Introduction

Mitosis is orchestrated by several intracellular signaling pathways to ensure proper cell division while maintaining genomic integrity. Errors during cell division, including chromosome mis-segregation or spindle defects, could lead to changes in chromosome content, producing aneuploid or polyploid progeny cells, which could be detrimental during development or lead to oncogenesis^1-3. Therefore, cells have evolved quality control mechanisms to ensure proper cell division during M phase. One such surveillance mechanism is known as the Mitotic Stopwatch Pathway (MSP). Foundational work from the Sluder lab in 2010⁴ first demonstrated a p53-dependent G1 arrest following prolonged mitosis. This was later expanded by three key studies published in 2016, which identified USP28 (ubiquitin-specific protease 28) and 53BP1 (p53-binding protein 1) as critical components of the pathway^5-7. During prolonged mitosis without centrosome loss, a ternary complex of 53BP1, USP28, and p53 forms and persists into the G1 phase, where it induces p21 transcription and enforces p53-dependent cell cycle arrest. In this pathway, 53BP1 and USP28’s deubiquitinase activity are required for p53 stabilization^8-10.

Proper mitotic arrest is critical for the efficacy of microtubule-targeting therapies such as taxanes (e.g., paclitaxel and docetaxel), which disrupt spindle formation and chromosome segregation. Pharmacological disruption of mitosis can induce cell death¹¹, as observed with paclitaxel treatement¹². Clinically, however, the effectiveness of taxanes is often limited, as many cancers develop resistance, including metastatic breast, ovarian, and non-small cell lung cancers¹³. This resistance is frequently associated with loss of MSP activity, for example due to defective p53 signaling^14-16. These observations underscore the urgent need to further elucidate the mechanisms underlying paclitaxel resistance in cancer.

Human germ cell-less protein-like 1 (GMCL1) is a putative substrate receptor of one of many CUL3-RING ubiquitin ligases (CRL3s). However, to date, the biological role of GMCL1 and its substrates have remained uncharacterized. GMCL1 received its namesake from its Drosophila melanogaster homolog, GCL (Germ Cell-Less), which plays essential roles in early embryonic development and germ cell determination^17,18. The role of GMCL1 in germ cell development appears to have been evolutionarily conserved as loss of GMCL1 expression in mice has been shown to cause defects in meiosis and spermiogenesis¹⁹, and altered GMCL1 expression was functionally associated with human asthenozoospermia²⁰. Although GMCL1 homologs have been primarily associated with germ cell biology, several databases (e.g., GTEx²¹ and GENT2²²) indicate that GMCL1 is expressed in somatic cells as well, suggesting broader biological functions beyond germ cell development. In contrast, the GMCL1 paralog GMCL2 is specifically expressed in germ cells.

Here, we present studies that clarify the role of GMCL1 in somatic cells. We show that, similar to its D. melanogaster counterpart²³, GMCL1 functions as a CRL3 substrate receptor. Furthermore, we identify 53BP1 as a bona fide substrate of CRL3^GMCL1 and demonstrate that its levels are regulated by GMCL1 during prolonged mitotic arrest.S By reducing mitotic 53BP1, GMCL1 inhibits the function of the USP28-p53-53BP1 mitotic stopwatch complex and limits p53 transmission to daughter cells. Based on these findings, we propose that GMCL1 inhibition may represent a potential novel approach to overcoming paclitaxel resistance in cancer cells with functional p53.

Results

Identification of 53BP1 as an interactor of GMCL1

We have shown that the GMCL1 ortholog in D. melanogaster, GCL, is a substrate recognition subunit of a CRL3 complex that is active specifically in mitosis²³. Therefore, we predicted the human GMCL1 to also behave as a CRL3 substrate receptor. GMCL1 contains a BTB (Broad-Complex, Tramtrack, and Bric-à-brac) and a BACK (BTB and C-terminal Kelch) domain (Supplementary Figure 1A), consistent with other CRL3 receptors. On its C-terminus, GMCL1 also contains a GCL domain (residues 379-515), which is distinct from Kelch domains commonly found in other CUL3 substrate receptors. This GCL domain is predicted to form a β-sandwich characterized by two opposing anti-parallel β-sheets, each made up of four β-strands^23,24 (Supplementary Figure 1A). A Dali search²⁵ of the GMCL1 C-terminal domain reveals that it has some structural homology to the MATH (meprin and TRAF homology) domain found in another CRL3 substrate receptor, SPOP²⁶, suggesting that the GCL domain in GMCL1 could potentially act as a protein-protein interaction motif to recruit substrates.

To investigate the role of GMCL1 in somatic cells, we used immunoprecipitation followed by mass spectrometry (IP-MS) to identify binding partners of GMCL1. Proteomics studies were performed by expressing and purifying the following FLAG-tagged proteins: (i) wild-type GMCL1 (GMCL1 WT), (ii) GMCL1 E142K (GMCL1 EK), which carries a mutation in the BTB domain that is predicted to disrupt its interaction with CUL3, and (iii) GMCL1 BTB/BACK-only (GMCL1 BBO) that lacks the GCL domain (Figure 1A-C). IP-MS analysis identified 1,765 potential binding partners that specifically interact with GMCL1 via its C-terminal domain. Using SAINT scores > 0.70 and FDR < 5%, this list was refined to 9 proteins that showed significant interaction with GMCL1 WT and GMCL1 EK, but not GMCL1 BBO, nominating 53BP1 as the most enriched protein (Figure 1B, Supplementary Table 1). This selective enrichment suggests that the C-terminal, “MATH-like” GCL domain of GMCL1 is critical for its interaction with binding partners, including 53BP1. The interaction between GMCL1, and endogenous 53BP1 and CUL3 was validated using immunoprecipitations (IPs) followed by immunoblotting (Figure 1C, D).

Figure 1.

To further study the direct interaction between GMCL1 and 53BP1, we used AlphaFold 3²⁷ to locate the positions of contact residues at their predicted binding interface. Consistent with our initial IP-MS experiment, Alphafold 3 model predicted that the C-terminal domain of GMCL1 would interact with 53BP1. Based on the predicted GMCL1-53BP1 complex structure, we identified Arg 433, which appears to be a solvent-exposed residue in GMCL1’s GCL domain that could interact with 53BP1 without impeding GMCL1 binding with CUL3. Thus, we generated the GMCL1 R433A (GMCL1 RA) point mutant and tested its binding to 53BP1 upon IP. As anticipated, compared to WT GMCL1, the R433A mutation completely abolished the binding of GMCL1 to 53BP1, but did not impact GMCL1’s binding to CUL3 (Figure 1D).

To determine which region of 53BP1 mediates its binding to GMCL1, we mapped the predicted GMCL1-binding site on 53BP1 and identified a conserved IEDI amino acid sequence within the Minimal Focus Forming region (MFF) of 53BP1²⁸ (Supplementary Figure 1B-E). Through a series of IPs, we demonstrate that a 53BP1 mutant either lacking the MFF region or containing the IEDI-to-AAAA mutation within this region lost its interaction with CRL3^GMCL1, suggesting that this conserved IEDI sequence within the 53BP1 MFF region forms a critical degron recognized by GMCL1 (Figure 1E). To examine the interaction between endogenous GMCL1 and endogenous 53BP1, we used CRISPR-Cas9 to knock in a FLAG tag at the C-terminus of GMCL1. Immunoprecipitation experiments confirmed that endogenous GMCL1 interacts with both endogenous CUL3 and 53BP1 (Figure 1F).

Finally, we sought to determine whether GMCL2, a GMCL1 paralog, is also able to interact with 53BP1. Immunoprecipitation of FLAG-tagged GMCL1 or GMCL2 from HEK293T cells revealed that while GMCL1 binds to 53BP1, GMCL2 does not (Supplementary Figure 1F), suggesting that GMCL1 and GMCL2 have distinct functions. Overall, our results suggest that GMCL1 is a CRL3 substrate receptor that interacts with 53BP1.

53BP1 is a bona fide substrate of GMCL1

To investigate whether GMCL1 regulates 53BP1 stability, we generated GMCL1 knock-out (KO) cells using CRISPR Cas-9²⁹ and compared 53BP1 levels between GMCL1 WT and two GMCL1 KO clones. RT-PCR analysis confirmed the loss of GMCL1 mRNA expression in the KO clones (Supplementary Figure 1G). We found that 53BP1 levels were significantly increased in our GMCL1 KO cells during M phase. While the whole cell extracts (WCE) showed modest differences in GMCL1 levels between the GMCL1 WT and KO clones, our fractionation experiments revealed that both 53BP1 and p53 mainly accumulated in the chromatin-bound fraction of GMCL1 KO M phase cells and this increase did not correspond to a decrease in 53BP1 levels in the soluble fraction (Figure 2A).

Figure 2.

To further explore the role of GMCL1 on 53BP1 stability, we stably reconstituted U2OS GMCL1 KO cells with either GMCL1 WT or binding mutants (i.e., GMCL1 EK and GMCL1 RA). Notably, the accumulation of 53BP1 in GMCL1 KO cells was rescued only upon re-expression of GMCL1 WT in the chromatin fraction. In contrast, both 53BP1 and p53 levels remained high in GMCL1 KO cells expressing either GMCL1 EK or GMCL1 RA (Figure 2B, Supplementary Figure 2A), emphasizing the importance of GMCL1’s ability to bind both CUL3 (through the E142 residue) and 53BP1 (through the R433 residue) to regulate 53BP1 levels. Of note, the R433A mutant was expressed at levels comparable to the WT protein. Interestingly, the E142K mutant showed reduced expression in mitotic cells, yet was the most abundantly expressed in asynchronous cells. Decreases in 53BP1 protein observed upon GMCL1 WT expression was not accompanied by an increase in the soluble fraction (Figure 2B, Supplementary Figure 2A), indicating that the reduction in chromatin-associated 53BP1 is not due to re-localization.

Next, we performed cycloheximide (CHX) to inhibit protein synthesis and to directly assess 53BP1 stability. In the chromatin-bound fraction, 53BP1 was more stable in GMCL1 KO cells compared to cells rescued with GMCL1 WT (Figure 2C). To further support the direct role of GMCL1 in regulating 53BP1 turnover, we co-expressed FLAG-tagged Trypsin-Resistant Tandem Ubiquitin-Binding Entity (TR-TUBE), a construct composed of tandem ubiquitin-binding motifs³⁰, together with either GMCL1 WT or the E142K mutant in HEK293T cells. TR-TUBE efficiently pulled down endogenous ubiquitinated 53BP1, with a marked increase in the presence of GMCL1 WT, but not the GMCL1 E142K mutant (Figure 2D and Supplementary Figure 2B). Collectively, these findings indicate that GMCL1 promotes 53BP1 ubiquitination and subsequent degradation during prolonged mitosis. Consequently, GMCL1-deficient cells retain elevated levels of 53BP1 and p53 following mitosis stress.

GMCL1-regulation of MSP affects cell cycle progression in daughter cells

During mitosis, 53BP1 helps monitor centrosome integrity and mitotic progression through the MSP complex (53BP1-USP28-p53), transmitting this information into daughter cells¹⁰. To investigate the effects of GMCL1 regulation on the MSP complex, we first analyzed post-mitotic GMCL1 KO daughter cells seven hours after release from prolonged nocodazole arrest. GMCL1 KO daughter cells reconstituted with GMCL1 WT, exhibited low levels of 53BP1, p53, and p21, along with reduced expression of apoptosis-related genes (Supplementary Figure 2C, D). In contrast, cells reconstituted with either the E142K or the R433A mutant displayed persistently elevated levels of 53BP1, p53, and p21, accompanied by increased expression of apoptosis-related genes (Supplementary Figure 2C, D).

To assess how GMCL1 levels affect cell cycle progression following mitotic delays, we performed FACS analyses of EdU incorporation in hTERT-RPE1 cells. Cells subjected to extended mitotic arrest were released by mitotic shake-off into fresh medium. After 30 hours, significantly less cells were found in S phase in the knockdown population, compared to control cells (Figure 3A, B and Supplementary Figure 3A). Importantly, co-depletion of GMCL1 with either TP53BP1 (hereafter 53BP1) or USP28 abolished this G1 arrest, and cells proceeded into S phase at levels comparable to or even exceeding those of control cells, effectively rescuing the phenotype caused by GMCL1 knockdown (Figure 3A,B and Supplementary Figure 3A).

Figure 3.

To further support our model, we overexpressed GMCL1 in hTERT-RPE1 cells and monitored cell cycle re-entry after mitotic delay. GMCL1-overexpressing cells progressed into S phase more rapidly than control cells, with notable entry observed as early as 18 hours post-release (Figure 3C, D and Supplementary Figure 4B). Importantly, in the absence of anti-mitotic drug treatment, GMCL1 KO cells exhibited no significant changes in baseline cell cycle profiles, regardless of reconstitution with WT or mutant GMCL1, compared to parental cells (Supplementary Figure 3C).

Together, these findings identify GMCL1 as a key regulator of cell fate under mitotic stress by acting upstream of the USP28-p53-53BP1 axis to influence cell cycle re-entry dynamics.

GMCL1 modulates taxane resistance in cancer cells

Taxanes, including paclitaxel, docetaxel, and cabazitaxel, are widely used chemotherapeutics that stabilize microtubules by preventing depolymerization³¹. However, resistance to taxanes commonly emerges through multiple mechanisms^13,32,33, including activation of pro-survival pathways such as destabilization of the 53BP1-USP28-p53 complex. Given our finding that GMCL1 controls 53BP1 stability during prolonged mitosis, we sought to investigate whether GMCL1 expression is associated with taxane resistance and 53BP1 protein abundance in cancer cells. To this end, we leveraged the PRISM (profiling relative inhibition simultaneously in mixtures) repurposing dataset, which quantifies the proliferation-inhibitory effects of 4,518 compounds across 578 cancer cell lines³⁴, and integrated this these data with DepMap proteomic and transcriptomic profiles (https://depmap.org). Since GMCL1 protein levels were not quantified in the dataset, we used GMCL1 mRNA expression as a surrogate to assess its association with resistance to taxanes (Figure 4A). Interestingly, we found that five cancer types (i.e., endometrial, breast, kidney pancreas, and upper aerodigestive tract cancers) with high levels of GMCL1 mRNA exhibited significant resistance to paclitaxel, cabazitaxel, and/or docetaxel (Figure 4B). In contrast, cell lines derived from 17 other cancer tissues with high GMCL1 mRNA expression did not show such significant correlation (Supplemental Figure 4A). Across multiple cancer types (Supplemental Figure 4B), we observed that in lung cancer cells with wild type p53, high GMCL1 expression combined with low 53BP1 levels was associated with significantly increased resistance to cabazitaxel and paclitaxel compared with cells showing low GMCL1 expression and high 53BP1 levels (Figure 4C). In contrast, this relationship was absent in p53-mutant lung cancer cells, where GMCL1 status did not correlate with taxane resistance (Figure 4C).

Figure 4.

To verify the impact of GMCL1 levels on paclitaxel sensitivity, we performed cell viability and apoptosis assays using cells with wild type or mutant p53. Paclitaxel treatment was chosen to mimic the conditions reported in DepMap. In p53 wild-type cells (MCF7 and U2OS), paclitaxel treatment led to a significant reduction in cell viability and an increase in apoptosis in GMCL1-depleted cells compared to cells transfected with non-targeting control siRNA (Figure 5A-D). However, GMCL1 knockdown did not affect cell viability or apoptosis in paclitaxel-treated cells with inactivated p53 (HeLa and HEC-1-A, respectively) (Figure 5E-H). Importantly, in hTERT-RPE1 cells, the reduction in cell viability and increase in apoptosis seen upon paclitaxel treatment in GMCL1 knockdown cells were rescued by simultaneous knockdown of either 53BP1 or USP28 (Figure 5I, J). These observations are consistent with the results in Figure 4 and suggest that paclitaxel resistance may, at least in part, be influenced by GMCL1 through the USP28-p53-53BP1 complex. Specifically, high GMCL1 expression appears to promote 53BP1 degradation, which in turn helps maintain lower p53 levels and reduces paclitaxel-induced cell death in cells with functional p53.

Figure 5.

Discussion

We identify GMCL1, a previously uncharacterized human CRL3 substrate receptor, as a regulatory component of the mitotic surveillance pathway. Specifically, we found that GMCL1 interacts with and mediates the degradation of 53BP1 during prolonged arrest in M phase. While 53BP1 is well known for its role in double-strand break (DSB) repair via non-homologous end joining (NHEJ)³⁵, it also participates in the so-called mitotic stopwatch composed of the 53BP1-USP28-p53 complex that stabilizes p53 during prolonged mitotic arrest^5-7,10. We showed that GMCL1 controls the levels of 53BP1 and, consequently, those of p53 in mitotic cells, thereby influencing p53 transmission to daughter cells (see model in Figure 5K).

We found that GMCL1 primarily regulates the levels of chromatin-associated 53BP1. A PLK1-dependent 53BP1 re-localization from chromatin to the nucleoplasm has been reported in mitosis^10,36,37. However, multiple GMCL1 KO clones display elevated chromatin-bound 53BP1 during M-phase arrest without a corresponding decrease in the soluble fraction, indicating the effect is not due to re-localization. The increased 53BP1 half-life in GMCL1 KO daughter cells supports this conclusion. Discrepancies in the reported localization of 53BP1 (chromatin vs. nucleoplasm) during mitosis may reflect differences in biochemical fractionation methods (e.g., differences in the concentrations of salt or detergent and/or sonication conditions).

Our findings suggest that GMCL1 functions as a regulator of mitotic stress response, with potential oncogenic properties in certain contexts. This underscores the role of GMCL1 in mitotic regulation and chromosome stability, which may vary based on tumor type, genetic background, and additional oncogenic mutations. Accordingly, we observed that a subset of cell lines exhibiting resistance to taxane-based agents, such as paclitaxel, cabazitaxel, and docetaxel, display elevated GMCL1 mRNA expression. Clinical application of our findings is currently limited by tumor heterogeneity and by the variable efficacy and cellular availability of taxanes during treatment.

GMCL1 has been primarily studied in the germ cells of D. melanogaster, where we have shown that it forms an active CRL3 complex during mitosis²³. Our new findings suggest a critical role for GMCL1 in mammalian somatic cell fate decisions by regulating 53BP1 stability during prolonged mitosis. We further show that GMCL1 loss sensitizes p53-wild-type, but not p53-mutant, cancer cells to paclitaxel-induced death, suggesting that GMCL1 inhibition may offer a selective therapeutic strategy for tumors with intact p53 function.

Materials and methods

Cell culture

Cell lines were purchased from ATCC and were routinely checked for mycoplasma contamination with the MycoStrip Mycoplasma Detection Kit (Invivogen). HEK293T (ATCC CRL-3216), HeLa (ATCC CCL-2) and hTERT RPE-1 (ATCC CRL-4000) were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco). U-2 OS (ATCC HTB-96), HCT-116 (ATCC CCL-247) and HEC-1-A (ATCC HTB-112.NM) cells were maintained in McCoy’s 5A medium (Gibco). MCF7 (ATCC HTB-22) were maintained in Eagle’s Minimum Essential Medium (EMEM) (ATCC). All media were supplemented with 10% fetal bovine serum (FBS) (Corning Life Sciences) and 1% penicillin/streptomycin/L-glutamine (Corning Life Sciences); however, MCF7 was further supplemented with human recombinant insulin (zinc solution; Gibco) to a final concentration of 11.2 µg/mL. All cell lines were maintained at 37°C and 5% CO2 in a humidified atmosphere.

Plasmids, siRNA, and transfection

Homo sapiens cDNAs were amplified by PCR using KAPA HiFi DNA Polymerase (Kapa Biosystems) and sub-cloned into a variety of vector backbones, including modified pCDNA3.1 and pLVX-PURO lentiviral vectors containing C-terminal Flag, HA or V5 tags,. Site-directed mutagenesis was performed using KAPA HiFi DNA Polymerase (Kapa Biosystems). All cell lines were transiently transfected using Lipofectamine 3000 (ThermoFisher Scientific) based on the manufacturer’s recommendation. siRNA oligo transfections were performed using RNAiMax (ThermoFisher Scientific) according to the manufacturer’s instructions.

Virus-mediated gene transfer

For the generation of lentivirus, HEK293T cells were transfected with pLVX constructs carrying the genes of interest, alongside the packaging plasmids pCMV-delta-R8.2 and pCMV-VSV-G. Viral supernatant was harvested 48 hours post-transfection, passed through a 0.45-μm sterile Millex-HV filter unit (Millipore Sigma), and supplemented with polybrene at a final concentration of 8 μg/ml (Sigma). Target cells were infected by replacing their culture medium with the virus-containing supernatant for an 8-hour incubation period. Selection of successfully transduced cells was performed using puromycin at a concentration of 1-2 μg/ml (Sigma).

CRISPR-Cas9 genome editing

CRISPR-Cas9 genome editing techniques were carried out as previously described³⁸ with modifications. In brief, to generate GMCL1-knockout U-2 OS cells, optimal gRNA target sequences closest to the start codon of the genes were designed using the Benchling CRISPR Genome Engineering tool (https://www.benchling.com). For transient Cas9 expression, gRNAs specific for GMCL1 gene was incorporated into the pRP [CRISPR]-Hygro-hCas9-U6 vector, which was obtained from VectorBuilder (https://en.vectorbuilder.com/). The following oligos were used to generate the proper gRNA in the vector: GMCL1 (5’-CGTGCCCCCACGTACCTTCG-3’). To generate GMCL1 2×Flag knock-in HCT116 cells, an optimal gRNA target sequence closest to the genomic target site and a ∼2 kb homologous recombination (HR) donor template was designed using the Benchling CRISPR Genome Engineering tool. The HR donor template was designed to introduce a 2×Flag tag in frame with the C terminus of GMCL1, in the following order: GMCL1-linker-FLAG-linker-FLAG-Stop codon and was purchased from VectorBuilder (https://en.vectorbuilder.com/). The following single gRNA sequence was used for the transient hCas9 expression vector: GMCL1 (5’-AAGTTACAGCAGATATATAA-3’).

Genomic DNA was collected using QuickExtract (Epicentre). Genotyping PCRs were performed with MyTaq HS Red Mix (Bioline), using primers surrounding the genomic target sites. The following primers were used for genotyping: GMCL1 (F: 5’-GCAGGCTTCTGATCTTCCCT-3’, R: 5’-ACTTGTCATCGTCGTCCTTGT-3’), and GMCL1 (F: 5’-GGGTGGGAGTTTGGAGAGTG-3’, R:5’-TCTGGATTTTCTGGGTGACGA-3’). The resulting PCR products were then purified and sequenced to determine the presence of insertion or deletion events. Clones positive for insertion or deletion events were then validated by western blot.

Antibodies

The following antibodies were used: β-actin (1:5,000, Sigma-Aldrich A5441), CUL3 (1:1,000, Bethyl Laboratories A301-109A), FLAG (1:2,000, Sigma-Aldrich F7425), GMCL1 (1:1,000, Proteintech 15575-1-AP), HA (1:2,000, Bethyl Laboratories A190-108A), Histone H3 (1:1,0000, Abcam, ab1791), p21 (1:1,000, Cell Signaling Technology 2947S), p53 (1:1,000, Proteintech 10442-1-AP), pHistone H3 (D2C8) (Ser10, 1:1,000, Cell Signaling Technology, #3377), Ubiquitin (p37) (1:1,000, Cell Signaling Technology, 58395S), USP28 (1:1,000, Proteintech 17707-1-AP), 53BP1 (1:2,000, Abcam ab36823), α-tubulin (1:5,000, Sigma-Aldrich T6074).

Drug treatment procedures

Where indicated, cells were treated with 400 ng/ml Nocodazole (Sigma-Aldrich M1404) for 16 hours, 100 nM paclitaxel for 48 hours, 10 μM MG132 for 3 hours, 2.5 μM MLN4924 for 3 hours, 100 μg/ml cycloheximide (CHX) for the indicated time.

Cell synchronization

Cells were synchronized using nocodazole. Cells were treated with 100 ng/ml nocodazole for 14 hours. Mitotic cells were then collected by shake-off (M phase cells), or washed three times with PBS, and replated in normal medium to allow them to resume cell cycle (to analyze subsequent cell cycle progression in the daughter cells).

qRT-PCR

Total RNA was purified using RNeasy mini kits (Qiagen). cDNA was generated using Double Primed EcoDry kits (Takara). The qPCR reaction was carried out using PowerUp SYBR Green (Applied Biosystems) and the Applied Biosystems QuantStudio 3 Real-Time PCR system in a 96-well format. ROX was used as a reference dye for fluorescent signal normalization and for well-to-well optical variations correction. Bar graphs represent the relative ratios of target genes to β-actin housekeeping gene values. For each biological sample, triplicate reactions were analyzed using absolute relative quantification method alongside in-experiment standard curves for each primer set to control for primer efficiency. The oligos used for qRT-PCR analysis were: β-actin (F: 5’-CATGTACGTTGCTATCCAGGC-3’, R: 5’-CTCCTTAATGTCACGCACGAT-3’), GMCL1 (F: 5’-GGAGATTCCTGACCAGAACATTG-3’, R: 5’-CGACTGGGCTTTATCAAGACAT-3), GMCL2 (F: 5’-CCACGCAGCGGGTCTGT, R: 5’-TGGATTTTCTGGGTGACGATTATTT), p21 (F: 5’-TGTCCGTCAGAACCCATGC-3’, R: 5’-AAAGTCGAAGTTCCATCGCTC-3’), NOXA (F: 5’-CCAAGCCGTGACCAAGGAC-3’, R: 5’-CGCCACATTGTGTAGCACCT-3’), PUMA delta (F: 5’-GCCAGATTTGTGAGACAAGAGG-3’, R: 5’-CAGGCACCTAATTGGGCTC-3’).

Fractionation and immunoprecipitation

For whole-cell lysates, cells were directly lysed in a buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.2% NP-40, 10% glycerol, 1 mM EDTA, 1 mM EGTA, 2 mM MgCl₂, and 1 mM dithiothreitol (DTT). Lysates were incubated on ice for 20 min and subsequently clarified by centrifugation at 20,000 × g for 15 min at 4 °C. When cellular fractionation was performed, it followed a previously established method³⁹. Briefly, cells were lysed in CSK buffer (10 mM HEPES, pH 7.4, 100 mM NaCl, 300 mM sucrose, 0.1% Triton X-100, 3 mM MgCl₂, and 1 mM EGTA) for 5 minutes. The soluble fraction was collected by centrifugation at 1,300 × g for 3 minutes at 4°C. Cell pellets were subsequently washed in CSK buffer and then lysed in chromatin extraction buffer (50 mM Tris-HCl, pH 7.4, 250 mM NaCl, 0.1% Triton X-100, 1 mM EDTA, 50mM NaF, 1 mM EGTA, 2 mM MgCl₂, and 250 U/mL Benzonase (Sigma-Aldrich)) for 30 minutes. Insoluble debris was removed by centrifugation at 20,000 × g for 15 minutes at 4°C. All buffers were supplemented with protease inhibitors (Complete ULTRA, Roche) and phosphatase inhibitors (Phosphatase Inhibitor Cocktail 2, Sigma-Aldrich).

For immunoprecipitation and affinity purification, samples were incubated with FLAG-M2 magnetic beads (Sigma-Aldrich) or anti-HA magnetic beads (Thermo Fisher Scientific) at 4°C for 2 hours. Beads were thoroughly washed with wash buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.2% NP-40, 1 mM EDTA, 1 mM EGTA, 2 mM MgCl₂, and 1 mM dithiothreitol (DTT), and protein elution was performed using either 3xFLAG peptide (Sigma-Aldrich) for mass spectrometry analysis or 1x Laemmli sample buffer for Western Blot analysis.

Immunoblotting

Western blotting was carried out as described previously³⁸. Protein samples were resolved under denaturing and reducing conditions on 4%–12% Bis-Tris gels (NuPAGE) and transferred onto PVDF membranes (Immobilon-P, Millipore). Membranes were blocked with 5% nonfat dried milk, incubated overnight at 4°C with primary antibodies, followed by washes and incubation with HRP-conjugated secondary antibodies (Amersham GE). Immunoreactive bands were visualized using enhanced chemiluminescence reagents (Pierce) and detected with a ChemiDoc MP imaging system (Bio-Rad). Each Western blot experiment was conducted at least three times to ensure reproducibility, with representative blots shown in the figures.

Cell cycle analysis by flow cytometry

EdU incorporation and propidium iodide staining were performed either on asynchronous or synchronized cells. Visualization of EdU and propidium iodide staining was performed following the instructions of the manufacturer³⁸. In brief, cells were pulsed with EdU (10 μM), fixed and permeabilized, and EdU was detected by copper-free click chemistry using the Click-iT™ Plus EdU Alexa Fluor™ 488 Flow Cytometry Assay Kit (Thermo Fisher Scientific). Flow cytometry analysis of cell cycle distribution was conducted using a CytoFlex Analyzer (Beckman Coulter) and data were processed with FlowJo v10 software (Becton Dickinson).

Mass spectrometry analysis of GMCL1 immunoprecipitations

The eluted anti FLAG-tag antibody purified protein complexes were reduced with 2 μl of 0.2 M DTT for 1 h at 57°C and subsequently alkylated with 2 μl of 0.5 M iodoacetamide (Sigma) for 45 min at room temperature in the dark. 250 ng of SP3 beads (Cytiva) were added proteins precipitated onto the beads by adding ethanol. Samples were placed in a thermomixer at 25°C for 10 min. Beads were washed three times with 80% ethanol and then digested overnight with 400ng of sequencing grade modified Trypsin (Promega) in 100 mM ammonium bicarbonate. Next, the samples were spun down 21,000 x g for 1 minute. The supernatant was transferred to a new tube while the beads were washed twice with 0.5% acetic acid. The washes were then combined with the supernatant collection. Samples were acidified with 10% TFA to pH1 and loaded onto a 0.1% TFA equilibrated Pierce C18 spin column using a microcentrifuge. The samples were rinsed twice with 0.1% TFA and twice more using 0.5% acetic acid. Peptides were eluted with 80% acetonitrile in 0.5% acetic acid. The organic solvent was removed using a SpeedVac concentrator and the sample reconstituted in 0.5% acetic acid.

An equal aliquot of each sample was loaded onto a trap column (Acclaim PepMap 100 pre-column, 75 μm × 2cm, C18, 3 μm, 100 Å, Thermo Scientific) connected to an analytical column (EASY-Spray column, 50 m × 75 μm internal diameter, PepMap RSLC C18, 2 μm, 100Å, Thermo Scientific) using the autosampler of an Easy nLC 1200 (Thermo Fisher Scientific) with solvent A consisting of 2% acetonitrile in 0.5% acetic acid and solvent B consisting of 80% acetonitrile in 0.5% acetic acid. The peptide mixture was gradient eluted using the following gradient: 5% solvent B for 5 minutes, 5-35% solvent B in 60 min, 35-45% solvent B in 10 min, followed by 45-100% solvent B in 10 min. The samples were acquired on the Orbitrap Eclipse using the following parameters: full MS spectra resolution of 120,000, an AGC target of 4e5, maximum ion time of 50 ms, scan range from 400 to 1,500 m/z. The MS/MS spectra were collected with the following parameters: a resolution of 30,000, an AGC target of 2e5, maximum ion time of 30 ms, one microscan, 2 m/z isolation window, normalized collision energy (NCE) of 27 and a dynamic exclusion of 30 s. To identify binding partners, all acquired MS2 spectra were searched against a UniProt human database using Sequest HT within Proteome Discoverer 1.4 (Thermo Fisher Scientific). Fixed modifications were set on cysteine (carbamidomethyl), variable modifications of oxidation on methionine, and deamidation on glutamine and asparagine. The resulting peptide spectra matches and proteins are filtered to better than 1% false discovery rate (FDR) and only proteins with at least two different peptides are reported. Proteins differentially expressed between GMCL1 WT and EK, as determined by SAINT scores⁴⁰ with a 5% FDR, were considered significantly enriched interactions when comparing to GMCL1 BTB.

Taxane resistance analysis

Taxane sensitivity data were obtained from the PRISM Repurposing dataset (DepMap 24Q2, www.depmap.org), which reports log2-fold changes (LFC) in cell viability across 578 cancer cell lines treated with various compounds, including paclitaxel, cabazitaxel, and docetaxel³⁴. Data preprocessing and integration: We integrated the GMCL1 RSEM-normalized mRNA expression (DepMap filename:

OmicsExpressionProteinCodingGenesTPMLogp1BatchCorrected.csv), with TP53BP1 protein abundance⁴¹ and TP53 mutation status (DepMap filename:

OmicsSomaticMutationsMatrixHotspot.csv) across all cell lines catalogued in the PRISM dataset into a harmonized dataset using R version 4.2.1 (Supplemental Table 2, 4A raw data). Cell lines with missing GMCL1 or TP53BP1 expression, taxane LFC data, or tissue type annotation were excluded from the analysis.

Stratification by expression levels: For tissue-level comparisons (e.g., breast, endometrium, kidney, pancreas, etc.), cell lines were stratified in parallel based on the median within tissue type of GMCL1 mRNA expression or TP53BP1 protein abundance into “high” and “low” GMCL1 or TP53BP1 groups, respectively. This allowed us to assess the relationship between baseline GMCL1 or TP53BP1 levels, TP53 mutation status, and taxane resistance under standardized, non-physiological screening conditions.

Statistical analysis: Differences in taxane sensitivity (LFC values) between high and low expression groups were assessed using two-sided, and unpaired Wilcoxon rank-sum tests between “high” vs “low” groups (* P < 0.05). To account for multiple testing across different tissue types and drug combinations, we applied the Benjamini Hochberg false discovery rate (FDR). Importantly, the dataset does not specifically isolate M-phase cells, but rather represents mixed populations, and the findings should be interpreted accordingly.

Cell viability and apoptosis assays

Cells (2500/well) were plated in a 96-well plate. The medium was replaced with 50 µL of medium containing the target siRNA (at a final concentration of 20 nM). After 24 hours, an additional 50 µL of medium containing either DMSO or paclitaxel (at a final concentration of 100 nM) was added. Luminescence was measured using BioTek Synergy Neo2 (Agilent) 48 hours post-treatment, using the CellTiter-Glo 2.0 Cell Viability Assay Kit (Promega) or RealTime-Glo Annexin V Apoptosis and Necrosis Assay Kit (Promega) following to the manufacturer’s recommendations.

Quantification and statistical analysis

Data analysis was performed using GraphPad Prism version 10.2.1. For comparisons involving three or more groups, One-way ANOVA followed by Bonferroni’s post hoc test or the Brown-Forsythe and Welch ANOVA test followed by Dunnett’S T3 multiple comparisons test was applied.

Supplemental figures

Supplemental Figure 1.

Supplemental Figure 2.

Supplemental Figure 3.

Supplemental Figure 4.

Data availability

We used taxane sensitivity from PRISM Repurposing (DepMap 24Q2) and merged it with DepMap 24Q2 transcriptome (GMCL1), TP53 mutation status, and TP53BP1 protein abundance.Original western blot images have been deposited at Mendeley at DOI:10.17632/gj3x6r263d.1 and are publicly available as of the date of publication. https://data.mendeley.com/preview/gj3x6r263d?a=1e452dfc-bf85-472c-83ce-0e997ba6fa40. The mass spectrometric raw files are accessible at https://massive.ucsd.edu under accession MassIVE MSV000097235 and at www.proteomexchange.org under accession PXD061458.

Acknowledgements

We would like to thank the NYU Proteomics Lab (supported in part by NYU School of Medicine and the Laura and Isaac Perlmutter Cancer Center Support grant P30CA016087 from the National Cancer Institute). We also thank the members of the Pagano Lab for helpful discussions. MP is an investigator with the Howard Hughes Medical Institute and his laboratory is supported by grant R35-GM136250 from the NIH. Y.K. is a recipient of the Japan Society for the Promotion of Science (JSPS) Postdoctoral Fellowships and the Uehara Memorial Foundation Postdoctoral fellowship. T.G.R. is grateful for funding from NIH Institutional training grant in Cell Biology (T32GM136542) and HHMI Gilliam Fellowship (GT15758). S.K. is supported by the K99 Career Development Award from NIGMS (1K99GM155613-01A1) and has been a Life Sciences Research Foundation (LSRF) awardee and an EMBO Long Term Postdoctoral Fellow. A. M is supported by US Department of Defense (DoD) (HT9425-24-1-0019) and National Cancer Institute (R01 CA296867-01A1). G.R. is supported by the Momentum Grant of the Hungarian Academy of Sciences (LP2023-15/2023), EMBO Installation Grant (IG5670-2024) and the HUN-REN Welcome Home and Foreign Researcher Recruitment Grant (KSZF-143/2023).

Additional information

Author contributions

Y.K., G.R. and M.P. conceived, supervised, and coordinated the study. T.G.R. performed all analyses from PRISM and DepMap data. Y.K. performed the mass spectrometric sample preparation, generated CRISIPR KO and KI cell lines, designed plasmid constructs, and performed the cell-based immunoprecipitations, fractionation and protein stability assays. A.M. and J.P. performed the initial mass spectrometric sample preparation and immunoprecipitations. B.M.U and J.O.P performed mass spectrometry analysis. S.F.G. performed structural predictions using AlphaFold3. The manuscript was written with inputs from all authors. Dr. Ruth Lehmann was included as a co-author in the original submission due to her intent to contribute to the manuscript’s preparation, but her name was later removed when she was unable to participate.

Funding

HHS | NIH | NIH Office of the Director (OD) (R35-GM136250)

• Michele Pagano

MEXT | Japan Society for the Promotion of Science (JSPS)

• Yuki Kito

Uehara Memorial Foundation (UMF)

• Yuki Kito

HHS | NIH | NIH Office of the Director (OD) (T32GM136542)

• Tania J González-Robles

Howard Hughes Medical Institute (HHMI) (GT15758)

• Tania J González-Robles

HHS | NIH | National Institute of General Medical Sciences (NIGMS) (1K99GM155613-01A1)

• Sharon Kaisari

Life Sciences Research Foundation (LSRF)

• Sharon Kaisari

European Molecular Biology Organization (EMBO)

• Sharon Kaisari

U.S. Department of Defense (DOD) (HT9425-24-1-0019)

• Antonio Marzio

HHS | NIH | National Cancer Institute (NCI) (R01 CA296867-01A1)

• Antonio Marzio

Magyar Tudományos Akadémia (MTA) (LP2023-15/2023)

• Gergely Róna

European Molecular Biology Organization (EMBO) (IG5670-2024)

• Gergely Róna

Hungarian Research Network (ELKH) (KSZF-143/2023)

• Gergely Róna

HHS | NIH | National Cancer Institute (NCI) (P30CA016087)

• Beatrix Ueberheide

Additional files

Supplementary Table 1

Supplementary Table 2