ABSTRACT
Ras genes are important oncogenes that are frequently mutated in cancer. Human oncogenic variants exhibit functional distinctions in terms of their representation in different cancer types, impact on cellular targets and sensitivity to pharmacological treatments. However, how these distinct variants influence and respond to the cellular networks in which they are embedded is poorly understood. To identify novel participants in the complex interplay between Ras genotype and cell interaction networks in vivo, we have developed and tested an experimental framework using a simple vulva-development assay in the nematode C. elegans. Using this system, we evaluated a set of Ras oncogenic substitution changes at G12, G13 and Q61. We found that these variants fall into distinct groups based on phenotypic differences, sensitivity to gene dosage and inhibition of the downstream kinase MEK and their response to genetic modulators that influence Ras activity in a non-autonomous manner. Together, our results demonstrated that oncogenic C. elegans Ras variants exhibit clear distinctions in how they interface with the vulva-development network and showed that extracellular modulators yield variant-restricted effects in vivo.
INTRODUCTION
Ras proteins are a class of highly conserved small GTPases that play a critical role in cell cycle regulation and development in animals (reviewed by Pylayeva-Gupta et al., 2011). In humans, the three genes encoding Ras proteins, i.e. KRAS, NRAS, HRAS, are important oncogenes and crucial drivers of many human cancers, especially pancreatic, colorectal and lung (Prior et al., 2012; Waters and Der, 2018). While the clinical importance of oncogenic Ras has been understood for over 40 years, therapeutics that specifically and effectively target Ras have not yet been established as standard-of-care. The overwhelming majority of sporadic oncogenic mutations affecting human Ras genes result in amino acid substitutions at just three positions (G12, G13 and Q61), but there is increasing awareness that different oncogenic Ras variants should be evaluated for their distinct properties, rather than considered as a single class of activated Ras mutations. This re-framing has yielded a number of important new research directions, including sub-classifying cancers on the basis of sensitivity to established, network-targeting therapeutics (McFall and Stites, 2021) and design of variant- or protein-specific inhibitors (Kim et al., 2023; Ostrem et al., 2013; Wang et al., 2022). While these developments are encouraging, clinical trials for FDA-approved G12C inhibitors report that only a subset of patients exhibit a response, many experience treatment-related adverse events and tumors rapidly evolve resistance (reviewed by Parikh et al., 2022; Rosen et al., 2023). Consequently, it is clear that the cellular networks in which oncogenic Ras proteins are embedded can influence outcomes. Understanding how different variants differentially engage effector and other proteins within a cancer cell, and how they may have distinct roles in the communication between cancer and non-cancer cells, are still open questions. Addressing these questions will be important for targeting Ras-driven cancers in a robust, multi-target manner that can increase drug efficacy and coincidentally reduce toxicity and the capacity of cells to evolve resistance.
Experimental animal models will be important in uncovering the complex relationships between Ras variants and other factors that influence Ras activity and disease progression. The nematode C. elegans provides a simplified system to study certain aspects of Ras biology in a multicellular context (reviewed by Cerón, 2023; Sundaram, 2013). The C. elegans genome includes a single Ras ortholog, i.e. let-60 (Fig. S1), and the gene has been shown to play an important role in development, homeostasis, nervous system function and other processes (Church et al., 1995; Hirotsu et al., 2000; Schutzman et al., 2001; Yochem et al., 1997; reviewed by Sundaram, 2013). let-60 function has been well-characterized in the development of the hermaphrodite vulva, a specialized epithelial structure that connects the gonad to the outside, permitting egg laying and mating with males (Beitel et al., 1990; Han and Sternberg, 1990). This structure is formed during larval development, where an EGF signal, LIN-3, from the anchor cell in the somatic gonad coordinates development and fate patterning of the six epithelial vulval precursor cells (VPCs) (reviewed by Sternberg, 2005). Normal vulval development occurs when three of six VPCs are induced to divide and produce vulval cell types, utilizing a canonical Ras/Raf/MEK/MPK signaling pathway. Importantly, classic genetic screens have identified gain-of-function alleles of let-60 that cause increased activity of Ras, resulting in division of up to six of the VPCs (rather than the wild type three), producing a multivulva (Muv) phenotype that causes animals to have visible bumps or protrusions on their ventral sides (Fig. 1B). These genetic mutants have provided the background for genetic dissection of the functional relationship between activated Ras and important effector proteins and downstream targets (Han et al., 1993; Lackner et al., 1994; Sieburth et al., 1998; Singh and Han, 1995; Wu and Han, 1994), for the screening of molecular inhibitors (Reiner et al., 2008; van der Hoeven et al., 2020) and discovery of new genes that mediate or modulate Ras activity (Bruinsma et al., 2002; Corchado-Sonera et al., 2022; Goldstein et al., 2006; Rocheleau et al., 2005; Sundaram and Han, 1995). While studies that used existing gain-of-function alleles have been crucial for understanding the signaling-network response to general Ras dysregulation, they were of limited use in a framework aiming to uncover variant-specific differences. Many C. elegans studies utilize gain-of-function alleles that cause a G13E substitution (Beitel et al., 1990), i.e. a variant not observed in human tumor samples (Fig. S1B; GCD database data release 38.0) (Grossman et al., 2016). Thus, it is not clear what the consequences of substitutions that mimic human tumor variants have in C. elegans and whether this powerful model system can be extended to variant-specific, rather than simply gene-specific, experimental discovery and chemical compound screening.
Here, we utilized CRISPR-mediated genome editing and controlled, single copy transposon insertion methods to evaluate the function of a set of oncogenic Ras variants in the C. elegans vulval development system. We show that, in contrast to the available let-60(G13E) mutants, a let-60(G13D) substitution did not grossly disrupt vulval development. Using insertions at a Mos1-mediated single-copy insertion (MosSCI) transposon landing site (Frøkjær-Jensen et al., 2014), we assessed the impact on let-60 gene function of several oncogenic substitution changes at G12, G13 and Q61. Importantly, we found that these substitutions have different effects, including whether they cause a Muv phenotype or whether animals are sensitive to gene dosage. Finally, we showed that the variants exhibit differential response to two non-autonomous modifiers (hpo-18 and szy-5) identified in a genetic screen using let-60(n1046[G13E]) (Corchado-Sonera et al., 2022). Together these results establish a systematic approach to assay oncogenic Ras variants and uncover functional distinctions among them using C. elegans as a model.
RESULTS
Isolation and characterization of let-60(G13D) mutants in C. elegans
With respect to sequence, the human variant most similar to the available C. elegans let-60(n1046[G13E]) is the smaller, but similarly charged G13D, which is a common substitution at G13 in cancer samples (Fig. S1) (McFall and Stites, 2021). One reason for the difference is suggested by the genomic sequence, as the predicted codon for amino acid 13 differs between C. elegans and human – i.e. GGA for C. elegans versus GGU or GGC for each human gene – with the G13D substitution requiring a two nucleotide change in C. elegans. Consequently, we first asked whether oncogenic variants cause a gain-of-function phenotype in C. elegans by directly engineering this substitution into the let-60 locus using CRISPR-mediated genome editing. We recovered three independent alleles, all of which result in grossly wild-type animals that exhibit no apparent vulval defects (Fig. 1). Specifically, unlike let-60(n1046[G13E]), the substitution does not confer the Muv phenotype (Fig. 1A-C,G). Further, in all larval let-60(gu287[G13D]) mutants evaluated, only three vulval precursor cells divided to produce vulval tissue as seen in wild type, and animals exhibit a vulva with typical morphology in the fourth larval stage (Fig. 1D-F,H). By contrast, in let-60(n1046[G13E]) mutants, between three and six cells divide – resulting in a normal vulval structure and in up to three other clusters of cells that form small vulva-like structures in animals at larval stage 4 (L4) – and undergo morphogenesis to produce bumps of tissue in adults. We conclude that, unlike the G13E substitution, substitution of G13D did not sufficiently disrupt LET-60 function to produce an activated Ras phenotype.
A MosSCI-based method to assess oncogenic Ras variant activity in C. elegans
We next aimed to test the activity of other oncogenic variants in the C. elegans model. After preliminary efforts to introduce alleles directly into the let-60 locus suggested many such alleles confer a lethal phenotype (not shown), we sought an alternate approach that would permit functional assessment of multiple variants in the vulval development model, where alterations to Ras activity (either activated and disrupted) confer specific outcomes that can be easily identified visually by using a dissecting microscope or similar optical methods. We utilized the MosSCI method of Mos1 transposon-mediated single-copy insertion to introduce one additional copy of the let-60 gene into a defined landing site in the C. elegans genome (Frøkjær-Jensen et al., 2014) that comprised 4 kb genomic DNA, including the let-60 coding exons as well as 1 kb each upstream and downstream, and including all of the annotated 5′ and 3′ UTR sequences. Since the inserted transposon can be made homozygous, this produced a strain with four let-60 gene copies: two at the endogenous genomic locus and two at the landing site. For simplicity, we will refer to the endogenous locus as let-60, the copies at the landing site as either transposon or Si(let-60), and the alleles or variants based on their impact on the predicted amino acid product. Complete genotypes, transposon names and strain nomenclature are listed in Table S1. Insertion of single copies and expression of each was confirmed by using PCR across the insertion locus and evaluating the sequence of let-60 cDNA derived from mRNA of each strain (Figs S2, S3).
When a wild-type copy of let-60 was introduced on the transposon, the animals are grossly normal and exhibit no Muv phenotype (Fig. 2A). This is in contrast to the case where multiple copies of the gene are introduced as part of an extrachromosomal array (Han and Sternberg, 1990). Introduction of a let-60(G13E) transposon caused a weak Muv phenotype, whereas introduction of wild-type Si(let-60(+)) into the let-60(G13E) background produced an intermediate phenotype. This demonstrates that the transposon-based gene copies produced a product that can influence vulval development. Since the Muv phenotype of these two genotypes was not identical, the activity of the transposon-based gene appeared lower than that at the endogenous locus, a phenotypic observation consistent with the relative abundance of the transgene-based variant in mRNA (Fig. S3). Finally, inclusion of the G13E variant at both loci enhanced the phenotype. Together, these results indicate that animals were sensitive to dosage of the G13E variant and that the effect does not dependent on the presence of the wild-type gene copy. However, for this variant, wild-type protein did compete with mutant to promote a more wild-type response in animals that express both. To ask whether the G13D variant exhibits a similar dosage effect, we produced an Si(let-60(G13D)) transgene and introduced it into the let-60(gu287[G13D]) background (Fig. 2B). These animals remained phenotypically wild type, indicating that increasing the dosage of this variant does not alter the effect observed in the mutant.
C. elegans vulval development exhibits distinct responses to different disease-associated Ras variants
We utilized the MosSCI method to introduce let-60 transposons encoding a set of additional oncogenic substitution mutations prevalent in human cancer samples, including G13R, G12D, G12C and Q61R (Fig. 3). Data for a G13S transgene that was also produced and analyzed but that exhibits higher expression, and may represent insertion of multiple gene copies, are included in Fig. S4. Like Si(let-60(G13D)), Si(let-60(G12C)) caused no gross defects to vulva development (Fig. 3C,F), whereas Si(let-60(G13R)), Si(let-60(G12D)) and Si(let-60(Q61R)) conferred a strong Muv phenotype on all or most animals (Fig. 3D,E,G). We next asked whether C. elegans animals exhibit a dosage sensitivity to the three variants that cause a phenotype (Fig. 4). We found that − whereas a single transposon copy, rather than two as in homozygotes, is sufficient to cause a high frequency of Muv for two of the variants, i.e. Si(let-60(G12D)) and Si(let-60(Q61R) – animals were sensitive to the dosage of Si(let-60(G13R)), and that most animals bearing only a single copy of the transgene exhibit a wild-type phenotype (Fig. 4). These results point to the capacity of each mutant variant to be subject to functional regulation and to compete with wild-type protein. Si(let-60(G13D)) and Si(let-60(G12C)) failed to interfere with wild type and, at least for let-60(G13D) (Fig. 1), essentially exhibits wild-type function. Excess wild-type protein was not sufficient to compete with Si(let-60(G12D)) or Si(let-60(Q61R)), but did compete with Si(let-60(G13R)) (Fig. 4).
Finally, we evaluated the sensitivity of each variant to the MEK inhibitor U0126 to test whether the distinctions among the different oncogenic variants in C. elegans reflects off-target effects, or the preferential engagement by some variants with different effectors (Fig. 5). Previous genetic work in C. elegans has defined that, in vulval development, let-60 acts through a canonical Ras/Raf/MEK/MPK kinase cascade (reviewed by Sternberg and Han, 1998) but that the developmental outcomes can be modulated by low ligand dosage and ‘effector switching’ to Ral/MAP4K/p38 (Zand et al., 2011). Likewise, wild-type and oncogenic human Ras proteins engage distinct effector proteins and target pathways – including Raf, Ral and PI3K – and different mutant variants are altered in their ability to engage different effectors (Céspedes et al., 2006; Hunter et al., 2015). To test how the mutant variants relate to downstream effects, we utilized the small-molecule MEK inhibitor U0126, which blocks an essential component of the canonical Ras/Raf/MEK/MPK pathway and functions in nematodes (Reiner et al., 2008). We found that, as with let-60(n1046[G13E]), the Muv phenotype caused by each of the oncogenic variants is sensitive to 60 μM of inhibitor (Fig. 5A), indicating that in each case the ultimate outcome is dependent on the canonical kinase signaling cascade. We next assessed each variant for sensitivity to U0126 at different doses (Fig. 5B). Although we found the EC50 for Si(let-60(G12D)) to be significantly higher than that for Si(let-60(Q61R)) and Si(let-60(G13R)), the response of these latter variants was statistically not different to each other. Altogether, these results show the oncogenic variants can be grouped into clear and discrete functional classes, based on whether they confer a gain-of-function phenotype, sensitivity to gene dosage and inhibition of downstream targets.
Non-autonomous modulators of activated C. elegans let-60/Ras exhibit differential effects on oncogenic variants
In a previous genetic screen, we have identified a set of non-autonomous modulators of activated let-60 (Corchado-Sonera et al., 2022). These genes, specifically when knocked down in mesodermal tissues, revert the Muv phenotype of animals bearing the genotype let-60(n1046[G13E]) to a more wild-type one. We did also find two genes, hpo-18 (encoding ATP synthase F1ε) and szy-5 (encoding a zinc finger protein), that mediate this effect in genetically chimeric animals carrying the let-60(n1046[G13E]) mutation only in cells derived from the embryonic precursor AB (including VPCs), and the mutant allele of each modifying gene only in cells derived from the embryonic precursor P1 (including anchor cell, somatic gonad and most body-wall muscles). To ask whether these modifiers are specific to let-60(n1046[G13E]) or whether they impact the activity of other mutant variants, we used the same experimental framework in this current study to evaluate the sensitivity of the transposon variants that cause the Muv phenotype to the non-autonomous modulators (Fig. 6A). First, we observed that inclusion of the oncogenic genotype only in cells derived from the AB precursor is sufficient to cause the Muv phenotype, as is the case for let-60(n1046[G13E]) (Fig. 6B). Next, we evaluated the sensitivity of each variant to mutation of hpo-18 or szy-5. We found that let-60(G12D) and let-60(Q61R) are insensitive to disruption of either gene in P1-derived cells. let-60(G13R) exhibits sensitivity, but to a lesser extent than that observed with let-60(n1046[G13E]). We concluded that the modifiers are variant- rather than gene-specific in their effect and, preferentially, alter the phenotype of variants with a weaker impact on pathway activation.
DISCUSSION
In this study we evaluated a set of oncogenic variants for their function in C. elegans by using CRISPR-mediated genome editing and controlled single-copy insertion methods. Aiming to classify and uncover unique sensitivities for the different variants, we found they fall into discrete groups (Table 1). G13D and G12C failed to disrupt signaling sufficiently to cause a Muv phenotype. G13R, Q61R and G12D exhibited distinctions based on their sensitivity to gene dosage, MEK inhibition and genetic modifiers. While Q61R is more similar to G13R regarding the sensitivity to MEK inhibition and to G12D regarding sensitivity to the dosage and genetic modifier, it is noticeable that the mutant variants can be roughly classified according to ‘strength’ of the phenotypic effect. We anticipate that other morphological and molecular readouts (Burdine et al., 1998; de la Cova et al., 2017) may identify signaling perturbations in animals with a grossly wild-type phenotype, as well as more quantitative distinctions among each of the tested variants, allowing further classification. Importantly, however, even when using a simple measure reflecting the gross morphological and simple qualitative (Muv/nonMuv) consequences of these mutations, clear functional distinctions are apparent. We, therefore, hypothesize that mutations, such as let-60(gu287[G13D]) – while they confer no overt phenotype on their own – may provide a sensitized genetic background for enhancement screening, just as the let-60(n1046[G13E]) background is amenable to suppressor screening. Mutants disrupted for gap-1 – which encodes one of two GAP proteins that negatively regulate LET-60/Ras activity in the vulval cells – have, similarly, provided a sensitized background for identification of genetic enhancers (Liu et al., 2017; Stetak et al., 2008). We conclude that distinct features of different Ras variants can be uncovered by using the C. elegans vulval development system, and by using simple morphological criteria that are easily adapted to moderate and high throughput screening.
Our experiments identified features of Ras biology suitable to model by using C. elegans and defined how oncogenic variants differentially impact them. Although Ras alters developmental-fate choices in vulval development – rather than cell cycle progression directly – it provides a sensitive in vivo assay. Indeed, genetic studies in C. elegans have been crucial in establishing the order in which proteins act with Ras, and in identifying many novel but conserved proteins that influence Ras activity (reviewed by Sternberg and Han, 1998). For this current study, we chose to introduce human oncogenic variants into the C. elegans gene – rather than to ectopically express a human protein (Das et al., 2021) – to ensure species-typical gene regulation and that engagement with effector proteins retains any co-evolutionary relationships. In this context, we compared our variant-specific results to studies using human proteins. Specifically, we noticed that the two variants that fail to confer the Muv phenotype (G12C and G13D) exhibit relatively more wild-type behavior in an assay of intrinsic GTPase activity and a quantitative assay of RAF affinity (Hunter et al., 2015). We interpret that, in the C. elegans system, proteins capable of engaging the canonical kinase pathway (via the standard interaction with RAF) and regulation via intrinsic GTPase activity behave in a sufficiently wild-type manner, so that normal development proceeds. In contrast, G12D and Q61R exhibit similar dosage-insensitive properties in our assays but differ with respect to their sensitivity to U0126-induced inhibition of MEK. Biochemically, Q61R has noticeably compromised GTPase activity (Burd et al., 2014), consistent with a direct role in catalysis for Q61 (Buhrman et al., 2010), whereas G12D retains an intrinsic GTPase activity comparable to G12C and G13D (Hunter et al., 2015). KRAS variants with a bulky side chain at G12 – including G12D – exhibit lower affinity for RAF (Hunter et al., 2015); moreover, experimentally, mammalian G12D mutants preferentially engage alternative downstream pathways, such as PI3K or Ral (Céspedes et al., 2006). We hypothesize that these differences may underlie the phenotypic distinctions between G12D and Q61R in our assays. A Ral-mediated response to Ras has been shown to participate in C. elegans vulval development (Shin et al., 2019; Zand et al., 2011) and, although the activity of G12D ultimately depends on the canonical cascade, differential engagement of this alternative pathway might contribute to the phenotypic distinctions we observed at low inhibitor dose. Future experiments using genetic mutants and more quantitative readouts will be important in dissecting the altered cellular consequences of each mutant.
A key distinction among the Ras variants is the sensitivity of animals to both dosage of the mutant gene and inhibition of the downstream kinase MEK. The let-60(n1046[G13E]) allele is dosage sensitive, as it is only weakly semi-dominant (Fig. 4) (see also Han and Sternberg, 1990) and animals bearing four copies have a stronger Muv phenotype than those bearing two (Fig. 2). Likewise, we found dosage sensitivity for Si(let-60(G13R)) but not for Si(let-60(G12D)) or Si(let-60(Q61R)). This sensitivity might reflect differences in the capacity of wild-type proteins from the endogenous let-60(+) to interfere with or impact the mutant variant activity within active Ras multimers or nanoclusters (reviewed by Simanshu et al., 2023). Alternatively, the results might reflect the importance of expression level on oncogenic ‘strength’ of a particular mutant variant. Many studies that evaluate differences among oncogenic variants (including this one) utilize techniques to control expression levels and other features across conditions to better focus on the differences in protein activity. However, protein abundance, whether influenced by transcription, translation or stability, is a critical component that influences the phenotypic outcome of different mutations in human and mouse cells. For example, differences in transcription levels determine whether a particular mutation causes tumor formation or apoptosis (Sarkisian et al., 2007), and translation codon representation (affecting translation rate) can affect the tumorigenicity of mutants with identical transcriptional characteristics (Lampson et al., 2013). Experimentally, a wide range of mutations can be ‘oncogenic’ in cells that are not found in human disease samples, and these results are interpreted to relate to a ‘sweet-spot’ between oncogenic Ras dosage and dysregulation to yield a disease outcome (Hidalgo et al., 2022; Hood et al., 2023; Li et al., 2018). This sweet-spot, however, arises from the complex interaction of protein abundance, oncogenic variant and cellular context (Le Roux et al., 2022). In C. elegans, inputs that modulate protein abundance have, likewise, been shown to reduce the phenotypic effect of the let-60(G13E) variant (Kramer-Drauberg et al., 2020). Our results demonstrate that the vulval development process is highly sensitive to the effects of Ras dosage and mutant variant, and can be used to dissect these fundamental inputs and the cellular processes that modulate them.
MATERIALS AND METHODS
Worm maintenance and genetics
Caenorhabditis elegans strains were grown on NGM plates seeded with Escherichia coli strain OP50 as a food source (Stiernagle, 2006). Strains were grown and experiments performed at 20°C, unless stated otherwise. The wild-type C. elegans used was strain N2 Bristol. Specific strains and genotypes are listed in Table S1. All strains not deposited with the Caenorhabditis Genetics Center are available upon request.
Production of let-60(G13D) mutant alleles
Three let-60(G13D) mutant alleles, i.e. let-60(gu282), let-60(gu283) and let-60(gu287), were generated using the method described by Arribere et al., (2014). Single-guide RNA (sgRNA) targeting the let-60 genomic locus was introduced into plasmid pDD162 (Dickinson et al., 2015), yielding plasmid pBP1 (for primers and plasmids see Tables S2 and S3). pBP1 (50 ng/μl) was then injected into the germline of wild-type C. elegans hermaphrodites together with a single-stranded DNA-repair template (20 ng/μl) and a fluorescent reporter (myo-2::mCherry, pCFJ90, 25 ng/μl) as a transformation marker (for primers and sequences, see Table S2). mCherry-positive F1 animals were selected and allowed to self-cross. After offspring on each plate founded by an F1 parent depleted the bacteria, a portion was saved and the remainder prepared to recover genomic DNA. Genomic DNA was used as template for PCR. The PCR product was then subjected to Sanger sequencing to identify plates founded by – typically heterozygous – candidate mutants. Animals were recovered from these candidate heterozygous plates, individually plated and allowed to self-cross. Plates containing their resulting offspring (F2) plates were also screened via Sanger sequencing to identify homozygous strains.
Plasmids and single-copy insertion transgene production and validation
A 4.1 kb genomic fragment containing all transcribed regions and the 1 kb upstream sequence of let-60 was produced using Phusion DNA polymerase (NEB) and wild-type genomic DNA template, and assembled into the XhoI site of plasmid pCFJ151 (Frøkjær-Jensen, 2015), which contains a selectable marker, i.e. Cbr-unc-119(+), and flanking sequences to allow MosSCI insertions at ttTi5605. Sequences used for primer and plasmid design were downloaded from Wormbase (Davis et al., 2022). The insert was confirmed by restriction enzyme digest and Sanger sequencing (yielding plasmid pLK24). While it is not known if this sequence includes all important cis regulatory elements for let-60, our transgene experiments (Figs 2, 3) and sequence of cDNA from L3 animals (Fig. S3) demonstrate that the transposon-based gene is expressed and can impact vulva development. Mutant variants were introduced into this template, using PCR-based site-directed mutagenesis (Bachman, 2013). Mutations were confirmed with Sanger sequencing. All plasmids are available upon request.
Injection mixes [each let-60 plasmid produced above (25 ng/μl; MosI transposase-encoding plasmid pCFJ601 (50 ng/μl; Addgene plasmid #34874) and pCFJ90 (myo-2::mCherry (20 ng/μl; Addgene plasmid #19327) encoding a marker to exclude extrachromosomal arrays)] were injected into C. elegans strain EG4322 (WormBase ID: WBStrain00006687; https://wormbase.org/species/c_elegans/strain/WBStrain00006687#03--10 ), which contains the ttTi5605 landing site on LG II and unc-119(ed9). NonUnc F1s were plated individually. Those that segregated nonUnc offspring that were negative for mCherry fluorescence were considered further, and offspring were plated individually to establish homozygous lines that segregate nonUnc offspring. These lines were screened using PCR to ensure homozygosity and transgene insertion at the landing site (Fig. S2), followed by PCR and sequencing of the let-60 copy in the landing site (using transgene-specific primers, Table S2), to ensure retention of the intended mutation.
Plasmids corresponding to wild type, G13E or 12 different oncogenic variants were initially produced. All plasmids were injected as above, to produce at least 50 F1 nonUnc offspring. While one or more insertions corresponding to the variants reported here – i.e. wild type, G13D, G13R, G12D, G12C and Q61R, as well as G12S as a potential multi-copy insertion (Fig. S4) – were obtained, no insertion lines were produced with plasmids encoding G12 V, G12A, G12R, G13C, Q61 L or Q61 K. While additional trials might yield such lines, the presence of many sterile and sickly animals among nonUnc F1 offspring following these injections suggests that these variants cannot produce fertile strains, even as single copy in the MosSCI landing site, and that alternative methods, such as using conditional and/or cell-specific expression, are necessary to evaluate the function of these variants in the C. elegans vulval development system.
Expression of the introduced transgene (Fig. S3) was evaluated by harvesting RNA from synchronized L3 animals (Trizol), production of first strand cDNA using random primers (Superscript III), followed by PCR and Sanger sequencing. Since the transgene sequences incorporate genomic DNA, the primers amplify cDNA derived from RNA produced by both the endogenous locus and the transgene. Primers (Table S2) were selected to ensure that the sequence reads were produced from spliced cDNA, rather than genomic DNA.
Microscopy and phenotypic analysis
The multivulva (Muv) phenotype (Figs 1–4, Fig. S4) was assessed by selecting hermaphrodite animals at L4 produced from well-fed parents. These L4 animals were aged overnight, and scored for presence of ectopic vulval protrusions the next day. Animals of heterozygous or single transgene genotype (Fig. 4) were generated by crossing homozygous hermaphrodites with males from strain FX17749, which contains the balancer hT2 that was dominantly marked with myo-2::gfp (qIs48). GFP-positive hermaphrodite F1 animals at L4 were selected for scoring. Animals were assessed for MEK sensitivity and dose response (Fig. 5) by treatment with U0126 as in (Dawes et al., 2017). Briefly, L1 animals were plated onto 35 mm NGM plates spread with either MEK inhibitor U0126 (AdooQ Bioscience) or DMSO (solvent control) in M9 buffer (Stiernagle, 2006) and seeded with OP50 1 day prior to plating. Animals were allowed to mature for 2–3 days before they were scored for the Muv phenotype. Inhibitor concentration is reported as the calculated concentration using the full volume of NGM in the plate. Vulval cell development and induction (Fig. 1) were evaluated in animals at L4 using DIC optics at 100× magnification as in (Chamberlin et al., 2020). EC50 values were calculated and displayed in Fig. 5B using the Quest Graph calculator (AAT Bioquest, Inc., https://www.aatbio.com/tools/ec50-calculator).
Production of genetic chimeras
Genetic chimeras (Fig. 6, Fig. S4) were generated as described by Artiles et al. (2019) and Corchado-Sonera et al. (2022). let-60 transposon strains were crossed with strains bearing transgene ccTi1594 – which causes overexpression of GPR-1 in the germline, yielding segregation of maternal and paternal DNA into separate blastomeres in the zygote – to produce maternal strains CM2783, CM2982, CM3042 and CM3043 (for strain details see Table S1). Hermaphrodites from these strains were crossed with control (N2) males or heterozygous mutant males, i.e. szy-5(tm810)/hT2(qIs48) or hpo-18(ok3436)/nT1(qIs51), bearing deletion alleles of genes that had previously been shown to revert the Muv phenotype associated with let-60(n1046[G13E]) when disrupted in P1-derived (non-VPC) cells (Corchado-Sonera et al., 2022). Chimeric offspring in which the maternal genome segregates into AB (and paternal into P1) exhibit a characteristic pattern of fluorescence (mCherry from hjSi20) in the anterior pharynx that is easy to identify under a dissecting microscope (Artiles et al., 2019). mCherry-positive chimeric offspring that lacked GFP – i.e. indicating animals that lack the paternal balancer chromosome hT2(qIs48) or nT1(qIs51) – were selected at larval stage 4 (L4) and allowed to mature to adulthood to evaluate the vulva phenotype.
Sequence analysis
Amino acid sequences used in Fig. S1 were obtained from NCBI (LET-60, NP_502213.3) or UniProt (KRAS, P01116-1; HRAS, P01112-1; NRAS, P01111) and aligned using ClustalW (Thompson et al., 1994). The variant frequencies in human cancer samples were obtained from the GDC data portal (Data Release 38.0) (Grossman et al., 2016). All variants of KRAS, HRAS and NRAS were summarized from all cancer types across all GDC datasets. These data disproportionately reflect KRAS variants as 67% of the total mutations identified to affect this gene.
Statistics
A threshold of α=0.05 was selected for all statistical tests. Proportional data (percent Muv) were evaluated using a two sample proportions, two-tailed Z-test (using the calculator at https://www.socscistatistics.com/tests/ztest/), with a standard (α/k) Bonferroni correction for multiple tests where appropriate. Pairwise comparison of EC50 values in Fig. 5B was evaluated using one-way ANOVA and Tukey's honest significance (HS) test (using the calculator at https://www.socscistatistics.com/tests/anova/default2.aspx) A two sample, two-tailed t-test was used for the data in Figure 1H (using the t-test function within Excel).
Acknowledgements
We thank A. Dawes, K. Freytag and C. Genova (Ohio State University) for comments on the manuscript. L. Kelley and B. Pew (Ohio State University) provided technical assistance in the early stages of this project. H.M.C. thanks C. E. Burd (Ohio State University), G. Leone, T. McFall, R. Urrutia and M. Zimmermann (all Medical College of Wisconsin) for the valuable discussions that motivated these experiments. pDD162 (Addgene plasmid #47549) was a gift from B. Goldstein (University of North Carolina). pCFJ151 (Addgene plasmid #19330), pCFJ601 (Addgene plasmid #34874) and pCFJ90 (Addgene plasmid #19327) were a gift from E. Jorgensen (University of Utah). Some C. elegans strains were supplied by the Caenorhabditis Genetics Center, which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). Sanger sequencing of plasmids and PCR products amplified from genomic DNA was performed by the Ohio State University Comprehensive Cancer Center (OSUCCC) Genomics Core Shared Resource which is subsidized by an NIH Cancer Center Support Grant (P30CA016058). This work was supported by an NIH award to H.M.C. (R21OD030067).
Footnotes
Author contributions
Conceptualization: H.M.C.; Methodology: H.L., H.M.C.; Investigation: H.L., H.M.C.; Resources: H.L.; Writing - original draft: H.M.C.; Writing - review & editing: H.L., H.M.C.; Supervision: H.M.C.; Funding acquisition: H.M.C.
Funding
This work was supported by National Institutes of Health (NIH) (R21OD030067). Open Access funding was provided by OSU: The Ohio State University. Deposited in PMC for immediate release.
Data availability
Strains and plasmids are available upon request. All data are incorporated into the article and its supplementary material.
References
Competing interests
The authors declare no competing or financial interests.