Journal of General Virology

Proteasome inhibitor induces nucleolar translocation of Epstein–Barr virus-encoded EBNA-5

Katja Pokrovskaja, Karin Mattsson, Elena Kashuba, George Klein and Laszlo Szekely

Microbiology and Tumor Biology Center, Karolinska Institute, S-171 77 Stockholm, Sweden

We have previously shown that Epstein–Barr virus (EBV)-encoded EBNA-5 is localized to PML bodies (PODs) in EBV-immortalized lymphoblastoid cell lines (LCLs). Here we have extended our study of the subnuclear localization of EBNA-5 and found a strict co-localization with PML in LCLs and in BL lines with an immunoblastic, LCL-like phenotype. Moreover, GFP–EBNA-5 accumulated in PML bodies upon transfection into LCLs. In contrast, transfection of cell lines of non-immunoblastic origin with an EBNA-5 expression construct showed preferential localization of the protein to the nucleoplasm. Since PML is involved in proteasome-dependent protein degradation, we investigated the total levels and sub-cellular localization of EBNA-5 upon inhibition of proteasome activity. We found that a proteasome inhibitor, MG132, induced the translocation of both endogenous and transfected EBNA-5 to the nucleoli in every cell line tested. The total EBNA-5 protein levels were not affected by the proteasomal block. EBNA-5 forms complexes with heat shock protein Hsp70. The proteasome inhibitor induced a rise in total levels of Hsp70 and dramatically changed its homogeneous nuclear and cytoplasmic distribution into nucleolar and cytoplasmic. This effect was EBNA-5-independent. The nucleolar localization of Hsp70 was enhanced by the presence of EBNA-5, however. EBNA-5 also enhanced the nucleolar translocation of a mutant p53 in a colon cancer line, SW480, treated with MG132. The coordinated changes in EBNA-5 and Hsp70 localization and the effect of EBNA-5 on mutant p53 distribution upon MG132 treatment might reflect the involvement of EBNA-5 in the regulation of intracellular protein trafficking associated with the proteasome-mediated degradation.

Epstein–Barr virus (EBV) is the most potent transforming virus known. EBV infection of resting B cells in vitro induces blast transformation and leads to the establishment of immortal lymphoblastoid cell lines (LCLs). This process is associated with the expression of nine latent viral proteins: the nuclear antigens, EBNA1–6, and three membrane proteins, LMP-1, -2a and -2b. Six of these proteins are required for immortalization: EBNA-1, -2, -3 (3A), -5 (LP), -6 (3C) and LMP-1 (Rickinson & Kieff, 1996).

EBNA-5 and EBNA-2 are the first virally encoded proteins expressed after B cell infection (Alfiery et al., 1991). They can drive gp340-activated primary B cells into the G₁ phase of the cell cycle (Sinclair et al., 1994). This suggests that EBNA-5 may play a role at the first steps of EBV transformation. EBNA-5 can co-operate with EBNA-2 in the activation of LMP-1 and Cp viral promoters (Harada & Kieff, 1997). The co-transfection of EBNA-3 and EBNA-5 into the DG75 line showed that EBNA-3 is tethered to the nuclear matrix fraction in the presence of EBNA-5 (Cludts & Farrell, 1998). This suggested that EBNA-5 may modify the intranuclear sorting of proteins.

After EBV infection of B cells, EBNA-5 is homogeneously distributed in the nucleus during the first 2 days. Later it accumulates in PML bodies or PODs (PML oncogenic domains; Szekely et al., 1995 b, 1996). These are distinct nuclear domains associated with the nuclear matrix. PML, in complex with SUMO-1, is required to form the POD structure (Ishov et al., 1999; Zhong et al., 2000 a). PODs are specifically disrupted in human acute promyelocytic leukaemia cells where PML is fused to a retinoic acid receptor alpha gene. A number of cellular proteins are localized to the PODs: SP100, INT6, CBP/p300, Hsp70, a fraction of Rb, Daxx and SUMO-1 (for review see Zhong et al., 2000 b). Herpes simplex virus type 1 infection abrogates the modification of PML by SUMO-1 (Muller & Dejean, 1999), leading to rapid PML protein degradation (Chelbi-Alix & de The, 1999) and disruption of the PODs. The EBV growth and transformation-associated EBNA-5 localizes to the PODs without disrupting them (Szekely et al., 1996).

PML expression is induced by interferons. Cellular response to interferon requires normal PML function (Quignon et al., 1998; Wang et al., 1998). PML was recently shown to regulate MHC expression in untransformed fibroblasts and to induce the expression of the proteins involved in antigen processing and presentation (Zheng et al., 1998). The ubiquitin–proteasome system is involved in the processing of MHC class I antigens, providing a link between the cellular degradation machinery and PML. The ubiquitin–proteasome system is the major pathway of selective protein degradation in eukaryotic cells. Initially, the target proteins are conjugated to the polypeptide ubiquitin through a lysine residue on the proteins. In the second step, the ubiquitin-conjugated proteins are recognized by the 26S proteasome and degraded (for review see Ciechanover, 1998). It is likely that nuclear POD structures are involved in this process. A protein genetically modified for rapid degradation can accumulate in the ubiquitinated form in the PODs upon proteasome inhibitor treatment. This was accompanied by the attraction of the proteasomes to the PODs, suggesting that the PODs may represent an intermediate reservoir for the ubiquitinated proteins targeted for degradation (Anton et al., 1999).

In an attempt to understand the role of EBNA-5 targeting to the PML bodies and in proteasome-mediated protein degradation, we monitored the changes in subcellular localization and in the total EBNA-5 levels upon proteasome inhibitor treatment. We have also continued to analyse the subnuclear localization of EBNA-5 in different cell types and found that EBNA-5 targets preferentially the PODs in the LCLs and in the BLs with an LCL-like phenotype. Nucleoplasmic localization was found in other cell types. Independently of EBNA-5 localization to the PODs or to the nucleoplasm, EBNA-5 translocated to the nucleoli when the cells were treated with proteasome inhibitor MG132.

Cell lines, culture conditions and transfections. All cell lines were grown in Iscove's cell culture medium supplemented with 10 % heat-inactivated FBS, 2 mM L-glutamine, 100 U/ml penicillin and 100 U/ml streptomycin. The cells were passaged every fourth day 1:5. Cultures were regularly tested for the absence of mycoplasma. Transfections were done using Lipofectamine Plus reagent (GibcoBRL) according to the manufacturer's instructions. MCF-7, a breast carcinoma line bearing wild-type p53, was obtained from M. Oren, The Weizmann Institute of Science, Israel. SW480 is a colorectal cancer line bearing mutant p53 (Arg to His 273 and Pro to Ser 309; Abarzua et al., 1995 and references therein). DG75 is an EBV-negative BL line (Ben-Bassat et al., 1977). IB-4 is an EBV-transformed LCL (King et al., 1980).

Clones of HeLa, MCF-7 and SW480 lines stably expressing EBNA-5 from a pBabe-EBNA-5 construct were generated using selection with 1 µg/ml puromycin (Sigma). Proteasome inhibitor MG132 was purchased from Calbiochem, diluted in DMSO and used at concentrations of 5–20 µM for the 6–16 h treatment of cells. The control cells were incubated with the same amounts of DMSO. GFP–EBNA-5 was made by cloning an EBNA-5-encoding BamHI–EcoRI fragment from pBabe-EBNA-5, containing four W repeats and the unique C-terminal region, into BglII–EcoRI-cleaved pEGFP-C1 (Clontech).

Antibodies, immunostainings and Western blotting. The following antibodies were used in this study for immunofluorescence and Western blotting: MAb D07 against p53 (DACO); MAb JF186 against EBNA-5 (Finke et al., 1987) and biotinylated JF186 (GibcoBRL biotinylation kit was used); MAb Ab-1 against Hsp72/73 (Oncogene Science); MAb pe2 against EBNA-2 (a gift from M. Rowe, University of Wales College of Medicine, UK); and MAb CS1-4 against LMP-1 (DACO). MAb against B23/nucleophosmin was a gift from P. K. Chan, Baylor College of Medicine, Houston, USA; rabbit serum against PML was a gift from H. de The, Institut d'Hematologie de l'Universite Paris VII, Paris, France. Cells were fixed in cold methanol–acetone, rehydrated for 30 min, stained with primary MAb for 1 h followed by three washes in PBS, incubated with secondary FITC or Texas red-conjugated antibodies, washed three times and mounted with 80 % glycerol solution in PBS containing 2.5 % 1,4-diazabicyclo-(2.2.2)octane (Sigma). Bisbenzimide (Hoechst 33258) was added at a concentration of 0.4 µg/ml to the secondary antibody for DNA staining. The double stainings for Hsp70 and EBNA-5 were done as follows: Ab-1 against Hsp70/rabbit-anti-mouse FITC-conjugated (DACO)/normal mouse serum/biotinylated JF186/Texas red-conjugated streptavidin (Vector). FITC-conjugated goat-anti-mouse and Texas red-conjugated goat-anti-rabbit (Vector) secondary antibodies were used for double EBNA-5/PML stainings.

The images were recorded on a DAS microscope Leitz DM RB with a Hamamatsu dual mode cooled CCD camera C4880. Alternatively, a fluorescence 3D microscope was used to reconstitute the image from a series of optical sections that were deblurred using the nearest neighbour algorithm with the help of the computer program TROOPER3, developed by us (Holmvall & Szekely, 1999). The original graphics files are available as 8 and 24 bit uncompressed TIFF images at our anonymous FTP site at ftp://130.237.124.100/E5LLL.

Total cell extracts were prepared by direct lysis in a hot Laemmli buffer. Proteins were separated on 10 % SDS–PAGE followed by transfer to nitrocellulose membrane (Schleicher & Schuell). Blocking and incubation with antibodies were performed in 5 % milk in PBS and 0.2 % Tween 20. Immunodetection using the ECL system (Amersham) was performed according to the manufacturer's instructions.

Both endogenously and exogenously expressed EBNA-5 co-localize with PML in the LCLs

We have previously shown that EBNA-5 co-localized with PML in LCL IB-4 and after EBV infection of peripheral B cells (Szekely et al., 1996). In order to see whether ectopically expressed EBNA-5 targets the same nuclear compartment, we have created a construct expressing a GFP–EBNA-5 fusion protein that can be directly detected in cells. LCL IARC171 was transfected with this construct by electroporation. IARC171 expressed low amounts of endogenous EBNA-5, detectable in 1–3 % of cells by immunostaining, localized to both nucleoplasm and the PODs (Fig. 1 A, upper row). Staining of the cells 2 days after GFP–EBNA-5 transfection with anti-PML MAb showed that GFP–EBNA-5 co-localized with PML in IARC171 (Fig. 1 B, upper row).

Fig. 1. Differential subnuclear localization of EBNA-5 in cell lines of different origin. (A) Double-staining of endogenously and exogenously expressed EBNA-5 and PML. LCL IARC171 expresses endogenous EBNA-5 that localized to both PML bodies and the nucleoplasm (upper row). BL line, DG75, and breast cancer line, MCF-7, were stably transfected with pBabe-EBNA-5. EBNA-5 was homogeneously distributed in the nucleoplasm and did not co-localize with PML (middle and lower row, respectively). (B) IARC171 (upper row), DG75 (middle row) and MCF-7 (lower row) were transfected with GFP–EBNA-5 and stained for PML 2 days later. GFP–EBNA-5 targeted PML bodies in IARC171, but not in DG75 or MCF-7. Red, PML; green, EBNA-5; blue, nuclei.

Partial co-localization of EBNA-5 and PML in B and non-B cells

BL DG75 was transfected with the pBabe-EBNA-5 construct and a puromycin-resistant pool was tested for the pattern of EBNA-5 expression. The protein localized to the nucleoplasm (Fig. 1 A, middle row). Transient transfection of BL lines DG75 and BL 28 with GFP–EBNA-5 showed nuclear localization and only seldom targeting of the PML bodies (Fig. 1 B, middle row and not shown). The EBV-positive BL lines cultured in vitro change their original, BL tumour phenotype to an immunoblastic, LCL-like phenotype, and activate the expression of EBNAs and LMPs (Rowe et al., 1987). In one such line, Namalwa, EBNA-5 was localized to both nucleoplasm and PODs (not shown).

We have established clones of SW480 (clones S2, S8, S9) and MCF-7 (clones M1, M2, M3 and M5) that ectopically express EBNA-5. The clones were double-stained for PML and EBNA-5. EBNA-5 was homogeneously distributed in the nucleoplasm, avoiding the nucleoli in all lines mentioned, and did not co-localize with PML. This is shown for MCF-7 clone M5 (Fig. 1 A, lower row) and MCF-7 transiently transfected with GFP–EBNA-5 (Fig. 1 B, lower row). Transiently transfected GFP–EBNA-5 often accumulates in conglomerates of irregular shape, seen in one of the cells in Fig. 1(B), which do not co-localize with PML.

EBNA-5 translocates to the nucleoli after treatment with proteasome inhibitor MG132

Clones M1, M2 and M3 of MCF-7 expressed EBNA-5 at moderate levels and showed primarily homogeneous nuclear staining as shown in Fig. 2(A, panel a) for M1. Treatment with 20 µM of MG132 for 6 h led to a nearly complete translocation of EBNA-5 to the nucleoli (Fig. 2 A, panel c). The phase contrast fields are shown in Fig. 2(A, panels b and d) to visualize the cells and the nucleoli. In order to prove the localization of EBNA-5 to the nucleolus, we double-stained M1 cells for EBNA-5 and nucleophosmin/B23 after MG132 treatment. Fig. 2(B) shows that EBNA-5 is localized deep inside the nucleoli and is surrounded by the B23 protein. The pattern of B23 staining was not changed by MG132 treatment (not shown).

Fig. 2. Exogenously expressed EBNA-5 translocated to the nucleoli in MG132-treated MCF-7 cells. (A) MCF-7 derivative clone M1 stably transfected with the pBabe-EBNA-5 construct cultured in the presence of DMSO (a, b) or 20 µM MG132 (c, d) for 6 h. Panels b and d represent phase contrast fields of panels a and c, respectively. (B) High magnification image of the double-staining for B23 (a, green) and EBNA-5 (b, red) of clone M1 treated with 20 µM of MG132. Overlap of green and red fluorescence is shown in panel c. Panel d shows phase contrast.

LCL IB-4 expressed endogenous EBNA-5 that was co-localized with PML (Szekely et al., 1996) and Fig. 3(A, left panels). EBNA-5 and PML partially dissociated and the former accumulated in the nucleoli after treatment of IB-4 with 5 µM of MG132 for 6 h (Fig. 3 A, right panels and C for EBNA-5 and phase contrast). We have also tested the localization of other EBV-encoded proteins and found that neither EBNA-2 (Fig. 3 B) nor EBNA-6 (not shown) nor LMP-1 (Fig. 3 B) changed their localization after MG132 treatment.

Fig. 3. EBNA-5 dissociates from PML bodies and co-localizes with Hsp70 in the nucleoli after MG132 treatment of LCL IB-4. (A) DMSO (left panel) and 20 µM MG132-treated (right panel) IB-4 cells were double-stained for EBNA-5 (red) and PML (green). The overlap shows the co-localization between the two proteins (yellow). The image is reconstituted from a series of 17 optical sections. (B) Neither EBNA-2 nor LMP-1 change their localization after MG132 treatment. (C) MG132-treated IB-4 cells were double-stained for EBNA-5 (red) and Hsp70 (green). The overlap shows the co-localization between the two proteins (yellow) in both nucleoli and nuclear dots. Phase contrast visualizes nucleoli.

Hsp70 co-localizes with EBNA-5 and accumulates in the nucleoli of MG132-treated cells

We have shown previously that EBNA-5 co-localized with constitutively expressed Hsp70 in IB-4 cells under normal conditions and upon heat shock (Szekely et al., 1995 a). EBNA-5 and Hsp70 co-localized in the nucleoli of IB-4 cells treated with MG132 for 6 h (Fig. 3 C).

Exogenously expressed EBNA-5 co-localized with Hsp70 in clone M5 in the nucleoplasm (MCF-7 transfected with EBNA-5, Fig. 4 A). This shows that localization to PML bodies is not a prerequisite for EBNA-5/Hsp70 co-localization. We noticed an increase in nuclear staining of Hsp70 in the clones of MCF-7 and SW480 expressing EBNA-5, as compared to the vector-transfected cells (Fig. 4 B, panel b and not shown). This suggests that EBNA-5 tethers Hsp70 to the nucleus.

Fig. 4. Concomitant changes in subnuclear distribution and total levels of EBNA-5 and Hsp70 in EBNA-5-expressing cell lines. (A) High magnification of EBNA-5 (red) and Hsp70 (green) double-staining of untreated MCF-7-derivative M5. (B) DMSO- and MG132-treated M5 cells that expressed low levels of EBNA-5 were double-stained for EBNA-5 (a and e, respectively) and Hsp70 (b and f, respectively). Overlap of green and red fluorescence is shown in panels c and g. Panel d, phase contrast of the field shown in panels a, b and c; panel h, phase contrast of panels e, f and g. Note a pronounced Hsp70 nuclear staining in DMSO-treated cells and a strong Hsp70 accumulation in the nucleoli upon MG132 treatment in EBNA-5 expressing cells. (C) Western blot of DMSO- or MG132-treated IB-4, MCF-7 clones M5 and M1, and HeLa clones H6 (EBNA-5-positive) and H4 (EBNA-5-negative) probed sequentially with antibodies against EBNA-5, p53 and Hsp70. Lower panel shows EBNA-5 and p53 protein expression levels in M1 after 16 h of MG132 treatment.

The antibody against Hsp70 recognizes both the inducible Hsp72 and the constitutive Hsc73 proteins. In untreated cells, Hsp70 showed homogeneous nuclear and cytoplasmic distribution (Fig. 4 B, panel b). Upon MG132 treatment, Hsp70 completely changed its localization: it accumulated in the cytoplasm and nucleoli (Fig. 4 B, panel f). EBNA-5 translocated to the nucleoli (Fig. 4 B, compare panels a and e). Clone M5, in which only 10 % of cells express high EBNA-5 levels, was chosen to demonstrate that the changes in the Hsp70 localization were independent of EBNA-5. We noticed, however, that cells expressing higher levels of EBNA-5 showed a stronger Hsp70 staining in the nucleoli (Fig. 4 A, panels e, f, g). The same phenomenon was observed in the MG132-treated S2 clone of SW480 that expressed EBNA-5.

Total protein levels were monitored by Western blotting (Fig. 4 C). EBNA-5 levels did not change after MG132 treatment in IB-4, the H6 clone of HeLa expressing EBNA-5 or clone M5 of MCF-7. Even 16 h treatment did not change the EBNA-5 levels in clone M1 (Fig. 4 C, lower panel). Hsp70 levels were increased and a faster migrating band, representing the inducible Hsp72 protein, appeared in all MG132-treated cells including the EBNA-5-negative HeLa clone H4. This again demonstrates that Hsp70 is induced in response to MG132 and that this induction is independent of EBNA-5.

In order to test the effect of proteasome inhibition, we probed the same membrane with antibodies against p53, a short-lived protein degraded through the proteasome pathway. The presence of EBNA-5 did not change the levels of wild-type p53 in MCF-7 or in HeLa cells. MG132 treatment led to the accumulation of wild-type p53 in IB-4 and MCF-7 cells with the appearance of ubiquitinated forms of the protein in accordance with previously published data (Kubbutat et al., 1997). In HeLa cells, p53 is complexed with human papillomavirus-encoded E6 protein, which targets p53 for degradation. The inhibition of proteasome activity has led to the accumulation of p53 in both EBNA-5-expressing clone H6 and vector control H4 Hela cells.

Presence of EBNA-5 enhances translocation of mutant p53 to the nucleoli of SW480 cells upon treatment with proteasome inhibitor

The colorectal cancer line SW480 expresses endogenous mutant p53 at high levels. p53 is localized to the nuclei, mainly avoiding the nucleoli (Fig. 5 A, panels c and e). P2 and S2 are pBabe vector-and pBabe-EBNA-5-transfected derivatives of SW480. The presence of EBNA-5 did not change the levels of mutant p53 (compare Fig. 5 A, panels c and e). Treatment with 20 µM of MG132 for 6 h led to a nearly complete translocation of EBNA-5 to the nucleoli of S2 (Fig. 5 A, panels a and b), to a partial accumulation of mutant p53 in the nucleoli in 10 % of the P2 cells (Fig. 5 A, panel f) and to a significant p53 accumulation in the nucleoli in 90 % of S2 cells (Fig. 5 A, panel d). Fig. 5(B) shows p53 inside the nucleoli in the MG132-treated S2 cells at high magnification. In order to substantiate this finding, we double-stained MG132-treated S2 cells for EBNA-5 and p53. Fig. 5(C) shows that p53 is preferentially accumulated in the nucleoli of cells when EBNA-5 is also in the nucleoli (arrows). In contrast, cells without EBNA-5 (square sign) or those where EBNA-5 has not changed its localization (asterisks) showed homogeneous p53 staining. These results show that EBNA-5 greatly enhances accumulation of mutant p53 in the nucleoli upon inhibition of proteasome activity.

Fig. 5. EBNA-5 enhances the nucleolar accumulation of mutant p53 in SW480 cells treated for 6 h with the proteasome inhibitor MG132. (A) EBNA-5 (a, b) and mutant p53 (c, d, e, f) in the DMSO- (a, c, e) or MG132-treated (b, d, f) derivatives of the SW480 colon cancer cell line. Clone S2 (a–d) stably expressed EBNA-5. P2 (e, f) is a vector control-expressing clone. (B) High magnification of the MG132-treated S2 cells stained for p53 (a). Phase contrast is shown in panel (c). Overlap of panels a and c shows that p53 is localized to the nucleoli (b). (C) Double-staining of the mutant p53 (green) and EBNA-5 (red) in MG132-treated S2 cells. Note that a cell without EBNA-5 (square sign) or those cells where EBNA-5 has not changed its localization (asterisks) showed homogeneous p53 staining, while the cells with nucleoli-localized EBNA-5 have also nucleoli-accumulated p53 (arrows).

The inhibition of wild-type p53 degradation by MG132 in MCF-7 cells led to the accumulation of the protein and its ubiquitinated forms (Fig. 4 C) in both nucleus and nucleolus (Fig. 6, panel d and phase contrast in panel f). Transient expression of GFP–EBNA-5 (Fig. 6, panels a and b) had no effect on the intranuclear distribution of wild-type p53 upon MG132 treatment (Fig. 6, panel d). Arrows point to the MG132-treated cell expressing GFP–EBNA-5 that has translocated to the nucleoli.

Fig. 6. Expression of GFP–EBNA-5 does not influence the proteasome inhibitor-induced nuclear and nucleolar accumulation of wild-type p53 in MCF-7 cells. MCF-7 cells, transfected with the GFP–EBNA-5 construct, were cultured 1 day after transfection in the presence of DMSO (a, c, e) or 10 µM of MG132 (b, d, f) for 16 h, and analysed for p53 expression by immunostaining. Panels a and b, GFP–EBNA-5 and DNA; panels c and d, wild-type p53; panel e, phase contrast of the field shown in panels a and c; panel f, phase contrast of panels b and d. Arrow indicates cell with GFP–EBNA-5 in the nucleoli.

We have found that the treatment of different EBNA-5-expressing cell lines with a proteasome inhibitor, MG132, led to a dramatic change in the distribution of EBNA-5, namely its almost complete translocation to the nucleoli. MG132 increased the levels of the cellular chaperon Hsp70 in agreement with previously published data (Bush et al., 1997; Kim et al., 1999). We have found in addition that MG132 induced a nucleolar accumulation of Hsp70.

The Hsp70 proteins are involved in protein folding, protein translocation across membranes and thermal tolerance. In non-stressed cells, Hsp70 associates transiently with nascent polypeptide chains, polypeptides unfolded for translocation or other aberrantly folded proteins (for review see Hightower, 1991). Both major members of the Hsp70 family, the stress-inducible Hsp72 and the constitutive cognate Hsc73, were found to associate with the soluble form of EBNA-5 in co-immunoprecipitation experiments. The W₁W₂-repeats in the EBNA-5 protein were required for binding (Mannick et al., 1995). Another study showed that the Y₂ C-terminal exon region of EBNA-5 formed complexes with Hsp70 (Kitay & Rowe, 1996). We have found previously that EBNA-5 is co-localized with Hsp70 in PODs in the LCL IB-4 under normal conditions and in the nucleoli upon heat shock (Szekely et al., 1995 a). Both Hsp72 and Hsc73 accumulate in the nucleoli in heat-shocked mammalian cells (for review see Hightower, 1991). We found that Hsp70 localized to the nucleoli upon MG132 treatment, independently of the presence or absence of EBNA-5. Considering the direct binding of EBNA-5 to Hsp70, it is possible that the latter transports EBNA-5 into the nucleoli. On the other hand, Hsp70 accumulated in the nucleoplasm of the EBNA-5-, but not the vector control-transfected clones of MCF-7 and SW480 (Fig. 4 A and not shown). It also showed a high degree of co-localization with EBNA-5 in the nucleoplasm, suggesting that EBNA-5 expression recruits Hsp70 into complex formation. The nucleolar localization of Hsp70 in proteasome inhibitor-treated cells was enhanced by the presence of EBNA-5 (Fig. 4 B), suggesting that there is an active component in the translocation of EBNA-5 to the nucleoli rather than passive transport by Hsp70.

EBNA-5 is strongly associated with the nuclear matrix (Szekely et al., 1995 a). It can also tether another EBV-encoded nuclear protein, EBNA-3, to the nuclear matrix upon co-transfection (Cludts & Farrell, 1998). This suggests that EBNA-5 can influence the subnuclear localization of other proteins and therefore it may play a scaffolding role by forming bridges between the nuclear matrix and nucleoplasmic proteins.

We found in this study that EBNA-5 greatly enhanced the accumulation of a mutant, but not wild-type, p53 in the nucleoli upon MG132 treatment. Mutations in p53 lead to misfolding of the protein. Mutant p53 forms complexes with Hsp70 (Pinhasi-Kimhi et al., 1986). The high affinity binding sites for these proteins were mapped to the hydrophobic core of the central DNA binding domain, not accessible in a wild-type conformation (Fourie et al., 1997). In the absence of EBNA-5, mutant p53 was only rarely found in the nucleoli of MG132-treated cells. Our study suggests that EBNA-5 targets mutant p53 to the nucleoli upon proteasome inhibitor treatment. GST–EBNA-5 could precipitate mutant p53 from SW480 and Namalwa cell lysates (Szekely et al., 1993). It is possible that EBNA-5 and Hsp70 form tri-component complexes with the misfolded mutant p53 that translocate to the nucleolus much more efficiently than the Hsp70–mutant p53 complex. The absence of the EBNA-5 effect on wild-type p53 distribution upon MG132 treatment supports this notion. Alternatively, EBNA-5 binds and translocates mutant p53 independently of Hsp70.

The nucleoli are special nuclear domains involved in rRNA synthesis (fibrillar compartment) and ribosome formation (granular compartment). Nucleolar proteins are involved either in rRNA synthesis and processing or in ribosome assembly (for review see Scheer & Hock, 1999). The involvement of the nucleoli in proteasome-mediated protein degradation has not been documented. Our data on proteasome inhibitor-induced nucleolar accumulation of Hsp70, EBNA-5 and p53 may suggest that nucleoli can be involved in proteasome-dependent protein degradation.

EBNA-5 accumulates in the PML bodies in cells with an immunoblastic phenotype but not in type I BLs or in non-B cells, suggesting the presence of EBNA-5-associated, B-blast-specific nuclear factors that target EBNA-5 into the PODs. Alternatively, it may require an additional EBV-encoded protein(s) or EBV genome in order to localize to the PODs.

It has been suggested that PML bodies are the nuclear analogues of the aggrosomes that feed proteasomes with ubiquitinated substrates (Anton et al., 1999). PML bodies also contain Hsp70 even in unstressed cells (Szekely et al., 1995 a), suggesting that misfolded nuclear proteins might accumulate in the PODs. It seems unlikely that EBNA-5 itself is degraded through the proteasome pathway since it does not contain lysines, which are the targets for ubiquitination. Also, our Western blot data showed no increase in EBNA-5 protein levels upon MG132 treatment in different cell lines even after 16 h of proteasomal block. It seems therefore unlikely that the PML bodies provide an intermediate reservoir for EBNA-5 on its way for degradation. On the other hand, our data suggest that POD-associated EBNA-5 might regulate the proteasome-mediated degradation of other proteins.

The PML bodies might be indirectly involved in transcription regulation by coordinating the degradation of transcription factors. Transcription co-activators such as CBP/p300, transcriptional repressors, such as Daxx and Tax, and the tumour suppressor protein Rb were detected in the PODs (for review see Zhong et al., 2000 b). EBNA-5 was found to enhance EBNA-2-mediated transactivation of EBV promoters (Harada & Kieff, 1997). The mechanism of this co-operation is not known. Our data raise the possibility that EBNA-5 might cooperate with EBNA-2 through modifying the degradation rate of transcription regulators.

The work was supported by Cancerfonden (Sweden) and The Cancer Research Institute (USA).

Abarzua, P., LoSardo, J. E., Gubler, M. L. & Neri, A. (1995). Microinjection of monoclonal antibody PAb421 into human SW480 colorectal carcinoma cells restores the transcription activation function to mutant p53. Cancer Research 55, 3490–3494.

Alfiery, C., Birkenbach, M. & Kieff, E. (1991). Early events in Epstein–Barr virus infection of human B lymphocytes. Virology 181, 595–608.

Anton, L. C., Schubert, U., Bacik, I., Princiotta, M. F., Wearsch, P. A., Gibbs, J., Day, P. M., Realini, C., Rechsteiner, M. C., Bennink, J. R. & Yewdell, J. W. (1999). Intracellular localization of proteasomal degradation of a viral antigen. Journal of Cell Biology 146, 113–124.

Ben-Bassat, H., Goldblum, N., Mitrani, S., Goldblum, T., Yoffey, J. M., Cohen, M. M., Bentwich, Z., Ramot, B., Klein, E. & Klein, G. (1977). Establishment in continuous culture of a new type of lymphocyte from a 'Burkitt like' malignant lymphoma (line D.G.-75). International Journal of Cancer 19, 27–33.

Bush, K. T., Goldberg, A. L. & Nigam, S. K. (1997). Proteasome inhibition leads to a heat-shock response, induction of endoplasmic reticulum chaperones, and thermotolerance. Journal of Biological Chemistry 272, 9086–9092.

Chelbi-Alix, M. K. & de The, H. (1999). Herpes virus induced proteasome-dependent degradation of the nuclear bodies-associated PML and Sp100 proteins. Oncogene 18, 935–941.

Ciechanover, A. (1998). The ubiquitin–proteasome pathway: on protein death and cell life. EMBO Journal 17, 7151–7160.

Cludts, I. & Farrell, P. J. (1998). Multiple functions within the Epstein–Barr virus EBNA-3A protein. Journal of Virology 72, 1862–1869.

Finke, J., Rowe, M., Ernberg, I., Rosen, A., Dillner, J. & Klein, G. (1987). Monoclonal and polyclonal antibodies against Epstein–Barr virus nuclear antigen 5 (EBNA-5) detect multiple protein species in Burkitt's lymphoma and lymphoblastoid cell lines. Journal of Virology 61, 3870–3878.

Fourie, A. M., Hupp, T. R., Lane, D. P., Sang, B. C., Barbosa, M. S., Sambrook, J. F. & Gething, M. J. (1997). HSP70 binding sites in the tumor suppressor protein p53. Journal of Biological Chemistry 272, 19471–19479.

Harada, S. & Kieff, E. (1997). Epstein–Barr virus nuclear protein LP stimulates EBNA-2 acidic domain-mediated transcriptional activation. Journal of Virology 71, 6611–6618.

Hightower, L. E. (1991). Heat shock, stress proteins, chaperones, and proteotoxicity. Cell 66, 191–197.

Holmvall, P. & Szekely, L. (1999). Computer programs that allow fast aquisition, visualization and overlap quantitation of fluorescent 3D microscopic objects by using nearest-neighbor deconvolution algorithm. Applied Immunohistochemistry & Molecular Morphology 7, 226–236.

Ishov, A. M., Sotnikov, A. G., Negorev, D., Vladimirova, O. V., Neff, N., Kamitani, T., Yeh, E. T., Strauss, J. F., III & Maul, G. G. (1999). PML is critical for ND10 formation and recruits the PML-interacting protein daxx to this nuclear structure when modified by SUMO-1. Journal of Cell Biology 147, 221–234.

Kim, D., Kim, S. H. & Li, G. C. (1999). Proteasome inhibitors MG132 and lactacystin hyperphosphorylate HSF1 and induce hsp70 and hsp27 expression. Biochemical and Biophysical Research Communications 254, 264–268.

King, W., Thomas-Powell, A. L., Raab-Traub, N., Hawke, M. & Kieff, E. (1980). Epstein–Barr virus RNA. V. Viral RNA in a restringently infected, growth-transformed cell line. Journal of Virology 36, 506–518.

Kitay, M. K. & Rowe, D. T. (1996). Protein–protein interactions between Epstein–Barr virus nuclear antigen-LP and cellular gene products: binding of 70-kilodalton heat shock proteins. Virology 220, 91–99.

Kubbutat, M. H., Jones, S. N. & Vousden, K. H. (1997). Regulation of p53 stability by Mdm2. Nature 387, 299–303.

Mannick, J. B., Tong, X., Hemnes, A. & Kieff, E. (1995). The Epstein–Barr virus nuclear antigen leader protein associates with hsp72/hsc73. Journal of Virology 69, 8169–8172.

Muller, S. & Dejean, A. (1999). Viral immediate-early proteins abrogate the modification by SUMO-1 of PML and Sp100 proteins, correlating with nuclear body disruption. Journal of Virology 73, 5137–5143.

Pinhasi-Kimhi, O., Michalovitz, D., Ben-Zeev, A. & Oren, M. (1986). Specific interaction between the p53 cellular tumour antigen and major heat shock proteins. Nature 320, 182–184.

Quignon, F., De Bels, F., Koken, M., Feunteun, J., Ameisen, J. C. & de The, H. (1998). PML induces a novel caspase-independent death process. Nature Genetics 20, 259–265.

Rickinson, A. B. & Kieff, E. (1996). Epstein–Barr virus. In Fields Virology, 3rd edn, pp. 2397–2436. Edited by B. N. Fields, D. M. Knipe & P. M. Howley. Philadelphia: Lippincott–Raven.

Rowe, M., Rowe, D. T., Gregory, D., Young, L. S., Farrell, P. J., Rupani, H. & Rickinson, A. B. (1987). Differences in B cell growth phenotype reflect novel patterns of Epstein–Barr virus latent gene expression in Burkitt's lymphoma cells. EMBO Journal 6, 2743–2751.

Scheer, U. & Hock, R. (1999). Structure and function of the nucleolus. Current Opinion in Cell Biology 11, 385–390.

Sinclair, A. J., Palmero, I., Peters, G. & Farrell, P. J. (1994). EBNA-2 and EBNA-LP cooperate to cause G0 to G1 transition during immortalisation of resting human B lymphocytes by Epstein–Barr virus. EMBO Journal 13, 3321–3328.

Szekely, L., Selivanova, G., Magnusson, K., Klein, G. & Wiman, K. G. (1993). EBNA-5, an Epstein–Barr virus encoded nuclear antigen, binds to the retinoblastoma and p53 proteins. Proceedings of the National Academy of Sciences, USA 90, 5455–5459.

Szekely, L., Jiang, W.-Q., Pokrovskaja, K., Wiman, K. G., Klein, G. & Ringertz, N. (1995 a). Reversible nucleolar translocation of Epstein–Barr virus-encoded EBNA-5 and hsp70 proteins after exposure to heat shock or cell density congestion. Journal of General Virology 76, 2423–2432.

Szekely, L., Pokrovskaja, K., Jiang, W.-Q., Selivanova, G., Löwber, M., Ringertz, N., Wiman, K. G. & Klein, G. (1995 b). Resting B-cells, EBV-infected B-blasts and established lymphoblastoid cell lines differ in their Rb, p53 and EBNA-5 expression patterns. Oncogene 10, 1869–1874.

Szekely, L., Pokrovskaja, K., Jiang, W. Q., de The, H., Ringertz, N. & Klein, G. (1996). The Epstein–Barr virus-encoded nuclear antigen EBNA-5 accumulates in PML-containing bodies. Journal of Virology 70, 2562–2568.

Wang, Z. G., Ruggero, D., Ronchetti, S., Zhong, S., Gaboli, M., Rivi, R. & Pandolfi, P. P. (1998). PML is essential for multiple apoptotic pathways. Nature Genetics 20, 266–272.

Zheng, P., Guo, Y., Niu, Q., Levy, D. E., Dyck, J. A., Lu, S., Sheiman, L. A. & Liu, Y. (1998). Proto-oncogene PML controls genes devoted to MHC class I antigen presentation. Nature 396, 373–376.

Zhong, S., Muller, S., Ponchetti, S., Freemont, P. S., Dejean, A. & Pandolfi, P. P. (2000 a). Role of SUMO-1-modified PML in nuclear body formation. Blood 95, 2748–2752.

Zhong, S., Salomoni, P. & Pandolfi, P. P. (2000 b). The transcriptional role of PML and the nuclear body. Nature Cell Biology 2, 85–90.

JGV Direct table of contents

© 2000 SGM

This article is now available in the February 2001 print issue of JGV (vol. 82, 345–358). The complete issue of the journal may be seen in electronic form on JGV Online.