Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2023 Oct 12;63(20):6412–6422. doi: 10.1021/acs.jcim.3c01212

Discovery of Hit Compounds Targeting the P4 Allosteric Site of K-RAS, Identified through Ensemble-Based Virtual Screening

Patricia Gomez-Gutierrez †,‡, Jaime Rubio-Martinez §, Juan J Perez †,*
PMCID: PMC10598794  PMID: 37824186

Abstract

graphic file with name ci3c01212_0011.jpg

Mutants of Ras are oncogenic drivers of a large number of human tumors. Despite being recognized as an attractive target for the treatment of cancer, the high affinity for its substrate tagged the protein as undruggable for a few years. The identification of cryptic pockets on the protein surface gave the opportunity to identify molecules capable of acting as allosteric modulators. Several molecules were disclosed in recent years, with sotorasib and adagrasib already approved for clinical use. The present study makes use of computational methods to characterize eight prospective allosteric pockets (P1–P8) in K-Ras, four of which (P1–P4) were previously characterized in the literature. The present study also describes the results of a virtual screening study focused on the discovery of hit compounds, binders of the P4 site that can be considered as peptidomimetics of a fragment of the SOS αI helix, a guanine exchange factor of Ras. After a detailed description of the computational procedure followed, we disclose five hit compounds, prospective binders of the P4 allosteric site that exhibit an inhibitory capability higher than 30% in a cell proliferation assay at 50 μM.

1. Introduction

The Ras family of proteins represents a group of small GTPases that function as molecular signal switches, cycling between the GDP-bound inactive and the GTP-bound active forms. The two forms exhibit different conformations, granting the protein the ability to recognize diverse substrates. Two types of proteins facilitate the exchange between the active and inactive forms. On the one hand, guanine exchange factors (GEFs) catalyze nucleotide exchange from GDP to GTP, and on the other hand, GTPase activating proteins (GAPs) facilitate the hydrolysis of the γ-phosphate of the bound GTP.1 The Ras family comprises different protein isoforms that share a high sequence identity including H-Ras, N-Ras, K-Ras4A, and K-Ras4B, the latter being two splicing isoforms of the same gene.2

This family of proteins engages and activates diverse downstream signaling pathways regulating different cellular responses, including proliferation, survival, and differentiation.3,4 Specifically, the Ras–Raf–MEK–ERK pathway controls cell proliferation by modulating the levels of diverse cell cycle regulators. Alternately, Ras can also promote proliferation and cell survival through the Ras–PI3K–Akt signaling pathway. In addition, Ras can activate other proteins like the Ral guanine nucleotide dissociation stimulator (RalGDS) or the T-cell lymphoma invasion and metastasis 1 (TIAM1), involved in the regulation of vesicle trafficking and cytoskeletal organization, respectively.5,6 Constitutive activation of Ras is the cause of severe pathologies including cancer.7 Actually, about 30% of human cancers are due to Ras constitutive activation, with the highest incidence in lung, colon, and pancreatic cancers.1 Dysregulation of Ras is mainly due to the occurrence of point mutations, most frequently in residues G12, G13, and Q61 that cover ca. 98% of all mutations associated with Ras,1 with K-Ras being the predominantly mutated isoform (85%).8 These mutations leave the protein in a constitutively active state due to a decreased intrinsic GTPase activity and/or to an increased resistance to GAP-mediated hydrolysis1,9 resulting in aberrant signaling.

Recognized as an attractive target for the treatment of cancer, the high cytosolic concentrations of GTP and GDP, together with their high affinity for Ras, rendered the design of competitive inhibitors elusive.10 Moreover, the lack of obvious alternative pockets in its crystal structure made the protein to be considered for a long time as undruggable.11 However, analysis of the protein dynamical behavior carried out by means of computational studies revealed the existence of transient pockets susceptible to be exploited for the design of allosteric inhibitors.12,13 As a proof of concept, diverse low-affinity inhibitors were subsequently identified and reported in the literature.14−18 Today, sotorasib19 and adagrasib,20 allosteric inhibitors specifically targeting the G12C mutant, are already approved for clinical use.

The structure of K-Ras consists of a globular GTPase domain (residues 1–166) and a hypervariable region (167–188). Furthermore, the globular domain is composed of two lobes including the N-terminal or effector lobe (residues 1–86), where all the downstream effectors and other major regulators such as GEF and GAP proteins make major contacts,21 and the C-terminal or allosteric lobe (residues 87–166) engaged in dimerization.22 Diverse allosteric sites were identified from computational studies,12,13 with four of them (P1–P4) subsequently validated experimentally.23,24 The P1 site or switch I/II pocket is located in the effector lobe between the β1−β3 strands and the α2 helix. Most of the allosteric inhibitors disclosed to date bind to this site including compounds described by Fesik’s group,14 Genetech,15 Kobe0065,16 0375-0604,25 Gorfe’s group,26 BI-2852,27 or Rabbitts’s group.28,29 The P2 site or switch II pocket is also located on the effector lobe between helices α2 and α3 and the P-loop. This is the binding site of irreversible inhibitors targeting the G12C K-Ras mutant like sotorasib,19 adagrasib,20 or ARS-853.30 Moreover, recent reports suggest that inhibition of K-Ras through this site can be achieved beyond the G12C mutants, opening the door for designing more general therapeutic agents.31 In this direction, a reversible ligand (TH-Z835) targeting the G12D mutant32 and a reversible pan K-Ras compound (BI-2865) were recently disclosed.33 P3 is located in the allosteric lobe between the C-terminus of helix α3, the L7 loop, and the C-terminus of helix α5. Compound KAL-21404358 is one of the few compounds disclosed shown to bind to this site.34 Interestingly, the compound Zn2+-cyclen binds with low affinity to a P3 subpocket termed P3b.35 Both compounds favor the state 1 versus the state 2 of the active, GTP bound K-Ras structure, characterized by the loss of hydrogen bonding interactions between the γ-phosphate and residues Thr35 and Gly60 and associated with a weaker affinity to effector molecules.34−36 Finally, the P4 site is located between the effector and the allosteric lobe close to the nucleotide binding site. It involves residues of the switch I, the α1 helix, and β2 and emerges when the switch I adopts an open conformation bound to GDP. Located at the interaction interface of some downstream effectors such as Raf, PI3K, or RalGDS,37 there is no definitively recognized small molecule binder yet. Andrographolide and its derivatives SRJ09 and SRJ23 were proposed to bind the site,38 although subsequent studies point to the switch I/II pocket as the most likely spot.39

Druggability of K-Ras by inhibition through the P4 site is puzzling because there is no direct evidence of a small molecule binder, although the antibody mimetic DARPin K27 binds to the site and inhibits K-Ras downstream signaling.40 This raises the question of whether the site is affordable only to macromolecule ligands. Aimed at exploring the viability of the site for designing novel small molecule K-Ras inhibitors, we report in the present work the results of a virtual screening study on the site and disclose novel hit molecules as prospective allosteric modulators. For this purpose, we first carried out a study of the dynamic behavior of the protein using molecular dynamics to better characterize the structural features of the P4 site. These results permitted defining a query that was used for screening the ZINC chemical library.41 As a result, we conclude this report by disclosing the structures of five molecular hits, prospective binders of the P4 site.

2. Methods

Identification of Transient Pockets in K-Ras

The computational procedure followed to identify transient sites on the protein is explained elsewhere.42 Specifically, protein was subjected to a 500 ns accelerated molecular dynamics (aMD) calculation within the classical (NVT) ensemble using the protocol implemented in Amber12.43 The solvent was treated explicitly using the TIP3P water model, and periodic boundary conditions were applied. Energy was computed using the Amber ff99SB force field,44 although parameters for GDP were taken from an alternative source.45 A cutoff distance of 10 Å was set for short-range noncovalent interactions, whereas electrostatic interactions were treated using the PME method.46 The temperature was kept constant at 300 K by means of a Langevin thermostat.47

The simulation was carried out on the structure of the human K-RasG12 V bound to GDP solved at 1.76 Å resolution (PDB ID: 4EPX),14 whose atomic coordinates were retrieved from the RCSB PDB website.48 Protonation states of the diverse residues were assigned using the Protonate3D function of the MOE program.49 The structure was soaked in a rectangular box of 12,876 water molecules, to which six Na+ ions were added to maintain the neutrality of the system. The system was energy minimized to relax the initial structure in several steps using the steepest descent method. In the first step, only side chains were free to move, applying to the protein backbone atoms and GDP-Mg a positional constraining harmonic potential with a force constant of 1 kcal/mol Å2. In the following steps, restrictions were gradually lifted, first on the protein backbone and then on the GDP-Mg system. Next, the temperature of the system was gradually increased using 100 ps of NVT molecular dynamics at a rate of 30 K every 10 ps. Subsequently, the system was subjected to an equilibration process consisting of 1 ns of molecular dynamics in the NPT ensemble and 1 ns of molecular dynamics in the NVT ensemble. The SHAKE algorithm was used to restrict the elongation motion of the bonds involving hydrogen atoms,50 which allowed using an integration time of 2 fs.

Conformational sampling was performed using aMD for 500 ns at 300 K. A biased potential was used in both the total and dihedral torsional energies, respectively, by adding a bias to the true potential to improve the sampling efficiency. Bias potentials were determined from the average potential energy (Ep = −12241 kcal/mol) and the average dihedral energy (Ed = 2394 kcal/mol) from a previous 50 ns classical molecular dynamics trajectory and the use of system size dependent tuning parameters αP = 8275 and αD = 120.51 After aMD calculations were completed, artifacts produced by the biased potential were removed using a 10th-order Maclaurin series reweighting process for each configuration to recover the canonical ensemble.51

Conformational analysis was carried out using 50,000 structures of the trajectory taken at regular intervals of 10 ps. Structures were aligned using the Cα coordinates of the residues included in the invariant nucleus (i.e., residues with a smaller displacement in regard to their initial position along the trajectory) to the first structure of the trajectory using the cpptraj module of the Amber12 program.43 To identify the diverse conformations the protein attains during the sampling process, we carried out a hierarchical cluster analysis by means of the average link algorithm52 using the backbone Cα root-mean-square deviation as a measure for the distance between two conformations. Subsequently, mapping of transient binding sites on the protein surface was carried out on each of the cluster representative structures using the FTMap procedure.53 It consists of sampling billions of positions of small organic molecules used as probes and scoring their poses using a detailed energy expression. Regions that bind several probe types serve to identify binding hot spots.

Cryptic Pocket Characterization

The druggability of each of the prospective binding sites was assessed by means of the SiteMap program.54 The procedure computes a druggability score for the target site. The computation includes diverse terms that promote ligand binding including adequate size, isolation from the solvent, and penalizes increasing hydrophilicity.

Virtual Screening

A pharmacophore was developed for any binding site of interest by means of the Site Finder program of MOE.49 In addition, an exclusion volume was included as restriction to disregard structures that, despite complying with the pharmacophore, may exhibit steric effects with the protein.

Virtual screening was carried out using the ZINC chemical library41 filtered to contain lead-like molecules only.55 Prior to undertaking the virtual screening process, up to 1000 conformations of each molecule were generated using the conformational import function of MOE.49 Molecules were first screened to determine the pharmacophore of the site of interest. Then, the compounds selected were subjected to a molecular docking process using the Glide program.56,57 Next, poses resulting from the docking process were filtered again for pharmacophore fulfillment using a lower degree of compliance. Finally, a diversity analysis of the set of compounds was performed by means of the canvas program58,59 using a molecular fingerprint that encodes the three-point pharmacophores that meet the 3D structures of the generated protein–ligand poses. Visual inspection of the complex structures permitted selecting those compounds to be procured from commercial vendors.

In Vitro Tests

Cell proliferation inhibition assays were conducted using the NIH/3T3 fibroblast cell line transformed with K-RAS oncogene DNA with the G12 V mutation by transfection. Cells were cultured in DMEM supplemented with 5% fetal bovine serum (FBS). The cultured cells were seeded in tissue culture microplates at a density of 2 × 103 cells/cm2 and incubated for 24 h at 37 °C in a humidified atmosphere containing 5% CO2. Test compounds were first dissolved in DMSO 100% and subsequently diluted to 50 μM, embedding a maximum 0.5% DMSO concentration, and incubated in the cell culture for another 72 h. After incubation, proliferation was quantified using the CellTiter 96 Aqueous Non-Radioactive Cell Proliferation Assay-MTS (Promega 10 #G5421) following manufacturer's instructions. Absorbance at 490 nm that is directly proportional to the number of living cells was recorded with a BMG Fluostar Optima Microplate Reader and normalized relative to control cells treated with the vehicle.

3. Results and Discussion

Identification of Transient Binding Pockets in K-Ras

Transient binding pockets on K-Ras were identified from the analysis of 50,000 structures taken from the aMD trajectory at regular intervals of 10 ps. The first step consisted of classifying the diverse conformations into groups by means of cluster analysis. This required the computation of a distance matrix for all pairs of structures using the root-mean-square deviation (rmsd) of the backbone Cα atoms once the structures were previously aligned using the less flexible elements of the protein, as explained below. In the second step, FTMap53 was used to identify hot spots on each of the diverse clusters characterized.

Figure 1 shows the structure of K-Ras color-coded according to residue flexibility, and Figure S1 shows the root-mean-square fluctuations of each residue. Inspection of both figures suggests that switch I is the most flexible structural motif, together with the N-terminus of the β2-strand (red in Figure 1). In contrast, the P-loop is one of the most stable structural elements during the simulation, together with the β protein core (β4−β6), the C-termini of the β1 and β3 strands, helix α1 N-terminus, helix α3 N-terminus, and the central part of helix α4 (blue in Figure 1). The rest of the elements, including switch II and loop L7, exhibit intermediate flexibility (green in Figure 1). The set of most stable structural elements was used to superimpose the 50,000 structures and to compute a distance matrix between each pair of structures using the backbone Cα rmsd. The subsequent cluster analysis was carried out by means of the average link algorithm52 that permitted to group the structures into six clusters. Clusters were represented by the structure with the closest root-mean-square deviation (rmsd) to the average value of the group. Figure S2 shows the representative structures of each cluster superimposed with the crystal structure used as starting point for the MD simulations in this work (PDB ID: 4EPX).14 Inspection of Figure S2 suggests that the differences between structures are located on the switch I and switch II regions, in addition to specific secondary elements linked to them like the β2 strand and the α3 helix, as well as the loops L4 and L7.

Figure 1.

Figure 1

Open in a new tab

Structural elements of K-Ras and their differential flexibility. The most flexible elements are depicted in red, and the least flexible elements are in blue, whereas those showing intermediate flexibility are depicted in green.

Mapping hot spots on the protein surface for each of the six cluster representative structures was performed using the FTMap program.53 For comparison purposes, we also mapped the crystallographic structure used in the present work. The aggregated results of the mapping on each of the six representatives as well as that of the crystal structure are shown in Figure 2. As can be seen, in addition to the nucleotide-binding site, eight different sites were identified and labeled P1–P8. Sites P1–P4 correspond to those well-characterized and validated sites described above.23,24 The rest of the sites identified (P5–P8), despite a few of them being described previously in computational studies,12,13,60 are not yet experimentally validated. Interestingly, only four of these sites (P1, P2, P3, and P7) appear in both the crystal and at least one of the representative structures, stressing that binding hot spot identification requires inclusion of the dynamical behavior of the protein.

Figure 2.

Figure 2

Open in a new tab

Mapping of molecular fragments from the FTMap calculations. (a) Using the crystallographic structure 4EPX as template and (b) aggregated results using each of the six cluster representatives.

A druggability profile of the eight sites identified was investigated using SiteMap.54 This procedure permits classifying sites into druggable, difficult, and undruggable by means of a druggability score, defined as a composite index comprising diverse properties like size, exposure, enclosure, hydrophobicity, and hydrophilicity.54 However, the threshold of the druggability index differentiating the diverse categories is fuzzy, and although it provides an overall view, values of other properties can be used to fine-tune the classification of a specific site.54

Two druggability indexes were computed in the present work and are shown in the first two columns of Table 1. Specifically, the Sscore is a composite index computed using parameters like size, enclosure, and hydrophilicity. On the other hand, the Dscore is computed using parameters that promote ligand binding like adequate size or isolation from solvent but also includes a penalizing term reflecting the hydrophilicity of the site.54 Inspection of the first two columns of Table 1 suggests that both indexes follow the same trends with small discrepancies, despite the difference between the Dscore and the Sscore for sites P7 and P5. Inspection of Table 1 suggests that none of the sites can be considered as druggable, with P3, P4, P1, P6, and P2 being difficult and the rest undruggable. The present results agree with the results of a previous analysis performed on sites P1–P4.23

Table 1. Druggability Characteristics of Sites P1–P8 Computed with SiteMap and Ordered by their Sscorea.

site Sscore Dscore size enclosure hydrophobicity hydrophilicity
P3 0.879 0.828 76 0.657 0.43 1.182
P4 0.835 0.821 53 0.699 1.258 0.867
P1 0.824 0.799 46 0.734 1.249 0.855
P6 0.804 0.774 40 0.743 1.409 0.818
P2 0.760 0.731 53 0.629 0.59 1.015
P7 0.706 0.573 39 0.670 0.406 1.278
P5 0.642 0.575 26 0.670 1.185 0.939
P8 0.501 0.514 16 0.400 0.5 0.312
Open in a new tab
a

Classification as druggable, difficult, and undruggable is loose, but as reference, druggable sites exhibit average Dscore values of around 1.1, difficult sites exhibit average Dscore values of around 0.9, and undruggable sites exhibit average Dscore values lower than 0.6. However, for a more accurate assessment, values of diverse specific parameters need to be considered.

Ligands targeting these sites have a potential role as allosteric regulators or protein–protein modulators depending on their specific binding location. However, it should be borne in mind that identification of a binding transient site does not guarantee that the site is necessarily functional.61 Concerning sites already validated, site P1 corresponds to the switch I/II pocket described above,14−18 the binding site of diverse ligands.14−16,25−29 There are also macromolecules described to bind to the site, acting as inhibitors, like the affimer K662 or the monobody JAM20.63 In the present study, the site is identified in five of the cluster representative structures (1, 2, 3, 5, and 6) as well as in the crystal structure. Moreover, it appears number three in Table 1, being considered as a difficult. Compounds binding the P1 site likely act as protein–protein interaction inhibitors preventing effectors to bind K-Ras or, alternatively, interfere with i-mediated GDP/GTP exchange.23

Site P2 corresponds to the switch II pocket, the binding site of diverse irreversible inhibitors targeting the G12C mutant19,20,30 ] including sotorasib and adagrasib.19,20 There are also macromolecules binding to the site acting as inhibitors, like the affimer K3.62 FTMap identifies the pocket in three of the cluster representative structures (1, 4, and 5), in addition to the crystal structure. The site appears listed as difficult in Table 1, giving expectations of finding noncovalent small molecule ligands for the site, in agreement with recent reports describing reversible binders to the site.31−34 Compounds binding to the site are thought to stabilize a non-SOS-recognizable switch II conformation and, consequently, inhibit GDP/GTP exchange.17

P3 is the binding site of compound KAL-21404358,34 as well as the Zn2+-cyclen in a P3b subpocket,35 as described above. The DARPins K13 and K19 also bind to the site.64 In the present study, the site is identified by FTMap in five cluster representative structures (1, 2, 4, 5, and 6), as well as in the crystal structure. However, the values listed in Table 1, together with other reports,23 point to the site as undruggable due to the low hydrobophicity indicator. The site P3b should also be considered as undruggable due to its small size. Binding to this site occurs preferentially in state 1 of the active Ras-GTP, providing the opportunity to stabilize the protein in this state that prevents the activation of the downstream cascade.37

Finally, P4 was previously proposed to be one of the prospective binding sites for andrographolide and its derivatives SRJ09 and SRJ23, together with P1.38,39 This uncertainty is probably due to the fact that P4 is located in the vicinity of P1. Moreover, the P4 site is the target of the antibody mimetic DARPin K27 that shows inhibition of downstream signaling.40 In the present work, the P4 site is identified in the cluster representative 4; however, cluster representatives 2 and 5 identify a subsite. In contrast, the site cannot be identified in the crystal structure. The difference between P4 in the structure of cluster 4 and in the structures of clusters 2 and 5 is a consequence of the conformation adopted by switch I. Whereas, in the former, it is widely open, in the latter, it adopts a partially closed form, making accessible only a subsite of P4. Interestingly, the structure of the switch I adopted in the representative of cluster 4 is very close to the conformation found in the structure of the K-Ras/SOS complex (PDB entry: 1nvv), and it is associated with an increased GDP off rate.40,65 The results of Table 1 suggest that the site can be considered as difficult, a classification that agrees with previous findings.23

Concerning the rest of the sites identified, P5 and P6 are located between the effector and the allosteric lobes, whereas P7 and P8 are located on the allosteric lobe. P5 is located between the helix α5 N-term, the helix α1 C-term, and the β2 strand. It appears in six cluster representatives but not in the crystal structure. The pocket was identified in previous computational studies.13,66 Subsequent directed mutagenesis experiments on residues Asn26 and Val45 confirmed that the pocket is part of the recognition interface of the cysteine-rich domain of the Raf protein,67 suggesting that binders to the pocket can act as protein–protein inhibitors of Raf. Site P6, not described previously, is located between the C-terminus of helix α5, the β1 strand, and the L3 loop. This site is identified in two cluster representative structures (4 and 6) and appears as the result of the movement of the L3 loop and the β-sheet coordinated with the movement of switch I. According to the results of Table 1, site P6 appears with similar druggability parameters as P1 that label it as difficult. Sites P7 and P8 were described in previous computational studies.13,60 Site P7 is located between the α4 helix, the N-terminal part of the α5 helix, and the β6 strand. The site was recently identified as the binding pocket of the monobody NS168 as well as the affimer K69.62 These proteins disrupt RAS dimerization, leading to the blocking of CRAF-BRAF heterodimerization and activation. P7 is analogous to a site previously described and associated with the calmodulin binding site.69 In the present study, P7 is identified in four cluster representative structures (2, 3, 4, and 6), being listed in Table 1 as undruggable due to its small size and hydrophobicity index. Finally, the P8 site is a small pocket located between helices α3 and α4 and was identified only in the cluster representative structure 4. Although the site may be considered as difficult according to the values of the Sscore and Dscore listed in Table 1, it should be considered as undruggable due to the low hydrophobicity index.

Virtual Screening of the P4 Site

A six-point pharmacophore was defined based on the K-Ras structure using the Site Finder module of MOE.49 The features that binders should exhibit include (Figure 3) an aromatic ring (Ph1) located in the vicinity of Tyr40 and Ile36; two hydrogen bond acceptor points (Ph2 and Ph4) in the vicinity of the NH group of the Asp33 backbone and the hydroxyl group of Tyr40, respectively; a hydrophobic moiety (Ph3) next to Ile21 and Thr20; and two hydrophobic or aromatic moieties (Ph5 and Ph6) adjacent to residues Leu56 and Ile21, respectively. In addition, three constraints were added, two (R2 and R4) to define the direction of the hydrogen acceptor points (Ph2 and Ph4) and another to specify the orientation of the aromatic ring (R1) of point Ph1, following the PHCD pharmacophore definition of MOE.49 In addition, an exclusion volume that defines the binding cavity was included to eliminate structures that, despite complying with the pharmacophore, overlap their structure with that of the protein.

Figure 3.

Figure 3

Open in a new tab

Pharmacophore of site P4. Color-coded spheres represent the features of the diverse pharmacophore points (Ph1–Ph6), as well as geometrical restrictions (R1–R3). Green represents hydrophobic/aromatic points; cyan represents proton accepting points; dark green represents hydrophobic points; navy blue represents a restriction on the aromatic ring conformation; and magenta represents a restriction regarding the direction of the proton accepting sites.

Screening of the ZINC compound library for pharmacophore fulfillment yielded a total of 1364 molecules and a total of 2372 conformations, with 11 of them fulfilling all restrictions. The resulting compounds were subjected to a molecular docking process using the Glide program.56,57 Next, to prevent the flexible docking process from taking ligands away from pharmacophore fulfillment, the diverse poses were filtered for their degree of compliance of at least six pharmacophoric points, reducing the size of the library to 336 compounds. Finally, to facilitate the selection of the final candidates, a diversity analysis was carried out using the canvas program of the Schrödinger software.56,57 For this purpose, we used a molecular fingerprint as a descriptor, encoding all three-point pharmacophores that meet the 3D structures of the generated protein–ligand poses. This permitted us to classify the hits into 30 clusters. Finally, a subset of 16 cluster representatives was selected according to their docking score and stereochemical complementarity with the binding site assessed by visual inspection. Of these, 13 compounds were purchased from commercial suppliers (A55001 to A55013) and submitted for in vitro tests.

Screening showed four compounds with a cell proliferation inhibition higher than 30% at 50 μM. This is an arbitrary cutoff to retain compounds for further investigation, loose enough not to miss any interesting hit and covering the maximal chemical diversity of the set.70,71 We then searched the corresponding diversity clusters of positive hits for additional compounds and selected six additional compounds for screening according to their docking score and stereochemical complementarity with the binding site, assessed by visual inspection. Four of these compounds were purchased (compounds A55014 to A55017) and submitted for in vitro tests, with one of them showing a cell proliferation inhibitory behavior at least 30% at 50 μM. Table 2 shows the results of the in vitro screening of all the compounds tested at 50 μM. Compounds A55003 and A55016 show inhibitions close to 50%, whereas compounds A55004, A55007 and A33013 inhibit cell proliferation at a 30–40% yield. The success rate of the screening process is about 30%, similar to the one observed in similar studies.72−76 The results shown in Table 2 reveal that the present compounds arrest proliferation. It is likely that the action is mediated through K-Ras, although it cannot be ruled out that they may enhance or activate other cellular mechanisms.

Table 2. Compounds with Inhibitory Profile >30%@50 μM in a Cell Proliferation Assaya.

graphic file with name ci3c01212_0010.jpg

Open in a new tab
a

Their commercial names, chemical formula, and inhibition activity are listed in the diverse columns (raw data on cell proliferation for the diverse assays are listed in Table S1 of the Supporting Information).

Figures 4–8 show the prospective bound conformations of the five compounds with cell proliferation inhibitory capacity higher than 30% at 50 μM. All the compounds fulfill pharmacophore point Ph1 by means of an aromatic ring that is placed perpendicular to the aromatic ring of Tyr40 and close to Ile36. The hydrophobic/aromatic point Ph6 is also fulfilled in all five compounds with moieties as diverse as a methyl group or a five-membered heterocyclic ring that interact with Pro34. Hydrophobic point Ph3 is fulfilled by means of a methyl or methylene group at a distance suitable for a van der Waals interaction with Ile21. All the compounds, except for A55003, fulfill the hydrophobic or aromatic pharmacophoric point Ph5. Specifically, compound A55004 fulfills the point by means of a benzene ring, compound A55007 fulfills the point by means of a methyl group, and compounds A55013 and A55016 fulfill the point by means of a piperidine ring. These moieties establish a van der Waals interaction with Leu56. Regarding hydrogen bond interactions, except for A55016, compounds fulfill the pharmacophoric hydrogen acceptor points Ph2 and Ph4. Specifically, compound A55003 establishes a double hydrogen bond interaction through the oxygen of an aliphatic hydroxyl group with the amide hydrogens of the Tyr32 and Asp33 backbone. The same interaction is established with the oxygen of an amide group in compound A55004, with the sulfur atom of a thiophene ring in compound A55007, and with the nitrogen in meta of the pyrazole ring of compound A55013, although the interaction is established only with the nitrogen of the Asp33 backbone in the latter. In contrast, the amide carbonyl of compound A55016 is far from both donor groups of the receptor to establish a hydrogen bond, despite it establishing a hydrogen bond with the hydroxyl group of Tyr40. All compounds, with the exception of A55013, have a central amide group that forms a hydrogen bond by means of the oxygen atom of the carbonyl group with the hydroxyl group of Tyr40. In contrast, compound A55013 stabilizes the interaction by means of an aliphatic-type hydroxyl group.

Figure 4.

Figure 4

Open in a new tab

Proposed conformation of compound A55003 bound to the P4 site. (a) Bound conformation showing K-Ras residues involved and (b) fulfillment of the pharmacophore of site P4.

Figure 8.

Figure 8

Open in a new tab

Proposed conformation of compound A550016 bound to the P4 site. (a) Bound conformation showing K-Ras residues involved and (b) fulfillment of the pharmacophore of site P4.

Figure 5.

Figure 5

Open in a new tab

Proposed conformation of compound A55004 bound to the P4 site. (a) Bound conformation showing K-Ras residues involved and (b) fulfillment of the pharmacophore of site P4.

Figure 6.

Figure 6

Open in a new tab

Proposed conformation of compound A55007 bound to the P4 site. (a) Bound conformation showing K-Ras residues involved and (b) fulfillment of the pharmacophore of site P4.

Figure 7.

Figure 7

Open in a new tab

Proposed conformation of compound A550013 bound to the P4 site. (a) Bound conformation showing K-Ras residues involved and (b) fulfillment of the pharmacophore of site P4.

After analysis of the prospective bound conformations of the diverse hits, we hypothesize that binding of compounds to the P4 site may exert a stabilizing action on switch I that facilitates its open conformation. This conformation is similar to the one adopted by K-Ras when bound to SOS, more widely open than that shown in the structures of Ras-GTP in state 1 and associated with an increased GDP dissociation rate. The fact that this type of open conformation was sampled during the simulation process suggests that the recognition between Ras and SOS could respond to a conformational selection mechanism. However, the two conformations are not identical showing some differences in the arrangement of diverse residues. Accordingly, a second step of induced fit is required between the interfaces of both proteins in the recognition process.

The compounds identified in the present work can be considered as peptidomimetics of the SOS helix αI77,78 interfering with the recognition between K-Ras and SOS and consequently arresting the GDP dissociation process. Inspection of the bound conformation of the active compounds showed that these compounds share a few interactions with the K-Ras/SOS complex. Specifically, all the compounds occupy in part the area of the SOS αI helix C-terminus in the K-Ras/SOS complex (Figure 9). Moreover, the interaction between their corresponding aromatic rings and Tyr40 in K-Ras mimics the interaction between Tyr40 and SOS His911 found in the K-Ras/SOS complex. Another interaction common to all compounds is the presence of an acceptor group forming a hydrogen bond with the backbone NH amide of Tyr32 and mimicking the interaction between Tyr32 and SOS Asn944 in the K-Ras/SOS complex.

Figure 9.

Figure 9

Open in a new tab

Superimposition of the SOS αI helix (green) and compound A550013 (cyan) bound to K-Ras (purple).

Conclusions

The present work reports the results of a modeling study aimed at identifying transient pockets of K-Ras as prospective allosteric sites to modulate protein activity. For this purpose, the dynamical behavior of the protein was investigated by means of a 500 ns molecular dynamics trajectory at 300 K using accelerated molecular dynamics as the sampling technique. Cluster analysis of the trajectory permitted us to identify six different conformations adopted by the protein. Each of the conformations was mapped for hot spots using diverse molecular probes by means of the FTMap program. The results permitted identification of eight cryptic pockets. Seven of them were previously described, with four of them already validated. We also analyzed the characteristics of the pockets.

Moreover, a virtual screening process was conducted on site P4 using the ZINC database. After a hierarchical process, we ended up with a short list of 16 compounds, 13 of which were purchased from commercial suppliers and submitted for in vitro testing. Four compounds exhibited cell proliferation inhibitory activity higher than 30% at 50 μM. In the second step, we selected an additional six compounds from the clusters of the positive hits for screening. Four of these compounds were purchased and submitted to in vitro screening. Only one of them showed cell proliferation inhibitory activity. Accordingly, the present work discloses five molecular hits with cell proliferation inhibitory activity higher than 30% at 50 μM, targeting the P4 site of K-Ras. Further studies need to be carried out to validate binding to the site and to assess their role as allosteric modulators of the K-Ras activity.

Acknowledgments

This study was supported in part by the Agència de Gestió d’Ajuts Universitaris i de Recerca (AGAUR)-Generalitat de Catalunya (2021SGR00350 and 2021SGR00342) and the Spanish Structures and Excellence María de Maeztu program (grant CEX2021-001202-M.)

Data Availability Statement

MD trajectories produced during the execution of this work can be obtained from the authors upon request. In-house scripts, input files, topologies, and PDB structures of all cluster representatives used for the analysis are placed in the public repository Github: https://github.com/JaimeRubioMartinez/KRas_P4_AllostericSite.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jcim.3c01212.

  • Detailed results regarding the root-mean-square fluctuations per residue and the structures in PDB format of the representatives obtained from the cluster analysis (PDF)

The authors declare no competing financial interest.

Supplementary Material

ci3c01212_si_001.pdf (696.2KB, pdf)

References

  1. Donovan S.; Shannon K. M.; Bollag G. GTPase activating proteins: critical regulators of intracellular signaling. Biochim. Biophys. Acta, Rev. Cancer 2002, 1602, 23–45. 10.1016/S0304-419X(01)00041-5. [DOI] [PubMed] [Google Scholar]
  2. Downward J. The Ras Superfamily of Small GTP-Binding Proteins. Trends Biochem. Sci. 1990, 15, 469–472. 10.1016/0968-0004(90)90300-Z. [DOI] [PubMed] [Google Scholar]
  3. Schubbert S.; Shannon K.; Bollag G. Hyperactive Ras in developmental disorders and cancer. Nat. Rev. Cancer 2007, 7, 295–308. 10.1038/nrc2109. [DOI] [PubMed] [Google Scholar]
  4. Engelman J. A.; Luo J.; Cantley L. C. The evolution of phosphatidylinositol 3-kinases as regulators of growth and metabolism. Nat. Rev. Genet. 2006, 7, 606–619. 10.1038/nrg1879. [DOI] [PubMed] [Google Scholar]
  5. Wolthuis R. M.; Bos J. L. Ras caught in another affair: the exchange factors for Ral. Curr. Opin. Genet. Dev. 1999, 9, 112–117. 10.1016/S0959-437X(99)80016-1. [DOI] [PubMed] [Google Scholar]
  6. Lambert J. M.; Lambert Q. T.; Reuther G. W.; Malliri A.; Siderovski D. P.; Sondek J.; Collard J. G.; Der C. J. Tiam1 mediates Ras activation of Rac by a PI(3)K-independent mechanism. Nat. Cell Biol. 2002, 4, 621–625. 10.1038/ncb833. [DOI] [PubMed] [Google Scholar]
  7. Simanshu D. K.; Nissley D. V.; McCormick F. RAS Proteins and Their Regulators in Human Disease. Cell 2017, 170, 17–33. 10.1016/j.cell.2017.06.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Hobbs G. A.; Der C. J.; Rossman K. L. RAS isoforms and mutations in cancer at a glance. J. Cell Sci. 2016, 129, 1287–1292. 10.1242/jcs.182873. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Lu S.; Jang H.; Gu S.; Zhang J.; Nussinov R. Drugging Ras GTPase: a comprehensive mechanistic and signaling structural view. Chem. Soc. Rev. 2016, 45, 4929–4952. 10.1039/C5CS00911A. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Scheffzek K.; Ahmadian M. R.; Kabsch W.; Wiesmuller L.; Lautwein A.; Schmitz F.; Wittinghofer A. The Ras-RasGAP complex: structural basis for GTPase activation and its loss in oncogenic Ras mutants. Science 1997, 277, 333–338. 10.1126/science.277.5324.333. [DOI] [PubMed] [Google Scholar]
  11. Cox A. D.; Fesik S. W.; Kimmelman A. C.; Luo J.; Der C. J. Drugging the undruggable RAS: Mission possible?. Nat. Rev. Drug Discovery 2014, 13, 828–851. 10.1038/nrd4389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Grant B. J.; Lukman S.; Hocker H. J.; Sayyah J.; Brown J. H.; McCammon J. A.; Gorfe A. A. Novel allosteric sites on Ras for lead generation. PLoS One 2011, 6, e25711 10.1371/journal.pone.0025711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Buhrman G.; O’Connor C.; Zerbe B.; Kearney B. M.; Napoleon R.; Kovrigina E. A.; Vajda S.; Kozakov D.; Kovrigin E. L.; Mattos C. Analysis of binding site hot spots on the surface of Ras GTPase. J. Mol. Biol. 2011, 413, 773–789. 10.1016/j.jmb.2011.09.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Sun Q.; Burke J. P.; Phan J.; Burns M. C.; Olejniczak E. T.; Waterson A. G.; Lee T.; Rossanese O. W.; Fesik S. W. Discovery of small molecules that bind to K-Ras and inhibit Sos-mediated activation. Angew. Chem., Int. Ed. Engl. 2012, 51, 6140–6143. 10.1002/anie.201201358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Maurer T.; Garrenton L. S.; Oh A.; Pitts K.; Anderson D. J.; Skelton N. J.; Fauber B. P.; Pan B.; Malek S.; Stokoe D.; Ludlam M. J.; Bowman K. K.; Wu J.; Giannetti A. M.; Starovasnik M. A.; Mellman I.; Jackson P. K.; Rudolph J.; Wang W.; Fang G. Small-molecule ligands bind to a distinct pocket in Ras and inhibit SOS-mediated nucleotide exchange activity. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 5299–5304. 10.1073/pnas.1116510109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Shima F.; Yoshikawa Y.; Ye M.; Araki M.; Matsumoto S.; Liao J.; Hu L.; Sugimoto T.; Ijiri Y.; Takeda A.; Nishiyama Y.; Sato C.; Muraoka S.; Tamura A.; Osoda T.; Tsuda K.; Miyakawa T.; Fukunishi H.; Shimada J.; Kumasaka T.; Yamamoto M.; Kataoka T. In silico discovery of small-molecule Ras inhibitors that display antitumor activity by blocking the Ras-effector interaction. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 8182–8187. 10.1073/pnas.1217730110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Ostrem J. M.; Peters U.; Sos M. L.; Wells J. A.; Shokat K. M. K-Ras(G12C) inhibitors allosterically control GTP affinity and effector interactions. Nature 2013, 503, 548–551. 10.1038/nature12796. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Spiegel J.; Cromm P. M.; Zimmermann G.; Grossmann T. N.; Waldmann H. Small-molecule modulation of Ras signaling. Nat. Chem. Biol. 2014, 10, 613–622. 10.1038/nchembio.1560. [DOI] [PubMed] [Google Scholar]
  19. Canon J.; Rex K.; Saiki A. Y.; Mohr C.; Cooke K.; Bagal D.; Gaida K.; Holt T.; Knutson C. G.; Koppada N.; Lanman B. A.; Werner J.; Rapaport A. S.; San Miguel T.; Ortiz R.; Osgood T.; Sun J. R.; Zhu X.; McCarter J. D.; Volak L. P.; Houk B. E.; Fakih M. G.; O’Neil B. H.; Price T. J.; Falchook G. S.; Desai J.; Kuo J.; Govindan R.; Hong D. S.; Ouyang W.; Henary H.; Arvedson T.; Cee V. J.; Lipford J. R. The clinical KRAS(G12C) inhibitor AMG 510 drives anti-tumour immunity. Nature 2019, 575, 217–223. 10.1038/s41586-019-1694-1. [DOI] [PubMed] [Google Scholar]
  20. Fell J. B.; Fischer J. P.; Baer B. R.; Blake J. F.; Bouhana K.; Briere D. M.; Brown K. D.; Burgess L. E.; Burns A. C.; Burkard M. R.; Chiang H.; Chicarelli M. J.; Cook A. W.; Gaudino J. J.; Hallin J.; Hanson L.; Hartley D. P.; Hicken E. J.; Hingorani G. P.; Hinklin R. J.; Mejia M. J.; Olson P.; Otten J. N.; Rhodes S. P.; Rodriguez M. E.; Savechenkov P.; Smith D. J.; Sudhakar N.; Sullivan F. X.; Tang T. P.; Vigers G. P.; Wollenberg L.; Christensen J. G.; Marx M. A. Identification of the Clinical Development Candidate MRTX849, a Covalent KRASG12C Inhibitor for the Treatment of Cancer. J. Med. Chem. 2020, 63, 6679–6693. 10.1021/acs.jmedchem.9b02052. [DOI] [PubMed] [Google Scholar]
  21. Gorfe A. A.; Grant B. J.; McCammon J. A. Mapping the nucleotide and isoform-dependent structural and dynamical features of Ras proteins. Structure 2008, 16, 885–896. 10.1016/j.str.2008.03.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Rudack T.; Teuber C.; Scherlo M.; Güldenhaupt J.; Schartner J.; Lübben M.; Klare J.; Gerwert K.; Kötting C. The Ras dimer structure. Chem. Sci. 2021, 12, 8178–8189. 10.1039/D1SC00957E. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Gupta A. K.; Wang X.; Pagba C. V.; Prakash P.; Sarkar-Banerjee S.; Putkey J.; Gorfe A. A. Multi-target, ensemble-based virtual screening yields novel allosteric KRAS inhibitors at high success rate. Chem. Biol. Drug Des. 2019, 94, 1441–1456. 10.1111/cbdd.13519. [DOI] [PubMed] [Google Scholar]
  24. Steffen C. L.; Kaya P.; Schaffner-Reckinger E.; Abankwa D. Eliminating oncogenic RAS: back to the future at the drawing board. Biochem. Soc. Trans. 2023, 51, 447–456. 10.1042/BST20221343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Xie C.; Li Y.; Li L. L.; Fan X. X.; Wang Y. W.; Wei C. L.; Liu L.; Leung E. L. H.; Yao X. J. Identification of a New Potent Inhibitor Targeting KRAS in Non-small Cell Lung Cancer Cells. Front. Pharmacol. 2017, 8, 823. 10.3389/fphar.2017.00823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. McCarthy M. J.; Pagba C. V.; Prakash P.; Naji A. K.; van der Hoeven D.; Liang H.; Gupta A. K.; Zhou Y.; Cho K. J.; Hancock J. F.; Gorfe A. A. Discovery of High-Affinity Noncovalent Allosteric KRAS Inhibitors That Disrupt Effector Binding. ACS Omega 2019, 4, 2921–2930. 10.1021/acsomega.8b03308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kessler D.; Gmachl M.; Mantoulidis A.; Martin L. J.; Zoephel A.; Mayer M.; Gollner A.; Covini D.; Fischer S.; Gerstberger T.; Gmaschitz T.; Goodwin C.; Greb P.; Häring D.; Hela W.; Hoffmann J.; Karolyi-Oezguer J.; Knesl P.; Kornigg S.; Koegl M.; Kousek R.; Lamarre L.; Moser F.; Munico-Martinez S.; Peinsipp C.; Phan J.; Rinnenthal J.; Sai J.; Salamon C.; Scherbantin Y.; Schipany K.; Schnitzer R.; Schrenk A.; Sharps B.; Siszler G.; Sun Q.; Waterson A.; Wolkerstorfer B.; Zeeb M.; Pearson M.; Fesik S. W.; McConnell D. B. Drugging an undruggable pocket on KRAS. Proc. Natl. Acad. Sci. U.S.A. 2019, 116, 15823–15829. 10.1073/pnas.1904529116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Cruz-Migoni A.; Canning P.; Quevedo C. E.; Bataille C. J. R.; Bery N.; Miller A.; Russell A. J.; Phillips S. E. V.; Carr S. B.; Rabbitts T. H. Structure-based development of new RAS-effector inhibitors from a combination of active and inactive RAS-binding compounds. Proc. Natl. Acad. Sci. U.S.A. 2019, 116, 2545–2550. 10.1073/pnas.1811360116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Quevedo C. E.; Cruz-Migoni A.; Bery N.; Miller A.; Tanaka T.; Petch D.; Bataille C. J. R.; Lee L. Y. W.; Fallon P. S.; Tulmin H.; Ehebauer M. T.; Fernandez-Fuentes N.; Russell A. J.; Carr S. B.; Phillips S. E. V.; Rabbitts T. H. Small molecule inhibitors of RAS-effector protein interactions derived using an intracellular antibody fragment. Nat. Commun. 2018, 9, 3169. 10.1038/s41467-018-05707-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Patricelli M. P.; Janes M. R.; Li L. S.; Hansen R.; Peters U.; Kessler L. V.; Chen Y.; Kucharski J. M.; Feng J.; Ely T.; Chen J. H.; Firdaus S. J.; Babbar A.; Ren P.; Liu Y. Selective Inhibition of Oncogenic KRAS Output with Small Molecules Targeting the Inactive State. Cancer Discovery 2016, 6, 316–329. 10.1158/2159-8290.CD-15-1105. [DOI] [PubMed] [Google Scholar]
  31. Vasta J. D.; Peacock D. M.; Zheng Q.; Walker J. A.; Zhang Z.; Zimprich C. A.; Thomas M. R.; Beck M. T.; Binkowski B. F.; Corona C. R.; Robers M. B.; Shokat K. M. KRAS is vulnerable to reversible switch-II pocket engagement in cells. Nat. Chem. Biol. 2022, 18, 596–604. 10.1038/s41589-022-00985-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mao Z.; Xiao H.; Shen P.; Yang Y.; Xue J.; Yang Y.; Shang Y.; Zhang L.; Li X.; Zhang Y.; Du Y.; Chen C. C.; Guo R. T.; Zhang Y. KRAS(G12D) can be targeted by potent inhibitors via formation of salt bridge. Cell Discovery 2022, 8, 5. 10.1038/s41421-021-00368-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kim D.; Herdeis L.; Rudolph D.; Zhao Y.; Böttcher J.; Vides A.; Ayala-Santos C. I.; Pourfarjam Y.; Cuevas-Navarro A.; Xue J. Y.; Mantoulidis A.; Bröker J.; Wunberg T.; Schaaf O.; Popow J.; Wolkerstorfer B.; Kropatsch K. G.; Qu R.; de Stanchina E.; Sang B.; Li C.; McConnell D. B.; Kraut N.; Lito P. Pan-KRAS inhibitor disables oncogenic signalling and tumour growth. Nature 2023, 619, 160–166. 10.1038/s41586-023-06123-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Feng H.; Zhang Y.; Bos P. H.; Chambers J. M.; Dupont M. M.; Stockwell B. R. K-RasG12D Has a Potential Allosteric Small Molecule Binding Site. Biochemistry 2019, 58, 2542–2554. 10.1021/acs.biochem.8b01300. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rosnizeck I. C.; Graf T.; Spoerner M.; Tränkle J.; Filchtinski D.; Herrmann C.; Gremer L.; Vetter I. R.; Wittinghofer A.; König B.; Kalbitzer H. R. Stabilizing a weak binding state for effectors in the human ras protein by cyclen complexes. Angew. Chem., Int. Ed. Engl. 2010, 49, 3830–3833. 10.1002/anie.200907002. [DOI] [PubMed] [Google Scholar]
  36. Meierhofer T.; Rosnizeck I. C.; Graf T.; Reiss K.; König B.; Kalbitzer H. R.; Spoerner M. Cu2+-cyclen as probe to identify conformational states in guanine nucleotide binding proteins. J. Am. Chem. Soc. 2011, 133, 2048–2051. 10.1021/ja108779j. [DOI] [PubMed] [Google Scholar]
  37. Qiu Y.; Wang Y.; Chai Z.; Ni D.; Li X.; Pu J.; Chen J.; Zhang J.; Lu S.; Lv C.; Ji M. Targeting RAS phosphorylation in cancer therapy: Mechanisms and modulators. Acta Pharm. Sin. B 2021, 11, 3433–3446. 10.1016/j.apsb.2021.02.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. 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, 10201–10206. 10.1073/pnas.1300016110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Quah S. Y.; Tan M. S.; Ho K. L.; Manan N. A.; Gorfe A. A.; Deb P. K.; Sagineedu S. R.; Stanslas J. In silico and saturation transfer difference NMR approaches to unravel the binding mode of an andrographolide derivative to K-Ras oncoprotein. Future Med. Chem. 2020, 12, 1611–1631. 10.4155/fmc-2020-0104. [DOI] [PubMed] [Google Scholar]
  40. Guillard S.; Kolasinska-Zwierz P.; Debreczeni J.; Breed J.; Zhang J.; Bery N.; Marwood R.; Tart J.; Overman R.; Stocki P.; Mistry B.; Phillips C.; Rabbitts T.; Jackson R.; Minter R. Structural and functional characterization of a DARPin which inhibits Ras nucleotide exchange. Nat. Commun. 2017, 8, 16111. 10.1038/ncomms16111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. 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, 1757–1768. 10.1021/ci3001277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Gomez-Gutierrez P.; Rubio-Martinez J.; Perez J. J. Identification of Potential Small Molecule Binding Pockets in p38α MAP kinase. J. Chem. Inf. Model. 2017, 57, 2566–2574. 10.1021/acs.jcim.7b00439. [DOI] [PubMed] [Google Scholar]
  43. Case D. A.; Darden T. A.; Cheatham T. E. III; Simmerling C. L.; Wang J.; Duke R. E.; Luo R.; Walker R. C.; Zhang W.; Merz K. M.; Roberts B.; Hayik S.; Roitberg A.; Seabra G.; Swails J.; Goetz A. W.; Kolossváry I.; Wong K. F.; Paesani F.; Vanicek J.; Wolf R. M.; Liu J.; Wu X.; Brozell S. R.; Steinbrecher T.; Gohlke H.; Cai Q.; Ye X.; Wang J.; Hsieh M.-J.; Cui G.; Roe D. R.; Mathews D. H.; Seetin M. G.; Salomon-Ferrer R.; Sagui C.; Babin V.; Luchko T.; Gusarov S.; Kovalenko A.; Kollman P. A.. AMBER 12. University of California: San Francisco, 2012. [Google Scholar]
  44. Hornak V.; Abel R.; Okur A.; Strockbine B.; Roitberg A.; Simmerling C. Comparison of multiple Amber force fields and development of improved protein backbone parameters. Proteins 2006, 65, 712–725. 10.1002/prot.21123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Meagher K. L.; Redman L. T.; Carlson H. A. Development of polyphosphate parameters for use with the AMBER force field. J. Comput. Chem. 2003, 24, 1016–1025. 10.1002/jcc.10262. [DOI] [PubMed] [Google Scholar]
  46. Darden T.; York D.; Pedersen L. Particle Mesh Ewald: An N·log(N) Method for Ewald Sums in Large Systems. J. Chem. Phys. 1993, 98, 10089–10092. 10.1063/1.464397. [DOI] [Google Scholar]
  47. Allen M. P.; Tildesley D. J.; Allen M. P.; Tildesley D. J.. Computer Simulation of Liquids: 2 Edition. 2017; pp 1–626Oxford Univeristy Press. [Google Scholar]
  48. 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, 235–242. 10.1093/nar/28.1.235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Molecular Operating Environment (MOE), 2013.08; Chemical Computing Group Inc.; 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A 2R7, 2015.
  50. Ryckaert J. P.; Ciccotti G.; Berendsen H. J. C. Numerical Integration of the Cartesian Equations of Motion of a System with Constraints: Molecular Dynamics of n-Alkanes. J. Comput. Phys. 1977, 23, 327–341. 10.1016/0021-9991(77)90098-5. [DOI] [Google Scholar]
  51. Hamelberg D.; Mongan J.; McCammon J. A. Accelerated Molecular Dynamics: a Promising and Efficient Simulation Method for Biomolecules. J. Chem. Phys. 2004, 120, 11919–11929. 10.1063/1.1755656. [DOI] [PubMed] [Google Scholar]
  52. Rokach L.; Maimon O.. Clustering methods. In Data Mining and Knowledge Discovery Handbook; Maimon O., Rokach L.; Eds.; Springer: New York, NY, 2005; Chapter 15, pp 321–352 10.1007/0-387-25465-X_15. [DOI] [Google Scholar]
  53. Brenke R.; Kozakov D.; Chuang G. Y.; Beglov D.; Hall D.; Landon M. R.; Mattos C.; Vajda S. Fragment-Based Identification of Druggable ‘Hot Spots’ of Proteins Using Fourier Domain Correlation Techniques. Bioinformatics 2009, 25, 621–627. 10.1093/bioinformatics/btp036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Halgren T. A. Identifying and Characterizing Binding Sites and Assessing Druggability. J. Chem. Inf. Model. 2009, 49, 377–389. 10.1021/ci800324m. [DOI] [PubMed] [Google Scholar]
  55. Oprea T. I. Property distribution of drug-related chemical databases. J. Comput. - Aided Mol. Des. 2000, 14, 251–264. 10.1023/A:1008130001697. [DOI] [PubMed] [Google Scholar]
  56. Friesner R. A.; Banks J. L.; Murphy R. B.; Halgren T. A.; Klicic J. J.; Mainz D. T.; Repasky M. P.; Knoll E. H.; Shelley M.; Perry J. K.; Shaw D. E.; Francis P.; Shenkin P. S. Glide: a new approach for rapid, accurate docking and scoring. 1. Method and assessment of docking accuracy. J. Med. Chem. 2004, 47, 1739–1749. 10.1021/jm0306430. [DOI] [PubMed] [Google Scholar]
  57. Halgren T. A.; Murphy R. B.; Friesner R. A.; Beard H. S.; Frye L. L.; Pollard W. T.; Banks J. L. Glide: a new approach for rapid, accurate docking and scoring. 2. Enrichment factors in database screening. J. Med. Chem. 2004, 47, 1750–1759. 10.1021/jm030644s. [DOI] [PubMed] [Google Scholar]
  58. Duan J.; Dixon S. L.; Lowrie J. F.; Sherman W. Analysis and comparison of 2D fingerprints: insights into database screening performance using eight fingerprint methods. J. Mol. Graphics Modell. 2010, 29, 157–170. 10.1016/j.jmgm.2010.05.008. [DOI] [PubMed] [Google Scholar]
  59. Sastry M.; Lowrie J. F.; Dixon S. L.; Sherman W. Large-scale systematic analysis of 2D fingerprint methods and parameters to improve virtual screening enrichments. J. Chem. Inf. Model. 2010, 50, 771–784. 10.1021/ci100062n. [DOI] [PubMed] [Google Scholar]
  60. Prakash P.; Hancock J. F.; Gorfe A. A. Binding hotspots on K-ras: consensus ligand binding sites and other reactive regions from probe-based molecular dynamics analysis. Proteins 2015, 83, 898–909. 10.1002/prot.24786. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Chatzigoulas A.; Cournia Zoe Rational design of allosteric modulators: Challenges and successes. Wiley Inter. Rev.: Comput. Mol. Sci. 2021, 11 (6), e1529 10.1002/wcms.1529. [DOI] [Google Scholar]
  62. Haza K. Z.; Martin H. L.; Rao A.; Turner A. L.; Saunders S. E.; Petersen B.; Tiede C.; Tipping K.; Tang A. A.; Ajayi M.; Taylor T.; Harvey M.; Fishwick K. M.; Adams T. L.; Gaule T. G.; Trinh C. H.; Johnson M.; Breeze A. L.; Edwards T. A.; McPherson M. J.; Tomlinson D. C. RAS-inhibiting biologics identify and probe druggable pockets including an SII-α3 allosteric site. Nat. Commun. 2021, 30 (12), 4045. 10.1038/s41467-021-24316-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Wallon L.; Khan I.; Teng K. W.; Koide A.; Zuberi M.; Li J.; Ketavarapu G.; Traaseth N. J.; O’Bryan J. P.; Koide S. Inhibition of RAS-driven signaling and tumorigenesis with a pan-RAS monobody targeting the Switch I/II pocket. Proc. Natl. Acad. Sci. U.S.A. 2022, 119, e2204481119 10.1073/pnas.2204481119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Bery N.; Legg S.; Debreczeni J.; Breed J.; Embrey K.; Stubbs C.; Kolasinska-Zwierz P.; Barrett N.; Marwood R.; Watson J.; Tart J.; Overman R.; Miller A.; Phillips C.; Minter R.; Rabbitts T. H. X1KRAS-specific inhibition using a DARPin binding to a site in the allosteric lobe. Nat. Commun. 2019, 10, 1–11. 10.1038/s41467-019-10419-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Chao F. A.; Dharmaiah S.; Taylor T.; Messing S.; Gillette W.; Esposito D.; Nissley D. V.; McCormick F.; Byrd R. A.; Simanshu D. K.; Cornilescu G. Insights into the Cross Talk between Effector and Allosteric Lobes of KRAS from Methyl Conformational Dynamics. J. Am. Chem. Soc. 2022, 144, 4196–4205. 10.1021/jacs.2c00007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Kapoor A.; Travesset A. Differential dynamics of RAS isoforms in GDP- and GTP-bound states. Proteins 2015, 83, 1091–1106. 10.1002/prot.24805. [DOI] [PubMed] [Google Scholar]
  67. Hu C. D.; Kariya K.; Tamada M.; Akasaka K.; Shirouzu M.; Yokoyama S.; Kataoka T. Cysteine-rich region of Raf-1 interacts with activator domain of post-translationally modified Ha-Ras. J. Biol. Chem. 1995, 270, 30274–30277. 10.1074/jbc.270.51.30274. [DOI] [PubMed] [Google Scholar]
  68. Spencer-Smith R.; Koide A.; Zhou Y.; Eguchi R. R.; Sha F.; Gajwani P.; Santana D.; Gupta A.; Jacobs M.; Herrero-Garcia E.; Cobbert J.; Lavoie H.; Smith M.; Rajakulendran T.; Dowdell E.; Okur M. N.; Dementieva I.; Sicheri F.; Therrien M.; Hancock J. F.; Ikura M.; Koide S.; O’Bryan J. P. Inhibition of RAS function through targeting an allosteric regulatory site. Nat. Chem. Biol. 2017, 13, 62–68. 10.1038/nchembio.2231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Abuasaker B.; Garrido E.; Vilaplana M.; Gómez-Zepeda J. D.; Brun S.; Garcia-Cajide M.; Mauvezin C.; Jaumot M.; Pujol M. D.; Rubio-Martínez J.; Agell Neus α4-α5 Helices on Surface of KRAS can Accommodate Small Compounds that Increase KRAS Signaling while Inducing CRC Cell Death. Int. J. Mol. Sci. 2023, 24, 748. 10.3390/ijms24010748. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Sadybekov A. V.; Katritch V. Computational approaches streamlining drug discovery. Nature 2023, 616, 673–685. 10.1038/s41586-023-05905-z. [DOI] [PubMed] [Google Scholar]
  71. Perez J. J. Managing molecular diversity. Chem. Soc. Rev. 2005, 34, 143–152. 10.1039/b209064n. [DOI] [PubMed] [Google Scholar]
  72. Gomez-Gutierrez P.; Campos Muelas P. M.; Vega M.; Perez González J. J. Identification of a Novel Inhibitory Allosteric Site of MAP Kinases. PLoS One 2016, 11, e0167379 10.1371/journal.pone.0167379. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Lupala C. L.; Gomez-Gutierrez P.; Perez J. J. New insights into the stereochemical requirements of the bradykinin B1 receptor antagonist binding. J. Mol. Graphics Modell. 2016, 68, 184–196. 10.1016/j.jmgm.2016.06.010. [DOI] [PubMed] [Google Scholar]
  74. Lupala C. L.; Gomez-Gutierrez P.; Perez J. J. New insights into the stereochemical requirements of the bradykinin B2 receptor antagonist binding. J. Comp.-Aided Mol. Design. 2016, 30, 85–101. 10.1007/s10822-015-9890-z. [DOI] [PubMed] [Google Scholar]
  75. Rasaeifar B.; Gomez-Gutierrez P.; Perez J. J. New Insights into the Stereochemical Requirements of the Bombesin BB1 Receptor Antagonists Binding. Pharmaceuticals 2020, 13, 197. 10.3390/ph13080197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Simanshu D. K.; Nissley D. V.; McCormick F. RAS Proteins gorfeand Their Regulators in Human Disease. Cell 2017, 170, 17–33. 10.1016/j.cell.2017.06.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  77. Perez J. J.; Corcho F.; Llorens O. Molecular modeling in the design of peptidomimetics and peptide surrogates. Curr. Med. Chem. 2002, 9, 2209–2229. 10.2174/0929867023368683. [DOI] [PubMed] [Google Scholar]
  78. Perez J. J. Designing Peptidomimetics. Curr. Top. Med. Chem. 2018, 18, 566–590. 10.2174/1568026618666180522075258. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

ci3c01212_si_001.pdf (696.2KB, pdf)

Data Availability Statement

MD trajectories produced during the execution of this work can be obtained from the authors upon request. In-house scripts, input files, topologies, and PDB structures of all cluster representatives used for the analysis are placed in the public repository Github: https://github.com/JaimeRubioMartinez/KRas_P4_AllostericSite.

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

RESOURCES