- Email: [email protected]
PCNA in synchronized 3T3 cells
Experimental
Ultrastructural
Cell Research 184 (1989) 81-89
lmmunolocalization of Cyclin/PCNA Synchronized 3T3 Cells’
in
IVAN RASKA,*,* KAREL KOBERNA,* MICHAL JARNiK,* VALBRIA PETRASOVICOVA,t JAN BEDNAR,* KAREL RASKA Jr-,$ and RODRIGO BRAVO” *Institute of Experimental Medicine, Czechoslovak Academy of Sciences, Lidovych Milici es-120 00 Prague, Czechoslovakia; tResearch Institute of Rheumatic Diseases, CS-I28 00 Prague, Czechoslovakia; $lnstitute of Physiology, Czechoslovak Academy of Sciences, CS-142 20 Prague, Czechoslovakia; OUMDNJ, Robert Wood Johnson Medical School, Piscataway, New Jersey 08854; and I’European Molecular Biology Laboratory, D-6900 Heidelberg, West Germany
61,
The immunolocalization of cyclin/PCNA’ in synchronized 3T3 cells was performed with human autoantibodies using an immunogold technique performed on thin cryosections. Previous immunofluorescent studies demonstrated that the DNA replication sites correspond to the localization of bound cyclin. We have found that in the early periods of S phase, the DNA replication sites (or sites potentially ready for the replication during the hydroxyurea DNA synthesis block) are situated in the perichromatin region and correspond to clustered gold particles present frequently over a morphologically distinct small nuclear area. Heavily labeled chromocenters, including perinucleolar condensed chromatin, exhibiting several such distinct areas were found in later periods of S phase. @ 1989 Academic
Press, Inc.
The aim of this study was to localize cyclin/PCNA [Ilr] in synchronized 3T3 cells during various periods of S phase using an immunogold technique. Cyclin is an auxilliary protein of the DNA polymerase delta [5, 61 and it was shown in the previous biochemical and light microscopical studies that the bound cyclin exhibiting the punctuate immunofluorescent pattern and DNA replication sites have identical localization during the S phase [3, 4, 7-103. Bound cyclin represents 2&30% of all cyclin molecules present in the cell, whereas free cyclin is extracted during the methanol fixation used for immunofluorescence [7]. The localization of cyclin was investigated in two different experiments. In the first experiment we immunolocalized cyclin in 3T3 cells which were stimulated to grow. In the second experiment, DNA synthesis in stimulated cells was blocked by hydroxyurea and the immunolocalization of cyclin was performed after its removal [8].
’ This paper is dedicated to the memory of Dr. Karel RaSka. 2 To whom reprint requests should be addressed. 3 PCNA, proliferating cell nuclear antigen. 81
Copyright @ 1989 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827/89 $03.00
82
RaSka et al.
0
4 0 12 16 20 WI WI WI (34 (~1 hours after stlmulatlon (sample)
UN
0 WI
2 4 6 0 10 (13th) WW VW (1w WI houra after hydroxyurea twmoval (sample)
24 WI
12 (14th)
Fig. 1. Synthesis of DNA after serum stimulation of quiescent 3T3 cells (A) and after hydroxyurea release (B) as followed by [3H]thymidine incorporation (for details see [SJ). EM investigations were performed on cell samples of serum-stimulated cells at 4-h intervals (samples 1 to 7) and on samples of hydroxyurea-treated cells at 2-h intervals (samples 8 to 14).
MATERIALS
AND METHODS
3T3 cells arrested in Go phase were stimulated to growth by addition of 20% fetal calf serum and analyzed at 4-h intervals during a period of 24 h (samples 1 to 7) [8]. In a parallel experiment, the incorporation of [3H]thymidine was determined (Fig. 1; see also [S]). The cells were fixed in 8% paraformaldehyde, preembedded in a fibrin clot [ll], infused in 2.1 M sucrose, and frozen in liquid nitrogen [12], and cryosections were prepared according to Griffiths et al. (1984). Bound (purified) human autoantibodies, which were shown to react in Western blots monospecially with cyclin [13], were revealed by means of 9-nm gold particles stabilized with protein A [14]. In the double-labeling experiments, a monoclonal antibody to DNA [IS] was used (kind gift of Boehringer Co., Mannheim) together with the autoantibody. The bound antibodies were revealed with the help of antimouse S-nm (DNA) and antihuman IO-nm (cyclin) gold adducts (kind gift of Janssen Pharmaceutics Co., Beerse). Control labeling experiments included omission of the primary antibody, saturation of sections with
EM immunolocalization of cyclin
83
Fig. 2. A cryosection of a 3T3 cell of the 5th sample exhibiting a labeled DA (arrows) in the perichromatin region. C, cytoplasm. ~77,000.
free secondary immunoglobulins or protein A prior to application of gold adducts, parallel application of antibodies with different specificities (e.g., RNA polymerase I, ribosomal proteins which were specially enriched in the nucleolar body of the nucleus), and, the double-labeling experiments, concomitant or consecutive application of antibodies. Morphometric estimation of surface-density labeling of cyclin in single-labeling experiments was performed on at least 30 prints of electron micrographs for each sample [16]. In the second experiment, hydroxyurea (5 mM) was added 8 h after stimulation of 3T3 cells to stop progression from G, to S phase. After 12 h, the hydroxyurea was washed out and the cyclin localization was investigated in cells harvested at 2-h intervals up to the 12th h (samples 8 to 14). The incorporation of [.‘H]thymidine was followed in a parallel experiment (Fig. 1; see also [S]).
RESULTS In the first three samples (0, 4, and 8 h after stimulation) the cell nuclei were characterized by the presence of rather homogeneously distributed individual gold particles. This kind of labeling pattern remained in subsequent samples. With the onset of the DNA synthesis (Fig. 1) in the 4th sample (12 h after stimulation) and especially in the 5th sample (16 h after stimulation), the intensity of the label was increased particularly in the perichromatin region, and the first clustered particles (i.e., several gold particles confined to a small nuclear area) appeared. Clustered particles frequently became confined to regions in which the morphology exhibited a “looser” texture and apparently lower DNA content than the condensed chromatin, their electron density usually being higher than that of the interchromatin regions (Figs. 2 and 3). We refer to these distinct
84
RaSka et al.
Fig. 3. A double-labeling experiment performed on a 3T3 cell from the 5th sample. Note that the DA (arrow) is enriched in the cyclin (larger gold particles) but not in the DNA (smaller gold particles) label. C, Condensed chromatin; CY, cytoplasm. x 145,300.
regions with clustered gold particles as distinct areas (DAs). In the 6th and 7th samples (20 and 24 h after stimulation), which exhibited the highest density of gold particles (increased threefold in comparison to the 1st sample), labeling within chromocenters including perinucleolar condensed chromatin was observed as well (Fig. 4). However, the label confined to areas delimited by chromocenters was not homogeneous and corresponded to the presence of several DAs (Figs. 4 and 5). The number of clustered particles represented, on average, 25% of the whole nuclear label. This, in fact, correponds well to the known percentage of bound cyclin [7]. The labeling of the nucleolar body was lower with respect to the overall nuclear label and consisted mainly of individual particles (Figs. 4 and 5). After the removal of hydroxyurea, the labeling picture in samples g-14 (0, 2, 4, 6, 8, 10, and 12 h after hydroxyurea removal) was different from that observed in cells not exposed to hydroxyurea. The difference was due mainly to the overall higher incidence of gold particles and to the fact that the clustered particles (and DAs) were already observed in the 8th sample when the hydroxyurea was just washed away. The intensity of labeling in the 8th sample was comparable to that in the 6th sample. The maximum level of labeling was already reached in the 10th and 1lth samples (fourfold increase in comparison to 1st sample). The label was qualitatively comparable to-but higher than-that of samples 6 and 7 (Fig. 6).
EM immunolocalization of cyclin
85
Fig. 4. A low-magnification picture of a 3T3 cell cryosection of the 7th sample exhibiting several labeled DAs (arrows) within the chromocenter and the perinucleolar condensed chromatin of two nucleoli. In a few DAs, gold particles follow the border of condensed chromatin. C, Cytoplasm: N, nucleolus; G, interchromatin granules. ~35,000.
86
RaSka et al.
Fig. 5. A higher magnificantion of the cryosection shown in Fig. 4. DAs (arrows) are well discernible within the perinucleolar condensed chromatin (P) at this magnification. In the nucleolar body (N), only a few individual particles are present. ~52,800.
DISCUSSION In this study we have attempted to relate functional events of DNA replication to immunocytochemical localization of cyclin. Because we used formaldehyde rather than methanol fixation we could detect both free and bound cyclin [7]. From the qualitative and quantitative character of the changes, and also compared to those seen in fluorescent studies (punctuate, granular, and nucleolar fluorescence; see also [7, S]), we propose that the clustered gold particles
EM immunolocalization
- --
_-.
of cyclin
81
--
Fig. 6 An intense labeling of perinucleolar condensed chromatin (P) in late S phase after the release of cells from hydroxyurea block. Note the uneven distribution of gold particles within the area delimited by the perinucleolar condensed chromatin. Clustered particles identify several DAs (arrowheads). The nucleolar body (N) is almost not labeled. ~46,200.
frequently present over DAs identify sites related to DNA synthesis. However, we do not claim that all replication sites can be identified as DAs. The high-resolution autoradiographic studies dealing with the DNA replication yield rather scattered data (for review see [17, 181). Whereas clustered particles (and DAs) were situated mainly in the perichromatin regions in the early periods of the S phase, an intensive but not homogeneous labeling (confined to DAs) within the chromocenters could be observed as well in the later periods of the S
88
RaSka et al.
phase in the present study. Our results are thus in agreement with those autoradiographic data [17, 181situating DNA replication sites in the regions of extended chromatin (euchromatin) during the early S phase, which then also spread to heterochromatin in the later S phase. Due to the low labeling density within the nucleolar body, we were not usually able to distinguish between the bound and the free cyclin within this nuclear organelle. This is understandable, as the bulk of nucleolar DNA is contined to the perinucleolar condensed chromatin, the concentration of DNA within the nuclear body being much lower (e.g., [ 191). The previously described nucleolar immunofluorescent staining in 3T3 cells [7, 81 apparently mimics the intensive labeling of the perinucleolar chromatin as observed in this study. In this sense, the replication of nucleolar DNA can be positioned at a later period of S phase. The labeled DAs in the 8th sample, in which the DNA synthesis was still suppressed, are interpreted as sites potentially ready for replication since the DNA synthesis is very rapid after the release of the hydroxyurea block (Fig. 1; see also [S]). Cyclin, the synthesis of which is not blocked by hydroxyurea treatment, may be present in structures with a lower DNA content [20], apparently the nuclear matrix [21, 221. This is supported by the observation that the concentration of DNA is diminished over DAs (cf. Fig. 3). The presence of DNA in DAs is not, however, ruled out, as the antibody used does not react substantially with the decondensed (relaxed) chromatin [I51 as one would expect for replicating DNA. The appearance of DAs illustrates that a functional event of DNA replication (or leading potentially to DNA replication) induces subtle morphological changes which have not been detected by ultrastructural methods until now. In addition, the existence of several DAs within individual chromocenters during the later periods of S phase may represent a morphological correlate in situ of several replication units present on one or several chromosomes. We thank Dr. Y. Takasaki for providing us with human serum (A.K.) anti-PCNA. The authors are obliged to Dr. G. Griftiths and Ruth Back (EMBL, Heidelberg) for their suggestions and critical help in preparing the manuscript. I.R. is particularly grateful to the Herbette Foundation which subsidized his stay at the Laboratory of Ultrastructural Analysis of the University of Lausanne where a part of this study was performed.
REFERENCES I. 2. 3. 4. 5. 6. 7. 8. 9. 10.
Miyashi, K., Fritzler, M. J., and Tan, E. M. (1978) J. Immunol. 121, 2228-2234. Tan, E. M. (1982) Adu. Immunol. 33, 167-240. Bravo, R. (1986) Exp. Cell Res. 163, 287-293. Celis, J. E., Madsen, P., Nielsen, S., and Celis, A. (1986) Leuk. Res. 10, 237-249. Bravo, R., Frank, R., Blundell, P. A., and Macdonald-Bravo, H. (1987) Nature (London) 326, 515-517. Prelich, Cl., Tan, C. K., Kostura, M. Mathews, M. B., So. A. G., Downey, K. M., and Stillman, B. (1987) Nature (London) 326, 517-520. Bravo, R., and Macdonald-Bravo, H. (1987) J. Cell Eiol. 105, 154Sl554. Bravo, R., and Macdonald-Bravo, H. (1985) EMBO J. 4, 665-661. Celis, J. E., Madsen, P., Celis, A., Nielsen, H. V., and Gesser, B. (1987). FEBS Letr. 220, 1-7. Celis, A., and Celis, J. E. (1985) Proc. Natl. Acad. Sci. USA 82, 3262-3266.
EM immunolocalization
of cyclin
89
11. Mandys, V., RaSka, I., Kostrouch, Z., Jamfk, M., PetraSevifova, V., Bednar, J., Kobema, K., and Jandejsek, J. (1988). Submitted for publication. 12. Grifliths, G., McDowall, A., Back, R., and Dubochet, J. (1984), J. Ulfrasrruct. Res. 89, 65-78. 13. Madsen, P., Ogata, K., and Celis, J. E. (1987) Leukemia 1, 220-225. 14. Slot, J. W., and Geuze, H. J. (1985) Eur. J. Cell. Biol. 38, 87-93. 15. Scheer, U., Messner, K., Hazan, R., RaSka, I., Hansmann, P, Falk, H., Spiess, E., and Franke, W. W. (1987) Eur. J. Gel/ Biol. 43, 358-371. 16. RaSka, I., Valouch, P., Armbruster, B. L., Hinterberger, M., Maly, A., Vorlicek, J., Smetana, K., and Kellenberger, E. (1985) Histochem. J. 17, 925-938. 17. Bouteille, M., Laval, M., and Dupuy-Coin, A. M. (1974) in The Cell Nucleus (Busch, H., Ed.), Vol. I, pp. 3-71, Academic Press, New York/London. 18. Fakan, S. (1978) in The Cell Nucleus (Busch, H., Ed.), Vol. V, pp. 3-53, Academic Press, New York/London. 19. Hadjiolov, A. A. (1985) in Cell Biology Monographs (Alfert, M., Beermann, W., Goldstein, L., Porter, K., Sitte, P. Eds.), Springer-Verlag, Vienna. 20. Madsen, P., and Celis, J. E. (1985) FEES Let?. 193, 5-11. 21. Carri, M. T., Micheli, G., Graziano, E., Pace, T., and Buongiomo-Nardelli, M. (1986) Exp. Cell Res. 164, 426A36.
22. Nelson, W. G., Pienta, K. J., Barrack, E. R., and Coeffey, D. S. (1986) Annu. Rev. Biophys. Chem. 15, 457475.
Received June 18, 1988 Revised version received March 23. 1989
Prmted
in Sweden
Ultrastructural
Cell Research 184 (1989) 81-89
lmmunolocalization of Cyclin/PCNA Synchronized 3T3 Cells’
in
IVAN RASKA,*,* KAREL KOBERNA,* MICHAL JARNiK,* VALBRIA PETRASOVICOVA,t JAN BEDNAR,* KAREL RASKA Jr-,$ and RODRIGO BRAVO” *Institute of Experimental Medicine, Czechoslovak Academy of Sciences, Lidovych Milici es-120 00 Prague, Czechoslovakia; tResearch Institute of Rheumatic Diseases, CS-I28 00 Prague, Czechoslovakia; $lnstitute of Physiology, Czechoslovak Academy of Sciences, CS-142 20 Prague, Czechoslovakia; OUMDNJ, Robert Wood Johnson Medical School, Piscataway, New Jersey 08854; and I’European Molecular Biology Laboratory, D-6900 Heidelberg, West Germany
61,
The immunolocalization of cyclin/PCNA’ in synchronized 3T3 cells was performed with human autoantibodies using an immunogold technique performed on thin cryosections. Previous immunofluorescent studies demonstrated that the DNA replication sites correspond to the localization of bound cyclin. We have found that in the early periods of S phase, the DNA replication sites (or sites potentially ready for the replication during the hydroxyurea DNA synthesis block) are situated in the perichromatin region and correspond to clustered gold particles present frequently over a morphologically distinct small nuclear area. Heavily labeled chromocenters, including perinucleolar condensed chromatin, exhibiting several such distinct areas were found in later periods of S phase. @ 1989 Academic
Press, Inc.
The aim of this study was to localize cyclin/PCNA [Ilr] in synchronized 3T3 cells during various periods of S phase using an immunogold technique. Cyclin is an auxilliary protein of the DNA polymerase delta [5, 61 and it was shown in the previous biochemical and light microscopical studies that the bound cyclin exhibiting the punctuate immunofluorescent pattern and DNA replication sites have identical localization during the S phase [3, 4, 7-103. Bound cyclin represents 2&30% of all cyclin molecules present in the cell, whereas free cyclin is extracted during the methanol fixation used for immunofluorescence [7]. The localization of cyclin was investigated in two different experiments. In the first experiment we immunolocalized cyclin in 3T3 cells which were stimulated to grow. In the second experiment, DNA synthesis in stimulated cells was blocked by hydroxyurea and the immunolocalization of cyclin was performed after its removal [8].
’ This paper is dedicated to the memory of Dr. Karel RaSka. 2 To whom reprint requests should be addressed. 3 PCNA, proliferating cell nuclear antigen. 81
Copyright @ 1989 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827/89 $03.00
82
RaSka et al.
0
4 0 12 16 20 WI WI WI (34 (~1 hours after stlmulatlon (sample)
UN
0 WI
2 4 6 0 10 (13th) WW VW (1w WI houra after hydroxyurea twmoval (sample)
24 WI
12 (14th)
Fig. 1. Synthesis of DNA after serum stimulation of quiescent 3T3 cells (A) and after hydroxyurea release (B) as followed by [3H]thymidine incorporation (for details see [SJ). EM investigations were performed on cell samples of serum-stimulated cells at 4-h intervals (samples 1 to 7) and on samples of hydroxyurea-treated cells at 2-h intervals (samples 8 to 14).
MATERIALS
AND METHODS
3T3 cells arrested in Go phase were stimulated to growth by addition of 20% fetal calf serum and analyzed at 4-h intervals during a period of 24 h (samples 1 to 7) [8]. In a parallel experiment, the incorporation of [3H]thymidine was determined (Fig. 1; see also [S]). The cells were fixed in 8% paraformaldehyde, preembedded in a fibrin clot [ll], infused in 2.1 M sucrose, and frozen in liquid nitrogen [12], and cryosections were prepared according to Griffiths et al. (1984). Bound (purified) human autoantibodies, which were shown to react in Western blots monospecially with cyclin [13], were revealed by means of 9-nm gold particles stabilized with protein A [14]. In the double-labeling experiments, a monoclonal antibody to DNA [IS] was used (kind gift of Boehringer Co., Mannheim) together with the autoantibody. The bound antibodies were revealed with the help of antimouse S-nm (DNA) and antihuman IO-nm (cyclin) gold adducts (kind gift of Janssen Pharmaceutics Co., Beerse). Control labeling experiments included omission of the primary antibody, saturation of sections with
EM immunolocalization of cyclin
83
Fig. 2. A cryosection of a 3T3 cell of the 5th sample exhibiting a labeled DA (arrows) in the perichromatin region. C, cytoplasm. ~77,000.
free secondary immunoglobulins or protein A prior to application of gold adducts, parallel application of antibodies with different specificities (e.g., RNA polymerase I, ribosomal proteins which were specially enriched in the nucleolar body of the nucleus), and, the double-labeling experiments, concomitant or consecutive application of antibodies. Morphometric estimation of surface-density labeling of cyclin in single-labeling experiments was performed on at least 30 prints of electron micrographs for each sample [16]. In the second experiment, hydroxyurea (5 mM) was added 8 h after stimulation of 3T3 cells to stop progression from G, to S phase. After 12 h, the hydroxyurea was washed out and the cyclin localization was investigated in cells harvested at 2-h intervals up to the 12th h (samples 8 to 14). The incorporation of [.‘H]thymidine was followed in a parallel experiment (Fig. 1; see also [S]).
RESULTS In the first three samples (0, 4, and 8 h after stimulation) the cell nuclei were characterized by the presence of rather homogeneously distributed individual gold particles. This kind of labeling pattern remained in subsequent samples. With the onset of the DNA synthesis (Fig. 1) in the 4th sample (12 h after stimulation) and especially in the 5th sample (16 h after stimulation), the intensity of the label was increased particularly in the perichromatin region, and the first clustered particles (i.e., several gold particles confined to a small nuclear area) appeared. Clustered particles frequently became confined to regions in which the morphology exhibited a “looser” texture and apparently lower DNA content than the condensed chromatin, their electron density usually being higher than that of the interchromatin regions (Figs. 2 and 3). We refer to these distinct
84
RaSka et al.
Fig. 3. A double-labeling experiment performed on a 3T3 cell from the 5th sample. Note that the DA (arrow) is enriched in the cyclin (larger gold particles) but not in the DNA (smaller gold particles) label. C, Condensed chromatin; CY, cytoplasm. x 145,300.
regions with clustered gold particles as distinct areas (DAs). In the 6th and 7th samples (20 and 24 h after stimulation), which exhibited the highest density of gold particles (increased threefold in comparison to the 1st sample), labeling within chromocenters including perinucleolar condensed chromatin was observed as well (Fig. 4). However, the label confined to areas delimited by chromocenters was not homogeneous and corresponded to the presence of several DAs (Figs. 4 and 5). The number of clustered particles represented, on average, 25% of the whole nuclear label. This, in fact, correponds well to the known percentage of bound cyclin [7]. The labeling of the nucleolar body was lower with respect to the overall nuclear label and consisted mainly of individual particles (Figs. 4 and 5). After the removal of hydroxyurea, the labeling picture in samples g-14 (0, 2, 4, 6, 8, 10, and 12 h after hydroxyurea removal) was different from that observed in cells not exposed to hydroxyurea. The difference was due mainly to the overall higher incidence of gold particles and to the fact that the clustered particles (and DAs) were already observed in the 8th sample when the hydroxyurea was just washed away. The intensity of labeling in the 8th sample was comparable to that in the 6th sample. The maximum level of labeling was already reached in the 10th and 1lth samples (fourfold increase in comparison to 1st sample). The label was qualitatively comparable to-but higher than-that of samples 6 and 7 (Fig. 6).
EM immunolocalization of cyclin
85
Fig. 4. A low-magnification picture of a 3T3 cell cryosection of the 7th sample exhibiting several labeled DAs (arrows) within the chromocenter and the perinucleolar condensed chromatin of two nucleoli. In a few DAs, gold particles follow the border of condensed chromatin. C, Cytoplasm: N, nucleolus; G, interchromatin granules. ~35,000.
86
RaSka et al.
Fig. 5. A higher magnificantion of the cryosection shown in Fig. 4. DAs (arrows) are well discernible within the perinucleolar condensed chromatin (P) at this magnification. In the nucleolar body (N), only a few individual particles are present. ~52,800.
DISCUSSION In this study we have attempted to relate functional events of DNA replication to immunocytochemical localization of cyclin. Because we used formaldehyde rather than methanol fixation we could detect both free and bound cyclin [7]. From the qualitative and quantitative character of the changes, and also compared to those seen in fluorescent studies (punctuate, granular, and nucleolar fluorescence; see also [7, S]), we propose that the clustered gold particles
EM immunolocalization
- --
_-.
of cyclin
81
--
Fig. 6 An intense labeling of perinucleolar condensed chromatin (P) in late S phase after the release of cells from hydroxyurea block. Note the uneven distribution of gold particles within the area delimited by the perinucleolar condensed chromatin. Clustered particles identify several DAs (arrowheads). The nucleolar body (N) is almost not labeled. ~46,200.
frequently present over DAs identify sites related to DNA synthesis. However, we do not claim that all replication sites can be identified as DAs. The high-resolution autoradiographic studies dealing with the DNA replication yield rather scattered data (for review see [17, 181). Whereas clustered particles (and DAs) were situated mainly in the perichromatin regions in the early periods of the S phase, an intensive but not homogeneous labeling (confined to DAs) within the chromocenters could be observed as well in the later periods of the S
88
RaSka et al.
phase in the present study. Our results are thus in agreement with those autoradiographic data [17, 181situating DNA replication sites in the regions of extended chromatin (euchromatin) during the early S phase, which then also spread to heterochromatin in the later S phase. Due to the low labeling density within the nucleolar body, we were not usually able to distinguish between the bound and the free cyclin within this nuclear organelle. This is understandable, as the bulk of nucleolar DNA is contined to the perinucleolar condensed chromatin, the concentration of DNA within the nuclear body being much lower (e.g., [ 191). The previously described nucleolar immunofluorescent staining in 3T3 cells [7, 81 apparently mimics the intensive labeling of the perinucleolar chromatin as observed in this study. In this sense, the replication of nucleolar DNA can be positioned at a later period of S phase. The labeled DAs in the 8th sample, in which the DNA synthesis was still suppressed, are interpreted as sites potentially ready for replication since the DNA synthesis is very rapid after the release of the hydroxyurea block (Fig. 1; see also [S]). Cyclin, the synthesis of which is not blocked by hydroxyurea treatment, may be present in structures with a lower DNA content [20], apparently the nuclear matrix [21, 221. This is supported by the observation that the concentration of DNA is diminished over DAs (cf. Fig. 3). The presence of DNA in DAs is not, however, ruled out, as the antibody used does not react substantially with the decondensed (relaxed) chromatin [I51 as one would expect for replicating DNA. The appearance of DAs illustrates that a functional event of DNA replication (or leading potentially to DNA replication) induces subtle morphological changes which have not been detected by ultrastructural methods until now. In addition, the existence of several DAs within individual chromocenters during the later periods of S phase may represent a morphological correlate in situ of several replication units present on one or several chromosomes. We thank Dr. Y. Takasaki for providing us with human serum (A.K.) anti-PCNA. The authors are obliged to Dr. G. Griftiths and Ruth Back (EMBL, Heidelberg) for their suggestions and critical help in preparing the manuscript. I.R. is particularly grateful to the Herbette Foundation which subsidized his stay at the Laboratory of Ultrastructural Analysis of the University of Lausanne where a part of this study was performed.
REFERENCES I. 2. 3. 4. 5. 6. 7. 8. 9. 10.
Miyashi, K., Fritzler, M. J., and Tan, E. M. (1978) J. Immunol. 121, 2228-2234. Tan, E. M. (1982) Adu. Immunol. 33, 167-240. Bravo, R. (1986) Exp. Cell Res. 163, 287-293. Celis, J. E., Madsen, P., Nielsen, S., and Celis, A. (1986) Leuk. Res. 10, 237-249. Bravo, R., Frank, R., Blundell, P. A., and Macdonald-Bravo, H. (1987) Nature (London) 326, 515-517. Prelich, Cl., Tan, C. K., Kostura, M. Mathews, M. B., So. A. G., Downey, K. M., and Stillman, B. (1987) Nature (London) 326, 517-520. Bravo, R., and Macdonald-Bravo, H. (1987) J. Cell Eiol. 105, 154Sl554. Bravo, R., and Macdonald-Bravo, H. (1985) EMBO J. 4, 665-661. Celis, J. E., Madsen, P., Celis, A., Nielsen, H. V., and Gesser, B. (1987). FEBS Letr. 220, 1-7. Celis, A., and Celis, J. E. (1985) Proc. Natl. Acad. Sci. USA 82, 3262-3266.
EM immunolocalization
of cyclin
89
11. Mandys, V., RaSka, I., Kostrouch, Z., Jamfk, M., PetraSevifova, V., Bednar, J., Kobema, K., and Jandejsek, J. (1988). Submitted for publication. 12. Grifliths, G., McDowall, A., Back, R., and Dubochet, J. (1984), J. Ulfrasrruct. Res. 89, 65-78. 13. Madsen, P., Ogata, K., and Celis, J. E. (1987) Leukemia 1, 220-225. 14. Slot, J. W., and Geuze, H. J. (1985) Eur. J. Cell. Biol. 38, 87-93. 15. Scheer, U., Messner, K., Hazan, R., RaSka, I., Hansmann, P, Falk, H., Spiess, E., and Franke, W. W. (1987) Eur. J. Gel/ Biol. 43, 358-371. 16. RaSka, I., Valouch, P., Armbruster, B. L., Hinterberger, M., Maly, A., Vorlicek, J., Smetana, K., and Kellenberger, E. (1985) Histochem. J. 17, 925-938. 17. Bouteille, M., Laval, M., and Dupuy-Coin, A. M. (1974) in The Cell Nucleus (Busch, H., Ed.), Vol. I, pp. 3-71, Academic Press, New York/London. 18. Fakan, S. (1978) in The Cell Nucleus (Busch, H., Ed.), Vol. V, pp. 3-53, Academic Press, New York/London. 19. Hadjiolov, A. A. (1985) in Cell Biology Monographs (Alfert, M., Beermann, W., Goldstein, L., Porter, K., Sitte, P. Eds.), Springer-Verlag, Vienna. 20. Madsen, P., and Celis, J. E. (1985) FEES Let?. 193, 5-11. 21. Carri, M. T., Micheli, G., Graziano, E., Pace, T., and Buongiomo-Nardelli, M. (1986) Exp. Cell Res. 164, 426A36.
22. Nelson, W. G., Pienta, K. J., Barrack, E. R., and Coeffey, D. S. (1986) Annu. Rev. Biophys. Chem. 15, 457475.
Received June 18, 1988 Revised version received March 23. 1989
Prmted
in Sweden