EXPERIMENTAL
CELL
RESEARCH
199,
262-268 (1992)
Cell Cycle-Dependent Migration of the DNA-Binding Protein Ku80 into Nucleoli LI-LAN Graduate Institute
of Microbiology
LI AND NING-HSING
and Immunology,
National
Inc.
INTRODUCTION
The distinct dynamic changes of nucleolar structure are connected to the different phases of the cell cycle and the overall growth rates of cells [ 11. Little is known about whether particular nucleolar components are responsible for these changes. We have previously demonstrated that the expression of nucleolin, the major nucleolar phosphoprotein, is dynamic and correlates well with the degree of cell proliferation. The mechanism of
$3.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
College, Taiwan,
MATERIALS
Republic of China
AND
METHODS
Cell cultures. Human cervical carcinoma HeLa, T leukemia CEM, and hepatoma HA22T cells were cultured in RPM1 1640 medium supplemented with 10% heat-inactivated fetal calf serum (FCS) (Boehringer-Mannheim, Mannheim, Germany). These cells were grown at 37°C with 5% CO,. HeLa and HA22T cells were passaged with trypsin/EDTA to maintain subconfluency. CEM cells were subcultured twice every week. For use in serum starvation studies, HeLa cells (5 X lo5 cells/ml) were placed in RPM1 1640 containing 0.5% heat-inactivated FCS and allowed to quiesce for 3 days. Synchronization procedures. To accumulate HeLa cells in mitotic phase [9], exponentially growing monolayers were maintained for 12 h in medium containing nocodazole (Sigma Chemical Co., St. Louis, MO) at 0.4 pg/ml, after which mitotic cells were obtained by “shake-
I To whom reprint requests should be addressed at: Graduate Institute of Microbiology and Immunology, National Yang-Ming Medical College, Shih-Pai, 11221 Taipei, Taiwan, Republic of China. Fax: (02) 821-2880. 0014-4827/92
Medical
this phenomenon was proven due to the self-cleaving activity of nucleolin and a putative proteolytic inhibitor, present only in proliferating cells, to prevent the cleavage [2]. The instability of nucleolin in nondividing cells may account in part for the degeneration of nucleoli in these cells. In attempts to search out more nucleolar proteins in relation to nucleolar dynamic changes, we generated monoclonal antibodies against purified nuclei. One antibody, termed LLl, particularly aroused attention since the immunofluorescence staining of cells showed variations in intensity within the nucleoli. Human cDNA clones were isolated by immunoscreening, and a partial nucleotide sequence was determined. The sequence was found to be identical to that of a previously reported cDNA encoding the 80-kDa subunit of the Ku DNA-binding protein [3, 41. Immunofluorescence staining revealed that Ku was distributed throughout the nucleoplasm and often also in the nucleoli of interphase cells [5, 61. However, Ku was rapidly dissociated from DNA and appeared at the periphery of the condensed chromosomes during mitosis [7]. Although a cell cycle-dependent distribution of Ku in the nucleoli was suggested [8], the direct experimental evidence was still lacking. We report here that nucleoli of cells at the Gl/S boundary had very small amounts of Ku80 protein. It appeared gradually in nucleoli during the cell cycle progression and reached a maximum at late S or G2 phase. Whether this cell cycledependent migration of Ku80 into nucleoli is responsible for nucleolar dynamic changes needs to be extensively studied.
The DNA-binding protein Ku (p7O/p80) was originally discovered through the use of human autoimmune sera. In attempts to search out nucleolar proteins in relation to nucleolar dynamic changes, we developed monoclonal antibodies against nuclear proteins. One antibody, termed LLl, received particular attention since asynchronous cells exhibited tremendous differences in their nucleolar fluorescence intensities after immunostaining. The LLl protein was proven to be the Ku subunit ~80 (Ku80) by cDNA cloning and sequencing. Possible correlations between the heterogeneous distribution of Ku80 in nucleoli and the cell cycle were examined. HeLa cells were synchronized at M phase by arrest with nocodazole, or at the GUS boundary by sequential treatments with thymidine and aphidicolin. These cells were then released by culturing in fresh medium to allow the cell cycle to progress synchronously. Immunofluorescent detection of Ku80 revealed that nucleoli of the cells at the GllS boundary had very small amounts of Ku80, which was mainly present in the nucleoplasm. Ku80 was gradually accumulated in nucleoli during S phase and reached the maximum at late S or G2 phase. Immunoblotting experiments showed that cell extracts prepared from different phases of the cell cycle had virtually identical amounts of Ku80. These results suggest that Ku80 migrates from nucleoplasm to 0 1992 Academic nucleoli in a cell cycle-dependent manner. Press,
Yang-Ming
YEH’
262
CELL
CYCLE-DEPENDENT
MIGRATION
off’. Interphase cells were collected by trypsin/EDTA treatment of monolayers after vigorous washing to remove all mitotic cells. About 85% of these mitotic cells showed condensed chromosomes as determined by DNA staining. These mitotic cells were washed with phosphate-buffered saline (PBS) and driven into interphase by culturing in fresh medium on glass slides. Cells were used for immunofluorescence staining at different time intervals. Cell extracts were also prepared from the above interphase, mitotic-phase, and Gl-phase cells (4 h after release from M phase). A second protocol was used to arrest cells at the Gl/S boundary by sequential treatments with thymidine and aphidicolin [lo]. HeLa cells were blocked with 2 n&f thymidine for 12 h. Cells were rinsed with PBS to remove thymidine and then exposed to medium in the presence of 2.4 X 10e5 M thymidine and deoxycytidine for 9 h. Aphidicolin (Sigma) was added to 5 pg/ml. After culturing for an additional 12 h, cells were blocked at the Gl/S boundary. Reversal of the Gl/S blockade was accomplished by washing the cells twice with PBS, and followed by suspension in complete medium. These cells were grown on glass slides and taken at various time points for immunofluorescence staining. Monoclonal antibodies. Balb/c mice were hyperimmunized with purified CEM nuclei (purity > 95%). Procedures for the preparation of nuclei were identical to those described previously [2]. Spleen cells of these mice were fused with NS-1 myeloma cells by standard techniques [ll]. Hybridoma supernatants were prescreened by a dot-immunobinding assay [ 121 to select clones secreting monoclonal antibodies against immunogens. Those positive hybrids were rescreened by indirect immunofluorescence microscopy. Monoclonal antibodies recognizing nucleolar components were retained; two of these, LLl and BL9, showed nucleolar positive reactions only in a proportion of the normal cell populations and therefore were used for further investigation. Zmmunoblotting. HeLa, CEM, HA22T, or synchronized HeLa cells were Iysed in extraction buffer [l% Triton X-100,0.2% deoxycholate, 0.6 M NaCl, 0.1% sodium dodecyl sulfate (SDS), 10 m&f TrisHCl, pH 8.0,20 mM 2-mercaptoethanol, 10 mM KCI, 0.5 mM MgCl,]. The lysates were sonicated to disrupt DNA and then clarified by centrifugation at 13,000g for 20 min. Samples were immediately stored at -70°C. Samples containing 100 pg total proteins were analyzed by SDS-7.5% polyacrylamide gel electrophoresis (PAGE). Proteins were transferred from gel to nitrocellulose membrane (Bio-Rad Laboratories, Richmond, CA) at 0.8 mA/cm* for 3 h with a NovaBlot electrophoretic transfer apparatus (LKB). Nonspecific protein binding was blocked by incubating the membrane in 50 n&f Tris-HCl, pH 7.5,150 mM NaCl (TBS) containing 3% nonfat milk for 2 h. The membrane was reacted with LLl or BL9 antibody for 1 h and then extensively washed with TBS. A 2000-fold diluted peroxidase-conjugated goat anti-mouse IgG (PXGAMIgG) (Bio-Rad) in TBS containing 3% nonfat milk was added to the membrane for 1 h. After washing with TBS, specific bands were visualized by reaction with a solution freshly prepared by dilution of the stock of 4-chloro-1-naphthol (3 mg/ml in methanol) (Merck & Co., Rahway, NJ) with 5 vol of TBS containing 0.02% H,O,. Isolation of cDNA clones. LLl antibody was used to screen a human hepatoma HA22T Xgtll cDNA library (kindly provided by Dr. T.-S. Su, Veterans General Hospital, Taipei, Republic of China). Methods for screening were based on the protocols reported by Young and Davis [13]. We modified only the last step for detection of positive plaques on nitrocellulose membrane by using PXGAMIgG and color reaction as described above, under “Immunoblotting.” Three positive cDNA clones, L3, L4, and L5, were plaque purified and characterized further by restriction mapping and sequencing. DNA sequencing. Three cDNA inserts from the L3, L4, and L5 clones were subcloned into the phagemid pBluescript II SK+ (Stratagene) in both orientations. Single-stranded DNA templates were rescued by infection of the recombinant phagemid-transformed Escherichia coli strain XLl-blue with helper phage VCSM13 according to the instructions of the manufacturer. Sequencing reactions were
OF Ku80
INTO
263
NUCLEOLI
carried out by the dideoxy chain termination method [14] with a Sequenase kit (United States Biochemical Corp.). Sequence analysis was performed with the PC/gene computer program (IntelliGenetics, Mountain View, CA). E. coli strain Y1089 was lysogenized Expression of fusionproteins. with the L3, L4, or L5 recombinant phage. Lysogens were grown in 2 ml LB medium containing 2 mM isopropyl thio-j?-D-galactopyranoside (Clontech), 2 mM MgCl,, 0.2% maltose, and 50 ag/ml of ampicillin, for 3.5 h at 37°C on a shaker. Bacteria were collected by centrifugation and lysed in 100 ~1 Laemmli reducing sample buffer [15]. Samples were boiled for 4 min and centrifuged to remove debris. Ten microliters of the supernatant was analyzed on SDS-PAGE. Indirect immunofluorescence. Cells grown on slides at various time intervals after synchronization were fixed with 2% formaldehyde in PBS for 20 min at room temperature. After washing once with PBS, cells were permeabilized with -2O’C acetone for 3 min and rinsed with PBS. Fifty microliters hybridoma supernatant was added onto cells and the slides were placed in a moisture chamber for 1 h. Cells were washed three times in PBS (5 min for each wash) and then incubated for 1 h with 50 81 of a 150-fold-diluted second antibody, Auorescein isothiocyanate-conjugated goat anti-mouse IgG (Cappel), in 1% bovine serum albumin/PBS. A DNA dye, Hoechst 33258 (Sigma), was added to 1 pg/ml during this second incubation. Cells were washed three times with PBS and then mounted in a solution containing 0.42% glycine, 0.021% NaOH, 0.5% NaCl, 0.03% NaN,, and 70% glycerol for observation with an immunofluorescence microscope. RESULTS
Heterogeneous Distribution of LLl Antigen of Asynchronous Cell Populations
in Nucleoli
Spleen cells of Balb/c mice hyperimmunized with purified CEM nuclei were used to develop hybridomas. When hybridoma supernatants were screened by a highly sensitive method, i.e., dot-immunobinding assay [12], approximately 25% of hybrids were identified as antibody-secreters with specificities to immunogens. Only about half of these positive supernatants had nuclear or nucleolar reactivities determined by immunofluorescence staining. One of the nucleolar reactive monoclonal antibodies, LLl, was particularly noted since its target LLl molecules were present in different amounts in nucleoli of asynchronous cells. Results using normally growing HeLa cells as an example are shown in Fig. 1. The LLl antigen was scattered throughout the nucleoplasm of all cells. However, cells displayed different fluorescence intensities in the nucleoli ranging from undetectable or low levels (as indicated by arrows) to moderate or very high levels. This heterogeneity with respect to concentrations of LLl antigen in nucleoli was consistently observed in all other cell lines tested (data not shown). Experiments with a control anti-nucleolin antibody did not show heterogeneous patterns in nucleoli of the same cell populations (data not shown), indicating that variations of LLl molecules in nucleoli were not due to fixation or permeation artifacts. Characterization To identify nal antibody,
of the LLl
Protein by Immunoblotting
the protein recognized by LLl monoclocell extracts of HeLa, CEM, and HA22T
264
Ll AND
YEH
‘I’ -2
Ku80 L3
f
L4 K
-
-0
-
L5
*
0.5 Kb
B
LLl kDa
1234
BL9 12
34
FIG. 1. Heterogeneous distribution of LLl antigen in nucleoli of asynchronous cell populations. HeLa cells were grown on slides. Cells were fixed with 2% formaldehyde in PBS and permeabilized with cold acetone. Indirect immunofluorescence staining was performed with LLl monoclonal antibody. Note the variations of fluorescence intensity in nucleoli ranging from very high to very low (indicated by arrows) levels.
were analyzed by immunoblotting. band
was
identified
in all
cell
A very sharp 85kDa extracts
under
reducing
conditions (Fig. 2B). Without reduction, extra diffuse forms migrating slightly faster than the 85kDa band in SDS gels were observed (Fig. 2A). This observation suggests that intramolecular disulfide bonds are present, resulting in the fast movement of the 85kDa protein. Identification
of LLl
Protein as the Ku80 Subunit
In order to further characterize the LLl protein, we isolated three LLl cDNA clones by immunoscreening
A kDa
123
B
FIG. 3. Identification of LLl protein as the Ku80 subunit. (A) Three CDNA clones (~3, ~4, and ~5) were isolated by immunoscreening with LLl monoclonal antibody. Four stretches of nucleotide sequences (indicated by arrows) were determined and found to be completely identical to those for the Ku80 subunit. Restriction sites on the cDNA inserts were determined by digestion with EcoRV (E), C/o1 (C), PstI (P), SnaI (S), andKpn1 (K). Results show the alignment of these three cDNA inserts as corresponding to the sequence of the known full length K&O. Open reading frame of Ku80 is boxed. (B) Escherichia coli Y1089 lysogenized with L3, L4, or L5 recombinant hgtll were induced by IPTG to express fusion proteins. The bacterial lysates were analyzed by immunoblotting and probed with monoclonal antibody LLl (left) or BL9 (right). Lanes l-4 indicate samples for bacterial control (lane 1) or lysogens containing L3, L4, or L5 cDNA (lanes 2-4, respectively). A bacterial protein cross-reacted with BL9 antibody is noted by the arrow.
123
FIG. 2. Characterization of the LLl protein by immunoblotting. Cell extracts of HeLa, CEM, and HA22T (lanes l-3, respectively) were subjected to SDS-PAGE under nonreducing (A) or reducing (B) conditions. Results show immunoblotting probed with LLl monoclonal antibody. Molecular weight markers are expressed in kDa.
the hepatoma cell line HA22T cDNA library. By gel electrophoresis, sizes of the cDNA inserts of clones L3, L4, and L5 were estimated to be 3.0, 1.4, and 1.8 kb, respectively. Each end of the cDNA inserts was sequenced by the dideoxy-nucleotide method [ 141. These nucleotide sequences were found to be identical to those for the 80-kDa subunit protein of the human autoantigen Ku (p7O/p80) reported previously [4]. A partial restriction map and the sequence data were used to align these three cDNA clones to corresnond with the full length Ku80 (Fig. 3A). All three clones have spanned regions through the translational termination-site predicted in the Ku80-cDNA sequence, but lack 5’-end coding regions in different lengths.
CELL
CYCLE-DEPENDENT
MIGRATION
Fusion proteins were induced and their apparent molecular weights were used to confirm that the LLl cDNA clones carry stop codons at the same site as that of K&O. The sizes of the fusion proteins were determined as 185,160, and 135 kDa for cDNA clones L3, L4, and L5, respectively (Fig. 3B). Since fusion proteins contain the llO-kDa P-galactosidase, the L3, L4, and L5 cDNA inserts were estimated to encode 75, 50-, and 25-kDa peptides, respectively. These sizes matched perfectly if the stop codon was assumed to be at the same site as Ku80 (see Fig. 3A). We have concluded that the LLl protein is the Ku80 subunit of the autoantigen Ku. One interesting finding was noted when bacterial extracts containing the fusion proteins were probed with a second monoclonal antibody, BL9, which has been proven as recognizing the same LLl protein (unpublished data). This antibody also detected a bacterial protein with a molecular weight of 50 kDa (indicated by arrow) since bacterial control without infection by the recombinant phage Xgtll also had this band (Fig. 3B, lane 1 of panel BL9). LLl protein shared at least one epitope with the bacterial 50-kDa protein. The significance of this observation is unknown.
Cell Cycle-Dependent Migration of LLl to Nucleoli Asynchronous cells immunostained with LLl monoclonal antibody exhibited tremendous differences in their nucleolar fluorescence intensities (Fig. 1). Questions were raised about whether these variations correlated with the different stages of the cell cycle. Experiments of cell synchronization were therefore performed. Mitotic-phase-arrested HeLa cells were obtained by nocodazole treatment and mechanical shakeoff [9]. Greater than 85% of these cells had condensed chromosomes (Fig. 4E and data not shown). LLl protein in mitotic cells appeared to be completely away from chromosomes (Fig. 4A). Four hours after release from metaphase, chromatin had not yet decondensed completely, representing the early Gl phase. LLl antigen at this stage r-as distributed throughout the nucleoplasm with a speckled pattern (Figs. 4B and 4F). At this time, mass nucleolar structure was not distinct [proven by immunostaining with monoclonal antibody to nucleolin (data not shown)], supporting the idea that cells were at early Gl phase. When cells were further incubated for up to 12 h chromatin was decondensed completely (Fig. 4G) indicating that cells were in S phase of the cell cycle. At this moment, nucleolar staining by the LLl antibody was apparent (Fig. 4C). The accumulation of LLl antigen in nucleoli was even more marked at 20 h of incubation when cells entered the late S phase (Fig. 4D). [Similar results were obtained (data not shown) when the same system was checked by monoclonal antibody BL9 which was known to recognize a different epitope on the same LLl protein.] In order to confirm the phenomenon that LLl indeed
OF Ku80
INTO
NUCLEOLI
265
migrated into nucleoli during cell-cycle progression, a second method of cell synchronization was used. Cells were arrested at the Gl/S boundary by sequential treatments with thymidine and aphidicolin [lo]. LLl protein was not detectable within nucleoli, but only in nucleoplasm of the Gl/S cells (Fig. 5A). Two h after removal of the drugs, LLl started to move into nucleoli (data not shown). This gradual accumulation became very apparent at 6 h when the immunofluorescence intensity of LLl in nucleoli was about the same as that in nucleoplasm (Fig. 5B). This result was similar to that of the cells at 12 h after releasing from M-phase synchronization (Fig. 4C). Once cells progressed further into late S phase, 10 h after releasing from the Gl/S boundary, the accumulation of LLl in nucleoli was remarkable (Fig. 5C). This observation was consistent with that of cells incubated for 20 h after releasing from M-phase arrest (Fig. 4D). After further incubation of the Gl/S synchronized cells (up to 13 h) some of the cells had undergone mitosis with the characteristic of condensed chromosomes (Fig. 5H). At this time point, most cells without condensed chromosomes were assumed to be at the late S or G2 phase, and their LLl protein still remained in nucleoli in large amounts (Fig. 5D). [Again, these results were confirmed by the BL9 antibody (data not shown).] It seems that migration of LLl into nucleoli is cell-cycle dependent.
Constant Expression of LLl Protein in Cell Cycle We then explored whether the overall expression of LLl protein was altered at different phases of the cell cycle. Cell extracts were prepared from asynchronous cells, interphase (Gl + S + G2) cells, or GO/Gl-, Gl-, and M-phase-arrested cells. These extracts, containing equal quantities of total proteins, were analyzed by immunoblotting. Results showed that all extracts prepared from cells in different phases of the cell cycle had virtually identical amounts of LLl protein (Fig. 6). Therefore, LLl protein remained constantly during the cell cycle, and only its subnuclear localization was changed. DISCUSSION The Ku nuclear protein was originally discovered with autoantibodies in sera of patients with scleroderma-polymyositis overlap syndrome [ 161, and more recently by monoclonal antibodies developed when mice were immunized with isolated nucleoli [8] or nuclei [5]. Anti-Ku antibodies precipitated two polypeptides with molecular weights of 80 and 70 kDa (~801~70) and a nucleic acid identified as DNA [5, 171. Further experiments showed that the Ku protein complex bound to free ends of linear double-stranded DNA [18] and unfolded nucleosomes present in the DNase I-sensitive re-
266
LI AND
YEH
FIG. 4. Sequential redistribution of LLl protein after cells released from M-phase arrest. HeLa cells were treated with nocodazole. Mitotic “round-up” cells were collected by mechanical shaking and cultured in normal medium on glass slides. At various times throughout the cell cycle, slides were fixed and processed for immunofluorescence staining as described under Materials and Methods. The timepoints are 0 h (A, E), 4 h (B, F), 12 h (C, G), and 20 h (D, H). A, B, C, and D are the FITC immunofluorescence micrographs, and E, F, G, and H are the corresponding Hoechst 33258 counterstain micrographs.
gions of chromatin [19]. These results imply that Ku may be involved in DNA repair and in maintenance of nucleosomes at the decondensed conformation to permit transcription and replication. The cDNA-derived amino acid sequences of both the 80- and the 70-kDa subunits have regions similar to the leucine zipper structure [3, 4, 201, coincident with characteristics of many known DNA-binding proteins for transcriptionactivation and transforming activities [21, 221. How-
ever, precise functions and biochemical mechanisms of the Ku complex are still unknown. We found that Ku80 was sorted gradually to nucleoli during cell-cycle progression, although the overall amount of Ku80 maintained constantly. These results suggest that additional functions of Ku80 may relate to the nucleolar activities. Nucleoli have no limiting membrane. It is not known how nucleoli of interphase cells maintain their normally compact structure in the highly
267
FIG. 5. Sequential redistribution of LLl boundary by treatments with thymidine and followed by fixation and immunofluorescence 13 h (D, H). A, B, C, and D are the FITC counterstain micrographs.
protein after cells released from the Gl/S boundary. HeLa cells were synchronized at the Gl/S aphidicolin. Cells were submitted to recovery by culturing in normal medium for various times staining with LLl monoclonal antibody. The timepoints are 0 h (A, E), 6 h (B, F), 10 h (C, G), and immunofluorescence micrographs, and E, F, G, and H are the corresponding Hoechst 33258
268
LI AND
kDa
12345
YEH
Ku70 and both proteins can be phosphorylated at serine residues [19]. Whether a cell cycle-dependent phosphorylationldephosphorylation of the Ku complex would control its nucleolar destination is an interesting question. It requires further studies at the biochemical level. This work was supported by Research Grants NSC 79-0412-BOlO50 and NSC 80-0412-BOlO-46 from the National Science Council of the Republic of China. We thank Dr. Tsung-Sheng Su for providing us with the hepatoma cell line HA22T cDNA library.
REFERENCES 1. 2. FIG. 6. Constant expression of LLl protein in cell cycle. HeLa cells were synchronized at M phase by nocodazole treatment followed by 4 h recovery at 37°C (Gl phase), or arrested at GO/G1 by serum starvation for 3 days. Cell extracts were prepared in extraction buffer as described under Materials and Methods; 100 pg of total protein from each of the samples was subjected to SDS-PAGE under reducing conditions. Immunoblotting was performed by using LLl monoclonal antibody. (Lane 1) asynchronous cells, (lane 2) M phase, (lane 3) interphase (Gl + S + G2), (lane 4) Gl phase, and (lane 5) GO/G1 boundary.
ordered arrangement of all nucleolar components including chromatin, nucleolar-specific ribonucleoproteins, preribosomes, and characteristic proteins. However, this nucleolar structure is rapidly disintegrated in prophase by unknown mechanisms [23]. It would be interesting to investigate whether Ku80 might be involved in nucleolar dynamic changes in terms of their disintegration due to the observation that the maximum accumulation of Ku80 in nucleoli was at late S and G2 phases (Figs. 4 and 5). This pattern was quite different from that of cyclin, the proliferating cell nuclear antigen, which was found to be absent in nucleoli throughout the cell cycle with the exception of a very short stage at which DNA synthesis was near a maximum [24]. Cyclin became undetectable in nucleoli thereafter. Based on our up-to-date knowledge, Ku80 is the only protein expressing the unique feature of gradual accumulation in nucleoli up to the maximum concentration prior to cells undergoing mitosis and disintegration of nucleoli. It is well known that the 70-kDa heat shock protein (HSP70) became concentrated in nuclei of early S phase cells [25], but accumulated within nucleoli only upon heat shock [26, 271. The presence of HSP70 in nucleoli correlated with the acceleration of recovery of normal nucleolar morphology and the export of ribosomes after heat shock [28]. The dynamic state and roles in nucleoli of HSP70 are therefore completely different from those of Ku80. Mechanisms of nucleolar targetting of Ku80 are unclear. Ku80 is known to be tightly associated with Received September 17, 1991 Revised version received November
14, 1991
3. 4.
5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.
Sommerville, J. (1986) Trends Biochem. Sci. 11, 438. Chen, C.-M., Chiang, S-Y., and Yeh, N.-H. (1991) J. Biol. Chem. 266,7754. Yaneva, M., Wen, J., Ayala, A., and Cook, R. (1989) J. Biol. Chem. 264, 13,407. Mimori, T., Ohosone, Y., Hama, N., Suwa, A,, Akizuki, M., Homma, M., Griffith, A. J., and Hardin, J. A. (1990) Proc. N&l. Acad. Sci. USA 87,1777. Reeves, W. H. (1985) J. Exp. Med. 161, 18. Francoeur, A.-M., Peebles, C. L., Gompper, P. T., and Tan, E. M. (1986) J. Immunol. 136, 1648. Reeves, W. H. (1987) J. Rheumatol. 14(Suppl. 13), 97. Yaneva, M., Ochs, R., McRorie, D. K., Zweig, S., and Busch, H. (1985) Biochim. Biophys. Acta 841, 22. Pines, J., and Hunter, T. (1989) Cell 58,833. Heintz, N., Sive, H. L., and Roeder, R. G. (1983) Mol. Cell. Biol. 3,539. Kohler, G., and Milstein, C. (1975) Nature 256, 495. Yeh, N.-H., Chen, C.-M., andHsieh, S.-L. (1988) Chin. J. Microbiol. Zmmunol. 21, 210. Young, R. A., and Davis, R. W. (1983) Proc. Natl. Acad. Sci. USA 80, 1194. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74,5463. Laemmli, U. (1970) Nature 227,680. Mimori, T., Akizuki, M., Yamagata, H., Inada, S., Yoshida, S., and Homma, M. (1981) J. Clin. Invest. 68, 611. Mimori, T., Hardin, J. A., and Steitz, J. A. (1986) J. Biol. Chem. 261,2274. Mimori, T., and Hardin, J. A. (1986) J. Biol. Chem. 261,10,375. Yaneva, M., and Busch, H. (1986) Biochemistry 26,5057. Reeves, W. H., and Sthoeger, Z. M. (1989) J. Biol. Chem. 264, 5047. Turner, R., and Tjian, R. (1989) Science 243,1689. Kouzarides, T., and Ziff, E. (1989) Nature 340,568. Goessens, G. (1984) Int. Rev. Cytol. 87,107. Celis, J. E., and Celis, A. (1985) Proc. Nutl. Acad. Sci. USA 82, 3262. Milarski, K. L., and Morimoto, R. I. (1986) Proc. Natl. Acad. Sci. USA 83,9517. Milarski, K. L., Welch, W. J., and Morimoto, R. I. (1989) J. Cell. Biol. 108, 413. Welch, W. J., and Feramisco, J. R. (1984) J. Btil. Chem. 259, 4501. Pelham, H. R. B. (1984) EMBO J. 3,3095.
CELL
RESEARCH
199,
262-268 (1992)
Cell Cycle-Dependent Migration of the DNA-Binding Protein Ku80 into Nucleoli LI-LAN Graduate Institute
of Microbiology
LI AND NING-HSING
and Immunology,
National
Inc.
INTRODUCTION
The distinct dynamic changes of nucleolar structure are connected to the different phases of the cell cycle and the overall growth rates of cells [ 11. Little is known about whether particular nucleolar components are responsible for these changes. We have previously demonstrated that the expression of nucleolin, the major nucleolar phosphoprotein, is dynamic and correlates well with the degree of cell proliferation. The mechanism of
$3.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
College, Taiwan,
MATERIALS
Republic of China
AND
METHODS
Cell cultures. Human cervical carcinoma HeLa, T leukemia CEM, and hepatoma HA22T cells were cultured in RPM1 1640 medium supplemented with 10% heat-inactivated fetal calf serum (FCS) (Boehringer-Mannheim, Mannheim, Germany). These cells were grown at 37°C with 5% CO,. HeLa and HA22T cells were passaged with trypsin/EDTA to maintain subconfluency. CEM cells were subcultured twice every week. For use in serum starvation studies, HeLa cells (5 X lo5 cells/ml) were placed in RPM1 1640 containing 0.5% heat-inactivated FCS and allowed to quiesce for 3 days. Synchronization procedures. To accumulate HeLa cells in mitotic phase [9], exponentially growing monolayers were maintained for 12 h in medium containing nocodazole (Sigma Chemical Co., St. Louis, MO) at 0.4 pg/ml, after which mitotic cells were obtained by “shake-
I To whom reprint requests should be addressed at: Graduate Institute of Microbiology and Immunology, National Yang-Ming Medical College, Shih-Pai, 11221 Taipei, Taiwan, Republic of China. Fax: (02) 821-2880. 0014-4827/92
Medical
this phenomenon was proven due to the self-cleaving activity of nucleolin and a putative proteolytic inhibitor, present only in proliferating cells, to prevent the cleavage [2]. The instability of nucleolin in nondividing cells may account in part for the degeneration of nucleoli in these cells. In attempts to search out more nucleolar proteins in relation to nucleolar dynamic changes, we generated monoclonal antibodies against purified nuclei. One antibody, termed LLl, particularly aroused attention since the immunofluorescence staining of cells showed variations in intensity within the nucleoli. Human cDNA clones were isolated by immunoscreening, and a partial nucleotide sequence was determined. The sequence was found to be identical to that of a previously reported cDNA encoding the 80-kDa subunit of the Ku DNA-binding protein [3, 41. Immunofluorescence staining revealed that Ku was distributed throughout the nucleoplasm and often also in the nucleoli of interphase cells [5, 61. However, Ku was rapidly dissociated from DNA and appeared at the periphery of the condensed chromosomes during mitosis [7]. Although a cell cycle-dependent distribution of Ku in the nucleoli was suggested [8], the direct experimental evidence was still lacking. We report here that nucleoli of cells at the Gl/S boundary had very small amounts of Ku80 protein. It appeared gradually in nucleoli during the cell cycle progression and reached a maximum at late S or G2 phase. Whether this cell cycledependent migration of Ku80 into nucleoli is responsible for nucleolar dynamic changes needs to be extensively studied.
The DNA-binding protein Ku (p7O/p80) was originally discovered through the use of human autoimmune sera. In attempts to search out nucleolar proteins in relation to nucleolar dynamic changes, we developed monoclonal antibodies against nuclear proteins. One antibody, termed LLl, received particular attention since asynchronous cells exhibited tremendous differences in their nucleolar fluorescence intensities after immunostaining. The LLl protein was proven to be the Ku subunit ~80 (Ku80) by cDNA cloning and sequencing. Possible correlations between the heterogeneous distribution of Ku80 in nucleoli and the cell cycle were examined. HeLa cells were synchronized at M phase by arrest with nocodazole, or at the GUS boundary by sequential treatments with thymidine and aphidicolin. These cells were then released by culturing in fresh medium to allow the cell cycle to progress synchronously. Immunofluorescent detection of Ku80 revealed that nucleoli of the cells at the GllS boundary had very small amounts of Ku80, which was mainly present in the nucleoplasm. Ku80 was gradually accumulated in nucleoli during S phase and reached the maximum at late S or G2 phase. Immunoblotting experiments showed that cell extracts prepared from different phases of the cell cycle had virtually identical amounts of Ku80. These results suggest that Ku80 migrates from nucleoplasm to 0 1992 Academic nucleoli in a cell cycle-dependent manner. Press,
Yang-Ming
YEH’
262
CELL
CYCLE-DEPENDENT
MIGRATION
off’. Interphase cells were collected by trypsin/EDTA treatment of monolayers after vigorous washing to remove all mitotic cells. About 85% of these mitotic cells showed condensed chromosomes as determined by DNA staining. These mitotic cells were washed with phosphate-buffered saline (PBS) and driven into interphase by culturing in fresh medium on glass slides. Cells were used for immunofluorescence staining at different time intervals. Cell extracts were also prepared from the above interphase, mitotic-phase, and Gl-phase cells (4 h after release from M phase). A second protocol was used to arrest cells at the Gl/S boundary by sequential treatments with thymidine and aphidicolin [lo]. HeLa cells were blocked with 2 n&f thymidine for 12 h. Cells were rinsed with PBS to remove thymidine and then exposed to medium in the presence of 2.4 X 10e5 M thymidine and deoxycytidine for 9 h. Aphidicolin (Sigma) was added to 5 pg/ml. After culturing for an additional 12 h, cells were blocked at the Gl/S boundary. Reversal of the Gl/S blockade was accomplished by washing the cells twice with PBS, and followed by suspension in complete medium. These cells were grown on glass slides and taken at various time points for immunofluorescence staining. Monoclonal antibodies. Balb/c mice were hyperimmunized with purified CEM nuclei (purity > 95%). Procedures for the preparation of nuclei were identical to those described previously [2]. Spleen cells of these mice were fused with NS-1 myeloma cells by standard techniques [ll]. Hybridoma supernatants were prescreened by a dot-immunobinding assay [ 121 to select clones secreting monoclonal antibodies against immunogens. Those positive hybrids were rescreened by indirect immunofluorescence microscopy. Monoclonal antibodies recognizing nucleolar components were retained; two of these, LLl and BL9, showed nucleolar positive reactions only in a proportion of the normal cell populations and therefore were used for further investigation. Zmmunoblotting. HeLa, CEM, HA22T, or synchronized HeLa cells were Iysed in extraction buffer [l% Triton X-100,0.2% deoxycholate, 0.6 M NaCl, 0.1% sodium dodecyl sulfate (SDS), 10 m&f TrisHCl, pH 8.0,20 mM 2-mercaptoethanol, 10 mM KCI, 0.5 mM MgCl,]. The lysates were sonicated to disrupt DNA and then clarified by centrifugation at 13,000g for 20 min. Samples were immediately stored at -70°C. Samples containing 100 pg total proteins were analyzed by SDS-7.5% polyacrylamide gel electrophoresis (PAGE). Proteins were transferred from gel to nitrocellulose membrane (Bio-Rad Laboratories, Richmond, CA) at 0.8 mA/cm* for 3 h with a NovaBlot electrophoretic transfer apparatus (LKB). Nonspecific protein binding was blocked by incubating the membrane in 50 n&f Tris-HCl, pH 7.5,150 mM NaCl (TBS) containing 3% nonfat milk for 2 h. The membrane was reacted with LLl or BL9 antibody for 1 h and then extensively washed with TBS. A 2000-fold diluted peroxidase-conjugated goat anti-mouse IgG (PXGAMIgG) (Bio-Rad) in TBS containing 3% nonfat milk was added to the membrane for 1 h. After washing with TBS, specific bands were visualized by reaction with a solution freshly prepared by dilution of the stock of 4-chloro-1-naphthol (3 mg/ml in methanol) (Merck & Co., Rahway, NJ) with 5 vol of TBS containing 0.02% H,O,. Isolation of cDNA clones. LLl antibody was used to screen a human hepatoma HA22T Xgtll cDNA library (kindly provided by Dr. T.-S. Su, Veterans General Hospital, Taipei, Republic of China). Methods for screening were based on the protocols reported by Young and Davis [13]. We modified only the last step for detection of positive plaques on nitrocellulose membrane by using PXGAMIgG and color reaction as described above, under “Immunoblotting.” Three positive cDNA clones, L3, L4, and L5, were plaque purified and characterized further by restriction mapping and sequencing. DNA sequencing. Three cDNA inserts from the L3, L4, and L5 clones were subcloned into the phagemid pBluescript II SK+ (Stratagene) in both orientations. Single-stranded DNA templates were rescued by infection of the recombinant phagemid-transformed Escherichia coli strain XLl-blue with helper phage VCSM13 according to the instructions of the manufacturer. Sequencing reactions were
OF Ku80
INTO
263
NUCLEOLI
carried out by the dideoxy chain termination method [14] with a Sequenase kit (United States Biochemical Corp.). Sequence analysis was performed with the PC/gene computer program (IntelliGenetics, Mountain View, CA). E. coli strain Y1089 was lysogenized Expression of fusionproteins. with the L3, L4, or L5 recombinant phage. Lysogens were grown in 2 ml LB medium containing 2 mM isopropyl thio-j?-D-galactopyranoside (Clontech), 2 mM MgCl,, 0.2% maltose, and 50 ag/ml of ampicillin, for 3.5 h at 37°C on a shaker. Bacteria were collected by centrifugation and lysed in 100 ~1 Laemmli reducing sample buffer [15]. Samples were boiled for 4 min and centrifuged to remove debris. Ten microliters of the supernatant was analyzed on SDS-PAGE. Indirect immunofluorescence. Cells grown on slides at various time intervals after synchronization were fixed with 2% formaldehyde in PBS for 20 min at room temperature. After washing once with PBS, cells were permeabilized with -2O’C acetone for 3 min and rinsed with PBS. Fifty microliters hybridoma supernatant was added onto cells and the slides were placed in a moisture chamber for 1 h. Cells were washed three times in PBS (5 min for each wash) and then incubated for 1 h with 50 81 of a 150-fold-diluted second antibody, Auorescein isothiocyanate-conjugated goat anti-mouse IgG (Cappel), in 1% bovine serum albumin/PBS. A DNA dye, Hoechst 33258 (Sigma), was added to 1 pg/ml during this second incubation. Cells were washed three times with PBS and then mounted in a solution containing 0.42% glycine, 0.021% NaOH, 0.5% NaCl, 0.03% NaN,, and 70% glycerol for observation with an immunofluorescence microscope. RESULTS
Heterogeneous Distribution of LLl Antigen of Asynchronous Cell Populations
in Nucleoli
Spleen cells of Balb/c mice hyperimmunized with purified CEM nuclei were used to develop hybridomas. When hybridoma supernatants were screened by a highly sensitive method, i.e., dot-immunobinding assay [12], approximately 25% of hybrids were identified as antibody-secreters with specificities to immunogens. Only about half of these positive supernatants had nuclear or nucleolar reactivities determined by immunofluorescence staining. One of the nucleolar reactive monoclonal antibodies, LLl, was particularly noted since its target LLl molecules were present in different amounts in nucleoli of asynchronous cells. Results using normally growing HeLa cells as an example are shown in Fig. 1. The LLl antigen was scattered throughout the nucleoplasm of all cells. However, cells displayed different fluorescence intensities in the nucleoli ranging from undetectable or low levels (as indicated by arrows) to moderate or very high levels. This heterogeneity with respect to concentrations of LLl antigen in nucleoli was consistently observed in all other cell lines tested (data not shown). Experiments with a control anti-nucleolin antibody did not show heterogeneous patterns in nucleoli of the same cell populations (data not shown), indicating that variations of LLl molecules in nucleoli were not due to fixation or permeation artifacts. Characterization To identify nal antibody,
of the LLl
Protein by Immunoblotting
the protein recognized by LLl monoclocell extracts of HeLa, CEM, and HA22T
264
Ll AND
YEH
‘I’ -2
Ku80 L3
f
L4 K
-
-0
-
L5
*
0.5 Kb
B
LLl kDa
1234
BL9 12
34
FIG. 1. Heterogeneous distribution of LLl antigen in nucleoli of asynchronous cell populations. HeLa cells were grown on slides. Cells were fixed with 2% formaldehyde in PBS and permeabilized with cold acetone. Indirect immunofluorescence staining was performed with LLl monoclonal antibody. Note the variations of fluorescence intensity in nucleoli ranging from very high to very low (indicated by arrows) levels.
were analyzed by immunoblotting. band
was
identified
in all
cell
A very sharp 85kDa extracts
under
reducing
conditions (Fig. 2B). Without reduction, extra diffuse forms migrating slightly faster than the 85kDa band in SDS gels were observed (Fig. 2A). This observation suggests that intramolecular disulfide bonds are present, resulting in the fast movement of the 85kDa protein. Identification
of LLl
Protein as the Ku80 Subunit
In order to further characterize the LLl protein, we isolated three LLl cDNA clones by immunoscreening
A kDa
123
B
FIG. 3. Identification of LLl protein as the Ku80 subunit. (A) Three CDNA clones (~3, ~4, and ~5) were isolated by immunoscreening with LLl monoclonal antibody. Four stretches of nucleotide sequences (indicated by arrows) were determined and found to be completely identical to those for the Ku80 subunit. Restriction sites on the cDNA inserts were determined by digestion with EcoRV (E), C/o1 (C), PstI (P), SnaI (S), andKpn1 (K). Results show the alignment of these three cDNA inserts as corresponding to the sequence of the known full length K&O. Open reading frame of Ku80 is boxed. (B) Escherichia coli Y1089 lysogenized with L3, L4, or L5 recombinant hgtll were induced by IPTG to express fusion proteins. The bacterial lysates were analyzed by immunoblotting and probed with monoclonal antibody LLl (left) or BL9 (right). Lanes l-4 indicate samples for bacterial control (lane 1) or lysogens containing L3, L4, or L5 cDNA (lanes 2-4, respectively). A bacterial protein cross-reacted with BL9 antibody is noted by the arrow.
123
FIG. 2. Characterization of the LLl protein by immunoblotting. Cell extracts of HeLa, CEM, and HA22T (lanes l-3, respectively) were subjected to SDS-PAGE under nonreducing (A) or reducing (B) conditions. Results show immunoblotting probed with LLl monoclonal antibody. Molecular weight markers are expressed in kDa.
the hepatoma cell line HA22T cDNA library. By gel electrophoresis, sizes of the cDNA inserts of clones L3, L4, and L5 were estimated to be 3.0, 1.4, and 1.8 kb, respectively. Each end of the cDNA inserts was sequenced by the dideoxy-nucleotide method [ 141. These nucleotide sequences were found to be identical to those for the 80-kDa subunit protein of the human autoantigen Ku (p7O/p80) reported previously [4]. A partial restriction map and the sequence data were used to align these three cDNA clones to corresnond with the full length Ku80 (Fig. 3A). All three clones have spanned regions through the translational termination-site predicted in the Ku80-cDNA sequence, but lack 5’-end coding regions in different lengths.
CELL
CYCLE-DEPENDENT
MIGRATION
Fusion proteins were induced and their apparent molecular weights were used to confirm that the LLl cDNA clones carry stop codons at the same site as that of K&O. The sizes of the fusion proteins were determined as 185,160, and 135 kDa for cDNA clones L3, L4, and L5, respectively (Fig. 3B). Since fusion proteins contain the llO-kDa P-galactosidase, the L3, L4, and L5 cDNA inserts were estimated to encode 75, 50-, and 25-kDa peptides, respectively. These sizes matched perfectly if the stop codon was assumed to be at the same site as Ku80 (see Fig. 3A). We have concluded that the LLl protein is the Ku80 subunit of the autoantigen Ku. One interesting finding was noted when bacterial extracts containing the fusion proteins were probed with a second monoclonal antibody, BL9, which has been proven as recognizing the same LLl protein (unpublished data). This antibody also detected a bacterial protein with a molecular weight of 50 kDa (indicated by arrow) since bacterial control without infection by the recombinant phage Xgtll also had this band (Fig. 3B, lane 1 of panel BL9). LLl protein shared at least one epitope with the bacterial 50-kDa protein. The significance of this observation is unknown.
Cell Cycle-Dependent Migration of LLl to Nucleoli Asynchronous cells immunostained with LLl monoclonal antibody exhibited tremendous differences in their nucleolar fluorescence intensities (Fig. 1). Questions were raised about whether these variations correlated with the different stages of the cell cycle. Experiments of cell synchronization were therefore performed. Mitotic-phase-arrested HeLa cells were obtained by nocodazole treatment and mechanical shakeoff [9]. Greater than 85% of these cells had condensed chromosomes (Fig. 4E and data not shown). LLl protein in mitotic cells appeared to be completely away from chromosomes (Fig. 4A). Four hours after release from metaphase, chromatin had not yet decondensed completely, representing the early Gl phase. LLl antigen at this stage r-as distributed throughout the nucleoplasm with a speckled pattern (Figs. 4B and 4F). At this time, mass nucleolar structure was not distinct [proven by immunostaining with monoclonal antibody to nucleolin (data not shown)], supporting the idea that cells were at early Gl phase. When cells were further incubated for up to 12 h chromatin was decondensed completely (Fig. 4G) indicating that cells were in S phase of the cell cycle. At this moment, nucleolar staining by the LLl antibody was apparent (Fig. 4C). The accumulation of LLl antigen in nucleoli was even more marked at 20 h of incubation when cells entered the late S phase (Fig. 4D). [Similar results were obtained (data not shown) when the same system was checked by monoclonal antibody BL9 which was known to recognize a different epitope on the same LLl protein.] In order to confirm the phenomenon that LLl indeed
OF Ku80
INTO
NUCLEOLI
265
migrated into nucleoli during cell-cycle progression, a second method of cell synchronization was used. Cells were arrested at the Gl/S boundary by sequential treatments with thymidine and aphidicolin [lo]. LLl protein was not detectable within nucleoli, but only in nucleoplasm of the Gl/S cells (Fig. 5A). Two h after removal of the drugs, LLl started to move into nucleoli (data not shown). This gradual accumulation became very apparent at 6 h when the immunofluorescence intensity of LLl in nucleoli was about the same as that in nucleoplasm (Fig. 5B). This result was similar to that of the cells at 12 h after releasing from M-phase synchronization (Fig. 4C). Once cells progressed further into late S phase, 10 h after releasing from the Gl/S boundary, the accumulation of LLl in nucleoli was remarkable (Fig. 5C). This observation was consistent with that of cells incubated for 20 h after releasing from M-phase arrest (Fig. 4D). After further incubation of the Gl/S synchronized cells (up to 13 h) some of the cells had undergone mitosis with the characteristic of condensed chromosomes (Fig. 5H). At this time point, most cells without condensed chromosomes were assumed to be at the late S or G2 phase, and their LLl protein still remained in nucleoli in large amounts (Fig. 5D). [Again, these results were confirmed by the BL9 antibody (data not shown).] It seems that migration of LLl into nucleoli is cell-cycle dependent.
Constant Expression of LLl Protein in Cell Cycle We then explored whether the overall expression of LLl protein was altered at different phases of the cell cycle. Cell extracts were prepared from asynchronous cells, interphase (Gl + S + G2) cells, or GO/Gl-, Gl-, and M-phase-arrested cells. These extracts, containing equal quantities of total proteins, were analyzed by immunoblotting. Results showed that all extracts prepared from cells in different phases of the cell cycle had virtually identical amounts of LLl protein (Fig. 6). Therefore, LLl protein remained constantly during the cell cycle, and only its subnuclear localization was changed. DISCUSSION The Ku nuclear protein was originally discovered with autoantibodies in sera of patients with scleroderma-polymyositis overlap syndrome [ 161, and more recently by monoclonal antibodies developed when mice were immunized with isolated nucleoli [8] or nuclei [5]. Anti-Ku antibodies precipitated two polypeptides with molecular weights of 80 and 70 kDa (~801~70) and a nucleic acid identified as DNA [5, 171. Further experiments showed that the Ku protein complex bound to free ends of linear double-stranded DNA [18] and unfolded nucleosomes present in the DNase I-sensitive re-
266
LI AND
YEH
FIG. 4. Sequential redistribution of LLl protein after cells released from M-phase arrest. HeLa cells were treated with nocodazole. Mitotic “round-up” cells were collected by mechanical shaking and cultured in normal medium on glass slides. At various times throughout the cell cycle, slides were fixed and processed for immunofluorescence staining as described under Materials and Methods. The timepoints are 0 h (A, E), 4 h (B, F), 12 h (C, G), and 20 h (D, H). A, B, C, and D are the FITC immunofluorescence micrographs, and E, F, G, and H are the corresponding Hoechst 33258 counterstain micrographs.
gions of chromatin [19]. These results imply that Ku may be involved in DNA repair and in maintenance of nucleosomes at the decondensed conformation to permit transcription and replication. The cDNA-derived amino acid sequences of both the 80- and the 70-kDa subunits have regions similar to the leucine zipper structure [3, 4, 201, coincident with characteristics of many known DNA-binding proteins for transcriptionactivation and transforming activities [21, 221. How-
ever, precise functions and biochemical mechanisms of the Ku complex are still unknown. We found that Ku80 was sorted gradually to nucleoli during cell-cycle progression, although the overall amount of Ku80 maintained constantly. These results suggest that additional functions of Ku80 may relate to the nucleolar activities. Nucleoli have no limiting membrane. It is not known how nucleoli of interphase cells maintain their normally compact structure in the highly
267
FIG. 5. Sequential redistribution of LLl boundary by treatments with thymidine and followed by fixation and immunofluorescence 13 h (D, H). A, B, C, and D are the FITC counterstain micrographs.
protein after cells released from the Gl/S boundary. HeLa cells were synchronized at the Gl/S aphidicolin. Cells were submitted to recovery by culturing in normal medium for various times staining with LLl monoclonal antibody. The timepoints are 0 h (A, E), 6 h (B, F), 10 h (C, G), and immunofluorescence micrographs, and E, F, G, and H are the corresponding Hoechst 33258
268
LI AND
kDa
12345
YEH
Ku70 and both proteins can be phosphorylated at serine residues [19]. Whether a cell cycle-dependent phosphorylationldephosphorylation of the Ku complex would control its nucleolar destination is an interesting question. It requires further studies at the biochemical level. This work was supported by Research Grants NSC 79-0412-BOlO50 and NSC 80-0412-BOlO-46 from the National Science Council of the Republic of China. We thank Dr. Tsung-Sheng Su for providing us with the hepatoma cell line HA22T cDNA library.
REFERENCES 1. 2. FIG. 6. Constant expression of LLl protein in cell cycle. HeLa cells were synchronized at M phase by nocodazole treatment followed by 4 h recovery at 37°C (Gl phase), or arrested at GO/G1 by serum starvation for 3 days. Cell extracts were prepared in extraction buffer as described under Materials and Methods; 100 pg of total protein from each of the samples was subjected to SDS-PAGE under reducing conditions. Immunoblotting was performed by using LLl monoclonal antibody. (Lane 1) asynchronous cells, (lane 2) M phase, (lane 3) interphase (Gl + S + G2), (lane 4) Gl phase, and (lane 5) GO/G1 boundary.
ordered arrangement of all nucleolar components including chromatin, nucleolar-specific ribonucleoproteins, preribosomes, and characteristic proteins. However, this nucleolar structure is rapidly disintegrated in prophase by unknown mechanisms [23]. It would be interesting to investigate whether Ku80 might be involved in nucleolar dynamic changes in terms of their disintegration due to the observation that the maximum accumulation of Ku80 in nucleoli was at late S and G2 phases (Figs. 4 and 5). This pattern was quite different from that of cyclin, the proliferating cell nuclear antigen, which was found to be absent in nucleoli throughout the cell cycle with the exception of a very short stage at which DNA synthesis was near a maximum [24]. Cyclin became undetectable in nucleoli thereafter. Based on our up-to-date knowledge, Ku80 is the only protein expressing the unique feature of gradual accumulation in nucleoli up to the maximum concentration prior to cells undergoing mitosis and disintegration of nucleoli. It is well known that the 70-kDa heat shock protein (HSP70) became concentrated in nuclei of early S phase cells [25], but accumulated within nucleoli only upon heat shock [26, 271. The presence of HSP70 in nucleoli correlated with the acceleration of recovery of normal nucleolar morphology and the export of ribosomes after heat shock [28]. The dynamic state and roles in nucleoli of HSP70 are therefore completely different from those of Ku80. Mechanisms of nucleolar targetting of Ku80 are unclear. Ku80 is known to be tightly associated with Received September 17, 1991 Revised version received November
14, 1991
3. 4.
5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.
Sommerville, J. (1986) Trends Biochem. Sci. 11, 438. Chen, C.-M., Chiang, S-Y., and Yeh, N.-H. (1991) J. Biol. Chem. 266,7754. Yaneva, M., Wen, J., Ayala, A., and Cook, R. (1989) J. Biol. Chem. 264, 13,407. Mimori, T., Ohosone, Y., Hama, N., Suwa, A,, Akizuki, M., Homma, M., Griffith, A. J., and Hardin, J. A. (1990) Proc. N&l. Acad. Sci. USA 87,1777. Reeves, W. H. (1985) J. Exp. Med. 161, 18. Francoeur, A.-M., Peebles, C. L., Gompper, P. T., and Tan, E. M. (1986) J. Immunol. 136, 1648. Reeves, W. H. (1987) J. Rheumatol. 14(Suppl. 13), 97. Yaneva, M., Ochs, R., McRorie, D. K., Zweig, S., and Busch, H. (1985) Biochim. Biophys. Acta 841, 22. Pines, J., and Hunter, T. (1989) Cell 58,833. Heintz, N., Sive, H. L., and Roeder, R. G. (1983) Mol. Cell. Biol. 3,539. Kohler, G., and Milstein, C. (1975) Nature 256, 495. Yeh, N.-H., Chen, C.-M., andHsieh, S.-L. (1988) Chin. J. Microbiol. Zmmunol. 21, 210. Young, R. A., and Davis, R. W. (1983) Proc. Natl. Acad. Sci. USA 80, 1194. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74,5463. Laemmli, U. (1970) Nature 227,680. Mimori, T., Akizuki, M., Yamagata, H., Inada, S., Yoshida, S., and Homma, M. (1981) J. Clin. Invest. 68, 611. Mimori, T., Hardin, J. A., and Steitz, J. A. (1986) J. Biol. Chem. 261,2274. Mimori, T., and Hardin, J. A. (1986) J. Biol. Chem. 261,10,375. Yaneva, M., and Busch, H. (1986) Biochemistry 26,5057. Reeves, W. H., and Sthoeger, Z. M. (1989) J. Biol. Chem. 264, 5047. Turner, R., and Tjian, R. (1989) Science 243,1689. Kouzarides, T., and Ziff, E. (1989) Nature 340,568. Goessens, G. (1984) Int. Rev. Cytol. 87,107. Celis, J. E., and Celis, A. (1985) Proc. Nutl. Acad. Sci. USA 82, 3262. Milarski, K. L., and Morimoto, R. I. (1986) Proc. Natl. Acad. Sci. USA 83,9517. Milarski, K. L., Welch, W. J., and Morimoto, R. I. (1989) J. Cell. Biol. 108, 413. Welch, W. J., and Feramisco, J. R. (1984) J. Btil. Chem. 259, 4501. Pelham, H. R. B. (1984) EMBO J. 3,3095.
