Transfer of mouse nuclear envelope specific proteins to nuclei of chick erythrocytes during reactivation in heterokaryons with mouse a9 cells

The chromatin of the mature avian erythrocyte is tightly condensed and inactive n DNA and RNA synthesis (Cameron & Prescott, 1963; Harris, 1965). However, if heterokaryons are formed by cell fusion between chick red blood cells and cells which are active in DNA replication and transcription, the inert chick red cell nuclei undergo enlargement and resume the synthesis of DNA and RNA. (For review of the work, see Harris (1974) and Ringertz & Savage (1976).) The enlargement of the erythrocyte nuclei is accompanied by an increase in dry mass, a dispersion of the condensed chromatin (Harris, 1967; Bolund, Ringertz & Harris, 1969) and the migration of nucleospecific proteins from the active cell into the erythrocyte nucleus (Ringertz, Carlsson, Ege & Bolund, 1971 ; Ege, Carlsson & Ringertz, 1971 ; Goto & Ringertz, 1974; Appels, Bolund & Ringertz, 1974a; Appels, Bolund, Goto & Ringertz, 1974,b). The migration of the signals activating the chick erythrocyte genome is selective. Enzymes and antigens characteristic of the mammalian cytoplasm as well as certain non-histone proteins of the nucleus are excluded from the enlarging erythrocyte nucleus (Ringertz et al. 1971; Appels et al. 1974a; Appels, Tallroth, Appels & Ringertz, 1975).

The total surface area of the erythrocyte nuclear envelope increases during reactivation. We studied the nuclear envelope during reactivation to see whether structural components of mammalian origin were integrated into the erythrocyte nuclear envelope and whether envelope alterations preceded the influx of nucleo-specific proteins required for the resumption of DNA and RNA synthesis.

These studies were made possible by using specific antisera directed against the major polypeptides of the pore complex fraction from rat liver (Ely, d’Arcy & Jost, 1978; Krohne et al. 1978). These antisera react strongly with various mammalian but not with chicken cells. The experiment show that the nuclear envelope specific antigens of mouse origin accumulate in the envelope of the reactivating erythrocytes and strengthen the view that protein influx also plays a major role during the growth of the erythrocyte envelope in heterokaryons.

Mouse A9 cells (Engel, McGee & Harris, 1969) were grown in Eagle’s MEM supplemented with 10% foetal calf serum and antibiotics. Chick embryo fibroblasts, obtained from trypsin-treated embryos were grown as monolayer cultures in Dulbecco’s MEM supplemented with 20% foetal calf serum and antibiotics. Chick erythrocytes were obtained from 14-day-old embryos as described (Bolund et al. 1969). They were fused with mouse A9 cells using u.v.-inactivated Sendai virus. After fusion, the heterokaryons were grown on coverslips. The relevant techniques have been described previously (Harris, Watkins, Ford & Schoefl, 1966; Harris, Sidebottom, Grace & Bramwell, 1969; Harris, 1967). The method used to isolate nuclei in nonaqueous media from 14-day-old chick erythrocytes was described by Kirsch et al. (1970). Antigen preparation, immunization and indirect fluorescence microscopy have been described previously (Ely et al. 1978).

The antisera against 3 major polypeptides of the pore complex-fibrous lamina fraction from rat liver used in this study have been described previously (Ely et al. 1978; Krohne et al. 1978). The antigens of apparent molecular weights of 71000, 67000, and 65000 were designated lamina proteins LP71, LP67, and LP65. Antisera to these proteins were raised in chickens and were designated anti-LP71, anti-LP67, and anti-LP65, respectively.

Antigen distribution in cells of a variety of different species has been examined by indirect immunofluorescence microscopy. Fig. 1A-D demonstrate the strong and specific fluorescence of the nuclear periphery in interphase cells of the mouse A9 line, when decorated with highly diluted anti-LP67. Control sera do not stain the cells specifically. A similar pattern of antigen distribution was also observed with anti-LP71 and anti-LP65 in interphase cells. All mammalian cells tested (including rat liver, rat hepatoma, rat embryo fibroblasts, marsupial kidney PTK2, mouse 3T3 and HeLa cells) showed strong envelope-specific fluorescence. Furtherlocalization of the antigen LP67 in sections of liver and mammary gland tissue and by electron microscopy using the ‘immunoperoxydase’ technique showed a strong concentration of the antigen at the boundary between inner nuclear envelope and peripheral chromatin (Krohne et al. 1978).

Primary chick embryo fibroblasts incubated with each of the sera anti-LP71, anti-LP67 and anti-LP65 at high dilutions could not be decorated. Fig. 1F shows the results obtained for anti-LP67. To visualize any perinuclear fluorescence in these cells undiluted (or a 1 : 5 dilution) antisera had to be used, in which case the fluorescence intensity was comparable to that observed in mouse A9 cells using a 100-fold dilution. However, the very weak perinuclear staining of the chicken cells with antisera and preimmune sera are different; the active sera decorated the area of the cell nucleus more than the preimmune control sera (Fig. 1B and F). The fluorescence obtained with preimmune sera apparently reflects the geometry of the flattened cells, with higher concentrations of cellular material near the nucleus staining non-specifically, whereas the weak fluorescence of nuclei decorated with anti-LPóy may reflect some cross-reaction of the antiserum with the chicken antigen.

Erythrocytes from 14-day-old chicks (Fig. 2A) were fused with mouse A9 cells. When the 3 antisera were reacted with unfused erythrocytes, no specific nuclear envelope-associated fluorescence was observed. Results obtained with serum anti-LP71, anti-LPóy and with a preimmune control serum are presented in Fig. 2B-G. The interior of the erythrocytes was accessible to antisera since they stained with antisera against chromatin (results not shown). Furthermore, nuclei isolated from 14-day-old erythrocytes did not stain with antigen LP67 (Fig. 2H-L).

Heterokaryons formed from erythrocytes and mouse cells were examined 18, 48 and 72 h after fusion using each of the 3 antisera. Some differences become apparent when heterokaryons were incubated with anti-LP71 or the 2 sera anti-LP67 and anti-LP65-Therefore, results obtained with anti-LP71 and anti-LP67 are presented, although the sera cross-react with each other.

During the early period of reactivation (0–18 h after fusion) antigen LP71 was exclusively associated with the mouse nucleus. At 18 h after fusion antigen was also detected in chick nuclei. Controls with preimmune sera revealed that this fluorescence was specific (Fig. 3A–D). The distribution of antigen was diffuse and not clearly associated with the erythrocyte envelope, even when the nuclei had undergone considerable enlargement (Fig. 3E, F). A more pronounced envelope-associated fluorescence as well as some fluorescence of the nuclear interior was observed 48 hours after fusion (Fig. 4A, B). In many cells an unstained halo separated the brightly stained envelope from stained parts of chromatin. Seventy-two hours after fusion, when erythrocytes were fully enlarged and chromatin was decondensed, antigen-LP71 was more clearly associated with the nuclear envelope (Fig. 4c, D). Antigens LP67 as well as LP65 also accumulate progressively in erythrocyte nuclei of heterokaryons. In contrast to LP71, they are detected somewhat later and are more clearly associated with the nuclear envelope, probably because the anti-LP67 serum is more specific (Fig. 5A–F). The appearance of antigen LP67 in erythrocytes reflects the degree of reactivation of erythrocyte nuclei and depends upon the number of erythrocyte nuclei in the heterokaryons. Only 10% of the total erythrocyte population investigated had resumed the incorporation of mouse antigen 18 h after fusion. Fluorescence was detected in 82% of erythrocyte nuclei 48 h after fusion, 76% of the cells were heavily decorated 3 days after fusion. Less antigen is generally present in heterokaryons with more than one erythrocyte nucleus than in those with only one. On the other hand, strong erythrocyte fluorescence was observed earlier in heterokaryons formed from erythrocytes and ‘giant’ A9 cells.

In some experiments protein synthesis was inhibited by adding cycloheximide to heterokaryons formed from irradiated A9 cells and chick erythrocytes 18 h after fusion. The influx of antigen was monitored over a period of 5 days and compared with the influx into untreated controls. Antigen LP67 was readily detected in enlarged erythrocytes of the untreated control 48 h after fusion (Fig. 5 A, B). In heterokaryons treated with cycloheximide from 18 to 48 h after fusion almost no antigen was detected even though the erythrocyte nuclei had undergone some enlargement (Fig. 5 c, D). Even 5 days after fusion little extra fluorescence could be detected in the erythrocyte nuclei, whereas in untreated heterokaryons, the fluorescence intensity of reactivated erythrocyte and irradiated A9 nucleus was identical (Fig. 5E-H; Fig. 6). This suggests that the synthesis of the new antigen rather than a redistribution determines the rate of antigen influx.

The precise nature of the antigens used in this study is not entirely clear. They are difficult to solubilize in high salt buffers and non-ionic detergents and remain associated with the nuclear periphery under conditions, which remove most of the proteins and the lipids of the envelope (Aaronson & Blobel, 1975; Cook, Brazell & Jost, 1976; Ely et al. 1978; Levin, Jost & Cook, 1978; Krohne, Franke & Scheer, 1978). The structural elements of the pore complex or the interporous connecting material resist extraction with high-salt buffers and non-ionic detergents and so are likely to contain the 3 antigens. The weak cross-reaction of anti-LPóy with chick fibroblasts (Fig. 1F) but not with chick erythrocytes (Fig. 2) favours the hypothesis that the antigens may be an element of the pore complex as the pore density varies considerably between these 2 cell types. However, electron-microscopic observations using antigen LP67 do not indicate an exclusive localization in pore complexes or interporous connecting fibres. Rather the antigen is associated with peripheral layer of chromatin (Krohne et al. 1978), as well as on the well defined nuclear rim (Fig. 4B). This suggests that the antigens are initially bound at 2 topologically different sites which form the same structure when chromatin is finally decondensed.

Despite these uncertainties regarding the location of the antigens, it is nevertheless evident that mouse macromolecules specific for the nuclear envelope are found in the erythrocyte envelope during reactivation in heterokaryons. After a lag of about 18 h the antigens accumulate rapidly as the nucleus grows in size suggesting that envelope enlargement during the early reactivation is a passive swelling rather than an active growth. In human-chick heterokaryons a similar lag precedes the accumulation of human nucleospecific antigens in the erythrocyte nucleus (Ege et al. 1971; Ringertz et al. 1971). On the other hand, erythrocytes can accumulate nucleospecific proteins immediately after heterokaryons are formed (Appels et al. 1974 a, 1975) before the envelope of the inert erythrocyte can be altered to any great extent, for example by the formation of new nuclear pores.

We wish to thank Claus Christensen for help with the photographs.