Skip to main content
NCBI home page
ACS Omega logoLink to ACS Omega
. 2025 Sep 26;10(44):52585–52597. doi: 10.1021/acsomega.5c05849

Proposal of Chemical Inhibitors That Compete with the Binding of RNA Polymerase II Subunits to Essential GTPases GPN Npa3 and Gpn1

Julio A Muñiz-Luna †, Ángel Santiago §, Gema R Cristóbal-Mondragón ∥, Mónica R Calera †,*, Roberto Sánchez-Olea †,*, Nina Pastor ‡,*
PMCID: PMC12612918  PMID: 41244487

Abstract

The Gpn1, Gpn2, and Gpn3 proteins are members of the GTPase GPN family; yeast Npa3 is an orthologue of the human Gpn1 protein. These proteins play a crucial role in the nuclear accumulation of RNA polymerase II, functioning as molecular chaperones. The crystallographic structures of Npa3 reveal open and closed conformations, which are dependent on the bound guanine nucleotide (GMPPCP or GDP, respectively). The open conformation of Npa3 exhibits a hydrophobic pocket proposed to be essential for the recognition and binding of specific peptides of RNA polymerase II, contributing to its biogenesis; however, structural data on these complexes remain unavailable. In this work, we present in silico models of the interactions between the crystallographic structure of monomeric Npa3 in its open conformation and yeast RNA polymerase II peptides, generated through flexible computational docking. To identify inhibitors of these interactions, potentially useful in understanding the molecular and cellular functions of these proteins, we performed molecular docking experiments using a designed library of FDA-approved compounds on both the Npa3 structure and a homology model of human Gpn1. Our analysis identified potential inhibitors, including atovaquone for both Npa3 and Gpn1 (docking scores: −14.4 and −13.5 kcal/mol, respectively) and tibolone for Npa3 (−13.6 kcal/mol), following flexible docking optimization. Additionally, our docking models suggest key residues in Npa3 such as F143 and W179, which may be critical for recognizing RNA polymerase II subunits and drug-like molecules. These findings can be further explored through biochemical and mutagenesis studies to assess their roles in RNA polymerase II recognition.

graphic file with name ao5c05849_0011.jpg

graphic file with name ao5c05849_0009.jpg

1. Introduction

GTPases constitute a superfamily of enzymes with the ability to bind guanine nucleotides and hydrolyze GTP, acting as molecular switches and transitioning between inactive (GDP-bound) and active (GTP-bound) conformations. This transition is regulated by guanine nucleotide exchange factors (GEFs) and GTPase activating proteins (GAPs). − GEFs facilitate the exchange of GDP for GTP, while GAPs promote the hydrolysis of GTP to GDP, leading to an inactive state of the GTPase. − GTPases share a common structural domain consisting of 170 residues distributed across five motifs (G1 to G5), crucial for GTP binding and hydrolysis. , During GTP hydrolysis, the γ phosphate is attacked by a nucleophilic water molecule, causing its release, resulting in the GTPase inactive state (GDP-bound). Additionally, there is an apo state, which is an alternative conformation with a disordered switch 1 region. In 2007, Gras et al. proposed the GTPase GPN-loop family, characterized by a highly conserved motif comprising glycine (G), proline (P), and asparagine (N) residues, located in the G domain, essential for stabilizing the phosphate ion posthydrolysis. This family includes human Gpn1, Gpn2, and Gpn3, while Npa3 in yeast (Saccharomyces cerevisiae) is the orthologue of Gpn1. In archaea, there is a single copy of the GPN gene. The three GPN genes are universally conserved in eukaryotic cells and indispensable for life. , Interactions among Gpn1, Gpn2, and Gpn3, leading to heterodimer formation, contribute to their functional roles. , The crystallographic structure of monomeric Npa3, reported in 2016, reveals two conformations depending on the bound guanine nucleotide: a closed form (GDP-bound) and an open form (GTP-bound, crystallized with a nonhydrolyzable analog of GTP). In addition, the open conformation of Npa3 exhibits an exclusive hydrophobic pocket with a cocrystallized laurate fatty acid molecule, which is not present in the closed conformation.

Npa3/Gpn1 plays a pivotal role in the nuclear accumulation of RNA polymerase II, the enzyme that transcribes all protein-coding genes and some noncoding RNA genes, such as miRNA, lncRNA, and snRNA. The structure of RNA polymerase II, a complex consisting of 12 subunits (Rpb1-Rpb12), has been extensively studied. ,,, The catalytic core, composed of 10 subunits, is organized into three subassemblies: the Rpb1 subassembly, which includes Rpb1, Rpb5, Rpb6, and Rpb8 subunits; the Rpb2 subassembly, formed by Rpb2 and Rpb6; and finally, the Rpb3 subassembly containing Rpb3, Rpb10, Rpb11, and Rpb12. The formation of these multisubunit complexes also requires intermediates and assembly factors. , A model proposed by Niesser et al. suggests that Npa3 can function as a molecular chaperone. This observation stems from qualitative binding assays that indicate that Npa3 interacts through its hydrophobic pocket with key peptides located at the interfaces of RNA polymerase II subunits. These interactions were described to promote the GDP to GTP exchange, proposing an allosteric regulation of nucleotide binding and GTPase activity.

However, understanding the interactions between the hydrophobic pocket of Npa3 and these interface peptides is hampered by a lack of structural information on the complexes. In this sense, computational modeling represents a valuable tool for understanding molecular interactions relevant for protein binding, as well as for the development and identification of bioactive compounds. − The requirements for searching for small molecules with biological activity have been extensively described by various authors, − emphasizing the importance of a detailed structural analysis of the receptor, the study of pocket properties, and characterization of those with potentially druggable properties. Computational techniques such as molecular docking are defined as essential tools for predicting models of the interactions between small ligands and macromolecules, as well generating protein–protein or protein–peptide complexes, ,− offering potential information about specific contacts, interaction energy, binding distances, etc. In this study, we explored the hydrophobic pocket in the open form of Npa3 and human Gpn1 (GTP-bound) in order to obtain information about its structural properties. With this information we performed flexible-type protein–peptide and protein–protein docking against RNA polymerase II peptides identified by Niesser et al. as being strong binders (Table ), in order to model these interactions in silico, providing information on the particular binding modes and specific contacts. To investigate potential chemical inhibitors that affect the function of these essential GTPases and to advance the understanding of their biological mechanisms, we conducted molecular docking experiments in this hydrophobic pocket of Npa3 and the homology model of human Gpn1 against a library of FDA-approved drugs. Finally, we propose potential inhibitors for future experimental evaluation as well as models for the recognition of the interaction of these GTPases with RNA polymerase II peptides. Given the biological relevance of Npa3, in combination with the existence of the hydrophobic pocket, its interaction with the peptides from the interfaces of RNA polymerase II subunits, and the limited understanding of its activity mechanism, Npa3 emerges as an ideal target in the search for potential chemical inhibitors able to affect its chaperone activity by disrupting the interactions with key peptides of RNA polymerase II subunits, preventing its assembly.

1. Strongest RNA Polymerase II-Interacting Peptides with Npa3 .

RNA polymerase II subunit peptide sequence peptide number position in the complete subunit ID UniProt/AlphaFold
Rpb1 FGHIDLAKPVFHVGF 21 81–95 P04050
KRIAFGFVDRTLPHF 194 773–787
ENSYLRGLTPQEFFF 201 801–815
YKQLVKDRKFLREVF 234 933–947
Rpb4 KNTMQYLTNFSRFRD 855 142–156 P20433
Rpb8 LNNLKQENAYLLIRR 1043 132–146 P20436
Rpb11 FAAYKVEHPFFARFK 1109 58–72 P38902
Open in a new tab

2. Results

2.1. Open Npa3 and Human Gpn1 Models Exhibit a Hydrophobic Pocket with Pharmacological Potential

The crystallographic structure of Npa3 and the human Gpn1 model in the open conformation (Figure A,B) display the GTPase core fold, characterized by a six-stranded parallel β-sheet surrounded by six α-helices. Notably, the hydrophobic pocket (laurate binding site) is conserved (Figures A and S1), suggesting the potential for targeting this conserved site across various species. The surface representation according to hydrophobicity for both the Npa3 and the human Gpn1 model structures are shown in Figure C,D, respectively. This pocket is predominantly composed of hydrophobic residues.

1.

1

Open in a new tab

3D structure of essential GTPases Npa3 and the homology model of human Gpn1 in the open conformation. Cartoon representation of the Npa3 crystal from S. cerevisiae (UniProt: P47122; PDB ID: 5HCN) and of the human Gpn1 model (UniProt: Q9HCN4) in the open conformation (GTP-bound, active form) are displayed in sand (A) and blue (B), respectively. Laurate and GMPPCP (nonhydrolyzable analogue of GTP) are displayed as spheres. The hydrophobic (gold) and hydrophilic (cyan) surface representation for Npa3 and Gpn1 is shown in (C, D) in the same orientation as in (A, B), respectively, with the scale bar for the molecular lipophilicity potential (MLP). The magnified area corresponds to the laurate binding site location.

To obtain information regarding the localization and properties of this pocket in open Npa3 and Gpn1, we used the bioinformatics tool DoGSiteScorer. , Notably, the pocket corresponding to the laurate binding site exhibited druggability scores of 0.72 and 0.8 for Npa3 and Gpn1, respectively; values closer to 1 suggest a high probability for acting as a ligand binding site. This score is based on properties such as the pocket volume, surface area, local hydrophobicity, and shape. Figure displays the hydrophobic pocket structure and the surrounding residues. This pocket in the Npa3 structure has a volume of 405.89 A3, as well as a 0.71/0.29 ratio of nonpolar and polar residues. All hydrophobic pocket descriptors for the Npa3 structure and the Gpn1 model are shown in Table S1. These structural descriptors suggest that the hydrophobic pocket could be a potential druggable target for in-depth study.

2.

2

Open in a new tab

Localization and constituent residues of the “laurate binding site” identified by DoGSiteScorer in the Npa3 structure. (A) Volume of the hydrophobic pocket calculated using the DoGSiteScorer tool shown as a green mesh, with laurate located internally. (B) Constituent residues of the hydrophobic pocket identified by DoGSiteScorer and their positions in relation to laurate.

2.2. Peptides in the Rpb1, Rpb4, Rpb8, and Rpb11 Subunits Interface Interact In Silico with Open Npa3 through Constituent Residues of the Hydrophobic Pocket

To evaluate the interactions between the Npa3 hydrophobic pocket and peptides from RNA polymerase II subunits displayed in Table , models were generated using the CABS-dock and HADDOCK 2.4 servers. The position of each peptide within the context of the complete subunit is displayed in Figure and the AlphaFold2 confidence scores for the predicted structures are shown in Figure S2. Ranking of models generated by CABS-dock and statistics of the HADDOCK score for the clusters generated for each peptide in its respective subunit are presented in Table S2 and Figure S3, respectively.

3.

3

Open in a new tab

Sequence and localization of peptides used for protein–protein flexible docking. In surface representation are the 15 amino acid peptides from Rpb1 (gray), Rpb4 (purple), Rpb8 (green), and Rpb11 (orange) subunits involved in the interaction with Npa3: (A) Peptides 21, 194, 201, and 234 of Rpb1, (B) 855 of Rpb4, (C) 1043 of Rpb8, and (D) 1109 of Rpb11 are highlighted in red. The position of each peptide within the context of the complete subunit models obtained from AlphaFold2, is indicated below its respective sequence.

To evaluate the most prominent interactions in the models generated using the CABS-dock and HADDOCK programs, residue–residue contacts were calculated. Figure presents heatmaps of the consensus residue–residue contacts from the best models, showing the interactions between all selected peptides from the Rpb1, Rpb4, Rpb8, and Rpb11 subunits and the residues comprising the Npa3 hydrophobic pocket. Using a 6 Å cutoff threshold to define carbon contacts, a total of 793 contacts were identified with CABS-dock, 144 with HADDOCK, and 101 shared contacts, i.e., those detected by both HADDOCK and CABS-dock.

4.

4

Open in a new tab

Residue contact heatmap from flexible docking models using CABS-dock and HADDOCK servers. Specific interactions between residues conforming to the hydrophobic pocket of open Npa3 with peptides 21 (A), 194 (B), 201 (C), and 234 (D) from Rpb1; 855 (E) from Rpb4; 1043 (F) from Rpb8; and 1109 (G) from Rpb11. Contacts present in CABS-dock models are shown in cyan, contacts in HADDOCK models are shown in red, and contacts shared in models from both servers are colored in green. A cutoff threshold of 6 Å was used to define the carbon–carbon contacts.

Among these interactions, peptide 21 showed the lowest number of shared contacts (in both CABS-dock and HADDOCK): F143 of Npa3 is paired with A87 of Rpb1 (Figure A). In contrast, peptide 194 in Rpb1 exhibited 24 shared contacts with Npa3 (Figure B), being the second highest. Peptide 1109 from Rpb11 showed the highest number of shared contacts, totaling 25 (Figure G). Considering the shared contacts, interestingly, F143 is the only Npa3 residue participating in the interaction with all the studied subunits, both in the context of the complete protein and at the individual peptide level. It participates in a total of 27 shared interactions, based on the evaluation of all peptides, contributing 26.7% of the total shared contacts. On the other hand, it shows 85 contacts according to HADDOCK, representing 24.3% of all the interactions proposed in the context of the complete subunits. In addition, this residue provides the highest number of contacts with a single subunit, a total of eight, including residues F777, G778, F779, V780, R782, L784, H786, and F787 of peptide 194 in Rpb1.

W179 residue in the Npa3 hydrophobic pocket showed the highest participation in shared interactions only behind F143, with a total of 13, establishing contact with all peptides except for 21 in Rpb1. In addition, it contributed the most interactions in the CABS-dock-generated docking, representing 90 of the 793 in total. These findings indicate that F143 and W179 residues may be crucial for the recognition of RNA polymerase II subunits by Npa3.

2.3. Identification of Compounds Docked to the Hydrophobic Pocket of Npa3 and Gpn1

To identify compounds that can function as potential inhibitors of Npa3-RNA polymerase II interactions, as an initial step we performed rigid docking of FDA-approved drugs over the Npa3 laurate binding pocket using AutoDock Vina. Previous validation by redocking laurate was performed. AutoDock Vina was able to reproduce the binding orientation of laurate cocrystallized with open Npa3 (Figure A), resulting in two hydrogen bonds from N137 and W179 with a laurate carboxylate oxygen (Figure B). In the crystal structure, the same two hydrogen bonds are observed between W179 and N137, interacting with a carboxylate oxygen in the fatty acid moiety from laurate. The binding distances for these interactions are 3.08 and 3.0 Å, respectively. In the redocking pose, these distances were 2.96 and 3.17 Å, respectively. F143 displayed hydrophobic interactions with laurate, suggesting that both F143 and W179 are important for ligand recognition. The resulting energy from redocking laurate into Npa3 was −6.1 kcal/mol, whereas that obtained from the Vina rescoring function for the crystallographic structure was −4.9 kcal/mol.

5.

5

Open in a new tab

Comparison of laurate binding modes: AutoDock Vina redocking against crystallographic structure of Npa3. (A) The structure of laurate is depicted in blue and green sticks, overlying the Npa3 surface for the crystallographic and redocking pose, respectively. (B) The specific 2D interaction diagrams are shown below.

The results of rigid docking for the FDA-approved compound library with the Npa3 and Gpn1 are shown in Figures S4 and S5, respectively, yielding a total of 8,930 model structures. Based on the docking results, the top five compounds with the best scores were selected for analysis. Specific interactions for the top five compounds were evaluated, including trifluperidol, novobiocin, atovaquone, prednisolone, and tibolone (Figure ). The specific interactions between Npa3 and the five compounds with the best energy are also listed in Table . Novobiocin presented the maximum number of hydrophobic interactions with Npa3, with a total of 13. Nine of these interactions were residues of the hydrophobic pocket (F212, L209, M208, F186, F183, D182, C151, M147, and S138), including W179 and F143 as in the CABS-dock and HADDOCK models. It also engages in one hydrogen bond with N137. The interaction energy calculated by AutoDock Vina for this interaction was −7.9 kcal/mol. Notably, trifluperidol was able to dock inside the hydrophobic pocket with an orientation similar to that of laurate, binding to Npa3 predominantly through hydrophobic interactions with an energy of −7.9 kcal/mol. It presented a total of 11 interactions, involving 10 residues from the hydrophobic pocket (Table ).

6.

6

Open in a new tab

Binding modes from rigid docking and 2D interaction diagrams for the specific contacts between open Npa3 and the five compounds with the best binding energy. Npa3 is shown in sand-colored surface representation, (A) trifluperidol compound docked inside the hydrophobic pocket is shown in yellow, (B) novobiocin shown in pink, (C) atovaquone shown in dark gray, (D) prednisolone in turquoise, and (E) tibolone in orange. For the 2D interaction diagrams, each compound representation corresponds to the assigned color in the 3D view.

2. Top 5 Compounds Docked with Open Npa3 (Rigid Docking).

ZINC ID ligand ΔG (kcal/mol) hydrogen bonding hydrophobic interactions
ZINC00538505 trifluperidol –7.9   L216, F212, L209, F186, F183, W179, V163, A150, M147, N146, and V132
ZINC03831234 novobiocin –7.9 N137 L216, F212, L209, M208, F186, F183, D182, W179, C151, M147, F143, S138, and N137
ZINC12504271 atovaquone –7.9   F186, F183, D182, W179, C151, L148, M147, and F143
ZINC03831370 prednisolone acetate –7.8 W179 and D182 F186, F183, D182, W179, F175, L172, M147, and F143
ZINC11616424 tibolone –7.8 W179 F186, F183, D182, and W179
Open in a new tab

The top five compounds with the best binding energy for the Gpn1 model include ketanserin, gliquidone, dicumarol, nefazodone, and atovaquone (Figure ). The specific interactions among the five compounds analyzed and Gpn1 are included in Table . Among these, gliquidone established the highest number of interactions, totaling 18. The interacting residues include L238, F234, L231, V230, M227, S226, L223, W202, M188, V186, F184, L177, C174, A173 M170, N169, M155, and Y153. Gliquidone docks inside the hydrophobic pocket using the sulfonylurea group, while the remaining molecule extends outside the pocket. Notably, gliquidone displays interactions mapping the portion of hydrophobic pocket such as L238, F234, L231, V230, W202, M188, V186, F184, L177, C174, A173, M170, and N169. For ketanserin, the 2-[4-(4-(p-fluorobenzoyl) piperidin-1-yl)­ethyl] group was inserted inside the pocket, while the quinazoline group was located at the pocket entrance. It exhibits a total of 16 hydrophobic interactions, involving L238, F234, L231, V230, F206, W202, M188, V186, C174, A173, M170, N169, F166, S160, M155, and Y153. It is noteworthy that only the last two are not considered components of the pocket. Dicumarol, on the other hand, was completely bound inside the pocket, involving 10 hydrophobic interactions with L238, F234, F206, W202, M188, V186, A173, M170, N169, and F166, all residues belonging to the hydrophobic pocket. Similarly, atovaquone displayed hydrophobic interactions with external residues of the hydrophobic pocket in Npa3 and Gpn1. Eight contacts were identified with Npa3 and nine with Gpn1. Atovaquone exhibits common interactions for rigid docking in both structures, involving residues F143, C151, W179 and F183 in Npa3, and its equivalents F166, C174, W202 and F206 in Gpn1.

7.

7

Open in a new tab

Binding modes from rigid docking and 2D interaction diagrams for the specific contacts between the Gpn1 model in open conformation and the five compounds with the best binding energy. The Gpn1 model is displayed in blue surface representation, while the docking of (A) ketanserin is in white, (B) gliquidone is in lime, (C) dicumarol is in cyan, (D) nefazodone is in yellow, and (E) atovaquone is in dark orange. For the 2D interaction diagrams, each compound representation corresponds to the assigned color in the 3D view.

3. Top 5 Compounds Docked with the Open Gpn1 Model (Rigid Docking).

ZINC ID ligand ΔG (kcal/mol) hydrophobic interactions
ZINC00537877 ketanserin –9.3 L238, F234, L231, V230, F206, W202, M188, V186, C174, A173, M170, N169, F166, S160, M155, and W153
ZINC01482077 gliquidone –9.1 L238, F234, L231, V230, M227, S226, L223, W202, M188, V186, F184, L177, C174, A173 M170, N169, M155, and W153
ZINC03869855 dicumarol –8.8 L238, F234, F206, W202, M188, V186, A173, M170, N169, and F166
ZINC00538065 nefazodone –8.6 L238, Y235, F234, L231, F209, F206, D205, W202, M188, V186, C174, A173, M170, M155, and Y153
ZINC12504271 atovaquone –8.6 L231, F209, F206, D205, W202, E201, F198, F166, and S160
Open in a new tab

Ketanserin, atovaquone, and dicumarol displayed interactions with F166 and W202 residues from the Gpn1 model, which are the equivalent residues F143 and W179 in the Npa3 structure, and they also were representative residues in all the analyzed models from CABS-dock and HADDOCK.

To account for the flexibility of protein, we conducted an additional refinement step of flexible docking. Binding energies resulting from flexible docking for trifluperidol, novobiocin, atovaquone, prednisolone, and tibolone on Npa3; and for ketanserin, gliquidone, dicumarol, nefazodone, and atovaquone with Gpn1 are shown in Table . In all cases flexible docking improves the energy score, suggesting better interactions of each compound with the proteins. Interestingly, the compound that exhibited the highest interaction energy in both Npa3 and Gpn1 was atovaquone, resulting in −14.4 and −13.5 kcal/mol for Npa3 and Gpn1, respectively. Figure details the specific interactions between the two proteins and atovaquone as well as the changes in the position of the side chains for the flexible residues. As can be appreciated in Figure , the improved energy score for atovaquone is related to a more relaxed position of the side chains in the hydrophobic residues.

4. Interaction Energies Generated by Flexible Protein Docking for the Five Compounds with the Best Rigid Docking Energies, Npa3 and Gpn1.

Npa3
 Gpn1
ZINC ID drug name ΔG (kcal/mol) ZINC ID drug name ΔG (kcal/mol)
ZINC12504271 atovaquone –14.4 ZINC12504271 atovaquone –13.5
ZINC11616424 tibolone –13.6 ZINC03869855 dicumarol –12.4
ZINC00538505 trifluperidol –10.8 ZINC01482077 gliquidone –11.3
ZINC03831370 prednisolone –9.6 ZINC00537877 ketanserin –10.3
ZINC03831234 novobiocin –8.7 ZINC00538065 nefazodone –9.4
Open in a new tab

8.

8

Open in a new tab

Flexible docking of atovaquone inside the hydrophobic pocket of open Npa3 and Gpn1. Atovaquone is displayed as spheres in salmon color. (A) For docking on the Npa3 structure, the side chains of residues indicated as flexible are colored in sand and are shown as sticks; the position after docking is shown in purple. For docking on the Gpn1 model in (B), the side chains established as flexible are shown in royal blue, whereas in its postdocking position they are colored in cyan. Shown to the right of each 3D representation are the 2D interaction diagrams. Hydrophobic interactions between the flexible Npa3 and Gpn1 model residues, with atovaquone following flexible docking, are indicated in red tabs.

Given the similarity in the pockets of Npa3 and Gpn1, it was surprising that only atovaquone was shared as the best ligand. To explore further the equivalence of the binding pocket in both proteins, we carried out flexible docking of the nonshared top five compounds in the other protein, with the same docking protocol. The resulting energies are presented in Table .

5. Cross-Docking Interaction Energies of Top Compounds from Flexible Docking between Npa3 and Gpn1.

Npa3
 Gpn1
ZINC ID drug name ΔG (kcal/mol) ZINC ID drug name ΔG (kcal/mol)
ZINC03869855 dicumarol –12 ZINC11616424 tibolone –12.6
ZINC01482077 gliquidone –9.9 ZINC00538505 trifluperidol –12
ZINC00537877 ketanserin –13 ZINC03831370 prednisolone –9.5
ZINC00538065 nefazodone –10.2 ZINC03831234 novobiocin –9.4
Open in a new tab

These ligands exhibit comparable binding energies in both proteins; the largest differences are found for trifluperidol (1.2 kcal/mol better for Gpn1), gliquidone (1.4 kcal/mol better for Gpn1), and ketanserin (2.7 kcal/mol better for Npa3). Furthermore, the binding energies are within the range of those in Table . Taken together, these results suggest that both binding pockets are equivalent and that these nine compounds are good candidates as inhibitors for either protein.

3. Discussion

The essential GTPases human Gpn1 and its yeast orthologue Npa3 in S. cerevisiae play a crucial role in the nuclear accumulation of RNA polymerase II. ,,, Despite their importance, the precise mechanism underlying their functions remains elusive. Npa3 in the open conformation features a hydrophobic pocket (“laurate binding site”) proposed to be relevant for the binding of peptides from RNA polymerase II subunits in biochemical experiments. In this work, we use a combination of computational techniques to evaluate the possible interactions of Npa3 and Gpn1 GTPases with RNA polymerase II subunit peptides experimentally evaluated by Niesser et al. and propose potential inhibitors of their RNA polymerase binding function, for future experimental evaluation.

The availability of the Npa3 crystallographic structure allowed the generation of a human Gpn1 model that revealed the conservation of this hydrophobic pocket (Figures and S1). Pocket analysis with DoGSiteScorer shows that the pocket containing the cocrystallized laurate fatty acid, i.e., the hydrophobic pocket proposed as the peptide binding site, has the highest volume and score in both Npa3 and Gpn1 models (Figure and Table S1). In addition, its druggability score is particularly important, considering that a higher druggability score is associated with a higher pharmacological potential of the pocket. In this sense, the druggability value was considered acceptable and was an indicator that the pocket exhibits advantageous characteristics, including an appropriate volume, accessibility, and physicochemical properties that make it suitable for ligand binding. Additionally, the laurate molecule cocrystallized in the open conformation of Npa3 structure is particularly crucial because it acts as a pocket marker, i.e., it binds to a potentially pharmacological site in the protein, correlating with the DoGSiteScorer results.

On the other hand, the CABS-dock and HADDOCK servers were a valuable instrument to determine the specific contacts between Npa3 and the peptides of the Rpb1, Rpb4, Rpb8, and Rpb11 subunits. This information is especially useful, given that no three-dimensional models are available for these interactions, which could be essential for the nuclear import of the RNA polymerase II complex. Flexible protein–peptide and protein–protein docking was performed with CABS-dock and HADDOCK 2.4 servers to evaluate the interaction of RNA polymerase II peptides with Npa3 through its hydrophobic pocket as individual peptide and in the conformations found in the full-length and folded subunits obtained from AlphaFold2 (Figures and S2). Residues most frequently involved in the interactions modeled by both servers were F143, W179, and N137 (Figure ). Overall, the number of interactions generated by individual peptides in CABS-dock models was approximately 5.5 times larger than those observed in HADDOCK, i.e., in the context of full-length subunits. These differences can be attributed to better accessibility of Npa3 to the subunit peptides in CABS-dock models, in contrast to the fully assembled RNA polymerase II subunits in HADDOCK models. Particularly for Rpb1 peptide 21 contacts, the best HADDOCK clusters revealed a single interaction between F143 of Npa3 and A87 in Rpb1; the limited interactions observed were a consequence of the presence of two α-helices located around the peptides defined as active for docking, decreasing the accessibility for Npa3. F143 contributes approximately a quarter of all interactions defined in consensus by CABS-dock and HADDOCK, representing the most crucial interaction in all HADDOCK docking results. F143 also participates in laurate binding in the crystal (Figure ). In our docking studies in the Npa3 structure, this residue shows participation in the binding of chemical compounds, such as novobiocin, atovaquone, prednisolone, and tibolone (Figure and Table ). For Gpn1, the equivalent residue corresponds to F166 (Figure S1) and participates in the interaction with ketanserin, dicumarol, and atovaquone (Figure and Table ). W179, in turn, is the second highest participant in the interactions shared by CABS-dock and HADDOCK (Figure ). In the Npa3 structure and Gpn1 model, it establishes a hydrogen bond with the laurate carboxylate, contributing to its binding and stabilization inside the hydrophobic pocket. In the docking of FDA-approved drugs, W179 (W202 in Gpn1) was involved in the binding of trifluperidol, novobiocin, prednisolone, and tibolone in Npa3, through hydrophobic interactions and hydrogen bonds (Figures and ). In Gpn1 its equivalent residue, W202, interacts hydrophobically with ketanserin, gliquidone, dicumarol, nefazodone, and atovaquone. Based on this information, we consider that residues F143 and W179 from the hydrophobic pocket in Npa3 could play a critical role in the recognition of RNA polymerase II subunits and drug-like molecules. These residues could be proposed to be evaluated by mutagenesis experiments to understand their role in the Npa3 function.

Other residues involved in the interaction of Npa3 with the evaluated ligands and the peptides at the interface of the RNA polymerase II subunits include F212 which contributes to the binding of trifluperidol and novobiocin and also interacts with peptides 201 and 234 in Rpb1 (Figures and ). Also, F183 participates in all evaluated interactions, with the above-mentioned exception of peptide 21, binding to Rpb1, Rpb4, Rpb8, and Rpb11, as well as to the top five compounds in the library.

In this study the presence of laurate cocrystallized with Npa3 inside the pocket was taken as a guide to validate the docking protocol with AutoDock Vina. The capability of AutoDock Vina to reproduce the laurate orientation inside the hydrophobic pocket indicates that AutoDock Vina was suitable for ligand docking of the FDA-approved drugs in Npa3 and human model Gpn1 (Figure ).

The rigid docking of the FDA-approved compounds resulted in novobiocin docked in the external portion of the hydrophobic pocket (Figure B), having the highest energy together with trifluperidol and atovaquone (Figure A,C, respectively). Novobiocin belongs to the aminocoumarin family and is an inhibitor of bacterial DNA synthesis by acting on DNA gyrase B. A precedent involving interaction between novobiocin and a molecular chaperone has been reported previously. They showed that this antibiotic binds to the C-terminal domain of the heat shock protein Hsp90, interfering with the association of the cochaperones Hsc70 and p23, causing its inhibition. In a previous work of the same group, it was reported that novobiocin binds to the ATP recognition site on gyrase B, competing with nucleotide binding. Furthermore, Hocker et al. demonstrated by molecular docking with AutoDock Vina, molecular dynamics simulations, and in vitro experimentation that the bicyclic diterpenoid lactone, andrographolide and its derivatives, bind to pockets in the small GTPase K-Ras, inhibiting GDP-GTP exchange and consequently its oncogenic signal.

Finally, with respect to flexible docking, we found a global increase in the interaction energies of trifluperidol, novobiocin, atovaquone, prednisolone, and tibolone on Npa3; and ketanserin, gliquidone, dicumarol, nefazodone, and atovaquone for the human Gpn1 model (Table and Figure ). These energy changes are probably the result of the induced fit allowed by flexible docking in contrast to rigid docking.

The flexible docking results (Table ) revealed an overlap between Npa3 and its human homologue Gpn1, with atovaquone consistently ranking among the top binders for both proteins, suggesting a conserved binding preference. To further assess the specificity of the remaining top compounds, cross-docking was performed with the four nonshared ligands (Table ). This analysis showed that all of these molecules exhibited favorable and comparable predicted affinities across both proteins. These findings suggest that beyond atovaquone, additional candidate ligands may display cross-reactivity within the GPN family, underscoring their potential as novel inhibitors of the RNA polymerase recognition site in GPN proteins and warranting further experimental validation.

Furthermore, it is plausible to consider that the binding of small molecules inside the hydrophobic pocket, or externally, potentially affect the recognition or hydrolysis of GTP, in agreement with the allosteric pocket model, suggesting the existence of pockets located in distinct sites from the catalytic one with affinity to small molecules, where their binding affects the function of the enzyme. Similarly, GTP/GDP binding can be affected by the interaction of key peptides from the interface of RNA polymerase II subunits with the constituent residues of the open Npa3 hydrophobic pocket. Also, it is essential to study the possible effect of these peptides on the GTPase activity of these essential enzymes. For this purpose, González-González et al. reported in 2017 a protocol for the purification of recombinant human Gpn1, allowing in the near future the assessment of the effect of these compounds in vitro.

4. Conclusions

In this work, we characterized the hydrophobic pocket in Npa3 proposed as essential for the recognition and binding of peptides at the interface of RNA polymerase II subunits Rpb1, Rpb4, Rpb8, and Rpb11. Based on this biochemical information, we modeled by flexible protein–peptide and protein–protein docking the interactions between the essential GTPase Npa3 from the yeast S. cerevisiae and the Rpb1, Rpb4, Rpb8, and Rpb11 subunits of the RNA polymerase II multienzyme complex. We identified the specific interactions between key residues of both Npa3 and these subunits. Additionally, we performed rigid molecular docking experiments between Npa3 and homology-modeled human Gpn1 against a library of FDA-approved compounds. In a second round of flexible docking, we determined the residues of both proteins that are most actively involved in interactions against the five chemical compounds with the highest interaction energy, resulting from the first rigid docking. For both GTPases, Gpn1 and Npa3, the top compound in common was atovaquone, exhibiting interaction energies of −7.9 and −8.6 kcal/mol, respectively. Based on the results of flexible molecular docking, we propose the compounds atovaquone, tibolone, ketanserine, dicumarol, and trifluperidol as potential allosteric inhibitors of the chaperone activity of Npa3 in the recognition of RNA polymerase II. The biological relevance of residues F143 and W179 in Npa3; and its equivalents F166 and W202 in Gpn1, which were constantly involved in interactions with the evaluated Rpb1 peptides and small molecules, requires additional examination by site-directed mutagenesis in future experiments. Based on these findings, we propose that these are key residues for the interaction of these GTPases with RNA polymerase II. We finally suggest these compounds for in vitro assays with particular interest in the exploration of their potential impact on the subcellular localization of the RNA polymerase II subunits, both in the yeast S. cerevisiae and in mammalian cells.

5. Materials and Methods

5.1. Selection and Energy Minimization of Open Npa3

This work was performed on the crystallographic structure of Npa3, the yeast ortholog of human Gpn1, in its open (GTP-bound) conformation. The structure is available in the Protein Data Bank (RCSB-PDB) with the PDB code 5HCN. The cocrystallized ligands with Npa3 include fatty acid laurate, GMPPCP (nonhydrolyzable analog of GTP), glycerol, and Mg2+, which were removed before structure optimization. The structure was minimized using the CHARMM36 potential and the charmm38b1 package. This process involved 50 steepest descent steps for all hydrogen atoms followed by a minimization of 50 conjugate gradient steps for all of the atoms; the purpose of these short energy minimizations is to eliminate clashes introduced by adding hydrogen atoms to the crystal structure. Geometrical and energetic assessments were conducted using Procheck, ProSa, and Molprobity. The final structure was used for subsequent analyses.

5.2. Human Gpn1 Model Generation

Human Gpn1 model was built as described by Cristóbal-Mondragón et al. Briefly, Gpn1 lacking N- and C-terminal regions (amino acids 19–270) was generated by homology modeling using the Modeler 9.19 program. The crystallographic structure of Npa3 in its open conformation was taken as a template. The energy minimization procedure applied to refine the model followed the same method as that described for the Npa3 crystallographic structure. The resulting optimized structure was used for subsequent analyses. All equivalences between the residues that constitute the hydrophobic pocket of Npa3 and Gpn1 are shown in Figure S1.

5.3. Pocket Identification and Hydrophobic Pocket Characterization of Open Npa3 and Gpn1 Models

The global properties of pockets in the structure of open Npa3 and Gpn1 were determined using the bioinformatics tool DoGSiteScorer server from the “Structure-Based Modeling Support Server-ProteinPlus”. It was configured for the calculation of pockets and subpockets in the full-length protein as well as for the prediction of the druggability score; values for this parameter are assigned between 0 and 1, the closer to 1 suggests that the cavity has a higher probability for acting as a ligand binding site. Pocket volumes were determined using default parameters in the DoGSiteScorer server. Briefly, DoGSiteScorer employs a grid-based algorithm and a difference-of-Gaussian (DoG) filter to detect potential binding cavities on the protein surface. The volume of each predicted pocket is then computed by counting the grid points comprising that pocket and multiplying by the grid box volume, yielding the pocket volume in cubic Å. The information obtained is presented as pocket descriptors, which include pocket volume, surface area, amino acid composition, hydrophobicity, shape, and others. This information allows the estimation of parameters such as pharmacological potential of the pocket. The hydrophobic surface representations for Npa3 and Gpn1 were generated using the “mlp” command in ChimeraX, with default settings and a coloring range of −20 to 20. Briefly, the “mlp” command calculates Molecular Lipophilicity Potential (MLP) maps for proteins, which are analogous to electrostatic potential maps. In this representation, positive potentials correspond to more hydrophobic areas (colored in dark gold), negative values indicate more hydrophilic areas (colored in dark cyan), while midrange values around 0 represent intermediate lipophilicity (colored in white). We relied on the standard protonation state assumptions from ChimeraX default MLP parameters; the MLP calculation thus reflects the standard charged or uncharged form of each amino acid side chain.

5.4. Flexible Docking between Peptides in the Interface of RNA Polymerase II Subunits with Open Monomeric Npa3

Guided by the biochemical assay performed by Niesser et al., peptides of 15 amino acids with a reported fluorescence intensity exceeding 4.0 were selected (see Table ). As a first step, we performed flexible docking of these peptides over the Npa3 structure using the CABS-dock server for flexible protein–peptide docking. The Npa3 structure was directly uploaded to the server, specifying the 15 amino acid sequence of each peptide. No other optional settings were defined on the server. As the second step, we considered the subunits within the context of their complete and folded structures. The full 3D models of S. cerevisiae Rpb1, Rpb4, Rpb8, and Rpb11 subunits (UniProt: P04050, P20433, P20436, and P38902, respectively) were obtained from the AlphaFold Protein Structure Database. Confidence intervals for the peptides located in each RNA polymerase II employed are shown in Figure S2. All of the structures were previously minimized following the same protocol described for the Npa3 and Gpn1 model structures. The structure of open Npa3 and the RNA polymerase II subunits were directly uploaded to the HADDOCK (High Ambiguity Driven protein–protein DOCKing) server version 2.4 for flexible protein–protein docking. For Npa3, active residues, i.e., those potentially involved in the interaction with these peptides and mapping the hydrophobic pocket (Y130, V132, N137, T142, F143, N146, M147, C151, L154, M161, V163, F165, W179, F183, M208, L209, F212, Y213, L216, and V218) were selected in all of the following flexible docking experiments, according to the findings from DoGSiteScorer. In the case of each RNA polymerase subunit, the 15 amino acids that showed an interaction with Npa3 in the biochemical assays were defined as active residues (Table ). All other docking parameters were maintained at default settings provided by the server.

From the complexes generated by the CABS-dock and HADDOCK servers, we selected the best for further analysis. CABS-dock models were chosen based on cluster ranking, considering structures that represent around 50% of the total models generated by the program (marked in red in Table S2). In the case of HADDOCK models, selection was based on HADDOCK energy scores (Figure S3), considering the ensemble of models with energies with overlapping standard deviations for each cluster. For the selected models, we calculated residue–residue contacts between the protein and peptide, using MDAnalysis (Python 3.7). Only carbon atoms were considered, and a distance cutoff threshold of 6.0 Å was applied to define contacts.

The generated contact lists were further filtered to retain those residue–residue contacts between the Npa3 hydrophobic pocket residues (as identified by DoGsiteScorer) and the 15 amino acids from each peptide. Residue–residue contacts were categorized in three groups: (1) exclusive to CABS-dock models, (2) exclusive to HADDOCK models, or (3) shared between both docking methods. Heatmap plots visualizing these interactions were created using Matplotlib (Python 3.7).

5.5. Chemical Library Preparation

The ligand library was designed using the ZINC 12 database for commercially available chemical compounds. Parameters such as oral bioavailability, blood-brain barrier crossing ability, human intestinal absorption, solubility, and low toxicity were assessed by applying Lipinski’s rule of 5, except for the molecular weight. Compounds exceeding a molecular weight of 600 g/mol were excluded. The following additional parameters were set: net charge ranging from −5 and 5, rotatable bonds between 0 and 50, polar surface area (Å2) from 0 to 200, polar desolvation (kcal/mol) from −400 to 1, and apolar desolvation (kcal/mol) between −100 and 40. The selection was focused on the FDA-approved drug catalog, resulting in 1786 compounds based on the search parameters.

5.6. Validation by Laurate Redocking and Rescoring of Cocrystallized Laurate

To validate the suitability of AutoDock Vina performance on this system, we carried out redocking experiments. The Npa3 structure in the open conformation was employed, and cocrystallized laurate was removed. The structure of laurate for redocking was obtained from the ZINC 15 database under ID ZINC1529498 (net charge −1, H donors = 0, hydrogen bridge acceptors = 2, tPSA = 40, rotatable bonds = 10, apolar desolvation = 9.86 and polar desolvation = 43.35). Molecular docking between laurate and Npa3 was performed using AutoDock Vina software, with the laurate molecule defined as flexible while Npa3 was maintained rigid. Box dimensions were based on the hydrophobic pocket location, using a grid spacing of 1 Å and including the total volume of the pocket. Ten binding poses for laurate were indicated using an exhaustivity of 20. Additionally, the binding energy of the cocrystallized laurate was calculated using the AutoDock Vina rescoring function, in order to subsequently be compared with the energy obtained from redocking.

5.7. Molecular Docking of the Chemical Library on the Open Npa3 and Gpn1 Hydrophobic Pocket

An initial molecular docking was performed using AutoDock Vina, involving the designed library of 1786 FDA-approved compounds against the minimized Npa3 and homology-modeled human Gpn1, both in the open conformation. The box size was set to 25 Å for the x, y, and z axes. Five poses for each compound were generated using an exhaustiveness of 20. The structures of Npa3 and Gpn1 were defined as rigid, whereas the compounds from the FDA library were defined as flexible. All proteins and ligands were transformed to the PDBQT format prior to use. Polar hydrogen atoms, partial charges, atom types, and torsional information for flexible ligands were considered during docking. Subsequently, we performed a second flexible receptor docking for the top 5 compounds that presented the best interaction energy in the rigid docking round, according to a cutoff energy criterion established as −7.5 kcal/mol against Npa3 and the human Gpn1 model. Flexible residues for Npa3 and Gpn1 correspond to those from the laurate binding site. The same settings employed in rigid docking were used for the sake of consistency. The five compounds with the best energy were selected for the final analysis. For the top five compounds identified through flexible docking for each protein, we performed additional flexible cross-docking of the nonshared compounds to evaluate their cross-interaction potential with both proteins. The same flexible docking parameters were applied as previously described. 2D interaction diagrams of specific contacts were generated with LigPlot+ software. The cutoff distance for determining hydrophobic interactions, hydrogen bridges, and salt bridges was set to default values as specified by the software.

Supplementary Material

ao5c05849_si_001.pdf (668KB, pdf)

Acknowledgments

This work was funded by the Consejo Nacional de Humanidades, Ciencias y Tecnologías (Conahcyt, currently SECIHTI) with grant number A1-S-21070 (RSO). We thank Consejo Nacional de Humanidades, Ciencias y Tecnologías (Conahcyt, currently SECIHTI) for supporting JAML by granting a doctoral fellowship with number 784721. G.R.C.-M. was supported by a Postdoctoral Fellowship from DGAPA, Universidad Nacional Autónoma de México (UNAM). We also thank the Laboratorio Nacional de Supercómputo del Sureste de México for providing access to the supercomputing system to perform all computational calculations through project number 202201035N (MRC).

Npa3 and Gpn1 input structures used in the docking experiments, structure of laurate and the redocked complex, FDA library in mol2 format, and top 5 ranked complexes resulting from flexible and docking assays (10.5281/zenodo.17117473).

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c05849.

  • Descriptors of the laurate binding site (Table S1); ranking of models generated by CABS-dock for Npa3 docking (Table S2); comparison of the hydrophobic pockets in Npa3 and Gpn1 (Figure S1); AlphaFold2 models for the docked peptides (Figure S2); HADDOCK scores of clusters of flexible docking (Figure S3); FDA-approved library bound to Npa3 (Figure S4); and FDA-approved library bound to Gpn1 (Figure S5) (PDF)

The authors declare no competing financial interest.

References

  1. Cherfils J., Zeghouf M.. Regulation of Small GTPases by GEFs, GAPs, and GDIs. Physiol. Rev. 2013;93(1):269–309. doi: 10.1152/physrev.00003.2012. [DOI] [PubMed] [Google Scholar]
  2. Reiner D. J.. Small GTPases. WormBook. 2018:1–65. doi: 10.1895/wormbook.1.67.2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bos J. L., Rehmann H., Wittinghofer A.. GEFs and GAPs: critical elements in the control of small G proteins. Cell. 2007;129(5):865–877. doi: 10.1016/j.cell.2007.05.018. [DOI] [PubMed] [Google Scholar]
  4. Bourne H. R., Sanders D. A., McCormick F.. The GTPase Superfamily: A Conserved Switch for Diverse Cell Functions. Nature. 1990;348:125–132. doi: 10.1038/348125a0. [DOI] [PubMed] [Google Scholar]
  5. Quilliam L. A., Khosravi-Far R., Huff S. Y., Der C. J.. Guanine Nucleotide Exchange Factors: Activators of the Ras Superfamily of Proteins. BioEssays. 1995;17(5):395–404. doi: 10.1002/bies.950170507. [DOI] [PubMed] [Google Scholar]
  6. Barrett T., Xiao B., Dodson E. J., Dodson G., Ludbrook S. B., Nurmahomed K., Gamblin S. J., Musacchio A., Smerdon S. J., Eccleston J. F.. The Structure of the GTPase-Activating Domain from P50rhoGAP. Nature. 1997;385(6615):458–461. doi: 10.1038/385458a0. [DOI] [PubMed] [Google Scholar]
  7. Geyer M., Wittinghofert A.. GEFs, GAPs, GDIs and Effectors: Taking a Closer (3D) Curr. Opin. Struct. Biol. 1997;7(6):786–792. doi: 10.1016/S0959-440X(97)80147-9. [DOI] [PubMed] [Google Scholar]
  8. Soundararajan M., Eswaran J.. Atypical GTPases as Drug Targets. Anti-Cancer Agents Med. Chem. 2012;12(1):19–28. doi: 10.2174/187152012798764705. [DOI] [PubMed] [Google Scholar]
  9. Bourne H. R., Sanders D. A., McCormick F.. The GTPase Superfamily: Conserved Structure and Molecular Mechanism. Nature. 1991;349:117–127. doi: 10.1038/349117a0. [DOI] [PubMed] [Google Scholar]
  10. Verstraeten N., Fauvart M., Versees W., Michiels J.. The Universally Conserved Prokaryotic GTPases. Microbiol. Mol. Biol. Rev. 2011;75(3):507–542. doi: 10.1128/MMBR.00009-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gras S., Chaumont V., Fernandez B., Carpentier P., Charrier-Savournin F., Schmitt S., Pineau C., Flament D., Hecker A., Forterre P., Armengaud J., Housset D.. Structural Insights into a New Homodimeric Self-Activated GTPase Family. EMBO Rep. 2007;8(6):569–575. doi: 10.1038/sj.embor.7400958. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Forget D., Lacombe A.-A., Cloutier P., Al-Khoury R., Bouchard A., Lavallée-Adam M., Faubert D., Jeronimo C., Blanchette M., Coulombe B.. The Protein Interaction Network of the Human Transcription Machinery Reveals a Role for the Conserved GTPase RPAP4/GPN1 and Microtubule Assembly in Nuclear Import and Biogenesis of RNA Polymerase II. Mol. Cell. Proteomics. 2010;9(12):2827–2839. doi: 10.1074/mcp.M110.003616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Minaker S. W., Filiatrault M. C., Ben-Aroya S., Hieter P., Stirling P. C.. Biogenesis of RNA Polymerases II and III Requires the Conserved GPN Small GTPases in Saccharomyces cerevisiae . Genetics. 2013;193(3):853–864. doi: 10.1534/genetics.112.148726. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Alonso B., Beraud C., Meguellati S., Chen S. W., Pellequer J. L., Armengaud J., Godon C.. Eukaryotic GPN-Loop GTPases Paralogs Use a Dimeric Assembly Reminiscent of Archeal GPN. Cell Cycle. 2013;12(3):463–472. doi: 10.4161/cc.23367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Niesser J., Wagner F. R., Kostrewa D., Mühlbacher W., Cramer P.. Structure of GPN-Loop GTPase Npa3 and Implications for RNA Polymerase II Assembly. Mol. Cell. Biol. 2016;36(5):820–831. doi: 10.1128/MCB.01009-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Calera M. R., Zamora-Ramos C., Araiza-Villanueva M. G., Moreno-Aguilar C. A., Peña-Gómez S. G., Castellanos-Terán F., Robledo-Rivera A. Y., Sánchez-Olea R.. Parcs/Gpn3 Is Required for the Nuclear Accumulation of RNA Polymerase II. Biochim. Biophys. Acta, Mol. Cell Res. 2011;1813(10):1708–1716. doi: 10.1016/j.bbamcr.2011.07.005. [DOI] [PubMed] [Google Scholar]
  17. Wild T., Cramer P.. Biogenesis of Multisubunit RNA Polymerases. Trends Biochem. Sci. 2012;37(3):99–105. doi: 10.1016/j.tibs.2011.12.001. [DOI] [PubMed] [Google Scholar]
  18. Boulon S., Pradet-Balade B., Verheggen C., Molle D., Boireau S., Georgieva M., Azzag K., Robert M. C., Ahmad Y., Neel H., Lamond A. I., Bertrand E.. HSP90 and Its R2TP/Prefoldin-like Cochaperone Are Involved in the Cytoplasmic Assembly of RNA Polymerase II. Mol. Cell. 2010;39(6):912–924. doi: 10.1016/j.molcel.2010.08.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Shoichet B. K., Bodial D. L., Kuntz I. D.. Molecular Docking Using Shape Descriptors. J. Comput. Chem. 1992;13(3):380–397. doi: 10.1002/jcc.540130311. [DOI] [Google Scholar]
  20. Nadendla R. R.. Molecular Modeling: A Powerful Tool for Drug Design and Molecular Docking. Resonance. 2004;9(5):51–60. doi: 10.1007/BF02834015. [DOI] [Google Scholar]
  21. Kharb M., Jat R. K., Parjapati G., Gupta A.. Review on Introduction To Molecular Docking Software Technique in Medicinal Chemistry. Int. J. Drug Res. Technol. 2012;2(2):189–197. [Google Scholar]
  22. Ferreira L. G., Dos Santos R. N., Oliva G., Andricopulo A. D.. Molecular Docking and Structure-Based Drug Design Strategies. Molecules. 2015;20(7):13384–13421. doi: 10.3390/molecules200713384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. An J., Totrov M., Abagyan R.. Comprehensive Identification of “Druggable” Protein Ligand Binding Sites. Genome Inf. 2004;15(2):31–41. doi: 10.11234/gi1990.15.2_31. [DOI] [PubMed] [Google Scholar]
  24. Pérot S., Sperandio O., Miteva M. A., Camproux A. C., Villoutreix B. O.. Druggable Pockets and Binding Site Centric Chemical Space: A Paradigm Shift in Drug Discovery. Drug Discovery Today. 2010;15(15–16):656–667. doi: 10.1016/j.drudis.2010.05.015. [DOI] [PubMed] [Google Scholar]
  25. Duhé, R. J. Drug Design. In Encyclopedia of Cancer; Schwab, M. , Ed.; Springer: Berlin, Heidelberg, 2014; p 1423. [Google Scholar]
  26. Hocker H. J., Cho K. J., Chen C. Y., Rambahal N., Sagineedu S. R., Shaari K., Stanslas J., Hancock J. F., Gorfe A. A.. Andrographolide derivatives inhibit guanine nucleotide exchange and abrogate oncogenic Ras function. Proc. Natl. Acad. Sci. U.S.A. 2013;110(25):10201–10206. doi: 10.1073/pnas.1300016110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kuntz I. D., Blaney J. M., Oatley S. J., Langridge R., Ferrin T. E.. A Geometric Approach to Macromolecule-Ligand Interactions. J. Mol. Biol. 1982;161(2):269–288. doi: 10.1016/0022-2836(82)90153-X. [DOI] [PubMed] [Google Scholar]
  28. Huang S. Y., Zou X.. Advances and Challenges in Protein-Ligand Docking. Int. J. Mol. Sci. 2010;11(8):3016–3034. doi: 10.3390/ijms11083016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Volkamer A., Griewel A., Grombacher T., Rarey M.. Analyzing the Topology of Active Sites: On the Prediction of Pockets and Subpockets. J. Chem. Inf. Model. 2010;50(11):2041–2052. doi: 10.1021/ci100241y. [DOI] [PubMed] [Google Scholar]
  30. Volkamer A., Kuhn D., Rippmann F., Rarey M.. Dogsitescorer: A Web Server for Automatic Binding Site Prediction, Analysis and Druggability Assessment. Bioinformatics. 2012;28(15):2074–2075. doi: 10.1093/bioinformatics/bts310. [DOI] [PubMed] [Google Scholar]
  31. Kurcinski M., Jamroz M., Blaszczyk M., Kolinski A., Kmiecik S.. CABS-dock web server for the flexible docking of peptides to proteins without prior knowledge of the binding site. Nucleic Acids Res. 2015;43(W1):W419–W424. doi: 10.1093/nar/gkv456. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Honorato R. V., Trellet M. E., Jiménez-García B., Schaarschmidt J. J., Giulini M., Reys V., Koukos P. I., Rodrigues J. P. G. L. M., Karaca E., van Zundert G. C. P., Roel-Touris J., van Noort C. W., Jandová Z., Melquiond A. S. J., Bonvin A. M. J. J.. The HADDOCK2.4 web server: A leap forward in integrative modelling of biomolecular complexes. Nat. Protoc. 2024;19(11):3219–3241. doi: 10.1038/s41596-024-01011-0. [DOI] [PubMed] [Google Scholar]
  33. Trott O., Olson A. J.. AutoDock Vina: Improving the Speed and Accuracy of Docking with a New Scoring Function, Efficient Optimization, and Multithreading. J. Comput. Chem. 2009;31(2):455–461. doi: 10.1002/jcc.21334. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Maxwell A.. The Interaction between Coumarin Drugs and DNA Gyrase. Mol. Microbiol. 1993;9(4):681–686. doi: 10.1111/j.1365-2958.1993.tb01728.x. [DOI] [PubMed] [Google Scholar]
  35. Marcu M. G., Chadli A., Bouhouche I., Catelli M., Neckers L. M.. The Heat Shock Protein 90 Antagonist Novobiocin Interacts with a Previously Unrecognized ATP-Binding Domain in the Carboxyl Terminus of the Chaperone. J. Biol. Chem. 2000;275(47):37181–37186. doi: 10.1074/jbc.M003701200. [DOI] [PubMed] [Google Scholar]
  36. Stank A., Kokh D. B., Fuller J. C., Wade R. C.. Protein Binding Pocket Dynamics. Acc. Chem. Res. 2016;49(5):809–815. doi: 10.1021/acs.accounts.5b00516. [DOI] [PubMed] [Google Scholar]
  37. González-González R., Guerra-Moreno J. A., Cristóbal-Mondragón G. R., Romero V., Peña-Gómez S. G., Montero-Morán G. M., Lara-González S., Hernández-Arana A., Fernández-Velasco D. A., Calera M. R., Sánchez-Olea R.. Human Gpn1 Purified from Bacteria Binds Guanine Nucleotides and Hydrolyzes GTP as a Protein Dimer Stabilized by Its C-Terminal Tail. Protein Expression Purif. 2017;132:85–96. doi: 10.1016/j.pep.2017.01.009. [DOI] [PubMed] [Google Scholar]
  38. Berman H. M., Westbrook J., Feng Z., Gilliland G., Bhat T. N., Weissig H., Shindyalov I. N., Bourne P. E.. The Protein Data Bank. Nucleic Acids Res. 2000;28(1):235–242. doi: 10.1093/nar/28.1.235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Huang J., MacKerell A. D. Jr.. CHARMM36 All-Atom Additive Protein Force Field: Validation Based on Comparison to NMR Data. J. Comput. Chem. 2013;34(25):2135–2145. doi: 10.1002/jcc.23354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Brooks B. R., Brooks C. L., Mackerell A. D., Nilsson L., Petrella R. J., Roux B., Won Y., Archontis G., Bartels C., Boresch S., Caflisch A., Caves L., Cui Q., Dinner A. R., Feig M., Fischer S., Gao J., Hodoscek M., Im W., Kuczera K., Lazaridis T., Ma J., Ovchinnikov V., Paci E., Pastor R. W., Post C. B., Pu J. Z., Schaefer M., Tidor B., Venable R. M., Woodcock H. L., Wu X., Yang W., York D. M., Karplus M.. CHARMM: The Biomolecular Simulation Program. J. Comput. Chem. 2009;30(10):1545–1614. doi: 10.1002/jcc.21287. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Williams C. J., Headd J. J., Moriarty N. W., Prisant M. G., Videau L. L., Deis L. N., Verma V., Keedy D. A., Hintze B. J., Chen V. B., Jain S., Lewis S. M., Arendall W. B. III, Snoeyink J., Adams P. D., Lovell S. C., Richardson J. S., Richardson D. C.. MolProbity: More and better reference data for improved all-atom structure validation. Protein Sci. 2018;27(1):293–315. doi: 10.1002/pro.3330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Cristóbal-Mondragón G. R., Lara-Chacón B., Santiago Á, De-la-Rosa V., González-González R., Muñiz-Luna J. A., Ladrón-de-Guevara E., Romero-Romero S., Rangel-Yescas G. E., Fernández Velasco D. A., Islas L. D., Pastor N., Sánchez-Olea R., Calera M. R.. FRET-Based Analysis and Molecular Modeling of the Human GPN-Loop GTPases 1 and 3 Heterodimer Unveils a Dominant-Negative Protein Complex. FEBS J. 2019;286(23):4797–4818. doi: 10.1111/febs.14996. [DOI] [PubMed] [Google Scholar]
  43. Webb B., Sali A.. Comparative Protein Structure Modeling Using MODELLER. Curr. Protoc. Bioinf. 2016;(June):1–37. doi: 10.1002/cpbi.3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Pettersen E. F., Goddard T. D., Huang C. C., Meng E. C., Couch G. S., Croll T. I., Morris J. H., Ferrin T. E.. UCSF ChimeraX: Structure visualization for researchers, educators, and developers. Protein Sci. 2021;30(1):70–82. doi: 10.1002/pro.3943. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Jumper J., Evans R., Pritzel A., Green T., Figurnov M., Ronneberger O., Tunyasuvunakool K., Bates R., Žídek A., Potapenko A., Bridgland A., Meyer C., Kohl S. A. A., Ballard A. J., Cowie A., Romera-Paredes B., Nikolov S., Jain R., Adler J., Back T., Petersen S., Reiman D., Clancy E., Zielinski M., Steinegger M., Pacholska M., Berghammer T., Bodenstein S., Silver D., Vinyals O., Senior A. W., Kavukcuoglu K., Kohli P., Hassabis D.. Highly Accurate Protein Structure Prediction with AlphaFold. Nature. 2021;596(7873):583–589. doi: 10.1038/s41586-021-03819-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Michaud-Agrawal N., Denning E. J., Woolf T. B., Beckstein O.. MDAnalysis: a toolkit for the analysis of molecular dynamics simulations. J. Comput. Chem. 2011;32(10):2319–2327. doi: 10.1002/jcc.21787. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Hunter J. D.. Matplotlib: A 2D Graphics Environment. Comput. Sci. Eng. 2007;9(3):90–95. doi: 10.1109/MCSE.2007.55. [DOI] [Google Scholar]
  48. Irwin J. J., Sterling T., Mysinger M. M., Bolstad E. S., Coleman R. G.. ZINC: A Free Tool to Discover Chemistry for Biology. J. Chem. Inf. Model. 2012;52(7):1757–1768. doi: 10.1021/ci3001277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Sterling T., Irwin J. J.. ZINC 15 - Ligand Discovery for Everyone. J. Chem. Inf. Model. 2015;55(11):2324–2337. doi: 10.1021/acs.jcim.5b00559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Laskowski R. A., Swindells M. B.. LigPlot+: Multiple Ligand-Protein Interaction Diagrams for Drug Discovery. J. Chem. Inf. Model. 2011;51:2778–2786. doi: 10.1021/ci200227u. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

ao5c05849_si_001.pdf (668KB, pdf)

Data Availability Statement

Npa3 and Gpn1 input structures used in the docking experiments, structure of laurate and the redocked complex, FDA library in mol2 format, and top 5 ranked complexes resulting from flexible and docking assays (10.5281/zenodo.17117473).

Articles from ACS Omega are provided here courtesy of American Chemical Society

RESOURCES