Persistence of major nuclear envelope antigens in an envelope-like structure during mitosis in Drosophila melanogaster embryos

The nuclear envelope is composed of inner and outer lipid bilayer membranes, which are separated by the perinuclear space. The two membranes join at the pore complexes, presumed passageways for the exchange of macromolecules between the nucleus and cytoplasm (e.g. see Feldherr et al 1984; Dworetzky and Feldherr, 1988). Underlying the inner membrane is the nuclear lamina (Franke, 1974), a proteinaceous layer of intermediate filament-like fibrils (Aebi et al. 1986). Major polypeptide components of the nuclear lamina, termed lamins, have been identified in a number of organisms (for reviews, see Gerace, 1986; Krohne and Benavente, 1986). In higher eukaryotes, the nuclear envelope breaks down at the onset of mitosis, only to be re-formed in each of the two daughter cells as mitosis is completed. This process has been termed open mitosis (see Franke, 1974). In contrast, lower eukaryotes such as yeast and myxomycètes, undergo mitotic chromosome segregation within an essentially intact nuclear envelope (see Heath, 1980). Nuclear division is accomplished by karyokinesis, a process not unlike cytokinesis in higher organisms. This has been termed closed mitosis.

Efforts from several laboratories have focused on the fate of vertebrate nuclear envelope components during mitosis (Ely et al. 1978; Gerace et al. 1978; Jost and Johnson, 1981; Krohne et al. 1978; Ottaviano and Gerace, 1985). Immunocytological studies performed in conjunction with cell fractionation experiments showed that during prophase, the polymeric lamina disintegrates into lamin oligomers that are dispersed uniformly in the cytoplasm. This pattern of protein distribution persists until telophase, at which time the lamina begins to reform at the centrosomal regions until it fully envelopes the chromatin in each of the daughter nuclei. Similar results were obtained with antibodies directed against vertebrate nuclear pore complex proteins including p62 (Davis and Blobel, 1986), gpl90 (Gerace et al. 1982) and a number of other pore complex glycoproteins (Snow et al. 1987). These results suggest that during mitosis in vertebrates, both lamina and pore complexes disassemble, coincident with the breakdown of the nuclear envelope observed microscopically.

Drosophila melanogaster is an organism well-suited to the study of nuclear envelope structure and function. Homologs of many of the well-characterized vertebrate nuclear envelope proteins have been identified (for a review, see Fisher, 1988). Moreover, the rapid early phases of Drosophila embryogenesis provide a unique perspective from which to view mitosis and the events surrounding it. The first 13 nuclear division cycles in the embryo occur rapidly and nearly synchronously within a syncitium. Each of the first nine cycles lasts approximately 10 min; the last four cycles are each approximately 15 min in length (Zalokar and Erk, 1976). Initial reports on the fate of the nuclear lamina during mitosis in Drosophila early embryos showed that, in prophase, the lamina apparently invaginated from both sides towards the central plane of mitosis. During metaphase, the lamina started to break down and lamins were apparently dispersed in the cytoplasm in irregularly shaped particles (Fuchs et al. 1983). On the other hand, detailed electron-microscopic analysis of the nuclear membranes in Drosophila early embryos (prior to cellularization) revealed that, during mitosis, the nuclear membranes broke at the spindle poles but remained fully in evidence elsewhere (Stafstrom and Staehelin, 1984). However, nuclear pore complexes associated with the envelope during interphase apparently disassembled and were lost during mitosis, leaving behind numerous fenestrations in the membrane. It was also shown in this study that in mitosis, a second layer of closely adherent membranous cisternae was acquired just outside the original membranes (Stafstrom and Staehelin, 1984).

To elucidate further the fate of the Drosophila nuclear envelope during mitosis in early embryos, we followed the fates of three different nuclear envelope proteins by indirect immunofluorescence. We show that at least a portion of both otefin (which is a transliteration of the Hebrew word meaning literally, ‘envelopes’), a 53K (K = 10³M_r) protein that we localized to the inner nuclear membrane and/or lamina, and lamin^* are retained in a structure that envelopes the entire mitotic apparatus including condensed chromosomes and spindle, while gpl88, a putative pore complex protein, is apparently released from the envelope during prophase and reinte-grates into the envelope during interphase. These results confirm the suggestion of Stafstrom and Staehelin (1984) that during the rapid early phases of embryogenesis, Drosophila nuclei undergo what may in effect be a form of closed mitosis similar to that seen in lower eukaryotes.

Monoclonal anti-Drosophila lamin antibody, 611A3A6 and monoclonal anti-otefin antibody 618A207 (Miller et al. 1985) were gifts from Dr Bruce Alberts. Monoclonal anti-Drosophila gpl88 antibodies AGP-26 and AGP-78 (Filson et al. 1985) were purified from tissue-culture supernatants and concentrated by ammonium sulfate precipitation as previously (McConnell et al. 1987). 5-nm gold-conjugated goat anti-mouse IgG was from Janssen Pharmaceutica (Piscataway, NJ). Affinity-purified Texas Red-conjugated rabbit anti-mouse IgG and rhodamine-conjugated goat anti-mouse IgG were from Jackson Laboratories (West Grove, PA).

About 5×10⁸ Schneider cells were washed once in RSB (10 mM-NaCl, 3mM-MgCl2, 10mM-Tris-HCl, pH 7.4, 0.5M-phenyl-methylsulfonyl fluoride (PMSF)), incubated for 10min on ice and Dounce homogenized (50 strokes with a tight pestle). Following centrifugation (750g, 10min, 4°C), the supernatant was collected (fraction SFI). The pellet was resuspended in RSB, loaded on a cushion of RSB containing 1.9M-sucrose and centrifuged (VTI 50.1 rotor, 24 000 revs min⁻¹, 30min, 4°C). The isolated nuclei were washed once in RSB containing 0.25 M-sucrose (RSBS), resuspended in 2ml RSBS and an equal volume of cold solution containing 2M-NaCl, 10 mM-EDTA, 10mM-Tris-HCl, pH7.4, was then added to the tube. Following 10 min incubation at room temperature, the tube was centrifuged (4500g, 15 min, 4°C), the supernatant was collected (fraction SFII) and the pellet was subjected to two washes in RSBS (fraction P). Protein lysates from the different subcellular fractions were prepared as described (McConnell et al. 1987). SDS-PAGE was according to Laemmli (1970). Blots were first incubated for 2h with 618A207 anti-otefin antibody (diluted 1:2), followed by 2h incubation with ^l25I-labeled sheep anti-mouse IgG antibody (Amersham, England) and exposed to X-ray film.

Collection, permeabilization and fixation were essentially as described by Karr and Alberts (1986). Drosophila melanogaster (Canton-S) embryos (0 – 3 h old) were collected at 25°C, rinsed in 0.4% NaCl, 0.03% Triton X-100 and dechorionated in a half-strength solution of commercial bleach. Dechorionated embryos were washed with 0.4% NaCl, 0.03% Triton X-100 to remove all traces of sodium hypochlorite and transferred to a 50 ml round-bottom flask containing 2 ml of 60mM-KCl, 15mM-NaCl, 0.15 mM-spermine, 0.5 mM-spermidine, 15 mM 2-mercaptoethanol, 15 mM-Tris-HCl, pH 7.4 (buffer A of Wallace et al. 1971), plus 8 ml of heptane. Taxol was added to a final concentration of 0.5 μ M, and after less than 30 s of vigorous shaking, 1 ml of freshly prepared 37% formaldehyde was added. Shaking was continued for an additional 10min. For devitellinization, the fixed embryos were collected, rinsed with PBS and transferred to a 50 ml round-bottom flask, previously cooled to − 70°C, which contained 5 ml of heptane, 4.5 ml of methanol and 0.5 ml of 50mM-EGTA, pH 7.0. After vigorous shaking for 10 min, the temperature was rapidly raised to approximately 23 °C by swirling the flask under a stream of hot tap water (Mitchison and Sedat, 1983). Devitellinized embryos were then rinsed with a methanol: EGTA solution minus the heptane and rehydrated by passage through solutions of methanol: PBS (4:1, 1:4 (v/v)) and finally, PBS alone. Indirect immunofluorescent staining of whole fixed embryos was also essentially as described by Karr and Alberts (1986). All incubations were done at room temperature, using gentle rotation in a humid chamber. Embryos were transferred into deep-depression slides containing PBS with 1% bovine serum albumin (BSA) and 0.1% Triton X-100 (PBSBT). Embryos were then incubated with primary antibody for 2h. When hybridoma tissue culture supernatant was used, it was diluted 1:1 (v/v) with PBSBT. Otherwise antibodies were diluted to a concentration of about 5 μ gml⁻¹ in PBSBT. Following a 3-h rinse in PBSBT, embryos were stained for 2 h with either rabbit or goat anti-mouse IgG coupled to Texas Red or rhodamine, respectively. Embryos were then rinsed for 2h in PBSBT. DNA was stained with 4,6-diamidino-2-phenylindole (DAPI) for 5 min in a 1 μ g ml⁻¹ solution of the dye in PBS. Following a 10 min rinse in PBS, the embryos were transferred successively into solutions of 1:4,2:3,3:3 and 4:1, glycerol: PBS (v/v) and finally mounted on a microscope slide in a solution of 2% w-propyl gallate in glycerol (Giloh and Sedat, 1982). Drosophila embryos were viewed in a Zeiss Universal microscope equipped with epifluorescence illumination and a 100 × /1.3 NA Planapo objective. Photographs were taken on Kodak Technical Pan 2415 film and developed with D19 or HCl10 developer.

Drosophila melanogaster (Canton-S) embryos (0-3.5 h old) were collected, dechorionated, fixed, devitelhmzed and reacted with monoclonal antibody as described above. All incubation steps were performed at room temperature. Following a 2-h incubation with antibody, the embryos were rinsed for 3 h in PBSBT and then incubated with 5-nm colloidal gold-conjugated goat anti-mouse IgG diluted in PBS with 1% BSA (PBSB). Embryos were next rinsed for 3 h with PBSB, then for 1 h in PBS, and then post-fixed in a solution containing PBS with 2% glutaraldehyde and 2% formaldehyde for 30 min. Following rinsing for 30 min with PBS, embryos were next fixed with 1% OsO₄. The post-fixed embryos were dehydrated in ethanol/propylene oxide and embedded in Epon. From the embedded embryos, 70 nm thick sections were cut, stained with uranyl acetate and lead citrate and viewed with a Phillips 300 transmission electron microscope at 60 kV.

Initial experiments with monoclonal antibody 618A207 suggested that it specifically recognized a 53K polypeptide, here termed otefin, that was localized to the periphery of the nucleus by indirect immunofluorescence (Miller et al. 1985). We confirmed these observations in both tissue culture cells (not shown) and embryos (see Fig. 3, for example). Immunoblot analysis performed on various developmental stages of Drosophila revealed that otefin is present in relatively large amounts during the embryonic stages and during pupation (Fig. 1A). We extended the analysis to include cell fractionation experiments and immunoelectron microscopy. These results are shown in Figs 1B and 2, respectively.

Following homogenization of Drosophila Schneider tissue-culture cells most or all of the otefin present in the crude homogenate (Fig. 1B, lane T) was recovered in the isolated nuclei (Fig. 1B, lane N). Also, most of the otefin in isolated nuclei was resistant to 1 M-NaCl extraction (Fig. 1B, lanes P and SFII). Identical results have previously been reported for the Drosophila lamins (Filson et al. 1985).

Results of immunoelectron microscopic experiments (Fig. 2) added to impressions obtained previously from indirect immunofluorescence analyses (Miller et al. 1985). Drosophila embryos from both the syncitial and cellular stages were fixed with formaldehyde in the presence of taxol (Karr and Alberts, 1986), incubated first with monoclonal anti-otefin antibody and then with 5-nm gold-conjugated goat anti-mouse IgG. After appropriate washing, the embryos were post-fixed with glutaraldehyde and OsO₄, embedded in Epon and sectioned for electron microscopy. Upon examination of hundreds of nuclei in several independent experiments, we found gold particles restricted to the inner membrane of the nuclear envelope (Fig. 2A and B). This result is consistent with localization either to the membrane itself or to the underlying structures of lamina and pore complexes. Tangential views of the nuclear envelope showed that the nuclear pore complexes were essentially unreactive with the anti-otefin antibody apparently ruling out pore complex localization (Fig. 2C). Probing of similar specimens with gold-conjugated goat anti-mouse IgG alone failed to reveal any nuclear envelope staining (not shown).

Indirect immunofluorescence microscopy was used to investigate the fate of otefin during the rapid mitotic cycles characteristic of early embryogenesis in Drosophila. These results are shown in Fig. 3. To determine the different stages of mitosis in the specimens being examined, we used the DNA-specific dye, DAPI. DAPI-stained specimens are shown in the left-hand panels; immunofluorescence micrographs of the same fields are shown on the right. In interphase, otefin was confined to the nuclear periphery and labeling of the nuclear envelope appeared to be diffusely granular (Fig. 3A). During prophase, otefin remained in a round envelope (Fig. 3B) and this morphology persisted through prometaphase (Fig. 3C). At this stage, spindle poles on opposite sides of the nucleus were also labeled, suggesting that some otefin may be associated with the spindle poles. At metaphase, a dramatic change in the immunofluorescence staining pattern obtained with anti-otefin antibody was observed. The envelope-like structure defined by immunofluorescent staining elongated perpendicular to the metaphase plate, enveloping the entire mitotic apparatus including the condensed chromosomes, spindle and spindle poles (Fig. 3D). During anaphase, this envelope-like structure was first elongated further and then became barbell-shaped (Fig. 3E). In telophase, the barbell shape was lost as otefin was distributed into each of the two daughter nuclei.

As the embryonic cell entered mitosis there was a reduction in the signal intensities and, in addition to the envelope-like staining seen with the anti-otefin antibody, there was an increase in the diffuse cytoplasmic background staining. This, in conjunction with the fact that envelope-like staining during mitosis was reduced relative to nuclear envelope staining seen with anti-otefin antibody during interphase, suggests that a portion of the otefin may be solubilized and redistributed during mitosis while a second portion remains associated with an envelope-like structure.

In comparing our current data on otefin localization during mitosis in early embryos with results of similar experiments performed with anti-lamin antibodies (Fuchs et al. 1983), we were struck by the apparent difference between the two antigens. As noted in the Introduction, Fuchs et al. (1983) reported that during mitosis the lamins were apparently dispersed in granular patches through the surrounding cytoplasm. There was no evidence for the persistence of an envelope-like structure surrounding the mitotic apparatus on the basis of staining with anti-lamin antibodies. In comparing our results for otefin with those of Fuchs et al. (1983) for lamin, one methodological difference between our current analysis and that of Fuchs et al. (1983) was noted, i.e. that we included taxol in our fixation step. We therefore repeated the staining of early embryos with anti-otefin antibody without inclusion of taxol at any point in the procedure. The results of this experiment are shown in Fig. 4.

Without taxol present during fixation, the results obtained with anti-otefin antibody were similar to those that had previously been reported for the lamin (Fuchs et al. 1983). During prophase, the envelope as revealed by otefin staining started to invaginate (Fig. 4A). During metaphase, the majority of the antigen was redistributed in the cytoplasm (Fig. 4B) and in telophase, most of the antigen was located over the centrosomes (Fig. 4C).

In light of the effects of taxol on the immunofluorescence staining pattern revealed with anti-otefin antibodies, we set about to reinvestigate the distribution of lamin during mitosis in early embryos. Embryos were fixed, permeabilized and probed with monoclonal anti-lamin anti-bodies exactly as described above for the experiments with anti-otefin shown in Fig. 3. Results of these analyses shown in Fig. 5 are essentially identical with those seen with anti-otefin antibody. During interphase, typical nuclear envelope staining was observed (Fig. 5A). In prophase, when the chromosomes started to condense, the anti-lamin staining remained in a rounded structure (Fig. 5B), which became elongated at metaphase perpen-dicular to the mitotic plate and included the spindle pole area (Fig. 5C). Early in anaphase, further elongation was observed (Fig. 5D), which become more pronounced during late anaphase (Fig. 5E).

In addition to the persistent envelope-like staining revealed with anti-lamin antibodies during mitosis, we also noted an increase in diffuse cytoplasmic background staining. This is again consistent with the notion that only a portion of the lamin remained in an envelope-like structure during mitosis while the remainder was redis-tributed through the surrounding cytoplasm.

Both lamin and otefin are apparently associated with the inner membrane of the nuclear envelope and neither is found in nuclear pore complexes. On the other hand, gpl88 is the Drosophila homolog of mammalian gpl90, a transmembrane glycoprotein localized to the rat liver nuclear pore complex by Gerace et al. (1982). Pore complex localization for gpl88 in Drosophila salivary gland nuclei has recently been confirmed (M. Berrios, personal communication). We therefore set about to investigate the distribution of gpl88 during mitosis in Drosophila early embryos using two monoclonal anti-bodies directed against this protein, AGP-26 and AGP-78 (Filson et al. 1985). Results of an experiment performed with AGP-26 are shown in Fig. 6. Similar results were obtained with AGP-78 (not shown). During interphase, nuclear envelope staining by anti-gpl88 antibodies (Fig. 6A) was similar to that observed with anti-lamin and anti-otefin antibodies. However, in contrast with otefin and lamin, gpl88 appeared to be diffusely localized throughout the cytoplasm during mitosis from prophase onwards (Fig. 6B and C). There was no immunocytological evidence for a persistent envelope-like structure during mitosis based on results with anti-gpl88 anti-bodies.

Stafstrom and Staehelin (1984) originally reported that during mitosis in Drosophila early embryos, a membranous envelope was detectable surrounding the mitotic apparatus. They termed this structure the spindle envelope, suggested that it derived at least in part from the nuclear envelope that was present during interphase, and noted that it differed from the nuclear envelope in at least two obvious ways. First, it lacked nuclear pore complexes but instead contained numerous fenestrations where pore complexes might have originally been. Second, the formation of a second layer of membranous cisternae was noted, located outside the first, so that the so-called spindle envelope was in effect, a double envelope.

In our current study, we used monoclonal antibodies directed against three major Drosophila nuclear envelope proteins, lamin, otefin and gpl88, to perform immunocytochemical analyses aimed at elucidating the fates of each during mitosis in Drosophila early embryos. We found that both lamin and otefin, components of the submembranous lamina and lamina and/or inner nuclear membrane, respectively, are detectable as components of the spindle envelope. In contrast, gpl88, a transmembrane glycoprotein that is a putative component of the nuclear pore complex, is apparently not associated with the spindle envelope, but rather redistributes completely into the surrounding cytoplasm during most or all of mitosis.

Although the fate of gpl88 is consistent with expectations for the behavior of a pore complex component, on the basis of the work of Stafstrom and Staehelin (1984), it is in some ways surprising. We might have anticipated that as an integral membrane protein, gp188 would remain associated with any membranous structures that persisted during mitosis. This observation raises the interesting possibility, therefore, that the spindle envelope is derived primarily from the inner membrane of the interphase nucleus while the outer nuclear membrane, including pore-complex components such as gpl88, is dispersed to the surrounding cytoplasm. Alternatively, it is possible that the spindle envelope is not derived from the nuclear envelope at all but rather originates somewhere in the cytoplasm and actually replaces the nuclear envelope during mitosis.

At this point, our impression is that only a portion of the lamin and otefin associated with the nuclear envelope during interphase remains associated with the spindle envelope in mitosis while the rest is redistributed through the cytoplasm. In the case of lamin, we have recently been able to show such redistribution in the final stages of oogenesis where germinal vesicle breakdown is accompanied by specific changes in lamin phosphorylation patterns and by lamin solubilization in a dimeric form; similar biochemical changes apparently take place during mitosis in Drosophila Schneider 2 tissue-culture cells (Smith and Fisher, 1989). Comparable studies remain to be performed for otefin.

It also remains to be determined whether the spindle envelope that is apparent during early embryogenesis when the nuclear division cycle is rapid is present during the more conventionally paced mitosis that occur later in development and in tissue culture cells. It was previously reported that during mitosis in Drosophila neural ganglion cells, lamin was redistributed uniformly throughout the mitotic cell as determined by indirect immunofluorescence (Berrios et al. 1985) and we have recently made similar observations with tissue-culture cells (Smith and Fisher, 1989). No evidence for a lamin-containing spindle envelope was noted in these studies. However, those experiments were performed in the absence of taxol. In light of our current observations on the effects of taxol in apparently stabilizing the spindle envelope in early embryos as well as previous suggestions by other workers (Strafstrom and Staehelin, 1984), we think that the question of its existence in other Drosophila cell types merits direct investigation. If the spindle envelope does exist in other cell types, particularly in tissue-culture cells, it may be possible to devise cell fractionation experiments directed first at quantitating the relative amounts of lamin and otefin that are solubilized and redistributed during mitosis versus those that remain membrane-associated and, subsequently, at elucidating the precise biochemical basis for the partitioning.

It is uncertain what role taxol plays in facilitating immunocytochemical visualization of the spindle envelope in Drosophila early embryos. Karr and Alberts (1986) demonstrated that addition of taxol was necessary to preserve microtubules when formaldehyde was used as a fixative in preparing Drosophila early embryos for immunocytochemistry. In their analysis, they found that taxol was without apparent effect on other cytoskeletal elements such as actin filaments; Karr and Alberts (1986) suggested that, at the concentrations used, the major effect of taxol was to stabilize pre-existing microtubules rather than induce the formation of new ones. Although we cannot rule out some direct effect of taxol on specific spindle envelope components, the simplest conclusion to be drawn from our current data is that the spindle envelope is dependent for its integrity on the integrity of the spindle itself and taxol thereby stabilizes the spindle envelope indirectly.

In conclusion, we would like to offer two hypotheses regarding the biological significance of the spindle envelope as it has been identified during mitosis in Drosophila early embryos. First, we think it possible that existence of the spindle envelope is secondary to the high rate of nuclear division during this stage in the Drosophila life cycle. Either the pre-existing nuclear envelope simply does not have time to break down completely during mitosis, or new nuclear envelopes begin to form from the large pool of soluble precursors immediately upon initiation of breakdown of the old one. In either case, the underlying hypothesis can be tested by detailed examination of mitosis in other Drosophila cell and tissue types where the cell cycle is much longer. The second hypothesis that we would like to offer is one that is not mutually exclusive with the first and concerns the functional significance of the spindle envelope. We think it may be advantageous for the organism to maintain some degree of separation between nuclear and cytoplasmic compartments during mitosis; for example, to conserve nuclear protein factors involved in gene expression. Such conservation may be more or less important during different developmental stages and the spindle envelope may be present to varying degrees as a result. As with our first hypothesis, deeper insight into this possibility will probably be forthcoming once we have been able to determine whether the spindle envelope exists in other Drosophila cells and tissues.

We thank Kathy Miller, Doug Kellogg and Bruce Alberts for monoclonal antibodies 611A3A6 and 618A207. This work was supported by Research grant 86-00023/1 from the US-Israel Binational Science Foundation.