CRISPR/Cas9-mediated knockout of the ubiquitin variant Ub^KEKS reveals a role in regulating nucleolar structures and composition

Ubiquitination, which stands for the covalent binding of a highly conserved 76 amino acids ubiquitin protein (Ub), is one of the most common post-translational modification (PTM) mechanisms (Deribe et al., 2010; Komander and Rape, 2012; Swatek and Komander, 2016). The isopeptide linkage between Ub C-terminal domain and the targeted protein's lysine residue is the result of an enzymatic cascade mediated by three classes of proteins: E1 activating enzymes, E2 conjugating enzymes and E3 ligases (Jin et al., 2007; Rotin and Kumar, 2009; Deshaies and Joazeiro, 2009; Ye and Rape, 2009; Schulman and Harper, 2009; Smit et al., 2014; Clague et al., 2015). The large number of proteins playing a role in this cascade leads to a powerful, multilayered communication system (Herhaus and Dikic, 2015; Akutsu et al., 2016; Yau and Rape, 2016). Such a system enables Ub to play a crucial role in numerous cellular mechanisms such as protein location, protein stability and even protein–protein interaction (Komander and Rape, 2012).

In addition to the ubiquitin genes encoding the canonical Ub protein, other variants of Ub encoded by genes that were initially annotated as pseudogenes have been reported (Dubois et al., 2020). Pseudogenes are usually described as an altered copy of genes that cannot produce a functional protein. However, Ub^KEKS is an example of a ubiquitin variant encoded by the pseudogene UBBP4 and acts as a fully functional PTM (Cheetham et al., 2019; Dubois et al., 2020). Interestingly, Ub^KEKS can specifically modify a large number of proteins that differ from the canonical Ub, thanks to four amino acid differences: Q2K, K33E, Q49K and N60S (Dubois et al., 2020). Cells lacking expression of UBBP4 by CRISPR/Cas9 knockout display enlarged nucleoli with an accumulation of lamin A (Dubois et al., 2020). Nucleoli are highly dynamic membrane-less organelles whose morphology and size correlate with nucleolar activity and can be used to detect cellular stress (Boulon et al., 2010; Jacob et al., 2013; Núñez Villacís et al., 2018). Therefore, these observations suggest a role for proteins modified by Ub^KEKS in regulating some of the many nucleolar functions, such as the production of ribosomes or proteins sequestration (Andersen et al., 2005; Núñez Villacís et al., 2018; Frottin et al., 2019; Iarovaia et al., 2019; Dubois et al., 2020).

In correlation with the high number of mechanisms taking place in the nucleolus, many modifications, either post-transcriptional or post-translational, serve as regulators (Shcherbik and Pestov, 2010; Tiku and Antebi, 2018; Tsekrekou et al., 2017; Sloan et al., 2017; Iarovaia et al., 2019; Cruz and Lemos, 2021). Ub and Ub-like proteins have already been shown to interfere with several key mechanisms of the nucleolus. For instance, Ub has been shown to promote protein trafficking between the nucleoli and other nuclear compartment upon cellular stress (Iarovaia et al., 2019). Under normal conditions, Bloom syndrome helicase (BLM) eases ribosomal RNA transcription in the nucleolus by interacting with both RNA polymerase I and DNA topoisomerase I (Grierson et al., 2012, 2013; Tangeman et al., 2016). Poly-ubiquitination of BLM, however, allows its translocation to the nucleoplasm to engage in double strand breaks DNA repair pathways (Tikoo et al., 2013). On another hand, SUMOylation of nucleolar proteins can disrupt ribosome biogenesis by targeting processing factors (Shcherbik and Pestov, 2010; Finkbeiner et al., 2011a). Modification of protein PELP1 by SUMO2 prevents the recruitment of key regulatory complexes to pre-60S ribosomal particles in the nucleolus, disrupting the maturation of 32S rRNA into 28S rRNA (Haindl et al., 2008; Finkbeiner et al., 2011a,b; Raman et al., 2016). Finally, it is now known that several PTMs belonging to the Ub-like protein family can build up complex regulatory interplays among ribosomal proteins (Bursac et al., 2014; El Motiam et al., 2019). RPL11 is a ribosomal protein that can be modified by both SUMO2 and NEDD8 (Xirodimas et al., 2008; El Motiam et al., 2019). NEDDylation of RPL11 allows its stabilization and nucleolar localization (Boulon et al., 2010; Bursac et al., 2014), where it plays a role in ribosome biogenesis (Iarovaia et al., 2019). In case of nucleolar stress, RPL11's NEDDylation is replaced by SUMOylation (El Motiam et al., 2019): RPL11 then translocates into the nucleoplasm, triggering a p53-dependent response pathway (Boulon et al., 2010; Hein et al., 2013; Bursac et al., 2014; Bailly et al., 2016; Núñez Villacís et al., 2018; Iarovaia et al., 2019; El Motiam et al., 2019). All these examples, however, show only a fraction of the vast regulatory system taking place in the nucleoli.

Here, we investigate the functional impacts of Ub^KEKS on the nucleolus using data-independent acquisition (DIA) mass spectrometry and microscopy. Starting from a whole-cell point of view, we worked our way down to a single-cellular-compartment scale to identify potential roles for the Ub^KEKS variant.

Ub^KEKS modifies different proteins compared to the canonical ubiquitin's targets (Dubois et al., 2020). To assess the impact of endogenous Ub^KEKS on the whole-cell scale, several Ub^KEKS knockout HeLa clones were generated using CRISPR/Cas9. Two distinct sets of guide RNAs (gRNAs) were used to create either a deletion of 715 bp or 1659 bp in the UBBP4 gene (Fig. 1A) (Dubois et al., 2020). Since generated Ub^KEKS-knockout cells (HeLa 2.7 and HeLa 4.3) are clonal populations, using two different strategies allows the identification of phenotypes specific to Ub^KEKS depletion (i.e. phenotypes common to both knockout clones), while disregarding the phenotypes which might be due to the original parental cell chosen during CRISPR/Cas9 manipulation. Whole cell proteomes of wild-type and knockout cells were mapped using DIA mass spectrometry (Fig. 1B). From this perspective, a complete and representative human spectral library was generated and used to increase the protein identification's quality and reproducibility between replicates (Figs S1 and S2A). Principal component analysis (PCA) confirmed that Ub^KEKS-knockout cells can be easily differentiated from wild-type cells on a proteomic level (Fig. S2B). More than 5300 proteins were successfully identified in each sample (Fig. 1C). The cellular impact of Ub^KEKS is well balanced with a similar number of proteins downregulated or upregulated in HeLa 2.7 and in HeLa 4.3 clones. This observation agrees with the observation that Ub^KEKS does not target proteins for degradation (Dubois et al., 2020). Among the 26 proteins similarly regulated between both Ub^KEKS-knockout clones, 10 were downregulated and 16 were upregulated (Fig. 1D, Table S1). Among the 96 proteins specifically regulated for each clone, 23 and 42 were upregulated and downregulated in HeLa 2.7, respectively, and nine and 22 were upregulated and downregulated in HeLa 4.3, respectively. A protein–protein interaction network was generated for all significantly modulated proteins (26 proteins common to both Ub^KEKS-knockout clones and 96 proteins specific to one clone) in Ub^KEKS-knockout cells (Fig. S2C). Among the 122 proteins analyzed, 55 proteins did not have any link with the others. Also, not every protein common to both HeLa 2.7 and HeLa 4.3 had a connection. Functional enrichment analysis was performed using the online tool ShinyGo (version 0.76) and no significant function or pathway could be identified, suggesting that Ub^KEKS potential cellular function(s) cannot be identified directly on such a large-scale analysis. However, since only a small number of proteins was found significantly modulated on a whole proteome scale (91 out of 5357 proteins for HeLa 2.7 and 57 out of 5357 proteins for HeLa 4.3), it was also deduced that Ub^KEKS does not play a role in protein stability or protein degradation.

The absence of Ub^KEKS does not impact the number of nucleoli per nucleus in HeLa cells (Fig. S3). However, HeLa cells lacking the Ub^KEKS protein were previously shown to have LMNA foci accumulation in the nucleoli, as well as an increase in the size of the nucleoli, suggesting a potential function for Ub^KEKS within this membraneless nuclear structure (Dubois et al., 2020). Nucleoli are composed of three distinct concentric sub-compartments (Fig. 2A). From the edges to the center, the granular component, the dense fibrillar component and the fibrillar center are separated by liquid–liquid phase separation (Lafontaine et al., 2021). In wild type, as well as in Ub^KEKS-knockout HeLa cells, nucleolar organization can be easily observed by confocal microscopy using immunofluorescence labeling (with endogenous NPM1 and fibrillarin antibodies), as illustrated in Fig. 2B and Fig. S4. To further investigate the difference in nucleolar morphology between wild-type and Ub^KEKS-knockout cells, a cleaner antibody for labeling the granular component was chosen (Nucleolin instead of NPM1 antibodies). Therefore, three different nucleolar proteins specific to each sub-compartments of the nucleolus were selected for immunofluorescence analysis. Nucleolin was used for detecting the granular component, fibrillarin for the dense fibrillar component and UBF for the fibrillar center (Fig. 3A-C). While there was no difference in the localization of either of them, the labeling showed that the shape of the nucleoli in the Ub^KEKS-knockout cells was more rounded with a clearly defined edge (Fig. 3A). To further study this observation, the circularity of the nucleoli was measured and quantified (Fig. 3D). Knockout cells showed significantly more round-shaped nucleoli than wild-type cells, and the reintroduction of Ub^KEKS protein through transient transfection rescued or reduced the phenotype in both knockout cell lines (clone 4.3 and clone 2.7 respectively; Fig. 3D, Fig. S5). The morphology of the nucleoli was further investigated using transmission electron microscopy (Fig. 4A). Interestingly, both Ub^KEKS-knockout clones showed a disorganized dense fibrillar component (DFC) and fibrillar center (FC) sub-compartments. While in the wild-type cells DFC formed a circular structure around the FC, cells lacking Ub^KEKS had misshapen structures either with distorted forms or less DFC surrounding the FC region. To address whether this structural disturbance is associated with ribosomal RNA production, quantitative PCR (qPCR) was performed for the 45S precursor as well as the 18S and 28S ribosomal RNAs transcribed by RNA Pol I. As a control for RNA Pol II transcription a long non-coding RNA MALAT1 levels have also been measured (Fig. 4B,C). None of the ribosomal RNAs showed difference of abundance in the Ub^KEKS-knockout cells, suggesting that ribosomal RNA transcription or processing is likely not affected by the disturbed nucleolar morphology observed. Therefore, a more general approach using mass spectrometry was chosen to identify possible functional consequences of such a disturbance in nucleolar organization.

Nucleoli from both wild-type and Ub^KEKS-knockout cells were successfully isolated using different sucrose cushions and successive centrifugations (Fig. 5A) (Li and Lam, 2015). Nucleoli isolated from Ub^KEKS-knockout cells display significantly larger nucleoli compared to wild-type HeLa cells (Fig. 5A,B). It is also worth mentioning that the difference of nucleoli's circularity shown previously is clearly visible in the entire cell pictures (Figs 3D and 5A). DIA mass spectrometry was used once again to identify the differences in protein composition which are responsible for enlarged nucleoli in Ub^KEKS-knockout cells. Similar to the total cell extracts analysis, a new spectral library was generated from nucleoli isolated from wild-type HeLa cells (Fig. 5A,bottom left; Fig. S6A). The resulting library displays an enrichment of nucleolar proteins (Fig. S6B, Table S2). Out of the 250 most abundant proteins, 53.2% are known to be located in the nucleolus and 22.40% to the nucleoplasm by current literature. Functional enrichment of those 250 proteins also highlights cellular pathways characteristic of nucleoli, mainly related to ribosome biogenesis, ribonucleoprotein complex biogenesis, non-coding RNA processing or ribosomal RNA maturation (Fig. S6C). Correlation coefficients were higher than 0.9 within each cell line, indicating that the isolation of nucleoli and the DIA mass spectrometry analysis were highly reproducible (Fig. S7A). Also, the protein composition of the nucleoli of each knockout clone can clearly be differentiated from wild-type nucleoli as shown by PCA analysis (Fig. S7B). Although the absence of Ub^KEKS leads to changes in protein composition, it does not result in an imbalance in the overall number of proteins in this organelle (Fig. 5C). Among significantly modulated proteins, seven were downregulated and 11 were upregulated in both HeLa 2.7 and HeLa 4.3 (Fig. 5D, Table S3). Interestingly, two additional proteins were found significantly modulated in both Ub^KEKS-knockout clones but showed opposite trends: ERCCL6 and NCK1 were upregulated in one of the clones while being downregulated in the other (Table S3). For protein–protein interactions and functional enrichment analyses, all 20 proteins which are common to HeLa 2.7 and HeLa 4.3 were considered. These proteins do not directly interact between one another (Fig. S7C) but seem to play a role in apoptosis regulation (Fig. 5E).

Among the proteins differentially regulated between nucleoli from KO cells and WT cells (Table S3), IFI16 and p14ARF were chosen due to their involvement in apoptosis pathway regulation (Kwak et al., 2003; Veneziano et al., 2016; Li et al., 2021) and used to validate the mass spectrometry data by western blot and immunofluorescence (Fig. 6A-C). The results confirmed the significant increase in the levels of both proteins in the nucleoli of Ub^KEKS-knockout clones. Interestingly, such increases are also detectable on a total cell extract level by immunoblotting (Fig. 6A) and partially correlate with data from the whole-cell proteome mapping: p14ARF was also identified as significantly upregulated by DIA mass spectrometry (Table S1) whereas IFI16 was labelled as non-significantly modulated. To further investigate functional consequences of these increases in apoptosis-related protein levels, cellular death was examined in wild-type and Ub^KEKS-knockout cells using Propidium Iodide (PI) and Annexin V labeling by flow cytometry (Fig. 6D). A total of four populations were identified due to this double labeling: viable, necrotic, early apoptotic or late apoptotic cells. The percentage of necrotic cells was not affected by the presence or absence of Ub^KEKS. However, Ub^KEKS-knockout clones showed a significantly higher percentage of apoptotic cells compared to wild type. A closer look at all four populations allowed us to determine that this apoptotic increase was due to an increase of early apoptotic cells numbers in Ub^KEKS-knockout clones. Impacts of a higher apoptotic rate on cell proliferation were monitored in real time by seeding 5 000 wild type or Ub^KEKS-knockout cells in XCELLigence RTCA E-plates and making regular cellular impedance measurements (Fig. 6E). The absence of Ub^KEKS strongly decreases cellular growth with notable differences that can be seen as soon as 40 h after seeding. Pictures of cells colored with Crystal Violet confirmed this proliferation delay due to the lack of Ub^KEKS (Fig. 6E). While 1 week of real-time monitoring allowed wild-type cells to be over confluent, this time was not even enough for Ub^KEKS-knockout cells to reach confluency.

For decades, Ub and Ub-like proteins regulate countless cellular mechanisms by forming complex messages through PTMs (Komander and Rape, 2012; Herhaus and Dikic, 2015; Akutsu et al., 2016; Yau and Rape, 2016). The discovery of Ub variants encoded by what was thought to be pseudogenes adds yet an additional level of regulation to this family of PTMs. Here, we first showed that deletion of Ub^KEKS does not result in a significant global change in the whole cellular proteome. This observation supports a different cellular role for Ub^KEKS than proteasomal degradation. Meanwhile, at a more precise scale, we also demonstrated that the nucleolus is a cellular compartment particularly impacted by the absence of Ub^KEKS. Changes in the morphology and internal structure of the nucleoli were observed in Ub^KEKS-knockout cells, including bigger and more circular nucleoli and structural disorganization of dense fibrillar component and fibrillar center. Previous studies have already linked nucleolar morphology to environmental stimuli and molecular composition (Boulon et al., 2010; Jacob et al., 2013; Núñez Villacís et al., 2018; Frottin et al., 2019; Lafontaine et al., 2021). Here, mass spectrometry analysis revealed that the absence of Ub^KEKS results in nucleolar protein composition changes, specifically pointing toward proteins involved in cell cycle regulation and stress response mechanisms. In consequences, we also observed an increase in basal levels of apoptosis and a significant cell proliferation delay in Ub^KEKS-knockout cells.

DIA mass spectrometry on total cell extracts showed that absence of Ub^KEKS does not impact protein stability or degradation. This confirms previous literature work where inhibition of the proteasome by MG132 failed to induce the accumulation of proteins modified by Ub^KEKS in cytoplasmic foci, contrarily to proteins modified with Ub (Dubois et al., 2020). Although proteasomal degradation is the most studied Ub function, many other mechanisms which do not impact protein levels are also regulated by Ub and Ub-like proteins, including cell-cycle control, DNA damage response and intracellular protein trafficking (Komander and Rape, 2012; Van der Veen and Ploegh, 2012; Akutsu et al., 2016; Pérez Berrocal et al., 2020). Since the absence of Ub^KEKS triggers the accumulation of LMNA in the nucleolus (Dubois et al., 2020) and changes of the nucleolar proteome (this paper), Ub^KEKS may act as a regulator for nucleolar protein trafficking.

Nucleolar structures heavily rely on liquid–liquid phase separation principle for its formation (Lafontaine et al., 2021). Multiphase liquid immiscibility is the result of biological molecules’ condensation in denser phase, stabilized by intra-macromolecular interactions (Keiten-Schmitz et al., 2021). Many factors, including intracellular pH and the content of proteins and RNA in the membrane-less compartment, can impact multiphase liquid immiscibility (Martin et al., 2015; Frottin et al., 2019; Stenström et al., 2020; Lafontaine et al., 2021; Keiten-Schmitz et al., 2021). Regarding macromolecules, protein concentration, protein–protein or protein–RNA interactions, and the presence of intrinsically disordered regions in protein and PTMs were characterized as key elements to gain correct nucleolar concentric structures (Martin et al., 2015; Stenström et al., 2020; Lafontaine et al., 2021; Keiten-Schmitz et al., 2021). As Ub^KEKS appears to be an important PTM for nucleolar organization (Dubois et al., 2020), its depletion results in nucleolar disorganization thus disrupting one of the main criteria of the liquid–liquid phase separation principle. Although nucleolar functions like protein sequestration or DNA repair do not require the correct concentric structure of this cellular compartment, ribosomal biosynthesis is highly dependable on it (Tsekrekou et al., 2017; Núñez Villacís et al., 2018; Iarovaia et al., 2019; Lafontaine et al., 2021). Here, we showed by qPCR that ribosomal RNA maturation does not seem impacted by the absence of Ub^KEKS and that only one ribosomal protein paralog (RPL22L1) was significantly modulated by Ub^KEKS as shown by DIA mass spectrometry. This protein, however, is not involved in ribosomal biogenesis either (Fahl et al., 2022). As mentioned previously, the hypothesis that Ub^KEKS could be a nucleolar protein trafficking regulator is quite appealing. Those new insights favour this assumption as impairment of protein exchange between nucleolus and nucleoplasm leads to changes in nucleolar composition, themselves leading to a disruption of the multiphase liquid immiscibility and disorganized nucleoli. Since the absence of Ub^KEKS has likely no impact on ribosomal biogenesis, a wider approach by DIA mass spectrometry was chosen to map the consequences on nucleolar proteins trafficking.

DIA mass spectrometry is a technique that fully developed over the last decade (Doerr, 2015). In contrast to the more common data dependent acquisition (DDA) mass spectrometry which only fragments and analyzes peptides with high intensity, DIA mass spectrometry process and investigates all peptides in a given sample (Doerr, 2015; Bilbao et al., 2015; Fenaille et al., 2017; Krasny and Huang, 2021). This allows the generation of more accurate and reproducible data (Bilbao et al., 2015). As DIA can analyze a complete mixture of peptides, it provides a precise, complete and unbiased map of the nucleolar proteome in the presence or absence of Ub^KEKS. Moreover, DIA uses a peptide library that is generated from a collection of samples, which can be very comprehensive but also very specific to a subset of proteins of interest. Here, several proteins whose nucleolar locations are impacted by Ub^KEKS were identified using such a library that was specifically generated from fractionated nucleoli. Functional enrichment of significantly modulated proteins in the absence of Ub^KEKS pointed toward the activation of apoptosis. The levels of p14arf and IFI16, two proteins involved in the regulation of P53-dependent apoptosis (Kwak et al., 2003; Veneziano et al., 2016; Li et al., 2021), were higher in the nucleoli of Ub^KEKS knockout cells. IFI16 can directly bind P53 and promote phosphorylation on its serine 15 or serine 392, leading to P53 stabilization and apoptosis (Johnstone et al., 2000; Ouchi and Ouchi, 2008; Li et al., 2021). p14arf is an important sensor of different types of cellular stresses, and is thus positively regulated in response to different oncogenic signals (Ozenne et al., 2010). p14arf is normally expressed at low levels because of N-terminal ubiquitination and proteasomal degradation, as well as sequestration within the nucleolus by interaction with NPM (Chen et al., 2010). p14arf can trigger apoptosis by binding and sequestering the ubiquitin ligases Mdm2 and ARF-BP1 in the nucleolus, thus stabilizing P53 in the nucleoplasm (Ozenne et al., 2010; Veneziano et al., 2016; Cilluffo et al., 2020), but also shows tumour-suppressive activities that are independent of p53 through interactions with TIP60, TOPO I and C1QBP (Karayan et al., 2001; Eymin et al., 2006; Itahana and Zhang, 2008). The nucleolar retention of p14arf is thought to prevent these tumour suppressor activities, but these mechanisms leading to the cellular redistribution of p14arf have not yet been elucidated. Overall, this suggests a p53-independent role in the cellular stress response in the absence of Ub^KEKS, since HeLa cells used in this study do not express a functional p53. The stabilization of these proteins in the absence of Ub^KEKS suggests that this modification is involved, either directly or indirectly, in regulating their expression and activation.

To conclude, this work demonstrates the potential role of ubiquitin variant Ub^KEKS in protein trafficking, more specifically between the nucleoli and nucleoplasm. By modifying its protein content, Ub^KEKS is required for maintaining the concentric structure of the nucleolus. Proteins release from or sequestration in nucleoli by Ub^KEKS is shown to have cellular-wide impacts such as apoptosis regulation in a p53-independent manner. This marks the first step for establishing Ub^KEKS as a new layer of regulation in the ubiquitination system. Yet, the expression among different tissues and the large number of targets of this ubiquitin variant remind us that many other potential functions of Ub^KEKS are waiting to be discovered.

HeLa cells which do not express Ub^KEKS were previously generated using the CRISPR/Cas9 method and obtained clones were validated by PCR and direct sequencing of those PCR products (Dubois et al., 2020). In this paper, Ub^KEKS-knockout cells are referred to as “HeLa 2.7” and “HeLa 4.3”, according to the chosen sgRNAs combination for invalidating UBBP4 pseudogene (Fig. 1A).

Wild-type (HeLa-CCL2, ATCC) and Ub^KEKS-knockout HeLa cell lines were cultured as adherent cells in Dulbecco's modified eagle medium (DMEM) which was supplemented with 10% fetal bovine serum and 100 U/ml penicillin/streptomycin.

For experiments using total cell extracts, exponentially growing wild-type and knockout HeLa cells were directly lysed into an 8 M urea/10 mM HEPES pH 8 buffer. Samples were sonicated on ice (Thermo Fisher Scientific Model 120 Sonic Dismembrator) and centrifuge at 16,000 g for 10 min (4°C) to remove cellular debris. In parallel, exponentially growing HEK293, HeLa, U2OS and HCT116 cells were harvested, sonicated and centrifuged following the same process in order to create a general DIA spectral library for human cells. All cleaned total cell extracts were then prepared for mass spectrometry according to the steps described below.

Wild-type and Ub^KEKS-knockout HeLa cells were seeded on glass coverslips in 24-well plates and cultivated under normal conditions until reaching 60% confluency. Cells were fixed with cold 4% paraformaldehyde (PFA) for 20 min and permeabilized with a 5 min incubation in cold 0.15% Triton X-100 in 1xPBS. Incubation in 10% goat serum (GS) in cold 1xPBS was used as a blocking step for 20 min. Coverslips were incubated with primary antibodies in 10% GS in 1xPBS overnight at 4°C. Next morning, cells were washed twice with ice cold 1xPBS before being incubated with secondary antibodies in 10% GS in 1xPBS for 1 h at room temperature. Cells were washed again twice with ice cold 1xPBS and then, incubated with DAPI solution (1 µg/µl) for 8 min in 1xPBS. After two final washes, coverslips were mounted on microscope slides and stored at 4°C in the dark until imaging.

NPM1 (ABclonal #A17983, 1:150) and fibrillarin (monoclonal 38F3, Abcam #ab4566, 1:1000) primary antibodies were used for the overnight incubation at 4°C. Anti-mouse (Invitrogen #A-11003, 1:800) and anti-rabbit (Invitrogen #A-11008, 1:800) secondary antibodies were used for the second incubation on the next morning.

For rescue experiments, cells were transfected with 100 ng HA-Ub^KEKS plasmid construct using JetPrime (Polyplus-transfection SA, France) according to the manufacturer's instructions. 24 h after transfection media was changed to fresh culture media and cells were further grown for 48 h. Cells were then prepared following the general protocol describe above. Primary antibodies against nucleolin (Abcam #ab136649, 1:2000), UBF (monoclonal F-9, Santa Cruz Biotechnology #sc-13125, 1:300) and fibrillarin (monoclonal 38F3, Abcam #ab4566, 1:1000) were used. Anti-mouse (Invitrogen #A11003, 1:800 was used as secondary antibodies.

Images were acquired on a Zeiss LSM 880 confocal microscope using a 40×1.4 NA plan Apo objective with Z-series stacks (step size 0.45 µm). Stacked images were subjected to maximum intensity projection prior analysis. Image analysis was performed using Fiji (version 1.53c) software (Schindelin et al., 2012). Nucleoli circularity was measured with Fiji through object identification using nucleolin immunofluorescence data. For each nucleolus, values on a scale between 0 and 1 were attributed, where 0 corresponds to a line and 1 to a perfect circle. Circularity values were analyzed in Graph Pad Prism version 9.0.0. (GraphPad Software, USA).

Both wild type and Ub^KEKS-knockout HeLa cells were grown in six-well plates to reach 90% confluency. Cells were washed with 1xPBS and first fixed in 1.5% glutaraldehyde-Na cacodylate solution (0.1 M, pH 7.4) for 30 min at room temperature followed by an overnight fixation step at 4°C in 2.5% glutaraldehyde- Na cacodylate solution (0.1 M, pH 7.4). Fixed cells were washed twice in 0.1 M Na cacodylate solution (pH 7.4) followed by a 1 h post fixation incubation in 1% OsO₄ - Na cacodylate solution (0.1 M, pH 7.4). Samples were washed twice in distilled H₂O and stained with 1% uranyl acetate at 4°C overnight in the dark. The next day, samples were washed twice in dH₂O and dehydrated in a sequential manner in 40-50-70-85-95-100% ethanol. Cells were coated with EPON epoxy resin under 25 mbar vacuum twice and polymerized at 60°C for 48 h. The specimens were then detached from the plastic dish and inverted into a new dish for embedding. After immersion in EPON epoxy resin samples were polymerized at 60°C for 48 h. After embedding, thin sections were prepared with ultramicrotome and were mounted on formvar/carbon supported copper grids (400 mesh size). Sections were contrasted with 2% uranyl acetate for 10 min and lead citrate for 5 min. Images were taken with a HITACHI H7500 electron microscope (Hitachi, Japan). All reagents were purchased from Electron Microscopy Sciences (Cedarlane, Hornby, Canada).

RNA was isolated from wild-type HeLa cells and Ub^KEKS-knockout clones 4.3 and 2.7 cells using TRIzol reagent (Invitrogen, USA) and reverse transcription was performed using ProtoScript II reverse transcriptase (New England Biolabs, USA) with oligo dT primers. Quantitative PCR (qPCR) was performed following the manufacturer's instruction using FastStart Essential DNA Green Master (Roche Molecular Systems, Switzerland). Target cDNAs were amplified using gene-specific primer pairs: 45S_ forward (FW): GAACGGTGGTGTGTCGTT, 45S_ reverse (RV): GCGTCTCGTCTCGTCTCACT; 28S_FW: AGAGGTAAACGGGTGGGGTC, 28S_RV: GGGGTCGGGAGGAACGG; 18S_FW: GATGGTAGTCGCCGTGCC, 18S_RV: GCCTGCTGCCTTCCTTGG; MALAT1_FW: GACGGAGGTTGAGATGAAGC, MALAT1_RV: ATTCGGGGCTCTGTAGTCCT (Sen Gupta and Sengupta, 2017) and C_t values were normalized to the expression of housekeeping genes (GAPDH_FW: TGATGACATCAAGAAGGTGGTGAA, GAPDH_RV: TCCTTGGAGGCCATGTGGGCCAT; HPRT_FW: TGTAGCCCTCTGTGTGCTCAAG, HPRT_RV: CCTGTTGACTGGTCATTACAATAGCT) and each gene was represented as 2^−ΔΔCt relative to the HeLa wild-type sample. The reactions were run in Optical 96-well Reaction Plates using a LightCycler 96 Instrument (Roche Molecular Systems, Switzerland) and results were analyzed using LightCycler 96 Software version 1.1.0.1320 (Roche Molecular Systems, Switzerland). Statistical analysis was performed using Graph Pad Prism version 9.0.0. (GraphPad Software, USA).

Nucleoli were isolated from exponentially growing wild-type and KO HeLa cells in 150 mm petri dish as previously described (Li and Lam, 2015). Briefly, cells were directly lysed in 3 ml of pre-chilled (−20°C) 0.5 M sucrose and 3 mM MgCl₂ solution (=solution SI). Lysates were sonicated using the same sonicator described previously and gently layered over 3 ml of a 1 M sucrose and 3 mM MgCl₂ solution (=solution SII). Samples were centrifuged at 1800 g for 10 min at 4°C. Pellets were resuspended into 3 ml of solution SI solution and layered on top of 3 ml of solution SII and centrifuged with the same parameters for an additional wash. Purity of the nucleoli samples was checked via phase-contrast microscopy (Axiovert 200M Zeiss microscope) and the size of isolated nucleoli was measured using Fiji software version 2.1.0/1.53c (Schindelin et al., 2012). Statistical analysis was performed using Graph Pad Prism version 9.0.0. (GraphPad Software, USA). Final pellets (nucleoli) were resuspended in a buffer of 8 M urea, 10 mM HEPES pH 8 and prepared for mass spectrometry.

Once in the 8 M urea, 10 mM HEPES pH 8 buffer, samples were quantified by BCA assay (Thermo Fisher Scientific #23225). DTT (5 mM final concentration) was added into 50 µg of proteins and samples were boiled for 2 min. After a 30-min incubation at room temperature, chloroacetamide (7.5 mM final concentration) was added to the mixture. Solutions were then incubated in the dark, at room temperature for 20 min. 50 mM of NH₄HCO₃ was added to each tube to reduce the final concentration of urea to 2 M. Peptide digestion was performed by adding 1 µg of trypsin (Trypsin Gold, Mass Spectrometry Grade, Promega Corporation, WI, USA) and incubating each sample overnight at 30°C. The samples were then acidified to a final concentration of 0.2% TFA. For samples from isolated nucleoli, a fraction of the digested sample was collected, at this step, to create the DIA spectral library generation and was processed as described in the next paragraph. On the other hand, remaining samples were cleaned using ZipTips C18 column (EMD Millipore, Burlington, VT), lyophilized in speedvac and finally resuspended in formic acid 1%. Peptides were quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific) at a wavelength of 205 nm. The peptides were then transferred into a glass vial (Thermo Fisher Scientific) and keep at −20°C until analysis by mass spectrometry.

Peptides collected for the DIA spectral library were desalted with ZipTips C18 column (EMD Millipore, Burlington, VT, USA), dried in speedvac and resuspended in 300 µl of 0.1% TFA. Those peptides were divided into several fractions using the Pierce High pH Reversed-Phase Peptide Fractionation kit (Thermo Fisher Scientific) according to the manufacturer protocol. Briefly, each column was initially centrifuged at 5000 g for 2 min at room temperature to remove the liquid and pack the resin material, followed by two washes with 300 μl of 100% ACN. The column was then conditioned with two washes of 0.1% TFA. Purified peptides were loaded on the column, centrifuged at 3000 g for 2 min at room temperature, followed by a wash with 300 μl of MS-grade water. Peptides were then eluted in eight fractions by successively loading 300 μl of eight different solutions containing 0.1% triethylamine and 5% up to 50% ACN. For each step, centrifugation at 3000 g for 2 min at room temperature was performed, with a new low-binding microtube to collect the fraction. Peptides were then concentrated by speedvac at 60°C until complete drying and then resuspended in 50 μl of 1% FA buffer. Peptides were assayed using a NanoDrop spectrophotometer (Thermo Fisher Scientific) and absorbance was measured at 205 nm. The peptides were then transferred to a glass vial (Thermo Fisher Scientific) and stored at −20°C until analysis by mass spectrometry.

250 ng of peptides from each fraction were injected into an HPLC (nanoElute, Bruker Daltonics) and loaded onto a trap column with a constant flow of 4 µl/min (Acclaim PepMap100 C18 column, 0.3 mm id×5 mm, Dionex Corporation) then eluted onto an analytical C18 Column (1.9 µm beads size, 75 µm×25 cm, PepSep). Peptides were eluted over a 2-h gradient of ACN (5-37%) in 0.1% FA at 400 nl/min, while being injected into a TimsTOF Pro ion mobility mass spectrometer equipped with a Captive Spray nano electrospray source (Bruker Daltonics). Data was acquired using data-dependent auto-MS/MS with a 100-1700 m/z mass range, with PASEF enabled with a number of PASEF scans set at 10 (1.17 s duty cycle) and a dynamic exclusion of 0.4 min, m/z dependent isolation window and collision energy of 42.0 eV. The target intensity was set to 20,000, with an intensity threshold of 2500.

The raw files were analyzed using MaxQuant (version 2.0.3.0) and the Uniprot human proteome database (version from March 2020 containing 75,776 entries). The settings used for the MaxQuant analysis (with TIMS-DDA type in group-specific parameters) were the following: all of the raw files were assigned with the same cell type name as well as a fraction set from 1 to 8; 1 miscleavage was allowed; fixed modification was carbamidomethylation on cysteine; enzyme was set as Trypsin (K/R not before P); variable modifications included in the analysis were methionine oxidation and protein N-terminal. A mass tolerance of 20 ppm was used for both precursor and fragment ions. Identification values “PSM FDR”, “Protein FDR” and “Site decoy fraction” were set to 0.05. Minimum peptide count was set to 1. Both the “Second peptides” and “Match between runs” options were also allowed. MaxQuant was run with a transfer q value of 0.3. The “peptides.txt”, “evidence.txt” and “msms.txt” files generated from this DDA analysis were subsequently used for DIA spectral library analysis. Library quality was validated by ensuring the location of identified proteins using the COMPARTMENTS resource in the Cytoscape software environment (Binder et al., 2014; Shannon et al., 2003). For the nucleoli library, functional analysis of the 250 proteins with the highest intensity was also performed through the ShinyGO online tool (version 0.76) (Ge et al., 2020).

Parameters used for DIA processing were almost identical as DDA processing, except that data were acquired using diaPASEF mode. Briefly, for each single TIMS (100 ms) in diaPASEF mode, we used one mobility window consisting of 27 mass steps (m/z between 114 to 1414 with a mass width of 50 Da) per cycle (1.27 s duty cycle). These steps cover the diagonal scan line for +2 and +3 charged peptides in the m/z-ion mobility plane.

Raw files were analyzed using MaxQuant (version 2.0.3.0) and the Uniprot human proteome database (version of the 21st of March 2020; 75,776 entries). The settings used for the MaxQuant analysis were almost identical as TIMS DDA analysis, except: TIMS MaxDIA type in group-specific parameters was selected, as well as MaxQuant as library type, followed by uploading the “peptides.txt”, “evidence.txt” and “msms.txt” files generated by MaxQuant from the specific library production. Classic LFQ normalization was also performed for every sample. Following the identification and normalization, proteins positive for at least either one of the “Reverse”, “Only identified by site” or “Potential contaminant” categories were eliminated, as well as proteins identified from a single unique peptide.

Refined tables obtained from MaxQuant were loaded into the ProStaR online interface (version 1.26.1) (Wieczorek et al., 2017) to be further analyzed. Only identified proteins with a unique peptide count equal to or higher than two were considered in all analyses of this article. Normalization step of ProStaR was skipped as it was already performed during the MaxQuant analysis. Partially observed missing values were imputated with the slsa algorithm. For value missing in the entire condition, the Det quantile algorithm (Quantile=1; Factor=0,2) was used to impute values. Differential abundance analysis was performed with a fold change and P-value cut-off adapted to each experiment. A Limma t-test (slim pi0 calibration) was applied to determine significantly downregulated or upregulated proteins in the Ub^KEKS-knockout clones. Volcano plots were drawn in GraphPad, based on the ProStaR differential analysis results table. Venn diagrams and the overlapping significant protein list were obtained using an online tool available at: https://bioinformatics.psb.ugent.be/webtools/Venn/. Interactions networks from significant proteins were generated using the STRING app running in the Cytoscape software environment (Shannon et al., 2003; von Mering et al., 2003). Functional analysis was also performed using the ShinyGO online visualization tool (version 0.76) with parameters adapted for each experiment (Ge et al., 2020).

We ran 25 µg of the total cell extracts or of purified nucleoli extracts on a 4%-20% gradient acrylamide gel (Novex™ WedgeWell™, Tris-Glycine, Invitrogen #XP04200BOX) and transferred on nitrocellulose membrane via semi-dry transfer. Two MS-identified upregulated proteins in HeLa knockout cells’ nucleoli were detected using IFI16 (ABclonal #A2007, 1:400) and p14arf (Abcam #ab216602, 1:1000) antibodies. Nucleolin (Abcam #ab136649, 1:2000) and ß-tubulin (Cell Signaling Technologies #2128S, 1:1000) antibodies were also used as loading control and purity control for isolated nucleoli respectively.

Wild-type and Ub^KEKS-knockout HeLa cells were seeded on glass coverslips in 24-well plates and prepared for confocal microscopy as described in the previous paragraph. IFI16 (ABclonal #A2007, 1:500) and p14arf (Abcam #ab216602, 1:500) antibodies were used to confirm MS results and complement immunoblotting observations. Anti-rabbit secondary antibodies (Invitrogen #A-11008, 1:800) were used for both IFI16 and P14arf primary antibodies. Image analysis was performed using Fiji (version 1.53c) software (Schindelin et al., 2012).

Exponentially growing wild-type and knockout HeLa cells were cultured in six-well plates until reaching 80-90% confluency. Supernatants were collected and centrifuged to collect dead cells and other cellular debris. Cells were washed twice with PBS 1x and harvested by trypsinization. Cell pellets were resuspended in 100 µl of Annexin V binding buffer (2.5 mM CaCl₂, 140 mM NaCl and 10 mM HEPES). Then, 5 µl of PE conjugated Annexin V (Biolegend #640907) and 10 µl of Propidium Iodide (0.5 mg/ml stock solution) were added. Each sample was incubated for 15 min in the dark prior having their volumes completed up to 500 µl with Annexin V binding buffer. Signals were acquired by flow cytometry (BD Fortessa cytometer, Becton Dickinson) and analyzed with the FlowJo version 10.8.1 (Becton Dickinson & Company). Percentages of cells for each of the populations went under statistical analysis in GraphPad. Significance was determined by a two-way ANOVA and Dunnett's multiple comparisons post hoc test.

Five thousand wild-type and Ub^KEKS-knockout HeLa cells were seeded and cultured as adherent cells in XCELLigence RTCA E-plates (Agilent E-plate VIEW 16 #300601140). Cells were incubated at 37°C under normal conditions. Blank with only medium was performed prior to time course measurements. Cellular impedance was measured every 15 min for 1 week, using a XCELLigence RTCA DP Analyzer. Growth curves were then generated from obtained raw data into Graph Pad Prism version 9.0.0. (GraphPad Software, USA). A two-way ANOVA coupled with Dunnett's multiple comparisons post hoc test was performed to assess the significance of observed differences. At the end of the experiments, cells were fixed in 1% glutaraldehyde for 5 min and stained with 1% crystal violet solution. Pictures were taken using phase-contrast microscopy (Axiovert 200M Zeiss microscope).

X.R. and F.-M.B. are both members of the FRQS-funded “Centre de Recherche du CHUS”.