a, Binding titration of NS1 displayed on yeast cells with the RAS isoforms bound to nucleotide using flow cytometry detection. Affinity values (K_D) are as follows: H-RAS GTPγS, 13.5±1.9 nM; H-RAS GDP, 15.7±2.7 nM; K-RAS GTPγS, 67±16 nM; K-RAS GDP, 60±11 nM; N-RAS GTPγS, not detectable; N-RAS GDP, not detectable. Error bars represent s.d. from N=3. The errors for the K_D values are s.d. from three independent measurements. b & c, Assessment of NS1 specificity in cells. b, Co-localization of CFP-NS1 (pseudo-colored red) co-localizes with YFP-tagged RAS isoforms (each pseudo-colored green). Scale bars, 10 um. c, Co-immunoprecipation of HA-tagged RAS isoforms by CFP-NS1. Full blot images for Fig. 1c are shown in

a, Effect of CFP-NS1 expression on EGF-stimulated ERK activation in HEK293 cells. CFP-NS1 and MYC-tagged ERK were co-expressed, and phosphorylation of MYC-tagged ERK was detected by Western blot with phosphospecific ERK antibodies. b, Cells transfected with the indicated oncogene along with CFP or CFP-NS1 were analyzed for ERK activation as in a. c, Effect of NS1 on AKT activation by RAS. Cells were transfected and analyzed as in b except HA-AKT was used in place of MYC-ERK. Quantification of results from b and c is provided in . d, NIH/3T3 transformation assay. Quantification of relative foci number for each oncogene is shown in the graphs. Results represent the ratio of foci number in presence of CFP-NS1 vs CFP alone and are the average of three independent experiments performed in triplicate +/− s.d. p values were determined by a Student’s t-test between CFP and CFP-NS1 for each oncogene. **, p<0.01, *, p<0.05. Effects of NS1 on oncogenic HER2/Neu, BRAF, and MEKK1 are shown in . e & f. Effects of NS1 on signaling and proliferation of human tumor cells. e. Titration of CFP-NS1 effects on ERK activation and proliferation (graph) in T24 bladder carcinoma cells (e) and A375 melanoma cells (f). Results in the graphs are the average of triplicate wells +/− s.e.m. shown by bars. Western blots in e and f are representative of at least 3 independent experiments. Full blot images for Figs 2a–c, e and f and , respectively.

a, Crystal structure of NS1:H-RAS-GDP. b, Chemical shift changes in H-RAS upon NS1 binding. Amino acids undergoing strong chemical shifts are highlighted red and weaker reacting residues highlighted orange in the H-RAS:RAF RBD structure (PDB ID: 4G0N). The most significant changes occurred in the β5, α4, β6, and α5 regions of H-RAS, which lie on the surface of RAS in opposition to the switch regions which bind effectors such as RAF, indicated in the figure. c, Amino acid sequences of the α4-β6-α5 region (residues 122–166) of the RAS isoforms. Asterisks mark amino acids whose mutation affects affinity to NS1. Dots mark amino acid side chains within 5 angstroms of amino acid side chains in NS1. d, Recognition of R135 in H-RAS by NS1. Polar interactions between R135 and surrounding amino acids and water (green sphere) are highlighted by dashed yellow lines. e. Effect of Arg135 to Lys mutation on H-RAS binding NS1 in vitro. Results are the average of three independent binding experiments +/− s.d. WT, 165.3 +/−2.9 AU; R135K, 0.44+/−0.29 AU.

a, Comparison of H-RAS crystal structures in the Protein Data Bank (PDB). Seventy-four of 80 active state H-RAS structures contained α4-α5 dimers but no α4-α5 dimers were present in inactive state structures. A protomer (shown in gray) of the α4-α5 dimers from six active state H-RAS structures (PDB ID, 5P21, 1AGP, 4L9W, 2C5I, 3K9L, and 3KUD) were superimposed using Pymol. SW1 and SW2 regions are highlighted yellow. b, NS1 monobody binds the α4-α5 dimer interface as highlighted in brown. The dimer interface shown is from PDB ID: 5P21. These regions overlap with the presumed dimerization interface identified in N-RAS. c. Comparison of NS1 binding surface (highlighted red) and α4-α5 dimer interface (denoted with black boarder) on H-RAS structure.

a, Model for NS1 inhibition of RAS-mediated signaling. b, The extent of nanoclustering of GFP-H-RAS(G12V) or GFP-K-RAS(G12V) in BHK plasma membrane sheets immunolabeled with α-GFP-4.5nm gold gold particles was evaluated in K-functions summarized as L_max. Data represent mean L_max values ± s.e.m. (Number of samples analyzed: H-RAS+RFP, n=19; H-RAS+RFP-NS1, n=23; K-RAS+RFP, n=16; K-RAS+RFP-NS1, n=24.; * indicates p < 0.05, bootstrap tests). c, The dimer/monomer ratio estimated from data in b. Analysis of the point patterns from b allowed calculation of dimer/monomer ratio shown as mean ± s.e.m. (The number of samples analyzed was the same as in b; * indicates p < 0.05, one-way ANOVA). d, Effect of NS1 on K-RAS(G12V) or N-RAS(G12V) dimerization measured by BRET. Each experiment was performed at least three times with values representing the average of three independent transfections +/− s.d. Additional controls for these experiments are provided in . e, Plasma membrane localization of GFP-K-RAS(G12V) and GFP-H-RAS(G12V) determined by counting total number of gold-labelled GFP particles per 1µm² area of intact plasma membrane sheets. Values are shown as mean ± s.e.m. (n>15; * indicates p < 0.05, one-way ANOVA).

a, Effect of CFP-NS1 vs CFP on binding of endogenous BRAF to either H-RAS(G12V) or K-RAS(G12V). Results are representative of at least 3 independent experiments. b, Effects of CFP-NS1 vs CFP on oncogenic H-RAS(Q61L)-induced heterodimerization of CRAF and BRAF measured by co-immunoprecipitation in HEK-293T cells. c, K-RAS(G12V) induced heterodimerization of BRAF and CRAF measured by BRET. Each experiment was performed at least three times with values representing the average of three independent transfections +/− s.d. Additional controls are provided in . d, Effect of NS1 on in vitro RAF kinase activity. Quantification of results from three independent experiments is shown in . Full blot images for Figs 6a, b, and d are shown in .