Nuclear-to-cytoplasmic Relocalization of the Proliferating Cell Nuclear Antigen (PCNA) during Differentiation Involves a Chromosome Region Maintenance 1 (CRM1)-dependent Export and Is a Prerequisite for PCNA Antiapoptotic Activity in Mature Neutrophils

Background: PCNA is exclusively cytoplasmic in neutrophils and inhibits their apoptosis.

Results: Nuclear export of PCNA is mediated by a CRM1-dependent nuclear export sequence (NES) and is pivotal for its antiapoptotic activity in mature neutrophils.

Conclusion: Cytoplasmic relocalization of PCNA is part of the terminal differentiation of neutrophils.

Significance: This process might be relevant to modulate neutrophil survival in inflammation or neutropenia.

Abstract

Neutrophils are deprived of proliferative capacity and have a tightly controlled lifespan to avoid their persistence at the site of injury. We have recently described that the proliferating cell nuclear antigen (PCNA), a nuclear factor involved in DNA replication and repair of proliferating cells, is a key regulator of neutrophil survival. In neutrophils, PCNA was localized exclusively in the cytoplasm due to its nuclear-to-cytoplasmic relocalization during granulocytic differentiation. We showed here that leptomycin B, an inhibitor of the chromosome region maintenance 1 (CRM1) exportin, inhibited PCNA relocalization during granulocytic differentiation of HL-60 and NB4 promyelocytic cell lines and of human CD34⁺ primary cells. Using enhanced green fluorescent protein fusion constructs, we have demonstrated that PCNA relocalization involved a nuclear export signal (NES) located from Ile-11 to Ile-23 in the PCNA sequence. However, this NES, located at the inner face of the PCNA trimer, was not functional in wild-type PCNA, but instead, was fully active and leptomycin B-sensitive in the monomeric PCNAY114A mutant. To test whether a defect in PCNA cytoplasmic relocalization would affect its antiapoptotic activity in mature neutrophils, a chimeric PCNA fused with the SV40 nuclear localization sequence (NLS) was generated to preclude its cytoplasmic localization. As expected, neutrophil-differentiated PLB985 cells expressing ectopic SV40NLS-PCNA had an increased nuclear PCNA as compared with cells expressing wild-type PCNA. Accordingly, the nuclear PCNA mutant did not show any antiapoptotic activity as compared with wild-type PCNA. Nuclear-to-cytoplasmic relocalization that occurred during myeloid differentiation is essential for PCNA antiapoptotic activity in mature neutrophils and is dependent on the newly identified monomerization-dependent PCNA NES.

Introduction

Neutrophils, the most abundant leukocyte in the circulation, are key actors in the first line of defense against bacterial and fungal infections. Their principal function is to migrate toward inflammation sites where they exert antimicrobial effects through the production and secretion of proteases, reactive oxygen species, antibiotic proteins, and proinflammatory mediators (1, 2). Neutrophils are terminally differentiated cells deprived of any proliferative capacity and have a very short lifespan (1). Recent in vitro and in vivo studies have demonstrated that neutrophil apoptosis and phagocytosis by macrophages is a highly regulated process that is pivotal for the inflammation resolution (3). Hence, controlling the fate of neutrophils obviously has numerous potential therapeutic applications, not only in the field of chronic inflammation, but also in immune modulation (4). Remarkably, neutrophils, which are nonproliferating cells, appear to use unusual pathways implicating cell cycle regulatory proteins such as cyclin-dependent kinases (5) to control their survival/apoptosis balance (6). We have recently reported that neutrophils express high amounts of the proliferating cell nuclear antigen (PCNA)³ (7). PCNA, which belongs to the family of DNA sliding clamps, is ideally suited to function as a moving platform for factors that act concomitantly with replication, but also with other cellular processes, such as DNA repair, cell cycle control, apoptosis, chromatin remodeling, and sister chromatid cohesion (8). PCNA is currently believed to have functions in proliferating cells only. Contrary to the dogma that PCNA functions are nuclear, we have discovered that in neutrophils, PCNA localizes exclusively in the cytoplasm and controls their survival (7). Notably, cytosolic PCNA levels changed with neutrophil survival rate. Moreover, PCNA overexpression rendered neutrophil-differentiated PLB985 cells significantly more resistant to tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)- or gliotoxin-induced apoptosis, and conversely, PCNA siRNA sensitized them to apoptosis (7). These results identify cytoplasmic PCNA as a key regulator of neutrophil lifespan, in a manner independent of the cell cycle, exerting its antiapoptotic activity by associating to procaspases to prevent their activation (7).

One of our salient observations was that PCNA nuclear-to-cytoplasmic relocalization occurred at the end of granulocyte differentiation. For instance, in CD34⁺ cells and in myeloblasts isolated from human bone marrow, PCNA was mainly detectable in the nucleus, whereas mature neutrophils were found to express PCNA exclusively within their cytoplasm. As little is known about PCNA shuttling during granulocytic differentiation, we first examined the molecular basis of PCNA nuclear-to-cytoplasm relocalization in an effort to understand whether the exclusive cytoplasmic localization observed in neutrophils could result from either decreased import to, or increased export from, the nucleus, or both. Concerning import, the PCNA sequence does not contain a classical nuclear localization signal (NLS) recognized by the importin-α–importin-β heterodimer. However, a nonclassical import signal localized to amino acids 101–120 recognized by importin-β only, whose deletion induced PCNA accumulation in the cytoplasm, has recently been identified (9, 10). Moreover, PCNA could use its cognate partners possessing an NLS as cargo proteins for active import. This is the case in p21^Cip1/Waf1, which has its own NLS and whose expression is down-regulated during granulocytic differentiation (11). As no mechanisms governing the active export of PCNA have been described so far, we investigated whether PCNA contains a functional CRM1 (chromosomal region maintenance 1)-dependent nuclear export signal (NES). We then investigated whether modulation of PCNA nuclear-to-cytoplasmic relocalization would directly affect granulocytic differentiation or its antiapoptotic activity in mature neutrophils.

EXPERIMENTAL PROCEDURES

Cell Line Culture and Transfection Procedure

Human neutrophils from healthy donors (Etablissement Français du Sang, Paris, France) were isolated from EDTA-anticoagulated blood, using density gradient centrifugation and Polymorphprep^TM (Nycomed) as described previously (7). Blood donors gave their written informed consent to participate in this study, which was approved by the INSERM Institutional Review Board and the Ethics Committee of Necker-Enfants Malades and Cochin Hospitals (Paris, France). Differentiation of CD34⁺ cells into granulocytes was induced as described previously (7, 12) with some modifications. Briefly, peripheral blood mononuclear cells from G-CSF-treated healthy donors (10 mg/kg for 5 days to induce hematopoietic stem cell mobilization) were collected after cytapheresis (Hematology Department, Necker Hospital, Paris, France). CD34⁺ cells were next isolated using the CD34 progenitor cell isolation kit (Miltenyi Biotec, Cologne, Germany) and were cultured with stem cell factor (SCF, 100 ng/ml), IL-3 (10 ng/ml), and IL-6 (100 ng/ml) in Iscove's medium Dulbecco's modified (Invitrogen) supplemented with 15% BIT 9500 (BSA-insulin-transferrin, Stem Cell Technologies), antibiotics (penicillin 100 units/ml, streptomycin 100 μg/ml), and l-glutamine 2 mm for 7 days. CD36⁺ cells were next negatively selected by incubation with monoclonal IgG1 anti-CD36 antibody (BD Biosciences) followed by anti-mouse IgG1 antibody coupled to magnetic beads (Miltenyi Biotec). The remaining cells were incubated with G-CSF (10 ng/ml), SCF (100 ng/ml), and IL-3 (10 ng/ml) in the same medium for 13 days to induce granulocytic differentiation. May Grunwald Giemsa staining was used at day 7 and at day 13 after G-CSF/SCF/IL-3) treatment to monitor differentiation (data not shown). When indicated, cells were treated with LMB (10 ng/ml for 6 h00) to inhibit the CRM1-dependent nuclear export.

PLB985, HL60, and NB4 promyelocytic cell lines were cultured at 37 °C in 5% CO₂ in Roswell Park Memorial Institute (RPMI) medium supplemented with 10% fetal calf serum and antibiotics (penicillin 100 units/ml and streptomycin at 100 μg/). NB4 or HL60 cells were induced to differentiate with all-trans-retinoic acid (ATRA) (1 μm, Sigma) for 5 days (13), and granulocyte differentiation was validated by CD11b expression, as described previously (14). PLB985 cell granulocyte differentiation was induced by exposure to 0.5% dimethylformamide (DMF) for 5 days, as described previously (7, 14). PLB985 cells were transfected using the AMAXA® system as indicated by the manufacturer. Briefly, cells (2 × 10⁶) were resuspended in 100 μl of cell line solution in the presence of 1 μg of plasmids. Cells were electroporated at the indicated voltage and transferred to culture plates. Transfected cells were selected on the basis of their resistance to neomycin (1 mg/ml).

HeLa cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum and antibiotics (penicillin at 100 units/ml and streptomycin at 100 μg/ml). To generate transient transfectants, HeLa cells were grown on cover glasses (8 × 10⁴ cells/well) and transfected the next day with 1 μg of DNA using Lipofectamine following the manufacturer's instructions (Invitrogen). When indicated, cells were treated with LMB (Sigma) to inhibit the CRM1-dependent nuclear export.

Western Blot Analysis

PCNA expression in the nucleus and cytosol was assessed in NB4 cells after cell fractionation using nitrogen cavitation to avoid contamination from granule proteases, as originally described (15). Briefly, cells were suspended at 20 × 10⁶ cells/ml in relax buffer (1 mm KCl, 30 mm NaCl, 35 mm MgCl₂, 100 mm PIPES, 1 mm ATP, pH 7.4) supplemented with antiproteinases (100 μm aprotinin, 400 μm leupeptin, 400 μm pepstatin, 4 mm PMSF, 1 mm orthovanadate, 1 mm EDTA) and were introduced into a cell disruption bomb (Parr Instrument Co., Moline, IL) with a nitrogen pressure at 350 p.s.i. 20 min at room temperature. Disruption of the cells was triggered by nitrogen decompression within the cavitation bomb. The lysed sample was collected and centrifuged at 3,000 rpm for 15 min to obtain the post-cavitation supernatant. The pellet containing the unbroken cells and the nuclei was resuspended in a hypotonic buffer (10 mm Hepes, 0.3 mm dithiothreitol, 400 μm leupeptin, 400 μm pepstatin, 4 mm PMSF, 1 mm orthovanadate, 1 mm EDTA, 1 mm EGTA) containing detergent (Nonidet-40 0.3%) and incubated for 15 min on ice to lyse intact cells. Before centrifugation, a microscopic observation with a hemocytometer was necessary to check that there were no intact cells. The suspension of lysed cells was next centrifuged at 13,000 rpm to obtain nuclei within the pellet. Nuclei were resuspended into a lysis buffer (PBS, Triton X-100 1%) to be analyzed by Western blot. The cytosolic fraction was obtained after centrifugation at 100,000 × g of the post-cavitation supernatant. For Western blot analysis, protein concentrations were measured using the BCA assay (Thermo Scientific), and samples were run on a 12.5% SDS-PAGE and transferred to nitrocellulose, as described previously (7). The mouse monoclonal PC10 antibody was used as the primary antibody followed by a horseradish peroxidase-coupled anti-mouse antibody. Western blots were developed using the enhanced chemiluminescence reagent (Thermo Scientific). Lamin and β-actin were used as control proteins for the nuclear and the cytosolic fractions, respectively, to validate the fractionation procedures as described previously (7).

Analysis of Apoptosis

Apoptosis was triggered by incubating PLB985 cells at 37 °C for 16 h with gliotoxin (1 μg/ml, Sigma) (16, 17) as described previously (7, 14). PLB985 cells were cytospun onto glass slides, and May Grunwald Giemsa staining (Hemacolor, Merck, Darmstadt, Germany) was performed to evaluate apoptotic cell morphology. Apoptosis analysis was also evaluated using flow cytometry as described previously (14). Briefly, mitochondrial depolarization was evaluated from the percentage of cells showing decreased fluorescence after DIOC6 labeling, and DNA fragmentation was measured based on the percentage of cells in the subG1 phase after propidium iodide labeling (7).

Plasmid Constructs

The 5′-3′ sequences of the oligonucleotides used to generate pNES1-EGFP were: GACGGTACCGATGATCCTCAAGAAGGTGTTGGAGGCACTCAAGGACCTCATCCGGATCCACCG (forward) and CGGTGGATCCCGGATGAGGTCCTTGAGTGCCTCCAACACCTTCTTGAGGATCATCGGTACCGTC (reverse), and for pNES2-EGFP: GACGGTACCGATGTTGATGGATTTAGATGTTGAACAACTTGGAATTCGGGATCCACCG (forward) and CGGTGGATCCCGAATTCCAAGTTGTTCAACATCTAAATCCATCAACATCGGTACCGTC (reverse). Oligonucleotides were synthesized with KpnI and BamHI ends. After annealing, the oligonucleotides were ligated into a XhoI-KpnI-digested pEGFP-N1 (Clontech) vector using T4 DNA ligase such that the insert was downstream of the N terminus of the EGFP. The same protocol was used to create pEGFP-NES1 and pEGFP-NES2, and the oligonucleotides were ligated into a XhoI-KpnI-digested pEGFP-C3 (Clontech). Combined point mutations within the PCNA sequence were also generated in NES1-EGFP at leucines 12, 16, 19, and 22 into alanine using the QuikChange mutagenesis kit (Stratagene). A mammalian expression plasmid encoding HA-PCNA was prepared by subcloning the full-length human PCNA cDNA into the pcDNA3-HA vector encoding an N-terminal HA epitope upstream of the cloning site (18). A fragment containing a human PCNA cDNA was prepared from bacterial expression plasmid PETPCNA (a gift from Dr. Bruce Stillman, Cold Spring Harbor, NY) by double digestion using NdeI and BamHI followed by treatment with the Klenow fragment. The human PCNA cDNA was inserted either into pcDNA3-HA or into pcDNA3.1 at the XbaI restriction site, also followed by treatment with the Klenow fragment. The oligonucleotides used for single mutations, combined point mutations, or deletions of PCNA were also constructed in the HA-PCNA sequence cloned in the pcDNA3 plasmid using the QuikChange mutagenesis kit following the manufacturer's instructions (Stratagene).

The plasmid construct pNLS(SV40)-EGFP-linker-PCNA (kind gift of Prof. Cardoso, Berlin, Germany) has been described previously (19). The subcloning of the EGFP-linker-PCNA cDNA was realized into the pcDNA3.1 mammalian expression plasmid after insertion of the HindIII restriction site in pEGFP-linker-PCNA by mutagenesis with sense (5′-GAAGAAGCGCAAAGCTTTGGTACCGGTCGCCACC-3′) and antisense (5′-GGTGGCGACCGGTACCAAAGCTTTGCGCTTCTTC-3′) oligonucleotides containing the restriction site. The plasmid construct pcDNA3-NLS(SV40)-linker-PCNA had been constructed with pcDNA3-PCNA using the oligonucleotides with the following sequences: 5′GGCCGCGCCATGCCGAAGAAGAAGCGCAAAGTAGGCGAAGGGCAAGGGCAAGGGCAAGGGCCGGGCCGCGGCTACGCGTATCGATCCC3′ and 5′TCGAGGGATCGATACGCGTAGCCGCGGCCCGGCCCTTGCCCTTGCCCTTGCCCTTCGCCTACTTTGCGCTTCTTCTTCGGCATGGCGC3′. The flexible and hydrophilic linker sequence was obtained from the pNLS(SV40)-EGFP-linker-PCNA plasmid (19). The dimerized oligonucleotides were directly ligated with pcDNA3-PCNA digested by NotI and XhoI. The GFP-PCNAΔ12–16 and Δ19–22 deletion mutants were obtained using the QuikChange mutagenesis kit. The monomeric PCNAY114A mutant was generated using a mutagenesis protocol with the following oligonucleotides: 5′-CCAGGAGAAAGTTTCAGACGCTGAAATGAAGTTGATGG-3′, 5′-CCATCAACTTCATTTCAGCGTCTGAAACTTTCTCCTGG-3′; 5′-CCAGGAGAAAGTTTCAGACAAAGAAATGAAGTTGATGG-3′; and 5′-CCATCAACTTCATTTCTTTGTCTGAAACTTTCTCCTGG-3′. All the plasmid sequences were checked by direct sequencing.

Immunofluorescence Analysis of PCNA Localization

PCNA immunofluorescence analysis of HL60, PLB985, and human neutrophils was performed after cytocentrifugation as described previously (7). Briefly, cells were fixed in PBS containing 3.7% formaldehyde (Sigma) for 20 min on ice and permeabilized with Triton X-100 (0.25%) for 5 min at room temperature followed by ice-cold methanol for 10 min, incubated with the rabbit polyclonal anti-PCNA (diluted 1/50, clone Ab5 Calbiochem) for 45 min followed by biotinylated rabbit IgG, diluted 1:100 (Dako Cytomation) for 30 min, and then followed by streptavidin-coupled Alexa Fluor 555, diluted 1:200 (2 μg/ml, Molecular Probes) for 30 min. For HA or CRM1 detection, HeLa cells were fixed as described for PCNA labeling and were incubated with a rabbit polyclonal anti-HA (diluted 1:100, Roche Applied Science) or anti-CRM1 (Santa Cruz Biotechnology) antibody, respectively, followed by an Alexa Fluor 555-coupled goat-anti rabbit IgG (diluted 1:100, Molecular Probes). The analysis of the subcellular localization of the NES1-EGFP and NES2-EGFP and EGFP-NES1 and EGFP-NES2 in HeLa cells was performed 24 h after transfection. The HeLa cells were washed with PBS, fixed with 3.7% paraformaldehyde, and then washed again with PBS. Nuclei were stained with Hoechst 33342 staining (Invitrogen) at 2 μg/ml for 10 min in PBS or with propidium iodide (Sigma) at 300 ng/ml and mounted in Fluoprep mounting medium (Biomerieux). Slides were mounted using Fluoprep medium and analyzed by fluorescence or confocal microscopy with a Leica TCS SP5 AOBS imaging microscope, model X63, with LAS AF version 1.8 software, as indicated. Quantification was carried out using the ImageJ software, version 1.42d (BD Biosciences). Image acquisition and image analysis were performed on the Cochin Institute Imaging Facility (INSERM U1016, Paris, France).

Statistical Analysis

Statistical analysis was performed using the Statview software package (SAS Institute Inc., Cary, NC). Comparisons were made using the Student's t or ANOVA test as indicated.

RESULTS

Evidence for a CRM1-dependent Nuclear Export of PCNA during Granulocytic Differentiation

Using immunofluorescence, we previously reported that PCNA relocalized from the nucleus to the cytosol in ATRA-induced granulocytic differentiation in the promyelocytic NB4 cells (7). We examined whether LMB, an inhibitor of the CRM1 exportin that requires a leucine-rich sequence referred to as an NES (20), could increase nuclear localization of PCNA during ATRA-induced granulocytic differentiation in the promyelocytic NB4 cells. Western blot analysis on cytosolic and nuclear extracts of NB4 cells showed a prominent cytosolic localization of PCNA in ATRA-differentiated NB4 cells with a low amount of nuclear PCNA, as described previously (7) (Fig. 1A). Notably, LMB treatment affected this PCNA intracellular distribution by decreasing cytosolic and increasing nuclear PCNA in LMB-treated NB4 cells as compared with the controls (Fig. 1A). The effect of LMB was also observed during ATRA-induced HL60 cell differentiation. PCNA immunolabeling showed again that PCNA relocalized from nucleus to cytosol during ATRA-induced differentiation (Fig. 1B, middle row) and that nuclear staining was significantly enhanced in LMB-treated ATRA-differentiated HL60 cells as compared with untreated cells (Fig. 1B, lower row). This was confirmed by nuclear fluorescence quantification (Fig. 1C). We previously reported that redistribution of PCNA from the nucleus to the cytoplasm occurred during granulocytic differentiation of human CD34⁺ primary cells (7). We confirmed here that 7 days after starting granulocytic differentiation with the combination of G-CSF/IL3/SCF treatment, PCNA exhibited a mixed cytoplasmic and nuclear distribution in some cells but was not exclusively cytosolic (Fig. 1D, upper panel), whereas on day 13, most cells had multilobular nuclei with PCNA located in the cytoplasm (Fig. 1D, lower panel). Accordingly, LMB treatment did not result in nuclear accumulation at day 7 because there was no functional nuclear export, as evidenced by quantification of nuclear fluorescence (Fig. 1E, upper panel). In contrast, at day 13, LMB significantly inhibited cytosolic localization (Fig. 1E, lower panel), thereby demonstrating that PCNA nuclear exclusion was involved in cytosolic localization of PCNA during granulocytic differentiation. The hypothesis of an active export in mature neutrophils was next tested. Notably, LMB treatment of neutrophils did not induce nuclear accumulation of PCNA (Fig. 2A), thus strongly suggesting that there was no active PCNA export. In contrast, cIAP1 nuclear accumulation was observed in LMB-treated mature neutrophils (Fig. 2B), thus providing a positive control for an active nuclear export (21). Taken together, these results strongly suggested the presence of an active PCNA nuclear export that occurred only during differentiation.

FIGURE 1.

PCNA nuclear-to-cytoplasmic relocalization during granulocytic differentiation. A, effect of LMB on PCNA in nucleus and cytosol in ATRA-differentiated NB4 cells. Four days after ATRA treatment, NB4 cells were treated with solvent (−) or with 20 ng/ml LMB (+) for 4 h, and PCNA protein was analyzed by Western blot in nuclear and cytosolic fractions (10 and 20 μg/well, respectively). β-Actin and lamin B were used as specific marker for cytosol and nuclei and as loading controls, respectively. This representative experiment has been performed three times with identical results. B, effect of LMB on PCNA relocalization in HL60 cells. Four days after ATRA treatment, HL60 cells were treated with solvent (−) or with 20 ng/ml LMB (+) for 4 h, and PCNA localization was analyzed with confocal microscopy after immunolabeling using the Ab5 rabbit polyclonal anti-PCNA antibody. The nuclei were visualized with Hoechst staining (original magnification ×630, scale bar = 20 μm). This representative experiment has been performed four times with identical results. C, quantification of nuclear fluorescence intensity with the ImageJ 1.42 software in HL60 cells after LMB treatment. Quantification of nuclear PCNA is expressed in arbitrary units, and the means ± S.E. of four independent experiments in which more than 100 cells in each condition have been counted are given (**, p < 0.01, Student's t test). D, effect of LMB on PCNA relocalization during granulocytic differentiation of human CD34⁺ cells. CD34⁺/CD36⁻ cells were treated with LMB (10 ng/ml or with solvent for 6 h) at day 7 (upper panel) or at day 13 (lower panel) after starting granulocytic differentiation with G-CSF/SCF/IL-3. PCNA localization was analyzed with confocal microscopy as in B. The nuclei were visualized with Hoechst staining (original magnification ×630, scale bar = 20 μm). This representative experiment has been performed three times with identical results. E, quantification of nuclear fluorescence intensity in CD34⁺/CD36⁻ cells at day 7 (upper panel) or at day 13 (lower panel) in the presence or absence of LMB as in C. Data are expressed in arbitrary units, and the means ± S.E. of three independent experiments are given (**, p < 0.01, Student's t test).

FIGURE 2.

Effect of LMB on PCNA and cIAP1 localization in mature neutrophils. Neutrophils were treated with solvent (−) or with 10 ng/ml LMB (+) for 5 h, and PCNA (A) and cIAP1 (B) localization was analyzed with confocal microscopy after immunolabeling. The nuclei were visualized with Hoechst staining (original magnification ×630, scale bar = 20 μm).

Identification of a Nuclear Export Signal in the PCNA Sequence

In the search for a putative NES, PCNA sequence analysis revealed two leucine-rich sequences, which constitutes a signature of conventional NES (22, 23) from Ile-11 to Ile-23 and from Leu-121 to Ile-128. They were named NES1 and NES2, respectively. Both sequences presented similarities with other functional NES identified in the cAMP-dependent protein kinase inhibitor (24), HIV-1 Rev (25), p53 (26, 27), BRCA (28), and CDC25B (18, 29) (Table 1), and were found highly conserved in PCNA sequences among vertebrates (Table 2). NES1 and NES2 sequences were fused to the amino terminus of the enhanced green fluorescent protein (EGFP) sequence using the pEGFP-N1 vector (30), leading to NES1-EGFP or NES2-EGFP protein that was expressed in transiently transfected HeLa cells. In the absence of the NES1 or NES2 sequence, the EGFP protein was equally distributed between the cytoplasm and the nucleus of the EGFP-N1-transfected cells (Fig. 3A, left upper panel). In contrast, the NES1-EGFP was exported from the nucleus and was almost entirely cytoplasmic (Fig. 3A, middle upper panel), whereas NES2-EGFP presented the same repartition as the EGFP protein alone (Fig. 3A, right upper panel). Likewise, the NES1 fused to the carboxyl-terminal part of the EGFP using the pEGFP-C3 vector, leading to the EGFP-NES1 protein was, again, mostly cytoplasmic (Fig. 3A, middle lower panel), and EGFP-NES2 was pancellular (Fig. 3A, right lower panel), as was the control (Fig. 3A, left lower panel). These results demonstrated that the NES1 sequence was active when fused to either the amino-terminal or the carboxyl-terminal part of the EGFP and was able to export EGFP from the nucleus to the cytoplasm. NES1-EGFP-transfected HeLa cells were next treated with LMB to prevent the nuclear exit of NES1-EGFP protein. In the absence of LMB, NES1-EGFP protein was cytoplasmic in 75% ± 3.4 of the cells, whereas after treatment, as expected, NES1-EGFP protein was retained in the nucleus in 53% ± 4.63 of the cells (Fig. 3, B and C). To identify the specific residues of NES1 involved in the active export, leucines 12, 16, 19, and 22 were mutated into alanine in the NES1-EGFP protein. EGFP fused with the mutated NES1 (MutNES1) showed prominent nuclear localization (Fig. 3, D and E), confirming that leucine residues were essential for NES1 functionality. We conclude that NES1 from the PCNA sequence that is conserved among species (Table 2) can mediate EGFP nuclear export in a CRM1-dependent manner.

TABLE 1.

Alignment and comparison of the putative PCNA NES with the consensus sequence for NES and with known NES

The sequences surrounding leucines 12, 16, 19, and 22 of NES1 and leucine 121, 126, valine 123, and isoleucine 128 of NES2 were compared with the consensus sequence for NES and the NES sequence of other tumor suppressors and the Rev protein and the protein kinase inhibitor (PKI) protein. Bold letters represent residues in the core region (LXL).

Consensus	FX(2,3) FX(2,3)-FXF (F = LIVFM)
NES1 PCNA(11–23)	ILKKVLEALKDLIN
NES 2 PCNA(121–128)	LDVEQLGI
HIV-1 Rev(75–84) (25)	LQLPPLERLTL
PKI(36–46) (24)	ELALKLAGLDI
p53(11–27) (27)	EPPLSQETFSDLWKLLP
p53(340–351) (26)	MFRELNEALEKLK
BRCA1(81–99) (28)	QLVEELLKIICAFQLDTGL
CDC25B(29–40) (18)	LPGLLLGSHGLL
CDC25B(52–65) (29)	VTTLTQTMHDLAGL

TABLE 2.

Alignment of the putative PCNA human NES with the PCNA sequence of vertebrate organisms

Human	ILKKVLEALKDLI	LDVEQLGI
Chimpanzee	ILKKVLEALKDLI	LDVEQLGI
Xenopus laevis	ILKKVLEALKDLI	LDVEQLGI
Mus musculus	ILKKVLEALKDLI	LDVEQLGI

FIGURE 3.

Localization and functional characterization of PCNA NES in HeLa cells. A, effect of NES1 and NES2 fusion on EGFP localization using either EGFP-N1 (upper panels) or EGFP-C3 (lower panels) plasmids, allowing us to fuse NES sequence at the amino- or carboxyl-terminal part of EGFP, respectively. HeLa cells were transfected with pEGFP-N1, pEGFP-C3, pNES1-EGFP, pEGFP-NES1, pNES2-EFGP, or pEGFP-NES2, and EGFP was detected by confocal scanning microscopy (original magnification ×630, scale bar = 20 μm). B, effect of LMB on NES1-EGFP localization. HeLa cells transfected with pNES1-EGFP were treated with either ethanol solvent (−) or 20 ng/ml LMB for 6 h prior to fixation and permeabilization. HeLa cells transfected with EGFP-N1 were used as controls. Nuclei were stained with propidium iodide (original magnification ×400, scale bar = 20 μm, left panel). C, the percentage of cells showing either strong nuclear localization (white bars, N>C) or strong cytoplasmic localization (black bars, N<C) was determined by direct counting. Counting was performed on a minimum of 100 transfected cells per experiment. The data are the means ± S.E. of three independent experiments, **, p <0.001, ANOVA test. D, effect of leucine into alanine mutations within NES1 on the NES1-EGFP localization. HeLa cells were transfected with pEGFP-N1, pNES1-EGFP, or pMutNES1-EGFP, and EGFP was detected with confocal scanning microscopy. Nuclei were visualized with Hoechst stain (original magnification ×400, scale bar = 20 μm, left panel). E, the percentages of cells showing nuclear and cytoplasmic EGFP localization were determined as in C. The data are the means ± S.E. of three independent experiments, **, p <0.001, ANOVA test.

The Export of Monomeric but Not Trimeric PCNA Is Dependent on NES1

To analyze whether NES1 could mediate the active export of the full-length PCNA, point mutations (L12A/L16A; L19A/L22A; L12A/L16A, and L19A/L22A) of leucine into alanine or deletions (Δ12–16; Δ19–22; Δ12–16, and Δ19–22) within NES1 were introduced into pcDNA3-HA-PCNA. HeLa cells were then transiently transfected with these vectors, and HA-PCNA localization was assessed by indirect immunofluorescence using an anti-HA mouse monoclonal antibody (Fig. 4A). Unexpectedly, a pancellular distribution of all the NES1 mutants was observed, in contrast to the wild-type HA-PCNA (Fig. 4B). These results were the exact opposite to what we had predicted because in the absence of a functional NES, PCNA should be retained within the nucleus.

FIGURE 4.

Effect of NES1 mutation on PCNA subcellular localization. A, analysis of PCNA localization by indirect immunofluorescence using mouse monoclonal anti-HA antibody in HeLa cells transfected with pcDNA3-HA-PCNA constructs containing different mutations or deletions within the PCNA NES1. The cells were visualized using epifluorescence microscopy (original magnification ×400, scale bar = 20 μm). B, the percentages of cells showing strong nuclear PCNA localization (white bars, N>C) and pancellular localization (gray bars, N=C) are plotted in histograms. Counting was performed on a minimum of 100 transfected cells per experiment in three independent experiments. The data are the means ± S.E. of the three independent experiments, ***, p <0.001 (ANOVA test).

Molecular modeling analysis showed that NES1 was located within the internal face of the PCNA trimer (Fig. 5A, left panel) and within the DNA-interacting domain of PCNA (31), thus strongly suggesting that NES1-deleted PCNA mutant would not be retained within nucleus. In addition, due to its localization within the trimer, NES1 could not bind CRM1 in trimeric PCNA (Fig. 5A). It has previously been reported that PCNAY114A has an inability to trimerize due to a local alteration at the monomer surface (32) (Fig. 5A, right panel). To test the hypothesis that NES1 could mediate nuclear export in monomeric PCNA, the monomeric PCNA mutant (PCNAY114A) was generated by mutating the conserved tyrosine 114 into alanine using the pcDNA3GFP-PCNA vector, which allows fluorescent PCNA to be observed directly. HeLa cells were then transfected with pcDNA3GFP-PCNA or pcDNA3GFP-PCNAY114A. As predicted, pGFP-PCNAY114A had a pancellular distribution, in contrast to wild-type GFP-PCNA, whose localization was restricted to the nucleus (Fig. 5B). Pertinently, NES1 deletion (Δ12–16 and Δ19–22) in GFP-PCNAY114A abrogated PCNA nuclear export, thus confirming that NES1 was essential for monomeric PCNA export. The same results were obtained with an HA-tagged form of PCNA or PCNAY114A (data not shown). These results were in contrast to the pancellular distribution of NES1-deleted-GFP-PCNA (Fig. 5B, lower row) already observed with NES1-deleted HA-PCNA (Fig. 4), confirming that NES1 was not functional in the trimeric conformation of PCNA. Accordingly, LMB treatment of HeLa cells expressing the PCNAY114A mutant relocalized the cytoplasmic monomeric PCNA to the nucleus as evidenced by a strong nuclear florescence, strongly suggesting that its nuclear export was mediated via the CRM1 exportin (Fig. 6A). Moreover, immunolabeling of CRM1 in HeLa cells expressing either wild-type EGFP-PCNA or EGFP-PCNAY114A provided evidence that CRM1 colocalized only with the monomeric PCNA. In fact, as shown on Fig. 6B, wild-type PCNA and CRM1 are both prominently located within the nucleus, but the merged fluorescence did not result in a total colocalization. It should be noted that some yellow dots can be observed because wild-type PCNA can be either trimeric or monomeric. In contrast, the PCNAY114A that is exclusively monomeric showed a pancellular localization as described (Fig. 6B) but also showed a significant colocalization with CRM1 as evidenced by the yellow fluorescence within nucleus in the merged pictures, strongly suggesting that PCNA and CRM1 are associated. Taken together, it can be concluded that NES1 was fully active in monomeric PCNA, thus revealing that PCNA cytoplasmic relocalization might be regulated by its monomerization.

FIGURE 5.

Nuclear localization of NES-deleted monomeric PCNA. A, tridimensional structure of trimeric and monomeric PCNA. Left panel: NES1 localization within the PCNA tridimensional structure (Protein Data Bank (PDB) ID: 1VYM). The interdomain connecting loop of PCNA is shown in green. The NES1 sequence is localized in the inner face of the PCNA trimer (cyan central helix). The protein is represented with graphics, and the NES1 atoms (Ile-11–Asn-23) are represented with balls colored following their type (nitrogen, blue; oxygen, red; carbon, cyan). Right panel: the Y114A mutation results in PCNA monomerization. The close-up of the mutated amino acids (orange balls) of the NES1 shows the following residues: Leu-12, Leu-16, Leu-19, and Leu-22. Both figures have been generated with PyMOL (42). B, analysis of GFP-PCNA localization by confocal scanning microscopy in HeLa cells transfected with pcDNA3-GFP-wtPCNA (wild-type PCNA), pcDNA3-GFP-PCNAY114A, pcDNA3-GFP-PCNAY114AΔ12–16NES1, or pcDNA3-GFP-PCNAΔ12–16NES1). Nuclei were visualized by Hoechst staining (original magnification ×630, scale bar = 20 μm).

FIGURE 6.

NES deletion in monomeric PCNA NES is functional in monomeric PCNA. HeLa cells were transfected with pcDNA3-GFP-wtPCNA (WT PCNA) or pcDNA3-GFP-PCNAY114A (PCNAY114A). A, effect of LMB on the monomeric PCNA mutant PCNAY114A. HeLa cells expressing PCNAY114A were treated with either ethanol solvent (−) or 20 ng/ml LMB for 6 h prior to fixation and permeabilization, and cells were analyzed by scanning confocal microscopy. HeLa cells transfected with pcDNA3-GFP-wtPCNA and treated with solvent were used as controls. B, colocalization of CRM1 with PCNAY114A but not with WT PCNA in HeLa cells. Immunofluorescence detection of CRM1 was performed using a rabbit polyclonal anti-CRM1 antibody followed by an Alexa Fluor 555-coupled goat anti-rabbit IgG, and cells were analyzed by scanning confocal microscopy. A and B show representative experiments that have been performed at least three times with similar results. The nuclei were visualized with Hoechst staining (original magnification ×630, scale bar = 20 μm).

The Nuclear-to-cytoplasmic Relocalization during Granulocyte Differentiation Is Pivotal for PCNA Antiapoptotic Activity in Mature Neutrophils

As described previously (7), the DMF-differentiated PLB985 cells considered to be “neutrophil-like cells” were used to study the impact of a modulation of PCNA localization on its antiapoptotic activity. PLB985 cells were stably transfected with a nuclear PCNA mutant referred to as SV40NLS-PCNA; indeed, as described previously, the fusion of PCNA with the SV40NLS resulted in its exclusive nuclear localization (19). As expected, overexpression of the SV40NLS-PCNA mutant resulted in PCNA nuclear localization in undifferentiated PLB985 cells, as evidenced by immunofluorescence after PCNA labeling (Fig. 7A). Likewise, an altered nuclear-to-cytoplasmic relocalization was observed in DMF-differentiated PLB985-SV40NLS-PCNA, as evidenced by a residual nuclear fluorescence as compared with PLB985-PCNA due to the inability of the ectopically expressed SV40NLS-PCNA to be retained in the cytoplasm (Fig. 7B). Quantification of the nuclear immunofluorescence confirmed the statistically significant increase in nuclear fluorescence before differentiation observed in PLB985-SV40NLS-PCNA as compared with PLB985-PCNA (Fig. 7C). Although the intensity of PCNA nuclear fluorescence decreased dramatically after neutrophil differentiation in both PLB985-SV40NLS-PCNA and PLB985-PCNA, reflecting the relocalization of endogenous wild-type PCNA in both cell lines, the level of nuclear PCNA was higher in PLB985-SV40NLS-PCNA as compared with PLB985-PCNA, thus validating our cellular model showing an altered cytoplasmic localization in PLB985-SV40NLS-PCNA during granulocytic differentiation. Western blot analysis of nuclear and cytosolic extracts obtained after nitrogen cavitation of PLB985-PCNA and PLB985-SV40NLS-PCNA provided evidence of an increased nuclear PCNA and a slight decrease in cytosolic PCNA in DMF-differentiated PLB985-SV40NLS-PCNA as compared with PLB985-PCNA (Fig. 7D). Of note, both cell lines contained a similar amount of PCNA (data not shown). Notably, despite a significant difference in nuclear PCNA after differentiation in PLB985-SV40NLS-PCNA as compared with PLB985-PCNA, no difference in the CD11b expression after DMF treatment was observed between PLB985-SV40-NLS-PCNA and PLB985-PCNA (Fig. 8A). We next studied whether this nuclear SV40NLS-PCNA mutant could still mediate its antiapoptotic activity in neutrophil-differentiated PLB985. Apoptosis was triggered by gliotoxin, a fungal toxin that inhibits NF-κB (16) and directly binds to Bak to trigger mitochondrial depolarization (17). We confirmed that PLB985 cells overexpressing wild-type PCNA had a decreased gliotoxin-induced apoptosis as compared with PLB985 cells transfected with the control plasmid. This was evidenced by a decreased percentage of cells with apoptotic morphology (Fig. 8B), with mitochondrial depolarization (Fig. 8C), or with DNA fragmentation (Fig. 8D) assessed after May Grunwald Giemsa staining, DIOC6 labeling, or measuring cells in the sub-G₁ phase, respectively. As predicted, neutrophil-differentiated PLB985 cells expressing the nuclear SV40NLS-PCNA mutant failed to display antiapoptotic activity and had apoptosis rates similar to those of controls transfected with the empty plasmid, thus confirming the essential role of PCNA cytoplasmic subcellular localization in its antiapoptotic activity in mature neutrophils.

FIGURE 7.

Localization of the SV40NLS-PCNA mutant in PLB985 cells before and after DMF-induced granulocytic differentiation. A and B, immunofluorescence analysis of PCNA localization in PLB985 cells transfected with pcDNA3-PCNA (wild-type WT PCNA) and pcDNA3-SV40NLS-PCNA before (− DMF) (A) and after (+ DMF) (B) granulocytic differentiation. Immunofluorescence analysis of PCNA was performed with the Ab5 rabbit polyclonal anti-PCNA antibody, and the cells were analyzed with scanning confocal microscopy. The nuclei were visualized with Hoechst staining (scale bar = 20 μm). This representative experiment has been performed at least three times with similar results. C, quantification of nuclear fluorescence intensity with the ImageJ 1.42 software. WT-PCNA- and SV40NLS-PCNA-transfected PLB985 cells are represented by black and gray bars, respectively. Fluorescence intensities have been determined in more than 100 cells in each of the four conditions and were expressed in arbitrary units. Data are means ± S.E. (**, p < 0.01, ANOVA test). AU, arbitrary units. D, Western blot analysis of PCNA protein in nuclear (20 μg/well) and cytosolic (10 μg/well) fractions. β-Actin and lamin B were used as specific marker for cytosol and nuclei and as loading control, respectively.

FIGURE 8.

Defect in antiapoptotic activity of the nuclear SV40NLS-PCNA mutant in neutrophil-differentiated PLB985 cells. A, CD11b membrane expression in neutrophil-differentiated PLB985 cells stably overexpressing wild-type PCNA or the nuclear SV40NLS-PCNA mutant as compared with control PLB985 cells (empty plasmid). PLB985 cells were treated with DMF for 5 days to induce granulocyte differentiation. CD11b expression was measured by flow cytometry and expressed as the percentage of CD11b-positive cells. B–D, neutrophil-differentiated PLB985 cells stably overexpressing wild-type PCNA or the nuclear SV40NLS-PCNA mutant or the control PLB985 cells were incubated with or without 1 μg/ml gliotoxin for 16 h to induce apoptosis. B, May Grunwald Giemsa staining showing cells with condensed nuclei indicative of an apoptotic morphology (shown by black arrows) in gliotoxin-treated cells (lower panels) as compared with untreated cells (upper panels). The typical experiment shown in B was representative of four independent experiments. The percentages of cells with an apoptotic morphology after gliotoxin treatment were 41.4% ± 8, 22.5% ± 7, and 46.9% ± 10 in PLB985-Empty plasmid, PLB985-PCNA, and PLB985-SV40NLS-PCNA, respectively. Data are means ± S.E. (more than 100 cells in each condition have been counted for each experiment). C, mitochondrial depolarization analysis after DIOC6 labeling in the neutrophil-differentiated PLB985 cells. The histogram in the left panel represents the percentage of apoptotic cells with decreased mitochondrial potential, and data are expressed as the means ± S.E. of nine independent experiments (*, p < 0.05, ANOVA test). Representative flow cytometry plots showing the percentage of cells with depolarized mitochondria with decreased DIOC6 labeling are depicted in the right panel. D, the percentage of neutrophil-differentiated PLB985 cells in the sub-G₁ phase showing DNA fragmentation after propidium iodide staining. The histogram on the left panel depicts the means ± S.E. of six independent experiments (*, p < 0.05, ANOVA test), and representative flow cytometry plots are shown in the right panel.

DISCUSSION

Our results demonstrated that PCNA nuclear-to-cytoplasmic relocalization during granulocytic differentiation was pivotal in its antiapoptotic activity in mature neutrophils. This PCNA relocalization involves an NES-mediated active export. Although low molecular mass proteins (below 60 kDa) may shuttle between nucleus and cytoplasm by passive diffusion through nuclear pores, regulated nuclear export has been suggested for several essential regulatory proteins (33–35). One of the salient findings is the identification of a novel functional NES within the PCNA sequence located between Ile-11 and Ile-23. Indeed, this sequence fused to the EGFP protein triggered the nuclear export of the chimeric EGFP in a CRM1-dependent manner because it was inhibited by LMB. Moreover, point mutations on the critical leucine residues within the sequence totally abolished this export. Unexpectedly, point mutations or deletions of this NES within the PCNA sequence did not result in nuclear retention, but in prominent cytoplasmic localization. The exclusive cytosolic localization of PCNA observed for all the PCNA mutants containing a deletion or a mutation in NES1 was surprising. For instance, it is unlikely that the two point mutations in the L12A/L16A would affect PCNA conformation or trimerization, as studied by molecular modeling.⁴ This increase in cytoplasmic localization of NES1-mutated PCNA might also suggest that the NES1 sequence is involved in PCNA nuclear import via its association with yet unknown NLS-containing proteins. Notably, molecular modeling analysis revealed that the residues Ile-11 to Ile-23 are buried in the trimer, and it is difficult to conceive how the export machinery could be accessible in this conformation. In contrast, when PCNA is monomeric, the same residues are exposed and should be accessible to the CRM1. This hypothesis was confirmed by the observation that the PCNAY114A monomeric mutant, which can expose NES1, had a preferred cytosolic localization when expressed in HeLa cells. Accordingly, NES mutations in the PCNAY114A monomeric mutant resulted in a relocalization of the cytosolic PCNAY114 to the nuclear compartment. Moreover, nuclear export of the monomeric PCNAY114A was reversed by LMB, and finally, colocalization between PCNA and CRM1 was observed only in HeLa expressing the monomeric PCNAY114A and not wild-type PCNA. Taken together, our data provided unambiguous evidence that NES1 could mediate monomeric PCNA export via a CRM1-dependent mechanism. It is noteworthy that this NES was localized precisely within the PCNA-DNA interaction domain (31) and, as such, should be masked when PCNA is bound to chromatin, therefore precluding its nuclear export. It is notable that two pools of PCNA are present within the nucleus depending its chromatin association (36). Each pool can be differentially assessed by using immunofluorescence in the presence or absence of Nonidet P-40 (37). The chromatin-bound PCNA will be Nonidet P-40-resistant, whereas soluble PCNA will be removed by Nonidet P-40 nuclear permeabilization. Pertinently, in contrast to wild-type EGFP-PCNA, NES1-deleted-EGFP-PCNA could not be detected by immunofluorescence after Nonidet P-40 treatment, thus suggesting that it was not chromatin-bound (data not shown). Taken together, our results show that the NES1, which is highly conserved between species, as shown in Table 2, can mediate chromatin association in trimeric PCNA, whereas it enhances nuclear export in monomeric PCNA. It is generally admitted that PCNA are active only in their trimeric form. For instance, PCNA mutants that were unable to form trimers have been isolated in a genetic screen as lethal mutants of Saccharomyces pombe PCNA (38). Likewise, in human, it has been well established that mutation of tyrosine 114 that is located within the intermonomer region is essential for both PCNA trimerization and biologic activity on polymerase-δ (32). Further investigations will thus be required to decipher the molecular events regulating PCNA monomer formation during granulocytic differentiation. Notably, in fibroblast, a selective delocalization of nuclear PCNA to cytosol was observed after serum starvation, whereas no change in subcellular localization was observed for three other nuclear antigens, namely DNA polymerase-α, c-Myc, and c-Fos, in different growth conditions (39), suggesting that PCNA nuclear-to-cytoplasmic relocalization might play a role in cells other than neutrophils.

The next relevant piece of information provided by this study is that PCNA nuclear-to-cytoplasmic relocalization occurring during granulocytic differentiation is an essential step for the antiapoptotic effect of PCNA in mature neutrophils. It has to be noted that it is impossible to transfect mature neutrophils, and the use of differentiated promyelocytic cell lines such as PLB985 has proven to be a useful (although not perfect) cellular model to perform molecular analysis of PCNA function in neutrophils (7). The constitutive nuclear form of PCNA was previously obtained by fusing the SV40NLS to PCNA, and PLB985 cells stably expressing this nuclear mutant showed an increased nuclear PCNA before and after differentiation. Nuclear mutant of PCNA did not act as a dominant negative form because nuclear-to-cytosolic relocalization of endogenous PCNA was observed in PLB985-SV40NLS-PCNA as in PLB985-PCNA. However, ectopically expressed SV40NLS-PCNA was still nuclear even after differentiation. Mechanisms involved in PCNA nuclear-to-cytoplasmic relocalization occurring during granulocytic differentiation might include other pathways than the NES1-mediated export reported here. There are certainly some possibilities that the nuclear-to-cytoplasmic relocalization of PCNA might be influenced by other as yet undetermined interacting partner proteins. Further analysis will be required to solve these issues, and a detailed analysis of the molecular partners of cytoplasmic PCNA in mature neutrophils might provide some clues. Whatever the mechanisms that are involved, PCNA cytosolic relocalization during differentiation constitutes a pivotal event to achieve its antiapoptotic function, which we have described previously (7).

Hence, this nuclear-to-cytoplasmic PCNA relocalization could be part of the specific features of the terminal neutrophil differentiation program (40), thus allowing the survival of the neutrophils to be finely tuned. It could be foreseen that a disturbance in this PCNA relocalization might severely impair neutrophil survival and might be implicated in neutropenia, in which an increased susceptibility to neutrophil apoptosis has been suggested (41). G-CSF, the main form of treatment of neutropenia, has been found to dramatically increase PCNA levels in neutrophils in vitro and in vivo (7), thus reinforcing the belief that PCNA plays a major role in neutrophil survival control. In conclusion, similarly to p53, which contains a NES within its tetramerization domain (27), we propose a model in which, during granulocytic differentiation, PCNA subcellular localization is established through a monomerization-regulated exposure of the NES1 to the export machinery. Our work opens new tracks of research on neutrophil survival that might have therapeutical application in neutropenia.