williams-Young, D. ; Ding, F. ; Lipparini, F. ; Egidi, F. ; Goings, J. ; Peng, B. ; Petrone, A. ; Henderson, T. ; Ranasinghe, D. ; Zakrzewski, V. G. ; Gao, J. ; Rega, N. ; Zheng, G. ; Liang, W. ; Hada, M. ; Ehara, M. ; Toyota, K. ; Fukuda, R. ; Hasegawa, J. ; Ishida, M. ; Nakajima, T. ; Honda, Y. ; Kitao, O. ; Nakai, H. ; Vreven, T. ; Throssell, K. ; Montgomery, J. A., Jr. ; Peralta, J. E. ; Ogliaro, F. ; Bearpark, M. J. ; Heyd, J. J. ; Brothers, E. N. ; Kudin, K. N. ; Staroverov, V. N. ; Keith, T. A. ; Kobayashi, R. ; Normand, J. ; Raghavachari, K. ; Rendell, A. P. ; Burant, J. C. ; Iyengar, S. S. ; Tomasi, J. ; Cossi, M. ; Millam, J. M. ; Klene, M. ; Adamo, C. ; Cammi, R. ; Ochterski, J. W. ; Martin, R. L. ; Morokuma, K. ; Farkas, O. ; Foresman, J. B. ; Fox, D. J. . Gaussian 16, Revision C.01; Gaussian Inc.: Wallingford CT, 2016. [Google Scholar]
  40. Wang J., Wolf R. M., Caldwell J. W., Kollman P. A., Case D. A.. Development and Testing of a General Amber Force Field. J. Comput. Chem. 2004;25:1157–1174. doi: 10.1002/jcc.20035. [DOI] [PubMed] [Google Scholar]
  41. Salomon-Ferrer R., Case D. A., Walker R. C.. An Overview of the Amber Biomolecular Simulation Package. WIREs Comput. Mol. Sci. 2013;3(2):198–210. doi: 10.1002/wcms.1121. [DOI] [Google Scholar]
  42. Hess B., Bekker H., Berendsen H. J. C., Fraaije J. G. E. M.. LINCS: A Linear Constraint Solver for Molecular Simulations. J. Comput. Chem. 1997;18(12):1463–1472. doi: 10.1002/(SICI)1096-987X(199709)18:12<1463::AID-JCC4>3.0.CO;2-H. [DOI] [Google Scholar]
  43. Bussi G., Donadio D., Parrinello M.. Canonical Sampling through Velocity Rescaling. J. Chem. Phys. 2007;126(1):014101. doi: 10.1063/1.2408420. [DOI] [PubMed] [Google Scholar]
  44. Bussi G., Parrinello M.. Stochastic Thermostats: Comparison of Local and Global Schemes. Comput. Phys. Commun. 2008;179(1–3):26–29. doi: 10.1016/j.cpc.2008.01.006. [DOI] [Google Scholar]
  45. Bussi G., Zykova-Timan T., Parrinello M.. Isothermal-Isobaric Molecular Dynamics Using Stochastic Velocity Rescaling. J. Chem. Phys. 2009;130(7):074101. doi: 10.1063/1.3073889. [DOI] [PubMed] [Google Scholar]
  46. Berendsen H. J. C., Postma J. P. M., van Gunsteren W. F., Dinola A., Haak J. R.. Molecular Dynamics with Coupling to an External Bath. J. Chem. Phys. 1984;81(8):3684–3690. doi: 10.1063/1.448118. [DOI] [Google Scholar]
  47. Van Der Spoel D., Lindahl E., Hess B., Groenhof G., Mark A. E., Berendsen H. J. C.. GROMACS: Fast, Flexible, and Free. J. Comput. Chem. 2005;26:1701–1718. doi: 10.1002/jcc.20291. [DOI] [PubMed] [Google Scholar]
  48. Harada R., Shigeta Y.. On-the-Fly Specifications of Reaction Coordinates in Parallel Cascade Selection Molecular Dynamics Accelerate Conformational Transitions of Proteins. J. Chem. Theory Comput. 2018;14(6):3332–3341. doi: 10.1021/acs.jctc.8b00264. [DOI] [PubMed] [Google Scholar]
  49. Hengphasatporn K., Harada R., Wilasluck P., Deetanya P., Sukandar E. R., Chavasiri W., Suroengrit A., Boonyasuppayakorn S., Rungrotmongkol T., Wangkanont K., Shigeta Y.. Promising SARS-CoV-2 Main Protease Inhibitor Ligand-Binding Modes Evaluated Using LB-PaCS-MD/FMO. Sci. Rep. 2022;12(1):17984. doi: 10.1038/s41598-022-22703-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Harada R., Morita R., Shigeta Y.. Free-Energy Profiles for Membrane Permeation of Compounds Calculated Using Rare-Event Sampling Methods. J. Chem. Inf. Model. 2023;63(1):259–269. doi: 10.1021/acs.jcim.2c01097. [DOI] [PubMed] [Google Scholar]
  51. Munei Y., Hengphasatporn K., Hori Y., Harada R., Shigeta Y.. Determination of the Association between Mesotrione Sensitivity and Conformational Change of 4-Hydroxyphenylpyruvate Dioxygenase via Free-Energy Analyses. J. Agric. Food Chem. 2023;71(24):9528–9537. doi: 10.1021/acs.jafc.3c01253. [DOI] [PubMed] [Google Scholar]
  52. Harada R., Nakamura T., Shigeta Y.. Automatic Detection of Hidden Dimensions to Obtain Appropriate Reaction Coordinates in the Outlier FLOODing (OFLOOD) Method. Chem. Phys. Lett. 2015;639:269–274. doi: 10.1016/j.cplett.2015.09.031. [DOI] [Google Scholar]
  53. Harada R., Nakamura T., Shigeta Y.. A Fast Convergent Simulated Annealing Algorithm for Protein-Folding: Simulated Annealing Outlier Flooding (Sa-Oflood) Method. Bull. Chem. Soc. Jpn. 2016;89(11):1361–1367. doi: 10.1246/bcsj.20160244. [DOI] [Google Scholar]
  54. Nakamura T., Kamidoi Y., Wakabayashi S., Yoshida N.. FlexDice: A Fast Clustering Method for Large High Dimensional Data Sets. IPSJ Trans. Databases. 2005;46(18):40–49. [Google Scholar]
  55. Chodera J. D., Noé F.. Markov state models of biomolecular conformational dynamics. Curr. Opin. Struct. Biol. 2014;25:135–144. doi: 10.1016/j.sbi.2014.04.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Scherer M. K., Trendelkamp-Schroer B., Paul F., Pérez-Hernández G., Hoffmann M., Plattner N., Wehmeyer C., Prinz J. H., Noé F.. PyEMMA 2: A Software Package for Estimation, Validation, and Analysis of Markov Models. J. Chem. Theory Comput. 2015;11(11):5525–5542. doi: 10.1021/acs.jctc.5b00743. [DOI] [PubMed] [Google Scholar]
  57. McGibbon R. T., Beauchamp K. A., Harrigan M. P., Klein C., Swails J. M., Hernández C. X., Schwantes C. R., Wang L. P., Lane T. J., Pande V. S.. MDTraj: A Modern Open Library for the Analysis of Molecular Dynamics Trajectories. Biophys. J. 2015;109(8):1528–1532. doi: 10.1016/j.bpj.2015.08.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Massova, I. ; Kollman, P. A. . Combined Molecular Mechanical and Continuum Solvent Approach (MM-PBSA/GBSA) to Predict Ligand Binding; Kluwer/Escom, 2000; Vol. 18. [Google Scholar]
  59. Valdés-Tresanco M. S., Valdés-Tresanco M. E., Valiente P. A., Moreno E.. Gmx_MMPBSA: A New Tool to Perform End-State Free Energy Calculations with GROMACS. J. Chem. Theory Comput. 2021;17(10):6281–6291. doi: 10.1021/acs.jctc.1c00645. [DOI] [PubMed] [Google Scholar]
  60. Crowe C., Nakasone M. A., Chandler S., Tatham M. H., Makukhin N., Hay R. T., Ciulli A.. et al. Mechanism of Degrader-Targeted Protein Ubiquitinability. Sci. Adv. 2024;10(41):eado6492. doi: 10.1126/sciadv.ado6492. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Kumar H., Sobhia M. E.. Interplay of PROTAC Complex Dynamics for Undruggable Targets: Insights into Ternary Complex Behavior and Linker Design. ACS Med. Chem. Lett. 2024;15(8):1306–1318. doi: 10.1021/acsmedchemlett.4c00189. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ci5c00102_si_001.pdf (902.9KB, pdf)

Data Availability Statement

3D structures of the proteins were downloaded from the Protein Data Bank (PDB). We used AmberTools20 and Gaussian 16 Rev C.01 for PROTAC preparation. AmberTools20 was used to prepare the MD system. GROMACS 2021.5 was used as the MD engine. Rms module in GROMACS was used to calculate Cα-RMSD. Alpha Shape Toolbox was used to calculate the α-shape area. MDTraj was used to calculate R g and the contact map. Prime module in Schrödinger Suite 2023–3 was used to calculate strain energy. PyEMMA was used to estimate MSM. Gmx_MMPBSA was used to calculate binding free-energy based on the MM/GBSA method. PyMOL was used for visualization. Scripts of our methodology and these conformational search results are deposited to Zenodo (DOI: 10.5281/zenodo.15307772).

Articles from Journal of Chemical Information and Modeling are provided here courtesy of American Chemical Society

RESOURCES