A subset of flavaglines inhibit RAS-GTP loading and MEK1/2 activation. (A) Different flavaglines (100 nM) were used for the RAS activation assay with the NanoBiT system for KRAS and CRAF-RBD. For KRAS wild-type (WT), cells were stimulated with EGF for 30 min after a 4 h incubation with the indicated flavagline. (B) For KRAS G12V mutation, the cells were treated with flavaglines for 4 h before adding the substrate. The value of DMSO-treated cells was set to 1 and each plot shows the relative value from four independent experiments. The bars represent mean±s.e.m. (C) HeLa cells were treated with flavaglines for 4 h and stimulated with EGF for 30 min followed by SDS-PAGE to check the phosphorylation of MEK1/2. The result from one representative experiment is shown. (D) The panel depicts the quantification of western blotting from three independent experiments as in C. The value of DMSO-treated cells was set to 1 and each plot shows the relative value from three independent experiments. The bars represent mean±s.e.m. (E) SAR analysis was done based on the efficiency of inhibition of KRAS GTP-loading. (F–H) NanoBiT assay for KRAS4B G12V, G13D and Q61K (F), NRAS G12V (G) and HRAS G12V (H) was performed as shown in Fig. 1B. Several flavaglines were used for the assay at the concentration of 100 nM and the cells were treated with respective compounds for 4 h in serum-free medium. DMSO-treated cells served as a control. In each plot, values from three independent experiments are shown. The bars represent mean±s.e.m.

A subset of flavaglines inhibit cell growth in vitro and in vivo. (A) Cells with the indicated KRAS mutations were treated with rocaglamide or FL42 for 24 h and cell viability was evaluated with an MTT assay. The IC₅₀ was calculated from a one-phase decay equation in Prism5. The bars represent mean±s.e.m., from three independent experiments. (B,C) HCT-116 and ASPC-1 cells were used for the soft agar colony formation assay. After the assay, the colonies were stained with Crystal Violet and counted using ImageJ software. The bars represent mean±s.e.m. **P<0.01 compared with DMSO (Welch's t-test in Excel). (D) A soft agar colony formation assay was performed in 96-well plates as indicated in the Materials and Methods section. After 1 week of cell culture in soft agar with compounds (100 nM), cells were treated with MTT and solubilized after 4 h of incubation. The value was obtained by measuring of absorbance (570 nm). Relative values are shown in the figure, and the value from DMSO treated cells served as control (set at 1). The bars represent mean±s.e.m. (E) DOX diet was given to the animals for 2 months. Data from the animals with normal diet is described as ctl. Then animals were treated with either rocaglamide (Roc, 2.5 mg/kg body weight) or DMSO by i.p. injections every 2 days for 6 weeks. Representative lung images are shown in the left panel. Nodules (arrowheads) were counted for all lungs. Each dot plot indicates the data from one individual animal and the bars represent mean±s.e.m.

Rocaglamide treatment inhibits KRAS nano-clustering. (A–C) Effect of rocaglamide on KRAS G12V, HRAS G12V and NRAS G12V nanoclustering as determined by FRET. HEK-293 EBNA cells were transfected with pmGFP-tagged KRAS G12V or HRAS G12V or NRAS G12V (1 μg) only to measure the lifetime of the donor fluorophore. For cells expressing the FRET pairs, cells were co-transfected with pmGFP-KRAS G12V (0.5 μg) and pmCherry-KRAS G12V (1.5 μg), or pmGFP-HRAS G12V (0.5 μg) and pmCherry-HRAS G12V (1.5 μg), or pmGFP-NRAS G12V (0.5 μg) and pmCherry-NRAS G12V (1.5 μg). After 24 h, cells were treated with 0.1% DMSO control or 25 nM or 50 nM of rocaglamide. After 24 h of treatment, cells were fixed with 4% PFA. Numbers on the bars indicate the number of analysed cells. ****P<0.0001; ns, not significant between control and treated samples (one-way ANOVA complemented with Tukey's comparison). The bars represent mean values±s.e.m. from three independent biological experiments. (D) A 1 µm² area of an intact apical PM sheet of a BHK cell expressing GFP-KRAS.G12V is shown with no treatment (CON) or with 200 nM rocaglamide. In each image, gold particles are colour-coded to indicate the extent of clustering: blue, monomer; yellow, dimer; orange, trimer; red, higher ordered multimers. (E) Univariate K-function analysis calculates the extent of nanoclustering of gold particles in the EM images. The extent of nanoclustering, L(r) – r, was plotted against the length scale r in nanometers. L(r) – r values above the 99% c.i. value of 0 indicate statistically significant clustering. The peak L(r) – r value, termed as L_max, describes the optimal clustering. The number of gold particles within the 1 µm² PM area was counted to estimate the extent of PM localization of GFP-KRAS.G12V. (F) For gold-labelled GFP-HRAS.G12V, the extent of nanoclustering, L(r) – r, was plotted against the length scale r in nanometers. The optimal clustering L_max, or gold numbers, is shown as a function of different Roc concentrations. (G) For gold-labelled GFP-NRAS.G12V, the extent of nanoclustering, L(r) – r, was plotted against the length scale r in nanometers. The optimal clustering L_max, or gold numbers, is shown as a function of different Roc concentrations. (D–G) For each condition, at least 15 EM images were imaged, pooled and calculated. All data are shown as mean±s.e.m. For the clustering analysis, statistical analysis was conducted via comparing our point pattern with 1000 bootstrap samples in bootstrap tests. For gold number counting, one-component ANOVA was used to evaluate the statistical significance.

Rocaglamide treatment inhibits KRAS curvature response. (A) Scanning EM images of a nanobar array showing individual nanobars with 250 nm width, 2 μm length, 300 nm height and 5 µm pitch. Scale bars: 2 μm (left), 1 μm (right). (B) When U-2OS cells were cultured on gelatine-coated nanobar arrays, CellMask Deep Red staining showed that plasma membrane wrapped evenly around the nanobar. Scale bars: 5 μm. (C) U-2OS cells transfected with KRAS-WT-GFP were cultured on nanobar arrays. The confocal image showed that KRAS-WT mostly distributed evenly along the nanobar (yellow arrows), similar to CellMask, seen as B, while it may accumulate weakly at the curved nanobar end (red arrows). After EGF treatment, KRAS-WT accumulated strongly at the ends of the nanobars (red arrow). U-2OS cells transfected with KRAS G12V-GFP, showed that KRAS G12V also had strong preference to accumulate at the ends of the nanobars (red arrows), similar to WT with EGF treatment. Averaged images of WT, G12V and WT with EGF treatment on 744–1052 nanobars are shown. Scale bars: 5 μm. (D) U-2OS cells transfected with KRAS WT or KRAS G12V were treated with rocaglamide (Roc) for 4 h before imaging. For the EGF-induced KRAS group, cells were pre-treated with rocaglamide (Roc) for 4 h before EGF stimulation overnight. Rocaglamide altered the preference for the curved ends of the nanobars for  KG12V and EGF-induced KRAS. Scale bars: 2 μm. (E) Averaged images of KRAS+DMSO, KRASG12V+DMSO, KRAS+DMSO+EGF and their rocaglamide (Roc) treatments on 86–417 nanobars. (F) End to centre intensity ratios of each protein/treatment were quantified by averaging over 200–784 data points. Error bar represents s.e.m. Statistical significance of WT vs G12V, WT vs WT+EGF, G12V vs WT+EGF, WT+Roc vs WT+DMSO, G12V+Roc vs G12V+DMSO, and WT+DMSO+EGF vs WT+Roc+EGF. ****P<0.0001, ns, not significant (unpaired Kolmogorov–Smirnov). (G) Frequency distribution of end to centre ratio of each protein with and without rocaglamide (Roc) treatment indicating that WT KRAS maintains its distribution with peak at 1.0 bin centre irrespective of rocaglamide treatment. However, EGF-induced KRAS shifted back to distribution with a 1.0 peak after rocaglamide (Roc) treatment compared to a distribution of EGF-induced KRAS WT with a peak at 1.5. KRAS G12V also showed left-shift to 1.0 with rocaglamide treatment, while KRAS G12V+DMSO showed its highest frequency at 1.5 bin centre. The significant difference of EGF-induced KRAS with and without rocaglamide treatment or KRAS G12V with and without rocaglamide showed that rocaglamide had a strongly inhibitive effect on EGF-induced KRAS WT and KRAS G12V clustering preference at curved nanobar ends.

The role of PHB1 in KRAS activation. (A) Cells were transfected with KRAS G12D plasmid or empty vector (EV) and after 2 days, the cells were treated with DSP for 30 min. After the reaction, cells were harvested and the lysate was obtained for the immunoprecipitation (IP). After IP, the sample was subjected to SDS-PAGE followed by western blotting analysis. (B) The cartoon shows the BiFC system to monitor the interaction between KRAS and PHB1. (C) HeLa cells were transfected with the PHB1 and KRAS BiFC pair. After 24 h, cells were harvested and FACS analysis was performed to check the level of YFP fluorescence signal. A dot blot shows the mean fluorescent intensity (MFI) from four independent experiments from each BiFC pair. The bars represent mean±s.e.m. relative to the value from EV control, set as 1. (D) A cellular thermal shift assay was performed using HEK-293T cell lysate. The lysate was prepared in 10% glycerol in PBS with proteinase inhibitor and incubated with 200 nM rocaglamide (Roc), FL1 or FL42 for 30 min on ice. After incubation, the cells were heated at 45 or 50°C for 6 min and then cooled on ice. After centrifugation, the supernatant was subjected to SDS-PAGE and western blotting. A representative example from two independent experiments is shown. (E) HeLa cells were transfected with siRNAs against PHB1 (siPHB1) or control siRNA (siCon) for 48 h followed by the stimulation with EGF (100 ng/ml) after 4 h of serum starvation. After stimulation, the cells were used for an active RAS pulldown assay. The sample was subjected to SDS-PAGE and western blotting. (F) Dot plot showing the relative values of GTP-loaded RAS against total RAS from three independent experiments as in E. The bars represent mean±s.e.m. (G) NanoBiT assay for KRAS G12V and CRAF-RBD with siRNA against PHB1. The NanoBiT plasmid and siRNA was co-transfected to HeLa cells and NanoBiT assay was performed 2 days later. The left panel shows a representative image of western blotting. The value from positive control pair from Promega (LgBit-PRKAR2A and SmBit-PRKACA) was used for normalization of the NanoBiT assay as siPHB1 transfection slightly inhibited the cell growth. Relative values were obtained from four independent experiments, with the value from siCon set to 1. The bars represent mean±s.e.m. *P<0.05 (Welch's t-test in Excel).

Rocaglamide treatment alters the phospholipid affinity of PHB1 and KRAS. (A) Purified PHB1 was used for the lipid strip binding assay. The lipid membrane was incubated with PHB1 protein overnight either with rocaglamide (10 µM) or DMSO. The signal intensity of the dot plot was obtained by Image J software and the data from three different experiments are shown as dot plots. The bars represent mean±s.e.m. *P<0.05, **P<0.01 (Welch's t-test in Excel). (B) Effect of rocaglamide on KRAS G12V and LactC2 (left), and KRAS G12V and Pass (right) co-clustering. For KRAS G12V/LactC2 FRET, HEK 293 EBNA cells were transfected with mGFP–LactC2 to measure the lifetime of the donor fluorophore, and for the FRET pair, cells were transfected with mGFP-LactC2 and mCherry-KRAS G12V in ratio of 1:3. For the KRAS G12V and Pass FRET pair, cells were transfected with mGFP-KRAS G12V to measure the lifetime of GFP only and for the FRET pair, cells were transfected with mGFP-KRAS G12V and mCherry-PASS in ratio of 1:3. Cells were treated with 0.1% DMSO control, 25 nM rocaglamide or FL1 for 24 h. Cells were fixed in 4% PFA. The apparent FRET efficiency was calculated from FLIM data (mean±s.e.m., n=3). The numbers in the bars indicate the number of analysed cells. ****P<0.0001; ns, not significant, between control and treated cells (one-way ANOVA tests).