Divergent Morphologies and Common Signaling Features of Active and Inactive Oncogenic RHOA Mutants in Yeast

Authors

Chenwei Wang

Shinsuke Ohnuki

Anna Savchenko

Hiroyuki Aburatani

Satoshi Yoshida

Riko Hatakeyama

Yoshikazu Ohya

Doi

10.3390/cells14181439

PMID: 24816253 · DOI: 10.3390/cells14181439 · Journal: Cells (2025)

TL;DR

This study uses Saccharomyces cerevisiae as a humanized model to functionally dissect five oncogenic [genes/RHOA|RHOA] hotspot mutants (R5Q, G17V, C16R, A161P, E40Q) by expressing them in place of the essential yeast homolog RHO1. Four variants (R5Q, G17V, C16R, A161P) complemented RHO1 loss while E40Q did not. All four viable mutants conferred myriocin resistance (suggesting TORC2 pathway activation) but showed comparable caspofungin sensitivity. High-content imaging with CalMorph (501 morphological parameters) plus PCA/LDA cleanly separated gain-of-function mutants (C16R, A161P) from loss-of-function mutants (R5Q, G17V), demonstrating that biochemically opposite RHOA mutations engage distinct cellular programs even in a simplified eukaryotic context.

Cohort & data

  • Model organism: S. cerevisiae (BY4743 diploid background); not a clinical cohort.
  • Strains: heterozygous diploids expressing wild-type human [[genes/RHOA|RHOA]] or one of four point-mutant alleles (R5Q, G17V, C16R, A161P) substituting for yeast RHO1, plus E40Q which failed complementation.
  • Controls: RHO1/RHO1, RHO1/RHO1 ade3, RHOA/RHOA, RHOA/rhoA∆.
  • Imaging readout: CalMorph v1.3 ([[methods/calmorph|CalMorph]]) — 501 morphological parameters across cell wall, actin, and nucleus; ≥200 cells per replicate × 5 biological replicates per strain.
  • Cancer-context relevance (per text): R5Q in diffuse-type gastric carcinoma ([cancer_types/DSTAD|DSTAD]) and Burkitt lymphoma ([cancer_types/BL|BL]); G17V in angioimmunoblastic T-cell lymphoma ([cancer_types/AITL|AITL], ~67% of cases) and other [cancer_types/PTCL|PTCL]; C16R and A161P predominantly in adult T-cell leukemia/lymphoma ([cancer_types/ATLL|ATLL]); E40Q in solid tumors including breast and head-and-neck squamous cell carcinoma.

Key findings

  • E40Q failed to complement RHO1 deletion on 5-FOA, while R5Q, G17V, C16R, and A161P all supported yeast viability (PMID:24816253, Figure 1B).
  • All four viable [[genes/RHOA|RHOA]] mutants conferred increased resistance to [[drugs/myriocin|myriocin]] (250–1000 ng/mL) relative to both RHOA/RHOA and RHOA/rhoA∆ controls, suggesting activation of TORC2 signaling (Figure 2B).
  • No clear differences in [[drugs/caspofungin|caspofungin]] sensitivity (50–200 ng/mL) across the four mutants (Figure 2C); cell wall integrity pathway not broadly activated.
  • Slt2 MAPK phosphorylation: only A161P reproducibly elevated Slt2 phosphorylation relative to RHOA/RHOA control across n=3 replicates; C16R showed occasional non-reproducible increases; G17V and R5Q indistinguishable from controls (Figure 2D).
  • Generalized linear model + one-way ANOVA on 501 CalMorph features identified 164 parameters significantly altered (FDR<0.05, Wald test with Storey’s correction) across the four mutants; A161P showed the greatest number of altered parameters. Breakdown: nuclear D* (70/164; 42.7%), actin A* (51/164; 31.1%), cell-shape C* (43/164; 26.2%).
  • Pairwise Pearson correlations among the four heterozygous mutants were comparatively high (0.649–0.842), while correlations between RHOA/rhoA∆ and each mutant were notably lower (0.221–0.271).
  • Linear discriminant analysis on top 17 PCs (>90% variance) cleanly separated three classes: RHOA/rhoA∆, gain-of-function (C16R, A161P), and loss-of-function (R5Q, G17V) along LD2 (Figure 5A).
  • Top LD1 loadings (separating mutants from rhoA∆): proportion of actin at bud neck (A9_A1B), cell length (C103_A1B). Top LD2 loading (separating GOF from LOF): variability in actin patch brightness (ACV122_C).
  • Hierarchical clustering grouped A161P, C16R, and G17V together despite G17V being dominant-negative, while R5Q (attenuated-output LOF) clustered closer to RHO1/RHO1 — indicating cluster membership tracks perturbation magnitude/pattern rather than the GOF/LOF dichotomy alone.

Genes & alterations

  • [[genes/RHOA|RHOA]]
    • R5Q — attenuated-output loss-of-function; reduces RhoA activation; recurrent in diffuse-type gastric carcinoma ([[cancer_types/DSTAD|DSTAD]]) and Burkitt lymphoma ([[cancer_types/BL|BL]]); diffuse large B-cell lymphoma implicated. Complements RHO1 in yeast; clusters with RHO1/RHO1 morphologically.
    • G17V — dominant-negative loss-of-function; abolishes GTP binding; present in ~67% of [[cancer_types/AITL|AITL]] and other [[cancer_types/PTCL|PTCL]]. Complements RHO1; clusters with C16R/A161P morphologically despite opposite biochemistry.
    • C16R — gain-of-function; accelerates GTP/GDP cycling; predominantly in [[cancer_types/ATLL|ATLL]]. Complements RHO1; clusters with A161P and G17V.
    • A161P — gain-of-function; predominantly in [[cancer_types/ATLL|ATLL]]; uniquely showed reproducible Slt2 phosphorylation elevation in yeast and the greatest number of altered CalMorph features.
    • E40Q — fails to complement RHO1 in yeast (no viable colonies on 5-FOA); recurs in solid tumors including breast cancer and head-and-neck squamous cell carcinoma ([cancer_types/HNSC|HNSC]); excluded from downstream morphological profiling.

Clinical implications

  • The paper makes no direct clinical claims and tests no therapeutic agent against patient samples. Translational implications stated by the authors:
    • Morphological features identified (actin bud-neck localization, cell length, actin patch variability) could serve as readouts for screening chemical or genetic modifiers that selectively affect gain- versus loss-of-function [[genes/RHOA|RHOA]] signaling, potentially informing class-specific therapeutic strategies for [[genes/RHOA|RHOA]]-driven cancers.
    • Shared myriocin resistance across mutants points to TORC2 as a common downstream node potentially exploitable across mutation classes.

Limitations & open questions

  • Yeast lacks higher-eukaryote RhoA effectors (e.g., mammalian formins beyond conserved homologs, ROCK kinases, RhoGEFs/GAPs); translational extrapolation is bounded.
  • E40Q phenotype is indistinguishable from a null under these conditions (no viable colonies on 5-FOA), so its specific oncogenic mechanism cannot be addressed by this system.
  • The hierarchy between TORC2 and Pkc1 downstream of RHOA variants is unresolved; authors note genetic epistasis with TORC2-specific reporters could clarify.
  • Slt2 phosphorylation was variable across biological replicates (n=3) and only A161P showed a reproducible signal — the molecular basis of allele-specific CWI activation in the presence of endogenous Rho1 remains unclear.
  • No quantitative cross-allele morphological comparison has been firmly established in mammalian systems under matched conditions (authors call this out as a gap).
  • Single-laboratory study; results were not replicated in mammalian cell lines or patient-derived models.

Citations from this paper used in the wiki

  • “All four variants conferred myriocin resistance, suggesting activation of the membrane stress response pathway” (Abstract).
  • “Gain-of-function (C16R and A161P) and loss-of-function (R5Q and G17V) mutants formed separate morphological clusters, indicating functional divergence” (Abstract).
  • “G17V, present in ~67% of AITL cases and other peripheral T-cell lymphomas, abolishes GTP binding ability” (Introduction).
  • “C16R and A161P are predominantly seen in ATL… whereas E40Q occurs in solid tumors such as breast cancer and head-and-neck squamous cell carcinoma” (Introduction).
  • “We identified 164 parameters that exhibited significant changes across the four mutants (FDR<0.05, Wald test with Storey’s correction)… A161P showed the greatest number of altered parameters” (Section 3.4).
  • “The first discriminant axis (LD1) differentiated RHOA heterozygotes from RHOA/rhoA∆, while the second axis (LD2) distinguished gain-of-function from loss-of-function mutants” (Section 3.4).

This page was processed by crosslinker on 2026-05-11.