The translation initiation factor 3 subunit eIF3K interacts with PML and associates with PML nuclear bodies

The promyelocytic leukemia protein (PML) is a tumor suppressor protein that regulates a variety of important cellular processes, including gene expression, DNA repair and cell fate decisions. Integral to its function is the ability of PML to form nuclear bodies (NBs) that serve as hubs for the interaction and modification of over 90 cellular proteins. There are seven canonical isoforms of PML, which encode diverse C-termini generated by alternative pre-mRNA splicing. Recruitment of specific cellular proteins to PML NBs is mediated by protein–protein interactions with individual PML isoforms. Using a yeast two-hybrid screen employing peptide sequences unique to PML isoform I (PML-I), we identified an interaction with the eukaryotic initiation factor 3 subunit K (eIF3K), and in the process identified a novel eIF3K isoform, which we term eIF3K-2. We further demonstrate that eIF3K and PML interact both in vitro via pull-down assays, as well as in vivo within human cells by co-immunoprecipitation and co-immunofluorescence. In addition, eIF3K isoform 2 (eIF3K-2) colocalizes to PML bodies, particularly those enriched in PML-I, while eIF3K isoform 1 associates poorly with PML NBs. Thus, we report eIF3K as the first known subunit of the eIF3 translation pre-initiation complex to interact directly with the PML protein, and provide data implicating alternative splicing of both PML and eIF3K as a possible regulatory mechanism for eIF3K localization at PML NBs.     

Epstein-Barr nuclear antigen 1 contributes to nasopharyngeal carcinoma through disruption of PML nuclear bodies

Latent Epstein-Barr virus (EBV) infection is strongly associated with several cancers, including nasopharyngeal carcinoma (NPC), a tumor that is endemic in several parts of the world. We have investigated the molecular basis for how EBV latent infection promotes the development of NPC. We show that the viral EBNA1 protein, previously known to be required to maintain the EBV episomes, also causes the disruption of the cellular PML (promyelocytic leukemia) nuclear bodies (or ND10s). This disruption occurs both in the context of a native latent infection and when exogenously expressed in EBV-negative NPC cells and involves loss of the PML proteins. We also show that EBNA1 is partially localized to PML nuclear bodies in NPC cells and interacts with a specific PML isoform. PML disruption by EBNA1 requires binding to the cellular ubiquitin specific protease, USP7 or HAUSP, but is independent of p53. We further observed that p53 activation, DNA repair and apoptosis, all of which depend on PML nuclear bodies, were impaired by EBNA1 expression and that cells expressing EBNA1 were more likely to survive after induction of DNA damage. The results point to an important role for EBNA1 in the development of NPC, in which EBNA1-mediated disruption of PML nuclear bodies promotes the survival of cells with DNA damage.     

Intranuclear Redistribution of SV40T, p53, and PML in a Conditionally SV40T-Immortalized Cell Line

We have previously reported that EBNA-5, one of the Epstein–Barr virus-encoded proteins, accumulates in the nuclear bodies containing PML, the promyelocytic leukemia associated protein. In this study, we examine the intranuclear distribution of SV40 large T-antigen (SV40T), the p53 tumor suppressor protein (p53), and PML in a conditionally immortalized cell line, IDH4. In IDH4 cells, the expression of SV40T is regulated by a dexamethasone (Dex)-driven promoter. Withdrawal of Dex results in down-regulation of SV40T and growth arrest, whereas addition of Dex to the growth-arrested cells results in up-regulation of SV40T and proliferation. In proliferating IDH4 cells, SV40T is concentrated in nuclear dots that are also positive for p53. Many of these dots are juxtaposed to PML positive structures but do not colocalize with them. After removal of Dex, SV40T–p53 dots gradually disappear, while the PML structures remain. Induction of SV40T in nonproliferating IDH4 cells causes a coordinated redistribution of SV40T and p53. The immunostaining for SV40T and p53 is first weak, then strong with a homogeneous distribution, and 3–4 days later becomes dot-like again. This reappearance of SV40T–p53 dots coincides with the recovery of proliferation in restimulated IDH4 cells. Also, the p53 pattern correlates with the SV40T pattern with regard to both morphology and intensity during both suppression and induction of SV40T. Taken together, our data suggest that (i) the level of p53 is coregulated with the level of SV40T in a dose-dependent fashion; (ii) the formation of SV40T–p53 nuclear dots correlates with the transformed phenotype; (iii) the SV40T–p53 dots localize preferentially to the neighborhood of PML bodies which are already present in normal cells.     

PML clastosomes prevent nuclear accumulation of mutant ataxin-7 and other polyglutamine proteins

The pathogenesis of spinocerebellar ataxia type 7 and other neurodegenerative polyglutamine (polyQ) disorders correlates with the aberrant accumulation of toxic polyQ-expanded proteins in the nucleus. Promyelocytic leukemia protein (PML) nuclear bodies are often present in polyQ aggregates, but their relation to pathogenesis is unclear. We show that expression of PML isoform IV leads to the formation of distinct nuclear bodies enriched in components of the ubiquitin-proteasome system. These bodies recruit soluble mutant ataxin-7 and promote its degradation by proteasome-dependent proteolysis, thus preventing the aggregate formation. Inversely, disruption of the endogenous nuclear bodies with cadmium increases the nuclear accumulation and aggregation of mutant ataxin-7, demonstrating their role in ataxin-7 turnover. Interestingly, beta-interferon treatment, which induces the expression of endogenous PML IV, prevents the accumulation of transiently expressed mutant ataxin-7 without affecting the level of the endogenous wild-type protein. Therefore, clastosomes represent a potential therapeutic target for preventing polyQ disorders.     

ZNF198, a zinc finger protein rearranged in myeloproliferative disease, localizes to the PML nuclear bodies and interacts with SUMO-1 and PML

The ZNF198/FGFR1 fusion gene in atypical myeloproliferative disease produces a constitutively active cytoplasmic tyrosine kinase, unlike ZNF198 which is normally a nuclear protein. We have now shown that the ZNF198/FGFR1 fusion kinase interacts with the endogenous ZNF198 protein suggesting that the function of ZNF198 may be compromised in cells expressing it. Little is currently known about the endogenous function of ZNF198 and to investigate this further we performed a yeast two-hybrid analysis and identified SUMO-1 as a binding partner of ZNF198. These observations were confirmed using co-immunoprecipitation which demonstrated that ZNF198 is covalently modified by SUMO-1. Since many of the SUMO-1-modified proteins are targeted to the PML nuclear bodies we used confocal microscopy to show that SUMO-1, PML and ZNF198 colocalize to punctate structures, shown by immunocytochemistry to be PML bodies. Using co-immunoprecipitation we now show that PML and sumoylated ZNF198 can be found in a protein complex in the cell. Mutation of the SUMO-1 binding site in wild-type ZNF198 resulted in loss of distinct PML bodies, reduced PML levels and a more dispersed nuclear localization of the PML protein. In cells expressing ZNF198/FGFR1, which also lack the SUMO-1 binding site, SUMO-1 is preferentially localized in the cytoplasm, which is associated with loss of distinct PML bodies. Recently, arsenic trioxide (ATO) was proposed as an alternative therapy for APL that was resistant to traditional therapy. Treatment of cells expressing ZNF198/FGFR1 with ATO demonstrated reduced autophosphorylation of the ZNF198/FGFR1 protein and induced apoptosis, which is not seen in cells expressing wild-type ZNF198. Overall our results suggest that the sumoylation of ZNF198 is important for PML body formation and that the abrogation of sumoylation of ZNF198 in ZNF198/FGFR1 expressing cells may be an important mechanism in cellular transformation.     

PML isoform II plays a critical role in nuclear lipid droplet formation

Lipid droplets (LDs) in the nucleus of hepatocyte-derived cell lines were found to be associated with premyelocytic leukemia (PML) nuclear bodies (NBs) and type I nucleoplasmic reticulum (NR) or the extension of the inner nuclear membrane. Knockdown of PML isoform II (PML-II) caused a significant decrease in both nuclear LDs and type I NR, whereas overexpression of PML-II increased both. Notably, these effects were evident only in limited types of cells, in which a moderate number of nuclear LDs exist intrinsically, and PML-II was targeted not only at PML NBs, but also at the nuclear envelope, excluding lamins and SUN proteins. Knockdown of SUN proteins induced a significant increase in the type I NR and nuclear LDs, but these effects were cancelled by simultaneous knockdown of PML-II. Nuclear LDs harbored diacylglycerol O-acyltransferase 2 and CTP:phosphocholine cytidylyltransferase α and incorporated newly synthesized lipid esters. These results corroborated that PML-II plays a critical role in generating nuclear LDs in specific cell types.     

The major immediate-early proteins IE1 and IE2 of human cytomegalovirus colocalize with and disrupt PML-associated nuclear bodies at very early times in infected permissive cells.

The major immediate-early (MIE) gene products of human cytomegalovirus (HCMV) are nuclear phosphoproteins that are thought to play key roles in initiating lytic cycle gene regulation pathways. We have examined the intranuclear localization pattern of both the IE1 and IE2 proteins in virus-infected and DNA-transfected cells. When HCMV-infected human diploid fibroblast (HF) cells were stained with specific monoclonal antibodies, IE1 localized as a mixture of nuclear diffuse and punctate patterns at very early times (2 h) but changed to an exclusively nuclear diffuse pattern at later times. In contrast, IE2 was distributed predominantly in nuclear punctate structures continuously from 2 to at least 12 h after infection. These punctate structures resembled the preexisting PML-associated nuclear bodies (ND10 or PML oncogenic domains [PODs]) that are disrupted and dispersed by the IE110 protein as a very early event in herpes simplex virus (HSV) infection. However, HCMV differed from HSV by leading instead to a change in both the PML and SP100 protein distribution from punctate bodies to uniform diffuse patterns, a process that was complete in 50% of the cells at 2 h and in 90% of the cells by 4 h after infection. Confocal double-label indirect immunofluorescence assay analysis confirmed that both IE1 and IE2 colocalized transiently with PML in punctate bodies at very early times after infection. In transient expression assays, introduction of IE1-encoding plasmid DNA alone into Vero or HF cells produced the typical total redistribution of PML into a uniform nuclear diffuse pattern together with the IE1 protein, whereas introduction of IE2-encoding plasmid DNA alone resulted in stable colocalization of the IE2 protein with PML in the PODs. A truncated mutant form of IE1 gave large nuclear aggregates and failed to redistribute PML, and similarly a deleted mutant form of IE2 failed to colocalize with the punctate PML bodies, confirming the specificity of these effects. Furthermore, both Vero and U373 cell lines constitutively expressing IE1 also showed total PML relocalization together with the IE1 protein into a nuclear diffuse pattern, although a very small percentage of the cells which failed to express IE1 reverted to a punctate PML pattern. Finally, the PML redistribution activity of IE1 and the direct association of IE2 with PML punctate bodies were both confirmed by infection with E1A-negative recombinant adenovirus vectors expressing either IE1 or IE2 alone. These results confirm that transient colocalization with and disruption of PML-associated nuclear bodies by IE1 and continuous targeting to PML-associated nuclear bodies by IE2 are intrinsic properties of these two MIE regulatory proteins, which we suggest may represent critical initial events for efficient lytic cycle infection by HCMV.     

The Epstein-Barr virus-encoded nuclear antigen EBNA-5 accumulates in PML-containing bodies.

EBNA-5 is one of the Epstein-Barr virus (EBV)-encoded nuclear proteins required for immortalization of human B lymphocytes. In the nuclei of EBV-transformed lymphoblastoid cell lines EBNA-5 is preferentially targetted to distinct nuclear foci. Previously we have shown (W.Q. Jiang, L. Szekely, V. Wendel-Hansen, N. Ringertz, G. Klein, and A. Rosen, Exp. Cell Res. 197:314-318, 1991) that the same foci also contained the retinoblastoma (Rb) protein. Using a similar double immunofluorescence technique, we now show that these foci colocalize with nuclear bodies positive for PML, the promyelocytic leukemia-associated protein. Artificial spreading of the chromatin by exposure to the forces of fluid surface tension disrupts this colocalization gradually, suggesting that the bodies consist of at least two subcomponents. Heat shock or metabolic stress induced by high cell density leads to the release of EBNA-5 from the PML-positive nuclear bodies and induces it to translocate to the nucleoli. In addition to their presence in nuclear bodies, both proteins are occasionally present in nuclear aggregates and doughnut-like structures in which PML is concentrated in an outer shell. Nuclear bodies with prominent PML staining are seen in resting B lymphocytes. This staining pattern does not change upon EBV infection. In freshly infected cells EBNA-5 antigens are first distributed throughout the nucleoplasm. After a few days intensely staining foci develop. These foci coincide with PML-positive nuclear bodies. At a later stage and in established lymphoblastoid cell lines EBNA-5 is almost exclusively present in the PML-positive nuclear foci. The colocalization is restricted to EBV-infected human lymphoblasts. The data presented indicate that the distinct EBNA-5 foci are not newly formed structures but the result of translocation of the viral protein to a specialized domain present already in the nuclei of uninfected cells.     

FIF [fibroblast growth factor-2 (FGF-2)-interacting-factor], a nuclear putatively antiapoptotic factor, interacts specifically with FGF-2

Numerous evidence indicates that some of the activities of fibroblast growth factor 2 (FGF-2) depend on an intracrine mode of action. Recently, we showed that three high molecular mass (HMM) nuclear forms of FGF-2 are part of a 320-kDa protein complex while the cytoplasmic AUG-initiated form is included in a 130-kDa complex. Consequently, the characterization of FGF endogenous targets has become crucial to allow the elucidation of their endogenous activities. Through the screening of GAL4-based yeast two-hybrid expression libraries, we have isolated a gene encoding a nuclear protein of 55 kDa, FIF (FGF-2-interacting-factor), which interacts specifically with FGF-2 but not with FGF-1, FGF-3, or FGF-6. In this system, FIF interacts equally well with the NH2-extended 24-kDa FGF form as with the 18-kDa form, indicating that the FIF-binding motif is located in the last 155 amino acids of FGF-2. Nevertheless, coimmunoprecipitation experiments showed an exclusive association with HMM FGF-2. The predicted protein contains a canonical leucine zipper domain and three overlapping hydrophobic heptad repeats. The region spanning these repeats is, together with a region located in the N-terminal part of the FIF protein, implicated in the binding to FGF-2. In contrast to the full-length FIF protein, several deletion constructs were able to transactivate a lac-Z reporter gene. Furthermore, the COOH-terminal part, but not the full-length FIF protein, has previously been shown to exhibit antiapoptotic properties. Thus we discuss the possibility that these activities could reflect a physiological function of FIF through its interaction with FGF-2.