Phenotypic characterization of the novel, non-hotspot oncogenic KRAS mutants E31D and E63K

Abstract

KRAS proto-oncogene, GTPase (KRAS) functions as a molecular switch at the apex of multiple signaling pathways controlling cell proliferation, differentiation, migration, and survival. Canonical KRAS mutants, such as those in codons 12 and 13, produce constitutively active oncoproteins that short-circuit epidermal growth factor receptor (EGFR)-initiated signaling, resulting in dysregulated downstream effectors associated with cellular transformation. Therefore, anti-EGFR therapy provides little to no clinical benefit to patients with activating KRAS mutations. Current genotyping procedures based on canonical mutation detection only account for ~40% of non-responders, highlighting the need to identify additional predictive biomarkers. In the present study, two novel non-hotspot KRAS mutations were functionally characterized in vitro: KRAS E31D was identified from a genetic screen of colorectal cancer specimens at the UP-National Institutes of Health. KRAS E63K is curated in the Catalogue of Somatic Mutations in Cancer database. Similar to the canonical mutants KRAS G12D and KRAS G13D, NIH3T3 cells overexpressing KRAS E31D and KRAS E63K showed altered morphology and were characteristically smaller, rounder, and highly refractile compared with their non-transformed counterparts. Filamentous actin staining also indicated cytoplasmic shrinkage, membrane ruffling, and formation of pseudopod protrusions. Further, they displayed higher proliferative rates and higher migratory rates in scratch wound assays compared with negative controls. These empirical findings suggest the activating impact of the novel KRAS mutations, which may contribute to resistance to anti-EGFR therapy. Complementary studies to elucidate the molecular mechanisms underlying the transforming effect of the rare mutants are required. In parallel, their oncogenic capacity in vivo should also be investigated.

Keywords: GTPase; KRAS proto-oncogene; carcinogenesis; colorectal cancer; epidermal growth factor receptor pathway; oncogene.

Figure 1.

KRAS E31D- and KRAS E63K-transfected cells show altered gross cellular morphology similar to canonical mutant controls (representative of three trials). (A) NIH3T3 cells were co-transfected with a pTargeT™ construct and empty pmiR-ZsGreen1, the latter to visualize transfected cells. The black arrows point to KRAS mutant transfectants with altered cell morphology. Scale bars: 50 µm. (B) NIH3T3 cells overexpressing canonical and novel KRAS mutants that exhibit apparent morphological alterations were counted and divided by the total number of fibroblasts in each view. The fraction of cells with morphological irregularities was significantly greater in populations overexpressing a mutant variant of the protein. *P<0.01. WT, wild-type. KRAS, KRAS Proto-Oncogene, GTPase.

Figure 2.

Cytoskeletal features associated with motility are prominent in cells transfected with KRAS E31D and KRAS E63K. Red arrows indicate polymerized actin of pseudopod extensions in cells transfected with the canonical (pTgTKRASG12D and pTgTKRASG13D) and novel KRAS (pTgTKRASE31D and pTgTKRASE63K) mutants. White arrows indicate filopodial actin. In addition to an overall diffused cytoskeletal actin, emergence of pseudopods and filopodial structures were observed in cells transfected with mutant KRAS constructs. KRAS, KRAS Proto-Oncogene, GTPase.

Figure 3.

NIH3T3 cells transfected with canonical and novel KRAS mutants increased cellular proliferation (representative of three trials). Experiment showing the effect of KRAS mutant overexpression on cell proliferation. (A) Cells fed with 10% serum showed comparable proliferation capacity. (B) Cells transfected with mutant KRAS, canonical and novel, showed significant increase in proliferative capacity at 72-h post-transfection when grown in serum-depleted conditions. *P<0.01. BCS, bovine calf serum; WT, wild-type; KRAS, KRAS Proto-Oncogene, GTPase.

Figure 4.

Overexpression of canonical and novel KRAS mutants resulted in increased cell migration in vitro. Scratch wound assays were conducted using transfected NIH3T3 cells grown in 2.5% serum. (A) Micrographs of wound fields directly after scratching the monolayer (0 h), and 16 h post-scratch. Compared with the controls (i.e., vector, pTgTKRASWT), narrower scratch gaps were noticeable for cells overexpressing KRAS mutant variants. Scale bars: 100 µm. Representative of three trials. (B) The distance between lines approximating the cell migration front was measured for each setup at time points 0 and 16 h, and the rate of front migration was calculated. Wound closure was significantly faster for cells overexpressing the mutant protein. *P<0.01. (C) NIH3T3 cells positively transfected with mutant KRAS constructs have an increased tendency to migrate into wound gaps. Cells used for scratch wound assays were co-transfected with a transfection marker, pmiR-ZsGreen1. Migrating fibroblasts were tracked directly after scratching the monolayer (0 h), and 40 h post-scratch to verify that cells facilitating the gap closure are positively transfected. Scale bars: 100 µm. WT, wild-type; KRAS, KRAS Proto-Oncogene, GTPase.

Figure 5.

ELK-TAD luciferase reporter assay showed increased activation of ELK-1 by KRAS mutants. **P <0.001 and ***P <0.0001. ELK-1, ELK1, ETS transcription factor; WT, wild-type; KRAS, KRAS Proto-Oncogene, GTPase.

Figure 6.

Superposition of modeled KRAS mutants with the solved crystal structure of human KRAS showed structural perturbations introduced by E31D and E63K mutations in a region critical to GTPase activity (PDB ID: 3GFT; cyan ribbon). Canonical and novel mutants are represented in red In each overlay image, the sphere represents the GTP-binding site of KRAS. Critical residues of the protein are highlighted in the wild-type schematics as follows: Yellow, phosphate binding-loop; green, switch I; black, switch II. Apparent deviations in the protein conformation brought about by the single amino acid mutations are denoted by arrows. Global RMSD values were also predicted via sequence alignment function for each mutant in comparison with the wild-type. RMSD, root-mean-square distance; KRAS, KRAS Proto-Oncogene, GTPase.

Figure 7.

Simulated interactions of GTP to wild-type and mutant KRAS yielded variable binding energies. The solved x-ray crystal structure of human KRAS. (A) PDB: 3GFT was used to build mutant homologues (B) G12D, (C) G13D, (D) E31D and (E) E63K. GTP binding was implemented using the CDOCKER algorithm. The ligand conformation with the lowest (i.e., most negative) CDOCKER energy is shown for each interaction diagram. For the mutant proteins, the amino acid change is highlighted in red and labelled accordingly. The magnesium ion cofactor is shown as a grey sphere above the bound GTP. WT, wild-type; KRAS, KRAS Proto-Oncogene, GTPase.