Association of Nuclear Matrix Proteins with Granular and Threaded Nuclear Bodies in Cell Lines Undergoing Apoptosis

EXPERIMENTAL CELL RESEARCH ARTICLE NO.

230, 325–336 (1997)

EX963415

Association of Nuclear Matrix Proteins with Granular and Threaded Nuclear Bodies in Cell Lines Undergoing Apoptosis MARINA ZWEYER,* BEAT M. RIEDERER,† ROBERT L. OCHS,‡ FRANK O. FACKELMAYER,§ TERUMI KOHWI-SHIGEMATSU,Ø RENATO BAREGGI,* PAOLA NARDUCCI,* ,1 AND ALBERTO M. MARTELLI* *Dipartimento di Morfologia Umana Normale, Universita` di Trieste, via Manzoni 16, I-34138 Trieste, Italy; †Institut de Biologie Cellulaire et de Morphologie, Universite´ de Lausanne, Rue du Bugnon 9, CH-1005, Lausanne, Switzerland; ‡W. M. Keck Autoimmune Disease Center, Department of Molecular and Experimental Medicine, Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, California 92037; §Department of Biology, Universitat Konstanz, Universitatsstrasse 10, D-78434 Konstanz, Germany; and ØLa Jolla Cancer Research Foundation, 10901 North Torrey Pines Road, La Jolla, California 92037

The granules which appear in the nucleolar area in apoptotic HL-60 cells after camptothecin administration (Zweyer et al., Exp. Cell Res. 221, 27–40, 1995) were detected also in several other cell lines induced to undergo apoptosis by different stimuli, such as MOLT-4 treated with staurosporine, K-562 incubated with actinomycin D, P-815 exposed to temperature causing heat shock, Jurkat cells treated with EGTA, U-937 growing in the presence of cycloheximide and tumor necrosis factor-a, and HeLa cells treated with etoposide. Using immunoelectron microscopy techniques, we demonstrate that, besides the already described nuclear matrix proteins p125 and p160, these granules contain other nucleoskeletal polypeptides such as proliferating cell nuclear antigen, a component of ribonucleoprotein particles, a 105-kDa constituent of nuclear spliceosomes, and the 240-kDa nuclear mitotic apparatus-associated protein referred to as NuMA. Moreover, we also found in the granules SAF-A/hn-RNP-U and SATB1 proteins, two polypeptides that have been reported to bind scaffold-associated regions DNA sequences in vitro, thus mediating the formation of looped DNA structures in vivo. Fibrillarin and coilin are not present in these granules or the PML protein. Thus, the granules seen during the apoptotic process apparently are different from coiled bodies or other types of nuclear bodies. Furthermore, these granules do not contain chromatin components such as histones and DNA. Last, Western blotting analysis revealed that nuclear matrix proteins present in the granules are not proteolytically degraded except for the NuMA polypeptide. We propose that these granules might represent aggregates of nuclear matrix proteins forming during the apoptotic process. Moreover, since the granules are present in several cell lines undergoing apoptosis, they could be considered a previously un1

To whom correspondence and reprint requests should be addressed. Fax: /39/40/639052.

recognized morphological hallmark of the apoptotic process. q 1997 Academic Press

INTRODUCTION

Apoptosis is a form of programmed cell death in which superflous or harmful cells are removed from an organism [1]. The term programmed cell death is usually reserved for physiological conditions during which disposal of cells takes place, such as development and morphogenesis, clonal selection of lymphocytes, and cell renewal in epithelia [2–4]. On the contrary, the term apoptosis indicates a form of active cell death triggered by pathological conditions such as inflammation, cancer, viral infections (including AIDS), or exposure to a wide variety of chemical and physical stimuli [5–7]. A well-known series of morphological modifications take place during the apoptotic process: chromatin margination and condensation, cell shrinkage, membrane blebbing, and finally disintegration of the cell into membrane-bound fragments called ‘‘apoptotic bodies’’ [8]. Several biochemical changes also occur: in most types of apoptosis there is internucleosomal DNA fragmentation, mediated by an as yet unidentified Ca2/and Mg 2/-dependent endonuclease, producing 200-bp oligonucleotides that have long been considered a ‘‘hallmark’’ of the phenomenon [9]. However, several reports now indicate that some types of apoptotic stimuli do not induce this kind of fragmentation [10–13]. On the other hand, genome fragmentation into 300- and 50kb pieces (effected by a different endonuclease activity) seems to occur in all types of apoptosis, and precedes endonucleosomal fragmentation [14–16]. Recently, attention has been drawn to the fact that during apoptosis proteins can be proteolysed by several types of proteases [17–21]. For example, the execution phase

325

AID

ECR 3415

/

6i16$$$721

01-20-97 17:19:05

0014-4827/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.

ecal

326

ZWEYER ET AL.

of apoptosis is triggered by a proteinase of the Ced-3/ ICE family (for a recent review see Ref. [22]). The interphase nucleus is thought to contain a mainly proteinaceous inner network referred to as nuclear matrix, scaffold, or skeleton [23, 24]. Numerous critical functions have been demonstrated to occur in close association with this matrix [24] and it provides a framework to maintain the overall size and shape of the nucleus. Moreover, the matrix acts as a structural attachment site for the DNA loops during the interphase [e.g., 25, 26]. It is important to emphasize that the calculated length of DNA loops is in a range of 10 to 200 kb, with an average of 86 kb [27, 28]. Evolutionarily highly conserved 300- to 1000-bp-long DNA sequences referred to as SARs (for Scaffold Associated Regions), have been identified that define the base of DNA loops, anchoring them to specific proteins [29–31]. Given the dramatic changes the nucleus undergoes during the apoptotic process, one would expect the matrix in apoptotic cells to show marked variations in comparison to untreated cells. However, the report by Arends et al. [32] ruled out any apparent changes in matrix structure and protein composition following steroid-induced apoptosis in thymocytes. Subsequently, Miller et al. [33] showed by immunofluorescence staining that matrix changes are indeed present and matrix proteins are released from apoptotic cells. Modifications at the light microscope level have also been reported by Tinnemans et al. [34]. Among nuclear proteins that undergo proteolysis are the abundant matrix components DNA topoisomerase IIa [35–37], poly(ADP-ribose)polymerase [38–40] that is cleaved by an ICE-like proteinase (referred to as prICE), and the nuclear lamins that form a meshwork located at the periphery of the nucleus, underlying the inner nuclear membrane [41, 42]. Lamins, however, are cleaved by yet another protease, referred to as Lamp, situated downstream from prICE in the apoptotic cascade [42, 43]. The action of an endonuclease coupled to protease activity appears to lead to detachment of DNA loops from the matrix [44]. Indeed, recent findings by Lagarkova et al. [45] indicate that cleavage occurs at precise sites reflective of a higher-order chromatin organization. We recently demonstrated that in HL-60 cells treated with the topoisomerase I inhibitor camptothecin, two nuclear matrix proteins were associated with fibrogranular nuclear bodies that originate from the nucleolus [46]. Here, we have extended our previous observations and we have characterized these granular structures in more depth. We have found that they contain several nuclear matrix proteins, and among these there are two polypeptides that have been described to bind SAR sequences, SAF-A/hnRNP-U and SATB1 [47, 48]. The fibrogranular nuclear bodies (fgNBs) are found in a variety of cell lines induced to undergo apoptosis by different agents. They seem to be

AID

ECR 3415

/

6i16$$$722

01-20-97 17:19:05

different from classic nuclear coiled bodies in that they do not contain markers for these structures such as fibrillarin or coilin. Moreover, they do not show the presence of the PML product, the hallmark of another class of nuclear bodies [49, 50]. Finally, the fgNBs do not contain chromatin components such as histones and DNA. Among several proteins present in the fgNBs, only NuMA showed proof of proteolytic degradation. We interpret our findings to mean that these nuclear bodies represent structures where proteins from the disassembled nuclear matrix aggregate prior to be extruded from the nucleus. Furthermore, we propose that they could be considered a previously unrecognized morphological hallmark of changes typical of apoptosis at the level of nuclear structure. MATERIALS AND METHODS Cell culture and induction of apoptosis. HL-60 human promyelocytic cells, K562 human erythroleukemia cells, U937 human leukemic myelomonocytic cells, MOLT-4 human T leukemic lymphocytic cells, Jurkat human T leukemic lymphocytic cells, and P-815 murine mastocytoma cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum. Human cervical carcinoma HeLa cells were cultured in Dulbecco’s modified minimum essential medium supplemented with 10% newborn calf serum. In all cultures, medium was renewed twice a week. All of the cultures were checked routinely for mycoplasma contamination. For induction of apoptosis, cells were treated as follows. HL-60 cells. HL-60 cells were synchronized at the G1/S border with 1 mg/ml aphidicolin, then released from the block and exposed for 4 h to 0.1 mg/ml of the DNA topoisomerase I inhibitor camptothecin (Sigma Chemical Co., St. Louis, MO), according to Del Bino et al. [51]. K562 cells. K562 cells were incubated for 24 h in the presence of 1 mg/ml of the RNA polymerase inhibitor actinomycin D (Boehringer, Mannheim, Germany) as reported by Kressel and Groscurth [52]. U937 cells. U937 cells were first treated for 2 h with 4 mM cycloheximide (CHX, Sigma) and then for an additional 3 h with 50 IU/ ml of tumor necrosis factor-a (TNF-a, Sigma, final purity ú95%) according to Cossarizza et al. [53]. MOLT-4 cells. MOLT-4 cells were exposed to 20 nM staurosporine (Boehringer), a protein kinase inhibitor, for 24 h [12]. P-815 cells. Apoptosis was induced by heat shock treatment (30 min at 447C) followed by recovery at 377C for 7 h [10]. Jurkat cells. Jurkat cells were incubated for 24 h in the presence of 5 mM EGTA, as described by Patterson et al. [54]. HeLa cells. HeLa cells were incubated for 3 h in the presence of 75 mM etoposide (VP-16, Sigma), an inhibitor of DNA topoisomerase II, as reported by Guano et al. [55]. Source of antibodies. The monoclonal antibodies to 125- and 160kDa inner matrix proteins (p125 and p160) and polyclonal antisera against SAR-binding proteins SAF-A/hnRNP-U and SATB1 have been described elsewhere [46 – 48, 56, 57]. Human autoantibody to fibrillarin and rabbit R288 antiserum to p80 coilin were as reported by Ochs et al. [58]. An antiserum recognizing a peptide of the p80 coilin-coding region from amino acids 397 to 407 was kindly gifted by Drs. A. I. Lamond and K. Bohmann, EMBL, Heidelberg, Germany [59]. Monoclonal antibodies to proliferating cell nuclear antigen (PCNA, or the auxiliary protein to DNA polymerase d) and PML product were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The monoclonal antibodies against 240-kDa NuMA [e.g., 60] were obtained from Oncogene Science (Cambridge, MA). According to the

ecal

327

NUCLEAR BODIES IN APOPTOTIC CELLS manufacturer, one of the antibodies (clone 204-41) recognizes an epitope located in the carboxy-terminal domain of the 240-kDa protein, whereas the other (clone 107-7) reacts with the amino-terminal domain. Monoclonal antibodies to the 105-kDa component of spliceosomes (clone 780-3) and to nuclear ribonucleoprotein particles (clone 58-15, RNP) were by Chemicon International (Temecula, CA) [61, 62]. Monoclonal antibody recognizing histone H1 and core histones (anti-histone pan) as well as monoclonal antibody to single- and double-stranded DNA were from Boehringer. Immunoelectron microscopy. A postembedding technique was used as a method for transmission immunoelectron microscopy [46]. Control cells and cells in the various stages of apoptosis were fixed with 1% glutaraldehyde in sodium phosphate buffer, pH 7.3, for 35 min at 47C, dehydrated up to 70% ethanol, and embedded in London Resin White (LRW) (Polyscience, Warrington, PA). Polymerization was performed overnight at 47C by accelerator. Part of the fixed samples was also dehydrated up to 100% ethanol, treated with propylene-oxyde, and embedded in araldite (Fluka, Switzerland). Sections were cut using an Ultracut ultramicrotome (Reichert Jung, Germany). To block nonspecific binding sites, the grids were treated with TBS buffer (20 mM Tris-HCl, pH 8.2, 150 mM NaCl containing 0.1% bovine serum albumin (BSA)) for 10 min at room temperature for samples treated with monoclonal antibodies. For samples treated with polyclonal antibodies, we used a modified TBS buffer (the abovementioned buffer, with 0.05% Triton and 0.05% Tween 20 added). Sections were incubated at 47C with the primary antibody diluted 1:5 in TBS for monoclonal antibodies to PML protein or histones, or 1:10 for monoclonal antibodies against p125, NuMA, RNP, spliceosomes, and PCNA. For incubation with polyclonal antisera, the primary antibodies were diluted in the modified TBS at a 1:50 dilution for anti-SATB1, or 1:100 for antibodies to fibrillarin, coilin, and SAFA/hnRNP-U. As a control for polyclonal antisera, normal serum or the preimmune serum (when available) was employed, while for monoclonal antibodies we used the secondary antibody. In both cases no significant labeling was evident (data not shown). Grids were washed several times in the same buffers and then incubated with the appropriate secondary antibody at dilutions ranging from 1:50 to 1:100 in TBS. Secondary antibodies were conjugated with 15-nm colloidal gold particles (Biocell, Cardiff, UK) and diluted in TBS. The grids were briefly stained with uranyl acetate (5 min) and subsequently with lead citrate (5 min). Sections were examined with a JEOL-JEM 100S electron microscope. DNA staining. Apoptotic HL-60 and U-937 cells were cytocentrifuged to glass slides, fixed with 4% paraformaldehyde in Dulbecco’s phosphate-buffered saline (PBS, pH 7.4), permeabilized in 0.2% Triton X-100 in PBS for 10 min, and stained for DNA with 0.5 mg/ ml 4*-6-diaminidino-2-phenylindole (DAPI), as previously indicated [46], to identify morphological changes typical of apoptotic cells [51]. Protein assay. Protein was assayed as reported by Bradford [63]. Western blot analysis. The method was as reported by Neri et al. [57]. Briefly, HL-60 and U937 cells (control and apoptotic) were incubated for 30 min at room temperature in 10 mM Tris-HCl, pH 7.4, 5 mM MgCl2 , 0.01% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 1 mg/ml aprotinin and leupeptin, 5 mg/ml pepstatin, 0.1 mM benzamidine, 5 mg/ml calpain I and II inhibitors, and 50 IU/ml of DNase I (Sigma). This step was included to digest chromosomal DNA thus improving resolution of gel electrophoresis. Proteins were ethanol-precipitated, dissolved in sample buffer [64], and separated on polyacrylamide–0.1% sodium dodecylsulfate gels, then transferred to 0.22-mm nitrocellulose paper in 192 mM glycine, 25 mM Tris, 20% methanol, pH 8.3. Equal strips were cut from the nitrocellulose pieces and saturated for 60 min at 377C in PBS containing 10% normal goat serum, and 4% BSA (saturation buffer). After a wash in PBS and 0.1% BSA (washing buffer), they were reacted for 3 h at room temperature in saturation buffer containing the primary antibody. After 4 washes as above, the strips were incubated for 60 min at room temperature with the appropriate secondary antibody, conjugated to alkaline phosphatase, diluted 1:1000 in washing buffer. Strips were

AID

ECR 3415

/

6i16$$$722

01-20-97 17:19:05

then washed and color was developed using p-nitrotetrazolium blue and 5-bromo-4-chloro-3-indolyl phosphate as substrates [57]. Densitometric scanning analysis of Western blots. This was performed by a densitometric computer program running on a Quantimet 970 image analyzer (Cambridge Instruments, UK) according to Martelli et al. [65].

RESULTS

fgNBs Contain Several Nuclear Matrix Proteins in HL-60 Cells Undergoing Apoptosis It should be emphasized that all the samples have been embedded both in LWR and in araldite [46]. However, since the antigenicity of all the proteins studied was well preserved with araldite, the results and the pictures presented in this paper refer to araldite-embedded samples because of their better morphology. Moreover, it is important to stress that our goal in this study was to characterize as thoroughly as possible the fgNBs that are seen in apoptotic cells only. The samples have not been osmicated because we wanted to preserve the antigenicity of the polypeptides as much as possible. Thus, cell membranes are not as easily detectable as in osmified samples, there is little staining in the nucleoplasm, and condensed chromatin appears less electron-dense than usual, and the same holds true for some typical features of the apoptotic cells, such as the crescentic caps, the micronuclei, and the chromatin component of the apoptotic bodies. However, under these conditions, fgNBs are always very easy to be identified, because of their intrinsic electron-density. Moreover, since colloidal gold particles often localized to the only electron-dense structures detectable in the samples (i.e., the fgNBs), we were forced to keep the background of the pictures as clear as possible, in order to better detect the particles. The nuclear bodies previously described [46] in the HL-60 cell line undergoing apoptosis by exposure to the topoisomerase I inhibitor camptothecin contained two proteins of the nucleoskeleton, referred to as p125 and p160. We wanted to verify whether other polypeptides known to be components of the nuclear matrix could be detected in these bodies. Indeed, all of the proteins we studied were detected by immunofluorescent staining and Western blotting in K562 cell nuclear matrix prepared according to a well-established procedure [29; manuscript in preparation]. As shown in Fig. 1A a 105-kDa splicing component behaved exactly like p125 and p160 nuclear proteins during nucleolar modifications [46]. In brief, besides the occasional appearance of a round electrontransparent inner area, some granular clusters appeared in the perinucleolar zone and in corresponding fibrillar centers. Afterward, small clusters of granules moved to condense in one or more larger bodies, asso-

ecal

328

AID

ZWEYER ET AL.

ECR 3415

/

6i16$$3415

01-20-97 17:19:05

ecal

NUCLEAR BODIES IN APOPTOTIC CELLS

ciated with the periphery of the residual nucleolus. Finally, the granular masses piled up to form threaded and coiled electron-dense structures. A RNP antigen (Fig. 1B), PCNA (Fig. 1C), and NuMA (clone 204-41) (Fig. 1D) protein associated with the fgNBs. Interestingly, these bodies also contained two polypeptides that bind to SARs, that is, SAF-A/hnRNP-U and SATB1. SAF-A/hnRNP-U concentrated in the nucleoli in the very early phase of apoptotic process, before the appearance of chromatin margination or nucleolar segregation. However, some of the antigen was present in cup-shaped chromatin marginations, as well as in micronuclei and in apoptotic bodies, especially in the condensed chromatin (Fig. 1E). The behavior of SATB1 was analogous to that of SAF-A/hnRNP-U as shown in Fig. 1F. A time course of the behavior of the different antigens during the apoptotic process showed that on the whole it was similar to the data reported by Zweyer et al. [46] for p125 and p160 (data not presented). fgNBs Are Detectable in Several Cell Lines Undergoing Apoptosis by Different Stimuli Next, we investigated whether fgNBs could also be detected in other cell lines induced to apoptosis by different stimuli. As a marker for ultrastructural immunocytochemical detection, we mainly employed the antibody to p125, because in our hands it always gave a very strong and specific labeling of the fgNBs. The results showed that apoptotic K-562, Jurkat, U-937, P-815, and HeLa cell lines behaved exactly as HL-60 cells treated with camptothecin (Figs. 2A – 2E). Even though for this series of experiments we preferentially used antibody to p125, it should be emphasized that in K562 cells the 105-kDa splicing component (Fig. 2F) and the RNP antigen (Fig. 2G) also were immunolocalized to these granules. On the other hand, apoptotic MOLT-4 cells showed slightly different features. As shown in Fig. 3A, in the very early stages the modifications were quite similar to those detected in other cell lines, featuring small granular clusters (labeled by antibody to p125) present both at the periphery of and within the nucleolus. Subsequently, the clusters spread all around in the nucleus (Figs. 3B and 3C). Finally, they sometimes condensed into large fgNBs located close to the modified nucleolus (Fig. 3D).

329

fgNBs Do Not Contain Markers Typical of Other Types of Nuclear Inclusions (Coilin, Fibrillarin, PML Protein) or of Chromatin (Histones and DNA) Nuclei are known to contain a variety of inclusions, referred to as coiled bodies or nuclear bodies [e.g., 66 – 68]. Proteins that are markers of these inclusions are now known and antibodies to them are available. Thus, we investigated whether or not the granular bodies seen in apoptotic HL-60 cells contain markers typical of either coiled bodies (p-80 coilin and fibrillarin) or of a particular class of nuclear inclusions containing the PML protein [49, 50]. Two different antisera against p-80 coilin did not label the fgNBs (Fig. 4A). However, the antigen was localized to the nucleoplasm in agreement with the suggestion about the existence of a pool of free coilin and also labeled typical coiled bodies, under the experimental conditions used in this paper in HeLa cells [68; data not shown]. Moreover, an antiserum to fibrillarin did not label these structures in apoptotic HL-60 cells either. In nonapoptotic cells, fibrillarin was mainly immunolocalized to the nucleolus, in agreement with the literature [58, 69; data not shown]. When nucleoli began to segregate in the early phases of apoptosis, fibrillarin was not present in the electron-dense granules (Fig. 4B). In the later phases of apoptotic process fgNBs were unlabeled (data not presented). The nuclear-body-associated PML protein could not be localized to the fgNBs, while it concentrated in the cup-shaped chromatin marginations (Fig. 4C). Lastly, we ascertained whether or not these bodies contained chromatin components such as histones and DNA. As presented in Fig. 4D, a monoclonal antibody recognizing histone H1 as well as core histones did not label the bodies present in HL-60 cells. On the whole, histones appeared to be distributed rather diffusely in apoptotic nuclei, and they did not show a specific localization in the crescent-shaped cups. DNA was not immunolocalized to the fgNBs (Figs. 4E and 4F) but instead was concentrated in cup-shaped chromatin marginations, in micronuclei, and in apoptotic bodies (Fig. 4E). The most interesting aspect of the behavior of DNA, as revealed by immunoelectron microscopy during apoptosis, was its selective presence in only one of the three components in which nucleoli segregate. The component so clearly labeled is the one which moves from the

FIG. 1. Immunolocalization of several nuclear matrix proteins in fgNBs of apoptotic HL-60 cells. (A) Immunogold localization of 105kDa splicing component. The protein is concentrated in coiled body-like structures (arrow). (B) A RNP component is present in the fgNBs contained within two micronuclei (arrows). (C) PCNA is selectively localized in fgNBs at the beginning of their formation (arrows). (D) The 240-kDa NuMa is also detectable in the fgNBs (arrows). In this case antibody from clone 204-41 was employed. (E) SAF-A/hnRNP-U is immunolocalized to both a fgNB and a condensed chromatin (arrows). (F) SATB1 polypeptide is present in the fgNBs and in condensed chromatin (arrows). Bar: 0.5 mm (B), 1.0 mm (A, C–F).

AID

ECR 3415

/

6i16$$$723

01-20-97 17:19:05

ecal

330

AID

ZWEYER ET AL.

ECR 3415

/

6i16$$3415

01-20-97 17:19:05

ecal

NUCLEAR BODIES IN APOPTOTIC CELLS

331

FIG. 3. Electron micrographs showing the behavior of p125 antigen in apoptotic MOLT-4 cells. (A) In the early phases of the process p125 localizes to small granular clusters, present both at the periphery of and within the nucleolus (arrows). (B and C) Afterward, the labeled clusters (arrows) spread all over the nucleoplasm. (D) The clusters may also end up in large fgNBs (arrow), usually located close to the modified nucleolus (Nu). Bar: 0.5 mm (A, D), 1.0 mm (B, C).

nucleolar center to the outer part of the nucleolus (Fig. 4F). This crescent-shaped component remains for some time close to the central one, which is totally unlabeled. Recher et al. [70] first showed this component to be derived from fibrillar centers. Most Matrix Proteins Present in fgNBs Are Not Degraded by Proteolysis Since it is now well established that some nuclear proteins are proteolytically cleaved during the apoptotic process, we wanted to verify if this applied also

to several of the polypeptides that end up selectively in the fgNBs. As a model, we chose HL-60 cells treated with camptothecin and U-937 cells incubated with TNF-a and CHX, because in these cell lines there is a high percentage of cells undergoing apoptosis. Indeed, DAPI staining of cells revealed that after 4 h treatment, in the HL-60 line 45% of cells showed morphological apoptotic aspects, while in the U-937 line the percentage was 72% after 3 h exposure. These data are in agreement with previous findings by other investigators [51, 53]. Thus, extracts were prepared from control and apoptotic cells, and the proteins separated by gel

FIG. 2. fgNBs are present in several cell lines undergoing apoptosis. (A) In K-562 cell line nucleoli undergo segregation of their components and the p125 nuclear matrix protein is localized in the granular part (arrow) and also in the fgNBs (arrowhead). (B) Immunogold localization in Jurkat cells of p125 to a fgNB in an apoptotic body (arrow). (C) fgNBs in apoptotic U-937 cells are clearly labeled by an antibody to p125 nuclear matrix protein (arrows). (D) p125 in a loosely organized fgNB in the P-815 cell line (arrow). (E) Cells of epithelial origin, like HeLa, also show nucleolar modifications quite similar to lymphocytic cells, and p125 protein has the same behavior during the apoptotic process, being present in the fgNBs. (F) Another protein tested in K-562 cells during apoptosis, the 105-kDa splicing component, has the same behavior as observed in HL-60 cell line, being associated with fgNBs (arrow). (G) Also a RNP component is immunolocalized to the fgNBs in apoptotic K-562 cells. Bar: 0.5 mm (A, C, E), 1.0 mm (B, D, G, F).

AID

ECR 3415

/

6i16$$$723

01-20-97 17:19:05

ecal

332

AID

ZWEYER ET AL.

ECR 3415

/

6i16$$3415

01-20-97 17:19:05

ecal

NUCLEAR BODIES IN APOPTOTIC CELLS

electrophoresis, blotted to nitrocellulose paper, and probed with antibodies. As shown in Fig. 5, of the polypeptides we investigated (i.e., p160, p125, PCNA, and NuMA) only NuMA showed proof of degradation in apoptotic HL-60 cells. Indeed, in control cells both of the antibodies recognized a 240-kDa band corresponding to the native protein, while in apoptotic cells the monoclonal directed to the carboxy-terminal domain (clone 204-41) revealed also the presence of a 200-kDa proteolytic fragment, while antibody 107-7 stained a band of approximately 40 kDa. Results obtained by visual inspection of Western blots were also supported by densitometric analysis, as illustrated in Table 1. DISCUSSION

We have observed fgNBs deriving from the nucleolus in several cells lines in which apoptosis has been triggered by various stimuli. Segregation of nucleolar components has been observed in the past also in cellular types different from those used for the present study (among the others, hepatocytes, Swiss 3T3 fibroblasts, and TG cells, which derive from a human tubaric carcinoma) [70–75]. However, the authors of these reports did not relate their observations to apoptosis, but only to nucleolar changes induced by some chemicals that are now known for causing an apoptotic cell death (i.e., camptothecin and actinomycin D). These granules had been interpreted by some investigators as masses of rRNAs and their precursors, deriving from the nucleolus, while others have shown that they also contain U3 RNA and a small amount of fibrillarin. Our results show that fgNBs do indeed contain proteins related to RNA metabolism, such as a splicing component, a RNP polypeptide, and SAF-A that is identical to hnRNP-U [48, 76]. Nevertheless, they also contain proteins that conceivably are not involved in RNA metabolism, such as PCNA [77], NuMA protein, and the SAR-binding factor SATB1. Although some of the stimuli other than actinomycin D and camptothecin could cause a block of rRNA synthesis (i.e., heat shock and etoposide; see Refs. [78, 79]), the granules appear after exposure to stimuli that are unlikely to interfere with rRNA production, such as treatment with EGTA or incubation

333

with CHX plus TNF-a, the latter producing only a translational block [80]. In any case, it should be stressed that in order to draw any conclusions about the possible relationship between fgNBs and components of RNA metabolism or DNA replication machinery, a more detailed study of a broader panel of proteins involved in these processes would be necessary; such experiments are now underway in our laboratory. The issue of the possible relationship between fgNBs and other kinds of nuclear bodies is still an open question. First of all, since coiled bodies might derive from nucleoli (according to some authors [67, 68]), we determined whether markers typical of coiled bodies could be immunolocalized to the fgNBs. As neither fibrillarin nor coilin was detectable in the fgNBs, we excluded such a relationship. Moreover, the bodies we have characterized appear to be different from the nuclear inclusions containing the PML protein [49, 50]. Indeed, PML protein was not present in the fgNBs, but was found to be highly concentrated in the marginated chromatin. At present, we have no possible explanation for the peculiar behavior of this polypeptide. The PML protein is a member of a family of proteins featuring the presence of a Cys/His-rich cluster referred to as a RING finger. Interestingly enough, an intranuclear redistribution of PML protein takes place in cells upon adenovirus infection [81, 82]. In this connection, it should be recalled that some viruses can induce apoptotic cell death (see Ref. [7] for a recent review). Thus, it might be that PML protein is involved in modifications occurring at the nuclear level during the apoptotic process, especially in the condensed chromatin. The contention that fgNBs are formed by aggregation of nuclear matrix proteins is strengthened by the observation that they do not contain chromatin components such as DNA and histones. Experiments are now underway to ascertain whether or not RNA is present. It could be argued that fgNBs represent structures with a high protein concentration that bind antibodies nonspecifically. However, it should be stressed that while monoclonal and polyclonal antibodies labeled these structures, others (anti-PML, anti-DNA, anti-histone, anti-fibrillarin, and anti-coilin) did not, indicating a

FIG. 4. Markers of some types of nuclear bodies as well as histones and DNA are not present in fgNBs detected in apoptotic HL-60 cells. (A) Coilin, a marker typical of coiled bodies, is not present in the coiled body-like structures from apoptotic cells. In this case, a polyclonal antiserum raised against a peptide comprising the amino acids 397–407 of p80 coilin, was employed. However, similar data were obtained using R288 antiserum (data not presented). (B) During apoptosis, fibrillarin is localized throughout the modified nucleolus (arrows), but not in the fgNBs (arrowhead). (C) A monoclonal antibody to the nuclear-body-associated PML protein very selectively labels cup-shaped chromatin marginations (arrow) but not the fgNBs (arrowhead). (D) Histones could be localized in nuclei during apoptosis, but they do not show any preferential association with condensed chromatin and are absent from fgNBs (arrow). (E) An apoptotic body containing a micronucleus with condensed chromatin and a fgNB. We can clearly observe the preferential immunolocalization of DNA to the condensed chromatin of the micronucleus (arrow) and its total absence from the fgNB (arrowhead). (F) DNA immunolocalization in a nucleolus during apoptosis. The modified nucleolus shows its segregated components. The small clusters of granules are not labeled. The periphery, presumably deriving from fibrillar centers, is selectively labeled by gold particles (arrows). Bar: 0.5 mm (B, E, F), 1.0 mm (A, C, D).

AID

ECR 3415

/

6i16$$$723

01-20-97 17:19:05

ecal

334

ZWEYER ET AL.

FIG. 5. Western blot analysis of proteins which selectively concentrate in fgNBs. (A) Lanes a, c, e, and g are control cells, while lanes b, d, f, and h are apoptotic cells. (B) Lanes a and c are control cells, while lanes b and d are apoptotic cells. (C) Lanes a and b are control cells, while c and d are apoptotic cells; in a and c antibody 20441 was used, whereas in b and d 107-7 was employed. The arrowhead points to 240-kDa NuMA, while the asterisk indicates the 40-kDa fragment. Polyacrylamide gels of 6% (A), 12% (B), and 10% (C) were used; 80 mg of protein was blotted to each lane.

high degree of specificity. Therefore, we believe that such an interpretation is unlikely. A very important point is the fact that fgNBs were present in all the apoptotic models we investigated, independent of the presence or absence of internucleosomal fragmentation, since they also occur in MOLT4 cells treated with staurosporine, in which no internucleosomal cleavage could be detected [12]. On the other hand, internucleosomal fragmentation was seen in HL60 cells exposed to camptothecin [51], in HeLa cells incubated with etoposide [55], and in Jurkat cells grown in the presence of EGTA [54]. Therefore, the formation of fgNBs could only require degradation of DNA in the large (300- and 50-kbp) fragments. On the contrary, Miller et al. [33] argued that changes observed in the localization of nuclear matrix proteins (lamin B, NuMA protein, and a polypeptide of nuclear pore complex) in MCF7 cells were strictly dependent on the type of apoptotic stimulus. In the case of adriamycin treatment that does not cause internucleosomal

AID

ECR 3415

/

6i16$$$723

01-20-97 17:19:05

DNA fragmentation, no modifications of the immunostaining patterns were seen. In this sense, it is also worth recalling that Tinnemans et al. [34] have shown that in the non-small-cell lung carcinoma cell line MR65 exposed to CHX, the apoptotic elements showed a dispersion throughout the nucleus of the immunoreactivity due to the nucleolar antigen Ki67, whereas we have shown that in apoptotic HL-60 cells the immunofluorescent pattern given by an antibody to nucleolar protein topoisomerase IIb did not change even during the very late stages of the process [46]. We believe that our findings are very interesting in light of the data published recently by Weaver et al. [83] indicating that the NuMA protein undergoes proteolysis very soon after treatment of thymocytes with dexamethasone (cleavage even precedes any substantial level of DNA fragmentation); such an observation has been confirmed by the report by Hsu and Yeh [84], who have detected a similar degradation of NuMA in apoptotic HeLa and HL-60 cells. Since Weaver et al. [83] used an antibody generated against an epitope residing in the C-terminal fragment of the protein, they concluded that the cleavage must have occurred within its N terminus, yielding the 200-kDa fragment and an undetectable 40-kDa fragment. Our data, obtained by means of two different antibodies to NuMA, confirm those findings and also demonstrate the existence of the 40-kDa fragment. Weaver et al. [83] reported changes in the distribution of NuMA in apoptotic thymocytes, as seen by fluorescence microscopy where condensed chromatin of apoptotic nuclei was not stained by antibody to NuMA polypeptide, while positivity was seen in the rest of the nucleoplasm. These findings are in good agreement with our own data, showing similar behavior of two nuclear matrix proteins (p160 and p125) in HL-60 cells treated with camptothecin [46], and support the contention of a progressive depolymerization of the nuclear matrix. All the proteins we have studied belong to the illTABLE 1 Densitometric Analysis of the Immunoblots Protein

HL-60 C

HL-60 A

U-937 C

U-937 A

p160 p125 PCNA 240-kDa NuMA (clone 204-41) 200-kDa NuMA (clone 204-41) 40-kDa NuMA (clone 107-7)

1.85 2.74 3.51

1.77 2.93 3.83

2.43 2.97 1.99

2.55 3.14 1.85

3.65

2.33

NA

NA

ND

1.71

NA

NA

ND

1.21

NA

NA

Note. C, control cells; A, apoptotic cells. The values are expressed as arbitrary units. Results are the mean from three different experiments. SD was less than 11%; ND, not detectable; NA, not assayed.

ecal

NUCLEAR BODIES IN APOPTOTIC CELLS

defined inner matrix network [24]. It is conceivable that the proteins undergoing proteolysis are key ‘‘structural’’ components of the scaffold (e.g., topoisomerase IIa, NuMA protein, lamins) and their cleavage would lead to matrix depolymerization, whereas those that are unaffected are ‘‘functional’’ components of the network (PCNA, p160, p125). However, it should be noted that poly(ADP-ribose)polymerase also is proteolytically cleaved, and one would expect this enzyme to play a functional and not a structural role. Tinnemans et al. [34] also showed that a protein of the inner nuclear matrix meshwork, a 13-kDa U1RNP particle, was not proteolytically cleaved while its immunolocalization changed in apoptotic cells. Taken together, results reported recently by several groups, including ours, unequivocally indicate that a degradation of the nuclear matrix takes place during apoptosis. This is at variance with Arends et al. [32]. These authors examined the protein composition of nuclear matrices from apoptotic thymocytes and found no significant differences in the electrophoretic profile when compared with control cells. Since there is no an overall degradation of matrix proteins during apoptosis, it might be that visual inspection of silver-stained gels failed to reveal dramatic differences that could indeed be revealed by immunochemical detection. A similar conclusion was also drawn by Patterson et al. [54] who analyzed protein synthesis in apoptotic Jurkat cells in comparison with control cells. Nevertheless, it is more difficult to explain why morphological differences were not observed by Arends et al. [32] in matrices prepared from apoptotic nuclei; at present we do not have an explanation for this apparent discrepancy. In any case, the common fate of matrix proteins during apoptosis would be to end up in the fgNBs. Therefore, we speculate that the fgNBs might represent aggregates of nuclear matrix proteins in which reciprocal interactions have been broken following the action of protease(s) acting on some critical targets. This might also indicate that disposal of the disassembled constituents of the nuclear matrix is not a disordered, fortuitous phenomenon, but, on the contrary, is reflective of some highly ordered process. In the future, it will be interesting to determine whether or not the lamins also associate with the fgNBs. Finally, because of their widespread presence, we propose that the fgNBs may be a new morphological hallmark of the apoptotic process. We thank Sonia Lach, Roberta Bortul, and Giovanna Baldini for helpful technical assistance and Drs. L. de Jong and R. van Driel, Institute for Biochemical Research, Amsterdam, The Netherlands, for the generous gift of antibody to p160. This work was supported by Italian MURST grants (40 and 60% to Universita` di Trieste and Fondi Associazione Italiana per la Ricerca sul Cancro), 1995 and 1996.

REFERENCES 1. Ellis, R. E., Yuan, J.-Y., and Horvitz, H. R. (1991) Annu. Rev. Cell Biol. 7, 663–698.

AID

ECR 3415

/

6i16$$$723

01-20-97 17:19:05

335

2. Cohen, J. J., Duke, R. C., Fadok, V. A., and Sellins, K. S. (1992) Annu. Rev. Immunol. 10, 267–293. 3. Abrams, J., White, K., Fessler, L. I., and Steller, H. (1993) Development 117, 29–43. 4. Kajstura, J., Mansukhani, M., Cheng, W., Reiss, K., Krajewski, S., Reed, J. C., Quaini, F., Sonnenblick, E. H., and Anversa, P. (1995) Exp. Cell Res. 219, 110–121. 5. Kerr, J. F. R., Wyllie, A. H., and Currie, A. R. (1972) Br. J. Cancer 26, 239–257. 6. Lockskin, R. A., and Zakeri, Z. (1991) in Current Communications in Cell and Molecular Biology (Tomei, L. D., and Cope, F. D., Eds.), Vol. 3, pp. 47–60, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 7. Bellamy, C. O. C., Malcomsom, R. D. G., Harrison, D. J., and Wyllie, A. H. (1995) Sem. Cancer Biol. 6, 3–16. 8. Wyllie, A. H., Kerr, J. F. R., and Currie, A. R. (1980) Int. Rev. Cytol. 68, 251–306. 9. Wyllie, A. H. (1980) Nature 284, 555–556. 10. Collins, R. J., Harmon, B. V., Gobe´, G. C., and Kerr, J. F. R. (1992) Int. J. Rad. Biol. 61, 451–453. 11. Mesner, P. W., Winters, T. R., and Green, S. H. (1992) J. Cell Biol. 119, 1669–1680. 12. Falcieri, E., Martelli, A. M., Bareggi, R., Cataldi, A., and Cocco, L. (1993) Biochem. Biophys. Res. Commun. 193, 19–25. 13. Oberhammer, F., Fritsch, G., Schmied, M., Pavelka, M., Printz, D., Purchio, T., Lassmann, H., and Schulte-Hermann, R. (1993) J. Cell Sci. 104, 317–326. 14. Oberhammer, F., Wilson, J. W., Dive, C., Morris, I. D., Hickman, J. A., Wakeling, A. E., Walker, P. R., and Sikorska, M. (1993) EMBO J. 12, 3679–3684. 15. Bicknell, G. R., Snowden, R. T., and Cohen, G. M. (1994) J. Cell Sci. 107, 2483–2489. 16. Beere, H. M., Chresta, C. M., Alejo-Herberg, A., Skladanowski, A., Dive, C., Larsen, A. K., and Hickman, J. A. (1995) Mol. Pharmacol. 47, 986–996. 17. Yuan, J. Y., Shaham, S., Ledoux, S., Ellis, H. M., and Horvitz, H. R. (1993) Cell 75, 653–660. 18. Miura, M., Zhu, H., Rotello, R., Hartweig, E. A., and Yuan, J. Y. (1993) Cell 75, 653–660. 19. Brancolini, C., Benedetti, M., and Schneider, C. (1995) EMBO J. 14, 5179–5190. 20. Martin, S. J., and Green, D. R. (1995) Cell 82, 349–352. 21. Zhivotovsky, B., Gahm, A., Ankarcrona, M., Nicotera, P., and Orrenius, S. (1995) Exp. Cell Res. 221, 404–412. 22. Earnshaw, W. C. (1995) Curr. Op. Cell Biol. 7, 337–343. 23. Cook, P. R. (1988) J. Cell Sci. 90, 1–6. 24. Berezney, R. (1991) J. Cell. Biochem. 47, 109–123. 25. Nelson, W. G., Pienta, K. J., Barrack, E. R., and Coffey, D. S. (1986) Annu. Rev. Biophys. Chem. 15, 457–475. 26. Gerdes, M. G., Carter, K. C., Moen, P. T., and Lawrence, J. B. (1994) J. Cell Biol. 126, 289–304. 27. Jackson, D. A., Dickinson, P., and Cook, P. R. (1990) Nucleic Acids Res. 18, 4385–4393. 28. Jackson, D. A., Dickinson, P., and Cook, P. R. (1990) EMBO J. 9, 567–571. 29. Mirkovitch, J., Mirault, M.-E., and Laemmli, U. K. (1984) Cell 39, 223–232. 30. Gasser, S. M., and Laemmli, U. K. (1986) EMBO J. 5, 511–518. 31. Izaurralde, E., Mirkovitch, J., and Laemmli, U. K. (1988) J. Mol. Biol. 200, 111–125. 32. Arends, M. J., Morris, R. G., and Wyllie, A. H. (1990) Am. J. Pathol. 136, 593–608.

ecal

336

ZWEYER ET AL.

33. Miller, T., Beausang, L. A., Meneghini, M., and Lidgard, G. (1993) Biotechniques 15, 1042–1047. 34. Tinnemans, M. M. F. J., Lenders, M.-H. J. H., ten Velde, G. P. M., Ramaekers, F. C. S., and Schutte, B. (1995) Eur. J. Cell Biol. 68, 35–46. 35. Berrios, M., Osheroff, N., and Fisher, P. A. (1985) Proc. Natl. Acad. Sci. USA 82, 4142–4146. 36. Kaufmann, S. H. (1989) Cancer Res. 49, 5870–5878. 37. Kaufmann, S. H., and Shaper, J. H. (1991) Exp. Cell Res. 192, 511–523. 38. Kaufmann, S. H., Brunet, G., Talbot, B., Lamarr, D., Dumas, C., Shaper, J. H., and Poirier, G. (1991) Exp. Cell Res. 192, 524– 535. 39. Lazebnik, Y. A., Kaufmann, S. H., Desnoyers, S., Poirier, G. G., and Earnshaw, W. C. (1994) Nature 371, 346–347. 40. Nicholson, D. W., Ali, A., Thornberry, N. A., Vaillancourt, J. P., Ding, C. K., Gallant, M., Gareau, Y., Griffin, P. R., Labelle, M., Lazebenik, Y. A., Munday, N. A., Raju, S. M., Smulson, M. E., Yamin, T.-T., Yu, V. L., and Miller, D. K. (1995) Nature 376, 37–43. 41. Oberhammer, F. A., Hochegger, K., Froschl, G., Tiefenbecher, R., and Pavelka, M. (1994) J. Cell Biol. 126, 827–837. 42. Lazebnik, Y. A., Takahashi, A., Moir, R. D., Goldman, R. D., Poirier, G. G., Kaufmann, S. H., and Earnshaw, W. C. (1995) Proc. Natl. Acad. Sci. USA 92, 9042–9046. 43. Lazebnik, Y. A., Takahashi, A., Poirier, G. G., Kaufmann, S. H., and Earnshaw, W. C. (1995) J. Cell Sci. Suppl. 19, 41–49. 44. Zhivotovsky, B., Wade, D., Gahm, A., Orrenius, S., and Nicotera, P. (1994) FEBS Lett. 351, 150–154. 45. Lagarkova, M. A., Iarovaia, O. V., and Razin, S. V. (1995) J. Biol. Chem. 270, 20239–20241. 46. Zweyer, M., Bareggi, R., Grill, V., Soranzo, M. R., Marugg, R. A., Riederer, B. M., Narducci, P., and Martelli, A. M. (1995) Exp. Cell Res. 221, 27–40. 47. Dickinson, L. A., Joh, T., Kohwi, Y., and Kohwi-Shigematsu, T. (1992) Cell 70, 631–645. 48. Fackelmayer, F. O., Dahm, K., Renz, A., Ramsperger, U., and Richter, A. (1994) Eur. J. Biochem. 221, 749–757. 49. Koken, M. H. M., Puvion-Dutilleul, F., Guillemin, M. C., Viron, A., Linares-Cruz, G., Stuurman, N., de Jong, L., Szostecki, C., Calco, C., Chomienne, C., Degos, L., Puvion, E., and De The´, H. (1994) EMBO J. 13, 1073–1083. 50. Weis, K., Rambaud, S., Lavau, C., Jansen, J., Carvalho, T., Carmo-Fonseca, M., Lamond, A., and Dejean, A. (1994) Cell 76, 345–356. 51. Del Bino, G., Lassota, P., and Darzynkiewicz, Z. (1991) Exp. Cell Res. 193, 27–35. 52. Kressel, M., and Groscurth, P. (1994) Cell Tissue Res. 278, 549– 555. 53. Cossarizza, A., Franceschi, C., Monti, D., Salvioli, S., Bellesia, E., Rivabene, R., Biondo, L., Rainaldi, G., Tinari, A., and Malorni, W. (1995) Exp. Cell Res. 220, 232–240. 54. Patterson, S. D., Grossman, J. S., D’Andrea, P., and Latter, G. I. (1995) J. Biol. Chem. 270, 9429–9436. 55. Guano, F., Bernardi, R., Negri, C., Donzelli, M., Prosperi, E., Astaldi Ricotti, G., and Scovassi, A. I. (1994) Cell Death Diff. 1, 101–107. 56. de Graaf, A., Meijne, A. M. L., van Renswoude, A. J. B. M.,

57.

58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73.

74. 75. 76. 77. 78. 79. 80. 81. 82.

83.

84.

Humbel, B. M., van Bergen en Henegouwen, P. M. P., de Jong, L., van Driel, R., and Verkleij, A. J. (1992) Exp. Cell Res. 202, 243–251. Neri, L. M., Santi, S., Marugg, R. A., Riederer, B. M., Capitani, S., Cataldi, A., and Martelli, A. M. (1994) Exp. Cell Res. 213, 275–285. Ochs, R. L., Lischwe, M. A., Spohn, W. H., and Busch, H. (1985) Biol. Cell 54, 123–134. Bohmann, K., Ferreira, J. A., and Lamond, A. I. (1995) J. Cell Biol. 131, 817–831. Zeng, C., He, D., Berget, S. M., and Brinkley, B. R. (1994) Proc. Natl. Acad. Sci. USA 91, 1505–1509. Clevenger, C. V., and Epstein, A. L. (1984) J. Histochem. Cytochem. 32, 757–765. Clevenger, C. V., and Epstein, A. L. (1984) Exp. Cell Res. 151, 194–207. Bradford, M. M. (1976) Anal. Biochem. 72, 248–254. Laemmli, U. K. (1970) Nature 227, 680–685. Martelli, A. M., Manzoli, L., Rubbini, S., Billi, A. M., Bareggi, R., and Cocco, L. (1995) Biol. Cell 83, 15–22. Brasch, K., and Ochs, L. (1992) Exp. Cell Res. 202, 211–223. Ochs, R. L., Stein, T. W., and Tan, E. M. (1994) J. Cell Sci. 107, 385–399. Bohmann, K., Ferreira, J., Santama, N., Weis, K., and Lamond, A. I. (1995) J. Cell Sci. Suppl. 19, 107–113. Ochs, R. L., and Smetana, K. (1991) Exp. Cell Res. 197, 183– 190. Recher, L., Chan, H., Briggs, L., and Parry, N. (1972) Cancer Res. 32, 2495–2501. Gajkowska, B., Puvion, E., and Bernhard, W. (1977) J. Ultrastr. Res. 60, 335–347. Hernandez-Verdun, D., Derenzini, M., and Bouteille, M. (1984) J. Ultrastr. Res. 88, 55–65. Puvion-Dutilleul, F., Mazan, S., Nicoloso, M., Pichard, E., Bachellerie, J. P., and Puvion, E. (1992) Eur. J. Cell Biol. 58, 149– 162. Yokoyama, Y., Niwa, K., and Tamaya, T. (1992) Exp. Cell Res. 202, 77–86. Farabegoli, F., Govoni, M., and Novello, F. (1992) Biol. Cell 74, 281–286. von Kries, J. P., Buck, F., and Stratling, W. H. (1994) Nucleic Acids Res. 22, 1215–1220. Bravo, R., and Macdonald-Bravo, H. (1987) J. Cell Biol. 105, 1549–1554. Laszlo, A. (1992) Cell Prolif. 25, 59–87. Daev, E., Chaly, N., Brown, D. L., Valentine, B., Little, J. E., Chen, X., and Walker, P. R. (1994) Exp. Cell Res. 214, 331–342. Leist, M., Gantner, G., Bohlinger, I., Germann, P. G., Tiegs, G., and Wendel, A. (1994) J. Immunol. 153, 1778–1788. Puvion-Dutilleul, F., Chelbi-Alix, M. K., Koken, M., Quignon, F., Puvion, E., and De The´, H. (1995) Exp. Cell Res. 218, 9–16. Carvalho, T., Seeler, J. B., Ohman, K., Jordan, P., Pettersson, U., Akusjarvi, G., Carmo-Fonseca, M., and Dejean, A. (1995) J. Cell Biol. 131, 45–56. Weaver, V. M., Carson, C. E., Walker, P. R., Chaly, N., Lach, B., Raymond, Y., Brown, D. L., and Sikorska, M. (1996) J. Cell Sci. 109, 45–56. Hsu, H.-L., and Yeh, N.-H. (1996) J. Cell Sci. 109, 277–288.

Received July 5, 1996 Revised version received October 21, 1996

AID

ECR 3415

/

6i16$$$724

01-20-97 17:19:05

ecal