Subnuclear localization of 20S and 26S proteasomes. Indirect immunofluorescence of HEp-2 cells with rabbit antibodies against 20S proteasomes purified from human placenta (B), and 26S proteasomes from rat muscle (F) revealed a speckled staining pattern within the nucleoplasm. Such a staining pattern was reproduced with mouse monoclonal antibodies to the alpha 4 subunit of 20S proteasomes (D). Colocalization of 20S proteasomes (H, green color) and splicing factor SC35 (H, red color) occurred in nucleoplasmic speckles (I, yellow color, inset). Note that cytoplasmic proteasome staining is weak due to confocal sectioning. (A, C, E, and H, gray color) represent corresponding differential interference contrast (DIC) images. Bars, 5 μm. (J) Fluorescence intensities of double-labeling experiments (I) were quantified with Metamorph analysis software according to MATERIALS AND METHODS. 20S proteasomes as well as SC35 are localized within the nucleoplasm (Nu) and excluded from the nucleolus (No). 34 ± 12.3% (mean ± SD) of SC35 colocalizes with 20S proteasomes within splicing speckles (*colocalization between SC35 and 20S proteasomes). (G) Characterization of the antibodies against 20S and 26S proteasomes by immunoblotting shows multiple bands between 20 and 30 kDa, representing proteasomal subunits. Molecular weight is given in kDa.

Mercury-induced colocalization of fibrillarin with proteasomes in nucleoplasmic sites. Double-labeling of untreated HEp-2 cells (first panel) with (A) antifibrillarin (green color) and (B) anti-20S proteasome antibodies (red color). The merged image shows no colocalization (C). In HgCl₂-treated HEp-2 cells (second panel) fibrillarin (E, green color) redistributes from the nucleolus to the nucleoplasm, and colocalizes with 20S proteasomes (F, red color) in nucleoplasmic aggregates (G, yellow color, inset). To visualize colocalization in more detail red and green channels were converted separately to greyscale images. Overlapping foci were determined by identification of pixels in which both red and green signals are within 30% of the maximum for that channel. Information of the pixel analysis is output in a superimposed channel showing no overlap in untreated HEp-2 cells (D) but colocalization of fibrillarin and proteasomes in distinct aggregations of fluorescent foci throughout the nucleoplasm in mercury-treated cells (H). Bar, 5 μm. Quantification of fluorescent data by means of Metamorph corroborated exclusive confinement of fibrillarin within nucleoli (No), and 20S proteasomes within the nucleoplasm (Nu) in untreated HEp-2 cells (I). However, upon mercury-treatment fibrillarin redistributes from the nucleolus (No) to the nucleoplasm (Nu), whereas 20S proteasomes maintain their subnuclear localization (J). Thus, no overlap could be observed in untreated cells, whereas 33 ± 16.7% (mean ± SD) of redistributed fibrillarin colocalizes with proteasomes within the nucleoplasm of HgCl₂-treated cells (K). *Colocalization between nucleoplasmic fibrillarin and 20S proteasomes.

Mercury-induced colocalization of proteasomes with fibrillarin is specific and does not occur with major nucleolar proteins B23 and nucleolin as revealed by double labeling immunofluorescence and confocal microscopy. (A) B23 (green color) is exclusively confined to the nucleolus in untreated control cells, whereas 20S proteasomes (red color) are localized within the nucleoplasm. Thus B23 does not overlap with nucleoplasmic proteasomes (merge, and graph). (B) On mercury-treatment a subfraction of B23 appears in numerous homogeneously distributed foci within the nucleoplasm (green color). However, no colocalization with proteasomes (red color) can be detected (merge, inset, and graph). (C) Nucleolin (green color) is exclusively localized to the nucleolus in untreated HEp-2 cells and does not colocalize with proteasomes (merge, and graph). (D) In contrast, a subfraction of nucleolin (green color) appeared in aggregates distributed throughout the nucleoplasm in HgCl₂-treated cells. However, no colocalization between nucleoplasmic nucleolin aggregates and proteasomes (red color) could be observed (merge, inset, and graph). The first column shows corresponding DICs. Bars, 5 μM. Quantification of the fluorescent data by Metamorph (graphs, right column) confirmed that nucleolar proteins B23 and nucleolin alter their subnuclear distribution but do not colocalize with 20S proteasomes either in untreated or in mercury-treated cells. In contrast, 20S proteasomes remain confined to the nucleoplasm. No, nucleolus; Nu, nucleoplasm. *Colocalization of B23 with 20S proteasomes in the nucleoplasm; ^#colocalization of nucleolin and 20S proteasomes in the nucleoplasm.

Colocalization of fibrillarin and 20S proteasomes in splenic cells from mercury-treated mice. B10.S mice were treated with saline (A and B) or subLCs of HgCl₂ (C and D) for 8 weeks. (A) Sera from saline-treated control mice did neither react in immunoblotting nor in indirect immunofluorescence of HEp-2 cells. (B) Splenic cells were isolated from these mice and subjected to double labeling and confocal microscopy: Fibrillarin (green color) is confined to the nucleolus, and thus no colocalization (merge, enlargements) could be observed with proteasomes (red). (C) Sera from mercury-treated mice reacted with a 34-kDa band corresponding to the molecular weight of fibrillarin by immunoblotting of HEp-2 cell lysates and showed an exclusive, clumpy staining of the nucleolus in indirect immunofluorescence of HEp-2 cells. (D) Double-labeling and confocal microscopy of splenic cells from HgCl₂-treated mice revealed a redistribution of fibrillarin from the nucleolus to the nucleoplasm (green color) and colocalization with proteasomes (red color) in distinct nucleoplasmic aggregates (yellow color, merge, enlargements). Bars, 5 μm. Molecular weight is given in kDa. fib, fibrillarin.

Proteasome-dependent processing of fibrillarin induced by mercury. (A) HEp-2 cells were treated with increasing concentrations of proteasome inhibitor lactacystin for 24 h and HgCl₂ where indicated and separated on SDS-PAGE, followed by immunoblotting. Antifibrillarin antibodies detected a 34-kDa band corresponding to the molecular weight of fibrillarin in untreated control cells (lane 1, filled arrowhead), and cells were treated with 0.1 μm lactacystin (lane 2). Simultaneous addition of increasing concentrations of lactacystin and HgCl₂ lead to accumulation of the 34-kDa band and to the appearance of additional slower migrating bands that represent the molecular weight of fibrillarin (34 kDa) plus multiples of ∼8.5 kDa (lanes 3–6, open arrowheads). (B) Immunoblotting of untreated HEp-2 cells (first column) and coimmunoprecipitates obtained with antiubiquitin antibodies (second column). Fibrillarin (upper panel) and topoisomerase I (lower panel) were coprecipitated with antiubiquitin antibodies. Ornithine decarboxylase (ODC), which is degraded by 26S proteasomes without being ubiquitinylated, served as a negative control and was not precipitated by antiubiquitin (middle panel). (C) For in situ accumulation studies HEp-2 cells were treated with 20 μM HgCl₂ and 1 μM lactacystin where indicated. Mercury-treatment induces redistribution of fibrillarin from the nucleolus to the nucleoplasm (second column) and accumulation of fibrillarin in the nucleoplasm of cells, which were additionally treated with lactacystin (lower panel). Fluorescence images from lactacystin-treated and control cells were recorded with identical settings of the confocal microscope. (D) Quantification of fluorescence intensities with Metamorph (see MATERIALS AND METHODS) revealed an accumulation of nucleoplasmic fibrillarin from 100 to 144 ± 14.8% (mean ± SD) in lactacystin-treated cells, whereas concentration of fibrillarin within the nucleolus remained constant at 105 ± 14.1% (mean ± SD, filled bars). Hg, mercuric chloride; Lc, lactacystin; Nu, nucleoplasm; No, nucleolus.