the translation initiation factor 3 subunit eif3k interacts with pml … · 2016. 1. 11. · pml...

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/yexcr

E X P E R I M E N T A L C E L L R E S E A R C H 3 1 9 ( 2 0 1 3 ) 2 5 5 4 – 2 5 6 5

0014-4827/$ - see frohttp://dx.doi.org/10.1

nCorresponding auFax: þ1 902 494 2519

E-mail address: d

Research Article

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

Jayme Salsmana, Jordan Pindera, Brenda Tsea, Dale Corkeryb, Graham Dellairea,b,n

aDepartment of Pathology, Dalhousie University, P.O. Box 15000, Halifax, Nova Scotia, Canada B3H 4R2bDepartment of Biochemistry and Molecular Biology, Dalhousie University, Halifax, Nova Scotia, Canada

a r t i c l e i n f o r m a t i o n

Article Chronology:

Received 11 December 2012Received in revised form8 August 2013Accepted 2 September 2013Available online 11 September 2013

Keywords:

PMLeIF3KNuclear structureLight microscopyYeast two-hybrid

nt matter & 2013 Elsevier016/j.yexcr.2013.09.001

thor at: Dalhousie [email protected] (G. Dellaire

a b s t r a c t

The promyelocytic leukemia protein (PML) is a tumor suppressor protein that regulates a varietyof 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 theinteraction and modification of over 90 cellular proteins. There are seven canonical isoforms ofPML, which encode diverse C-termini generated by alternative pre-mRNA splicing. Recruitment ofspecific cellular proteins to PML NBs is mediated by protein–protein interactions with individualPML isoforms. Using a yeast two-hybrid screen employing peptide sequences unique to PMLisoform 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 furtherdemonstrate that eIF3K and PML interact both in vitro via pull-down assays, as well as in vivowithin human cells by co-immunoprecipitation and co-immunofluorescence. In addition, eIF3Kisoform 2 (eIF3K-2) colocalizes to PML bodies, particularly those enriched in PML-I, while eIF3Kisoform 1 associates poorly with PML NBs. Thus, we report eIF3K as the first known subunit of theeIF3 translation pre-initiation complex to interact directly with the PML protein, and provide dataimplicating alternative splicing of both PML and eIF3K as a possible regulatory mechanism foreIF3K localization at PML NBs.

& 2013 Elsevier Inc. All rights reserved.

Introduction

The human promyelocytic leukemia (PML) protein is an importanttumor suppressor protein, and mice in which the PML gene isdisrupted are more susceptible to carcinogen-induced cancers[5,13,56]. The PML protein acts as the molecular scaffold for theformation of promyelocytic leukemia nuclear bodies (PML NBs),which are discrete, dynamic subnuclear structures found withinmammalian cells [3]. These heterogeneous multi-protein com-plexes are implicated in a wide spectrum of cellular functions

Inc. All rights reserved.

ity, Departments of Pathol

).

including tumor suppression, DNA repair, apoptosis, antiviralresponse, cellular senescence, and transcriptional regulation[4,13,37,45,60]. While PML is the major structural component ofthe PML NB, more than 90 different proteins can associate withPML NBs, either transiently or constitutively [15,40,54] and manyof these proteins themselves are intimately linked to the variousfunctions ascribed to PML NBs. One of the prominent functionsassociated with PML and PML NBs is tumor suppression, since lossof the PML protein and/or PML NBs has been associated withacute promyelocytic leukemia [12,32] and various human solid

ogy, P.O. Box 15000, Halifax, Nova Scotia, Canada B3H 4R2.

dx.doi.org/10.1016/j.yexcr.2013.09.001dx.doi.org/10.1016/j.yexcr.2013.09.001dx.doi.org/10.1016/j.yexcr.2013.09.001http://crossmark.crossref.org/dialog/?doi=10.1016/j.yexcr.2013.09.001&domain=pdfmailto:[email protected]/10.1016/j.yexcr.2013.09.001

E X P E R I M E N T A L C E L L R E S E A R C H 3 1 9 ( 2 0 1 3 ) 2 5 5 4 – 2 5 6 5 2555

tumors including carcinomas of the colon, lung, breast andprostate [25,55].

The PML gene transcript is generated from nine exons, of whichexons 6–9 are spliced alternatively to yield at least seven PMLisoforms with unique C-terminal sequences [20,29]. Exons 1–3 areconserved in all PML isoforms and encode a tripartite motif(TRIM) containing a RING finger, two B-boxes and an α-helicalcoiled-coil motif [44] which are involved in mediating homo- andhetero-association between PML isoforms [6,22]. Several motifshave also been identified in the C-terminus of PML, including anuclear localization signal in exon 6 [30] and a nuclear exportsignal in exon 9 of PML isoform I (PML-I) [26,29]. Relative to theother PML isoforms, PML-I is more highly conserved betweenmouse and humans, and has the highest expression level [10],suggesting that this isoform is a significant contributor to PML NBfunctions. It is becoming increasingly clear that isoform-specificPML protein–protein interactions help regulate PML biology. Forexample, PML-IV has been shown to interact with p53 [21,24],whereas PML-II has recently been implicated in expression of theMHC class II gene cluster through specific recruitment of CIITA[53]. Additionally, PML-I interacts with acute myeloid leukemia 1(AML1) [41] in an isoform-specific manner and PML-III interactswith TRF1 to facilitate the alternative lengthening of telomeres(ALT) in U2OS cells [58]. Collectively, these examples suggest thatindividual isoforms through their protein–protein interactions canalter PML NB function, which could occur in a tissue-specific

Fig. 1 – Domain mapping of the interaction of PML-I with eIF3K. (Awhich the protein fragments indicated in (C) are fused to either thtranscriptional activation domain (Gal4AD) and expressed in a yeaseIF3K indicating the locations of the sequence encoded by exon 2the aa positions of the truncation points for the deletion series uswere patched onto solid YPDA medium and incubated for 7 days areporter gene accumulate a dark red pigment (�, darker shades),(þ or þþ, lighter shades).

manner or in response to various stimuli [10]. Thus, it is likely thatthe diversity of cellular functions attributed to the PML NB arise,at least in part, via equally diverse protein–protein interactionsmediated by the various PML isoforms.In this study, we used a yeast two-hybrid screen to identify a

novel protein–protein interaction between peptide sequencesunique to PML-I, encoded by exon 9 of the PML gene, and subunitK of the eukaryotic translation initiation factor 3 (eIF3K). Inaddition, we show that eIF3K also interacts with the PML proteinin an exon 9-independent manner, and localizes to PML NBs.Furthermore, the recruitment of eIF3K to PML NBs appears to beregulated by alternative splicing, as eIF3K isoform 2, which weterm eIF3K-2, is preferentially recruited to PML NBs and is furtherenriched at bodies by overexpression of PML-I. This is the firstsubunit of the eIF3 translation pre-initiation complex shown tointeract directly with PML and, in addition to eIF3E, is the secondsubunit of eIF3 known to associate with PML nuclear bodies.

Results

eIF3K interacts with a peptide encoded by exon 9 of PML-I

In order to identify proteins that specifically interact with PML-I andnot with other PML isoforms we performed a yeast two-hybridscreen using the unique exon 9 region of human PML-I [20,29],

) Schematic of a yeast two-hybrid protein interaction assay, ine yeast Gal4 DNA-binding domain (Gal4DBD) or the Gal4t strain encoding the ADE2 reporter gene. (B) Diagram of(absent in eIF3K isoform 2), the HAM and WH domains anded in (C). (C) Yeast expressing the indicated fusion proteinst 28 1C. Yeast lacking detectable expression of the ADE2while yeast expressing ADE2 appear pink or light pink


encoding amino acids 604–882 as a bait to screen a human fetalbrain cDNA library [16] (Fig. 1A). One of the interacting peptidesidentified was an N-terminal fragment of subunit K of the eukaryotictranslation initiation factor 3, (eIF3K, aa 1–97) (Supplementaryinformation Fig. 1). The human eIF3 complex contains at least 13subunits, of which eIF3K is non-essential, and may therefore play arole in regulating the translation initiation of specific transcripts[2,7,27,36,38,42]. There are three well-defined domains within theeIF3K sequence (Fig. 1B). The first is the N-terminal HEAT (Hunting-ton, Elongation factor 3, A subunit of protein phosphatase 2A,Target-of-rapamycin) repeat-like HAM (HEAT analogous motif) (aa24–121) [57]. HEAT domains often mediate protein–protein interac-tions [1]. The second domain is the winged-helix-like (WH) domainin the C-terminus (aa 132–191), which could mediate protein–RNAinteractions [57]. In addition, the HAM and WH domains formelements of the PCI (26S proteasome lid, COP9 signalosome,eukaryotic initiation factor 3) domain (aa 60–218). The fragmentof eIF3K identified in the yeast two-hybrid screen (aa 1–97) encodes�75% of the HAM domain (summarized in Fig. 1B).To further characterize the interaction between PML-I and

eIF3K, yeast two-hybrid interaction assays were performed usingvarious truncated peptides encoding fragments of PML-I andeIF3K. The amino acid sequence encoded by exon 9 of PML-Iincludes a predicted exonuclease III-like fold (aa 604–750)required for targeting PML-I to nucleolar fibrillar centers duringstress or senescence [9]. To determine whether the exonucleaseIII-like domain of PML-I was required for interaction with eIF3K,we tested a fragment encoded by exon 9 of PML-I (aa 750–882),

Fig. 2 – In vitro interaction between PML-I and eIF3K. MBP, MBP-Pexpressed in E. coli and purified with amylose (MBP) or glutathionbound to GST beads, followed by incubation with MBP–PML-I-E9Δdetected by SDS-PAGE and Western blotting. (B) MBP or MBP–PMLincubation with GST-eIF3K. Recovery of GST-eIF3K with MBP–PMLFor (A) and (B) lanes 1–3 were loaded with the equivalent of 10% ofproteins after purification.

which lacks the exonuclease III-like domain, for interaction witheIF3K using a yeast two-hybrid assay. This fragment of PML-I wasfound to interact with eIF3K, indicating that the PML exonucleaseIII-like domain was not required for interaction with eIF3K(Supplementary information Fig. 1).

In order to map the region of eIF3K required for interactionwith amino acids 750–882 of PML-I, several eIF3K deletionmutants were also generated, and tested for interaction withPML-I (aa 750–882) in a yeast two-hybrid assay (Fig. 1C). Thesmallest fragment of eIF3K found to interact with PML-I containedresidues 57–97 that correspond to a region within the HAMdomain of eIF3K. The C-terminal region of eIF3K (aa 97–218)absent in the sequence identified from the yeast two-hybridscreen did not interact with PML-I (aa 750–882), suggesting thatresidues 57–97 may contain the only region required for interac-tion with this PML-I fragment.

During the cloning of the full-length human eIF3K cDNA fromHeLa total mRNA, we isolated two different forms of eIF3K, onematching the previously reported isoform ([52]; NCBI ReferenceID: NM_013234), as well as an unreported isoform (Supplemen-tary information Fig. 2). The novel eIF3K isoform (eIF3K isoform 2;GenBank ID: JX870647) is similar to a human (EAW56803) and amouse EST (AAH27638), suggesting that this isoform is conservedin mammals. Isoform 2 of eIF3K (eIF3K-2) contains an in-framedeletion of exon 2 (encoding aa 21–53) resulting in deletion ofpart of the HAM domain. Removing the residues encoded by exon2 from the fragment of eIF3K isolated in the yeast two-hybridscreen (clone 1–20, 54–97) appeared to weaken but did not

ML-I-E9Δexo (750–882 aa), GST and GST-eIF3K proteins weree Sepharose (GST) beads. (A) GST or GST-eIF3K proteins wereexo. Recovery of MBP–PML-I-E9Δexo with GST-eIF3K was-I-E9Δexo proteins were bound to amylose beads, followed by-I-E9Δexo was detected by SDS-PAGE and Western blotting.the sample used for pull-downs. Lanes 4 and 5 show the eluted

Fig. 3 – eIF3K interacts with PML in vitro and in vivo. (A) GST orGST-eIF3K proteins (eIF3K-1 and eIF3K-2) were bound to GSTbeads, followed by incubation with 6HIS-tagged PML-I orPML-IV from insect cell lysates. Recovery of PML-I or PML-IVfollowing GST purification was detected by SDS-PAGE andWestern blotting. (B) HEK-293 cell lysates were used forco-immunoprecipitation of endogenous PML and eIF3Kproteins with mouse anti-eIF3K monoclonal antibody.Immunoprecipitated (IP) proteins were separated by SDS-PAGEand the co-immunoprecipitated PML protein was assessed byWestern blotting analysis with rabbit anti-PML polyclonalantibodies. Input lanes were loaded with the equivalent of 10%of the sample used for immunoprecipitation.


abolish interaction with PML-I (aa 750–882), suggesting bothisoforms of eIF3K interact with this isoform of PML (Fig. 1C).This weaker two-hybrid interaction could reflect a preference ofPML-I for isoform 1 over isoform 2 of eIF3K. We also observedthat the eIF3K-2 fragment was not as highly expressed in yeast asthe isoform 1 fragment (Supplementary information Fig. 3), thusexpression level differences could also contribute to the differ-ences observed in the reporter assay.

PML and eIF3K interact in vitro and in vivo

To test whether the interaction between PML-I and eIF3Kobserved in our yeast assay was direct, we performed in vitroprotein interaction assays using affinity-purified recombinantglutathione-S-transferase (GST)-tagged eIF3K-2 and maltose bind-ing protein (MBP)-tagged PML-I (aa 750–882), which lacks theexonuclease III domain. This shorter exon 9 fragment was used inthese assays due to the propensity of the full-length peptideencoded by exon 9 to form insoluble protein aggregates whenexpressed in E. coli (data not shown). In vitro pull-down assaysrevealed that MBP–PML-I750–882 bound to GST-eIF3K but not GSTalone (Fig. 2A). The reciprocal MBP co-purification assay demon-strated that GST-eIF3K bound to MBP–PML-I750–882 but not MBPalone (Fig. 2B). Collectively these data support the interaction ofPML-I directly with eIF3K in vitro, which is consistent with theobserved yeast two-hybrid interaction assay (Fig. 1C).

To address the possibility that eIF3K could interact with otherregions of PML-I, hexahistadine-tagged full length PML-I and PML-IVwere expressed in insect cells and assessed for the ability toco-purify with GST-tagged eIF3K-1 or eIF3K-2 (Fig. 3A). PML-IVcontains the same exon arrangement as PML-I, but lacks exon 9 andinstead contains a short sequence of 13 aa at the C-terminus that areunique to PML-IV [10,29]. After GST-purification, both PML-I andPML-IV were recovered with eIF3K isoforms 1 and 2, but not withGST alone (Fig. 3A), suggesting that both isoforms of eIF3K alsointeract with PML independent of the sequence encoded by exon 9.

To determine whether PML and eIF3K interact in human cells weperformed co-immunoprecipitation assays. Consistent with dataobtained from the yeast two-hybrid and in vitro protein interactionassays, endogenous PML was recovered when endogenous eIF3Kwas immunoprecipitated from human HEK-293T cells (Fig. 3B),indicating that these two proteins can also associate in vivo.

eIF3K colocalizes with PML NBs

Although several reports indicate that eIF3K is primarily localizedto the cytoplasm, the precise subcellular localization of eIF3K isnot well established. Different reports indicate eIF3K can be foundassociated with intermediate filaments [35] or sites of cell–cellcontact [31] in epithelial cells as well as adopting either peri-nuclear [31] or pan-cytoplasmic [35] localization in other celltypes. Cytoplasmic localization is also consistent with the abilityof eIF3K to associate with the translation pre-initiation complex.However, other translation initiation factors, such as eIF3E andeIF4E, also have nuclear functions and have been reported toassociate with PML and PML NBs [17,33]. Similarly, although PMLis primarily a nuclear localized protein, at least one isoform (PML-VIIb) lacks the nuclear localization signal [29] and is localized tothe cytoplasm where it appears to play a role in TGF-β signaling[34]. PML isoforms also localize to both the nucleus and

cytoplasm in early G1 [14,19]. In addition, PML-I contains apredicted nuclear export signal (NES), potentially allowing shut-tling to and from the nucleus, although the precise regulation of


this shuttling and its functional consequences have yet to bedetermined [26]. Therefore, we wanted to determine the cellularcompartment in which PML and eIF3K interact, which may thenaffect the interpretation of the possible functional consequences oftheir interaction.To determine the cellular localization of PML and eIF3K, U2OS

cells were immunostained for endogenous eIF3K and PML

Fig. 4 – eIF3K colocalizes with PML-I and PML-IV in PML NBs. (A) U(green) and endogenous PML (red). (B) U2OS cells stably expressinimmunostained for endogenous eIF3K (red). Examples of colocalizindicated with arrows. DNA is visualized with DAPI (blue, MergeþDof cells displaying either strong (black) or partial (gray) colocalizacells as indicated. (D) Western blot of 293A, U2OS and HeLa cell ly�20 kDa in U2OS and HeLa lysates observed with longer exposure(E) U2OS cells were transfected with empty vector, eIF3K-1 or eIF3K-2 ewith anti-eIF3K antibodies. Untransfected GFP–PML-I- and GFP–PML-I

(Fig. 4A). Endogenous eIF3K is localized primarily to the cytoplasmin perinuclear structures consistent with a previous report [31].The specificity of the antibody used in our studies for eIF3K-1 andeIF3K-2 was confirmed by the co-detection of ectopicallyexpressed epitope-tagged eIF3K isoforms by immunofluroesenceand Western blotting techniques (Supplementary informationFig. 4). Conversely, PML was primarily found in the nucleus of

2OS cells were fixed and immunostained for endogenous eIF3Kg GFP-tagged PML-I or PML-IV (green) were fixed andation (merge) between eIF3K and PML nuclear bodies areAPI). Scale bars¼10 μm. (C) Bar graph indicating the percentagetion between eIF3K and PML in PML-I- or PML-IV-expressingsates with anti-eIF3K antibody. Arrow indicates a protein of(Long Exp) that is of similar molecular weight as eIF3K-2.xpression plasmids and cell lysates were analyzed by Western blotV-expressing U2OS cells were similarly prepared and analyzed.


cells within punctate nuclear domains known as PML nuclearbodies (NBs). Thus, PML and eIF3K do not appear to colocalize tothe same discrete structures under normal growth conditions inU2OS cells. However, with the exception of a small number ofconstitutively associated PML NB components such as SP100,SUMO1 and the PML isoforms [10,51] the majority of PMLNB-associated proteins interact with PML NBs in a transientfashion either during the cell cycle or in response to diversecellular stimuli and cell stress [5,13]. Often these interactions canbe enhanced by increased expression of PML, the putativeinteracting protein, or both. In order to enhance the probabilityof observing eIF3K at PML NBs, we used U2OS cell lines thatexpress N-terminally GFP-tagged PML-I or PML-IV [18]. PML NBstypically contain all nuclear PML isoforms, with PML-I being thepredominant isoform [10]. In these cell lines, PML NBs increase innumber and are enriched in the PML isoform that is over-expressed (Fig. 4B, Supplementary information Fig. 5).

In contrast to untransfected U2OS cells, endogenous eIF3Kassociated with PML NBs in a subset of cells expressed GFP-tagged PML-I or PML-IV (Fig. 4B). The percentage of cells in whichendogenous eIF3K colocalized with PML nuclear bodies wasquantified (Fig. 4C). Colocalization was scored as either strong,partial or negative (Supplementary information Fig. 6). In cellswith strong colocalization every PML nuclear body had an obviouseIF3K punctum associated with it (Supplementary informationFig. 6). Partial colocalization was defined as a cell having eithersome, but not all PML bodies associated with eIF3K or havingpoorly defined eIF3K-PML body colocalization (Supplementaryinformation Fig. 6). In U2OS cells expressing GFP–PML-I, 71% ofcells (29% strong, 43% partial colocalization) showed some degreeof endogenous eIF3K localized to PML-I enriched NBs (Fig. 4C).Similarly, in U2OS cells expressing GFP–PML-IV, 91% of cells (47%strong, 44% partial colocalization) showed some degree of eIF3Kassociation with PML-IV enriched NBs (Fig. 4C). These resultsindicate that endogenous eIF3K isoforms can associate with PMLNBs enriched in either PML-I or PML-IV, which is consistent withour ability to co-purify full-length recombinant PML-I and PML-IVwith eIF3K in vitro (Fig. 3A).

While the data presented in Fig. 4B indicate that endogenouseIF3K interacts with PML-I- or PML-IV-enriched PML NBs, it isunclear which eIF3K isoform(s) might be involved in the associa-tion. To begin to identify which isoform(s) of eIF3K might beassociated with PML NBs we first analyzed cell lysates from U2OS,293 and HeLa cells by Western blotting (Fig. 4D). In all three celllysates, a protein of �24 kDa was detected with the anti-eIF3Kantibody (Novus Biologicals, NB100-93304, Fig. 4D). There is somediscrepancy about the molecular weight of eIF3K, with variousreports detecting eIF3K at between 24 and 28 kDa [31,48]. Sincethis antibody was confirmed to detect epitope-tagged eIF3K-1 andeIF3K-2 by Western blotting (Supplementary information Fig. 4C)we also tested the ability to detect untagged eIF3K-1 and eIF3K-2by expressing these proteins in U2OS cells (Fig. 4E). Compared tovector-transfected cells, eIF3K-1-transfected cells showed a sub-stantial increase in the �24 kDa band, suggesting that eIF3K-1 isthe predominant isoform in the cell lines we tested (Fig. 4E).Expression of untagged eIF3K-2 revealed a protein of �20 kDa(Fig4E) which is similar in size to proteins identified in U2OS andHeLa lysates upon longer exposure (Fig. 4C, arrow), suggestingthat eIF3K-2 may comprise only a small fraction of total eIF3K incells. In addition, detection of eIF3K in GFP–PML-I and GFP–PML-

IV expressing U2OS cells (Fig. 4E) was similar to untreated U2OS(Fig. 4D) or vector-transfected U2OS cells (Fig. 4E), suggesting thateIF3K-1 is the predominant isoform in these cells and thatexpression of PML-I or PML-IV does not significantly alter theeIF3K isoform distribution. Given that only a portion of theendogenous eIF3K immnuofluorescence signal is associated withPML-NBs (Fig. 4B) it remained unclear which eIF3K isoform(s) wasbeing detected at the PML NB. It is possible that only eIF3K-2, aminor component of eIF3K in the cell, is specifically recruited toPML NBs. Alternatively, only a small portion of total eIF3K-1 maybe localized to bodies. Finally, since both isoforms of eIF3Kinteract with PML in vitro, it is possible that both isoforms ofeIF3K associate with PML NBs.In order to more directly test which isoform(s) of eIF3K can

associate with PML bodies, C-terminally FLAG-tagged eIF3K isoforms1 and 2 were expressed in U2OS cells stably expressing GFP–PML-Ior GFP–PML-IV, and the ability to associate with PML NBs wasassessed (Fig. 5A). Isoform 1 of eIF3K showed a pan-cellularlocalization pattern and was generally not observed to associatewith PML NBs in either the PML-I- or PML-IV-expressing cells.Isoform 2 of eIF3K adopted a similar pan-cellular localizationpattern, however in contrast to isoform 1, it could additionally befound at PML NBs (Fig. 5A). In contrast to endogenous eIF3K, thisassociation was more apparent in PML-I-enriched bodies thanPML-IV-enriched bodies (Fig. 5A). Similar localization experimentsperformed with untagged or HA-tagged eIF3K-1 and eIF3K-2 yieldedsimilar results, with eIF3K-2 showing increased association withPML-I-enriched PML NBs (Supplementary information Figs. 7 and 8).The percentage of transfected cells showing strong or partialcolocalization with PML NBs was determined (Fig. 5B) using thesame scoring criteria as in Fig. 4 (Supplementary information Fig. 9).When FLAG-tagged eIF3K-1 was expressed in PML-I- or PML-IV-expressing U2OS cells, there was no colocalization between eIF3K-1and PML NBs. (Fig. 5B). In contrast, FLAG-tagged eIF3K-2 waspartially (20.5%, gray bars) or strongly (0.7%, black bars) associatedwith PML-IV-enriched NBs in 21.2% of the eIF3K-2-transfected cells(Fig. 5B). In PML-I-expressing U2OS cells, eIF3K-2 was partially (35%)or strongly (26.1%) associated with nuclear bodies in 61.1% oftransfected cells (Fig. 5B). A similar trend was observed withuntagged and HA-tagged eIF3K (Fig. 5B).These results suggest thateIF3K-2, but not eIF3K-1, associates with PML NBs enriched in PML-Ito a greater degree than those bodies enriched in PML-IV. Unlike thelocalization pattern observed for endogenous eIF3K (Fig. 4B), inwhich the isoform(s) of eIF3K being detected are uncertain, theseresults identify eIF3K-2 as a specific isoform of eIF3K thatinteracts with PML NBs. Furthermore, these results implicate theexon 9 encoded peptide, found only in PML-I, as an importantdeterminant of eIF3K-2 interaction with PML NBs in cells. Similarcolocalization of eIF3K-2 and PML-I was also observed whenHA-tagged eIF3K-2 and FLAG-tagged PML-I were expressed inU2OS cells (Supplementary information Fig. 10). The association ofeIF3K-2 and PML-I was also confirmed by co-immunoprecipitationfrom cellular lysates from cells expressing FLAG-tagged PML-I andHA-tagged eIF3K-2 (eIF3K-2-HA) in transfected HEK-293A cells(Fig. 5C). eIF3K-2-HA was not recovered from cell lysates whenFLAG-PML-I was not present or in the rabbit IgG control IP samples(Fig. 5C), indicating that eIF3K-2-HA was specifically recovered withPML-I. Taken with the experiments in Fig. 3B, these data alsoindicate that eIF3K-2 and PML can be reciprocally immunoprecipi-tated, further supporting the specificity of this interaction in cells.

Fig. 5 – eIF3K isoform 2 colocalizes with PML-I-containing nuclear bodies. (A) FLAG-tagged eIF3K was expressed in U2OS cellsstably expressing GFP–PML-I or GFP–PML-IV. Cells were fixed and immunostained for eIF3K (anti-FLAG). Examples of colocalizationbetween eIF3K and PML is indicated by arrows. DNA is visualized with DAPI (blue, merge with DAPI). Scale bars¼10 μm. (B) Bargraph indicating the percentage of transfected cells displaying either strong (black bars) or partial (gray bars) colocalizationbetween eIF3K and PML in GFP–PML-I- or GFP–PML-IV-expressing U2OS cells. For each cell line, eIF3K-1 or eIF3K-2 was expressedwith either no epitope tag (�) or with an N-terminal FLAG (F) or HA (H) tag as indicated. Values represent the mean þ/� standarderror, n¼3–5. (C) FLAG-PML-I and eIF3K-2-HA expression plasmids were transiently expressed, alone or in combination, in 293Acells. Following immunoprecipitation (IP) with anti-FLAG resin or mouse IgG control beads, recovered proteins were identified byWestern blotting (WB) for PML-I-FLAG and eIF3K-2-HA as indicated. Input lanes were loaded with the equivalent of 10% of thesample used for immunoprecipitation and 80% of the post-IP elution was loaded in the IP lanes.



Discussion

Loss of the PML protein is associated with genomic instability andcancer of various histological origins [25,59]. There are multipleisoforms of the PML protein that localize to PML NBs [10,29] yethow each individual PML isoform contributes to the variousfunctions attributed to PML NBs is poorly understood. The distinctC-terminal domains of PML isoforms can coordinate specificprotein–protein interactions and contribute to the regulation ofPML NB function. Although more than 90 proteins are confirmedto associate with PML NBs [15], the majority of interactions aretransitory in nature [13], and only a few isoform-specific protein–protein interactions have been reported. Furthermore, the major-ity of PML NB interactions reported to date have been described inrelation to PML-IV. In this study, we have identified the eukaryotictranslation initiation factor 3 subunit K (eIF3K) as a novelinteracting protein of PML. At least two other translation factors,eIF3E and eIF4E [17,33] have been reported to colocalize with PMLNBs; however, no PML isoform-specific interactions have beendescribed for these proteins nor have specific isoforms of transla-tion factors been shown to preferentially interact with PML.The interaction between PML and eIF3K appears to be multi-factorial with one interaction domain unique to PML-I (aa 750–882, encoded by exon 9) and at least one other interaction domainfound in exons 1–8, which are conserved between PML -I andPML-IV. Yeast two-hybrid assays indicated that the association ofeIF3K with exon 9 of PML-I requires a portion of the HAM domainof eIF3K. The HAM domain of eIF3K mediates other proteininteractions, such as with cytoplasmic keratin 18 and the 5-HT7receptor [11,35]. Although the PCI domain of eIF3K may alsomediate protein-protein interactions, as has been seen for otherPCI domain-containing proteins [43], this domain was notrequired for association with aa 750–882 of PML-I in a yeasttwo-hybrid assay (Fig. 1B and C).

In the course of our experiments we also identified a novelhuman isoform of eIF3K, which we have designated as isoform 2(eIF3K-2). eIF3K-2 is encoded by an alternatively spliced mRNAthat results in an in-frame deletion of exon 2, and which can alsobe identified in cDNA libraries of humans and mice (Supplemen-tary information Fig. 2). In addition, analysis of expression taggedsequences (ESTs) within Genbank encoding eIF3K suggests theremay be several other isoforms of eIF3K that could vary in theirability to interact with PML.

We have shown by cell-free and tissue culture methods thateIF3K and PML can interact with each other. In vitro, eIF3K-1 andeIF3K-2 both interact with the C-terminal domain of PML-I (exon9, Fig. 1) and with full-length recombinant PML-I and PML-IV(Fig. 3). The interaction between aa 750–882 of PML-I and eIF3K islikely directly based on our in vitro pull-down assays usingpurified recombinant proteins (Fig. 2). In addition, we demon-strate that eIF3K and PML form protein complexes in vivo that canbe co-immunoprecipitated from human HEK-293 cell lysates(Figs. 3 and 5C).

Although both eIF3K isoforms studied here can interact withPML in vitro, we observed differential abilities to associate withPML NBs in cells. Expression of GFP-tagged PML-I or PML-IV inU2OS cells promoted an association of endogenous eIF3K withPML NBs (Fig. 4). Although eIF3K-1 appears to be the predominantisoform in these cells (Fig. 4E) it is not clear which endogenous

eIF3K isoform is detected at PML NBs. However, when specificisoforms of eIF3K were introduced in these cells, only eIF3K-2 wasfound to be strongly associated with PML NBs, and eIF3K-2 wasassociated with PML-I-enriched NBs to a greater degree thanPML-IV-enriched NBs. Our in vitro data also indicated that eIF3Kisoforms can interact with PML-I through the unique exon 9encoded C-terminus, and through another domain that might becommon between all PML isoforms. However, in cells PML-I ismore efficient at recruiting eIF3K-2 to PML NBs. The discrepancybetween the in vitro and cell line data may reflect the differentsecondary and tertiary structure of PML within nuclear bodies incells versus soluble PML used for in vitro studies. PML NBs formthrough SUMO-dependent heteromultimerization of nuclear bodyassociated proteins like DAXX and SP100 that are SUMOylatedwith several PML protein isoforms via their SUMO interactionmotif (SIM) domains encoded in exon 7 of isoforms I to VI[28,47,50]. In addition, multimerization of PML protein isoformsis driven by interactions between the RING finger B-box coiledcoil (RBCC) domain found in all PML isoforms [29]. The highlymultimeric state of PML within PML NBs is one factor in theirresistance to solubilization, which is in contrast to recombinantsoluble PML protein which is either monomeric or in a reducedmultimeric state, and is not associated with other proteins such asDAXX and SP100. Thus, PML protein within PML NBs is likely toform tertiary structures that may partly occlude protein interac-tions that might occur between eIF3K isoforms and the commonN-terminus (encoded by exons 1–7) that encodes the RBCC andSIM domains found in PML isoforms I to VI. As such, the uniqueC-terminus of PML-I may serve as a more accessible docking sitefor eIF3K isoforms in cells. Nonetheless, it remains unclear whyonly eIF3K-2, but not eIF3K-1, colocalizes to PML bodies since bothisoforms of eIF3K were able to interact with PML exon 9 in vitro.It is possible that the inclusion of exon 2 in eIF3K-1 could facilitatean interaction with a cellular factor, not present in our in vitrointeractions studies, that excludes it from interacting with PMLbodies. Alternately, secondary structure changes catalyzed byexclusion of peptide sequences encoded by exon 2 within eIF3K-2 may facilitate better interaction with PML-I within nuclearbodies. Another eIF3 subunit, eIF3E, also colocalizes with PML NBsindirectly via interaction with the Ret finger protein in anisoform-independent manner [39]. Therefore, a third possibility,although less likely given our in vitro data demonstrating directinteraction between eIF3K and the C-terminus of PML-I, is that anadapter protein might interact with eIF3K-2, but not eIF3K-1, tofacilitate association with PML-NBs. Alternatively, while eIF3K-2clearly showed stronger association with PML NBs than eIF3K-1when these proteins were ectopically expressed (Fig. 5), it ispossible that endogenous eIF3K-1 (the more abundant eIF3Kisoform) might also have an affinity for PML NBs when it isexpressed at endogenous levels (Fig. 4B). However, since only asmall proportion of total eIF3K is at PML bodies, it is possible thateIF3K-2, which represents a small proportion of total eIF3K, is stillthe major endogenous eIF3K isoform at PML bodies. This latterinterpretation would be consistent with the observation thateIF3K-2 is more strongly associated with PML NBs when ectopi-cally expressed.Finally, in response to various cellular stresses (e.g. DNA

damage) many proteins transiently interact with PML bodiesand post-translational modifications often regulate or result fromthese PML body interactions [5,13]. In the case of another eIF3


subunit, eIF3F, its localization to the nucleus is regulated viaphosphorylation by the cyclin dependent kinase 11 (CDK11p46)[49]. Although our in vitro data suggests that eIF3K interacts withPML in the absence of post-translational modifications, we cannotexclude the possibility that the association of eIF3K isoforms withPML NBs in cells may be regulated by post-translational mod-ification of eIF3K and/or the PML protein itself. Nonetheless, givenour novel finding that eIF3K-2 localizes to PML NBs in an isoform-dependent manner, it is tempting to speculate that alternativesplicing of eIF3K transcripts may regulate the localization of eIF3Kat PML NBs, for example between different cell types or undervarious physiological conditions that promote the alternativesplicing of eIF3K.The significance of the association of eIF3K with PML NBs is

unclear since the biological role(s) of eIF3K remain poorlycharacterized. In the cytoplasm, the main function of eIF3K ispresumably the regulation of translation initiation; however,eIF3K does not appear to be an essential subunit of the eIF3complex [36] and general translation does not appear to beaffected by eIF3K silencing [35]. As well, eIF3K, along with theeIF3E and eIF3F subunits, are not conserved in budding yeast[2,38,42]. Even in the cytoplasm, eIF3K has been reported to playother roles beyond translation, including regulation of apoptosis,through an interaction with keratin 18 [35], and 5-HT7 receptortrafficking and stability [11]. In addition, eIF3K has been shown tocolocalize with cyclin D3 in the cytoplasm [48] and dynein at sitesof cell–cell contact in epithelial cells [31]. Our study is the first todescribe eIF3K-2 as a novel isoform of eIF3K, and eIF3K-2 is thepredominant isoform associated with PML NBs within cells.Previous studies attributing various cellular functions to eIF3Khave investigated isoform 1 [11,48], an isoform that we havedemonstrated to be weakly associated with PML NBs in cells, orendogenous eIF3K, where the isoform(s) being studied areunclear [31,35]. Therefore, it will be important to determine whatfunctions are shared between eIF3K isoforms, and what specificfunctions can be attributed to eIF3K-2, before the functionalconsequences of the localization of this isoform to PML NBs canbe fully explored.

Materials and methods:

Plasmid construction

The human PML isoform I, exon 9 (PML-I-E9) sequence encodingamino acids 604–882 (GenBank Accession NM_033238.2) and asmaller fragment of the human PML-I-E9 sequence lacking theexonuclease III domain (amino acids (aa) 750–882) was clonedinto pGBT9 or pGBKT7 in frame with the coding region for theDNA binding domain of the yeast Gal4 transcription factor(Gal4DBD). Fragments of eIF3K were cloned into pACT2 (Clontech)in frame with the coding region for the activation domain of theGal4 transcription factor (Gal4AD). An alternative splice isoform ofeIF3K (eIF3K isoform 2, see Supplementary information Fig. 2)was isolated from total HeLa RNA extracts, reverse transcribedand cloned into the mammalian expression plasmid pcDNA3.1 (-)containing an influenza hemagglutinin (HA) epitope tag on theC-terminus (eIF3K-Iso2-HA). For expression of GST-eIF3K in E. coli, thecoding sequence for eIF3K was cloned into pGEX-5�1 vector (GE LifeSciences). Maltose binding protein (MBP)–PML-I recombinant plasmid

was created by cloning the sequence encoding PML-I-E9 lacking theexonuclease III domain (750–882 aa) into pMal-p4G vector (NewEngland Biolabs).

Yeast strains and media

Yeast strains and media—Standard methods were used for growthand manipulation of yeast and bacteria [8,46]. Yeast were culturedin YPDA (1% yeast extract (BioShop Canada), 2% Peptone (BioShopCanada), 0.003% adenine, 2% glucose), synthetic defined (0.67%yeast nitrogen base without amino acids (BioShop Canada), 2%glucose, 0.003% adenine, 0.002% uracil, and all required amino acids).Yeast were transformed by the LiAc/SS-DNA/PEG method [23].

Cell culture

HEK-293A, U2OS and HeLa cell lines used were cultured inDulbecco's modified Eagle's medium (Sigma) supplemented with10% fetal calf serum, 1% penicillin/streptomycin at 37 1C with 5%CO2. Human osteosarcoma (U2OS) cell lines stably expressinggreen fluorescent protein (GFP)–PML-I and GFP–PML-IV [18] weremaintained as above with the addition of 800 (PML-I) or 1600(PML-IV) mg/ml of G418.

Yeast two-hybrid assay

Yeast two-hybrid screens were performed according to instruc-tions described in the Matchmaker Gold yeast two-hybrid manual(Clontech). PML-I-E9 bait plasmid (pGBKT7) was introduced intoSaccharomyces cerevisiae PJ696 yeast two-hybrid reporter strainand mated with an isogenic yeast strain transformed with ahuman adult brain cDNA library cloned into pACT2 (Clontech).For domain mapping, plasmids directing expression of the indi-cated Gal4AD-tagged fragments of eIF3K and Gal4DBD-tagged PML-I(aa 750–882) were transformed into yeast strains Y2HGold andY187, respectively. Transformants were mated, and diploids wereisolated on selective medium lacking leucine and tryptophan andsubsequently patched to solid rich medium containing a limitingamount of adenine (YPDA).

Expression and purification of recombinant eIF3K and PML

GST-eIF3K and MBP–PML-I (aa 750–882) recombinant proteinswere expressed in Escherichia coli Rosetta (DE3) cells (Novagen),while His6-tagged full-length PML-I and PML-IV were expressedin insect cells.

For expression of GST-tagged eIF3K in E. coli, the codingsequence for eIF3K (isoform 1 or 2) was cloned into the GSTexpression vector pGEX-5X-1, and the plasmid transformed intoRosetta cells. Transformants were grown to OD600 0.5–1.0 in LBmedium and were induced with 0.1 mM IPTG for 3 h at 37 1C.Cells were harvested, aliquoted and stored at �80 1C. Cells(�80 μL wet pellet volume) were resuspended in 800 μL ofpurification buffer (50 mM Tris pH 8.0, 150 mM NaCl, 0.1%NP-40, 1 mM PMSF) containing 0.4 mg lysozyme and incubated onice for 20 min. Cell suspensions were sonicated for five rounds of10 s each at 30% output on a Misonex Sonicator 3000 andcentrifuged at 21 000g for 15 min at 4 1C. Protein concentration inthe supernatant was quantified by a Bradford assay and diluted to afinal volume of 750 μL in purification buffer. 60 μL of glutathione


Sepharose 4B beads (GE Life Sciences) were added to the solublefraction and the suspensions rocked for 30 min at room tempera-ture. Beads were washed three times in purification buffer.

For expression of full-length PML-I and PML-IV protein in insectcells, the Bac-to-Bac baculovirus expression system was usedaccording to the manufacturer's instructions (Life Technologies).The coding sequence for PML-I and PML-IV were cloned intoFastbac expression vectors and transformed into Max efficiencyDH10Bac competent cells (Life Technologies). Bacmids weretransfected into sf-9 cells with Cellfectin (Life Technologies) toproduce recombinant baculovirus particles. Virus was collectedfrom the culture medium and used to infect sf-9 cells for 48 h.Infected cell lysates were prepared by resuspending cell pellets inten volumes of insect cell lysis buffer (50 mM Tris pH 7.4, 250 mMsucrose, 150 mM NaCl, 0.1% Triton X-100, and 2 mM PMSF).Suspensions were sonicated for six rounds of 10 s at 30% output.Lysates were supplemented with 15% glycerol, aliquoted andstored at �80 1C. Suspensions were centrifuged at 21,000g for30 min. Protein concentration in the supernatant was quantifiedby a Bradford assay, and volumes of supernatants containingequal levels of PML-I and PML-IV were diluted to a final volume of400 μL in insect cell lysis buffer.

In vitro binding assays

In vitro affinity purification assays were performed by pre-clearing2 μg of protein lysates with glutathione Sepharose 4B or amyloseresin for GST and MBP tagged proteins, respectively and incubatedwith GST, GST-eIF3K or MBP and MBP–PML-I (750–882 aa) boundbeads for 3 h at 4 1C. The beads were washed three times with washbuffer (PBS supplemented with 0.1% Triton X-100 and 50 mM NaCl)and once with PBS. The co-purified proteins were eluted by boilingin 2� SDS loading buffer and analyzed by Western blotting.

For testing interaction of GST-eIF3K with His6-PML-I and His6-PML-IV, 400 μL of diluted insect cell lysate was added to 20 μL ofgluthione Sepharose beads containing bound GST or GST-eIF3Kand rocked for 1 h at room temperature. Beads were washed fourtimes for 5 min at room temperature in 1 mL purification buffercontaining 300 mM NaCl. Co-purified proteins were eluted byboiling for 5 min in 60 μL of protein gel loading buffer.

Co-immunoprecipitation assay and western blotting

For co-immunoprecipitation of endogenous proteins, untrans-fected human embryonic kidney (HEK-293T) cells were lysedice-cold lysis buffer (50 mM Tris–HCl pH 8, 300 mM NaCl, 0.25%Nonidet P-40, 1� protease inhibitor cocktail). Cleared lysateswere incubated overnight at 4 1C with mouse anti-eIF3K antibodythat was pre-bound to PureProteome™ Protein G magnetic beads(Millipore). The beads were washed 3 times with wash buffer(50 mM Tris–HCl pH 8, 150 mM NaCl, 0.1% Nonidet P-40, 5 mMEDTA) followed by one wash with PBS. The bound proteins wereeluted by boiling in 2� SDS sample buffer and analyzed byWestern blotting. For co-immunoprecipitation of tagged PMLand eIF3K, HEK-293A cells were transfected with N-terminalFLAG tagged PML-I and C-terminal HA-tagged eIF3K expressionplasmids using Lipofectamine 2000 (Life Technologies) followingthe manufacturer's instructions. At 48 h post transfection, cellswere lysed with ice-cold lysis buffer (50 mM Tris–HCl pH 8,150 mM NaCl, 0.5% Nonidet P-40, 1� protease inhibitor). Cleared

lysates were mixed with mouse monoclonal anti-FLAG M2magnetic beads (Sigma) or mouse IgG bound to PureProteome™

Protein G magnetic beads (Millipore) and incubated overnight at4 1C. Beads were washed four times with lysis buffer and elutedwith 2� protein sample buffer and boiling. Recovered proteinswere analyzed by SDS-PAGE and Western Blotting with rabbitanti-ECS (DDDDK) antibody (Bethyl Laboratories, A190-102A) andrabbit anti-eIF3K (Novus Biologicals, NB100-93304).

Immunofluorescence

Protein colocalization was visualized by fluorescence microscopy.Cells were plated onto coverslips in a 6-well plate and transfectedwith the relevant plasmids. Forty-eight hours post-transfection, cellswere washed with PBS and fixed in 3% paraformaldehyde. Cells werepermeabilized with 0.5% Triton X-100 in PBS, and blocked with 4%BSA in PBS. Cells were then immunolabeled with primary antibodies(1 h, diluted in blocking buffer), washed with PBS and incubatedwith fluorescently labeled secondary antibodies (45 min, diluted inblocking buffer). Following washes in PBS, cells were incubated with1 μg/mL of 4',6-diamidino-2-phenylindole (DAPI) stain to visualizeDNA. Fluorescent images were captured with Zeiss Axiovert 200MMicroscope and Hamamatsu Orca R2 Camera or with Zeiss CellObserver Microscope under a 63� immersion oil objective lens.Images were processed using only linear adjustments (e.g. bright-ness/contrast) with Slidebook (Intelligent Imaging Innovations,Boulder, CO) and Adobe Photoshop CS5. Quantification of eIF3Kcolocalization with PML nuclear bodies was determined by count-ing 50–100 eIF3K positive cells per experimental condition andcalculating the percentage of cells with no, partial or completecolocalization between eIF3K and PML nuclear bodies. The meansand standard errors from at least three independent experimentswere calculated and statistical significance was determined usingthe student's t-test in Microsoft Excel 2007.

Acknowledgments

This research was funded by an operating grant to G.D. from theCanadian Institutes of Health Research (CIHR, MOP-84260). J.S.,D.C. and J.P. are supported by trainee awards from the BeatriceHunter Cancer Research Institute with funds provided as part ofthe Terry Fox Foundation Strategic Health Research TrainingProgram in Cancer Research at CIHR. B.T. received a studentshipfrom the Nova Scotia Health Research Foundation.

Appendix A. Supporting information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.yexcr.2013.09.001.

r e f e r e n c e

[1] M.A. Andrade, C. Petosa, S.I. O'Donoghue, C.W. Muller, P. Bork,Comparison of ARM and HEAT protein repeats, J. Mol. Biol. 309(2001) 1–18.

[2] K. Asano, T.G. Kinzy, W.C. Merrick, J.W. Hershey, Conservationand diversity of eukaryotic translation initiation factor eIF3, J.Biol. Chem. 272 (1997) 1101–1109.

dx.doi.org/doi:10.1016/j.yexcr.2013.09.001http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref1http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref1http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref1http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref2http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref2http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref2


[3] C.A. Ascoli, G.G. Maul, Identification of a novel nuclear domain,J. Cell Biol. 112 (1991) 785–795.

[4] R. Bernardi, P.P. Pandolfi, Role of PML and the PML-nuclear bodyin the control of programmed cell death, Oncogene 22 (2003)9048–9057.

[5] R. Bernardi, P.P. Pandolfi, Structure, dynamics and functions ofpromyelocytic leukaemia nuclear bodies, Nat. Rev. Mol. Cell Biol.8 (2007) 1006–1016.

[6] K.L. Borden, RING fingers and B-boxes: zinc-binding protein-proteininteraction domains, Biochem. Cell Biol. 76 (1998) 351–358.

[7] K.S. Browning, D.R. Gallie, J.W. Hershey, A.G. Hinnebusch,U. Maitra, W.C. Merrick, C. Norbury, Unified nomenclature for thesubunits of eukaryotic initiation factor 3, Trends Biochem. Sci. 26(2001) 284.

[8] D. Burke, D. Dawson, T. Stearns, Methods in Yeast Genetics: ACold Spring Harbor Laboratory Course Manual, Cold SpringHarbor Laboratory Press, Woodbury, New York, 2000.

[9] W. Condemine, Y. Takahashi, M. Le Bras, H. de Thé, A nucleolartargeting signal in PML-I addresses PML to nucleolar caps instressed or senescent cells, J. Cell Sci. 120 (2007) 3219–3227.

[10] W. Condemine, Y. Takahashi, J. Zhu, F. Puvion-Dutilleul,S. Guegan, A. Janin, H. de Thé, Characterization of endogenoushuman promyelocytic leukemia isoforms, Cancer Res. 66 (2006)6192–6198.

[11] K. De Martelaere, B. Lintermans, G. Haegeman, P. Vanhoenacker,Novel interaction between the human 5-HT7 receptor isoformsand PLAC-24/eIF3k, Cell. Signal. 19 (2007) 278–288.

[12] H. de Thé, C. Lavau, A. Marchio, C. Chomienne, L. Degos, A.Dejean, The PML-RAR alpha fusion mRNA generated by the t(15;17) translocation in acute promyelocytic leukemia encodes afunctionally altered RAR, Cell 66 (1991) 675–684.

[13] G. Dellaire, D.P. Bazett-Jones, PML nuclear bodies: dynamicsensors of DNA damage and cellular stress, Bioessays 26 (2004)963–977.

[14] G. Dellaire, R.W. Ching, H. Dehghani, Y. Ren, D.P. Bazett-Jones,The number of PML nuclear bodies increases in early S phase by afission mechanism, J.Cell Sci. 119 (2006) 1026–1033.

[15] G. Dellaire, R. Farrall, W.A. Bickmore, The nuclear protein database(NPD): sub-nuclear localisation and functional annotation of thenuclear proteome, Nucleic Acids Res. 31 (2003) 328–330.

[16] G. Dellaire, E.M. Makarov, J.J. Cowger, D. Longman, H.G. Suther-land, R. Luhrmann, J. Torchia, W.A. Bickmore, Mammalian PRP4kinase copurifies and interacts with components of both the U5snRNP and the N-CoR deacetylase complexes, Mol. Cell Biol. 22(2002) 5141–5156.

[17] C. Desbois, R. Rousset, F. Bantignies, P. Jalinot, Exclusion of Int-6from PML nuclear bodies by binding to the HTLV-I Tax onco-protein, Science 273 (1996) 951–953.

[18] C.H. Eskiw, G. Dellaire, D.P. Bazett-Jones, Chromatin contributesto structural integrity of promyelocytic leukemia bodies througha SUMO-1-independent mechanism, J. Biol. Chem. 279 (2004)9577–9585.

[19] R.D. Everett, P. Lomonte, T. Sternsdorf, R. van Driel, A. Orr, Cellcycle regulation of PML modification and ND10 composition, J.Cell Sci. 112 (Pt 24) (1999) 4581–4588.

[20] M. Fagioli, M. Alcalay, P.P. Pandolfi, L. Venturini, A. Mencarelli,A. Simeone, D. Acampora, F. Grignani, P.G. Pelicci, Alternativesplicing of PML transcripts predicts coexpression of severalcarboxy-terminally different protein isoforms, Oncogene 7(1992) 1083–1091.

[21] V. Fogal, M. Gostissa, P. Sandy, P. Zacchi, T. Sternsdorf, K. Jensen,P.P. Pandolfi, H. Will, C. Schneider, S.G. Del, Regulation of p53activity in nuclear bodies by a specific PML isoform, EMBO J. 19(2000) 6185–6195.

[22] P.S. Freemont, The RING finger. A novel protein sequencemotif related to the zinc finger, Ann. N. Y. Acad. Sci. 684 (1993)174–192.

[23] R.D. Gietz, R.H. Schiestl, A.R. Willems, R.A. Woods, Studies on thetransformation of intact yeast cells by the LiAc/SS-DNA/PEGprocedure, Yeast 11 (1995) 355–360.

[24] A. Guo, P. Salomoni, J. Luo, A. Shih, S. Zhong, W. Gu, P.P. Pandolfi,The function of PML in p53-dependent apoptosis, Nat. Cell Biol. 2(2000) 730–736.

[25] C. Gurrieri, P. Capodieci, R. Bernardi, P.P. Scaglioni, K. Nafa,L.J. Rush, D.A. Verbel, C. Cordon-Cardo, P.P. Pandolfi, Loss of thetumor suppressor PML in human cancers of multiple histologicorigins, J. Natl. Cancer Inst. 96 (2004) 269–279.

[26] B.R. Henderson, A. Eleftheriou, A comparison of the activity,sequence specificity, and CRM1-dependence of different nuclearexport signals, Exp. Cell Res. 256 (2000) 213–224.

[27] J.W. Hershey, K. Asano, T. Naranda, H.P. Vornlocher, P. Hanachi,W.C. Merrick, Conservation and diversity in the structure oftranslation initiation factor EIF3 from humans and yeast,Biochimie 78 (1996) 903–907.

[28] A.M. Ishov, A.G. Sotnikov, D. Negorev, O.V. Vladimirova, N. Neff,T. Kamitani, E.T. Yeh, J.F. Strauss III, G.G. Maul, PML is critical forND10 formation and recruits the PML-interacting protein daxx tothis nuclear structure when modified by SUMO-1, J.Cell Biol. 147(1999) 221–234.

[29] K. Jensen, C. Shiels, P.S. Freemont, PML protein isoforms and theRBCC/TRIM motif, Oncogene 20 (2001) 7223–7233.

[30] T. Kamitani, K. Kito, H.P. Nguyen, H. Wada, T. Fukuda-Kamitani,E.T. Yeh, Identification of three major sentrinization sites in PML,J. Biol. Chem. 273 (1998) 26675–26682.

[31] S. Karki, L.A. Ligon, J. DeSantis, M. Tokito, E.L. Holzbaur, PLAC-24is a cytoplasmic dynein-binding protein that is recruited to sitesof cell–cell contact, Mol. Biol. Cell 13 (2002) 1722–1734.

[32] M.H. Koken, F. Puvion-Dutilleul, M.C. Guillemin, A. Viron,G. Linares-Cruz, N. Stuurman, L. de Jong, C. Szostecki, F. Calvo,C. Chomienne, The t(15;17) translocation alters a nuclear body ina retinoic acid-reversible fashion, EMBO J. 13 (1994) 1073–1083.

[33] H.K. Lai, K.L. Borden, The promyelocytic leukemia (PML) proteinsuppresses cyclin D1 protein production by altering the nuclearcytoplasmic distribution of cyclin D1 mRNA, Oncogene 19 (2000)1623–1634.

[34] H.K. Lin, S. Bergmann, P.P. Pandolfi, Cytoplasmic PML function inTGF-beta signaling, Nature 431 (2004) 205–211.

[35] Y.M. Lin, Y.R. Chen, J.R. Lin, W.J. Wang, A. Inoko, M. Inagaki,Y.C. Wu, R.H. Chen, eIF3k Regulates apoptosis in epithelial cellsby releasing caspase 3 from keratin-containing inclusions, J.CellSci. 121 (2008) 2382–2393.

[36] M. Masutani, N. Sonenberg, S. Yokoyama, H. Imataka, Reconsti-tution reveals the functional core of mammalian eIF3, EMBO J. 26(2007) 3373–3383.

[37] G.G. Maul, Nuclear domain 10, the site of DNA virus transcriptionand replication, Bioessays 20 (1998) 660–667.

[38] G.L. Mayeur, C.S. Fraser, F. Peiretti, K.L. Block, J.W. Hershey,Characterization of eIF3k: a newly discovered subunit of mam-malian translation initiation factor elF3, Eur. J. Biochem. 270(2003) 4133–4139.

[39] C. Morris-Desbois, V. Bochard, C. Reynaud, P. Jalinot, Interactionbetween the Ret finger protein and the Int-6 gene product andco-localisation into nuclear bodies, J. Cell Sci. 112 (Pt 19) (1999)3331–3342.

[40] D. Negorev, G.G. Maul, Cellular proteins localized at and inter-acting within ND10/PML nuclear bodies/PODs suggest functionsof a nuclear depot, Oncogene 20 (2001) 7234–7242.

[41] L.A. Nguyen, P.P. Pandolfi, Y. Aikawa, Y. Tagata, M. Ohki,I. Kitabayashi, Physical and functional link of the leukemia-associated factors AML1 and PML, Blood 105 (2005) 292–300.

[42] L. Phan, X. Zhang, K. Asano, J. Anderson, H.P. Vornlocher,J.R. Greenberg, J. Qin, A.G. Hinnebusch, Identification of atranslation initiation factor 3 (eIF3) core complex, conserved inyeast and mammals, that interacts with eIF5, Mol. Cell. Biol. 18(1998) 4935–4946.

http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref3http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref3http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref4http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref4http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref4http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref5http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref5http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref5http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref6http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref6http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref7http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref7http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref7http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref7http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref8http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref8http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref8http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref9http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref9http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref9http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref10http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref10http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref10http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref10http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref11http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref11http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref11http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref12http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref12http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref12http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref12http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref13http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref13http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref13http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref14http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref14http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref14http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref15http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref15http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref15http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref16http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref16http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref16http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref16http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref16http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref17http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref17http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref17http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref18http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref18http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref18http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref18http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref19http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref19http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref19http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref20http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref20http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref20http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref20http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref20http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref21http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref21http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref21http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref21http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref22http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref22http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref22http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref23http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref23http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref23http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref24http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref24http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref24http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref25http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref25http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref25http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref25http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref26http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref26http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref26http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref27http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref27http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref27http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref27http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref28http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref28http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref28http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref28http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref28http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref29http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref29http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref30http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref30http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref30http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref31http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref31http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref31http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref32http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref32http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref32http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref32http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref33http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref33http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref33http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref33http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref34http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref34http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref35http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref35http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref35http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref35http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref36http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref36http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref36http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref37http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref37http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref38http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref38http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref38http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref38http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref39http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref39http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref39http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref39http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref40http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref40http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref40http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref41http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref41http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref41http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref42http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref42http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref42http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref42http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref42


[43] E. Pick, K. Hofmann, M.H. Glickman, PCI complexes: beyond theproteasome, CSN, and eIF3 Troika, Mol. Cell 35 (2009) 260–264.

[44] A. Reymond, G. Meroni, A. Fantozzi, G. Merla, S. Cairo, L. Luzi, D.Riganelli, E. Zanaria, S. Messali, S. Cainarca, A. Guffanti,S. Minucci, P.G. Pelicci, A. Ballabio, The tripartite motif familyidentifies cell compartments, EMBO J. 20 (2001) 2140–2151.

[45] P. Salomoni, P.P. Pandolfi, The role of PML in tumor suppression,Cell 108 (2002) 165–170.

[46] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: ALaboratory Manual, Cold Spring Harbor Laboratory Press, 1989.

[47] T.H. Shen, H.K. Lin, P.P. Scaglioni, T.M. Yung, P.P. Pandolfi, Themechanisms of PML-nuclear body formation, Mol. Cell 24 (2006)331–339.

[48] X. Shen, Y. Yang, W. Liu, M. Sun, J. Jiang, H. Zong, J. Gu,Identification of the p28 subunit of eukaryotic initiation factor 3(eIF3k) as a new interaction partner of cyclin D3, FEBS Lett. 573(2004) 139–146.

[49] J. Shi, J.W. Hershey, M.A. Nelson, Phosphorylation of the eukar-yotic initiation factor 3f by cyclin-dependent kinase 11 duringapoptosis, FEBS Lett. 583 (2009) 971–977.

[50] T. Sternsdorf, K. Jensen, B. Reich, H. Will, The nuclear dot proteinsp100, characterization of domains necessary for dimerization,subcellular localization, and modification by small ubiquitin-likemodifiers, J. Biol. Chem. 274 (1999) 12555–12566.

[51] T. Sternsdorf, K. Jensen, H. Will, Evidence for covalent modifica-tion of the nuclear dot-associated proteins PML and Sp100 byPIC1/SUMO-1, J. Cell Biol. 139 (1997) 1621–1634.

[52] R.L. Strausberg, E.A. Feingold, L.H. Grouse, J.G. Derge,R.D. Klausner, F.S. Collins, L. Wagner, C.M. Shenmen, G.D. Schuler,S.F. Altschul, B. Zeeberg, K.H. Buetow, C.F. Schaefer, N.K. Bhat,R.F. Hopkins, H. Jordan, T. Moore, S.I. Max, J. Wang, F. Hsieh,L. Diatchenko, K. Marusina, A.A. Farmer, G.M. Rubin, L. Hong,M. Stapleton, M.B. Soares, M.F. Bonaldo, T.L. Casavant,T.E. Scheetz, M.J. Brownstein, T.B. Usdin, S. Toshiyuki, P. Carninci,C. Prange, S.S. Raha, N.A. Loquellano, G.J. Peters, R.D. Abramson,S.J. Mullahy, S.A. Bosak, P.J. McEwan, K.J. McKernan, J.A. Malek,P.H. Gunaratne, S. Richards, K.C. Worley, S. Hale, A.M. Garcia,

L.J. Gay, S.W. Hulyk, D.K. Villalon, D.M. Muzny, E.J. Sodergren,X. Lu, R.A. Gibbs, J. Fahey, E. Helton, M. Ketteman, A. Madan,S. Rodrigues, A. Sanchez, M. Whiting, A. Madan, A.C. Young,Y. Shevchenko, G.G. Bouffard, R.W. Blakesley, J.W. Touchman,E.D. Green, M.C. Dickson, A.C. Rodriguez, J. Grimwood,J. Schmutz, R.M. Myers, Y.S. Butterfield, M.I. Krzywinski,U. Skalska, D.E. Smailus, A. Schnerch, J.E. Schein, S.J. Jones,M.A. Marra, Generation and initial analysis of more than 15,000full-length human and mouse cDNA sequences, Proc. Natl. Acad.Sci. USA 99 (2002) 16899–16903.

[53] T. Ulbricht, M. Alzrigat, A. Horch, N. Reuter, A. von MikeczV. Steimle, E. Schmitt, O.H. Kramer, T. Stamminger, P. Hemmerich,PML promotes MHC class II gene expression by stabilizing the classII transactivator, J.Cell Biol. 199 (2012) 49–63.

[54] E. Van Damme, K. Laukens, T.H. Dang, X. Van Ostade, A manuallycurated network of the PML nuclear body interactome reveals animportant role for PML-NBs in SUMOylation dynamics, Int. J. Biol.Sci. 6 (2010) 51–67.

[55] M. Vernier, V. Bourdeau, M.F. Gaumont-Leclerc, O. Moiseeva,V. Begin, F. Saad, A.M. Mes-Masson, G. Ferbeyre, Regulation ofE2Fs and senescence by PML nuclear bodies, Genes Dev. 25(2011) 41–50.

[56] Z.G. Wang, L. Delva, M. Gaboli, R. Rivi, M. Giorgio, C. Cordon-Cardo, F. Grosveld, P.P. Pandolfi, Role of PML in cell growth andthe retinoic acid pathway, Science 279 (1998) 1547–1551.

[57] Z. Wei, P. Zhang, Z. Zhou, Z. Cheng, M. Wan, W. Gong, Crystalstructure of human eIF3k, the first structure of eIF3 subunits,J. Biol. Chem. 279 (2004) 34983–34990.

[58] J. Yu, J. Lan, C. Wang, Q. Wu, Y. Zhu, X. Lai, J. Sun, C. Jin, H. Huang,PML3 interacts with TRF1 and is essential for ALT-associated PMLbodies assembly in U2OS cells, Cancer Lett. 291 (2010) 177–186.

[59] S. Zhong, P. Hu, T.Z. Ye, R. Stan, N.A. Ellis, P.P. Pandolfi, A role forPML and the nuclear body in genomic stability, Oncogene 18(1999) 7941–7947.

[60] S. Zhong, P. Salomoni, P.P. Pandolfi, The transcriptional role ofPML and the nuclear body, Nat.Cell Biol. 2 (2000) E85–E90.

http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref43http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref43http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref44http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref44http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref44http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref44http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref45http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref45http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref46http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref46http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref47http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref47http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref47http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref48http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref48http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref48http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref48http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref49http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref49http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref49http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref50http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref50http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref50http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref50http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref51http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref51http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref51http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref52http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref53http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref53http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref53http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref53http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref54http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref54http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref54http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref54http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref55http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref55http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref55http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref55http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref56http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref56http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref56http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref57http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref57http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref57http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref58http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref58http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref58http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref59http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref59http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref59http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref60http://refhub.elsevier.com/S0014-4827(13)00382-0/sbref60

The translation initiation factor 3 subunit eIF3K interacts with PML and associates with PML nuclear bodiesIntroductionResultseIF3K interacts with a peptide encoded by exon 9 of PML-IPML and eIF3K interact in vitro and in vivoeIF3K colocalizes with PML NBs

DiscussionMaterials and methods:Plasmid constructionYeast strains and mediaCell cultureYeast two-hybrid assayExpression and purification of recombinant eIF3K and PMLIn vitro binding assaysCo-immunoprecipitation assay and western blottingImmunofluorescence

AcknowledgmentsSupporting informationReference

the translation initiation factor 3 subunit eif3k interacts with pml … · 2016. 1. 11. · pml...

Documents