|Home | About | Journals | Submit | Contact Us | Français|
Pf1, also known as Phf12 (plant homeodomain [PHD] zinc finger protein 12), is a member of the PHD zinc finger family of proteins. Pf1 associates with a chromatin-interacting protein complex comprised of MRG15, Sin3B, and histone deacetylase 1 (HDAC1) that functions as a transcriptional modulator. The biological function of Pf1 remains largely elusive. We undertook the generation of Pf1 knockout mice to elucidate its physiological role. We demonstrate that Pf1 is required for mid- to late gestation viability. Pf1 inactivation impairs the proliferative potential of mouse embryonic fibroblasts (MEFs) and is associated with a significant decrease in bromodeoxyuridine incorporation; an increase in senescence-associated β-galactosidase (SA-β-Gal) activity, a marker of cellular senescence; and elevated levels of phosphorylated H2AX (γ-H2A.X), a marker associated with DNA double-strand breaks. Analysis of transcripts differentially expressed in wild-type and Pf1-deficient cells revealed the impact of Pf1 in multiple regulatory arms of the ribosome biogenesis pathways. Strikingly, assessment of the morphology of the nucleoli exposed an abnormal nucleolar structure in Pf1-deficient cells. Finally, proteomic analysis of the Pf1-interacting complexes highlighted proteins involved in ribosome biogenesis. Taken together, our data reveal an unsuspected function for the Pf1-associated chromatin complex in the ribosomal biogenesis and senescence pathways.
Pf1, also known as Phf12 (plant homeodomain [PHD] zinc finger protein 12), is a member of the PHD zinc finger family of proteins. PHD domains are small, 50- to 80-amino-acid-long domains often found in clusters of two or three and/or in the proximity of other chromatin-interacting domains, such as bromo- or chromodomains. Consistently, many of the PHD-containing proteins are nuclear proteins that interact with chromatin. Increasing evidence suggests that PHD domains are capable of recognizing modified and unmodified histone tails and that PHD domain-containing proteins act as epigenetic readers (1).
The Pf1 gene is conserved throughout evolution, and the Pf1 protein, like its Saccharomyces cerevisiae yeast homolog, Rco1, contains two PHD domains in its N terminus. Mammalian Pf1 was first identified in a yeast two-hybrid screen for proteins interacting with the paired amphipathic helix 2 (PAH2) domain of Sin3A and shown to function as a transcriptional repressor (2). A later report identified Pf1 to be one of the components of the human MRG15 complex, together with Sin3B but not together with its close homolog, Sin3A (3). The PHD domains of Pf1 are important for the interaction with the MRG domain of MRG15, which relies mainly on hydrophobic interactions (3,–5). The interaction of Pf1 with a complex containing Sin3B but not Sin3A was also confirmed in experiments identifying associations between Sin3B, histone deacetylase 1 (HDAC1), MRG15, and Pf1. While its precise function in transcriptional regulation remains unclear, the Pf1-MRG15-Sin3B-HDAC complex appears to modulate RNA polymerase II (RNAPII) progression at actively transcribed regions, similar to what has been reported for the Rpd3S complex, its yeast homolog (6).
Reports assessing the biological function of Pf1 point to roles in phosphoinositide signaling (7), nervous system development (8), and the epithelial-to-mesenchymal transition and maintenance of a stem cell phenotype (9). To our knowledge, no comprehensive study elucidating the in vivo function of the Pf1 protein has been conducted so far.
Mouse strains with genetic inactivation of Sin3B, HDAC1, or MRG15, the three components of the protein complex repeatedly shown to also contain the Pf1 protein, have previously been described. Sin3B−/− embryos reveal a global retardation in size, a pale color, an abnormal hematopoietic picture in the liver, and a reduction in bone deposition at later embryonic stages and die shortly before birth (10). Sin3B knockout mouse embryonic fibroblasts (MEFs) are refractory to quiescence and to oncogene-induced senescence (OIS), a stable cell cycle arrest in response to oncogene activation (10,–13). Knockout of HDAC1 leads to embryonic lethality before day 10.5 of gestation, and HDAC1−/− embryos examined at earlier embryonic stages are extremely growth retarded as a result of a defect in cell proliferation (14, 15). The loss of HDAC1 in embryonic stem (ES) cells leads to impaired proliferation that is associated with upregulated expression of cyclin-dependent kinase inhibitors p21 and p27 (14), but conditional HDAC1 knockout in MEFs does not affect cell proliferation under normal conditions (16). Similarly, genetic inactivation of MRG15 results in embryonic lethality between embryonic day 14.5 (E14.5) and birth. The null embryos are smaller than the wild-type (wt) and heterozygous embryos, with the most prominent defects being seen in the heart, lung, liver, and keratinocyte-epidermal layer of the skin. The growth potential of Mrg15−/− MEFs is greatly reduced (17).
Here, we describe the generation of Pf1 knockout mice to define the physiological role of Pf1. We report that Pf1 inactivation leads to embryonic lethality. Moreover, we show that Pf1 inactivation impairs the proliferative potential of the cells and is associated with a strong nucleolar stress. Taken together, our data demonstrate that Pf1 is crucial for embryogenesis and nuclear functions.
To investigate the biological functions of Pf1 in a physiologically relevant system, we generated a mouse strain harboring a genetic inactivation of the Pf1 locus. Specifically, we identified a mouse ES cell clone harboring a retroviral insertion in the 5′ region of Pf1, which results in the disruption of a splice acceptor site (Fig. 1A). We then used this ES cell clone to derive a Pf1+/− mouse strain. Germ line transmission of the mutant Pf1 allele was verified by PCR, and heterozygous mice were intercrossed to obtain Pf1−/− embryos (Fig. 1B). Genetic inactivation of Pf1 in the Pf1−/− embryos was verified by quantitative reverse transcriptase PCR (qRT-PCR) (Fig. 1C), and embryonic development was followed. While Pf1 heterozygous mice developed normally, no live Pf1−/− mouse was born. Temporal analysis of embryonic development revealed that Pf1−/− embryos die at mid- to late gestation with diverse developmental defects, including edema and internal hemorrhage (Fig. 1D and andE).E). Moreover, macroscopic analysis of the Pf1−/− embryos indicated a global growth retardation, where E14.5 Pf1−/− embryos were comparable in size to E11.5 wild-type embryos (Fig. 1F and data not shown). Pf1−/− embryos also presented impaired development of the skeleton, of the associated skeletal muscle, and of the brain (Fig. 1F). To determine if the global growth retardation observed at stage E14.5 results from a decrease in cell proliferation or, alternatively, from an increase in cell apoptosis, we performed immunohistological staining on the whole embryo at stage E14.5 for Ki67, a cell proliferation marker, and caspase-3 (Casp3), an effector caspase and a marker of the cell commitment to apoptosis. No difference in the levels of Ki67 was observed at this stage (Fig. 1G). In contrast, a localized increase in Casp3 levels was observed in the livers of Pf1−/− embryos (Fig. 1H). Overall, these observations indicate that Pf1 is essential for normal embryonic development and Pf1 protects from aberrant apoptosis in a tissue-specific manner.
We next generated mouse embryonic fibroblasts (MEFs) from E14.5 Pf1+/+, Pf1+/−, and Pf1−/− embryos to assess the cellular function of Pf1. Following a standard 3T3 protocol, we observed no significant differences in the ability of Pf1+/+ and Pf1+/− MEFs to proliferate (Fig. 2A). In contrast, Pf1−/− MEFs lost their proliferative capacity as early as during the first cell passage. This result suggests that Pf1 functions to maintain a cellular proliferative state under culture conditions. The decreased proliferative potential of Pf1−/− MEFs was associated with a significant decrease in bromodeoxyuridine (BrdU) incorporation and an increase in senescence-associated β-galactosidase (SA-β-Gal) activity, a marker of cellular senescence (Fig. 2B and andC,C, respectively). Cellular senescence can be triggered by the accumulation of irreparable DNA damage (18). Thus, we assessed the levels of phosphorylated H2AX (γ-H2A.X), a marker associated with DNA double-strand breaks, in Pf1−/− and Pf1+/+ MEFs. Consistent with the increased senescence detected in Pf1-null cells, Pf1−/− cells exhibited a strong activation of the DNA damage response, as visualized by an increase in the γ-H2AX signal compared to that for control cells (Fig. 2D). Thus, our results suggest that Pf1 is essential to prevent premature cell cycle exit and entry into cellular senescence.
To identify the molecular mechanisms underlying the defects elicited by genetic inactivation of Pf1, we next examined the impact of Pf1 inactivation on the MEF transcriptome. RNA extracted from wild-type and Pf1-null MEFs was profiled by hybridization to a whole-genome microarray. As shown in Fig. 3A, Pf1 depletion led to specific changes in gene expression, with 149 genes being upregulated and 323 genes being downregulated by a 2-fold change (P < 0.05) in Pf1−/− MEFs. Gene ontology (GO) analysis by analysis of the transcripts differentially expressed in wild-type and Pf1-deficient cells with the Database for Annotation, Visualization, and Integrated Discovery (DAVID) revealed the impact of Pf1 in multiple regulatory arms of the RNA biogenesis and maturation pathways (Fig. 3B). Specifically, genetic inactivation of Pf1 resulted in a strong and significant increase in the abundance of a large number of transcripts involved in ribosome biogenesis and rRNA processing (Fig. 3B and andC).C). This finding was further validated by qRT-PCR for a subset of genes, including the genes for Dimt1 (dimethyladenosine transferase 1 homolog, an rRNA methyltransferase), Pes1 (pescadillo ribosomal biogenesis factor 1, a component of PeBoW complex with roles in pre-rRNA processing and 60S ribosomal subunit maturation), and Pa2G4 (proliferation-associated 2G4, involved in ribosome assembly and the regulation of intermediate and late steps of rRNA processing) (Fig. 3D). Thus, our results suggest that Pf1 plays an important role in regulating the levels of transcripts involved in ribosomal biosynthesis and rRNA processing.
The nucleolus is the site of rRNA synthesis and nascent ribosome assembly (19). Previous studies pointed to a link between nucleolar stress and cellular senescence (20, 21), although the molecular bases for this link remain elusive. To confirm our observations that Pf1 deletion affects expression of the rRNA processing machinery, we analyzed by immunofluorescence the levels of fibrillarin, SENP3, and UBF, three integral components of the nucleolus that participate in rRNA processing. Fibrillarin is an rRNA 2′-O-methyltransferase involved in the early stages of rRNA processing (22,–25). SENP3 is a SUMO-isopeptidase involved in the conversion of the 32S rRNA intermediate to mature 28S rRNA in mammals (26,–28). SENP3 also tightly associates with the nucleophosmin NPM1, another central factor in ribosome biogenesis (29, 30). Fibrillarin and SENP3 exert their function and are present in the outer two layers of the nucleoli: the dense fibrillar component and the granular component. UBF is essential for RNA polymerase I (Pol I) preinitiation complex binding to enhancer regions of ribosomal DNA (rDNA) sequences; it exerts its functions in the fibrillar center of the nucleolus and also contributes to rRNA processing (22). As shown in Fig. 4, genetic inactivation of Pf1 alters the shape and size of the nucleolus. Indeed, the nucleolus of the wild-type Pf1 MEFs visualized by fibrillarin and SENP3 immunofluorescence appears as multiple small units (Fig. 4A and andB).B). The staining pattern for these markers was strikingly different in Pf1−/− MEFs, with an increase in the fluorescence intensity and a denser distribution in fewer spots being seen within the nucleus, indicating a reorganization of the nucleolus in these cells. However, disruption of the nucleolar structure was not generalized since no difference in the intensity or distribution of the UBF signal was observed between Pf1+/+ and Pf1−/− MEFs (Fig. 4C). This result suggests that Pf1 controls the expression of specific components of the rRNA processing machinery and affects the outer nucleolar structure.
To identify the molecular basis underlying the impact of Pf1 on nucleolar structure and function, we aimed to identify novel Pf1-interacting proteins. To do so, we generated Flag-tagged constructs of wild-type Pf1 and two Pf1 variants with point mutations unable to interact either with MRG15 or with histone H3 (the Pf1F210A and Pf1D57N mutants, respectively [5, 31]). We then stably expressed the three constructs in HeLa S3 cells, extracted nuclear proteins, immunoprecipitated protein complexes containing the Pf1 construct using anti-Flag beads, and analyzed the protein composition of the precipitated complexes by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Silver-stained electrophoresis gels of the precipitated proteins revealed a discrete band at about 130 kDa, which corresponds to the apparent molecular mass of Pf1, that was present in Pf1, Pf1F210A, and Pf1D57N samples but absent in the control HeLa cell lysate infected with the pBabe empty vector (Fig. 5A). Moreover, a second prominent band at about 38 kDa, which corresponds to the apparent molecular mass of MRG15, was observed only for wild-type Pf1 and the Pf1D57N mutant but not the Pf1F210A mutant, as expected (Fig. 5A). The precipitated proteins were then analyzed by nano-LC-MS/MS to identify the Pf1-associated proteins. As shown in Table 1, this analysis identified most of the proteins previously known to be components of the mammalian homolog of the Rpd3S complex and, to a lesser extent, of the canonical Sin3-HDAC complex. Notably, the enrichment confirmed the interaction between Pf1 and the histone H3 lysine 4 (H3K4)-specific demethylase KDM5A, EMSY, GATAD1, Sin3B, HDAC1, and MRG15 (5, 6). Thus, this result further confirmed a physical basis for the functional involvement of Pf1 in transcriptional regulation and the stable association between Pf1, Sin3B, and a discrete group of proteins previously known to be the components of the mammalian homolog of the Rpd3S complex (3, 6, 32, 33).
Our analysis also points to the requirement of a Pf1 association with MRG15 for Pf1 interaction with some proteins, including components of the canonical Sin3-HDAC complex. Indeed, Table 2 lists the proteins that we identified to be Pf1 interactors dependent upon the Pf1 association with either MRG15 or histone H3 (proteins coimmunoprecipitated with the wt Pf1 construct but not with the mutant Pf1F210A or Pf1D57N construct, respectively). Consistent with its known function in transcriptional regulation, the proteins most abundantly coimmunoprecipitated with wt Pf1 are involved in chromosome and chromatin organization, protein complex biogenesis, transcription, cell cycle control, and nuclear transport (Table 3). Importantly, LC-MS/MS analysis also revealed the association of Pf1 with several proteins not previously identified to interact with Pf1. Table 4 lists some of these newly found possible Pf1 interactors, grouped according to the biological process annotations. Interestingly, LC-MS/MS analysis of Pf1-associated complexes highlighted proteins with roles in ribosome biogenesis, similar to the gene ontology terms associated with transcripts expressed at higher levels in Pf1 knockout MEFs than wt Pf1 MEFs (Fig. 3B). These proteins identified by LC-MS/MS include ribosomal biogenesis protein LAS1L (LAS1L); nucleolar proteins 56 and 58 (NOP56 and NOP58, respectively); proline-, glutamic acid-, and leucine-rich protein 1 (PELP1); nucleophosmin 1 (NPM1); fibrillarin (FBL); testis expressed 10 (TEX10); WD repeat domain 3 and 18 (WDR3 and WDR18, respectively); BMS1 homolog, ribosome assembly protein (BMS1); as well as ribosomal proteins S6, S7, S14, and S16 (RPS6, RPS7, RPS14, and RPS16, respectively) and other proteins (Table 4). A complex comprising PELP1, TEX10, WDR18, and SENP3 has already been described (27). Additionally, the same report demonstrated that PELP1 and LAS1L are SUMOylated by the SENP3 isopeptidase (27). SENP3 is a marker of the outer two layers of the nucleoli (the dense fibrillar and granular components), and we have shown that anti-SENP3 antibody staining of the nucleoli is more intense in Pf1−/− MEFs than in control cells, marking a few large and clearly visible intranuclear structures in Pf1−/− MEFs, as opposed to more diffuse, multiple small puncta in the wt cells (see above and Fig. 4B). Finally, other examples of newly identified Pf1-associated functional protein clusters include components of the anaphase-promoting complex or proteins involved in transport through the nuclear envelope (Table 4).
We thus conducted coimmunoprecipitation (co-IP) experiments using tagged constructs of Pf1 and selected potential novel Pf1 interaction candidates identified in our LC-MS/MS experiment, as well as an additional component of the anaphase-promoting complex, namely, ACP8. Compared to the results obtained with our negative control, SOSS complex subunit B1 (NABP2) (Fig. 5E), all the co-IPs performed showed a clear interaction between Pf1 and the proteins tested (Fig. 5B to toD).D). The interaction with Pf1 was particularly strong for PELP1 (Fig. 5B) and Sec13 (Fig. 5D). Altogether, these results corroborate the findings of our LC-MS/MS analysis and suggest that Pf1 directly contributes to the maintenance of nucleolar functions through its interaction with proteins that participate in ribosomal biogenesis. Along with the demonstration that Pf1 regulates the transcription of genes involved in the rRNA pathway, our results indicate that Pf1 may coordinate transcriptional and posttranscriptional steps of ribosomal biogenesis.
We undertook the generation and characterization of Pf1 knockout mice in order to define the physiological role of Pf1. Pf1 was first identified to be a component of the Sin3-HDAC complex. Specifically, Pf1 appears to interact directly with the Sin3 scaffold protein Sin3A or Sin3B, depending on the experimental and/or cellular context. Biochemical studies have now revealed that Pf1 serves as an integral component of a small HDAC1/2-containing complex that comprises the Sin3B, HDAC1/2, and Mrg15 proteins (2,–5). Pf1, like Sin3B, Mrg15, or HDAC1/2, is essential for embryonic development. However, the cellular phenotypes elicited upon Pf1 inactivation that we report here differed drastically from those resulting from the genetic deletion of Sin3B. Indeed, we have previously demonstrated that Sin3B inactivation in primary mouse embryonic fibroblasts does not impact cellular proliferation under normal culture conditions but impairs cell cycle exit triggered by proquiescence or prosenescence signals (10,–12). In contrast, Pf1-null MEFs are unable to sustain proliferation, and they enter spontaneous premature senescence under normal culture conditions. The discrepancy between Pf1- and Sin3B-null MEF phenotypes may be attributed to the partial redundancy between Sin3B and its close paralog, Sin3A. Indeed, Sin3A inactivation is incompatible with cellular proliferation. While our previous report suggested that the interaction between Sin3B and Pf1 is tighter than the one between Sin3A and Pf1 (6), it is important to note that our proteomic study clearly identified Sin3A to be a Pf1 interactor. Interestingly, the phenotypes elicited upon Mrg15 inactivation are reminiscent of those uncovered in Pf1-null cells. Indeed, Tominaga and colleagues observed that Mrg15−/− embryos present growth retardation and delayed development of many organs and tissues, as well as defective cell proliferation and differentiation (17). Interestingly, Mrg15-null cells adopt senescent phenotypes prematurely and cease to proliferate earlier than their wild-type counterparts, a phenotype that is shared by Pf1-null MEFs. While Mrg15 is an integral part of two distinct chromatin-modifying complexes, the Sin3-Pf1-HDAC complex and the NuA4 complex, the phenotypes elicited upon Mrg15 inactivation are strikingly similar to those that we report here in Pf1-null embryos and cells. Thus, it is tempting to speculate that the resembling phenotypes elicited upon Mrg15 or Pf1 inactivation reflect the functional interaction between the two proteins. In that aspect, it is important to note that Eaf3, the yeast homolog of Mrg15, is dispensable for a functional NuA4 complex (34).
Cellular senescence is a stable cell cycle arrest in response to various cellular stresses (35,–37). Many stimuli trigger senescence in primary cells. This includes oxidative stress, DNA damage, the expression of activated oncogenes (oncogene-induced senescence), or serial passaging, which drives the cell to replicative senescence (38). Thus, senescence is often seen as a mechanism to prevent damaged or mutated cells from proliferating uncontrollably (39). Interestingly, senescent cells often present morphological changes in the nucleolus, showing a single, very prominent nucleolus instead of a few smaller ones (20, 21). Our results indicate that Pf1 controls specific steps of ribosome biogenesis, such as splicing, covalent modifications, and maturation of the pre-rRNA transcripts, and that the changes observed in the nucleoli of Pf1−/− MEFs are unlikely to be an unspecific, secondary consequence of the changes in the rate at which the Pf1−/− cells proliferate. Interestingly, it was reported earlier that premature aging in yeast Werner helicase Sgs1 mutants correlates with alterations of the nucleolar structure, reminiscent of what we observed in Pf1−/− MEFs (40). Given the functional link between senescence and aging in mammals (41), it is tempting to speculate that Pf1 engages a program that coordinates nucleolar integrity and prevention of premature aging. In that aspect, it is intriguing to note the recent demonstration that perturbation of ribosomal biogenesis results in the activation of a senescence program in mammalian cells (42).
It is also worth mentioning that the yeast Rpd3-Sin3 complex has been shown to catalyze histone H4 deacetylation at rDNA chromatin through this mechanism to control RNA polymerase I localization, rDNA transcription, and, as a consequence, the shape and size of the nucleolus in yeast cells (43). The results of experiments conducted in mammalian cells suggest that NoRC, an SNF2h-containing nucleolar chromatin remodeling complex, silences the rDNA locus by targeting the Sin3 corepressor complex to rDNA promoters (44). Given our observation that only the outer layers of the nucleolus are altered in Pf1−/− MEFs, it is unlikely that Pf1 alters ribosomal biogenesis directly through the modulation of rDNA transcription. Indeed, we were unable to detect an alteration in the amount of transcripts corresponding to rRNA in the absence of Pf1 (data not shown). Together, these observations point to a role of Pf1 in posttranscriptional events linked to ribosomal biogenesis.
Our coimmunoprecipitation experiments confirmed the interaction of Pf1 with PELP1, LAS1L, and NPM1. Previous studies showed that inactivation of these proteins can, directly or not, induce senescence, similar to what we observed during Pf1 inactivation. First, NPM1 was recently identified to control a p53-mediated cellular senescence. Using a model of colorectal cancer, Wong and colleagues showed that the suppression of NPM1 activity reduces viability and enhances senescence and cell cycle arrest (45). Second, PELP1 silencing was also shown to promote senescence and inhibit proliferation, colony formation, migration, invasion, and xenograft formation in the same colorectal cancer model (46). Finally, LAS1L inactivation has been showed to result in G1 arrest, linking LAS1L to cell cycle progression (47). Moreover, SENP3, which we showed is more abundant in the Pf1−/− MEFs than in the control cells, interacts with NPM1 (26, 27, 30) and forms a nucleolar complex with PELP1 and LAS1L (47). If the results of our experiments cannot completely rule out the possibility of an indirect role of Pf1 in ribosomal biogenesis, the results of our LC-MS/MS and subsequent co-IP analyses suggest that the effect of Pf1 on nucleolar functions might not be mediated or at least might not be mediated solely by the Pf1 association with the Sin3-MRG15-HDAC complex. It will therefore be particularly important to perform an exhaustive characterization of the different protein complexes in which Pf1 is present, in order to understand how Pf1 executes its function in ribosome biogenesis and to discriminate between this function in ribosome biogenesis from the one that has already been characterized to be a transcriptional regulator. Indeed, we have previously shown that the Pf1 complex modulates RNAPII-driven transcription. In that aspect, it is important to note that recent findings have demonstrated that specific factors can coordinate the transcriptional regulation of ribosomal proteins and rRNA maturation (48). Our results suggest that Pf1 and its associated chromatin-modifying activities may also coordinate several steps in the generation of functional ribosomal subunits.
In conclusion, the results of our study point to a central function for a chromatin-associated protein in the regulation of ribosomal biogenesis and senescence. Whether this property of Pf1 contributes to the alteration of nucleolar functions and the accumulation of senescent cells in physiological settings, such as replicative aging or the response to exogenous stress, remains to be investigated.
Pf1−/− mice were derived using the gene-trapping technique. Briefly, this included the identification of an embryonic stem (ES) cell clone with a ROSA26geo cassette integrated into the intron between exons 2 and 3 of the Pf1 locus, subsequent microinjection of the selected ES cells into blastocysts to generate chimeric mice, and derivation of Pf1+/− mice. The heterozygous animals were intercrossed, and embryos were collected at different time points. All procedures involving the use of animal experimentation were approved by the New York University (NYU) IACUC and Institutional Biosafety Committee (IBC).
Mouse embryos were fixed overnight in 10% formalin (Thermo Fisher Scientific, Waltham, MA) and processed for paraffin embedding. For histology, deparaffinized sections (5 μm) were stained with Gill's hematoxylin (Richard-Allan Scientific) and eosin Y, followed by an alcohol dehydration series and mounting (Permount; Thermo Fisher Scientific, Waltham, MA). Trichrome staining was performed at the NYU School of Medicine Histopathology Core Facility. Paraformaldehyde-fixed, paraffin-embedded, 4-μm sections of tissue were stained using unconjugated polyclonal rabbit anti-mouse cleaved caspase-3 (Asp-175) (catalog number 9661S; lot number 42 RRID:AB_331440; Cell Signaling Technology) and unconjugated rabbit anti-mouse Ki67 clone SP7 (catalog number RM-9106; lot number 1308g RRID:AB_2335745; Lab Vision).
Mouse embryonic fibroblasts (MEFs) were generated from embryos at 13.5 and 14.5 days postcoitus. The head and the red organs were removed and used for genotyping. The torso was then minced and dispersed in 0.1% trypsin (45 min at 37°C). Cells were subsequently grown in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Atlanta Biologicals, Flowery Branch, GA), 50 μM β-mercaptoethanol (Sigma-Aldrich, St. Louis, MO), and penicillin-streptomycin and subcultured 1:4 upon confluence. The cultures were maintained in 6% O2, 5% CO2 at 37°C.
The HeLa S3 cell line was maintained in serum-like modified Eagle medium supplemented with 10% fetal bovine serum, 2 mM glutamine (Corning Inc., Corning, NY), and 1% nonessential amino acids (NEAA; Corning Inc., Corning, NY). The cultures were maintained in spinner flasks at 3 × 105 to 9 × 105 cells/ml in 5% CO2 at 37°C.
293T cells were cultured in DMEM supplemented with 10% donor calf serum (DCS; Atlanta Biologicals, Flowery Branch, GA) and penicillin-streptomycin (Corning Inc., Corning, NY) and subcultured 1:4 upon confluence. The cultures were maintained in 5% CO2 at 37°C. 293T cells were transfected using a standard calcium phosphate transfection protocol. The viral particles produced were collected for three consecutive days as a suspension in cell culture medium. The virus suspension supplemented with 6 μg/ml Polybrene (Millipore-Merck KGaA, Darmstadt, Germany) was filtered through a Millex-HV syringe filter unit (pore size, 0.45 μm; Millipore-Merck KGaA, Darmstadt, Germany) and used to infect the HeLa S3 cells.
Growth curve experiments were performed essentially as described previously (49). Briefly, 25,000 cells per well were plated into 12-well plates. At the times indicated above, cells were washed with phosphate-buffered saline (PBS), fixed in 10% formalin, and rinsed with distilled water. Cells were stained with 0.1% crystal violet (Sigma-Aldrich, St. Louis, MO) for 30 min, rinsed extensively, and dried. Cell-associated dye was extracted with 2.0 ml 10% acetic acid. Aliquots were diluted 1:4 with H2O and transferred to 96-well microtiter plates, and the optical density at 590 nm was determined. Values were normalized to the optical density at day 0 for the appropriate cell type. Within an experiment, each point was determined in triplicate; each growth curve was performed at least twice.
Total RNA was isolated from Pf1+/+ and Pf1−/− MEFs at passage 2 or 3 (Pf1+/+ lines 7706-5 and 7706-6 and Pf1−/− lines 7706-2 and 7706-4) using the TRIzol reagent (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer's protocol. The RNA samples were examined on an Affymetrix GeneChip mouse genome 430A (version 2.0) array (Affymetrix, Santa Clara, CA). Data were analyzed using GenePattern software (50) and Database for Annotation, Visualization, and Integrated Discovery (DAVID) tools (51, 52). cDNA was synthesized using Moloney murine leukemia virus reverse transcriptase (M-MLV RT; Promega, Madison, WI), and quantitative RT-PCR analyses were performed using Maxima SYBR green–carboxy-X-rhodamine 2× quantitative PCR (qPCR) master mix (Thermo Fisher Scientific, Waltham, MA), by following optimized manufacturer protocols.
Subconfluent cultures were labeled for 2 h with 30 μM BrdU (Amersham). Cells were detached with trypsin, fixed in 4% paraformaldehyde, treated with 4 N HCl for 10 min at room temperature and 0.1 M borate (pH 8.5) for 2 min at room temperature (with PBS washes between each step), and, finally, permeabilized with 0.1% Triton X-100–3% bovine serum albumin (BSA)–PBS-Tween 20 (PBS-T) for 5 min. After subsequent washing steps in PBS-T, the cells were then incubated with BrdU antibody (dilution, 1:100 in PBS-T) for 80 min and then with a fluorescein isothiocyanate (FITC)-conjugated secondary antibody (dilution, 1:500 in PBS-T; Calbiochem-Merck KGaA, Darmstadt, Germany). The cells were counterstained with DAPI (4′,6-diamidino-2-phenylindole) to identify all nuclei, and the percentage of BrdU-labeled cells (cells labeled with FITC and DAPI) was quantified using a fluorescence microscope. At least 200 cells were counted per sample; each experiment was performed at least three times.
Cells were washed in PBS, fixed for 3 to 5 min (room temperature) in 3% formaldehyde, washed, and incubated at 37°C with fresh X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) solution, consisting of 1× citric buffer (25 mM Na2HPO4, 7.4 mM citric acid, pH 6.0), 150 mM NaCl, 2 mM MgCl2, 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, and 1 mg/ml X-Gal. Cells were incubated at 37°C overnight. The percentage of SA-β-Gal-positive cells was quantified using a phase-contrast inverted microscope. At least 200 cells were counted per sample; each experiment was performed at least three times.
Cells were grown on coverslips to subconfluence, washed with PBS, and fixed in 3% paraformaldehyde–PBS for 10 min at room temperature. Fixed cells were permeabilized with Triton X-100 buffer (0.1% Triton X-100, 20 mM HEPES-KOH, pH 7.9, 50 mM NaCl, 3 mM MgCl2, 300 mM sucrose) for 5 min and blocked in 0.5% (wt/vol) BSA (Thermo Fisher Scientific, Waltham, MA) in PBS for 30 min at room temperature. The cells were then incubated with primary antibody in 0.5% (wt/vol) BSA in PBS for 2 h at room temperature, followed by three 5-min washes with PBS. Fluorescently labeled secondary antibodies in PBS were then added for 1 h at room temperature, and coverslips were washed 3 times with PBS. The primary antibodies used were anti-γ-H2A.X (catalog no. 05-636; Upstate, Millipore-Merck KGaA, Darmstadt, Germany), antifibrillarin (catalog no. 2639S; Cell Signaling, Danvers, MA), anti-UBF (catalog no. sc-9131; Santa Cruz Biotechnology, Inc., Dallas, TX), and anti-SENP3 (catalog no. 5591S; Cell Signaling, Danvers, MA). The cells were counterstained with TO-PRO3 (Thermo Fisher Scientific, Waltham, MA) by following the manufacturer's protocol and visualized by confocal microscopy.
Transfections and coimmunoprecipitations were performed essentially as described previously (6), using an EZview red anti-Flag M2 affinity gel (Sigma-Aldrich, St. Louis, MO) during the overnight incubation step. Constructs of Pf1 wt and mutant proteins were generated in the David lab. The tagged constructs of APC6, APC7, and APC8 used in the coimmunoprecipitation experiments were obtained from D. Izawa (Cambridge, UK). Plasmids expressing PELP1, LAS1L, and NPM1 were a gift from S. Muller (Frankfurt, Germany, and Cambridge, UK), and plasmids expressing Sec13 and NABP2 were obtained from M. Pagano (New York, NY). Pf1 immunoprecipitation samples were reduced, alkylated, and run into a gel to remove any detergents and other mass spectrometry-incompatible reagents. The gel plugs were excised, digested with trypsin in gel, extracted, and desalted (53). Following desalting, the peptide mixtures were gradient eluted directly into a Thermo Scientific Q Exactive mass spectrometer. Data were searched using the Sequest program within the Proteome Discoverer software suite for peptide and protein identifications.
The raw data from the microarray experiment have been deposited in the GEO database with GEO accession number GSE86398.
We thank the staff of NYU Langone School of Medicine, B. Ueberheide, and the Proteomics Resource Center for assistance with the proteomics experiments, A. Heguy and the Genome Technology Center for assistance with the microarray experiments, and Esthelle Hoedt for help with proteomics data analyses. We also thank Beatriz Fontoura, John Goodier, Daisuke Izawa, Stefan Muller, and Celine Verheggen for the generous gift of reagents.
The work reported here was supported by the National Institutes of Health (R01CA148639 and R21CA155736 to G.D.), the Samuel Waxman Cancer Research Foundation (to G.D.), and a Feinberg NYU individual grant (to G.D.). K.M. was supported by departmental training grant 5-T32CA9161-40. The mass spectrometric experiments were in part supported by Cancer Center support grant P30CA016087 at the Laura and Isaac Perlmutter Cancer Center. The Proteomics Resource Center and the Genome Technology Center are supported in part by NIH/NCI support grant P30CA016087.