Structural basis for LZTR1 recognition of RAS GTPases for degradation

Abstract

The RAS family of small guanosine triphosphatases (GTPases) are tightly regulated signaling molecules that are further modulated by ubiquitination and proteolysis. Leucine Zipper-like Transcription Regulator 1 (LZTR1), a substrate adapter of the Cullin-3 RING E3 ubiquitin ligase, binds specific RAS GTPases and promotes their ubiquitination and proteasomal degradation. We present structures of LZTR1 Kelch domains bound to RIT1, MRAS, and KRAS, revealing interfaces that govern RAS isoform selectivity and nucleotide specificity. Biochemical and structural analyses of disease-associated Kelch domain mutations revealed three types of alterations: impaired substrate interaction, loop destabilization, and blade-blade repulsion. In cellular and mouse models, mutations disrupting substrate binding phenocopied LZTR1 loss, underscoring its substrate specificity. These findings define RAS recognition mechanisms by LZTR1 and suggest a molecular glue strategy to degrade oncogenic KRAS.

Fig. 1.. Structural basis for LZTR1 interactions with GDP-bound RAS proteins.

A. Domain architecture of LZTR1 and the RAS GTPases RIT1, MRAS, and KRAS analyzed in this study. Biophysical and structural studies focused on the N-terminal Kelch domain of LZTR1. RIT1 and MRAS contain 11 and 10 additional N-terminal residues, respectively, compared to KRAS, requiring residue numbering adjustments to identify equivalent switch residues for comparison. B. Bar graph comparing the binding affinities (K_D) of LZTR1 with RAS family members RIT1, MRAS, KRAS, HRAS, and NRAS in their GDP- and GMPPNP-bound forms. Error bar: errors of the fit. C. Overall structure of RIT1(GDP) in complex with LZTR1 shown as a cartoon representation. D. Overall structure of MRAS(GDP) in complex with LZTR1 depicted as a cartoon representation. The individual Kelch motifs of LZTR1 and the switch regions of RIT1 and MRAS are colored using the same color coding as in panel 1A. E. Top view of LZTR1 with Kelch motifs and blades numbered from the RIT1-LZTR1 structure. F. Bar graph comparing the binding affinities of LZTR1 with wild-type KRAS and its T35A, E62A, and double-mutant (T35A/E62A) variants, highlighting the effects of these mutations on binding. Error bar: error of the fit. G. Overall structure of KRAS-T35A/E62A(GDP) in complex with LZTR1 shown as a cartoon representation. H. Structural superposition of MRAS-GMPPNP (PDB: 9B4R, cartoon representation) onto MRAS-GDP (cartoon representation) bound to LZTR1 (surface representation) highlights distinct conformations of the switch regions, elucidating the structural basis for LZTR1's specificity for GDP-bound MRAS.

Fig. 2.. Molecular basis of RIT1-LZTR1 interactions.

A. Schematic overview of the RIT1-LZTR1 interaction interface. The lines represent different types of Interactions: hydrogen bonds (solid blue), salt bridges (solid red), and non-bonded contacts (striped orange), with the width of the striped lines proportional to the number of atomic contacts. B. Enlarged view of the RIT1-LZTR1 complex showing the interaction interface, with Kelch motifs (surface) and RIT1 switch regions (cartoon) color-coded as in Figure 1A. C. Close-up of interactions between the switch-I region of RIT1 (cartoon) and the Kelch IV and V motifs of LZTR1 (surface). D. Detailed view of the α2-helix in the switch-II region of RIT1 (cartoon) interacting with the Kelch IV motif of LZTR1 (surface). E. Interaction interface highlighting the switch-II region of RIT1 (cartoon) and the Kelch I, II, and III motifs of LZTR1 (surface). Hydrogen bonds and salt bridge interactions are depicted as dashed lines. F. Bar graph showing the binding affinity (K_D) measured using ITC for point mutants of RIT1(GDP) interface residues with LZTR1. Error bar: error of the fit. Residues mutated to alanine are in bold in panels C, D and E. G. Immunoblot of proteins in lysates and glutathione pulldowns from 293T LZTR1-ALFA cells ectopically expressing GST-fused RIT1 mutants. WCL: whole cell lysates. H. Immunoblot of indicated proteins from a cell-based degradation assay co-expressing GST-tagged RIT1 mutants and HA-LZTR1 in 293T cells. I. Bar graph showing the binding affinity (K_D) measured using ITC for point mutants of LZTR1 interface residues with RIT1 (GDP). Error bar: error of the fit. J. Representative immunoblot and Coomassie staining of in vitro pulldown assays between recombinant GST-RIT1 and lysates from 293T cells ectopically expressing the indicated HA-LZTR1 mutations. WCL: whole cell lysates. K. Immunoblot of indicated proteins from a cell-based degradation assay co-expressing GST-tagged RIT1 and indicated HA-LZTR1 mutations in 293T cells. L. Sequence alignment of RIT1, MRAS, and KRAS, with LZTR1-interacting residues highlighted in brown. Fully conserved residues among the three GTPases are shown in bold, and residues within the switch regions are underlined below the alignment. M. Structural superposition of RIT1, MRAS, and KRAS complexes with LZTR1, highlighting similarities and subtle differences in the conformations of switch-I. RAS GTPases are depicted as cartoons, while LZTR1 is shown as a surface representation. N. Structural superposition of RIT1, MRAS, and KRAS complexes with LZTR1, highlighting similarities and subtle differences in the conformations of switch-II. RAS GTPases are depicted as cartoons, while LZTR1 is shown as a surface representation. O. Quantification of KRAS protein levels in HEK293T-Flp-In cells stably expressing KRAS WT or E62A in the presence (sgAAVS1) or absence (sgLZTR1) of LZTR1 (n=3); Student’s t-test.

Fig. 3.. Mechanistic insights into LZTR1 disease-associated variants.

A. Enlarged view of the RIT1-LZTR1 interaction interface showing the switch-II residues mutated in NS. B. Bar graph showing the impact of disease-associated mutations in RIT1 and LZTR1 on their binding affinities. Error bar: error of the fit. C. Representative immunoblot and Coomassie staining of in vitro pulldown assays used to assess the effects of disease-associated variants on RIT1 and MRAS binding. GST-RIT1 and lysates from 293T cells ectopically expressing the indicated HA-LZTR1 variants (two that lead to loss of function and two that show no effect in this assay are shown). WCL: whole cell lysates. D. Immunoblot of indicated proteins from a cell-based degradation assay used to assess the effects of LZTR1 disease-associated variants on RIT1 and MRAS degradation. 293T cells co-expressing GST-tagged RIT1 or MRAS and indicated HA-LZTR1 variants (two that lead to loss of function and two that show no effect in this assay are shown). E. RIT1 interaction assay with the indicated HA-LZTR1 variants. Densiometric quantification and normalization to LZTR1 WT binding to recombinant GST-RIT1 were used to assess the effect of each variant (n=3). AU: arbitrary units. Error bar: SEM. F. In-cell RIT1 degradation assay with the indicated HA-LZTR1 variants. Densiometric quantification and normalization to GST-RIT1 co-expressed with empty vector (EV) were used to analyze the effect of each variant (n=3). AU: arbitrary units. Error bar: SEM. G. Schematic representation of loss-of-function mechanisms for LZTR1 mutations based on the LZTR1-RIT1 structure. H. Enlarged view of the RIT1-LZTR1 complex illustrating the loss-of-function mechanisms of LZTR1 mutations. Type I mutations disrupt direct interactions with RIT1. Type II mutations destabilize Kelch loops essential for RIT1 binding. Type III mutations disrupt hydrophobic Kelch-Kelch interfaces, altering loop conformations required for RIT1 engagement.

Figure 4.. The type I LZTR1 mutation G248R phenocopies LZTR1 loss in cells and mice.

A. Enlarged view of the RIT1-LZTR1 interface showing LZTR1 G248 (left) and its steric clash with RIT1 D56 upon mutation to arginine (right). B. Immunoblot of indicated proteins in lysates from 293T LZTR1-ALFA cells wildtype (+/+), heterozygous (+/GR), homozygous G248R (GR/GR), and knockout (−/−) LZTR1 treated for the indicated times with 20 ng/mL of EGF. C. SPRY2 mRNA levels assessed by qPCR in cells from b, treated for 2 hours with or without EGF (20 ng/mL) (n=3). D. Volcano plot obtained from TMT-based quantitative proteomics in lysates from 293T LZTR1-ALFA isogenic cells described in panel B. Highlighted in red are proteins that were significantly upregulated or downregulated, as defined by a q-value ≤0.05 (False Discovery Rate (FDR) of <5%). The most significantly upregulated protein in GR/GR and −/− cells, RIT1, is labelled. Source data are provided in Data S1. E. Percentages of obtained genotypes upon weaning (21 days of age) the offspring of indicated heterozygous mutant intercrosses, using Lztr1^G248R (n=88) and Lztr1^null (n=92) mice. Estimated Mendelian rates are shown and were used to calculate the indicated Chi-square statistics. F. Representative images of hematoxylin-eosin (H&E) and cleaved caspase-3 (cl-C3) immunohistochemistry staining in livers from E16.5 embryos with the indicated genotypes. G. Immunoblot of indicated proteins in lysates from E13.5 primary MEFs derived from individual embryos with indicated genotypes.