Introduction

Although the human body contains trillions of cells that could potentially be targeted by oncogenic mutations, cancers arise infrequently throughout a human lifetime, suggesting the presence of effective resistant mechanisms (Lowe et al., 2004). One key mechanism involves the induction of DNA replication stress by these oncogenic mutations, leading to cellular senescence or death (Hills and Diffley, 2014; Kotsantis et al., 2018). This mechanism is mediated through the activation of the DNA damage response (DDR) pathway and the tumor suppressor protein p53, which are triggered by DNA double-strand breaks. Only the cells that escape senescence and death, such as those harboring p53 mutations, can undergo transformation by oncogenic mutations, thereby initiating tumorigenesis (Gaillard et al., 2015; Igarashi et al., 2024; Macheret and Halazonetis, 2015). Among the most frequently mutated oncogenes in human cancers are the RAS genes (KRAS, NRAS, and HRAS), which mutations are present in 20-30% of all cancer cases (Gimple and Wang, 2019; Sanchez-Vega et al., 2018). These mutations often occur at codons 12, 13, and 61, resulting in RAS small GTPases being locked in a constitutively active, GTP-bound state (Moore et al., 2020). Notably, previous studies have shown that oncogenic HRAS^G12Vcan induce DNA replication stress via the DDR pathway, ultimately leading to cellular senescence (Di Micco et al., 2006; Serrano et al., 1997).

The Ras oncogene at 85D gene (Ras85D, hereafter referred to as Ras) in Drosophila exhibits high homology to human RAS genes (Neuman-Silberberg et al., 1984). Our previous research demonstrated that oncogenic Ras^G12V triggers cell death in Drosophila ovarian nurse cells (Zhang et al., 2024), a type of post-mitotic germ cells that undergo G/S endoreplication to become polyploid (Hammond and Laird, 1985) (Figure 1A). Notably, oncogenic Ras^G12V promotes the division of mitotic germ cells, the precursors to nurse cells. Furthermore, monoallelic deletion of the cyclin A (cycA) or cyclin-dependent kinase 1 (cdk1) gene (cycA^+/- or cdk1^+/-) suppresses the nurse cell death induced by oncogenic Ras^G12V (Zhang et al., 2024). While CycA, a key cyclin promoting cell division, is not expressed in normal nurse cells (Lilly et al., 2000; Lilly and Spradling, 1996), its ectopic expression in nurse cells can trigger their death (as observed in this study). These findings suggest that the nurse cell death induced by oncogenic Ras^G12V is primarily due to aberrant promotion of cell division, rather than DNA hyper-replication. Drosophila ovarian nurse cells thus offer a valuable model for studying this specific form of cellular stress.

Figure 1.

Ubiquitination not only mediates protein degradation through the proteasome or lysosome but also plays crucial regulatory roles in various biological processes (Dikic, 2017; Kwon and Ciechanover, 2017). These functions are largely determined by the topological structures of the polyubiquitin chains, where ubiquitin can be linked to another ubiquitin via any of its seven lysine (K) residues or its first methionine residue. Among these, K11- and K48-linked polyubiquitin primarily mediates the proteasomal degradation. In contrast, K63 linkage facilitates the lysosomal degradation of protein aggregates and damaged organelles through autophagy. Additionally, K63 linkage is involved in non-degradative processes such as DNA repair, kinase activation, and protein transport (Kwon and Ciechanover, 2017). Ubc13, a ubiquitin-conjugating (E2) enzyme (Ubc), is responsible for the K63 linkage, with its activity requiring a Ubc variant (Uev, E2) protein as a cofactor (Hofmann and Pickart, 1999; McKenna et al., 2001). These E2 enzymes have been shown to play regulatory roles in various biological processes, including cell death (Ma et al., 2013), DNA repair (Broomfield et al., 1998), DDR, and innate immunity (Andersen et al., 2005; Bai et al., 2020; Zhou et al., 2005).

However, it still remains unknown whether they also possess proteasomal proteolytic functions and whether such functions hold significant biological roles.

In this study, to investigate the intrinsic protective mechanisms against the nurse cell death induced by oncogenic Ras^G12V in Drosophila, we conducted a genome-wide genetic screen and identified Uev1A as a crucial protector. Mechanistically, Uev1A works in conjunction with the anaphase-promoting complex or cyclosome (APC/C) to degrade the essential cyclin Cyclin A (CycA) via the proteasome, highlighting the critical role of its proteolytic function. Additionally, Uev1A and its human homologs, UBE2V1 and UBE2V2, also counteract oncogenic Ras-driven tumorigenesis in diploid cells, inhibiting the overgrowth of Drosophila germline tumors and human colorectal tumor xenografts in nude mice, respectively. These findings suggest that upregulation of UBE2V1/2 could represent a promising therapeutic approach for human colorectal cancers with RAS mutations.

Results

Genetic screen identifies Uev1A as a crucial protector against the nurse cell death induced by oncogenic Ras^G12V

While oncogenic Ras^G12V triggers cell death in nurse cells (Zhang et al., 2024), it does not have the same effect in cystocytes, the precursor cells before differentiating into nurse cells (Figure 1A, D, E). To confirm Ras^G12V expression in cystocytes, we generated a UASz-flag-Ras^G12V transgenic fly strain, which phenocopied the effects observed with the previously used UASp-Ras^G12V. RAS small GTPases need to be anchored to the inner cell membrane for their proper function (Willingham et al., 1980). Using bam-GAL4-VP16 (Chen and McKearin, 2003), we overexpressed Ras^G12V specifically in cystocytes (bam>flag-Ras^G12V) at 29°C and observed membrane-localized Flag signals, validating the normal expression of Ras^G12V (Figure 1B). Also, Flag signals were detected in some early-stage nurse cells, although these signals gradually diminished over time (Figure 1B). However, nurse cell death remained very low in this context (Figure 1D, E). This may be attributed to either insufficient levels of the oncoproteins or the presence of a protective mechanism.

To explore this potential protective mechanism, we conducted a genetic modifier screen by introducing individual genome-wide Deficiencies (361 lines) into bam>Ras^G12V flies (bam>Ras^G12V; Deficiency/+, Figure 1C). Of note, ovaries from bam>Ras^G12V; Df#7584/+ flies exhibited the highest incidence of degrading egg chambers with dying nurse cells per ovary (see Source data 1). One gene deleted in this deficiency line is uev1a, and RNAi targeting uev1a reproduced the phenotype seen with Df#7584 (Figure 1D, E). In this and all subsequent experiments, we quantified the nurse-cell-death phenotype using the percentage of degrading to total egg chambers per ovary (Figure 1E), a method that is more precise than our approach in the genetic screen. To further verify the role of uev1a, we generated a UASz-uev1a transgenic fly strain. Overexpression of Uev1A successfully rescued the nurse-cell-death phenotype in bam>Ras^G12V; Df#7584/+ fly ovaries (Figure 1D, E), confirming that Uev1A is a crucial protector against the nurse cell death induced by oncogenic Ras^G12V.

Protective role of Uev1A against the nurse cell death induced by direct overexpression of Ras^G12V

In contrast to the scenario above (bam>Ras^G12V), direct overexpression of Ras^G12V in nurse cells using nos-GAL4-VP16 (Rørth, 1998; Van Doren et al., 1998) (nos>Ras^G12V) resulted in substantial cell death even at 25°C (Figure 1F, I) (Zhang et al., 2024). Such observation prompted us to investigate the role of Uev1A in this context. Notably, the incidence of dying nurse cells was markedly elevated in nos>Ras^G12V+uev1a-RNAi ovaries compared to the nos>Ras^G12V+GFP-RNAi control ovaries (Figure 1F, I). To further validate this, we generated two uev1a mutants using the CRISPR/Cas9 technique: uev1a^Δ1 and uev1a^Δ2. Both mutations consist of small deletions that cause frame shifts within the second coding exon of the uev1a gene (Figure 1G). Of note, flies with the uev1a^Δ1/Δ1, uev1a^Δ1/Δ2, uev1a^Δ1/Df#7584, uev1a^Δ2/Δ2, or uev1a^Δ2/Df#7584 genotype could not survive into adults, suggesting that both mutations are strong loss-of-function alleles. Consistent with the effects observed with uev1a-RNAi, either mutation markedly increased the incidence of dying nurse cells in nos>Ras^G12V ovaries (Figure 1H, I).

We next investigated whether upregulating Uev1A could mitigate the nurse cell death induced by direct overexpression of Ras^G12V. Remarkably, compared to the nos>Ras^G12V+lacZ control ovaries, the incidence of dying nurse cells was significantly reduced in nos>Ras^G12V+Uev1A ovaries (Figure 1K, L). Uev1A has two human homologs, UBE2V1 and UBE2V2, which share 67% (85%) and 67% (86%) sequence identities (similarities) with Uev1A, respectively (Figure 1J). We generated transgenic fly strains for both UASz-UBE2V1 and UASz-UBE2V2. Notably, overexpression of either UBE2V1 or UBE2V2 significantly mitigated the nurse cell death induced by oncogenic Ras^G12V, similar to the effects observed with Uev1A overexpression (Figure 1K, L). Taken together, these results further confirm the protective role of Uev1A against the nurse cell death induced by oncogenic Ras^G12V.

Then, we were intrigued by the potential of Uev1A to protect against the nurse cell death induced by other oncogenic mutations. Yorkie (Yki), a key oncoprotein in the Hippo pathway (Huang et al., 2005), has a hyperactive form known as Yki^3SA (Oh and Irvine, 2009). Its overexpression using nos-GAL4-VP16 (nos>Yki^3SA) at 29°C, but not at 25°C, could induce nurse cell death (Figure 1—figure supplement 1), albeit much less severe than that induced by nos>Ras^G12V at 25°C (compared with Figure 1F, I). Of note, the nurse cell death induced by oncogenic Yki^3SAwas significantly alleviated by Uev1A overexpression (Figure 1—figure supplement 1), implying a broad role of Uev1A in this process. However, due to the mild effect of Yki^3SA on triggering nurse cell death, we focused on Ras^G12V in our subsequent studies.

Oncogenic Ras^G12V triggers nurse cells to die through activating the MAPK pathway

To investigate how Uev1A protects against the nurse cell death induced by oncogenic Ras^G12V, we first sought to further explore the mechanisms driving this cell death. Since egg chambers contain both nurse cells and oocytes (Figure 1A), we considered the possibility that the death signal could originate from the oocytes. The Bicaudal D (BicD) gene is essential for oocyte determination, and egg chambers deficient in it fail to specify oocytes (Suter and Steward, 1991). We confirmed this by using nos>BicD-RNAi (Figure 2—figure supplement 1A). Importantly, nurse cell death was much more pronounced in nos>BicD-RNAi+Ras^G12Vegg chambers than in nos>BicD-RNAi+GFP control ones (Figure 2—figure supplement 1A, B), indicating that the death signal is intrinsic to the nurse cells rather than originating from the oocytes.

Figure 2.

Our previous research demonstrated that oncogenic Ras^G12V promotes excessive proliferation in germline stem cells (GSCs) by activating the mitogen-activated protein kinase (MAPK) pathway (Zhang et al., 2024). This finding led us to investigate whether the same pathway influences the nurse cell death induced by oncogenic Ras^G12V. Notably, knockdown of downstream of raf1 (dsor1, encoding Drosophila MAPKK) (Tsuda et al., 1993) or rolled (rl, encoding Drosophila MAPK) (Biggs and Zipursky, 1992) significantly mitigated the nurse cell death induced by oncogenic Ras^G12V (Figure 2A, B), highlighting the pivotal role of the MAPK pathway in this process.

CycA and the DDR pathway orchestrate the nurse cell death induced by oncogenic Ras^G12V

The Ras/MAPK pathway is well known to promote cell cycle progression (Gimple and Wang, 2019). Our previous work demonstrated that reducing the gene dosage of CycA or Cdk1 suppresses the nurse cell death induced by oncogenic Ras^G12V (Zhang et al., 2024). In addition to CycA and Cdk1, Cyclin B (CycB) and String (Stg) are two other essential regulators that drive cell division (Figure 2C) (Edgar and Lehner, 1996). To investigate whether these regulators could individually induce nurse cell death upon overexpression, we tested each and found that only CycA triggered this phenotype (Figure 2C, D). Notably, co-overexpression of CycA with Uev1A significantly suppressed this effect (Figure 2C, D), suggesting a genetic interaction between the two.

Oncogenic RAS can also induce DNA replication stress in diploid cells through forced cell cycle progression, thereby activating the DDR pathway and p53 to trigger cellular senescence or death (Di Micco et al., 2006; Serrano et al., 1997). To investigate whether DDR plays a similar role in Ras-induced death of polyploid nurse cells, we targeted two key DDR genes: telomere fusion (tefu, encoding Drosophila ATM) (Oikemus et al., 2004; Silva et al., 2004; Song et al., 2004) and loki (lok, encoding Drosophila Chk2) (Masrouha et al., 2003; Xu et al., 2001). Strikingly, knockdown of either gene markedly enhanced this cell death (Figure 2—figure supplement 2A, B), revealing a protective role for the DDR pathway. By contrast, knockdown of p53 suppressed this cell death (Figure 2—figure supplement 2A, B), demonstrating opposing functions for DDR and p53 in this system.

Given that RNAi targeting tefu or lok phenocopied that of uev1a, we hypothesized that Uev1A may protect against the nurse cell death induced by oncogenic Ras^G12V through the DDR pathway. Since DDR is triggered by oncogenic Ras^G12V- induced stress in this context, we speculated that Uev1A should be upregulated as a response. To test this, we initially attempted to generate an antibody against Uev1A using its full protein sequence as the antigen; however, this approach was unsuccessful. As an alternative, we created a uev1a-flag knock-in fly strain by inserting the coding sequence of a 3xFlag tag immediately before the stop codon ("TAG") of the uev1a gene (Figure 2—figure supplement 2C). These knock-in flies were homozygous viable and fertile, indicating that the Flag insertion did not disrupt Uev1A’s normal function. Previous studies have shown that Uev1A can activate JNK signaling (Ma et al., 2013), which acts in the surrounding follicle cells to promote nurse cell removal during late oogenesis (Timmons et al., 2016). Indeed, we detected Uev1A-Flag signals in such follicle cells (Figure 2—figure supplement 2D), confirming that these signals reflect endogenous Uev1a expression in Drosophila ovaries. Surprisingly, Uev1A-Flag signals were low in both nos>Ras^G12V; uev1a-flag/+ and uev1a-flag/+ control germ cells, including nurse cells, with no significant differences between the two groups (Figure 2—figure supplement 2E). This suggests that Uev1A expression remains at basal levels, rather than being upregulated by oncogenic Ras^G12V-induced stress in this context. Although the possibility that Uev1A may play a role in the DDR pathway at these basal levels cannot be ruled out, we were then prompted to explore its potential proteolytic function as an E2 enzyme.

Uev1A collaborates with the APC/C complex to mitigate the nurse cell death induced by oncogenic Ras^G12V

We first investigated whether Uev1A is directly involved in degradation of the Ras^G12V oncoproteins. To address the challenges of analyzing membrane-localized signals in nos>flag-Ras^G12V nurse cells due to severe cell death, we performed this assay in bam>flag-Ras^G12Vovaries at 29°C. Notably, similar membrane-localized Flag signals were detected in both bam>flag-Ras^G12V+Uev1A and bam>flag-Ras^G12V+GFP control germ cells, including nurse cells (Figure 2—figure supplement 3). This result suggests that Uev1A does not directly degrade the Ras^G12V oncoproteins.

APC/C complex primarily functions as an E3 ligase to facilitate the degradation of CycA during cell cycle progression (Sudakin et al., 1995). A compelling alternative model to explain our findings is that Uev1A collaborates with the APC/C complex to degrade CycA. Cell division cycle 27 (Cdc27, Drosophila APC3) is an essential part of the substrate recognition TPR lobe within the APC/C complex (Yamano, 2019). In bam>Ras^G12V+cdc27-RNAi ovaries, we observed dying nurse cells, a phenotype that was exacerbated with mutations in either uev1a^Δ1or uev1a^Δ2 (Figure 2E, F). Furthermore, knocking down cdc27 could increase the incidence of dying nurse cells in bam>Ras^G12V+uev1a-RNAi ovaries (Figure 2E, F). Additionally, we investigated Fizzy- related (Fzr), the Drosophila homolog of Cdh1, which plays a critical role as an activator of APC/C-dependent proteolysis and functions as a tumor suppressor in tumorigenesis (Greil et al., 2022; Sigrist and Lehner, 1997). Remarkably, nearly all egg chambers in bam>Ras^G12V+uev1a-RNAi+fzr-RNAi ovaries exhibited degradation (Figure 2E, F).

We also knocked down three additional APC/C complex genes in bam>Ras^G12V+uev1a-RNAi ovaries: shattered (shtd, encoding Drosophila APC1) (Tanaka-Matakatsu et al., 2007), morula (mr, encoding Drosophila APC2) (Kashevsky et al., 2002), and lemming A (lmgA, encoding Drosophila APC11) (Nagy et al., 2012). Among these factors, Shtd is a critical component of the scaffolding platform, while Mr and LmgA are essential components of the catalytic modules within the APC/C complex (Yamano, 2019). However, no significant enhancement in nurse cell death was observed (data not shown). This may be due to the relatively mild phenotype of nurse cell death in bam>Ras^G12V+uev1a-RNAi ovaries. To address this, we switched to knocking down these genes in nos>Ras^G12V ovaries. Notably, knockdown of any of them could significantly exacerbated the nurse cell death induced by oncogenic Ras^G12V (Figure 2G, H), similar to the effect observed with Uev1A downregulation. Collectively, these findings provide genetic evidence that Uev1A collaborates with the APC/C complex to mitigate the nurse cell death induced by oncogenic Ras^G12V.

Then, we investigated whether Uev1A and the APC/C complex protect nurse cells from death during normal oogenesis. Given that the G2/M-promoting CycA is absent in the G/S-endocycling nurse cells (Lilly et al., 2000; Lilly and Spradling, 1996), we used the cystocyte-specific driver, bam-GAL4-VP16, to conduct the assays at 29°C. Similar to nurse cells, Uev1A expression was at basal levels in cystocytes, with some expression detected in inner sheath cells and epithelial follicle cells (Figure 2—figure supplement 2E). Remarkably, nearly all egg chambers in bam>fzr-RNAi ovaries exhibited degradation (Figure 2—figure supplement 4), underscoring Fzr’s critical regulatory role in this context. Uev1A, lacking the conserved cysteine residue in its presumed active site for ubiquitin binding (Sancho et al., 1998; Zhou et al., 2005), typically partners with Bendless (Ben), the Drosophila homolog of mammalian Ubc13 (Muralidhar and Thomas, 1993; Oh et al., 1994), to perform the E2 enzyme function. Ovaries with the bam>cdc27-RNAi, bam>cdc27-RNAi+uev1a-RNAi, or bam>cdc27- RNAi+ben-RNAi genotype showed minimal degradation of egg chambers (Figure 2—figure supplement 4). However, degraded egg chambers were prevalent in bam>cdc27-RNAi+ben-RNAi; uev1a^Δ1/+ and bam>cdc27-RNAi+ben-RNAi; uev1a^Δ2/+ ovaries, where Uev1A, Ben, and Cdc27 are all downregulated (Figure 2—figure supplement 4). These results suggest that Uev1A, Ben, and the APC/C complex also work together to protect nurse cells from death during normal oogenesis.

Uev1A and the APC/C complex work together to degrade CycA via the proteasome

The APC/C complex is known to collaborate with the E2 enzymes UBE2C (Drosophila homolog: Vihar) and UBE2S (Drosophila homolog: Ube2S) to facilitate the proteasomal degradation of CycA during cell cycle progression (Greil et al., 2022; Yamano, 2019). This degradation involves the assembly of branched ubiquitin chains, incorporating K11, K48, and K63 linkages through a two-step mechanism (Meyer and Rape, 2014). However, the involvement of Uev1A in this process remains unexplored. To explore it, we performed biochemical assays in cultured Drosophila Schneider 2 (S2) cells, overexpressing tag-fused proteins using act-GAL4 to enhance expression efficiency. Co-immunoprecipitation (co-IP) assays revealed that Uev1A interacts with several components of the APC/C complex, including Mr, Cdc16 (Drosophila APC6), and Cdc23 (Drosophila APC8). In addition, Cdc27 was found to interact with CycA (Figure 3A, B and Figure 3—figure supplement 1). Remarkably, RNAi targeting uev1a, ben, or cdc27 significantly stabilized CycA proteins after the treatment with cycloheximide (CHX), an inhibitor of protein synthesis (Figure 3C-E). Since the K63 linkage primarily mediates lysosomal degradation via autophagy (Kwon and Ciechanover, 2017), we tested CycA stabilization with a chloroquine (CQ, a lysosome inhibitor) or MG132 (a proteasome inhibitor) treatment. Notably, MG132, but not CQ, treatment significantly stabilized CycA proteins (Figure 3F). These results suggest that Uev1A, Ben, and Cdc27 cooperate to promote CycA degradation through the proteasome rather than the lysosome.

Figure 3.

To directly validate this, we performed ubiquitination assays for CycA in S2 cells. Cdc27 significantly enhanced CycA ubiquitination, while this effect was markedly reduced upon the knockdown of uev1a, ben, or both (Figure 3G). This result indicates that both Uev1A and Ben are essential for APC/C-mediated proteasomal degradation of CycA. Furthermore, we explored the roles of the seven lysine (K) residues in ubiquitin for polyubiquitin chain formation. The 7KR (lysine to arginine) mutation completely abolished Cdc27-promoted polyubiquitination of CycA. Intriguingly, the 6KR+K48 mutation, in which all K residues except K48 were mutated to R residues, failed to restore polyubiquitination. In contrast, either 6KR+K11 or 6KR+K63 mutation significantly restored polyubiquitination (Figure 3H). These findings suggest that the K11 and K63 linkages are primarily responsible for CycA polyubiquitination in Drosophila cells. Together with previous studies (Hofmann and Pickart, 1999; McKenna et al., 2001), we propose that Uev1A and Ben mediate the K63 linkage in this process.

Uev1A inhibits the overgrowth of Drosophila germline tumors driven by oncogenic Ras^G12V

Given the absence of cell division in normal polyploid nurse cells (Hammond and Laird, 1985), their death induced by division-promoting Ras^G12V represents an artificial stress. Notably, Uev1A was not upregulated in response to this stress (Figure 2—figure supplement 2E), and it executed the function through degrading CycA (Figure 2 and 3). These findings prompted us to investigate whether Uev1A also counteracts oncogenic Ras-driven tumorigenesis in diploid cells, which undergo normal cell division. Our prior research demonstrated that oncogenic Ras^G12V markedly promotes the overgrowth of diploid bam-deficient germline tumors (Zhang et al., 2024), which are highly resistant to cell death (Zhang et al., 2023; Zhao et al., 2018). Intriguingly, knocking down uev1a significantly enhanced the overgrowth of these tumors, while overexpressing Uev1A suppressed it (Figure 4A, B). These results indicate that Uev1A also plays a role in counteracting oncogenic Ras-driven tumorigenesis in diploid cells.

Figure 4.

Considering the therapeutic potential of gene upregulation in inhibiting tumor growth, a critical concern is its impact on normal physiological processes. In this study, we examined the impact of Uev1A overexpression on Drosophila oogenesis and GSC maintenance. Notably, the nos>Uev1A flies remained fertile, and their ovaries appeared morphologically similar to those of the nos>lacZ control flies (Figure 4C), indicating normal oogenesis. Furthermore, each germarium in the ovaries of 14-day-old nos>Uev1A flies contained a similar number of GSCs as those in the nos>lacZ control flies (Figure 4D, E). These results suggest that Uev1A overexpression does not disrupt normal oogenesis and GSC maintenance.

UBE2V1 and UBE2V2 inhibit the overgrowth of human colorectal tumors driven by oncogenic KRAS

Our findings in Drosophila prompted us to explore the tumor-suppressive effects of UBE2V1 and UBE2V2 on human tumor growth. Using the Kaplan-Meier plotter (https://kmplot.com/analysis), we first evaluated the correlation between UBE2V1/2 expression and prognosis in several types of RAS-mutant cancer patients, including melanoma, myeloma, lung cancer, and colorectal cancer. Among them, higher expression levels of UBE2V1/2 were significantly associated with improved overall survival and relapse-free survival in KRAS-mutant colorectal cancer patients (Figure 5A). RNA-seq data from The Cancer Genome Atlas (TCGA) showed that UBE2V1 and UBE2V2 are transcribed at similar levels in colorectal cancer patients with oncogenic RAS mutations (including KRAS, HRAS, and NRAS) as in those without such mutations (Figure 5B). These findings suggest that UBE2V1 and UBE2V2 are not upregulated in response to oncogenic RAS mutations, paralleling the behavior of Uev1A in Drosophila ovarian nurse cells (Figure 2—figure supplement 2E).

Figure 5.

To investigate the potential tumor-suppressive effects of UBE2V1 and UBE2V2 on colorectal cancer cells, we overexpressed them in human SW480 cells (carrying the KRAS^G12V mutation) and HCT116 cells (carrying the KRAS^G13D mutation), respectively. Overexpression efficiency was confirmed by Western blot analysis (Figure 5—figure supplement 1A). Proliferation assays using 5-ethynyl-2’-deoxyuridine (EdU) demonstrated that overexpression of either UBE2V1 or UBE2V2 significantly inhibited the proliferation of KRAS-mutant colorectal cancer cells in vitro (Figure 5C, D). This inhibitory effect was further corroborated by clonogenic and cell viability (CCK8) assays, which showed that UBE2V1 or UBE2V2 overexpression markedly impaired cell colony formation and survival (Figure 5E–G).

To validate the tumor-suppressive effects of UBE2V1 and UBE2V2 in vivo, we generated stable SW480 cell lines overexpressing either of them, with overexpression efficiency confirmed by Western blot analysis (Figure 5—figure supplement 1B). Subcutaneous tumorigenesis assays in Balb/c nude mice showed that UBE2V1 or UBE2V2 overexpression significantly suppressed tumor formation (Figure 5H, I). Immunohistochemical staining further revealed a marked reduction in Cyclin A expression and Ki-67 positivity rates in tumors overexpressing either UBE2V1 or UBE2V2 (Figure 5J, K). Notably, the tumor-suppressive effect of UBE2V1 was more pronounced than that of UBE2V2, aligning with their ability to protect against the nurse cell death induced by oncogenic Ras^G12Vin Drosophila (Figure 1K, L). These results provide strong evidence for the tumor-suppressive effects of UBE2V1 and UBE2V2 on KRAS-mutant colorectal cancer.

Discussion

It is well established that oncogenic Ras can induce DNA replication stress, leading to cellular senescence or death (Hills and Diffley, 2014; Kotsantis et al., 2018). However, our previous (Zhang et al., 2024) and current findings demonstrated that oncogenic Ras^G12V can also trigger cell death in polyploid Drosophila ovarian nurse cells through aberrantly promoting their division. In this study, we performed a genome-wide genetic screen and identified the E2 enzyme Uev1A as a crucial protector against this specific form of cellular stress. Mechanistically, Uev1A collaborates with the APC/C complex (E3) to facilitate the proteasomal degradation of CycA, which overexpression alone can also trigger nurse cell death. Furthermore, Uev1A and its human homologs, UBE2V1 and UBE2V2, counteract oncogenic Ras-driven tumorigenesis in diploid Drosophila germline and human colorectal tumor cells, respectively (Figure 4F). These findings highlight the critical role of Uev1A in counteracting oncogenic Ras stimuli in both polyploid and diploid cells.

Our studies reveal that the DDR pathway protects against the nurse cell death induced by oncogenic Ras^G12V (Figure 2—figure supplement 2A, B), suggesting that this cellular stress both activates and is subsequently suppressed by the DDR. In contrast, p53 promotes nurse cell death under the same conditions (Figure 2—figure supplement 2A, B). Interestingly, previous research has also demonstrated distinct roles of p53 and Lok, a crucial DDR regulator, in regulating nurse cell death during mid-oogenesis. While Lok overexpression triggers nurse cell death, p53 overexpression does not. Furthermore, Lok-induced nurse cell death remains unaffected by p53 mutation, indicating that its mechanism operates independently of p53 (Bakhrat et al., 2010). The underlying mechanisms driving these differences warrant further investigation.

Previous studies have shown that Uev1A promotes cell death through activating JNK signaling (Ma et al., 2013). If this mechanism also applied to the nurse cell death induced by oncogenic Ras^G12V, we would expect a reduction in Uev1A levels to suppress this cell death. However, our observations revealed the opposite effect (Figure 1F-I). Additionally, we did not detect the expression of puc-lacZ, a reporter for JNK signaling (Glise et al., 1995), in nos>Ras^G12V nurse cells (data not shown). These findings suggest that Uev1A’s protective role in this context is independent of JNK signaling.

Of note, the suppressive effects of Uev1A on Ras^G12V; bam^RNAi germline tumors in Drosophila were less pronounced than that of UBE2V1/2 on oncogenic KRAS^G12V-driven human colorectal tumor xenografts in nude mice (compare Figure 4A, B with Figure 5H, I). Our previous research has shown that Ras^G12V; bam^-/- germline tumor cells divide infrequently (Zhang et al., 2024). This suggests that the relatively weak suppression observed in these tumors may be due to the fact that Uev1A’s mechanism of action is cell cycle-dependent: the more frequently tumor cells divide, the more effectively Uev1A can suppress tumor overgrowth. We did not observe significant negative effects of Uev1A overexpression on GSC maintenance in Drosophila ovaries (Figure 4D, E). This is likely because GSCs also divide infrequently (Morris and Spradling, 2011). Therefore, these findings underscore the importance of considering the potential negative effects of Uev1A/UBE2V1/2 overexpression in other contexts, particularly when cells divide rapidly.

RAS oncoproteins have long been considered "undruggable", due to the lack of deep-binding, targetable pockets (Moore et al., 2020). To date, targeted therapies have been approved only for KRAS^G12C (Canon et al., 2019; Ou et al., 2022; Skoulidis et al., 2021). Recently, small-molecule pan-KRAS degraders, developed using the proteolysis-targeting chimeras (PROTACs) strategy, have shown promise (Popow et al., 2024), though their clinical effectiveness and safety remain to be fully validated. Alternatively, targeting key regulators in the RAS signaling pathway—such as SOS, SHP2, Farnesyltransferase, Raf, MEK, ERK, and PI3K—has emerged as an attractive therapeutic approach (Moore et al., 2020). Recent research has also explored the strategy of hyperactivating oncogenic RAS to trigger cell death as a potential means to suppress the overgrowth of human colon tumor xenografts in nude mice (Dias et al., 2024). Additionally, targeting the cellular dependencies driven by oncogenic RAS may trigger RAS-specific synthetic lethality, presenting a promising therapeutic approach (Moore et al., 2020). Our findings highlight the tumor-suppressive effects of UBE2V1 and UBE2V2 on human colorectal tumors driven by oncogenic KRAS (Figure 5). Notably, both of these two E2 enzymes are relatively small, with UBE2V1 consisting of 147 amino acid residues and UBE2V2 145 residues (Figure 1J). This raises the exciting possibility that their upregulation, through mRNA delivery or small-molecule agonists, could offer a promising therapeutic approach for human cancer.

Materials and Methods

Key resources table

Data availability

The results of the deficiency screening are provided in Source data 1. All genotypes are listed in Source data 2, and the raw quantification data can be found in Source data 3.

Fly husbandry

Cross experiments were conducted at 25°C, except for some performed at 29°C as noted.

Transgenic flies

UASz-flag-Ras^G12V, UASz-uev1a, UASz-UBE2V1, and UASz-UBE2V2

The coding sequences (CDSs) of flag-Ras^G12V, uev1a-RA, UBE2V1, and UBE2V2 were cloned into the pUASz1.0 vector (DeLuca and Spradling, 2018). These plasmids were then microinjected into fertilized fly embryos with the genotype “y w nos-int; attP40”, resulting in the generation of transgenic fly strains. The DNA sequence encoding 3xFlag was as follows: ATGGACTACAAAGACCATGACGGTGATTATAAAGATCATGACATCGATTACAAGGAT GACGATGACAAGCTT.

uev1a ^Δ1 and uev1a ^Δ2 mutants

The gRNA targeting the second coding exon of the uev1a gene was designed with the DNA sequence “GATCCAGCTCCTCCAGTAAGCGG” and cloned into the pCFD3 vector (Port et al., 2014). The plasmids were then microinjected into fertilized fly embryos with the genotype “y v nos-int; attP40” to generate transgenic fly strain. These transgenic flies were subsequently crossed with “nos-cas9; FRT2A” flies to generate uev1a knockout mutants, which molecular information was identified by Sanger sequencing.

uev1a-flag knock in

The gRNA targeting the second-to-last coding exon of the uev1a gene was designed with the DNA sequence “GTTATAAACCGGAGCGTTGGTGG” and cloned into the pU6-BbsI-chiRNA vector. The homology arms consisted of 991 bp upstream of the stop codon ("TAG"), followed by a 3xFlag tag sequence, the “TAG” stop codon, and 979 bp downstream of “TAG”. To prevent cleavage by the gRNA, the targeting sequence within the upstream homology arm was mutated to “CCTACCCTGCGCTTCATTA”, ensuring that the amino acid sequence remained unaltered. These DNA sequences were synthesized into the pUASz1.0 vector between the BamHI and KpnI sites and used as a double-stranded DNA (dsDNA) donor for homologous recombination-mediated repair. The pU6-uev1a-gRNA-2 (100 ng/μL) and pUASz-uev1a-donor (100 ng/μL) plasmids were then co-injected into fertilized fly embryos with the genotype “nos-cas9 (on X)”. The resulting uev1a-flag knock-in fly strain was identified by PCR and confirmed by Sanger sequencing.

Immuno-fluorescent staining

Fly ovaries were dissected in the PBS solution, fixed by the 4% paraformaldehyde solution (diluted in PBS) for 3 hours, washed by the PBST solution (0.3% Triton X-100 diluted in PBS) for 1 hour at room temperature (RT), then incubated with the primary antibodies (diluted in PBST) overnight at 4°C, washed by the PBST solution for 1 hour at RT, incubated with the secondary antibodies and 0.1 μg/mL DAPI (diluted in PBST) overnight at 4°C, washed by the PBST solution for 1 hour at RT, and subsequently mounted using 70% glycerol (autoclaved). The primary antibodies were used at the concentrations as follows: mouse anti-Flag (Sigma, #F1804) at 1:500 and mouse anti-α-Spectrin at 1:100. The Alexa Fluor conjugated secondary antibodies were used at 1:2,000.

Construction of plasmids used in S2 cells

The CDSs of the genes were amplified from the cDNAs derived from either S2 cells or adult wild-type (w¹¹¹⁸) flies and subsequently cloned into the pUASt-attB vector. N- or C-terminal tags were incorporated based on prior studies or structural predictions from AlphaFold. The following proteins were tagged at the N-terminus: APC4, Ben, CDC16, CDC23, Cdc27, CycA, Fzr, Fzy, Ida, Mr, Ub, Ub^7KR, Ub^6KR+K11, Ub^6KR+K48, Ub^6KR+K63, and Uev1A. In contrast, APC7 was tagged at the C-terminus.

Culture and transfection of S2 cells

S2 cells were maintained in insect culture medium supplemented with 10% FBS and incubated at 27°C without CO₂. For transient transfection, the X-tremeGENE HP DNA transfection reagent was used. Cells were harvested 36 hours post-transfection for further analysis.

RNAi assay in S2 cells

Two double-stranded RNAs (dsRNAs) targeting distinct regions of the relevant genes were synthesized using the T7 RiboMAX™ Express RNAi System. S2 cells were seeded in 6-well tissue culture plates and treated with 15 μg of dsRNA in the culture medium, followed by incubation for 24 hours. Expression plasmids were then transiently transfected into the cells, after which an additional 10 μg of dsRNA was added. The cells were cultured for another 36 hours before harvesting. As a negative control, dsRNA targeting the AcGFP gene was used. The templates for dsRNA synthesis were generated by PCR amplification from S2 cell genomic DNA using the following primers:

uev1a#1-F: GAATTAATACGACTCACTATAGGGAGAACGGAATTTCCGCTTACTG

uev1a#1-R: GAATTAATACGACTCACTATAGGGAGACGGACCGATGATCATGCC

uev1a#2-F: GAATTAATACGACTCACTATAGGGAGACACTAAAGATCGAGTGCG

uev1a#2-R: GAATTAATACGACTCACTATAGGGAGATGCCAGCTTCAGGTTCTC

ben#1-F: GAATTAATACGACTCACTATAGGGAGACCACGTCGCATCATCAAG

ben#1-R: GAATTAATACGACTCACTATAGGGAGAAGTCTTCGACGGCATATTTC

ben#2-F: GAATTAATACGACTCACTATAGGGAGACAGATCCGGACCATATTG

ben#2-R: GAATTAATACGACTCACTATAGGGAGATCAGTCTTCGACGGCATATTTC

cdc27#1-F: GAATTAATACGACTCACTATAGGGAGATCGCCCAGGATCTGATTAAC

cdc27#1-R: GAATTAATACGACTCACTATAGGGAGAGCAGCGACAGATCCTTCTTC

cdc27#2-F: GAATTAATACGACTCACTATAGGGAGAGATGATGGGCAAAAAGCTAAAG

cdc27#2-R: GAATTAATACGACTCACTATAGGGAGACCATCGGCCGATTGTTTC

AcGFP-F: GAATTAATACGACTCACTATAGGGAGATGCACCACCGGCAAGCTGCCTG

AcGFP-R: GAATTAATACGACTCACTATAGGGAGAGGCCAGCTGCACGCTGCCATC

Immunoprecipitation (IP), immunoblotting (ib), in vivo ubiquitination and protein stability assays in S2 cells

IP assay

The transfected S2 cells were lysed using NP-40 lysis buffer supplemented with protease inhibitors at 4°C for 30 minutes. The supernatants were incubated with primary antibodies at 4°C for 3 hours, followed by incubation with Protein G Sepharose for 2 hours. The beads were washed five times with NP-40 lysis buffer and subsequently analyzed by Western blotting. The following antibodies were used: mouse anti-Flag (Utibody, #UM3009) at 1:200, mouse anti-Myc at 1:200, and mouse anti-HA at 1:200.

IB assay

After 36 hours of transfection, the S2 cells were harvested and lysed in 2×SDS protein loading buffer [0.25 M Tris-HCl (pH 6.8), 78 mg/mL DTT, 100 mg/mL SDS, 50% glycerol, and 5 mg/mL bromophenol blue]. The lysates were then boiled for 5 minutes and subjected to western blot analysis. The following antibodies were used: mouse anti-Flag (Utibody, #UM3009) at 1:5000, rabbit anti-Flag at 1:5000, rabbit anti-Myc at 1:3000, rabbit anti-HA at 1:3000, mouse anti-β-Actin (Zenbio, #200068-8F10) at 1:5000, and rabbit anti-β-Tubulin at 1:5000.

in vivo ubiquitination assay

30 hours post-transfection, S2 cells were treated with MG132 at a final concentration of 50 μM for 6 hours. Proteins were subsequently enriched through immunoprecipitation and analyzed by Western blot analysis.

Protein stability assay

S2 cells were treated with 20 µg/mL of the protein synthesis inhibitor CHX for specified time intervals prior to harvesting.

Human cell culture

The human colon cancer cell lines SW480 and HCT116, as well as the human embryonic kidney cell line 293T, were purchased from the American Type Culture Collection (ATCC) and cultured in Dulbecco’s Modified Eagle Medium supplemented with 10% FBS. The cultures were maintained at 37°C with 5% CO₂ in a humidified incubator. The lentiviral vector pCDH-CMV was used for gene overexpression, while pMD2.G and psPAX2 served as lentiviral packaging plasmids. Target plasmids, pCDH-CMV-UBE2V1 and pCDH-CMV-UBE2V2, were synthesized by the GENEWIZ company (Suzhou, China). Plasmids were transfected into 293T cells using Lipofectamine 2000 to generate lentiviral particles, according to the manufacturer’s protocol. SW480 and HCT116 cells were then transduced with these lentiviruses to establish stable cell lines expressing UBE2V1 and UBE2V2, respectively. To select for stably transduced cells, the cultures were maintained in medium containing 2 µg/mL puromycin.

EdU incorporation assay

Cells were seeded in 24-well plates at a density of 100,000 cells per well. Cell proliferation was assessed using the EdU incorporation assay kit, following the manufacturer’s instructions. Briefly, EdU reagent was added to the culture medium, and cells were incubated for 2 hours to label DNA-synthesizing cells. After incubation, the medium containing EdU was removed, and cells were washed with PBS, fixed with fixative solution for 10 minutes, and then subjected to a click reaction for fluorescence labeling. Fluorescence microscopy was used to capture images and calculate the percentage of EdU⁺ cells.

Colony formation assay

Cells were seeded in 6-well plates at a density of 500 cells per well and cultured at 37°C in a 5% CO₂ incubator for 10 days without medium change. After 10 days, cells were gently washed with PBS to remove non-adherent cells, then stained with 0.5% crystal violet solution (containing 10% methanol and 1% acetic acid) for 15 minutes. The plates were subsequently rinsed with water to remove excess stain and air-dried. Colonies were photographed and counted using image analysis software.

CCK8 cell viability assay

Cells were seeded in 96-well plates at a density of 2,000 cells per well. Cell viability was assessed over five days using a CCK8 kit following the manufacturer’s instructions. Measurements were taken on Day 0 (baseline) and daily from Day 1 to Day 4. At each time point, 10 µL of CCK8 solution was added to each well, and the plates were incubated for 2 hours at 37°C in a humidified incubator. Absorbance at 450 nm was measured using a microplate reader to assess cell viability and proliferation dynamics throughout the experimental period.

Western blot analysis

Cells were lysed with RIPA buffer containing protease inhibitors, and protein concentrations were determined using a BCA Protein Assay Kit. Protein samples (20 µg per lane) were separated by SDS-PAGE and transferred to PVDF membranes. Membranes were blocked with 5% non-fat milk and incubated overnight at 4°C with primary antibodies: mouse anti-β-actin (Abmart, #M20011) at 1:5,000, rabbit anti-α-Tubulin at 1:5,000, rabbit anti-UBE2V1 at 1:2,000, and mouse anti-UBE2V2 at 1:2,000. Following washes, membranes were incubated with HRP-conjugated secondary antibodies and developed using ECL. Band intensities were quantified by densitometry.

Immunohistochemistry assay

Tissue sections were deparaffinized with xylene, rehydrated through a graded ethanol series, and subjected to antigen retrieval in citrate buffer (pH 6.0) at 98°C for 10 minutes. Endogenous peroxidase activity was blocked using 3% hydrogen peroxide, and nonspecific binding was blocked with 5% BSA. Sections were incubated overnight at 4°C with primary antibodies: rabbit anti-CycA at 1:200 and rabbit anti-Ki67 at 1:2,000. After incubation with HRP-conjugated secondary antibodies, sections were developed using DAB, counterstained with hematoxylin, and mounted. Images were captured using a light microscope.

Animal experiment

Male Balb/c nude mice (7-week-old) were used to assess the tumorigenic potential of SW480 colon cancer cells. The mice were housed under specific pathogen-free conditions with a 12-hour light/dark cycle and had ad libitum access to food and water. SW480 cells were stably transfected with either the empty vector pCDH-CMV (control), pCDH-CMV-UBE2V1 (UBE2V1-OE), or pCDH-CMV-UBE2V2 (UBE2V2-OE), and 5×10⁶ cells were subcutaneously injected into the right flank of each mouse. Each group consisted of six mice. Tumor growth was monitored approximately one week after injection. Tumor volumes were measured every four days using calipers and calculated using the formula: volume = length × width²/2. The experiment was terminated when tumors reached ∼1,000 mm³, and mice were humanely euthanized according to institutional and national ethical guidelines.

Image collection and processing

Fluorescent images were captured using a Zeiss LSM 710 confocal microscope (Carl Zeiss AG, BaWü, GER) and processed with ZEN 3.0 SR imaging software (Carl Zeiss) and Adobe Photoshop 2022 (San Jose, CA, USA).

Acknowledgements

We gratefully thank Eric H. Baehrecke, Michael Buszczak, Zheng Guo, Ruth Lehmann, Erika Matunis, Duojia Pan, Zhaohui Wang, Wei Xiao, Addgene, ATCC, BDSC, CEMCS, DGRC, GenBank, and THFC for providing plasmids, cell lines, and fly strains. This study was supported by grants 32270841 and 32070871 from the National Natural Science Foundation of China (NSFC) to Shaowei Zhao.

Additional information

Author contributions

S.Z. conceived the project. S.Z., S.W., and H.Z. supervised the studies. Q.Z., Yun.W., X.F., Z.W., Y.Z., L.Y., Yue.W., M.Y., D.S., and R.Z. performed the experiments. S.Z. wrote the manuscript and all authors commented on it.