
==== Front
ACS Chem Biol
ACS Chem Biol
cb
acbcct
ACS Chemical Biology
1554-8929
1554-8937
American Chemical Society

39194017
10.1021/acschembio.4c00077
Article
Identification and Characterization of Novel Small-Molecule Enhancers of the CUL3LZTR1 E3 Ligase KRAS Complex
Piech Sophie †
Brüschweiler Sven ‡
Westphalen Josepha †
Siess Katharina M. ‡
García Murias Julio †
Konrat Robert ‡⊥
Bigenzahn Johannes W. *†§
https://orcid.org/0000-0002-0570-1768
Superti-Furga Giulio *†∥
† CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, 1090 Vienna, Austria
‡ MAG-LAB GmbH, 1030 Vienna, Austria
§ Department of Laboratory Medicine, Medical University of Vienna, 1090 Vienna, Austria
∥ Center for Physiology and Pharmacology, Medical University of Vienna, 1090 Vienna Austria
⊥ Department of Structural and Computational Biology, University of Vienna, 1030 Vienna, Austria
* Email: johannes.bigenzahn@meduniwien.ac.at.
* Email: gsuperti@cemm.oeaw.ac.at. Tel: +43 1 40160 70 001. Fax: +43 1 40160 970 000.
28 08 2024
20 09 2024
19 9 19421952
02 02 2024
19 08 2024
02 08 2024
© 2024 The Authors. Published by American Chemical Society
2024
The Authors
https://creativecommons.org/licenses/by-nc-nd/4.0/ Permits non-commercial access and re-use, provided that author attribution and integrity are maintained; but does not permit creation of adaptations or other derivative works (https://creativecommons.org/licenses/by-nc-nd/4.0/).

The RAS family of GTPases is among the most frequently mutated proteins in human cancer, creating a high clinical demand for therapies that counteract their signaling activity. An important layer of regulation that could be therapeutically exploited is the proteostatic regulation of the main RAS GTPases KRAS, NRAS, and HRAS, as well as the closely related members, MRAS and RIT1, by the leucine zipper-like transcriptional regulator 1 cullin 3 RING E3 ubiquitin ligase complex (CUL3LZTR1). Genetic inactivation of LZTR1, as observed in different cancer entities and Noonan syndrome leads to enhanced RAS GTPase abundance and altered MAPK pathway activation state. Novel therapeutic approaches to interfere with hyperactive RAS signaling, thereby complementing existing treatments, are highly sought after. Motivated by the growing arsenal of molecular glue degraders, we report the identification of novel chemical fragments that enhance the protein–protein interaction (PPI) of the KRAS-LZTR1 complex. We established a split-luciferase-based reporter assay that monitors the RAS GTPase-LZTR1 interaction in a scalable format, capable of capturing chemical, as well as mutational perturbations. Using this screening system, in combination with a small fragment library, we identified two fragments, C53 and Z86, that enhance the interaction of the KRAS-LZTR1 complex in a dose-dependent manner. Further orthogonal validation experiments using proximity biotinylation (BioID), thermal shift assays, and NMR spectroscopy demonstrated fragment-dependent enhanced recruitment of endogenous LZTR1 and physical engagement of KRAS. The two fragments, which potentiate the KRAS-LZTR1 interaction, serve as starting points for fragment-based drug discovery. Additionally, the assay we introduced is amenable to high-throughput screening to further explore the pharmacological modulation of the CUL3LZTR1-RAS GTPase complex.

Ã&#150;sterreichischen Akademie der Wissenschaften 10.13039/501100001822 NA Magistrat der Stadt Wien 10.13039/501100010716 MUW-AP21005MWF Medizinische UniversitÃ¤t Wien 10.13039/501100005788 NA Austrian Science Fund 10.13039/501100002428 FWF SFB F4711 document-id-old-9cb4c00077
document-id-new-14cb4c00077
ccc-price
==== Body
pmcIntroduction

RAS family GTPases are known as one of the most significant oncogenes, with the KRAS gene displaying the highest frequency of mutations within the main RAS GTPase group (85%), followed by NRAS (11%) and HRAS (4%).1 These mutations are genetic drivers in numerous cancers and predominantly occur at the glycine 12 (G12) and 13 (G13) residues, constituting 81% and 14% of all KRAS mutations, respectively.2,3 The resulting structural changes hinder the arginine finger of associated GTPase activating proteins (GAPs) to hydrolyze GTP to GDP, shifting the KRAS pool toward its active GTP-bound state.4 Active KRAS facilitates its growth-promoting function by interacting with several downstream effector proteins, with RAF1 being a predominant member, leading to enhanced MAPK signaling.5 Despite its importance as a cancer driver, and potential as a therapeutic target, KRAS was long considered undruggable because it lacks any large hydrophobic pockets or an active site suitable to classical drug discovery-based approaches.3,6 In the past decade, a multitude of new drug discovery efforts have been launched to target KRAS, stirred by the discovery, and recent approval of covalent G12C-mutation targeting small molecules.7,8 However, moderate clinical efficacy, and the high frequency of resistance upon treatment with these agents, demonstrate the need for additional novel targeting approaches.9,10

Apart from mutations affecting the RAS GTPase itself, or its associated GAP and GEF interacting partners leading to enhanced RAS activity, alteration of the proteostatic regulation of individual RAS GTPase proteins has been identified as an additional important regulatory layer leading to MAPK pathway hyperactivation.11−13 The leucine zipper-like transcriptional regulator 1 (LZTR1) protein serves as a substrate receptor (SR) for the cullin 3 RING E3 ubiquitin ligase (CUL3LZTR1) complex leading to ubiquitination and altered abundance of the main RAS GTPases KRAS, NRAS, and HRAS as well as the closely related GTPases MRAS and RIT1.11−13 Structurally, LZTR1 belongs to the BTB/BACK domain-containing protein family, harboring an N-terminal Kelch domain followed by two BTB/BACK domains, whereby the Kelch domain acts as the substrate binding site and the BTB/BACK domains as interaction unit with the CUL3 scaffold.14 LZTR1-mediated RAS ubiquitination leads to reduced abundance, attenuating the activation and downstream signaling of the MAPK pathway. As a result, LZTR1 acts as a negative modulator of the RAS-MAPK signaling pathway by keeping RAS levels in check. Additionally, mutations of LZTR1 that have been identified in and associated with different human diseases, interfere with CUL3LZTR1 complex assembly and/or recruitment of RAS itself.12 Consequently, the loss of LZTR1 is associated with increased RAS GTPase protein levels, overactivation of RAS, and hyperactivation of the MAPK signaling pathway.11

In theory, if one could favor complex formation between LZTR1 and KRAS by using compounds acting like molecular glues, then one would open a new pharmacological strategy for the regulation of KRAS activity. Molecular glue degraders are one embodiment of targeted protein degradation (TPD) drugs that can strengthen preexisting PPIs or induce neo-PPIs with ubiquitin ligases, resulting in desired target protein degradation and therapeutic effects.15,16 In this study, we aimed to discover fragments that chemically modulate the LZTR1-KRAS PPI in a molecular glue-like manner, resulting in enhanced E3 ligase-substrate complex formation. We hypothesized that such fragments, when optimized for favorable kinetic properties and bioavailability, could lead to reduced RAS abundance and signaling output, as well as decreased downstream MAPK pathway activation. To this end, we aimed to develop a robust assay that could monitor changes in LZTR1-KRAS PPI formation. Through an iterative exploratory and optimization campaign, we designed a high-throughput screening-compatible split-luciferase-based PPI reporter assay to monitor LZTR1-KRAS recruitment in live cells. Screening of a fragment library at high concentrations led to the identification of two candidate fragments able to enhance the natural propensity of LZTR1 to bind KRAS. Orthogonal validation using proximity biotinylation (BioID), thermal shift assays, as well as NMR further corroborated our findings, providing a first proof of concept for the feasibility of small-molecule-based CUL3LZTR1-KRAS complex modulation in cells.

Results

Split-Luciferase-Based Assay (SLA) Allows for the Detection of RAS GTPase-LZTR1 PPI

To design a PPI assay that is high-throughput screening-compatible, we tested whether the NanoBiT system was suitable for the detection of the KRAS-LZTR1 interaction. The assay separates the small luciferase NanoLuc into two entities: a small tag (SmBiT, 11 amino acids) and a large tag (LgBiT, 18 kDa).17 Given that these two subunits have a weak affinity towards each other (Kd ∼ 190 μM),18 they only come into proximity when fused to two proteins that have sufficient affinity for each other to support complex formation, even if transiently. This PPI-induced proximity results in the complementation of the luciferase resulting in luminescent signal emission upon substrate addition.17 We tested detecting the interaction of LgBiT-tagged KRAS4A (LgBiTKRAS4A) with SmBiT-tagged LZTR1 (SmBiTLZTR1) or RAF1 (RAF1SmBiT) upon transient transfection of HEK293T cells (Figure 1A,B). Initially, we tried all orientation combinations of the tags, N- and C-terminally, except with RAS, where we solely placed the LgBiT tag on the N-terminus with the understanding that its C-terminal post-translational processing allows for localizing to the cellular membrane required for proper LZTR1 interaction.19 Cells transfected with a vector expressing the LgBiT-tagged mCherry fluorescent protein (LgBiTmCherry) served as a negative control. Cotransfection of LgBiTKRAS4A with SmBiTLZTR1 or RAF1SmBiT led to robust expression and a pronounced increase in luminescence compared to LgBiTmCherry indicative of successful PPI formation (Figure 1C and Supplementary Figure 1A). Moreover, we observed robust SLA luminescence induction with both KRAS splice isoforms (LgBiTKRAS4A and LgBiTKRAS4B) upon coexpression with SmBiTLZTR1 compared to LgBiTmCherry (Supplementary Figure 1B). LZTR1 mutations identified in Noonan syndrome patients as well as in different cancer entities have been shown to interfere with efficient KRAS proteostatic regulation. We therefore tested whether the LZTR1-KRAS SLA was able to capture mutational interference. Indeed, overexpression of the frequent LZTR1 G248R mutation led to a reduced PPI signal in comparison to WT protein (Figure 1D and Supplementary Figure 1C). Since LZTR1 has been shown to regulate several RAS GTPase family members, we tested whether the SLA was able to capture these interactions (Figure 1B). All tested RAS GTPases displayed a robust interaction signal, with LgBiTHRAS being the strongest one in our assay, while the negative control LgBiTmCherry as well as the non-LZTR1 substrate GTPase RAC1 only showed little signal increase above background (Figure 1E and Supplementary Figure 1D). The difference in signal strength in our assay could be at least partially a consequence of different expression levels, attributable to the use of transient transfection (Supplementary Figure 1D). With the understanding that LZTR1 also regulates the ubiquitination and degradation of the small GTPase RIT1, a member of the RAS family of GTPases, and that RIT1 mutations escape LZTR1-mediated proteostatic regulation, we next applied the same method and cotransfected HEK293T cells with LgBiTRIT1 WT and the M90I mutant (Supplementary Figure 1E).14 Correspondingly, LgBiTRIT1 M90I showed a reduced interaction signal compared to that of the WT protein (Figure 1F). Interestingly, we observed a similar pattern comparing LgBiTKRAS4A WT and the commonly observed oncogenic G12D mutant indicating that at least some KRAS mutants might escape LZTR1-based abundance regulation (Figure 1G and Supplementary Figure 1F). Furthermore, we assessed whether pharmacological perturbation of the CUL3 E3 ligase complex itself would lead to an altered interaction signal between LgBiTKRAS4A and SmBiTLZTR1. However, following treatment of transfected HEK293T cells with the E1 inhibitor TAK-243, the COP9 signalosome inhibitor CSN5i-3, or the neddylation inhibitor MLN4924, we did not observe a similar change in the interaction signal compared to that observed with mutational perturbation (Supplementary Figure 1G).

Figure 1 Split-luciferase assay (SLA)-based detection of RAS family GTPase-LZTR1 PPIs. (A) Scheme of SLA assay where two proteins of interest, either KRAS and LZTR1, or KRAS and RAF1, are fused to LgBiT or to SmBiT tags, respectively. After coexpression of the constructs in HEK293T cells and the addition of the live substrate, PPIs are detected. mCherry fused to LgBiT was used as a negative control. (B) Schematic outline of LgBiT- and SmBiT- containing fusion constructs drawn to scale based on their length. Numbering represents amino acid length (excluding stop codon). (C) HEK293T cells were cotransfected with LgBiTKRAS4A and SmBiTLZTR1 or RAF1SmBiT and afterwards analyzed by SLA. The expression of the LgBiTmCherry construct with SmBiTLZTR1 or RAF1SmBiT served as negative control. n = 3. (D) HEK293T cells were cotransfected with LgBiTKRAS4A and SmBiTLZTR1 or SmBiTLZTR1 G248R and analyzed by SLA. LgBiTmCherry + SmBiTLZTR1 served as negative control. n = 3. (E) HEK293T cells were cotransfected with different RAS GTPase family members (LgBiT-fused KRAS4A, NRAS, HRAS, MRAS, and RIT1) and SmBiTLZTR1 and analyzed by SLA. LgBiTmCherry and LgBiTRAC1 constructs coexpressed with SmBiTLZTR1 served as negative control. n = 3. (F, G) HEK293T cells were cotransfected with LgBiTKRAS4A WT G12D and SmBiTLZTR1 (F) as well as LgBiTRIT1 WT or M90I and SmBiTLZTR1 (G) and analyzed by SLA. LgBiTmCherry and SmBiTLZTR1 cotransfection served negative control. n = 3. Statistical significance was calculated with a two-way ANOVA with Dunnett’s multiple comparisons test correction (E) or two-tailed t test (D, F, and G). ns, nonsignificant; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001.

Fragment-Based Screen Identifies C53 as a Modulator of the LZTR1-KRAS Interaction

Fragment-based drug discovery (FBDD) aims at the identification of low-molecular-weight molecules (MW < 300 g/mol) and is particularly applicable for challenging target classes, including modulation of PPIs.20,21 In contrast to high-throughput compound libraries, fragment libraries are smaller, with low-potency molecules that can then be grown and optimized to form larger lead compounds.22 Based on the observation that the LZTR1-KRAS SLA can monitor mutation-induced interaction changes, we aimed to identify fragments that could increase the signal between LgBiTKRAS4A and SmBiTLZTR1, and thus potentially act as enhancers of this PPI in a molecular glue-like manner (Figure 2A). We screened a library of around 450 commercially available fragments at a concentration of 50 μM and identified fragment C53 (PC–C53-N; Z-score = 3.20), which showed a prominent increase in luminescence signal compared to the DMSO control (Figure 2B). Structurally, C53 is an N-acetyl-7,8-dichlorotetrahydroisoquinoline (Figure 2C).

Figure 2 Identification of fragments that enhance the KRAS-LZTR1 interaction. (A) Schematic of fragment screen using SLA in a 384-well plate format. (B) Luminescence for all fragments screened in cells coexpressing LgBiTKRAS4A WT and SmBiTLZTR1 using the SLA. Each dot represents an individual fragment with the hit C53 highlighted. (C, D) Chemical structure of fragment hits C53 and Z86. (E) Dose–response bar graph in HEK293T cells coexpressing LgBiTKRAS4A WT and SmBiTLZTR1 upon treatment with increasing concentrations of C53 or Z86 (10, 20, 30, 40, 50, and 60 μM, 0 μM corresponds to DMSO negative control treatment). n = 4. (F) Luminescence for all fragments and analogues screened in cells coexpressing LgBiTKRAS4A G12D and SmBiTLZTR1 using the SLA with C53 and Z86 highlighted. (G) Dose–response bar graph in HEK293T cells coexpressing LgBiTKRAS4A G12D and SmBiTLZTR1 upon treatment with increasing concentrations of C53 or Z86 (10, 20, 30, 40, 50, and 60 μM, 0 μM corresponds to DMSO negative control treatment). n = 3. Statistical significance was calculated with a two-way ANOVA with Dunnett’s multiple comparisons test correction. ns, nonsignificant; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001.

As the initial library that we screened was designed for diversity, with as little structural similarity among the fragments as possible, we wanted to further explore 50 analogues of C53. For this purpose, we designed a virtual make-on-demand library from Enamine, utilizing its REAL Space structure–activity relationship (SAR) search tool for small-molecule characterization and scaffold-hopping.23 The chosen analogues were screened in the same cellular model configuration, and we detected fragment Z86 (Z6466689386) displaying a modest improvement in PPI-modulating propensity (Supplementary Figure 2A). Z86 differed from C53, by containing a carboxylic acid (Figure 2D). Further single-well validation experiments of C53 and Z86 confirmed our initial observations from the screen revealing a dose-dependent signal increase of cells coexpressing SmBiTLZTR1 and LgBiTKRAS4A (Figure 2E). Interestingly, we observed that several of the selected analogues, contrary to C53 and Z86, led to a reduced interaction signal. To rule out cytotoxicity as a potential confounding factor in the analogue library screen, we evaluated the impact of our analogues on cell viability. Reassuringly, most of the screened analogues did not demonstrate any substantial antiproliferative effect, apart from fragment Z6739291816 (Supplementary Figure 2B). We speculate that analogues decreasing the luminescent signal in our SLA likely represent fragments that interact with one of the two targets but are sterically not tolerated in the PPI interface and thereby prevent SmBiTLZTR1 from engaging with LgBiTKRAS4A.

Following our observation that the KRAS G12D mutant showed reduced binding propensity to LZTR1, we rescreened both fragment libraries (original 450 fragments and 50 C53 analogs) in the context of LgBiTKRAS4A G12D and SmBiTLZTR1, to test whether the mutant KRAS protein would display a similar or altered fragment preference (Figure 2F). Interestingly, C53 (Z-score = 2.46) and Z86 (Z-score = 4.64) showed again a marked increase in luminescence signal compared to the DMSO control (Figure 2F). Moreover, C53 and Z86 also showed a dose-dependent increase in the signal in single-well validation experiments, corroborating the screen result (Figure 2G). Given that we were able to validate fragment activity in our SLA for both SmBiTLZTR1-LgBiTKRAS4A WT and SmBiTLZTR1-LgBiTKRAS4A G12D cells, we decided to pursue orthogonal validation experiments.

Enhanced Recruitment of the LZTR1-KRAS Complex in the Presence of C53

To test whether C53 would be able to enhance the recruitment of endogenous LZTR1, we used the cellular proximity biotinylation assay (BioID). Previously, we successfully used BioID to map the association of the four main RAS GTPases (fused to BirA*) with endogenously expressed LZTR1 in chronic myeloid leukemia (CML) cells.11 We stably expressed KRAS4A fused to miniTurboID, a BirA* derivative enzyme with shorter labeling times and improved sensitivity, in the CML cell line K-562.24 Empty vector-transduced cells, as well as cells expressing miniTurboID-GFP, served as a negative control. Cells were treated with C53 overnight, followed by the addition of biotin labeling for 2 h. As expected, we found an association of miniTurboID-fused KRAS4A with endogenous LZTR1 in the unperturbed state. Treatment with C53 led to enhanced recruitment and labeling of LZTR1, strongly supporting the notion that C53 acted by favoring the formation of the LZTR1-KRAS complex (Figure 3A).

Figure 3 Orthogonal validation experiments detect enhanced recruitment and fragment engagement. (A) Proximity biotinylation (miniTurboID) analysis in K-562 cells lentivirally transduced with the indicated miniTurboID-FLAG constructs. Cells were treated with 50 μM of C53 (+) or DMSO (−) overnight, followed by 50 μM biotin treatment for 2 h, cell lysis, StrepTactin enrichment, and subsequent SDS-PAGE and immunoblot analysis. Immunoblots are probed with the indicated antibodies. miniTurboID-GFP fusion protein served as negative control. BioID results shown are representative of two independent biological experiments (n = 2). (B) Thermal shift assay analysis of C53. HEK293T cells coexpressing LgBiTKRAS4A and SmBiTLZTR1 were lysed and treated with C53 (50 μM) for 1 h. The samples were separated by SDS-PAGE following immunoblotting analysis. Immunoblots show the thermostability of endogenous RAS and LgBiTKRAS4A monitored with anti-pan-RAS antibody staining, following heat treatment at a temperature range of 40.9–61.1 °C, in the absence (−) or presence (+) of C53. (C) Band intensity of the thermal shift immunoblot films was quantified and normalized to the band intensity of the α-tubulin control. The fold change of endogenous RAS signal is displayed in the presence and absence of C53 between 40.9 and 49.1 °C. (D–E) Same as (B) and (C) except treatment with Z86. Thermal shift assay quantifications are based on two independent and representative sets (n = 2).

Fragment Engagement with KRAS

A ligand-target interaction can be monitored by thermal shift assay, where ligand-engagement thermodynamically stabilizes its target, delaying temperature-dependent denaturation.25 We tested whether the thermal shift assay would allow target engagement detection of the fragments with either LZTR1 or KRAS. For this purpose, we lysed HEK293T cells coexpressing LgBiTKRAS4A and SmBiTLZTR1, treated the lysates with 50 μM C53, Z86, or DMSO for 1 h, and subsequently incubated the samples at an increasing temperature range before immunoblotting analysis. Compared to vehicle control, we detected thermal stabilization of the endogenous RAS pool within the applied temperature range (40.9 to 53.1 °C) in C53 or Z86 fragment-treated cell lysates (C53, Figure 3B and C; Z86, Figure 3D and E). Interestingly, we did not observe a thermal shift to the same extent for the LgBiTKRAS4A fusion protein, which could potentially be attributed to an altered thermal stabilization property of the LgBiT-tagged KRAS4A. These data provide further evidence for target engagement of both fragments with endogenous (K-)RAS. However, in contrast to our BioID findings, we failed to detect any stabilization of endogenous and SmBiT-tagged LZTR1 upon fragment incubation.

NMR Reveals Binding of Fragments to the Switch I/II Pocket of KRAS G12D

15N-labeled GDP-bound KRAS G12D was utilized to further validate fragment binding with 2D NMR spectroscopy. To determine ligand binding, we monitored the changes in the 1H and 15N chemical shifts in 1H–15N HSQC spectra. The addition of compounds C53 and Z86 to GDP-bound KRAS G12D led to chemical shift perturbations (CSPs) of a similar set of peaks, indicative of an identical binding site for the two compounds (Figure 4A and Supplementary Figure 3A). Amide backbone chemical shifts were available for GDP-KRAS G12D, and the chemical shift changes upon fragment addition were used to map the binding site (Figure 4B and Supplementary Figure 3B). Among the peaks that showed no overlap in the 2D spectra and could be unambiguously assigned, L56, R73, T74, and G75 showed CSPs more than twice the standard deviation of the mean upon C53 addition (Figure 4B). These residues coincide with a conserved RAS surface pocket referred to as the switch I/II pocket.26−28 The magnitudes of CSPs induced by Z86 were smaller; however, for residues S39, T74, and G75, significant CSPs were observed (Supplementary Figure 3B).

Figure 4 NMR-based C53 binding validation and characterization. (A) Overlay of 1H–15N HSQC spectra of 100 μM GDP-KRAS G12D in the absence (red) and presence (black) of 2.5 mM C53. The four largest amide backbone CSPs are indicated by black arrows. (B) Histogram showing CSP values for backbone amide groups of 15N-labeled GDP-KRAS G12D in the presence of 2.5 mM C53. Horizontal dotted gray and black lines indicate one and two standard deviations of the CSP values, respectively. Residues without a bar were not assigned or overlapped in the 2D spectra. (C) Ribbon and surface representation of KRAS GDP (PDB ID 4EPW) with nitrogen atoms shown as blue spheres for residues that showed CSP values larger than 2 standard deviations of the mean value of all CSPs in the NMR titration experiment and are located in the SI/II pocket. (D) Binding site docking pose of C53, the compound is shown in yellow sticks representation and the protein (PDB ID 4EPW) is represented as in (C).

Binding Pose Prediction of C53 by Molecular Docking

To generate a model of the potential binding pose of C53, a binding site docking simulation, based on the CSP data, was performed using AutoDock Vina.29 In two of the docking poses, the benzene ring and the chloro groups of C53 are inserted into the SI/II pocket, which is formed by V7, L56, and Y71, as well as the aliphatic parts of the K5 and T74 side chains (Figure 4D and Supplementary Figure 3C). These ligand orientations are similar to the ligand poses in the X-ray structures determined by Maurer et al. and Sun et al. (Supplementary Figure 3D,E). These docking poses suggest that the piperidine and propanone motifs of C53 may be mediating the interaction with LZTR1.

Discussion

In this study, we aimed to design a high-throughput chemical screening method to exploit the proteostatic interaction between the CUL3LZTR1 complex and the RAS family of GTPases. This approach seeks to identify small molecules capable of altering the equilibrium of the naturally occurring PPI between these proteins, providing new therapeutic avenues to mitigate cancerous RAS signaling. Our first step was to establish a screening modality that, once scaled up to an industrial level, could warrant a drug discovery campaign. We validated this approach by identifying two small-molecule fragments, C53 and Z86, that enhanced the KRAS-LZTR1 interaction, with evidence for enhanced complex formation using BioID, thermal shift assays as well as NMR spectroscopy.

Diverse therapeutic strategies have been pursued in the past to target RAS for cancer therapy. Given the difficulty of targeting RAS itself, the initial focus has been on interfering with its post-translational processing, a prerequisite for its efficient signaling activity, as exemplified by the development of farnesyl transferase inhibitors. However, following farnesyl transferase inhibition, especially KRAS and NRAS mutated cancers failed to demonstrate encouraging activity due to alternative geranylgeranyl transferase modification.30 Other non-covalent small binding molecules have been discovered by means of NMR-based fragment screening, yielding some promising compounds.6 Unfortunately, these compounds also prevent the normal function of RAS and might have little effect on RAS mutants, as well as potentially result in enhanced toxicities within healthy tissues.6 Only recently, notable milestones have been reached with the development of Sotorasib (AMG-510), a covalent ligand, that binds to the distinct KRAS mutant allele G12C, showing the first promising results in clinical trials of KRAS mutant cancers.31 Nevertheless, clinical data has illustrated that even direct targeting of KRAS G12C only provides an intermittent control of tumor growth, requiring potential additional combinatorial treatments for sufficient long-term disease control.31

Considering that an imbalance of protein turnover is associated with cancer development and can create selective therapeutic vulnerabilities,32 the growing field of targeted protein degradation (TPD) offers a compelling new approach for targeting RAS-driven cancers. First TPD methods have attempted to degrade KRAS with bifunctional molecules (PROTACs) hijacking the E3 ligase adapter proteins such as Cereblon (CRBN)33 and Von Hippel-Lindau Tumor Suppressor (VHL).34,35 Despite the promise of PROTAC-based strategies, their clinical utility can be limited by three key factors: their large size, poor drug-like properties, and the fact that they do not exploit endogenous substrate-ligase pairs as molecular glue-like small molecules do.

Therefore, as an alternative strategy, we directed our efforts toward developing proximity-inducing fragments to chemically target the native substrate-ligase interaction of KRAS, specifically the KRAS-LZTR1 PPI interface. Similar efforts to chemically enhance the existing affinity of a PPI interface have been made, such as the identification of the small molecule NRX-252114 by Simonetta et al. NRX-252114 strengthens the interaction between the oncogenic transcription factor β-Catenin and its associated E3 ligase, SCFβ-TrCP, resulting in enhanced ubiquitination and degradation of mutant β-Catenin.36 Interestingly, recent efforts to identify novel molecular glue degraders with antiproliferative effects in transformed cells have revealed several candidates that potentiate the weak affinity between substrate proteins and E3 ligase complexes. Notably, Kozicka et al. identified small-molecule degraders that enhance the minimal affinity between CDK12-cyclin K and the DDB1-CUL4-RBX1 E3 ligase complex, leading to substrate degradation.37

Despite the robustness and reproducibility of the screening setup presented here, several aspects warrant further investigation and improvement. First, although we successfully employed the SLA constructs using transient transfection, we were unable to achieve stable expression of these constructs in cells with a screening-compatible signal-to-noise ratio. Additionally, the assay variability observed in some instances might be due to expression variability resulting from transient transfection. Second, we have not detected any proteostatic consequences of fragment action, potentially due to the low affinity of our current fragments. Finally, our study has been limited by the lack of physical structural data on the LZTR1 protein. Despite our efforts, we have not been able to obtain suitable amounts of soluble, properly folded LZTR1 protein to sufficiently assess ternary complex formation, evaluate direct LZTR1 target engagement in vitro, and enable further hit expansion to increase the potency of our fragments.

Nevertheless, the effects of C53 and Z86 observed in living cells demonstrate the engagement of both endogenous LZTR1 and RAS as well as the frequently mutated KRAS G12D variant. This provides an attractive starting ground for developing a novel class of RAS GTPase-targeting agents that engage the CUL3-based E3 ligase complex CUL3LZTR1. We are convinced that our study will advance the biochemical toolbox for screening proximity-inducing drugs and pave the way for optimizing fragments that could proteostatically regulate KRAS. Ultimately, enhanced degradation of KRAS, facilitated by potentiating the KRAS-LZTR1 PPI interface, could bolster existing RAS and MAPK pathway-focused pharmacological interventions, leading to a more durable treatment of RAS-driven cancers.

Materials and Methods

Chemicals

Compound libraries and fragments C53 (PC-C53-N, Z1861995405) and Z86 (Z6466689386) were obtained from Enamine. The structures of C53 and Z86 are shown in Figure 2C and D. Additional reagents were used as follows: TAK-243 (S8341, Selleckchem, Houston, TX), CSN5i-3 (HY-112134, MedChemExpress, Monmouth Junction, NJ), and MLN4924 (S7109, Selleckchem). All chemicals were dissolved in DMSO (D5879, Sigma-Aldrich, St. Louis, MI).

Cell Lines

HEK293T cells were obtained from ATCC (Manassas, VA) and K-562 cells were from DSMZ (Braunschweig, Germany). HEK293T cells were cultured in DMEM (D5796, Sigma-Aldrich, St. Louis, MO) and K-562 cells were cultured in RPMI1640 medium (R8758, Sigma) both supplemented with 10% (v/v) FCS (S1810-500, Biowest, Riverside, MO) and antibiotics (100 U/mL penicillin and 100 mg mL–1 streptomycin; P4333, Sigma). Cells were cultured at 37 °C and 5% CO2, authenticated by STR profiling, and checked for mycoplasma infection by PCR or ELISA regularly.

Plasmids

For split-luciferase-based reporter assays, the LgBiT cDNA, derived from pBiT1.1-N (Promega, Madison), containing a short linker sequence and a FLAG tag, was cloned in frame with mCherry or the indicated human GTPase-encoding cDNAs into the LEIH (pRRL-EF1a-IRES-HygroR) expression vector using the NEBuilder HiFi DNA Assembly Master Mix (NEB, Ipswich, MA). The SmBiT tag sequence was added at the N-terminus of LZTR1 and the C terminus of RAF1 using the Q5 Site-Directed Mutagenesis Kit (NEB). SmBiT-LZTR1 and RAF1-SmBiT cDNAs in gateway-compatible pDONR221 or pDONR223 vectors were transferred by LR recombination (11791100, Thermo Fisher Scientific, Waltham, MA) into the LEgwSHIB (pRRL-EF1a-gateway-StrepHA-IRES-BlastR) expression vector. pDONR221 or pDONR223 entry plasmids containing the coding sequence of mCherry or human KRAS4A, KRAS4B, NRAS, HRAS, RIT1, RAC1, LZTR1, and RAF1 have been described previously.11 KRAS4A G12D, RIT1 M90I, and LZTR1 G248R mutations were performed using the Q5 Site-Directed Mutagenesis Kit (NEB).

For proximity biotinylation experiments, stable lentiviral expression vectors were generated by insertion of the miniTurboID cDNA, derived from pcDNA3-V5 miniTurbo-NES (Addgene plasmid #107170), with a FLAG tag-gateway (gw) cassette into the LEIH (pRRL-EF1a-IRES-HygroR) vector using the NEBuilder HiFi DNA Assembly Master Mix (NEB). GFP and KRAS4A cDNAs in gateway-compatible pDONR221 were transferred into LEmTIDFgwIH (pRRL-EF1A-miniTurboID-FLAG-gateway-IRES-HygroR) by LR recombination.

Lentiviral Transduction

To generate ecotropic receptor (EcoR)-expressing K-562 cells, HEK293T were transiently transfected with LERZIE (pRRL-EF1a-rtTA3-P2A-ZeoR-IRES-EcoR) as well as psPAX2 (Addgene plasmid #12260) and pMD2.G (Addgene plasmid #12259) packaging plasmids using polyethylenimine (PEI) as previously described.11 24 h post transfection, medium was replaced, and virus-containing supernatant was harvested after 48 h, filtered through 0.45 μm sterile filters (6780-2504, Cytiva, Marlborough, MA), supplemented with 8 μg/mL protamine sulfate (Sigma) and added to K-562 target cells followed by selection of transduced cells using zeocin (ant-zn-1, InvivoGen, San Diego, CA).

For proximity biotinylation experiments, lentiviral supernatant was prepared as mentioned above using the respective empty vector, LEmTIDFgwIH-GFP or -KRAS4A lentiviral expression vectors, and pEcoEnv envelope plasmid instead of pMD2.G. Virus-containing supernatant was applied to K-562EcoR cells followed by selection of transduced cells using hygromycin (ant-hg-1, InvivoGen).

Split-Luciferase Assay (SLA)

HEK293T cells were seeded into 6-well plates. After 24 h, SmBiT- and LgBiT-fusion constructs were transiently cotransfected using polyethylenimine (PEI). SmBiT- and LgBiT-constructs were cotransfected in equal amounts (3 μg total plasmid DNA). 48 h after transfection, cells were washed with phosphate-buffer saline (PBS) (D8537, Sigma), detached using trypsin (T3924, Sigma), and harvested in DMEM supplemented with 5% (v/v) FCS and antibiotics. Following cell counting-based normalization, equal cell amounts were seeded in 384-well flat clear-bottom white assay plates (3765, Corning, Corning, NY), Nano-Glo Live Cell Reagent (N2014, Promega, Madison, WI) was added, and after a 10 min incubation time, luminescence was recorded on a SpectraMax i3x microplate reader (Molecular Devices, San José, CA).

Fragment-Based High-Throughput Screen

HEK293T were transiently transfected with either LgBiTKRAS4A WT and SmBiTLZTR1 or LgBiTKRAS4A G12D and SmBiTLZTR1 as described above and seeded at a concentration of 20,000 cells per well in 384-well flat clear-bottom white assay plates containing the spotted compounds at 50 μM or DMSO as negative control. Following incubation for 18 h, Nano-Glo Live Cell Reagent was added and after 10 min incubation, luminescence was recorded on a SpectraMax i3x microplate reader (Molecular Devices). Hit thresholds were defined by >20% (LgBiTKRAS4A WT+SmBiTLZTR1) and >15% (LgBiTKRAS4A G12D+SmBiTLZTR1) enhancement of interaction compared to DMSO control as well as a Z-score threshold of ≥2.5 for mean percent luminescence normalized to DMSO control.

Immunoblotting

Cells were lysed using Nonidet-40 lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.5% NP-40, and one tablet of Roche EDTA-free protease inhibitor cocktail (Sigma-Aldrich) per 50 mL) for 10 min on ice. Lysates were cleared by centrifugation (13000 rpm, 10 min, 4 °C), and proteins were subsequently quantified and normalized with Bradford assay using γ-globin as a standard (Bio-Rad, Hercules, CA). Cell lysates were resolved by SDS-PAGE and transferred to Protran BA 85 nitrocellulose membranes (GE Healthcare, Little Chalfont, UK). The membranes were immunoblotted with indicated antibodies, and bound antibodies were visualized with horseradish peroxidase-conjugated secondary antibodies using the ECL Western blotting system (Thermo Fisher Scientific).

Antibodies used were: mouse monoclonal anti-RAS clone RAS10 (Millipore, 05-516, 1:500 dilution), mouse monoclonal anti-LZTR1 (E-12) (Santa Cruz, SACSC-390166, 1:200 dilution), rabbit monoclonal anti-c-RAF (Cell Signaling, 9422S, 1:1000 dilution), mouse monoclonal anti-FLAG M2 (Sigma, F1804, 1:1000 dilution), and mouse monoclonal anti-α-tubulin (Abcam, ab7291, 1:10000 dilution). The secondary antibodies used were goat anti-mouse HRP (115-035-003, Jackson ImmunoResearch, West Grove, PA) and goat anti-rabbit HRP (111-035-003, Jackson ImmunoResearch).

Thermal Shift Assay

LgBiT and SmBiT constructs were transiently expressed in HEK293T cells, as described above. After 48 h, cells were harvested and washed with PBS prior to lysis in Nonidet-40 lysis buffer and lysates were prepared as described above. Equal amounts of cell lysates were incubated with 50 μM of the respective fragment or DMSO as negative control for 1 h on ice. After compound incubation, 30 μL of the lysates were heated in individual tubes in a thermocycler at different temperatures for 6 min and then cooled for 3 min on ice. Subsequently, the samples were centrifuged at 13 000 rpm for 40 min at 4 °C, and supernatants were transferred to new tubes, mixed with 4× Laemmli buffer, and analyzed by SDS-PAGE and immunoblotting with the indicated antibodies.

Proximity Biotinylation (miniTurboID)

K-562EcoR cells expressing empty vector, miniTurboID-eGFP or -KRAS4A were treated with 50 μM C53 for 18 h followed by treatment with 50 μM biotin for 2 h, washed with PBS, and subsequently lysed with Nonidet-40 lysis buffer. Protein concentration was determined using Bradford assay and 15 mg of cell lysates were incubated with StrepTactin sepharose beads (2-1201-010, IBA Lifesciences, Göttingen, Germany) for 2 h at 4 °C. Beads were recovered by centrifugation and washed three times with lysis buffer, bound proteins were eluted by addition of 4x Laemmli buffer and boiling for 5 min before analysis by SDS-PAGE and immunoblotting.

Protein Expression and Purification of KRAS4B G12D

Uniformly 15N-labeled KRAS4B G12D (residues 1–169) was expressed in Escherichia coli BL21 (DE3) with an N-terminal His6-tag followed by a TEV (tobacco etch virus) protease cleavage site. Cells were grown at 37 °C in M9 minimal media containing 15NH4Cl as a sole nitrogen source in the presence of kanamycin until OD600 ≈ 0.6, then the temperature was lowered to 18 °C, and after 45 min protein synthesis was induced by adding isopropyl-β-D-thiogalactopyranoside (IPTG) to a final concentration of 0.4 mM. Expression was carried out overnight. Cells were lysed by sonication in lysis buffer containing 20 mM Tris-HCl (pH 7.5), 300 mM NaCl, 10 mM imidazole, 1 mM dithiothreitol (DTT), 1 mM phenylmethylsulfonyl fluoride (PMSF) and afterward clarified by centrifugation. Subsequently, the supernatant was loaded onto a HisTrap FF crude (Cytiva, Marlborough, MA) column. After elution, the buffer was exchanged on a HiPrep Desalting (Cytiva) column to 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 1 mM TCEP, then the His–tag was cleaved by incubation with TEV. The His-tag was removed by passing the sample through a second HisTrap FF crude (Cytiva) column. The protein solution was afterward supplemented with 1 mg GDP per 20 mg protein and 5 mM MgCl2. In a final purification step, GDP-KRAS4BG12D was purified to homogeneity by size exclusion chromatography using a Superdex 75 (Cytiva).

NMR Spectroscopy

NMR spectra were recorded at 25 °C on an Avance III HD 800 MHz (18.8 T) spectrometer. Data were processed using TopSpin 4 (Bruker BioSpin) and analyzed using CcpNmr.38 C53 titration experiments were carried out on a 100 μM 15N uniformly labeled GDP-KRAS G12D solution in PBS, pH 7.4, 1 mM TCEP in 5% D2O/95% H2O with 50 mM C53 stock solutions in dimethyl sulfoxid-d6 (DMSO-d6). Z86 titration experiments were carried out on a 50 μM 15N uniformly labeled GDP-KRAS G12D solution in 100 mM HEPES buffer, pH 7.4, 150 mM NaCl, 1 mM TCEP in 5% D2O/95% H2O with 50 mM Z86 stock solutions in dimethyl sulfoxid-d6 (DMSO-d6). The ligands were titrated to final concentrations of 0.500, 1.0, 1.5, and 2.5 mM. Total DMSO-d6 concentration was kept constant at 5% for all NMR measurements. At each titration step chemical shift changes upon ligand addition were monitored with 1H–15N HSQC experiments. The combined chemical shift perturbation of 1HN and 15NH was calculated as . The GDP-KRAS G12D assignment with BMRB ID 27719 was used for the CSP analysis.

Molecular Docking Simulations of C53

In the docking calculation, the crystal structure of GDP-KRAS (PDB ID 4EPW) was employed as the receptor. Before protonation at physiological pH and conversion to the PDBQT file format using AutoDockTools,39 all HETATM lines were removed from the PDB file. The docking search space in the simulation was centered on the switch I/II pocket, encompassing the residues that exhibited chemical shift perturbations in the NMR experiment. C53 was prepared using ChemDraw 22.2.0 (Revvity Signals) and converted to the PDBQT format with AutoDockTools. It was then docked into the rigid receptor using AutoDock Vina 1.20.29 The calculation utilized default settings with an exhaustiveness level of 32. The docking poses aligned best with the X-ray structures from Sun et al. (PDB ID 4EPW) and Maurer et al. (PDB ID 4DST), featured the benzyl group and chlorine atoms inserted into the switch I/II pocket and exhibited the second and fourth lowest estimated free binding energies.

Data Analysis

The band intensity of the thermal shift immunoblots was analyzed using Image Lab Software (version 6.1, Bio-Rad). Data organization and calculations were performed using Microsoft Excel (Microsoft, Redmond, WA) unless otherwise stated. Fragment screen analysis was performed using the R programming environment within RStudio (Posit PBC, Boston, MA). Luminescent signal bar graphs were analyzed in GraphPad Prism 9 (version 9.4.0). Experiments were carried out in typical independent triplicate sets (n = 3) unless otherwise stated. Statistical significance of individual split-luciferase assay results was calculated with a two-tailed t test or for comparison of multiple conditions a two-way ANOVA with Dunnett’s multiple comparisons test correction. ns, nonsignificant; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acschembio.4c00077.Additional immunoblot and fragment screening as well as NMR results (PDF)

Supplementary Material

cb4c00077_si_001.pdf

Author Contributions

S.P., J.W.B., and G.S.-F. designed research; S.P., J.W.B., J.W., and J.G.M. performed research; S.B., K.M.S., and R.K. designed, performed, and analyzed the NMR research; S.P., J.W., J.G.M., J.W.B., and G.S.-F. analyzed and interpreted the data; S.P., J.W.B., and G.S.-F. wrote the paper with contributions from all other coauthors.

The authors declare the following competing financial interest(s): G.S.-F. is co-founder and owns shares of Proxygen GmbH and Solgate GmbH. G.S.-F. receives research funding from Pfizer. R.K. is co-founder and scientific head of MAG-LAB GmbH. The other authors declare no competing financial interest.

Acknowledgments

The authors acknowledge the members of the Superti-Furga laboratory for critical discussions and suggestions. They thank the Molecular Discovery Platform at CeMM for assistance with the fragment-based screen. They are grateful to G. Winter and his laboratory for reading the manuscript, providing critical feedback, and providing reagents. This work was supported by the Austrian Academy of Sciences (to G.S.-F., S.P., J.W., J.G.M.), the Medical University of Vienna (to J.W.B.), the Austrian Science Fund (FWF SFB F4711 to G.S.-F.), and the Medical Scientific Fund of the Mayor of the City of Vienna (MUW-AP21005MWF to J.W.B.). Plasmids obtained through Addgene were a gift from D. Trono and A. Ting.
==== Refs
References

Hobbs G. A. ; Der C. J. ; Rossman K. L. RAS isoforms and mutations in cancer at a glance. J. Cell Sci. 2016, 129 (7 ), 1287–1292. 10.1242/jcs.182873.26985062
Moore A. R. ; Rosenberg S. C. ; McCormick F. ; Malek S. RAS-targeted therapies: is the undruggable drugged?. Nat. Rev. Drug Discovery 2020, 19 (8 ), 533–552. 10.1038/s41573-020-0068-6.32528145
Uras I. Z. ; Moll H. P. ; Casanova E. Targeting KRAS Mutant Non-Small-Cell Lung Cancer: Past, Present and Future. Int. J. Mol. Sci. 2020, 21 (12 ), 4325 10.3390/ijms21124325.32560574
Jiao D. ; Yang S. Overcoming Resistance to Drugs Targeting KRAS(G12C) Mutation. Innovation 2020, 1 (2 ), 100035 10.1016/j.xinn.2020.100035.32939510
Herdeis L. ; Gerlach D. ; McConnell D. B. ; Kessler D. Stopping the beating heart of cancer: KRAS reviewed. Curr. Opin. Struct. Biol. 2021, 71 , 136–147. 10.1016/j.sbi.2021.06.013.34303932
McCormick F. Progress in targeting RAS with small molecule drugs. Biochem. J. 2019, 476 (2 ), 365–374. 10.1042/BCJ20170441.30705085
Hong D. S. ; Fakih M. G. ; Strickler J. H. ; Desai J. ; Durm G. A. ; Shapiro G. I. ; et al. KRAS(G12C) Inhibition with Sotorasib in Advanced Solid Tumors. N. Engl. J. Med. 2020, 383 (13 ), 1207–1217. 10.1056/NEJMoa1917239.32955176
Skoulidis F. ; Li B. T. ; Dy G. K. ; Price T. J. ; Falchook G. S. ; Wolf J. ; et al. Sotorasib for Lung Cancers with KRAS p.G12C Mutation. N. Engl. J. Med. 2021, 384 (25 ), 2371–2381. 10.1056/NEJMoa2103695.34096690
Awad M. M. ; Liu S. ; Rybkin I. I. ; Arbour K. C. ; Dilly J. ; Zhu V. W. ; et al. Acquired Resistance to KRASG12C Inhibition in Cancer. N. Eng. J. Med. 2021, 384 (25 ), 2382–2393. 10.1056/NEJMoa2105281.
Weinberg R. A. It took a long, long time: Ras and the race to cure cancer. Cell 2024, 187 (7 ), 1574–1577. 10.1016/j.cell.2024.02.042.38552606
Bigenzahn J. W. ; Collu G. M. ; Kartnig F. ; Pieraks M. ; Vladimer G. I. ; Heinz L. X. ; et al. LZTR1 is a regulator of RAS ubiquitination and signaling. Science 2018, 362 (6419 ), 1171–1177. 10.1126/science.aap8210.30442766
Steklov M. ; Pandolfi S. ; Baietti M. F. ; Batiuk A. ; Carai P. ; Najm P. ; et al. Mutations in LZTR1 drive human disease by dysregulating RAS ubiquitination. Science 2018, 362 (6419 ), 1177–1182. 10.1126/science.aap7607.30442762
Castel P. ; Cheng A. ; Cuevas-Navarro A. ; Everman D. B. ; Papageorge A. G. ; Simanshu D. K. ; et al. RIT1 oncoproteins escape LZTR1-mediated proteolysis. Science 2019, 363 (6432 ), 1226–1230. 10.1126/science.aav1444.30872527
Abe T. ; Umeki I. ; Kanno S.-i. ; Inoue S.-i. ; Niihori T. ; Aoki Y. LZTR1 facilitates polyubiquitination and degradation of RAS-GTPases. Cell Death Differentiation 2020, 27 (3 ), 1023–1035. 10.1038/s41418-019-0395-5.31337872
Stanton B. Z. ; Chory E. J. ; Crabtree G. R. Chemically induced proximity in biology and medicine. Science 2018, 359 (6380 ), eaao5902 10.1126/science.aao5902.29590011
Du X. ; Li Y. ; Xia Y. L. ; Ai S. M. ; Liang J. ; Sang P. ; et al. Insights into Protein-Ligand Interactions: Mechanisms, Models, and Methods. Int. J. Mol. Sci. 2016, 17 (2 ), 144 10.3390/ijms17020144.26821017
Cooley R. ; Kara N. ; Hui N. S. ; Tart J. ; Roustan C. ; George R. ; et al. Development of a cell-free split-luciferase biochemical assay as a tool for screening for inhibitors of challenging protein-protein interaction targets. Wellcome Open Res. 2020, 5 , 20 10.12688/wellcomeopenres.15675.1.32587898
Dixon A. S. ; Schwinn M. K. ; Hall M. P. ; Zimmerman K. ; Otto P. ; Lubben T. H. ; et al. NanoLuc Complementation Reporter Optimized for Accurate Measurement of Protein Interactions in Cells. ACS Chem. Biol. 2016, 11 (2 ), 400–408. 10.1021/acschembio.5b00753.26569370
Hancock J. F. ; Parton R. G. Ras plasma membrane signalling platforms. Biochem. J. 2005, 389 (Pt 1 ), 1–11. 10.1042/BJ20050231.15954863
Erlanson D. A. ; McDowell R. S. ; O’Brien T. Fragment-Based Drug Discovery. J. Med. Chem. 2004, 47 (14 ), 3463–3482. 10.1021/jm040031v.15214773
Wang Z. Z. ; Shi X. X. ; Huang G. Y. ; Hao G. F. ; Yang G. F. Fragment-based drug discovery supports drugging ’undruggable’ protein-protein interactions. Trends Biochem. Sci. 2023, 48 (6 ), 539–552. 10.1016/j.tibs.2023.01.008.36841635
Murray C. W. ; Rees D. C. The rise of fragment-based drug discovery. Nat. Chem. 2009, 1 (3 ), 187–192. 10.1038/nchem.217.21378847
Grygorenko O. O. ; Radchenko D. S. ; Dziuba I. ; Chuprina A. ; Gubina K. E. ; Moroz Y. S. Generating Multibillion Chemical Space of Readily Accessible Screening Compounds. iScience 2020, 23 (11 ), 101681 10.1016/j.isci.2020.101681.33145486
Branon T. C. ; Bosch J. A. ; Sanchez A. D. ; Udeshi N. D. ; Svinkina T. ; Carr S. A. ; et al. Efficient proximity labeling in living cells and organisms with TurboID. Nat. Biotechnol. 2018, 36 (9 ), 880–887. 10.1038/nbt.4201.30125270
Molina D. M. ; Jafari R. ; Ignatushchenko M. ; Seki T. ; Larsson E. A. ; Dan C. ; et al. Monitoring drug target engagement in cells and tissues using the cellular thermal shift assay. Science 2013, 341 (6141 ), 84–87. 10.1126/science.1233606.23828940
Maurer T. ; Garrenton L. S. ; Oh A. ; Pitts K. ; Anderson D. J. ; Skelton N. J. ; et al. 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 (14 ), 5299–5304. 10.1073/pnas.1116510109.22431598
Sun Q. ; Burke J. P. ; Phan J. ; Burns M. C. ; Olejniczak E. T. ; Waterson A. G. ; et al. Discovery of small molecules that bind to K-Ras and inhibit Sos-mediated activation. Angew. Chem., Int. Ed. 2012, 51 (25 ), 6140–6143. 10.1002/anie.201201358.
Kessler D. ; Gmachl M. ; Mantoulidis A. ; Martin L. J. ; Zoephel A. ; Mayer M. ; et al. Drugging an undruggable pocket on KRAS. Proc. Natl. Acad. Sci. U. S. A. 2019, 116 (32 ), 15823–15829. 10.1073/pnas.1904529116.31332011
Eberhardt J. ; Santos-Martins D. ; Tillack A. F. ; Forli S. AutoDock Vina 1.2.0: New Docking Methods, Expanded Force Field, and Python Bindings. J. Chem. Inf. Model. 2021, 61 (8 ), 3891–3898. 10.1021/acs.jcim.1c00203.34278794
Gilardi M. ; Wang Z. ; Proietto M. ; Chillà A. ; Calleja-Valera J. L. ; Goto Y. ; et al. Tipifarnib as a Precision Therapy for HRAS-Mutant Head and Neck Squamous Cell Carcinomas. Mol. Cancer Ther. 2020, 19 (9 ), 1784–1796. 10.1158/1535-7163.MCT-19-0958.32727882
Skoulidis F. ; Li B. T. ; Dy G. K. ; Price T. J. ; Falchook G. S. ; Wolf J. ; et al. Sotorasib for Lung Cancers with KRAS p.G12C Mutation. N. Engl. J. Med. 2021, 384 (25 ), 2371–2381. 10.1056/NEJMoa2103695.34096690
Schlierf A. ; Altmann E. ; Quancard J. ; Jefferson A. B. ; Assenberg R. ; Renatus M. ; et al. Targeted inhibition of the COP9 signalosome for treatment of cancer. Nat. Commun. 2016, 7 (1 ), 13166 10.1038/ncomms13166.27774986
Zeng M. ; Xiong Y. ; Safaee N. ; Nowak R. P. ; Donovan K. A. ; Yuan C. J. ; et al. Exploring Targeted Degradation Strategy for Oncogenic KRAS(G12C). Cell Chem. Biol. 2020, 27 (1 ), 19–31.e6. 10.1016/j.chembiol.2019.12.006.31883964
Bond M. J. ; Chu L. ; Nalawansha D. A. ; Li K. ; Crews C. M. Targeted Degradation of Oncogenic KRAS(G12C) by VHL-Recruiting PROTACs. ACS Cent Sci. 2020, 6 (8 ), 1367–1375. 10.1021/acscentsci.0c00411.32875077
Popow J. ; Farnaby W. ; Gollner A. ; Kofink C. ; Fischer G. ; Wurm M. Targeting cancer with small molecule pan-KRAS degraders bioRxiv, 2023.
Simonetta K. R. ; Taygerly J. ; Boyle K. ; Basham S. E. ; Padovani C. ; Lou Y. ; et al. Prospective discovery of small molecule enhancers of an E3 ligase-substrate interaction. Nat. Commun. 2019, 10 (1 ), 1402 10.1038/s41467-019-09358-9.30926793
Kozicka Z. ; Suchyta D. J. ; Focht V. ; Kempf G. ; Petzold G. ; Jentzsch M. ; et al. Design principles for cyclin K molecular glue degraders. Nat. Chem. Biol. 2024, 20 (1 ), 93–102. 10.1038/s41589-023-01409-z.37679459
Vranken W. F. ; Boucher W. ; Stevens T. J. ; Fogh R. H. ; Pajon A. ; Llinas M. ; et al. The CCPN data model for NMR spectroscopy: development of a software pipeline. Proteins 2005, 59 (4 ), 687–696. 10.1002/prot.20449.15815974
Morris G. M. ; Huey R. ; Lindstrom W. ; Sanner M. F. ; Belew R. K. ; Goodsell D. S. ; Olson A. J. AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. J. Comput. Chem. 2009, 30 (16 ), 2785–2791. 10.1002/jcc.21256.19399780
